The Routledge Handbook of Liberal Naturalism - Caro M., Macarthur D.

1. Aristotle on (second) nature, habit and character
1. Introduction
2. Aristotle's ethical "naturalism"

3. Hume and liberal naturalism
1. Introduction

4. Philosophy and Hume's therapeutic liberal naturalism

4. Kant on nature and humanity
1. Introduction

2. Theoretical knowledge of nature
Sensibility and understanding
How nature is possible
Theoretical reason

Cosmos, mind, body, freedom
3. Practical reason
Freedom of the will

God and immortality as postulates of practical reason
Symbolism, analogy and religion

6. Husserlian phenomenology and liberal naturalism
1. Introduction

7. Merleau-Ponty and liberal naturalism
1. Introduction
2. Liberal naturalism

8. Classical pragmatism and liberal naturalism
1. Introduction
2. Theoretical consequences of evolutionary theory

3. Practical consequences of evolutionary theory

2. Logic
3. Regimentation and explication

10. Wilfrid sellars and liberal naturalism
1. Introduction
2. Sellarsian naturalism, round one

3. Naturalism and the two images of humanity-in-the-world

6. Putting humanity into the scientific image

11. Philippa Foot's liberal naturalism in ethics
1. Philippa Foot's position in ethical naturalism

4. Foot's liberal naturalism and the explanation of normativity

12. Bernard Williams's liberal naturalism
1. Part I

13. Price's subject naturalism and liberal naturalism
1. Introduction

3. The subject-naturalist rejection of object naturalism

3.1. Blocking a route to object naturalism

3.2. Undermining a presupposition of object naturalism
3.3. Vindicating the correctness of assertions?

4. Does Price's rejection of object naturalism depend on subject naturalism?

5. Subject naturalism and liberal naturalism

6. Subject naturalism and McDowell's naturalism of second nature

14. Relaxed naturalism
1. A relaxed philosophy of nature

15. Liberal or radical naturalism?
1. Introduction

4. Two-dimensional biological normativity

5. Liberal and radical conceptions of normativity as "natural"

16. Naturalism as a stance
1. Scientific naturalism

3.2. Simplicity in metaphysical reasoning

4. Liberal naturalism and the naturalistic stance

17. Liberal naturalism
1. The placement problem

2. When science divorced from the ordinary view of the world

$3. Kant and the fracture between the subjective and the objective views$ 4. Scientific naturalism and the placement problem

5. Liberal naturalism and the reconciliation problem

18. Liberal naturalism and God
1. Introduction

19. Taylor and liberal naturalism
1. Introduction

20. Can selves be naturalised? The problem of temporal perspective
1. Introduction

3. Temporal perspective and reidentifiability

21. Liberal naturalism, ontological commitment and explanation
1. Introduction

3. Liberal naturalism and moderate naturalism about metaphysics

22. Naturalism with Chinese characteristics
1. Introduction
2. Beginnings

7. The comparison with liberal naturalism

23. Liberal naturalism and aesthetics
1. Introduction

2. Can scientific naturalism be nonreductive?

3. Triangulation and "thick" explanations

5. Must we go subpersonal in aesthetics?

8. Cavell's problem of the other and film spectatorship
Notes

24. Liberal Naturalism, Aesthetic Reflection, and the Sublime
1. Introduction

4. Non-perceptually represented perceptual properties

25. Philosophy of perception and liberal naturalism
1. Introduction

2. The traditional problems of perception

3. How does philosophy of perception interact with questions about naturalism?

3.1. Naïve realism and liberal naturalism

3.2. "Indirect" views of perception and liberal naturalism

3.3. Representational views and liberal naturalism

26. Ethics and liberal naturalism
1. Introduction

3. A naturalistic understanding of nature

4. A naturalistic understanding of value in general

5. A naturalistic understanding of ethical normativity

27. Kantian constitutivism and the naturalistic challenge
1. Introduction

3. Ethical naturalism and the paradox of punishment

5. The constitutivist claim
6. Conclusion

28. The rational wolf. Moral philosophy as key to John McDowell's liberal naturalism
1. Our conceptual capacities and their place in nature
2. Animality and rationality

3. McDowell's case for liberal naturalism (I)

4. McDowell's case for liberal naturalism (II)

5. Some conclusions regarding liberal naturalism

1. Realism, constructivism and the method of avoidance

2. Naturalism, the original position and the evidential basis of the theory of justice

30. Scientific naturalism and normative explanation
1. The historical and philosophical context

2. Scientific naturalism
2.1. Naturalism and the drive for intellectual economy

2.2. Two conceptions of the Razor as requiring preference for the simpler
3. Ontological pluralism
3.1. Abstract entities

3.2. The normative domain
3.3. Property identity as an element in ontological reduction

5. Moral realism in the intuitionist tradition

6. Normative explanation
6.1. Normative properties as consequential on natural ones

6.2. The causal element in moral perception

7. Normativity and the scientific habit of mind
7.1. Objectivity

7.2. Religion, politics, and scientific objectivity

31. Scientism and liberal naturalism
1. Introduction
2. What is this scientism thing, anyway?

3.1. Scientism as an unwarranted trump card against embarrassing questions

32. The foundations of psychoanalysis and liberal naturalism
1. Introduction

3. Naturalism and the foundations of psychoanalysis

4. A liberal naturalist notion of the Freudian unconscious and the method to its discovery

5. The unconscious is manifest in free association

33. Actualism as a form of liberal naturalism
1. What's in a name?

3. The indexicality of the actual world and locating minds from within it

4. Redefining metaphysics with an actualist's account of modality
Notes

34. Critical naturalism for the human sciences
1. Naturalism descends on the human sciences

3. Can we make liberal naturalism more productive?

4. Why the human sciences cannot be fully naturalised

35. Habermas and liberal naturalism
1. Introduction
2. Vivat imperator scientia naturalis

3. Habermas's transcendental pragmatism in "knowledge and human interests"

36. Strawson and non-revisionary naturalism
1. Introduction

Metaphilosophical Naturalism
Epistemological naturalism
Ontological naturalism

3. "Reductive" vs. "non-reductive" naturalism

6. Relaxed realism and liberal naturalism

37. Putnam and liberal naturalism
1. Introduction

Author: Caro M. Macarthur D.

Tags: philosophy ethics naturalism

ISBN: 978-0-815-38182-2

Year: 2022

Similar

Minds and Persons

Handbook of Political Theory

Orientalism Western Conceptions of the Orient

Objectivity, Relativism, and Truth

Text

THE ROUTLEDGE HANDBOOK OF
LIBERAL NATURALISM

The central question of naturalism – the relation of philosophy to science – was one of the defining
strands of twentieth-century thought and remains a major source of debate and controversy. Today
many argue that philosophy should fold itself into the sciences, especially the natural sciences. Liberal
naturalists argue that such scientific naturalism demands reductive and Procrustean conceptions of
knowledge and reality. Moreover, many philosophical problems are beyond the scope of the sciences,
such as the nature of persons, the normativity of the space of reasons, and how best to understand the
peculiar mix of objectivity and subjectivity of ethics and art.
The Routledge Handbook of Liberal Naturalism is the first collection to present a comprehensive
overview of liberal naturalism, a philosophical outlook that lies between scientific naturalism and
supernaturalism. Comprising 37 chapters by an international team of contributors, it examines
important cutting-edge topics including:

•
•
•
•
•
•

what is liberal naturalism?
is metaphysics a viable project?
naturalism in the history of philosophy, including Hume, Dewey, and Quine
contemporary liberal naturalists such as P.F. Strawson, John McDowell, Hilary Putnam, and John
Rawls
related kinds of naturalism, including subject naturalism, common-sense naturalism and biological naturalism
the bearing of liberal naturalism on contemporary debates in epistemology, philosophy of mind,
ethics and aesthetics.

Essential reading for students and researchers in all areas of philosophy, this volume will be of particular
interest for those studying philosophical naturalism, philosophy of science, metaphysics, epistemology,
philosophy of mind, ethics and aesthetics.
Mario De Caro is Professor of Moral Philosophy at Roma Tre University, Italy, and regularly Visiting
Professor at Tufts University, USA. He has published five volumes in Italian and edited Interpretations
and Causes. New Perspectives on Donald Davidson’s Philosophy (1999), Hilary Putnam’s volume Naturalism,
Realism, and Normativity (2016) and, with Maria Silvia Vaccarezza, Practical Wisdom: Philosophical and
Psychological Perspectives (Routledge, 2021).
David Macarthur is an Associate Professor in the Philosophy Department at the University of Sydney,
Australia. He has co-edited three collections with Mario De Caro: Naturalism in Question (2004), Naturalism
and Normativity (2010), and Philosophy in an Age of Science: Physics, Mathematics, and Skepticism (2012). He is
editor of Pragmatism as a Way of Life: The Lasting Legacy of William James and John Dewey (2017).

ROUTLEDGE HANDBOOKS IN PHILOSOPHY

Routledge Handbooks in Philosophy are state-of-the-art surveys of emerging, newly refreshed and
important fields in philosophy, providing accessible yet thorough assessments of key problems,
themes, thinkers and recent developments in research.
All chapters for each volume are specially commissioned, and written by leading scholars in the
field. Carefully edited and organised, Routledge Handbooks in Philosophy provide indispensable
reference tools for students and researchers seeking a comprehensive overview of new and exciting topics in philosophy. They are also valuable teaching resources as accompaniments to
textbooks, anthologies and research-orientated publications.
The Routledge Handbook of Philosophy and Improvisation in the Arts
Edited by Alessandro Bertinetto and Marcello Ruta
The Routledge Handbook of Idealism and Immaterialism
Edited by Joshua Farris and Benedikt Paul Göcke
The Routledge Handbook of Philosophy of Economics
Edited by Conrad Heilmann and Julian Reiss
The Routledge Handbook of Logical Empiricism
Edited by Thomas Uebel and Christoph Limbeck-Lilienau
The Routledge Handbook of Philosophy of Agency
Edited by Luca Ferrero
The Routledge Handbook of Propositions
Edited by Adam Russell Murray and Chris Tillman
The Routledge Handbook of Liberal Naturalism
Edited by Mario De Caro and David Macarthur
For more information about this series, please visit: https://www.routledge.com/RoutledgeHandbooks-in-Philosophy/book-series/RHP

THE ROUTLEDGE
HANDBOOK OF LIBERAL
NATURALISM

Edited by Mario De Caro and David Macarthur

Cover image: © Getty Images
First published 2022
by Routledge
2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN
and by Routledge
605 Third Avenue, New York, NY 10158
Routledge is an imprint of the Taylor & Francis Group, an informa business
© 2022 selection and editorial matter Mario De Caro and David Macarthur; individual
chapters, the contributors
The right of Mario De Caro and David Macarthur to be identified as the authors of the
editorial material, and of the authors for their individual chapters, has been asserted in
accordance with sections 77 and 78 of the Copyright, Designs and Patents Act 1988.
All rights reserved. No part of this book may be reprinted or reproduced or utilized in any
form or by any electronic, mechanical, or other means, now known or hereafter
invented, including photocopying and recording, or in any information storage or
retrieval system, without permission in writing from the publishers.
Trademark notice: Product or corporate names may be trademarks or registered
trademarks, and are used only for identification and explanation without intent to
infringe.
British Library Cataloguing-in-Publication Data
A catalogue record for this book is available from the British Library
Library of Congress Cataloguing-in-Publication Data
Names: De Caro, Mario, editor. | Macarthur, David, editor.
Title: The Routledge handbook of liberal naturalism / edited by Mario De Caro and
David Macarthur.
Description: Abingdon, Oxon ; New York, NY : Routledge, 2022. | Series: Routledge
handbooks in philosophy | Includes bibliographical references and index.
Identifiers: LCCN 2021042322 (print) | LCCN 2021042323 (ebook) | ISBN
9780815381822 (hbk) | ISBN 9781032205106 (pbk) | ISBN 9781351209472 (ebk)
Subjects: LCSH: Naturalism. | Knowledge, Theory of. Classification: LCC B828.2 .R67
2022 (print) | LCC B828.2 (ebook) | DDC 146--dc23
LC record available at https://lccn.loc.gov/2021042322
LC ebook record available at https://lccn.loc.gov/2021042323
ISBN: 978-0-815-38182-2 (hbk)
ISBN: 978-1-032-20510-6 (pbk)
ISBN: 978-1-351-20947-2 (ebk)
DOI: 10.4324/9781351209472
Typeset in Bembo
by MPS Limited, Dehradun

CONTENTS

Notes on Contributors

Introduction
Mario De Caro and David Macarthur

PART I

Historical naturalisms and their relation to liberal naturalism
1 Aristotle on (second) nature, habit and character
Riccardo Chiaradonna and Flavia Farina

5
7

2 Spinoza and liberal naturalism
Alex Douglas

3 Hume and liberal naturalism
Benedict Smith

4 Kant on nature and humanity
Allen Wood

5 Nietzsche’s naturalism: neither liberal nor illiberal
Brian Leiter

6 Husserlian phenomenology and liberal naturalism
Andrea Staiti

Contents

7 Merleau-Ponty and liberal naturalism
Jack Reynolds

8 Classical pragmatism and liberal naturalism
Steven Levine

PART II

Theoretical cousins of liberal naturalism
9 Quine’s naturalism: neither “reductive” nor “liberal”
Garry Ebbs

95
97

10 Wilfrid Sellars and liberal naturalism
Willem A. deVries

110

11 Philippa Foot’s liberal naturalism in ethics
Gabriele De Anna

122

12 Bernard Williams’s liberal naturalism
Sophie-Grace Chappell

134

13 Price’s subject naturalism and liberal naturalism
Lionel Shapiro

152

14 Relaxed naturalism: a liberating philosophy of nature
Daniel Hutto

165

15 Liberal or radical naturalism
Joseph Rouse

177

16 Naturalism as a stance
Jack Ritchie

190

PART III

Challenges for liberal naturalism

203

17 Liberal naturalism: origins and prospects
Mario De Caro

205

18 Liberal naturalism and God
Fiona Ellis

218

Contents

19 Taylor and liberal naturalism
Nicholas Smith

227

20 Can selves be naturalised? The problem of temporal perspective
Patrick Stokes

235

21 Liberal naturalism, ontological commitment and explanation
Matteo Morganti

245

22 Naturalism with Chinese characteristics
Barry Allen

255

PART IV

Applications of Liberal Naturalism

265

23 Liberal naturalism and aesthetics: art up close and personal
David Macarthur

267

24 Liberal naturalism, aesthetic reflection, and the sublime
Jennifer McMahon

281

25 Philosophy of perception and liberal naturalism
Thomas Raleigh

299

26 Ethics and liberal naturalism
Hans Fink

320

27 Kantian constitutivism and the naturalistic challenge
Carla Bagnoli

329

28 The rational wolf. Moral philosophy as key to John McDowell liberal
naturalism
Sofia Miguens

339

29 Rawls and liberal naturalism
Paul Patton

349

30 Scientific naturalism and normative explanation
Robert Audi

358

31 Scientism and liberal naturalism
Massimo Pigliucci

371

vii

Contents

32 The foundations of psychoanalysis and liberal naturalism: the Freudian
unconscious and the manifest image
Talia Morag

383

33 Actualism as a form of liberal naturalism
Paul Redding

411

34 Critical naturalism for the human sciences
Daniel Andler

419

35 Habermas and liberal naturalism
Paul Giladi

429

36 Strawson and non-revisionary naturalism
Hans-Johan Glock

441

37 Putnam and liberal naturalism
Massimo Dell’Utri

455

Index

464

viii

CONTRIBUTORS

Barry Allen – Distinguished University Professor of Philosophy, McMaster University.
Daniel Andler – Emeritus Professor of Philosophy, Université Paris IV, Sorbonne.
Robert Audi – John O’Brien Professor of Philosophy, University of Notre Dame.
Carla Bagnoli – Professor of Theoretical Philosophy, University of Modena and Reggio
Emilia.
Sophie-Grace Chappell – Professor of Philosophy, The Open University.
Riccardo Chiaradonna – Professor of Ancient Philosophy, Roma Tre University.
Gabriele De Anna – Professor of Philosophy, University of Bamberg & Associate Professor of
Political Philosophy, University of Udine.
Mario De Caro – Professor of Moral Philosophy, Roma Tre University and Visiting
Professor, Tufts University.
Massimo Dell’Utri – Professor of Theoretical Philosophy and Philosophy of Language,
University of Sassari.
Willem De Vries – Professor of Philosophy, University of New Hampshire.
Alex Douglas – Senior Lecturer in Philosophy, University of St Andrews.
Gary Ebbs – Professor of Philosophy, University of Indiana, Bloomington.

Contributors

Fiona Ellis – Professor of Philosophy and Director of the Centre for Philosophy of Religion,
University of Roehampton.
Flavia Farina – Post-doctoral Fellow, Roma Tre University.
Hans Fink – Emeritus Professor of Philosophy, Aarhus University.
Paul Giladi – Senior Lecturer in Philosophy, Manchester Metropolitan University.
Hans-Johan Glock – Professor of Philosophy, University of Zürich.
Daniel Hutto – Senior Professor of Philosophical Psychology, University of Wollongong.
Brian Leiter – Karl N. Llewellyn Professor of Jurisprudence and Director of the Centre for
Law, Philosophy and Human Values, University of Chicago.
Steven Levine – Professor of Philosophy, University of Massachusetts, Boston.
David Macarthur – Associate Professor of Philosophy, University of Sydney.
Jennifer McMahon – Professor of Philosophy, University of Adelaide.
Sofia Miguens – Professor of Philosophy, University of Porto.
Talia Morag – Lecturer in Philosophy, University of Wollongong.
Matteo Morganti – Associate Professor of Philosophy, Roma Tre University.
Paul Patton – Hongyi Chair Professor of Philosophy, Wuhan University & Professor of
Philosophy, Flinders University.
Massimo Pigliucci – Professor of Philosophy, City University, New York.
Thomas Raleigh – Associate Professor of Philosophy of Mind and Artificial Intelligence,
University of Luxembourg.
Paul Redding – Emeritus Professor of Philosophy, University of Sydney.
Jack Reynolds – Professor of Philosophy, Deakin University.
Jack Ritchie – Senior Lecturer, University of Cape Town.
Joseph Rouse – Hedding Professor of Moral Science, Wesleyan University.
Lionel Shapiro – Professor of Philosophy, University of Connecticut.
Nicholas Smith – Professor of Philosophy, Macquarie University.

Contributors

Benedict Smith – Assistant Professor of Philosophy, Durham University.
Andrea Staiti – Associate Professor of Moral Philosophy, University of Parma and Boston
College.
Patrick Stokes – Associate Professor of Philosophy, Deakin University.
Allen Wood – Ruth Norman Halls Professor of Philosophy, Indiana University,
Bloomington.

INTRODUCTION
Mario De Caro and David Macarthur

Liberal Naturalism is a recent arrival in the history of philosophical naturalisms, a tradition which
extends at least as far back as the Ionian School (sixth century BC), and which includes the
naturalisms of Aristotle (fourth century BC), Spinoza (seventeenth century), Hume (eighteenth
century), Dewey (early twentieth century) and Quine (mid twentieth century). Yet it has some
claim to being an important game-changer in debates over naturalism, one that holds the
potential to radically change our way of seeing both the tradition and contemporary issues. The
key issue concerns the relation between science and philosophy and, in particular, the importance of providing a philosophical platform from which to combat the rise of scientism in
the late twentieth and early twenty-first centuries – where “scientism” is, let us say, the idea that
appropriate respect for the successful sciences dictates that one accepts a scientific metaphysics
(i.e. the world is nothing over and above the scientific image of the world) and scientific
epistemology (i.e. scientific knowledge is the only genuine knowledge that there is). Of course,
if a scientific naturalist regards the successful sciences as the natural sciences then she needs to
make appropriate adjustments: the world is nothing other than the natural scientific image of the
world; knowledge is nothing other than natural scientific knowledge.
The philosophical outlook of liberal naturalism is at once nonscientistic and antisupernatural. Its image of the world is neither the minimal composite scientific image of the
successful sciences – natural (and social?) –, nor is it the maximal supernatural image of superstitious religion (which is not to say all religion takes this metaphysically inflationary form).
What liberal naturalism brings into view is the often forgotten realm of the manifest image or
what some philosophers conceive of as the “common sense” world (Moore) or the “life-world”
[Lebenswelt] (Husserl).1 Just how the manifest image should be conceived is one of the significant open questions we aim to address in this volume.
The manifest image is also characterisable as the nonsupernatural nonscientific realm – in the
sense that the objects of the manifest image are encounterable nonsupernatural things and, at the
same time, not mere posits of scientific theory; nor exhaustively explicable in scientific terms.
Persons and artworks are paradigm nonscientific nonsupernatural entities. Liberal naturalism is a
philosophy that takes the manifest image that includes persons and artworks seriously; in particular, by denying the traditional large-scale clash that Sellars and many others after him have
imagined to exist between the manifest and scientific images of the world – to the detriment of
the manifest image.2
DOI: 10.4324/9781351209472-1

Mario De Caro and David Macarthur

Liberal naturalism makes room in its vision of the world for nonscientific realities (that are
not posits of successful scientific explanations) and nonscientific knowledge or understanding.
But for all that it is still a form of naturalism on the ground that it refuses to admit the supernatural. From this perspective, “natural” is not identified with the scientific image of the
world but is, rather, equivalent to the anti-supernatural. By focusing attention on the category
of the supernatural, liberal naturalism endorses a long-standing theme of traditional naturalisms:
namely, the mission to combat a sense that some philosophical or other position has lost touch
with reality. But what do we mean by “losing touch with reality”? Or, in other words, what do
we mean by supernatural? This is not easy to explain since it is widely admitted that science does
not explain everything and some new previously unexplained phenomenon rather than automatically being banished as supernatural might rather show gaps or shortcomings in our
current scientific understanding of the world that needs correction. What is “supernatural”
today (in the sense of what cannot be explained by current science) may be part of tomorrow’s
new science. Think of the discovery of the phenomenon like quantum entanglement, which
from the older point of view of classical physics appears supernatural. It is a virtue of liberal
naturalism to clarify the question of the supernatural as a vital topic for further research.
Liberal naturalism also calls attention to a second way in which a philosophy can lose touch
with aspects of reality or nature, namely, by way of an allegiance to a misguided scientistic
naturalism, e.g. strict physicalism, the view that the only entities and events that there are, are
those posited by the successful explanations of physics.3 A scientistic naturalism is a reductive or
eliminative scientific naturalism, which treats the full extent of nature as exhausted by, say, the
scientific image of the world; at its most extreme, the image of the world of physics alone.4
From a scientistic point of view what is putatively a part of nature can be threatened from two
directions: (1) it can be denied its natural status by being considered supernatural; or (2) it can
be denied autonomous status as something legitimate to admit into our worldview in its own
right by being eliminated as a “false imaginary glare” (to borrow a phrase from the supernaturalist Berkeley)5 or by being reduced to something that passes scientific muster, say, the
posits of the successful natural sciences. In this way the manifest image is threatened with
unreality or illegitimacy.
What is “liberal” in liberal naturalism is a more inclusive view of nature or reality in
comparison to the scientific image – which is best thought of as the composite image of all the
successful sciences. The “naturalism” in liberal naturalism is, as we have seen, its commitment
to anti-supernaturalism. This need not be seen as a dogmatic or question-begging commitment
but a reflection of a sense that our worldview, including our knowledge of the world, must be
suitably tethered to the empirical realm. It is based on the plausible idea that the meanings of
our terms and the scope of our inquiries are in various ways based on, or conditioned by,
human experience. What it is to be suitably based on or conditioned by experience is another
open question for further research but that there is some kind of empirical tethering at stake is a
key theme of both of the leading inspirations for the liberal naturalist outlook, namely, Hilary
Putnam and John McDowell.6
Philosophers in the scientific age we are currently living in have been so focused on the
sciences, particularly the natural sciences, and how philosophy relates to them, that they have
overlooked the world under their own noses. One of the prime virtues of liberal naturalism is to
inaugurate a philosophy of the manifest image, taking the everyday world of human experience
– its practices and their conditions and presuppositions – as a new and fitting topic for philosophical reflection. To date philosophy has an impoverished conception of the manifest image
which has often been discussed in terms of the dubious notion of “common sense” beliefs or
knowledge; as well as the highly prejudicial notion that our cognitive grasp of the manifest
2

Introduction

image – including our relations with each other – is simply proto-science (e.g. “folk psychology”, “folk physics”, etc.) to be replaced by genuine science.7 Given the controversial
assumptions of the common sense tradition of philosophy (e.g. Moore’s view that we all know
common sense propositions with certainty) as well as those of the phenomenological tradition
(e.g. Husserl’s essentialist views about mental states), just how to fruitfully approach the manifest
image philosophically is a contentious and open question.8
The aim of this volume is to address liberal naturalism from various points of view to give a
broad survey of its historical situatedness, its game-changing power and promise in application
to contemporary issues, as well as exploring particular objections or problems arising within
these points of view in some depth. The handbook will be divided into the following four
sections.
1.

History: The relation between liberal naturalism and various historically important forms of
naturalism such as those of Aristotle, Spinoza, Hume, Nietzsche and the classical pragmatists. We are particularly interested in asking whether specific forms of naturalism in the
history of philosophy might be profitably looked at as crucial first steps on the way to
liberal naturalism; or if not, why not?
Related Naturalisms: Consideration of some close theoretical cousins of liberal naturalism
and what distinguishes them from it. Here are included: (1) broad scientific naturalisms that
do not limit their conception of nature to the image of the natural sciences but happily
include the image of the social sciences too (e.g. Hutto’s relaxed naturalism); (2) allegedly
non-supernaturalist forms of theism (e.g. Ellis’s expansive naturalism) and (3) forms of
naturalism based on adopting a biological or anthropological stance towards language or
reason including Price’s subject naturalism and Rouse’s radical naturalism.
Challenges: Theoretical challenges for liberal naturalism such as how a pluralistic ontology
does justice to its dependency on the physical world, the issue of God and selves, the scope
of natural nonscientific items and the need to acknowledge the importance of the natural
processes of becoming at the center of Chinese philosophy.
Applications: Fruitful applications of liberal naturalism to contemporary debates across a
wide range of topics that demonstrate its power and promise. Of particular importance is
the idea that liberal naturalism, rather than focusing on the philosophical implications of
the sciences, becomes a philosophy of the manifest image, capable of doing justice to
rational normativity in various areas of our thought and talk, as well as our experience of
the values of things. This section also includes discussion of philosophers such as Habermas
and Rawls who can be counted as a liberal naturalists even if they do not identify
themselves in these terms.9

Liberal naturalism allows one freedom from metaphysical projects of naturalisation as well as
from myths of science and scientific metaphysics such as ontological monism, causal fundamentalism and the nomological model of science – but without endorsing any form of supernaturalism.10 In calling renewed philosophical attention to our everyday lives without
reductive or eliminative tendencies, liberal naturalism puts various nonscientific concepts at the
center of philosophical attention, for example: persons and their responsibilities and commitments; artifacts including the built environment; the space of first- and second-person relations
including rational relations; moral and political values; and (our experience of) artworks and
their significance.

Mario De Caro and David Macarthur

Notes
1 G.E. Moore “A Defence of Common Sense”, in J.H. Muirhead (ed.) Contemporary British Philosophy
(London: Allen & Unwin, 1925), 192–233; E. Husserl, The Crisis of European Science and Transcendental
Phenomenology. Trans. D. Carr (Evanston, IL: Northwestern University Press, 1970), 33–34.
2 Wilfrid Sellars, “Philosophy and the Scientific Image of Man”, in K. Sharp & R. Brandom (eds.) In the
Space of Reasons: Selected Essays of Wilfrid Sellars (Cambridge, MA: Harvard University Press, 2007),
369–410.
3 For example, see David Armstrong, “Naturalism, Materialism, and First Philosophy”, The Nature of
Mind (St. Lucia, Queensland: University of Queensland Press, 1980), 149–165.
4 For the (mis)identification of naturalism with strict physicalism see David Papineau, Philosophical
Naturalism (Oxford: Blackwell, 1993).
5 G. Berkeley, Philosophical Writings, ed. D.M. Clarke (Cambridge University Press, 2008), 193.
6 Hilary Putnam, “The Content and Appeal of Naturalism”, in M. De Caro and D. Macarthur (eds.)
Naturalism in Question (Cambridge, MA: Harvard University Press, 2004), 59–70; John McDowell,
“Two Sort of Naturalism”, in Mind, Value and Reality (Cambridge, MA: Harvard University Press,
1998), 167–197. It is important to distinguish inspiration from doctrinal fidelity. It should not be
supposed that either Putnam’s or McDowell’s versions of naturalism are ideal or without difficulties.
7 P. Pettit & F. Jackson, “In Defence of Folk Psychology”, Philosophical Studies 59/1 (1990): 31–54.
8 G.E. Moore, “A Defence of Common Sense”, in T. Baldwin (ed.) G.E. Moore: Selected Writings
(London: Routledge, 1993), 106–133; E. Husserl, Ideas Pertaining to a Pure Phenomenology and to a
Phenomenological Philosophy – First Book: General Introduction to a Pure Phenomenology [1913], trans. F.
Kersten (The Hague: Nijhoff, 1982).
9 Another example of a philosopher who is plausibly interpreted as a liberal naturalist, even if he does not
work under this label, is Tim Scanlo, What We Owe to Each Other (Cambridge, MA: Harvard
University Press, 1998), 11f.
10 See D. Macarthur, “Taking the Human Sciences Seriously”, in, M. De Caro, and D. Macarthur (eds.)
“Realism, Common Sense, and Science”. The Monist, 28, (2015)197–214.

PART I

Historical naturalisms and
their relation to liberal naturalism

1
ARISTOTLE ON (SECOND)
NATURE, HABIT AND
CHARACTER
Riccardo Chiaradonna and Flavia Farina

1 Introduction
This paper focuses on Aristotle’s so-called “naturalist” approach to agency (Section 1) and,
more precisely, on the similarities and differences subsisting between nature and character.
Sections 2 and 3 are devoted to the notions of nature (physis) and habit (hexis). Two senses of
nature can be distinguished in Aristotle, i.e. a technical and a common or broad sense. The
notion of hexis has a multi-faceted background too and can be traced back to Aristotle’s
Categories as well as to his physics and ontology. Section 4 investigates how the possession of
habits determines the general quality of moral actions but, at the same time, does not completely undermine the bilateral and symmetrical conception of that which depends on us.
According to the present interpretation, regularity does not directly affect the outcome, the
particular course of actions, but the moral quality of the action. Section 5 investigates differences and similarities between character and nature, in order to clarify whether we can ascribe a
notion of second nature to Aristotle’s ethical theory.

2 Aristotle’s ethical “naturalism”
In Nicomachean Ethics (NE) 1152a29–33, Aristotle states that an agent’s character is hard to
change, since it becomes like nature.
For it is easier to change one’s habit than one’s nature. Even habit is hard to change
just because it is like nature, as Evenus says: I say that habit’s but long practice, friend,
And this becomes men’s nature in the end. (NE 1152a29–33, trans. Barnes and Kenny
in Aristotle, 2014)
This statement raises questions as to whether character can be said to be a second nature and as
to the possibility of ascribing a naturalist ethical perspective to Aristotle.
Some contemporary naturalist philosophers, such as G.E.M. Anscombe (1958), Philippa
(2001) and Hursthouse (1999) argued that their theories are grounded in (or inspired by)
Aristotle’s views on nature. Along these lines, in his essay “Two Sort of Naturalism”,

DOI: 10.4324/9781351209472-3

Riccardo Chiaradonna and Flavia Farina

McDowell explores the concept of second nature, thus seeking to frame Aristotle’s ethical
outlook within an obviously naturalistic perspective, while at the same time rejecting a certain
kind of naturalism which aims to validate the necessity of virtues by appealing to the necessity of
natural, extra-ethical facts.2 McDowell argues that nature can be distinguished into a mere
nature and into a nature whose realisation directly implies transcending nature itself. This is the
concept of second nature McDowell endorses: second nature, while springing from our natural
capacities (such as reason, the imagination and so on), makes human beings capable of distancing themselves from their natural (in the sense of “first nature”) motivational impulses, thus
enabling them to even turn against first nature itself. Virtue and practical reason don’t merely
offer reasons to act in certain ways (as first nature does): they entail reflective questioning and
the agent’s distancing from her first-natural motivational impulses. However, McDowell says,
first nature still matters as it provides human beings with the necessary natural endowment that
enables them to shape their second nature. Also, first nature matters because “first natural facts
can be part of what reflection takes into account” (McDowell 1998, 185). McDowell describes
his conception as a naturalist one insofar as “a formed state of practical reason is one’s second
nature, not something that dictates to one’s nature from outside”.3
Against McDowell, David Forman argues that this account follows Aristotle’s account of
second nature as something that is achieved in the course of normal human upbringing, but
overlooks the fact that according to Aristotle’s account of second nature our practical responses
arise as natural responses.4 While McDowell’s account focuses on how something natural comes
to transcend nature, Forman argues, Aristotle’s account aims to show how something becomes
similar to nature in certain relevant respects.
Here we shall focus on those features which make nature comparable to moral character. We
will also focus on habit in order to establish why character is said to come close to nature.
Through a comparative analysis of the notions of nature and habit, we will try to clarify
whether an agent’s character could be considered her “second nature” and, if so, how this
second nature should be understood.

3 Nature (physis)
Aristotle uses the word “nature” – and all the adjectives, adverbs and locutions related to it – in
different argumentative contexts: in Aristotle’s own terms, nature can be said in many ways.5
Two general senses or meanings can be identified: (1) a technical and (2) a broad sense.
According to this technical sense, nature is the internal starting point of movement and rest (see
Physics [Phys.] 2.1 and Metaphysics [Metaph.] 5.4). The broad sense of nature corresponds instead
to what is “common” (koinon) as opposed to “private” (idion): what is natural corresponds to
what belongs to everybody, i.e. to what is common to everybody. Indeed, certain emotions are
said to be natural insofar as they occur in every human being, without implying a strict reference to the technical sense of nature. The distinction between the technical and the broad
sense notwithstanding, we can identify some common features belonging to nature in general
(i.e. in both its technical and broad sense). Starting from these common features, some characteristics will be highlighted in order to explain the relationship between nature and character.
In Phys. 2.1, Aristotle defines nature, stating that everything which has a nature has an internal
starting point of movement and rest (Phys. 193a30–31), unlike the products of arts and crafts,
which have an external moving principle, namely the artist or craftsman. Nature is also form, thus
constituting the essential element of a given substance. Aristotle refers the expressions “by nature”
or “according to nature” to what belongs to something in virtue of what it is: for instance, fire’s
property of going upwards belongs to fire by nature. In Eudemian Ethics [EE] 1094b4–8, Aristotle
8

Nature, habit and character

investigates what occurs “by nature” by distinguishing two cases: on the one hand, everything
which occurs in a given substance from its generation (e.g. being biped in the case of man) is by
nature; on the other, what is also by nature is that which develops if its generation is not impeded
(e.g. reason in human begins, which exists potentially in children). Aristotle hence provides both
an essentialist definition of nature and a characterisation of it in terms of dynamic processes.6
According to its broader sense, nature corresponds to what is “common” as opposed to
“private”. The word physis is used with this meaning in EE 1228b24–26, where Aristotle distinguishes between a relative and an absolute sense according to which things can be said to be
frightening. In a relative sense, some things do not appear frightening to the brave agent while, at
the same time, they can be very frightening to a cowardly agent: in this case, different perceptions
are explained through the difference in habits between the two agents. Some things, however, are
absolutely frightening, because they are frightening to the majority of human beings, both cowardly and brave. They are, so to speak, frightening to human nature itself. Here, the word nature
is clearly not used according to its technical meaning, i.e. as the starting point of movement and
rest or as form. Aristotle is rather saying that those things which appear frightening to everyone, or
at least to the majority of humans, can be said to be frightening to human nature in general. It is
thus common to everyone to experience fear in certain circumstances. Fear in the absolute sense
does not occur to a certain individual and in a certain situation because everyone acknowledges
that some things, actions or events are capable of generating pain and, consequently, fear.
Despite their differences, at least one common feature can be found both in the broad and
the technical sense of nature: in both cases nature is characterised in terms of “frequency” and
“regularity”. What is natural occurs uniformly, as it is present in every individual of a given
group and occurs in a regular manner through time. For instance, every apple tree produces
apples and every human being is scared of certain things. Moreover, the same apple tree will not
produce apples in the springtime and pears in the summer and, in this sense, what is natural is
not susceptible to variability.
As to the technical notion, Aristotle states that whatever occurs by nature occurs in a
uniform and regular manner:
[…] for if anything is of a certain character naturally, it either is so invariably and is not
sometimes of this and sometimes of another character (e.g. fire, which travels upwards
naturally, does not sometimes do so and sometimes not) or there is a ratio in the
variation. (Phys. 252a17–19, trans. Hardie and Gaye in Aristotle, 1984)
A natural process always tends to have the same outcome unless the irregularity is explained
through a violent cause which prevents the natural course of the process. Natural processes are
thus one-sided. An apple tree can only produce apples and not apples and/or pears; fire moves
naturally upwards and not sometimes upwards and sometimes downwards, unless some violent
cause interferes. In a natural process, therefore, X and not-X are asymmetrical: for instance, a
human being does not have the same possibility of aging and not aging. She will age always and
necessarily unless a violent cause interferes, e.g. a premature death impeding the occurrence of
the aging process. As to the broad sense of nature, things capable of generating fear absolutely,
and thus frightening to human nature, will not be frightening to a certain individual X and will
not be frightening at all to a certain individual Y.
Two features can thus be identified:
a.

What is by nature occurs regularly and uniformly. From this perspective, it is possible to
explain why Aristotle uses the word nature to mean “common”: some passions are said to
9

Riccardo Chiaradonna and Flavia Farina

be natural precisely because they are common to all human beings, i.e. they are present
regularly in every individual of the human species.
Natural processes are one-sided.

If character is said to become like nature, it is necessary to understand whether both nature and
character share these features.

4 Habit (hexis)
As is the case with nature, the notion of habit (hexis) has a multi-faceted background. In the
Categories [Cat.] Aristotle explains that habits or states (hexeis) and dispositions (diatheseis) form a
genus (the first one) within the category of quality. Habits are different from dispositions insofar
as they are stable and permanent:
One kind of quality let us call states and conditions. A state differs from a condition in
being more stable and lasting longer. Such are the branches of knowledge and the
virtues. For knowledge seems to be something permanent and hard to change if one
has even a moderate grasp of a branch of knowledge, unless a great change is brought
about by illness or some other such thing. So also virtue; justice, temperance, and the
rest seem to be not easily changed. It is what are easily changed and quickly changing
that we call conditions, e.g. hotness and chill and sickness and health and the like.
(Cat. 8b26–37, trans. Ackrill in Aristotle, 1984)
Aristotle refers the expressions “nature” or “by nature” to attributes which belong to certain
substances in virtue of what these substances are. The same cannot hold for habits which are
qualities and do not fall under the category of substance. In the Categories, Aristotle sets out a
division of “things that are (ta onta)”. The division includes particular substances (what Aristotle
calls “primary substances”: a particular human being, a particular horse), universal substances
(what Aristotle calls “secondary substances”: these are the genera and species under which
particular substances fall, e.g. human being, animal), particular properties (e.g. the color white
which belongs to Socrates) and universal properties (e.g. color, quality). Habits are neither
universal nor particular substances: rather, they are qualities inhering in substances. According
to Aristotle’s terminology in the Categories, habits or states inhere in substances as in a subject: so
they are not parts of substances and they cannot exist independently of them (Cat. 1a24–25).
In the Categories quality is defined as follows: “By a quality I mean that in virtue of which
things are said to be qualified somehow” (Cat. 8b25–26, trans. Ackrill). So Aristotle starts from
the fact that we call things “qualified” (e.g. we might say that a human being is white: is it
worth noting that according to Aristotle “white” refers both to the color of a particular thing
and to that particular thing insofar as it is colored). As a consequence of this, we regard the
quality (e.g. white) as that item in virtue of which things are said to be qualified.
These remarks suggest that nature and habit have a different status: for nature, unlike habit,
refers to what something is and not to some extrinsic property. Nonetheless, Aristotle’s division
of beings in the Categories allows for borderline cases. This happens with the so-called specific
differences (biped or rational vis-à-vis human being). Differences are not substances (they are
neither particular substances nor species and genera under which particular substances fall) but
share some features with substances, since they do not inhere in substances as properties inhere
in their subjects (Cat. 3a20–25). For example, the difference biped is not a mere external quality
belonging to any particular human being; this difference is part of the very nature of any
10

Nature, habit and character

particular falling under a species or a genus: “e.g. man is an animal of a certain quality because
he is two-footed, and the horse is so because it is four-footed; and a circle is a figure of
particular quality because it is without angles” (Metaph. 1020a34–35, trans. Ross in Aristotle,
1984). On the other hand, the difference is no genus or species under which particular substances fall (the status of biped is different from that of human being or animal). The difference
is actually a quality, but unlike other types of quality the difference is, so to speak, internal to
substance: it is an essential quality. Aristotle does not suggest that habits have the same status as
differences, for habits are genuine qualities and do not specify the essence or nature of their
subjects (a substance is not what it is in virtue of its habits). That said, since habits are durable
and difficult to remove from their subjects, their status comes close to that of differences. For,
unlike dispositions and other qualities, habits are stable and enduring characterisations of substances: so habits do not explain what a substance is (e.g. being virtuous does not explain what a
human being is) but, at the same time, they are more closely linked to substances than other
types of properties.
Other Aristotelian passages further expound the distinctive status of habits. Physics 7.3 is
particularly interesting. In this difficult chapter, Aristotle shows that alteration (i.e. change according to quality) only pertains to perceptible qualities and does not include other types of
properties such as habits (both physical and psychic habits). This view is in tension with the
Categories, where habits are included in the list of qualities: in the Physics the generation of habits is
separated from generation according to quality. Aristotle first argues that habits are to be identified
with virtues (or excellences) and vices (Phys. 246a11–12). This entails that habits as such have an
evaluative character: a habit is always a good or a bad habit and, from this fact, we can further infer
that habits are never neutral. As Ursula Coope (2012: 60) remarks, Aristotle “is assuming here that
a hexis is, by definition, something that contributes, either positively or negatively, to a thing’s
functioning”. This view does not emerge from the Categories, but a passage from the Metaphysics
lends support to it: “‘habit’ means a disposition according to which that which is disposed is either
well or ill disposed, either in itself or with reference to something else, e.g. health is a having; for
it is such a disposition” (Metaph. 1022b10–14, trans. Ross).
In Phys. 7.3 Aristotle connects habits with completions as different from alterations: acquiring a habit means acquiring a perfection or completion. For example, Aristotle says that it
would be odd to regard the completion of a house (the coping and tiling) as an alteration. This
entails both that a house is not altered when it acquires its coping or tiling, and that the coping
and tiling themselves are not alterations (Phys. 246a18–19: see Coope, 2012: 63). In addition to
this, Aristotle links habits to relations (Phys. 246b3–4): this means that habits are somewhat
constituted by (or depend on: the issue is controversial) relations holding between their underlying subjects (e.g. bodily excellences such as health and fitness consist in the blending of hot
and cold elements in due proportion). So habits come to be and pass away in the body or in the
soul when the body or the soul undergo appropriate changes. Even if habits are not alterations
in themselves, then, they are based on alterations which take place in the underlying subjects. In
virtue of these processes, the body or the soul are put in a certain condition regarding their parts
or external objects, and this is the habit (it is controversial whether “is” means strict identity or
some kind of dependence).
Aristotle regards virtues or excellences as completions or perfections. So he says that “when
anything acquires its proper excellence we call it perfect, since it is then really in its natural state:
e.g. a circle is perfect when it becomes really a circle and when it is best” (Phys. 246a13–15,
trans. Hardie and Gaye). When something acquires an excellence (a good habit), it comes to
have a stable disposition to function well (by having a healthy body or good character traits).
These remarks connect the notion of habit to that of nature, insofar as habits have an evaluative
11

Riccardo Chiaradonna and Flavia Farina

aspect in themselves: the virtuous habit (the excellence) holds when something is complete or
perfect. This does not mean that habits refer to what something is (a human being is not what
she is in virtue of her good habits); nonetheless, habits are some sort of natural completion or
perfection. They are stable conditions which entail that their subjects are complete and function well.

5 The unidirectionality of habits
In NE 1152a31, Aristotle states that character, once acquired, becomes like nature. Rather than
focusing on the process of acquiring habits, we will focus on an already morally formed agent.
In NE 1113b7–8 Aristotle also states that if it depends on the agent to act, it also depends on the
agent to refrain from action: here emerges a first significant difference between actions and
nature. Unlike a natural process, which is one-sided, actions are two-sided: if acting depends on
the agent, not acting depends on the agent too. Natural processes are not under the agent’s
control: so, e.g. the realisation of the aging process does not depend on the agent. However
Aristotle also states:
[…] for the same is not true of the sciences and the capacities as of states of character.
For it seems that whereas the same capacity or science deals with contraries, it is contrary
states of character which deal with contraries – for instance, as a result of health we do
not produce contraries but what is healthy; for we say a man walks healthily, when he
walks as a healthy man would. (NE 1129a11–15, trans. Barnes and Kenny)
The agent’s character states are acquired through the constant repetition of actions characterised
by the same quality as the character state which the agent will come to acquire: agents become
courageous by acting courageously and cowardly by acting cowardly. Virtue, once acquired, is a
habit, as is also the case with vice. Aristotle says that while sciences are concerned with both
contraries, habits, being contraries themselves, are not about contraries. Trying to make this point
clearer, we can say that while medical science is about health and sickness and can produce both
health and sickness in a patient, virtue is not about virtue and vice; as a consequence, it seems that
virtue cannot produce vice in an agent, because it is already a contrary.7 Aristotle also says:
[…] but actions and states are not voluntary in the same way: we control our actions
from the beginning to the end if we know the particular facts; but as for states, though
we control their origin, the particular way in which they develop is not known (any
more than it is in illnesses), and yet because it was in our power to act in this way or not
in this way, the states are voluntary. (NE 1114b30–1115a4, trans. Barnes and Kenny)
Two issues need to be investigated: the relation between habits and contraries; the different
ways in which habits and actions are said to depend on the agent.
Aristotle argues that the agent cannot bring about actions that are not consistent with the
quality of her character states: so habits are unidirectional and do not have a symmetrical relation with contraries. These remarks, applied to vices and virtues, suggest that an agent, having
formed a sound set of habits defining her moral character, is bound to act virtuously if she is
virtuous, and badly if she is bad. Hence, habits determine the moral quality of actions, thereby
excluding a whole class of actions contrary to the agent’s habits.
Stating that a habit is not about contraries does not imply, however, that habits determine a
single course of action. If habits determined actions exactly, the deliberative process could not
12

Nature, habit and character

take place: for the possibility of choosing between (at least two) alternatives is indeed a necessary
condition for deliberation itself. We suggest the following interpretation: the agent is still required to deliberate whether one alternative is more suitable or, roughly, better than another,
given the circumstances, but she will not choose between vice or virtue, because she will always
act coherently and consistently with her habits.8
According to the present reading, the moral qualification of the means is determined from the
start by the agent’s hexeis. However, the agent will still be in need to deliberate about the suitability
and correctness of the means, given the specific circumstances of the action. The agent will find
herself (1) not deliberating about the moral qualification of her actions, while, at the same time, (2)
deliberating about what could be the best course of action to take, given the circumstances.
Concerning (1), the agent will act consistently with her habits: for instance, as a healthy
human being will walk healthily on account of her healthy state, in the same way a brave
commander will act bravely on account of the hexis of bravery she possesses. She will not
cowardly run away from the battlefield or go hide in the woods, scared by the enemy. The
moral qualification of her action is already given.
However, concerning (2), she will have to consider the circumstances of her action and to
acquire the relevant information about the context in which the action will be accomplished:
she will still be in need of deliberating about the military strategy, the best time to act, the best
position to ensure that the army will be protected against side attacks and so on.
In conclusion, the agent does not deliberate about the quality of her action: an agent does
not consider courses of action characterised by contrary qualitative ranges as suitable alternatives, because habits, once acquired, are unidirectional (no more than a good-tempered agent
will consider homicide or violence an alternative when trying to settle a disagreement). This,
however, does not undermine the principle implied in the bilateral and symmetrical conception
of that which depends on us: if doing or not doing X depends on an agent, an alternative is
already given (without necessarily taking into account the moral qualification of the action).
Finally, Aristotle focuses on the dependency relation holding between actions and habits. An
agent is in control of her actions from beginning to end: if the accomplishment of an end requires
a series of intermediate actions, the agent is in control of the occurring or not occurring of the
intermediate actions. So if an agent wants to pay back a debt, she is in control of all the actions
needed in order to reach the goal, e.g. making or saving money, giving the sum that is owed back
and so on. As to habits, however, the agent is in control of their beginning (i.e. the process of
building up the moral character itself) but she is not in control of them from beginning to end, i.e.
the agent cannot give up a habit at her discretion and she is not in control of the moral qualification of her actions, once the habit is acquired. This situation is similar to that in which a
subject who has become sick is not in control of the manifestation and progression of the illness,
e.g. in the case of a chronic disease due to an unhealthy lifestyle. Hence, a brave agent will always
and constantly tend to act courageously, i.e. consistently with her habits. The soundness and
firmness of habits are indeed features that make people reliable in everyday relations9: a soldier is
charged with leading an attack in a battle precisely since it is conceivable that she will continue to
act courageously because she has already acted bravely in other circumstances. We can then
recognise that the moral qualification of actions depending on habits is necessarily consistent with
them, because habits are only open to courses of actions consistent with habits themselves.

6 Character as a second nature?
It is now possible to focus on elements of convergence and divergence between nature and
character. A first difference emerging between the agent’s moral conduct and nature is due to
13

Riccardo Chiaradonna and Flavia Farina

the fact that while nature is unilateral, human agency is bidirectional: an agent can X and not X,
can act and can refrain from acting. What depends on an agent is symmetrical, thus differing
from natural one-sidedness, which is asymmetrical. Nonetheless, the agent’s character is coherent and consistent if we look at the quality of the actions which follow from her habits. It is
important to note that this regularity does not directly affect the outcome, the particular action
performed by the agent, but rather its moral quality (i.e. regularity affects the type of action and
not its particular instance): the agent is still in control of her actions from beginning to end and
she can X and not-X.
There are two senses in which an agent can be said to act or refrain from acting. On the one
hand, we can consider the alternatives to be characterised by contrary qualities, i.e. vice or
virtue. In this sense, X and not-X are not both available to an already formed agent, because
habits are unidirectional. If, however, we consider X and not-X as alternatives characterised by
the same quality, e.g. as two alternative courses of action which are both courageous, then the
agent can be said to X or not-X. In this sense, there is an unbridgeable gap between the onesidedness of natural processes and the two-sidedness of what depends on the agent. But since
alternative outcomes are both characterised by the same quality, nature and character are similar: if the actions following from character are consistent with the agent’s habits, actions are
characterised by a certain kind of uniformity and regularity. A brave agent will regularly act in a
brave fashion and she will bring about actions characterised by courage in a constant and regular
manner.
We can thus compare the regularity of the moral qualification of actions to the regularity of
nature. A truly virtuous agent does not act virtuously sometimes and badly at other times, no
more than fire goes sometimes upwards and other times downwards.
That is why it is thought the mark of a more courageous man to be fearless and
undisturbed in sudden alarms than to be so in those that are foreseen; for it must have
proceeded more from a state of character, because less from preparation; for acts that
are foreseen may be chosen by calculation and reason, but sudden actions in accordance with one’s state of character. (NE 1117a17–22, trans. Barnes and Kenny)
Habits, once acquired, become a kind of moral instinct, which spontaneously determine the
moral quality of actions which follow from habits.
There is a further feature with respect to which character and nature can be said to be
similar: resistance to change, i.e. the feature mentioned in the NE passage quoted in the
opening of this chapter, where Aristotle states that changing one’s character is difficult as it
becomes like a nature (NE 1152a29–33). An agent, having formed her character, will act in
accordance with her habits. Once habits have been acquired, Aristotle says, it is no longer
possible to cease to be the persons we have become, as someone who has become sick, cannot
go back to being healthy simply by willing it. In order to change one’s character, the agent
should desire to be a different person, but Aristotle seems to disregard the possibility of desiring
a radical moral change, because habits are constant ways of expressing desires.10 Hence, it is
difficult to imagine a bad agent wanting to change her habits and becoming virtuous, because
habits shape the course of the desire itself. Secondly, making oneself a different kind of agent
through habituation requires time and perseverance: it requires a person to change her old
habits into new ones. However, the mature ethical agent will always act consistently with her
habits: a bad agent will act badly or poorly. How, then, could this agent get used to behaving
for a long time and constantly in a way that is contrary to her habits, which, on their part, are
unidirectional and hence give rise to actions consistent with themselves?
14

Nature, habit and character

Resistance to change and exactness in determining the quality range of actions are thus two
features which make nature and character similar: as nature determines exactly its outcomes, so
habits determine the quality of the actions following from them. In this respect, it is possible to
understand why character can be said to become a kind of nature that is difficult to change.
Despite the similarity, however, there is indeed a gap between nature and character, and this
makes an exact overlap between the two impossible: while nature is unidirectional, moral
agents can act or refrain from acting and this is, at least, one reason why the similarities between
character and nature in Aristotle do not amount to sameness.

Notes
1
2
3
4
5
6
7

Riccardo Chiaradonna is the author of Section 3; Flavia Farina is the author of Sections 1, 2, 4, 5.
McDowell (1998, 167).
McDowell (1998, 185).
Forman (2008, 571).
Morel (1997, 132).
Morel (1997, 132).
NE 1129a11–15 is a controversial passage. Natali (2017, 99) has expressed some reservations on
whether Aristotle is speaking on his own behalf here. The argument occurs in a methodological and
definitional passage. Furthermore, in the text we find the verb “dokei”, which could be a sign that
Aristotle is not expressing his own view on how habits work. However, even if it is true that the verb
“dokei” often signals a shared view or a shared starting point from which Aristotle develops his own
argument, this is not the only occurrence and meaning of the verb. Sometimes Aristotle uses the term
“dokei” in order to corroborate his own thesis through the appeal to a general agreement (NE 1174b8;
NE 1173a1). Indeed, the argumentative context to which Aristotle is referring, i.e. the distinction
between sciences and habits, can be viewed against an Aristotelian background such as Metaph. 9.2–5;
Cat. 8–9.
8 Bobzien (2014, 93) argues that an agent, being the person she is, and the circumstances in which the
agent finds herself are all the causal factors we need to bring about the agent’s action. Apparently,
Bobzien still argues that the agent’s action and deliberation are causally determined by her moral
character and by the circumstances at hand in a quasi-mechanistic process à la Furley (1967, 219). We
would argue that both Furley and Bobzien neglect (1) the nonhomogeneity of the action’s causal chain
(Natali 2004, 190) and (2) Aristotle’s statement that an agent can always be said to be in control of
accomplishing her action or refraining from acting. Furthermore, since we can think of more than one
course of action which could be consistent with the agent’s habits in the given circumstances, whether
she will accomplish an action or another cannot only be determined by the circumstances themselves
and by the interaction between circumstances and habits.
9 On this point, see Frede (2014, 49).
10 Natali (1999, 473).

References
Anscombe, G.E.M. Modern Moral Philosophy. Philosophy, 33, (1958). 1–19.
Aristotle. J. Barnes and A. Kenny (eds.), Aristotle’s Ethics: The Complete Writings. Princeton University
Press. (2014).
Aristotle. In J. Barnes (ed.), The Complete Works of Aristotle, 2 vols. Princeton University Press. (1984).
Bobzien, S. Choice and Moral Responsibility (NE III 1-5). In R. Polansky, The Cambridge Companion
to Aristotle’s Nicomachean Ethics. 81-109. Cambridge University Press. (2014).
Bonitz, H. Index Aristotelicus. De Gruyter. (1961).
Coope, U. Commentaries on the Six Sections of the Chapter: 46a10–246b3. In S. Maso, C. Natali, and G.
Seel (eds.), Physics VII.3: “What Is Alteration?”, pp. 57–72. Parmenides Publishing. (2012).
Donini, P. Ethos: Aristotele e il determinismo. Edizioni dell’Orso. (1989).
Donini, P. Abitudine e saggezza. Edizioni dell’Orso. (2014).

Riccardo Chiaradonna and Flavia Farina
Farina, F. La natura e gli stati abituali nell’Etica Nicomachea – Analogie e differenze. In G. Bagnati, M.
Cassan, and A. Morelli (eds.), Le varietà del naturalismo. Philosophica, Vol. 4, pp. 31–46. Ca’ Foscari
Edizioni. (2019).
Foot, P. Natural Goodness. Clarendon Press. (2001).
Forman, D. Autonomy as Second Nature: On McDowell’s Aristotelian Naturalism. Inquiry, 51(6), (2008).
563–580.
Frede, D. Free Will in Aristotle?. In P. Destrée, R. Salles, and M. Zingano (eds.), What Is Up to Us? Studies
on Causality and Responsibility in Ancient Philosophy, pp. 39–57. Akademia Verlag. (2014).
Furley, D.J. Two Studies in the Greek Atomists. Princeton University Press. (1967).
Hursthouse, R. On Virtue Ethics. Clarendon Press. (1999).
Kelsey, S. Aristotle on Interpreting Nature. In M. Leunissen (ed.), Aristotle’s Physics – A Critical Guide,
pp. 31–45. Cambridge University Press. (2015).
McDowell, J. Two Sort of Naturalism. In Mind, Value and Reality, pp. 167–197. Harvard University
Press. (1998).
Morel, P.-M. L’habitude: une seconde nature?. In Aristote et la notion de nature. Enjeux épistémologiques et
pratiques, pp. 131–148. Presses Universitaires. (1997).
Natali, C. L’action efficace: études sur la philosophie de l’action d’Aristote. Peeters. (2004).
Stavrianeas, S. Nature as Principle of Change. In M. Leunissen (ed.), Aristotle’s Physics – A Critical Guide,
pp. 46–65. Cambridge University Press. (2015).

2
SPINOZA AND LIBERAL
NATURALISM
Alex Douglas

Although I and others have argued in the past that Spinoza is not well-described as a naturalist
in most contexts (Douglas 2015a; Schliesser 2017), he does seem to pass the basic test of liberal
naturalism. Liberal naturalism focusses on the rejection of supernatural entities, defined as those
that are “above nature”, or “can violate the laws of nature” (De Caro and Voltolini 2010, 73).
Such entities are “in principle unaccountable by science, ineliminable from our ontology, and
contradictory to scientific knowledge” (De Caro and Voltolini 2010, 74).
Spinoza appears to overtly reject supernatural entities at various points in his writing. In Part
Two, Chapter 25 of his Short Treatise on God, Man, and His Wellbeing, he denies the existence of
devils (G1.107–108).1 In Chapter Six of his Theologico-Political Treatise, he explains miracles as
natural events whose causes we do not understand, rather than violations of the laws of nature
(G3.81–96). In the Appendix to Part One of his Ethics, he insists that God never acts against or
suspends the laws of nature (G2.77–83). And in the Preface to Part Three of the same work
there is a passage that can sound to contemporary ears like a rallying call for liberal naturalists:
There is nothing in nature that can be ascribed to any deficiency in it. For nature’s
virtue and power of acting is the same always and everywhere. That is, the laws and
rules of nature, according to which everything happens and by which one form is
changed into another, are everywhere and always the same. And our understanding of
the reason for things of every variety – the universal laws and rules – should likewise
be always the same. (G2.138)2
The context of the passage is that Spinoza is opposing those who treat humans as “a kingdom
within a kingdom”, who have the power to “greatly perturb the order of nature” – i.e. we
might say, as supernatural. Nothing, Spinoza assures us, perturbs the order of nature in this way.
The laws and rules of nature are everywhere the same. Nothing escapes them, not even God
himself, who, on some readings, is the laws of nature (Curley 1988, 1.16, esp. 42). This sounds
like a sort of naturalism, and many have described it as such (Bennett 1984, sect. 9.10; Della
Rocca 2008, 5–12).
Should we join them? I maintain my reservations. My main reason is this: while contemporary naturalism is typically explained in terms of some concept of the laws of nature, the
concept is grounded in natural science. For example, in the definitions of “supernatural” cited
DOI: 10.4324/9781351209472-4

Alex Douglas

above there is a graceful slide from the idea of violating the laws of nature to that of being unaccountable by science. If pressed to expand upon the phrase, “the laws of nature”, a contemporary
philosopher would be motivated to break “nature” down into a list of special sciences: the laws
of physics, of chemistry, of biology and so on. But in referring to the laws of nature – or to the
order of nature – Spinoza does not seem to intend anything discovered and established by the
natural sciences (Schliesser 2017). He intends something akin to a system of mathematical
objects, established and discovered through a form of deductive reasoning. His order of nature
is grasped by quasi-mathematical demonstrations that are “the eyes of the mind” (E5p23s,
G2.296),3 not dissected by the peeping and probing of the empirical scientist.
A guidebook on Spinoza, misguided in my view, suggests that he provided: “a systematic and
coherent vision of how the world might look to someone who is willing to suspend belief in
revealed religion and take very seriously the claims of the newly emerging natural science” (Cook
2008, 7). To cast doubt upon this, I give the example of the sole axiom to Part Four of the Ethics:
“Nothing is given among the singular things in Nature such that another more powerful and
stronger is not also given. Rather, for each thing there is given another, stronger, by which it can
be destroyed” (G2.210). I believe the language here to be deliberately mathematical. The discussion is not in terms of objects found in the world but rather of singular things, given, such that
others are given. This is language that one might find in a mathematics textbook, of the general
form x y:Ryx, as we are told for instance that for any given prime number there is given
another, larger prime number. And the consequence of Spinoza’s axiom is a fact – or a proposed
fact – that would startle those who view the world from a scientific outlook. The axiom tells us,
on the basis of no experimental evidence, that not only is there an infinity of concrete items in the
world but also that they form an unending series of increasing power.
While the terms are vague, an apparent application of the axiom to a particular case is this:
given that there is the Sun, there necessarily exists something more powerful and stronger than the Sun, and
something again more powerful and stronger than that, and so on without end. There is a question about
how to interpret “more powerful and stronger” in the axiom, but in some cases the ascending
hierarchy that it determines to exist might proceed into a domain of creatures that violate some
upper bounds set by the laws of physics: a star too dense and hot to have not burst or collapsed
for example. In this case Spinoza’s Nature will be populated by an infinite number of supernatural entities. This conclusion might be avoided by implicit restrictions on the meanings of
“more powerful and stronger” (an ascending hierarchy of ever-hotter things does not seem to
violate the laws of physics as such). But the real point here is that the attitude behind it involves
no naturalistic humility, waiting upon the authority of natural science to settle questions of
fundamental ontology. Quite the contrary; Spinoza, without leaving his chair, bodies forth an
infinite collection of astronomical discoveries.
Few commentators discuss the dramatic consequences of this axiom. Fewer still question
why Spinoza thinks we should believe it. One exception is Elizabeth Anscombe, who gives a
plausible speculation:
We may be reminded of St Anselm’s answer to Gaunilo, who treated Anselm’s argument for something greater or better than everything else in the world; Anselm says:
‘You may point to something which you think is greater than all the other things –
but I can then think of something that would be greater still.’ But, for Spinoza, what can
consistently be thought of must exist, being itself a possible thing; everything possible
exists necessarily, because God can cause it and he necessarily causes everything that is in
his power to cause.
(Anscombe 2011, 86)
18

Spinoza and liberal naturalism

This is very much in line with the premises Spinoza establishes up through Ethics 1p16: “From
the necessity of the divine nature must follow infinitely many things in infinitely many ways –
that is, everything which can fall under an infinite intellect”.
That proposition and its demonstration, in my view, destroy a standard philosophical distinction made between mathematical objects and concrete objects. In the case of the former,
most philosophers will accept that there is at least a prima facie case for saying that they exist
insofar as they can be consistently thought of. As Michael Dummett puts it, when it comes to a
system of mathematical objects, “[i]t must…be from the possibility of our knowing its existence
a priori that the necessity of its existence derives; and this is sufficient…to justify the assertion of
its existence” (Dummett 1991, 307–308). The same is not true of concrete objects, according to
most philosophers. But it is according to Spinoza, since his necessarily existing God necessarily
gives rise to everything which can fall under an infinite intellect. Proving the actual existence of
a concrete object is then a mere matter of proving its conceivability by an infinite intellect – for
this we have no better proxy than conceivability by our own finite intellects. The discoveries of
the natural scientists, whatever they may be, have no force to hold back this ontological tide,
since violations of the contingent laws they claim to discover are no less conceivable than
adherences.
If Spinoza’s “order of nature” imposes any restriction upon ontology, therefore, it can only
be an order whose existence is grounded on its conceivability rather than one uncovered by the
piecemeal empirical work of natural science. Immediately we run into the difficulty that there
appear to be many conceivable “orders” – many different and apparently inconsistent systems
determining the ordering and connection of beings in nature. The solution most consistent
with Spinoza’s commitments in E1p16 is, in my view, to deny the mutual inconsistency of
these various orders and have them existing alongside one another – perhaps in real possible
worlds or Whiteheadian “Cosmic Epochs”. The theory that the specific sort of order we
observe in our portion of the universe – its apparent adherence to certain laws of physics, for
instance – is a merely local phenomenon that is not present elsewhere in the universe might
underlie Spinoza’s puzzling statement that he ascribes “neither beauty nor ugliness, nor order
nor confusion” to the universe as a whole (Letter 32 to Henry Oldenburg, G4.170a).
It is interesting that Leibniz uses the word “naturalism” to refer to Spinoza’s view that
everything conceivable must exist. He finds its roots in Descartes. After quoting Ethics 1p16, he
goes on:
This view is quite false, and makes the same mistake that Descartes insinuated, that
matter successively accepts all shapes. Spinoza begins where Descartes leaves off: in
naturalism.
(Leibniz 1989, 277, “Comments on Spinoza’s Philosophy”)
We can gather from his other criticisms that Leibniz’s complaint here is directed towards
Spinoza’s denial of a central role for God’s free choice in deciding which possibilities to actualize. No choice is left, of course, if God’s nature requires that everything conceivable must
actually exist. Descartes already reduces the role of divine choice in requiring, as he appears to
do in Part Three of the Principles of Philosophy (sect. 47), that matter form itself sooner or later
into every possible shape. Since shape largely defines body for Descartes, this leaves God with
no choice over which bodies to create. He must realize every geometrically possible body at
some point and at best can choose only when and where each comes into being and how many
times. Spinoza’s God does not seem to even choose these things. Every possible spatial and
temporal ordering of every possible body must obtain somewhere. Think of any series of shapes
19

Alex Douglas

that matter could successively take, over a period of any length you like, and that series will be
instantiated somewhere in the infinite history of the universe. After all, such a series is conceivable by a finite intellect, and a fortiori by an infinite one. The personal will of God is reduced
down to a very simple algorithm: actualize every geometrically possible series of material
formations.
This is what Leibniz calls “naturalism”: replacing the voluntary personhood of God with the
automatic grinding out of mathematical possibility. There is some kinship with naturalism in
the contemporary sense. Speaking loosely, a naturalistic philosopher in the contemporary sense
is likely to deny any sort of end-directed teleology or divine purposefulness at work in the
universe. So is a naturalist in Leibniz’s sense. But there is also a crucial difference.
Contemporary naturalism is not committed to the view that every purely mathematical or
logical possibility is actualised. It will be committed to that view if the natural sciences provide
reasons to be. But this will not then be a function of naturalism qua naturalism. By contrast,
commitment to that view is part of the definition of “naturalism” as Leibniz takes it.
Spinoza’s ontology also violates the implicit epistemology behind naturalism, even liberal
naturalism. De Caro and Voltolini note the following:
an advocate of supernaturalism may conceive of supernatural entities or forces that are
utterly detached from the natural world and therefore do not interfere in any way with
natural causal processes (Parmenides’s eternal and unchanging Being, the absolutely
self-sufficient God characteristic of some forms of Neoplatonism, or even the
Buddhist Nirvana, conceived as the complete extinction of the flame of the self may
be examples of this view). Unsurprisingly, however, these views appeal to special
cognitive powers – which typically include some extreme forms of mystic illumination – in order to account for the human capacity to grasp those noncausal and supernatural entities or forces. For the purpose of our discussion, it is crucial to notice
that those cognitive powers are absolutely irreconcilable with anything we could
intuitively regard as natural forms of understanding. This is enough to rank the views
that appeal to such kinds of entities within supernaturalism.
(De Caro and Voltolini 2010, 74–75)
Spinoza’s God seems to belong in the same broad category of metaphysical posits as
Parmenides’s Being or the God of Neoplatonism. It is hopeless to try to distinguish Spinoza
from Parmenides or the Neoplatonists by holding that his God, which he sometimes calls
Nature, is an object discovered and investigated through the empirical sciences. This is because
Spinoza, as I have argued elsewhere (Douglas 2015b, 87–92), insists that God is best known a
priori. God is grasped via deductive demonstrations – the eyes of the mind – a process not
altogether unlike Parmenides’s Way of Ideas. Spinoza’s metaphysics, grounded as it is on his
idea of God, seems as much a violation of naturalism as Parmenides’s philosophy of Being. If
naturalism rules out whatever cognitive powers would be required to grasp Parmenides’s Being,
it should also rule out whatever cognitive powers are required to grasp Spinoza’s God, and
therefore Spinoza’s entire ontology.4 In addition there is the infinite hierarchy of beings with
ever greater power found through the axiom referred to above. This infinite group also must be
grasped through nonempirical means. Again there is dependence upon a cognitive power not
recognised by naturalism.
It is important, however, to consider why the grasping of such objects should constitute a
violation of naturalism. The problem is the causal inertness of the objects in question. A naturalistic account of knowledge will want to say that objects come to be known to us by acting
20

Spinoza and liberal naturalism

upon us, leading to a problem about our cognitive access to abstract objects outside the causal
order of concrete nature. Platonism about mathematical objects is uncomfortable to naturalists
for this reason. Even if we make an exception for numbers, we should not go on making
exceptions for Parmenidean Being, or the God of the Neoplatonists – or that of Spinoza.
Spinoza certainly concedes that God is in a causal relation to us, as well as everything else
(E1p18, p25). But it seems unlikely that this causal relation is what naturalists think of as a
“natural causal process”. There is reason to believe that Spinoza believes that natural causal
processes can only involve modes of God rather than God in himself (Laerke 2012). And in any
case there is no obvious way to explain how God’s causal relationship to us explains our
knowledge of him in a way that would make sense to a naturalist. Spinoza’s own explanation of
why the human mind has adequate knowledge of God – or, as he says, of God’s eternal and
infinite essence – does not refer to any causal relation between the mind and God but rather to
the fact that the idea of God’s essence is logically involved in all other ideas (E2p46d).
How does the mind have access to its objects? This question receives a very different
treatment in the context of Cartesian philosophy – Spinoza’s context – than it does in the
contemporary context.
In the Cartesian context it is natural to believe that whatever is thinkable has, ipso facto, a
certain sort of being, and a fairly robust one at that. The relation between the being of such an
object and our thoughts about it cannot depend on any natural causal process. Johannes
Clauberg (1622–1665) was the first philosopher to attempt to develop a system of Cartesian
logic. He also attempted to bring Cartesian ideas to bear on traditional questions of ontology.
According to him, the absence of contradiction in our idea of something is enough to show that
that thing has a real existence:
If that which we think does not imply any contradiction in our thought…to the point
that we judge that this thing is in nature or at least that it could be, then we attribute to
it not only objective being but also real being, and we…call it not only noeton, intelligible, but also eton, some real thing and properly ‘something’.
(Clauberg 1968, 1.285)5
Here “objective being” refers, roughly, to representational content in an idea.6 Clauberg is
asserting that if our idea of something is noncontradictory, then the object has some being
beyond this representational content. Such an object may only possibly exist in nature – “in
rerum natura…esse posse”. But it is no less real for that, and our minds can access such an object
simply by considering whether its idea of that object is not self-contradictory. Beyond the realm
of nature – rerum natura – there is the realm of objects with real being. Our minds can access this
realm. And its access cannot depend on any natural causal process.
Descartes had proposed that the mind can know “eternal or immutable forms or essences”
that, again, do not exist in nature:
when, for example, I imagine a triangle, even if perhaps a figure of this kind exists and
will exist nowhere outside my mind, nor will exist, still there is determined some
nature or essence or immutable and eternal form of it, which is neither made by me
nor depends on my mind. (Meditation Five, AT 7.64)7
These objects are known through innate ideas. But in his Comments on a Certain Broadsheet
(1647), Descartes points out to Regius that innate ideas do not come from external objects but
rather from “the faculty of thinking that is in me”:
21

Alex Douglas

I have never written or judged that the mind needs innate ideas as something distinct
from its faculty of thinking. Rather, I have taken note of those thoughts in me that
come neither from external objects nor from the determination of my will, but only
from the faculty of thinking that is in me. (AT 8b.357)8
If we read these two passages together we can piece together a suggestion that the mind, by the
mere exercise of its inner power of thinking, can come to grasp immutable essences or forms
that are not mere products of that power – since Descartes asserts that they do not depend on his
mind. This is perhaps then the origin of Clauberg’s notion of real beings outside nature.
From a contemporary point of view, this is puzzling. Why should the fabrication of an idea
by our mind’s intrinsic powers be construed as the grasping of an object outside of it? Clauberg
quite clearly interprets Descartes’s theory of “intellectual perception” or “cognition” as a way of
grasping objects outside of the mind, one that is distinct from sensory perception. On this,
Massimiliano Savini comments as follows:
The distinction between perceptions conditioned by the corporeal regime and perceptions that have their origin in the pure intellect grounds the distinction employed by
Clauberg between two types of knowledge: cognitio sensualis and cognitio
intellectualis…The difference between cognitio intellectualis and cognitio sensualis is not…
due to a difference in the objects brought before them, but rather to an intrinsic difference in the perceptive acts that sustain the one and the other species of knowledge. In
the third meditation Descartes had clarified this point with the example of the sun,
whereby we can have two types of ideas of the same object. One is constituted starting
from the representation offered by the senses. The other derives from innate ideas that
the soul possesses in itself. Clauberg’s commentary establishes the difference between
these two ideas on the basis of the heterogeneity of the perceptions that sustain them….
(Savini 2011, 229, my translation)
It might sound strange to describe an idea that is derived from innate ideas, which the soul
possesses or generates within itself, as a perception. But Descartes does describe it this way, and
Clauberg takes the description seriously. In fact, deriving our idea of the sun from our own
innate ideas is, Clauberg insists, the only way to perceive the sun clearly and distinctly
(Clauberg 1968, 1.256).9
The next important point is that Clauberg would most likely not regard any consistent
thought of a thing as enough to guarantee the “real being” of the thing. Descartes, in his replies
to Gassendi concerning Meditation Five, had made an important distinction:
it should be noted that those ideas that do not contain true and immutable natures, but
rather are merely fictitious and assembled by the intellect, can be divided by that same
intellect, not only by abstraction but by a clear and distinction operation. Therefore those
that the intellect cannot thus divide were undoubtedly not assembled by it (AT 7.117).10
If, therefore, we encounter a clear and distinct idea that is not divisible by the intellect, we
know that it must contain a true and immutable nature. That the mind brings an idea into being
through the exercise of its own power does not entail that it has constructed the content of the
idea. That consequence holds only in cases where the intellect can break down the content it
has assembled together. But it is presumably only ideas that contain true and immutable natures
that guarantee real being for Clauberg.
22

Spinoza and liberal naturalism

Clauberg’s implicit reasoning, I believe, is as follows. There must be something to explain
the unified content of an idea that contains a true and immutable nature. This cannot be the
mind’s activity in constructing the idea. And so it must be the real being of something outside
the mind. That said, the manner in which this real being explains the content of the idea remains
mysterious. The idea is generated entirely by the mind’s power of thinking, not by the causal
influence of the real being. The perfection of the mind or the goodness of its creator might be
invoked to explain why the ideas that spring forth from its own power of thinking correspond
to real beings.
It is still not clear why we should construe this as the mind grasping or perceiving the object
rather than generating an idea that fortuitously or providentially matches up with it. But the
historical point here is that it is quite natural in the Cartesian context to speak of the mind
“perceiving” objects with which it has no direct causal contact. This distances the Cartesian
theory of mind from the sort of theory that must stand in the background of contemporary
naturalism. The Cartesian theory already contains a notion of nonsensory, intellectual perception. No new, abnormal cognitive powers must be added to explain how the mind could
grasp causally inert objects such as “real beings” which exist outside of rerum natura.
Spinoza, taking the demonstrations to be the eyes of the mind, simply develops the Cartesian
theory of intellectual perception in a particular direction. In a certain sense we can say that his
theory does not countenance the existence of anything beyond what we can observe. But this
holds only if “observation” is extended to embrace intellectual perception. This seems a position quite distant from that of contemporary naturalism. Again I stress that Spinoza undermines the standard distinction between abstract and concrete objects. The natural world, for
him, consists of objects whose existence can, like that of mathematical objects, be detected
through some sort of inward intellectual operation distinct from sensory experience.
It is not that perceiving with “the eyes of the mind” is, for Spinoza, exactly equivalent to
perceiving with the senses. The analogy is with the apprehension of mathematical truth.
Dummett offers two criticisms of the Platonist notion that apprehending mathematical truth is
somehow equivalent to perceiving physical objects (Dummett 1978, 202–203). First, he argues,
physical objects are perceived by causally acting upon us, unlike mathematical objects. Second,
mathematical truths are, at least some of the time, discovered through a process of deduction:
“To describe the transition from premisses to conclusions in some step within a deductive
argument as a report of mental observation of logical connections is simply to blur a valuable
distinction” (Dummett 1978, 203). But neither Spinoza nor Descartes before him would disagree with Dummett on either point.
On the first, intellectual perception, we have seen, is distinct from sensory perception and
thus requires no causal contact with the object. The second point is trickier. Leslie Beck notes
that, on the one hand, Descartes does seem to reduce deduction to mere mental observation:
[T]he perfect deduction is reducible, if the inference is evident, to a true intellectual
intuition and this latter is a direct and immediate vision of the truth. Given the proper
conditions of attention, &c., there is no need of elaborate formulas; it is merely a
question of seeing.
(Beck 1952, 109)
Yet, as Beck goes on:
Like Spinoza, Descartes does not conceive intellectual intuition, or what the Ethics
calls Scientia Intuitiva, as an immediate knowing exclusive of reasoning. To see is, in a
23

Alex Douglas

sense, to reason: “the eyes of the mind, with which it sees and observes, are the
demonstrations themselves”.
(Beck 1952, 109–110)
Dummett’s distinction is not, therefore, blurred. Observing with the senses requires no reasoning. Observing with the eyes of the mind does require it. The fact that the latter depends on
reasoning does not stop it from being a type of observation.
This is not to say that Spinoza and Descartes agree on what this second sort of “intellectual”
observation requires. Olivier Dubouclez proposes that, according to Descartes, out of the two
species of mathematical reasoning – analysis and synthesis – only the first allows for the actual
discovery of objects; he calls this the “apophatic power of analysis” (Dubouclez 2013, 264–266).
Synthesis, on the other hand, is the method of demonstration favored by Spinoza in the Ethics.
Some Cartesians, for example Christoph Wittich, saw in this the Achilles heel of his system
(Wittich 1690; Verbeek 2005; Douglas, 2015a). Perhaps, then, there was disagreement on
which sort of reasoning should really count as a species of observation, capable of directly
grasping and even discovering a certain special class of objects.
But in any case Spinoza was working within a philosophical context in which the idea of the
mind being able to cognitively grasp objects outside of itself by some sort of nonsensory intellectual operation was fairly uncontroversial. He did not, therefore, face anything like the sort
of explanatory challenge faced by a contemporary naturalist who wishes to posit noncausal
entities. To understand this is to understand the most crucial point of difference between
Spinoza’s naturalism and the contemporary variety. Spinoza believed in a natural, inviolable
order. This was akin to a logical or mathematical system, accessible to some rational faculty and
populated by entities that are discoverable by the same faculty. The existence of a God (also
revealed through rational demonstration) who realises every conceivable possibility destroys the
distinction between abstract and concrete entities. The whole of nature is transparent to a single
intellectual operation – the test of consistent conceivability.
Spinoza’s philosophy is naturalistic in the sense of seeing in the universe an inviolable natural
order, and in replacing the providential divine will of traditional theology with something more
algorithmic. But in Spinoza’s case this is the automatic actualisation of every logical possibility
rather than, as in some forms of contemporary scientism, the playing out of a deterministic
system seeded by some brute-contingent initial event. Spinoza is a naturalist in the Leibnizian
sense, not in the contemporary sense. Spinoza’s naturalism is built on two related features that
are make more sense in his early modern context than in ours. The first is a blurring of the
abstract/concrete distinction. The second is belief in a rational faculty capable of intellectually
perceiving real beings. Can liberal naturalism stretch to accommodate these?

Notes
1
2
3
4

“G” refers to (Spinoza 1925).
All translations of Spinoza are mine.
“E5p23s” means: Ethics, Proposition 23, Scholium.
Although Spinoza heretically claims that God is extended, this does not entail that his God is a physical
object amenable to study by empirical physics. Extended objects might not be physical objects in that
sense – and Alison Peterman gives reason to think that for Spinoza they are not (Peterman 2015).
5 “Si illud, de quo cogitamus, nullam involvit in cogitatione nostra repugnantiam…adeo ut judicemus id esse in
rerum natura aut saltem esse posse, tunc ei non modo esse objectivum, verum etiam esse reale attribuimus, nec solum
νοητόν, intelligibile, sed etiam ἐτόν, reale quid et proprie Aliquid,…appellamus” – Metaphysica De Ente 3.18.

Spinoza and liberal naturalism
6 Do not be confused by the more recent use of “objective” to contrast with “subjective” and to mean
something “real outside the mind”. Here “objective” expresses the property of being an object of
thought and means almost the opposite of “objective” in the more recent sense.
7 “cum, exempli causa, triangulum imaginor, etsi fortasse talis figura nullibi gentium extra cogitationem meam existat,
nec unquam extiterit, est tamen profecto determinata quaedam ejus natura, sive essentia, sive forma, immutabilis &
aeterna, quae a me non efficta est, nec a mente mea dependet.” “AT” refers to (Descartes 1897) – all translations
are mine.
8 “Non enim unquam scripsi vel judicavi, mentem indigere ideis innatis, quae sint aliquid diversum ab ejus facultate
cogitandi; sed cum adverterem quasdam in me esse cogitationes, quae non ab objectis externis, nec a voluntatis meae
determinatione procedebant, sed a sola cogitandi facultate, quae in me est”.
9 Corporis animae in homine coniunctio, 51.6.
10 “advertendum est illas ideas, quae non continent veras & immutabiles naturas, sed tantum fictitias & ab intellectu
compositas, ab eodem intellectu non per abstractionem tantum, sed per claram & distinctam operationem dividi
posse, adeo ut illa, quae intellectus sic dividere non potest, procul dubio ab ipso non fuerint composita”.

References
Anscombe, G.E.M. How Can and Man Be Free? Spinoza’s Thought and That of Some Others. In From
Plato to Wittgenstein: Essays by G. E. M. Anscombe, pp. 92–103. Imprint Academic. (2011).
Beck, L.J. The Method of Descartes: A Study of the Regulae. Oxford University Press. (1952).
Bennett, J. A Study of Spinoza’s Ethics. Hackett Publishing. (1984).
Clauberg, J. Opera Omnia Philosophica. Olms. (1968).
Cook, J.T. Spinoza’s Ethics: A Reader’s Guide. Bloomsbury Academic. (2008).
Curley, E. Behind the Geometrical Method: A Reading of Spinoza’s Ethics. Princeton University Press. (1988).
De Caro, M. and Macarthur, D. (eds). Naturalism and Normativity. Columbia University Press. (2010).
De Caro, M. and Voltolini, A. Is Liberal Naturalism Possible? In M. De Caro and D. Macarthur (eds.),
Naturalism and Normativity, pp. 69–86. Columbia University Press. (2010).
Della Rocca, M. Spinoza. Routledge. (2008).
Descartes, R. C. Adam, and P. Tannery (eds.), Oeuvres de Descartes. Cerf. (1897).
Douglas, A. Christoph Wittich’s Anti-Spinoza. Intellectual History Review, 24(2), (2015a). 153–166.
Douglas, A. Was Spinoza a Naturalist? Pacific Philosophical Quarterly, 96(1), (2015b). 77–99.
Dubouclez, O. Descartes et la voie de l’analyse. Presses Universitaires de France. (2013).
Dummett, M. Platonism. In Truth and Other Enigmas, pp. 202–214. Harvard University Press. (1978).
Dummett, M. Frege: Philosophy of Mathematics. Harvard University Press. (1991).
Laerke, M. “Deus quatenus”… Sur l’emploi des particules réduplicatives dans l’Éthique. In M. Delbraccio,
P.-F. Moreau, and C. Cohen Boulakia (eds.), Lectures Contemporaines de Spinoza. Presses Universitaires
de Paris-Sorbonne. (2012).
Leibniz, G.F. R. Ariew, and D. Garber (eds., trans.), Philosophical Essays. Hackett. (1989).
Peterman, A. Spinoza on Extension. Philosophers’ Imprint, 15(14), (2015). 1–23.
Savini, M. Johannes Clauberg: Methodus cartesiana et ontologie. Vrin. (2011).
Schliesser, E. Spinoza and the Philosophy of Science. In M. Della Rocca (ed.), The Oxford Handbook of
Spinoza, pp. 155–189. Oxford University Press. (2017).
Spinoza, B.C. Gebhardt (eds.), Opera. Carl Winter. (1925).
Verbeek, T. Wittich’s Critique of Spinoza. In Schmaltz, T. (ed.), Receptions of Descartes: Cartesianism and
Anti-Cartesianism in Early Modern Europe, pp. 103–116. Routledge. (2005).
Wittich, C. Anti-Spinoza, Sive Examen Ethices Benedicti de Spinoza, et Commentarius de Deo et Eius Attributis.
Wolters. (1690).

3
HUME AND LIBERAL
NATURALISM
Benedict Smith

1 Introduction
The work of David Hume (1711–1776) is regarded as one of the most influential articulations
of a naturalistic approach in philosophy, interpreted by many as a key inspiration for naturalist
projects in a range of areas. His A Treatise of Human Nature (1739/40), An Enquiry Concerning
Human Understanding (1748) and An Enquiry Concerning the Principles of Morals (1751),1 amongst
other works, have been interpreted as resolute expressions of a philosophical attitude guided by
the methods, metaphysics and epistemology of natural science. As is often noted, Hume’s
naturalistic method seems evident from the subtitle of the Treatise: “Being an Attempt to
Introduce the Experimental Method of Reasoning into Moral Subjects”. Hume’s ambition was
to investigate the foundations of our thought and practice with the aim of explaining our
beliefs, ideas, emotions and behavior in ways that rendered them intelligible by science. But
characterising the role of science and naturalism in Hume’s philosophy is not straightforward.
According to some readings, Hume’s stated task of understanding ourselves and our ideas involved placing relevant phenomena into the explanatory frameworks of the natural sciences, an
expression of orthodox philosophical naturalism. But other readings emphasise quite different
aspects to Hume’s naturalism and clarifying these shows the close connections Hume’s approach
has with more “liberal” forms of naturalism.
One of the many things that makes Hume’s views distinctive is how he attempted to incorporate our rational capacities into a relevant naturalistic account. Our capacity for judgement
as well as our emotional and biological drives are by “an absolute and uncontroulable necessity”
determined by nature (T SBN, 183). Thus rationality is not conceived as a faculty radically
distinct from our more animal characteristics as it would be by, say, a Cartesian perspective.
Hume’s naturalism, understood one way, insists that every aspect of our bodily and cognitive
lives, as well our everyday interpersonal interaction including moral thought and agency, is explicable as phenomena within the scientifically described world. Under a particularly austere
version of this interpretation, Hume’s naturalism is a form of scientism: roughly, the dogmatic
assertion that the natural sciences have complete and exclusive authority over what can be said to
exist and the methods by which we can achieve genuine knowledge. However, there are elements at the core of Hume’s naturalism that motivate a quite different understanding of his work,
elements that situate him amongst other liberal naturalists and as a therapeutic liberal naturalist in
26

DOI: 10.4324/9781351209472-5

Hume and liberal naturalism

some respects. In the next section I lay out how Hume’s view is understood as a version of
scientific naturalism before turning to those elements that suggest a different interpretation.
Overall I discuss the relation between two cardinal themes in Hume’s philosophy, naturalism and
scepticism, focusing on the Treatise and An Enquiry Concerning Human Understanding.

2 Hume as scientific naturalist?
Interpretations of Hume have often presented him as articulating a clearly recognisable form of
scientific naturalism. By this I mean a broadly familiar commitment to the priority of the natural
sciences: philosophical inquiry should be constrained by the metaphysics, epistemology and
methodology of science. As contemporary proponent Alex Rosenberg puts it: “Naturalism is
the philosophical theory that treats science as our most reliable source of knowledge and scientific method as the most effective route to knowledge” (2013, 32). And, at least to many
modern readers, the way that Hume describes his project in the opening pages of the Treatise is a
clear expression of this kind of naturalism. There Hume explains that a proper “science of man”
or “science of human nature” will provide explanations of our ideas and behavior, a process
founded on “experience and observation” (T SBN, xvi).
Some have interpreted Hume’s naturalism as principally a metaphysical commitment about
the nature of the world. For example, David Lewis describes Hume’s world as “a vast mosaic of
local matters of particular fact, just one little thing after another” (1986, ix) and John McDowell
writes that “[t]he tendency of the scientific outlook is to purge the world of meaning” describing Hume as the “prophet” of this tendency (1998, 174).2 Such a scientistic outlook leaves
the world “disenchanted” in Max Weber’s sense and leads to a conception of nature as an
“ineffable lump” (McDowell 1998, 178). That characterisation of Hume’s naturalism is, I think,
a view that is possible only from a contemporary perspective. Hume describes his project as one
that takes human nature as continuous with the natural world, not as somehow standing over
against a disinterested objective reality or one that seeks to expel meaning or normativity from
the world. The aim of Hume’s science of man was to illuminate our ideas, thoughts, feelings
and agency. It is difficult to understand how these normative aspects of our nature would relate
to or be contained “in” a world conceived as a disenchanted normless lump implying that
human beings, if we wanted to account for the relevant kinds of normativity, would need to be
“enchanted wholly from within” to use Bilgrami’s phrase (2010, 31). Instead of starting with a
conception of the world as devoid of meaning and then conducting his science of man against
that background, Hume describes the science of man in the following way. After reflecting on
the limitations to the science of man, noting that the “ultimate principles” of human nature
cannot ever be known, Hume tells us that:
We must…glean up our experiments in this science from a cautious observation of
human life, and take them as they appear in the common course of the world, by
men’s behaviour in company, in affairs, and in their pleasures. (T SNB, xviii–xix)
The question of the starting point for Hume’s naturalism is important. In the above passage
from the introduction to the Treatise, Hume emphasises that the science of man takes as its
subject matter human life – our ordinary social existence and our ordinary everyday beliefs.
Furthermore Hume writes that “the science of man is the only solid foundation for the other
sciences” (T SBN, xvi). This emphasis, however, is problematically related to some of Hume’s
other philosophical commitments, particularly his empiricism. On one view Hume’s philosophical method began, as it were, with what can be detected on the inside of a subject’s mental
27

Benedict Smith

life, the internally accessible objects – “perceptions” in Hume’s sense – that a person can
allegedly identify by introspection. This partly expressed Hume’s imagistic theory of the relation between impressions and ideas and the subsequent psychological activities that supposedly
provide us with an intelligible world.
According to Stroud the “naturalism” of the twentieth and twenty-first Centuries is characterised by an attempt to explain the processes whereby we come to understand and know the
things we do, even if that attempt makes use of the kind of knowledge that we are trying to
explain, thus taking for granted forms of understanding that are then articulated in naturalistic
terms (Stroud 2016, 23). And the attempt is not restricted to a naturalistic conception of
knowledge. An influential trend in contemporary philosophy seeks to naturalise all aspects of
our cognitive and experiential lives including meaning and intentionality. Stroud takes such a
naturalistic project to be “continuous with” Hume’s conception of his science of man and at
least indirectly inspired by it (Stroud 2016, 23). However, Stroud indicates ways that Hume’s
naturalism contrasts with contemporary versions in the resources that are admissible at the
outset of naturalistic enquiry. Hume “started with what he thought human beings start with as
knowers: what they perceive in sense-experience”. The content of what knowers have in
sense-experience are transient atoms: “from these materials alone, Hume thought…human
beings construct their elaborate conception of the world and their place in it” (Stroud 2016,
23). So this emphasises a constructive natural psychological process that begins from what is
given in sense-experience. The “world” on this account is generated out of mental particulars.
Hume’s empiricist view is arguably unconvincing and stands in a difficult relation to other
components in Hume’s philosophy. But I will put these tensions, important as they are, to one
side. What is relevant is the aligning of Hume’s naturalism with a reductive form of psychological explanation. The alignment is encouraged by a modern view of what philosophical
naturalism amounts to, one expressed by Quine’s proposal to naturalise epistemology: our task
ought to be explaining the processes whereby we come to believe the things we do and that,
according to Quine, was the proper domain of psychology with Hume taken as an early
champion of this approach (Quine 1969).
But the psychological-naturalist reading of Hume ignores other dimensions to how the
science of man is characterised and thus other dimensions to his naturalism. As Annette Baier
has observed, for example, in setting out his science of man:
Hume was initiating not the science (in our sense) of psychology, either introspective
or experimental, but a broader discipline of reflection on human nature, into which
Charles Darwin and Michel Foucault, as much as William James and Sigmund Freud,
can be seen to belong.
(Baier 1991, 25)3
Baier emphasises the interpersonal dimension of Hume’s work, not in the sense that Hume was
just interested in trying to apply the discoveries of the science of human nature to the interpersonal world, but rather as regarding that world, our world, as the context within which that
science can operate. Yet the social word as well as other taken-for-granted aspects of human life
such as belief in the persistence of a self, the belief that ordinary everyday objects exist when
unperceived, that events are causally related and so on, can come under the scrutiny of reason
and disturbed from their usual place in our lives. It is this critical scrutiny of our natural thought
and agency that is an origin of scepticism, and I will come back to say a bit more about that and
its relation to Hume’s liberal naturalism below. In any case it is by no means obvious what
Hume’s naturalism amounts to. But it is misguided to portray it as a primary commitment to a
28

Hume and liberal naturalism

disenchanted conception of the world or as exclusively an early form of introspectionist
psychology.
As stated Hume’s overarching naturalism aimed to incorporate our rational capacities as well
as other kinds of behavior: “Nature…has determin’d us to judge as well as to breathe and feel”
as he puts it in the Treatise (T SBN, 183). A core theme in the Treatise is the project of clarifying
the epistemic status of our beliefs. According to Hume, rational reflection aimed at uncovering
the justificatory grounds of, for example, causal belief and inductive inference, reveals these to
be entirely without rational justification. Through reflection we detach ourselves from our
beliefs in the sense that we acknowledge their unjustifiable status yet the detachment cannot be
total. In some sense we remain existentially committed to them. By this I mean something
similar to Norman Kemp Smith’s influential interpretation of Hume’s naturalism in which he
describes our “natural beliefs” as those that we possess irrespective of the consequences of
rational scrutiny and without the use of reason or evidence (Kemp Smith 1949). This sets up an
important tension between the conclusions of philosophical reflection and the fact that the
“current of nature” nevertheless determines us to think and act according to these beliefs (T
SBN, 269). The tension is a difficult one to resolve, at least initially. In Book 1 of the Treatise,
Hume warned of the dangers of “intense reflection”:
[S]ceptical doubt, both with respect to reason and the senses, is a malady, which can
never be radically cur’d, but must return upon us every moment, however we may
chace it away…sceptical doubt arises naturally from a profound and intense reflection
on those subjects, it always encreases, the farther we carry our reflections, whether in
opposition or conformity to it. (T SBN, 218)
By the conclusion of Book 1 the malady had become something resembling the symptoms of a
psychiatric illness:
The intense view of these manifold contradictions and imperfections in human reason
has so wrought upon me, and heated my brain, that I am ready to reject all belief and
reasoning, and can look upon no opinion even as more probable or likely than another. Where am I, or what? From what causes do I derive my existence, and to what
condition shall I return? Whose favour shall I court, and whose anger must I dread?
What beings surround me? and on whom have I any influence, or who have any
influence on me? I am confounded with all these questions, and begin to fancy myself
in the most deplorable condition imaginable, inviron’d with the deepest darkness, and
utterly depriv’d of the use of every member and faculty. (T SBN, 268–269)
Immediately after this passage Hume writes:
Most fortunately it happens, that since reason is incapable of dispelling these clouds,
nature herself suffices to that purpose, and cures me of this philosophical melancholy
and delirium, either by relaxing this bent of mind, or by some avocation, and lively
impression of my senses, which obliterate all these chimeras. I dine, I play a game of
back-gammon, I converse, and am merry with my friends; and when after three or
four hour’s amusement, I wou’d return to these speculations, they appear so cold, and
strain’d, and ridiculous, that I cannot find in my heart to enter into them any farther.
(T SBN, 269)

Benedict Smith

A reintegration into ordinary life from a period of detached spectatorial reflection enables us to
re-inhabit, so to speak, our ordinary natural beliefs providing a form of antidote to scepticism.
The subject matter of Hume’s naturalism is human nature, and the content of the science of
man is what we in fact do, feel and think in ordinary life. As Strawson notes there (at least) two
Humes: the sceptic and the naturalist (Strawson 1987, 12). Scepticism emerges as a result of
“intense reflection” on the rational standing of our ideas, and nature, including human nature,
is articulated partly as a way to demonstrate the status of scepticism. That is to say that, for
Hume, getting clearer about human nature shows the proper place that detached reflection has
in our lives.4 In my view this is what indicates Hume’s therapeutic liberal naturalism. To
anticipate: Hume’s naturalism incorporates detached philosophical thought thus not regarding it
as a rival competitor perspective but as one expression of the diverse kinds of agency that
constitute human beings in the world.

3 Hume as liberal naturalist
“Liberal naturalism” has been characterised in a number of ways, emphasising different forms
that the process of liberalisation could take.5 A common factor in the different articulations of
liberal naturalism is an acknowledgement of the diversity of scientific understanding and,
crucially, to give a persuasive account of normative phenomena that are essential to our
thought, discourse and action.6 Such an account, however, could be in tension with an understanding of what is constituted by “scientific understanding”. If this means assuming
something like a disenchanted conception of the world, then what we are aiming to understand
will exclude normative phenomena. And I take it as uncontroversial that at least a core aspect of
Hume’s ambition was to account for normative phenomena such as our striving for and
achieving knowledge and understanding.
The label “liberal naturalism” was first used by John McDowell in 1999 to articulate a form
of naturalism that contrasted with what he called “restrictive naturalism” (2009, 261–262). This
contrast is aimed, partly, to acknowledge that thinking and knowing cannot be reduced to or
incorporated into how modern science conceives of what is natural. According to McDowell
the success of modern science can tempt us to equate what is natural with what can be exclusively understood by subsuming phenomena under scientific laws of nature, the “realm of
law” (McDowell 1996, 73). But for Sellarsian reasons thinking and knowing are achievements
that are intelligible only in terms articulated within a “logical space of reasons”, a normative
space that contrasts with the realm of law (Sellars 1997, 76). So if we insist on equating what is
natural with what is intelligible only in terms of the realm of law, then thinking and knowing
will be conceived of as supernatural. But regarding thought and knowledge as supernatural is
unattractive as well as deeply confused. To avoid this, McDowell suggests, we should regard
thinking and knowing as “part of our way of being animals” (2009, 262). This way of being
animals, for mature humans at least, incorporates our rational capacities. Our agency is infused
with rationality along with, for conceptualists like McDowell, our perceptual capacities. The
point here is not to pretend that Hume had a similar view of the role of concepts in experience
but to press the broader idea that both McDowell and Hume insist on the nonsupernatural
character of agency and thought. The idea that these are not supernatural might seem entirely
unremarkable. But from a scientific naturalist perspective what is needed to make sense of this is
to provide a naturalistic account of thinking and knowing whilst respecting their irreducibly
normative character and the worldly context, including interpersonal relations, that conditions
thought and agency in the first place. And that is an impossible task if the relevant kind of
naturalism can respect only what is intelligible in causal-explanatory terms.
30

Hume and liberal naturalism

To avoid supernaturalism about thinking and knowing we should be careful about what we
take the natural sciences to have achieved. According to McDowell “what the modern scientific revolution yielded was clarity about the realm of law, and that is not the same thing as
clarity about nature” (McDowell, 2013, 261). So being a naturalist about knowing and thinking
in a liberal sense opposes reducing or translating these achievements into the realm of law; a
liberal naturalist can point to the surely incontrovertible fact that thinking and knowing are
amongst the many things that we do, part of “our mode of living” in McDowell’s words
(McDowell, 2013, 261). There is nothing supernatural about thinking and knowing and their
naturalistic character cannot be made clear by understanding them in terms of the realm of law
or somehow as operations in a disenchanted world. One of the key elements in McDowell’s
presentation of liberal naturalism is the attempt to steer past what might otherwise be an intractable bifurcation between our being part of nature and our exercising rationality. And this is
clearly a Humean aim as well.
A liberal naturalist like a scientific naturalist denies supernatural phenomena including forces,
objects and agencies which are typically invoked by theistic accounts of the world. A difference,
however, is that for the liberal naturalist what is deemed to be the province of the natural is not
determined by only what natural science tells us. For the liberal naturalist persons, for example,
are a proper object of reflection and understanding. Persons cannot be understood except as
regarding them as embedded and engaged in a meaningful world with other persons, amongst
many other aspects. Persons are not, I am assuming, supernatural but just like the supernatural
forces, objects and agencies invoked by the world’s religions, persons would be excluded from
the scientifically described world. Liberal naturalism seeks to account for nonscientific, nonsupernatural forms of understanding and the worldly features to which those forms are answerable. Hume’s contribution to liberal naturalism lies in his rejection of supernatural
explanations and his commitment to the science of human nature as a form of enquiry that takes
“common life” as its subject matter. Our common life is constituted by a set of beliefs and
commitments about the existence of the external world, about the self, about other persons and
of course much else besides. These beliefs and the things they are about are not supernatural but
understanding them is not intelligible by trying to translate them into the framework of scientific naturalism. The beliefs and phenomena in question are irreducibly normative and
nonsupernatural. That makes liberal naturalism an appropriate perspective from which to understand them.
H.O. Mounce has drawn a distinction between a metaphysical and an epistemological way
to characterise Hume’s naturalism (Mounce 1999, 11–12). Metaphysical naturalism means
something along the lines of what McDowell assumed when describing Hume as the prophet of
disenchantment (McDowell 1998, 174), the scientistic idea that reality is identical to nature
understood through the categories of natural science. By “epistemological naturalism” Mounce
does not mean a Quinean view of giving a account of thinking and knowing in descriptive
psychological terms. What is meant, rather, is the idea that our human nature, the target of the
science of man, is constituted by our being foundationally and pre-philosophically related to the
world through forms of thought, feeling and action. These relations are not the outcome of
reasoning and inference or some kind of explicit cognitive achievement but are presupposed by
those things (Mounce 1999, 4). Hence there is a way that we inhabit the world that is primordially meaningful and that makes use of forms of understanding. In part Hume’s naturalism
aims to capture these pre-rational relations and to that extent shares its form with themes in
Wittgenstein.7 In a way that runs counter to elements in his empiricism, Hume’s naturalism
understood as epistemological in Mounce’s sense emphasises how our knowledge has its source
not merely in individualistic experience and inference from it. Such experience and reasoning is
31

Benedict Smith

only possible because we have a world already “given to us in natural beliefs” (Mounce 1999, 2).
We cannot generate the world and what it contains from private experience and inference. So
this suggests a quite different naturalistic starting point to that described by Stroud above.
Recall: the idea was that Hume started with what he thought human beings start with as
knowers, what they perceive in sense-experience. Stroud is right to say that Hume started with
persons as knowers. But in my view what that means implies a quite different point of departure
than imagining subjects of experience being affected by sensory input and constructing a world
from fleeting momentary impressions. Again for somewhat Sellarsian reasons, to start as a
knower is to start already within a world-involving normative context, a position that can be
characterised in naturalistic terms although not those of orthodox scientific naturalism. From
what Hume tells about how he characterises the science of man it is persons engaged in the
world and with others that is a principal focus despite the presence of his contrasting empiricist
commitments.

4 Philosophy and Hume’s therapeutic liberal naturalism
When Hume frames his science of man as resting on the cautious observation of common life,
he is not proposing that we somehow step outside nature, as if the relevant observational stance
is somehow detached from the natural world. He is proposing that, at least at an initial stage, we
step outside of custom and ordinary experience. He cannot be proposing that we step outside
nature for there is no such place. Rational reflection and adopting a detached critical stance is
one expression of our human nature and the process of coming to see how detached reflection
is related to common life is an important theme in Hume’s work. In the Enquiry Hume suggests
that a scepticism that is “antecedent to all study and philosophy”, a form of extreme Cartesian
doubt, is impossible for us. Were this doubt attainable, it would be “entirely incurable; and no
reasoning could ever bring us to a state of assurance and conviction upon any subject”. But
Hume does suggest that:
[T]his species of scepticism, when more moderate, may be understood in a very
reasonable sense, and is a necessary preparative to the study of philosophy, by preserving a proper impartiality in our judgments, and weaning our mind from all those
prejudices, which we may have imbibed from education or rash opinion. (E
SBN, 150)
So this is a “scepticism” geared to help the process of responsible belief-formation, not as a
strategy or a set of judgements aimed at undermining claims to knowledge.
Common life involves a pre-rational belief that our senses provide access to a real, independent world:
It seems evident, that men are carried, by a natural instinct or prepossession, to repose
faith in their senses; and that, without any reasoning, or even almost before the use of
reason, we always suppose an external universe, which depends not on our perception, but would exist, though we and every sensible creature were absent or annihilated. (E SBN, 151)
“But”, Hume continues, “this universal and primary opinion of all men is soon destroyed by
the slightest philosophy” (E SBN, 152). Philosophical reflection teaches us that, in fact, what
we have access to are only ever “perceptions in the mind”. So the natural realism of common
32

Hume and liberal naturalism

life is in tension with a philosophically informed phenomenalist anti-realism. Yet this philosophical position cannot itself be defended by reason. Thus “philosophy finds herself extremely
embarrassed”. Natural beliefs do not withstand critical assessment yet neither does the philosophical stance itself: “to justify this pretended philosophical system, by a chain of clear and
convincing argument, or even any appearance of argument, exceeds the power of all human
capacity” (E SBN, 152). External world scepticism resulting from philosophical reflection is
“contrary to natural instinct” but that cannot be a settled perspective, partly because the
scepticism itself “carries no rational evidence with it” (E SBN, 155).
As already described in the Treatise, the sceptical attitude arising from philosophical reflection
is undermined by engagement in ordinary life and thought. Hume writes in the Enquiry that “[t]
he great subverter of…the excessive principles of scepticism, is action, and employment, and
the occupations of common life” (E SBN, 158–159). What does the subverting, then, is not any
particular argument or operation of reason. Rational reflection produced the excessive scepticism in the first place. Our acting is what subverts scepticism. So if this is a form of refutation
then it will be of a distinctive type; not the outcome of reasoning but an embodied and enacted
repudiation of philosophical conclusions.
Engaging in the affairs of common life is not just a distraction from philosophy, a psychologically necessary relief from the disturbing predicament brought on by philosophical
reflections. Common life has also a normatively constraining role with regard to philosophical
activity. Common life is not a disenchanted life. But it is not enchanted wholly from within, to
borrow Bilgrami’s phrase again, occurring somehow within a pre-established and disenchanted
world. And neither is it a way of living that lacks all critical resource; it constrains philosophy
not just because we cannot exist as sceptics and are thus necessarily propelled back into our
everyday mindless routines. Rather, ordinary life shows us the limits of philosophical reflection
and thus reveals its character in more perspicuous terms.8 Excessive scepticism seems, quite
literally, pointless: “the chief and most confounding objection to excessive scepticism, [is] that
no durable good can ever result from it…We need only ask such a sceptic, What his meaning is?
And what he proposes by all these curious researches? He is immediately at a loss, and knows not what
to answer”. (E, 159–160). But excessive scepticism is not simply impotent and ultimately
useless. It is by working through excessive scepticism that a more “moderate” scepticism
emerges: “a more mitigated scepticism…which may be both durable and useful, and which may,
in part, be the result of this…excessive scepticism”. And crucially for the point I am trying to
make here the “undistinguished doubts” of excessive scepticism “are, in some measure, corrected by common sense and reflection” (E SBN, 161). So in order for excessive scepticism to
be “subverted” and “corrected” common life must be able to apply normative pressure on the
conclusions of philosophical thought. And the relation between common life and mitigated
scepticism is mutually illuminating:
Those who have a propensity to philosophy, will still continue their researches; because they reflect, that, besides the immediate pleasure, attending such an occupation,
philosophical decisions are nothing but the reflections of common life, methodized
and corrected. But they will never be tempted to go beyond common life, so long as
they consider the imperfection of those faculties which they employ, their narrow
reach, and their inaccurate operations. (E SBN, 162)
By the end of the Enquiry Hume describes a situation in which a person who has passed through
the “amazement and confusion” (E SBN, 160) of excessive scepticism, the unconstrained
application of philosophical scrutiny to our beliefs, can reach a way of thinking “both durable
33

Benedict Smith

and useful” by recognising the mutually illuminating character of common life and philosophy
(E SBN, 161). Excessive scepticism becomes mitigated by utilising an insight about the limitations of critical reflection, an insight that is provided by an enlightened attitude towards our
ordinary beliefs constitutive of common life. In some sense this captures a process of naturalising
philosophical activity though not of course by making philosophy a chapter of psychology in
Quine’s sense.
Hume’s liberal naturalism is expressed in the attempt to understand reason and rational
reflection as part of our natural mode of living; the transformation from excessive sceptical
disintegration to mitigated scepticism “comes from nature alone” as Stroud puts it (2016, 31). It
is this sense, then, that I think reading Hume as a therapeutic liberal naturalist is appropriate. All
I mean by therapeutic here is that idea that through entirely natural processes we come to a
form of understanding of both philosophical activity and of our ordinary “daily practice and
experience” (E SBN, 162). This understanding is the result of seeing our common life anew, an
understanding achieved by, partially, employing reason; we need to inhabit the “amazement
and confusion” of excessive scepticism reached through rational reflection as a stage on the way
to understanding. The transition from that state to one that is more “durable and useful” is not
the direct outcome of reason but of an insight once we come to see that the content and
significance of our ordinary lives “subvert” and “correct” the conclusions of philosophical
reflection.9
The doctrines of liberal naturalism do not form a precisely delineated set. But among the
significant themes is a view about the perspective from which naturalistic enquiry ought to
begin, or the context within which naturalism can proceed. If proponents of liberal naturalism
are right, orthodox scientific naturalism assumes that, for any given phenomena to be regarded
as genuinely part of the world and for us to secure genuine knowledge of that phenomena, the
metaphysics and methods of natural science are privileged. In his presentation of philosophical
(or scientific) naturalism, Rosenberg claims that what he calls the “interpretive disciplines” such
as history and other humanities, whilst important, cannot provide “real knowledge” since only
scientific naturalism is able to do that (2013, 41). It is of course no part of liberal naturalism to
deny that scientific naturalism is a vital source of knowledge and understanding. But liberal
naturalism does deny that this is to be regarded as the only or the best form of understanding
irrespective of what our subject matter is. Cases where the subject matter of what we are trying
to understand cannot even be identified without preserving its normative character exemplify
the issue. Hume’s science of human nature adopts different perspectives and one of these
perspectives takes human thinkers and agents as engaged in ordinary life and frames that subject
matter as basic and irreducible. Human life is also the context within which philosophical
reflection is initiated and, as Hume explains, to which it returns in some sense. Hume’s liberal
naturalism, then, sheds light on the character of ordinary life and of philosophical reflection.

5 Conclusion
In the opening sections of the Enquiry, Hume notes how the “mere philosopher” and the
“mere ignorant” lead an equally impoverished life (E SBN, 8). Hume proposes that another
way of living is preferable, one illuminated by a form naturalistic enquiry. Hume writes that we
are “reasonable”, “social” and “active” beings and that “nature has pointed out a mixed kind of
life as most suitable to human race. Indulge your passion for science, says she, but let your
science be human” (E SBN, 8–9).10 The diversity of human nature as revealed through Hume’s
science of man suggests that we need diverse means and methods for understanding ourselves. It
would seriously distort the character of Hume’s philosophy quite generally and his naturalism in
34

Hume and liberal naturalism

particular if we insist that the only naturalism present in Hume’s work is a forerunner of
scientific naturalism. Hume’s naturalism is more liberal than some interpretations have supposed
and appreciating this can enhance our understanding both of Hume and of the character and
prospects of liberal naturalism.

Notes
1 In referring to these works below, I will use “T” to refer to the Treatise followed by the page number
contained in the edition prepared by L.A. Selby-Bigge and P.H. Nidditch abbreviated to “SBN”. And
similarly with the Enquiry referred to as “E”.
2 See Smith (2016, 311).
3 In making this point Baier refers the reader to the work of Gilles Deleuze who, in his 1953 book on
Hume Empiricism and Subjectivity: An Essay on Hume’s Theory of Human Nature, writes that “Hume is a
moralist and a sociologist, before being a psychologist…one must be a moralist, sociologist, or historian
before being a psychologist, in order to be a psychologist” (1991, 21–22). See Smith (2016) for further
discussion.
4 Hume’s attempt to demonstrate the limits of scepticism and reason is an important theme in Mounce’s
interpretation of Hume’s naturalism (e.g. 1999, 15–16).
5 For previous examples of the ways that liberal naturalism has been motivated and defended, see the
essays collected in De Caro and Macarthur (2004) and (2010).
6 See, for example, De Caro and Macarthur (2010, 9).
7 See, for example, Pears (1990, xii) and Smith (2018).
8 The theme of Hume’s naturalism revealing the limits of reason and human understanding is important
to Mounce’ s account. See, for example, Mounce (1999, 22–23).
9 This transition is emphasised in more detail and framed narrative terms in Livingstone’s Hume’s
Philosophy of Common Life (1984, 28–31). See also Stroud (2016) for an account of the move from
excessive to mitigated scepticism.
10 See also Smith (2018, 256).

References
Baier, A. A Progress of Sentiments: Reflections of Hume’s Treatise. Harvard University Press. (1991).
Bilgrami, A. The Wider Significance of Naturalism: A Genealogical Essay. In M. De Caro and D.
Macarthur (eds.), Naturalism and Normativity, pp. 23–54. Harvard University Press. (2010).
De Caro, M. and Macarthur, D. Naturalism in Question. Harvard University Press. (2004).
De Caro, M. and Macarthur, D. Naturalism and Normativity. Columbia University Press. (2010).
Kemp Smith, N. The Philosophy of David Hume. Macmillan. (1949).
Lewis, D. Philosophical Papers, Vol. 2. Oxford University Press. (1986).
Livingstone, D. Hume’s Philosophy of Common Life. Chicago University Press. (1984).
Macarthur, D. Liberal Naturalism and the Scientific Image of the World. Inquiry, 62(5), (2019). 565–585.
McDowell, J. Mind and World. Harvard University Press. (1996).
McDowell, J. Two Sorts of Naturalism. In Mind, Value and Reality, pp. 167–197. Harvard University
Press. (1998).
Mounce, H.O. Hume’s Naturalism. Routledge. (1999).
Pears, D. Hume’s System: An Examination of the First Book of his Treatise. Oxford University Press. (1990).
Quine, W.V.O. Epistemology Naturalized. In Ontological Relativity and Other Essays. Columbia University
Press. (1969).
Sellars, W. Empiricism and the Philosophy of Mind. Harvard University Press. (1997).
Smith, B. Wittgenstein, Hume and Naturalism. In K. Cahill and T. Raleigh (eds.), Wittgenstein and
Naturalism. Routledge. (2018).
Smith, B. Naturalism, Experience and Hume’s ‘Science of Human Nature’. International Journal of
Philosophical Studies, 24(3), (2016). 310–323.
Stroud, B. Naturalism and Scepticism in the Philosophy of Hume. In P. Russell (ed.), The Oxford
Handbook of Hume, pp. 21–31. Oxford University Press. (2016).
Strawson, P. F. Skepticism and Naturalism: Some Varieties. Methuen & Co. (1987).

4
KANT ON NATURE
AND HUMANITY
Allen Wood

1 Introduction
Can we apply to Kant our terms for contemporary philosophical “-isms” (“naturalism”, “scientism”, “liberal naturalism”, etc.)? We could try, but to do so would be anachronistic in his
case, as in the case of any other eighteenth century philosopher. What I propose in this essay is
something different. It might be instructive to survey the position Kant takes on a series of issues
that were already live in his time and that bear some relation to the issues that might lead us to
classify our philosophical contemporaries using the “-isms” listed above.
Kant began his career as what we would now call a “natural scientist” – the name in his day
would have been “natural philosopher”. His doctoral thesis, De igne, was about combustion and
electricity, and was supervised by a physicist, Johann Gottfried Teske (1704–1772). Kant’s early
writings were mostly about physics, astronomy and geology, or what we would now call “earth
sciences” (a field in whose invention Kant can even claim some share). It was only slowly that
Kant began to combine the interest in natural science and its foundations with other areas we
might consider “philosophical”. Kant became a philosopher in our sense of the word only when
he began to reflect critically on the metaphysical and epistemological foundations of these
natural sciences and their relation to the dominant Wolffian philosophy of the time. Starting in
the 1770s, after Kant finally secured a professorship in Königsberg, he took up the study of
human nature – what he called “pragmatic anthropology” – another discipline to whose invention he also contributed. It is only in the 1780s and 1790s that Kant’s writings come to
include interests in morality, religion and politics; toward the end of his life, these interests even
became central, even dominant – but without ever displacing his broader concerns with natural
science and its foundations, which he pursued to the end of his life.
Let us begin with a sketch of Kant’s conception of what Kant thinks we can cognise through
our theoretical faculties – those of natural science or natural philosophy. Next we will consider
practical reason, the way Kant thinks we should view our predicament as agents, especially
regarding freedom of the will, morality and religion, including our attitude toward the ideas of
God and immortality of the soul. Finally, we should look at Kant’s conception of our
knowledge of ourselves, our human nature and predicament. This procedure would also follow
the order of Kant’s four famous questions, presented in the introduction to his logic lectures as
edited by Jäsche in 1800:
36

DOI: 10.4324/9781351209472-6

Kant on nature and humanity

“What can I know?
What ought I to do?
What may I hope?
What is the human being?” (VL 9, 25)
Kant then adds that the first three questions all have reference to the fourth.

2 Theoretical knowledge of nature
Sensibility and understanding
Kant sees himself as distinguished from his modern predecessors in holding that we have two
distinct cognitive faculties – sensibility and understanding – that must work together to provide
any genuine cognition. Kant thinks earlier modern philosophers attempted to reduce these
faculties to a single one: the rationalists (or “noologists”) by treating sensibility as a confused or
inferior form of thinking, and the empiricists by trying to reduce thinking to merely an association of sensations according to mechanistic laws (KrV A854/B892). By contrast, Kant
holds that sensibility and understanding each affords a separate source of cognition, and that the
two faculties must work together to provide cognition of objects (KrV A50–52/B74–76). At
the same time, each of these two sources places determinate conditions that limit our cognition
of objects. Kant calls these “transcendental” conditions, because they go beyond any empirical
data that could be given to our faculties, and constitute sources of a priori cognition (KrV
A56/80).

How nature is possible
Through sensibility, on Kant’s account, we intuit objects – that is, we come into immediate
cognitive contact with individual objects. The forms of our sensibility are space and time; what
we can cognise is therefore limited to what can be given to us through the senses as spatiotemporal (KrV A19–21/B33–35). Through our understanding we combine intuitions into the
cognition of a single coherent world, about which we can make objectively valid judgments
(A66–83/B91–116). Understanding organises our cognition through concepts, whose most
general forms (the 12 “categories”) are specified by the conditions under which judgments can
form a coherent system. The unity of the world we cognise is conditioned by the active selfawareness of any cognising subject in what Kant calls the transcendental unity of apperception (KrV
A95–118, B131–143). Nature, for Kant, is the unified objective world constituted by this most
fundamental condition of our empirical cognition (KrV A 125–128). Although nature is
cognised as real independently of us, it is our faculties that make possible for us the concept of
nature in general. They are the transcendental conditions of nature as we can know it, and also
conditions of the limits of our possible cognition of nature. These limits for Kant can and
should be experienced by us as tightly constraining in relation to the excessive aspirations of
both natural science and of metaphysics, especially when they aspire to the cognition of more
than our limited faculties permit.

Theoretical reason: the limits of our cognition of nature
In addition to sensibility and understanding, however, there is for Kant a third and still higher
faculty: reason. It provides no separate source of cognition, but it actively directs the inquiry into
37

Allen Wood

nature according to principles, seeking unconditioned completeness in all the ways knowledge
may be given to us. In the title Critique of Pure Reason, the word “critique” means “judge” (or
“critic”, from the Greek word κριτής) (KrV Axi–xii). The genitive in the title is simultaneously
subjective and objective: Reason is to judge (and criticise) its own powers, establishing what it can
know and also the limits of what it can know. Reason seeks completeness in its cognition of
nature, but it is an essential part of Kant’s conception of reason’s powers that this completeness
is never our actual possession: our knowledge of reality, and even of nature, is limited. It is our
peculiar fate, as Kant says in the very first sentence of the Critique, that our faculties assign us to
ask questions that we can never answer (KrV Avii).
The short way to put this result is that natural science cannot explain everything, and the quest for a
supernatural metaphysics represents to us only a kind of knowledge that we cannot avoid seeking but can
never acquire. Our practical predicament as free agents drives us to hope for, and even have faith
regarding, what lies beyond nature. However, our knowledge is confined entirely to nature, and
even our knowledge of the natural world, and especially of ourselves as part of it, is painfully
limited. The subject of cognition itself cannot, as such, be given as a natural object, or its true
nature cognised by us (KrV A341–405/B399–432). Even the relation of mind or consciousness
to our natural embodiment can never be an object of cognition (A381–396). Our knowledge of
nature as a whole is also limited and even paradoxical to us (KrV A405–567/B432–595). The
natural world itself, as our reason must inquire into it, is constituted by a set of regressive series
of conditions: future time is conditioned by past time, any portion of the spatial world is
conditioned by a larger enclosing portion; every causal power of a natural substance is conditioned by the powers of other substances, and the existence of any natural substance is
conditioned by the existence of others. These series of conditions can never be completed
within our experience, though reason can project that completion in the form of ideas –
concepts generated by reason itself, to which no corresponding object can ever be given to our
senses. These ideas are that of the thinking subject, of the beginning of the world in time and its
limit in space, of a cause that needs no prior cause but acts entirely from itself, and of a necessary
being in nature grounding the existence of all others.
Further, reason infers the idea of a being in which all positive properties or realities are
present: an ens realissimum – the abstract metaphysical conception of what religions call “God”
(KrV A567–583/B595–611). However, this metaphysical object of thought can never be an
object of cognition for us; it is only an idea – or ideal: that is, the a priori concept of an
individual object which, however, can never be cognised by us. Theoretical reason must
forever remain agnostic about even its existence (KrV A583–632/B611–670).
Further, to ascribe personal characteristics to this ideal – to think of it as a Deity for the
purposes of religion – requires thinking that Kant calls symbolic: we must draw from experience
of ourselves the conceptions of understanding and will, remove from them the limits we recognise in ourselves as finite beings and think about these attributes in ways that have moral and
aesthetic significance for human life (KU 5, 383). Such thinking is continuous with our agency
as moral beings who are concerned with an ultimate destiny in the world about which they may
have rational hopes, but which they know they can never comprehend as knowers.
For theoretical reason, all ideas serve only as ways of directing our cognitive striving and
organising its results, but they also constitute limits to our cognition (KrV A643–704/
B670–732). We can never cognise their objects, or even know whether they have any. Those
objects would be transcendent or supernatural, and in them we also recognise the focus of
familiar supernatural metaphysical aspirations and religious beliefs: in a simple, immortal soul, a
free will, a Deity. Since we must, from a theoretical standpoint, be agnostic about all these
objects, Kant denies the possibility of any theoretical cognition of the supernatural.
38

Kant on nature and humanity

Cosmos, mind, body, freedom
Our cognitive capacities limit us to natural objects. However, not even everything that happens
within nature is cognisable by us, since some questions we may ask about these objects also
transcend our cognitive powers. We cannot know whether this cosmos exists of necessity or
contingently – whether it is the only actual one of many possible worlds. We cannot grasp the
natural cosmos as a whole – its limit in space, its beginning (or lack of one) in time and the series
of causal necessity through which it is structured. The unknowable within nature includes
ourselves, both as knowers and agents. We cannot cognise our free agency, we cannot know
whether our consciousness is a function of our material body; if it is not, we can never know
the relation between this immaterial soul and our material body.

3 Practical reason
For Kant we are not only cognising beings but also, and even more fundamentally, active
beings. An action is anything we freely choose as a means to an end we have freely set for
ourselves (G 4, 417). Our actions are subject to three basic principles of reason: (1) instrumental
or technical reason, which requires us to take the necessary means to any end we set; (2)
prudential or pragmatic reason, which tells us to combine the objects of our empirical desires
into a greatest attainable whole, under the idea of our own happiness; and (3) moral reason,
which subjects all action to an unconditional or categorical imperative or moral law (G 4,
414–421). It is significant that moral reason is not the only kind of practical reason, since in all
its applications practical reason in a sense takes us “beyond nature” – that is, it depends on our
free agency and takes us beyond what is given to us passively through the senses.
All three principles of practical reason for Kant go “beyond nature” in the sense that they
rationally constrain our pursuit of satisfactions arising from sensible desires or inclinations.
Instrumental reason subjects action to the constraints imposed by our setting of ends based on
these desires; prudential reason counsels us to set a comprehensive end and prefer it to all partial
or limited satisfactions when the two conflict. However, it is in regard to moral reason that
Kant thinks we are rationally justified in assenting to some propositions that go beyond what
our theoretical cognitions affords us, and even some matters that could be considered as going
beyond nature. Kant presents the moral imperative as a system of different formulas, each with
its own function within the system: the form of the imperative is universal validity, which serves
as a standard of judgment requiring us to act in a way we can will to be a universal law for all
rational beings. The matter of the imperative is humanity or rational nature as an end in itself,
which we must respect and whose interests we must care about in all our actions. Furthermore,
the combination of the form and matter of the moral principle is autonomy, the idea of the will
of every rational being, thought of as giving universal law to itself and all others in an ideal
community of all rational beings as a realm of ends (G 4, 421–436).

Freedom of the will
Kant also holds that practical reason commits us to certain kinds of assent regarding matters that
transcend the capacities of theoretical reason. The most basic of these concerns freedom of the
will – the idea of a cause that acts from itself rather than being externally determined by natural
causes outside it. It is a disputed question in Kant interpretation how far Kant thinks we are
justified in assent to supernatural metaphysical claims regarding our freedom. Some even think
Kant holds that we are to think of our freedom as a power possessed by a supernatural
39

Allen Wood

(“noumenal” or “intelligible”) agent outside space and time. My own view is that Kant, on pain
of incoherence or self-contradiction, could never be committed to any such thing. As I see it,
Kant uses the transcendent idea of such an agent only for the purpose of arguing that we cannot
show there is any logical impossibility in our being free while at the same time subject in our
actions to the natural causal mechanism of nature (KrV A532–558/B560–586). We cannot
accept, comprehend or even conceptualise any positive theory whatever regarding free agency
(G 4, 461–463). These are supernatural matters forever uncognisable by us. The most we can
show is that there is no contradiction or logical impossibility in our actions being both free and
subject to natural causal necessity (KrV A556–558/B584–586). Kant therefore holds that from a
theoretical standpoint we must remain completely agnostic regarding even the real possibility of
free will. However, in every action, and even every act of theoretical judgment, we unavoidably presuppose our own freedom (G 4, 448), so that from a practical standpoint we have
rational grounds to settle the theoretically undecidable question of freedom in favor of assent to
the claim that we are free. We can therefore provide a “deduction” – that is, a warrant for
applying the concept of freedom to ourselves, but only for practical purposes (KpV 5, 42–57).
We can settle issues about freedom – which agents are free and for which deeds – on empirical
grounds; but this does not settle the metaphysical issue of how freedom and nature are ultimately related, or whether the freedom we necessarily presuppose is an illusion (KrV
A802–803/B830–831). Our freedom is undeniable in practice, and assent to it is entirely rational, but we can never even distinctly conceptualise, much less gain any insight into, how
freedom is possible (G 4, 459). The question of free will is one on which we must admit that
our cognition of nature, and of ourselves as parts of nature, is limited, and about which
questions that we cannot help asking are nevertheless forever unanswerable by us.

Practical assent or belief
Kant has an interesting, distinctive and usually underappreciated theory about how purely
practical considerations might justify assent in matters where theoretical proof or evidence is
lacking. It is a basic principle of instrumental reason that we can rationally choose to set as an
end only what we conceive to be at least possible of attainment and think of our own actions as
possibly contributing to its actuality. Unless we can know that some object is impossible, or that
we can do nothing toward making it actual, it is still rational to set it as an end even if we are
uncertain regarding its possibility or the effectiveness of our action toward making it actual. But
in such cases, if we do set it as an end, we are rationally committed to assent to its possibility and
to the conditions of its possibility. This means that our purposive action can justify assent for
practical purposes even to matters about which theoretical proof or evidence, considered by
itself, would require complete suspense of judgment. Assent of this kind Kant calls “belief” or
“faith” (Glaube). Since this is a rather technical sense of the word “belief” (or the German world
“Glaube”), I will refer to it as “Belief” (capitalised) (KrV A820–831/B848–859; cf. KpV 5,
142–148; KU 5, 467–474).
Kant illustrates Belief by what he calls “pragmatic Belief”. Suppose a physician is called upon
to treat a patient with some respiratory symptoms. She cannot determine on the grounds of
theoretical evidence alone what the correct diagnosis is. But she cannot exclude the possibility
that she can do something to help the patient. For this, however, she needs to assent, for
practical purposes to the best available diagnosis she can come up with (in other words to
Believe this diagnosis). Here practical considerations, the choice to set the end of helping the
patient, justifies an assent (a Belief) that cannot be justified on purely theoretical grounds (KrV
A823–825/B851–853).
40

Kant on nature and humanity

God and immortality as postulates of practical reason
Kant applies this reasoning in drawing a connection between moral reason and certain kinds of
supernatural or religious Beliefs. He argues that our vocation as moral beings demands that we
set as an end our conception of a world that morality would create if it could, and to see our
fulfilment of our duties as contributions toward the achievement of this highest moral end,
which Kant calls the “highest good” (or summum bonum). We cannot know on purely theoretical grounds whether the highest good is possible, or whether fulfilling our moral duties will
contribute to its actuality. But we cannot exclude these as impossible either. Therefore it is
entirely rational, and even belongs to our moral vocation, to set the highest good as an end.
Kant then argues that we can conceive the possibility of the highest good only if we think of
nature as subject to a supremely good wise and powerful being – in other words, a God. Setting
the highest good as an end therefore provides a distinctively practical ground for assent to God’s
existence, or Belief in God (KrV A828–831/B856–859; KpV 5, 110–161; KU 5, 447–474).
Kant applies the same reasoning to Belief in immortality of the soul by thinking of the highest
good as containing either the happiness of which we have made ourselves worthy, to be enjoyed in a future life (KrV A797–819/B825–847), or else as involving our endless moral
progress, also in an eternal future (KpV 5, 122–124). In my view, the application of moral Belief
to immortality tends to recede in importance for Kant as his thinking about these topics develops. Although Belief in immortality does not disappear, it seems to receive less emphasis, and
its practical justification is increasingly marginalised.
For those to whom Kant’s conception of the highest good seems alien or unfamiliar, it
should be possible to think of many of our larger ends for humanity – world peace, the free or
the just society, even the survival of our species in the face of human-made threats such as
global climate disruption or nuclear war – in the same way that Kant thinks of the highest good,
and therefore to realise that setting these as ends might involve the same practical commitment
to Belief in the conditions of their possibility. It might not be so unreasonable for people setting
ends such as world peace, a free or democratic society, etc., also to think that these ends can be
reached only with God’s help, and in this way some might see his moral argument for God’s
existence as applicable even without sharing his precise concept of the highest good.
In the Critique of Practical Reason, Kant calls God and immortality “postulates of practical
reason”, using the term “postulate” in a sense drawn from his logic. When we set ourselves a
practical task, we can articulate the condition of its possible success in two ways. We call one of
these condition a “problem” when we see it as requiring (and as capable of) demonstration. But
when we see it as immediately known through the very concept of the end, then we call it a
“postulate.” Assent to (Belief in) such a postulate can therefore be justified if we are justified in
setting the end. And Kant thinks of the existence of God and immortality of the soul as postulates in relation to our task of pursuing highest good through fulfilling our moral duties.

Symbolism, analogy and religion
For Kant, not only our cognition but even our thinking can have significance for us only if it
combines concepts with something sensible as their content. This is true even of ideas whose
objects are theoretically uncognisable, such as the idea of an ens realissimum or Deity. Concepts
can be applied to cognisable objects of nature only through what Kant calls schemata, representations of our imagination that can enable us to recognise in experience examples of
unities, pluralities totalities, realities, negations, substances, causes and so on. Even the idea of
God can have meaning for morality or religion only when it is given a kind of sensory content.
41

Allen Wood

This cannot be a schema, however, but must assume the character of a symbol. Symbols are
drawn from our sensible experience, but applied to supersensible objects of thought through
analogy. Analogy for Kant is not an imperfect similarity between objects but rather a sameness in
our way of thinking about objects that may be wholly dissimilar. When religions represent God
as commanding and judging human beings, or as being forgiving, loving and merciful to them,
they think about an utterly uncognisable object of reason in ways they think about human
beings (P 4, 357; cf. R 6, 64–65n, 136; VpR 28, 1023; KU 5, 351–354; Anth 7, 191–192).
Symbolic thought for Kant is closely related to our aesthetic experiences of the beautiful and
the sublime, and our aesthetic ideas – sensible representations so rich in content that no
concepts can ever adequately capture them. For Kant, religious thinking can have human
meaning and value when it takes the form of symbols, and when religious people accept the
responsibility for interpreting the meaning of these symbols in terms of what matters to them in
human life. The greatest danger posed by religious thinking for Kant is found in anthropomorphism: representing God not symbolically but literally as like a human being, a cosmic
tyrant whose arbitrary will we obey out of hope for his favors and fear of his displeasure (R 6,
141–142; VpR 28, 1001–1002). This corruption of religion, which Kant regards as deplorably
common in ecclesiastical institutions and popular cults, tends toward the representation of
God’s acts, such as those presented in religious scriptures, in literal terms rather than as symbols
through which we can make meaningful our moral vocation and our relations to other human
beings. Religious abuse takes the form of delusion (Wahn) – the confusion of a symbolic representation with what it represents. Anthropomorphism leads to a series of evils and abuses in
religion: idolatry – the confusion of the symbol with what it symbolises; counterfeit service of God –
substituting traditional rituals for good life-conduct; priestcraft – or the tyranny of an aristocracy of
religious authorities over the minds and consciences of human beings; fetishism – believing that by
ritual acts we can conjure up supernatural effects to our own advantage; and superstition – which
for Kant means not merely false and irrational beliefs, but the employment of our understanding in
violation of its own essential laws (R 6, 167–185). All these delusions – corruptions found in
popular and institutionalised religious faith – are theoretical errors arising more fundamentally out
of common human moral vices.

4 The empirical study of human nature
The courses Kant offered most often as a professor in Königsberg were a pair devoted to the
empirical study of human beings in their environment. Starting in the early 1770s, every
summer he offered a course on Physical Geography, concerning the varied natural or physical
environments around the globe in which they live. Physical geography included the ways
people produced the material conditions of their life, and it therefore included the study of their
material culture, often also their customs and even their religion, so that Kant’s “physical
geography” involved a rudimentary version of what we would now call both physical and
cultural “anthropology”. In the winter, Kant offered the course he called “Anthropology”,
which dealt with the social environment of human beings, their rational faculties and what they
are capable of making of themselves as free and rational agents. Kant’s impetus for the latter
course was Kant’s dissatisfaction with the “physiological” approach taken by Ernst Platner in his
treatise Anthropology for Physicians and Philosophers (1772). Like Platner, Kant was interested in
the role played by human physiology – what we can learn about the human body as a living
organism – in understanding ourselves. However, in place of Platner’s attempt to understand
human nature exclusively through the physiological approach, Kant devised an empirical science of human nature he called “pragmatic” anthropology, which deals with “what the human
42

Kant on nature and humanity

being as a free agent makes, or can and ought to make, of himself” (Anth 7, 119). One might
think from this description that Kantian “anthropology” would have primarily, perhaps even
exclusively, moral preoccupations. But that would be a serious mistake. Although we do learn
some important things from Kant’s anthropological writings about his conception of the application of moral principles, the term “pragmatic” as applied to Kant’s anthropology actually
carries the connotations of “useful” and “prudential”. Kant is interested more in how human
beings can use their faculties to achieve their extremely varied ends in life, and especially how
they can achieve these ends through their social interaction with others.
Kant’s lectures on anthropology were intended for a popular audience, and drew more
auditors than any other lecture course he gave. His sources for “pragmatic anthropology” were
extremely varied, and exhibit the wide range of Kant’s own sensibilities as a polymath and
representative of enlightenment culture. Kant certainly does cite natural scientific and physiological sources, such as Descartes, Buffon, Brown, Linnaeus, Moscati, Camper, Blumenbach,
Soemmering, Clavius, Helmont and Haller, as well as economists (especially Adam Ferguson
and Adam Smith) and psychologists such as Pietro Verri. But he even more often refers to
literary and historical sources: to Haller’s poetry as well as his physiology, to historians, classical
as well as modern historical writings of Hume, Voltaire and Robertson, to travel narratives (of
which Kant was an avid consumer), to novelists, poets and dramatists such as Sterne, Swift,
Butler, Fielding, Richardson, Shakespeare and Molière, to the memoirs and published correspondence of men and women of letters, and essayist, such as Kant’s favorite Montaigne. To put
it somewhat anachronistically in the terminology of more recent culture, Kant’s study of human
nature drew at least as much on the humanities as it did on the sciences.
Kant’s anthropology involves a complex individual and social psychology, focusing on the
traits of discontent, competitiveness and a propensity to pretense and self-deception. This
positive theory itself helps to underwrite some of Kant’s doubts about the possibility of human
self-knowledge. For it forces us not only to the conclusion that the laws governing human
behavior extremely are variable, but also compels us to admit that their discovery is blocked by
obstacles thrown up by human nature itself. Kant denies that we can know even in our own
case the principles on which we act. Kantian anthropology says that human beings have a strong
tendency to conceal and disguise the truth about themselves “The human being has from nature
a propensity to dissemble” (Ak 25, 1197). If someone notices we are observing him, then he
will either become embarrassed, and hence unable to show himself as he really is, or else he will
deliberately dissemble, and refuse to show himself as he is (Ak 7, 121, Ak 25, 857–859).
To see human nature as it truly is, we would have to observe behavior which is unselfconscious. But human nature in its full development occurs only in civilisation, and it is one
of the effects of civilisation to make people more vulnerable to the opinions of others, hence
more sensitive to the way others perceive them. “In crude people, their entire humanity is not
yet developed”, but if we observe more cultivated people, “then [we] run into the difficulty
that the more educated (gebildet) the human being is, the more he dissembles and the less he
wants to be found out (erforscht) by others” (Ak 25, 857).
Kant distinguishes between merely “noticing” oneself (which we do haphazardly all the
time) and “observing” oneself (in a methodical way). The latter (he claims) would be necessary
for a scientific anthropology, but it is inherently untrustworthy. When a person is being observed by others, “he wants to represent himself and makes his own person into an artificial
illusion” (Anth 7, 132). It is just the same when we study ourselves: “Without noticing what we
are doing, we suppose we are discovering within us what we ourselves have put there” (Anth 7,
133). Kant is thus very much in agreement with Nietzsche’s critique of “naive empiricism”: the
“inner” world of our sensations and feelings is even less trustworthy and more “phenomenal”
43

Allen Wood

than the world of external objects. Hence those who have sought to make a meticulous record
of their inner lives usually record only lies and self-deceptions; in fact, extreme zeal in selfhonesty, especially as found in religion, leads sooner to enthusiasm and madness than to truth.
For those who undertake “this hard descent into the Hell of self-knowledge” (Ak 25, 7),
coming to know the deeper truth about oneself usually produces only anguish and despair,
which unfits them equally for knowledge and for action (Anth 7, 132–133). “Nothing is more
harmful to a human being than being a precise observer of himself” (Ak 25, 252); “All selfscrutinizers fall into the gloomiest hypochondria” (Ak 25, 863; cf. 25, 477–478, 865).
Kant’s view that we are psychologically opaque connects his theory of human nature with a
set of ideas usually associated with later thinkers, such as Nietzsche and Freud. Kant holds that
most of our mental life consist of “obscure representations”, that is, representations which are
unaccompanied by consciousness; if we ever learn about them at all, we must do so through
inference (Anth 7, 135–137). This is partly because many representations are purely physiological in origin, and never need to reach consciousness. But in some cases, Kant thinks, we have
a tendency to make our representations obscure by pushing them into unconsciousness. “We
play with obscure representations and have an interest, when loved or unloved objects are
before our imagination, in putting them into the shadows” (Anth 7, 136). The paradigm example of this, he thinks, is the way people deal with their sexual thoughts and desires.
Kant is interested in the similarities and differences between human beings and other animal
species, and places the empirical study of human nature within the framework of biological
theory. He views human freedom itself, as well as the human propensity to social antagonism or
competition, within the context of nature’s provision for the development of the species capacities of natural kinds of organisms (I 8, 19–22). But he agrees with Rousseau that the chief
characteristic of human nature is perfectibility: its species capacities are endlessly self-developing.
For Kant, the traditional definition of the human species as “rational animal” means not so
much that human beings behave rationally (for they do this only very imperfectly) as that their
capacity for reason makes human nature extremely variable and unpredictable (Anth 7,
321–322). Rationality must be viewed as a problem set for human beings by their nature, for
whose solution not nature but human beings are responsible. The traditional definition is also
defective in that it belongs to rational capacities to open our nature to modification by being the
source of perfectibility. Reason, regarded as an empirical sign of our freedom, is precisely our
capacity for an indeterminate mode of life, one that is open-ended and self-devised, in contrast to
the life of other animals, which is fixed for them by instinct (MA 8, 111–115). So understood,
the traditional definition is only a confession that human nature is in principle indefinable in the
way the natures of other living beings are.
For all other living species, Kant thinks, their mode of life is determined for them by natural
instinct, and their innate faculties are suited to that mode of life. But human beings must invent
their own relationship to nature, and Kant is struck by the wide variety of such relationships
found in different climates and situations on the earth’s surface. In domestic society human
beings must pass on their perfected capacities from one generation to another through education.
(Kant realises, here at least, that the difference between human beings and other creatures is
really only a matter of degree: from Linnaeus he draws the observation that young birds must be
taught by their parents the songs characteristic of their species, and that the songs of finches and
nightingales differ from one country to another, thus showing that animals are capable of “a
tradition, as it were” [Anth 7, 323 and note].) Human beings are also capable of determining for
themselves the form of their social interactions with one another, by adopting shared principles
for the government of social wholes.

Kant on nature and humanity

The uniqueness of humanity among animal species has the consequence for Kant that it
makes the empirical study of human beings different from that of other animals, and the rest of
nature. Kant thinks that empirical psychology (whose treatment by A.G. Baumgarten served as
the textbook used for his anthropology lectures) is therefore inherently problematic. It can
never be transformed into a true natural science using the mathematical methods that have been
so successful in physics (MAN 4, 471).
The problematic character of human nature, human freedom and human self-making, indicate for Kant that our relation to ourselves should be more the preoccupation of our practical
than our theoretical reason. We are never merely passive observers of human nature, but always
active participants in making it. Our relation to what it has been is the business of historians; our
relation to what we are now is not that of external observers but that of free, active beings who
must take responsibility for what we are in relation to rational principles: instrumental, prudential, cultural, political and moral.
We human beings are among the objects of our natural science, but more fundamentally,
both as inquirers and as agents, we are subjects, who must assume responsibility for the rational
principles according to which we inquire into nature and for our own lives as individuals and as
legislative members of social wholes. Our task is to understand ourselves as part of nature, but
also to direct ourselves toward ends of our own devising as autonomous beings. In relation to
our practical task, Kant regards both aesthetic and religious thinking as indispensable to our full
humanity. Having begun his intellectual career as a natural scientist, and never having abandoned that vocation, Kant’s interests developed and broadened in ways that made art, morality,
religion, politics and the study of human culture their central focus during the last decade or so
of his life as a researcher and philosopher.
So where does Kant stand regarding contemporary “-isms”: naturalism, scientism, liberal
naturalism and so forth? All these terms are artifacts of late twentieth and early twenty-first
century philosophy. So as I said at the outset, it is anachronistic to try to locate Kant anywhere
among them. During the nineteenth century, however, there was a brand of philosophy, or
antiphilosophy, that was usually called “materialism” and either put forward a metaphysics
based on natural science or rejected all metaphysics (all philosophy) in favor of its own understanding of the results of natural science. The cause of “Neo-Kantianism” in nineteenth and
early twentieth century philosophy was in part to respond to this movement by insisting that
there is still crucial role in human knowledge for philosophy – for a stage of transcendental
reflection that is required to ground natural science and also to criticise it and be aware of its
limits. It was already somewhat anachronistic even then to ask how far Kant himself partook of
this movement, which saw him as its inspiration. But I tend to think this aspect of those who
wanted to go “back to Kant” – both in the late nineteenth century and also today in the early
twenty-first century – was saying something that needed to be said. It was also seeing something
in Kant’s philosophy that is really there and that is still needed. A correction is required to what
now usually calls itself “naturalism” and is often described by its foes as “scientism”. If this
makes Kant not himself a “liberal naturalist” but at least a precursor of what now goes by that
name, then that may be the best way to locate him in relation to these options.

Abbreviations
Ak

Immanuel Kants Schriften. Ausgabe der königlich preussischen Akademie der
Wissenschaften (Berlin: W. de Gruyter, 1902–). Unless otherwise footnoted,
writings of Immanuel Kant will be cited by volume, page number in this edition.

Allen Wood

Anth
Ca

G
I
KpV
KrV
KU
MA
MAN
MS
P
R
VL
VpR

Anthropologie in pragmatischer Hinsicht (1798), Ak 7. Anthropology from a pragmatic point of
view, Ca Anthropology, History and Education
Cambridge Edition of the Writings of Immanuel Kant (New York: Cambridge University
Press, 1992–) Specific works will be cited using the following system of abbreviations (works not abbreviated below will be cited simply as Ak volume, page).
Grundlegung zur Metaphysik der Sitten (1785), Ak 4. Groundwork of the metaphysics of
morals, Ca Practical Philosophy
Idee zu einer allgemeinen Geschichte in weltbürgerlicher Absicht (1784), Ak 8. Idea toward a
universal history with a cosmopolitan aim, Ca Anthropology History and Education
Kritik der praktischen Vernunft (1788), Ak 5. Critique of practical reason, Ca Practical
Philosophy
Kritik der reinen Vernunft (1781, 1787). Cited by A/B pagination. Critique of pure reason,
Ca Critique of Pure Reason
Kritik der Urteilskraft (1790), Ak 5. Critique of the power of judgment, Ca Critique of the
Power of Judgment
Mutmaßlicher Anfang der Menschengeschichte (1786), Ak 8. Conjectural beginning of human
history, Ca Anthropology History and Education
Metaphysische Anfangsgründe der Naturwissenschaft, Ak 4. Metaphysical First Principles of
Natural Science, Ca Theoretical Philosophy after 1781
Metaphysik der Sitten (1797–1798), Ak 6. Metaphysics of morals, Ca Practical Philosophy
Prolegomena zu einer jeden künftigen Metaphysik (1783), Ak 4. Prolegomena to Any Future
Metaphysics, Ca Theoretical Philosophy after 1781
Religion innerhalb der Grenzen der bloßen Vernunft (1793–1794), Ak 6. Religion within the
boundaries of mere reason, Ca Religion and Rational. Theology
Vorlesungen überLogik, Ak 9, 24. Lectures on Logic, Ca Lectures on Logic
Vorlesungen über die philosophische Religionslehre, Ak 28. Lectures on the Philosophical
Doctrine of Religion, Ca Religion and Rational Theology

5
NIETZSCHE’S NATURALISM:
NEITHER LIBERAL NOR
ILLIBERAL
Brian Leiter

It is no longer controversial that Nietzsche is some kind of philosophical naturalist, a view I
argued for throughout the 1990s and then systematically in my 2002 book (Leiter 2002). As one
scholar wrote subsequently: “Most commentators on Nietzsche would agree that he is in a
broad sense a naturalist in his mature philosophy” (Janaway 2007, 34).1 This marked a dramatic
change from an earlier consensus from the 1960s through the 1990s when resolutely antinaturalistic readings (sometimes obvious misreadings) by Martin Heidegger, Jacques Derrida,
Richard Rorty and Alexander Nehamas were particularly influential. But Nietzsche is also not a
“liberal naturalist”: while he rejects both the supernatural (as any naturalist must) and physicalism (in this regard he is not an “illiberal” naturalist), he countenances the reality only of that
which is explicable by the various Wisssenschaften (sciences). And these sciences, on Nietzsche’s
view, undermine the objectivity (or mind- or attitude-independence) of values, the firstpersonal point of view, and much of our common-sense or “folk” picture of the world (Leiter
2019, 17–111). “Liberal naturalists” with their tolerance for objective values and reasons, and
much of the “manifest image”, are from Nietzsche’s standpoint still in thrall to the same impulses that gave us belief in God: they want human beings to be “special”, while Nietzsche says
the philosopher’s task is to repudiate the “dignified verbal pageantry” and “the false old finery,
debris, and gold dust of unconscious human vanity” in order to recognize again “the terrible
basic text of homo natura [der schreckliche Grundtext homo natura]”:
To translate humanity back into nature; to gain control of the many vain and fanciful
interpretations…that have been scribbled and drawn over that terrible eternal basic
text of homo natura so far; to make sure that, from now on, the human being will stand
before the human being, just as he already stands before the rest of nature today,
hardened by the discipline of science. —with courageous Oedipus eyes and sealed up
Oedipus ears, deaf to the lures of the old metaphysical bird catchers who have been
whistling to him for far too long: “You are more! You are higher! You have a different origin!” (BGE 230)
The familiar modern slogan of naturalism – “what really exists and what we know are scientific,
not philosophical, questions” – confronts a translation problem when it comes to German.
Nietzsche agrees with the naturalist slogan, but he, of course, understands Wissenschaft (science)
DOI: 10.4324/9781351209472-7

Brian Leiter

more broadly than, say, Quine, who was in any case a bad naturalist in practice, even though he
officially endorsed the slogan. (Quine retained allegiance to failed scientific programs – physicalist reductionism and behaviorism – long after they were scientifically discredited.)
Wissenschaften are any forms of inquiry into the world with rigorous and well-defined methods
that generally secure the reliability of their findings: history, classics (Nietzsche’s own field),
psychology, biology, physics are all Wissenschaften when rightly practiced. To this extent,
Nietzsche can seem a liberal naturalist, except that he thinks (correctly in my view) that the
successful Wissenschaften actually explain away so much of what liberal naturalists want to
preserve. Liberal naturalism is, from a Nietzschean perspective, a fig leaf for anti-naturalism and
its sobering portrait of the world and the human situation.
Nietzsche’s naturalism, as I have argued (Leiter 2015, 5–9), is centrally methodological (hereafter
M-Naturalism), calling for continuity with the methods of successful sciences (in the nineteenthcentury, this meant especially physiology and biology, although as we will see psychology was of
central importance to Nietzsche). This continuity entails some substantive commitments, such as
the denial of supernatural entities, which play no explanatory role in the successful empirical sciences, as well as skepticism about freedom of the will, which Nietzsche, like many others in the
nineteenth-century, took to be undermined by the sciences. Crucially, though, M-Naturalism
requires the philosopher seeking to understand human beliefs, attitudes and behavior to develop a
speculative psychology of human beings and human nature. This aligns Nietzsche quite closely
with Hume, as many scholars have now noted (cf. Kail 2009), though Hume had only Newtonian
science as a paradigm, while Nietzsche had the benefit of extensive familiarity with developments
in nineteenth-century science, especially biology, on which to draw, both substantively and
speculatively (Emden 2014). As we know from Barry Stroud’s seminal work (1977, 3–4), Hume
modeled his theory of human nature on Newtonian science by trying to identify a few basic,
general principles that would provide a broadly deterministic explanation of human phenomena,
much as Newtonian mechanics did for physical phenomena. Yet the Humean theory is still speculative, because its claims about human nature are not confirmed in anything resembling a scientific
manner, nor do they even win support from any contemporaneous science of Hume’s day.
Nietzsche’s speculative M-Naturalism obviously differs from Hume’s in some respects:
Nietzsche, for example, appears to be a skeptic about nomic determinism based on his professed
(if not entirely cogent) skepticism about laws of nature (cf. BGE 21–22). Yet Nietzsche, like
Hume, has a sustained interest in causal explanations of why human beings act and feel as they do.
The crux of this speculative naturalism derives from ideas popular among German Materialists in
the 1850s and after that human beings are fundamentally bodily organisms, creatures whose
physiology explains most or all of their conscious life and behavior (Leiter 2015, 50–56).
Nietzsche adds to this Materialist doctrine the proto-Freudian idea that the unconscious psychic
life of the person is also of paramount importance in the causal determination of conscious life and
behavior.2 Nietzsche’s commitment to the importance of psychological explanation is central to
his disagreement with many of the German Materialists, a point to which we return.
Nietzsche accepts what I have called (Leiter 1998) a “Doctrine of Types”, according to
which each person has a more-or-less fixed psycho-physical constitution, which defines him as
a particular type of person.3 Call the relevant psycho-physical facts “type-facts”. Type-facts, for
Nietzsche, are either physiological facts about the person, or facts about the person’s unconscious
drives or affects. Nietzsche’s claim is that each person has certain physiological and psychic traits
that constitute the “type” of person he or she is. While this is not Nietzsche’s precise terminology, the underlying ideas are omnipresent in his writings.
A typical Nietzschean form of “debunking argument” (as I shall call it) runs as follows: a
person’s theoretical beliefs are best explained in terms of his moral beliefs; and his moral beliefs
48

Nietzsche’s naturalism

are best explained in terms of natural facts about the type of person he is (i.e. in terms of typefacts). So Nietzsche says, “every great philosophy so far has been…the personal confession of its
author and a kind of involuntary and unconscious memoir”; thus, to really grasp this philosophy, one must ask “at what morality does all this (does he) aim” (BGE 6)? But the “morality”
that a philosopher embraces simply bears “decisive witness to who he is” – that is, who he
essentially is – that is, to the “innermost drives of his nature” (BGE 6). Indeed, this explanation
of a person’s moral beliefs in terms of psycho-physical facts about the person is a recurring
theme in Nietzsche. “[M]oralities are…merely a sign language of the affects” (BGE 187), he
says. “Answers to the questions about the value of existence…may always be considered first of
all as the symptoms of certain bodies” (GS P:2). “Moral judgments”, he says are, “symptoms
and sign languages which betray the process of physiological prosperity or failure” (WP 258).
“[O]ur moral judgments and evaluations…are only images and fantasies based on a physiological process unknown to us” (D 119), so that “it is always necessary to draw forth…the
physiological phenomenon behind the moral predispositions and prejudices” (D 542). A
“morality of sympathy”, he claims is “just another expression of…physiological overexcitability” (TI IX:37). Ressentiment – and the morality that grows out of it – he attributes to an
“actual physiological cause [Ursache]” (GM I:15). Nietzsche sums up the idea well in the preface
to the Genealogy: “our thoughts, values, every ‘yes’, ‘no’, ‘if’ and ‘but’ grow from us with the
same inevitability as fruits borne on the tree – all related and each with an affinity to each, and
evidence of one will, one health, one earth, one sun” (GM P:2).
Like Hume, Nietzsche proffers a speculative psychology, but Nietzsche’s speculations, unlike
some of Hume’s, seem to fare rather well in light of subsequent research in scientific psychology
(Leiter 2019, 162–180). But the crucial point is that his speculative psychology (as well as the
occasional physiological explanations he offers in passing) give us causal explanations for various
human phenomena, which, even if not law-governed, seem to have a deterministic character,
that is, they presuppose that the phenomena in question have causal determinants. Readers should
remember how omnipresent causal claims are in Nietzsche’s philosophical writing. When he says
in Daybreak, for example, that “[O]ur moral judgments and evaluations...are only images and
fantasies based on a physiological process unknown to us” (D 119), so that “it is always necessary
to draw forth...the physiological phenomenon behind the moral predispositions and prejudices” (D
542), he is making a causal claim, that is, the claim that certain physiological processes cause moral
judgments through some presumably complicated process that yields them as “images” and
“fantasies” brought about by these causes. When he says in the Genealogy that ressentiment, and the
morality that grows out of it, has an “actual physiological cause [Ursache]” (GM I:15), his meaning
is, of course, unmistakable. When he devotes an entire chapter of Twilight of the Idols to what he
calls “the four great errors”, errors that almost entirely concern causation – “confusing cause and
effect”, the “error of false causation”, the “error of imaginary causes” he calls them – it is clear that
he wants to distinguish genuine causal relations from the mistaken ones that infect religious and
moral thinking. When he returns to the same theme in The Anti-Christ, he again denounces
Christianity for trafficking in “imaginary causes” and for propounding “an imaginary natural
science”, one that depends on anthropocentric concepts and that lacks, as Nietzsche puts it, “any
concept of natural cause” (A 15; cf. A 25) – science consisting, on his account, of “the healthy
concepts of cause and effect” (A 49). Psychological and physical causation are essential for
Nietzsche’s entire moral psychology and revaluation of values.4
The autonomy of causal psychological explanation is of special importance for Nietzsche. He
concludes the first chapter of Beyond Good and Evil, one of the major works of his mature
period, by declaring that “psychology is again the path to the fundamental problems” (BGE 23).
This comes at the end of a chapter in which Nietzsche lampoons the “prejudices” of
49

Brian Leiter

philosophers, such as their beliefs in the unconditional value of truth over falsehood, in “the
will” and the Cogito, in the “synthetic a priori”, in “absolute certainties” and laws of nature.
Nietzsche offers a psychological diagnosis of all of these philosophical beliefs that follows the
strategy of the “debunking argument” mentioned already: belief in these ideas “bear decided
and decisive witness to who he [the philosopher] is – which means, in what order of rank the
innermost drives of his nature stand with respect to each other” (BGE 6): in other words,
philosophers advance these claims not because they have epistemic support, but because they
satisfy psychological needs the philosophers have. As Nietzsche and everyone who has studied
the history of philosophy knows, no philosophical arguments are successful, and almost any
position can be given a “philosophical” defense: all we can do, on Nietzsche’s view, is diagnose
and explain why philosophers profess to believe what they do (Leiter 2018). But it is crucial for
this debunking approach that psychological explanation (especially in terms of unconscious
drives and desires) be an autonomous form of explanation, and this is part of the reason why
Nietzsche is such a vigorous critic of the physical reductionism he founds among some German
Materialists (GS 373, GM III: 16; cf. Leiter 2015, 18–21). Here again Nietzsche betrays his
prescient naturalistic judgment: as Jerry Fodor famously argued in our own time (Fodor 1975,
9–26), reduction of the “special” sciences to the “basic” sciences has rarely happened, instead
the domain of successful psychological explanation has expanded apace, absent alleged vindicatory reduction by physical or biological sciences.
One upshot of psychological explanation, for Nietzsche, is that there are no objective facts
about what is morally right and wrong, good or bad. Nietzsche is a thorough-going anti-realist
about value, including, importantly, his own value judgements (Leiter 2019, 49–66). His arguments for this conclusion are explanatory in nature: one can always explain judgments of value
by reference to psycho-physical facts about those who make the judgments, without any reference to attitude-independent features of the world (Leiter 2019, 29–38). Nietzsche, in this
respect, is part of the familiar tradition of value anti-realists who are also sentimentalists, like
Hume and, in the German tradition, Johann Gottfried von Herder (with whom Nietzsche was
familiar) – that is, philosophers who think the best explanation of our moral judgments is in
terms of our emotional responses to states of affairs in the world, responses that are, themselves,
explicable in terms of psychological facts about the judger. Of course, if our emotional judgments had cognitive content – if they were, in fact, epistemically sensitive to the putatively moral
features of the world – then sentimentalism would be compatible with moral realism. But that is
not Nietzsche’s view. He understands our basic emotional or affective responses as brute artifacts of psychology in the first instance (feelings of aversion and attraction), although sometimes with a cultural overlay, as in this illuminating example he gives in Daybreak:
The same drive evolves into the painful feeling of cowardice under the impress of the
reproach custom has imposed upon this drive; or into the pleasant feeling of humility if
it happens that a custom such as the Christian has taken it to its heart and called it good.
That is to say, it is attended by either a good or a bad conscience! In it itself it has, like
every drive, neither this moral character nor any moral character at all, nor even a
definite attendant sensation of pleasure or displeasure: it acquires all this, as its second
nature, only when it enters into relations with drives already baptized good or evil or
is noted as a quality of beings the people has already evaluated and determined in a
moral sense… (D 38)
In this example, the same drive is said to have the potential to give rise to two different moral
feelings, that of cowardice (which has an unpleasant valence) or humility (a pleasant valence),
50

Nietzsche’s naturalism

depending on the cultural context. The drive in question must be something like, a disposition to
avoid offending dangerous enemies, which, if experienced by a Homeric Greek would then give rise
to feelings of self-contempt for being a coward and, if experienced by a Christian, would be
experienced as the admirable virtue of humility. Cultures, through mechanisms of parental and
institutional inculcation, teach individuals to have particular affective responses to the very same
drive on this picture. We thus have two layers of affective responses: first, there is the affect of
aversion towards offending dangerous enemies which is produced by the drive itself, but then there is
the distinctively moral affect of feeling ashamed (as the Homeric Greek does) or proud (as the
Christian does) of that affective response. The feeling of being a coward, on this account, represents
the combination of a feeling of aversion towards offending a dangerous enemy, conjoined with a metafeeling of contempt or disgust for having that original feeling. The meta-feeling is, at bottom, a
feeling of aversion away from the underlying affect, though perhaps to individuate it correctly
we will need to add some kind of belief about why that basic feeling of aversion is contemptible:
for example, the belief that Homeric men slay their offending enemies, rather than cower
before them. This is probably what Nietzsche is getting at when he writes: “[B]ehind feelings
there stand judgments [Urtheile] and evaluations [Werthschätzungen] which we inherit in the
form of feelings (inclinations, aversions). The inspiration born of a feeling is the grandchild of a
judgment – and often of a false judgment!” (D 35).
Since for Nietzsche morality is not objective and moral judgments are the causal product of,
at bottom, non-cognitive affective responses to states of affairs – responses which are themselves
explicable in psychological terms – it is probably not surprising that he has a similarly debunking
view of cognate concepts like freedom and moral responsibility. The feeling of “freely willing”
an action is epiphenomenal according to Nietzsche (BGE 19), not evidence of the causal
structure of action; actions arise from unconscious and physiological mechanisms we do not at
all understand introspectively and over which we have no control (Leiter 2019, 115–146). In
Daybreak, he writes:
We laugh at him who steps out of his room at the moment when the sun steps out of
its room, and then says: “I will that the sun shall rise”; and at him who cannot stop a
wheel, and says: “I will that it shall roll”; and at him who is thrown down in wrestling,
and says: “here I lie, but I will lie here!” But, all laughter aside, are we ourselves ever
acting any differently whenever we employ the expression “I will”? (D 124)
As Nietzsche puts the same point elsewhere, the faculty of the will “no longer ‘acts’ or ‘moves’”
(A 14): it is causally inert. Thus, there remains no conceptual space even on Nietzsche’s view
for the compatibilist idea that the right kind of causal determination of the will is compatible
with responsibility for our actions, since the will is epiphenomenal. As Nietzsche’s Zarathustra
says: “thought is one thing, the deed is another, and the image of the deed still another: the
wheel of causality does not roll between them” (Z I, “On the Pale Criminal”). This is a pithy
statement of the point of the D 124 passage, and it rules out desperate compatibilist attempts to
save moral responsibility: I may well identify with my “thoughts” or my will, but if they do not
cause my actions, how could I possibly be responsible for them? As Nietzsche writes at the
conclusion of “The Four Great Errors” chapter of Twilight of the Idols:
Today we no longer have any pity for the concept of ‘free will’: we know only too well
what it really is—the foulest of all theologians’ artifices, aimed at making mankind
“responsible” in their sense…[T]he doctrine of the will has been invented essentially for
the purpose of punishment, that is, because one wanted to impute guilt. (TI VI:7)
51

Brian Leiter

Once we abandon this “error of free will” we should, Nietzsche thinks, abandon the concepts
picking out the reactive attitudes whose intelligibility depends on it, concepts like “guilt”.
Zarathustra describes the required revision to our thinking about freedom and responsibility
that results: “‘Enemy’ you shall say, but not ‘villain’; ‘sick’ you shall say, but not ‘scoundrel’;
‘fool’ you shall say, but not ‘sinner’” (Z I: “On the Pale Criminal”). The concepts we are
supposed to abandon – that of villain, scoundrel and sinner – are all ones that require freedom
and responsibility that would license blame, while the substitute concepts Nietzsche proposes
(enemy, sick and fool) merely describe a person’s condition or character, without supposing
anything about the agent’s responsibility for being in that condition or having that character.
Almost nothing of our “manifest image” of moral responsibility would survive Nietzsche’s
naturalistic attack.
To be sure, Nietzsche sometimes still uses the language of “freedom” but in a highly revisionary and deflationary sense, one that does not underwrite a conception of free will (and
moral responsibility) that would be recognizable in the philosophical tradition (Leiter 2019,
147–161). For Nietzsche, the resolute naturalist, we are neither free nor morally responsible; we
are artifacts of our psycho-physical natures and our circumstances. But how plausible is
Nietzsche’s naturalistic moral psychology?
Philosophical naturalists, of course, always incur an evidential burden that most philosophers
do not: their claims must answer to the facts as they unfold in the course of systematic inquiry in
the various Wissenschaften. Kantians make up their moral psychology from their sanctimonious
armchairs, invoking an interest only in the “concept” or “possibility” of moral motivation, but
naturalists actually care about how human beings really work. Hume arguably does not fare well
by this more demanding evidential standard (though he fares better than Kant, to be sure), since
much of his speculation about human nature involves wishful (or, more charitably, “optimistic”) thinking about human moral propensities. Certainly many people do have sympathetic
dispositions in certain circumstances, but at the same time, as Nietzsche likes to emphasize, the
history of the world is the history of the pleasure that people take in cruelty. Nietzsche is
certainly not prone to wishful thinking, but how does his speculative M-Naturalism look more
than a century later?
One important reason that naturalistic philosophers should take Nietzsche seriously is because he seems to have gotten, at least in broad contours, many points about human moral
psychology right. Consider:
1.

Nietzsche presupposes that heritable type-facts are central determinants of personality and
morally significant behaviors, a claim well-supported by extensive empirical findings in
behavioral genetics since (Knobe and Leiter 2007).
Nietzsche claims that consciousness is a “surface” and that “the greatest part of conscious
thought must still be attributed to [non-conscious] instinctive activity” (BGE 3), theses
overwhelmingly vindicated by recent work by psychologists on the role of the unconscious
(e.g. Wilson 2002) and by philosophers who have produced synthetic meta-analyses of
work on consciousness in psychology and neuroscience (e.g. Rosenthal 2008; cf. Leiter
2019, 135–146).
Nietzsche claims that moral judgments are post-hoc rationalizations of feelings that have an
antecedent source, and thus are not the outcome of rational reflection or discursiveness, a
conclusion in sync with the findings of the ascendant “social intuitionism” in the empirical
moral psychology of Jonathan Haidt (2001) and others (cf. Leiter 2019, 67–83).

Nietzsche’s naturalism

Of course, Nietzsche himself did not practice the methods of contemporary empirical psychology, but we need to distinguish between what counts as confirmation of a theory from what
might lead a genius like Nietzsche to have perceived a possible truth about human moral
psychology. Empirical psychology has evolved methods for testing and confirming hypotheses
that were not in use in the nineteenth-century5 – hence the need for a naturalistically-minded
philosopher like Nietzsche to speculate. Yet Nietzsche certainly did not lack evidence on which to
base his speculative moral psychology, and this is crucial for understanding Nietzsche’s own
naturalism, which in this regard is more liberal than the illiberal kind associated with Quine and
his progeny.
Nietzsche’s evidence appears to have been of three primary kinds: first, his own observations, especially of the behavior of others; second, the psychological observations of other
writers recorded in a wide array of historical, literary and philosophical texts over long periods
of time, observations which, in some respects, tended to reinforce each other (consider, e.g. the
realism about human motivations detailed by Thucydides in antiquity and, in the modern era,
in the aphorisms of La Rochefoucauld, both authors whom Nietzsche admired); and third, his
reading about contemporaneous scientific developments, many of which – even if amateurish
or simply wrong by today’s standards – did represent systematic attempts to bring wissenschaftlich
methods to bear on the study of human beings and which, in some of their broad outlines, have
been vindicated by subsequent developments. By the standards of our contemporary methods
in the human sciences, such as they are, we would not deem insights arrived at based on this
evidence to be well-confirmed, but that certainly does not mean it is not, in the hands of a
genius like Nietzsche, adequate for insights that survive scrutiny by our contemporary approaches. This is precisely one of the reasons why Nietzsche is a great speculative M-Naturalist in
the history of philosophy: with unsystematic data and methods he could nonetheless arrive at
hypotheses that turn out to win support by more systematic data and methods. And the resulting naturalism is neither illiberal in the manner of Anglophone reductionism or eliminativism of the Quine variety, nor “liberal” in the manner of the McDowell variety that
reflects the basic Christian impulse Nietzsche always opposed to preserve the “special” (dare we
say, “God-given”?) status of human beings.6

Notes
1
2

3
4
5

Janaway, in an otherwise illuminating study, makes several errors in his discussion of my version of
Nietzsche’s naturalism; in the end, his view of Nietzsche’s naturalism is just a version of mine. For an
extended discussion, see Leiter (2015, 244–264).
Nietzsche’s “official” view seems to be that physiology is primary, but he mostly concentrates on
psychological claims, most obviously because he is no physiologist! There were, of course, other
anticipations of the Freudian idea, ones that Nietzsche likely encountered, e.g. Fechner (1848).
(Thanks to Dan Telech for calling the Fechner article to my attention.)
Are individuals born with these traits or do they acquire them? Nietzsche’s texts are unclear on this
point. But it is clear that particular traits wax and wane in importance in an individual’s life depending
on circumstances.
On the central role of causation in Nietzsche, contra skeptics, see the discussion in Leiter (2013,
587–592).
While the use of systematic experimentation, rather than mere introspection, seems a rather clear
advance, parts of empirical psychology have been plagued by failures of replication, and the social and
medical sciences more generally have suffered from a specious notion of “statistical significance.” See,
e.g. the devastating critique in Colquhoun (2014).
My thanks to Joshua Fox for characteristically excellent research assistance.

Brian Leiter

References
Nietzsche’s works are cited as follows, unless otherwise noted: roman numerals refer to major parts or
chapters in Nietzsche’s works; Arabic numerals refer to sections, not pages. I use the standard abbreviations for Nietzsche’s works, as follows: The Antichrist (A); Beyond Good and Evil (BGE); Daybreak
(D); Ecce Homo (EH); The Gay Science (GS); Thus Spoke Zarathustra (Z); Twilight of the Idols (TI); The Will
to Power (WP). I have consulted a variety of translations, but have sometimes emended them slightly based
on the Colli & Montinari edition of the Sämtliche Werke (Berlin: de Gruyter, 1980-).
Colquhoun, D. An Investigation of the False Discovery Rate and the Misinterpretation of P-Values. Royal
Society Open Science (November 19), http://rsos.royalsocietypublishing.org/content/1/3/140216. (2014).
Emden, C. Nietzsche’s Naturalism: Philosophy and the Life Sciences in the Nineteenth Century. Cambridge
University Press. (2014).
Fechner, G.T. Über das Lustprinzip des Handelns. Zeitschrift für Philosophie und philosophische Kritik, 19,
(1848). 163–194.
Fodor, J. The Language of Thought. Harvard University Press. (1975).
Haidt, J. The Emotional Dog and its Rational Tail: A Social Intuitionist Approach to Moral Judgment.
Psychological Review, 108, (2001). 814–834.
Janaway, C. Beyond Selflessness. Oxford University Press. (2007).
Kail, P. Nietzsche and Hume: Naturalism and Explanation. Journal of Nietzsche Studies, 37, (2009). 5–22.
Knobe, J. and Leiter, B. The Case for Nietzschean Moral Psychology. In B. Leiter and N. Sinhababu
(eds.), Nietzsche and Morality, pp. 83–109. Oxford University Press. (2007).
Leiter, B. The Paradox of Fatalism and Self-Creation in Nietzsche. In C. Janaway (ed.), Willing and
Nothingness: Schopenhauer as Nietzsche’s Educator, pp. 217–257. Oxford University Press. (1998).
Leiter, B. Nietzsche on Morality. Routledge. (2002).
Leiter, B. Nietzsche’s Naturalism Reconsidered. In K. Gemes and J. Richardson (eds.), The Oxford
Handbook of Nietzsche, pp. 576–598. Oxford University Press. (2013).
Leiter, B. Nietzsche on Morality, 2nd edition. Routledge. (2015).
Leiter, B. The History of Philosophy Reveals that ‘Great’ Philosophy is Disguised Moral Advocacy: A
Nietzschean Case Against the Socratic Canon in Philosophy. In M. van Ackern (ed.), Proceedings of the
British Academy 214: Philosophy and the Historical Perspective, pp. 185–199. Oxford University
Press. (2018).
Leiter, B. Moral Psychology with Nietzsche. Oxford University Press. (2019).
Rosenthal, D. Consciousness and Its Function. Neuropsychologia, 46, (2008). 829–840.
Stroud, B. Hume. Routledge. (1977).
Wilson, T. Strangers to Ourselves: Discovering the Adaptive Unconscious. Harvard University Press. (2002).

6
HUSSERLIAN PHENOMENOLOGY
AND LIBERAL NATURALISM
Andrea Staiti

1 Introduction
In this chapter, I will argue that Husserlian phenomenology is compatible with liberal naturalism. By “compatible” I don’t mean that one position entails the other, that they say the
same things using different words, or that one can be a phenomenologist and a liberal naturalist
at the same time. This is not the place to elaborate on what it takes for two philosophical
currents to be compatible, although it would be a thoroughly interesting issue in its own right.
What I have in mind is roughly that Husserlian phenomenology and liberal naturalism have in
scientific or bald naturalism a “common enemy” (Husserl 1994, 178), as Husserl once put the
point in a letter to Heinrich Rickert, and, as the old saying has it, the enemy of my enemy is my
friend. My view is that liberal naturalism and Husserlian phenomenology share an antiscientistic philosophical agenda, but they pursue different strategies to concretise it. For reasons
to be explained in what follows, I also believe that the Husserlian strategy is philosophically
superior to the liberal naturalist strategy.
I will proceed in four steps. In the first section I will address Husserl’s critique of naturalism
and argue that it resonates with criticism of scientific naturalism by self-styled liberal naturalists.
In the second section I will distinguish the agenda and the strategy of liberal naturalism. In so
doing I will consider the idea of re-enchantment of nature as a commonly pursued philosophical strategy among liberal naturalists, and point out what strikes me as problematic in it.
I will argue that re-enchantment of nature leads straight to the restoration of natural teleology,
which I consider unpalatable. In the third section I will present Husserl’s views on teleology and
the scope of teleological reasoning: I will show that he rejects a readmission of teleological
reasoning in the investigation of nature. In the fourth section I will turn to Husserl’s views on
nature and insist that his strategy involves an austere concept of nature coupled with an argument that nature is an abstraction; by contrast, reality in the most concrete sense is our human
and culturally constituted life-world, a view that liberal naturalists should find attractive. I will
conclude with a brief recapitulation and a statement of what strikes me as promising about
Husserl’s strategy, provided that one is not a priori allergic to any form of transcendental
idealism.

DOI: 10.4324/9781351209472-8

Andrea Staiti

2 Husserl’s critique of naturalism
The argument that Husserlian phenomenology is compatible with liberal naturalism might
sound perplexing. Even if one pays attention to the specific formulation of the claim, namely
that Husserlian phenomenology is compatible with and not itself a form of liberal naturalism, such
that the two philosophical approaches remain distinct, and yet potentially cooperative, initial
skepticism about the suggestion of compatibility seems justified. Husserl is considered, for good
reasons, one of the staunchest opponents of naturalism in twentieth century philosophy
(Pearson/Protevi 2016, 35–36; Moran 2008). He even goes as far as to label naturalism “the sin
against the Holy Ghost of philosophy” (Husserl 2008a, 173). How could Husserl’s work be
compatible with any kind of naturalism, liberal or otherwise?
I presume that for a good number of phenomenologists the matter might be easily settled.
Naturalism and phenomenology are simply incompatible and adversarial, period. By contrast,
phenomenologists who are not comfortable with this conclusion are faced with the difficult task
of identifying those sources, in their philosophical tradition, that can help articulate a different
position about naturalism and phenomenology. Thereby, it is understandable that they often do
not consider Husserl a particularly promising intellectual partner. For instance, in a recent
contribution Jack Reynolds defends what he calls a “potential compatibility of phenomenology
and naturalism” (Reynolds 2019, 20), once we are prepared to admit that “a certain dream of
transcendental phenomenology is over” (Reynolds 2019, 26). Reynolds favors MerleauPonty’s more empirically-based approach to phenomenology and rejects Husserl’s transcendentalism as incompatible with the typically naturalistic refusal of a First Philosophy grounding
the sciences. In a similar vein, Steven Crowell has recently argued for the compatibility of
transcendental phenomenology and what he calls “soft” naturalism (Crowell 2013, 148) provided
that the model of transcendental subject is no longer Husserl’s “consciousness”, but the embodied, existentially involved self that Husserl partly delineates in his analyses of the human
person and that Heidegger’s Dasein articulates even more coherently.
However, if we look a bit closer at Husserl’s charges against naturalism and the corresponding tenets of his transcendentalism, it turns out that his complaints are fundamentally in
line with the ones typically voiced by self-styled liberal naturalists against the so-called scientific
(or “bald”, or, perhaps better, “scientisitic”) naturalists. The locus classicus for Husserl’s antinaturalism, Philosophy as a Rigorous Science (1981 [1911]), criticises both the naturalisation of
ideas and the naturalisation of consciousness. In sum (see Staiti 2012 for further detail), Husserl
complains that naturalism reduces all reality to either the physical or the psychical and therefore
construes ideas as necessarily psychical, since they are patently not physical; however, the
psychological interpretation of ideas completely misconstrues their distinctive trait: validity or
invalidity, which have nothing to do with factual occurrences in the psyche of this or that
individual. The simple fact that an idea is instantiated as a thought in someone’s mind says
nothing about its validity or invalidity. In order to begin to understand validity and invalidity
we need to treat ideas as robust objects in their own right standing in objective, mindindependent relations. Individual thinkers can grasp the validity of such objective relations and
when that happens, ideas exert a normative force on thinking; however, talk of objectively valid
ideas is not merely shorthand for normative relations obtaining at the level of factual thinking.
As for consciousness, Husserl complains that naturalists treat consciousness as a mere appendage of the physical having no distinctive structure of its own and governed by the same
kind of laws that hold in the physical universe. So, for instance, late nineteenth century experimental psychologists adopted a chemical model to investigate consciousness: they set out to
determine what consciousness is in itself, that is, as distinct from its introspective appearance,
56

Phenomenology and liberal naturalism

and they imagined this an sich reality of consciousness as a conglomeration of psychical atoms
(sensations) that are held together by purely mechanical laws of association. Husserl, by contrast,
calls for an approach to consciousness that doesn’t foist onto it models borrowed from the
investigation of physical nature. The distinctive trait of consciousness is intentionality, which is
not a natural relation, but the meaningful reference to an object. It is the instantiation of ideal of
meanings that governs the intentionality of conscious acts, and no projected naturalisation of
consciousness can reduce meaning to physical reality as described by the natural sciences.1 Thus,
consciousness properly (i.e. non-naturalistically) understood has a distinctive and essential share
in the ideal dimension of meaning. These two interconnected lines of criticisms of naturalism
taken together delineate Husserl’s transcendentalism, that is, the view that the existence and
validity of, respectively, things and ideas refer back to non-naturalised consciousness as the
venue where intentional references as meaning-governed phenomena are constituted.
If we now consider, as an authoritative source, the entry on liberal naturalism in The
Cambridge Dictionary of Philosophy, we find the statement that this version of naturalism denies
“that the third-personal scientific stance adequately accounts for the self – particularly selfconsciousness and rational agency” (De Caro and Macarthur, 2013). This seems to be precisely
the same complaint that motivates Husserl’s critique of naturalism summarised above. The self,
or consciousness, is not a swarm of psychical particles appended to a swarm of physical particles,
but a genuine domain of reality in its own right, whose governing principle is intentionality.
Intentionality, in turn, involves the reality of ideas, both as entities in their own right and as
instantiated in experienced meanings. This picture provides a solid framework to make sense of
the phenomenon of normativity, which, as De Caro and Macarthur further point out, is “not
explainable in purely causal or nomic terms” (De Caro and Macarthur, 2013). This is because,
Husserl would explain, normativity is entirely grounded in ideality as a dimension of being in its
own right. Husserl and liberal naturalists, then, seem to share the very same concerns. Steven
Crowell is therefore right when he writes: “What Husserl means by ‘naturalism’ is essentially
what John McDowell calls ‘bald naturalism’” (Crowell 2013, 42), that is, scientific, viz., scientistic naturalism. Consequently, Husserl’s critique of naturalism is actually a critique of “bald”
or scientific naturalism, and, in this regard, it is fully resonant with liberal naturalism.

3 Liberal naturalism: its agenda and its strategy
Can we conclude that liberal naturalism and Husserlian phenomenology are two versions of
one and the same intellectual enterprise? In order to answer this question correctly we need to
distinguish between the agenda of liberal naturalism and the philosophical strategy that liberal
naturalists typically adopt. The common criticism of scientific naturalism provides evidence that
liberal naturalism and Husserlian phenomenology share, broadly speaking, the same agenda: that
is, the rehabilitation of forms of cognition that do not share the methods of the natural sciences,
such as the human sciences, and the acceptance of entities or properties, such as values or
meanings and their normative implications, that do not seem to belong in the ontology
of nature. In Husserl, this agenda is part of the much broader and more ambitious project of
renewing philosophy and raise it to the level of a genuine science, but we can leave that broader
project aside for the purpose of the present chapter. Let us ask, instead, how is the hoped-for
rehabilitation of non-natural-scientific forms of knowledge and entities to be carried out
specifically? This is where, I maintain, the two philosophical currents diverge, that is, they
pursue different strategies. The main point of difference is the way Husserlian phenomenology
and contemporary liberal naturalism carve out the concept of nature. This is not merely a verbal
dispute about what deserves the notoriously slippery label “nature”; rather, as I aim to show, it
57

Andrea Staiti

is a genuine conceptual difference in philosophical strategies. To spell out this point, let me
now turn to what I take to be the typical strategy adopted by liberal naturalists against scientific
naturalists. In the ensuing sections I will address the Husserlian strategy and highlight their
difference.
Since there doesn’t seem to be a definite list of theses that all and only liberal naturalists
share, let us take as an exemplary and illustrious representative of the liberal naturalist strategy
John McDowell. I do not intend to provide anything beyond a “standard” picture of
McDowell’s views, and I am aware of the fact that he has reservations about specifying the
theoretical content of “liberal naturalism” and cognate phrases beyond what is required in the
dialectical context of critique of scientific or bald naturalism.2 In very brief compass, McDowell
traces the historical roots of “bald” naturalism to the disenchantment of the world (a famous
phrase by Max Weber) operated by the rise of modern science. He points out that the disenchanted world of natural science is the correlate of a definite intellectual stance, one that has
enormous and undeniable merits (McDowell 1998, 182); however, McDowell argues, it would
be unwarranted to contend that “a putative operation of the intellect can stand up to reflective
scrutiny only if its products can be validated on the basis of the facts of nature, conceived in the
disenchanted way that is encouraged by modern science” (McDowell 1998, 175). By contrast,
we should recognise the reality of another dimension of nature that is not captured by the
natural-scientific stance. Hans Fink spells out this conception effectively as follows:
The manmade, the artificial, the cultural, the historical, the ethical, the normative, the
mental, the logical, the abstract, the mysterious, the extraordinary, are all examples of
ways of being natural rather than examples of ways of being nonnatural. Nature is
never mere nature. That which is more than mere nature is nature, too (Fink
2008, 67).
Besides physical nature there is the dimension that the Aristotelian tradition calls “second nature”: our
habits, our ethical standards and, more generally, everything that falls into what McDowell famously calls
“the space of reasons”. The space of reasons, in turn, is not metaphysically separate from first nature. For
rational beings nature itself “stands” in the space of reasons and is caught up in the play of our justificatory
practices, as is clearly shown by the role of perceptual judgments. In Mind and World McDowell describes
this position as “a call for a re-enchantment of nature” (McDowell 1996, 74) and this is the key point to
understand the difference with Husserlian phenomenology. Re-enchanting nature means carving out a
concept of nature that includes, as part of the same ontological whole, all the non-natural-scientific entities
listed by Fink in the above quote. It amounts an “over-arching or all-inclusive use of the concept of nature”
(Fink 2008, 66) that includes the normative, the logical, the ideal and the conscious within its scope.
Construing nature as an over-arching concept that includes the physical, the mental and the
normative is a perfectly legitimate philosophical move and an effective way to contrast bald
naturalism; however, I worry that if we spell out the implications of this philosophical strategy
we end up with consequences that I take to be unpalatable. This is shown with admirable
intellectual honesty by Thomas Nagel in his much-criticised book Mind and Cosmos (2012). In
sum, if we take seriously the idea that consciousness, cognition and value are real facts in the
natural world and if we believe that the natural world is at bottom governed by natural laws,
there is no way to avoid the thought of restoring some kind of natural teleology, that is, as
Nagel puts it, the idea “that in addition to physical law of the familiar kind there are other laws
of nature that are ‘biased toward the marvelous’” (Nagel 2012, 92). Nagel hastens to add that
what he has in mind would amount to a “teleology without intention” (Nagel 2012, 93), that
is, a kind of intrinsic tendency of nature to develop in such a way as to give rise to complex
58

Phenomenology and liberal naturalism

systems that display life, consciousness and, therefore, value. In other words, if nature is the
overarching concept, including nonphysical entities that do not seem to be strictly necessary in
order for the universe to hang together in the familiar way explored by the natural sciences,
then it seems consistent to conclude that nature must lean toward the creation of such entities for
some other reason. Nagel’s guess is that it must be the intrinsic value of the existence of such
entities: “The idea of teleology implies some kind of value in the result toward which things
tend, even if teleology is separated from intention, and the result is not the goal of an agent who
aims at it” (Nagel 2012, 97). Once we implant values in nature (as the all-inclusive concept of
nature requires), it is only consistent to view them as poles toward which the rest of nature
somehow tends. It seems philosophically unsatisfactory to lump together under one rubric, that
is, “nature”, the physical, the mental and the normative without then trying to develop a story
about their interrelation, and that story seems to be necessarily teleological. Since there is no
intrinsic necessity for purely physical nature to give rise to conscious beings who are responsive
to meaning and value, and therefore themselves valuable, nor is it probabilistically or evolutionarily likely that such beings would come to exist, positing a bias of nature toward such
beings due to an intrinsic value in their existence seems to be a reasonable option. If this is
correct, then there is a straightforward logical path leading from McDowell’s re-enchantment
and Fink’s all-inclusive concept of nature to Nagel’s restoration of natural teleology. Perhaps
this is why McDowell is hesitant about formulating “the positive doctrine” (McDowell 2008,
216, see footnote 1) associated with liberal naturalism and cognate phrases beyond the dialectical context of critique of scientific naturalism. If we start unpacking the implications of the
positive doctrine, we end up with the problematical concept of natural teleology, one that
Nagel admirably chooses not to shy away from, but that will sound unpalatable to many.
Part of the reason why the prospect of restoring natural teleology does and should sound
unpalatable to many is that it would imply the necessity to rethink the scope, the methods, and
the theoretical aims of the natural sciences from the bottom up; however, it seems that (1) we
don’t have the faintest idea of how such reconceived natural science would have to look like
(Nagel himself is adamant on this point) and, more importantly, (2) there doesn’t seem to be
any independent reason to believe that the natural sciences need teleology. There’s a lot that
needs to be fixed about the natural sciences, such as the funding schemes of research agencies,
bad publication practices and the tendency of scientists to feel qualified to philosophise, but
their methodological outlook and their overall attitude toward the world seem to be as
powerful, reliable and promising as they have ever been. It seems unwarranted to think that in
order to prevent natural-scientific thinking from degenerating into bad metaphysics (i.e., in
order to push the liberal naturalist agenda) we should end up with a restoration of natural
teleology, whose elimination was precisely what gave rise to modern natural science and allowed it to thrive. Before we readmit natural teleology and concoct reforms of natural science
we should, at least, look for alternatives. Husserlian phenomenology offers one such alternative,
or so I will argue.

4 Husserl on natural teleology
Where does Husserl stand about the issues we have been discussing so far? In particular, what is
his position about the re-enchantment of nature and the restoration of natural teleology? In a
recent contribution, John Drummond attributes to Husserl a peculiar form of naturalism, which
he presents as fundamentally continuous with McDowell’s. Drummond writes about Husserl’s
putative naturalism:

Andrea Staiti

[It] has certain similarities to McDowell’s notion of second nature. It points to an
enchanted world wherein nature is best understood as the correlate of what is discovered when we adopt what Husserl calls the ‘personalistic attitude,’ an attitude in
which we attend to how the world reveals itself in the ordinary, everyday experiences
of persons (Drummond 2015, 139).
I have two reservations about Drummond’s attribution to Husserl of a naturalism of reenchanted nature. First, the attribution to Husserl of any kind of naturalism seems to be at odds
with his outspoken adversity to this intellectual trend. Recall that the argument I am trying to
develop in the present chapter is that Husserlian phenomenology is compatible with liberal
naturalism, not that it is itself a version of liberal naturalism. Second, and more substantively, for
Husserl the personalistic attitude is precisely the attitude that does not reveal nature, but rather
the full-blown world of culture and psyche (Geist)3 or what in his later texts goes under the
name of life-world. This is why at the beginning of Ideas II, which is devoted to a phenomenology of material nature, Husserl recommends to assume the naturalistic, rather than the
personalistic attitude, that is, the attitude of natural science (Husserl 1989, 3–4). If we take the
personalistic attitude we do not attend to an “enchanted” nature, as Drummond would have it.
Rather, we do not attend to nature at all. I will sketch out the way Husserl determines the
concept of nature vis-à-vis McDowell in the next section. In this section, I want to address the
claim that Husserl’s (putative) naturalism points to an enchanted understanding of nature as
revealed to persons. Recall that the “re-enchantment” proposed by McDowell and Fink
amounts to an overarching concept of nature that includes values, norms, etc. as part of the
same ontological whole and, as I have argued with Nagel, that concept leads straight to the issue
of natural teleology. Let us then see what Husserl has to say about natural teleology. It will
become clear that he strongly rejects such notion as a conflation of the human and the natural
world. Therefore, it seems to me that attributing to Husserl a re-enchantment of nature of sorts,
as Drummond’s quote seems to suggest, does not get his perspective right. It conflates Husserl’s
and liberal naturalism’s common agenda with a commonality of strategies. Like McDowell,
Husserl wants to rehabilitate the world of culture and the forms of cognition that pertain to it
against the threat of scientism. Both philosophers are critical of bad, scientistic metaphysics and
the hegemonic tendency of bald naturalism. But their strategies differ. McDowell proposes, at
least implicitly, an expanded, all-inclusive concept of nature, which, however, if spelled out in
detail, leads straight to Nagel’s teleology, while Husserl defends an austere concept of nature,
whose objectual correlate only comes into view when we bracket out value and meaning, but
he points out that nature is just one dimension or abstract layer in the concrete world of human
culture, which, in turn, is the only reality that there is. This strategy allows him to concretise
the antiscientistic agenda he shares with liberal naturalists without having to buy into the
problematical concept of natural teleology.
In his recently published lectures Introduction to Philosophy (1916/20) (Husserl 2012), Husserl
includes a helpfully detailed section on the contrast between teleological and natural-scientific
worldview. He argues that the “ateleological description of the world” inaugurated by natural
science is “required by reason and justified” (Husserl 2012, 106) and that it constituted enormous progress vis-à-vis the Aristotelian conception of nature as a teleologically structured
whole. In other words, it was necessary, in order to properly investigate nature according to its
essence, to abstract it from the totality of the concrete world of human culture by way of
stripping things of their value-predicates. Husserl even goes as far as to quote approvingly
Auguste Comte, founder of positivism, and argue that

Phenomenology and liberal naturalism

a teleological worldview has its psychological sources in the naïve animism that we are
all still familiar with from our infancy, i.e. the very natural tendency to follow the
crudest analogy and interpret all things as humans or animal creatures, that is, to look
at all of them as living bodies endowed with a psychical interiority and to look at all
processes of nature as brought about by a psyche (Husserl 2012, 190).
For Husserl, the proper place of teleological reasoning is not the investigation of the natural
world, but the investigation of the human world: “Only the world of spirit [Geist] is normable
and normable throughout. Nature is the realm of the un-spiritual and therefore it is not
normable” (Husserl 2004, 145). By “normable” Husserl means suitable for quid juris-type of
questions, which, in turn, presuppose the experience of values as the standards against which we
can measure facts and establish their legitimacy. None of this can happen with regard to nature.
When we carry out “normative” evaluations, we are entirely operating in the human world,
and we are viewing facts in terms of their distance from or proximity to certain value-standards,
thereby engaging in a form of teleological reasoning. This is how Husserl presents the legitimacy, and even necessity of teleological reasoning in history:
A world of the type of our world, a world with humans qua surrounding world for its
humans possesses an a priori necessary teleology in itself. Humans, wakeful subjects and
subjects of an awakened reason, cannot help but recognize the ethical institution of
goals and, accordingly, submit themselves to it. No matter how they err, no matter
how ethically evil they are, no matter how dumbly they shape their world in single
cases, over and above all intelligent and dumb activity the idea of a perfect human
world holds sway, i.e. the idea of a world onto which an ethical humanity has impressed its ideal form and continues to do so in progressively higher levels of value
(Husserl 2012, 180).
In sum, Husserl is saying that when we consider the human world we cannot help but do so by
reference to some highest standard of what a good, perfectly functioning human community
would be like. This standard, in turn, is not something we can randomly construct, because it is
defined by what human beings are, that is, rational beings striving towards fulfillment of their
intentions, both theoretical and practical. Teleology is, to speak Kantian, a way of looking at the
human world, but, to add a Husserlian twist, one that is grounded in and justified by what
the human world actually is. Rational subjects are beholden to norms and norms have their
ultimate source of validity in the norm of all norms, which is the ethical imperative to live up to
one’s human, rational nature. Christine Korsgaard argues in the same line as Husserl in her book
Self-Constitution when she writes: “A teleological conception of the world is essential to our
functioning as agents” (Korsgaard 2009, 37). Her argument is that when we consider ourselves
as agents, we cannot help but consider the world around us as inhabited by objects and situations that are relevant to our goals and purposes, objects whose very unity is understood in
terms of their intrinsic teleology (animal, vegetable, artefact, etc.), rather than swarms of particles and force fields. A world considered in teleological terms is, in turn, a world in which
morality is at stake, that is, the way in which we can be good or bad at constituting ourselves as
agents and handling the things around us.
While for Korsgaard the teleological worldview in the agent perspective is pervasive, Husserl
hastens to clarify, more moderately,

Andrea Staiti

that this way of looking teleologically at the surrounding world does not amount to
what people usually call a ‘teleological worldview’. The ethical perspective on the
world is, in fact, not really so universal as to encompass the whole surrounding world.
It only reaches as far as the surrounding world is seriously included in the practical
sphere of influence of humans. We cannot move suns and stars from their orbits, we
cannot change practically the processes in the Earth’s inner core; thus, infinities of
nature remain impervious to our human teleology (Husserl 2012, 186–187).
This statement should make it clear that, as I already argued, Husserl is not in the business of reenchanting nature, not even for the sake of our self-understanding as agents.4 Nonetheless, and
unlike Max Weber, who coined the phrase “disenchantment of the world”, Husserl doesn’t
believe that the legitimate disenchantment of nature can be extended to our human world,
which is founded in, but not reducible to, its natural underpinnings. In order to make scientific
sense of the human world we need norms, goals and values, and we need them because they are
real aspects of it. The dimension of normativity, or McDowell’s space of reasons, results from
our human interaction with nature and among ourselves; however, to say that these dimensions
of the human world are real does not mean to say that they are natural, that is, cut from the same
ontological cloth as rocks and rivers, let alone protons and force fields. Liberal naturalists should
be satisfied with this conclusion, which creates a more than hospitable intellectual space for
distinctively human activities and forms of cognition, such as ethics, that resist naturalisation. As
John McDowell writes: “We can hold that free responsiveness to reasons is extra-natural, in a
sense of ‘natural’ that suits scientific naturalism, without thereby implying that it is supernatural”
(McDowell 2008, 218). However, instead of limiting the sense of “natural” in which free
responsiveness to reasons is extra-natural to the one that suits scientific naturalism, a Husserlian
phenomenologist would point out that talk of “second nature” is, in any sense, misleading.
Husserl is explicit on this point: “The fundamental trait of moral life is only that of a ‘second
nature’, rather than first nature; however, obviously, talk of nature is not appropriate, since
morality belongs entirely to the realm of spirit [Geistigkeit] and not to the realm […] of psychophysical nature” (Husserl 2004, 165). As we have seen above, consciousness does not have a
structural affinity with the things of nature, and while it is localised in a psychophysical body
that belongs to the natural nexus, calling consciousness “natural” is only attractive if we consider this label a kind of honorific title, whose absence in a given subject-matter amounts to a
loss of intellectual respectability. But isn’t that sort of chauvinism precisely what liberal naturalists want to counter? So, why talk of second nature? The same goes for values and norms.
Rejecting the idea that they are somehow supernatural doesn’t automatically imply that they are
natural. For the sake of clarity, we would do better to admit that there is a significant amount of
things in the world that are extremely real, such as mortgages or regrets, without therefore
being natural. In fact, to a closer look, it is “nature” and “the natural” that have a kind of halo of
irreality, if we look at them from the only perspective that is available to us, that is, the human
perspective. Nature is the correlate of an unnatural, that is, wholly artificial attitude, one that
has to be actively implemented, rather than a readily available domain of reality that may or may
not be expanded to include other domains. The Husserlian strategy is to reverse our perspective. Instead of starting off with a readily available concept of nature and proposing to
expand it, he starts out with a genuinely all-inclusive concept, that is, the concept of the human
world of culture, or life-world, and proceeds to carve out the concept of nature as resulting
from an abstractive impoverishment of such world. “Abstractive impoverishment” has a negative ring to it, but it is actually an imperfect translation of the value-free German word Abbau,
which does not have a pejorative connotation (see Staiti 2010). For Husserl, the kind of
62

Phenomenology and liberal naturalism

abstractive impoverishment we need in order to visualise nature as an enclosed domain of being
and start to investigate it the way its essence requires is a thoroughly positive procedure. In the
next (and last) section I want to sketch out Husserl’s concept of nature.

5 Husserl on nature
There are several venues in Husserl’s oeuvre where he presents his views on nature. The most
widely known is arguably the first chapter of Ideas II (Husserl 1989, 3–29), but other texts,
unfortunately unavailable in English translation, are even more helpful and clear to grasp Husserl’s
thoughts on this matter. I am thinking particularly of the body of research manuscripts titled The
Reality-Structure of the Life-World: Nature as Abstract Core-Layer of the World in volume thirty-nine of
Husserliana (Husserl 2008b, 259–306) and the excursus on Nature and Spirit in the lectures
Introduction to Ethics, in Husserliana thirty-seven (Husserl 2004, 259–320), both written in the early
1920s. It would exceed the scope of the present chapter to follow the details of Husserl’s analyses,
so I will draw liberally on these materials to paint a broad picture of Husserl’s views on nature.
The larger philosophical framework in which Husserl develops his views on nature is the
inquiry about how a world in general is necessarily structured. This, in Husserlian idiom, is an
eidetic analysis: it is about essence, not fact. Instead of focusing on this existing universe with its
factually valid natural laws, he takes this existing universe as a mere example of a world in
general and asks what the latter must necessarily contain. The way to proceed is through a
method of variation: we are invited to carry out a kind of mental experiment where we vary the
factual content of the existing universe in order to figure out what cannot be thought away if
the whole in question must continue to be a world, as opposed, for instance, to a random
collection of entities that do not hang together as worlds do. In order for this kind of inquiry to
yield the desired insights, it is key to make sure that one starts out as presupposition-less as
possible, otherwise one might inadvertently smuggle into the putative necessary content of a
world in general features or laws that pertain only to the factual universe as we know it. For
Husserl, then, the starting point of the variation has to be the concrete world of human experience,
and necessarily so, for the following reason. We can contrast scientific theorising with human
experiencing. Scientific theories change with time, they can be successful or unsuccessful, wellsupported or purely speculative, etc.; however, regardless of their level of success, and regardless
of how far-fetched they might be, scientific theories are about something, and that something is
not itself the product of scientific theorising. Rather, it is presupposed by all theorising: it is the
world of human experience, that is, the colorful, odorous, haptic, etc. surrounding world that
invariably shows up in the experience of a wakeful subject. Furthermore, the world of human
experience is not just given to us through the five senses as an object of contemplation. It is a
world we act upon, and things of the world receive all sorts of determinations and enduring
properties from our creative activities. Things in the world are endowed with properties that
stem from human activity. Such properties are, to a large extent, value-properties: they embody
the values that things have for us based on the feelings that their natural properties awaken in us.
The world of human experience is also populated by other fellow humans. We experience
them as having a psychological experience like ours and as capable of both registering valueproperties things already possess and infusing things with new value-properties through their
creative activity.
Two key ideas emerge, for Husserl, if we think through the implications of this analysis.
First, while we can very well imagine a world without humans, it is impossible to imagine a
world that doesn’t show up in the experience of some subject. This isn’t because it is subjects
who somehow maintain the world in existence, but because a world, as opposed to some other
63

Andrea Staiti

thinkable arrangement of items, is necessarily conceived as lawfully ordered, and that order is
only conceivable concretely as showing up in orderly sequences of appearances in a subject’s
consciousness. Thus, subjects and worlds are not only contingently but rather essentially related.
This view doesn’t exclude the possibility or some other form of co-existence of entities that
does not add up to a world. But the moment we think of a collection of entities that co-exist in
lawful terms, such that every entity necessarily has a determined position vis-à-vis all other
entities and a state that is determined by its interactions with them, we are necessarily led back to
a manifold of orderly subjective experiences in which the lawful ordering of the collection at
issue shows up. This is, in Husserl’s terms, the most general sense of “world”, as opposed to
some other thinkable arrangement where the entities that fall into it are simply juxtaposed or
mutually indifferent. In other words: how do we tell apart a world from any other possible
random collection of entities? For Husserl the answer necessarily involves subjects and the
orderly series of their experiences. And since other kinds of collections of entities, including
purely formal mathematical sets, can be conceived as “impoverished” variants of a world, but
not the other way around, the world is the most fundamental notion of totality, to which all
other conceivable notions of totality must refer back.
Second, a world thus conceived will have a layered structure that needs to be worked out a
priori, that is, not through empirical investigation but through eidetic analysis. A concrete
human world includes a variety of different entities belonging in structurally different domains
of reality. A method is thus needed to isolate abstractively the different domains building up a
world and investigate them in their own right, as well as their interrelation. “Nature” will have
to be identified and demarcated as an enclosed domain within every possible world. What is
required is a transition from the informal sense in which we talk about nature, for instance, with
reference to the blossoming trees in a man-made orchard to a rigorous sense of nature that
highlights its function in the overarching whole of a world. Husserl writes:
The orientation toward [Einstellung auf] the ‘practical’ surrounding world yields a
universal a priori 4 structures that belong to a culturally significant world and a surrounding world in general (qua necessarily culturally significant); however, in this
orientation we necessarily include in our inquiry the subjects as active as well as their
activities, which are themselves worldly. Nature, too, is a structure that belongs in the
investigation, although it hasn’t been purified through ‘impoverishment’ [Abbau].
There is constantly talk of nature, however, it hasn’t yet displayed its pure, enclosed
a priori (Husserl 2008b, 264).
In order to transition from the informal talk of (unpurified) nature in the context of the human
surrounding world to a rigorous talk of nature as a self-enclosed and a priori structured domain,
we need a deliberate method: “an abstractive impoverishment [Abbau] of the concrete world,
whose primary final goal we call a removal [Abbau] of all intentional cultural and subjective
realities in order to abstractively attain mere nature” (Husserl 2008b, 265). Mere nature in this
fundamental sense is the “core-structure of this world” (Husserl 2008b, 279).
So, how do we get from the full-blown, practical and culturally significant human world to
mere or pure nature? Husserl’s Abbau has three steps. First, we need to bracket out all those
meaning-properties of things that stem from our human activity. These include both properties
that stem from theoretical activity (e.g. copper is a good conductor) and value-properties that
stem from practical activity (e.g. that is a challenging cliff!) (see Husserl 2004, 293–294). The
second step is an abstraction from all feeling qualities attaching to things. Husserl notes that
there is no experience and, correlatively, no object of experience that doesn’t come with some
64

Phenomenology and liberal naturalism

“emotional coloring” (Husserl 2004, 294; Husserl 2008b, 269). Things always matter to us in
some way; however, we deliberately decide to “turn off the faucets” of feeling, as it were, and
see what remains of the world. Husserl’s answer is: “What remains is the ‘mere natural object’
with its purely physical properties, i.e. a determinate ‘something’ that has its spatial extension, its
inner properties that spread over the spatial figure, endure in time, change, and, in their change,
stand under causal laws” (Husserl 2004, 295). Third, and final, in order to be focusing exclusively on the natural object thus identified, we also need to bracket out the subjective
manifold of appearances of the physical properties, that is, those lived-experiences that, as we
argued above, are necessarily tied up with the very notion of a world, but that do not belong in
the sphere of nature qua experiences. At the level of pure nature there is no experience qua
experience; all genuine, first-personal subjectivity is bracketed out and there is only room for
psychical phenomena as causally inert appendages of physical bodies.
It should be noted that when Husserl talks about “physical properties” in this context he
does not mean properties like atomic number and electric charge, but those properties that
belong to the physical object as it shows up in intuitive sensory experience, such as color, shape,
(felt) weight, (felt) temperature. The abstract layer of pure nature that lies at the core of every
possible world is still an object of direct experience (Husserl 2004, 294). It takes the idealising and
mathematising mode of thinking of contemporary science to construct the perfect circles of
geometry out of the roughly circular shapes of intuitive experience, and to create the idea of an
all-encompassing and thoroughly deterministic causality on the basis of the imperfect “causal
style” (Husserl 1970, 31) of the world, which only amounts to the vague experience that the
momentary state of a thing functionally depends on its surroundings, and leaves plenty of room
for sheer chance (Husserl 2008b, 301–302). The details of how natural-scientific thinking
constructs a world in itself behind the world of sensory appearance through idealisation and
mathematisation are provided in Husserl’s famous last work The Crisis of the European Sciences
and Transcendental Phenomenology (Husserl 1970 [1936]). Discussing Husserl’s Crisis here would
lead us too far afield.5 It is nonetheless noteworthy that Husserl’s “pure nature” fits neither the
“scientific image”, nor the “manifest image”, to use Sellars’ famous distinction. The fundamental layer of pure nature is no longer the full-blown world of culture, populated by values
and people, but it is neither the world of unobservables and quantum mechanics. Pure nature
still falls within the purview of sensory experience, but its legitimate scientific investigation
requires the introduction of idealising and mathematising procedures. Pure nature is, so to speak,
the go-between between the human world of culture and the mathematised world of physics
and, at the same time, the dimension that guarantees that despite their fundamental differences
the two “worlds” do not and cannot drift apart completely. No matter how abstract and
counterintuitive it becomes, natural science is necessarily about nature, a dimension of reality that
natural science doesn’t create, but finds as given as an abstract layer within the human world. In
an appendix to his lectures on ethics, Husserl summarises effectively his view as follows:
Nature is manifestly a mere structure of the full-blown world given to us, and thus to
every subjectivity. Scientific nature is not the one true world, it does not exhaust the
true world. Natural science is just a system of sciences, i.e. the ones that focus
on the lowermost structure which can be accessed at any time through the elimination
of the functions of valuing and willing that we have described (Husserl 2004, 336).
To return to the main line of thought, how do we distinguish between properties that belong in
the cultural, emotional or broadly subjective sphere of Geist and those that are genuinely
natural? Husserl suggests, again, a method of variation. Natural objects are constantly caught up
65

Andrea Staiti

in change. We can reproduce that ceaseless becoming in our imagination and ask what
properties, if modified or thought away, affect the object in its very essence and what properties
fail to do so. Clearly, the cultural value-properties can change without affecting an object in its
essence. A pre-historic spear was experienced as having the value-property “good for hunting”
by pre-historic folks, but is now experienced as having the value-property “precious archeological item”. The change in cultural value-properties did not affect the object in its essence.
The same goes with feeling-properties. The subjective emotional coloring of a landscape can
vary with time for the person contemplating it, without therefore causing a change in essence
to, say, the mountain that is part of it. By contrast, the natural properties are the identitypreserving properties, that is, the ones whose variation causes changes in the individual essence of a
thing. If, for instance, we modify in thought the physical property “solid” and change it into
“liquid”, the spear ceases to be a spear. If we try to strip away “extension” altogether from a
physical thing, the very essential unity of the thing is destroyed, unlike what happens if we strip
away “cultural significance” or “emotional coloring”, as Husserl’s method of Abbau requires.
Nature is thus a self-enclosed, abstract and at bottom purely physical layer in our concrete
human world. In order to remain focused on nature, for Husserl, we have to switch from our
everyday, personalistic attitude and adopt what he calls the “naturalistic attitude”. Husserl points
out that “nature is precisely the intentional correlate of experience as carried out in this attitude” (Husserl 1989, 4), which confirms my reservations about Drummond’s talk of nature
“as the correlate of what is discovered when we adopt what Husserl calls the personalistic
attitude” (Drummond 2015, 139). The naturalistic attitude is the mindset of practicing scientists: it is the habitualised, and mostly unaware, version of the Abbau discussed above. In the
naturalistic attitude scientists “see” only nature, such that they tend to either eliminate or reduce
to their natural underpinnings all the properties and entities that are not straightforwardly
physical. However, just as much as nature is an abstraction from the full-blown world of human
culture, the naturalistic attitude is an abstraction from the personalistic attitude:
Upon closer scrutiny, it will even appear that there are not here two attitudes with
equal rights and of the same order, or two perfectly equal apperceptions which at once
penetrate one another, but that the naturalistic attitude is in fact subordinated to the
personalistic, and that the former only acquires by means of an abstraction or, rather,
by means of a kind of self-forgetfulness of the personal Ego, a certain autonomy
(Husserl 1989, 193).
The naturalistic attitude that reveals nature is thus subordinated to the personalistic attitude that
discloses the concrete world of human culture and the living subjects that inhabit it. In the same
sense in which it would be absurd to require that the naturalistic attitude somehow “expand” to
include the personalistic, as if natural scientists ought to readmit typically personalistic items
such as goals, values, consciousness, etc. in their disciplinary purview, it is, on closer inspection,
absurd to imagine to “expand” nature in order to include in it those dimensions of reality that
build up the human world.
In sum, the phenomenological strategy to concretise the anti-scientistic agenda it shares with
liberal naturalists does not contemplate the extension of the concept of nature. Rather, the
concept of nature should be kept as narrow and austere as the things themselves require, but
without forgetting that nature is not a readily available and self-standing domain of reality. It is,
rather, an abstraction obtained when a specific theoretical interest is directed at the only
concrete world that there is: our cultural and value-laden human world. Nature is the abstract
core-layer of the human world, and it is revealed and explored by concerns that are thoroughly
66

Phenomenology and liberal naturalism

human. The overarching ontological whole is not nature, but the concrete world of humans, in
which nature is integrated as a necessary fundamental layer. This strategy is more attractive than
the one suggested by the expansion of the concept of nature, among other things, because it
blocks from the beginning unwarranted speculations about natural teleology, without therefore
giving up the reality of the only domain where talk of teloi is legitimate and even necessary: the
human world.

6 Conclusion
Let us recapitulate the results I take this chapter to have achieved. (1) Husserlian phenomenology and liberal naturalism are compatible insofar as they share a common philosophical
agenda: overcoming scientific (a.k.a. scientistic) naturalism. (2) Liberal naturalism sets out to
overcome scientific naturalism through an expansion of the concept of nature, which, however,
raises a question about the necessity to restore natural teleology, as Thomas Nagel shows. (3)
Husserlian phenomenology, by contrast, rejects the thought of natural teleology and defends an
austere concept of nature; however, (4) Husserl shows that nature is not a self-standing and
readily available domain in its own right: nature is an abstract core-layer within the only
concrete world that deserves to be called “real”: the human world. Nature is revealed through a
method of abstractive impoverishment from the world of culture, and it is thus the correlate of a
thoroughly unnatural attitude, namely, the naturalistic attitude, which, in turn, is subordinated
to the personalistic attitude. The fact that the naturalistic attitude is subordinated to the personalistic attitude is not meant to delegitimise natural science; rather, it is meant to ground in
the most radical way possible its legitimacy within its limits, in keeping with the critical impetus
of the best transcendental philosophy since Kant.
By way of conclusion, let me note that Husserl’s views are, in fact, not without a deliberate
tinge of transcendental idealism, which, however, is hardly avoidable by anyone seriously trying
to offer an alternative to scientism, and is thus a good thing, rather than a shortcoming. As I
mentioned in the previous section, we have to take the relationship between “world” and
subjectivity to be more intimate and constitutive than a naïve realist would have it. There may
be lots of things going on “out there” without subjects, but in order for there to be a world as a
distinctive, orderly, and layered form of totality, subjects are needed. Moreover, as a consequence, we have to take the relationship between attitudes and dimensions of reality revealed
through them as constitutive. The priority of the personalistic over the naturalistic attitude
reflects a priority in the dimensions of reality revealed through them, such that it is not “just”
our human, contingent perspective that determines the fact that nature is an abstract layer of
reality. The point of phenomenology is that our perspective as subjects is the only perspective
from which other perspectives, including the ones that efface subjects from view, make sense
and matter in the first place. I take this last statement to condense the deep intellectual kinship
between Husserlian phenomenology and liberal naturalism, despite the difference in their respective philosophical strategies.

Acknowledgments
I presented sections of this chapter at University College Dublin and at the University of
Helsinki. I would like to thank Ruth Boeker, Jim O’Shea, Brian O’Connor, Sara Heinämaa,
Mirja Hartimo and all the other participants in my talks for their helpful feedback. Andrea
Cimino and Emiliano Trizio generously read and commented on an earlier draft, which led to
significant improvements. The final version is entirely my responsibility.
67

Andrea Staiti

Notes
1

3
4

For the sake of simplicity, I leave it open whether ideal meanings are instantiated in the conscious acts
themselves, as Husserl contends in the Logical Investigations, or in the intentional objects as correlates of
conscious acts, as Husserl redefines his view from Ideen, onwards. In both cases, the possibility of intentionality is essentially grounded in the phenomenon of meaning.
In a response to a paper by Hans Fink (2008) McDowell writes about labels such as “liberal naturalism”, “Aristotelian naturalism”, “relaxed naturalism” etc.: “Perhaps I should not have used such
labels at all. In any case, I want to insist that they do the work I want them to do only in a certain sort
of dialectical context, and one falsifies what I wanted to do with them if one lifts them out of the
context and presses for a formulation of a positive doctrine they might be names for” (McDowell
2008, 216).
Geist is a notoriously difficult term to translate. It includes the world of culture and the full scope of
consciousness, including the low-level phenomena of sensation and association (passive Geistigkeit) and
the high-level personal phenomena of position-taking, valuing, etc. (active Geistigkeit).
It should nonetheless be noted that in later writings Husserl did leave room for the possibility of a
teleological perspective on natural history, even though he was never tempted to plant teloi in nature at
the ontological level. Viewing the world as a whole in teleological terms, however, amounts to viewing
the world from the perspective of God, for whom, by definition, it is the whole world that is included
in his practical sphere of influence, since he (or she, or it) is supposedly the one who created it. There
is, indeed, a deliberately theological dimension to the unpublished writings of the later Husserl, which,
however, is largely programmatic and never infringes upon his phenomenological perspective on nature,
that is, the one outlined in this chapter. Husserl believed that, in due time, phenomenology would
provide the foundations for a rational metaphysics dealing with vexed but urgent questions such as the
existence of God and the immortality of the soul. This is an area of his thought that is still largely
unexplored and it is unclear to what extent it is meant to be continuous and consistent with the strictly
phenomenological dimension. For an illuminating discussion of these issues, which largely exceed the
scope of this chapter, see Trizio (2018).
But see Staiti (2014, 222–263), Staiti (2018) and Staiti (2020) for further details.

References
Crowell, S. Normativity and Phenomenology in Husserl and Heidegger. Cambridge University Press. (2013).
De Caro, M. and Macarthur, D. Liberal Naturalism. In R. Audi (ed.), The Cambridge Dictionary of
Philosophy. Cambridge University Press. (2013).
Drummond, J. Neo-Aristotelian Ethics: Naturalistic or Phenomenological. In J. Bloechl and N. de
Warren (eds.), Phenomenology in a New Key: Between Analysis and History, pp. 135–149.
Springer. (2015).
Fink, H. Three Sorts of Naturalism. In J. Lingaard (ed.), John McDowell: Experience, Norm, and Nature,
pp. 52–71. Wiley Blackwell. (2008).
Husserl, E. The Crisis of the European Sciences and Transcendental Phenomenology. Northwestern University
Press. (1970).
Husserl, E. Philosophy as a Rigorous Science. In P. McCormick and F. Elliston (eds.), Husserl: Shorter
Works, pp. 166–197. Notre Dame University Press. (1981).
Husserl, E. Ideas Pertaining to a Pure Phenomenology and to a Phenomenological Philosophy. Second Book: Studies
in the Phenomenology of Constitution. Edmund Husserl – Collected Works III. Kluwer. (1989).
Husserl, E. Briefwechsel – Band V: Die Neukantianer. Kluwer. (1994).
Husserl, E. Einleitung in die Ethik. Vorlesungen Sommersemester 1920 und 1924, Husserliana XXXVII.
Kluwer. (2004).
Husserl, E. Introduction to Logic and Theory of Knowledge. Lectures 1906/1907, Edmund Husserl – Collected
Works XIII. Springer. (2008a).
Husserl, E. Die Lebenswelt. Auslegungen der vorgegebenen Welt und ihrer Konstitution. Texte aus dem Nachlass
(1916–1937), Husserliana XXXIX. Springer. (2008b).
Husserl, E. Einleitung in die Philosophie. Vorlesungen 1916–1920, Husserliana Materialien IX.
Springer. (2012).
Korsgaard, C. Self-Constitution: Agency, Identity, and Integrity. Oxford University Press. (2009).
McDowell, J. Mind and World, Harvard University Press. (1996).

Phenomenology and liberal naturalism
McDowell, J. Two Sorts of Naturalism. In Mind, Value, and Reality, pp. 167–197. Harvard University
Press. (1998).
McDowell, J. Response to Hans Fink. In J. Lingaard (ed.), John McDowell: Experience, Norm, and Nature,
pp. 214–219. Wiley Blackwell. (2008).
Moran, D. Husserl’s Transcendental Philosophy and the Critique of Naturalism. Continental Philosophy
Review, 41, (2008). 401–425.
Nagel, T. Mind and Cosmos: Why the Materialist Neo-Darwinian Conception of Nature is Almost Certainly False.
Oxford University Press. (2012).
Pearson, K.A. and Protevi, J. Naturalism in the Continental Tradition. In K.J. Clark (ed.), The Blackwell
Companion to Naturalism, pp. 34–48. Wiley Blackwell. (2016).
Staiti, A. Abbau. In H.H. Gander (ed.), Husserl Lexicon, pp. 17–18. WBG. (2010).
Staiti, A. Unforgivable Sinners? Epistemological and Psychological Naturalism in Husserl’s Philosophy as a
Rigorous Science. Rivista Internazionale di Filosofia e Psicologia, 3(2), (2012). 147–160.
Staiti, A. Husserl’s Transcendental Phenomenology: Nature, Spirit, and Life. Cambridge University
Press. (2014).
Staiti, A. Pre-Predicative Experience and Life-World: Two Distinct Projects in Husserl’s Late
Phenomenology. In D. Zahavi (ed.), Oxford Handbook of the History of Phenomenology, pp. 155–172.
Oxford University Press. (2018).
Staiti, A. Giovanni Piana’s Conversazioni and Some Recent Controversies on Husserl’s Krisis.
Phenomenological Reviews, Special Issue I, https://reviews.sdvigpress.org/pub-257431. (2020).
Trizio, E. The Telos of Consciousness and the Telos of World History, HUMANA.MENTE. Journal of
Philosophical Studies, 11(34), (2018). 77–103.

7
MERLEAU-PONTY
AND LIBERAL NATURALISM
Jack Reynolds

1 Introduction
Maurice Merleau-Ponty (1908–1961) was a phenomenologist. His Phenomenology of Perception
(1945) is one of the classic texts of that tradition. He also seriously engaged with a range of
sciences throughout his career, from psychology through to biology and physics. Some philosophers even proclaim him to be one of the first cognitive scientists, avant la lettre.1 Although
he said critical things about naturalism, along with many other phenomenologists, he also
accepted a “truth in naturalism”, even if it could not be said to be the truth that constrains all
others. As neither a classical naturalist nor a non-naturalist, then, he appears to be a moderate or
liberal naturalist (hereafter LN).2 But can a phenomenologist really be a naturalist, even a liberal
one? A lot hinges on how we tease this out, both as to whether it is plausible to claim MerleauPonty as an LN (I argue it is), and as to whether it is an attractive and coherent position. Indeed,
despite its important challenges to orthodox naturalism, there are arguably two traps for LN to
avoid. If it becomes too liberal, we get: dualism or an ontological pluralism that is difficult to
distinguish from a constructivism; or, in seeking to sidestep that metaphysical dilemma, there
might be an insistence on an overly neat methodological separation between description/understanding and explanation that is belied in practice (both scientific and philosophical). It is
doubtful that such positions can legitimately claim to be naturalist in orientation, liberal or not.
Merleau-Ponty’s version of LN avoids these traps, however, and it is thus a useful resource for
contemporary work trying to navigate between scientific naturalism and non-naturalism.

2 Liberal naturalism
LN refers to a diverse group of philosophers, including John McDowell, Hilary Putnam and
Huw Price, and others like Strawson and Rawls are sometimes co-opted (see, e.g. De Caro and
Macarthur 2010). They are unified in opposing scientific naturalism, however, and we can
hence define LN in opposition to what it rejects. Primarily, this is “strong” versions of naturalism, which involve claims to exclusivity of some kind: that is, only scientific theories approximate to the truth; only the knowledge gleaned through scientifically reputable methods
warrants the name; the only entities that exist (or are real) are those posited or required by our
best sciences. LN aims to weaken or pluralise these exclusivity claims, whether in the
70

DOI: 10.4324/9781351209472-9

Merleau-Ponty and liberal naturalism

methodological, epistemic, or ontological registers, without succumbing to non-naturalism or
dualism. This is the dialectical difficulty of any “neither/nor”/“middle-way” strategy. It is liable
to be criticised by the opposing sides for failing to give them what they want, or, if it is
embraced by either side, it risks losing its point of difference.3 In addition, there are more than a
few philosophers who contend not just that a particular middle-way option does not work, but
that there is no viable intermediate position between dualism and physicalism (Kim 1999), or
idealism and materialism (Gardner 2007).
Epistemically, LN is committed to the idea that, as Mario De Caro puts it, “there are legitimate ways of knowing and understanding that are neither reducible to, nor incompatible
with, scientific method” (De Caro 2014, and forthcoming). Such a claim is not obviously antiscientific, but it needs to be established what those ways of knowing are, and in what ways they
are irreducible to scientific method, which might be characterised as the “observing,
hypothesis-forming, testing, formal modeling, axiomatizing and explaining that have been so
successful in the natural, biological, social and mathematical sciences” (Railton 1997, 4).4 If
LN’s methods are reducible to scientific methods and modes of explanation, then there is no
reason for the naturalist to consider weakening their exclusivist epistemic commitments,
especially given they will often appeal (perhaps problematically) to reducibility in principle
rather than in fact.
But LN must also avoid the non-naturalist horn of the dilemma. As such, any allegedly
irreducible methods and ways of knowing must not smuggle in dualist intuitions, or special
reasoning powers that are outright contradicted by well-corroborated empirical studies.
Prosecuting that case also requires a move of demarcation concerning illegitimate ways of
knowing. If nothing is ruled out, LN’s epistemological openness threatens to collapse into an
“anything goes” non-naturalism. This is arguably also the case for any stipulative move that
simply declares different methods are at stake and that there is no significant epistemic problem
concerning the diversity of the sorts of knowledge claims produced. That strategy pluralises
ways of knowing, but replaces objectivism with relativism.
It is clear, however, that LN is open to the possibility that methods like phenomenology,
hermeneutics, transcendental reasoning, etc., might count as legitimate ways of knowing, which
is not the case for scientific naturalism. As such, Merleau-Ponty’s epistemic commitments might
be brought within the fold, if we can show that they are not reducible to the scientific, and if
there are sufficient criteria and exclusions to establish his naturalist credentials.
Ontologically, LN holds that there are entities (or structures, processes, etc.) that are not part
of the basic scientific furniture of the world, but which are real or exist. If the “super-natural”
move is to invoke Platonic forms, creationism, magic, God or vital teleological forces without
physical instantiation, LN insists on the ontological reality of more everyday middle-sized “dry
goods” as J.L. Austin puts it. These are the more mundane entities that constitute, or are
presupposed by, what Wilfrid Sellars calls the manifest image. They are not “spooky” in any
obvious sense, since all human societies are structured around them (e.g. persons, artefacts,
norms), but questions remain as to their ultimate compatibility with the scientific image in
terms of its methods and its objects/processes/structures. Are persons, artworks, along with all
of the so-called “4Ms” – Mind, Meaning, Morality and Modality (cf. Price 1997) – reducible to
scientific objects, and thus what naturalists call the “causal closure” of the physical?
A weak way to argue for the possibility of LN is to deny any incompatibility with science,
whether in terms of its methods or its objects. A more modally committed way is to show that
science (and scientific naturalism) constitutively depends on the manifest image, and/or on
what phenomenologists like Husserl and Merleau-Ponty call the Lebenswelt (life-world), which
is argued to be both prior to, and presupposed by, the idealising of natural science. These
71

Jack Reynolds

methodological and epistemic arguments aim to show that any project of elimination or reduction is untenable because it smuggles in what it purports to explain (e.g. the 4Ms, subjectivity, temporality, etc.). Any contention that there is a necessary presupposition of the
manifest image (or the life-world) typically remains at a level of abstraction, however. It enables
a kind of reconciliation by fiat, especially if it is also held (with Husserl) that any transcendental
investigation can learn nothing from the sphere of the “positive” and empirical science – indeed, that something like this would be a category mistake or nonsensical.5 Showing the
consequences of this solution is a much more complex matter in any more restricted domain of
inquiry, especially where there is the appearance of conflict between scientific claims and others
that are more philosophical and/or common-sensical. The naturalist is right to ask for some
details concerning their inter-relation, since otherwise it is a solution from “on high”, transcendentally secured perhaps, but not one that speaks to the conflict on the ground.
Despite this explanatory demand, the scientific naturalist is also sceptical of the global reconciliationist ambitions at the heart of LN (see De Caro 2014). Although various naturalist
programs aim to reduce everyday objects and ideas to scientific objects (or properties, theories,
etc.), the reduction conserves that which has been reduced in only a weak sense. After all, the
core claim of any reduction is that x is nothing more than y, or x is nothing over and above y
(e.g. Smart 1959). It is not eliminativism, because they do not deny that x is real. They simply
hold that x is ultimately y, at its most basic level, and the more basic level (or the reduced
theory) is considered to have the greater epistemic warrant. In practice, this means that orthodox naturalism has an antipathy towards attempts to “have it both ways”, which aim to unify
ordinary justification and our life-world with scientific inquiry (see, e.g. Kornblith 2014, 166).
In the end, if the basic theory or properties are defined by the norms and methods of natural
science, then as Sellars famously puts it: “In the dimension of describing and explaining the
world, science is the measure of all things, of what is that it is, and of what is not that it is not”
(Sellars, 1963, Sect. 41). This scientia mensura principle, as it is known, appears to endorse strong
versions of methodological and ontological naturalism. It opens up the prospect that the scientific image might eliminate or replace the manifest one, even if Sellars himself also advocates
the need for a “stereoscopic view” in ways that have been taken up by “left” Sellarsians like
McDowell, Brandom, Putnam and others.6
Non-naturalists also reject liberal projects of reconciliation. They do not seek to integrate
scientific and non-scientific ways of understanding, nor to render them “continuous” in either
methods, ontologies, or “results”. In negative vein, the non-naturalist will place pressure on any
meta-philosophy of “mutual constraint” between science and philosophy, pointing out that this
clarion call is often platitudinous and short on details, refrains that are very similar to those
coming from the naturalist (cf. Zahavi 2009). The non-naturalist’s positive solution, if it does
not explicitly countenance metaphysical dualism, involves a “methodological separatism” and a
division of labor in which each is allocated to their own separate tasks/domains (Reynolds
2018). Such a move has a long and complex philosophical history that I cannot do justice
to here.7
Husserl is an important figure to briefly consider, however. Although there are resources
within Husserl’s thought that can be utilised for LN (see Staiti this volume), his philosophy is
based on a strong methodological distinction between philosophy and science, and between
phenomenological philosophy (utilising the reduction) and ordinary reasoning, episteme and
doxa. Husserl shares this view with the orthodox naturalist,8 but he reverses what constitutes
episteme and doxa. Doxa paradigmatically includes any unreflective and naïve science. This is
“naturalism” in the pejorative sense, which he describes as the original philosophical “sin”
(Husserl, cited in Moran 2013, 92).
72

Merleau-Ponty and liberal naturalism

Husserl instead defended a version of phenomenology as first philosophy. In his lecture
course of that name, he says:
Mind (Geist) in nature, the adaptation of mind to its nature, the evolution of cognitive
minds, the evolution of the sciences and forms of human culture in general – all this
also has its philosophical sides, but these philosophical sides do not belong in the
theory of knowledge, to First Philosophy. I would say that they rather belong to “last
philosophy”.9
Husserl is quite clear here: the findings of science, and philosophical questions about the findings
of science, are last philosophy. It is difficult to imagine stronger non-naturalist statements than
these. It is also important to recognise the dismissal of any “best fit” approach to philosophy that
aims to integrate the variety of knowledge claims in an overall picture of how “things hang
together”, to invoke Sellars, since that sort of coherentist picture would be insufficiently epistemically grounded. In regard to the Agrippa trilemma, Husserl thus chooses foundationalism
over coherentism, notwithstanding that his later work famously proposed to address a crisis in the
sciences by reawakening their forgotten phenomenological bases (e.g. Crisis).
By contrast, naturalists generally argue that any such theory of knowledge, qua first philosophy, is a foolhardy ambition. Without expressly targeting Husserl, Sider captures the
prevailing sentiment, when he says: “it would be foolish to require generally that epistemological foundations be established before substantive inquiry can begin. Mathematics did not
proceed foundation-first. Nor did physics. Nor has ethics, traditionally” (Sider 2001, xiv). Or,
as Kornblith puts it:
the constraints that science presents for philosophical theorizing should be welcomed,
for philosophical theorizing unconstrained by empirical fact loses its connection with
the very phenomena which we, as philosophers, seek to understand. Philosophy is an
autonomous discipline, in the sense that it addresses a distinctive set of questions and
concerns, and in this respect it is no more nor less autonomous than physics or
chemistry or biology. It is surely all the autonomy we should want.
(Kornblith 2002, 27)
Who is right, Kornblith or Husserl? While I am much more indebted to Husserl in terms of my
own topical interests and philosophical inheritance, I agree with Kornblith regarding the autonomy of philosophy. Unlike Kornblith, however, I think that these theses regarding mutual
constraint are compatible with phenomenology, and in fact make phenomenological reflection
akin to a necessary but not sufficient condition, once we recognise that phenomenology is
always in dialectic with non and extra-phenomenology. Systematically showing that it is beyond what can be achieved in this chapter, I will sketch out the key aspects of Merleau-Ponty’s
thought that make the LN ascription reasonable.

3 Merleau-Ponty: perception and embodiment
Merleau-Ponty begins his first major work, the Structure of Behavior, by wondering if there is
anything in the naturalism of science that might find a place “in” transcendental philosophy,
rather than simply being bracketed away as a theoretical elaboration of what phenomenologists
call the “natural attitude”, which we might gloss as simply our common-sense beliefs and
opinions. As he puts it:
73

Jack Reynolds

And once the criticism of realistic analysis and causal thinking has been made, is there
nothing justified in the naturalism of science—nothing which “understood” and
transposed, ought to find a place in transcendental philosophy?
(Merleau-Ponty 1963, 4)
There is a lot to unpack in this remark, but Merleau-Ponty answers the question that he poses
in the affirmative, concluding that there is such a place and a “truth of naturalism”. Indeed, his
programmatic statements to this effect, at the beginning and end of the book, are important for
many empirically-minded phenomenologists (e.g. Gallagher 2017).10 Accepting naturalism (and
science) as playing a role in, as in “within” or internal to transcendental phenomenology, would
be a “category mistake” for some other phenomenologists, however. Indeed, this “category
mistake” style retort arguably remains the orthodox phenomenological response to MerleauPonty’s question.
While Merleau-Ponty is committed to phenomenology, he has a deflated conception of
transcendental phenomenology in comparison to Husserl, one that is more provisional and
fallibilistic. This might be a pro or a con, of course, but it is less suspect to a naturalist. Although
Merleau-Ponty embraces the phenomenological idea that there is a world of perception that has
its own sense and forms of intentionality that are irreducible to the causal, propositional and
linguistic, his Preface to Phenomenology of Perception famously contends that the most important
lesson of the phenomenological reduction is its incompleteability (Merleau-Ponty 2013, lxxvii).
As defended by Husserl, the phenomenological reduction aims to bracket away the successes of
our various sciences, as well as culture, history and metaphysics, in order to return to our
experience of the things themselves, without presupposition. On Merleau-Ponty’s view,
however, this is a heuristic that can bring certain phenomena to light that were previously
neglected (or where theory has overdetermined the observation/experience), rather than a
move that facilitates access to a presuppositionless starting point for a rigorous science of
consciousness. These methodological differences have some important consequences for
Merleau-Ponty’s views on embodiment and perception, which are pivotal to his own middleway between naturalism and non-naturalism and important for LN more generally.
In Phenomenology of Perception, Merleau-Ponty focuses on the body as a problem for empiricism and intellectualism, an antimony that he contends has characterised large parts of the
history of Western philosophy. This embodied turn is not unique to Merleau-Ponty’s thought,
being also present in a complicated way in Sartre’s Being and Nothingness. Merleau-Ponty also
borrows many of his analyses of embodiment directly from Husserl’s phenomenology, particularly from Ideas 2, which Merleau-Ponty was an early reader of in the Husserl archives in
Leuven. But Husserl and Merleau-Ponty’s treatments of embodiment have a quite different
methodological (and ontological) status. We have already seen Merleau-Ponty’s dialectical
focus, in which the role given to the body (and perception) in the empiricist and intellectualist
traditions is found wanting, both descriptively and explanatorily, thus motivating a rethinking
of it. This differs from the way in which Husserl usually motivates his own phenomenology of
embodiment.
In addition, we must consider their respective investment in, or denial of, the transcendental
ego. For Husserl, although our embodiment gives us a worldly structure, it is also constituted
within the field of consciousness by the “transcendental ego”, which Husserl argues is required
to understand both the empirical ego and the unity of the body-subject (cf. Staiti 2016, 131).
The body is not identical with the transcendental ego; rather, the latter is the ground and
condition of possibility of the former. For reasons related to this, Husserl is able to entertain
thought experiments concerning the destruction of the world but the persistence of the
74

Merleau-Ponty and liberal naturalism

transcendental subject (e.g. Sect. 49 in Ideas). For Merleau-Ponty, by contrast, embodiment is
paradoxically both constituting and constituted. What he calls the body-subject is the most basic
enabling condition, albeit situated within a what he calls a phenomenal or transcendental field
(this includes the sociohistorical, and it is not a field of pure consciousness). Any reduction that
aimed to bracket away our own embodiment would, on this view, be a Cartesian residue, and
Merleau-Ponty argues that some of the key phenomena of interest to Husserl in his later work
are presupposed by, rather than accessed through, the phenomenological reduction.11 Given
that naturalism typically has a suspicion of transcendental reasoning, an allegiance to Ockham’s
razor (e.g. an injunction not to multiply entities unnecessarily), and raises questions about the
capacity of reason to be autonomous, Merleau-Ponty’s embodied phenomenology appears to
be the more naturalist position, even if questions remain about how his nonmechanistic and
noncausal understanding of embodiment fits with physicalism (see Reynolds 2020).
Although Merleau-Ponty insists that the perceptual world needs to be appreciated on its
own terms as much as possible, his view of perception is again methodologically distinct from
Husserl’s, even if doctrinally similar. For example, Merleau-Ponty agrees with Husserl that any
perception occurs against a horizonal (or figure-ground) structure, and likewise emphasises a
complex co-imbrication of perception and motility (whether possible or actual). While there
are some technical differences between their views concerning the “co-presented” or “appresentational” aspect of an object we perceive (its backside, say), and their treatment of our
perception of other people (c.f. Overgaard 2017), the main difference between them is an
epistemic foundationalism that remains part of Husserl’s analyses.
For Husserl, the presentational aspect of perception makes it privileged, not just phenomenologically but also epistemologically. Husserl is interested in the justificatory force of experience, and perception is the “gold” standard. Without being able to consider potential
counter-examples like hallucination here (and debates concerning disjunctivism), Husserl’s
basic claim is that perception presents the relevant objects in the flesh, unlike say imagining or
thinking about those objects. For Husserl, that which is presented in this way (in a fulfilled
intuition) is justificatory in itself, rather than in virtue of anything else (e.g. the game of giving
and asking for reasons). Berghofer and Wiltsche (2019) give a useful example to help clarify
what is at stake. On having commenced a coffee elsewhere in their university, they wonder if
perhaps they had left their office door open, when they absent-mindedly ambled out together
immersed in a conversation. Any belief that Berghofer or Wiltsche might have that the door is
open or closed, prior to seeing it, gets its justification from other reasons we (or they) might
provide. For them, this might include a memory of perception, which is not quite goldstandard, and might elicit questions about the reliability of memory that appeal to other factors.
They might also draw on the testimony of a colleague who walked past the door, or engineering knowledge that the door is weighted in such a way that it tends to close of its own
accord. This gives general probabilistic knowledge that the door is likely to be shut, whether
they closed it or not. But that differs from the knowledge given in, and only in, the direct
perception of the door. To compress Berghofer and Wiltsche’s rich discussion, we have an
asymmetry here, which might be expressed by saying that seeing is believing (sometimes), but
believing is not seeing. Or, to the point in a more Merleau-Pontyian frame, with Claude
Romano: “to perceive is not to believe, it is to have no need to believe” (Romano 2016, 311).
Most (if not all) phenomenologists would agree with some version of these sentiments. But
on the Husserlian construal they are attached to the idea of “originary givenness” or “selfgivenness” of perception, which also plays an important role in Husserl’s famous phenomenological “principle of all principles”, which he outlines as follows:

Jack Reynolds

No conceivable theory can make us err with respect to the principle of all principles: that
every originary presentive intuition is a legitimizing source of cognition, that everything
originally offered to us in ‘intuition’ is to be accepted simply as what it is presented as
being, but also only within the limits in which it is presented there. (1982, Sect. 24).
This principle establishes something akin to an epistemic tribunal in which perception, as an
“originary presentive intuition”, has evidentiary priority. Depending on how we understand
the italicised remark (c.f. Berghofer 2017), other evidence is also comprehended in terms of its
potential fulfillment in any such presentive intuition. Although this is debated, such a view
appears to lead to scientific anti-realism about microscopic and theoretical entities in science
that are not directly perceivable (in regular vision), and perhaps to a denial that the expert
perceives what they think they perceive in an x-ray, say, or through a microscope (c.f. Wiltsche
2012; Berghofer 2017; Reynolds 2018, Chap. 3). Here the phenomenological epistemology
appears to license conclusions that are ontological or metaphysical in character.
While there are various possible neo-Husserlian rejoinders, Merleau-Ponty provides an
alternative construal of the primacy of perception. He contends that perception always involves
latency, depths, invisibility and a persistent ambiguity. Perception has a primacy, but he would
dispute that it is akin to a gold-standard that institutes a normative framework for reason and
evidence, as well as the associated idea that science must (ultimately) be based in a self-present
intuition of the object for it to be belief-worthy. While phenomenology can help us to avoid
what he calls the “experience error”, in which we make perceptions out of the things perceived
(and retrospectively read theory back onto our experience), perception remains intrinsically
paradoxical, rather than an epistemic ground from which we might build up knowledge a la
empiricism. Perception is ambiguous in both the natural attitude and in our philosophical reflection on it. It has its own sense, but it does not inaugurate an epistemic tribunal of reason
against which we can judge or measure rationality per se. Despite having himself used this very
term, Merleau-Ponty thus ultimately avoids a “phenomenological positivism” about perception, as well as the idea of phenomenological epistemology as first philosophy.
In his final, unfinished work, The Visible and The Invisible, Merleau-Ponty returns to this
“perceptual faith”. He argues that this “faith” is not a reason or judgement, but refers to an
openness (1964, 88) or “contact with being prior to reflection, a contact that makes reflection
itself possible” (65). While “we can live it, we can neither think it nor formulate it, nor set it up
in theses…And it is this unjustifiable certitude of a sensible world common to us that is the seat
of truth within us” (11). For Merleau-Ponty, this faith also involves an ambiguous realism about
the connection between perception and that which is perceived, which is both a part of ordinary experience and complicates any attempted phenomenological reduction. This perceptual
faith is likely to be suspect to the naturalist, who maintains that the true is the objective. Perhaps
it is ultimately little better than an illusion. Merleau-Ponty wants to guard against that conclusion, but not by simply accepting the opposite thesis, that the true is restricted to (or merely
an elaboration of) the perceptual faith. As Merleau-Ponty puts it:
if the philosopher questions, and hence feigns ignorance of the world and of the
visions of the world which are operative and take form continually within him, he
does so precisely in order to make them speak, because he believes in them and
expects from them all his future science (1964, 4).
Note that Merleau-Ponty does not posit any sort of radical rupture between the pretheoretical
experience of the perceptual faith and future science. Rather, he speaks of the paradoxes of the
76

Merleau-Ponty and liberal naturalism

vision and their “incompossible details”. Whatever challenges science poses to our ordinary
experiences of the world, and there are many, it assumes the background of an ongoing perceptual faith (otherwise we have skepticism, or some kind of Platonic heaven, both of which
appear immanently problematic for modern science and Merleau-Ponty alike12). But we cannot
reach further back to an epistemic ground, where that faith is no longer a faith but instead
becomes strictly evidentiary and the basis for a rigorous science. To argue for the necessary
presupposition of this perceptual faith, paradoxical as it is, is therefore not to limit science to the
evidences of experience, not to commit the sort of justificatory mistake that Sellars calls the
“myth of the given”.13
The Visible and the Invisible also offers some interesting reflections regarding ontological
naturalism. According to Merleau-Ponty, “no ontology is exactly required by the thought
proper to physics at work” (1964, 17). Note that this formulation does not insist on a neoHeideggerian account of the ontological difference between ontic sciences and any properly
ontological philosophy, nor a neo-Husserlian version of phenomenological first philosophy in
which there is no encroachment between the philosophical and the scientific. On the contrary,
Merleau-Ponty’s remark allows that physics might constrain our ontology, but he also claims
that if it does, it does not do so in any reductive manner: no ontology, he says, is “exactly
required”, which is not to rule out some form of constraint. Contemporary physics does not
entail physicalism, we might say today, and various deflationary naturalists like David Papineau
and Arthur Fine might agree, along with those like John Dupré who adopt a process ontology.
In a related vein, Merleau-Ponty says that Einstein effectively postulates that: “what is is not
that upon which we have an openness, but only that upon which we can operate” (1964, 18). Recall
that he defines the perceptual faith as “that upon which we have an openness”. Certain versions
of philosophical science (i.e. scientific naturalism), then, ignore this openness and this faith. And
whether or not Merleau-Ponty is fair in regard to Einstein, we might think here of Quine’s
famous remark, “to be is to be the value of a bound variable”. This is a version of ontological
naturalism in which the furniture of the world is any items expressly posited or required for our
most mature sciences to do their work, in this case including mathematics.
For Merleau-Ponty, however, this proceeds from and depends upon the perceptual faith, but
does not dissipate its obscurity. Any challenge that science poses to the perceptual faith will take for
granted some dimensions of experience, observation and connection with the world. As he puts it:
The binocular perception is not made up of two monocular perceptions surmounted;
it is of another order. The binocular images are not in the same sense that the thing
perceived with both eyes is. They are the phantoms and it is the real; they are the prethings and it is the thing (1964, 7).
Here he reaffirms the ontological dimension of the perceptual faith. Any account of what is
must be pluralised to include, rather than eliminate, this particular faith.
Some inheritors of the naturalist mantle have a very different view. Consider, by contrast,
what contemporary theorists of predictive processing and binocular rivalry would say about the
remarks from Merleau-Ponty we have quoted. Binocular rivalry occurs in experimental settings
when different images are presented to each eye and an experiential oscillation between them is
induced. Without being able to address the details of the predictive processing program (see
Hohwy 2013), it explains this visual experience by drawing on the idea of the brain as a
hypothesis engine that seeks to minimise prediction error. Some interpretations of this are more
“internalist” than others, but the Bayesian-like operations of the brain involve a form of
guessing (albeit informed by “priors”) on the causes of a given stimuli, which then sets up
77

Jack Reynolds

models of the world. On this view, however, perception is inferential and there is a genuine
sense in which “we” (or our brains, at least) are quite radically separated from the world and
groping in the dark. Our embodied and worldly sense of perceiving and interacting with the
world of ordinary objects is indeed akin to a faith, in a pejorative sense, which science can help
to explain if not dispel. The problem with this sort of “user illusion” idea that accompanies
illiberal naturalism, is that it appears to jeopardise the grounds from which they want to proclaim their own scientific theories as true, which are experiential and observational, and also
complexly related to the practical life of this or that scientist (c.f. Zahavi 2018). Perhaps the
appeal to mathematics and Bayesian formalism can assist the predictive processing theorist to
resolve this prima facie problem, or perhaps they might go more radically embodied and enactive in their view of cognition. Without resolving that here, my point is simply that scientific
realism and manifest-image irrealism is not an easy position to coherently hold. As Zahavi
suggests, there is an ordinary and everyday realism about phenomenology (2018), but in my
view the prospects for reconciliation of a perceptual realism with (rather than against) scientific
realism are better with a Merleau-Ponty-inspired version of phenomenology.

4 Conclusion: Merleau-Ponty as liberal naturalist
Are Merleau-Ponty’s transformations to the phenomenological project sufficient to make him
an LN? It is clear that Merleau-Ponty is not a scientific naturalist, since he is not committed to a
strong (i.e. exclusive) reading of either ontological or methodological naturalism. He thus meets
condition 1 for being an LN.
Whether Merleau-Ponty meets condition 2, and avoids non-naturalism, is more difficult to
establish. This is because of a lack of clarity around what non-naturalism might be, and the
reasonable concern that many accusations of “spookiness” are a “straw-man” argument that the
naturalist uses to dismiss a range of view. Indeed, virtually no philosopher explicitly accepts
supernaturalism, or explicitly posits entities or ways of knowing “whose existence or truth would
contradict the laws of nature insofar as we know them” (De Caro and Macarthur 2010, 12).
Nonetheless, the issue is whether or not Merleau-Ponty implicitly contravenes norms of naturalism and a more nuanced understanding of condition 2 that really does apply to at least some
positions. If we rule this out, we seem to sidestep naturalism, effectively saying there is no
genuine issue to address. Sidestepping naturalism may or may not be the right philosophical
strategy, but it does not allow a philosopher to establish that their position avoids supernaturalism and constitutes a genuine middle-way.
In the epistemic and methodological register, an allegation of spookiness is likely to accompany the ascription of implausible powers of cognition to the armchair philosopher. In this
respect, Merleau-Ponty engages in forms of reasoning that many naturalists would regard as
suspect. Notably, he is part of a tradition of transcendental phenomenology, even if unlike most
in that tradition he insists that the distinction between philosophy and science is imprecise and
that, insofar as they can be distinguished, they exist in something like a relationship of “mutual
constraint”, as Francisco Varela and others have emphasised in recent times (1996, 343).
Merleau-Ponty’s work is also arguably unique in this tradition by proceeding through systematic engagement with a variety of sciences – including Gestalt psychology, physics, anatomical biology, morphogenesis and beyond – in ways that respect the epistemic credentials of
science and do not set limits upon what science might achieve from the philosophical armchair.
He is not a methodological dualist, or a methodological separatist. His project is not to rigorously distinguish the empirical from the transcendental, or the a priori from the a posteriori,
and assign to each their “relative right” and autonomy (see Staiti 2016; cf. Reynolds 2018).
78

Merleau-Ponty and liberal naturalism

Of course, without that distinction in principle, it might be thought to be a “slippery slope”
to the Sellarsian view that science is the measure of all things, and to a strong version of
scientific naturalism. Again, this is the dilemma that confronts purveyors of middle-ways, including Merleau-Ponty. While I have claimed he is not a scientific naturalist, that does not free
him from the horns of this dilemma, since the worry might be that there is no coherent and
stable “middle” space to occupy. For Merleau-Ponty, however, it is the dialectical aspect of his
thought that is central to his solution. It means that justification is holistic more than foundationalist, and it enables him to endorse something akin to a “results continuity” thesis regarding philosophy and the relevant empirical sciences over the long haul, albeit not necessarily
deference to the current scientific orthodoxy. He also claims that the actual methods of each are
complexly intertwined, with something like an a priori lab and an a posteriori arm-chair, rather
than being strictly autonomous. By my lights, then, Merleau-Ponty both avoids scientific
naturalism and has enough constraints built into his philosophy to rule out first philosophy and
other forms of non-naturalism, especially in his early work.
Ontologically, however, how readily might Merleau-Ponty avoid the charge of supernaturalism? This charge will be levelled at any hint of dualism, any area where the regular
spatio-temporal order of cause and effect does not apply. There can be a naturalist “black-mail”
about this, of course, in which consciousness itself might become suspect and we end up with a
rather austere and disenchanted view of the world. But is there an implicit ontological dualism
in Merleau-Ponty’s thought, when he suggests that “perception is not an event in nature”? Or,
is there a tacit dualism, if he effectively advocates an emergentist view, contrary to the standard
naturalist acceptation of the causal closure of the physical, as I contend elsewhere (Reynolds
2020)? Both cases commit him to dualism from an orthodox naturalist perspective, but he is
expressly endeavoring to set out a middle-way. So this is not a simple question to adjudicate,
without presupposing this or that epistemology and metaphysics. It is clear that, like Husserl,
Merleau-Ponty does not think that everything real is explicable in terms of a physical law, best
stated at the level of physics. While this view is controversial, it does not entail that he is a nonnaturalist, and/or anti-science. On the contrary, both the thesis of the unity of science, and
various strong ontological construals of naturalism (e.g. physicalism and the causal closure of the
physical), go beyond anything a scientist would be liable to proclaim (cf. Ritchie 2005).
Moreover, a temporal and biological naturalism remains a viable prospect (see Reynolds 2020;
Morris 2018) that is less exclusionary in terms of what it rules out. For example, it accords a
holistic explanatory priority to organisms, rather than subscribing to the sort of “smallism” in
which what is real (causally and ontologically) is the micro-physical, the neurological or genetic
(i.e. DNA). In other words, taking the naturalist challenge seriously does not entail that we
must give in to it. There is room for ontological work beyond saying that our current best
sciences just give us our ontology. If that ontology is promulgated through reflection on those
sciences, including lacunae in their own self-understanding,14 then we face a difficult interpretive question. Is any resultant philosophy a form of speculative metaphysics, which the
naturalist will rule out, and might also be taken as a bridge too far for even a liberal naturalism?
Does it re-enchant nature, or are there ways of rethinking nature that are not super-naturalist?
These questions are especially pressing in regard to The Visible and the Invisible, where
Merleau-Ponty elaborates a new philosophy of nature (cf. Gallagher 2018; Morris 2018). It is
not meant to be dualist, of course, but to what extent is this ontology of nature, replete with
terms like chiasm, flesh, etc., compatible with naturalism? It is fair to say, with Gardner (2017),
that not many naturalist philosophers are interested in these sorts of ideas, which are not based
on concepts part of any mature science. There is certainly a rethinking of nature at stake here, if
not a re-enchanting. For many naturalists, any re-enchanting of nature that conceives of
79

Jack Reynolds

teleology and purpose (say) as part of nature, or enactivist reconstruals of the relationship between life and mind (like Thompson 2007) are prima facie suspect. Why? Nature is conceived
of in a certain way, primarily inherited from Newton, involving objective forces, fields, molecules (or fermions and bosons in recent times), and governed by laws. This understanding is
dominated by “philo-physics” and the idea of a hierarchy of sciences with potential intertheoretic reduction – via bridging laws – between them. But this is not the only way to think
about nature, even within physics, and especially if we draw on other sciences (cf. Gallagher
2017; Gallagher 2018), and perhaps biology in particular (cf. Reynolds 2020). Likewise, if we
proclaim the diversity of methods and objects of the sciences and deny any thesis of the unity of
science (e.g. with the work of Dupré), there is again more philosophical room to move.
In Merleau-Ponty’s work, both early and late, certain themes persist that promise a way
beyond the dilemma of a primacy of perception and a faith in everyday objects, on the one
hand, versus a primacy of scientific objects and theories on the other hand, the latter of which is
standardly accompanied by reductionist programs that seek to explain the whole compositionally. LN is right to rethink this and explore new options between naturalism and nonnaturalism. In my view, Merleau-Ponty is an essential dialogic partner for this project. LN
needs phenomenology to articulate the relationship between the manifest and scientific images
in a way that embraces an ambiguous perceptual realism that can be reconciled with the scientific image. Husserlian phenomenology purports to achieve this reconciliation through a
strict methodological separation and an epistemic priority of perceptual presence at the expense
of the abstractions of the scientific world. Husserl argues that our increasingly abstract sciences
are in crisis and in need of renewal via a phenomenological undertaking that uncovers the
perceptual and phenomenological bases from which our sciences have developed and since lost
contact (c.f. Husserl 1970). His proposed reconciliation is radical, but being foundational rather
than dialectical, it is hard to motivate with contemporary naturalism. Merleau-Ponty’s phenomenology presents as a more likely and propitious partner for LN in the work ahead. That
said, it might also be that the preferable theoretical resource for versions of LN that aim for
metaphysical neutrality and/or quietism is a version of descriptive Husserlian phenomenology
that is shorn of the foundationalist commitment to revivifying science through phenomenology.15 Just as phenomenology bifurcates on this question of metaphysics, so too does LN
between quietest and realist versions (see De Caro, 2015), and it is Merleau-Ponty’s oeuvre that
can best assist with that second task.

Notes
1 Merleau-Ponty is an important influence on the interdisciplinary and empirically-oriented research
that has come to be called “4e” cognition – embodied, embedded, extended and enactive cognition
(c.f. Gallagher 2017).
2 Other related positions include Hutto and Satne’s “relaxed naturalism” (cf. chapter 14) and Lynne
Rudder Baker’s “near naturalism”. De Caro (2014) contends that that the suffix “near” in Baker’s
expression signals that her view is neutral regarding the possible existence of supernatural entities
(i.e. God).
3 Merleau-Ponty’s philosophy depends on a similar strategy. A neither empiricism nor intellectualism
trajectory is fundamental to Phenomenology of Perception, and a neither reflection nor intuition move
frames his methodological reflections in The Visible and the Invisible. His “indirect ontology” also
proceeds by distinguishing itself from what it is not, neither monism nor dualism (see Reynolds and
Roffe, 2018).
4 Appealing to intuition or ordinary language on such matters may not be sufficient, potentially begging
the question against the naturalist or allowing a dualism to sneak back into the picture. This is arguably
a question for versions of LN committed to a Wittgenstein quietism and/or the idea of a metaphysics-

Merleau-Ponty and liberal naturalism

5
6
7

8
9
10
11
12
13
14
15

free Lebenswelt. For phenomenology the life-world is prior to scientific idealisations but it is not itself
pre-predicative or prior to naïve metaphysics (Staiti 2018).
Husserl often makes remarks like this. See Hua VI, 208, and discussion in Overgaard (2004, 41). It is
also true that elsewhere in his oeuvre he says quite different things, for example that every transcendental analysis has a parallel analysis that might be performed in regard to the “constituted”.
“Left” Sellarsians either drop the scientia mensura principle, or construe it in more liberal fashion,
retaining a space for normativity, meaning and the first-person perspective. See Sachs (2014) and
Macarthur (2019).
It might be thought that if any philosophical position can effect a synthesis and reconciliation of the
empirical and the philosophical, the causal and the rational, it will come from the Kantian and postKantian traditions, including phenomenology. But judgments on this diverge widely. For orthodox
naturalists, like Kornblith, the fate of post-Kantian reconciliatory projects are central to their
scepticism.
At times, Husserl appears to agree with the orthodox naturalist that we should not be re-enchanting
nature (see Staiti 2016). At other times he undertakes his own attempted re-enchantment, that is, in his
critique of the post-Galilean mathematicising of nature in Crisis.
Husserl (1956, 385). See also Staiti (2016, 137).
He concludes the book by calling for a redefinition of transcendental philosophy that makes it pay heed
to the real world (Merleau-Ponty 1963, 224).
Merleau-Ponty has in mind motor-intentionality, temporality, the life-world and intersubjectivity.
Rather than directly criticise Husserl, Merleau-Ponty appeals to Husserl’s “unthought” to support his
own views.
While many think skepticism is a sign of something having gone wrong with a philosophical worldview, see Hohwy (2017) for an alternative argument.
Sellars criticises any version of the given that has a dual role, as a basic epistemic given akin to sensedata, while also being justificatory in regard to the game of giving and asking for reasons (see
Sachs 2014).
Although it was incomplete at the time of his death, his plans for The Visible and the Invisible went in
this direction, and his Nature course notes provide some indication of what this might have looked like.
This is basically what Gilbert Ryle argues about Husserl. Ryle is another figure worth co-opting
for LN.

References
Berghofer, P. and Wiltsche, H. The Co-Presentational Character of Perception. In C. Limbeck and F.
Stadler (eds.), The Philosophy of Perception and Observation. De Gruyter. (2019).
Berghofer, P. Transcendental Phenomenology and Unobservable Entities. Perspectives, 7(1), (2017). 1–13.
De Caro, M. and Macarthur, D. Naturalism and Normativity. Columbia University Press. (2010).
De Caro, M. Two Forms of Non-Reductive Naturalism. Phenomenology and Mind. Firenze University
Press. (2014).
De Caro, M. Putnam’s Liberal Naturalism. In M. Frauchiger (ed.), Themes from Putnam. Lauener Library of
Analytical Philosophy. (forthcoming)
De Caro, M. Realism, Common-Sense, and Science. Monist, 98(1), (2015). 197–214.
Gallagher, S. Enactivist Interventions. Oxford University Press. (2017).
Gallagher, S. Rethinking Nature: Phenomenology and a Non-reductionist Cognitive Science. Australasian
Philosophical Review, 2(2), (2018). 125–138.
Gardner, S. The Limits of Naturalism and the Metaphysics of German Idealism. In E. Hammer (ed.),
German Idealism: Contemporary Perspectives, pp. 19–49. Routledge. (2007).
Gardner, S. Merleau-Ponty’s Phenomenology in the light of Kant’s Third Critique and Schelling’s Realidealismus. Continental Philosophy Review, 50(1), (2017). 5–25.
Hohwy, J. The Predictive Mind. Oxford University Press. (2013).
Hohwy, J. How to Entrain Your Evil Demon. In T. Metzinger and W. Wiese (eds.), Philosophy and
Predictive Processing, Vol. 2. MIND Group. (2017).
Husserl, E. Erste Philosophie. Hua VII. Nijhoff. (1956).
Husserl, E. The Crisis of European Sciences and Transcendental Phenomenology. Trans. David Carr.
Northwestern University Press. (1970).

Jack Reynolds
Husserl, E. Ideas Pertaining to a Pure Phenomenology and to a Phenomenological Philosophy. Trans. F. Kersten.
Nijhoff. (1982).
Kim, J. Making Sense of Emergence. Philosophical Studies, 95, (1999). 3–36.
Kornblith, H. Knowledge and Its Place in Nature. Clarendon Press. (2002).
Kornblith, H. A Naturalistic Epistemology: Selected Papers. Oxford University Press. (2014).
Macarthur, D. Liberal Naturalism and the Philosophy of the Manifest Image. In A. Gare and W. Hudson
(eds.), For a New Naturalism, pp. 50–65. Telos Press Publishing. (2017).
Macarthur, D. Liberal Naturalism and the Scientific Image of the World. Inquiry: An Interdisciplinary Journal
of Philosophy, 62(5), (2019). 565–585.
Merleau-Ponty, M. Structure of Behaviour. Trans. A.L. Fisher. Beacon Press. (1963).
Merleau-Ponty, M. Signs. Trans. R. McCleary. Northwestern University Press. (1964).
Merleau-Ponty, M. Phenomenology of Perception. Trans. D. Landes. Routledge. (2013).
Merleau-Ponty, M. The Visible and the Invisible. Trans. A. Lingis. Northwestern UP. (1964).
Moran, D. ‘Let’s Look at It Objectively’: Why Phenomenology Cannot be Naturalised. Royal Institute of
Philosophy Supplement, 72, (2013). 89–115.
Morris, D. Merleau-Ponty’s Developmental Ontology. Northwestern UP. (2018).
Overgaard, S. Husserl and Heidegger on Being and World. Kluwer. (2004).
Overgaard, S. Other Minds Embodied. Continental Philosophy Review, 50(1), (2017). 65–80.
Price, H. Naturalism and the Fate of the M-Worlds. Proceedings of the Aristotelian Society, 71, (1997).
247–267.
Railton, P. Made in the Shade: Moral Compatibilism and the Aims of Moral Theory. Canadian Journal of
Philosophy, 21, (1997). 79–106.
Reynolds, J. Phenomenology, Naturalism and Science. Routledge. (2018).
Reynolds, J. Temporal Naturalism: Reconciling the “4Ms” and Points of View Within a Robust Liberal
Naturalism. Phenomenology and Cognitive Sciences, 19(1), (2000). 1–21.
Reynolds, J. and Roffe, J. Neither-Nor: Merleau-Ponty’s Ontology in “The Intertwining/The Chiasm”.
In A. Mildenberg (ed.), Understanding Merleau-Ponty, Understanding Modernism, pp. 100–114.
Bloomsbury. (2018).
Ritchie, J. Understanding Naturalism. Routledge. (2005).
Romano, C. At the Heart of Reason. Northwestern UP. (2016).
Sachs, C. Intentionality and the Myths of the Given: Between Pragmatism and Phenomenology.
Routledge. (2014).
Sellars, W. Science, Perception and Reality. Routledge. (1963).
Sider, T. Four Dimensionalism. Oxford University Press. (2001).
Smart, J. Sensations and Brain Processes. The Philosophical Review, 68(2), (1959). 141–156.
Staiti, A. The Relative Right of Naturalism: Reassessing Husserl on the Mind/Body Problem. In B. Centi
(ed.), Tra corpo e mente: Questioni di confine, pp. 125–150. Le Lettere. (2016).
Staiti, A. Pre-Predicative Experience and Life-World: Two Distinct Projects in Husserl’s Late
Phenomenology. In D. Zahavi (ed.), Oxford Handbook of the History of Phenomenology, pp. 155–172.
Oxford University Press. (2018).
Thompson, E. Mind in Life. Harvard University Press. (2010).
Varela, F. Neurophenomenology: A Methodological Remedy to the Hard Problem. Journal of
Consciousness Studies, 3(4), (1996). 330–349.
Wiltsche, H. What Is Wrong with Husserl’s Scientific Anti-Realism? Inquiry, 55(2), (2012). 105–130.
Zahavi, D. Review of Evan Thompson’s Mind in Life. Husserl Studies, 25(2), (2009). 159–168.
Zahavi, D. Brain, Mind, World: Predictive Coding, Neo-Kantianism, and Transcendental Idealism.
Husserl Studies, 34(1), (2018). 47–61.

8
CLASSICAL PRAGMATISM AND
LIBERAL NATURALISM
Steven Levin

1 Introduction
In this chapter I examine the relationship between classical pragmatism and liberal naturalism.
While it is clear that all three of the classical pragmatists, Charles S. Peirce (1839–1914),
William James (1842–1910) and John Dewey (1859–1952), are naturalists in one sense or
another, it is an open question whether the pragmatist’s brand of naturalism is a form of liberal
naturalism.1
To properly assess the pragmatist’s naturalism we must first put it in the context of their
engagement with the development of evolutionary theory in the second half of the nineteenth
century.2 All of the classical pragmatists took evolutionary theory to open up a new epoch in
philosophy and intellectual life generally, having deep and pervasive consequences for both
theory and practice. In sections two and three I lay out some of these consequences. I then, in
section four, compare pragmatic naturalism and contemporary scientific naturalism. In section
five I put pragmatic and liberal naturalism into relation, and then, in section six, I isolate the
main factors by which they are distinguished.

2 Theoretical consequences of evolutionary theory
For all three of the classical pragmatists, evolutionary theory gives us a new picture of nature,
one where it can no longer be understood as tending toward fixed and final ends, as the classical
tradition posits, nor as a well-regulated clock-like mechanism, as the Enlightenment posits.
Both of these views accept a dualistic picture of nature, one in which physical changes are seen
either as expressing hidden and unchanging essences, or as governed by timeless and perfectly
balanced natural laws.
Evolutionary theory, in contrast, gives us a single-level picture of nature in which nothing
timeless is seen as standing behind its manifest changes. Nature is no longer understood as
economical, symmetrical, orderly and benevolent, but as mutable, contingent, wasteful and
overproductive. While we can distinguish the relatively stable aspects of nature from those that
are changing, stability is itself an evolutionary development of a mutable and unstable world.
This is the upshot of Peirce’s synechist and tychist metaphysics, where continuity and chance
are seen as irreducible yet interdependent features of nature.3 According to this metaphysics, the
DOI: 10.4324/9781351209472-10

Steven Levine

laws of nature are not timeless regularities but “habits” that themselves evolve through time.
Like bodily habits, natural laws have stability, but just as bodily habits can be modified when a
changing environment requires new patterns of behavior, so too the laws of nature can (and do)
change in the face of the contingency and chance found in the universe.
Dewey also argues, in his naturalistic metaphysics, that existence is marked by chance and
continuity. It is, in his terms, both precarious and stable.4 While natural existence is marked by
chance transactions and changes, it is also continuous in the sense that none of these transactions
or changes can be explained by the “appearance upon the scene of a totally new outside force”
(LW 12, 31). In light of this, Dewey outlines a “principle of continuity”, which “excludes
complete rupture on the one side and mere repetition of identities on the other; it precludes
reduction of ‘higher’ to the ‘lower’ just as it precludes complete breaks and gaps” (LW 12, 30).
This has significant ontological ramifications. If one rejects the reduction of the higher to the
lower across the board, then the really real is not to be identified with a single level of being, the
microphysical for instance. Dewey instead espouses an evolutionary pluralism in which matter,
life and mind are distinct yet continuous “plateaus of existence”, which are all equally real.5
Dewey’s pluralism has limits, however, for if there are no complete breaks and gaps in nature, if
there are no genuine ruptures, then there is no space for extra- or supernatural items, items that
are not continuous with the rest of nature. This bears on how Dewey’s pragmatic naturalism
deals with the so-called placement problems, the problems of placing items that seem recalcitrant to naturalistic explanation – consciousness, intentionality, meaning, value, etc. – in
nature. Such items are in fact part of the continuous continuum of nature and admit of being
explained in nonreductive yet naturalistic terms.6
This altered ontology necessitates a new epistemology and logic. On the pre-Darwinian
account, knowledge could only be ascertained through the rational apprehension of unchanging essences. In light of this, the modes by which we apprehend the changing natural
world, that is, our sensing and manipulating it, were either dismissed altogether (rationalism), or
thought to only put us into contact with “mere appearances” (empiricism). But if there are no
underlying essences to apprehend, if nature exists on a single level, then, the pragmatists
conclude, something is what it does, and coming to know this something will be achieved by
grasping the manifest patterns of its behavior. But how is this knowing to be acquired?
For the pragmatists, unlike contemporary philosophers, this question is not primarily answered by epistemology, understood as a standalone discipline, but by logic. This is because for
them logic concerns not only deductive inference, but also, and more importantly, the synthetic
inferences at play in inquiry and discovery.7 Logic in its synthetic aspect is therefore not sealed
off from the development of the empirical sciences; rather, “each chief step in science has been
a lesson in logic” (Peirce 1992a, 111). Evolutionary theory has been a lesson in logic because it
introduces the idea that to grasp an unknown something requires not subsuming it under a
universal essence or law, but rather giving a “historical” account of the development of its
behavior, an account written in the language of probability and statistics. This historical account
is filled out not by our intuiting an object’s timeless essence, nor by our passively perceiving its
properties, but by our undertaking a temporally extended inquiry. The knowing that is achieved
through inquiry is not second-rate, a grasp of mere “appearances”, but nor is it an indubitable
knowing of something that cannot be otherwise than it is.
At the most general level, inquiry is an attempt to replace doubt with belief, to arrive at a
belief we have no reason to doubt.8 To arrive at such a belief – which for the pragmatists is a
stable disposition to act – we must undertake an experimental action-cycle in which we develop, through abductive inference, a hypothesis about the consequences of interacting with an
object, act in light of that hypothesis and observe the consequences of so acting. This
84

Pragmatism and liberal naturalism

observation may accord with our hypothesis, leading us to exit the experiment by inductively
forming generalizations about the object – generalizations that shape our beliefs and determine
the nature of our future interaction with the object – or it may lead, if unsuccessful, to a new
cycle of inference, action, perception and inference. Here, to paraphrase Kant, we put nature to
the question by instituting a course of directed action in light of hypothetical and conditional
ideas. While knowledge on this picture is attainable, it is, like scientific inquiry, fallible, potentially requiring revision in light of altered and recalcitrant experience.

3 Practical consequences of evolutionary theory
While changes in our theoretical picture of the world brought about by evolutionary theory were
substantial, the changes in practical life were even more significant. While the sciences of the early
modern period had already disenchanted the picture of nature in which it contains qualities and
final ends – locating them in a subjectivity that projects them onto nature – the evolutionary view
of nature thoroughly delegitimized the natural theologies that, post-Enlightenment, continued to
support the idea that nature is an orderly whole fashioned by a transcendent designer. Besides
undermining the providential picture of the universe provided by these theologies, evolutionary
theory also undermined the conception we have of ourselves as free and spiritual beings. For
instance, Spencer – who was as influential on the pragmatists as Darwin himself – argued that the
mind and its powers are completely explainable by the psychological laws of association, which
are themselves the product of evolution. Insofar as the human mind is a purely passive product of
natural processes, it lacks all freedom and spontaneity.9
In the latter half of the nineteenth century these ideas provoked a widespread moral and
spiritual crisis, one well represented biographically by James’s famous mental breakdown.10
James’s philosophical response to this crisis is most clearly found in his book Pragmatism. There,
James attempts to do justice to two intellectual temperaments found in late modernity, which
he calls the “tough-minded” and the “tender-minded”. While these temperaments have a
perennial quality, the tough-minded temperament has become far more prevalent in the wake
of the Early Modern and Darwinian scientific revolutions. We are now, James says, “almost
born scientific” (James 1979a, 14). According to this temperament:
Man is no law-giver to nature, he is an absorber. She is who stands firm; he it is who
must accommodate her. Let him record truth, inhuman though it may be, and submit
to it! The romantic spontaneity and courage are gone, the vision is materialistic and
depressing. Ideas appear as inert by-products of physiology; what is higher is explained
by what is lower and treated forever as a case of ‘nothing but’ – nothing but something
else of a quite inferior sort. You get, in short, a materialistic universe.
(James 1979a, 15)
The tender-minded, who wish to maintain the importance of moral value, freedom and the
transcendent, recoil from this conclusion, feeling that these items can find no place in a materialistic universe. They are either seen as nothing but something else, that is, reduced to
something else, or eliminated altogether. But most persons, James agues, have mixed temperaments, wanting “a system that will combine…the scientific loyalty to facts and willingness
to take account of them…but also the old confidence in human values and the resultant
spontaneity” (James 1979a, 17). Pragmatism, he argues, is the philosophical view that can satisfy
both of these wants, and in so doing ameliorate the crisis brought about by Darwinian
revolution.
85

Steven Levine

For James, pragmatism is a semantic thesis, a thesis about the meaning of our concepts. It is
based in the pragmatic maxim, which Peirce in his paper “How to Make our Ideas Clear” defines
in this way: “Consider what effects, that might conceivably have practical bearing, we conceive
the object of our conception to have. Then, our conception of these effects is the whole of our
conception of the object” (Peirce 1992a, 132). The maxim says that if one makes explicit one’s
conception of the effects of an object on other things and oneself, a conception which is formed
by our perception of the object and our acting upon it, then one will have made the content of
the concept of that object as clear as it can be.11 James interprets this doctrine as saying that the
meaning of a concept of an object is cashed out by the consequences the concept’s use has for our
lives, that is, the perceptions and actions that its use brings about. If the use of a concept has such
consequences, if it makes a difference in one’s experience, then it has pragmatic meaning. So while
the application of the pragmatic maxim shows that certain concepts are spurious by not making
any difference in experience, it also shows that certain concepts that we may have taken to be
spurious in fact have meaning. For instance, James concludes that value concepts, like the concepts of the secondary qualities, are not “inert by-products” of our physiological responses to
things, but genuinely meaningful. While these concepts and their objects are “subjective”, they,
in having consequences for our experience, are real and not second-rate. We must attend to
experience and show scientific loyalty to the facts, as the tough-minded say, but we must do
justice to all the facts that are experienced, including those that the tough-minded wish to
eliminate. This is the path to meeting our disparate temperamental needs.
Dewey was also centrally concerned with the moral and spiritual crisis brought about by
development of the natural sciences, a “crisis” due to the perceived “incompatibility between
the conclusions of natural science about the world in which we live and the realm of higher
values, of ideal and spiritual qualities” (LW 4, 33). His strategy to meet this crisis is different
from James’s, however, because, like Peirce, he does not interpret the pragmatic maxim as
establishing the genuineness of a concept merely through its having consequences for an individual subject’s life and experience. He instead advocates for the compatibility of natural
science and the realm of value by arguing that the mathematically based natural sciences, physics
and chemistry, concern certain kinds of relations, whereas the moral sciences concern qualities, of
which value is a sub-species. While grasp of relations emerges through a process of theoretical
abstraction, qualities and values are only ascertainable within experience. But this does not
make them “subjective” – because experience for Dewey is nothing subjective.
Experience for Dewey is not what it was for the classical empiricists, namely, “a name for the
ensemble of Cartesian cogitationes, Lockean ideas” (Rorty 1979, 150). While Dewey, like Peirce
and James, takes himself to be a kind of empiricist, it is not because he thinks that all concepts
have their origin in sense-impressions. It is rather because he takes experiment, creatively acting
beyond the given, to stand at the center of experience. Experience in its most basic sense is, for
Dewey, a temporally extended learning process through which an organism interacts with an
environment through simultaneously acting on and being acted on by it. It “is a matter of
simultaneous doings and sufferings” (MW 10, 8). Experience is a learning process because
through acting on the world and undergoing the consequences of such acting, agents come,
through time, to reflectively grasp the connections that exist between doing and suffering with
respect to an object, a grasp that feeds back into their system of habits and bodily skills, preparing the way for subsequent and enriched experience.
How does this bear on Dewey’s account of quality and value? The point is simple: an
organism’s experiential interaction with the environment is itself within nature, and phenomena
that recur owing to this process, phenomena like something being red, frightening, joyful or
good, are themselves as natural as anything else. It is only the pervasive subjectivism of modern
86

Pragmatism and liberal naturalism

philosophy’s account of experience, itself a compensatory response to an objectivism that drains
nature of quality and value, which blocks us from grasping this simple point.
To sum up, we can say that the pragmatists attempt to undermine the felt tension between
natural science, value, and free human thought and agency by arguing that philosophers and
intellectuals in general have not properly understood the importance of the Darwinian revolution, incorrectly reading into it the mechanistic determinism of Enlightenment science and
thought. The pragmatists argue that value and free human thought and agency are not only
compatible with the evolutionary picture of nature but are in fact made possible by it. It is only
because the universe is open, unfolding and uncertain that it is a field that can be changed and
modified by human thought and action in light of reflectively chosen values and goods. The
Darwinian conception of nature therefore does not lead to nihilism, as the tough-minded think,
but to a melioristic vision in which human action, itself risky and uncertain, can, if intelligent,
bring about conditions in which human flourishing is possible.

4 Scientific and pragmatic naturalism
Now that we have outlined the classical pragmatist’s naturalism, we need to give an account of
its proximity to and distance from liberal naturalism. Before doing so, however, it would be
helpful to briefly compare pragmatic naturalism and contemporary scientific naturalism.
It is common to distinguish between two forms of scientific naturalism, one ontological, the
other methodological. Ontological naturalism argues that all genuinely real things are either
identical to spatiotemporally located physical entities, or at least supervenient upon them. What
drives the position is the idea, formed through the progress of natural science, that only physical
entities can have causal effects on others things.12 Based on this, the scientific naturalist concludes that mental phenomena or social phenomena, insofar as they have causal effects, must be
identical to arrangements of physical entities, or supervenient upon them. Given this, the
ontological naturalist holds that the common-sense objects of the manifest image – that is,
values, meanings, numbers, etc. – do not “limn the true nature of reality”. This sets the stage for
contemporary projects of naturalization, which attempt to give reductive, eliminativist or
nonfactualist accounts of these objects.
Methodological scientific naturalism is directly concerned not with what there is, but with
the methods by which we gain knowledge of what there is. Its main claim is that genuine
knowledge of the world results either from inquiries that utilize the methods of the successful
natural sciences, or from inquiries that are informed by these methods. This rules out, as an
independent source of knowledge, introspection, intuition and, most importantly, a priori
knowledge. Given that philosophy was traditionally taken to be an a priori science whose
concepts ground the concepts of the other sciences, this has ramifications for the place of
philosophy in the system of knowledge. For if, as Quine famously argued, we can no longer
systematically distinguish between a priori and a posteriori truths, then we are no longer in a
position to identify truths that are purely philosophical. As a result, the methodological scientific naturalist blurs the line between philosophy and empirical science.13
How do the classical pragmatists stand with respect to these two forms of naturalism? As we
saw above, the pragmatists reject any reductive ontological program. They embrace an evolutionary ontological pluralism in which nature is varied and mutable yet continuous. They do,
however, accept a form of methodological naturalism.14 But which exact form?
In his famous paper “The Fixation of Belief ”, Peirce distinguishes the scientific method
from three insufficient methods of belief fixation: the method of tenacity, authority and the a
priori method. The scientific method is sufficient to fix belief, unlike these other methods,
87

Steven Levine

because it fixes belief in light of recalcitrant experience rather than in light of what one wants to
believe (the method of tenacity), what an authority believes (the method of authority), or what
it is agreeable to reason to believe (the a priori method). While the three nonscientific methods
in fact fix belief – the method of authority being the prime method of belief fixation
throughout history – such methods are insufficient because the beliefs produced by them are
subject to being dislodged when confronted with contrary beliefs from other persons, cultures
or, in the case of the a priori method, other philosophers. This is not the case with the method
of science, however, because it aims to have our beliefs answer to the way things are rather than
to other attitudes and beliefs. The pragmatists are methodological naturalists in the sense that
they think that any inquiry, for it to produce beliefs that we have no reason to challenge, must
be informed by the scientific method.
This method is more an attitude or spirit of belief-fixation than any kind of specialized
procedure. “That which is essential” for science, Peirce says, “is the scientific spirit, which is
determined not to rest satisfied with existing opinions, but to press on to the real truth of
nature” (Peirce 1935, 428). Expanding on the point, Dewey says:
On its negative side, it is freedom from control, routine, prejudice, dogma, unexamined tradition, sheer self-interest. Positively, it is the will to inquire, to examine,
to discriminate, to draw conclusions only on the basis of evidence after taking pains to
gather all the evidence. It is intention to reach beliefs, and to test those that are
entertained, on the basis of observed fact, recognizing also that facts are without
meaning save as they point to ideas. It is, in turn, the experimental attitude which
recognizes that while ideas are necessary to deal with facts, yet they are working
hypotheses to be tested by the consequences they produce. Above all, it is the attitude
which is rooted in the problems that are set and questions that are raised by conditions
of actuality. The unscientific attitude is that which shuns such problems, which runs
away from them, or covers them up instead of facing them. (LW 13, 273)
This attitude informs not only the work of those we call scientists, but that of any “body of
persons who deal intelligently and openly with the objects and energies of the common environment” (LW 13, 271). Scientific inquiry for the pragmatists is any self-correcting enterprise
undertaken by inquirers who have the right attitude or spirit of belief-fixation.
Inquiry that adheres to the scientific method requires more than this, however, for inquiry, as
we mentioned above, has a kind of logical structure. But this logical structure is so general that it
does not entail any given procedure for a science. Cultural anthropology, for example, satisfies the
logical criteria for being a science. One can undertake inquiry in cultural anthropology (fieldwork)
by engaging in an action cycle involving quite complex interactions between doing, observing and
interpreting, which involves abductive, inductive and deductive inferences, and this inquiry can be
informed by the correct attitude of belief-fixation. This, clearly, has very little to do with the
nature of contemporary research in physics, yet both are examples of scientific inquiry for the
pragmatists. Going beyond this, Dewey famously argues that we are able to undertake inquiry in
the value spheres of ethics and politics.15 While the scientific attitude informs these inquiries, and
they adhere to the general logical form of inquiry, they have quite different contents and procedures than inquiries in the natural sciences. In this sense, the pragmatists are method pluralists,
even thought any inquiry, to be science, must involve the scientific attitude.
How does this bear on the pragmatist’s view of the relationship between philosophy and
science? The pragmatists are committed to bringing an experimentalist attitude into philosophy,
rejecting its image of itself as an a priori science, and they are committed to the idea that
88

Pragmatism and liberal naturalism

philosophy is itself a form of inquiry. Philosophy must be informed by the results of the other
sciences, and its results must be able to feed back into the sciences. There is a kind of circular
relationship between philosophy and the other sciences, not an asymmetrical grounding relationship. In line with their method pluralism, they take it that there are many forms of
philosophical inquiry. The classical pragmatists engage directly in natural scientific forms of
inquiry, historical genealogy, therapeutic philosophy, abductive metaphysics, conceptual and
classificatory analysis, and phenomenological analyses of lived experience. All of these forms of
philosophy are legitimate insofar as they potentially advance inquiry.

5 Pragmatic and liberal naturalism
Let us now take up the question of the relationship between pragmatic and liberal naturalism.
Liberal naturalism is the thesis that there are facts that are natural, and so not supernatural, which
nonetheless are not explicable by the natural sciences.16 Liberal naturalists often focus on
normative facts, for example, the fact that a belief is justified by another belief given a rule of
reasoning, or the fact that an action is right given a standard of conduct. The liberal naturalist
holds that these normative facts are, in McDowell’s terms, sui generis.17 Such facts are not
explicable by the explanatory schemas of the natural sciences, because such schemas are concerned with what generally happens, not with what ought to happen given a rule or standard.
Such facts are not supernatural, because to find them so depends on endorsing the prior thesis
that nature is equivalent to what is explicable by the natural sciences. But this thesis is a metaphysical thesis, one that the liberal naturalist contests.
There are two major liberal naturalist strategies to contest this thesis, one quietist, the other
quasi-transcendental. The quietist strategy, most famously given by McDowell, says that the
liberal naturalist does not need to demonstrate that normativity is sui generis, but only to remind us
of the obvious fact that normativity is an element of our developing second nature. Nature
includes more than what is subject to natural scientific explanations; it includes our acquisition
of a second nature. It is natural for creatures like us to develop a responsiveness to reasons and
for this to inform our thought and action. It is the forgetting of this obvious fact that makes it
seem as if normativity must be explained either in reductive naturalistic terms or be seen as
something outside of nature.18
The pragmatic naturalist rejects the quietist strategy, for two reasons. First, reminders of the
obvious are not enough to break the grip of scientific naturalism, because what is obvious to the
liberal naturalist is not necessarily obvious to those who are not. For example, the toughminded think that intellectual integrity requires them to courageously face up to the disenchanted world that has been detailed for us by the natural sciences. It’s not that they have
forgotten the concept of second nature and need to be reminded of it; it’s rather that they
actively reject the concept, given their background understanding of how things are and must
be. Given this understanding, a reminder that second nature is natural is simply not a reminder
about something that is obvious. Second, to not address the mystery of how the normativity that
suffuses our second nature is related to first nature leaves open an explanatory gap that has been,
and will continue to be, filled in by the reductive natural-scientific explanatory schemas that
liberal naturalism means to replace. Intellectual life, like nature, abhors a vacuum, and to not fill
this gap correctly opens the door for it to be filled incorrectly.
McDowell argues that this gap cannot be filled, because normative facts cannot be grasped
from “the sideways-on”, that is, from the third-person point of view assumed by the natural
sciences. From this explanatory perspective, which is foreign to the perspective of the selfconscious subjects who operate within the space of normative facts, the so-called space of
89

Steven Levine

reasons, the rational connections that comprise this space disappear. Dewey agrees with
McDowell that we must take the subject’s point of view very seriously. As we saw above,
quality and value, while not subjective, nevertheless appear from within this point of view. But
Dewey would question McDowell’s assumption that there is a dichotomy between this point of
view and the sideways-on view. For there are, according to Dewey, two kinds of sideways-on
views, those that are completely detached from the point of view of self-conscious subjects, for
instance chemistry and physics, and those that are not, for instance biology and ethology in
certain of their modalities, and anthropology, developmental psychology, history, sociology,
etc. While the point of view taken by the latter sciences is not directly based in the everyday
experience of an individual self-conscious subject, their insights are still based in experience, in a
collective learning process that has our form of life as its object. Unlike chemistry or physics, we
are both the subject and object of these sciences, and their goal is to help us become selfconscious about the kind of natural yet rational creature we are. From the pragmatist’s perspective, McDowell’s liberal naturalism is shaped by a dichotomy between subject and object
that is a remnant of the transcendental tradition.19
This leads us to the quasi-transcendental strategy. This strategy argues that natural science
itself calls upon normative resources in its explication of things that cannot be naturalized.
While this argument is found in non-naturalists like Kant and Husserl, one can find versions of
it in certain liberal naturalists, for instance, Putnam.20 Putnam endorses the so-called indispensability arguments, which he thinks have their origin in the classical pragmatists.21
According to this kind of argument, “normative discourse – talk of right and wrong, good and
bad, better and worse – is indispensable in science and in social and personal life as well”
(Putnam, 1994, 154). Normative discourse is indispensable for science because in developing
theories and guiding practice we continually use concepts like “simple”, “elegant”, “plausible”,
“reasonable”, etc. Scientific inquiry is, as Sellars puts it, “fraught with ought”.22 If this is so,
then the scientific naturalist is in a bind, for they can either admit that natural science has
supernatural conditions of possibility, or they can follow the liberal naturalist and hold that
normativity is nothing outside of nature. Putnam argues for the latter conclusion by claiming
that the fact that normative concepts are indispensable is not a metaphysical necessity, but a
contingent fact about how our practices have come to be structured over time. We find that
these concepts are indispensable for our practice, and if we take the “primacy of practice”
seriously, then we cannot eliminate them from our image of the world and ourselves.
Pragmatic naturalists are of two minds about the quasi-transcendental strategy. On the one
hand, Putnam is right that the classical pragmatists, because they accept the primacy of practice,
accept indispensability arguments. But, on the other hand, if the pragmatists are not to violate
their first rule of reasoning to “not block the way of inquiry”, they must take such arguments to
be merely provisional.23 While the pragmatic naturalist agrees with the liberal naturalist that
normativity is indispensable for the natural sciences as well as everyday life, they wish to leave
open the question of whether a naturalist account of normativity is possible. Given the centrality of continuity to their view, we can anticipate their answer to this question: we will
eventually be able to give a naturalist yet nonreductive account of normativity.

6 Conclusion
In thinking that it can give a naturalistic account of the central concepts of theoretical and
practical philosophy, an account that eschews both reductionism and supernaturalism, while
also avoiding the moral crises that naturalistic accounts of these concepts are thought to bring
about, pragmatic naturalism is clearly more in proximity to liberal naturalism than to the various
90

Pragmatism and liberal naturalism

scientific naturalisms that circulate in contemporary philosophy. However, as we have seen,
there are also significant differences between pragmatic and liberal naturalism.
From the pragmatic point of view, the main difference stems from the fact that the liberal
naturalist mostly accepts, rather than contests, the scientific naturalist’s image of nature and
science. The liberal naturalist, for instance, does not think that their argument that nature
includes our second nature requires a fundamental rethinking of first nature.24 We can keep the
scientific naturalist account of first nature in place while simply reminding ourselves that it is
incomplete as an account of nature as a whole. The pragmatist, in contrast, thinks we must give
a fundamentally different account of nature than the scientific naturalist, who still, the pragmatist thinks, has a kind of pre-evolutionary view. We do not merely need to expand our
picture of nature to include second nature; we need to reconceive of nature in such a way that
value, meaning, reason and freedom can be seen as emergent features of it.
With respect to science, the liberal naturalist argues that science, while powerful in its
domain, has important limits. There are facts central to our lives, normative facts that cannot be
made the object of science.25 But in making this argument the liberal naturalist does not
fundamentally contest the image of science given by the scientific naturalist. They take over this
image and then say that, so understood, it has limits. The pragmatist, on the other hand,
challenges the view of science typically found in scientific naturalist accounts. On their account
of inquiry, many objects that escape the net of science for both the scientific and liberal naturalists, for example, value and meaning, in fact fall within it. Perhaps there are objects that in
principle escape the pragmatist’s capacious account of science. But before definitively answering
this question one must have the right account of science. From the point of view of the classical
pragmatist, this is precisely what the liberal naturalist does not have.

Notes
1 There are significant philosophical differences between the classical pragmatists. In this paper, however, I try to develop a unified account of pragmatic naturalism. I point out some of their major
differences in the notes.
2 See Weiner (1949) and Pearce (2019a, 2019b).
3 See the papers in Peirce’s Monist Metaphysical Series, in Peirce (1992a). For more on this and Peirce’s
naturalism generally, see Dea and Haydon (2019).
4 See Chapter 2 of Experience and Nature (LW 1). All references to Dewey are to The Collected Works of
John Dewey 1882–1953 (Early Works, Middle Works, and Late Works).
5 For Dewey’s emergentist evolutionary pluralism, see Experience and Nature (LW 1). For more on this,
see Bernstein (forthcoming). For James’s pluralism, see James (1981, Chap. 21) and James (1977). Also
see Dupré (1993) for a contemporary biologically informed pluralism.
6 For Dewey this conclusion applies to God as well. It is not clear whether Peirce and James agree. In
The Will to Believe and Other Essays in Popular Philosophy (James 1979b), James famously argues for the
right to believe in God, but not directly for the existence of a transcendent God. In A Pluralistic
Universe he posits a God, but he takes it that God is a finite feature of the universe rather than a
transcendent being. So it is not clear that James posits anything not in some way continuous with the
rest of nature. Peirce, in contrast, argues that a genuine pluralism must make space for the possibility of
the transcendent, but his argument for this is abductive and fallible. For more, see Anderson (2004).
7 While Peirce and Dewey agree on this point, Peirce accuses Dewey’s logic of being psychologistic. For
an analysis of this criticism, see Sleeper (1986).
8 For summary pragmatic accounts of inquiry, see Peirce’s “The Fixation of Belief ” in Peirce 1992a and
Dewey’s “The Pattern of Inquiry” in LW 12.
9 For the relationship between the pragmatists and Spencer, see James’s “Remarks on Spencer’s
Definition of Mind as Correspondence” in James (1978) and Godfrey-Smith (1996).
10 See Perry (1935).

Steven Levine
11 To be precise, one will have elucidated the concept’s “pragmatic” meaning, which for Peirce is one of
three forms of meaning, the others being denotative and connotative meaning.
12 See Papineau (2007).
13 See “Two Dogmas of Empiricism” in Quine (1953). For more on this, see Friedman (1997).
14 For pragmatism and methodological naturalism, see Gava (2019) and Kitcher (2018).
15 Here we find another significant difference between Peirce and Dewey, as Peirce argues that genuine
inquiry can only be found in theoretical and not practical contexts. See Peirce (1992b). For an interpretation of Peirce in which his view allows for inquiry in these domains, see Misak (2004).
16 For the variety of liberal naturalist views, see the essays in De Caro and Macarthur (2004, 2010).
17 See McDowell (1996).
18 See Chapter 4 of McDowell (1996).
19 For more on this, see Godfrey-Smith (2010) and chapter six of Levine (2019).
20 One can find this kind of argument in another liberal naturalist, Habermas. See Habermas (2003).
21 See Misak (2011).
22 See Putnam (2004) for his version of this idea.
23 For this rule, see Peirce (1992b).
24 See McDowell (2000).
25 For a liberal naturalist argument that this conclusion ultimately applies to the human as well as the
natural sciences, see Macarthur (2010).

References
Anderson, D. Peirce’s Common Sense Marriage of Religion and Science. In C. Misak (ed.), The Cambridge
Companion to Peirce. Cambridge University Press. (2004).
Bernstein, R.J. Pragmatic Naturalism: John Dewey’s Living Legacy. Graduate Faculty Philosophy Journal.
(2020).
Dea, S. and Haydon, N. From the Experimentalist Disposition to the Absolute: Peirce’s Pragmatic
Naturalism. In P. Giladi (ed.), Responses to Naturalism: Critical Perspectives from Idealism and Pragmatism.
Routledge Press. (2019).
De Caro, M. and Macarthur, D. (eds.), Naturalism in Question. Harvard University Press. (2004)
De Caro, M. and Macarthur, D. (eds.), Naturalism and Normativity. Columbia University Press. (2010).
Dewey, J. The Collected Works of John Dewey 1882–1953. Southern Illinois University Press. (1969–1990).
Dupré, J. The Disorder of Things: Metaphysical Foundations of the Disunity of Science. Harvard University
Press. (1993).
Friedman, M. Philosophical Naturalism. Proceedings and Addresses of the American Philosophical Association,
71(2), (1997). 7–21.
Gava, G. Peirce and Methodological Naturalism. In P. Giladi (ed.), Responses to Naturalism: Critical
Perspectives from Idealism and Pragmatism. Routledge Press. (2019).
Godfrey-Smith, P. Complexity and the Function of Mind in Nature. Cambridge University Press. (1996).
Godfrey-Smith, P. Dewey, Continuity, and McDowell. In M. De Caro and D. Macarthur (ed.),
Naturalism and Normativity. Columbia University Press. (2010).
Habermas, Jürgen Truth and Justification. MIT Press. (2003).
James, W. The Principles of Psychology. Harvard University Press. (1981).
James, W. A Pluralistic Universe. Harvard University Press. (1977).
James, W. Pragmatism. Harvard University Press. (1979a).
James, W. The Will to Believe and Other Essays in Popular Philosophy. Harvard University Press. (1979b).
James, W. Essays in Philosophy. Harvard University Press. (1978).
Kitcher, P. Deweyan Naturalism. In M. Bagger (ed.), Pragmatism and Naturalism: Scientific and Social Inquiry
After Representationalism. Harvard University Press. (2018).
Levine, S. Pragmatism, Objectivity, and Experience. Cambridge University Press. (2019).
Macarthur, D. Taking the Human Sciences Seriously. In M. De Caro and D. Macarthur (eds.), Naturalism
and Normativity. Columbia University Press. (2010).
McDowell, J. Mind and World. Harvard University Press. (1996).
McDowell, J. Response to Gubeljic, Link, Muller, and Osburg. In M. Willaschek (ed.), John McDowell:
Reason and Nature. Lit Verlag. (2000).

Pragmatism and liberal naturalism
Misak, C. Peirce on Vital Matters. In C. Misak (ed.), The Cambridge Companion to Peirce. Cambridge
University Press. (2004).
Misak, C. American Pragmatism and Indispensability Arguments. Transactions of the Charles S. Peirce Society,
47, (2011). 261–273.
Papineau, D. Naturalism. Stanford Encyclopedia of Philosophy, http://plato. stanford.edu/entries/naturalism/.
(2007).
Pearce, T. Dewey, Darwinism, and Directed Variation. In A. Daratos and P. Walter (eds.), Penser
l’évolution: Nietzsche, Bergson, Dewey. Vrin. (2019a).
Pearce, T. James and Evolution. In A. Klein (ed.), The Oxford Handbook of W. James. Oxford University
Press. (2019b).
Peirce, C.S. Collected Papers of Charles Sanders Peirce, Vol. 6. C. Hartshorne, P. Weiss, and A. Burks (eds.).
Harvard University Press. (1935).
Peirce, C.S. The Essential Peirce, Volume 1: Selected Philosophical Writings 1867–1893. N. Houser and J.K.
Kloesel (eds.). Indiana University Press. (1992a).
Peirce, C.S. Reason and the Logic of Things: The Cambridge Conferences Lectures of 1898. K.L. Ketner (ed.).
Harvard University Press. (1992b).
Perry, R.B. The Thought and Character of W. James, Vol. 1 and 2. Oxford University Press. (1935).
Putnam, H. Words and Life. Harvard University Press. (1994).
Putnam, H. The Collapse of the Fact/Value Dichotomy and Other Essays. Harvard University Press. (2004).
Quine, W.V.O. Two Dogmas of Empiricism. In From a Logical Point of View. Harvard University
Press. (1953).
Rorty, R. Philosophy and the Mirror of Nature. Princeton University Press. (1979).
Sleeper, R.W. The Necessity of Pragmatism: John Dewey’s Conception of Philosophy. University of Illinois
Press. (1986).
Weiner, P. Evolution and the Founders of Pragmatism. Harvard University Press. (1949).

PART II

Theoretical cousins
of liberal naturalism

9
QUINE’S NATURALISM: NEITHER
“REDUCTIVE” NOR “LIBERAL”
Gary Ebbs

“A doctrine is not judged at all until it is judged in its best form”.
– J.S. Mill
W.V. Quine describes his naturalism as “the recognition that it is within science itself, and not
in some prior philosophy, that reality is to be identified and described”.1 Many readers take
Quine’s use of “science” here to encompass only the natural sciences and discourses we can
reduce to or explain in terms of the natural sciences. This reading seems well supported by some
of Quine’s own explanations of his views. For instance, when summarising his critique of the
methodology of the idioms of propositional attitude and his closely related thesis of the indeterminacy of translation, respectively, Quine writes,
If we are limning the true and ultimate structure of reality, the canonical scheme for us
is the austere scheme that knows no quotation but direct quotation and no propositional attitudes but only the physical constitution and behavior of organisms.2
[T]heory in physics is an ultimate parameter. There is no legitimate first philosophy,
higher or firmer than physics, to which to appeal over physicists’ heads...the indeterminacy of translation is not just inherited as a special case of the underdetermination of our theory of nature. It is parallel but additional.3
Such passages suggest that Quine’s naturalism is the thesis that it is only within the established
natural sciences that reality is to be identified and described. I shall call this the standard reading
of Quine’s naturalism.
Philosophers who accept the standard reading typically dismiss Quine’s naturalism as “reductive” and espouse more “liberal” naturalisms according to which our ordinary discourses
about propositional attitudes, meaning and translation are naturalistic and respectable despite
not being reducible to natural science.4
My central goal here is to show that the standard reading of Quine’s naturalism, though
superficially plausible, is incorrect. Quine’s naturalism, I shall argue, is neither “reductive” nor
“liberal”. My goal is not to defend it, but to explain it on its own terms, in hopes of increasing
the likelihood that it will eventually be evaluated in its best form.
DOI: 10.4324/9781351209472-12

Gary Ebbs

1 Working from within
The most obvious and basic problem with the standard reading of Quine’s naturalism is that
Quine says repeatedly, in many different ways, that he uses “science” broadly to include both
the “hard sciences” and “the softer sciences, from psychology and economics through sociology
to history”.5 “We have no word with the breadth of Wissenschaft”, he writes, “but that is what I
have in mind”.6 Quine inherits this encompassing use of “science” from his mentor Rudolf
Carnap, who, in a paper about the unity of the sciences, writes
We use the word ‘science’ here in its widest sense, including all theoretical knowledge,
no matter whether in the field of natural sciences or in the field of the social sciences and
the so-called humanities, and no matter whether it is knowledge found by the application of special scientific procedures, or knowledge based on common sense in everyday life…What usually is called science is merely a more systematic continuation of
those activities which we carry out in everyday life in order to know something.7
In similar spirit, Quine writes:
We imbibe an archaic natural philosophy with our mother’s milk. In the fullness of
time, what with catching up on current literature and making some supplementary
observations of our own, we become clearer on things. But the process is one of growth
and gradual change: we do not break with the past, nor do we attain to standards of
evidence and reality different in kind from the vague standards of children and laymen.8
What Quine describes here as “the process…of growth and gradual change” is what he elsewhere calls the method of “working from within”.9 It is this method that Quine highlights
when he writes:
Neurath has likened science to a boat which, if we are to rebuild it, we must rebuild
plank by plank while staying afloat in it. The philosopher and the scientist are in the
same boat.10
The sense in which the naturalistic philosopher and the scientist are in the same boat, according
to Quine, is just that, like the scientist,
The naturalistic philosopher begins his reasoning within the inherited world theory as a
going concern. He tentatively believes all of it, but believes also that some unidentified
portions are wrong. He tries to improve, clarify, and understand the system from within.11
These passages make plain Quine’s fundamental commitment to the method of working from
within, which I summarise as follows:
Working from within: We can do no better in our pursuit of truth than to start with the
world theory we inherit, including our ordinary talk of physical things, our talk of
meaning, propositional attitudes, and our talk of translation, and try to improve our
understanding of these many different ways of theorizing, by relying on already accepted beliefs and inferences, and applying our best methods for reevaluating particular statements, beliefs, and inferences, and arriving at new ones.
98

Quine’s naturalism

2 Logic
Quine’s efforts to improve and clarify our inherited world theory from within are unified and
disciplined by his commitment to expressing it in a canonical notation with a first-order
grammar comprising just three basic constructions: predication, quantification and the truthfunctions.12 “The doctrine”, he writes, “is that all traits of reality worthy of the name can be set
down in an idiom of this austere form if in any idiom”.13 Quine emphasises that the doctrine,
though it of course requires that our logical grammar be extensional, does not legislate on
which predicates we may use, or what we may quantify over.14 By combining Cantor’s diagonal method with Grelling’s paradox, Quine also proves that there can be no “fundamental
set of general terms on the basis of which all traits and states of everything could in principle be
formulated”.15 In short, the doctrine provides no support for the standard reading of Quine’s
naturalism, and even implies that some versions of the standard reading are logically
inconsistent.
Quine’s doctrine is inextricable from his conception of first-order logic as the most general
science. He rejects Carnap’s view that logic is analytic, and adopts in its place a doctrine of
gradualism, according to which logic is supported by observation in the same indirect way that
the most general and systematic aspects of natural science are. In addition to the important
technical point that, unlike mathematics, first-order logic has a complete proof procedure,
Quine highlights three “salient traits” of logic, as distinguished from other general and systematic sciences.
First, “the logical truths, being tied to grammar and not to the lexicon, will be among the
truths on which all speakers are likeliest to agree…For it is only lexicon, not grammar, that
registers differences in background from speaker to speaker; and logical truths stay true under all
lexical substitutions”.16 The logical truths are therefore the ones to which speakers are most
likely to be able to appeal when they wish to clarify and resolve their disagreements.
Second, unlike the generalisations of any of the special sciences, such as physics and
mathematics, to formulate logical generalisations, Quine argues, we need to use a truth predicate. We express logical truths by saying, for example, “Every sentence of the form ’p → p’ is
true”, where the sentences we generalise on are sentences in Quine’s preferred canonical
notation and we define the truth predicate by Tarski’s methods so that it entails instances of the
disquotational paradigm “‘_____’ is true if and only if _____”.17 In accepting particular instances of this disquotational paradigm, Quine takes for granted our practical ability to use our
words, in the sense that contrasts with mentioning them, an ability that he regards as both
fundamental and indispensable to logic and to theorising more generally.
Third, first-order logic has “universal applicability” due to “the invariance of logical truth
under lexical substitutions”.18 Thus logic for Quine is the most general science, the framework
for all the others. It is therefore not reducible to, or expressible in terms of, any of the less
general sciences, including the established natural sciences, such as physics.

3 Regimentation and explication
To explain how we may express our theories in a canonical notation with a first-order logical
grammar, Quine develops a pragmatic method of regimentation. The idea is that to regiment a
given natural language sentence S, as one used it on a particular occasion O, is to adopt a certain
kind of linguistic policy – to decide to use a sentence S′ with a perspicuous logical grammar in
place of S, as one used it on occasion O – for the purpose of clarifying and facilitating one’s
inquiries. When a speaker decides to use a sentence S′ of an artificial first-order language in
99

Gary Ebbs

place of a sentence S of his ordinary language, the relation of S′ to S “is just that the particular
business that the speaker was on that occasion trying to get on with, with the help of S among
other things, can be managed well enough to suit him by using S′ instead of S”.19 If a speaker
judges that by using a regimented sentence S′ in place of an ordinary language sentence S, he
can more easily achieve his goals, he is entitled to use S′ in place of S.
Quine’s method of explication is an application of his more general method regimentation.20
To explicate a linguistic expression e that one finds useful in some ways yet problematic in others
is to decide to use, in place of e, a different linguistic expression e′ that preserves and clarifies what
one takes to be useful about e yet avoids what one takes to be the problems with e.21 Quine gives
several examples of what he regards as successful explications. Some concern ontology (e.g. his
proposed explications of “ordered pair” as sets of a certain kind and of “mental events” as “bodily
states”22); and some only concern logical symbols (e.g. his proposed explication of “if p then q” as
“p → q”, as explained by the standard two-valued truth-table and his explication of identity in
terms of the basic predicates of a language23). For Quine the proper question to ask about a given
explication is not whether it is faithful to a concept, if any, expressed by the expression it is to
replace, but whether the explication furthers our goals as inquirers. If the answer to this latter
question is “yes”, then we are within our rights to reject some of the consequences associated
with the expression we aim to explicate.
Quine emphasises that a synonymy constraint on explications would be out of place for the
sorts of paraphrases that matter to science, and to philosophy when it is viewed as continuous
with science, “even if the notion of synonymy…were in the best of shape”.24 If we seek above
all to provide the best explanations of observed phenomena, we should use our best explications
of terms previously in use and not concern ourselves with conceptual analyses of them, even if
we suppose that synonymy is an objective relation.
Quine notes that when they are successful, explications eliminate the problems that motivate
us to introduce them. He also notes that our sense of why the expressions we explicate matter to
us and what they mean cannot but be shaped by our decisions about how to explicate them. His
method of working from within commits him to rejecting the very idea of a “cosmic exile,” a
point of view on what we now mean or what we meant in the past, that is independent of our
own best judgments about how to formulate our theories.25 He concludes that the distinctions
between elimination and explication, or between reduction and clarification, are unreal.26
This conclusion challenges the widespread assumption among Quine’s critics that there is a
clear and important distinction between “reductive” and “liberal” naturalisms. P.F. Strawson,
for instance, argues that, on the one hand, from our commonsense standpoint, a given table
may feel smooth and hard, and on the other, from an “objective” scientific standpoint it is just
an aggregate of particles, hence not smooth or hard. He thinks we cannot rest with the second
view, which he calls “reductive” or “scientific naturalism”, since it implies that the commonsense view of objects is illusory. He therefore recommends that we accept a “liberal”
naturalism that accepts an “ultimate relativity” in our conception of the “real” properties of
objects.27 In a review of the book in which Strawson presents this argument for “liberal”
naturalism, Quine writes,
I see no need of contrasting standpoints in the case of the table. The molecular
physicist can agree with a percipient that the table is smooth and hard; he just goes on
to provide a surprising microphysical detail of what constitutes smoothness and
hardness. He reveals the fine structure. This way of relating the physicist to the
percipient restores a welcome unity of outlook.28

100

Quine’s naturalism

Quine’s reasoning here is of a piece with his criticism of the philosophers who, “in steadfast
laymanship…deplore…departures from ordinary usage”.29 What Strawson and others miss,
Quine thinks, is that “it is precisely by showing how to circumvent the problematic parts of
ordinary usage that we show the problems to be purely verbal”.30

4 Ontology
Quine also applies his method of explication to dissolve the apparent incompatibility between our
commonsense conception of middle-sized objects, such as desks, on the one hand, and the news
from physics that such objects are “swarms of vibrating molecules”, on the other.31 He argues that to
explain how these two parts of our inclusive conceptual scheme fit together, we need to specify
linguistic behaviors that clarify the boundaries between the desk (or river or mountain) we are
referring to, and other middle-sized objects of our possible attention. Among these behaviors are the
speaker’s application of sentences of the form “This is the same as that”, where the demonstratives
are applied to observable portions of the nearby scene. In this way, “The concept of identity is seen
to perform a central function in the specifying of spatio-temorally broad objects by ostension”.32 His
commitments in logic lead him to seek a unified ontology, one that reconciles our ordinary talk of
objects with physics. His solution is to explicate our ordinary individuative predicates, such as
“table”, “desk”, “river” and “mountain”, so that they are true or false of aggregates of molecules that
are more or less singled out by our uses of such sentences as “this is the same as that”.
This is not a reduction of ordinary objects to aggregates of molecules, at least not in the sense
of “reduction” that requires that we be able to specify exactly which aggregates of molecules are
to be identified with a given table, desk, river or mountain. No stipulation that we can formulate and no independent fact of the matter settles “how much to include in the table in the
way of superficial or hovering molecules”.33 Our identification of a table with an aggregate of
molecules “is neither a matter of convention nor a matter of inscrutable but objective fact”.34
What drives us to identify the table with an aggregate of molecules is our commitment to
bivalence in logic and our goal of regimenting ordinary talk of objects in a canonical notation
that quantifies over aggregates of molecules, among other things. To the objection that this
undermines our ordinary conception of objects, Quine will answer that the distinction between
replacing our ordinary conception of objects and clarifying it is unreal.35
In sum, it is Quine’s conception of logic as the framework for all other sciences, his methods
of regimentation and explication, and his goal of reconciling our ordinary talk of objects with
what physics teaches us about them, not any one of these aspects of his philosophy in isolation,
that together lead Quine to identify ordinary objects with aggregates of molecules.
Some philosophers are so committed to the standard reading of Quine’s naturalism that they
criticise Quine for straying from what they take to be its consequences. John Burgess, for
instance, argues that Quine’s interest in ontology, regimentation and explication is “foreign to
the scientific culture”.36 This criticism is rooted in the misreading of Quine’s naturalism that I
highlighted in section 1. Quine’s use of “science” is much broader and more inclusive that
Burgess seems to realise. It encompasses regimentation and explication, which Quine regards as
integral to logic and other established sciences, whose methods, in turn, are continuous with the
methods of commonsense discourse about middle-sized objects and other people.37
Even granting that Burgess’s objection is off-track, one might still think that Quine’s central
aims are to limit the things we quantify over to the particles of physics and aggregates of these
and to explain all the laws of our conceptual scheme in terms of the laws of logic and physics.
On this version of the standard reading, Quine starts with a metaphysical vision of the world
according to physics and seeks to fit everything else into it.
101

Gary Ebbs

In fact, however, Quine’s commitment to reconciling all aspects of our world theory with
what we learn from physics is not based on a metaphysical vision of the world according to
physics. It is based, instead, on a minimalist and nonreductive assumption about the methodological role of physics in our total world theory. Quine explains this assumption as follows:
Why…this special deference to physical theory?…The answer is not that everything worth
saying can be translated into the technical vocabulary physics; not even that all good science
can be translated into that vocabulary. The answer is rather this: nothing happens in the
world, not the flutter of an eyelid, not the flicker of a thought, without some redistribution
of micro physical states. It is usually hopeless and pointless to determine just what micro
physical states lapsed and what ones supervened in the event, but some reshuffling at that
level there had to be; physics can settle for no less. If the physicist suspected there was any
event that did not consist in the redistribution of the elementary states allowed for by his
physical theory, he would seek a way of supplementing his theory. Full coverage in this
sense is the very business of physics, and only of physics. Anyone who will say, “Physics is
all very well in its place” – and who will not? – is then already committed to a physicalism
of at least the nonreductive, non-translational sort stated above.38
In this passage Quine denies both that everything worth saying can be translated into the
technical vocabulary physics and that all good science can be so translated. His deference to
physics is primarily methodological. Just as for Quine logic is the most general science, so for
Quine physics is the natural science that aims at “full coverage”, in the sense that any evidence
that there are events that do not consist in the redistribution of the elementary states allowed for
by current physical theory is evidence that some part of the theory needs to be revised.39
Quine methodological point that physics aims at “full coverage” does not commit him to a
physicalism according to which all objects, whether reduced to physics or not, are located in
space and time. He argues that mathematics, as represented by set theory, is indispensible to the
developed natural sciences, and concludes that among the values of the variables of our total
conceptual scheme are sets, which are not located in space or time.40

5 Epistemology
Another problem for the standard reading of Quine’s naturalism is that it is incompatible with
Quine’s naturalised epistemology. Despite Quine’s notorious claim, quoted at the start of this
paper, that we do not need the idioms of propositional attitude to describe “the true and
ultimate structure of reality”, his considered view is that there is no way to naturalise epistemology without relying at key points on our uses of the idioms of propositional attitude.
To see why, recall first (from section 1) that for Quine we can do no better in our pursuit of
truth than to work from within the inclusive conceptual scheme that we inherit. As Quine
explains, his naturalised epistemology is rooted in this inclusive conceptual scheme:
It is by thinking within this unitary conceptual scheme itself, thinking about the
processes of the physical world, that we come to appreciate that the world can be
evidenced only through stimulation of our senses. It is by thinking within the same
conceptual scheme that we come to appreciate that language, being a social art, is
learned primarily with reference to intersubjectively conspicuous objects, and hence
that such objects are bound to be central conceptually. Both of these aperçu are part of
the scientific understanding of the scientific enterprise; not prior to it.41
102

Quine’s naturalism

As I shall now argue, Quine’s development of these aperçu commits him to accepting irreducible
applications of the methodology of the idioms of propositional attitude.
Quine distinguishes between the doctrinal and the conceptual tasks of epistemology – the
tasks, respectively, of constructing good theories and of clarifying meanings.42 His main contribution to the doctrinal side of epistemology is to elucidate the norms of scientific method.
When he writes that “it is within science itself, and not in some prior philosophy, that reality is
to be identified and described”, he is endorsing the norms embodied in scientific method, as we
actually accept and apply it, and rejecting the traditional philosophical project of validating these
norms.43 He is also expressing his agreement with Hume that, in Quine’s words, “the methodology [of science] is ultimate and irreducible to deductive logic, and the effort to find a
formulation below and beyond science itself is vain and doomed to failure”.44 Philosophers
such as Barry Stroud who accept the central goals of traditional epistemology see Quine’s
agreement with Hume on this point as a capitulation to radical skepticism.45 In Quine’s view,
however, skepticism about the traditional epistemologist’s project, “far from being antithetical
to science is decidedly in the scientific spirit”.46 Put another way, Quine’s point is that scientific
method, broadly construed in the ways I described above, is “the last arbiter of truth”.47
The task of the conceptual side of epistemology for Quine is therefore not what he regards as
the hopeless one of trying to validate the judgments we have arrived at via our applications of
scientific method. The task, instead, is to clarify the relation of these judgments to sensory evidence, guided by the two aperçu noted above. Quine reformulates the first of these as the
“motivating insight” that “we can know external things only through impacts at our nerve
endings” and the second as the observation that “where it matters socially” there is “an objective
pull” toward a “uniformity” in the use of language keyed to “intersubjectively conspicuous
circumstances of utterance”.48 Quine further refines and clarifies these points by adopting a
behavioristic account of how sentences are related to sensory stimulation according to which
I

In the sense of “meaning” relevant to the conceptual task of epistemology, the meanings of
our sentences and words are settled, if at all, by the totality of facts about how sentences of
our inclusive conceptual scheme are variously associated with one another and linked to
impacts at our nerve endings by the mechanism of conditioned response.49

To investigate such associations between sentences and links to impacts at nerve endings Quine
introduces a series of definitions, starting with the following two:
The affirmative (or negative) stimulus meaning of sentence S (for speaker A at t) is the
range of stimulations (“global” triggerings of A’s sensory receptors) that would prompt
A’s assent to (or dissent from) S.50
An occasion sentence (for example, ‘It’s raining’, ‘That’s red’, ‘Lo, a rabbit’) is a
sentence to which speakers assent or dissent only if queried after an appropriate
prompting stimulation.51
Quine stresses that stimulus meanings are “the objective reality that the linguistic has to probe”
when translating occasion sentences.52 The linguist “translates not by identity of stimulus
meanings, but by significant approximations of stimulus meanings”.53 Based on this idea Quine
defines “observation sentence”, as follows:

103

Gary Ebbs

An observation sentence is an occasion sentence on which all speakers of the language
give the same verdict when given the same prompting stimulation.54
This explication presupposes what I will call his first conjecture, namely, that the patterns of neural
stimulations that two speakers receive at their sensory surfaces when they assent to or dissent
from a given sentence may be objectively compared.
Quine realises from the start that this conjecture is of no practical use to the linguist, who
typically has no information about the patterns of neural stimulation at a speaker’s sensory
surfaces that prompt the speaker’s assent to or dissent from a given sentence. He emphasises,
instead, that to translate an occasion sentence that a given speaker assents to in circumstances
that the linguist witnesses, the linguist does not concern himself with patterns of neural stimulation, but simply tries to find an occasion sentence of his own that he also assents to in those
circumstances. From 1960 through 1990 Quine nevertheless also relies on his first conjecture to
clarify and explain in physiological terms the traditional idea that observation sentences are “the
intersubjective tribunal of scientific hypotheses”.55
Quine eventually realises, however, that his speculation about intersubjective similarity of
stimulations rests on “anatomical minutiae” that “ought not to matter here” and that all the
work of clarifying the notion of intersubjectivity of observation, both in translation and language learning, must rest on our own judgments, as translators and teachers, of whether we and
others jointly witness the same circumstances.56 If we judge, for instance, that another speaker’s
utterance of an occasion sentence is prompted by her witnessing an occasion that we would
describe by saying “It’s raining”, we may conclude that she perceives that it’s raining, and decide
on that basis to translate her sentence as “It’s raining”. In so concluding and deciding, we rely
on what Quine calls “empathy” – the method of imagining ourselves in the other person’s
perceptual situation and judging how things are from her point of view.57
It is integral to this change in Quine’s definition of observation sentences that he now
accepts mentalistic predicates that may be used to ascribe de dicto propositional attitudes, such as
Alex perceives that it’s raining. He takes sentences as the objects of the proposition attitudes,
“treating ‘that’ as a quotation mark initiating a name that comes after it”.58 He treats the
quotations as concatenations of letters and spaces. Once quotations are dissolved in this way, “a
single language, regimented in predicate logic, can take them [the attitudes de dicto] in stride
along with natural science”.59 The ascriptions of these mentalistic predicates depend irreducibly
on empathy, however, even if their logical form is extensional. Thus the mentalistic predicates
describe “irreducibly mental ways of grouping physical states and events”.60
To accommodate the change in his definition of observation sentence Quine also offers
what I will call his second conjecture, according to which our successful applications of empathy to
identify situations we jointly witness with other speakers can be explained by a “preestablished
harmony of subjective scales of perceptual similarity”. The idea is that
If A and B jointly witness two events, and A’s neural intakes on the two occasions are
perceptually similar by A’s standards, then B’s intakes will likewise tend to be similar
by B’s.61
Quine speculates that this preestablished harmony is “accounted for by natural selection”.62
This second conjecture, if true, would partly explain how one speaker’s disposition to assent to
or dissent from a given occasion sentence may come to be in harmony with another speaker’s
disposition to assent to or dissent from it. Even if the second conjecture is true, however, it
leaves us with no physiological correlate of the assumed intersubjectivity of circumstances that
104

Quine’s naturalism

observation sentences describe. It therefore does not show us how to eliminate, or do without,
our use of empathy to define observation sentences.
Observation sentences are closely linked by our linguistic dispositions to impacts at our
nerve endings. Quine observes that nonobservational, or theoretical, sentences, by contrast, are
linked by our linguistic dispositions to sensory stimulation in complex and indirect ways that
depend on our dispositions to accept other theoretical sentences.63 There is therefore no
prospect of clarifying the meanings of theoretical sentences by identifying stimulations that
would prompt assent to them, taken singly. This is another important sense in which the
conceptual side of Quine’s epistemology is not reductive. We may speculate about how our
linguistic dispositions relate them to sensory stimulation, but when doing epistemology we do
better to learn the theories of which they are part and to use logic to draw observational
consequences from them.

6 Indeterminacy
These points presuppose Quine’s behaviorist explication of meaning, as summarised by (I)
above. From (I) and our unproblematic, intersubjective uses of the idioms of propositional
attitude to identify and translate observation sentences, Quine infers that such uses of propositional attitudes, and the related translations, are “covered” by physics in his nonreductive
sense, so that
(II) The translation of observation sentences is determinate and factual in Quine’s nonreductive
sense.64
Considerations sketched in the previous paragraph imply
(III) Theoretical sentences are linked by our linguistic dispositions to sensory stimulation in
complex and indirect ways that depend on our dispositions to accept other theoretical
sentences.65
From (III), Quine infers
(IV) The theoretical sentences of a typical language may be translated into theoretical sentences
of that same language, or another one, in inequivalent ways compatibly with all the facts
described in (I).66
From (I), (III) and (IV), he infers that apart from our uses of propositional attitude idioms to
identify and translate observation sentences, most of statements we make by using sentences
containing the idioms of propositional attitude, being dependent on translations of theoretical
sentences that are not uniquely determined by linguistic dispositions, are not “covered” by
physics, and are therefore not objective or factual even in Quine’s nonreductive sense.67 Such
statements are nevertheless pragmatically indispensable, so Quine includes them as a “secondgrade” part of his total world theory.68
In sum, contrary to the standard reading, Quine’s indeterminacy thesis does not rest on the
assumption that all facts about meaning are reducible to physics. It is instead a natural conclusion
of Quine’s efforts to clarify our translational practices from within, while reconciling them as
best he can with relevant discoveries of the developed sciences and with the methodological
point that physics aims at “full coverage”.
105

Gary Ebbs

7 Meaningfulness
One point on which many liberal naturalists agree is that Quine’s naturalism leads to the absurd
conclusion that our sentences are meaningless, so that we cannot use them to make assertions.
(See for instance, McDowell (1994), Putnam (1985) and Strawson (1985).) H.P. Grice and P.F.
Strawson encapsulate the criticism as follows, “If talk of sentence-synonymy is meaningless,
then it seems that talk of sentences have a meaning at all must be meaningless too”.69 The
Grice–Strawson criticism, which is directed at Quine’s arguments in “Two Dogmas of
Empiricism”, looks even more compelling after one has studied Quine’s later argument for the
indeterminacy of translation. For if theoretical sentences have no determinate meanings, then it
seems I cannot use any of my theoretical sentences to make assertions.
This apparently formidable criticism rests on what Quine calls the “fallacy of subtraction”,
whereby it is argued that if we start with a sentence we can use to make assertions and we
subtract our assumption that sentence-synonymy is an objective relation, we are left with a
meaningless sentence, that is, one that we cannot use to make assertions.70 Working from
within our inclusive conceptual scheme, we of course take for granted our practical ability to
use our words to make assertions. Our efforts to clarify and improve our conceptual scheme are
always understood to be compatible with, and to presuppose, it. Quine’s commitment to (I) of
section 5, for instance, is part of his effort to clarify the relation between sentences and sensory
stimulation; it is not an analysis or explanation of our practical ability to use our sentences to
make assertions. We can accept the indeterminacy of translation, he thinks, while still “acquiescing in our mother tongue and taking its words at face value”.71 He expresses the same
basic attitude toward our practical ability to use our words to make assertions when he explains
his commitment to regimenting our statements so that they do not imply that meanings or
attributes, construed as abstract objects, exist. He writes, “I feel no reluctance toward refusing to
admit meanings, for I do not thereby deny that words and statements are meaningful”.72 By
adopting this and other similar explications, Quine subtracts what he regards as doubtful and
unclear assumptions from our understanding of our use of sentences to make assertions, leaving
his minimalist conception of language use as the difference.73 Here, as elsewhere, he regards the
dichotomy between elimination and clarification, or between “reductive” and “liberal” naturalisms, as unreal.

Notes
1
2
3
4
5
6
7
8
9
10
11
12
13

Quine (1981a, 21).
Quine (1960, 221).
Quine (1969a, 303).
See for example Fogelin (1997), McDowell (1994), Price (2007), Putnam (1985), Rorty (1972),
Strawson (1985) and Weir (2014).
Quine (1995, 49). Verhaegh (2018, Sect. 4.5) also emphasises this key point.
Quine (2016 [1995], 34).
Carnap (1949 [1938], 410).
Quine (1976a [1955], 229–230).
Quine (1976b [1955], 252). Verhaegh (2018) argues that Quine deliberately applied this method to
work out the distinctive details of his naturalism.
Quine (1960, 3).
Quine (1981d [1975], 72).
Names and function terms, which are usually included among the constructions of first-order logic, are
definable in terms of the above three basic constructions and identity.
Quine (1960, 228).

106

Quine’s naturalism
14 Quine (1960, 228 and 232, respectively). According to David Macarthur, “The element of first
philosophy in Quine’s system is his scientism, his commitment to the idea that it is the system of
science in particular – that is, ‘total science’ – that limns the true nature of reality; or, as he also puts it,
‘all traits of reality worthy of the name can be set down in an idiom of this austere form [i.e. a properly
scientific vocabulary] if in any idiom’” Macarthur (2019, 183). This passage combines the misreading
of Quine’s use of ‘science’ that I explained in section 1 with the misreading of Quine’s doctrine
concerning the form of a canonical notation that I highlight here.
15 Quine (1960, 231–232).
16 Quine (1986b, 102–103).
17 Quine (1960, 273–274); Quine (1986b, 10–12).
18 Quine (1986b, 102–103).
19 Quine (1960, 160).
20 Gustafsson (2014, 18) observes that “there are examples of Quinean explication where the substitutions
involved have little to do with making the notation more ‘canonical’ ” . From this he infers that
Quine’s method of explication should not be viewed as a special application of his more general
method of regimentation. In fact, however, to regiment a given sentence for Quine is not simply to
rewrite it in terms of a first-order logical notation; it is always also to decide how best to construct
one’s overall world theory. It is therefore no fundamental change of method for Quine when one
replaces a sentence or expression that is already in canonical notation with another sentence or expression in canonical notation that one takes to further one’s goals better.
21 Ibid., 53.
22 Ibid., 258–259; 264–265.
23 Ibid., 259, 230.
24 Ibid., 208.
25 Quine (1960, 275–276).
26 Ibid., 265. Gustafsson (2006) also highlights this consequence of Quine’s method of explication.
27 Strawson (1985, 44).
28 Quine et al. (2008a [1985], 207).
29 Quine (1960, 261).
30 Ibid., 261. Verhaegh (2018, Sect. 4.5), makes a similar point and quotes Quine’s remark in Quine et al.
(2008a [1985]) that Strawson’s dichotomies between “reductive” and “liberal” naturalisms “waiver and
dissolve” under scrutiny.
31 Quine (1976b [1955], 246).
32 Quine (1953c, 68).
33 Quine (1981b, 34).
34 Ibid., 36.
35 Quine (1960, 265).
36 Burgess (1998, 213).
37 James Andrew Smith, Jr. develops a similar criticism of the claim in Burgess and Rosen (1997) that
Quine’s predilection for nominalism is incompatible with his naturalism.
38 Quine (1981b [1978], 98).
39 Since Quine does not assert a reductive physicalism, he can accept, for instance, the sentences of
biology as factual whether or not they are reduced to physical terms, so long as he accepts that there are
no biological differences without physical differences. Gubelmann (2019, at p. 360 and throughout)
misses this key point.
40 How does quantifying over sets fit with Quine’s “full coverage” standard for facts? Quine writes:
Mathematics is integral to our system of the world. Its empirical support is real but remote, mediated
by the empirically supported natural science that the mathematics serves to implement. On this score I
ought to grant mathematics a fact of the matter. But …my suggested standard for facts of the matter is
directed rather at concrete situations, and pales progressively as we move upward and outward. Quine
(1986a, 430).
41 Quine (1976b [1955], 252–253).
42 Quine (1969b, 69).
43 Quine (1981a, 21).
44 Quine (2008a[1946], 135–136).
45 Stroud (1984, 224).
46 Quine (2008a [1946], 135–136).

107

Gary Ebbs
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73

Quine (1960, 23). I develop and defend these points in Ebbs (2019).
Ibid., 2, 8.
Ibid., 11, 73.
Ibid., 32–36,
Ibid., 9.
Ibid., 39.
Ibid.
Ibid., 10; (1969b, 87).
Quine (1969a, 87).
Quine (1992, 40),
Quine (1992, 46).
Quine (1992, 68).
Ibid., 72.
Ibid.
Quine (2016 [1995], 35).
Ibid., 34.
Quine (1960, 11, 73).
Gubelmann (2019) rejects (II), but his argument for doing so rests on the false assumption that Quine is
a reductive physicalist. See note 39.
Quine (1960, 68–72).
Ibid., 73–79; Quine (1992, 47–48).
Quine distinguishes between indeterminacy of reference and indeterminacy of truth value. He regards
the first as a trivial consequence of (I) and (II), and the second as a substantive yet plausible conjecture,
given (I), (III) and (IV).
Quine (1969d, 24).
Grice and Strawson (1956, 146).
Quine (1960, 206).
Quine (1969c, 48).
Quine (1953a, 11).
I develop and defend these points in Ebbs (2011).

References
Burgess, J.P. Occam’s Razor and Scientific Method. In M. Schirn (ed.), The Philosophy of Mathematics
Today, pp. 195–214. Oxford University Press. (1998).
Burgess, J. and Rosen, G. A Subject with No Object. Oxford University Press. (1997).
Carnap, R. Logical Foundations of the Unity of Science. First published in the International Encyclopedia of
Unified Science, Vol. I, No. 1 (1938); reprinted in H. Feigl and W. Sellars (eds.), Readings in Philosophical
Analysis, pp. 408–432. Appleton-Centry-Crofts, Inc. ([1938]/1949).
Ebbs, G. Quine Gets the Last Word. The Journal of Philosophy, 108(11,November), (2011). 617–632.
Ebbs, G. Quine on the Norms of Naturalized Epistemology. In R. Sinclair (ed.), Science and Sensibilia by
W.V. Quine: The 1980 Immanuel Kant Lectures, pp. 115–136, Palgrave-Macmillan. (2019).
Fogelin, R. Quine’s Limited Naturalism. The Journal of Philosophy, 94(11, November 1972), (1997).
534–563.
Grice, H.P. and Strawson, P.F. In Defense of a Dogma. The Philosophical Review, 65(2), (1956). 141–158.
Gustafsson, M. Quine on Explication and Elimination. Canadian Journal of Philosophy, 36(1),
(2006). 57–70.
Macarthur, D. Naturalism from the Mid-Twentieth Century to the Present: Quine’s ‘Hegelianism’,
Armstrong’s Empiricism, and the Rise of Liberal Naturalism. In Kelly Becker and Iain D. Thomson
(eds.), The Cambridge History of Philosophy: 1945–2015, pp. 171–188. Cambridge University
Press (2019).
McDowell, J. Mind and World. Harvard University Press. (1994).
Mill, J.S. Collected Works. In J.M. Robson (ed.), Essays on Ethics, Religion and Society, Vol. 10. University
of Toronto Press. ([1832–1874]/1969).
Price, H. Quining Naturalism. The Journal of Philosophy, 104(8, August), (2007). 375–402.

108

Quine’s naturalism
Putnam, H. A Comparison of Something with Something Else. First published in New Literary History, 17:
61–79; reprinted In J. Conant (ed.), Words and Life, pp. 330–350. Harvard University Press. (1985).
Quine, W.V. On What There Is. Review of Metaphysics, 2, (1948). 21–38; reprinted In From a Logical Point
of View, pp. 1–19. Harvard University Press. (1953a)
Quine, W.V. Two Dogmas of Empiricism. In From a Logical Point of View. Harvard University Press.
(1953b). 20–46.
Quine, W.V. Identity, Ostension, and Hypostasis. In From a Logical Point of View. Harvard University
Press. (1953c). 65–79.
Quine, W.V. From a Logical Point of View. Harvard University Press. (1953d)
Quine, W.V. The Scope and Language of Science. First published in 1955; reprinted In The Ways of
Paradox, pp. 228–245, revised and enlarged edition. Harvard University Press. ([1955]/1976a).
Quine, W.V. Posits and Reality. In The Ways of Paradox, pp. 246–254, revised and enlarged edition.
Harvard University Press. ([1955]/1976b).
Quine, W.V. The Ways of Paradox, revised and enlarged edition. Harvard University Press. (1976c)
Quine, W.V. Word and Object. MIT Press. (1960)
Quine, W.V. Reply to Chomsky. In D. Davidson and J. Hintikaa (eds.), Words and Objections,
pp. 302–311. D. Reidel Publishing Company. (1969a).
Quine, W.V. Epistemology Naturalized. In Ontological Relativity and Other Essays, pp. 69–90. Columbia
University Press. (1969b).
Quine, W.V. Ontological Relativity. In Ontological Relativity and Other Essays, pp. 26–68. Columbia
University Press. (1969c).
Quine, W.V. Speaking of Objects. In Ontological Relativity and Other Essays, pp. 1–25. Columbia
University Press. (1969d).
Quine, W.V. Ontological Relativity and Other Essays. Columbia University Press. (1969e).
Quine, W.V. Things and Their Place in Theories. In Theories and Things, pp. 1–23. Harvard University
Press. (1981a).
Quine, W.V. What Price Bivalence? In Theories and Things, pp. 31–37. Harvard University Press. (1981b).
Quine, W.V. Goodman’s Ways of Worldmaking. In Theories and Things, pp. 96–99. Harvard University
Press. (1981c).
Quine, W.V. Five Milestones of Empiricism. In Theories and Things, pp. 67–72. Harvard University Press.
(1981d).
Quine, W.V. Theories and Things. Harvard University Press. (1981e).
Quine, W.V. Reply to Putnam. In L. E. Hahn, and Paul A. Schilpp (eds.), The Philosophy of W. V. Quine,
pp. 427–431. Open Court. (1986a).
Quine, W.V. Philosophy of Logic, 2nd edition. Harvard University Press. (1986b).
Quine, W.V. Pursuit of Truth, rev. ed. Harvard University Press. (1992).
Quine, W.V. From Stimulus to Science. Harvard University Press. (1995).
Quine, W.V. Lectures on David Hume’s Philosophy. In D. Follesdal and D. B. Quine (eds.), Confessions of
a Confirmed Extensionalist and Other Essays. Harvard University Press. ([1946]/2008a) 36–136.
Quine, W.V. Four Hot Questions in Philosophy. A review of P.F. Strawson’s Skepticism and Naturalism:
Some Varieties that first appeared in the New York Review of Books in 1985; reprinted In D. Follesdal
and D. B. Quine (eds.), Quine in Dialogue, pp. 206–215. Harvard University Press. ([1985]/2008b).
Quine, W.V. Response to Gary Ebbs (written in 1995). In F. Janssen-Lauret and G. Kemp (eds.), Quine
and His Place in History, pp. 33–36. Palgrave Macmillan. ([1995]/2016).
Rorty, R. Indeterminacy of Translation and of Truth. Synthese, 23(4 ), (1972). 443–462.
Smith, Jr., James Andrew Quine on Naturalism, Nominalism, and Philosophy’s Place within Science.
Synthese 198 (2), (2021). 1549–1567
Strawson, P.F. Skepticism and Naturalism: Some Varieties. Columbia University Press. (1985).
Stroud, B. The Significance of Philosophical Scepticism. Oxford University Press. (1984).
Verhaegh, S. Working from Within. Oxford: Oxford University Press. (2018).
Weir, A. Quine’s Naturalism. In G. Harman and E. Lepore (eds.), A Companion to W. V. O. Quine,
pp. 114–147. Wiley–Blackwell. (2014).

109

10
WILFRID SELLARS AND LIBERAL
NATURALISM
Willem A. deVries

1 Introduction
Wilfrid Sellars – one of the foremost philosophers of the twentieth century – was a naturalist,
and there is good reason to believe he was a naturalist of distinctively liberal stripe. But there is
good reason as well to deny that he was a liberal naturalist. It is not because Sellars was simply
inconsistent or confused. He was a widely-read, well-trained and original philosopher who
tried, like most such philosophers, to walk a fine line, trying to find a balance among the many
insights, both inherited and original, he acknowledged allegiance to. As is the case with other
creative, synthesising philosophers, tensions remain in his work, straining against each other. To
the extent to which Sellars’ position holds together, it is an admirable and sophisticated philosophical structure; to the extent the center does not hold, its pieces can still form useful
resources for others. In this essay, I will explore Sellars’ naturalism, especially those aspects that
count for and against its liberality.
The plan of this essay is pretty straightforward. It will be, as befits an examination of Sellars,
dialectical in structure. In section 1, I offer evidence for the obvious and review why there is no
question that Sellars was a naturalist. I will also explore what he took naturalism to be. Section 2
examines Sellars’ conception of what he calls “the manifest image of man-in-the-world”. The
point I want to make there is that Sellars gives the manifest image [MI] a naturalistic interpretation, but it is clearly a liberal naturalism. In section 3, we turn to Sellars’ conception of the
scientific image of humanity-in-the-world [SI], which he claims clashes with the MI and will
eventually supersede it. Sellars was a committed scientific realist; put that together with his
belief that the SI will displace the MI, and there seems to be good reason to think he is a very
illiberal naturalist. Finally, in section 4, we will look more closely at the “synoptic vision”
Sellars really endorses, which preserves central aspects of the MI. The pendulum thus seems to
swing back to favor a liberal naturalism.

2 Sellarsian naturalism, round one
Sellars calls himself a naturalist. His book, Naturalism and Ontology, makes that clear. He calls
himself a naturalist, however, with some reluctance, for he fears the term is “wishy-washy and
ambiguous”. “One could believe almost anything about the world and even some things about
110

DOI: 10.4324/9781351209472-13

Wilfrid Sellars and liberal naturalism

God, and yet be a Naturalist” (NAO, Introduction paragraph 7, 9–10).1 He would, he points
out, prefer the label “non-reductive materialism”, but “materialism” has been so abused by
others that he sticks with “naturalism”. This label is acceptable to him only because “if it does
not entail scientific realism, [it] is at least not incompatible with it” (NAO, Introduction
paragraph 7, 10). His allegiance to scientific realism cannot be pushed aside.
We can get a better fix on Sellars’ conception of naturalism by noting the company it keeps
in his thought. There are three important associations to be found in his talk of naturalism. The
first is an association between naturalism and empiricism. Empiricists have more consistently
rejected the supernatural, the supersensible and the “metaphysical” than rationalists have. For
Sellars, it is crucial that experience is itself natural. Empiricism is primarily an epistemological
position concerned with how we acquire concepts and knowledge, so the association with
empiricism favors a methodological form of naturalism. What it is we acquire knowledge of is a
more difficult question for empiricism.
Second, Sellars also associates naturalism with materialism, as in the introduction to
Naturalism and Ontology. Sellars became wary of “materialism”, however, and in general the
term “materialism” has lost popularity. We no longer think of physics as all about matter – there
are fields and forces as well, at very least. Quantum mechanics has made even matter seem less
material than it used to be. The slightly less specific “physicalism” has taken the place of
“materialism”. The association of naturalism with materialism or physicalism gives Sellarsian
naturalism a substantive dimension to complement the methodological dimension noted above.
Let me take a moment here to clarify an important feature of Sellars’ naturalistic ontology.
Sellars recognises two methodologically distinct forms of ontological investigation. There is a
formal, analytic/interpretive inquiry that develops the theory of the ontological commitments
of representational and especially linguistic systems. This is a higher-order, meta-conceptual
investigation generally undertaken only by philosophers. This is the enterprise that occupies
NAO, GE and his essays on abstract entities: AE, CAE, EAE, inter alia. Sellars strives to elaborate a sophisticated nominalism in these essays,2 because he believes that nominalism is a
requirement for naturalism. Cognitive activities do not require commitment to the existence of
non-natural entities such as abstracta. This formal, analytical aspect of ontology applies to both
the MI and the SI; his nominalistic semantics is an important part of his naturalism.
The second form of ontological inquiry is simply first-order empirical science. Empirical
science is the rigorous investigation of what there is as well as how it operates. Proper, that is,
naturalistic, methodology is the way to understand the true nature of things.
The third association that is distinctive about Sellarsian naturalism is his deep concern with
the argument between ethical naturalism and non-naturalism. In an early article, “Mind,
Meaning and Behavior”, Sellars argues that “3.3221 The common presupposition of Naturalist
and Non-naturalist is causal reducibility implies logical reducibility. We rejected this presupposition” (MMB, 87). Those who are familiar with Jim O’Shea’s discussions of Sellars’
“naturalism with a normative turn” will recognise the ideas.3 For those who are not, I will
unpack the claim further.
My interpretation of the claim that normative claims are causally, but not logically, reducible to
physicalistic descriptions is that (1) for any situation in which a norm is being obeyed or realised
there is possible a true, purely physicalistic description of it; (2) such a physicalistic description
contains no normative terminology; (3) the situation will evolve temporally in accordance with
causal law; and (4) that description – as far as the strict physical chronology of the situation is
concerned – leaves nothing out. The description is not “gappy” in terms of the fundamental ontology of the natural world; neither will any empirically well-established causal principles be
violated.4 It thus constitutes a causal reduction of that particular norm-involving situation. But the
111

Willem A. deVries

norm-involving description of the situation is logically irreducible to the physical: the physical
description of the situation is neither synonymous with nor entailing of any norm-involving
description of it, nor is any other logical relation going to connect the two. The norm-involving
description expresses something that cannot be expressed in the purely physical language, despite
the completeness of the physical description on its own terms.
Sellars’ most extensive discussion of this crucial point occurs in his essay, “Counterfactuals,
Dispositions, and the Causal Modalities” [CDCM], particularly paragraphs 79–80. He argues
there against descriptivism, the notion that all nonlogical concepts are descriptive in a strict sense.
He also ascribes to naturalism the thesis “that the world, including the verbal behavior of those
who use the term ‘ought’ – and the mental states involving the concept to which this word
gives expression – can, ‘in principle,’ be described without using the term ‘ought’ or any other
prescriptive expression” (CDCM, Sect. 79, 282). That is, naturalism is committed to the idea
that there is a way the world is, independently of any way the world ought (or ought not) to be.
In fact, he thinks that naturalism is more generally committed to the idea that the world can be
described without any modal terminology.
The idea that the world can, in principle, be so described that the description contains
no modal expression is of a piece with the idea that the world can, in principle, be so
described that the description contains no prescriptive expression. For what is being
called to mind is the ideal of a statement of “everything that is the case” which,
however, serves, through and through, only the purpose of stating what is the case.
(CDCM, Sects. 79–80)
His descriptive ideal expresses a commitment to the idea that all modality comes to the world
from within; there is no supernatural basis for normativity or even causation.
Sellars declares his scientific realism in the so-called scientia mensura: “…in the dimension of
describing and explaining the world, science is the measure of all things, of what is that it is, and
of what is not that it is not” (EPM, IX, Sect. 41; in SPR, 173). But CDCM makes it clear that
describing and explaining are different enough enterprises that their vocabularies differ in nontrivial ways. The explanatory enterprise employs a distinctive modal vocabulary in which talk of
physical necessities and entailments is indispensable. But the descriptive enterprise, at least in its
pure form, need make no use of such vocabulary. Sellars spends a great deal of time analysing the
semantics of the modalities and of predication itself precisely in order to be able to claim that the
verbiage employed in the explanatory enterprise (and also any new vocabulary employed in the
prescriptive enterprises of action and deliberation) does not add any new objects to the world –
the ontology remains unchanged.
Sellars tells us,
[A]lthough describing and explaining (predicting, retrodicting, understanding) are
distinguishable, they are also, in an important sense, inseparable. It is only because the
expressions in terms of which we describe objects, even such basic expressions as
words for the perceptible characteristics of molar objects5 locate these objects in a
space of implications, that they describe at all, rather than merely label.
(CDCM, Sect. 108)
We have to recognise that descriptive language is not (and cannot be) entirely self-sufficient and
that many other linguistic functions are requisite for description to be possible. Then we can
see that,
112

Wilfrid Sellars and liberal naturalism

once the tautology ‘The world is described by descriptive concepts’ is freed from the
idea that the business of all nonlogical concepts is to describe, the way is clear to an
ungrudging recognition that many expressions which empiricists have relegated to
second-class citizenship in discourse, are not inferior, just different.
(CDCM, Sect. 79)
There are two important points here. One is that Sellarsian naturalism is committed to there
being a way the world is that is logically independent of and prior to any way the world ought to be.
Nonetheless, our knowledge of, even our ability to think of the world cannot be divorced from
our ability to employ normative concepts. The second point is that when ontology is done
properly, recognition of the different distinctive uses of language is perfectly compatible with
naturalism, which, therefore, need not try to force all language into the Procrustean bed of
reductive materialism.

3 Naturalism and the two images of humanity-in-the-world
Sellars further unfolded his naturalism in his 1962 essay “Philosophy and the Scientific Image of
Man” [PSIM]. There he distinguishes between “the manifest image” and “the scientific image”
of humanity in the world. This distinction has taken on a life of its own in the literature, so let’s
see how Sellars himself understood it. The manifest image [MI] is a conceptual framework
embedded in the practices and natural languages developed by humans in terms of which they
make sense of and cope with the world and their place therein. It is an idealisation that abstracts
away from the many differences of detail across languages and cultures. But it is not so idealised
that it can be regarded as fixed and immutable. What is most important about the MI is the fact
that it makes available the possibility of self-reference and metaconceptual representation; the
manifest image is “the framework in terms of which man came to be aware of himself as manin-the-world” (PSIM, paragraph 14; in SPR, 6). Its origins may be cloaked in the mists of time,
but they are, in Sellars’ view, to be explained causally. Whatever its origins, because of its selfreferential and metaconceptual capacities, it has been further developed in dynamic and even
sometimes rational ways.
Sellars thinks that the fundamental structures of the manifest image have been best captured
by such philosophers as Aristotle and P.F. Strawson. It portrays a world in which persons and
things are the basic kinds of objects, though the emphasis is on persons. In the MI, “object” is a
derivative category.6 That the category of person remains fundamental in the manifest image
means that the explanatory forms appropriate to personhood cannot be entirely overridden or
discarded in the framework. We explain human activity by reference to a person’s habits,
character, impulses, etc.
It is a striking exaggeration to say of a person, that he is a “mere creature of habit and
impulse”, but in the early stages of the development of manifest image, the world
includes truncated persons which are mere creatures of habit, acting out routines,
broken by impulses, in a life which never rises above what ours is like in our most
unreflective moments. Finally, the sense in which the wind “did” things was pruned,
save for poetic and expressive purposes—and, one is tempted to add, for philosophical
purposes—of implications pertaining to “knowing what one is doing” and “knowing
what the circumstances are”.
(PSIM, paragraph 36; in SPR, 13)

113

Willem A. deVries

The lesson, I take it, is that object makes sense in the MI only against the background of and
contrast to persons. Further, Sellars does not believe that the manifest image is essentially
dualistic (see PSIM, paragraph 31; in SPR, 10–11); the person/object contrast is not to be
understood as a contrast between material and immaterial items. Thus, the manifest image
is thoroughly permeated by the categorial (and therefore the explanatory) structures associated with persons, however minor a part of the universe is made up of full-fledged
persons.
It seems obvious that Sellars portrays the MI as a liberally naturalistic framework. Sellars
acknowledges that there are pressures that some have responded to by separating bodies from
minds or spirits, but he argues that this is not necessary. “[T]he essential dualism in the manifest
image is not that between mind and body as substances, but between two radically different
ways in which the human individual is related to the world” (PSIM, paragraph 31; in SPR, 11).
Neither of the ways the human individual is related to the world (sense and intentionality),
however, removes the person from the world or nature.
Though Sellars believes that the MI can (and ought to) be given a naturalistic analysis, he does
not believe that the MI is ultimately a stable and enduring framework. It will have to be superseded, for there are questions prompted by the MI that it will never be in a position to answer.
That brings us to the framework Sellars believes will replace the MI, the scientific image.

4 The scientific image
The scientific image is also a highly idealised construct; it is so idealised that we cannot claim to
know its fundamental structures yet. The distinguishing mark of the scientific image is its
employment of postulational methods; it adds to our conception of the world by invoking, not
only hitherto unobserved objects, but new kinds of unobservable objects. In fact, in Sellars’
view, science does not simply add to our ontology; ultimately, it challenges the ontology and the
explanatory structures constitutive of the manifest image. Most obviously, the challenge takes
the form of installing a new set of fundamental or basic objects, but it also reconfigures the
acceptable explanatory forms, because the new objects are categorially distinct from any old
ones. Though we do not know yet what the basic objects of the scientific framework will be
(quarks? Strings? Sellars’ favored absolute processes?), we can already surmise that persons will
not be basic objects in the manifest image.7
The scientific image is an idealisation with both synchronous and dynamic dimensions. It
idealises away from the fact that there are different sciences that engage the world at different
levels of grain with different tools. (See PSIM, paragraph 55; in SPR, 20.) It is also an idealisation
that could exist only in some hypothetical future in which we’ve avoided blowing ourselves up or
destroying our environment long enough to answer all the interesting scientific questions.
Sellars is clear that the SI is methodologically dependent on the MI; it could not get off the
ground unless the MI were already in place. He denies that this entails any substantive priority of
the MI over the SI. Quite the contrary: the SI is a better image of the world, more highly refined
and detailed, better supported by rigorously acquired evidence and reasoning, capable of
epistemological closure in a way that the MI is not. The SI
purports to be a complete image, i.e. to define a framework which could be the whole
truth about that which belongs to the image. Thus although methodologically a development within the manifest image, the scientific image presents itself as a rival
image. From its point of view the manifest image on which it rests is an “inadequate”

114

Wilfrid Sellars and liberal naturalism

but pragmatically useful likeness of a reality which first finds its adequate (in principle)
likeness in the scientific image.
(PSIM, paragraph 56; in SPR, 20)
Sellars also distinguishes between unifying the postulated entities of two sciences and unifying
the sciences themselves. The objects of investigation in chemistry are all objects of physical
investigation or composites thereof, but chemistry and physics are different sciences with different methodologies, instruments and vocabularies. This is a model as well for the relations
between the sciences of “behavioristics” and neurophysiology. Thus, Sellars believes, “the
correlational content of behaviouristics points to a structure of postulated processes and principles which telescope together with those of neurophysiological theory” (PSIM, paragraph 70;
in SPR, 25). Sellars’ view, then, is that psychology can be an autonomous science with vocabulary, methods, and instruments peculiar to it, but the entities it deals with must consist of
organised structures of the entities treated in the “lower” sciences.
But here’s where things start to get sticky in Sellarsland. Do we have good reason to believe
that the mental entities psychology worries about will have the proper relations to the entities of
neurophysiology, chemistry and physics? Yes, no and yes, according to Sellars. That is, Sellars
thinks (1) intentional states will not be a significant barrier to aligning the entities of psychology
with the entities of neurophysiology and even physics. (2) Sensations, on the contrary, do pose a
significant challenge to such an alignment, but (3) appropriate adjustments that naturalise
sensation states will be able to be made – but at the level of fundamental physics. This summary
now needs to be unpacked.
First, of course, we need an adequate analysis of the entities and properties involved in
psychology as it appears in the MI, for any science of psychology will, like all other sciences,
start with the appearances familiar to us in the MI. Sellars takes dead seriously the distinction
between the sensory and the conceptual or intentional; he thinks that such a distinction is built
into the MI, and it informs the fundamental structure of his philosophy of mind.
Our concept of an intentional state, according to Sellars, is modeled on the concept of a
covert act of “inner speech” – so we think of intentional states as having a subject (or thinker),
an attitude and a content. Sellars spent a great deal of effort developing his functionalist semantics, according to which giving the meaning of a linguistic utterance or the intentional
content of an intentional state is a matter of classifying it functionally.8 Such a functional
classification conveys the function of the item in a complex system of meaningful interactions
with one’s environment and fellow creatures. Such functional classification is usually done by
displaying or citing a relevantly similar utterance or content, not by giving a complex description of the function in question. Sellars believes his functionalist semantics is thoroughly
naturalistic. (1) It makes no reference to and allows no room for non-natural abstracta to play an
explanatory role in the constitution or understanding of meaningful utterances or intentional
states. (2) It applies to the states and activities of material bodies, requiring no allusions to
immaterial spirits. Sellars exploits the computer model of the mind at times to reinforce his
naturalistic analysis of the meaningful (e.g. PSIM, paragraph 93; in SPR, 33; BBK, passim).
Sellars therefore thinks that the pure descriptions developed in the SI9 will encompass our
intentional states without a glitch, though it will turn out, as we’ll see, that they will not be able
to express everything contained in the concept of intentionality. There remain, of course, difficult and serious questions about the way(s) in which intentional states are embodied in the
neurophysiology of living organisms, but those are empirical questions, to be answered by
empirical investigations in cognitive neuroscience. Sellars’ naturalistic analysis of intentionality
is materialistic in one sense, yet not strongly reductive. People’s psychological states must be
115

Willem A. deVries

understood as two-tracked: they are physiological states describable in the language of biology
and biochemistry, but they lead a “double life”, for they play significant roles in the social and
personal activity of their bearers. Performing those roles requires another vocabulary that
cannot be reduced to the vocabulary of biochemistry. Such states are causally reducible but
logically irreducible.
Sensation states, however, pose a very different challenge, according to Sellars.10 To understand his point, let’s discuss his take on identifying complex entities with systems of more
basic entities. Sellars enunciates a principle of reducibility in PSIM:
If an object is in a strict sense a system of objects, then every property of the object must
consist in the fact that its constituents have such and such qualities and stand in such
and such relations or, roughly, every property of a system of objects consists of
properties of, and relations between, its constituents.
(PSIM, paragraph 74; in SPR, 27)11
Sellars applies this principle to the sensory qualities of objects. Sellars does not think there is any
plausible way to think that a manifest property like being pink can be identified with a property (or
relation among properties) of any finer grained system of uncolored objects. The status of color
properties is a problem, whether we are working within the manifest or the scientific image.
Pink does not seem to be made up of imperceptible qualities in the way in which
being a ladder is made up of being cylindrical (the rungs), rectangular (the frame),
wooden, etc. The manifest ice cube presents itself to us as something which is pink
through and through, as a pink continuum, all the regions of which, however small,
are pink. It presents itself to us as ultimately homogeneous.
(PSIM, paragraph 73; in SPR, 26)
There isn’t time or space here to run through the entire, complex dialectic that constitutes
Sellars’ argument for his position,12 so I will cut to the chase here.
Sellars thinks that the status of the sensory is neither resolved nor resolvable in any satisfactory way within the MI itself. Stories about “being-for-sense” and awkward distinctions
between primary and secondary qualities litter the history of philosophy. Indeed, the difficulties
of finding the right place for sensory qualities is a driving factor in Sellars’ thinking about the SI
as well, for, as long as we hold on to a “particulate” image of the ultimate reality of the world,
he believes, the sensory qualities will be no more at home in the SI than in the MI (see PSIM,
paragraph 107; in SPR, 37).
The problems posed by sensory qualities do not inspire Sellars to abandon his naturalism.
Instead, Sellars argues that, under the pressure of the need to find a place in the natural world
for qualitative content (the existence of which pace Dennett, he believed to be undeniable),
science itself will need to adjust. Sellars argues that the sciences will move to an ontology of
“absolute processes” in place of the object-centered ontology currently assumed. Absolute
processes are subjectless; there is no object to which the process belongs. Absolute processes are
not like John runs or The wine ferments, but more like It’s raining. The subject term is a mere
dummy. In this framework,” objects” and object-bound processes would, in traditional terminology, be “logical constructions” out of, that is, patterns of “absolute processes” (FMPP, III,
paragraph 112, 85). So, the ultimate physical picture would include regular physical objects,
now construed as patterns of absolute processes, but in addition “the domain of absolute

116

Wilfrid Sellars and liberal naturalism

processes would include σ-ings (e.g. C#ings, reddings), the transposition of sensa into the
framework of absolute process” (FMPP, III, paragraph 115, 85).13 Sellars concludes
whereas the objects of contemporary neuro-physiological theory are taken to consist of
neurons, which consist of molecules, which consist of quarks,…-all physical2 objects-an ideal successor theory formulated in terms of absolute processes (both φ2-ings and σings) might so constitute certain of its ‘objects’ (e.g. neurons in the visual cortex) that
they had σ-ings as ingredients, differing in this respect from purely physical2 structures.
(FMPP, III, paragraph 124, 86)14
Thus, Sellars argues that mental states, whether intentional states or sensory states, will find a
place in ultimate science, although they currently pose real challenges to the naturalism implicit
in the MI. Other purported challenges to naturalism – ghosts or gods – simply don’t seem to be
taken seriously by him.
It is because Sellars was devoted to scientific realism that the liberality of his naturalism can
be put into serious question.

5 Scientific realism: how liberal can it be?
The argument from scientific realism to reductionist naturalism and therefore against liberal naturalism seems fairly direct. Suppose that science truly is the measure of what is, and, further, that
science is developing a new and better language for the description and explanation of the items and
events in the world that will justifiably replace the commonsense, “manifest image” language we
have operated with so far. Then, since, in Sellars’ view, “no picture of the world contains as such
mentalistic expressions functioning as such” (SM, Ch. V, paragraph 78, 143), “in this sense there are no
mental acts” (loc. cit). It is this aspect of Sellars’ view that has encouraged people like Richard Rorty
(in some of his guises) and Paul Churchland to move beyond reductive to eliminative materialism,
the least liberal form of naturalism. Sellars was sure he could avoid eliminativism, but it can be
difficult to see how he can pull it off. For Sellars “agreed with Kant that the world of common sense
is a ‘phenomenal’ world, but suggested that it is ‘scientific objects’, rather than metaphysical unknowables, which are the true things-in-themselves” (SM Ch. V, paragraph 79, 143).
The notion of picturing involved here is Sellars’ bow to the idea that representations
(whether linguistic or mental) correspond to things in the world. In his view, picturing is a
nonintentional (ergo, natural) relation that holds between some representings (namely, atomic,
singular descriptive statements in special circumstances in their guise as natural objects or events)
and the objects of the surrounding environment. It is a mapping relationship or isomorphism
that permits of degrees of accuracy, and, in Sellars’ view, a principal goal of science is to enable
the construction of ever finer-grained, more accurate pictures of various parts of the world.
Sellars’ conception of picturing has given even the most astute of Sellars’ commentators difficulty;
unfortunately, we cannot get deeper into details here.15
Sellars says that there is a sense in which there are no mental acts, but he also claims that it is a
sense that is nonetheless compatible with the “indispensibility and logical irreducibility of
mentalistic discourse” (SM Ch. V, paragraph 78, 143). Sellars has therefore told us that: (1)
There are, in some important sense, no mental acts. (2) Sensory states will be revealed as
physical, though that will require a change in the categorial structure of future science. (3)
Intentional acts can be explained in an extensional language – presumably the language of an
adequate science. (4) Yet mentalistic discourse is indispensable and logically irreducible. It is a
strange brew. Let’s see how Sellars hopes to pull off this complex position.
117

Willem A. deVries

6 Putting humanity into the scientific image
Sellars seems to think that such concepts of the personal as “character”, “impulse” and “decision”
can be reconstructed as “extraordinarily complex defined [physical] concepts” (PSIM, paragraph
109; in SPR, 38). But even if this is possible, the idea that we could reconstruct the categories
pertaining to personhood adequately in the language of ideal science “fails decisively on another
count” (PSIM, paragraph 110; in SPR, 38). We can see why this is so when we recognise that,
To say that a certain person desired to do A, thought it his duty to do B but was forced
to do C, is not to describe him as one might describe a scientific specimen. One does,
indeed, describe him, but one does something more. And it is this something more
which is the irreducible core of the framework of persons.
(PSIM, paragraph 111; in SPR, 39)
The “something more” isn’t hard to understand.
To think of a featherless biped as a person is to think of it as a being with which one is
bound up in a network of rights and duties. From this point of view, the irreducibility
of the personal is the irreducibility of the ‘ought’ to the ‘is’.
(PSIM, paragraph 112; in SPR, 39)
Persons exist only where normativity exists, but describing the use of “ought” will never
substitute for feeling the call of duty. In a triumphant SI, we may be able to describe the world
(and ourselves) in ideal fashion, but from that description, no prescriptions and thus no normativity can be derived. Whence cometh normativity?
Sellars’ answer is that normativity arises within communities.
The fundamental principles of a community, which define what is “correct” or “incorrect”, “right” or “wrong”, “done” or “not done”, are the most general common
intentions of that community with respect to the behaviour of members of the group.
(PSIM, paragraph 113; in SPR, 39)
Persons exist insofar as there is a community of beings that thinks thoughts of the form, “We
(one) shall do (or abstain from doing) actions of kind A in circumstances of kind C”. Sellars tells
us that “[t]o think thoughts of this kind is not to classify or explain, but to rehearse an intention”
(PSIM, paragraph 113; in SPR, 40). Norms are the communal creations of groups that exercise
the ability to legislate for themselves.16
Sellars thinks that his approach means that
the conceptual framework of persons is not something that needs to be reconciled with the
scientific image, but rather something to be joined to it. Thus, to complete the scientific
image we need to enrich it not with more ways of saying what is the case, but with the
language of community and individual intentions, so that by construing the actions we
intend to do and the circumstances in which we intend to do them in scientific terms, we
directly relate the world as conceived by scientific theory to our purposes, and make it our
world and no longer an alien appendage to the world in which we do our living.
(PSIM, paragraph 114; in SPR, 40)

118

Wilfrid Sellars and liberal naturalism

This is how Sellars hopes to preserve and naturalise the categories of personhood.
Let me raise some doubts about Sellars’ strategy here. He talks about “joining” the language
of community and individual intentions to the scientific image of the world, but that seems to
imply that the two are fairly independent of each other, capable of separate existence. But
science is a human activity; it may portray the world without thereby portraying norms as items
in the world, yet the scientific image could not come to exist independently of such normativity. It is too facile to dismiss this complaint as pointing to a merely methodological priority of
the MI over the SI, for the members of the Peircean community that achieves ideal science will
not be able to “discard the (normative) ladder” they climbed to reach their goal. If they did,
they would no longer be able to make sense of what they have done. Because science itself is
such a human, norm-governed activity, the separation between the norm-infested MI and an
anormative SI seems artificial. Rather than separable domains joined together, it is much more
likely to be the case that the MI and the nascent SI co-evolve.17 But then it isn’t clear that their
“clash” needs to result in a wholesale replacement of one by the other.
Sellars also portrays the ultimate dénouement in terms of our coming to construe “the
actions we intend to do and the circumstances in which we intend to do them in scientific
terms, [so] we directly relate the world as conceived by scientific theory to our purposes” (loc.
cit.). But, for one thing, a major point of scientific investigation is to bleach out of our portrayal
of the world that aspect of the portrayal that is conditioned on its being our portrayal and tied to
our needs and interests. At the same time, we, of course, need to be able to refer to and
characterise the world in convenient ways applicable in real time to our purposes and plans. I
see little reason to believe that advanced science will afford us opportunities to develop
scientifically-based, “purified” concepts of the artifacts and objects that we use and encounter in
everyday life (or in the laboratory). There is unlikely to be a pure, norm-free, scientific description or “reconstruction” of the concept of hammer, for instance, not just because hammers
can vary widely in their design, material and construction, but because hammers are things that
are to-be-pounded-with. Normativity is built into their very concept.
For reasons like these, the relationship between the norm-infested manifest image and the
norm-free scientific image will turn out to be much more dialectical than Sellars envisaged.
Exploring this dialectic should enable us to develop a richer naturalism as well. Latter-day
Sellarsians like Brandom, McDowell and Huw Price have been sensitive to these issues, exploiting a linguistic pluralism much like Sellars’. Their liberal naturalism comes at the cost of
Sellars’ scientific realism. For the orthodox Sellarsian, however, trying to unify scientific realism
with liberal naturalism remains the ideal.

Notes
1 I will cite Sellars’ works parenthetically in the text, using the industry-standard abbreviations. For a list,
see the reference list at the end of the essay.
2 The complexity of his nominalism is interestingly discussed in Kraut 2010.
3 See O’Shea (2007) and (2009).
4 One thinks of the cartoon by Sidney Harris showing two scientists in front of a blackboard filled with
equations on the left and right, in the middle of which sits the phrase “then a miracle occurs”. I can’t
(unfortunately) reproduce it here, but it can be found at http://www.sciencecartoonsplus.com/
gallery/math/index.php#.
5 [Sellars’ footnote:] For an elaboration of this point, see my essay “Empiricism and the Philosophy of
Mind” (EPM, 31; in SPR, 53), particularly Sects. 35–38 (103–113).
6 Sellars mentions, speculatively, the notion of an “original image” (one in which persons is the sole
ontological category, so that being an object is always a way of being a person) out of which the MI

119

Willem A. deVries

8
9

10
11

12
13
14

developed by truncating the notion of a person to produce a new ontological category: (mere) object
or thing. But I am not convinced this idea can be coherently unified with other aspects of his thought.
See deVries (2005, 284, n. 7).
Sellars says that, assuming “the correlational content of behaviouristics points to a structure of postulated processes and principles which telescope together with those of neurophysiological theory,…
the scientific image of man turns out to be that of a complex physical system” (PSIM, paragraph 70; in
SPR, 25).
See MFC or SM, Chapter III.
Sellars writes, “If (as I do not believe) it should turn out, for example, that the behavior of persons
requires for its description and explanation ‘mental acts’ having an ‘intentionality’ which can not be
explicated in terms of the forms and categories of an extensional logic, then it would be odd to include
these ‘mental acts’ as part of the subject matter of a ‘physical theory’, and to speak of them as ‘physical’
events” (SSIS, 439).
Sellars distinguishes the mind–body problem, which involves intentional states, from the
sensorium–body problem, centered on the sensory. See FMPP, lecture III, “Is Consciousness
Physical?”, paragraphs 2–4, 66.
Notice that if he sticks to this principle, then even if a person is a (as quoted above) “complex physical
system” (PSIM, paragraph 70; in SPR, 25), they are not so in the “strict sense”, for many of the
properties of persons do not consist of properties of and relations between their constituents. I own a
Toyota, but that property of mine depends on a number of fairly large-scale social structures and
relations.
I have given my best effort to understand Sellars on sensation deVries (2005, ch. 8). But see also Jay
Rosenberg, “The Place of Color in the Scheme of Things: a Roadmap to Sellars’ Carus Lectures” in
his (2007).
Sellars uses the Greek letter σ as a variable ranging over sensings.
In “The Concept of Emergence”, jointly authored with Paul Meehl, Sellars defined two notions of the
physical.

Physical1: an event or entity is physical1 if it belongs in the space-time network.
Physical2: an event or entity is physical2 if it is definable in terms of theoretical primitives
adequate to describe completely the actual states though not necessarily the potentialities of the
universe before the appearance of life. (CE, 252)
15 For treatments of Sellarsian picturing, please see my Wilfrid Sellars, James O’Shea’s identically titled
work, my (2010), and Jay Rosenberg, “Sellarsian Picturing” in his 2007.
16 For much more detail on the Sellarsian treatment of norms I recommend Wolf and Koons (2016) and
Koons (2019).
17 It is a shame that Sellars says very little about biological norms, which might offer a path to a deeper
naturalisation of the social normativity operating among humans.

References
deVries, W.A. Wilfrid Sellars. Philosophy Now Series. Acumen Publishing and McGill-Queen’s University
Press. (2005).
deVries, W.A. Naturalism, the Autonomy of Reason, and Pictures. International Journal of Philosophical
Studies, 18(3), (2010). 395–413.
Koons, J.R. The Ethics of Wilfrid Sellars. Routledge Studies in American Philosophy. Routledge. (2019).
Kraut, R. Universals, Metaphysical Explanations, and Pragmatism. Journal of Philosophy, 107, (2010).
590–609.
O’Shea, J.R. Wilfrid Sellars: Naturalism with a Normative Turn. Key Contemporary Thinkers series.
Blackwell/Polity Press. (2007).
O’Shea, J.R. On the Structure of Sellars’s Naturalism with a Normative Turn. In W. deVries (ed.),
Empiricism, Perceptual Knowledge, Normativity, and Realism: Essays on Wilfrid Sellars. Oxford University
Press. (2009).

120

Wilfrid Sellars and liberal naturalism
Rosenberg, J.F. Wilfrid Sellars: Fusing the Images. Oxford University Press. (2007).
Wolf, M.P. and Koons, J.R. The Normative and the Natural. Palgrave Macmillan. (2016).
Sellars Citations

MMB
CE

Mind, Meaning, and Behavior, Philosophical Studies 3 (1952): 83–95.
The Concept of Emergence (with Paul Meehl), in Minnesota Studies in the
Philosophy of Science, Vol. I: The Foundations of Science and the Concepts of Psychology
and Psychoanalysis, edited by H. Feigl e M. Scriven (Minneapolis: University of
Minnesota Press, 1956)
EPM
Empiricism and the Philosophy of Mind”, in Minnesota Studies in the Philosophy of
Science, Vol. I: The Foundations of Science and the Concepts of Psychology and
Psychoanalysis, edited by H. Feigl e M. Scriven (Minneapolis: University of Minnesota
Press, 1956): 253–329. Reprinted as Chapter 5 in SPR. Republished as a book with
an introduction by Richard Rorty, and a study guide by Robert Brandom
(Cambridge, Mass.: Harvard University Press, 1997).
CDCM Counterfactuals, Dispositions, and the Causal Modalities”, in Minnesota Studies in the
Philosophy of Science, Vol. II, edited by H. Feigl, M. Scriven and G. Maxwell
(Minneapolis: University of Minnesota Press, 1957): 225–308.
GE
Grammar and Existence: A Preface to Ontology, Mind 69 (1960): 499–533.
Reprinted in SPR and ISR.
BBK
Being and Being Known, Proceedings of the American Catholic Philosophical
Association (1960): 28–49. Reprinted in SPR and ISR.
PSIM Philosophy and the Scientific Image of Man, in Frontiers of Science and Philosophy,
edited by Robert Colodny (Pittsburgh: University of Pittsburgh Press, 1962): 35–78.
Reprinted in SPR and ISR.
EAE
Empiricism and Abstract Entities, in The Philosophy of Rudolf Carnap (The Library of
Living Philosophers), edited by P. (La Salle, Illinois: Open Court, 1963): 431–468.
AE
Abstract Entities, Review of Metaphysics 16 (1963): 627–671.
CAE
Classes as Abstract Entities and the Russell Paradox, Review of Metaphysics 17
(1963): 67–90.
SPR
Science, Perception and Reality (London: Routledge and Kegan Paul, 1963).
Republished by Ridgeview Publishing Company, Atascadero, California, 1991.
SM
Science and Metaphysics: Variations on Kantian Themes, The John Locke Lectures for
1965–1966 (London: Routledge and Kegan Paul, 1968). Republished by Ridgeview
Publishing Company, Atascadero, California, 1992.
SSIS
Science, Sense Impressions, and Sensa: A Reply to Cornman, Review of Metaphysics 24,
3, (1971): 391–447.
MFC
Meaning as Functional Classification, Synthese 27 (1974): 417–437
NAO Naturalism and Ontology (Reseda, California: Ridgeview Publishing Co., 1980).
FMPP Foundations for a Metaphysics of Pure Process (The Carus Lectures) The Monist 64
(1981): 3–90.
ISR
In the Space of Reasons: Selected Essays of Wilfrid Sellars, edited by K. Scharp, R. B.
Brandom. (Cambridge, Mass.: Harvard University Press, 2007).

121

11
PHILIPPA FOOT’S LIBERAL
NATURALISM IN ETHICS
Gabriele De Anna

1 Philippa Foot’s position in ethical naturalism
Ethical naturalism is the general view that an ethical statement expresses a cognitive content
which tracks (a) natural property(ies). Of course, this definition is general: there are many ways
of understanding “cognitive content”, “to track” and “natural properties”, and there are different consistent ways of combining instances of interpretations of those expressions. Hence,
many different versions of ethical naturalism are possible.
Ethical naturalism has been a discredited philosophical position in the English-speaking
philosophical world since the beginning of the twentieth century, when Moore objected to it
with his famous open question argument. The main targets of his argument were the exponents
of Social Darwinism, such as Spencer and Huxley, who had tried to derive ethical prescriptions
from the alleged direction of evolution. Persuaded by Moore’s argument, philosophers typically
divided between those who rejected the view that ethical statements express cognitive contents
and those who rejected the view that ethical sentences are meant to track natural properties.
The former, noncognitivists, took ethical statements to be expressions of emotions or subjective
perspectives, the latter, intuitionists, took moral properties to be nonnatural objects of a sui
generis human intuition.
In the second half of the twentieth century, the interest for ethical naturalism has risen again.
There have been and there still are two mainstreams interested in that general view: evolutionary ethics, which is a comeback of Social Darwinism after Edward O. Wilsons’s Sociobiology.
The New Synthesis (1975), and analytical neo-Aristotelianism, which has grown following in the
footsteps of Elizabeth Anscombe’s work and has been represented, for example, by ethicists like
Philippa Foot, Alasdair Macintyre, Rosalind Hursthouse and John McDowell. Evolutionary
ethicists are normally cognitivists (i.e. they hold that ethical statements can be true or false), but,
a part from a few exceptions – that is, evolutionary-Aristotelian, for example, Casebeer (2003)
and Cornell Realists, for example, Richard Boyd, Nicholas Sturgeon and David Brick –, they
usually claim that ethical theory can only explain human behavior, not justify it (Ruse 2006).
This line of reasoning has led many philosophers to error theory (i.e. the view that all ethical
statements are false) or skepticism (the view that the truth-values of ethical statements cannot be
decided) about moral knowledge (Joyce 2006, Street 2006). Mainstream evolutionary ethicists
meet Moore’s challenge by denying that the natural properties tracked by ethical statements
122

DOI: 10.4324/9781351209472-14

Liberal naturalism in ethics

justify morality. Neo-Aristotelians, by contrast, are more positive about the possibility of justifying ethics: in their view, the Aristotelian notion of virtue does the job, since ethical
statements track traits of character that lead to human flourishing, that is, to the fulfillment of
human nature. In order to meet Moore’s challenge, they often deny that the relevant human
nature is the object of scientific investigation and claim that it is a feature of ourselves that we
understand in our experience of living a human life. In these ways, however, it is often contended that their position is not really a kind of naturalism.
Philippa Foot’s ethical naturalism is a very interesting case, since her very well-known and
widely read book Natural Goodness (2001) has often been read as grounding the virtues on a
biological notion of human nature and human flourishing. Under this reading, her position in
contemporary ethical naturalism would have her laying across the two mainstreams mentioned
above and thereby posing a frontal challenge to Moore’s argument. She would ground the
virtue on human nature to justify ethics and at the same time see nature as an object of scientific
enquiry. We shall see that this reading of her book is wrong, but her account of the role of
human nature in ethical theory (Section 2) leads her to a position which promises to be a more
genuine form of naturalism (although liberal), in comparison to those of other neo-Aristotelians
(Section 3). At the end, I will question whether her naturalism is liberal enough to allow a
convincing account of normativity (Section 4).

2 Foot’s defence of ethical naturalism
The position of Foot in the debate on ethical naturalism remained unchanged from early stages
of her scientific career, throughout her entire philosophical life. In the introduction to her first
collection of papers (Foot 1977), which spans over twenty years, she wrote: “Two themes run
through many of the essays: opposition to emotivism and prescriptivism, and the thought that a
sound moral philosophy should start from a theory of the virtues and vices” (Foot 1977, xiii).
Already at the beginning of her philosophical production (Foot 1958, 1959), Foot complained against the anticognitivist claims of emotivists and perspectivists. Her contention was
that the meaning of a moral statement must have also a descriptive component and that that
component cannot be anything whatsoever, but it has to be appropriate to the emotion or the
perspective expressed by the statement. Her arguments had a Wittgensteinian flavor. She
analysed the grammar of ethical terms and showed that there are some contexts in which they
break down, as when, for example, one tries to say that it is wrong to go around trees right
handed or to look at hedgehogs at moonlight (1958, 107). The upshot is that “it is quite
impossible to call anything you like good or harm” (1959, 120). However, she was ready to go
even further in her criticism to emotivism and perspectivism and to deny that the contents of
ethical statements have an evaluative component that is logically unrelated to the descriptive
component. According to emotivism and persepctivism, the emotional component of the
meaning of a moral statement exercise a guiding or directing force on the hearer. Foot
countered that the full evaluative meaning of a moral statement can be grasped even by
someone who does not feel the relevant emotion or share the relevant perspective, that is,
someone who does not feel in any way under an obligation commanded by the statement.
“I can speak of someone else as having the virtue of courage, and of course recognise
it as a virtue in the proper sense, while knowing that I am a complete coward, and
making no resolution to reform. I know that I should be better off if I were courageous and so have a reason to cultivate courage, but I may also know that I will do
nothing of the kind (Foot 1959, 124).
123

Gabriele De Anna

When it comes to what Foot takes the cognitive content of ethical statements to be, we are led
back to the second theme that she acknowledges as characteristic of her work in the introduction to her fist collection of papers: the thesis that “moral philosophy should start from a
theory of the virtues and vices”. The interest in virtues and vices derives from Foot’s constant
reading of Aristotle and Thomas Aquinas, which she had begun at a very early age under the
influence of her Oxford colleague Elizabeth Anscombe. Foot and Anscombe met daily for years
to discuss philosophy: they had divergent opinions on many issues and Anscombe won usually
their disputes. Aristotle and Aquinas were constant references of Anscombe and Foot was introduced to their writings in this way (Hacker-Wright 2013, 3–4; Hursthouse 2018, 25–27).
Already in the Fifties, Foot held that talk about the virtues is guiding and normative (Foot 1959,
119), but also that it is loaded with cognitive content. One cannot ascribe virtues and vices
regardless from what characteristics the evaluated subject has. The problem is what the criteria
to be used in this evaluation are. According to Foot, “virtues are in general beneficial characteristics, and indeed ones that a human being needs to have, for his own sake and that of his
fellows” (1977, 2). Hence, the evaluation of virtues and vices does not rest on the intuition of
special, sui generis properties, but on judgments concerning what is beneficial to human beings.
“It is surely clear that moral virtues must be connected with human good and harm”, and this
explains why “it is quite impossible to call anything you like good or harm” (1959, 120). In this
way, by starting moral philosophy from a theory of virtues and vices, Foot distances herself from
intuitionists and endorses a form of naturalism: the virtues connect the good to what benefits
human beings.
As I mentioned in the first section, by rejecting noncognitivism and intuitionism at the same
time, one faces the challenge of Moore’s open question argument. Foot was ready to take that
challenge from the start of her writing career, and she carried on the task until the end, when
she proposed the most worked-out version of her ethical naturalism in Natural Goodness (Foot
2001). Then, she acknowledged that her intention had been to propose “a naturalistic theory of
ethics: to break really radically both with G.E. Moore’s antinaturalism and with the subjectivist
theories such as emotivism and prescriptivism that have been seen as clarifications and developments of Moore’s original thought” (2001, 5).
Before considering Foot’s complaints against Moore, it may be worth revising the main
point of the open question argument. The argument contends that, if we identify goodness
with (or reduce it to) a certain natural property N, and if we then ask ourselves “Are all and
only N things good?”, it will always be possible to imagine that there might be Ns that are not
good or good things that are not N. That means that the question “Are all and only N things
good?” is open, but this should not be the case, if good were identical to (or reducible to) N
(Moore 1903, 98 and following). The upshot is that goodness is not identical to, nor reducible
to, any natural property.
In the Fifties, Foot shared Prior’s and Frankena’s thesis that Moors’s argument is problematic: if it is not explained what a “natural property” is, the argument runs the risk of becoming
a truism. Moore’s outlook lacked such an explanation and “a natural property tended to become one not identical with goodness, and the naturalistic fallacy that of identifying goodness
with ‘some other thing’” (Foot 1959, 100). According to Foot, the antinaturalist argument can
really take off only if goodness is thought of as involving a nonfactual, evaluative component.
But then it becomes contentious that goodness cannot be explained in natural terms, since there
are factual conditions for the application of evaluative terms, such as, for example, “rude” (Ibid.,
101–104). It is so that, in an attempt to offer a full answer to the antinaturalist challenge, in the
early Sixties, Foot turns to an analysis of the grammar of “good”, in order to understand what
factual conditions allow deploying that term.
124

Liberal naturalism in ethics

Foot analyses various uses of “good”, and she concludes that there is an ordinary sense of that
word, according to which the choices of the speakers are not among the conditions for the use
of it. She claims that, in its ordinary use, the word “good” does not stand on its own, but it
acquires a meaning in connection to a name to which it must be attached. The features of the
referent of the name set standards and criteria which fill the word “good” with a definite
content. She considers artifacts (e.g. “knife”), functional nouns (e.g. “father”) and organs (e.g.
“roots”) as examples. In all these cases, something is a good sample of the relevant sort, if it
fulfills some requisites set by the end which items of that sort are meant to serve. In order to be
good, a knife must cut well, a father must care for his children, a root must absorb water and
minerals. In all cases of this kind, “good” is an evaluative term, but the application of it depends
entirely on factual, objective conditions of the item. The emotions, perspectives or attitudes of
the speaker who evaluates the item are indifferent. I must recognise that a certain knife is good,
for example, even if I would not chose that particular knife myself, since I do not really cook
and I just want a green objects which matches the curtains in my kitchen. These remarks on the
grammar of “good” show, in Foot’s view, that evaluative terms (and the moral judgments to
which they contribute) have a cognitive content which shapes their evaluative character.
Hence, the antinaturalist challenge can be resisted.
The subsequent phases of the development of Foot’s thought represent a period of rest, from
the point of view of an understanding of her naturalism. Already in her 1961 essay, Foot had
noticed that some uses of the word “good” do not work as the cases related to functional words.
The criteria of evaluation of “good man”, for example, assuming that it has any, are not set by
the end served by men. A man’s organs have the life of the man as an end. However, what is the
end of the entire man? Initially, she was not troubled by this remark, and she was more impressed by the thought that evaluative language works independently from the choices of
speakers. This last observation made her temporarily abandon Anscombe’s notion of action, that
she had been following up to that point, and embrace a more Humean conception (HackerWright 2013, 59–61). If a speaker can acknowledge the moral judgment “X is good” without
choosing X, then one could be a sound ethical reasoner and still be immoral. Hence, in her new
view, morality required more than sound reason: it required also the existence of desires in the
subject, which do the “pumping” and set the agent and her otherwise motivationally inert
moral psychology in motion.
In a reconstruction of Foot’s philosophical itinerary, Hursthouse (2018, 31–33) suggested
that Foot’s new convictions about practical reason caused her to stall for twenty years. In those
years, she wrote on other topics (e.g. on abortion and art), but gave up the attempt to explain
how nature can be a criterion for action and why “good” works differently when jointed to
“man” than when jointed to “knife”. She did some work on Hume on moral judgment,
possibly in the hope of finding a solution to that problem, but she did not really focus on the
problem. She also worked on Kant, in the hope of finding a criterion of right action, but ended
up with a solution which sounded Kantian but was not really so. In her still famous paper
“Morality as a System of Hypothetical Imperatives” (Foot 1972), a paper that she subsequently
disowned (2002, 2), Foot claims that, as rational agents, we build moral systems by balancing
imperatives. However, no imperative is categorical; they are all hypothetical, in the sense that
they are conditional upon the subsistence of relevant desires. This is the upshot of her Humean
take on action: reason by itself is ineffective and needs emotional “pumps” to motivate actions.
In the Eighties, Foot went back to the observation that “good man” works differently from
“good knife”, and to its significance for an understanding of the factual presuppositions of
judgments about goodness. Foot’s interactions with Michael Thompson, whom she was supervising at the University of California at Los Angeles, favored this change in the direction of
125

Gabriele De Anna

her philosophy. Thompson was already developing the conception of Aristotelian natural
goodness that he would publish later (1995, 2008). According to Thompson, a proper account
of the concept of life calls for a new logical category, different from Fregean objects, first order
functions o second order functions. Even in order to identify a certain thing as a living organism, we need to recognise that it is engaged in a certain activity that is typical of the lifeform of its kind. For example, we understand that a certain quantity of matter is a living
organism since it is breathing, and we can do that, since we know that animals breathe. Hence,
we can identify living things, since some universal, necessity statements are true of them, for
example “dogs breathe”, “dogs have four legs”, “dogs are hairy” and so on. These universals,
that Thompson call Aristotelian necessities, are not traditional freegan universally quantified
statements: they are true even if some instances lack the relevant property. For example, it is
true that “dogs have four legs”, even if there are dogs with less than four legs (or even if there
were only dogs with a missing leg left).
Aristotelian necessities ground a form of normativity, Thompson notices. Since dogs are
furry and Fido is a dog, then Fido should be furry. Fido is a defective dog, if it is not furry. In
general, “Fs are Gs” and “a is an F” implies “a should be G”, and “a is good if it is G”.
However, this is just a thin kind of normativity, “natural normativity” in Thompson’s words.
That is to say, it is a kind of normativity that does not guide human action by itself, independently from the occurrence of relevant interests or desires in the agent. Hence, for example, plants of a certain species must have roots (in the natural sense of normativity) but this is
does not bound a certain agent to look after the roots of a plant, unless that agent has some
interest in the life that plant.
Philippa Foot acknowledged that Thompson’s Aristotelian necessities and natural normativity could account for the meaning of “good” in “good man” and could help her explain the
role of facts in the normativity of virtue-language (Foot 2002, 27–32). Concerning the meaning
of “good”, Aristotelian necessities set criteria that individuals of a living species should meet in
order to be good, even if members of that species are not meant to serve a particular end, like
artifacts, organs or functional roles. Aristotelian necessities express what good samples of the
relevant living species should do. Concerning the normativity of virtue language, virtues (and
vices) can be defined as traits of character that benefit (or harm) humans in the sense that they
enable them to act (or prevent them from acting) according to the requirements of Aristotelian
necessities typical of the human species. Foot concludes that
for all the differences that there are, as we shall see, between the evaluation of plants
and animals and their parts and characteristics on the one hand, and the moral evaluation of humans on the other, we shall find that these evaluations share a basic
logical structure and status.
(Foot 2001, 27)
Here Foot refers to living things as a distinct logical category underpinning Aristotelian necessities and suggests that evaluations (i.e. our use of the word “good”) in context different from
functional objects depend on the natural normativity implied by Aristotelian necessities that are
typical of each form of life. For this reason, the logical structure of the evaluation of a plant as
good since it does what plants of that species normally do is the same as the logical structure of
the evaluation of a good man.
The evaluation of humans, however, is different from those of other life forms in an important way (Foot 2001, 39–51). When we evaluate a good plant as a good plant we consider
the activities and the well-functioning of all its organs. In the case of humans, however, things
126

Liberal naturalism in ethics

are different. We would not deny that some human is good on the basis of the fact that her
lungs or her eyes do not work properly. We say that a human is not good, only if she acts badly,
if her will fails. Hence, the logical structure of the evaluation is like that of a plant, but it seems
that with humans there is a break in what we consider a typical life: the typical life of human is
not the exercise of bodily functions, but the activities of her reason and her will. A human being
is good if she employs her will to make choices that make her life fully fulfilled as a human life,
including all the relations with others that one needs to flourish. Hence, murder or promise
breaking are wrong to the extent that they jeopardise such a fulfillment of life.
Aristotelian necessities are the strongest answer to Moore’s antinaturalist challenge that Foot
gave in her philosophical work. Moore contends that goodness cannot be identified with or
reduced to any natural property. Foot replies that “good” works differently in different contexts
and that when it is applied to living things it expresses the fulfillment of the Aristotelian necessities associated with the species to which each evaluated individual belongs. Since
Aristotelian necessities express factual truths about living forms that are also criteria for evaluation, this answer meets the challenge. The antinaturalist, however, could resist this answer. If
the open question arguments risks becoming trivial if “natural property” is not properly defined, Foot’s answer risks begging the question by calling “natural” properties which are really
loaded with values. Are Aristotelian necessities really respectable natural properties?

3 How liberal is Foot’s naturalism?
What is a natural property? One’s conception of nature and of our access to it determines how
one answers this question. If one assumes the thesis that the natural sciences exhaust our
epistemic possibilities in the understanding of nature, then a natural property will be a scientific
property. That thesis, however, is far from self-evident. It was typically assumed by positivists
and it was widespread at the time of Moore. As mentioned, his argument was addressed primarily against Social Darwinists. When he wrote about natural properties he had in mind what
Social Darwinist took (mistakenly) to be a property of nature discovered by science: the
evolution of species toward better forms. Since the Positivist era, and after the contributions of
Neo-Positivism to our understanding of the structure of scientific knowledge, however, the
thesis that natural properties are just scientific properties cannot be simply assumed.
The thesis that natural properties are just scientific properties seems to fly in the face of the
widespread acknowledgement that normativity is everywhere, even in the sciences. In fact,
scientific theories have logical structures and logic is prominently normative discipline. For
example, if (if α then β) and α, then it must be true that β. Since normativity is constitutive of
science and science is a natural mode of knowledge, normativity must be somehow natural.
However, can science explain and justify normativity? This is an open question, and for this
reason the thesis that natural properties are scientific properties cannot be assumed, but must be
argued for. Philosophers who want to support the thesis that natural properties are scientific
properties (i.e. scientific naturalists) often attempt to reduce normative properties to scientific
properties, by claiming that our experience of normativity is illusory (error theories), or by
denying that normative statements have a cognitive content.
Many philosophers, however, are dissatisfied with reductive attempts, and, while holding on
to naturalism, want to claim that there are normative properties in nature, although these are not
scientific properties and cannot be investigate through natural sciences. They are liberal naturalists:
What makes Scientific Naturalism and Liberal Naturalism both versions of naturalism is
that neither countenances the supernatural, whether in the form of entities (such as
127

Gabriele De Anna

God, spirits, entelechies, or Cartesian minds), events (such as miracles or magic), or
epistemic faculties (such as mystical insight or spiritual intuition). The importance of
this for the philosophical approach to normativity is that any form of naturalism will
be opposed to Platonism about norms, where this is understood as the view that
normative facts hold wholly independently of human practices (say, of reason giving)
and are, as it were, simply there anyway waiting to be discovered. For similar reasons
it will be opposed to a Moorean non-naturalism that holds that our access to normative facts is by way of a sui generis epistemic faculty of intuition directed at just this
kind of fact. And of course it will be opposed to any theistic foundation for normative
facts or our access to them.
(De Caro and Macarthur 2010, 3)
This characterisation of liberal naturalism is broad: it aims at capturing a wide range of different,
possible views that can be grouped under that label. It makes clear that liberal naturalists share an
epistemic and an ontological concern. The epistemic concern is to widen the possible methods
of knowledge beyond the strictly scientific ones, without leaping into the “mystical”, or
without introducing “spiritual” or “sui generis” intuitions. Hence, liberal naturalism accepts
methods different from those of the natural science, but requires also that all acceptable methods
deploy only our customary empirical and rational capacities. The ontological concern of liberal
naturalists is not to admit the existence of entities that are incompatible with what we know
about the natural world through the physical sciences. Of course, in principle, the nonscientific
methods that liberal naturalists accept could highlight aspects of reality that the natural sciences
overlook, but this ontological extension should be compatible with what the natural sciences
tell us about nature. In the above passage, the exclusion of God, spirits, entelechies, Cartesian
minds, miracles, magic, Platonic ideas and norms and Moorean moral properties could seem
gratuitous, but it is justified by the assumption that only entities that do not interfere with the
laws of nature are compatible with the natural sciences:
There may be entities (e.g. numbers) that do not and cannot causally affect the world
investigated by the sciences and are both irreducible to and ontologically independent
of entities accountable by science but are not supernatural either, since they cannot
violate any laws of nature (Ibid., 12).
Different versions of liberal naturalism can be differentiated on the basis of how they understand
laws of nature and violations of them. Personally, I believe that the laws of nature are just
descriptions of regularities depending on the exercise of causal powers, and from that point of
view it is questionable that God, spirits, Cartesian minds, Platonic ideas and Moorean moral
properties are all equally incompatible with natural sciences. However, I cannot discuss my
view here (cf. De Anna 2018). I want just to claim that where the divide between the natural
and the supernatural rests is a contentious matter within liberal naturalism.
One of the contentious issues, which weigh in determining where the divide between the
natural and the supernatural falls, is the explanation of value, that is, of the normativity of
human action. Human action is normative in the sense that we do not take all one might want
to do to be to be equally good or acceptable. Fiona Ellis has pointed out that attempts to
account for this intuition in naturalistic terms have called for a progressive enlargement of what
philosophers count as natural. Scientific naturalists had a hard time explaining human action in
purely physical terms, since the physical and neural processes underpinning human action are
too complex to be reduced. For that reason, some philosophers accepted an expanded naturalism,
128

Liberal naturalism in ethics

which was open also to the results of human sciences, like psychology (Railton 1986).
However, Ellis noticed that expanded naturalism can account for our desires and choice
processes, but not for the criteria that allow us to take a normative stance towards them and
evaluate actions as good or bad.
In order to address this problem, some philosophers have accepted an even more relaxed
form of naturalism that Ellis calls expansive naturalism. John McDowell, for example, claimed
that we have two “kinds of intelligibility” to understand nature: the intelligibility through
natural laws “sought by natural science” and “the kind we find in something when we place it
in relation to other occupants of ‘the logical space of reasons’”, that rational beings can share
(McDowell 1994, 70–71). With the latter kind of intelligibility, we can recognise values in
reality that sciences cannot acknowledge. However, this is not to deny the scientific image (in
Wilfrid Sellars’s terms), since values not are incompatible with these existence of laws of nature.
McDowell explains the location of values in Kantian terms and – following Wilfrid Sellars –
refers to a “space of reasons”, which is different form the space of laws (McDowell 1994,
70–86). How to understand the relations between the two spaces and the position of humans in
both is a problem which led many to question the coherence and genuine nature of the ensuing
version of naturalism (Pietroski 1996, Neta 2007).
Now, let us go back to our problem: what kind of naturalism does Foot endorse? Foot and
Thompson’s Aristotelian necessities are not the objects of natural science, so that her view is not
a kind of scientific naturalism. At the same time, such necessities concern living organism and
the natural species to which they belong. Aristotelian necessities belong to the very entities that
natural sciences investigate. Does that mean that Foot is not committed to the existence of
separate “spaces” (unlike McDowell) and that she does not hid references to nonnatural realms
of reality? If that were the case, one could suppose that her naturalism is not as expansive as
McDowell’s and that she holds an expanded naturalism which is open to the ontology of human
and social sciences, but not to the existence of features of reality that the sciences cannot
investigate.
Some philosophers, who have interpreted her views along these lines, have convincingly
argued that such a reading of her position makes it an unpersuasive account of normativity. If
Foot’s naturalism does not make room for an ontology richer than that of the natural sciences,
how can it account for the normativity of human action? By taking natural facts to be just
scientific fact, she is it pushed back into the grip of the open question argument. Joseph Millum
(2006), for example, claims that her account of functions is just a linguistic shortening of longer
evolutionary accounts, but then it does not ground normativity, since functions selected
through evolutionary processes have no particular claim to be fulfilled. Alasdair MacIntyre
(2002) makes an akin remark: if all there is to Foot’s functions is their being part of nature as
conceived by the sciences, then there is nothing telling against the possibility that an atypical
member of a species might act in ways different from the typical members. For example, a
person with a tendency to free ride would have no reason not to follow her instincts, provided
that most members of her group collaborate to common ends.
Other interpreters have countered that Foot’s naturalism is not open to these objections: her
conception of functions should not be taken as a loose phrasing of a biological account. John
Hacker-Wright (2009) has suggested that contra – Millum and MacIntyre – the natural function
which plays the key role in Foot-s account of normativity is not a biological function, but
human rationality. Rationality is a property that we do not grasp by considering humans as
biological organism, but by reflecting on our life from the point of view of our agency and our
practical rationality. “These are what we might call ‘internal observations’ or interpretations;
they are assessment made by a human of the human situation rather than observations that are
129

Gabriele De Anna

merely of the human made from a position of scientific detachment” (Ibid., 320). I agree with
this defence of Foot’s naturalism. Foot and Thompson’s appeal to Aristotelian necessities is not
meant to point to scientific truths about humans, but to a fundamental logical categories. In
their view, living things are only intelligible on the basis of Aristotelian necessities, which are
not the objects of natural science, but conditions for the possibility of the very identification of
the objects of science (i.e. living organisms) by us. However, this remark rises doubts on the
strength of Foot’s naturalism. Are we not now drown back to a “space of reasons” that can be
hardly accounted for in natural terms? So conceived, are not Aristotelian necessities pointing to
facts beyond the facts that sciences can access?
Foot resisted attempts to take her references to rationality as pointing to nonnatural facts,
albeit facts concerning the very objects investigated by the sciences, and eventually reducing
them to natural facts. As Hursthouse has pointed out, “Foot was always a Wittgensteinian,
through and through” (2018, 25). For Wittgenstein, the purpose of philosophy is to get a clear
view of our use of words and “the particular purpose in Foot’s case has always been to get
clearer about our use of words when we are making moral judgments” (Ibid., 25–26).
Following Wittgenstein, Foot holds that philosophy does not enlarge our experience and or
ontology beyond scientific and everyday experience. The fact that we speak about rationality
and about Aristotelian necessities does not imply the existence of special references of our
words. Hence, we are not required to locate those referents in a special sphere (e.g. the space of
reasons) of reality or to reduce them do scientifically acceptable entities. Philosophy only enlightens how our language works – our moral language, in this case.
I would conclude that Foot took herself to be less liberal a naturalist than other Aristotelians,
like for example McDowell. This interpretation is supported by textual evidence. In Natural
Goodness, Foot disagreed with McDowell’s attempt to reconcile the requirements of a life or
virtue with the possibility of suffering for the virtues, by claiming that “happiness” refers to a
state of fulfillment of human life which is independent of the feelings of the subject.
Interestingly, she disagrees with McDowell, since she sees the sense of “happiness” as necessarily
expressing “sentiments of pleasure, pride and honor” and complaints against McDowell’s account of the matter on the basis on Wittgenstein’s dictum that we should “bring words back
form their metaphysical to their everyday use” (Foot 2002, 97–98). As suggested by
Hursthouse, Foot really seems to hope that Wittgenstein’s methodology can allow her to describe the grammar of goodness while holding on to a not too liberal form of naturalism and
while avoiding metaphysical commitments.

4 Foot’s liberal naturalism and the explanation of normativity
Some philosophers, including myself, who are generally sympathetic to a naturalist approach in
metaethics and to many arguments put forward by Foot, have complained that Foot’s account
of normativity is ultimately unsatisfying, even if it can be defended from the objection considered in the last section. In their views, her account fails to explain why the knowledge of
Aristotelian necessities (i.e. natural goodness), including the knowledge that rationality is the
mark of a human life, should bind agents (Hacker-Wright 2018, 10–13; Frey 2018, 58–68; De
Anna 2020a, 109). The existence of natural goodness and the recognition that developing
reason is the fulfillment of human life do not by themselves explain why agents should care
about these facts and elect the realisation of natural goodness as the end of their actions. In order
to explain – as she wishes: see Section 2, above – why the “descriptive component” of the
meaning of a moral statement cannot be “anything whatsoever”, she seems to miss an explanation of the structure of agency, which makes the attainment of natural goals a requirement
130

Liberal naturalism in ethics

for agents. Such an explanation can be found in the analysis of the virtues, including the
intellectual virtues, carried out by Aristotle, and in the elaboration of that analysis suggested by
Aquinas in his theory of the powers implicated in human agency. From this point of view,
Foot’s account of normativity can only be made cogent if it can be completed with an akin
account of the powers of human agency (Hacker-Wright 2020). Now, I would like to make
some comments on the consequence of this line of defence of Foot’s position for the strength of
the underling naturalism.
An enquiry into the powers of human action concerns the capacities that we deploy in
making our decisions, and these include both reasoning capacities and appetitive capacities.
Such an investigation is not necessarily incompatible with a Wittgensteinian methodology
implying the study of the grammar of language, the analysis of language games and the highlighting of features of our form of life. We cannot study our cognitive and appetitive powers by
looking at human action from a third person perspective, since we need to appreciate the
reasons of agents, in order to understand and analyse their actions. Hence, we should analyse
action from the first person perspective. How can we do that? Introspection is not a candidate,
since the powers we deploy in action cannot be identified or re-identified as objects of our
awareness. The required analysis calls for a reflection on the reasoning processes that we carry
out when we make decisions and on the considerations that we bear to weigh when we justify
our deeds or when we criticise those of others. As I have claimed somewhere else, the
Wittesgenteinian analysis of language games and of forms of life can actually be the best way to
carry out a project of this kind (De Anna 2020b). This would be one way to fulfill Foot’s desire
to use a Wittgensteinian account of the grammar of “good” to explain why “it is quite impossible to call anything you like good or harm” (1959, 120) – again, cf. Section 2 above.
If one follows the line I suggest, however, one commits to metaphysics more than Foot
intends to, when she recalls Wittgenstein’s methodology. For the sake of a metanormative
theory, the investigation of our cognitive and appetitive powers must identify a normative
structure in those powers: it must explain what a correct actualisation and articulation of those
powers is. This presupposes that the analysis of our powers exhibits features of an essence of
humanity that all humans share. Indeed, in order for such an account of normativity to take off,
humans must be seen as entities with a sort of second order disposition: a disposition to develop
certain (first order) cognitive and appetitive dispositions. This second order disposition was called
“inclination” by Aquinas (cf. Frey 2018, 69–71). One might suspect that a metaphysical commitment of this kind is incompatible with Wittgenstein’s methodology, as Foot seems to imply.
In my opinion, Wittgenstein’s methodology is not really as antimetaphysical as Foot suggests. Certainly, no one can doubt Wittgenstein’s intention to free us from ontological bulimia,
that is, the simplistic view that whenever we use a word to refer to something, the denoted
thing belongs to a species identified by a real essence expressed by the word we used. He
famously claimed that there is no essence of meaning, since words come in different kinds and
for each kind having a meaning means something different. For example, in some cases having a
meaning means having a referent, in other cases it does not. From his point of view, philosophy
can free us from the false assumption that there are essences of anything we talk about, an
assumption which cases us a lot of conundrums. However, to claim that there is not always an
essence does not mean that there is never one: one can see the very goal of Wittgenstein’s
analysis of grammar as an attempt to understand when the things we refer to with our words are
underpinned by essences. Indeed, in the Philosophical Investigations he claims that “essence is
expressed in grammar” (371). This being so, to claim that we have “to bring words back form
their metaphysical to their everyday use” does not necessarily means that we have to reject
metaphysics altogether. It can also mean that we should test our metaphysical assumptions and
131

Gabriele De Anna

negotiate our ontological commitments by bringing them to account for our overall experience
of the world, that is, by considering how they can express an all-encompassing account of our
form of life and of our linguistic and social practices. It cannot be denied a priori that the output
of this scrutiny processes may be that some metaphysical structures are conditions of the possibility of our linguistic and social practices.
If the points about an explanation of normativity that I have mentioned – and assumed – in
this section are correct, it may be that a Wittgensteinian analysis of human action requires us to
commit to an ontology of the inclination of the powers of agency. If that were the case, we
would have to accept a version of naturalism more liberal than that actually endorsed by Foot.
Such a liberal naturalism would be different from McDowell’s, since it would be open to a
metaphysics of human essence, rather than a metaphysics of reasons. Marie McGinn complained
against McDowell’s interpretation of Wittgenstein’s naturalism that, according to Wittgenstein,
normativity is not just a characteristic of second nature (i.e. the cognitive and appetitive features
that humans develop through their upbringing by being embedded in a culture), but of first
nature (i.e. primitive and cross-cultural human responsiveness) (McGinn 2010, 338). I take this
point a step forward in an essentialist direction. The ensuing position would still be a kind of
naturalism, in the sense that it would not counter scientific facts and scientific explanations. It
would just grant that human beings share a nature which grants them different sorts of properties and powers: some which can be investigated with the methods of the natural sciences and
others (i.e. inclinations) that emerge only in our experience of normativity. To that extent, the
suggested position challenges Moore’s open question augment. The metaphysical commitment
of such a position, however, would not be gratuitous: it would originate from the attempt to
keep together the knowledge of the world that we gain through the natural sciences with our
ethical experience, that is, the experience of the normativity of human action. The ensuing
naturalism is liberal beyond the limits that Foot would have been keen to accept. However, I
want to suggest that, if we want to make Foot’s ethical naturalism succeed, this is the kind of
liberal naturalism that we need accept and that she should have accepted.

Acknowledgments
I am very grateful to John Hacker-Wright for his comments on a previous draft of this essay.

References
Casebeer, W. Natural Ethical Facts: Evolution, Connectionism, and Moral Cognition. MIT Press. (2003).
De Anna, G. Value, Transcendence, Analogy. European Journal for Philosophy of Religion, 10, (2018).
105–129.
De Anna, G. Authority and the Metaphysics of Political Communities. Routledge. (2020a).
De Anna, G. Normativity, Volitional Capacities and Rationality as a Form of Life. Philosophy and Social
Criticism, 46(2), (2020b). 152–161.
De Caro, M. and Macarthur, D. Science, Naturalism, and the Problem of Normativity. In M. De Caro
and D. Macarthur (eds.), Naturalism and Normativity, pp. 1–19. Columbia University Press. (2010).
Foot, P. Moral Arguments. Mind, 67, (1958). 502–513. Now in P. Foot, Virtues and Vices. Oxford and
New York: Oxford University Press, 1977, 96–109.
Foot, P. Moral Beliefs. Proceedings of the Aristotelian Society, 59, (1959). 83–104. Now in P. Foot, Virtues and
Vices, pp. 110-131. Oxford University Press. (1977).
Foot, P. Goodness and Choice.In Aristotelian Society Supplementary Volumes, 35, (1961): 45–80.
Foot, P. Morality a System of Hypothetical Imperatives. The Philosophical Review, 81, 305–313. (1972).
Now in P. Foot, Virtues and Vices. Oxford and New York: Oxford University Press, 1977, 132–147.
Foot, P. Virtues and Vices.Oxford University Press. ([1977]/2002).

132

Liberal naturalism in ethics
Foot, P. Natural Goodness. Oxford University Press. (2001).
Foot, P. Moral Dilemmas. Oxford University Press. (2002).
Frey, J. How to Be an Ethical Naturalist. In John Hacker-Wright (ed.), Philippa Foot on Goodness and
Virtue, pp. 47–84. Palgrave Macmillan. (2018).
Hacker-Wright, J. What is Natural about Foot’s Ethical Naturalism? Ratio, 22, (2009). 308–321.
Hacker-Wright, J. Philippa Foot’s Moral Thought. Bloomsbury. (2013).
Hacker-Wright, J. Introduction: From Natural Goodness to Morality. In John Hacker-Wright (ed.),
Philippa Foot on Goodness and Virtue, pp. 1–24. Palgrave Macmillan. (2018).
Hacker-Wright, J. Passions, Virtue, and Rational Life. Philosophy and Social Criticism, 46(2), (2020).
131–140.
Hursthouse, R. The Grammar of Goodness in Foot’s Ethical Naturalism. In John Hacker-Wright (ed.),
Philippa Foot on Goodness and Virtue, 25–46. Palgrave Macmillan. (2018).
Joyce, R. The Evolution of Morality. MIT Press. (2006).
McDowell, J. Mind and World. Harvard University Press. (1994).
MacIntyre, A. Virtues in Foot and Geach. The Philosophical Quarterly, 52, (2002). 621–631.
McGinn, M. Wittgenstein and Naturalism. In Mario De Caro and David Macarthur (eds.), Naturalism and
Normativity, pp. 322–351. Harvard University Press. (2010).
Millum, J. Natural Goodness and Natural Evil. In Ratio, 19, (2006). 199–213.
Moore, G. E. Principia Ethica. Cambridge University Press. (1903).
Neta, R. Review of Naturalism in Question by M. De Caro and D. Macarthur (eds.), The Philosophical
Review, 116, (2007). 657–663.
Pietroski, P. M. Critical note on Mind and World by J. McDowell. Canadian Journal of Philosophy, 26(4),
(1996). 613–636.
Railton, P. Moral Realism. The Philosophical Review, 95, (1986):163–207.
Ruse, M. Is Darwinian Metaethics Possible (And If It Is, Is It Well Taken)?. In Giovanni Boniolo and G.
De Anna (eds.), Evolutionary Ethics and Contemporary Biology, pp. 13–26. Cambridge University
Press. (2006).
Street, S. A Darwinian Dilemma for Realist Theories of Value. Philosophical Studies, 127, (2006). 106–166.
Thompson, M. The Representation of Life. In Hursthouse, R. , Lawrence, G. and W. Quinn (eds.),
Virtues and Reasons. Philippa Foot and Moral Theory, pp. 247–296. Clarendon. (1995).
Thompson, M. Life and Action: Elementary Structures of Practice and Practical Thought. Harvard University
Press. (2008).
Wilson, E.O. Sociobiology: The New Synthesis. Harvard University Press. (1975).

133

12
BERNARD WILLIAMS’S LIBERAL
NATURALISM
Sophie-Grace Chappell

1 Part I
“Naturalism” is a word that has meant many things in its history. Not all the things that
philosophers have meant by “naturalism” are both sensical and sensible; some are neither. For
instance, there has been a tendency to state naturalism as the thesis that “Only the natural exists”
(or: “can be known”), where “the natural” means either “the physical”, or “the scientific”, or
“part of complete science” (whatever that may mean), or all of these. Down this road, serious
difficulties quickly emerge for naturalism – as for its familiar bedfellows physicalism and scientism – about stabilising definitions for any of these terms that will credibly serve the role that
the argument is supposed to give them; false-or-trivial dilemmas loom.1
Relatedly, it is hard to believe that there are no beliefs, as some “eliminativist” naturalists
have sometimes urged us to (most famously, Churchland (1981)). That is not merely because of
the apparent self-refutation involved in even trying to disbelieve in belief.2 It is also because the
rationales that have been given for abandoning as unscientific – and therefore (?) fraudulent –
the concept of belief, along with the rest of “folk psychology”, seem either implausible or
confused. Whether or not folk psychology “is a theory” (an issue that often seems at best a sideissue; though see some remarks at the end of Part III about the status of theories in ethics and
science), the claim that folk psychology lacks predictive power just looks spectacularly false.3
On the face of it, folk psychology is corroborated every time we post a letter, solve a cryptic
crossword, or indeed publish a philosophy essay. Again, as I’ve pointed out myself, the claim
that folk-psychological explanations somehow conflict with other explanations, and in particular with neuroscientific ones, looks rather like the muddle illustrated by imagining that there
is somehow a conflict between “The clock is chiming four because of the interaction of its cogs
and gears”, “The clock is chiming four because the workshop made it that way”, “The clock is
chiming four because I wound it up last night”, “The clock is chiming four because I set it last
night”, and “The clock is chiming four because it is four”4; or by imagining, perhaps under the
spell of what in Chappell (2015) I called “the curse of the definite article”, that just one of these
because-clauses has to be taken as somehow the explanation, all the others being in some sense
bogus or illusory.
Such reflections have justified a suspicion whether “naturalism” is a worthwhile philosophical term at all. But Bernard Williams, for one, calls it “a useful word”5:
134

DOI: 10.4324/9781351209472-15

Bernard Williams’s liberal naturalism

A naturalistic view of ethics was previously contrasted with a supernaturalistic view,
and it meant a view according to which ethics was to be understood in worldly terms,
without reference to God or any transcendental authority.
This on its own suggests a naturalism that by any normal standards is extremely liberal, meaning
“generously and humanely inclusive”. (“Liberal” is ambiguous between this “generous and
humane” sense and a political sense that is irrelevant here.)6 If naturalistic ethics simply means
ethics without God or transcendental authority, then counter-intuitively, Immanuel Kant’s
enlightenment-humanistic ethics of autonomy counts as a form of naturalism. So too, equally
counter-intuitively, does Jonathan Dancy’s recent particularism – which is taken by most
commentators, including Dancy himself, to be a clear instance of non-naturalism.7
However, this is not Williams’s only indication of what he intends by “naturalism”. In the
next chapter of the same book, he homes in rather more closely on one contrast in particular:
that between ethics and science.
[I]n relation to ethics there is a genuine and profound difference to be found, and
[this] is enough to motivate some version of the feeling…that science has some chance
of being more or less what it seems, a systematised theoretical account of how the
world really is, while ethical thought has no chance of being everything it
seems…The basic difference lies…in our reflective understanding of the best hopes
we could coherently entertain for eliminating disagreement in the two areas. It is a
matter of what, under the most favourable conditions, would be the best explanation
of the end of disagreement: the explanation—as I shall say from now on—of
convergence.8
In the best case people will of course come to agree in both ethics and science. And they will
come to agree, not because they come to share some muddle or mistake, but because their
understanding is as good as it could be. How good is that?
In a scientific enquiry there should ideally be convergence on an answer, where the
best explanation of the convergence involves the idea that the answer represents how
things are; in the area of the ethical, at least at a high level of generality, there is no
such coherent hope.9
As we might put it: there is progress in science, there is only consensus in ethics. For in science
there is something to progress toward, even though it is of course an ideal. The targeted ideal is
what Williams calls “the absolute conception of the world” meaning a full and completely
accurate “systematised theoretical account of how the world really is”.10 Science has that to
converge on, but there is nothing equivalent for ethics to aim at. And that is why Williams also
said, in some of the last and most interesting remarks that he published on how to understand
(his) naturalism, that to be a naturalist is just “to adopt an appropriately suspicious rule of
method: never explain the ethical in terms of something special to ethics if you can explain it in
terms that apply to the non-ethical as well”.11
Furthermore, says Williams in the same discussion, the scientist’s own understanding of what
she herself is doing as a scientist is, or can be, part of this progress toward convergence. She can
correctly understand her own present scientific practice as directed – intentionally, now, by her
– toward finding the truth. At the same time, she can also correctly understand the history of
how scientists like her have come to practise science as she now practises it, as a history of
135

Sophie-Grace Chappell

convergence on methods that are preferred because they are good (not necessarily perfect, but
good) methods for finding the truth.
In this sense science explains the world and everything in it, including science itself.12 And it
is the nature of scientific understanding, not only that it explains scientific practice but also that
it vindicates it: “It is an important feature of modern science that it contributes to explaining how
creatures with our origins and characteristics can understand a world with properties that this
same science ascribes to the world”.13 By contrast, ethical thought faces “serious problems”
about “the way in which it can understand its own nature and the extent to which it can
consistently appear to be what it really is”.14
More specifically: because in ethics there is (says Williams) nothing like the absolute conception to aim at, four crucial consequences follow, all of which provide stark contrasts between
ethics and science. First, convergence in ethics is not necessarily, even in the best case, convergence by rational means on “the way things really are”; not at least in the way it is in science.
Secondly, a full understanding of our own ethical practice can be explanatory, but will not be
vindicatory, or not at least in the way that a scientific understanding of scientific practice is
vindicatory; in fact it is likely to be subversive in important ways. Thirdly, and because of these
first two points: though Williams accepts that there is something worth calling ethical knowledge, still the main place in which we are likely to find that ethical knowledge is in the past or in
remaining fragments of the past – in “hypertraditional societies” or the closest extant approximations to them15; and in the psychological equivalents of these political structures, namely
innocent people of good character, who have virtuous dispositions but have not reflected on
what it is to have virtuous dispositions.16 Fourthly, and as a consequence of all three prior points:
whereas increased understanding and reflection in science typically lead to increases (both
qualitative and quantitative) in scientific knowledge, increased understanding and reflection in
ethics can have exactly the opposite effect: it can lead to decreases in the quantity and the quality
of the ethical knowledge that is available to us. “There is some [ethical knowledge], and in the
less reflective past there has been more”17; but in ethics, “reflection” can “destroy knowledge”.18
Williams counts as a naturalist because, first, he takes the considerations just rehearsed to be
the essential ones in setting the stage for the basic question of metaethics, which he frames as the
question how and where “the ethical” fits in with the rest of our world-view. And he counts as
a naturalist because, second, he thinks that answering this question is, so to speak, the payment
of an explanatory debt that humane understanding owes to science. For any naturalist, there is
such an explanatory debt, and it runs in that direction: from humane understanding to science.
It is not that the debt-relation is real but the other way around, as some philosophers have
argued. (Perhaps Heidegger’s views on technology and human life, in Being and Time and
elsewhere, are the clearest example.) Nor is it that there is no explanatory debt-relation in either
direction, as some other philosophers have apparently thought. (Maybe the later Wittgenstein
thought this; certainly some of his quietist followers have believed it.) Rather, Williams thinks
that humane understanding, and specifically ethics, has a genuine explanatory debt to science,
and that this cannot be paid in full without some subversion of our prephilosophical assumptions about ethics.
Thus there are a number of reasons why “ethical thought has no chance of being everything
it seems”. Perhaps the most basic reason of all concerns our “thick” ethical concepts, the
descriptive-cum-prescriptive concepts that frame and articulate our whole ethical world-view.
Williams’s thesis is that we cannot fully understand how and why we have come to have the
particular thick ethical concepts that we do without coming to see them as at least to some
extent historically contingent, perhaps arbitrary too, and in either case not deeply based in “the
nature of things” in the way that the basic concepts of science are.
136

Bernard Williams’s liberal naturalism

However, Williams counts as a liberal naturalist because he denies emphatically that paying
ethics’ explanatory debt to science means unmasking it as a misleading or fraudulent discourse
from top to bottom. My second paragraph mentioned how some naturalists have taken this
iconoclastic view of “folk psychology”. Other naturalists have been no less iconoclastic about
ethics. Streumer (2017), for example, is prepared to argue that not just all ethical but all
normative propositions without exception are false. Streumer concedes that his thesis is not one
that anyone can believe, and does not take that to be an objection to discussing his view.
Presumably, however, he also thinks that there is no reason to discuss his view; so I’ll stop there.
Less recently, and less extremely, J.L. Mackie has famously argued that moral discourse as we
have it is systematically falsified by its untenable, quasimagical presuppositions. The vigour with
which Mackie (1978)’s Chapter 1 argues for this negative thesis is decidedly overshadowed by
the enthusiasm with which Chapters 2–10 argue their way back toward, and into, the positive
project of “invention” mentioned in his subtitle, the project of (re)constructing something very
like moral discourse as we have it, without the magic and on the basis, roughly, of a contractualist consensus.
Williams, anyway, seems not to be this kind of iconoclast. In most of his work, and on the
whole – and with some reservations that we’ll come to in part III of this essay – his naturalism is
much less radical and destructive than the naturalism of Churchland or Mackie or Streumer (or
Quine, perhaps the real original begetter, at least in post-war analytic philosophy, of iconoclastic naturalism). Indeed it is just as apt to call Williams a conservative naturalist as a liberal one –
once more, in a non-political sense: the sense that one of his main questions, throughout his
ethical philosophy, is “Given our other knowledge, how much of our ethical knowledge can
we keep, or reform enough to salvage it?”
It is with this conservative question in mind that we should read Williams’s epigraph to
Ethics and the Limits of Philosophy, from Wallace Stevens’ poem Esthétique du Mal, which speaks
both of “the mortal no” and “the cold vacancy When the phantoms are gone”, but also of “the
imagination’s new beginning”, “the yes of the realist spoken because he must say yes”, and the
“passion for yes” that underlies “every no” and “that ha[s] never been broken”.19 The same
question is audible in Williams’s often-quoted and highly quotable remark that “Theory typically uses the assumption that we probably have too many ethical ideas, some of which may
well turn out to be mere prejudices. Our major problem now is actually that we have not too
many but too few, and we need to cherish as many as we can”.20
This conservative “What can be salvaged?” question is pervasive, together with a reasonable
degree of cautious optimism about answering it, in Making Sense of Humanity (Williams (1995)),
the last volume of his collected essays on ethics published in his lifetime. In most of these papers,
the underlying question is indeed how much of our humanity we can make sense of – a
question for which, as Williams demonstrates in winning detail, we need the resources of
psychology, sociology, history, and the humanities, as well as those more usually deployed in
philosophical inquiry. It is central also to Williams’s last monograph, Truth and Truthfulness
(Williams (2002)), which involves a particularly ambitious constructive project, namely a
vindicatory genealogy – again, as far as possible – of the two things in its title.
The question is there too in Williams’s earlier campaigns against what he believes cannot be
salvaged – his critical engagements with two test-pieces for meta-ethical naturalism, the internal
reasons thesis and neo-Aristotelian eudaimonism.21 The internal reasons thesis – Williams 1981
is the canonical formulation – is the thesis that we can have no justification for action that is not
intelligibly connected to our pre-existing motivations: crudely but vividly, our reasons depend
on our wants. Many moral philosophers, notably utilitarian ones, have had the ambition to
identify reasons for all of us that go beyond any possible intelligible extension of anyone’s wants;
137

Sophie-Grace Chappell

or they have said, as contemporary Kantians quite often do, that our wants depend on our
reasons, not vice versa. Williams thinks that this philosophical ambition, to identify external
reasons and to build an edifice of normative theory and moral psychology upon such reasons, is
a “phantom”. The idea of an external reason is an idea that cannot be sustained as really
intelligible, given serious and careful investigation in the light of what we know about ourselves, both from science and from the humanities. As Diamond would put it, following
Anscombe, the attempt to articulate such an idea “peters out into nothing”.22
The same, says Williams, goes for neo-Aristotelian eudaimonism, the view that the ethical
good for humans is to be understood as the natural flourishing of human beings, considered as a
biological (but also a sociological and political) species. Williams thinks that this is the closest
approach we have to a feasible ethical theory: “it represents the only intelligible form of ethical
objectivity at the reflective level”.23 But it does not come close enough to be actually feasible. For
the eudaimonist’s philosophical programme to be really intelligible to us, we would need both to
have a better grip than is actually possible on the notion of an external reason – see above – and
also to have retained beliefs from Aristotelian teleological science that no serious or reflective
person can possibly retain, once she has come to appreciate the essentially magical character of
Aristotelian teleology, and also the drastically non-teleological nature of modern science.

2 Part II
How might we challenge Williams’s naturalism, and the relativism that we have seen results
from it? I begin with two very simple objections that can be applied generically, not only to
Williams’s argument, but to most naturalism-based arguments for relativism.
First is the charge that Williams commits a non sequitur in arguing from disagreement about
ethics to relativism in ethics: that people have conflicting views whether p shows nothing even
about the truth or falsity of p, let alone its absolute or relative status. But to this familiar
complaint there is an equally familiar answer, and it is certainly the answer that Williams would
give: the relativist’s point is not merely that there is disagreement, but that there is – in some
worrying way – ungrounded disagreement. (For some evidence that ethical disagreement is not,
at any rate, firmly or directly grounded in any very convincing way, see the quotation from
Williams coming in two paragraphs’ time.)
A second generic objection raises the empirical question whether there is, in fact, enough
transcultural variation in ethical thinking to justify the relativist worry. The real problem, according to (for example) Derek Parfit and Peter Singer, and before them Richard Hare, is not
that people disagree when they think ethically.24 It is that they never get as far as thinking
ethically in the first place: their thoughts are pre-empted by prejudice, tradition, religion, and
other ideological elements that Parfit and these others take to be essentially irrational, and
indeed antirational, in a way that ethics is not. Clear away these distractions, and immediately
we will move a lot closer toward ethical agreement – and toward rationally motivated agreement:
so toward convergence and not mere consensus.
To judge by the briskness with which he sometimes dismisses this second objection to
relativism, Williams evidently sees it as resting on whiggish, meliorist, and intellectually imperialist presuppositions that are themselves decidedly ideological. But the objection deserves
more than a brisk response; and Williams does sometimes give it that. Hence, for instance, this
comment on the French Revolution:
When we consider how [politically liberal] forms of argument came to prevail, we can
indeed see them as having won, but not necessarily as having won an argument. For
138

Bernard Williams’s liberal naturalism

liberal ideas to have won an argument, the representatives of the ancien régime would
have had to have shared with the nascent liberals a conception of something that the
argument was about, and not just in the obvious sense that it was about the way to live
or the way to order society. They would have had to agree that there was some aim,
of reason or freedom or whatever, which liberal ideas served better or of which they
were a better expression, [but] the relevant ideas of freedom, reason, and so on were
themselves involved in the change. If in this sense the liberals did not win an argument, then the explanations of how liberalism came to prevail are not vindicatory…a
crisis [in science] is agreed by all parties to be a crisis of explanation, and…to a
considerable extent there has come to be agreement [on what will count as a successful
explanation]. But…the processes by which the old political and ethical order has
changed into modernity…were not…crises of explanation. They were crises of
confidence or of legitimacy, and the story of how one conception rather than another
came to provide the basis of a new legitimacy is not on the face of it vindicatory.25
Whether this is a successful rebuttal of more whiggish, and more rationalistic, views of the
history of ethics like Parfit’s is not something that I have space to settle here. And it is precisely
because there is so much to say about it that the exchange is worth noting. The question about
the extent of actual ethical disagreement, and the suggestion that the roots of most major
disagreements in ethics are ideological rather than rational – these both need careful and detailed
investigation not only by philosophers but also by historians and sociologists. From philosophers
at least, they do not usually get this. It is a merit of Williams’s work that he does more than most
to address this neglect, and to take seriously the idea that philosophical views always can and
often should be answerable to historical and sociological evidence.
Leaving aside these two generic objections to relativism, I turn to objections specifically to
Williams’s relativism. In this “the basic idea”, as Williams himself says, “is very simple”: science,
he holds, is truth-guided and truth-targeting in a way that ethics is not and cannot be.26 A
reflective understanding of this contrast cannot fail to have at least some subversive effects on
our ethical consciousness. Not everything can be salvaged, and ethics, to say it again, “has no
chance of being everything it seems”.
So the “basic idea” is a contrast, and the most obvious objections to that contrast are three:
1
2
3

pessimism about scientific realism – the denial that science is truth-guided and truth-targeting;
optimism about ethical realism – the claim that ethics is truth-guided and truth-targeting
too; and
optimism about the contrast – a claim that the contrast between science and ethics, genuine
though it is, does not really have any subversive impact.

All three of these responses (to Williams, and to others who offer a rather similar outlook) are
represented in the literature. (1) Richard Rorty is happy to say that science too is just another
form of human conversation, with no distinctively strong relation to “ultimate truths” or “how
the world is anyway” or “the absolute conception of the world”; none of these are things that
Rorty believes in anyway.27 (2) Nicholas Sturgeon is an example of an ethical realist who thinks
that ethical facts are just as “hard” and “objective” as scientific ones: thus Sturgeon argues that
the badness of slavery is a fact with causal powers no less real or definite than, say, the fact that
an object is made of silicon or has a mass of 33 kg.28 (3) By far the largest number of responses
to Williams have been along the lines that ethical and scientific understanding are indeed quite
different, but that the differences do not undermine the prospects for a good deal more ethical
139

Sophie-Grace Chappell

truth and ethical knowledge than Williams allows: this is, for instance, the position of Jardine,
Hookway, McDowell, Thomas and Wiggins.29
As this lightning tour of the literature already shows, it is hard to spell out (3) without its
turning into, or ending up conjoined with, some variant of (1) or (2). On the other hand, many
defences of (1) and (2) that do not also involve an acknowledgement of something like (3) have
been rather implausible. It is unconvincing, for instance, to claim as Rorty (ultimately) does
that, for all its high prestige in our culture, there is nothing epistemically special about science,
and so that (in the end) not even the success of science can be explained by reference to the way
the world is – that success too is merely an oscillation in our folkways. (These days we are
noticing political and social reasons, too, why hardline relativism about truth in science is not
such a great idea.)
A different challenge to Williams that is broadly in the style of (3), but borrows also from (1)
and (2), begins from Williams’s key notion of a thick ethical concept: the notion of a
descriptive-cum-prescriptive ethical concept that frames and articulates our whole ethical
world-view. (Paradigm examples are things like BRAVERY, LYING, PROMISING,
COURTESY, CRUELTY.30) Williams tells us that the doubt raised by thick ethical concepts
is simply that, as any reflective person can see, they manifestly have histories; that these histories
must at least to some extent be histories of accident and contingency; and that different thick
concepts have different histories in different societies. This history can of course continue into
our own experience: we as individuals can both acquire and lose a grip on particular thick
concepts. Children – and adults – can be and are taught COURTESY or GENTLENESS or
COURAGE; at the first of his three trials Oscar Wilde said, with apparent sincerity, that
“blasphemous” “is not a word of mine”31, though as a boy Wilde was undoubtedly taught to
have BLASPHEMY as one of his thick concepts.
Whatever changes they may undergo, the thick ethical concepts that we have at any time
are, so to speak, the only lenses through which we can then see, frame, articulate, and understand the ethical world at all. So accident and contingency are an ineliminable part of our
ethical vision. The way we are bound to see things ethically is, to be sure, necessary for us; yet it
is not necessary. It is a product of our history, would be different if our history had been
different, and it need not be consistently translatable32 into the terms of anyone else’s ethical
vision at all. Hence, the problem that Williams sees about the idea of convergence on ethical
knowledge. If and when I cotton on to your application of your thick concepts, that will not be
because I come to be more guided by the world in how I see things. It will be because I come to
be more house-trained in the ways of your culture. Our coming to agree on the verdict that, say,
Boris has been discourteous to Nicola is not us converging on a fact in the world; it is us
coming into consensus about how to see the world.
Here we revisit the two generic objections to relativism that I began with, only this time in
versions more customised to thick concepts. Does a relativist worry really follow from the mere
fact that different cultures have conflicting thick ethical concepts? (Answer: “Yes: at any rate a
worry follows – a problem but not necessarily an insoluble problem – if the differences are rationally
groundless”.) And do different cultures really have conflicting thick ethical concepts? Mightn’t
there really be a good deal of convergence in their ethical concepts, once we isolate the strictly
ethical from the detritus of tradition and superstition with which it has become contaminated? The
answer to this second question is, as before, that it is a large, interesting, and not exclusively
philosophical issue. On the philosophical side of the issue, one thing that certainly needs registering
is a doubt about the idea of “isolating the strictly ethical” from ideological “contaminants”.
But I won’t explore that doubt here, because there is a different objection that I want to
raise. I want to ask: What if our basic scientific concepts are just as much cultural products as
140

Bernard Williams’s liberal naturalism

our thick ethical concepts? Won’t those scientific concepts then be just as much groundless, and
just as much in at least potential rational conflict, as the thick ethical concepts?

3 Part III
Take the case of disagreement (ethical disagreement, if that is what it is) about slavery. To us
today, the very idea of chattel slavery is a paradigm instance of the thick ethical concept
INJUSTICE33; where the institution of slavery has a racial and indeed racist aspect, as it clearly
did in the US Confederacy, that makes things ethically worse, not better. But notoriously, the
Founding Fathers of the USA saw things differently. It was possible for them both to be slaveowners themselves, and to co-sign the ringing and historic declaration that “We hold these
truths to be self-evident, that all men are created equal, that they are endowed by their Creator
with certain unalienable Rights” – et cetera.
How exactly did the US Founding Fathers see things differently? A proper answer to that
question would take us back to the historical and sociological inquiries that, I have said, need to
go alongside, and be joined up with, our philosophical thinking about the issue of relativism.
But for the sake of the present argument, let me suggest (in defiance no doubt of at least the de
dictos of literal history) that their thought can be represented like this: they had a thick ethical
concept of INJUSTICE which projected differently from ours.
To see what I mean by that, take some three cases of slavery C1, C2, C3 to each of which
we apply our thick ethical predicate UNJUST; and suppose that Jefferson agrees with us in
these applications – he too calls all three of C1–C3 unjust. Like us, moreover, his basic reason
for calling them unjust is that he thinks they are cases of injustice because they are instances of a
general truth that looks something like this:
(G) To enslave a man is unjust.
But now take C4, in which it is a black person who is enslaved. We say that C4 too, sharing
as it does all the morally relevant properties of C1–C3, is unjust, and also that it is one way
worse than C1–C3: namely, it is an instance of another thick ethical property, RACISM. But
Jefferson goes in exactly the opposite direction about C4. He says (let us suppose) that what
happens in C4 is not unjust, precisely because it makes a racial discrimination.
We protest to Jefferson that he is treating C4 inconsistently with his treatment of C1–C3.
He acts inconsistently, we tell him, because he is not “going on in the same way” with the thick
ethical concept INJUSTICE: he applied that concept to C1–C3, so it is straightforward that he
should project his concept to C4 as well, because there is no relevant difference in C4. To this
Jefferson simply replies that there is a relevant difference in C4, namely, a difference of race.
Given that difference, he says, he is completely justified in treating C4 differently; it simply isn’t
a case of the same kind. (Compare the ways in which we might deny, for instance, that I can be
either just or unjust to myself, or to a statue. These might look like cases for a straightforward
projection of our concept of (IN)JUSTICE at first sight – but not after further reflection.)
So for us, “going on in the same way” with the relevant thick ethical concept INJUSTICE
means treating C4 the same as C1–C3; but for Jefferson it means the opposite. And what that
shows, apparently, is two decidedly relativistic results. First, we have a different thick concept
from Jefferson, and his concept and ours conflict. Secondly, there is no obvious way of discriminating between Jefferson’s rationality in his deployment of his concept, and our rationality
in deploying our own. By our lights, to be sure, Jefferson has failed to project the thick ethical
predicate UNJUST, to “go on” using that predicate “in the same way”. So he is certainly
irrational if he is using the same thick ethical predicate as we are. But that’s exactly the point: he is not
operating inconsistently with our thick ethical predicate, according to which the addition of
141

Sophie-Grace Chappell

RACISM to the mix in no way “silences” INJUSTICE – on the contrary, it amplifies it.
Rather, he is operating consistently with his thick ethical predicate, according to which the
addition to the mix, not indeed of our thick concept of RACISM, but of some eighteenthcentury concept about race of Jefferson’s own, does “silence” INJUSTICE. Hence the crucial
point: Jefferson’s way of projecting is no more rationally ungrounded than ours (it is also, of
course, no less rationally ungrounded than ours). As far as rationality goes, there is nothing to
choose between his concept and our concept.
So far, so supportive of Williams-style, thick-ethical-concepts-based ethical relativism. But
now notice that we can describe Jefferson’s beliefs and behaviour in another way. We have
supposed so far that Jefferson, while apparently agreeing with us on (G) “To enslave a man is
unjust”, parts company with us in cases like C4 because his thick ethical concept of INJUSTICE projects differently from ours. However, Jefferson could equally part company with us in
cases like C4 for a quite different reason. Jefferson could uphold both his commitments, to
being a slave-owner and to the generalisation (G) “To enslave a man is unjust”, not by projecting his thick ethical concept INJUSTICE differently from us, but by projecting his concept
of MAN differently from ours, so that black people do not come under that concept.34
Now Jefferson’s behaviour obviously has ethical consequences – and pretty abhorrent ones
too. Still, the concept MAN is not in any obvious way a thick ethical concept. On the face of it,
MAN – HUMANITY as we now more correctly say – is a scientific concept: it is, in fact, a
typical example of a working scientific concept that is also a working concept of ordinary life.
So we can get relativistic results about the rational underdetermination of our concepts not only
in ethics, but in science too. Certainly the original case-study for this “Kripkensteinian” argument about the projectibility of concepts was not ethical; it was the mathematical concept
PLUS. But Kripke 1982’s argument that there is no difference as to rationality between working
with different concepts that may look the same, but project differently, is an argument that is
meant to be completely general. So it applies to ethical concepts, and to scientific concepts too.
Kripke’s argument is not, of course, the only kind of argument that has been advanced for a
skeptical doubt about the possibility of giving any kind of rational privilege to any particular
way of projecting our concepts. There are also Goodman’s famous doubts about which classes
of entities scientific induction is supposed to generalise over, and with which predicates (are
emeralds green, for example, or are emerires bleen?), doubts that lead us toward a general
scepticism about the notion of there being unique classes of sequences in the world that are the
inductive regularities.35 There are Quine’s doubts about the indeterminacy of translation and
reference, based on the thought that there is always, in any natural language, going to be some
indeterminacy as to what kind of entities its referring terms actually refer to.36 And there is
Putnam’s “model theoretic argument”, which perhaps is not noticed as much as it should be
because nontechnical philosophers like me are put off by Putnam own highly technical
statement of it.37 But the argument can be informally stated with at least approximate accuracy.
This is Drew Khlentzos’ summary:
According to the Model-Theoretic Argument, there are simply too many ways in
which our mental symbols can be mapped onto items in the world. The consequence
of this is a dilemma for the realist. The first horn of the dilemma is that s/he must
accept that what our symbols refer to is massively indeterminate. The second horn is
that s/he must insist that even an ideal theory, whose terms and predicates can demonstrably be mapped veridically onto objects and properties in the world might still
be false, i.e., that such a mapping might not be the right one, the one ‘intended’.
Neither alternative can be defended, according to anti-realists. Concerning the first
142

Bernard Williams’s liberal naturalism

alternative, massive indeterminacy for perfectly determinate terms is absurd. As for the
second, for realists to contend that even an ideal theory could be false is to resort to
unmotivated dogmatism, since on their own admission we cannot tell which mapping
the world has set up for us. Such dogmatism leaves the realist with no answer to a
skepticism which undermines any capacity to reliably represent the world, anti-realists
maintain.
(https://plato.stanford.edu/entries/realism-sem-challenge/#3.5)
Alongside these famous examples from the recent history of philosophy, a fifth kind of sceptical
doubt about the ontology of science concerns how we describe evolutionary processes. We are
readily impressed by examples of convergent evolution, that is, cases where we are inclined to
say that “the same environmental problem” has been resolved by creatures with quite different
phylogenetic ancestries via “analogous adaptational solutions”. This way of talking is natural
enough, and no doubt unobjectionable in itself. But it understates how much, ontologically and
conceptually speaking, is up for grabs in evolutionary processes and, therefore, in evolutionary
explanation. To describe something as a problem for some creature that is solved by adaptations
in it is a way of talking that presupposes a particular scale and a particular resolution in our
picture of things. But it is only at that scale, and at that resolution, that either the problem or the
solution, or indeed the creature, so much as exists. And where an environment is resistant and
unhelpful at that scale and resolution, one thing that evolutionary processes can and often do try
is simply to get round the resistance by changing the scale and the resolution at which they
operate. To talk about “convergent evolution” is to choose to focus on a particular scale and
resolution at which biological events are happening; but that choice is precisely that, a choice,
and at other scales and resolutions quite different events, and therefore quite different “convergences” and “divergences”, would come into view. (And perhaps something parallel can be
said about discovery in science: maybe scientific discovery proceeds most spectacularly when it
stops trying to answer the question as set, and instead reframes the question and the entire
conceptual apparatus that surrounds it.)
Given these various kinds of consideration, the right conclusion seems obvious: rational
ungroundedness in our conceptual scheme, of the kind that bothers Williams in ethics, is not a
feature of ethics alone. The concepts and the ontology of science too are a welter of rational
indeterminacy, of cases where it is a matter of little more than historical happenstance that we
have the concepts that we do and not others. If Williams is a relativist about the thick ethical
concepts because they are historically accidental and not rationally grounded, then he has exactly that
same reason to be a relativist about the working concepts of science also.
Though Williams never to my knowledge addresses this problem explicitly, he perhaps sees
it coming38:
…the absolute conception will…be a conception of the world that might be arrived at
by any investigators, even if they were very different from us. What counts as a
relevant difference from us, and indeed what for various levels of description will
count as “us”, will, again, be explained on the basis of the conception itself; we shall
be able to explain, for instance, why one kind of observer can make observations that
another kind cannot make. It is centrally important that these ideas relate to science,
not to all kinds of knowledge. We can know things whose content is perspectival: we
can know that grass is green, for instance, though green, for certain, and probably grass
are concepts that would not be available to every competent observer of the world,
and would not figure in the absolute conception. (As we shall see, people can know
143

Sophie-Grace Chappell

things even more locally perspectival than that.) [Williams is plainly thinking of claims
that involve thick ethical concepts.] The contrast with value should be expressed not
in terms of knowledge but of science. The aim is to outline the possibility of a
convergence characteristic of science, one that could meaningfully be said to be a
convergence on how things (anyway) are.
In this passage, despite its careful distinction between knowledge and science, Williams makes
some large and surprising concessions about the nature and the content of the absolute conception – concessions that we probably would not expect of someone who is in so many other
respects, as I have noted, a decidedly liberal naturalist. For one thing, and in a particularly austere
moment, he makes it very clear here that there is a certain sense in which he actually agrees
with the “eliminativists” with whom I started this essay. For Williams too, the mental and the
folk-psychological are not part of “how things (anyway) are”, because they are not part of the
absolute conception. Nor are colours; nor is grass; nor, apparently, are most of the stock of
“medium-sized dry goods” – trees, rocks, animals, artifacts – that we intuitively assume to be
the contents of the world, and to the understanding of which a large amount of working science
and human science (zoology, cognitive psychology, sociology, economics) is actually addressed.
Williams disagrees with writers like Quine and Churchland about what we can know: he thinks
we can know minds and their contents, for instance, though that knowledge is perspectival.
Still, from the standpoint not of knowledge but of science, for him too the basic reality of the
world is as resolutely austere as it is for any eliminativist.
What then is part of Williams’s absolute conception, if not minds, colours, ordinary quotidian objects, or anything like that? Here Williams seems caught between three different and
conflicting aspirations. First and most obviously, he wants the absolute conception to be
completely non-perspectival: that is the point of calling it absolute. But he also wants the absolute conception to play two other roles. Second, he wants it to be the answer to the question
“What is the account of fundamental reality that any successful inquirer at all, no matter how
different from us, will arrive at?” And third, closely related to a point of his about science that I
mentioned in footnote 1: he wants the absolute conception to explain both itself, and the
possibility of all the other nonabsolute conceptions of reality that somehow exist alongside it. As
he put this last point in his Descartes book, “the substance of the absolute conception…lies in the
idea that it could nonvacuously explain how it itself, and the various perspectival views of the
world, are possible”.39
I am not the first commentator to observe that it is hard to see how these three aspirations
can possibly be satisfied together. In Putnam’s words, “the notion of an absolute conception of
the world is either just the ordinary notion of an objective description or a we-know-notwhat”.40 Any account of the absolute conception that is rich enough to do all the explaining
required for a genuinely liberal naturalism cannot also be austere enough to be both completely
non-perspectival, and also part, as it were, of the final consensus of all intelligent aliens.41 Even
without the problems about the rational underdetermination of our conceptual frameworks that
I have been focusing on, it is easy to see how there might be not much more to the “final
consensus of all intelligent aliens” than fundamental mathematics (if even that). But if that is
really all there is to the absolute conception of the world, then the absolute conception is
absolute all right; but it is no longer all clear how it is a conception of the world. In this sense
(non-fundamental) science, which unlike the absolute conception can involve perspectival
knowledge, looks set to do more of the actual work of explaining things than the absolute
conception. But then, unfortunately, one of the many things that non-fundamental science
seems better placed to explain than the absolute conception, is how we get from the absolute
144

Bernard Williams’s liberal naturalism

conception to perspectival knowledge. If indeed that can be explained at all, as both Blackburn and
Putnam have doubted:
[T]here is no evident reason why [the possible convergence of any sufficiently advanced investigators of a world on post-Newtonian physics and mathematics] should
equip them to explain how our perspectival view of the world is possible, if only
because they may not be equipped to understand our view of the world or to know
what it is. Indeed, Williams’s well‐known and highly developed sense of history
suggests that his view ought to be that often they will not be equipped to understand
some of our social, political, and ethical concepts, precisely because these are the
contingent growths of our peculiar history, and need to be understood in historical
terms. Williams himself says as much (1978: 301–2). There is a tension here in his
thought, or even an outright inconsistency, and Hilary Putnam and others are right to
notice it.42
However, if the arguments of this essay (which to some extent follow Putnam) are right, things
are actually even more problematic for the absolute conception. For one thing: if Kripke is right
about plus and quus, it evidently follows that not even the absoluteness of mathematics will be
absolute enough to escape rational underdetermination. For another: given the rational underdetermination of our scientific concepts, it is far from clear that there need be any one thing
worth calling “the final consensus of all intelligent aliens”. All such aliens will be in the same
position as us about rational underdetermination. So their historically-contingent conceptual
schemes could be various in all sorts of rationally ungrounded ways; and what we and they are
prepared to count as a consensus can vary too in ways that defy any intelligible ambition for
rational uniqueness. For connected reasons, rational underdetermination equally threatens the
idea of uniquely correct explanations in science; if there is an ungrounded and ineliminable
looseness and contingency in the foundations of our conceptual schemes, the same looseness
will go over both into what we will feel as a question that needs an answer, and into what we
will see as an explanation that answers that question.
In the best exposition and defence that I know of a stance at least very close to Williams’s on
these matters, Adrian Moore rejects the above line of criticism as the pricking of a bubble that is
an implausible inflation of what Williams actually says.43 When (says Moore) we set aside the
hyperbole that his critics unfairly read into him, Williams’s crucial point is simply that when we
compare the notion of a scientific point of view with that of an ethical point of view, we find the
following key difference. Science can (a) say what is in the view from that viewpoint, and can
also (b) explain why that is what is in the view from there, in such a way that (a) and (b) do not
conflict. Ethics, by contrast, cannot do this. Perhaps in conjunction with the social sciences and
the history of ideas, ethics can certainly provide both (a) descriptions of perspectives and (b)
explanations of why those are the perspectives that we have. However – as is shown by
Williams’s own distinguished work on genealogies in societies, and dispositions in individuals –
what ethics says under heading (b) is necessarily disruptive of what it says under (a). To put it as
Williams sometimes puts it, the view from within an ethical disposition cannot be the same as
the view from outside it: in particular, it is not ethical in the same way – the disposition praises
and blames, for example, whereas the view of the disposition does neither. By contrast, the view
from outside a scientific perspective is not disruptive of that perspective. It explains the perspective and what is seen from it; and it too is scientific.
I think Moore’s reading makes good sense of many things that Williams says.44 In particular,
it is illuminating about what Williams thought he was doing by emphasising the difference
145

Sophie-Grace Chappell

between the views from within and from outside a disposition. However, I am not convinced
that all the difficulties are banished. In both ethics and science, we can distinguish talk about
viewpoints or perspectives from talk about conceptual framings. The two things go closely together, of
course, because any ethical or scientific perspective is at least partly constituted by the way we
frame and conceptualise its subject-matter. Still, they are distinguishable, and the distinction
between them, in the context of Williams’s argument, quickly begins to look like an ambiguity.
Thus it is certainly true that what one sees from within the disposition of, say, generosity is
different from what one sees from outside it, and that this difference is not the same at all as the
difference between, for example, looking at the stars through a telescope, and looking at the stars
on a celestial “map” or “globe” that has been constructed using, inter alia, data obtained by
looking through telescopes. The celestial map (or other formalisation of astronomy) explains what
we see through the telescope, and explains it completely – or something on the way to completely,
for example, “completely satisfactorily” (to us now given our astronomical knowledge).
That seems right, in general, about perspectives in science. Is the situation at all analogous
with perspectives in ethics? If Williams is right it should not be analogous at all. But in fact there
is both a negative and a positive analogy. On the negative side, it is not true that every scientific
explanation of how a partial perspective fits into a larger view is obviously vindicatory rather
than subversive. One familiar example, which Blackburn discusses a bit, is colour experience. It
is famously problematic for natural science to make good sense of what happens, subjectively, to
Mary the colour scientist.45 In the current state of that extremely vexed debate, it is not at all
obvious that the ultimate winner is going to be the side that makes our colour experience look
anything more than a kind of evolutionarily-advantageous mass hallucination. Another example
is mentioned by Moore in the course of his very accomplished defence of Williams’s notion of
an absolute conception: this is the case of our temporal experience and knowledge, which as
Moore himself implies, is very naturally seen as perspectival in a way strikingly reminiscent of
the perspectivality of ethical experience.46
On the positive side too, there is at least some analogy between perspectival views in ethics
and in science. Think here about the relation of the deliverances of a particular virtuous disposition, such as generosity, to virtue overall. One can imagine a good person wanting to
balance the impulses or motivations that come from generosity with those that come, for
example, from justice (“I want to give the beggar this bread, but it’s not mine”), or with those
that come from loyalty (“I want to give the beggar this money, but I mean it for my daughter’s
birthday present”). This is the sort of thing that good people have to think about all the time;
and it really does involve standing outside our own ethical dispositions, and the viewpoints that
those dispositions suggest, and trying to assess them from outside those dispositions. Such
conflicts might arise within a single agent; but – crucially – they might also arise between an
agent’s present dispositions, and the possibility of a disposition that the agent does not yet (fully)
possess, but sees is available in a society that the agent has just joined. (“I would turn these strangers
away back in England; but now I live in Iran, I am aware of the local custom/ virtue of
hospitality; and I see its attractions”.)
There are then serious doubts about Williams’s right to say that, in ethics, we can’t find
anywhere non-subversive to stand outside a disposition to criticise it or evaluate it or seek to
explain how it grips those whom it does grip. We certainly can, when that disposition is part of
a larger nexus of sensibility into which it needs to be adequately integrated; and such adequate
integrations, where successful, may very well be vindicatory. To this, perhaps, Williams will
object “The point is, you can’t make vindicatory sense of dispositions from outside all dispositions”; but to that the rejoinder is obvious. Neither can you make vindicatory sense of scientific
perspectives from outside all science.
146

Bernard Williams’s liberal naturalism

So Williams’s contrast of “inside” and “outside” for ethics and for science looks unconvincing
when we take it as a point about perspectives. Does it fare any better, taken as a point about
conceptual framing? No; it looks even more unconvincing. As I have already argued above,
science itself does not explain why science works with just that very stock of concepts that it does
work with, any more than ethics does. In both cases it looks very much like contingent history is
at least partly relevant: just because they have a quite different history from us, intelligent aliens
will quite certainly also have a quite different conceptual repertoire from earthlings both in ethics
and in science. Of course we can demonstrate, from within science’s conceptual framework, what a
marvellous conceptual framework it is, and just how remarkably explanatory and illuminating it
is. (And rightly so: it really is.) But we can do that in ethics too. And we cannot stand outside our
science’s conceptual framework and demonstrate its merits, from this outside standpoint, any
more than we can in the case of ethics. Just this is the conclusion of the original objection that I
developed above. And as far as I can see, it has not been answered.
What remains standing at the end of this line of inquiry is, I think, not very much of
Williams’s notion of the absolute conception. But at least two positive (or more or less positive)
remarks can still be made, with which I close.
First: this is not, as Putnam’s arguments are, a rejection of “metaphysical realism”47 nor, as
Rorty’s are, of a science-centered theory of knowledge. Nothing I have said has challenged, or
been intended to challenge, the eminently reasonable broadly naturalist idea that science has a
genuinely special epistemic status, and that ethics owes science a genuine explanatory debt. My
target is Williams’s articulation of that idea, not the idea itself. There always did seem to be a
certain irony in the spectacle of Williams, of all people, building up an absolute conception of
anything; even science. If Williams is, as I believe, right in many of the other illuminating things
he has to say about the questionable effects on our self-understanding of such absolutising
constructs as systematic moral theories,48 perhaps it is not entirely surprising if science too ends
up best seen as a human practice; a very special one, of course, but still a human practice.
Perhaps what we discover here is that, while (as above) ethics certainly owes an explanatory
debt to science, it is also true, conversely, that science owes an explanatory debt – of a very
different kind – to ethics.
Second, and last, consider this from Kripke49:
The entire “game” that we have described—that the community attributes a concept
to an individual so long as he exhibits sufficient conformity, under test circumstances,
to the behaviour of the community—would lose its point outside a community that
generally agrees in its practices…On Wittgenstein’s conception, such agreement is
essential for our game of ascribing rules and concepts to each other (see [Philosophical
Investigations I,] 240). The set of responses in which we agree, and the way they
interweave with our activities, is our form of life. Beings who agreed in consistently
giving bizarre quus-like responses would share in another form of life. By definition,
such another form of life would be bizarre and incomprehensible to us…[But they]
could play the game of attributing rules and concepts to each other as we do…There is
no “objective fact”—that we all mean addition by “+”, or even that a given individual
does—that explains our agreement in particular cases. Rather our licence to say of
each other that we mean addition by “+” is part of a “language game” that sustains
itself only because of the brute fact that we generally agree.
Underlying our use and deployment of the concepts that we have, says Kripke – whether
ethical, mathematical, or scientific – is not any kind of ur-rational ur-grounding such as we
147

Sophie-Grace Chappell

might believe in if we were in the grip of a “rationalistic theory of rationality” (Williams 1985),
but simply the “brute fact” of communal consensus, what Philosophical Investigations calls
“agreement in form of life”: “This is simply what we do”.50 Or as Stanley Cavell equally
famously, and more colourfully, puts it51:
We learn and teach words in certain contexts, and then we are expected, and expect
others, to be able to project them into further contexts. Nothing ensures that this
projection will take place (in particular, not the grasping of universals nor the grasping
of book of rules), just as nothing insures that we will make, and understand, the same
projections. That on the whole we do is a matter of our sharing routes of interest and
feeling, senses of humour and of significance and of fulfilment, of what is outrageous,
of what is similar to what else, what a rebuke, what forgiveness, of when an utterance
is an assertion, when an appeal, when an explanation – all the whirl of organism
Wittgenstein calls “forms of life”. Human speech and activity, sanity and community,
rest upon nothing more, but nothing less, than this. It is a vision as simple as it is
difficult, and as difficult as it is (and because it is) terrifying.
We might find that we wanted a way of describing the condition of someone who understands
this “simple” but “terrifying” point that, both in ethics and beyond, her reasons and her
concepts rest in this way, not on some further reason for having those reasons or concepts, but
ultimately on something like consensus: on having a life together. We might even decide that
the word we want to describe this person’s condition, at least when things go well with her and
she is not too shaken by the whirl of organism, is Williams’s own word: confidence.52

Notes
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21

Crane and Mellor (1990).
Baker (1987) and Boghossian (1990).
Fodor (1987, Chap. 1).
Chappell (2020).
Williams (1985, 121).
On the nature of liberal naturalism, see De Caro and Macarthur (2004, 2010).
Dancy (2004).
Williams (1985: 135)
Ibid., 136.
Ibid., 139; cp. 111; Williams (1978, 246).
Altham and Harrison (1995, 204).
Williams also says that the absolute conception explains the world and everything in it, including the
absolute conception itself. Perhaps he thinks the two claims are equivalent. On why they are not, and
on why the difference matters, see below.
Williams (1985, 139–140).
Ibid., 155.
Ibid., 143.
Ibid., 199–200.
Ibid., 155.
Ibid., 167.
Ibid., x.
Ibid., 117.
A little confusingly Aristotelian eudaimonism, which Williams rejects, is one of the normative-ethical
positions often called ethical naturalism. Ethical naturalism as a normative-ethical position is the view
that what is good and bad ethically speaking is (at least fairly) smoothly continuous with what is “good
for us” zoologically speaking. This sort of naturalism is easily allied with naturalism in the metaethical

148

Bernard Williams’s liberal naturalism

22
23
24
25
26
27
28
29
30
31
32

34
35
36
37
38
39
40
41

42
43
44

sense in which I use the word in this essay, to denote the view that it needs to be explained how ethics
fits in to the larger scientific world-picture. The two senses of “naturalism” are nonetheless different
from each other.
Diamond (2019) and Anscombe (1959).
Williams (1985, 153).
See in particular Parfit’s ringing pronouncement, in the last paragraph of Reasons and Persons (1987),
that properly philosophical ethics has barely begun, because it has barely begun to be secular.
Williams (2006, 190–191)
Williams (1985, 136).
Rorty (1979).
Sturgeon (1985).
Jardine, Hookway, and McDowell in Altham and Harrison (1995); Thomas (2007); and
Wiggins (1987).
I take the convention of capitalising the names of thick concepts from Kirchin (2013).
https://famous-trials.com/wilde/346-literarypart
The word often used in this connection is “commensurability”, but that seems a distracting misnomer.
“Commensurable with” is a polyadic relation between two or more items x, y,… and some metric M,
such that x and y are commensurable under M iff x > y, or y > x, or y = x. This has little to do with the
issue at hand, which seems much better understood as a matter of mutual intelligibility or, in closer
alignment with what I say in the main text, (inter-)translatability.
Strictly, Williams (1985, 234) takes justice to be “intermediate” between the thick and thin concepts.
(RIGHT and OUGHT he takes to be paradigm thin concepts.) My discussion ignores this point of
Williams’s because, for at least two reasons, I think it is wrong: while certainly some concepts are
thicker than others, there are no (naturally-occurring) purely thin concepts. First, a concept that had no
function except to direct would not be a concept at all. Second, the reasons for taking the paradigm thick
concepts to be thick are (a) their clearly culture-specific histories, (b) their clear possession or implication of particular descriptive contents, and (c) the usual hankering of analytic philosophy for a
place that is safe from history, which in this context the thin concepts are supposed to supply. But (c) is
misguided, and (a) and (b) are just as true of JUST and RIGHT and OUGHT as they are of CRUEL
and LIE. See further my essay in Kirchin (2013).
Or indeed by projecting his concept of ENSLAVEMENT differently, but I pass over that here.
Goodman (1955).
Quine (1960).
Putnam (1985).
Williams (1985, 139).
Williams (1978, 246).
Putnam (2009, 102).
We are close here to two large and intriguing questions: “How different could Intelligent Alien science
be from Earthling science, yet still be science?”, and “What must any version of science involve, to
count as science at all?” Imaginative exploration of these questions is a large part of what can make
science fiction so philosophically exciting. For three relevant philosophical explorations, see Hacking
(2000) [science in general] and Radick (2005) [genetic biology specifically]; also Weston (1988) [on
the question how we could tell if we did make contact with scientifically sophisticated aliens]. For
illuminating and helpful discussion of these issues online, many thanks to a host of friends (especially
Ian Kidd and Bill Wringe, for these references).
Blackburn (2012, 245–246).
Moore (1997, 2019).
Aside from their obvious merits anyway, Adrian Moore’s (1997) and (2019) defences of the absolute
conception have the following special claim on our attention. Altham and Harrison (1995) includes
sparkling essays on the absolute conception by Nicholas Jardine and Christopher Hookway (I have
found both essays enormously helpful for this essay). In his “Replies”, Williams (Altham and Harrison
(1995, 202–210)) concedes that some of his earlier claims about the absolute conception were overbold; but he unfortunately doesn’t find space to give a detailed response to either Jardine’s or
Hookway’s objections. He apologises for this brevity in a footnote (223, n.22), promising to say more
in Truth and Truthfulness. Alas, Williams did not live to redeem this promise: Truth and
Truthfulness–finished in a hurry under the stress of terminal illness – has just one six-line footnote about
the absolute conception (2003, 295, n19), and it says nothing about Jardine’s or Hookway’s critiques.

149

Sophie-Grace Chappell

45
46
47
48

49
50
51
52

However, what Williams does do in that footnote is endorse and recommend Moore (1997) as a
defence of the absolute conception. For this reason I rather regret not having the space here to engage
more directly with Moore’s defence. However, I do know the upshot of my engagement would be
that, while I am impressed by his argument, and while I think it makes a forceful case for one particular
framing of the issue, I don’t think that that framing addresses or answers all the problems that I raise
here. I hope I will be able to take these issues further some other time.
Jackson (1982).
Moore (2019, 178).
Notwithstanding the Wittgensteinian and Cavellian remarks that are coming. Those remarks are
certainly not expressions of antinaturalism; in the truest sense, they are naturalism. Nor, in the true
sense, are they antirealist. I take my view to be a shaken realism, perhaps. But not a stirred one.
Is there a connection between Williams’s rejection of moral theory and his advocacy of the absolute
conception? Is his anti-theory approach to ethics in fact driven by the thought that ethics is not science –
which does have a theory? Except for the phrase from 1985: 136 that I quote above, “a systematised
theoretical account of how the world really is”, I see little clear explicit evidence for that. But it does
often look, implicitly at least, like that is what is going on.
Kripke (1982, 98–99).
Wittgenstein (1951, Sect. 241, 217).
Cavell (1969, 52).
Williams (1985, 170–171). Thanks for comments and encouragement to my many helpful Facebook
philosophy friends; also to Alice Crary and Adrian Moore, neither of whom is responsible for any of
my mistakes.

References
Altham, J.E.J. & Harrison, R. World, Mind, and Ethics: Essays on the Ethical Philosophy of Bernard Williams.
Cambridge University Press. (1995).
Anscombe, G.E.M. An Introduction to Wittgenstein’s Tractatus. Basil Blackwell. (1959).
Baker, L.R. Saving Belief. Princeton University Press. (1987).
Blackburn, S. The Absolute Conception: Putnam vs Williams. In Practical Tortoise Raising and Other Essays.
Oxford University Press. (2012).
Boghossian, P. The Status of Content. Philosophical Review, 99, (1990). 157–184.
Cavell, S. Must We Mean What We Say? Cambridge University Press. (1969).
Chappell, S.G. Knowing What To Do. Oxford University Press. (2015).
Chappell, S.G. The Glass Onion. In A. Coles and J. Collicutt (eds.), The Neurology of Religion. Cambridge
University Press. (2020).
Churchland, P. Eliminative Materialism and the Propositional Attitudes. Journal of Philosophy, 78(2),
(1981). 67–90.
Crane, T. and Mellor, D.H. There is No Question of Physicalism. Mind 99(394), (1990). 185–206.
Dancy, J. Ethics without Principles. Clarendon Press. (2004).
De Caro, M. and Macarthur, D. Introduction: The Nature of Naturalism. Naturalism in Question. Harvard
University Press. (2004).
De Caro, M. and Macarthur, D. Introduction: Science, Nature and the Problem of Normativity.
Naturalism and Normativity. Columbia University Press. (2010).
Diamond, C. Reading Wittgenstein with Anscombe, Going on to Ethics. Harvard University Press. (2019).
Fodor, J. Psychosemantics. MIT Press. (1987).
Goodman, N. Fact, Fiction, and Forecast. Harvard University Press. (1955).
Hacking, I. How Inevitable Are the Results of Successful Science? Philosophy of Science, 67(Supp.),
(2000). 58–71.
Kirchin, S. (ed.) Thick Concepts. Oxford University. (2013).
Kripke, S. Wittgenstein on Rules and Private Language. Harvard University Press. (1982).
Mackie, J.L. Ethics: Inventing Right and Wrong. Penguin. (1978).
McDowell, J. Non-Cognitivism and Rule-Following. In S. Holtzman and C.M. Leich (eds.), Wittgenstein:
To Follow A Rule. Routledge. (1981). 141–162.
Moore, A. Points of View. Oxford University Press. (1997).

150

Bernard Williams’s liberal naturalism
Moore, A. Realism and the Absolute Conception. First published in Thomas (2007); reprinted with
revisions In Language, World, and Limits: Essays in the Philosophy of Language and Metaphysics. Oxford
University Press (2019).
Parfit, D. Reasons and Persons. Oxford University Press. (1987).
Putnam, H. Realism and Reason: Vol. 3 of Philosophical Papers. Cambridge University Press. (1985).
Putnam, H. Realism With a Human Face. Harvard University Press. (1992).
Putnam, H. Renewing Philosophy. Harvard University Press. (2009).
Quine, W.V.O. Word and Object. MIT Press. (1960).
Radick, G. Other Histories, Other Biologies. In A. O’Hear (ed.), Philosophy, Biology and Life. Cambridge
University Press. (2005). 21–47.
Rorty, R. Philosophy and the Mirror of Nature. Princeton University Press. (1979).
Streumer, B. Unbelievable Errors. Oxford University Press. (2017).
Sturgeon, N. Moral Explanations. In D. Copp and D. Zimmerman (eds.), Morality, Reason, and Truth.
Rowman and Allanheld. (1985). 49–78.
Thomas, Alan (ed.) Bernard Williams: Contemporary Philosophy in Focus. Cambridge University Press. (2007).
Weston, M. Radio Astronomy as Epistemology: Some Reflections on the Contemporary Search for
Extraterrestrial Intelligence. The Monist 71(1), Philosophical Problems of Space Exploration, (1988).
88–100.
Wiggins, D. Needs, Values, Truth. Oxford University Press. (1987).
Williams, B. Descartes: the Project of Pure Enquiry. Penguin. (1978).
Williams, B. Internal and External Reasons. In Moral Luck, pp. 101–113. Cambridge University
Press. (1981).
Williams, B. Ethics and the Limits of Philosophy. Fontana. (1985).
Williams, B. Making Sense of Humanity. Cambridge University Press. (1995).
Williams, B. Truth and Truthfulness. Princeton University Press. (2002).
Williams, B. and Moore, Adrian (eds.), Philosophy as a Humanistic Discipline. Cambridge University
Press. (2006).
Wittgenstein, L. Philosophical Investigations. Trans. G.E.M. Anscombe. Macmillan Publishing
Company. (1953).

151

13
PRICE’S SUBJECT NATURALISM
AND LIBERAL NATURALISM
Lionel Shapiro

1 Introduction
According to one generically pragmatist line of thought, metaphysical perplexities can be
overcome by turning attention to our uses of words and concepts. Over the last three decades,
Huw Price has motivated a version of this program by distinguishing two ways of adhering to
naturalism in philosophy. Price argues that by pursuing naturalistic inquiry into ourselves as
subjects, as speakers and thinkers, we can undermine naturalistic doctrines about the objects of our
talk and thought, doctrines that give rise to metaphysical puzzles. In this sense, he opposes
“object naturalism” from the standpoint of “subject naturalism”.1
In this chapter, I’ll first seek to identify how Price conceives of object naturalism and subject
naturalism, as well as his reasons for holding that object naturalism can be undermined by
subject-naturalistic inquiry. Then I’ll address five questions about how his project bears on the
prospects for a liberal naturalism.
1
2
3
4
5

Does Price’s strategy depend on his requirement that the relevant inquiry into human
discourse and thought be conducted in natural-scientific terms?
Is Price’s strategy even compatible with that requirement?
Does the worldview Price arrives at amount to a liberal naturalism, i.e. a naturalism that
could be obtained by relaxing a more restrictive version?
Is Price’s strategy consistent with a liberal naturalism?
Should a proponent of Price’s strategy accept a liberal naturalism?

I’ll present considerations in favor of negative answers to (1), (2) and (3), and affirmative answers to (4) and (5). Crucially, the liberal naturalism I’ll recommend to the subject naturalist
will be a view about inquiry into discourse and thought, not a metaphysical view about the
extent of the natural realm.
To clarify the resulting view of the terrain, I’ll then contrast Price’s distinction between two
kinds of naturalism with that drawn by John McDowell, who distinguishes between a “restrictive naturalism” and his own “liberal naturalism” (2009c, 261–262). Both Price and
McDowell seek to avoid philosophical anxieties by focusing on human subjects as natural
beings. But there’s an important difference. McDowell maintains an ontological naturalism about
152

DOI: 10.4324/9781351209472-16

Price’s subject naturalism

(e.g.) moral values by advocating a position that’s liberal in its naturalism. By contrast, Price
argues that restrictively naturalistic inquiry into human subjects can vindicate a position that’s
liberal (though not naturalistic) in its ontology.2 The version of subject naturalism I’ll recommend
represents a middle ground: it holds that liberally naturalistic inquiry into human subjects can
vindicate a position that’s liberal (though not naturalistic) in its ontology.

2 What is subject naturalism?
According to Price, “philosophical naturalism” or “naturalism per se” is just the view that “in
some areas, philosophy properly defers to science” (2011d, 184, 199). This may be an overly
broad formulation, since it allows believers in the supernatural to count as naturalists.
Nonetheless, it helpfully calls attention to two respects in which varieties of naturalism can
differ from each other. They can differ in respect of the domains to which they apply, as well as
in respect of the kind of deference to science they require.
Price points out that naturalism is usually regarded as applying to all reality, and as requiring
deference to science with regard to both ontological commitment and epistemology. On this
view, which he calls “object naturalism” (2011d, 185),
(ON) “all there is is the world studied by science, [and] all genuine knowledge is scientific
knowledge”.
Object naturalists holds that “the only facts there are are the kind of facts recognized by
natural science” (2011c, 4–5; 2013, 24, 168). Such facts concern objects, properties and happenings that are in principle describable in the vocabulary of the natural sciences, and they are
knowable, if at all, by methods of natural-scientific inquiry.3
Price rejects (ON), concluding that philosophy’s “debt…to science” (2011d, 198) is sufficiently acknowledged by a view he calls subject naturalism. Subject naturalism “is the philosophical viewpoint that begins with the realization that we humans (our thought and talk
included) are surely part of the natural world” (2011c, 5). This might appear to be a restriction
of object naturalism to a particular domain: thought and talk done by human subjects. Thus
construed, subject naturalism would countenance facts outside the reach of natural science, as
long as these don’t pertain to human subjects. But this misidentifies the sense in which subject
naturalism is the more liberal view. After all, Price acknowledges normative facts about human
beings.4 Yet he doesn’t hold that such facts can be established by natural-scientific means.
In fact, subject naturalism can’t be distinguished from object naturalism in terms of its domain of
application. Rather, it’s a view about the nature of the proper deference to science when philosophising
about any domain. Introducing the label, Price says that according to subject naturalism,
philosophy needs to begin with what science tells us about ourselves. Science tells us that
we humans are natural creatures, and if the claims and ambitions of philosophy
conflict with this view, then philosophy needs to give way. (2011d, 186)
This formulation subsumes two theses about the “relevance of science to philosophy” (2011c,
184, 186). The first imposes a constraint. Philosophical claims must be consistent with “the
basic (‘subject naturalist’) premise that the creatures employing the language in question are
simply natural creatures, in a natural environment” (2011c, 9; 2011d, 187, 194, 198). This rules
out philosophical claims crediting us with supernatural faculties, faculties that can’t be understood as acquired in the development of an organism that arose by biological evolution.
(Alleged examples of supernatural faculties might be immediate awareness of universals, or
rational intuition into truths.) However, this constraint doesn’t yet underwrite the second thesis
153

Lionel Shapiro

in Price’s formulation, namely that what science reveals about ourselves can and should serve as
a place for philosophy to begin.
Central to Price’s subject naturalism is the view that some philosophical problems can only
be resolved via inquiry into human language and thought. These problems concern, in the
well-known words of Wilfrid Sellars (1963, 1), how “things in the broadest possible sense of
the term hang together in the broadest possible sense of the term”. How can things as heterogeneous as physical objects, mathematical structures, colors, moral and aesthetic values or
norms, possibility, probability, causation and meaning be incorporated into a coherent
worldview? Price addresses such “location problems” or “placement problems” as they are
posed from the point of view of an object naturalist (2011c, 4–8; 2011d, 186–188; 2013,
26–29).5 Faced with a placement problem, we seem to have a choice of locating the relevant
item within the “sparse world apparently described by science”, or embracing some version of
eliminativism, fictionalism, or projectivism according to which the items are illusory. Either
way, we’d be doing what Price calls “metaphysics”: addressing philosophical perplexities
about how things hang together by investigating the existence and nature of the things in
question.6
Subject naturalism is advertised as a way of sidestepping this choice, and thus avoiding
metaphysics (in this sense) altogether.
(SN1) Placement problems can be sidestepped via inquiry into how humans, as natural
creatures, use language and thought.
Price describes such inquiry as “subject naturalistic” (e.g. 2011d, 196, 193–194). And he imposes a constraint on it:
(SN2) The inquiry mentioned in (SN1) should be conducted using methods drawn
from the natural sciences, specifically from the “scientific perspective of a linguistic
anthropologist” (2011c, 11).
I’ll be understanding Price’s subject naturalism as encapsulated by (SN1) and (SN2).7

3 The subject-naturalist rejection of object naturalism
According to Price, subject-naturalistic inquiry helps sidestep the problems resulting from
object naturalism by undermining (ON) itself. That’s because (ON) carries presuppositions
about language users that could be refuted by naturalistic inquiry.
If the presuppositions of object naturalism turn out to be suspect, from this selfreflective scientific standpoint, then subject naturalism gives us reason to reject object
naturalism. (2011d, 186).
This yields a sense in which “[s]ubject naturalism…comes first” in relation to object naturalism.
Subject-naturalistic inquiry should enjoy a priority over object-naturalistic attempts to resolve
the placement problems, because those attempts should be placed on hold pending subjectnaturalistic “validation” of object naturalism’s presuppositions.
In fact, Price expects that subject-naturalistic inquiry will call into question the credentials of
object naturalism:

154

Price’s subject naturalism

Invalidity Thesis: There are strong reasons for doubting whether object naturalism
deserves to be “validated”—whether its presuppositions do survive subject naturalist
scrutiny. (2011d, 187)
However, it isn’t always clear how Price aims to substantiate the Invalidity Thesis. I understand
him as offering two very different lines of arguments. In the rest of the section, I’ll distinguish
these, and contrast them with one he doesn’t offer.

3.1 Blocking a route to object naturalism: representationalism
One of Price’s lines of argument seeks to show that subject-naturalistic inquiry undermines one
possible strategy for validating object naturalism. That strategy would rest on an assumption Price
calls “representationalism”. This is the view that our employment of linguistic and mental items
in contentful talk and thought must be explained in terms of their bearing a relation of
“standing for” or “representation” to objects, properties, relations and/or states of affairs.
Suppose we embrace representationalism as part of a naturalistic explanation of language and
thought. In that case, the representation relation will itself have to be describable in naturalscientific terms. Hence, also the relata on the worldly end of this relation, for example, normative properties and states of affairs, must be so describable.
Given a naturalistic conception of speakers, the addition of a representationalist
conception of speech makes [object naturalism] almost irresistible. Term by term,
sentence by sentence, topic by topic, the representationalist’s semantic ladder leads us
from language to the world, from words to their worldly objects. Somehow, the
resulting multiplicity of kinds of entities—values, meanings, and the rest—needs to be
accommodated within the natural realm. To what else, after all, could natural speakers
be related by natural semantic relations? (2011d, 198–199; 2011c, 4)
In short, provided it supports representationalism, subject-naturalistic inquiry will validate
object naturalism.
On the other hand, if subject-naturalistic inquiry calls representationalism into question, this
semantic strategy for validating object naturalism will be unavailable. Price argues that this is
indeed the case: subject-naturalistic inquiry into the function of semantic vocabulary supports a
deflationary view that debars “refers”, “true of ” and “true” from playing the explanatory role
representationalists take them to play. Consequently, “semantic deflationism…blocks a particular route to location problems – a route that otherwise carries a lot of traffic” (2011f,
258–259, 262–264). Once we embark on that route, the route that starts with a naturalistic
version of representationalism, the only way to maintain realism about the objects of our
thought and talk is to “place” them in the “natural realm”. Hence, a subject-naturalistic vindication of deflationism about semantic vocabulary can block one route to object naturalism.
Notice that this is a route on which placement problems are not just an upshot of object naturalism, but also source of object naturalism.
At times, Price appears to regard the strategy of validating object naturalism via representationalism as more than merely one possible source of object naturalism. “For someone
who takes science seriously, the only route to object naturalism is…to concede that the problem
begins at the linguistic level, and to defend the representationalist view”, at least for discourse
they don’t dismiss as non-factual (2011d, 196 [my emphasis]). This looks like an unnecessary
commitment. For one thing, it excludes as unmotivated any object naturalism that rejects
155

Lionel Shapiro

representationalism (Knowles 2018, 297–300), a position whose availability Price has since
acknowledged (2011c, 5; 2013, 25n4). In addition, even a representationalist may take it that
their representationalism only yields placement problems given a pre-existing commitment to
object naturalism.8

3.2 Undermining a presupposition of object naturalism: functional uniformity
According to a second line of argument, subject-naturalistic inquiry does more than just undermine one strategy for validating object naturalism. It actually serves to invalidate object
naturalism.
Here the role of subject-naturalistic inquiry is to reject what Price sees as the object naturalist’s presupposition that the vocabulary we use to talk about (say) values is “in the same line
of work” as the language used in natural-scientific description and explanation. Semantic deflationism is important here too: it rules out the answer that the alleged common line of work is
that of representing the world or making true claims (e.g. Price 2011f, 264).
According to Price, subject-naturalistic inquiry reveals that (e.g.) normative and semantic
vocabularies serve very different functions in our linguistic economy from those served by the
vocabularies employed in the natural sciences. He argues that the resulting “functional pluralism” (2011b, 136) about our thought and talk should have the effect of “blocking reductionist
moves” (O’Leary-Hawthorne and Price 2011, 124, Price 2011b, 146).
[T]he functional standpoint threatens to undercut the motivation for reductionism:
once we have an adequate explanation for the fact that the folk talk of Xs and Ys and
Zs, an explanation which distinguishes these activities from what the folk are doing
when they do physics, why should we try to reduce the Xs and Ys and Zs to what is
talked about in physics? (2011e, 78)
In other words, functional pluralism undermines both reductionist proposals and the eliminativist alternative, according to which the conclusion that values or meanings are recalcitrant
to object-naturalistic reduction implies that there are no values or meanings.
With regard to its opposition to reductionism and eliminativism, Price emphasises, his
functional pluralism resembles nonfactualist/noncognitivist versions of expressivism. But while
traditional expressivism preserves a contrast between (e.g.) ethical discourse and genuinely descriptive discourse, Price argues that the anti-reductionist upshot of his functional pluralism doesn’t
depend on the privileging of any kind of discourse (e.g. that of natural science) as genuinely
descriptive.9 Instead, “the philosophically interesting work of non-cognitivism – the work of
blocking reductionist moves, in particular – is done by the functional characterization” that reveals the heterogeneity of vocabularies (O’Leary-Hawthorne and Price 2011, 126; cf. 123).

3.3 Vindicating the correctness of assertions?
Robert Brandom offers a third interpretation of how Price’s subject naturalist counters the
object-naturalist’s doctrine that (e.g.) normative talk remains suspect pending placement of its
objects in the world described by the natural sciences.
[W]e describe how the use of the vocabulary is taught and learned. If there is nothing
mysterious about that, and we can say in our favoured [natural-scientific] terms just
what one needs to do in order to use the vocabulary correctly, Price argues, then the
156

Price’s subject naturalism

vocabulary should count as naturalistically acceptable.… (Brandom 2013, 86 [my
emphasis]; cf. 2008, 25; 2015, 93)
For now, I’ll ignore Brandom’s questionable assumption that Price aims to reveal (e.g.) moral
vocabulary as “naturalistically” acceptable, rather than just as acceptable without reductionist or
eliminativist analysis.
Thus construed, the subject-naturalist’s approach to placement problems would require
vindicating the correctness of uses of the target vocabulary. As Brandom notes (2013, 88), this
presupposes that it’s legitimate for subject-naturalistic inquiry to employ a notion of correctness.
Here, it’s crucial that “correct” not mean true. On the deflationary approach to truth Price
shares with Brandom, a subject-naturalistic explanation of the functioning of the target vocabulary won’t specify truth conditions. The truth condition of “Eating meat is wrong” can be
specified, in English, by using that normative sentence itself. Instead of a truth condition, the
kind of correctness Price makes use of in subject-naturalistic explanations is what he calls a
“subjective assertibility condition”. Here is one such possible condition: “The utterance ‘X is
good’ is prima facie appropriate when used by a speaker who approves of (or desires) X” (Price
2011g, 82–83; 2011e, 73).
But it’s unclear that placement problems for (say) moral properties or states of affairs can be
undercut via an account of subjective assertibility conditions for sentences of moral discourse.
Suppose a subject naturalist offers such an account, as well as an explanation of how moral
discourse qualifies as making truth-evaluable claims. That would still leave them free to advocate, on object-naturalist grounds, an error theory on which ascriptions of moral properties
achieve only subjective assertibility and never truth. Pending an explanation of how to leverage
a naturalistically acceptable explanation of a vocabulary’s “correct use” into a way to escape
placement problems, Brandom’s interpretation fails to capture Price’s subject-naturalist
program.

4 Does Price’s rejection of object naturalism depend on subject naturalism?
We’ve seen two ways in which subject-naturalistic inquiry into human language and thought
might undercut object naturalism: by blocking the representationalist route to object naturalism
and by supporting functional pluralism. I now turn to questions (1) and (2), which can be
reformulated as follows. Is there a reason why either of these strategies must be carried out via
natural-scientific explanations of the functioning of linguistic or mental items? And is there
reason to expect that either could be carried out that way?
Regarding question (1), the anti-representationalist strategy doesn’t appear to require
natural-scientific inquiry into our uses of semantic expressions such as “true” and “refers”. I see
no reason why the deflationary explanations of semantic talk must be presented from the
“scientific perspective of a linguistic anthropologist” (2011c, 11). The same goes for the
functional pluralist strategy. As Knowles (2011, 79) observes, “if ultimately there is no necessity
about using scientific conceptions in the logic of expressivist explanation”, then “it becomes
very unclear what the naturalism really amounts to here” (also Shapiro 2014, 502).10 In short,
Price’s strategies for undermining object naturalism don’t appear to require (SN2).
Regarding question (2), there are reasons to doubt that Price’s critique of object naturalism
can rest on explanations of language and thought that employ the conceptual and methodological tools Price regards as proper to the natural sciences.11
First, both the deflationist strategy and the functional pluralist strategy involve semantic
ascent: rather than theorise about the nature of moral goodness or the nature of reference, we
157

Lionel Shapiro

examine the functioning of the words “good” and “refers”. But why should we care, in this
connection, about the functioning of those particular English words? The obvious answer is
that we’re trying to explain the difference between we we’re doing when we speak of what’s
good, and of what refers to what, and what we’re doing when we engage in scientific description
and explanation. We aim to
explain in naturalistic terms…what role the different language games play in our
lives—what differences there are between the functions of talk of value and the
functions of talk of electrons, for example.
(Price 2011d, 199)
Price’s interest in how the English words “good” and “refers” function thus rests on the fact
that these are words English-speakers use to talk of being good and of referring. They’re words
we use in order to say that something is good, or say that it refers to something else. And this
means that the target of our subject-naturalistic inquiry must be linguistic expressions considered insofar as they enjoy a semantic characterisation (by which I don’t mean a characterisation in terms of reference or truth, but a characterisation in terms of what they’re used
to say).12 If that’s right, the functional explanations Price uses to undercut object naturalism
must employ semantic locutions tied to propositional-attitude ascriptions (Shapiro 2014,
504–505).13 But by Price’s own lights (2011b, 133n3; 2011a, 205n2, 209, 219–220), we have
no reason to expect that the facts stated using such locutions can be reconstructed from within
the austere perspective of the natural sciences (cf. Macarthur 2014, 76–77).
There’s a further reason for wondering whether natural-scientific inquiry is suited to revealing the distinctive role in our discursive economy of at least some of the vocabularies that
give rise to placement problems. Inspired by Kant and Sellars, Brandom argues that in the case
of both modal and normative vocabularies, the role that distinguishes them from “empirical
descriptive vocabulary” is their “framework-explicitating function”:
A central observation of Kant’s is that what we might call the framework of empirical
description—the commitments, practices, abilities and procedures that form the necessary practical background within the horizon of which alone it is possible to engage
in the cognitive theoretical activity of describing how things empirically are—
essentially involves elements expressible in words that…do not perform the function of
describing (in the narrow sense) how things are.
(Brandom 2013, 105; cf. 2015, 35)
If this is right, the pragmatist approach to the target vocabularies should proceed via the
identification of necessary features of the framework of empirical description. Identifying such features wouldn’t seem to call for natural-scientific investigation from the viewpoint of Price’s
empirical “linguistic anthropologist”. Rather, it would be a task for what Sellars calls
“transcendental linguistics”, which seeks an “analytic account of the resources a language
must have to be the bearer of empirical meaning” (2002b, 281; 2002a, 268). The project
could remain one of undermining object naturalism about normative properties and facts (and
any modal ones excluded from natural-scientific discourse) by examining the distinctive
framework-explicitating functions of normative (and modal) expressions. But whether or not
that examination should be viewed as analytic, it would no longer be natural-scientific
inquiry.

158

Price’s subject naturalism

5 Subject naturalism and liberal naturalism
Can Price’s subject naturalism be viewed as a species of “liberal naturalism”? On one understanding of this term, it applies to positions that agree with (ON) in insisting that the only facts
are natural facts, but depart from (ON) in embracing “a broader, more expansive conception of
nature that makes room for a class of nonscientific, but nonetheless nonsupernatural, entities”
(De Caro and Macarthur 2010, 4).
When liberal naturalism is understood this way, the answer to our question is “no”.14 But
it’s important to consider why. It might be thought that in rejecting (ON), Price advocates an
object non-naturalism about the objects, properties and facts responsible for placement problems.
Against this, he clarifies that while his subject naturalist “may agree that moral properties are not
natural properties”, this would be a misleading way of putting the upshot of subject-naturalistic
inquiry, which doesn’t serve to establish any metaphysical thesis. The true upshot is “put more
clearly by shifting explicitly to the meta-linguistic frame, and saying that moral terms and
concepts are in a different ‘line of work’ to the terms and concepts of natural science” (2019,
139). Price’s naturalism, as encapsulated by (SN1) and (SN2), isn’t a metaphysical doctrine.
When it comes to the objects of metaphysical perplexity, then, Price defends neither naturalism,
whether restrictive or liberal, nor non-naturalism (2013, 169).
On the other hand, when it comes to the methodology of his functional explanations of our
talk and thought, Price insists on a restrictive naturalism on which the only concepts invoked
are those from the natural sciences. His claim is that privileging natural science in this explanatory project doesn’t entail privileging it in matters of ontology (2011b, 142; 2011c, 30–31;
2013, 59–60).15 In short, then, no aspect of Price’s position can be described as advocating a
looser conception of the natural.
Still, no aspect of his position is incompatible with such a conception. All he rules out is a
looser conception of how the functioning of the target vocabularies is to be explained. Earlier, I argued
that Price’s philosophical aims are compatible with, and may even require, such a looser
conception. The functional explanations may need to use intentional descriptions, and they
may be justified by transcendental arguments. To this, I can now add that such a conception
should still count as naturalistic. After all, instances of speech and thinking wouldn’t be characterised in a way that requires any supernatural capacities. In summary, I’ve answered questions
(3), (4) and (5) as follows.
•
•
•

Price’s subject naturalism doesn’t amount to a liberal naturalism.
Yet his commitment to a restrictive naturalism concerning the kind of inquiry that’s
supposed to undercut placement problems doesn’t itself rule out a liberal naturalism.
Moreover, there is good reason for a subject naturalist to reject that commitment. This
would entail embracing a liberal naturalism about inquiry into the thought and talk of
human subjects.

Pursuing the subject naturalist project using a more liberal conception of what counts as naturalistically respectable wouldn’t, however, amount to a liberal naturalism that regards as natural
the objects that give rise to placement problems. Subject naturalism gives us no reason to recognise, let alone answer, a substantial question concerning the extent of the “natural realm”.
While it may remain convenient to speak of “natural facts” as those recognised by the natural
sciences, Price’s subject naturalism doesn’t furnish any way to understand this as a substantive
metaphysical thesis rather than a definitional truth.

159

Lionel Shapiro

Though I’ve argued that Price adopts an excessively restrictive naturalism, my reasons differ
from those of critics who take him to hold that we can only understand human beings, or
engage in philosophy, using natural-scientific methods. Redding (2010, 272) objects to Price
that “the idea that we can only learn about ourselves from science is, many will think, ludicrous”. Yet it would be wrong to attribute that view to Price: as mentioned earlier, he thinks
we can learn normative facts about ourselves by nonscientific means. Likewise, contrary to
Macarthur (2014, 73), there’s little reason to think Price would insist that “persons qua rational
agents are fully understandable, or completely explicable, in scientific terms”.16 And Price’s
assertion of science’s privileged role in the explanation of the functions of expressions and concepts
needn’t amount to asserting that “the scientific framework does have a special kind of priority
because it is from this framework that one practices philosophy” (Macarthur 2014, 80; cf.
Redding 2010, 271–272). Nor does it reveal a “significant bias against acknowledging, in one’s
theoretical voice, the existence of abstract items that are not part of a causal structure studied by
science” (Macarthur 2014, 83, [my emphasis]). There’s no reason to attribute to Price the view
that the functional explanations by which metaphysical puzzles can be sidestepped are the sole
kind of philosophy or theorising. Surely he would agree that his own defenses of (SN1) and (SN2)
are examples of philosophical theory.

6 Subject naturalism and McDowell’s naturalism of second nature
To bring out what’s distinctive in Price’s subject naturalism, it helps to compare it with an
outlook with which it has much in common, starting with its rejection of (ON) based on an
embrace of (SN1). This is McDowell’s “naturalism of second nature” (1994, 84–84,
1998b, 194).
Like Price, McDowell denies that the “dispassionate and dehumanized stance for investigation” rightly taken in natural science should be viewed as “conforming to metaphysical insight
into the nature of reality taken as such” (1998b, 175). In particular, he agrees that (ON)’s rejection
of irreducible facts about moral value can’t be validated by a scientific understanding of the
features of human nature on which the functioning of moral concepts are grounded. Science
needn’t show how our judgments of moral value track the facts, on pain of otherwise exposing
them as merely “subjective responses to a world that contains nothing valuable” (1998a, 166).
Moreover, McDowell and Price explain this common commitment in similar ways.
McDowell puts it in terms of
a relation…between our concepts and the facts of nature that underlie them. The
concepts would not be the same if the facts of (first) nature were different, and the
facts help to make it intelligible that the concepts are as they are, but that does not
mean that correctness and incorrectness in the application of the concepts can be
captured by requirements spelled out at the level of the underlying facts”.
(1998b, 193)
Price too has long been at pains to distinguish the “external” or “explanatory perspective”, from
which natural-scientific facts about human beings are invoked in understanding a discursive
practice, from the “internal” or “participant’s perspective” on that practice from which the
truth of its claims can be assessed (1988, 154, 157, 163).
Finally, just as McDowell (1998b, 186) advocates an “expansion of the notion of the world”
to accommodate facts that have posed placement problems, Price argues that the “e-world”,
consisting of the facts “visible only from within science” and its external perspective, is “properly
160

Price’s subject naturalism

contained in” a more expansive “i-world” consisting of facts visible from the internal perspectives of diverse forms of discourse not limited to science (2013, 52–56).17
Given so much agreement, are there important differences between Price’s subject naturalism and McDowell’s “liberal naturalism” (2009c, 262)? Two related contrasts seem the most
significant. First, Price holds that natural-scientific investigations of discourse can suffice to undermine the mindset that gives rise to placement problems about moral reality. McDowell
disagrees. To be sure, he admits that anthropological “reflections about the benefits of cooperation and social order go some distance towards…making it intelligible that we inculcate
ethical sensibilities in our young” (1998a, 166), and even that scientific investigation of ethical
thought can help “alleviate [a] sense of mystery” (1998a, 165).18 Crucially, though, he doesn’t
think it can yield a fully satisfying “diagnosis and exorcism” of the object-naturalist metaphysical
picture that renders unreduced ethical facts problematic (1998a, 166). Rather, he insists that
undermining (ON) requires conceiving of human subjects from a perspective other than that of
the natural scientist.
To see why, we need to turn to a second contrast. McDowell accords the status of “natural”
to some of the facts of Price’s i-world that fall outside the narrower e-world. “Natural phenomena” in McDowell’s liberal sense include not only phenomena recognised by object
naturalists, but also “manifestations of a second nature acquired in acquiring command of a
language” (2009b, 247) and a proper upbringing, such as rational agency and moral virtue. And
McDowell makes clear that natural phenomena also include whatever phenomena are “open to
view” to one who has acquired such a second nature: “there is nothing against bringing this
richer reality under the rubric of nature too” (1998b, 192).
Does it matter whether we call some facts outside Price’s e-world “natural”? In other words,
does anything rest on whether we follow McDowell to a liberal naturalism, or instead follow Price
to an ontological liberalism that doesn’t insist on recognising as natural any phenomena outside the
purview of natural science? McDowell’s affirmative answer depends on transcendental considerations. His thinking about naturalism is driven by a desire to escape from “transcendental
anxiety” about “the very possibility of thought’s being directed at the objective world” (2009b,
243). According to McDowell, objective thought requires that states of affairs be open to view to
subjects with the appropriately formed second nature. In the case of moral thought, we can only
understand its contentfulness if we see ourselves as having our minds opened by our upbringing to
moral states of affairs. And this understanding of contentfulness can only be available from the
perspective of one who engages in moral thinking. Contrary to Price, then, McDowell denies
that we can understand ourselves as moral thinkers from an external perspective that doesn’t
simultaneously allow us to lay claim to moral knowledge. Finally, as we saw, it’s the fact that
moral states of affairs are open to view by human agents that qualifies them as natural for
McDowell (cf. Macarthur 2014, 83–84). Understanding us as the natural beings we are requires
invoking, as part of nature, the moral reality we naturally have access to.
I’ll conclude by proposing that there is room for a liberal naturalism that constitutes a middle
ground between Price’s subject naturalism and McDowell’s version of liberal naturalism. Earlier, I
argued that if the subject naturalist’s perspective doesn’t allow us to understand ourselves as even
deploying contents in the target discourses, then subject naturalism won’t help with Price’s project of
sidestepping placement problems. And Price’s restrictions on subject naturalism do, by his lights,
put the contentfulness of our talk outside the perspective of subject-naturalistic inquiry. But is
there reason to replace Price’s version of subject naturalism with McDowell’s alternative, on
which acquisition of a second nature makes available direct access to “natural” moral facts?
I’m not convinced this is necessary, though addressing McDowell’s views about contentfulness
would exceed this chapter’s scope. My suggestion here has been that Price’s project can, and should,
161

Lionel Shapiro

be carried out without hewing to his austere strictures: subject naturalist explanations of vocabularies
should be allowed to traffic in propositional-attitude locutions.19 Despite this liberalisation, the
explanations could still be given from a detached, external perspective. They needn’t exhibit
statements using the target vocabulary as cases of truly stating, let alone knowing, the propositions
they state. The result would be a liberal naturalism about human thought and language that seeks to
undermine metaphysical puzzles about (e.g.) moral facts, but one which declines to join McDowell
in extending the realm of the natural to include these moral facts themselves.20

Notes
1 While that terminology first appears in Price (2011d, published 2004), the distinction and its use are
clearly anticipated in the introduction and the concluding section of Price (1988, 4–5, 215–216).
2 For similar distinctions, see Macarthur (2014) and Christias (2019, 509–510). Unlike Macarthur,
Christias sees himself as drawing a distinction within “liberal naturalism”.
3 There are important questions about how “natural science” is delimited; here I can only note that
Price construes it broadly enough to include at least some “human sciences”, specifically the kind of
“anthropological” inquiry with which is concerned (2011c, 29–30). Cf. Redding (2010, 373) and
Macarthur (2014, 73).
4 According to Macarthur (2014, 81), Price doesn’t acknowledge moral truths on a “face-value understanding”. That differs from Price’s own assessment: he insists that his position allows moral talk “to
be taken at face value, but without the metaphysical spooks” (2011a, 147). Macarthur appears to hold
that taking moral truths at face value requires taking moral values to figure in “causal explanations of
our talk of moral values”.
5 The term “location problem” derives from Jackson (1998). As his source for “placement”, Price
(2011d: 6n1) cites an exchange between Simon Blackburn and John McDowell (see Blackburn 1993,
163, whose language is picked up by McDowell 1998a, 162–166). Though I have introduced placement problems as concerned with placing the things we talk of, Price agrees with Blackburn that they
ultimately derive from a concern with placing our talk of those things (see note 8).
6 Price allows that there can also be legitimate inquiry directed at questions of this sort (2011e, 264;
2019, 137). I’ll have little to say about what distinguishes the philosophical perplexities he views as
giving rise to metaphysics, other than that metaphysically reductionist or eliminativist answers are typically couched as views about what the items in question “really” are, or whether they “really” exist
(Macarthur and Price 2011, 235–236). Metaphysics in Price’s sense thus presupposes an “absolute,
theory-independent ontological viewpoint” (2011a, 134). Price’s narrow conception of metaphysics is
criticised by Legg and Giladi (2018).
7 Viewed this way, Price’s subject naturalism aligns closely with the approach Sellars himself announces
in an early paper (1949: 290–92). Here, Sellars aims to show naturalists a way to avoid a “failure of
nerve”. The problem they face is the apparent dilemma between maintaining that mathematical,
normative, and modal concepts are “pseudo-concepts”, and claiming that they are “included within
the scope of empirical science”. His “pragmatist” alternative involves insisting (here in the mathematical case) “that there is no aspect of mathematical inquiry as a mode of human behavior which requires a
departure from the categories of naturalistic psychology for its interpretation”.
8 Elsewhere in the same paper, Price describes his “linguistic conception” of the “origins of placement
problems” differently: “Roughly, we note that humans…employ the term ‘X’ in language, or the
concept X, in thought. In light of a commitment to object naturalism,…we come to wonder how what
these speakers are thereby talking or thinking about could be the kind of thing studied by science”
(2011d, 188 [first emphasis mine]).
9 Indeed, Price argues that subject-naturalistic inquiry calls into question the traditional expressivist
“bifurcation” of assertoric discourse into domains that are, and aren’t, genuinely descriptive. His
functional pluralism can thus be understood as a “global expressivism” (2011f, 2011c, 8–9; 2103, 2019;
Macarthur and Price 2011).
10 Brandom argues, against Price’s “descriptivist subject naturalism”, that subject-naturalistic vindications
of a vocabulary’s respectability should be allowed to avail themselves of any metavocabulary for which that
same task can be performed (2015, 93–95; 91n57). On this more liberal view, what remains of naturalism?
Brandom’s own answer rests on his account of how a vocabulary’s “respectability” is vindicated. That’s

162

Price’s subject naturalism

11
12
13

15
16
17

18
19
20

achieved by showing how the vocabulary can be correctly deployed on the basis of abilities involved in
using “empirical descriptive vocabulary, whether that of common sense, the special sciences, or
fundamental physics”. On this version of subject naturalism, scientific vocabulary enjoys no privileged
role in the explanatory project.
Brandom (2013, 88) expresses doubts about this, based on his view that the explanations may require
using normative vocabulary. See also Redding (2010) and Macarthur (2014).
As Brandom says, “[w]hat the subject naturalist wants is a naturalistic account of the discursive practices
of using the target vocabulary as meaningful in the way that it is meaningful” (2015, 91).
As we saw, Price’s strategy does preclude explaining the functions of expressions by simply specifying
what they are used to “describe” or “represent”, but this needn’t rule out employing intentional
vocabulary in other ways. I’ve argued elsewhere that Price’s expressivist explanations invariably employ such vocabulary (Shapiro 2014, 500).
Redding (2010, 271) suggests that Price is attempting to secure “a ‘soft’ or liberal naturalism in the
culture generally” by “the strategy of adopting a strict or scientistic naturalism within philosophy”. I’ll
argue that Price is doing neither of these things. Macarthur (2014, 76) recognises that Price’s “subject
naturalism is not a form of liberal naturalism”.
Macarthur (2014) and Christias (2019) criticise Price by defending that entailment; Macarthur rejects
the entailment’s conclusion whereas Christias accepts it. I lack space to address their reasons (but see
note 4).
Admittedly, Price sometimes seems to restrict “explanatory projects” to using the vocabulary of natural
science (2010, 179). But perhaps he could say that understanding humans as rational agents isn’t an
explanatory project in the relevant sense.
Against the object-naturalist doctrine that “there can be no facts other than those that would figure in a
scientific understanding of the world”, McDowell insists that “[w]e have no point of vantage on the
question what can be the case, that is, what can be a fact, external to the modes of thought and speech
we know our way around in” (1998a, 164). Price agrees: attempts to use any conception of a “totality
of all the facts” to delimit factual discourse ignore that “what facts we take there to be depends on what
kinds of assertoric claims our language equips us to make” (2013, 54).
Price (2015) criticises McDowell for holding that puzzlement about (e.g.) ethics can’t be addressed by
substantive philosophical theorising. Perhaps the disagreement here concerns only what should count
as “substantive philosophy”.
Of course, the viability of this middle ground depends on the explanatory autonomy of a perspective
on human beings that includes intentional characterisations but not moral ones.
For another discussion of Price’s subject naturalism with attention to liberal naturalism, see Beasley
(2020), which appeared after this chapter was submitted.

References
Beasley, B. Naturalism Without a Subject: Huw Price’s Pragmatism. Inquiry, forthcoming. (2020).
https://doi.org/10.1080/0020174X.2020.1820903
Blackburn, S. Errors and the Phenomenology of Value. In Essays in Quasi-Realism. Oxford University
Press. ([1985]/1993).
Brandom, R. Between Saying and Doing. Oxford University Press. (2008).
Brandom, R. Global Anti-Representationalism. In H. Price (ed.), Expressivism, Pragmatism and
Representationalism. Cambridge University Press. (2013).
Brandom, R. From Empiricism to Expressivism. Harvard University Press. (2015).
Christias, D. Towards a Reformed Liberal and Scientific Naturalism. Dialectica, 73, (2019). 507–534.
De Caro, M. and Macarthur, D. Naturalism and Normativity. Columbia University Press. (2010).
Jackson, F. From Metaphysics to Ethics. Clarendon Press. (1998).
Knowles, J. Naturalism, Pragmatism and the Retreat from Metaphysics: Scientific versus Subject
Naturalism. In J. Knowles and H. Rydenfelt (eds.), Pragmatism, Science and Naturalism. Peter
Lang. (2011).
Knowles, J. Representationalism, Metaphysics, Naturalism: Price, Horwich, and Beyond. In K. Cahill and
T. Raleigh (eds.), Wittgenstein and Naturalism. Routledge. (2018).

163

Lionel Shapiro
Legg, C. and Giladi, P. Metaphysics: Low in Price, High in Value: A Critique of Global Expressivism.
Transactions of the Charles S. Peirce Society, 54, (2018). 64–82.
Macarthur, D. Subject Naturalism, Scientism and the Problem of Linguistic Meaning: Critical Remarks on
Price’s ‘Naturalism Without Representationalism’. Análisis, 1, (2014). 69–85.
Macarthur, D. and Price, H. Pragmatism, Quasi-Realism, and the Global Challenge. Naturalism Without
Mirrors. Oxford University Press. ([2007]/2011).
McDowell, J. Mind and World. Harvard University Press. (1994).
McDowell, J. Projection and Truth in Ethics. In Mind, Value, and Reality. Harvard University Press.
([1987]/1998a).
McDowell, J. Two Sorts of Naturalism. In Mind, Value, and Reality. Harvard University Press. ([1996]/
1998b).
McDowell, J. Conceptual Capacities in Perception. In Having the World in View. Harvard University Press.
([2006]/2009a).
McDowell, J. Experiencing the World. In The Engaged Intellect. Harvard University Press. ([2000]/2009b).
McDowell, J. Naturalism in the Philosophy of Mind. In The Engaged Intellect. Harvard University Press.
([1999]/2009c).
O’Leary-Hawthorne, J. and Price, H. How to Stand Up for Non-cognitivists. Naturalism Without Mirrors.
Oxford University Press. ([1996]/2011).
Price, H. Facts and the Function of Truth. Blackwell. (1988).
Price, H. Reply to Knowles. In M. Miłkowski and K. Talmont-Kaminski (eds.), Beyond Description:
Naturalism and Normativity. College Publications. (2010).
Price, H. Immodesty Without Mirrors: Making Sense of Wittgenstein’s Linguistic Pluralism. In Naturalism
Without Mirrors. Oxford University Press. ([2004]/2011a).
Price, H. Naturalism and the Fate of the M-Worlds. In Naturalism Without Mirrors. Oxford University
Press. ([1997]/2011b).
Price, H. Naturalism Without Mirrors. Oxford University Press. (2011c).
Price, H. Naturalism Without Representationalism. In Naturalism Without Mirrors. Oxford University
Press. ([2004]/2011d).
Price, H. Semantic Minimalism and the Frege Point. In Naturalism Without Mirrors. Oxford University
Press. ([1993]/2011e).
Price, H. The Semantic Foundations of Metaphysics. In Naturalism Without Mirrors. Oxford University
Press. ([2004]/2011f).
Price, H. Two Paths to Pragmatism. In Naturalism Without Mirrors. Oxford University Press. ([1991]/
2011g).
Price, H. Expressivism, Pragmatism and Representationalism. Cambridge University Press. (2013).
Price, H. Idling and Sidling Toward Philosophical Peace. In S. Gross, N. Tebben, and M. Williams (eds.),
Meaning Without Representation. Oxford University Press. (2015).
Price, H. Global Expressivism by the Method of Differences. Royal Institute of Philosophy Supplement, 86,
(2019). 133–154.
Redding, P. Two Directions for Analytic Kantianism: Naturalism and Idealism. In M. De Caro and D.
Macarthur (eds.), Naturalism and Normativity. Columbia University Press. (2010).
Sellars, W. Language, Rules and Behavior. In S. Hook (ed.), John Dewey: Philosopher of Science and Freedom.
The Dial Press. (1949).
Sellars, W. Science, Perception and Reality. Routledge and Kegan Paul. (1963).
Sellars, W. Ontology, the A Priori and Kant. In J. Sicha (ed.), Kant’s Transcendental Metaphysics.
Ridgeview. ([1970]/2002a).
Sellars, W. Some Remarks on Kant’s Theory of Experience. In J. Sicha (ed.), Kant’s Transcendental
Metaphysics. Ridgeview. ([1967]/2002b).
Shapiro, L. Linguistic Function and Content: Reflections on Price’s Pragmatism. Philosophical Quarterly,
64, (2014). 497–506.

164

14
RELAXED NATURALISM: A
LIBERATING PHILOSOPHY OF
NATURE
Daniel Hutto

1 A relaxed philosophy of nature
Naturalisms, of whatever stripe, are quintessential philosophical offerings. Each brand of naturalism – be it restrictive, liberal or relaxed – embeds and articulates its own philosophy of
nature. Such philosophical offerings provide direction with respect to how we should demarcate what belongs to nature and what does not. A naturalistic philosophy also gives guidance
about which methods and means are appropriate for understanding the various aspects of
nature. The success or otherwise of any given philosophy of nature ultimately depends on how
well it stands up as an account of how things hang together. Each such proposal is assessed not
only with respect to what it includes and excludes, but also the coherency of its account of what
is included and excluded.
At one extreme end of the spectrum we find Scientific Naturalism. Scientific Naturalism is
strict in its claim that determining what is included in and excluded from nature is solely the
preserve of the natural sciences. Rosenberg (2014) a fierce defender of this brand of naturalism
describes it as follows:
Naturalism [of the strictly scientific variety] is the label for the thesis that the [only]
tools we should use in answering philosophical problems are the methods and findings
of the mature sciences – from physics across to biology and increasingly neuroscience.
It enables us to rule out answers to philosophical questions that are incompatible with
scientific findings (17).
Relaxed Naturalism, in contrast to Scientific Naturalism, assumes that the mature natural sciences are neither our sole nor even always our best or primary guide to knowing and deciding
what there is (Hutto and Satne 2015, 2018a, 2018b). Rejecting that thesis, Relaxed Naturalism
holds that humanistic disciplines, and their diverse methods and findings, in conjunction with
those of the sciences can be relevant to investigating and understanding what we find in nature
(Hutto 2020).
Relaxed Naturalism resists reductionism yet seeks to draws upon and harmoniously integrate
the discoveries from a wide range of sciences and humanistic disciplines. It maintains that to
fully understand certain features of the natural world – features which include our
DOI: 10.4324/9781351209472-17

165

Daniel Hutto

understanding of our own human practices and activities – requires synthesising findings from,
inter alia, anthropology, developmental psychology, comparative psychology, cognitive archaeology and social neuroscience. Beyond this, a Relaxed Naturalism also holds that, depending on what we are seeking to understand, it is not only appropriate but, indeed,
sometimes necessary to synthesise findings that span not only the hard, natural sciences but also
humanistic disciplines, such as philosophy, history and literature.
In these respects, the approach Relaxed Naturalism recommends agrees substantially with
the core commitments of Liberal Naturalism, as formulated by De Caro and Macarthur (2004,
2010). Going by the official word, Liberal Naturalism seeks to strike a virtuous mean in advancing its philosophy of nature, avoiding the extreme vices of scientism and supernaturalism.
Its three central tenets are as follows: (1) deny that reality is supernatural or that it contains
supernatural beings, powers, forces, and so on; (2) deny that reality is exhausted by what can be
gleaned from composite scientific depictions of reality and (3) affirm that there is a plurality of
non-scientific and scientific forms of knowing and understanding (see Macarthur 2018, 46).
Liberal Naturalism alerts us to and warns us against the tendency to obscure important facts
about our natural situation – facts that will be lost from sight or misconstrued when we become
attached to either of two more extreme philosophies of nature, those that are overly restrictive
or overly permissive in their views about methods and metaphysics. In philosophical circles of
the current secular age, a strong version of Scientific Naturalism has become attractive and poses
the greater threat.
In several key respects, Liberal Naturalism and Relaxed Naturalism are on the same page.
Both resist the idea that the natural sciences are the sole arbiters for determining what there is in
nature and how we can know it. Both of these naturalisms are united in promoting the view
that we need more than the methods and findings of the natural sciences in order to recognise
all that we find in nature and to understand it.
Viewing nature through the lens of Relaxed Naturalism directs us to do more than simply
reject scientistic reductionism and embrace a moretolerant pluralism. It seeks to help us to
understand how philosophy, other humanistic disciplines and various natural and social sciences
can, using their own special methods, take an interest in the same subject matter and contribute
to our understanding of it in complementary, non-competing ways.
The self-styled negative and quietist program of Liberal Naturalism reveals it to be primarily
an attempt to remind us than there is more in nature than dreamt of in the natural sciences. As
such, Liberal Naturalism’s main ambition can be understood as seeking to re-enchant our
conception of nature, as McDowell (1994) recommends, ensuring that the sort of hard-nosed,
mad-dog disenchanted naturalism, of the sort which Rosenberg (2014) promotes, fails to
attract.
A viable, well-rounded naturalism requires something more than taking up a liberal attitude
about the kinds of entities that belong on the list of the legitimately natural. It requires more
than acknowledging that we cannot shed total or complete light on the character of all natural
phenomena if we only employ the methods of the natural sciences.
Adopting an appropriate naturalistic attitude, requires much more than simply liberalising
our thinking so as to become more tolerant to admitting more and more items into an open and
pluralist metaphysics. To achieve a well-balanced yet appropriately relaxed naturalism requires
engaging in the crucial philosophical work of debunking the attractions of misleading pictures
of nature. Such work is needed to free us from certain compelling tendencies of thought that
not only impede how we think about nature and what counts as natural, but also impede us
when it comes to thinking about which methods are legitimate for understanding certain
important aspects of the natural world.
166

Relaxed naturalism

In short, we will only arrive at an appropriate naturalistic attitude if we liberate our thinking
by questioning standard assumptions – including assumptions about what a properly naturalistic
philosophical enterprise looks like – when constructing a viable philosophy of nature.
Importantly, the task of liberating our thinking in this way goes beyond merely adopting a
Liberal Naturalistic attitude.
Consider Gallagher’s (2017) musings about what is required of a satisfactory philosophy of
nature. He holds that an adequate philosophy of nature should take “seriously the results of
science, and its claims remain consistent with them, but it can reframe those results to integrate
them with results from many sciences” (Gallagher 2017, 22).
Gallagher’s proposal, captured in the brief statement above, may seem anodyne enough and
on the right track as an expression of a suitably Relaxed Naturalism. Yet, in one way,
Gallagher’s proposal does not go far enough while, in another way, it goes too far. It does not
go far enough because it fails to mention that the synthesising work needed to understand all we
find in nature will require integrating more than just the findings of various sciences, natural or
otherwise. Such synthesising work also requires taking seriously what can be gleaned from
various humanistic disciplines and adding their contributions into the mix as well.
Yet, elsewhere, Gallagher’s proposal about what an adequate philosophy of nature might
look like goes too far. It suggests that to liberally conceive of nature is to see it as comprising
aspects that are beyond the recognition and reach of natural science and to adopt methods
thatcompel us to “rethink science” (see Gallagher 2017, 127). TheRelaxed Naturalist’s synthesising project is content to let the various sciences and humanistic disciplines take their own
courses, seeking ways to understand how they harmoniously relate without thereby seeking to
“rethink” these sciences or disciplines in an effort to unify them in some stronger fashion.
Though it has some synthesising ambitions, Relaxed Naturalism is content to leave ragged what
we find ragged in the world.
Where Gallagher (2017) and Relaxed Naturalists strongly agree is in thinking that developing and advancing “a philosophy of nature is not to do science…a philosophy of nature is neither
natural philosophy (in the traditional sense) nor the kind of naturalistic philosophy that is
necessarily continuous with science” (23, emphasis added).
The mature natural sciences are, apparently, not in the business of doing the kind of synthesising work needed for articulating any kind of philosophy of nature. It is a job for philosophy, not the hard sciences, to provide such accounts: including engaging in the self-reflective
task of explicating philosophy’s own role in providing such accounts. This is hardly surprising
given it is the job of philosophy, as opposed to that of hard natural sciences, to stand back and
evaluate “the situation in which we find ourselves” (Williams 2006, 182). Such self-reflective
work is philosophical work par excellence.
For the reasons just mentioned, it would be difficult to deny that the articulation of a
philosophy of nature is, at its heart, a philosophical project. Still, not everyone will agree that
articulating a philosophy of nature is, as Gallagher (2017) proposes, “a different kind of intellectual project from science” (22). Those attracted to the view that philosophy reduces to
science or, more softly, that it is best understood as indistinguishably part of, or continuous
with, science, as Quineians and Scientific Naturalists hold, will disagree.

2 Two contrasting visions of nature
The preceding section reminds us why a major part of the work involved in producing an
adequately complete philosophy of nature requires getting clear about the fundamentally
philosophical character of that project. Thus, to be complete, a philosophy of nature
167

Daniel Hutto

unavoidably requires us to understand how philosophy coherently fits into its picture of
nature.1 Deciding which philosophy of nature we ought to adopt thereby requires us to engage
in the meta-philosophical project of deliberating about the point, purpose and methods of
philosophy.
At this juncture, we reach a crucial crossroads. Relaxed and Liberal Naturalists, on the one
side, as opposed to Scientific Naturalists, on the other, fundamentally disagree about what
methods are needed for conducting philosophical work.
This disagreement between naturalists over methods is clear if we consider Williams (2006)
claim that the philosophical work of understanding and making sense of our own situation
reveals philosophy to be a humanistic discipline that is historical at heart. On Williams’ view,
philosophy is always “in continual confrontation with its history, as a resource for comprehension and critique” (Moran 2017, 185).
By Williams’s lights, it is an inevitable and central part of philosophical work that we attend
to the history of how particular arguments and argumentative forms were used in the practice of
doing philosophy down the ages. The purpose of attending to such a history is that it provides
us with an understanding of the contingencies that brought us to where we are now, and which
inform our current thinking. In other words, such a history provides “the story of how one
conception rather than another came to provide the basis of a new legitimacy” (Williams 2006,
191). Crucially, attention to this sort of history is needed, Williams (2006) maintains, “if we are
to know what reflective attitude to take to our own conceptions” (191).
Ultimately, this kind of historical outlook is inescapable, Williams maintains, because:
…in seeking to understand ourselves—we need concepts and explanations which are
rooted in our more local practices, our culture, and our history, and these cannot be
replaced by concepts which we might share with very different investigators of the
world
(Williams 2006, 186–187)2
The idea that philosophy in an inescapably humanistic discipline flies in the face of the Scientific
Naturalism that dominates much of today’s philosophy. Many post-Quinean naturalistic philosophers are inclined to suppose that all proper philosophising must take the sole form of
theorising that aims to add to the compendium of scientific knowledge. Variations on this
theme include actively embracing the idea that philosophy’s only legitimate business is that of
advancing theoretical explanations that are empirically testable, at least indirectly.
This scientistic vision of philosophy could not be more perfectly opposed to the humanistic
vision promoted by Williams. This is highlighted by the fact that Rosenberg (2018), a chief
champion of Scientific Naturalism, holds that history – at least of the narrative sort – is not a
credible candidate to play an indispensable part in any sound philosophical work. This is because, according to Rosenberg, narratively-based history is “not enough to understand anything, ... in fact it’s not even a good way, let alone the best or only way, to understand
anything” (Rosenberg 2018, 5).
The apparent trouble with historical narratives in general is that “they fail to identify the real
causal forces that drive events” (Rosenberg 2018, 6). Because of this feature, historical narratives provide us with no genuine explanatory or predictive grip.
According to a familiar analysis, since the natural sciences are solely concerned with predictive and explanatory power, it is unsurprising that “when it comes to physics, geology and
the other natural sciences, the specialists don’t care about history much at all” (Rosenberg
2018, 7).
168

Relaxed naturalism

It is well known that interpretive disciplines, of the sort that narrative history epitomises, do
not deliver definitive, incontestable accounts of what someone thought and why they thought
it. For this reason, the interpretative disciplines show no prospect of supplying anything like the
kind of theoretically driven explanatory power which is the hallmark of the hard, natural
sciences. By contrast, what explains the explanatory successes of the hard, natural sciences, at
least according to this familiar story, can be put down to the fact that “explanations that survive
the winnowing process are better than the ones that don’t in their predictive success and their
technological applications” (Rosenberg 2018, 8).
In short, and in sum, for the purpose of getting at the underlying workings of the world we
must look to the natural sciences. And, when we look to the natural sciences, we discover that:
“Science is not stories; it’s theories, laws, models, findings, observations, experiments”
(Rosenberg 2018, 4).
To discover the underlying workings of the world – what really moves things – we need an
accurate, ahistorical theory not a narrative history of events. Where does this leave us with
respect to philosophical practice? How might we extend this reasoning to its natural conclusion
and apply it to the activity of doing philosophy itself ?
This is not something Rosenberg comments on directly. Yet we can assume that, in sticking
to the principles of Scientific Naturalism, he would reject Williams’ conclusion that history is
pivotal for conducting proper philosophical work. If philosophy is thought of being continuous
with science and narrative history thought to be an impediment to science, then we can assume
it will be unimportant for philosophy as well.
What methods should philosophy employ in order to conform to Scientific Naturalism? A
hardcore response is that philosophers should restrict themselves entirely to theory- and modelbuilding and to abandon many of their traditional pursuits in line with that injunction.
Yet, in practice, even staunch supporters of Scientific Naturalism – those philosophers who
believe philosophy to be part of science – have not typically limited their targets or methods to
those that would meet the stringent test of being recognisably scientific in character. It is far
from obvious that all of the philosophical questions they seek to address can be answered by
working only with the tools of the hard, natural sciences. It is even less obvious that all
questions of interest can be meaningfully posed in the terms made available by such sciences.3
Recently, philosophical practice has come under intense scrutiny andarguments have been
advanced against traditional forms of philosophical practice that fail to be sufficiently naturalistic –
properly empirical and scientific – in character. In particular, there have been calls on philosophers
to revise and reform old school variants of conceptual analysis in order to bring philosophical
methods in line with what is expected of them within a wholeheartedly naturalist framework.
To shine more light on these issues, the next section examines Machery’s (2017) recent
proposal for revising conceptual analysis to render it properly naturalistic. Machery’s proposal
provides a revealing, concrete test case – one that can help to advance our thinking about the
disagreement between Williams and Rosenberg, and to see more clearly the reason for favoring
a Relaxed Naturalism over its stricter naturalistsic rival.

3 Naturalised conceptual analysis: a test case
At the close of the twentieth century, the brand of analytic conceptual analysis associated with
the Canberra Plan held the mantle of being the “most influential self-proclaimed naturalistic
approach in contemporary philosophical literature in metaphysics” (Ismael 2014, 7).
Today the Canberra approach to analytic metaphysics has fallen out of favor with many
naturalists because it embraces questionable methodological assumptions about how to decide
169

Daniel Hutto

between rival possibilities. Machery (2017) exposes the method of cases fortesting philosophical
proposals as promoting a modally immodest form of philosophy – one that would require us to
have super-natural cognitive capabilities – cognitive capabilities that far outstrip those we are in
fact gifted with in the actual world.
Against this backdrop, in Philosophy in its Proper Bounds, Machery (2017) offers up a new way
of thinking about conceptual analysis in wholly naturalistic terms. Machery’s cry is: “Conceptual
analysis is dead! Long live conceptual analysis!” According to Machery’s blueprint, this new breed
of naturalised conceptual analysis is as much a descriptive as it is a normative project.
The descriptive aspect of naturalised conceptual analysis is to articulate the content of
people’s actual ways of thinking about certain topics. The prescriptive aspect of such work
involves evaluating and “reforming lay concepts that are of philosophical, logical, mathematical,
or scientific importance” (Machery 2017, 215). And the motivation for such revision and
reform is “that existing concepts may be poorly suited to some philosophical or nonphilosophical needs” (Machery 2017, 217).
In Machery’s words, naturalised conceptual analysis, as he conceives of it, is “similar to
explication or to conceptual engineering” (Machery 2017, 209). Crucially, in line with the
agenda of eschewing the supernatural commitments of its philosophical forerunner, Machery’s
naturalised conceptual analysis “calls for an empirical, including experimental, methodology”
(Machery 2017, 210).
What features must concepts be assumed to have such that they can be engineered in the
ways Machery has in mind? Principally, they are assumed to have specific contents that underwrite and drive certain inferential tendencies. It is because we can gain a sense of those
contents and inferential tendencies – because we can reveal the way we inclined to think about
Xs – that we may seek to reform our thinking about Xs so as to change those tendencies.
Fundamentally, in this is way, Machery holds that the descriptive task precedes and grounds the
prescriptive task of his version of naturalised conceptual analysis: “We want to reform a concept, and we want to reform it in a particular way, because we have some grasp of its current
content” (Machery 2017, 218).
At the very heart of Machery’s reformist vision of conceptual analysis is an apparently wholly
scientifically-inspired conception of concepts. According to his official description of this
project, in opposition to a longstanding tradition in philosophy, concepts are assumed to be
“psychological entities and the distinction between what is constitutive of a concept and what is
not is drawn in psychological terms” (Machery 2017, 210).
Machery conceives of concepts as a way of thinking about some specified class of things.
Thus, a concept X is a way of thinking about Xs. More precisely, Machery understands
concepts to be “bodies of information about” individuals, classes, substances, or events. Or,
invoking what he takes to be an equivalent notion, Machery (2017) tells us that a concept
“consists in a subset of people’s belief-like states…about an individual, a class, a substance, or an
event-type” (211).
By Machery’s (2017) lights, “concepts have some content: their referent (what they are
about: dogs, triangles, etc.), and their cognitive content (the body of information about dogs,
triangles, etc.). Concepts are about entities and can characterise these more or less accurately” (211).
The above characterisation of concepts is meant to reflect what psychologists have empirically discovered about the nature of concepts. He maintains that we can be confident that
the best psychological theories warrant characterising concepts as bodies of knowledge that are
stored in long-term memory and used by default in processes underlying competences that
result in judgments about the references of concepts (Machery 2009, 4, 7, 12).4
170

Relaxed naturalism

Concepts – so conceived – are entities in the heads or brains of individuals that drive fast,
automatic and context-independent inferences.5 Because we do not have transparent access to
our concepts, we can be mistaken about their content and character. Moreover, the actual
concepts that drive our thinking may be far from ideal: there is no requirement that our
concepts capture the properties believed to be central – or necessary – to the extension of a
concept. Summing up this view, Machery (2017) tells us:
Concepts are the tracks our minds prefer to travel on. They underwrite a particular
kind of inference: inferences the mind is disposed to draw, that, so to speak, spring to
mind, that it only resists when attention is drawn to particular facts that defeat this
disposition…concepts underwrite inferences the mind draws by default
(Machery 2017, 223)
These commitments about the nature of concepts present a challenge for Machery’s view.
Consider what he has to say about how his version of conceptual analysis might come into play
in rethinking or re-engineering our concept of race. He writes:
An important controversy about the concept of race is whether people conclude by
default that two people are biologically different from the fact that they are judged to
belong to two distinct races. If people’s mind does not preferentially travel on these
tracks, the concept of race can straightforwardly be deployed in critical projects. If it does,
then the concept of race must be modified since, as a matter of fact, races do not differ
biologically: The concept of race leads us to draw erroneous conclusions from true
premises, and at the very least this concept must be reformed”.
(Machery 2017, 224–225, [emphases added])
Machery (2017) repeatedly employs talk of “the” concept of race. But he is not entitled to do so
on his view of what concepts are – especially as he eschews supernatural conceptions of
concepts that regard them to be ideal, abstract denizens of a third realm. But, then, whose
concept of race is at hazard? Is it X’s, Y’s or Z’s concept of race the one under scrutiny?
Presumably, each will have different inferential profiles. But perhaps we can help ourselves to
the thought that some significant portion of certain populations will effectively share the same
inferential tendencies and thus, in effect, share a concept of race. But, if that is so, we will want
to know how and why it is so?
Machery (2017) gives us a clue when he tells us that “the social theorist should not refrain
from describing her work as changing social institutions, perhaps by engineering new concepts”
(Machery 2017, 231). Yet if concepts are, as Machery (2017) maintains, the default tracks along
which individual minds prefer to travel, then it is more likely that we will only be able to reengineer those tracks by re-engineering social institutions that shape the thinking of individuals,
rather than working the other way around. What, after all, would it mean to re-engineer the
belief-like states and contents that underwrite default tendencies of individual minds first? How
would we even begin such a task? And whose concepts, exactly, would we focus upon?
These sorts of practical problems are only compounded if we think about the prescriptive task
of a naturalised conceptual analysis of the sort Machery proposes. How would we begin to decide
if a concept – say, “the” concept of race – is empirically valid or fit for our political purposes if
concepts do not exist in the public sphere but instead are only entities in the heads of various
individuals? Alternatively, we might suppose that our shared concepts are first and foremost
embedded and enshrined in our public practices and institutions. To accept this is to agree that:
171

Daniel Hutto

Ideas and philosophies have histories that are constantly in the making. We have little
chance of coming up with new ideas fit for new times unless we understand the ideas
and time with which they are contiguous. Western democracy, for example, cannot
simply be exported or imposed on countries with very different histories or cultures.
For democracy to travel it must adapt.
(Baggini 2018, xviii)
Baggani’s (2018) example the concept of democracy and questions of the tenability of transporting into new contexts reminds us that if we are to understand and evaluate shared concepts
we need, of necessity, to attend to the forms of life – the practices and institutions – that
surround and foster them. To imagine any significant change to our concepts requires imagining many other connections in our practices to be different. Evaluating conceptual change
demands that we imagine many things that we take for granted to be different than they are –
not everything, of course, but quite a lot. For example, to change any concept that is significant
to our lives would change an indefinite number of related practices, even if we cannot anticipate an exact list of which changes would be necessary.
Chalmers (2020) likens engineering concepts to engineering a bridge. In exploring the
puzzle of whether such engineering should be regarded as a case of de novo engineering or reengineering he focuses on the hard case of the bridge that now replaces the original Tappan Zee
Bridge, both of which span the Hudson river in roughly the same location in upstate
New York.
What makes Chalmers’s question hard to answer – is this a new bridge or not? – is that both
bridges are artificial structures that are designed more or less to fulfill precisely the same role in
the lives of the bridge users. Replacing one bridge for the other may raise metaphysical puzzles
of the kind of concern to Chalmers, but it does not raise significant questions when it comes to
evaluating the fitness for purpose of the two bridges.
Consideration of the case of these two bridges provides us with a salient reminder. Most of
our significant concepts are not the products of artificial design that can be replaced like the
Tappan Zee Bridge. They are interwoven into the fabric of our lives in complex ways such that
adjusting or replacing them will require detailed and careful attention to the organic nature and
histories of our original concepts. For these reasons, we ought to be as cautious of conceptual
engineering our living ideas as we are about genetic engineering living systems.6
Taking stock of all of these points, we have definitive reason for thinking that when it comes
to answering evaluative and prescriptive questions about whether and how we should reengineer significant concepts we cannot proceed by simply looking exclusively to the findings
of hard natural sciences.
The reason is simple. We simply cannot look to scientific theories and findings alone, stripped
of all cultural content and baggage, in pursuing such philosophical projects. This is because the
first step in getting a handle on the shared concepts under scrutiny is to get an insider understanding of such concepts – an “understanding obtained as part of language acquisition and of
enculturation, that is, of the immersion into a shared linguistic practice” (Glock 2017, 85).
This is why we need to attend to a concept’s history when evaluating its fitness for purpose.
We need to situate concepts against their background in our lived practices. Conceptual
clarification is “ultimately an attempt to articulate participatory knowledge of a shared language” (Glock 2017, 86). That is to say, “one must start out by characterizing how these
expressions function in established practice” (Glock 2017, 87). Yet to evaluate the fitness of our
concepts we need to know something about how they, and why they, came to occupy the
positions that they do within our practices.
172

Relaxed naturalism

These reminders put us in a position to see why Rosenberg’s (2018) hope of avoiding
history and focusing solely on the content of current thoughts,wants, and desires when doing
philosophy is forlorn. Such an approach is hopeless if “the content of our concepts is a contingent historical phenomenon” (Williams 2006, 191).
Fundamentally, we cannot so much as articulate or explicate what we currently believe
without looking to the sedimented history of the concepts that inform and influence out
thinking. It is because our concepts have a history – a public history – that historical investigations
are necessary, indispensable and nonnegotiable for any kind of conceptual engineering project.
Engaging with the history of ideas is necessary in our philosophical pursuits because the content of
our thoughts and beliefs – what we might possibly think in the here and now – does not stand free
from that history. Past notions, and deeply ingrained ways of thinking, can hold sway over us,
whether we know it or not. The fact is that when thinking about our situation, “we are not
unencumbered intelligences selecting in principle among all possible outlooks” (Williams 2006, 194).
Arguably, one of the main sources of encumbering pictures comes from the history of those
ideas themselves. We are affected and influenced by that history without always knowing it.
Thus, with reference to concepts such as “freedom”, and “reason” Williams (2006) reminds us
that “philosophers cannot altogether ignore history if they are going to understand our ethical
concepts at all” (191).7
Expanding on this point, Williams (2006) remarks that
there are some…virtues, such as authenticity or integrity of a certain kind, which are
as a whole a manifestly contingent cultural development; they would not have
evolved at all if Western history had not taken a certain course…the reflective understanding of our ideas and motivations, which I take to be by general agreement a
philosophical aim, is going to involve historical understanding. Here history helps
philosophical understanding or is part of it (192).
Williams makes a crucially important point, but his lesson does not only apply to ethicopolitical
concepts. There are clear cases in which we need to look to the history of ideas in order to help
better understand how philosophical pictures that took root in the past can illicitly and unhelpfully hold sway over our thinking now, also without our knowing or be aware of such
influence. Indeed, we are perhaps nowhere more susceptible to subtle metaphors, and the
metaphysical siren songs of picture-driven theorising to which they give rise, than when it
comes to thinking about the nature of minds – including how we tend to think about the
nature of concepts themselves.
For example, as Machery’s (2017) work reveals, the idea that mental processes are operations
over inner mental objects remains popular in today’s sciences of the mind. That is to say many
philosophers and scientists of the mind are committed to a picture according to which “concepts
are…literally mental objects” (Fodor 2003, 14). Arguably, this idea can trace its ancestry back to
concepts of “mind” and “idea” that arose in Western philosophy during the early modern period
– from the mid-seventeenth to the mid-eighteenth century – with the rise of empiricism. A new
concept of “mind” as a kind of container of mental happenings was first forged at this time,
displacing the previously dominant Aristotelian conception of mind (Fischer 2011, 35). As Fischer
(2011) emphasises, empiricist ideas – which were new at the time – are neither products of pure,
untainted scientific theorising nor unproblematic snippets of common-sense. Instead, they are,
decidedly philosophical in origin. Yet once the philosophical picture was articulated it gave rise to
“fresh intuitions…shared, without explicit argument, by many early modern thinkers” (Fischer
2011, 37). To the extent that philosophers and psychologists today remain attached to remnants of
173

Daniel Hutto

the core of the early modern empiricist picture of mind, they may have fallen into a trap that lies
in our language and the history of our ideas.
Ultimately, the contingent, cultural character of our concepts is what makes it the case that
philosophical work of clarifying, evaluating and possibly reforming our concepts has a necessarily and unavoidably historical dimension. Contra Rosenberg, and the extreme Scientific
Naturalists, it turns out that predictive and explanatory power are neither the only, nor even the
primary things that should interest us in pursuing our philosophical ends.
Philosophers inspired by Wittgenstein who assume a more relaxed, naturalistic attitude than
their exclusive scientific cousins are equally critical of the traditional analytic version of conceptual analysis. They too see the job of naturalistic philosophy as being to “root out, diagnose
and treat metaphysics” (Macarthur 2018, 42). They too seek to “overcome the urge to provide
metaphysical explanations of phenomena in general from the armchair” (Macarthur 2018, 45).
Moreover, they too view the primary task of philosophy to be that of “elucidating existing
concepts” (Glock 2017, 81). And they recognise that, “in addition to the application and the
elucidation of concepts, there is conceptual construction or concept-formation, the devising of
novel conceptual structures” (Glock 2017, 80).
Pivotally, however, the Relaxed Naturalist way of conceiving of concepts is as existing as part
of the living fabric of our public practices rather than being confined to the heads of individuals.
They do not look solely to the natural sciences to decide how we should think about the
properties of concepts. Relaxed Naturalists are prepared to recognise that our concepts are bound
up with the “ordinary use of words as they function in the everyday lives of human beings
considered as natural beings in a natural world” (Macarthur 2018, 42). In agreeing with
Wittgenstein in this key respect, the approach adopted by Relaxed Naturalists, despite resisting
Scientific Naturalism, also “deserves the name of naturalism” (Macarthur 2018, 42).

4 Conclusion
If we are to clarify and evaluate our concepts, we must delve into and attend to the history of
ideas when doing philosophy. Exposing the deep-rooted ways that we tend to think about
specific topics can be part and parcel of what is needed to clarify, liberate and advance our
thinking about those topics.
The preceding analysis has given reason for thinking that to succeed in a project of constructing an adequate philosophy of nature we need to liberate our thinking about the nature of
philosophy itself, freeing it from purely scientistic conceptions.
Relaxed Naturalism provides a framework for doing just that – one that respects scientific
findings but does not defer wholly to the hard, natural sciences, giving them sole and final say
about what the natural world contains and how we should study it.

Notes
1

Gallagher (2017) proposes that enactivism might provide a satisfactory philosophy of nature. He thinks
it is a real contender because “enactivism involves not only a rethinking of the nature of mind and
brain but also a rethinking of the concept of nature itself ” (Gallagher 2017, 126). Yet, even though
enactivism may well provide ***crucial insights into the links between and, perhaps, even the origins
of life and mind, it falls short of providing a complete and adequate philosophy of nature. That is
because, as just noted, to be complete such an account requires, inter alia, giving an account of the
nature of philosophy’s synthesising work itself. In this respect, it is not clear that the core tenets of
enactivism have much to say on that score directly. Of course, there are versions of enactivism that are
consistent with and can sponsor an adequately coherent and complete philosophical vision of nature.

174

Relaxed naturalism
2
3

Elaborating this point, Williams (2006) tells us that, “If one is to understand our own view of such
things…one must try to understand why they take certain forms here rather than others, and one can
only do that with the help of history” (Williams 2006, 192).
Reporting on the rise of experimental philosophy, D’Oro and Overgaard (2017) take stock of pressure on
naturalistic philosophers bring their methods more in line with their stated views. As these authors observe,
Despite countless substantive differences in terms of style, general outlook, and much more, most
philosophers – whether ‘analytic’ or ‘continental’ – arguably used to agree on one thing: that the
activity of philosophizing differed in significant ways from the typical activities of empirical scientists…In recent years, however, this erstwhile consensus – or near consensus – has shattered (1).

Although Machery (2009, 2017) claims that the above characterisation of concepts derives wholly from
empirically focused and experimentally supported psychological theorising, it will be clear to informed
onlookers that the above conception of concept is, at best, an impure product of philosophical and
psychological thinking: in freely invoking the notions of “content”, “reference”, “aboutness” and
“accuracy” it has philosophical fingerprints all over it.
Crucially, Machery (2017) reports that as he conceives of its conceptual analysis concerns empirically
divining the concepts that we actually use in thinking about a given topic and in this respect it “does
differ from the empirical study of what concepts are about. The former involves psychology, the
second not! The former involves studying how we think about [Xs], the latter only studies what [Xs]
are, not how we think about [X]s” (Machery 2017, 226).
The concern to ensure that conceptual engineering projects respect the living character of our concepts
recalls a standard worry raised about the Canberra Plan’s philosophical project. Braddon-Mitchell and
Nola (2009) present the worry as follows: “Since Canberra is a planned city founded originally as the
seat of the Federal Government of Australia, its detractors complain that it lacks the features that arise
in cities that grow organically and have diverse inhabitants who are not largely government bureaucrats” (1). The emphasis on understanding the organic history of our concepts is closer in spirit to the
Sydney Plan, which promotes a style of naturalism that “brings together different strands of pragmatism” (Ismael 2014, p. 7). “a positive, anti-representationalist understanding of linguistic and mental
phenomenon” (Ismael 2014, p. 7). Thus,
Where the Canberra Plan asks, ‘what do x beliefs represent?’ the Sydney Plan raises the question
‘what facts about ourselves and the world jointly support the formation of x beliefs and the role
they play in our lives?’. Where the Canberra Plan conceives of the relationship between everyday concepts and the Absolute structures described by a fundamental theory in terms of
semantic notions (reference, truth, and satisfaction), the Sydney Plan substitutes a fully detailed
side-on account of use that may or may not take the form of a traditional theory of reference.
There will be a story about agents and their relation to the world and how the concepts facilitate
their interaction, but there won’t be anything that looks recognizably like reference to an
independently well-defined feature of the landscape (Ismael 2014, 7).

Of course, as Baggini (2018) reminds us, the
relationship between classical philosophical texts and the ‘folk philosophy ‘of a people is clearly
not a simple one…Most Americans and Europeans, for example, assert the value of individual
freedom and liberty without any deep knowledge of how these concepts have been justified and
explained by their philosophers (xvi).

References
Baggini, J. How the World Thinks: A Global History of Philosophy. Granta. (2018).
Braddon-Mitchell, D. and Nola, R. (eds.). Conceptual Analysis and Philosophical Naturalism. MIT
Press. (2009).

175

Daniel Hutto
Chalmers, D. What is Conceptual Engineering and What Should it Be? Inquiry. DOI: 10.1080/0020174
X.2020.1817141. (2020).
D’Oro, G. and Overgaard, S. (eds.). The Cambridge Companion to Philosophical Methodology. Cambridge
University Press. (2017).
De Caro, M. and Macarthur, D. (eds.). Naturalism in Question. Harvard University Press. (2004).
De Caro, M. and Macarthur, D. (eds.). Naturalism and Normativity. Columbia University Press. (2010).
Fischer, E. Philosophical Delusion and Its Therapy: Outline of a Philosophical Revolution. Routledge. (2011).
Fodor, J.A. Hume Variations. Oxford University Press. (2003).
Gallagher, S. Enactivist Interventions. Oxford University Press. (2017).
Glock, H.-J. Impure Conceptual Analysis. In G. D’Oro and S. Overgaard (eds.), The Cambridge Companion
to Philosophical Methodology, pp. 77–100. Cambridge University Press. (2017).
Hutto, D.D. and Satne, G. The Natural Origins of Content. Philosophia, 43(3), (2015). 521–536.
Hutto, D.D. and Satne, G. Naturalism in the Goldilock’s Zone: Wittgenstein’s Delicate Balancing Act. In
K.M. Cahill and T. Raleigh (eds.), Wittgenstein and naturalism, pp. 145–197. Routledge. (2018a).
Hutto, D.D. and Satne, G. Wittgenstein’s Inspiring View of Nature: On Connecting Philosophy and
Science Aright. Philosophical Investigations, 41(2), (2018b). 141–160.
Hutto, D.D. Philosophers’ Magazine, 88, pp. 75–82. (2020).
Ismael, J. Naturalism on the Sydney Plan. In M. Haug (ed.), Philosophical Methodology: The Armchair or the
Laboratory, pp. 86–104. Routledge. (2014).
Rosenberg, A. Disenchanted Naturalism. In B. Bashour and H.D. Muller (eds.), Contemporary Philosophical
Naturalism and its Implications, pp. 17–36. Routledge. (2014).
Rosenberg, A. How History Gets Things Wrong: The Neuroscience of Our Addiction to Stories. MIT
Press. (2018).
Macarthur, D. Wittgenstein’s Liberal Naturalism of Human Nature. In K.M. Cahill and T. Raleigh (eds.),
Wittgenstein and Naturalism, pp. 89–144. Routledge. (2018).
McDowell, J. Mind and World. Harvard University Press. (1994).
Machery, E. Doing Without Concepts. Oxford University Press. (2009).
Machery, E. Philosophy in its Proper Bounds. Oxford University Press. (2017).
Moran, R. The Philosophical Imagination: Selected Essays. Oxford University Press. (2017).
Williams, B. In A.W. Moore. (ed.), Philosophy as a Humanistic Discipline. Princeton University
Press. (2006).

176

15
LIBERAL OR RADICAL
NATURALISM?
Joseph Rouse

1 Introduction
Liberal naturalism, as presented in De Caro and Macarthur’s (2004, 2010) anthologies and other
work by their contributors, is not a unified credo but a loosely affiliated group of revisionist
responses to the dominant conceptions of philosophical naturalism in the Anglophone world.
That lack of unity complicates the task of assessing the relation between liberal naturalism and
other revisionist reconceptions of naturalism in philosophy, such as the critical analysis of
philosophical debates over naturalism I advanced in How Scientific Practices Matter (Rouse 2002),
and the more constructive re-conception of naturalism I developed in Articulating the World
(Rouse 2015).
My intervention into philosophical discussions of naturalism has important affinities with
many of the canonical contributions affiliated under the liberal naturalist banner. We share a
more pluralistic conception of scientific understanding than is characteristic of orthodox naturalisms, and reject conceptions of nature that would require error-theoretic, reductionist, or
nontruth-conducive treatments of conceptual, epistemic, moral/political, or aesthetic normativity. I endorse liberal naturalists’ emphasis upon “anti-supernaturalism” as the most definitive
naturalist commitment, and my view has some overlap with the primacy some liberal naturalists
(e.g. Price 2004, 2011) accord to understanding human conceptual and epistemic capacities as
natural phenomena (“subject naturalism”) over seeking scientific imprimatur for a physicalist or
other scientistic metaphysics. Some recognised liberal naturalists also share my concerns over
the naturalistic credentials of orthodox conceptions, whether due to the extended “placement
problem” of accounting for the normativity of scientific understanding in orthodox terms, to
their violations of naturalistic scruples against “first philosophy”, or to their natural-theological
conceptions of scientific understanding as aspiring to a gods-eye view from “sideways on”.
The differences between a liberal naturalist orientation and my approach are at least as
important as the similarities, however, and they are also less well understood. These differences
readily fall into three groups. First, most liberal naturalisms begin from a concern to accommodate a more adequate conception of mind, language, ethics, social and political criticism,
aesthetics, or epistemology within an understanding of human life as a natural phenomenon,
while still retaining appropriate respect for natural scientific understanding. My project begins
instead with work in the philosophy of science and interdisciplinary science studies that engages
DOI: 10.4324/9781351209472-18

177

Joseph Rouse

closely with scientific research, and seeks a better understanding of what the sciences accomplish, how they do that, and how to bring those practices and achievements into more effective
conversation with philosophical work on other topics. Second, I find many liberal naturalist
accounts too deferential to orthodox naturalist conceptions of natural scientific methods and
achievements. Too often, liberal naturalists concede “the scientific image” to orthodox naturalists, and seek only a more expansive conception of nature. Third, I find some versions of
liberal naturalism to be too tolerant, underestimating the extent and overlooking some of the
ways in which scientific inquiry bears upon their primary philosophical concerns. For the
purposes of understanding these differences and their importance, I propose to expand the
liberal naturalists’ taxonomy, to distinguish both orthodox and liberal naturalism from the more
radical naturalism that I advocate.

2 Orthodox and liberal naturalisms
Orthodox naturalism also encompasses a variety of positions that characterise in different ways
the philosophical sufficiency or completeness they ascribe to scientific understanding. The
distinction between ontological naturalism – the entities that play a role in scientific explanations are the only entities there are – and methodological naturalism – scientific inquiry
and its methods provide the only genuine or reliable basis for knowledge – marks one familiar
division among these alternatives. Orthodox naturalists also divide over the unity of science:
physicalists insist that what there is can ultimately be reduced to or supervenes upon physical
entities, or that the methods of the “special sciences” are dependent upon or legitimated by an
understanding of their physical basis; pluralists recognise the ontological or methodological
autonomy of astronomy, chemistry, biology, the neurosciences and perhaps geology or the
environmental sciences. Differences also arise concerning what kinds of explanatory constituents or strategies are taken over from their favored sciences: do the sciences uncover singular causes, causal structure, counterfactually robust regularities, governing laws or symmetries,
or some combination thereof, or do naturalists simply defer to scientific authority by accepting
whatever explanatory strategies the sciences invoke?
All of these views deploy a dual conception of “the scientific image”, as both an image of
nature as conceived and explained scientifically, and an image of what a genuinely scientific
understanding of nature would be. Ontological naturalists are committed to some version of
what Teller (2001) tellingly and critically depicts as “the Perfect Model Model” of scientific
understanding. Methodological naturalists presume some way of demarcating genuine scientific
methods or achievements, and perhaps of specifying scientific “methods” in ways that can be
distinguished from their entanglement with theoretical understanding of their domains.
Orthodox naturalists also respond in various ways to recognition of the fallibility and incompleteness of current scientific understanding and methodological choices. Whether a naturalistic conception of what there is or of how it can legitimately or reliably be known appeals
to the ontological commitments or methodological choices of some presumably more adequate
future science, or insists that philosophy can do no better than to work within the limitations of
the best contemporary science, orthodox naturalism must accommodate the revisability of the
scientific image.
Orthodox naturalists almost invariably insist upon the empirical accountability of what they
take to be the scientific image of nature, but they rarely acknowledge their own empirical
assumptions about how scientific inquiry is conducted, what forms its theoretical understanding
takes, or what considerations are germane to scientific decision-making. They now have
available more than a half century of careful, detailed work by philosophers of science and
178

Liberal or radical naturalism?

interdisciplinary science studies scholars who insist upon close attention to scientific practice
and scientific understanding in practice. Although that body of work has profoundly revised
familiar philosophical conceptions of scientific work and its achievements, orthodox philosophical naturalists have made almost no reference to that literature, and made no serious effort to
accommodate or account for its findings, to the detriment of their aspirations to naturalism.
The various strands of liberal naturalism resist easy generalisation, apart from their resistance
to the orthodox versions of ontological or methodological naturalism. I will emphasise three
broad features characteristic of many of the canonical contributors to liberal naturalism, which
allow for important contrasts to the more radical naturalism advanced in Articulating the World.
Liberal naturalists differ in the extent or character of their exemplification of these features,
which partially accounts for some of the overlap between my views and those of liberal naturalism. First, apart from endorsing a more pluralistic conception of the sciences and scientific
understanding, liberal naturalists typically accept an image of natural science consonant with
those endorsed by orthodox naturalists. For McDowell (1994), natural scientific understanding
is the realm of law. For Davidson (1980, 1984) it is a closed, “homonomic” causal system. For
Price, it is “the sum of all we take to be the case” (2011, 28). Liberal naturalists also match the
orthodox in their inattentiveness to the more complex portrayals of scientific practice and
understanding in recent philosophy of science and science studies.1
A second characteristic feature of liberal naturalisms is that their acceptance of an orthodox
scientific image with restricted scope does important contrastive work. Liberal naturalists typically conceive of nature as more inclusive than the scientific image, with the former rather
than the latter as their touchstone for their commitment to anti-supernaturalism. Thus,
McDowell (1994) contrasts acculturated second nature to a law-governed first nature,
Davidson’s (1980) anomalous monism allows only token identity between events understood
within a closed system of laws and those same events interpreted under the constitutive ideal of
rationality, Price (2011) emphasises the functional roles of other linguistic practices and their
autonomy from natural scientific determinations of what is the case, and Macarthur (2010) calls
for an expansive naturalism that “takes the human sciences seriously”.
A third characteristic feature of liberal naturalism is a propensity to defend the autonomy of
ethics and politics, rational thought and conversation, human sciences and culture from detailed
accountability to the natural sciences. Liberal naturalists are united, for example, in denying that
due respect for scientific understanding requires the elimination or reduction of normative
discourse, or its relegation to a diminished cognitive or semantic status, as in various forms of
non-cognitivism or error theories. Liberal naturalists nevertheless hold widely variant constructive views in these domains, and they are typically able to do so because their shared
commitment to a liberal anti-supernaturalism places minimal constraints on normative discourse, apart from rejecting any conception of normative facts as independent of human
practices or as accessible through a sui generis faculty of intuition (De Caro and Macarthur 2010,
3). Liberal naturalists may argue for a specific conception of nature that places more stringent
constraints on normative discourse or its content, authority and force, but such constraints are
not endemic to liberal naturalism as such.

3 Scientific understanding in practice
A more radical naturalism begins at home, in the natural sciences. A viable naturalism must be
able to show how scientific understanding is a scientifically intelligible natural phenomenon.
Most orthodox and liberal naturalisms do not satisfy this criterion, although I cannot defend that
claim here. My more limited aim is to indicate how my more radical naturalism proposes to live
179

Joseph Rouse

up to it. Two mutually supportive tasks must be undertaken. The conceptual capacities exercised in scientific work and many other aspects of human life must be adequately situated
within the best scientific understanding of the world and our place within it. This explicative
and explanatory project must in turn be shown to exemplify an empirically and normatively
responsible account of scientific work, which itself exemplifies its proposed scientific understanding of human conceptual capacities. The motivating insight of Articulating the World is that
we must revise our familiar philosophical understanding of both the sciences and of human
conceptual capacities and practices in order to satisfy this dual criterion.
I begin by summarising some central themes of the re-conception of scientific understanding
put forward in Articulating the World. This summary must nevertheless leave out much of its
supporting detail, not only as presented in the book, but drawing upon other work by many
philosophers, historians, and scholars in feminist and other cultural studies of the sciences. The
first theme is to replace philosophers’ epistemological preconception of “the scientific image”
with an account of scientific understanding as embedded in ongoing research practices. Both
orthodox and liberal naturalists identify a scientific conception of the world with a body of
justified or reliable scientific knowledge. Researchers, however, understand their domains in
ways that exceed and revise what can be codified as established knowledge, and are oriented
toward its further refinement, extension and revision. The scientific review literature does not
compile a comprehensive summary of what is already known, but instead reorganises past
achievements as relevant background oriented toward subsequent research that would revise
their sense and significance. When scientists do participate in summarising an epistemic consensus, as in the reports of the Intergovernmental Commission on Climate Change, researchers’
own understanding of the field is always already directed beyond the consensus view. The
difference can be readily expressed in Sellarsian terms. Orthodox and liberal naturalists identify
“the scientific image” as a position within the space of reasons, a body of claims that have been
justified and accepted scientifically, or as I earlier quoted Price, “the sum of all we take to be the
case”. Scientific understanding in practice is instead an ongoing reconfiguration of the space of
reasons, of what can count as intelligible and significant projects, defensible positions, reasons
for or against them, and possible ways of extending or revising them. Science offers not a single
“image” of the world, but a conceptual space of research opportunities and intelligible
disagreements.
The second theme in my re-conception of scientific understanding is to recognise that it is
not confined to the verbal or mathematical development of theoretical models and claims. One
important aspect of this expansion is to recognise the conceptual significance of experimental
systems, procedures and skills.2 Scientists reconfigure the space of reasons in significant part by
rearranging things in ways that make previously inaccessible aspects of the world discernible and
intelligible. Their apparatus, materials, procedures and phenomena are integral to how scientists
understand their domain; understanding the capacities, limits and opportunities embedded in
their material systems and skills is central to scientific understanding and the ways it projects
beyond any current codification of knowledge. These facilities and the practices and skills that
deploy them also involve discipline-building: opening conceptual domains in the sciences is not
separable from the institutional realisation, financial support and pedagogical renewal of the
disciplines that sustain a research domain.3
A third theme is a reconception of the modal character of scientific understanding. Other
naturalists have mostly started from philosophical pre-conceptions of natural laws or causal
relations and their necessity. My more radical naturalism starts instead by asking what work the
lawfulness of scientific concepts and capacities does in scientific practice. Building on Lange
(2000, 2007) and Haugeland (1998, 2013), I emphasizeemphasise that lawfulness plays essential
180

Liberal or radical naturalism?

roles in inductive confirmation, experimental design and counterfactual reasoning more generally. Lawful patterns do not acquire the requisite modal significance individually, but only as
belonging to interconnected, open-ended, conceptual-domain-constitutive groups of such
patterns, which together remain invariant under any counterfactual supposition consistent with
the whole group. The reliability and resilience of the instruments, skills, and practices that
sustain a research field are integral to this holistic invariance under relevant counterfactual
perturbation. To undertake empirically accountable scientific research requires implicit commitment to the lawful invariance of the conceptual relations and material practices of a research
field.4 The counterfactual stability of a conceptual domain is not a synchronic status, but an
ongoing issue for a temporally extended research practice. Lawful patterns, the conceptual
relations they encompass, the skills to recognise and characterise their instances, and the normative concerns that orient inquiry within that domain are open to ongoing extension, revision
and repair to maintain a coherent domain of inquiry. The resilience and reliability of those
capacities for conceptual and practical adjustment are an integral part of the domain’s characteristic forms of counterfactually extended invariance (necessity).
The resulting modal pluralism of scientific domains provides a more fully developed conception of scientific pluralism that often accords with liberal naturalists’ challenges to orthodox
forms of scientism. The fourth and final theme from my re-conception of scientific understanding
nevertheless tempers that pluralism by recognising the character and importance of accountability
across conceptual domains. Adapting and revising terms from Davidson (1980), I argue that
scientific understanding develops both homonomically and heteronomically. Scientific domains
are always open to more extensive or intensive homonomic (“internal”) conceptual development,
which is both enabled and constrained by the need for empirically accountable adjustments of
their concepts, skills and standards while maintaining the counterfactual invariance that allows
their inductive projectibility, explanatory significance and other inferential capacities.
Scientific domains are also heteronomically accountable in less systematic ways to practices
and achievements in other domains, including practices in other aspects of human life. An
entirely self-contained domain of concepts, claims and practices would be, in McDowell’s
picturesque phrase, a “frictionless spinning in a void” (1994, 66).5 The conceptual significance
of a scientific domain arises in part from the ways in which it makes a difference to what is said
and done elsewhere. Sciences differ in their ranges of counterfactual invariance, the standards
and skills for the application of their concepts and methods, and how they matter to other
aspects of what we say and do, but they are only partially autonomous. They cannot license
claims in conflict with those of other practices without requiring adjudication; they may only
answer to relatively gross or insensitive features of the more fine-grained conceptual discriminations within other fields, but such conflicts are not thereby irrelevant. New tools, skills
or conceptual issues drawn from other domains can also be disruptive or productive for them.
Moreover, the evidential standards governing scientific claims within a discipline are also
heteronomically accountable to the broader consequences of accepting a hypothesis, choosing a
methodology, or interpreting data in one way rather than another. The range of normative
considerations relevant to choices at every stage of scientific practice extends well beyond
narrowly epistemic considerations, because the material and inferential consequences of those
choices are also more far-reaching (Biddle and Kukla 2017; Rouse 2018).6 As we shall see in the
next section, however, such heteronomic accountability is multi-directional: ethical, political,
economic and cultural practices are also answerable to material and conceptual developments in
the sciences. A more radical naturalism thereby steers between the scientistic imperialism of
orthodox naturalism, and the nearly unfettered autonomy from natural scientific understanding
that some liberal naturalists and social theorists too readily ascribe to other aspects of human life.
181

Joseph Rouse

This re-conception of the scientific image highlights the ways in which the sciences produce
not a unified representation of the world as a whole, but an ongoing reconfiguration of the
conceptual capacities and possibilities that constitute the space of reasons, including the heteronomic relations through which scientific and other conceptual practices answer to one
another. I now turn to how recent scientific developments have opened new ways to understand how our scientific and other conceptual capacities and the ways of life they make
possible are recognisable as scientifically intelligible aspects of nature.

4 Two-dimensional biological normativity
Human beings are organisms. Our ways of life are both enabled and constrained cosmologically,
thermodynamically and more specifically bioenergetically and biochemically, but we now best
understand the forms those processes take in developmental and evolutionary biological terms.
The emergence of ecological-developmental biology, niche construction theory and other
aspects of the extended evolutionary synthesis (Pigliucci and Müller 2010) not only more
adequately conceptualises our evolutionary history, but also our evolved ways of life. Organisms
are not self-contained entities. They – we – are thermodynamically open systems, sustained by
ongoing exchange of energy, other resources and waste products with their environments. The
environment of an organism is not its physically specifiable surroundings, but rather “a spatial
and temporal juxtaposition of bits and pieces of the world that produces a surrounding for the
organism that is relevant to it,…[a] space defined by the activities of the organism itself”
(Lewontin 2000, 52–53). The converse is also true: the organism’s traits and way of life are ways
of responding to and acting on its environment. Put another way, organism and environment
are not two distinct entities that interact, but a pattern of “intra-action” (Barad 2007) in which
each only becomes determinately bounded through their intra-active involvement.
Neo-Darwinist evolutionary theory treated an organism’s environment as an external causal
influence driving natural selection. The extended synthesis recognises developmental and
evolutionary processes as causally bidirectional, because organisms change their environments in
ways that affect selection pressures on their lineages. Niche construction encompasses not only
physical reconstruction of environments and migration to new ones, but also behavioral patterns that are integral to the developmental reconstruction and evolutionary maintenance of
those patterns and their descendants. Language and other conceptual capacities are exemplary
forms of behavioral niche construction. Human beings normally develop, and only develop
normally, in conceptually articulated environments. Spoken and written language, ostensive
and vocative performances, images, musical expression, dramatic performances, and equipmental complexes are pervasive, salient and developmentally scaffolded features of human
environments, which have coevolved with human bodies and ways of life.
De Caro and Macarthur rightly insist that “the debate over which form of naturalism is best
will depend to a considerable extent on which provides the best account of core normative
phenomena such as reasons and values” (2010, 9). My more radical naturalism understands
normativity as a natural, biological phenomenon. Organisms and their lineages are Aristotelian
energeia, goal-directed processes whose goal is internal to the continuing process (Okrent 2007,
2017).7 Organisms develop and lineages evolve, but what is biologically at stake in those
changes is whether that process continues to sustain and reproduce itself, in whatever form it
takes under changing conditions, some of which it helps bring about. The emergence of
language and other partially autonomous practices in the hominin lineage brings about more
complex forms of biological normativity, however. Performances of these practices are
proximately open to assessment in relation to other such performances, for example, in the case
182

Liberal or radical naturalism?

of utterances, their audibility, discernibility, grammaticality, semantic contentfulness, conversational appropriateness and other intralinguistic issues. Both those performances and the
larger pattern to which they belong are then also open to assessment for their contribution to
their more encompassing way of life: here arise issues of justification and truth, among others.
This two-dimensional normativity arises because of the interdependence of so much of what
we do with the performances of others and the circumstances in which they are situated. This
interdependence has itself evolved through thousands of generations of ratcheted niche construction, and adaptation to and development within those re-constructed environments.8
Speaking a language, producing and selling goods, teaching or taking a class, performing an
experiment, attending a meeting, having sex, riding the subway, and so much else we do are
performances that depend upon what others do; they would not “make sense” and could not be
the same kind of action apart from belonging to a larger pattern of practice. All too often,
however, others do not fully cooperate, or the circumstances are recalcitrant. Facing a misalignment of performances with their social and material circumstances, we adjust what we do:
acting differently, remonstrating with others, persisting in the face of incongruity, rearranging
the setting, or giving up. Often we try to say what has gone wrong or where adjustments seem
called for, but these discursive performances are themselves part of the pattern of practice at
issue, as well as part of a partially autonomous linguistic practice. Others respond to the same
mis-alignments, although they may understand or describe them differently. Issues arise
wherever performances or circumstances are misaligned in ways that seem to require adjustments. What is at stake in how those issues are resolved is both whether and how that pattern of
practice continues, and trade-offs are always possible between them.
“Issues” and “stakes” are anaphoric expressions for the normative accountability of what
organisms do to the larger environmentally intra-active way of life to which they belong. Issues
arise within an ongoing process or practice wherever things don’t go smoothly or well with
respect to their constitutive goals. For most organisms, those issues are threats to the continuation of life or lineage. Organisms and populations respond to those issues in various ways,
but what is at stake in those responses is one-dimensional: life or death, reproduction or extinction. The niche constructive emergence of multiple sustained, partially autonomous patterns of practice in the hominin lineage enables a more complex accountability. It matters to us
not only whether a practice continues, but also how it continues, in ways that encompass what
kinds of life we can thereby lead, and who we thereby become. We have evolved a rich
normative language to articulate what is at issue and at stake in how our patterns of practice
continue: virtues, goods, reasons, rights, truth, duties, norms as well as much “thicker” normative vocabularies more closely connected to the practices in which they arise. That normative articulation coevolves with the practices themselves, as integral to a two-dimensionally
normative organismic way of life. Practices and their subpatterns can go extinct; like the organismic way of life to which they belong, they only exist in continuing to be reproduced, and
they die out if they fail to adjust to changing circumstances. In the case of practices, however,
those circumstances are the other situated practices that currently articulate human ways of life.
Scientific practices, and the normative issues that animate them, are themselves forms of
developmental-evolutionary niche construction. The novel phenomena produced in laboratories and other scientific sites, their discursive articulation in diagrams, models, equations and
other theoretical conceptualisations, the disciplined and institutional ways of life in which those
manipulative and articulative capacities are developed and sustained, and the issues and stakes
with respect to which their performances and achievements are assessed, are complex reconstructions of the environments in which we now live and develop our capacities. These
practices and capacities are also part of a larger pattern of niche constructive transformation on a
183

Joseph Rouse

global scale. Scientific practices help shape the terms in which we live, and in which we assess
and revise how we live. They also consequentially transform our material surroundings, the
ways we make and sustain our living, and the issues we confront in maintaining that way of life.
The languages in which we describe and understand the world are not abstract propositional
contents that represent independent objects as their truth-makers. These terms and the claims
and performances they enable only acquire their sense and their use as part of a transformative
material practice that belongs to a biologically and historically specific way of life. They do not
represent the world as a more or less complete “picture” or image of it.9 They are instead
integral components of that world, contributing to the discursive articulation of our environmentally intra-active, two-dimensionally normative, organismic way of life.
This more radical naturalism thereby fulfills the indispensable desideratum for any viable
philosophical naturalism, which is to show how our capacities for scientific understanding are
themselves a scientifically intelligible natural phenomenon. Conceptual capacities and their
exercise in practice, including scientific practice, are a biologically intelligible natural phenomenon, whose two-dimensional normativity is an evolutionary novelty in the hominin
lineage. These capacities are materially realised in an ongoing, niche constructive transformation of our developmental and selective environments. Understanding scientific and other
conceptual capacities as a form of developmental-evolutionary niche construction in turn exemplifies this naturalistic reconception of our conceptual capacities and their normative accountability. This reconception is not a revised scientific image of human life from a “god’s-eye
view” or from “sideways on” (McDowell 1994), but is instead an open-ended reconfiguration
of the conceptual space within which we understand ourselves and our capacities. I do not
advocate a niche constructive understanding of our conceptual capacities as a sketchy “first
draft” of a final scientific image of humanity-in-the-world, but as part of an ongoing reconstruction of our conceptually articulated way of life that is constitutively open to further
transformative development.

5 Liberal and radical conceptions of normativity as “natural”
The previous section sketched a more radically naturalistic re-conception of scientific understanding and its two-dimensional normative accountability within massively nichereconstructed human ways of life. We can now attend to some of the most significant differences between this reconception of normative accountability, and some liberal naturalist
conceptions. As always, such generalisations are complicated by the diversity of liberal naturalist
views, but I find three important themes with respect to which liberal naturalists often diverge
from my more radically naturalistic approach. These concern the scientific intelligibility of
various forms of normativity, the supposed autonomy of a broad range of normative concerns
from our biological ways of life, and the temporally extended externalism of our normative
accountability.
A more radical naturalism understands normative assessment and the discursive practices
within which it occurs as a scientifically intelligible natural phenomenon. Discursive practices
and the adjudication of the normative issues that arise within them are appropriately understood
as integral to the developmental evolution of human ways of life, and the issues that arise within
them. Many liberal naturalists propose to expand the concept of nature in ways that extend
beyond the domains of natural scientific understanding, so that our acculturated second nature,
the “social world”, or the Sellarsian space of reasons is thoroughly natural, but outside the
domain of the natural sciences. I argue for a more thoroughgoing fusion of the scientific and
manifest “images” of humanity-in-the-world: the developmental and selective environments of
184

Liberal or radical naturalism?

human organisms are coextensive with the space of reasons as intra-active practices of normative assessment. These practices are thoroughly material, situated engagements with the
world, by living agents whose way of life is at issue and at stake in its ongoing evolution. To be
sure, the various scientific practices have their own partially autonomous conceptual domains,
whose homonomic articulation answers to their own characteristic ranges of counterfactual
invariance and distinct considerations of normative assessment. These practices are never fully
self-contained, however, but also answer heteronomically to the concerns that arise in other
aspects of human life.
A second theme concerns how the philosophical import of this rejection of any principled
distinction between a scientific domain of causes or laws and a broader conception of second
nature or culture, or between the scientific image and the space of reasons, is manifest on both
sides of these alleged divides. From the side of the natural sciences, there are at least three forms
of heteronomic accountability that are both integral to scientific understanding and inseparable
from the broader social and political entanglements of scientific practices. First, scientific understanding, embedded in research practices directed toward the further extension and refinement of their conceptual domains, incorporates judgments of scientific significance. Such
judgments guide not only the direction of subsequent research and conceptual refinement but
also the determination of which issues can be or should be relegated as tolerable noise or
conceptual discontinuity,10 ceteris paribus considerations, or irrelevant to the scientific domain.
These judgments also draw upon and respond to a broad range of cultural and political concerns. Second, the normative concerns that more specifically govern scientific methods and
hypothesis acceptance cannot be confined to narrowly “epistemic” concerns, because their
fallibility renders them sensitive to the epistemic and phronetic risks thereby undertaken (Biddle
and Kukla, 2017). Scientific choices at every stage of research practice have broader consequences (conceptually, technologically, environmentally, diagnostically and politically,
among others), those choices ought to take into account the possible consequences of error in
these respects, and the inevitable tradeoffs between risks of “false positive” and “false negative”
outcomes of those choices means that those broader normative concerns cannot be circumvented. Third, the methods and outcomes of experimental, observational, clinical or fieldinvestigative practices and forms of theoretical modeling cannot be disentangled from the niche
constructive issues raised by building and sustaining scientific practices over time and space:
communicative considerations of language (e.g. Gordin 2015), format, or credibility; institutional considerations of discipline-formation, material provision and financial support; and
pedagogical considerations of how best to recruit and train new practitioners, including
maintaining an appropriate balance between tradition and innovation, among others.
The inseparability of first and second nature or biology and culture is also evident from the
side of culture or second nature, whose supposed autonomy from natural scientistic imperialism
has been a primary concern of many liberal naturalists. As rational agents, language speakers, or
acculturated persons, humans are also organisms whose cognitive and affective capacities and
propensities are evolved and materially constrained, in ways that raise issues for both humanscientific understanding and the shapes of how we can live together. The modern evolutionary
synthesis once seemed to ratify a convenient separation of a human anatomy, physiology and
cognitive architecture established genetically in our early speciation from the forms and processes of subsequent historical and cultural transformation that take place against that background (Ingold 1995). Ecological-developmental biology and the evolutionary significance of
niche construction undermine any such separation, and encourage more integrated conceptions
of what Donna Haraway (2003) influentially dubbed “natureculture”.

185

Joseph Rouse

Such integration is reinforced by any serious reflection on a wide range of critical cultural,
political, or ethical issues that resist any compartmentalisation of their natural-scientific and
sociopolitical dimensions: the entangled issues of climate change, energy provision, agricultural
production/distribution and resource depletion; changing forms of disease, health, aging, reproduction/kinship and health-care provision; the effects of nuclear, chemical or biological and
other forms of weaponry on warfare, security, diplomacy and human prospects; environmental
degradation and loss of biological diversity; the effects of automation and computation upon
work, communication, privacy and the forms of sociality and political power; and a host of
other niche constructive transformations of human biopolitics. Liberal naturalists’ otherwise
salutary efforts to battle creeping scientism may be focused on the wrong struggles on the wrong
front.11
The third and final theme that partially differentiates more radical from liberal conceptions
of normativity as naturalistic concerns the temporality of normative determinations. My view
joins liberal naturalists in recognising that normative discourse and the practices that it articulates are truth-conducive. The biological environments we inhabit are now replete with a
wide range of normative standings open to reasoned critical assessment and possible endorsement. The discursive practices that articulate and sustain these normative relationships nevertheless extend beyond declarative utterances that answer to already extant normative statuses.
Normative discourse includes prescriptive and imperative utterances and other performances
through which we hold one another accountable, and the vocative and recognitive aspects of
those practices through which those holdings are directed toward, and acknowledged and taken
up by organisms as responsive and responsible agents.12 Crucially, however, these normative
determinations are constituted and sustained by the practices of which they are a part, and those
practices are temporally extended. The niche constructive character of normative discourse and
practice thereby commits my more radical naturalism to a temporal externalism about the
content, authority and force of normative determinations.
Temporal externalism (Rouse 2015) is a feature of normative determinations themselves as
well as their discursive articulation.13 On the one hand, we can have obligations and rights,
exhibit virtues or vices, provide justifications for what we say or do, use words meaningfully,
and acquire other normative standings, because our lives are already caught up in a variety of
practices, in which what we can do and how we can live depends upon the patterns of discursively articulated practice that make up our developmental and selective environments. On
the other hand, these normative relationships only exist through their ongoing uptake and
reproduction in practice. Their content, authority and force depend upon whether and how the
practices to which they belong continue over time, including their future development.
Whether we are obligated, virtuous, justified or intelligible depends upon the place of what we
say and do within a temporally extended and futurally open pattern of situated performances.
The result is that normative discourse is anaphoric in character. Efforts to ascribe normative
significance are mutually accountable to what is at issue and at stake in the temporally extended
pattern of practice to which they belong. Agents and their circumstances can be answerable to
the same issues and stakes, even if they have no neutral or shared formulation of those issues.
The question of how, truthfully, to characterise what is at issue and at stake in some ongoing
pattern of practice is itself part of what is at issue.
This conception of normative discourse and practice as anaphoric and temporally externalist
is not inconsistent with the core tenets of liberal naturalism, and some liberal naturalists hold
related views of the ways normative standing and its articulation are embedded in discursive and
other practices. My view is nevertheless more radically naturalistic than most liberal naturalisms
in that I present the niche constructive, anaphoric and temporally extended character of
186

Liberal or radical naturalism?

normative standings and their discursive articulation to be obligatory for naturalists. As organisms that have evolved a two-dimensionally normative, discursively articulated way of life,
what we say and do is accountable both to whether and to how our species and our ways of life
reproduce themselves. Recognition of and appropriate response to the increasing tension between these two dimensions of our biological normativity, between how we live and whether
those ways of life are viable, is one of the most compelling reasons for a more radical naturalism.

Notes
1 John Dupre (1993) is the contemporary philosopher of science most commonly cited by recent
proponents of liberal naturalism, but Dupre also explicitly ventures into the broader naturalist terrain,
and liberal naturalists are more inclined to cite his accounts of scientific disunity than his underlying
metaphysical pluralism or his detailed studies of evolutionary biology or microbiology.
2 I use the term “experiment” to stand in for a wider range of material practices. Not all scientific work
is experimental, but field observations, clinical research, computer simulations, other modeling
practices, and the like share with experimentation the need to change aspects of the world to allow for
its intelligibility.
3 For exemplary discussions of how conceptual articulation is entangled with discipline-building, see
Bechtel (1993), Lenoir (1997, chap. 2) and Ankeny and Leonelli’s (2016) conception of scientific
repertoires.
4 One consequence is that scientific domains have their own distinctive forms of counterfactual invariance (physical, chemical, functional-biological, evolutionary, medical, etc., necessity), with autonomous counterfactual ranges and constitutive standards of precision, accuracy, relevance and ceteris
paribus variability.
5 Hacking (1992) defends the supposedly “self-vindication” of mutually tailored theories and experimental practices that thereby become nearly irrefutable, without recognising that he thereby, as
McDowell concluded a parallel criticism of Davidson, “manages to be comfortable with his coherentism…because he does not see that emptiness is the threat” (1994, 68).
6 The heteronomic inferential significance of scientific practices and concepts has been extensively
explored in work by anthropologists, feminist scholars and cultural historians of science. Rouse (1993)
provided an early explication of this work, and some of its differences from the social studies of science
with which many philosophers are more familiar.
7 This understanding is complicated, but not fundamentally revised, by the recognition that all eukaryotic organisms are multigenomic holobionts rather than genetically specifiable individuals
(Bordenstein and Theis 2015).
8 Tomasello et al (1993) describe as “the ratchet effect”, the ways in which practices and skills are stably
reproduced in ways that allow their subsequent elaboration and refinement.
9 The point of this claim is not to deny that some linguistic expressions in the sciences have a representational role, but only to insist that their diverse representational uses are accountable to a wider
range of normative concerns arising from their places in various practices, and that their representational uses do not fit together into a unified representation of the world.
10 Mark Wilson (2006) provides many instructive examples of such discontinuities in mechanics and
materials science that are not taken to threaten the conceptual unity of the domain: “property dragging”, “lifts”, or conceptual gaps between “theory facades”.
11 Akeel Bilgrami’s (2010) contribution to an earlier volume on liberal naturalism provides an instructive
sketch of a genealogy of the cultural, political and theological stakes in the historical emergence of
current conceptions of scientific understanding, rationality, and the place of human ways of life in
nature. In that context, my account of a more “radical” naturecultural naturalism has some continuity
with what Bilgrami describes as the “Radical Enlightenment”. Bilgrami, however, seems to take it as
now settled which conceptions of nature and of scientific rationality emerged from that prehistory of
the debates over naturalism, and his objections to those conceptions lead him to repudiate naturalism. I
think that conceptual developments in the sciences still provide much richer and indispensable conceptual resources for political criticism. In this respect, my radical naturalism is joined by many advocates of a feminist materialism (Barad 2007; Alaimo and Hekman 2008; Pitts-Taylor 2016) and other
work that I have previously construed as “cultural studies of science” (Rouse 1993).

187

Joseph Rouse
12 Kukla and Lance (2009) discuss the wide ranges of ways in which discursive practices can introduce,
sustain, assess and uphold or reject normative statuses, through the pragmatic significance of prescriptive, imperative, vocative and recognitive performances.
13 The phrase “temporal externalism” was introduced as an extension of the semantic externalism advocated by Putnam (1975) and Burge (1979). Semantic temporal externalism is the claim that the
semantic significance of the uses of linguistic expressions depends not only upon their material and
social environment, but also upon the temporally extended pattern of past and future uses of those
expressions. Normative temporal externalism extends this line of argument beyond the semantic accountability of uses of words to understand the multiple forms of normative accountability of what we
say and do, including ethical and political as well as semantic and epistemic normativity.

References
Alaimo, S. and Hekman, S. (eds.). Material Feminisms. Indiana University Press. (2008).
Ankeny, R. and Leonelli, S. Repertoires: A Post-Kuhnian Perspective on Scientific Change and
Collaborative Research. Studies in History and Philosophy of Science, 60, (2016). 18–28.
Barad, K. Meeting the Universe Halfway. Duke University Press. (2007).
Bechtel, W. Integrating Disciplines by Creating New Disciplines: The Case of Cell Biology. Biology and
Philosophy, 8, (1993). 277–279.
Biddle, J. and Kukla, R. The Geography of Epistemic Risk. In K. Elliott and T. Richards, (eds.), Exploring
Inductive Risk, pp. 215–237. Oxford University Press. (2017).
Bilgrami, A. The Wider Significance of Naturalism: A Genealogical Essay. In M. De Caro and D.
Macarthur (eds.), Naturalism and Normativity, pp. 23–54. Columbia University Press. (2010).
Bordenstein, S. and Theis, K. Host Biology in Light of the Microbiome: Ten Principles of Holobionts and
Hologenomes. PLOS Biology. (2015). doi: 10.1371/journal.pbio.1002226.
Brandom, R. Making It Explicit. Harvard University Press. (1994).
Davidson, D. Essays on Action and Events. Oxford University Press. (1980).
Davidson, D. Inquiries into Truth and Interpretation. Oxford University Press. (1984).
De Caro, M., and Macarthur, D. (eds.). Naturalism in Question. Harvard University Press. (2004).
De Caro, M., and Macarthur, D. (eds.). Naturalism and Normativity. Columbia University Press. (2010).
Dupre, J. The Disorder of Things. Harvard University Press. (1993).
Gordin, M. Scientific Babel. University of Chicago Press. (2015).
Hacking, I. The Self-Vindication of the Laboratory Sciences. In A. Pickering (ed.), Science as Culture and
Practice, pp. 29–64. University of Chicago Press. (1992).
Haraway, D. The Companion Species Manifesto. Prickly Paradigm Press. (2003).
Haugeland, J. Having Thought. Harvard University Press. (1998).
Haugeland, J. Dasein Disclosed. Harvard University Press. (2013).
Ingold, T. ‘People Like Us’: The Concept of the Anatomically Modern Human. Cultural Dynamics, 7,
(1995). 187–214.
Lange, M. Natural Laws in Scientific Practice. Oxford University Press. (2000).
Lange, M. Laws and Theories. In S. Sarkar and A. Plutynska (eds.), A Companion to the Philosophy of
Biology, pp. 489–505. Blackwell. (2007).
Lenoir, T. Instituting Science. Stanford University Press. (1997).
Lewontin, R. The Triple Helix. Harvard University Press. (2000).
Macarthur, D. Taking the Human Sciences Seriously. In M. De Caro and D. Macarthur (eds.), Naturalism
and Normativity. Columbia University Press, pp. 123–141. (2010).
McDowell, J. Mind and World. Harvard University Press. (1994).
Okrent, M. Rational Animals. Ohio University Press. (2007).
Okrent, M. Nature and Normativity. Routledge. (2017).
Pigliucci, M. and Müller, G. Evolution – The Extended Synthesis. MIT Press. (2010).
Pitts-Taylor, V. (eds.). Mattering. NYU Press. (2016).
Price, H. Naturalism without Representationalism. In M. De Caro and D. Macarthur (eds.), Naturalism in
Question. Harvard University Press, pp. 71–88. (2004).
Price, H. Naturalism Without Mirrors. Oxford University Press. (2011).
Rouse, J. What Are Cultural Studies of Scientific Knowledge? Configurations, 1(1), (1993). 1–22.
Rouse, J. How Scientific Practices Matter. University of Chicago Press. (2002).

188

Liberal or radical naturalism?
Rouse, J. Temporal Externalism and the Normativity of Discursive Practice. Journal of the Philosophy of
History, 8, (2014). 20–38.
Rouse, J. Articulating the World. University of Chicago Press. (2015).
Rouse, J. Epistemic Risk, Scientific Significance, and Conceptual Normativity. Unpublished talk to the
7th Biennial Meeting of the Society for the Philosophy of Science in Practice, Ghent, Belgium, 30
June–2 July. (2018).
Teller, P. Twilight of the Perfect Model Model. Erkenntnis, 55, (2001). 393–415.
Tomasello, M., Kruger, A., and Ratner, H. Cultural Learning. Behavioral and Brain Sciences, 16, (1993).
495–552.
Wilson, M. Wandering Significance. Oxford University Press. (2006).

189

16
NATURALISM AS A STANCE
Jack Ritchie

Science should be used not mentioned.
Mark Wilson

1 Scientific naturalism: a troubled doctrine?
My aim in this paper is to try to get some handle on the idea of “liberal naturalism”. In
particular, in what is supposed liberality consists. I’ll begin with some criticisms directed at the
way Macarthur and De Caro introduce their concept of liberal naturalism and then offer my
own preferred understanding of naturalism and how it should be put to work in philosophy. I’ll
end with some brief reflections on what it might mean to be a more or less liberal naturalist, if
my understanding of the term is accepted.
Liberal naturalism is defined by its authors in opposition to another (and what they take to be
a bad) form of naturalism, what they call scientific naturalism.1 According to De Caro and
Macarthur, the scientific naturalist holds one or both of two doctrines:
(OD) Ontological Doctrine
The world consists of nothing but the entities to which successful scientific explanation
commits us
(MD)2 Methodological Doctrine
Scientific inquiry is our only legitimate form of inquiry (De Caro and Macarthur 2010, 5).
Liberal naturalism, by contrast, is liberal in allowing that there may be methods other than
scientific ones and entities other than those posited in scientific explanation. It remains naturalistic in rejecting any supernatural methods or ontology.
Something like the ontological doctrine is held by naturalists who call themselves physicalists. According to a standard sort of view, these thinkers believe the basic ontology of the
world has been identified by physics and there is philosophical work to do in showing how
apparently nonphysical things like minds or numbers or morality fit into that physicalist ontology. Frank Jackson’s (1998) work is a very good example of this.

190

DOI: 10.4324/9781351209472-19

Naturalism as a stance

Explicit advocates of the second doctrine seem to me harder to find. But perhaps claims like
those of Quine that “philosophy is continuous with natural science” are supposed to hint at
such a view. Quine’s dictum seems to imply that legitimate methods in philosophy are or ought
to be the same as those used in science. If we add to this the thought that science and philosophy are the only legitimate forms of theoretical inquiry, something it might be plausible to
think Quine held, then we get a claim close to MD.
OD and MD are described by De Caro and Macarthur as doctrines and so I suppose are
meant to be the kind of thing that might be asserted, believed or denied. But stated in this very
bald way, they strike me as deeply problematic. One set of problems arises when we ask the
eminently reasonable question: why should we believe either doctrine? Obvious answers seem
to founder on incoherence. Where has the claim that scientific inquiry is the only legitimate
form of inquiry come from? Not, presumably, from scientific inquiry. The complete forms of
legitimate inquiry are not, as far as I am aware, a topic of any scientific research. Then, the
methodological doctrine must be legitimated by some nonscientific investigation. Well, in that
case, its purported justification must undermine its own content.
A similar, but less immediate problem arises with the ontological doctrine. If it has a justification, then again it must come from science, perhaps from reflection on the warranted outputs of
our current scientific theory. But while of course, scientists do offer justifications for the existence
of particular entities, for example the existence of the Higgs boson, there does not seem to be in
the sciences themselves the kind of completeness claim articulated in the Ontological Doctrine.
Indeed, if we look at current science, especially current physics, it wears its incompleteness on its
sleeve. We lack a complete understanding of quantum gravity, dark matter or energy and many
other aspects of the universe. As a result, we expect current physics and its ontology to be revised
in future ways, difficult to discern from our current epistemic position, to deal with these unsolved problems. So either the ontological doctrine says something which is false, the world
consists of nothing but the entities to which current successful scientific explanation commits, or it
is not really an ontological doctrine at all since it lacks any way to clearly articulate what the
ontology of a future completed physics will say about the world.
I suppose it might be claimed that a scientific naturalist could hold one or both of these
doctrines dogmatically and insist that they require no further justification. For example, the
ontological naturalist might list several entities posited by current science and insist that there is
this much and no more. Likewise, the methodological doctrine could be given content by a list
of approved methods and a similar dogmatic judgment that these methods and no others are
acceptable. But while of course such a position is imaginable, and not strictly incoherent, it
surely runs counter to the very spirit of naturalism. Naturalists, whatever else they are, are
impressed by science, its methods, products and progress. Science, at its best, is undogmatic and
open-ended. We have seen and continue to expect both the theories and the methods of
science to develop and even sometimes be discarded over time. Any philosophical view which
takes inspiration from science ought to take seriously this aspect of its history. Any philosophical
thesis which rests on dogmas about scientific ontology or methodology will violate this.3
These observations are bad news not only for scientific naturalism but liberal naturalism too.
If there is no clear formulation of OD or MD which is not obviously false or does not rest on
unjustified claims, then the force of the contrast with liberal naturalism seems lost. Is anyone
who rejects OD and MD (which I have argued should be everyone) and the supernatural, a
liberal naturalist? That would make liberal naturalism amount to no more than a denial of the
supernatural, hardly an interesting philosophical view. To make progress, we need a different
way to think of naturalism, only then might we have a position which is defensible and might
be properly contrasted with others. In the next section I try to do just that.
191

Jack Ritchie

2 Naturalism as a stance
My criticism of a dogmatic version of so-called scientific naturalism complained of the way
it would violate the spirit of naturalism. The spirit of a philosophical position is not,
whatever else it is, some straightforward thesis. As I suggested above, the spirit of naturalism
in part consists in something like an admiration for the sciences. I suggest, if one wants to
understand properly what naturalism is, then it is this spirit that one should focus on and not
a doctrine or dogma. One good reason to make this shift, of course, is to avoid the problems
noted above with OD and MD but another equally good reason is that this seems to be the
way many reflective naturalists both in the past and in the present have thought of their
position.
An interesting historical example of this can be found in many of the papers published in the
1944 collection Naturalism and the Human Spirit. This volume includes work by some figures
well-known and some less so to a modern audience, but whether famous or obscure all express
similar views of naturalism:
The naturalist is one who has respect for the conclusion[s] of natural science (John
Dewey, 2)
The least common denominator of all historic naturalism, therefore, is not so much a set
of specific doctrines as the method of scientific or rational empiricism. (Sidney Hook, 45)
The career of naturalism…is the history of the slow growth of an attitude rather than a
specific philosophical doctrine. (Harold A. Larrabee, 319)
[N]aturalists exhibit… a community of temper, of method, and even of general
outlook. (John Herman Randall, 455)
Methods, tempers, attitudes of respect are not doctrines. We see similar sentiments expressed by
current naturalists. For example, Penelope Maddy, who offers perhaps the most detailed and
interesting account of naturalism in contemporary philosophy, describes her naturalism, what
she calls Second Philosophy, as follows:
Second Philosophy as I understand it, isn’t a set of beliefs, a set of propositions to be
affirmed; it has no theory. (Maddy 2007, 1)
Ritchie (2008, 196) describes naturalism as an attitude, and Ladyman and Ross (2007, 64) in
their wide-ranging book on structuralism begin that work by describing their own naturalism
which they call, following Bas van Fraassen (2002), not a thesis but a stance.
According to van Fraassen, a stance is an “attitude, a commitment, a cluster of such –
possibly including some propositional attitudes such as belief…but cannot be simply equated
with having beliefs”. (47–48). The exact details of what a stance is has been the subject of some
recent interesting debate4 but I think van Fraassen’s brief description and the statements of
naturalists above give us enough to characterise some of the main elements of naturalistic stance.
Let’s begin with the most basicelement, already articulated:
1 Admiration for the sciences
All naturalism must share this. Something which is more distinctive of philosophical
naturalism is:

192

Naturalism as a stance

2 Antifoundationalism (a rejection first philosophy)
This is an explicit part, of course, of Quine’s naturalism and is well articulated in his favorite metaphor:
Neurath has likened science to a boat which, if we are to rebuild it, we must rebuild
plank by plank while staying afloat in it. (Quine 1960, 3)
Point 2 is motivated, in large part by the failure of foundationalist epistemologies. A failure
which reaches from the very ambitious views of Descartes and Kant to ground all scientific
knowledge to the more modest attempts of the logical empiricists like Reichenbach to vindicate
the straight-rule of induction.5 All of these attempts have failed either on their own terms or
contain presuppositions overturned by future scientific knowledge.
1 & 2, I take to form a kind of Ur-naturalism. It is difficult to see what it would mean to be a
naturalist, if either of these attitudes were rejected. A physicalist-naturalist like Jackson shares
these attitudes and indeed takes them (erroneously in my view) to motivate the idea that
physicalism is a naturalised form of metaphysics; metaphysical theorising begins here with a
picture of the world given to us by empirical investigation.
To these two attitudes, I would add the following, which is more distinctive of the kind of
naturalism offered by Maddy:
3

A conviction that reflection on the sciences can help solve or dissolve various philosophical
problems.6

This part of the naturalistic stance is in some ways a development of the rejection of first
philosophy. Once we have rejected the idea of sure philosophical foundations, and so are forced
to begin philosophising in the midst of our ongoing attempts to understand the world, philosophical problems, if they have any solution, must be answerable to and by other parts of
inquiry. This attitude is nicely displayed by Maddy’s second philosopher when she says:
[The] Second Philosopher is equally at home in anthropology, astronomy, botany,
chemistry, linguistics, neuroscience, physics, physiology, psychology, sociology…and
even mathematics once she realises how central it is to her ongoing effort to understand the world…[s]he is fully capable of addressing a wide range of questions we
would just as typically regard as “philosophical”. (Maddy, 2)
That we have formulated naturalism as a stance rather than a doctrine does not absolve us from the
responsibilities of justification. It is perfectly reasonable to ask why anyone ought to have these attitudes. I have offered some reasons for accepting 1 & 2 but 3, which I take to be the most distinctive
part of philosophical naturalism, at least of my stripe, must prove itself in action. It must be shown that
philosophers who adopt this attitude really do make progress with some philosophical problems.
It is my aim below to give a brief illustration of this kind of naturalism at work. I will show
how from the naturalistic stance we may criticise the methods and thereby the claims of some
contemporary analytic metaphysics.

3 A stance in action: criticism of analytic metaphysics
How are we to pursue metaphysical knowledge? Contemporary metaphysicians, perhaps in a
naturalistic spirit, have a common answer. Ted Sider, John Hawthorne and Dean Zimmerman,
put it nicely when they say:
193

Jack Ritchie

[M]etaphysicians use standards for choosing theories that are like the standards used by
scientists (simplicity, comprehensiveness, elegance, and so on)7 (2008, 8)
The thought here is that science proceeds using a broadly hypothetic-deductive method.
However since many theories will be compatible with the data, we need extraempirical criteria,
sometimes called theoretical virtues, like those listed above to choose the correct theory.8
Metaphysicians then claim what is good for science, is good for philosophy too. So those same
criteria can be used to select among rival metaphysical theories.
Here then is an interesting philosophical question. Is this a viable method of theory selection? One obvious immediate objection that arises is that it is unclear why any of these criteria
listed above have anything to do with truth. Why, for example, should the fact that one theory
is more elegant than another give us any reason to think it is more likely to be true?
In the following I’m going to concentrate on just one of these so called virtues – simplicity –
and consider to what extent the claims of the metaphysicians are justified. First I’ll illustrate the
role simplicity judgments seem to play in some metaphysical disputes and then I will turn to
consider what role, if any, they play in science.
It’s not difficult to find (or at least reconstruct) cases in metaphysics where the comparative
simplicity of one view against a rival is offered as reason to believe that theory. For example, in
the dispute over the nature of properties, objects and states of affairs, trope theorists will claim
their theory is superior to realism about universals since in order to explain both what an
individual object is and what a property is trope theorists need only invoke one kind of entity –
an abstract particular – whereas realists will require the comparatively more extravagant theory
positing both an individual and a universal to explain the same facts (Maurin 2018).
Similar sorts of arguments can be seen in the debate between nihilists and realists about ordinary objects. Most of us believe that objects like chairs, peanut butter jars and other middlesized dry goods exist. We are realists about the everyday stuff. Nihilists claim there are no such
things. Only atoms arranged in various ways. Nihilists typically claim their view is simpler since it
demands only the existence of atoms, not atoms and tables, say (Brenner 2015). Since it is simpler,
it is better and so I suppose, if we believe the methodology stated above, closer to the truth.
We even find parsimony arguments enlisted to support doctrines which look at first glance
extravagant. David Lewis (1979) claims his modal realism is more parsimonious than views
which take possibilities to be basic or possible worlds abstract since although he postulates a
concrete possible world corresponding to every genuine possibility, these worlds are all one
type of thing – concrete worlds. Ersatzists must believe in both the concrete actual world and
the merely abstract or otherwise not fully concrete possible worlds.
These cases illustrate one very serious problem with simplicity arguments which I won’t
discuss here, namely judgments of relativity simplicity are not straightforward. A trope theorist
might claim their theory is simpler since one type of object, a trope, is required to explain
qualitative identity among distinct individualsbut a realist might reasonably reply that their theory
is simpler since only one object is required to explain the qualitative identity, the universal,
whereas the trope theory requires two distinct individuals to exist, two exactly resembling tropes.
Similar sorts of worries are easy to imagine with Lewis’s parsimony argument for modal realism.
Let us set this worry aside and focus now on whether simplicity plays a role in scientific
theorising, what kind of role it plays, and what connections there are between simplicity and
truth. Once we have answers to these questions we can finally turn to think about whether
simplicity might play the same role in metaphysical theorising that it plays in scientific thinking.

194

Naturalism as a stance

3.1 Simplicity in science
When we turn to the writings of the great and dead (physicists, anyway) we seem to find
attitudes and methodological pronouncements in line with those of the metaphysicians:
Consider for example one of Newton’s famous rules of reasoning:
Rule I We are to admit no more causes of natural things than such as are both true and
sufficient to explain their appearances. To this purpose the philosophers say that Nature
does nothing in vain, and more is in vain when less will serve; for Nature is pleased with
simplicity, and affects not the pomp of superfluous causes. (Newton 1726)
More vaguely but perhaps also more poetically consider Einstein’s claim:
Our experience hitherto justifies us in believing that nature is the realisation of the
simplest conceivable mathematical ideas. I am convinced that we can discover by
means of purely mathematical constructions the concepts and the laws connecting
them with each other, which furnish the key to the understanding of natural phenomena. (Einstein 1933, quoted in Norton ms)
Even so, our basic question of what the connection between simplicity and truth is remains and
so we must move beyond appeals to authority to look at how simplicity judgments are used in
science and how they can be justified in giving us some insight into the workings of the world.
Before we begin to consider, how, if at all, simplicity might be a guide to truth, it is worth
clearing the field of other less controversial senses in which simplicity is playing some kind of
role in science. No one will deny that sometimes there are pragmatic justifications for simplicity. Simpler theories are often easier to understand or easier to use in making predictions.
For example, climate scientists may prefer to work with a simpler model of the Earth’s global
climate because it allows a manageable number of runs in a finite time on a supercomputer or
because it is easier to keep track of the major factors responsible for the dominant outcome
trends. Rocket scientists prefer (in most cases) to work with Newtonian models since these
make calculations more manageable, even though, of course, they must believe Newtonian
mechanics is strictly false. Whatever role simplicity plays here, it has nothing to do with truth.
Another role simplicity has been claimed to play, which has some connection with truth, but
it is not relevant for our discussion, is as an efficient heuristic. For example, Popper (1959)9
counsels that we ought to try first to falsify the simplest hypothesis before moving to more
complicated versions. In Popper’s writing this is connected to his idea of degree of falsifiability
and can be illustrated by a simple curve fitting process. A linear function is easier to falsify than a
quadratic since just three data point are necessary (ignoring error) in order to falsify the hypothesis whereas we need at least four for a quadratic. As the curve gets more complex, yet
more data is needed. A similar idea worked out with greater sophistication can be found in
Kevin Kelly’s (2007) work. He shows given certain assumptions that the most efficient method
of arriving at the truth is to start with the simplest hypothesis.
Here it is the empirical evidence which is the ultimate guide to truth. Simplicity itself is not a
reason to favor one hypothesis over another; it is a method for efficiently organising your
testing. Clearly, this sort of heuristic is inapplicable to metaphysical theorising, where there is no
analogue to the kind of empirical testing procedure described here.

195

Jack Ritchie

Likelihoods and base rates
The two sorts of cases described above do not offer a direct connection between truth and
simplicity but there are other cases though where there does, prima facie, appear to be a direct
epistemic role for simplicity as argued by both metaphysicians and the great dead scientists. I am
going to try to describe these simplicity involving modes of inference in a roundabout way.
First I will begin with a toy example. I’ll try to explain in the example the intuitive reason we
have for thinking the simpler of two hypothesis is true and then I will offer an example where
the same pattern of reasoning is involved in a real scientific case.
Let’s begin with a nice case invented by John Norton (ms). Suppose we are walking along
the beach and we notice in the sand a neat row of bird footprints. What can explain this fact?
Consider two hypotheses:
H1: A two-legged bird hopped along the sand leaving behind these footprints
H2: Two one-legged birds hopped along the sand, each leaving one half of the observed footprints.
We would all agree that H1 is the better, that is to say more likely hypothesis and we might
gloss that by saying H1 is the simpler hypothesis.
What is going on here? As Norton notes it seems implausible to think we are making use of
general rule that X is simpler so X is more likely to be true as we can see when we consider a
similar sort of case. Imagine now we walk further down the beach and we see a zigg-zagging
mess of bird prints.
One hypothesis (call it H3) which could explain this observation is that flock of birds landed
in this area and then flew off. A second explanation (call it H4) is that a single bird walked and
flew on a bit and walked and flew and created the mess of prints by itself. H4 is in one sense
more parsimonious since only one cause, a single bird, is postulated but our intuition does not
seem to suggest to us this hypothesis is more likely.
Part of what is going on when we judge H1 is a better hypothesis than H2 is very easy to
explain. We know that there are many more two-legged birds than one-legged birds and this
fact by itself makes it more likely that the pattern is caused by a two-legged bird. In fact, almost
any hypothesis involving a two-legged bird is more probable by this rationale.
Certain judgments made in science conform to this sort of pattern of reasoning. For example, if a doctor is confronted with a patient who has flu like symptoms, one possible hypothesis might be the patient has flu, another is that she has the bubonic plague. (The early
symptoms of plague are very similar to flu.) It is more reasonable to believe the former, since the
flu is a much more common disease, fortunately, than bubonic plague.10
We can make some of what is going on here explicit in Bayesian terms. According to
Bayesians, we ought to update our belief in an hypothesis by Bayes’ rule:
P (H|e ) =

P (e|H ) P (H )
,
P (e )

where H is our hypothesis and e is the evidence.
If we are comparing two hypotheses judged against the same evidence only two things will
matter to our assessment of which is more probable. The prior probabilities, P (H ) and the
likelihoods, P (e|H ). In the case of both the examples of disease and birds discussed above, we
judge that the prior for the intuitively simpler hypothesis to be higher. But really here it is not
196

Naturalism as a stance

simplicity but our background knowledge of the frequency of one compared to two legged
birds or flu suffers to plague suffers that is doing all the epistemic work.
Norton’s bird case though seems to hang on more than just the priors. H1 would seem to be a
better hypothesis than H2 even if one-legged bird were as common as two and that is because at an
intuitive level the one-legged hypothesis does not account for the data as well. Two one-legged birds
might hop around the sand in any old fashion but one two-legged bird is constrained by it shape to
make a pattern of close together prints. In Bayesian terms that is just to say that P(e/H2legs)>P(e/H1leg).
Many cases where simplicity plays a role fit this likelihood pattern. Consider another toy example. All the lights in your building suddenly go out. One explanation is the simultaneous failure
of all bulbs. Another is that there has been a power cut. Intuition suggests the second explanation is
clearly better and we might also say clearly simpler, since only one common cause is appealed to
against many independent causes. Our intuition is well explained by appeal to relative likelihoods of
the two hypotheses. It obviously follows immediately form the fact a power cut occurs that all the
lights will go out. A common cause immediately explains the correlation. Separate causes, as offered
in the second explanation, are coincidental and improbable.11
In real science, we will find similar kinds of common cause explanations, which again might
be glossed in terms of simplicity. For example, in biology under certain assumptions, including
that the trait is not an adaptation, we may reasonably infer that a common trait between species
is evidence of a recent common ancestor.12 But again what explains why this is a good inference, is not a general rule about inferring to the simplest explanation but our background
knowledge of the relative likelihoods.
One way to emphasise this last point is to see that different assumptions can undermine the
idea that a correlation is evidence of a common cause. A nice example of Elliott Sober’s illustrates this well. If we live in a society in which primogeniture holds, then if two individual
males are wealthy that is evidence against their wealth having a common cause. So again, what’s
really important is not simplicity but background knowledge.

Curve fitting and model selection
We have uncovered two ways in which simplicity judgments can be seen to playing an epistemic role in science and we have explained that role in probabilistic terms. Simplicity judgments are a proxy for base rates or likelihoods. Curve fitting creates a third, and interestingly
different case, where simplicity seems to be playing some kind of role.
Curve fitting creates a special class of model selection problems in science. One problem we
encounter is the following: we collect data but many curves will fit the data, some more
precisely than others. How are we to choose the best? The simplest answer is that we choose or
ought to choose the curve that goes through all the data points. But in any real case it is likely
there will be error in our data. If our curve is too well-accommodated to the data it will be
overfitted; it will be responsive not just to the genuine facts (the signal) but in part the noise too.
Statisticians have developed various criteria to help determine the best curve, some of which
involve simplicity considerations. Probably the simplest is the so called Akaike Information
criterion (AIC). The AIC is meant to offer a way of ranking models in terms of their future
performance. More specifically it gives us a way to choose which is the best model to rely on if
we wish to make predictions about interpolating points in our data. Best here means that
choosing that model will in the long run produce the most accurate results.
AIC assesses competing models according to the following formula:
Ln(L) − 2k,
197

Jack Ritchie

where k is the number of parameters and L is the maximum value of the likelihood function.
The model with the highest AIC score is the best. As can be seen by inspection, there is a
penalty for complexity. The more complex the curve, the more that negatively affects its AIC
score.
AIC is only a good rule for model selection if certain assumptions are in place. One we have
mentioned already; the AIC does not give us grounds for extrapolating beyond the end points
of our data. But in addition to that we must also assume that our data come from the same
underlying reality, that repeated attempt to provide the best estimate of a parameter should have
a normal distribution and that one of the selected models is true (or very nearly so).
There are other model selection criteria, like the Bayesian Information Criterion, which also
make use of simplicity. The BIC has slightly different and stronger assumptions, most notably
those involved in setting the priors for the various models, than AIC; and perhaps not surprisingly offers a slightly different way of scoring models. In addition there are other model
scoring methods like the so-called cross one out method (CV) and the Takeuchi Information
Criteria (TIC), which do not involve parsimony considerations but can be shown under certain
assumptions to be equivalent to AIC.13
There are interesting questions to be posed about how these different criteria relate; and
perhaps under what exact circumstances which one is the best to use. Nevertheless, model
selection practice and theory does indeed seem to show that, given certain assumptions simplicity considerations play an important epistemic role.

3.2 Simplicity in metaphysical reasoning
I have identified three patterns of inference in which we might reaonsably argue simplicity is
playing a genuine role in scientific practice. First, simplicity is sometimes a proxy for base rates
or prior probabilities. Second, simplicity judgments are implicit appeals to likelihoods, as in
common cause reasoning and finally parsimony considerations have a genuine role to play
model selection.
These examples demonstrate, I think, that there is no simple inference from simplicity to
likely truth. In each of these three patterns of inference we must identify further background
facts which legitimate the inference that the simpler hypothesis is more likely to be true; and in
each case it is easy to see how further facts might undermine that inference.
If this is the way simplicity arguments work in science, could they work the same way in
metaphysics? I think it is straightforward and uncontroversial that the very interesting work of
statisticians on model selection can have no bearing on metaphysics. We cannot construct rival
metaphysical views as rival models of data. Let us then turn to the base rate and likelihood
justifications. Can these be made to do work in metaphysics? In an abstract or formal way, a
metaphysician might claim yes. If we are strictly subjective Bayesians, we ought to interpret the
probabilities as credences, subjective degrees of belief. It might seem then for the metaphysician
simplicity arguments are playing a role in fixing our priors. So the thought, say, that trope
theory is simpler could be reconstructed as claim that it has a higher prior or indeed given the
data that it is needed to explain, qualitative similarity or difference, it has a higher likelihood
than its rivals and so ought to have a higher posterior probability.
While this would be formally analogous to the way some simplicity arguments were construed above, it would be no more than that. In the case of arguments involving appeals to prior
probabilities, epistemically sound versions of those arguments are anchored in knowledge of
base rates. Similarly likelihood reasoning is persuasive only when we have reasonable grounds
for comparing the likelihood of two rival hypotheses. In many common cause cases this can be
198

Naturalism as a stance

done, as the examples above indicate. But how on earth are we to make any even approximate
judgment about the relative likelihood of, say possible worlds realism versus ersatzism? I see no
way this can be done and so at best the metaphysician will have a purely formal parallel to some
of the patterns of inference in science which make appeal to simplicity.
I conclude therefore that although in science there are forms of reasoning which make
appeal to simplicity, these kinds of argument cannot be applied in metaphysics since they either
involve assumptions which have no analogue in metaphysical debate (as in the model selection
case) or they lack anyway to objectively ground the relevant probabilistic judgments for which
the simplicity claims are proxies.14
My argument is modest. There are many responses a metaphysician might give. It could be
that there are other kinds of simplicity considerations in science that would have more than a
merely formal analogue to metaphysical reasoning. It could be that there is a special metaphysical way in which simplicity guides us to the truth; something which only truly applies in
metaphysical reasoning. I (and any naturalist) ought to agree to this possibility but the onus is on
the metaphysician to uncover or explain these other forms of argument from simplicity. It is
notable that there are no such concrete suggestions in the metaphysics literature. It could also be
claimed that the criticism of simplicity arguments offered above is irrelevant since metaphysicians misunderstand their own methodology. Maybe simplicity is not playing a role in theory
choice in metaphysics. Perhaps the other virtues, elegance, fruitfulness and so on are doing the
heavy epistemic lifting. In that case the metaphysician owes us an account of how either singly
or jointly these virtues are guides to truth. The naturalist will again recommend looking at the
role they play, if any, in science to help think through the matter. Finally, perhaps Zimmerman
and others completely mischaracterise the methods of metaphysics. Maybe something other
than theoretical virtues guides the metaphysician to truth. In that case the naturalist would like
to know what that is. The above argument can’t conclusively prove (at least all by itself) that
much metaphysics is methodologically bankrupt but it does put pressure on the metaphysician
to do better than make airy appeals to the methods of science or so-called theoretical virtues.

4 Liberal naturalism and the naturalistic stance
So this, as I see it, is the naturalistic stance and this is the naturalistic way to pursue philosophy.
Where does liberal naturalism fit into this picture of philosophy?
Liberal naturalism is often expressed explicitly as an ontological doctrine: the world contains
irreducibly, nonsupernatural entities not posited in any of the sciences (see De Caro and
Macarthur 2010, p. 10-12) . But if what I said in the opening section is correct that view of
naturalism relies on a contrast with a position that makes no real sense. I would suggest if there
is a viable philosophical position called liberal naturalism, it ought too to be a stance.
Although that does not appear to be the official line from De Caro and Macarthur certain
attempts to illuminate the character of liberal naturalism clearly suggest that it is something
other than a metaphysical doctrine. Consider for example De Caro’s more detailed articulation
of what Putnam’s liberal naturalism involves in a recent collection of essays. Among several
points the following are listed:
5 It is our epistemic duty to try to resolve the contradictions and conceptual tensions
between the commonsense and the scientific worldviews
11 …a notion that one should use with caution if one does not want to end up
with the reconciliation problem is that of causation. (Putnam and De Caro
2016, 9–10)
199

Jack Ritchie

Duties and counsels of caution are clearly not doctrines and certainly not metaphysical claims.
So I see in statements of liberal naturalism there is already movement towards the idea it is a
stance.
If we shift our thinking about liberal naturalism from an ontological doctrine to a stance,
how might it differ from my position? I think one uncontentious thing that can be said about
De Caro and Macarthur and many of the thinkers they wish to label liberal naturalists, in
comparison to me (and probably other naturalists like Maddy, Ladyman and Ross too) is that
they have a very different philosophical focus. In this paper and elsewhere I have been interested mostly in broad matters of epistemology and metaphysics. Liberal naturalists, whatever
else is true of them, have a more human-centered focus. They are concerned with matters
directly or indirectly connected to ethical and perhaps political norms. By itself, I do not take
such a focus to involve a different conception of naturalism. It would be perfectly coherent, and
in my view sensible, to approach philosophical problems more intimately connected to ethical
or political issues from the stance I described above. For example, it seems eminently reasonable
to think that aspects of social psychology, like work on implicit bias, will be highly relevant to
ethical questions in understanding human reasoning and our susceptibility to prejudice. I think
work done by experimental philosophers which suggests problems with appeals to intuition is
relevant to work in ethical theorising; and even the claims I have made here about simplicity, I
think have some relevance since we occasionally find ethicists making an appeal to simplicity as
a guide to truth in theory building.
Moreover, when we switch to talk of values or more broadly meta-ethical theorising, other
sciences will be relevant. As an illustration, consider De Caro’s summary of the principles of
Putnam’s liberal naturalism: “Contrary to a common opinion, not all objective knowledge
should be seen as object-based”. (9) I couldn’t agree more. Part of our attempt to understand
and investigate such matters ought to lead us to linguistics and other empirical studies of language in which these matters can be properly theorised and tested. Thomas Hofweber (2016),
for example, has done this for arithmetic. Drawing on work in linguistics, he argues that
number terms are not referential. Such studies can and have been extended to our ethical and
other normative speech.15 All of this is of a piece with the naturalistic stance as I understand it.
If liberal naturalists accepted this way of characterising the contrast with me, as one concerning matters of interest, rather than the fundamentals of the stance described above, then
perhaps we should understand the “liberal” in liberal naturalism like the “liberal” in liberal arts,
as indicating a broad range of philosophical interests. I find this conception of liberal naturalism
quite appealing (no doubt, because it means liberal naturalists fundamentally agree with me) but
I worry it is not quite right. More important, perhaps, than just a different focus is another
attitude. Liberal naturalists are suspicious of the claim that reflection on scientific theories or
methods will really help them address the questions they are most interested in.
There is a naturalist-friendly way to understand this suspicion as a concern about the reliability of the relevant sciences, especially psychology. For example, many key claims in social
psychology, including some related to implicit bias,16 have not been successfully reproduced.
Given that it would be foolish to place too much trust in such results and so equally foolish to
answer philosophical questions by appeal to such work. All naturalists would surely agree. We
follow the evidence where it leads.
It may be that liberal naturalism involves a more radical position, one in which there are in
principle reasons to think that the mixture of theoretical and empirical tools developed in the
sciences cannot be applied to addressing certain philosophical questions. I’m not sure what the
arguments for that more radical position would be; I have highlighted here some work which I
200

Naturalism as a stance

think addresses some of the standard examples. But like all good naturalists and all good liberals,
I have an open mind and I am willing to be persuaded.

Notes
1 This is a view both authors have attributed to me. As we shall see, I don’t recognise their description of
scientific naturalism as the naturalism I hold.
2 OD and MD are my inserts for convenience.
3 It should be obvious that my argument here draws heavily on van Fraassen’s (lecture 2, 2002) discussion of a dogmatic empiricism, as does my discussion of a stance below.
4 See for example the essays in Monton (2007) and the Special Issue edited by Rowbottom and
Bueno (2010).
5 One of De Caro and Macarthur’s favorite examples of a liberal naturalist is a very important figure
here. Putnam’s (1990) refutation of Reichenbach’s vindication of induction is an extremely important
and conclusive undermining of a first philosophical program.
6 Jackson’s naturalism diverges at this point. Physicalism poses a problem for the philosopher: how can
the mind or numbers or whatever fit into the natural world. But it is a distinctively philosophical job,
with distinctively philosophical tools, conceptual analysis, to answer this question.
7 See also Paul (2012), Sider (2007) for similar statements. Often such writers will talk in general about
theoretical virtues and inference to the best explanation. When pressed to explain what makes one
explanation best or what the theoretical virtues are, they will give a list similar to the one quoted
above. This work owes a debt to the Quine and Ullian (1975) conception of scientific method.
8 Putnam (2002) (someone who has accepted the label liberal naturalism) endorses this view of science
and uses it argue against the fact-value distinction. As will become clear, I think this is an erroneous
view of science and Putnam’s argument against the fact/value distinction does not work.
9 Popper also has a more ambitious theory of simplicity connecting it to degree of corroboration. It is
widely accepted this does not work. For a nice, clear recent discussion see Sober (2015, p. 97ff ).
10 Obviously this would not be a good inference in Medieval Europe.
11 This of course can be made more explicit making use of Reichenbach’s theory of the common cause.
The details of this (and its limitations) are well set out in Sober (2015).
12 See Sober (2015, Chap. 3) and Sober (2008, Chap. 4) for some of the complications in real arguments of
this kind in biology. Here I have focused on common cause explanations since the connection to
simplicity is obvious – one cause is simpler than many. A more nebulous notion of simplicity, sometimes
called elegance, can also be connected to relative likelihoods. For example, it is often claimed that
Copernicus’s theory was simpler than Ptolemy’s. One thing this can mean is that Copernicus’s theory
accounts for the data better. For example, it falls straight out of a heliocentric picture that Jupiter will
regress at certain times of the year. This require special contrivance in a geostationary view. In probabilistic terms P( Jupiter regression/Cop) = 1; P( Jupiter regression/Ptol) < 1. There are I think no other
senses of epistemically relevant simplicity considerations.
13 AIC is a special case of TIC; CV is asymptotically equivalent to AIC.
14 There is one other bad simplicity argument, which is so bad I have not bothered to put it in the main text.
It is a truth of the probability calculus that P(a) ≥ P(a & b). So it must be the case that P (atoms) > (P(atoms
and middle-sized goods). This probability claim is, of course, correct but it does not support a metaphysical position like nihilism over a commonsense ontology. Nihilism is the altogether bolder hypotheses
that only atoms exist. But it is not true that there is a simple probability argument that P(a &~b) ≥ P(a & b).
15 Maddy’s (2007) on logic is another good example.
16 See Hermanson (2017) for some criticism of the way the psychology literature has been used by
philosophers.

References
Brenner, A. Merelogical Nihilism and Theoretical Unification. Analytic Philosophy, 56(4), (2015).
318–337.
De Caro, M. and Macarthur, D. Naturalism and Normativity. Columbia University Press. (2010).
Einstein, A. On the Methods of Theoretical Physics. In Ideas and Opinions, pp. 270–276. Bonanza
([1933]/1954).

201

Jack Ritchie
Hermanson, S. Implicit Bias and Philosophy (Vols. 1 and 2), edited by M. Brownstein and J. Saul. Philosophy,
92(2), (2017). 315–332.
Hofweber, T. Ontology and the Ambitions of Metaphysics. Oxford University Press. (2016).
Jackson, F. From Metaphysics to Ethics: A Defence of Conceptual Analysis. Clarendon Press. (1998).
Kelly, K. A New Solution to the Puzzle of Simplicity. Philosophy of Science, 74(5), (2007). 561–573.
Krikorian, Y.H. Naturalism and the Human Spirit, 2nd edition. Columbia University Press. (1945).
Ladyman, J., Ross, D., Spurrett, D., and Collier, J.G. Every Thing Must Go: Metaphysics Naturalized.
Oxford University Press. (2007).
Lewis, D. Possible Worlds. In M.J. Loux (ed.), The Possible and the Actual: Readings in the Metaphysics of
Modality. Cornell University Press. (1979).
Maddy, P. Second Philosophy: A Naturalistic Method. Oxford University Press. (2007).
Maurin, A.S. Trope, The Stanford Encyclopedia of Philosophy. In Edward N. Zalta (ed.), Summer 2018
edition (2018). https://plato.stanford.edu/archives/sum2018/entries/tropes/.
Monton, B.J. Images of Empiricism: Essays on Science and Stances, With a Reply From Bas C. van Fraassen.
Oxford University Press. (2007).
Newton, I. Mathematical Principles of Natural Philosophy. 3rd edition, trans. Andrew Motte, rev. Florian
Cajori. University of California Press. [1726]/(1962).
Norton, J. (ms) The Material Theory of Induction. Available here: https://sites.pitt.edu/~jdnorton/
papers/material_theory/Material_Induction_March_14_2021.pdf
Paul, L. Metaphysics as Modeling: The Handmaiden’s Tale. Philosophical Studies: an International Journal for
Philosophy in the Analytic Tradition, 160(1), (2012). 1–29.
Popper, K.R. The Logic of Scientific Discovery. Basic Book, Inc. (1959).
Putnam, H. The Meaning of the Concept of Probability in Application to Finite Sequences. Routledge. (2011).
Putnam, H. and De Caro, M. Naturalism, Realism, and Normativity. Harvard University Press. (2016).
Quine, W.V. From a Logical Point of View: None Logico-Philosophical Essays, 2nd edition. Harvard University
Press. (1961).
Quine, W.V. and Ullian, J.S. The Web of Belief. Random House. (1970).
Ritchie, J. Understanding Naturalism. Acumen. (2008).
Rowbottom, D. and Bueno, O. Stance and Rationality: A Perspective. Synthese, 178(1), (2011). 1–5.
Sider, T., Hawthorne, J., and Zimmerman, D.W. Contemporary Debates in Metaphysics. Blackwell. (2008).
Sober, E. Evidence and Evolution: The Logic Behind the Science. Cambridge University Press. (2008).
Sober, E. Ockham’s Razors: A User’s Manual. Cambridge University Press. (2016).
Van Fraassen, B.C. The Empirical Stance. Yale University Press. (2002).
Wilson, M. Wandering Significance: An Essay on Conceptual Behavior. Oxford University Press. (2008).

202

PART III

Challenges for liberal naturalism

17
LIBERAL NATURALISM:
ORIGINS AND PROSPECTS
Mario De Caro

1 The placement problem
Saying that contemporary science is deeply at odds with the ordinary view of the world, and
that the most spectacular instance of this fracture is offered by physics, is little more than stating
a truism. As Gregg Rosenberg (2004, 241) writes:
Science has already shown us in many ways – from the relativity, responsiveness and
surprising geometry of space and time to the randomness, indeterminacy, nonlocality
and uncertainty principle of quantum mechanics – that commonsense intuition breaks
down at the fundamental level of the world.
However, from the philosophical point of view, the counterintuitive nature of contemporary
science, and particularly of physics, has relevant consequences also regarding the ordinary view
of the world itself. If one makes the natural sciences – and particularly physics, with all its
oddities – the only judges of what is real and what is not (as many philosophers advocating
“scientific naturalism” tend to do today), the common features of the ordinary view of the
world – such as secondary properties, moral phenomena, freedom, self-consciousness, qualia
and intentionality – look ipso facto queer,1 if not absurd altogether. Consequently, there is an
array of emerging questions that philosophy must face. In particular, notwithstanding their
reputed heterogeneity with respect to scientific features, are these features fully reducible to the
physical components of nature? Or are they mere illusions? Or are they mysteries that humans
won’t ever be able to grasp?
Unsurprisingly, how to conceive of the features of the ordinary view – under the assumption
that our legitimate ontology is entirely dictated by natural science – is one of the most discussed
problems in contemporary philosophy. As every serious philosophical problem, it has also a
name, and actually two: the “placement problem” (Price 2004) and the “location problem”
(Jackson 1998). Here is how Huw Price presents it:
If all reality is ultimately natural reality, how are we to ‘place’ moral facts, mathematical facts, meaning facts, and so on? How are we to locate topics of these kinds
within a naturalistic framework, thus conceived? (Price 2004, 74).
DOI: 10.4324/9781351209472-21

205

Mario De Caro

Also John Searle (2007, 4–5) expresses the problem with great clarity:
How can we square a conception of ourselves as mindful, meaning-creating, free,
rational, etc., agents with a universe that consists entirely of mindless, meaningless,
unfree, nonrational, brute physical particles?
Right, how can we? As we will see, scientific naturalists have put forward several strategies for
solving the placement problem. Liberal naturalists, however, tend to be very skeptical about
these attempts and have proposed some promising ways of reconceptualising the entire question, as is evident from this very collection. However, in order to determine precisely the
nature of the placement problem and the best options to solve (or dissolve) it, it is helpful to first
look at its genealogy.

2 When science divorced from the ordinary view of the world
The fracture between science and common sense became particularly evident with the advent
of general relativity and quantum mechanics in the twentieth century. Its origins, however, can
be traced back to the second half of the sixteenth century, when the process of mathematisation
of physics began. Evidence of this is offered, for example, , by Jacopo Mazzoni – a professor of
philosophy at the University of Pisa and a colleague and friend of Galileo’s –, who in 1597
discussed the question of the application of mathematics to physics, by comparing the views of
two opposite philosophical schools of the time. On the one side stood the Platonic party (of
which Galileo would soon become the leader), according to which such application was legitimate, or even indispensable; on the other side, stood the Aristotelian party that defended the
opposite position.2 Mazzoni (1597, 188) wrote:
There is no other question which has given place to more noble and beautiful speculations…than the question whether the use of mathematics in physical science as an
instrument of proof and a middle term of demonstration, is opportune or not; in other
words, whether it brings us some profit, or on the contrary is dangerous and harmful…
It is well known that Plato believed that mathematics was quite particularly appropriate
for physical investigations, which was the reason why he himself had many times recourse to it for the explanation of physical mysteries. But Aristotle held a quite different
view and he explained the errors of Plato by his too great attachment to mathematics.
Thus, while the Platonists attributed epistemological primacy to mathematical physics (with its
required substantial level of abstraction), the Aristotelians claimed that knowledge could only be
based on perception. The two groups also defined their respective ontologies in light of these
epistemological views. For the Aristotelians, space was a hierarchical, value-laden, anisotropic,
anthropocentric and continuous structure, and conceiving terrestrial bodies as mathematical
entities would be “dangerous and harmful”. For the Platonists, space was Euclidean and isotropic, and atoms (whose existence implied the reality of vacuum, which was instead denied by
the Aristotelians) were the basic components of both celestial and terrestrial entities.
Consequently, the Aristotelians maintained the traditional rigid distinction between astronomy
(the mathematical science studying the motions of the celestial bodies) and physics (the nonmathematical science of terrestrial movements). Conversely, the Platonists refused that distinction, presenting mathematics (and more specifically Euclidean geometry) as the tool that
made it possible to interpret the natural world in its entirety.
206

Liberal naturalism: origins and prospects

The mathematisation of the physical space defended by the Platonists, and in particular by
Galileo, was arguably the most important premise of the scientific revolution. According to the
Aristotelians, quantitative (i.e. measurable) properties have metaphysical priority over qualitative ones (those that cannot be measured, such colors, sounds, smells and so on). According to
Galileo, on the contrary, only the intrinsically geometrical spatial properties of bodies are real,
while qualities are mere appearances generated by the perceiving apparatus of the perceiver. In
two famous passages of The Essayer, he writes:
[The real structure of the world] is written in this very big book [the book of nature],
that is always open under our eyes; but we cannot understand it, if we do not learn to
understand the language and recognize the letters with which this book is written. It is
written in a mathematical language, and the letters are triangles, circles and other
geometrical figures (Galilei 1623, 232).
In external bodies there are only size, figure, number and motion; if ears, tongues,
noses are taken away, what remains are only figures, numbers and motion. Smells,
tastes, and sounds outside living animals are instead pure names, just as a tickle is,
when armpits are taken away (Galilei 1623, 264).
Drawing on a tradition initiated by Democritus and continued by Descartes, Locke and
Newton among others, Galileo’s strategy gave rise to the mathematical turn in the history of
physical sciences. It did not take long, however, to understand that the mathematisation of the
physical world had created a much more general problem than that of dislodging the qualitative
properties from the natural world. To use P.F. Strawson’s (1985) terminology, the “subjective
view” (according to which the mental is irreducible to the physical and intentional actions are
categorically different from happenings) and the scientific view of the world (Strawons’s
“objective view”) started to be seen as constitutively heterogeneous.3 As is well known,
Descartes’s radical attempt to solve the problem by appealing to the mind-body ontological
dualism was very soon seen as unsatisfactory; thus, many other attempts were made to harmonise the relationship between the subjective and the objective views. In this light, some
philosophers (including Hobbes and other materialists) privileged the scientific view, others
(spiritualists, materialists and empiricists) privileged the subjective view, while Spinoza developed a neutral ontological monism according to which the ultimate reality is neither mental nor
physical. However, none of these proposals was convincing. As a result, the problem of the odd
relationship between the two views of the world kept lurking, until Kant gave its most
memorable rendition with the Third antinomy of the Critique of Pure Reason.

3 Kant and the fracture between the subjective and the objective views
Once the whole reality is identified with the subject matter of physics, the subjective view is
under severe threat. Kant saw that very clearly: in the physical (which he called “phenomenal”)
world there is no space for the most important components of our self-conception, such as
morality, mental causation, personal identity, agency, and, most importantly, freedom(in
which, according to Kant, all the other items are grounded). More specifically, in the Third
antinomy, Kant highlighted that freedom appears to be both real and impossible. Seeing it one
way or the other depends on the perspective one takes between the subjective and the objective
view.4 On the one hand, as Kant (1781/1787) argued, “a causality through freedom” is to be
conceived as the foundation of moral responsibility. In the antinomy this claim represents the
207

Mario De Caro

thesis: we cannot but believe in our freedom, that is, in an original spontaneity that escapes the
nomological network of natural causation. Freedom, in short, is a necessary condition of any
attribution of moral responsibility. On the other hand, physics – which for Kant means
Newtonian mechanics – teaches us that human beings, as physical bodies, can only obey the
ubiquitous and irrevocable deterministic laws of nature. This is the antithesis of the antinomy: if
free will existed, allowing human beings to self-determine, it would represent “a lawless faculty”, an intolerable rupture of the natural order producing “disorder and incoherence” (414).
From this perspective, therefore, “there is no freedom; everything in the world takes place
solely in accordance with laws of nature”. In sum, for Kant, the problem of free will is that
freedom appears to us as both real and impossible.
As is well known, Kant developed a “transcendental” solution of the antinomy, which in his
view proved the possibility (not the reality) of freedom as “free or unconditioned causality” –
that is, as the capacity of humans to act without being caused. In that light, however, agents are
not to be conceived as phenomenal subjects but as noumena: that is, as intelligible beings that are
not framed in space-time – and so are not touched by natural laws, but freely self-determined as
causa sui. Today, however, Kant’s attempted solution of the antinomy has very few followers
since it faces three formidable objections. First, Kant believes that, in order to account for the
“imputability” of action, freedom has to be conceived as “absolute spontaneity”, that is, as the
power to start new causal chains without any conditioning – and no naturalists, of any sort,
would accept this idea. Second, there is the problem of how Kant’s noumenal self, which by
definition would be timeless, could enter into a causal relationship. By definition, causal relationships presuppose a temporal asymmetry between the event-cause and the event-effect (or,
at most, the temporal coincidence of the two events in the problematic case of “instantaneous
causation”). Third, by accepting determinism, Kant concedes that, at the phenomenal level,
every human action has a cause (or, more correctly, a set of causes) that is its sufficient condition; and this, put in the language of contemporary metaphysics, means that every action is
necessitated by the spatiotemporal events that cause it. However, once granted that, at the
phenomenal level, all actions have sufficient causes, what additional causal role could the
noumenal self play in the determination of any specific action?5
For all these reasons, Kant’s attempted solution of the third antinomy is currently out of
favor. The situation with his antinomian reading of the problem of freedom – and more
generally of the entire subjective view – is very different. Today, in fact, several influential
authors decline the question in ways that closely recall the original Kantian formulation. In his
classic A View from Nowhere, for example, Thomas Nagel (1986) writes that, on the one hand,
freedom appears incompatible with the order of nature, but on the other hand it is indispensable
for us: it “is something we can’t get rid of, either in relation to ourselves or in relation to others.
We are apparently condemned to want something impossible”. (113). And, more recently,
Peter van Inwagen (2002, 169) has echoed Nagel’s words: “Free will seems…to be impossible.
But free will also seems to exist. The impossible therefore seems to exist”.
As said, however, the problem of freedom is only one expression (if, arguably, the most
relevant expression) of the more general problem of the fractured relationship between the
subjective and the objective view of the world. Let’s now briefly consider how scientific
naturalists approach this problem.

4 Scientific naturalism and the placement problem
Scientific naturalism is defined by three main claims:

208

Liberal naturalism: origins and prospects

1
2

Reality consists of nothing more than the entities to which the successful explanations of
the natural sciences (if not physics alone) commit us;
Scientific inquiry is our only genuine source of knowledge; all other alleged forms of
knowledge (e.g. ordinary perception, a priori knowledge, introspection) are either reducible in principle to scientific knowledge or illegitimate;
Philosophy must be continuous with science as to its contents, methods and purposes.

In this light, the natural sciences (if not physics alone) are our only genuine source of
knowledge; consequently, they have the final saying regarding every ontological issue; and this
means that the whole of nature coincides with the subject matter of natural sciences (or of
physics alone). Alex Rosenberg (2009) is one of the most intrepid advocates of this view:
What is the world really like? It’s fermions and bosons, and everything that can be
made up of them, and nothing that can’t be made up of them. All the facts about
fermions and bosons determine or “fix” all the other facts about reality and what exists
in this universe or any other if, as physics may end up showing, there are other ones.
In effect, scientism’s metaphysics is, to more than a first approximation, given by what
physics tell us about the universe. The reason we trust physics to be scientism’s
metaphysics is its track record of fantastically powerful explanation, prediction and
technological application. If what physics says about reality doesn’t go, that track
record would be a totally inexplicable mystery or coincidence.6
Given these assumptions, scientific naturalists are presented with three possible strategies for
dealing with the placement problem: reductionism, eliminativism and mysterianism. According
to the first strategy, that of reductionism, one must understand whether the controversial
phenomena postulated by the subjective view are ontologically genuine to the extent that they
are reducible to scientifically acceptable phenomena. A good example of this strategy is offered
by the attempts by some compatibilists and event-causal libertarians try to reduce free will to the
causation of decisions and actions by mental (that is, physical) states. Another classic example is
that of the type-identity theory in philosophy of mind, according to which all types of mental
events are identical to types of physical events (Kim 1992, 2007; Smart 2007; Gozzano and Hill
2012; Polger and Shapiro 2016). Alternatively, one can think of the flourishing “neuroaesthetics” movements, aiming to reduce aesthetic properties to neurological ones (Chatterjee
2013; Zeki et al. 2020). One could provide many more examples across different fields, but
these three should suffice to give an idea of the reductionist approach.
Not all scientific naturalists, however, are happy with the reductionist strategy for solving the
placement problem. Going back to our examples, today a number of philosophers convincingly
object that such reductions of free will are exposed to the so-called “disappearing agent” objection. The underlying idea is that, if agents are nothing over and above their physical and mental
components, there is no way that they control the actions they perform (Pereboom 2004, 2017).
Based on this objection, the third antinomy cannot be solved by simply reducing free will to
forms of event-based mental causation. As to the type-identity theory in philosophy of mind,
several serious objections have been presented. In particular, it has been persuasively claimed that
the new versions of this theory are not immune to the traditional functionalist objection (Endicott
2005, 2017). Finally, coming to neuroaesthetics, the attempts to reduce aesthetic properties to
neurological ones leave out what is most specific to them (Gallagher 2011; Hutto 2015).
Given such difficulties, many scientific naturalists opt for the eliminativist approach.
According to them, the recalcitrant phenomena of the subjective view of the world are
209

Mario De Caro

constitutively incompatible with the natural sciences and consequently should be eliminated
from our ontology (as in the past happened with phlogiston, the epicycles of Ptolemaic astronomy, and the alleged magical properties of witchcraft). To return to our examples:
Pereboom (2004, 2017) and Caruso (2015) claim that free will is nothing more than an illusion;
Wallace (2016) goes beyond the well-known proposals by Stich (1983) and Churchland (1989)
in defending eliminativism about the mental; Machery (2015) claims that the same idea of
“concept” should be abandoned; and Benovsky (2019) argues in detail that aesthetic objects do
not exist.
The eliminativist attitude of some scientific naturalists is remarkably ambitious. For instance,
this is Rosenberg’s (2009) menu of the eliminable items:
Science forces upon us a very disillusioned ‘take’ on reality. It forces us to say “No” in
response to many questions to which almost everyone hopes the answers are “Yes”.
These are the questions about purpose in nature, the meaning of life, the grounds of
morality, the significance of consciousness, the character of thought, the freedom of
the will, the limits of human self-understanding, and the trajectory of human history.
In this quote there is a strange mixture of ideas that today no naturalist of any kind would
consider worthy of consideration (the purpose of nature, the trajectory of human history) with
others that are absolutely central to the ordinary worldview as if they were all at the same level
of plausibility. In any case, the fundamental question that faces eliminationism is a generalisation
of the thesis of Kant’s Third antinomy: can we really conceive of our world without the central
ideas of the subjective view, such as free will, consciousness, morality and so on? This is an
objection that also a minority of scientific naturalists – the so-called “mysterians” – formulate.
In their opinion, the features of the subjective view cannot be reduced to scientifically acceptable ones, but cannot be eliminated either since they play essential roles in our selfconception and lives. In this perspective, the placement problem cannot be solved (McGinn
1993, Nagel 1997; Pinker 1997; van Inwagen 2017). The inspiration for this view comes from
Noam Chomsky’s famous distinction between “problems” and “mysteries”: the former are
questions that we are in principle able to solve; the latter are questions that, because of our
limitations, we are and always will be unable to solve. In this sense – mysterians argue – how to
locate the recalcitrant features of the manifest image in the natural world will always be a
mystery for us, in the same sense in which Pythagoras’s theorem will always be a mystery for
dogs: something that we do not really understand. Among these alleged mysteries, McGinn
(1993) includes freedom (see also van Inwagen 2017), meaning, consciousness, the self, the a
priori and knowledge. Considering this list, it is clear that not much is left for philosophers to
work on; unsurprisingly, then, McGinn (1993, 197) calls philosophy a “futile” activity.
Mysterianism is an intellectually honest position insofar as it admits that the attempts at explaining in scientific terms, or explaining away, the features of the subjective view encounter
enormous challenges. However, this is also a defeatist view as long as it claims that we cannot
understand (and a fortiori cannot solve) some of the philosophical issues surrounding our selfconception. In this regard, one is reminded of Lynne Baker’s slogan that “we should not lend faith
to metaphysics that render ordinary but significant phenomena unintelligible”. (Baker 2013, 73).

5 Liberal naturalism and the reconciliation problem
From the point of view of liberal naturalism, scientific naturalism is based on a serious misconception that can labelled “the myth of monism”, to use John Dupré’s (2004) term. The
210

Liberal naturalism: origins and prospects

monistic attitude of scientific naturalism has an ontological face and an epistemological one:
natural sciences (if not physics alone) are our only genuine source of knowledge; therefore, they
have the final say on every ontological issue.
On the background of this view, there is the idea that nature is nothing more than the
subject matter of the natural sciences. This is a radical change compared with the versions of
naturalism held in the previous ages up to John Dewey (1925, 66), who wrote:
Mind and matter [are] different characters of natural events, in which matter expresses
their sequential order, and mind the order of their meanings in their logical connections and dependencies.
Dewey advocated a much broader view of nature than contemporary scientific naturalists
would concede. For him, besides being the subject matter of the natural sciences, nature expands to the normative components that belong to the subjective view. In turn, this implies a
form of constitutive pluralism, according to which there are different and mutually irreducible
ways of understanding a reality that in itself is irreducibly variegated.
A clearly pluralistic attitude like Dewey’s is central to contemporary liberal naturalism.7 This
view can be defined by the following three theses:
1
2

Some real entities are irreducible to, but not incompatible with, the entities that are part of
the domain of a science-based ontology.
Some legitimate forms of understanding (for example, a priori reasoning, conceptual
analysis and introspection) are neither reducible to scientific understanding nor incompatible with it.
There are issues in dealing with which philosophy is not continuous with science as to its
content, method and purpose, although it should not be at odds with it.

Considering these three theses, it is evident that liberal naturalism does not face the placement
problem: there is nothing to be “placed” in the scientific world, and thereby legitimated,
because the features of the subjective view simply do not need any legitimation. That said,
however, liberal naturalism encounters a different problem, which can be called the
“Reconciliation problem” (De Caro 2020). This concerns the relationship between the phenomena of the subjective view and those of the scientific view as well as how one should
evaluate the conflicts that frequently raise between these two views.
Consider the case of moral properties: what is the relationship between human beings when
they are judged through normative statements (“You must do that”, “One should not lie”, “I
was wrong to make that choice”) and when they are described in biological terms? Relying on
McDowell’s (1994, 1995) distinction, the question is: what is the relationship between the space
of reasons and the space of natural laws?
Liberal naturalists have developed various responses to these questions, but three are the
main ones. The first response posits a categorical distinction between the subjective and the
objective view such that they are totally unrelated (Strawson 1985; Bilgrami 2006; White 2007;
Macarthur 2019). According to this view – inspired by Kant and Wittgenstein –human beings
can be looked at from two different perspectives, which respectively rely on completely different conceptual apparatuses that identify real but unrelated features of the world. The advantage of this position is that the ordinary worldview and the scientific worldview can be
understood according to their own respective criteria, without questioning their relationship, so
that the reconciliation problem ipso facto dissolves. The disadvantage, however, is that the gap
211

Mario De Caro

between their respective conceptual apparatuses becomes unbridgeable – and possibly mysterious. Thus, the advocates of this view have to prove that, in their conception, the very
plausible idea that the two views relate to the very same world does not vanish altogether.
The second liberal naturalist response to the reconciliation problem appeals to emergentism,
the view according to which every organic complex is characterised by “emergent” properties
(Dupré 1995). These properties depend on the occurrence of specific physical and chemical
conditions but, crucially, cannot be predicted or explained by the properties of the parts that
constitute that organic complex. In in this perspective, then, , the whole cannot be reduced to
the sum of its parts since the features of the higher ontological level are unpredictable and
inexplicable by the features of the lower level, even if the former depend on the latter for their
existence. In this way, the problem of the relationship between the two worldviews is solved:
while the scientific view deals with the more fundamental properties, the subjective view deals
with the higher-level emergent properties. Even in this case, however, there are considerable
difficulties. The first is the classic objction according to which the notion of emergence is
obscure and that it does not explain what needs to be explained. More specifically, it remains
utterly unclear why in nature there are “jumps” that make it impossible to account for the
higher-level features of the entities in terms of the lower-level ones that constitute them. To
put it in a different way: according to this objection, emergentism just points toward a fundamental phenomenon of nature (that of emergence), but without explaining at all why that
phenomenon exists and, in essence, what its nature is. The emergentists, respond to this criticism by stating that the emergence of certain features on others is a “brute phenomenon”, in
itself inexplicable. This response, however, is widely considered unsatisfactory because it merely
acknowledges the existence of the phenomenon rather than illuminating it. Another common
criticism is that, in itself, emergentism is not able to make justice to the specificity of the
subjective view because it accounts (perhaps and at most) for the genesis of that view but
doesn’t say anything about its essential intentional and normative components.
The third answer that liberal naturalism can offer to the reconciliation problem resorts to the
notion of global supervenience (McDowell 2006; Putnam 2008). This is a relation of ontological dependence but weaker than the relation of reduction, which as said is rejected by liberal
naturalists as inapplicable to the subjective view. In its basic version, global supervenience is a
relation of covariance according to which if two worlds are identical regarding their subvenient
features (e.g. the physical ones), they are also identical as to their supervenient features (e.g.
those of the subjective view). In this case, if two subjects are different regarding their mental
features, then some physical features of their two respective worlds must also differ (in the heads
of the subject or in the respective external environments with which they causally interact).8 By
contrast, it is not necessary that a difference in the physical features corresponds to a difference
in the mental features. To give a more concrete example: the property of believing that
Tegucigalpa is the capital of Honduras and the property of believing that Managua is the capital
of Nicaragua necessarily correspond to two different physical configurations, but different
physical configurations may correspond to each of these two mental properties. In short, global
supervenience is a many-one relation: different subvenient features (those of the lower level,
which is represented by the whole physical world) can correspond to the same supervenient
features (those of the higher level), while different supervenient features always correspond to
different subvenient features. Identical physical worlds will therefore be identical also as to their
mental features (and to all the other supervenient features).
By appealing to global supervenience, the ontological nexus between the different levels of
features is guaranteed. However, this nexus, as mentioned, does not imply that higher properties (especially those belonging to the subjective view) must be reducible to lower ones.
212

Liberal naturalism: origins and prospects

There is no way, for example, to determine once and for all the whole set of physical features to
which a mental feature corresponds because that set is open: there can always be, for example,
new physical configurations underlying the mental property of believing that Tegucigalpa is the
capital of Honduras. And because the subvenient basis of higher-level phenomena is potentially
represented by the whole physical world, appealing to the notion of emergence – which requires that the complexity of the internal organisation of the specific components of a whole is
so complex that it produces an ontological jump – does not make much sense.
Predictably, scientific naturalists object that, from an ontological point of view, global supervenience is too weak a notion, because it fails to account for the unity of the world, that is,
for the idea that everything that exists is nothing more and above the fundamental physical
components of reality. This objection, however, does not trouble liberal naturalists at all.
Indeed, it merely reiterates a thesis they firmly disagree with: namely, that there is an ontological hierarchy that goes from the bottom to the top of reality. For a liberal naturalist who
appeals to global supervenience, the world is structured in levels: the lower levels are necessary
for the existence of the higher levels, and any change in the higher levels presupposes a change
in the lower levels; but this does not mean that the lower levels exhaust what needs to be said
with respect to the higher levels.
Also, some liberal naturalists, in particular those who see the subjective and the objective view
of the world as categorically different, refuse to appeal to supervenience for several reasons. First,
they claim that this appeal would be a dangerous concession to scientific naturalism. As John
McDowell (2006) has noticed, however, even a nonnaturalist such a G.E. Moore (1922, 263)
appealed to supervenience in regard to the relationship between values and nonnormative facts:
so, this objection misses its target. Another objection comes from John Dupré (2010, 291-293),
who has argued that the supervenience thesis is empirically vacuous. This objection, however, is
valid only if one agrees with Dupré’s very strong empiricist assumption that every belief has to be
grounded in empirical facts. Finally, other liberal naturalists claim that the supervenience thesis is
“unassessable” since the attempt to relate, say, evaluative facts to nonevaluative facts hinges upon a
categorical mistake. However, as noted by McDowell (2006, 71), if one appeals to the notion of
supervenience in the right way, one will notice that it is not philosophically problematic: “Given
an action that is, say, despicable, and another exactly like it in all nonevaluative respects, the
second must be despicable”, and this is an “innocuous claim”.
That said, scientific naturalists can still mount an attack against liberal naturalists by claiming
that their idea that philosophical beliefs have to be compatible with our best scientific beliefs is
too vague. For example, one can say that ontological dualism and creationism are logically
compatible with the scientific view of the world, and nobody should consider these theories as
naturalist, of course.
It should be noted, however, that liberal naturalists are appealing to a notion of compatibility
between philosophy and science that is not to be understood just in a logical sense, but in a
broader one. Not only is that one should not accept philosophical conceptions that openly
contradict our best scientific theories but also that one should not accept conceptions that
expand our ontology in areas that are already well accounted for by the explanations and entities
postulated by science alone. The notorious theory called “intelligent design” is a good example
in this sense. According to this conception, biological evolution is guided by a superior intelligence. Clearly this view is logically compatible with the neo-Darwinian theory of evolution
and – its advocates proclaim – actually fills its (alleged) gaps. However, if the former part of this
claim is correct, the latter is not. In order to explain biological evolution, there is no need to
postulate the role of a superior intelligence because the process of natural selection discovered
by Darwin, together with genetic mechanisms, does not have serious gaps and, in fact, offers an
213

Mario De Caro

excellent account of the evolution of species (Dennett 2006). Therefore, even liberal naturalists,
like radical naturalists, can easily reject intelligent design, with its antinaturalistic ontological
implications.
Differently from scientific naturalists, however, liberal naturalists hold that traditional philosophical methods may be legitimate even when they are irreducible to the methods of the
natural sciences, as long as they are not incompatible with them (in the broad sense of
“compatibility” discussed above). Thus, for the liberal naturalist, while mystical intuition is not
legitimate – insofar as it is “broadly incompatible” with the scientific worldview – conceptual
analysis, phenomenological inquiry, or the transcendental method are. Actually, this set of
methods is essential to the philosophical enterprise and characterises it as an autonomous domain. In this perspective, then, if the infamous Heideggerian saying that “science does not
think” should be rejected as misleading and ideological, so too is the attitude of scientific
naturalism, according to which, in essence, only science thinks (with the corollary that philosophy can “think” only to the extent that it mimics science). The question to ask, at this
point, is what the relationship between science and philosophy is when they conflict regarding
some issues.
For liberal naturalists, science does not necessarily have the last word (even if, of course, this
is frequently the case). There are in fact cases in which philosophy can clarify some aspects of
scientific investigation, and sometimes even help settle some disputes between scientists (vice
versa there are also situations in which science can contribute to philosophical discussions).
Consider, for example, the cases in which scientists themselves often reflect with philosophical
tools. Good examples of this are the classical discussions on infinitesimal calculus or on set
theory or, more recently, the debates on the interpretations of quantum mechanics or string
theory, or even the investigations on the neurophysiological bases of mental phenomena. In
these domains, philosophical reflection and scientific investigation often contribute to each
other.
At this point, scientific naturalists might reframe their charge. In particular, they could claim
that liberal naturalism, by assuming that natural sciences cannot explain the whole of reality, is
still too inclusive from an epistemological point of view. The reason is that it leaves room for
conceptions that are clearly irreconcilable with naturalism, including the most radical cultural
relativism or postmodernism, which substantially devalue the status of natural science, going so
far as to proclaim the lack of objectivity of its theories. However, in adopting a perspective that
accommodates both scientific and ordinary worldviews, liberal naturalism is immune to this
charge: this conception asserts, in fact, that reality limits the legitimacy of our interpretations,
insofar as it determines the truth conditions of the claims we make on the basis of our different
cognitive sources. It is obvious that all of our judgments about the world are fallible and will
always be; but this does not at all mean that there are no objective canons of truth as cultural
relativism, postmodernism and similar positions claim (Boghossian, 2006).
Another, more insidious, critique of liberal naturalism acknowledges its conceptual legitimacy, but declares its implausibility. This critique has been expressed in different ways. One of
such ways is based on the “burden of proof argument”. According to this argument, scientific
naturalism is the default naturalistic conception. Therefore, it is up to the liberal naturalist to
prove that, in principle, the natural sciences cannot account for some aspect of reality. Put in
another way: it is a task for the liberal naturalist to show that some of the real properties of the
world will never be reducible to the properties accepted by the natural sciences (Macdonald
2006, 231) – and this sounds very much as an attainable task.
Not even this argument, however, is convincing. First, scientific naturalists, who by definition reject the possibility of nonempirical arguments, are not entitled to ask the liberal
214

Liberal naturalism: origins and prospects

naturalist to demonstrate that some properties of the world are irreducible to his scientific
properties in priniciple. Moreover, by the same yardstick, the liberal naturalist might argue that
the burden of proof falls on those who claim that phenomena such as free will, consciousness
and so on can be explained by the natural sciences.
A different way of raising the charge of implausibility against liberal naturalism is based on
the so-called “great success of science argument”. According to this argument, starting with the
scientific revolution, natural science has progressively explained an astonishing number of
phenomena that had previously appeared indecipherable, making it possible to predict and
control them; therefore, it is rational to infer that natural sciences can also explain the problems
of free will, consciousness, personal identity and so on.
Not even argument, however, works. First, the inductive inference on which it is based is
very problematic: why should it necessarily be possible to explain phenomena in a domain with
kinds of explanations that have worked in other domains? Moreover, as has been noted (Crane
and Mellor 1995), it is not clear to which scientific theories the argument of the great success of
science refers. Certainly, it cannot apply to today’s theories since they are incapable of solving
the problems of free will, consciousness and so on. On the other hand, we really have no idea
what kind of theories (if any) could be able to explain such problems. Wouldn’t it be better,
then, to confine this kind of speculative reasoning to the (already crowded) sphere of overly
ambitious philosophical fantasies? Moreover, when employed against liberal naturalism, the
argument of the great success of science actually sounds question-begging. Liberal naturalism, in
fact, hinges on the idea that it is rational to believe that some important features of the world are
ineliminable and not reducible to the features of the natural sciences. Merely proclaiming that
such properties can be eliminated or reduced because science is inherently capable of doing so
seems very much like a petitio principii.
A final way of denying plausibility to liberal naturalism has already been mentioned and
consists in emphasising the indubitability of ontological and epistemological monism
(Churchland 1996; Schaffer 2018). From this point of view, all properties of the world that now
seem to us ineliminable and irreducible to properties acceptable by the natural sciences, in fact,
are necessarily either eliminable or reducible. The emphasis on monism can sometimes be
justified as a methodological criterion; if, however, monism is considered as an ontological
principle, not only is it at odds with the ordinary worldview but also with the most common
scientific practice. Today, in fact, pluralism is a very widespread conception even within the
natural sciences (Ludwig and Ruphy 2021). And the monistic cause becomes even less promising when the human and social sciences are also taken into consideration. In fact, the idea
that pluralism poses a threat to the scientific conception of the world is an anachronistic idea
that it would be time to abandon.

Notes
1
2

3
4
5

The term “queer” has been first used in the philosophical context by Mackie (1977, 36).
It should be noted that the late Renaissance saw the flourishing of many forms of Platonism and
Aristotelianism, often very different from each other. The two forms discussed by Mazzoni, however,
were two of the most influential. On the discussion on the mathematisation of physics, see
De Caro (2017).
On the distinction between the subjective and the objective views, see also Nagel (1979).
In this respect, one should remember that Strawson was deeply influenced by Kant: see his (1966).
Wood (1999, 173) writes that the distinction between phenomenal and noumenal causation is “a
desperate expedient”.

215

Mario De Caro
6
7
8

Remarkably, Rosenberg uses the term “scientism”– which is normally used with derogatory
connotations – to label his own view.
On liberal naturalism, see Strawson (1985); McDowell (1995); De Caro and Macarthur (2004, 2010,
forthcoming); and Putnam (2015).
The reference to the environments is aimed at accounting for the idea of semantic externalism.

References
Baker, L.R. Naturalism and the First-person Perspective. Oxford University Press. (2013).
Benovsky, J. Eliminativism, Objects, and Persons: The Virtues of Non-Existence. Routledge. (2019).
Bilgrami, A. Self-Knowledge and Resentment. Harvard University Press. (2006).
Bilgrami, A. Some Philosophical Integrations. In Macdonald C. and Macdonald G. (eds.), McDowell
and His Critics, pp. 50–66. Blackwell. (2006).
Boghossian, P. Fear of Knowledge: Against Relativism and Constructivism. Oxford University Press. (2006).
Caruso, G. Exploring the Illusion of Free Will and Moral Responsibility. Lexington Press. (2015).
Chatterjee, A. The Aesthetic Brain How We Evolved to Desire Beauty and Enjoy Art. Oxford University
Press. (2013).
Churchland, P.M. A Neurocomputational Perspective: The Nature of Mind and the Structure of Science. MIT
Press. (1989).
Churchland, P.M. The Engine of Reason, the Seat of the Soul. MIT Press. (1996).
Crane, T. and Mellor, D.H. There Is No Question of Physicalism. Mind, 99, 185–206. (1995).
De Caro, M. On Galileo’s Platonism, Again. In R. Pisano et al. (eds.), Hypotheses and Perspectives in the
History and Philosophy of Science, pp. 85–104. Springer. (2017).
De Caro, M. Realism, Common Sense, and Science. The Monist, 98(2), (2015). 197–214.
De Caro, M. The Indispensability of the Manifest Image. Philosophy and Social Criticism, 46(2), (2020).
162–172.
De Caro, M. and Macarthur, D. (eds.) Naturalism in Question. Harvard University Press. (2004).
De Caro, M. and Macarthur, D. (eds.) Naturalism and Normativity. Columbia University Press. (2010).
De Caro, M. and Macarthur, D. (eds.) forthcoming. Liberal Naturalism. Harvard University Press.
Dennett, D. The Hoax of Intelligent Design and How It Was Perpetrated. In J. Brockman (ed.), Intelligent
Thought: Science versus the Intelligent Design Movement, pp. 33–49. Vintage. (2006).
Dewey, J. Experience and Nature. Allen and Unwin. (1925).
Dupré, J. The Disorder of Things: Metaphysical Foundations of the Disunity of Science. Harvard University
Press. (1995).
Dupré, J. The Miracle of Monism. In M. De Caro and D. Macarthur (eds.), Naturalism in Question,
pp. 36–58. Harvard University Press. (2004).
Dupré, J. How To Be Naturalistic without Being Simplistic in the Study of Human Nature. In M. De
Caro and D. Macarthur (eds.) Naturalism and Normativity, pp. 289–303. Columbia University
Press. (2010).
Endicott, R.P. Multiple Realizability. In T. Gale (ed.), The Encyclopedia of Philosophy, 2nd edition,
pp. 427–432. Macmillan Reference. (2005).
Endicott, R.P. The Counter-Revolution over Multiple Realization. Metascience, 26(2), (2017). 229–232.
Galilei, G. Il Saggiatore. Engl. trans. (partial), The Essayer. In S. Drake (ed.), Discoveries and Opinions of
Galileo, pp. 231–280. Doubleday & Co. ([1623]/1957).
Gallagher. Aesthetics and Kinaesthetics. In H. Bredekamp and J.M. Krois (eds.), Sehen und Handeln,
pp. 99–117. Akademie Verlag. (2011).
Gallagher, S. Why We Are Not All Novelists. In P.F. Bundgaard and F. Stjernfelt (eds.), Investigations Into
the Phenomenology and the Ontology of the Work of Art, pp. 129–144. Springer. (2015).
Gozzano, S. and Hill, C. New Perspectives on Type Identity: The Mental and the Physical. Cambridge
University Press. (2012).
Hutto, D. Enactive Aesthetics: Philosophical Reflections on Artful Minds. In A. Scarinzi (ed.), Aesthetics and
the Embodied Mind: Beyond Art Theory the Cartesian Mind-Body Dichotomy, pp. 211–227. Springer. (2015).
Jackson, F. From Metaphysics to Ethics. Oxford University Press. (1998).
Kant, I. Kritik der reinen Vernunft. Engl. transl. Critique of Pure Reason. St. Martin’s Press. ([1781]/1965).
Kim, J. Multiple Realization and the Metaphysics of Reduction. Philosophy and Phenomenological Research,
52(1), (1992). 1–26.

216

Liberal naturalism: origins and prospects
Kim, J. Physicalism, or Something Near Enough. Princeton University Press. (2007).
Ludwig, D and Ruphy, S. Scientific Naturalism. In E. Zalta (ed.), Stanford Encyclopedia of Philosophy,
(2021). https://plato.stanford.edu/entries/scientific‐pluralism/
Macarthur, D. Liberal Naturalism and the Scientific Image of the World. Inquiry, 62(5), (2019). 565–585.
Machery, E. Doing Without Concepts. Oxford University Press. (2009).
Mackie, J. Ethics: Inventing Right and Wrong. Penguin. (1977).
Macdonald, C. and McDonald, G. (eds). McDowell and His Critics. Blackwell. (2006).
Mazzoni, J. In Universam Platonis et Aristotelis Philosophiam Praeludia, sive de comparatione Platonis et Aristotelis.
Venezia. (1597).
McDowell, J. Mind and World. Harvard University Press. (1994).
McDowell, J. Two Sorts of Naturalism. Repr. in Id. Mind, Language, and Reality, pp. 167–197. Harvard
University Press. ([1995]/1998).
McDowell. Response to Bilgrami. In C. Macdonald and G. MacDonald (eds.), McDowell and His Critics,
pp. 66–72. Blackwell (2006).
McGinn Problems in Philosophy: The Limits of Inquiry. Blackwell. (1993).
Moore, G.E. Philosophical Studies. Routledge. (1922).
Nagel, T. Subjective and Objective. In Id. Mortal Questions, pp. 207–222. Canto. (1979).
Nagel, T. A View from Nowhere The View from Nowhere. Oxford University Press. (1986).
Nagel, T. The Last Word. Oxford University Press. (1997).
Pereboom. D. Is Our Conception of Agent-Causation Coherent? Philosophical Topics, 32, (2004).
275–286.
Pereboom, D. Free Will, Agency, and the Meaning in Life. Oxford University Press. (2014).
Pereboom, D. Living without Free Will. Cambridge University Press. (2017).
Pinker, S. How the Mind Works. Norton. (1997).
Polger, T.W. and Shapiro, L.A. (eds.). The Counter-Revolution Over Multiple Realization. Oxford University
Press. (2016).
Price, H. Naturalism Without Representationalism. In M. De Caro and D. Macarthur (eds.), Naturalism in
Question, pp. 71–105. Cambridge: Harvard University Press.
Putnam, H. Reply to White. European Journal of Analytic Philosophy, 4(2), (2008). 29–32.
Putnam, H. Naturalism, Realism, and Normativity. Repr. in M. De Caro (ed.), Naturalism, Realism, and
Normativity, pp. 21–43. Harvard University Press. (2004).
Rosenberg, G. A Place for Consciousness. Probing the Deep Structure of the Natural World. Oxford University
Press. (2004).
Rosenberg, A. The Disenchanted Naturalist’s Guide to Reality. (2009). https://nationalhumanitiescenter.
org/on-the-human/2009/11/the-disenchanted-naturalists-guide-to-reality/.
Schaffer, J. Monism. In E. Zalta (ed.), Stanford Encyclopedia of Philosophy. (2018). https://plato.stanford.
edu/entries/monism/.
Smart, J.J.C. The Mind/Brain Identity Theory. In E. Zalta (ed.), Stanford Encyclopedia of Philosophy. (2007).
https://plato.stanford.edu/entries/mind-identity/.
Stich, S. Folk Psychology to Cognitive Science: The Case Against Belief. MIT Press. (1983).
Strawson, P.F. The Bound of Sense: An Essay on Kant’s Critique of Pure Reason. Methuen. (1966).
Strawson, P.F. Skepticism and Naturalism: Some Varieties. Columbia University Press. (1985).
Van Inwagen, P. Free Will Remains a Mystery. In Id. Thinking about Free Will, pp. 90–110. Cambridge
University Press. ([2000]/2017).
Wallace, M. Saving Mental Fictionalism from Cognitive Collapse. Res Philosophica, 93(2), (2016).
405–424.
White, S. Empirical Psychology, Trascendental Phenomenology, and the Self. In M. Marraffa, M. De
Caro & F. Ferretti (eds.), Cartographies of the Mind. Philosophy and Psychology in Intersection, pp. 243–254.
Springer. (2007).
Wood, A.W. Kant’s Ethical Thought. Cambridge University Press. (1999).
Zeki S., Bao Y., Pöppel E. The Art, Science, and Brain Tryptich. Psych Journal, 9(4), (2020). 427–428.

217

18
LIBERAL NATURALISM
AND GOD
Fiona Ellis

1 Introduction
In my book God, Value, and Nature, I defend a form of liberal or expansive naturalism.1 The
naturalist dimension of my position is pretty orthodox from the point of view of contemporary
Anglo-American philosophy, for naturalism is the dominant programmatic approach, and it is
made clear that this is what we ought to be.2 Why? Because it is the only way of avoiding the
metaphysical and epistemological difficulties of the opposing supernaturalist position.
Supernaturalism involves the postulation of weird entities and/or realms of being – things like
immaterial souls, Platonic forms and divine beings, all of which stand outside the natural world,
and cannot be understood to be a part of it.3 The naturalist protests that there is no way of
explaining how this supernatural realm relates to our ordinary familiar world, that there is an
equal mystery of how we gain knowledge of it, and that we can explain what needs to be
explained without making this problematic metaphysical detour. Naturalism is a form of in this
respect: the natural world is the only world there is, and there is nothing beyond it– no God, no
Platonic form of Goodness, no heaven, above us only sky.4
I defend a form of naturalism which can accommodate God. That is to say that on my
position the natural world is God-involving. This is highly unorthodox from the point of view
of naturalistic philosophy, for most naturalists (and many theists too!) take these positions to be
logically incompatible.5 By contrast, I want to say that they can both be true – that one can be a
naturalist and a theist. The naturalism at issue here takes us beyond the scientistic paradigm
which has defined the position until more recently.6 That is to say, it involves a rejection of the
contestable idea that there is no more to the natural world than what the scientist can comprehend. It counts as liberal in this respect, and the world thus understood is a value-involving
world. The naturalist will see this as a covert way of reintroducing supernaturalism, but Iris
Murdoch describes such a position as true naturalism, claiming that “the true naturalist…is one
who believes that as moral beings we are immersed in a reality which transcends us and that
moral progress consists in awareness of this reality and submission to its purposes”.7 Murdoch is
a true naturalist in this sense, she describes her position as a form of Platonism, and contemporary liberal or expansive naturalists – myself included – are indebted to her vision.
The typical liberal naturalist – Murdoch included – opposes theism, and does not engage
with the question of the relation between naturalism and God – understandably so if theism and
218

DOI: 10.4324/9781351209472-22

Liberal naturalism and God

naturalism are logically incompatible. I shall argue that such an attitude presupposes a contestable conception of God, and an equally contestable conception of the moral reality which,
on my position, has a theistic dimension. This much casts doubt upon some prevalent assumptions amongst naturalists and atheists, but my position invites equal worries from the
theist’s camp, and I shall take as my focus two related objections which were articulated and
discussed at a recent workshop on naturalism and supernaturalism.8
The first objection – expressed by Lynne Baker – is that if any form of naturalism is true,
then there is nothing outside the natural world. This means that “reality stops with the
mundane” and “nothing is transcendent”. The second related objection comes from John
Cottingham who objects that on my position I’ve got to deny that there is anything “external”
or “transcendent”, whereas he wants to say that “even after discarding silly ideas of “another
place” or a destination where we will be issued with harps (or pitchforks), we still need
something “external”.9 He then refers to the “radical immanentism” of my own position,
distinguishing it from his own commitment to a kind of “divine externality”.
I’ve said already that naturalism as I understand it can accommodate God. I take this to mean
that it can accommodate the transcendent. Yet if Baker is right that “nothing is transcendent”
on the naturalist picture, then we have a justification for Cottingham’s claim that I am committed to “radical immanentism”, and that radical immanentism excludes God, assuming that
God requires reference to the transcendent. I shall argue that this objection presupposes a
contestable metaphysical framework, and that a properly theistic position must reject it. We
shall see that there is a knife-edge between theism thus conceived and Murdoch’s true naturalism. There are important implications here for an understanding of the limits of liberal
naturalism, assuming that true naturalism in this sense comes under its umbrella.

2 Liberal naturalism
The naturalist denies that there is anything beyond nature, and it is a common enough refrain
amongst naturalists that this world is the only world there is. Thus, John Hermann Randall,
writing in 1944, claims that “naturalism” can be defined negatively “as the refusal to take
‘nature’ or ‘the natural’ as a term of distinction”.10 “Nature” he continues, has become “the allinclusive category”. In Randall’s hands, this is a reductive claim, for he commits to an exclusively scientific conception of nature.11 This is scientific naturalism, and it is unclear what
reason could be given for insisting that nature is to be measured in these terms alone. As John
McDowell puts it, “scientism is a superstition”, and we should “discourag(e) this dazzlement by
science” which leads us to suppose that “genuine truth is restricted to what can be validated by
their methods”.12
McDowell defends a form of liberal naturalism which rejects the offending scientistic
strictures, and it promises to accommodate the idea that there are values in the world which
make normative demands upon us. It is similar to Murdoch’s “true naturalism” in this respect,
and it involves nothing spooky or other-worldly, except in so far as we have moved beyond the
limits of the world as scientistically conceived. We might even go so far as to describe moral
reality as “supernatural” in this respect, “supernatural” being the logical complement of
“natural” in the scientific sense.
Liberal naturalism in this sense involves a form of moral realism, and the moral judgments we
make in this context are assessable as true or false. This is not to deny that there will be genuine
dilemmas and conflicts, and Murdoch talks of the slowness of moral change and achievement,
and of the “infinite difficulty of the task of apprehending a magnetic but inexhaustible reality”.13 She refers in this context to the clear vision which comes from imagination, effort and
219

Fiona Ellis

attention – a vision in which the will becomes a matter of “obedience”, and reality is revealed
“to the patient eye of love”.14 Her aim is to articulate a moral philosophy in which “the
concept of love…can once again be made central”, and she talks in this context of an “ideal
limit of love or knowledge which always recedes”.15
Some of this imagery sets Murdoch apart from McDowell, and there is a question of
whether she has erred in the direction of a more suspect supernaturalism. But what does it mean
to be appropriately suspect in this context? It is not enough to say that the offending position
involves reference to a weird realm of being, for the liberal naturalist commits this error from
the viewpoint of the scientific naturalist, and scientific naturalism is itself suspect.16 So there’s
nothing weird to the idea that there are dimensions of nature which elude science, and which
count as supernatural in this innocuous sense, and contemporary expansive naturalists are
adamant that the natural world is the only world there is. Witness James Griffin:
[v]alues do not need any world except the ordinary world around us…An otherworldly realm of values just produces unnecessary problems about what it could
possibly be and how we could learn about it. All that seems right to me right. But to
defend it, one does not have to adopt a reductive form of naturalism.17
What of Murdoch’s claim that moral reality has an infinite elusive character, and that we face
the task of apprehending a magnetic but inexhaustible reality? After all, Murdoch is a Platonist,
and Platonism involves reference to a realm of forms which is grounded in the ultimate form of
Goodness. McDowell defends Plato against the suspect supernaturalist charge by describing him
as a naturalist “with a penchant for vividly realized pictorial presentations of his thought”.18 He
distinguishes this down to earth Platonism from the “rampant” variety in which moral reality
lies in some inaccessible beyond.19
The implication here is that Plato’s position can be shorn of the picture-thinking, and that it
is equivalent to a liberal form of naturalism in this respect. But how are the limits of such
naturalism to be understood? What if the picture-thinking has a point which goes beyond the
purely ornamental? And what if the so-called rampant variety is itself just a pictorially vivid way
of making this point? I am thinking here of the idea – so important to Murdoch (and Plato too)
– that the truths at issue in this context lie at the blurry limits of our capacity to comprehend.
There is no obvious reason for insisting that such epistemic humility is suspect; nor does it
require reference to another world except in so far that such talk is just a vivid way of giving
expression to our limitations.20
McDowell uses the imagery of darkness to refer to that which exceeds the limits of his own
liberal conception of nature, he makes clear that “natural” as he understands it is “not supernatural (not occult, not magical)”, and adds:
There is no need for me to take a stand on whether everything is natural in that sense
(thereby, among other things, giving needless offence to people who think respect for
modern science is compatible with a kind of religious belief that preserves room for
the supernatural.21
He seems to be suggesting that the relevant mysterious dimension, such as it is, can have no
bearing upon an understanding of the natural world and our natural human being. It is, after all,
“occult” and “magical”, and seemingly sealed off from anything to which we could be receptive by virtue of our natural human being, including, of course, the moral reality with which
we engage at this level.
220

Liberal naturalism and God

The implication here is that there is nothing intrinsically mysterious about moral reality,22
and that any darkness should be relegated to that which is supernatural in a more suspect sense.
We are to suppose that it is in this context that we ascend to the level of religious reality. The
idea that moral reality is unmysterious takes us some distance from Murdoch’s true naturalism,
and Murdoch takes herself to be defining and defending an authentic form of religion in this
context – religion without God as she sees it.23 So both McDowell and Murdoch are in the
business of articulating an atheistic liberal naturalism, but Murdoch takes the moral reality at
issue to be infinite in its mystery, depth and religious significance, whereas McDowell seems to
want to relegate any mystery to the realm of the supernatural, the supernatural in this context
having a religious significance which has no bearing upon morality.
The assumption here is that religious reality takes us into the realm of suspect supernaturalism, and that it does so by committing us to a second, supernatural, realm in addition to
the natural world. The charge is familiar, as is the response, for we can ask again whether there
isn’t a more sympathetic way of interpreting the idea that religious reality is supernatural, and,
equally to the point, what the argument is for insisting that it is to be dualistically opposed to the
moral. Murdoch herself insists that moral philosophy should attempt to retain a central concept
which has all of the characteristics traditionally associated with God, where God “was (or is) a
single perfect transcendent non-representable and necessarily real object of attention”.24 This concession is
significant and the ambivalence palpable, but Murdoch agrees that suspect supernaturalism is to
be avoided, reserving the complaint for theism. Theism counts as such in the sense that its
defining beliefs involve reference to supernatural phenomena (God is a supernatural person,
heaven is a supernatural place),25 and this supernaturalism is problematic not only because it
detracts attention from what really matters (i.e. morality), but because it points in the opposite
direction. It does so by pandering to our egoistic desires, when, for example, we are motivated
to be moral for the sake of heavenly rewards.26 A clear argument for concluding that theism
stands opposed to the moral, but Murdoch’s egoistic interpretation can be contested, and her
aspirations for moral philosophy pose a challenge to the idea that a properly moral life must
dispense with God.

3 Rethinking God
I have tried to put pressure on the idea that the God/nature and God/morality distinctions are
as absolute and unambiguous as the typical naturalist assumes. Murdoch makes a move in this
direction by defending a naturalism in which the transcendent plays a fundamental role.
However, she dissociates this naturalism from theism, identifies theism with suspect supernaturalism, and recommends that we believe instead in “the unique sovereign place of goodness
or virtue in human life”.27 Goodness in this context is the “magnetic center toward which love
naturally moves”, and it is in this sense that we have a moral philosophy in which the concept of
love is central.
Murdoch’s true naturalism poses a challenge to morally deficient forms of religion and
theism, and we can agree – as many theologians have agreed – that the conception of God as a
supernatural person raises difficulties, and that God thus understood is easily side-lined when set
against the intraworldly loving relations which are so obviously central to a properly human life.
Yet those who take seriously such a conception can insist that we are up against the limits of
language in this context, and that the description is not intended to be taken in these literalist
terms – as if God is an infinitely remote super-being with no bearing upon our loving relations
with others. The point is familiar from my previous defence of picture thinking, but there is a
more interesting response in this context – one which grants the relevant conceptual obstacles
221

Fiona Ellis

whilst insisting that there are better and worse pictures in this context, and some which can help
us to make better sense of the idea that God and nature (and hence, God and morality) are
inextricably tied.
This diagnostic approach is to be found in John Robinson’s famous 1963 book Honest to
God.28 Robinson’s book is indebted to various German theologians, all of whom seek to move
beyond dualistic supernaturalism and reductive naturalism. The naturalist critique of supernaturalism allows us to tear down an idol – namely, the idea of God as a distant supernatural
being – whilst theism allows us to “challenge the naturalist’s assumption that God is merely a
redundant name for nature or for humanity” and to safeguard God’s transcendence.29 But what
does it mean to introduce a transcendent God? Robinson answers this question with the help of
Paul Tillich. Tillich tells us that:
To call God transcendent […] does not mean that one must establish a ‘superworld’ of
divine objects. It does mean that, within itself, the finite world points beyond itself. In
other words, it is self-transcendent.30
Robinson takes this “great contribution to theology” to involve:
the reinterpretation of transcendence in a way which preserves its reality while detaching it from the projection of supranaturalism [Robinson’s term for suspect supernaturalism]. ‘The Divine’, as he sees it, does not inhabit a transcendent world above
nature; it is to be found in the ‘ecstatic’ character of this world, as its transcendent
Depth and Ground. Indeed, as a recent commentator has observed, supranaturalism
for Tillich actually represents ‘a loss of transcendence’.31
The themes, claims and images are familiar from Murdoch’s true naturalism, and Robinson
likewise gives center stage to the concept of love, taking as his starting point Ludwig
Feuerbach’s claim that the true atheist “is not the man who denies God, the subject; it is the
man for whom the attributes of divinity, such as love, wisdom and justice, are nothing”.32
Robinson grants that this is very near to his own position in the sense that he is wanting to
interpret theological assertions as assertions about human life, but he insists that this is NOT a
form of atheism, for love thus understood is grounded in God. It is in this sense that we can
agree with Dietrich Bonhoeffer that “God is the ‘beyond’ in the midst”.33 As for the worry that
this picture involves a denial of God’s transcendence, the proper response is to point out – with
Tillich – that it is the opposing dualistic picture which carries this implication. It does so by
reducing God to this-worldly categories – he becomes “one object among other objects” as
Karl Rahner put it – and on a level with any other thing in this respect.34
The idea of God as love has been developed more recently by Paul Fiddes.35 Fiddes argues
that it offers a real alternative to the offending dualistic picture, for it allows us to say that God’s
being is irreducibly relational, and that we partake in this reality when we stand in loving
relations to others. Understood from this perspective the distinction between God and world is
no longer to be modeled on two externally related items between which there is an insurmountable gap. The picture is rather of a circle within a circle – the large circle corresponding to God’s infinite love, and the small one capturing the sense in which we ourselves are
capable of partaking in this love by standing in loving relations to others. Fiddes adds that it is an
implication of this position that “the presence of God will always be hidden in the sense that it
cannot be observed or known as an object of perception, but can only be participated in…God
is not the object of desire but the one in whom we desire the good”.36
222

Liberal naturalism and God

I am gesturing toward a position which requires extensive thought and elucidation, but what
little I have said suggests that its structure offers a way of avoiding a conjunctive conception of
the God/world relation, and that it has much in common with Murdoch’s true naturalism. We
are immersed in a reality which transcends us, love is central to this framework, and love is that
by virtue of which we move – however falteringly – toward this infinite reality. As for the
worry that Murdoch’s Goodness is too impersonal to admit of theistic characterisation, it should
be clear from what has been said that there is a real and unresolved issue concerning what it
means to describe God in personal terms, and whether those terms are admissible. We are
reminding here of Feuerbach’s important point that the true atheist is “is not the man who
denies God, the subject; it is the man for whom the attributes of divinity, such as love, wisdom,
and justice, are nothing”.

4 Responding to Baker and Cottingham
I have made theistic sense of Murdoch’s true naturalism, and I want now to spell out the
implications for an assessment of the objections advanced by Baker and Cottingham. Baker’s
worry is that, if any form of naturalism is true, then there is nothing outside the natural world.
She takes this to mean that “reality stops with the mundane” and “nothing is transcendent”.
Cottingham protests in similar vein that the naturalist has to deny that there is anything “external” or “transcendent”, and that all that remains is a “radical immanentism”.
Naturalism in the sense with which I am concerned is to be distinguished from radical
immanentism as Cottingham understands it, for Cottingham’s radical immanentism is a form of
atheism, whereas the natural world as I understand it involves God. Furthermore, it involves
God in a sense that should be acceptable to Cottingham, for I agree with him that there are
intimations or traces of God to be found in the natural world, and that they are manifest, for
example, “in the compelling power of our moral sensibilities”.37 Cottingham objects that a
naturalist has to deny that there is anything “external” or “transcendent”, whereas he wants to
say that “even after discarding silly ideas of ‘another place’ or a destination where we will be
issued with harps (or pitchforks), we still need a kind of divine externality”. This suggests that
Cottingham agrees with me that God is not externally related to the world in the way that
things within the world are spatially related, but he thinks that a liberal naturalist picture cannot
accommodate divine externality. That is to say, Cottingham thinks that on my position God is
reducible to the world.
I have explicitly denied that God is reducible to the world, and have argued that the
proposed framework offers the prospects for safeguarding God’s transcendence rather than
eliminating it. It does so by guaranteeing that God is irreducible to any finite measure, and
allowing that God’s infinite reality has depths which exceed our powers of love and knowledge.
We are immersed in a reality which has an ever-receding limit in this respect, and, being so
immersed, we must give up on the idea that the relation between God and the world is a
conjunctive relation – as if God and the world add up to two, as Herbert McCabe puts it in the
context of making a similar anti-dualistic point.38 Indeed, I am happy to describe this liberal
naturalism as a radical immanentism, provided that it is made clear that the force of “radical” in
this context, rather than signifying a thoroughgoing atheism, serves to capture the position
which becomes available once the transcendent/immanent dualism has been put to rest. This, I
would contend, is the truly radical approach.
What I have said addresses Baker’s worry that naturalism fails to accommodate transcendence, and it also offers a response to her objection that if naturalism is true, then there is
nothing outside the natural world. Certainly, there is nothing outside the natural world if this
223

Fiona Ellis

involves denying that God lies beyond or somehow in competition with the natural world. As
I’ve made clear, however, it does not follow from this denial that God is reducible to the world,
and to suppose that it does is simply to propagate the offending dualistic framework – one
according to which God must either be out there in the manner of a supernatural being or
squeezed out of the picture altogether. Neither of these pictures can accommodate God’s
transcendence.
What of the worry that if naturalism is true then reality must stop with the mundane? Liberal
naturalists like Griffin make a point of describing the natural world as “ordinary” and “familiar”,
no doubt to fend off the worry that the position involves anything remotely weird. Perhaps
there is a similar motive at work in McDowell’s unwillingness to allow that moral reality could
be mysterious in a religiously significant sense. McDowell’s position is open to challenge in this
respect, and from a purely phenomenological view it seems absurd to deny that the world can
be strange, extraordinary, enchanting, terrifying, ecstasy-inducing and all of the other things
which are excluded if we settle for no more than the ordinary and the familiar. Perhaps the
limits of the ordinary and the familiar are suitably expanded in Griffin’s more liberal scheme of
things, but I take it that Baker’s reference to the mundane is intended to confine us to a
disenchanted nature, and the naturalist can resist these terms of debate.

5 Moving ahead
I have defended a conception of naturalism which is inspired by Murdoch’s true naturalism, but
which is to be comprehended theistically. I am not suggesting that this position is mandatory;
the point is simply that it is worth taking seriously, and that some standard objections to it can
be overcome. The existence of the present volume testifies to the pliability of the term
“naturalism”, and although the orthodox liberal approach is atheistic, this atheism is premised
upon a contestable and highly problematic conception of God. I have argued that we should
reject the offending conception, whilst allowing (with the typical liberal naturalist) that the
reality in which we are immersed is a value-involving world. The idea that it points in the
direction of God becomes infinitely less weird if we remind ourselves that this does NOT mean
that there is a weird super-being at the end of the journey. As a wise man once put it, perhaps
we are already deep in God.39

Notes
1 Ellis (2014). Some of the material of the present paper appears in simplified form in a piece published
for Think: Philosophy for Everyone (issue 56, 2020). Thanks to the Royal Institute of Philosophy and
Cambridge University Press for permission to use this material.
2 See Bernstein (1995, 58) and De Caro and Macarthur (2004).
3 See, for example, De Caro and Macarthur (2004), Dupré (2010), Stroud (2004). Older versions of
these criticisms can be found in the papers in Krikorian (1944), and are well summed up in John
Randall (1944).
4 Clear statements of this antisupernaturalist dialectic are to be found in Stroud (2004), and
Dupré (2010).
5 See, for example, Plantinga (2011).
6 See the introduction to De Caro and Macarthur (2004) for a detailed exposition of scientific naturalism
and some responses to it. See also Ellis (2014, chap. 1).
7 Murdoch (1956).
8 This workshop – which took place at Heythrop College in October 2017 – was part of a project
entitled Supernaturalism and Naturalism: Beyond the Divide which I directed with Mario De Caro. Lynne

224

Liberal naturalism and God

9
10
11
12
13
14
15
16

17
18
19
20

sadly passed away during this time, and her paper – “Beyond Naturalism” – was presented by Kate
Sonderegger.
Cottingham’s worries were expressed in an email exchange following the relevant workshop.
Randall (1944, 357).
Randall (1944, 358). As he puts it: “[t]here is no ‘realm’ to which the methods for dealing with nature
cannot be extended. This insistence on the universal and unrestricted application of ‘scientific method’
is a theme pervading every one of these essays [in the volume for which his essay was the epilogue]”.
McDowell (2002, 295).
Murdoch (1970, 42).
Murdoch (1970, 37–40).
Murdoch (1970, 28).
In any case, there is nothing weirder than the things discovered by science. As Mark Platts puts it: “The
world is a queer place. I find neutrinos, aardvarks, infinite sequences of objects, and (most pertinently)
impressionist paintings peculiar kinds of entities; but I do not expect nuclear physics, zoology, formal
semantics or art history to pay much regard to that”, “Moral reality and the end of desire”, in Platts
(1980, 72).
Griffin (1996, 44).
McDowell (1998, 177, footnote 19).
McDowell (1994, 21).
McDowell seems to concede this point when he tells us that

The remoteness of the Form of the Good is a metaphorical version of the thesis that
value in not in the world, utterly distinct from the dreary literal version that has
obsessed recent moral philosophy. The point of the metaphor is the colossal difficulty
of attaining a capacity to cope clear-sightedly with the ethical reality that is part of our
world. Unlike other philosophical responses to uncodifiability, this one may actually
work towards moral improvement; negatively, by inducing humility, and positively,
by an inspiring effect akin to that of a religious conversion.

21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39

He adds in a footnote that this view of Plato is beautifully elaborated by Murdoch, “Virtue and
Reason”, in McDowell (1998, 73).
McDowell (2009, 218).
But see the caveat in footnote 20.
See, for example, “The Ontological Proof ”, in Murdoch (1992, 419–425).
Murdoch (1992, 55).
“The Ontological Proof ”, in Murdoch (1992, 419–425).
“The Ontological Proof ”, 426.
“The Ontological Proof ”, 426.
Robinson (1963).
Robinson (1963, 32).
Systematic Theology, Vol. II, 8. Quoted in Robinson (1963, 34).
Robinson (1963, 34).
Robinson (1963, 30).
Robinson (1963, 32).
Rahner (1978, 61).
Fiddes defends this conception of God in Fiddes (2002, 35–60). His more recent Fiddes (2017) raises
some important issues for the epistemological points I shall be raising.
Fiddes (2002, 55).
See Cottingham (2018, 39).
“Creation”, in McCabe (1987, 6).
The wise man in question is John McDade who made this point at the end of his wonderful little paper
McDade (2001).

225

Fiona Ellis

References
Bernstein, R. Whatever Happened to Naturalism? Proceedings and Addresses of the American Philosophical
Association, 69(2), (1995).
Cottingham, J. Transcending Science: Humane Models of Religious Understanding. In F. Ellis (ed.), New
Models of Religious Understanding. Oxford University Press. (2018).
De Caro, M. and D. Macarthur, D. Introduction: The Nature of Naturalism. In Naturalism in Question.
Harvard University Press. (2004).
Dupré, J. How to be Naturalistic Without Being Simplistic in the Study of Human Nature. In M. De
Caro and D. Macarthur (eds.), Naturalism and Normativity. Columbia University Press. (2010).
Ellis, F. God, Value, and Nature. Oxford University Press. (2014).
Fiddes, P.S. The Quest for a Place Which Is “Not-a-Place”: The Hiddenness of God and the Presence of
God. In O. Davies and D. Turner (eds.), Silence and the Word, pp. 35–60. Cambridge University
Press. (2002).
Fiddes, P.S. ‘God is love: love is God’. A cutting-edge issue for the theology of love. 2017. https://
loveinreligionorg.files.wordpress.com/2017/02/fiddes-god-is-love.pdf
Griffin, J. Value Judgement: Improving our Ethical Beliefs. Clarendon Press. (1996).
Krikorian, Y.H. (ed.) Naturalism and the Human Spirit. Columbia University Press. (1944).
McCabe, H. Creation. In God Matters. Continuum Press. (1987).
McDade, J. Heaven, Then and Now. New Blackfriars, 83(971), (2001). 42–48.
McDowell, J. Mind and World. Harvard University Press. (1994).
McDowell, J. Two Sorts of Naturalism. In Mind, Value, and Reality. Harvard University Press. (1998).
McDowell, J. Response to Charles Lamore. In N.H. Smith (ed.), Reading McDowell on Mind and World.
Routledge. (2002).
McDowell, J. Reply to Fink. In J. Lindgaard (ed.), John McDowell: Experience, Norm, and Nature.
Blackwell. (2009).
Murdoch, I. Vision and Choice in Morality. In Proceedings of the Aristotelian Society, Supplementary
Volumes, Vol. 30, pp. 32–58. Dreams and Self-Knowledge. (1956).
Murdoch, I. The Sovereignty of the Good. Routledge. (1970).
Murdoch, I. Metaphysics as a Guide to Morals. Chatto and Windus. (1992).
Plantinga, A. Where the Conflict Really Lies: Science, Religion, and Naturalism. Oxford University
Press. (2011).
Platts, M. (ed.) Reference, Truth, and Reality. Routledge. (1980).
Randall, J.H. Epilogue: The Nature of Naturalism. In Y.H. Krikorian (ed.), Naturalism and the Human
Spirit. Columbia University Press. (1944).
Rahner, K. Foundations of Christian Faith: An Introduction to the Idea of Christianity. Trans. William V. Dych.
Darton, Longman, and Todd. (1978).
Robinson, J. Honest to God. SCM Press. (1963).
Stroud, B. The Charm of Naturalism. In M. De Caro and D. Macarthur (eds.), Naturalism in Question.
Harvard University Press. (2004).

226

19
TAYLOR AND LIBERAL
NATURALISM
Nicholas Smith

1 Introduction
Just as some cousins resemble each other more than they do members of their immediate family,
so there are liberal naturalists who appear to have more in common with certain relatives than
they do with themselves. Compare two of the most prominent self-conscious proponents of
liberal naturalism (LN), Richard Rorty and John McDowell, with one of the most prominent
self-conscious proponents of philosophical hermeneutics, Hans-Georg Gadamer. Both Rorty
and McDowell are impressed by Gadamer’s formulation of the thought that truth cannot be
reduced to procedures, and in particular by his account of why the objective procedures of
modern science are partial or incomplete in their delivery of truth (Rorty 1980; McDowell
1994; Gadamer 1993 [1960]). But whereas Rorty takes the partiality or incompleteness of
scientific truth, as shown by Gadamer, to amount to the limited utility of the scientific method,
which itself is quite heterogeneous, McDowell takes it to reflect the broader scope of truth as a
goal of enquiry, a goal not to be tied exclusively to the codifiable procedures of the modern
sciences, but which is shared by other more or less formalisable discourses and forms of understanding. Whereas Rorty reads Gadamer as a protopragmatist showing us how to get by
without a notion of “the Truth”, McDowell reads him as a neo-Aristotelian showing us how to
find truth in places the codifiable methods of the modern sciences cannot reach. Whoever is
right on this matter, most readers of Truth and Method would agree that McDowell’s views on it
are much closer to Gadamer’s than they are to Rorty’s. From this angle at least, McDowell
resembles his hermeneutic cousin more than his liberal naturalist sibling.
Gadamer is not the only member of the hermeneutic family who could easily be mistaken
for a liberal naturalist. Another is Charles Taylor. All the distinctive features of LN seem to be
there in Taylor’s writings. Taylor wholeheartedly endorses the liberal naturalist conception of
reality as containing more kinds of thing than the kind posited by successful scientific explanations. Indeed, Taylor is one of the foremost critics of the opposing thesis, the reductive
naturalist one, that the only “really real” things are those of a kind that are posited by natural
science. Not only are we led to posit the existence of many more kinds of thing than a physicist
or biologist would typically posit in seeking to understand and explain human affairs, Taylor
argues, but also the attempt to restrict our “posits” to those of a kind a physicist or biologist
would accept can distort our understanding of the phenomena at hand and blind us to the best
DOI: 10.4324/9781351209472-23

227

Nicholas Smith

explanations of them. Taylor is famous for his view that we should allow our ontology, the
things we take to be real, to be guided by our “best account” of the phenomena, without any a
priori restriction on the kind of concept or type of vocabulary that should feature in that
account (Taylor 1989). If, in accounting for human actions and the course of human lives, the
best we can do is draw on moral concepts and common sense, we shouldn’t flinch at the
conclusion that moral concepts and common sense really do inform us about human reality. We
should allow our ontology of the human to be shaped by them. And throughout his work
Taylor has sought to show that the best accounts we can offer of many aspects of human life do
indeed need to be couched in broadly speaking moral concepts, concepts that relate us to “the
good”, and that without such concepts we would be all at sea in attempting to understand and
explain these aspects of human life. Arguably, no one has done more than Taylor to expose the
inadequacy of attempts at explaining human history, politics, culture, religion and so on independently of such concepts; he is as consistent and vehement a critic of the folly of modeling
the humanities exclusively on the hard sciences as any avowed liberal naturalist (Taylor 1985a,
1985b). In his insistence on the equal epistemological standing of the humanities and the hard
sciences, and in this sense the compatibility between the “manifest image” projected by the
former and the “scientific image” constructed by the latter, again Taylor looks like a liberal
naturalist.
As a critic of reductive scientific naturalism (SN), Taylor’s liberal naturalist credentials are
impeccable. Taylor’s distinctive contribution to the liberal naturalist outlook, I will suggest, lies
largely in his distinctive critique of SN, the naturalism in opposition to which LN defines itself.
As just mentioned, Taylor shares the liberal naturalist repudiation of the reductive, monistic
epistemology and ontology of SN, and he endorses the more open-ended, pluralist conception
of knowledge and reality characteristic of LN. But Taylor has his own reasons for rejecting SN,
and in particular, he has his own account of why SN seems to “hold us captive” despite the
many criticisms that can be made of its central epistemological and ontological claims (Taylor
1995). Taylor is surely right in his view that, if the critique of SN is to be effective, it can’t rest
content with a list of that kind of naturalism’s errors and mistakes. It must also account for the
persistent appeal of SN, an appeal it has despite the shortcomings liberal naturalists and others
persistently draw attention to. We can call critique of the latter type, critique that does not just
criticise but also shows up the appeal of the criticised standpoint and the motivation behind it,
“genealogical” critique. Genealogical critique is typically historical because the motivation
behind a deep-rooted philosophical standpoint is typically deep-rooted historically: the motivation, the source of the appeal, is not a sudden, out-of-the-blue occurrence, but has come
from somewhere, it has a history. Taylor’s distinctive contribution to LN, I am suggesting,
consists largely in his genealogical critique of SN, in his attempt to historicise the naturalism in
opposition to which LN defines itself. We will return to this critique shortly.
But LN does not just define itself in opposition to SN: it is an opposition to SN which is also
opposed to “supernaturalism”. The liberality of the liberal naturalist allows her to admit into her
ontology the kind of thing we refer to when talking common sense, competently doing humanities subjects, or making and criticising works of art (De Caro and Macarthur 2004). But the
liberal naturalist acknowledges the need for some border control: there is no place for “supernatural” entities in her ontology. Likewise, while the liberal naturalist is by disposition
epistemologically relaxed, and comfortable with affording epistemic status to a variety of discourses and practices in the humanities and the arts as well as the natural sciences, she stiffens up
when it comes to superstition or practices that systematically violate the principles of modern
natural science. The supernatural and the superstitious are as much out of bounds for the liberal
naturalist as they are for the scientific naturalist. The same holds for philosophers in the
228

Taylor and liberal naturalism

hermeneutic tradition such as Taylor and Gadamer. Philosophical hermeneutics is in no way
motivated by an urge to reinstate the enchanted conception of nature that preceded the
Galilean transformation of science, as if the veil of disenchantment could be lifted by some
epistemological conjuring trick, allowing the world of spirits and demons to reveal itself once
more. As indicated above, philosophical hermeneutics, as developed by Gadamer and Taylor
amongst others, is a form of scientific realism, which means that it takes reality to be (in part)
what the most successful sciences of nature, namely the post-Galilean ones, show it to be
(Dreyfus and Taylor 2015). But in not ruling out enchantment by other means, in embracing
realism not just about science but about the humanities, the arts, ethics and even forms of
religious self-understanding, Taylor might appear to be letting the “supernatural” in through
the back door, so to speak, and in that light might look like a covert “supernaturalist”. The
worry is that while Taylor’s attempt at “retrieving realism” looks fine and consistent with LN
when it comes to the claims of the empirical sciences and common sense, it is too accommodating of supernaturalism, and perhaps superstition, when it comes to other kinds of claim,
such as general philosophical claims about the self or the nature of human agency, and those
pressed in works of moral, artistic and religious imagination where what Taylor calls “constitutive goods” and “moral sources” are at stake (Taylor 1989, 2011).
The suspicion that Taylor’s hermeneutics is a form of covert supernaturalism, and thus not a
form of LN, is no doubt reinforced by Taylor’s many avowals of his Catholic faith since the
publication of Sources of the Self (Taylor 2007, 2011). And this might even lead us to question
the bona fide character of the critique of SN conducted in that book, which I am suggesting is
Taylor’s most distinctive contribution to LN. If, as is now apparent to everyone, Taylor’s
genealogy of SN had a retrieval of theistic moral sources in view, does that not compromise it
from a liberal naturalist standpoint? And more generally, does Taylor’s alternative to SN,
considered independently of his Catholic confession, show tell-tale signs of an underlying
supernaturalism?

2 Historicising naturalism
To answer these questions, let’s first consider the thrust of Taylor’s genealogical critique of SN.
As noted above, genealogical critique targets the underlying appeal of a philosophical outlook,
an appeal that need not be diminished by well-considered objections to the outlook’s distinctive
epistemological and metaphysical claims. Such critique becomes apt when the original motivation for the outlook is forgotten, or when the outlook becomes so culturally entrenched that
alternatives to it, which must once have existed, become difficult to formulate. Taylor believes
that SN is such a culturally entrenched outlook, the original motivation for which has been
forgotten even – indeed especially – by its most ardent defenders. The point of the critique is to
retrieve, by way of a historical reconstruction, the original appeal and underlying motivation
behind SN. If successful, the critique will show that the appeal of SN properly comes to view
only once the standpoint of SN is abandoned. Since LN understands itself as the most viable
alternative to SN – the rival to SN that doesn’t rely on supernaturalism – such an historicising
strategy looks well-suited to the liberal naturalist project.
It seems clear that the appeal of SN has a lot to do with the power and prestige of modern
natural science and technology. One might challenge this or that epistemological or metaphysical thesis put forward by the scientific naturalist, but one cannot argue with the success of
modern science or the effectiveness of modern technology. If, in challenging SN, one seems to
put in doubt the success and effectiveness of science and technology, one has already lost the
argument. This is part of what it means to say that SN is deeply entrenched in the culture of
229

Nicholas Smith

modernity, the culture that provides the background for both SN and LN. But while the power
and prestige of modern science and technology is beyond doubt and beyond dispute, the exact
source of that prestige and the reasons for its emergence are more open to question. What
exactly is it about modern science and technology that commands esteem? What was the
character of the achievement involved in the transition from a pre-modern to a modern scientific
outlook? The obvious answer is that science and technology enabled human beings to take
better control of their environment, to transform their environment in a way that made it better
suited for the realisation of human purposes. The more scientific understanding developed, the
more effectively human beings were able to interact with nature and get the things they wanted
from it. The heightened prestige of the sciences would then be explained by the heightened
power over nature the sciences made possible, an increase in power indicative of cognitive
gains. The primary achievement in the transition to the modern scientific outlook, understood
this way, is cognitive/technical (Habermas 1984).
This is the way that scientific naturalists themselves tend to answer the question of the source
of the prestige of modern science. Liberal naturalists of a pragmatist persuasion (such as Rorty
and Habermas) share that tendency. The answer is congenial to champions of both those
standpoints because it does not invoke any suspicious supernatural entities: the purposes served
by science are continuous with those revealed by science. In the case of SN (but not LN), that
continuity is further explicated in terms of the reducibility of the former to the latter. The
prestige of science is thus itself part of the natural order as described by science. Now Taylor has
no argument with the view that science does vastly increase the “recipes for action” available
for humans in their dealings with their environment (Taylor 1985b). For him, it does indeed
represent a gain in the cognitive/technical dimension. What he takes issue with is the view that
this suffices to explain the prestige of science. For Taylor, the prestige of science and technology
is more fundamentally a matter of the admiration in which it is held, which is to say the
achievement of spirit it represents. To use Taylor’s terminology, the emergence, rise and
eventual triumph of post-Galilean science and modern technology is part of a broader cultural
transition, one marked by a change in “strong evaluations” and “orientation to the good”.
The story of this change is recounted in the central chapters of Sources of the Self. The main
message is that the adoption of a disengaged stance towards the world, the stance required for
arriving at successful scientific explanations (those of the post-Galilean sort) was in the first
instance a matter of moral/religious reform, a means of getting things right in a moral or
spiritual sense. The scientific outlook later championed by the Enlightenment, Taylor shows,
had its roots in this moral/religious reform, and progressed in stages through various forms of
rationalised Christianity and Deism. At the time of its emergence, scientific naturalism – or
what Taylor calls “Enlightenment naturalism” – was seen as a step in the moral evolution of the
species, not just as a method for getting more things done, being more technically proficient, or
meeting morally neutral human ends. The “instrumentalisation” and “disenchantment” of the
world, which the scientific stance epitomised, was not just useful for realising purposes human
beings shared with other natural beings, “natural” purposes such as a need for food and prevention of disease. It also realised a higher purpose, a capacity of self-control and selfresponsibility, that distinguished humans from other parts of nature, but which hitherto lay
dormant or inadequately developed. By developing this capacity human beings stood to realise
more fully their dignity as self-responsible agents. Furthermore, the capacities unleashed
through science and technology could be utilised to ameliorate the condition of humankind
generally, by eliminating hunger, disease and so on. Indeed, it was a duty, well-ensconced in
the Protestant ethic, for each human being to do what they could through hard work and selfdiscipline to promote that end. In this way, the scientific outlook and technological prowess get
230

Taylor and liberal naturalism

interwoven in what Taylor calls the modern “identity”, that is, the orientation to the good that
came to dominate the culture of modernity, and it is above all owing to this that science and
technology have prestige in that culture.
If Taylor is right, then SN, the sort of naturalism LN seeks to correct, is ultimately driven by
a historically contingent conception of what it is to be a fully human agent. The genealogical
critique serves as a reminder of this. We need to be reminded of it because SN lacks “articulacy” about its own motivation. It presents itself as driven by purely epistemological and
metaphysical considerations, by a conception of knowledge and reality modeled exclusively on
the modern natural sciences. Within this model, the concepts of strong value, distinctions of
worth (for example, between the admirable and the contemptible), an orientation to the good,
and so forth have no application. But these are the concepts we need to understand the underlying appeal of SN, concepts that flesh out what it means to be a fully human agent. In order
to grasp the appeal of SN, we thus need to abandon the standpoint of SN. We should adopt
instead a standpoint from which the central features of human agency appear for what they are.
And from that standpoint, SN will appear as one interpretation amongst others of what it means
to be a full human agent, one with a history and various paths of development.

3 Philosophical anthropology
The task of bringing the central, historically enduring features of human agency clearly into
view falls to what Taylor calls “philosophical anthropology”. Liberal naturalists have used the
term “anthropology” in a variety of ways, but in Taylor’s usage philosophical anthropology
refers to reflection on the constitutive features of human reality in its full scope. The constitutive features of human reality are the features that make human reality the kind of reality it
is. Possession of a soul, a mind, self-consciousness and free-will have been proposed as such
human-making features. But, not unlike many scientific naturalists, Taylor rejects such “ontic”,
substance-like proposals. Instead, he follows the hermeneutic tradition in taking human reality
to be what it is on account of the full range of capacities human beings have to make sense of the
world. Human reality is the reality of beings who confront “inescapable questions” about the
meaning of that reality. This gives rise to a conception of the human that is at odds with SN,
but what about LN?
First, let’s consider human perceptual capacities. Taylor’s view, drawn from Merleau-Ponty,
is that perception is fundamentally a capacity for orientation within an environment containing
items with various degrees of significance. In this respect, human perceptual capacities resemble
the perceptual capacities of other animals. Like other animals, humans must respond appropriately to the opportunities and obstacles provided by the surrounding world. They must cope
with the world they find themselves in and the exercise of perceptual capacities is crucial to this.
The world first appears as a perceptual field, with a foreground and a background, a “within
reach” and an “out of reach”, invitations to explore and warnings of danger, things to be
equipped with and things “in the way” and so on. It is by coping with an environment as it is
disclosed in a perceptual field that humans first acquire knowledge. But this knowledge is not
merely their representation of reality; it is their reality insofar as they have to cope with it.
Furthermore, in Taylor’s view, human knowledge and human reality never fully loses its
character as embodied coping.
Why is this view at odds with SN? First, because it undermines the key epistemological
claim of SN, that the scientific method is the one true source of knowledge. Perception qua
embodied coping informs us about the world, it yields knowledge, and indeed stands in the
background of all subsequent objective, scientific enquiry. That it provides such a background
231

Nicholas Smith

undermines another of SN’s epistemological claims, namely that scientific knowledge is selfsufficient. If objective knowledge of the kind science provides makes sense only against a
background of pre-objective perceptual understanding, the realm of science must be limited,
finite and dependent on something else. But because perception qua embodied coping is not so
much a way of representing a world as a way of inhabiting one, because human reality becomes
what it is through immersion in a field of significances, the view also undermines a central
ontological claim of SN, namely that reality is exhausted by the posits of successful scientific
explanations. Note that none of these reasons involves recourse to supernatural entities, superstitious beliefs, or violations of scientific principles. On the face of it, then, this aspect of
Taylor’s philosophical anthropology is compatible with LN. But it is only compatible with
those versions of LN that do not respond to SN by (1) positing some special ontic feature such
as mind, free-will or self-consciousness; (2) assimilating the pre-objective world inhabited by
embodied agents to the object-world as it appears either to science or reflective common sense;
(3) restricting the world-disclosive function of perception to the exercise of conceptual capacities
on the part of the perceiver; or (4) denying a world-disclosive function to perceptual experience
at all.
If human beings first come to grips with reality through perception and embodied coping, if
they first come to know and inhabit reality in this way, they also have a range of other capacities
at their disposal for extending this knowledge and deepening their responsiveness to the demands reality makes on them. In particular, human beings have linguistic capacities which,
when exercised in their full range, distinguish human beings from other animals. Taylor does
not deny the existence of linguistic capacities in some nonhuman animals; indeed, for Taylor
the continuity between the linguistic powers of nonhuman and human animals is a crucial fact
about them and something that any plausible philosophy of language must account for. It is
undeniable that many species have developed powers for intra-species communication as a
mechanism for survival, and it may also be true that some species have evolved powers to
communicate for its own sake, or for ends not directly related to the perpetuation of the species.
But the ends served by the exercise of linguistic capacities in the human case, while including
the exigencies of survival, are far more extensive. That SN draws the limits of these powers in
the wrong place, and explains them in the wrong way, is a point of agreement between Taylor
and LN. But Taylor has a different view to many liberal naturalists about the right way to
understand them.
One major point of difference has to do with the role of truth-conditional theories of
meaning. The idea that meaning is essentially a property of sentences, tied essentially to the
conditions under which sentences are true (facts that are grasped when the sentence is understood), is central to the approach to language taken by some of the most prominent advocates of LN (such as Rorty and McDowell). The approach is congenial to liberal naturalists
because it respects the autonomy of the “semantic”– the irreducibility of meanings to natural
causes, of logic to natural psychological processes – without positing a realm of meaningcontents independent of nature. It thereby offers an alternative both to SN and supernaturalism.
But in Taylor’s view it remains encumbered by problematic assumptions that can be shared by
SN and supernaturalism alike. Central amongst these is the idea that linguistic meaning is
fundamentally a matter of “designation” (Taylor 1985a, 2016). In their earlier iterations, theories of meaning based on this assumption took the vehicle of designation to be the word (the
word that names) and the object of designation to be separately existing objects, items like tables
and trees. The fundamental linguistic capacity was thus thought to be the power to describe a
nonlinguistic, independently existing reality. Although truth-theoretic semantics takes a different view of the vehicle and object of designation, it remains committed, Taylor argues, to the
232

Taylor and liberal naturalism

primacy of the designative function. That the primary designating vehicle is the sentence rather
than the word, and the primary designated item is the fact rather than the object, does not
fundamentally alter the structure of the theory. It does not mark a fundamental departure from
earlier theories that took the primary linguistic capacity to be that of designation.
What’s wrong with such theories? Two things, in Taylor’s view. First, they define the
human linguistic capacity too narrowly, so that the full shape of this capacity becomes distorted.
They screen out the diversity of language forms, such as body language, what Taylor calls the
“footings” by which human relations are established and maintained, ritual, music, narrative,
metaphor, and generally what he calls non-assertoric “portrayals” (Taylor 2016). These are just
as much instances of the linguistic capacity at work as are the making of assertions, the utterance
of grammatically well-formed sentences, the formulation of literal truths, the construction of
scientific theories and so forth. Second, the capacity the designative theories do focus on and
identify as first in the order of explanation concerning meaning is a “late arrival” that presupposes the development of more basic meaning-making powers. The capacity to name an
object correctly and to describe a fact accurately, which undoubtedly human beings possess and
exercise when uttering common-sense truths as well as formulating successful scientific theories,
realises a potential inherent in a more basic capacity for linguistic disclosure. The exercise of this
capacity is simultaneously a “bringing about” and an “opening to”, the production of something meaningful as well as a response to something meaningful. For example, it is only once a
feeling is linguistically expressed in a certain way, say through a dance or a piece of music or a
poem, that its full meaning becomes manifest. The expression shapes the feeling but is also
responsive to it. And for Taylor this isn’t just true of internal states. The meaning of the order in
which human beings are set, the larger reality they are a part of, also has this character. It has a
meaning that can be disclosed, accessed and understood only through productive acts of linguistic expression. Such disclosure is not the business of the “fact-establishing” discourses of
science and common sense. Rather, as the founding figures of the Romantic movement were
the first fully to appreciate, it is the calling of poetry.

4 Moral sources and supernaturalism
Taylor takes seriously the thought that success in the formulation of a poem, the composition of
a piece of music, or the production of a work of art, can be just as much a matter of articulating
a truth, or responding appropriately to reality, as success in the construction of a scientific
theory can be. A successful work of art, on this view, is a work of understanding, a bringing into
view of something that would otherwise remain hidden. The Romantic poets are a key reference point for Taylor because they showed how the productive imagination can be mobilised to bring into view something that remains hidden or obscured in the world as
represented by the modern natural sciences. What goes missing in the world as represented by
modern natural science, so the Romantics thought, is nature revealing itself as a “moral source”.
A moral source is Taylor’s term for a reality contact with which, or reflection on which, moves
us, and in moving us helps us to realise some higher, more fully human end. We get a prereflective apprehension of nature as a moral source when we feel elevated at the sight of a
beautiful landscape, fulfilment in an experience of wilderness, and the like. To have a sense of
wholeness through a connection with nature of this sort is to have a sense of nature as a
constitutive good, contact with which, or connection with which, brings about a good in me
(my sense of wholeness). In such moments, a certain empowering attunement with nature
seems to be manifest. Can that sense of empowering attunement and connection with something beyond me withstand reflection? It can, the Romantics thought, but only through
233

Nicholas Smith

portrayals of nature that resonate within the individual and resist the neutralising, reifying
tendencies of reflection.
The Romantics were not the only ones to think that art provided a locus for the manifestation of moral sources. And it is not only nature that has been thought to be disclosed as a
moral source in art: God and human powers of self-responsibility and creative imagination have
also featured as “frontiers of exploration” (Taylor 1989). In Taylor’s view, the exploration in art
of God, nature and human powers as moral sources is no less valid a form of enquiry than the
exploration of nature through scientific observation and theorising. Of course, the results of
these searches for understanding do not have the same kind of validity. A successful portrayal of
nature as a moral source in a work of art, for example, does not establish a fact; it does not
possess or aspire to objective validity. But neither is it just subjective fancy. It is answerable to
something: nature insofar as it is able to move us and constitute a higher or fuller mode of life.
Does this view of the powers of linguistic disclosure, powers that stretch to the manifestation
of moral sources, entangle Taylor in supernaturalism, and hence mark a departure from LN? I
think it does – certainly it’s a view that liberal naturalists such as Rorty and Habermas cannot
endorse (Rorty 1998; Habermas 1993). But it also poses a challenge to LN, for the following
reason. We saw that LN makes a point of granting art and art criticism a place in its ontology
and epistemology. The motivation behind this move is to retrieve a sense of the proper significance of art from subjectivistic and trivialising construals of the “aesthetic” that typically
accompany SN. Taylor’s understanding of art as a locus of moral sources shares that motivation.
But in the LN version, the powers of art are constrained from the outside by an ex ante
exclusion of the supernatural. There is no place for the supernatural or for mystery – this is what
makes it a form of naturalism. But if, by imposing this constraint, LN nullifies our sense of the
mysteriousness of the world as disclosed by art, and indeed of the world as it may be prereflectively experienced, then it is at risk of reproducing the very subjectivistic interpretation of
art that motivates its repudiation of SN in the first place. It may be that in order for LN to carry
through on its commitment to the ontological purport of art it may have to drop or modify its
commitment to antisupernaturalism.

References
De Caro, M. and Macarthur, D. Naturalism in Question. Harvard University Press. (2004).
Dreyfus, H. and Taylor, C. Retrieving Realism. Harvard University Press. (2015).
Gadamer, H.G. Truth and Method. Trans. J. Weinsheimer and D. Marshall. Sheed and Ward.
([1960]/1993).
Habermas, J. The Theory of Communicative Action, Vol. I. Trans. T. McCarthy. Beacon (1984).
Habermas, J. Justification and Application. Trans. C. Cronin. Polity. (1993).
McDowell, J. Mind and World. Harvard University Press. (1994).
Rorty, R. Philosophy and the Mirror of Nature. Blackwell. (1980).
Rorty, R. Truth and Progress. In Philosophical Papers 3. Cambridge University Press (1998).
Taylor, C. Human Agency and Language. In Philosophical Papers 1. Cambridge University Press. (1985a).
Taylor, C. Philosophy and the Human Sciences. Philosophical Papers 2. Cambridge University Press.
(1985b).
Taylor, C. Sources of the Self. Harvard University Press. (1989).
Taylor, C. Philosophical Arguments. Harvard University Press. (1995).
Taylor, C. A Secular Age. Belknap Press of Harvard University Press. (2007).
Taylor, C. Dilemmas and Connections. Selected Essays Belknap Press of Harvard University Press. (2011).
Taylor, C. The Language Animal. Belknap Press of Harvard University Press. (2016).

234

20
CAN SELVES BE NATURALISED?
THE PROBLEM OF TEMPORAL
PERSPECTIVE
Patrick Stokes

1 Introduction
It’s been noted that many discussions of selfhood proceed without ever giving their reader a
clear sense of what the term “self” even refers to (Zahavi 2014, chap. 1) Yet one thing people
seem to agree on is that if scientific naturalism is true, anything that might answer to the name
“self” is in trouble. As Tom Whyman, discussing P.F. Strawson, has recently put it:
Since according to Strawson hard naturalists deny or denigrate the reality of whatever
cannot be successfully incorporated within or reduced to the terminology of the
natural sciences, they are liable to find themselves obliged to deny things like the
objectivity of ethical claims; freedom of the will; even the existence of the self. All of
these things are ‘manifestly’ real, to borrow and lightly bastardise a phrase from Wilfrid
Sellars: they are crucial components of the world as we find it. But scientistically
minded philosophers are notorious if not for denying the existence of these manifestly
real things outright, then at least for treating them with suspicion – indeed, this could
serve as a good general characterization of the bulk of the canonical problems of
analytic philosophy. (Whyman 2018, 4–5)
In fact, the history of the concept of the self from Descartes to the present, via the decisive if
divergent interventions of Hume and Locke, is sometimes presented as a process of progressive
naturalisation (e.g. Martin and Barresi 2000). The Cartesian subject, having been conceived as
an immaterial, extensionless substance, could never become a properly integrated part of the
naturalistic explanatory schemas of science that emerged in Descartes’ wake. It’s worth remembering though that this naturalisation wasn’t simply driven by a rejection of supernaturalism per se. Locke’s innovation was not so much to shift focus from soul-substances to
extended consciousness – after all, Descartes’ soul was primarily a res cogitans, a thing that thinks
essentially – as to recognise that immaterial substances even if they exist do no real explanatory
work in grounding the diachronic extension of personal identity, which depends instead on
extension of consciousness (Locke 1975). Souls first drop out of the philosophical inventory not
because they are supernatural, but because they are surplus to requirements.

DOI: 10.4324/9781351209472-24

235

Patrick Stokes

So far, this progression will sound very amenable to those committed to liberal naturalism.
As David Macarthur has argued, liberal naturalism rejects the supernatural; that is, it rejects
“commitments to entities (e.g. spiritual agencies) or forms of understanding that are neither part
of the manifest or the scientific image of the world” (Macarthur 2018, 10), and indeed it does so
in large part precisely because, like Locke with souls, such entities are explanatorily otiose
(Macarthur 2018, 10 n. 22).
With the arrival of more stringent forms of naturalism, however, selves and persons (two
terms Locke took to be coterminous) were thrown into a more precarious position. First, we
can argue that selves or persons are identical with other, less mysterious entities, such as individual human animals, or bodies, or brains, or brain stems. The “I” in “I am sitting here
writing this chapter” thus refers simply to a specific human animal (or body, or brain, or…), not
to anything more mysterious or nebulous. We can thus avoid outrageously expansive ontologies according to which there is a body and a human animal and a self sitting in this chair
typing these words. Beyond that, we could reduce talk of selves or persons further, to underlying subpersonal facts about psychological or physical states; the locus classicus being Parfit
(1984). Or, we could simply do away with the idea of selves or persons altogether (e.g. Giles
1997; Metzinger 2003).
Of course, these revisionary options are likely available for any macroscopic object. The
specifics may vary, but whether the universe contains any persons is not fundamentally that
dissimilar a question to whether the universe contains any cats or umbrellas or football teams or
nebulae. It’s also worth noting here that like these other entities, persons may well play important explanatory roles in various sciences – for instance behavioral psychology – regardless of
whether we take the entities studied by such sciences to be reducible to more basic entities
posited in more basic sciences. Hence we might well expect that the problem of whether a
scientific worldview could acknowledge the existence of persons looks, at first glance, like it is
simply one more instance of the sorts of “placement problem”, to use Huw Price’s (2011) term,
that liberal naturalists grapple with.
Yet there is, I think, nonetheless a unique challenge here for liberal naturalists. In what
follows I want to suggest that the way we speak about selves and persons is not unitary. We exist
as both selves and persons – a distinction I’ll explain below. Moreover, as the history of trying to
incorporate selves into metaphysics suggests, persons may be far more amenable to at least the
more liberal forms of naturalism than selves. In particular, I will argue that selves are not merely
essentially first-personal entities, but essentially present-tense ones. Their temporality is such
that attempting to speak of their persistence or transience amounts to a category mistake. And
that, in turn, raises doubts as to just how liberal a liberal naturalism would have to be in order to
admit selves to its ontology – or just how naturalist a view that did so would remain.

2 Persons as objects
The voluminous literature on personal identity is largely, though not entirely, built around
reidentification (albeit in a metaphysical rather than merely epistemic sense): what is it for person
x at time t1 to be the same person as person y at time t2? The answer one gives to this question
will be mutually dependent upon the definition one gives of what persons are. Pointing to the
intuitions derived from thought experiments about mind-swaps and the like, a neo-Lockean
will say that persons are essentially self-conscious subjects, and that their persistence conditions
consist in unbroken chains of psychological states. An animalist will insist we are each an animal,
and that “person” either picks out that animal or is a phase sortal applying to one stage of the
animal’s life (namely, the stage in which the animal is conscious and capable of the sort of self236

Can selves be naturalised?

awareness characteristic of neo-Lockean persons). The animalist will then claim that we persist
just so long as a human animal persists as a human animal. Both approaches then get bogged
down in fairly familiar problem cases, which we do not need to discuss here. Other options
include identifying persons with their bodies, with their brains, with some proper part of their
brains and so on.
A somewhat different approach instead asks about the characterisation of persons. In Marya
Schechtman’s terms, it asks “what it means to say that a particular characteristic is that of a given
person” (Schechtman 1996, 73). This approach is more responsive to practical identity concerns
than metaphysical ones, though some versions explicitly aim at a maximal version that is both
practical and metaphysical, securing both practical and literal identity (e.g. Rudd 2012;
Schechtman 2014). This “practical” approach to personal identity finds expression in “narrative” views of personal identity (Schechtman 1996), as well as the more Kantian approach of,
for example, Christine Korsgaard (2009). In some ways, these approaches are closer to Locke
than neo-Lockeans are: for Locke, self was a “Forensick [sic] Term appropriating Actions and
their Merit” (Locke 1975, 346), one intrinsically linked to moral imputability. Far from the
impersonal thought experiments about brain transplants and personal fission that characterised
twentieth-century analytic thought on personal identity, Locke’s concern was practical, moral
and ultimately, soteriological (see Stokes 2015, chap. 1).
Part of the reason for this diversity of approaches to what should, presumably, be a fairly basic
part of the furniture of our world, is just that the concept “person” does a very great deal of work,
and that finding one object that corresponds to all the functions we ascribe to personhood is
accordingly tricky. We want to say that the protagonist of one of those 1980s body-swap movies
has woken up in someone else’s body, and that our childhood friend is now tragically in a
persistent vegetative state. We want to say that Vladimir Lenin is no more and that you can visit
Lenin in his mausoleum any day except Monday and Friday. We want to say that I’m not the man
I used to be and that I’m liable for things I did decades ago. All these ways of talking about persons
involve contradictions, and whichever theoretical account of persons we settle on will rule some
uses in while ruling some out as metaphorical at best and nonsensical at worst.
This hasn’t stopped philosophers who might be counted as (or declare themselves to be)
liberal naturalists from welcoming persons into the fold. Lynne Rudder Baker, for instance,
adopted what she called a “big tent metaphysics” that includes not merely the objects accepted
by scientific naturalists, but those entities that are important to us. It is a metaphysics built on
the premise that “What we consider to be real should not be independent of what we consider
to be important. Else, why bother with metaphysics?” (Baker 2008, 10). This metaphysics
explicitly allows for a range of everyday entities to exist via relationships of constitution (Baker
2007), and in particular, allows for persons to exist (Baker 2000). Yet persons are a very particular kind of entity within the big tent: unlike all other entities, they have a first-person
perspective, something that makes a radical and indeed transformative difference.
First-person perspectives are sometimes claimed to present a unique form of “placement
problem”. Placement problems arise when we confront a phenomenon that is “both practically
indispensable in our everyday lives and yet apparently absent from the scientific image” and are
thus confronted with “the problem of how to find a “place” for such things […] within the
scientific image of the world?” (Macarthur 2018, 2). Persons, as we’ll discuss below, appear to
be just such an indispensable thing even from the perspective of science. But first-personal
properties refuse to act the same way as other properties do. To use Thomas Nagel’s (1986)
well-worn example, for instance, if I know every fact there is to know about the universe,
including every fact about every person within that universe, it appears there is still one further
fact that I want to know, and which cannot be incorporated into that universal inventory: which
237

Patrick Stokes

one of those persons is me. (I might know that one of them is “Nagel”, but knowing that
“Nagel is Nagel” is not the same thing as knowing that “I am Nagel”). Accommodating such
facts within the third-personal frame of science seems incredibly difficult. If Locke is right that
each of us is a being that is, crucially, “self to it self” (Locke 1975, 336), then as Baker puts it
starkly, “The question here is whether these first-person phenomena can fit into a non-firstperson ontology” (Baker 2013, 48).

3 Temporal perspective and reidentifiability
One thing we need to note here is that both the reidentification question and the characterisation question are essentially diachronic. The reidentification question asks about the identity
conditions of a person at two times. The characterisation question, less obviously, asks about
what unifies characteristics over time; the sort of practical orientation towards the concept of
personhood it embodies makes no real sense without persons persisting over time. This is not to
say that neither approach says anything about the synchronic identity conditions of persons;
indeed, they must. But both are also centrally concerned with tracking the persistence of
persons across time. We can describe these temporal trajectories without specifying the point in
time from which we do so: like “x at t1 is the same person as y at t2”, “the life of Napoleon
takes place between 1769 and 1821” is true regardless of when the statement is uttered (even if
this phrase were somehow uttered years before Napoleon was born, or during his life). A
statement like “The woman sitting on that throne was once a baby” can be recast in atemporal
terms (“The woman sitting on the throne in Westminster Abbey on 2 June 1953 is identical
with a baby born in Mayfair in April 1926”) without loss.
First-person perspectives, however, are intrinsically temporally indexical. We see the world,
unavoidably, from a present moment, which in turn makes some moments past and others
future, and time appears to “flow” such that future moments become present and then past.
Such indexicality has been taken to be both an irreducible feature of human existence and
something that liberal naturalism can accommodate:
Lived and intrinsic time appear to be a necessary condition of the first-person perspective and minimal selves, but this view need not be spookily invested in supernaturalism or buttressed to any special human capacity, reflective or otherwise […] As
such, it can be minimally ‘naturalized’ via biology, psychology, and other relevant
sciences, in a way that involves thinking within rather than outside of time. (Reynolds
2017, 118)
The irreducibility of this temporal dimension becomes apparent in certain phenomena of
egocentric concern which only make sense from specific temporal perspectives. For instance,
the same event may be the subject of fearful anticipation or relief that it is over. But it cannot be
both simultaneously; I can only feel relief if the painful event is in the past, and feel fearful
anticipation if it is in the future. From the implicitly third-personal (or rather impersonal)
viewpoint of scientific naturalism, notoriously, these temporally-indexical features become
invisible. Worse, as (Parfit 1984) shows, our temporal biases may in fact cause us to form
preferences that strike us as incorrigibly reasonable and yet from another perspective utterly
bewildering. For instance, in Parfit’s famous example, most of us would prefer to learn we’d
had a long, painful surgery yesterday than to learn we will have a shorter, painful surgery later
today (Parfit 1984, 165–166). But that preference makes no sense from a “side-on” perspective
(Stokes 2014, 2013) where we view our lives as a whole, from which we would be concerned
238

Can selves be naturalised?

to optimise the value of our life as a whole. Why? Because from that perspective, it is clearly
irrational, all things being equal, to prefer a life with more pain in it, and yet from here-andnow, such a preference seems irresistable. You might well insist here that all things are not equal
if we’re comparing past and future pains. But from the viewpoint of scientific naturalism, there
is no relevant difference to their status qua pains. Ten units of pain is a greater amount of pain
than two units, regardless of where in time the pain-events happen to fall.
If liberal naturalism can solve these temporal placement problems, it will be able to do so in
virtue of the fact that “now” can always be coordinated to a point in time on a fourdimensionalist/eternalist/block universe view of time. In McTaggart’s (1908) well-worn terminology, A-series descriptors (“now”, “yesterday”, “future” etc.) and B-series descriptors
(“before the 20 September 1997”, “three years after the Battle of Hastings” etc.) have common
referents. For any moment we pick out with an A-series term like “now”, there’s a (or rather an
infinite number of) corresponding B-series description(s). The question then becomes simply
whether A-series terms are unreal, or at best merely epiphenomenal upon B-series terms. A
liberal naturalist might insist that we need the A-series descriptors for various forms of explanatory activity, even if they disappear from the sub specie aeternitatis viewpoint of scientific
naturalism. But part of their claim to be taken seriously is at least in part grounded in their Bseries respectability, so to speak. It’s only insofar as they supervene upon B-series terms that the
more “subjective” A-series terms can be taken seriously at all.
Presumably, this also applies to entities such as persons as liberal naturalists like Baker present
them. First-person-perspectives may well seem unplaceable in the same way as temporal
properties like “now” are, yet they are also linked to (in Baker’s model, they help to compose)
entities that are locatable at specific times and places. Huw Price describes placement problems thus:
On the face of it, a typical placement problem seeks to understand how some object,
property, or fact can be a natural object, property, or fact. Ignoring for present purposes the distinction between objects, properties, and facts, the issue is thus how some
thing, X, can be a natural thing—the sort of thing revealed by science (at least in
principle). (Price 2011, 188)
Placement problems so construed presuppose a more-or-less stable “thing” that needs to be
placed. On what does the stability of this “thing” depend? It depends upon other stable things,
into which scientific naturalism threatens to decompose the objects we seek to place. Hence
mental states depend upon brain states or observed behaviors or reports of introspection,
temporal properties like “nowness” will depend upon moments in time to bear such A-series
predicates and so on. As Nagel (1986) notes, taking inventory of the world from a “view from
nowhere” might make it invisible which of all the persons in that inventory is me, and this is
precisely the sort of placement problem liberal naturalists will grapple with. But it’s precisely
insofar as “me” is anchored to a person, locatable and diachronically persistent, that the work of
placement can get going at all. (Whether that person can be losslessly reduced to subpersonal
facts is then a further question).
Putting first-personal and temporal indexicality together, consider a statement like “I look
back on my youth with regret and my future with fear”. To say this is to make a first-personal
and temporally embedded report, but it is one anchored in certain facts that can be viewed
atemporally and third-personally: facts about the persistence over time of a particular body, or
of a chain of appropriately connected mental states, or of the functioning of a human animal, as
per your theoretical commitments or philosophical intuitions. The argument between realists,
239

Patrick Stokes

reductionists and eliminativists about the self, which partly maps onto that between liberal and
scientific naturalists, is then about whether the personal statements (“I look back…”) simply
reduce down to these sub-personal facts or not.
The point I want to take from all this is that reidentifiability plays a role in even those
putatively liberal naturalist views of the self that take first-person perspectives seriously. The
view of persons they deliver is of reidentifiable, diachronically extended entities that have firstperson-perspectives “attached” to them somehow. The “here-and-now-me-ness” of a self, to
coin an admittedly cumbersome phrase, supervenes upon and tracks a particular set of intersubjectively accessible person-facts. Precisely how it does so may be unclear or controversial, but
that it does so is intrinsic to the project of trying to take persons seriously as part of a liberal yet
nonspooky ontology. To be part of that ontology, persons have to be locatable and reidentifiable things that persist through time. And that, I want to suggest, may point to a serious
problem for that project.

4 Selves and persons
To say that persons can be accommodated in a liberal naturalist ontology is in part to say that
persons are reidentifiable objects, whatever other perspectival oddities they may have. Yet far
from reconciling our subjective experience with the scientific viewpoint, this may look like
we’re simply shoehorning the phenomenology of temporal self-experience into a framework
where it doesn’t really fit. Indeed, that is one aspect of Husserl’s critique of naturalism: because
naturalism views humans as “filled-out extension”, “the duration of a man’s spirit is taken as an
objective duration” (Husserl 1970, 315–316; Moran 2013, 98). This is a problem because, as
Dermot Moran puts it, “the peculiar syntheses of our temporal consciousness are not taken into
account in the objectivist understanding of temporality in nature” (Moran 2013, 98).
There have of course been other important attempts to incorporate the significance of firstperson perspectives into a metaphysics of persons. Barry Dainton (2008) for instance, begins
with the first-personal datum that we cannot imagine getting out of our stream of consciousness: just try to imagine running ahead of it or stepping outside of it, and you’ll see such a
thing to be unimaginable. From this, he concludes that we persist so long as our stream of
consciousness does: that is, as long as a stream of overlapping experiential specious presents do.
But what happens then when we fall asleep? Do we cease to exist – and on the assumption that
entities can’t persist across gaps (i.e. for an object p to persist between t1 and tn there can be no
time between t1 and tn where p does not exist) does that mean a wholly new person comes into
existence every time we awake from dreamless sleep or anesthesia? To counter this, Dainton
comes to identify us not with our consciousness itself but with the subphenomenal base of those
moments of consciousness, which would have produced continuous consciousness were we not
asleep or drugged. In that way, what starts as a phenomenological attempt to explore personhood or selfhood ends up positing selves as physical objects with more-or-less specifiable
physical persistence conditions. In a similar way, Galen Strawson (Strawson 2009, 2005, 1999)
argues that we experience ourselves as first-personal subjects of experience, the thing that is
figured at the center of each moment of consciousness. But instead of determining subphenomenal persistence conditions for such an entity, Strawson concludes that their duration is
in fact vanishingly small. Instead of one self, we have an innumerable number of ephemeral
selves throughout the course of a human lifetime.
As I’ve argued previously (Stokes 2014) the problem with both these attempts to make selves
or persons metaphysically respectable is that they end up abandoning the first-person perspective: they turn, as it were, “side-on” and try to build a third-personal metaphysics of
240

Can selves be naturalised?

personhood out of first-personal materials. In the attempt to fit selves into an ontology, they
abandon what is intrinsic to the self as a phenomenally figured entity: it’s here-and-nowness.
Here, I’d suggest, we reach the point where it is useful to disambiguate the terms self and
person, in a way that has been suggested in various places, such as Johnston (2010). Take the person
to be the diachronically extended object of various sorts of judgments. It’s persons who we identify
and re-identify, persons to whom we attribute moral blameworthiness and praiseworthiness,
persons who carry debts, keep promises and maintain friendships. Regardless of whether we take
the continuity of persons to be psychologically or physically grounded (or something else, or a
combination), what is essential is their diachronic character. Persons can be spoken of both firstand third-personally. So when I say “Napoleon lived to be fifty-one”, “My arm hasn’t been right
since I broke it at the age of nineteen”, and “If you’re no better by tomorrow I’ll take you to do
the doctor” the referents are all persons (Napoleon, me, you, the doctor).
By contrast, the thing picked out by certain uses of “I” or “me”, as well as by Strawson’s
more artificial if less ambiguous “me*” and cognates (“I∗”, “mine∗” etc.), is not the diachronic
person. Rather, the referent is what I experience myself as being right now, the subject of this
experience – and it is this which I am designating a “self”. The self is disclosed in intrinsically
first-personal experiences, such as the Nagelian worry that a complete inventory of all the
people in the world doesn’t tell me which one is me, or realising that the disheveled figure over
there is in fact my own reflection – that I look like that (Baker 2013, 38). It also presents itself to
us as somehow independent of any of the persistence conditions we appeal to when we try to
reidentify persons, as Nagel notes:
When I consider my own individual life from the inside, it seems that my existence in
the future or the past – the existence of the same “I” as this one – depends on nothing
but itself. […] My nature then appears to be at least conceptually independent not only
of bodily continuity but also of all other subjective mental conditions, such as memory
and psychological continuity. It can seem, in this frame of mind, that whether a past or
future mental state is mine or not is a fact not analysable in terms of any relations of
continuity, psychological or physical, between that state and my present state. The
migration of the self from one body to another seems conceiveable, even it is not in fact
possible. So does the persistence of the self over a total break in psychological continuity
– as in the fantasy of reincarnation without memory. (Nagel 1986, 33)
Selves, as I’m using the term here, have just this sense of incorrigibility that Strawson describes,
the same sense, I’d suggest, at work in Dainton’s intuition that we can’t imagine “getting
outside of” our stream of consciousness. This sense of “here-now-ness”, so to speak, also allows
for certain forms of alienation, such as that described by Strawson in feeling no sense that the
person he remembers being in the past or anticipates being in the future is him∗. What’s missing
in the past and future case, at least in some exercises of memory and anticipatory imagination, is
precisely the sense of immediate presence that is constitutive of the self in this sense.
That immediacy has the effect of making selves, in this sense, intrinsically present-tense. Nowness is an irreducible part of the experience of self. We can only make something that is
intrinsically present-tense have duration, and thus be reidentifiable, at the cost of a fairly serious
category mistake (Stokes 2017). And that is precisely the error that bedevils both naturalist and
metaphysical accounts of the self: they turn the self into something countable and with an at
least theoretically measurable duration, even if, as in Strawson, that duration is vanishingly
small. If such selves cannot be spoken of as having duration, then, as Johnston (2010, 175)
argues, they cannot be reidentified; “these sortals fail to implicate any determinate conditions of
241

Patrick Stokes

success or failure when it comes to such things are believing that I, the same consciousness, or
self, or occupant of the central position in the same arena of presence, will be around in any
anticipated future”. For Johnston, that pretty much rules out “self” as a serious category.1 Yet
the incorrigibility of the self means this category is, if we take the phenomenology of selfhood
seriously, not so easily corriged. We come, then, to a critical decision-point for liberal naturalists. Should these selves be admitted into the liberal naturalist world-inventory? Can a
naturalism be liberal enough to incorporate entities that by definition cannot be reidentified?

5 Natural selves?
Let’s consider how card-carrying liberal naturalists have approached personhood. In recent
work, David Macarthur has distinguished between the “theoretical scientific image” and the
“practical scientific image”. The theoretical scientific image “includes the posits of the best
explanations in physics, chemistry, biology, psychology, sociology, economics, etc”. whereas
the practical scientific image “accepts all of this and, in addition, acknowledges persons (e.g.
scientists and laboratory assistants), ordinary artifacts (e.g. chairs, tables and experimental
equipment) and meaningful language […] The practice of doing science thus presupposes these
everyday phenomena and our ordinary relation to them”. (Macarthur 2018, 5). The sense that
these things are somehow incompatible with or superfluous to the scientific image – a sense that
largely underlies the familiar Sellarsian distinction between the manifest and scientific images of
the world – depends, according to Macarthur, on ignoring the realities of scientific practice, in
particular focusing on “the objects of scientific study and forgetting about the subjects conducting
these studies, their artifacts and their normative relations to each other” (Macarthur 2018, 6). In
other words, persons don’t, or don’t solely, earn their place in the liberal naturalist inventory by
virtue of their inclusion as objects of study in fields such as psychology, anthropology and the
other “human” sciences. Indeed, we might add that if those were the only grounds for their
inclusion, the scientific naturalist might reply that either the “persons” figured in these sciences
are at least theoretically reducible to more basic units (e.g. those of neurology) without explanatory loss, or, more cynically, that the sciences they figure in aren’t the “best” sciences.
Luckily for persons, then, they pay the cost of their admission in another way too: through
being necessary posits for the practice of science itself, not just its findings. Scientific exploration
and explanation do not happen purely on paper or in the abstract. Hypothesising, discovering
and knowing are all activities performed by persons.
So Macarthur thinks persons, just insofar as they are essentially agents, are a solid example of
an indispensable, nonsupernatural and yet nonscientific entity. Because persons are the sort of
entities to which we can and must ascribe reasons for acting, they cannot be reduced to the
purely causal descriptions of the sciences, even while they are a necessary feature of the practical
scientific image (science can’t happen without scientists deciding to do science). Drawing on
Putnam, he argues that persons are indispensable in the sense that “They are so deeply implicated in our worldview that they and our worldview would seem to stand and fall together –
which means it is hard to see how we could ever be rationally required to give them up”
(Macarthur 2018, 13).
Notice, however, the sort of entity we’re talking about here. Macarthur glosses
“person” thus:
A person is, minimally, an embodied rational agent acknowledged as such by others
whose actions are locatable in the logical space of reasons. It is appropriate to ask of a
person why they acted in the way that they did and expect a reason in reply – where a
242

Can selves be naturalised?

‘reason’ is a consideration in favor of so acting from the agent’s point of view. As
Locke famously recognized, ‘person’ is a forensic term, ‘appropriating actions and
their merit; and so belongs only to intelligent agents capable of a law, and happiness,
and misery.’ That is, it is essential to our concept of a person that their actions are
intelligible, to the extent that they are, from the point of view of reason in deontological terms such as responsibility, praise and blame. But, even so, a person also falls
within the logical space of scientific intelligibility in so far as persons are causal agents
who open doors, spill coffee, cause emotional reactions, contribute to GDP and global
warming and so on. (Macarthur 2018, 11)
This concept of a person is already, Macarthur acknowledges, a long way from anything a
biochemist or histologist would invoke (Macarthur 2018, 11). As the reference to Locke indicates though, this is quite close to the sort of account of persons we meet in practically
minded philosophers of personal identity such as Baker and Schechtman (and also Christine
Korsgaard, Peter Goldie and Alasdair Macintyre). Moreover, Macarthur’s persons are, taking
their cue from Peter Darwall (and ultimately Hegel!), constitutively second-personal, that is to say
intersubjective, entities: “To be a person is to have a certain self-conception as such and to be
taken or acknowledged as such by other persons and to be aware of that fact. One’s status as a
person is a matter of mutual acknowledgment” (Macarthur 2018, 16). As Schechtman (2014)
puts it, persons exist in, and are constituted in, “person-space”, a discursive and conceptual
space in which practices of personhood are applied, without this collapsing (as even many of
Locke’s early critics thought) into a mere fictionalism.
So persons, even the relatively thick and wooly persons favored by philosophers who view
person primarily as a moral and practical term (a la Locke), seem fairly unproblematic for liberal
naturalism. But what about selves as distinguished above?
Both persons as figured in the human sciences and persons as preconditions of scientific
practice are essentially diachronic. These selves are primarily agents, and as such their being is
necessarily stretched across time. To engage in scientific inquiry, for example, involves various
temporally-embedded attitudes and actions: looking at past observations to construct hypotheses that can be tested in future experiments, abandoning once-present hypotheses on the
basis of those same now-past experiments, and so on. None of this is possible for creatures that
lack duration in their very being. The self each of us experiences ourselves as being right here and
now slips through this description of agency. We may well experience ourselves as being such an
agent (“I am a scientist and I am looking at old data to generate new hypotheses I can test if I
ever get the funding!”) but the temporal specificity of our self-apprehension as such escapes the
sort of description of persons liberal naturalism is concerned with. What is essentially presenttense and first-personal for me now becomes just another object in the field of objects through a
second-personal, intersubjective view of persons. The center of the universe becomes just
another node.
Naturalists of different stripes might come to very different views on these sorts of selves. A
scientific naturalist might conclude that neither Strawson’s nor Dainton’s selves deserve any
place in a maximal ontological inventory. Strawson’s whispy, unlocatable selves, with durations
so short as to be unverifiable, are unavailable to any sort of empirical study and so a clear nonstarter. But Dainton’s selves, insofar as they appear to be reducible to a set of subphenomenal
facts (in the standard case at least, facts about brain states) are also surplus to requirements. What
would the liberal naturalist say? They might well concede Dainton’s selves to be useful, just
insofar as they answer to an entity that plays a role in the explanatory practices of, for example,
psychology or social science. Strawson’s selves, by contrast, might still fail to find a home. It’s
243

Patrick Stokes

hard to see what explanatory practices might need such short-lived entities, entities with no real
reidentification conditions. Nagel’s “I” would presumably fail to pass muster on similar
grounds. It’s hard to see how anything that is intrinsically non-reidentifiable could play a role in
such explanatory practices at all.
We still find ourselves, then, stuck with incorrigible features of self-experience that we can’t
seem to fit into what we take to be our best conception of the world. I’m not yet convinced this
gap is unbridgeable by liberal naturalists. But the gap does suggest that the gulf between naturalism and phenomenology will resist bridging in more places than we might have hoped.

Note
1

For a response to Johnston, see Baker (2013, 160–168).

References
Baker, L.R. Persons and Bodies: A Constitution View. Cambridge University Press. (2000).
Baker, L.R. The Metaphysics of Everyday Life: An Essay in Practical Realism. Cambridge University
Press. (2007).
Baker, L.R. Big-Tent Metaphysics. Abstracta, 4(1), (2008). 8–15.
Baker, L.R. Naturalism and the First-Person Perspective. Oxford University Press. (2013).
Dainton, B. The Phenomenal Self. Oxford University Press. (2008).
Giles, J. No Self to Be Found: The Search for Personal Identity. University Press of America. (1997).
Husserl, E. The Crisis of European Sciences and Transcendental Phenomenology. Northwestern University
Press. (1970).
Johnston, M. Surviving Death. Princeton University Press. (2010).
Korsgaard, C.M. Self-Constitution: Agency, Identity, and Integrity. Oxford University Press. (2009).
Locke, J. An Essay Concerning Human Understanding. Oxford University Press. (1975).
Macarthur, D. Liberal Naturalism and the Scientific Image of the World. Inquiry, 9, (2018). 1–21.
Martin, R. and Barresi, J. Naturalization of the Soul: Self and Personal Identity in the Eighteenth Century.
Routledge. (2000).
McTaggart, J.E. The Unreality of Time. Mind, 17(68), (1908). 457–474.
Metzinger, T. Being No One: The Self-Model Theory of Subjectivity. MIT Press. (2003).
Moran, D. ‘Let’s Look at It Objectively’: Why Phenomenology Cannot be Naturalized. In H. Carel and
D. Meacham (eds.), Phenomenology and Naturalism, pp. 89–115. Cambridge University Press. (2013).
Nagel, T. The View From Nowhere. Oxford University Press. (1986).
Parfit, D. Reasons and Persons. Oxford University Press. (1984).
Price, H. Naturalism Without Mirrors. Oxford University Press. (2011).
Reynolds, J. Phenomenology, Naturalism and Science: A Hybrid and Heretical Proposal. Routledge. (2017).
Rudd, A. Self, Value, and Narrative: A Kierkegaardian Approach. Oxford University Press. (2012).
Schechtman, M. The Constitution of Selves. Cornell University Press. (1996).
Schechtman, M. Staying Alive: Personal Identity, Practical Concerns, and the Unity of a Life. Oxford University
Press. (2014).
Stokes, P. Will It Be Me? Identity, Concern and Perspective. Canadian Journal of Philosophy, 43(2), (2013).
206–226.
Stokes, P. Crossing the Bridge: The First-Person and Time. Phenomenology and the Cognitive Sciences, 13(2),
(2014). 295–312.
Stokes, P. The Naked Self: Kierkegaard and Personal Identity. Oxford University Press. (2015).
Stokes, P. Temporal Asymmetry and the Self/Person Split. Journal of Value Inquiry, 51(2), (2017). 203–219.
Strawson, G. The Self and the SESMET. Journal of Consciousness Studies, 6(4), (1999). 99–135.
Strawson, G. The self? Wiley-Blackwell. (2005).
Strawson, G. Selves: An Essay in Revisionary Metaphysics. Oxford University Press. (2009).
Whyman, T. P.F. Strawson’s Soft Naturalism: A Radicalisation and Defence. International Journal of
Philosophical Studies, 26(4), (2018). 561–581.
Zahavi, D. Self and Other: Exploring Subjectivity, Empathy, and Shame. Oxford University Press. (2014).

244

21
LIBERAL NATURALISM,
ONTOLOGICAL COMMITMENT
AND EXPLANATION
Matteo Morganti

1 Introduction
There has been a resurgence in the interest of philosophers in naturalism and related notions
recently. On the one hand, strong forms of naturalism have been developed which recommend
a restriction of the range of meaningful philosophical questions similar to that urged by neopositivists (see, e.g. Ritchie 2008 or Maddy 2007). In some cases, this takes the form of an
examination of science with a view to extracting philosophical conclusions directly from
contemporary scientific theories (so leading to a radical naturalisation of, e.g. metaphysics, as in
Ladyman and Ross (2007)). On the other hand, milder and more sophisticated forms of naturalism have been put forward, with the aim of finding a stable middle ground between antinaturalism, which pays basically no heed to science, and the strong naturalism just described.
The most popular of these nonreductive varieties of naturalism is no doubt the so-called “liberal
naturalism”, discussed and defended by several authors (see, e.g. Price 2004; De Caro and
Macarthur 2004, 2012; De Caro and Voltolini 2010; Macarthur 2019) based on insights of
Putnam (1990), Nagel (1986), McDowell (1994) and Davidson (2001) among others.1
Liberal naturalism is essentially based on the idea that the domain of the natural is not
exhausted by the sort of entities that are (or may be) inquired into by science, hence strong
naturalism – which liberal naturalists often refer to as “scientific naturalism” – is untenable.
Things like the mental, consciousness, agency, normativity and meaning, it is claimed, fully
qualify as natural, as they play a central role in our everyday experience of the world. Indeed, in
many cases they seem to be presupposed by science itself – if not conceptually, from the
viewpoint of scientific practice. Yet, these entities, according to liberal naturalists, are not
amenable to scientific analysis. Hence, there are admissible forms of rational inquiry – primarily
concerned with what Sellars (1962) called the “manifest image” of the world – that differ in
fundamental respects from science, and must be acknowledged as equally important even in a
naturalistic context.
While the leading thought and the rationale underpinning liberal naturalism is clear and prima
facie compelling, however, what exactly belongs to the domain of the “non-scientific-yetnatural” envisaged by liberal naturalists remains not entirely clear – or so I will argue in this paper.
When not simply invoking pluralism about possible descriptions of the world (which is hopelessly
vague), liberal naturalists emphasise two elements. On the constructive side, there is the
DOI: 10.4324/9781351209472-25

245

Matteo Morganti

abovementioned thought that some things that play a central role in our daily experience of the
world are not reducible to scientific entities and theories hence should be included in our fundamental ontology. On the “negative” side, liberal naturalists maintain that some entities and
processes are in conflict with science and also not part of our common sense view of the world,
hence should be discarded. But how should (ir)reducibility and/or (lack of) conflict be established? Also, do these two general criteria account for all relevant entities? In my opinion, more
needs to be said about the motivations underlying liberal naturalism as well as the criteria that are
to be employed for determining the liberal naturalist’s precise ontological commitments.
In view of this, here I will suggest ways of further elaborating on the liberal naturalists’ basic
insights (as I understand them, at any rate). I will submit that the truly interesting dichotomy is not
so much that between the natural and the supernatural but, rather, that between supernatural (i.e.
metaphysical) entities, properties, processes etc. that play a genuine explanatory role in connection to
current science2 and those that do not. The emphasis on explanation is, of course, not new: liberal
naturalism is explicitly aimed to explain our common sense experience. Yet, rather than specific
indications for assessing the explanatory power of the specific ontological hypotheses and commitments at hand, liberal naturalists typically argue – at a rather general level – that, in order to
have satisfactory explanations at all, one should include certain families of nonscientific items in
one’s hypotheses. On this basis, they then go on to providing putatively paradigmatic examples of
natural, supernatural and nonscientific-yet-natural entities. The problem with this is, I think, that
while it gives the impression that the relevant distinctions are based on clear-cut and fixed
boundaries, this is not the case. Upon scrutiny, more fine-grained distinctions turn out to be
required. Once these are in place,3 I will argue, a less “monolithic” treatment of nonscientific
entities becomes plausible: one according to which, in particular, not everything that belongs to
our common sense view of the world is dealt with in the same way and, moreover, more than the
common sense view of the world may plausibly be regarded as acceptable from the liberal naturalist’s perspective. In particular, at least some of the concepts and categories employed by
analytic metaphysicians need not be rejected as supernatural.
The plan of the paper is the following. In section 1, the main section, I will first discuss
liberal naturalism, suggesting problems with its current formulation and motivation. I will
propose a few amendments to the liberal naturalist viewpoint that, to my mind, make the
position stronger and less prone to criticism. In section 2, I will more briefly argue in favor of
the addition of at least a proper part of current analytic metaphysics to the liberal naturalist
conception. In particular, I will give a sketch of the recently proposed “moderate naturalism
about metaphysics” (Morganti 2013; Morganti and Tahko 2017), and argue that – although put
forward independently of liberal naturalism – it naturally completes the liberal naturalist perspective. In section 3, I will summarise the results of the discussion.

2 Liberal naturalism
Liberal naturalism is the view that, while it is correct to accept the authority of science and its
primacy in providing knowledge of the natural world, scientific ontology and methodology do
not exhaust the domain of what there is and the range of ways that are available to us for getting
to know what the world is like. To the contrary, there is more in conceptual space than the
extremes constituted by
(1) radical, strong, scientific naturalism – the view that everything is, at roots, made by the kind
of entities posited by scientific theories, hence analysable and explicable in purely scientific
terms, and
246

Liberal naturalism

(2) supernaturalism/anti-naturalism – the contrary view that there are forms of understanding and
existence that are not reducible to, and in fact are more fundamental than, scientific ones
(philosophy, religion or mysticism, for instance, consequently being in no way dependent
on, and less significant than, science).
Accordingly, liberal naturalists acknowledge the existence of a third domain of entities different
from both supernatural entities and natural entities intended as those certified by our best
current science. Some entities, they claim, are not scientific, as they do not appear in our
physics, biology, etc., yet are perfectly acceptable because
(1) they are not in conflict with science, and
(2) they are an essential part of our experience of the world, hence of nature.
Thus, liberal naturalism is clearly in disagreement with, for instance, Quine’s idea that our ontological inventory should be dictated by science and, ideally, philosophy should adopt the methods
of science. More specifically, this anti-Quinean stance is motivated by the thought that there is a
portion of reality that is not situated at the level of the third-person objectivity based on physical
causality sought by science, but rather at the level of first-person facts such as self-consciousness or
perceptual experience, and/or second-person facts related to intersubjectivity and mutual recognition
by agents in communities of humans.4 Not only do these latter entities and facts, argues the liberal
naturalist, belong to a domain – the abovementioned “manifest image”, which escapes scientific
reduction. They are in fact presupposed by science. For, scientific knowledge and, more importantly,
scientific practice, are based on there being a community of rational agents who find themselves in
particular mental states, share a meaningful language, act on the basis of certain shared norms and so
on. They are also based, more simply, on a prescientific conception of macroscopic objects such as
computers or tables. Macarthur, for instance, writes: “The practice of doing science […] presupposes these everyday phenomena and our ordinary relation to them – something that philosophy has a perennial tendency to dismiss or overlook” (2019, 569).
Some critics (most notably, Neta 2007) do not see space for a third option of this sort, and
contend that whatever is not part of science is supernatural. In particular, either something that
may at first appear nonscientific is in fact scientific, or it is not naturally acceptable (I will call this
“Neta’s dilemma” in what follows). Supporters of liberal naturalism typically react to Neta’s
dilemma by arguing that it presupposes what is being questioned, namely, a scientific conception
of naturalism. Something, they say, may play an explanatory role in a naturalistic context even
though it is not explicitly posited by science. Especially so in domains that seem to escape a
scientific reconstruction, such as those mentioned a moment ago, for example, those characterised
by values and norms, those having to do with intentional states and irreducibly subjective facts, or
those pertaining to language and meanings, etc. Even granting that there is a way out of Neta’s
dilemma, however, an important question remains concerning the criteria that liberal naturalists
have in mind for precisely defining the domain of the nonscientific-yet-natural.

2.1 The upward path: naturalness, presupposition and irreducibility
In the course of the foregoing discussion, it has emerged that the criteria for ontological
commitment employed by liberal naturalists are essentially two:
(1) something being presupposed by science, and
(2) something being irreducible to science.
247

Matteo Morganti

Liberal naturalists have often used these two notions almost interchangeably. And indeed both
aspects are involved in discussions of what is known as the “location problem” or “placement
problem”, that is, the problem of finding a place in a science-based picture of the world for
things that appear not to lend themselves to a purely scientific account. Yet, it seems to me that
there is a significant ambiguity, and a potential tension, here. It is one thing to say that some x is
presupposed by science and is irreducible; and it is another thing to contend that x is to be
considered irreducible because it is presupposed by science. There is a frequent tacit shift in the
literature from the latter to the former claim, and this tends to hide the fact that being presupposed by science simply does not entail being irreducible to the scientific ontology and
vocabulary: x may turn out to be fully analysable in scientific terms after all!5 Putting it slightly
differently, if irreducibility is at most probable, hence conjectural, it is not surprising that radical
naturalists typically argue that certain things may well be presupposed by science and the
manifest image, yet can be, or will eventually be, made sense of on the basis of science alone.
Liberal naturalists are right in pointing out that this means either reduction or elimination, but
only based on unwarranted philosophical presuppositions and expectations concerning future
science.uch analysis. However, lacking further argument, certainty concerning future reduction
is as much an article of faith as certainty of irreducibility. More generally, the set of things that
are presupposed by science, the set of things that belong to the manifest image and the set of
things that are not reducible to (currently accepted or purely hypothetical) scientific items are
not coextensive (let alone necessarily coextensive), and this is a fact liberal naturalists should pay
more attention to.6
In the light of this, I think the best move for the liberal naturalist is to qualify their claims in
two ways. On the one hand, claims of irreducibility should not just depend on the current state
of our scientific knowledge, but also, crucially, on the very nature of the phenomena at hand.
This intuition seems shared for instance by Putnam, when he claims: “I […] deny that the
world can be completely described in the language game of theoretical physics; not because
there are regions in which physics is false, but because, to use Aristotelian language, the world
has many levels of form, and there is no realistic possibility of reducing them all to the level of
fundamental physics” (Putnam 2012, 65, italics mine). On the other, arguments should be
provided to the effect that elimination is not a sensible option. Perhaps this is what liberal
naturalists had in mind all along, but it must be made explicit anyway. For, once it is done,
important distinctions emerge. Consider, for instance, the difference between the contents of
subjective conscious experience (“qualia”) and strong free will intended in the libertarian sense.
The possibility of unexpected, radical changes in our conceptual schemes notwithstanding,7 it is
plausible to think that qualia, because of their essentially subjective nature, are in principle irreducible to science and at the same time cannot simply be eliminated from our worldview. A
weaker, compatibilist conception of free will could instead be perfectly fine, in the sense of i)
being positively grounded in scientific results and ii) not entailing a(n equally) radical conflict
with our common sense perspective (obviously enough, I am not claiming that compatibilism is
true, but only using it as an example) . Thus, at least prima facie, qualia satisfy (1) and (2) above,
while free will may fail to satisfy (2).
If the above discussion went at least partly in the right direction, then it must be acknowledged that not all items belonging to the manifest image are on a par, and it is much wiser
to proceed on a case-by-case basis – in particular, by distinguishing cases in which ontological
reduction/elimination is (or appears to be) altogether impossible, and cases in which it is (or
may be) simply undesirable. In any case, a stricter definition of the relevant domains of entities,
processes etc. is in order.

248

Liberal naturalism

2.2 The downward path: delimiting the supernatural
Sometimes, liberal naturalists focus not so much on what should legitimately be included in the
domain of the natural, but rather on what should be excluded, hence regarded as supernatural.
The typical claim is that the supernatural comprises all the entities, processes, etc. that – in spite
of the fact that they are purported to explain certain facts – are in conflict with our best current
science. That science gets priority in cases of conflict is no doubt a claim on which liberal and
scientific naturalists agree – in fact, it seems to be the central tenet of naturalism. In this case too,
however, something more must be said, as mere talk of conflict won’t do.
First, conflict is relatively easy to recognise if a nonscientific hypothesis entails consequences –
be it in terms of existential claims or properties ascribed to things – that are logically incompatible
with established scientific “truths”. However, cases of straightforward contradiction between
scientific and nonscientific claims are rare and – from the naturalistic perspective – easily solved.
For example, the creationist theory according to which all existing animal species (and humans)
have been created by God a few thousand years ago, and have remained unchanged ever since, is
simply to be discarded, as Darwin’s evolutionary theory is currently regarded as the correct
scientific account of the origin of animal species, and the latter’s way of accounting for the
empirical evidence is squarely incompatible with that of creationism. In at least some (and I
suspect most) interesting cases, though, the perceived conflict is of a different sort. For instance,
clearly the claim that the Big Bang theory is true and God was the cause of the entire process is
not equally in conflict with the most popular theory in today’s cosmology, as it includes it as one
of the two components of a consistent conjunction. Halvorson (2016) goes as far as to say that a
creating God is fully compatible with naturalism,8 as it contributes to the fundamental aim of
providing coherent, general explanatory schemata. If the liberal naturalist disagrees with this, s/he
must provide reasons, and, obviously enough, simply stating that the explanans is not ok because
it is supernatural won’t do, the definition of the supernatural being exactly what is at stake.
One could claim that certain explanantia (God, in this case) are not acceptable from the
viewpoint of scientific methodology because they escape empirical verification. But this won’t
do either: for this also applies to items that liberal naturalists are keen to include in their
ontology, such as meanings, norms or values. Couldn’t I be wrong in thinking that there
literally are normative facts out there as much as my neighbor in thinking that there is an
omnipotent, omniscient and benevolent deity? The issue, then, concerns not so much the
things that do the explaining, but rather the nature of the explanations provided on the basis of
them. I believe the point here is best put in terms of explanatory usefulness/superfluousness, and
in particular in terms of three distinct requirements. A naturalist, I contend, should prefer not to
believe in the existence of x insofar as:
a
b
c

the existence of x is neither implied by science nor presupposed by it,
the existence of x is not a necessary component of the manifest image9 and
whatever explanatory role x is supposed to play, it only provides ad hoc additions to the
explanations already provided by science alone.10

I think that (a), (b) and (c) are individually necessary and jointly sufficient for picking out the
supernatural. Without entering a detailed discussion, let me just offer some scattered remarks
and examples in support of this claim: without (b) free will could arguably be regarded as
supernatural; and without (c), a Big-Bang-creating God would be ok. As for (a), it expresses the
basic tenets of liberal naturalism: if x is implied by science, it ipso facto goes into any naturalist’s
ontological inventory; if it is not implied but it is presupposed by science, then either it is
249

Matteo Morganti

amenable to reduction or it is at any rate nonscientific-yet-natural in the sense determined by
criteria b. and c.
The upshot is that considerations having to do with explanatory power and theory-choice
play a more fundamental role than acknowledged so far by liberal naturalists, who – as I stated
earlier – tend to simply offer rather crystallised list of entity-types, together with general – and
to some extent generic – claims concerning pluralism and the manifest image. However, there
is more to say.

3 Liberal naturalism and moderate naturalism about metaphysics
With the above discussion in mind, let us get back for a moment to Neta’s dilemma. In the
terminology used a moment ago, current liberal naturalists seem to think that criteria (a) and (b)
above are separately sufficient for delimiting the natural: for either something is natural because
it is postulated by our best current science, or because it is a necessary component of the
manifest image and, as such, is presupposed at least indirectly by science. It is this that provides a
way of the dilemma - as something might qualify as natural without being strictly speaking
scientific. Now, I wish to suggest that – once it is acknowledged that it is the kind of explanations that are provided that counts, hence (c) must also be considered necessary for supernaturalness – there is room for acknowledging things that satisfy both (a) and (b) but qualify
as natural nonetheless.
To begin with, consider an example. How can we establish whether or not Platonism about
numbers is ok? For sure, mathematics per se does not require one to take a stance on philosophical issues concerning abstract entities and existence. Nor is the existence of numbers as
self-subsistent entities part of the manifest image of the world. Hence, there is a sense in which
Platonic entities are not explanatorily necessary from a naturalistic viewpoint. However, and
this was Quine’s point concerning mathematical entities, it looks as though in cases like this
science together with additional philosophical considerations (in this case, about indispensability
in scientific practice) leads to explanations that are overall more comprehensive. That is, to explanations that may add little or nothing to science per se and are not grounded in common
sense, yet contribute to defining a larger picture of reality. Independently of how compelling
one finds indispensability arguments, surely it should be recognised that Platonism about
numbers provides at least one possible explanation of the nature of numbers, and one that
contributes to a more comprehensive worldview. Now: isn’t this the very same insight that
underlies the typical liberal naturalist idea that we should acknowledge nonscientific explanations of, say, facts about consciousness or libertarian free action, even in spite of the existence of
deflationary/eliminativist scientific accounts? If this is so, then Platonism, which certainly is not
part of the manifest image, could plausibly be regarded as part of our explanation and understanding of Nature broadly understood. And similarly for other – clearly not all! – metaphysical
posits and hypotheses.
On the one hand, notions such as – say – causality, identity, object, event and even universals or tropes may ground explanations of aspects of reality that are distinct from those
inquired into by science – including those having to do with the first- and second-person
perspective liberal naturalists care so much about. Thus, there is no reason for not including
them in the conception of nature endorsed by liberal naturalists. Especially so when one looks at
what at least some liberal naturalists say about the manifest image. Consider, for instance, De
Caro’s claim that “a ‘second nature’[…] also exists, which is distinct from the nature that is
investigated by the natural sciences. ‘Second nature’ stands for the world of culture, into which
we enter by way of education, and this is a world that is still ‘natural’, even if it cannot be
250

Liberal naturalism

accounted for by the natural sciences” (2015, 72–73, emphasis added). Or Macarthur’s idea that
“[l] iberal naturalism thus equates nature with the manifest image or, better, with the critical
manifest image that is the result of subjecting the manifest image to critical scrutiny, which
includes how well it hangs together with the scientific image of the world” (Macarthur 2019,
574, emphasis added). Surely, metaphysics is part of the culture that at least some humans
acquire by way of education, and a tool (the very tool?) for the definition of a critical image,
that is, of better explanations of the world we live in. Besides, isn’t liberal naturalism itself
ultimately a metaphysical thesis, based on specific metaphysical presuppositions?
On the other hand, and here we get back to the idea of items that are “presupposed” by
science, it can be plausibly argued that metaphysical presuppositions (besides not being sharply
distinguishable from scientific ones, as is well-known) are (1) always in play in actual scientific
theorising and also (2) required for understanding what science is actually saying about the world,
that is, for interpreting science. Indeed, how could one even start assessing the relationship between
the scientific image and the manifest image if the former is not interpreted, and is just presented as
a more or less formalised set of axioms and laws? And, doesn’t the interpretation of scientific
theories necessarily require the use of notions and concepts which are not themselves scientific
and, at least in some cases, are also importantly discontinuous with respect to common sense?11
To put it slightly differently, and more succinctly: if a more elaborated conception of explanation is put into play, there are good reasons for the liberal naturalist to accept more than
they usually accept, that is, to be even more liberal.12 In particular, there are good reasons for
accepting at least some of those entities, properties and processes, typically regarded as supernatural, that we call “metaphysical”.
The foregoing reveals an interesting link between liberal naturalism and another form of
philosophical naturalism, which I have recently proposed Morganti (2013). The underlying
reasoning was essentially the following:
(1) There is no sharp dividing line between metaphysics and science in terms of questions,
scope, generality, verifiability or what have you, and at least a proper part of contemporary
analytic metaphysics seeks answers to questions concerning the same reality inquired into
by the empirical sciences;
(2) Metaphysical theses are underdetermined by the evidence, but empirical data underdetermine scientific theories as well; hence, there is at best a difference of degree (of
connection with the world) between science and metaphysics, not one of kind;
(3) Metaphysics is needed for understanding (i.e. interpreting) science as much as science is
needed for giving substance, as it were, to metaphysics – in particular, by using metaphysical concepts for understanding what science says about reality, one ipso facto carries
out “indirect tests”13 of metaphysical hypotheses;
(4) Therefore, one should endorse “moderate naturalism about metaphysics”: naturalism about
metaphysics, that is, is best understood as the view that metaphysics should preserve its
autonomy, but be studied in parallel with science, being put to the test of empirical
evidence while at the same time defining the tools for the interpretation of science itself.
This form of naturalism seems very much in line with what was said in the previous sections.
Indeed, it is arguably a particular form of liberal naturalism, specifically concerned with the
relationship between scientific and philosophical methodology. Since it is intended to apply to
metaphysics, moreover, it meshes well with the liberal naturalist emphasis on the distinction
between natural and supernatural ontological posits. Additionally, since it is primarily a methodological thesis concerning how science and metaphysics should be carried out and put together,
251

Matteo Morganti

it promises to provide clear criteria for determining which nonscientific theoretical posits are to
be deemed acceptable from a naturalistic viewpoint. Not surprisingly (in Morganti 2013 as in
Morganti and Tahko 2017), moderate naturalism about metaphysics has been developed exactly
with a view to developing evaluation proceduresfor assessing various metaphysical hypotheses
in their connection with science, and more generally for critically evaluating and comparing
alternative ways of describing reality on the basis of both empirical factors and the so-called
nonempirical virtues.

4 The upshot: liberal naturalism 2.0?
Based on the preceding discussion, we can now close by summarising the main points of the
discussion, and defining a sort of updated version of liberal naturalism (while also putting the
two forms of naturalism discussed above – liberal naturalism and moderate naturalism about
metaphysics – together). In particular, I think that the take-home lesson of the present paper is
essentially the following:
1

The distinction between natural, supernatural and natural-yet-nonscientific entities boils
down to the – perhaps more customary – distinction between entities posited by science
and metaphysical entities broadly understood, the latter in turn to be divided into entities
that – given the current state of our knowledge – contribute to an overall explanatory increase
and those that do not;
The generic talk of “conflict with science” should be made more precise: besides overt
contradictions, what counts is whether or not nonscientific additions contribute to an
overall explanation in which science plays a central role yet does not provide by itself all
the explanations that can be given;
The original emphasis on items of the Sellarsian manifest image that are presupposed by
science should be enlarged so as to accept all those entities that account for aspects of either
the manifest image or the scientific image itself – in the latter case by interpreting theories.
This includes, most notably, (some) metaphysical items and hypotheses;
At the same time, some restrictions might be in order. A particularly thorny issue concerns
cases in which something seems to be part of our manifest image, and possibly presupposed
by scientific practice, but also to be amenable to scientific reduction or even elimination.
When exactly is it that replacing our common sense view of things like, say, action or the
mental, with metaphysically “thinner” substitutes is ok? On what basis should sciencebased deflationary or error-theoretic accounts be deemed (un)acceptable by liberal naturalists? I suspect answering this sort of questions will be vital for the future development of
(liberal) naturalism.

Notes
1 Also see Rosenberg (2004) for the use of this terminology.
2 Adding the term “current” to “science” is particularly important: as philosophers interested in issues
surrounding naturalism know well, there are good reasons for thinking that the relevant notions and
dichotomies are much more contextual and time-relative than one may think.
3 At least in their general outline: there will be no space here to get into all the relevant details.
4 Price (2004) distinguishes “subject naturalism” from “object naturalism”, and claims that the latter is
untenable. For the idea that the second-person perspective is central in the context of liberal naturalism, see Macarthur (2019).
5 The following passage is telling:

252

Liberal naturalism
Furthermore, to take seriously the scientific presupposition of mind, meaning and morals is to
see that scientific naturalists are, in spite of themselves, ontologically committed to nonscientific
phenomena. […] The most powerful objection to scientific naturalism is that the truths about
mind, meaning and morals are considered to be irreducible to truths about the causal items
recognised by the sciences. (Macarthur 2019, 570; boldface italic added)
6 For a partial illustration, think about temporal notions. Surely practicing scientists need to presuppose a
temporal dimension, if only to carry out experiments. And time is no doubt fundamental in our
common sense experience. But it is undisputable that recent physical science has strongly questioned
our entrenched conception of time, and even provided reasons for regarding time (and space) as an
emergent entity, or even for denying the reality of time altogether. On a slightly different note,
cognitive scientists have offered putatively exhaustive analyses of temporal experience purely in
physiological terms.
7 This should always be borne in mind, lest one do with other notions what Kant did when he argued
for the necessary truth of Euclidean geometry as the geometry of the physical world.
8 In Halvorson’s terminology, theism is compatible with “methodological naturalism”, even though it is
not compatible with “metaphysical naturalism”.
9 I say “necessary” here because the belief in God is widespread and possibly plays a central role in the
worldview of the majority of human beings. Yet, this appears to be a contingent, fully cultural fact, and
not something due to the very constitution of our biological apparatuses (which is itself a contingent
fact, of course, but in a different sense: we simply cannot change our basic biological features).
10 There is clearly no space here for a complete discussion of ad hoc-ness. For present purposes, suffice it
to notice that – in our example – the idea of a creating God that kickstarts the Big Bang calls into play
an alleged nonphysical first cause where physical theory simply doesn’t look for one. And, of course,
whether postulating such a first cause is truly satisfactory, or instead one is led into an infinite regress, is
also open to discussion. If this is correct, then the theist’s only hope is to argue (as in Halvorson’s case,
mentioned a moment ago) that there is no (vicious) regress and there is an overall net gain in explanation if one postulates the existence of God.
11 Examples include the use of notions such as primitive identity, identity of the indiscernibles, weak
discernibility, etc. to interpret quantum entities as individual or nonindividual objects; the reference to
theories such as presentism, eternalism and the growing block view to understand the nature of time
(and space) in the context of relativity theory and, more generally, contemporary physics; and the
application of concepts such as essential property, resemblance and dispositional conception of laws of
nature in discussion of natural kinds, especially in biology.
12 It is barely worth pointing out that this is a general claim that is completely independent of one’s views
concerning the ontology of mathematical entities.
13 For instance, using the first of the examples listed in footnote 11, it is not the case that just by looking
at quantum mechanics one learns that the microscopic domain is constituted by entities that differ from
microscopic ones in such and such a way. Rather, if one follows Leibniz in believing that any two
numerically distinct entities must differ in some monadic property, then one will interpret quantum
entities as different in kind from the individual objects that we interact with in our everyday experience; but the are other options leading to ontologies of individuals, of structures of events and
much more.

References
Davidson, D. Essays on Action and Events. Oxford University Press. (2001).
De Caro, M. Two Forms of Non-Reductive Naturalism. Phenomenology and Mind, 7, (2015). 71–83.
De Caro, M. and Macarthur, D. (eds.) Naturalism in Question. Harvard University Press, 2004.
De Caro, M. and Macarthur, D. (eds.), H. Putnam, Philosophy in an Age of Science. Physics, Mathematics and
Skepticism. Harvard University Press (2012).
De Caro, M. and Voltolini, A. Is Liberal Naturalism Possible? In M. De Caro and D. Macarthur (eds.),
Normativity and Naturalism, pp. 69–86. Columbia University Press (2010).
Halvorson, H. Why methodological naturalism? In K. J. Clark (ed.), The Blackwell Companion to Naturalism.
John Wiley & Sons. (2016).

253

Matteo Morganti
Ladyman, J. and Ross, D. (with Spurrett, D. and Collier, J.) Every Thing Must Go. Metaphysics Naturalised.
Oxford University Press. (2007).
Maddy, P. Second Philosophy: A Naturalistic Method. Oxford University Press. (2007).
Macarthur, D. Liberal Naturalism and the Scientific Image of the World. Inquiry, 62, (2019). 565–585.
McDowell, J. Mind and World. Harvard University Press. (1994).
Morganti, M. Combining Science and Metaphysics. Contemporary Physics, Conceptual Revision and Common
Sense. Palgrave Macmillan. (2013).
Morganti, M. and Tahko, T.E. Moderately Naturalistic Metaphysics. Synthese, 194(7), (2017). 2557–2580.
Nagel, T. The View from Nowhere. Oxford University Press. (1986).
Neta, R. Review of De Caro, M. and Macarthur, D. (eds.), Naturalism in Question. The Philosophical
Review, 116, (2007). 657–663.
Price, H. Naturalism without Representationalism. In De Caro, M. and Macarthur, D. (eds.), Naturalism in
Question, pp. 71–88. Harvard University Press. (2004).
Putnam, H. From Quantum Mechanics to Ethics and Back Again. In M. De Caro, and D. Macarthur
(eds.), H. Putnam, Philosophy in an Age of Science. Physics, Mathematics, and Skepticism, pp. 51–71.
Harvard University Press. (2012).
Putnam, H. Realism with a Human Face. Harvard University Press. (1990).
Ritchie, J. Understanding Naturalism. Acumen. (2008).
Rosenberg, G. A Place for Consciousness: Probing the Deep Structure of the Natural World. Oxford University
Press. (2004).
Sellars, W. Philosophy and the Scientific Image of Man. In R. Colodny (ed.), Frontiers of Science and
Philosophy, pp. 35–78. University of Pittsburgh Press. (1962).

254

22
NATURALISM WITH
CHINESE CHARACTERISTICS
Barry Allen

1 Introduction
Naturalism is a polemical term, and the polemic does not translate well into the terms of
traditional Chinese thought.1 If these thinkers were “naturalistic”, it would be unconsciously.
Can an unconscious presumption be a philosophy? I would say not: philosophy requires making
a philosophical decision. Yet it is not pointless to consider these works in thinking about our
naturalism. The closest analogy to naturalism that I find in traditional Chinese thought is an idea
of immanence, the ontological immanence of nature. All the conditions, principles, or causes of
the cosmos operate immanently in the cosmos. Whatever else naturalism is, it is this idea of
immanence. The universe is self-contained, without an external principle or source of unity.
This is the “nothing supernatural” moment of naturalism. Different versions introduce further
conditions, but naturalism is always a thought about immanence, and that gives us a point of
entry into the Chinese material.

2 Beginnings
At the beginning of Chinese civilisation, in the Shang and early Zhou dynasties (second millennium BCE), people worshiped a supreme deity named Tian. This is a word now translated
almost indifferently as “heaven” or “nature”. At the beginning, however, Tian was a kind of
personality-god, the Lord in Heaven, who takes a close interest in human affairs, directing,
scrutinising and responding to deeds with oracles. This Lord in Heaven was the foundation and
guardian of the ancient political order and had to be consulted prior to major undertakings.
Success and failure were viewed as a balance between human action and Heaven’s will.2 It
would be wrong to describe this divine Tian as supernatural. It is one of many powers active in
nature, one especially concerning to human beings but otherwise no less “natural” than anything in the sky.
The late Zhou dynasty was a time of crisis, climaxing with dynastic collapse. The crisis
seriously challenged the idea of a heavenly approved political order. From this time complaints
are heard against Heaven’s cruelty, the untrustworthiness of the deity, and the need to seek
solutions elsewhere rather than trying to discern the will of this inscrutable spirit. They started
to accustom themselves to thinking and acting godlessly, and Heaven’s decree faded into
DOI: 10.4324/9781351209472-26

255

Barry Allen

obsolescence. Not all at once, of course. But in the two centuries between the crisis-ridden
onset of the Spring and Autumn period (722 BCE) and the activity of Confucius (551–479),
Tian transformed from a watchful, interacting deity to an impersonal natural order with no role
in political life. Tian became a name for the sky, or the order observed in the sky. This celestial
order cannot be flouted and constrains what people do, but it is nothing personal, cannot be
addressed as a person, and takes no interest in human affairs. The mandarins retained the
ceremonial facade: the king still invokes Heaven and performs the old ceremonies. But the
belief, the trust, the expectation of divine favor is gone.
The result is not exactly atheism. There is still a kind of vague belief in deities and spirits.
But atheism is another polemical word, and traditional China did not have the polemic.
There being no one to argue with, what happened more resembles what Nietzsche calls the
death of God. The idea of a willful Lord in Heaven came to seem irrelevant, an unappealing
waste of devotion better spent on something with a more tangible prospect. Rulers should
concentrate on human affairs and not seek the support of a deity, about which their counselors were discreetly skeptical. Solutions to the problems of kingship lie in the secular realm
and the natural order, depending on knowledge and wisdom more than being right
with God.

3 Confucianism
On the evidence of the Analects Confucius was not entirely there yet. He speaks of offending
Heaven; Heaven’s curse; Heaven’s intention.3 But also of Heaven as an impersonal source of
loss and gain that is not under human control.4 The ceremonies he treasures are not practiced to
appease spiritual beings, whom we are told to keep at a distance.5 He does not say they do not
exist, but that they do not matter, and have nothing to do with the rites and ceremonies
Confucius revers. The importance of Confucian ceremony is to spiritualise life, especially social
life. Their most important function is harmony, which they achieve by making every action,
especially every interaction, feel meaningful, appropriate, natural and fair.6
This argument becomes clear in the “Discourse on Tian” by the late-classical Confucian
Xunzi.7 He explains how earth responds to heaven’s regular course. For instance, night comes
and the earth cools, winter comes and the earth turns stony. When the stars announce spring,
the earth warms and becomes fertile. The response is consistent and without harm. Socrates
urged his friends to agree that a supernatural mind must be behind such effects. To Xunzi, they
confirm the intricate interrelatedness of things. Nothing happens without response; every
change responds to and incites another.
The question Xunzi poses in this treatise is whether people’s actions can incite a response
from heaven. What effect do our actions, including ceremony, have on events? Especially
urgent is to know what pattern of human action results in harmony between heaven, earth and
humanity. Xunzi directs readers to practice the traditional rituals, which Confucians revere, but
as Confucius said we should not be superstitious, or think that in revering these traditions we
interact with deities. What then are we doing, and why is it effective?
Xunzi refers to a traditional rain sacrifice, a ritual performed in time of drought. He
explains that the ritual cannot produce rain; more generally, that ritual is not any kind of
instrument or technology. The rain sacrifice is not a way of manipulating heaven or coaxing
rain from the sky. He says the ritual is a celebration and worship of Tian, performed for the
satisfaction of its form, and to affirm, in a situation pregnant with disorder, a non-superstitious
confidence in nature’s immanent order. The point of the ritual is to sustain social harmony, to
keep people calm.
256

Naturalism with Chinese characteristics

We are admonished not to misunderstand unusual events like a comet, earthquake or
eclipse. These are natural occurrences, not judgments of heaven on people’s doings. The
pattern escapes us, but they belong to an immanent cosmic order and are without moral significance. We may marvel at them but should not fear them or interpret them superstitiously.
He speaks of heaven’s constancy: it does not act specially, neither for the good nor against the
wicked, nor does constancy exclude events like flood or drought. Whether such events are
disasters depends on how we respond. If superiors are enlightened and government stable,
people can survive such disruptions. But if these human conditions are unmet heaven cannot
help us. Order and chaos are due to rulers and governments. The disasters we properly fear, like
poor agriculture, high prices, heartless policies and war, do not come from heaven. They come
from bad government.
Mencius, the second name in ancient Confucianism, develops a further side of the idea of
immanence. He cites a line from the ancient Book of Songs: “If there is a thing, there is a norm”.
He then comments, “Kongzi said, ‘The one who composed this ode understood the dao!’
Hence, if there is a thing, there must be a norm”.8 The reasoning is unexpected, but it can be
worked with. Why must there be a norm?
Immanence implies that what a thing is has no substantial, intrinsic determination or
identity. Identity is instead a matter of relations in ramifying networks and process without end.
That implies an optimal way for things to coexist, a condition of mutual adaptation under
which things avoid provoking a destructive response. That is the norm. It does not make things
indestructible. Instead, it allots each thing its time, bestowing a duration whose power to absorb
disturbance contributes to the stability of an entire ecology. In the words of the Confucian
classic Zhongyong: “The ten thousand things are nurtured at the same time, and yet do not harm
each other. The ways are traveled at the same time, and yet are not contrary to each
other…These things are what make Heaven and Earth great”.9
What a thing is is the difference it makes to an environment, and what it ought to do or how
it ought to be handled belong together, grow from the same conditions, and are not so indifferent to each other as modern logic assumes. The Confucian idea is that we cannot understand what a thing is without understanding its relation to others upon whose existence it
depends, and which define the norm of its interaction. These are norms; flout them at your
peril. Yet they are immanent in nature, deriving from the nature’s relational ontology. They are
not an example of an “ought” from an “is”; for they are, to use Kant’s terms, merely hypothetical maxims. If you want to avoid untimely destruction and enjoy your allotted span,
harken to the order that sages find in nature.

4 The Daodejing
So far, I have discussed the theme of immanence in ancient Confucianism. It was their way of
remaining serious about rites and ceremonies without a superstitious interpretation of nature,
which they had learned was futile. Daoists are a second philosophical lineage, with two classical
works, the Daodejing, or Classic of the Dao and Its Power, attributed to the legendary sage
Laozi, and a work known after its only partly legendary author, Zhuangzi, or The Book of
Master Zhuang. Modern scholars do not agree about which came first. It seems simplest to
think of them as approximately contemporary, from around the mid-fourth century BCE,
China’s Warring States period, about a century after Confucius.10
To being with the Daodejing and explore what this work says to the theme of immanence, I
single out three themes: (1) Complementarity of opposites; (2) Fecundity of emptiness and (3)
Immanent naturalism.
257

Barry Allen

Complementarity of opposites. Opposites do not merely generate each other; they
complement each other’s action. The greatest effectiveness, the greatest strength, comes
from weakness.
To have and to lack generate each other.
Long and short offset each other.
High and low incline into each other.
Note and rhythm harmonize with each other.
Before and after follow each other.11
Fecundity of emptiness. An image from the Daodejing likens the dao to the hub of a wheel.
It is empty; there is nothing there. Yet without this nothing, what is there – the spokes and
rim – would be useless. The wheel is made useful by the emptiness at the center. The
image of emptiness, an emptiness that is a source of power and enables what surrounds it to
be effective, recurs throughout this work: emptiness at the center of a wheel; the emptiness
of a bowl or jar; the empty space of a room enclosed by walls; the emptiness of a doorway
offering access to the room inside. It is the source of every power, every use, every being.
Thirty spokes are joined in the hub of a wheel.
But only by relying on what is not there, do we have the use of
the carriage.
By adding and removing clay we form a vessel.
But only by relying on what is not there, do we have use of the
vessel.
By carving out doors and windows we make a room.
But only by relying on what is not there do we have the use of
the room.
And so, what is there is the basis for profit;
What is not there is the basis for use.12
Immanent naturalism is a plausible step from these considerations. Nothing transcends
nature, nothing beyond or outside nature. The power of nature is a power in nature, a
power of the formless, the nameless, engendering form. Formlessness is an expression for
the power of becoming. The formless is not the amebic, spectral or gaseous; it is the event
of form, its becoming, as the nameless is the not yet mature development of the named.
The power of nature is this original process in advance of form. That is the dao and the
source of its power, the power perceptible in becoming, expressed in change, an original
passage, an energetic pulse that formlessly becomes form.

Dao is not an entity, not a being with a form, shape, or presence in the world or beyond it.
Judged from the perspective of beings, dao is not a being, and may therefore be described as
nothing. Yet this nonentity silently governs the world, enfolding all things into an infinitely
complex interdependence. Plato said that becoming is for being; the telos of becoming is a
finished form. Not in China, where to the contrary the becoming is original, and form is but a
perception. The causes of form are immanent, but form does not cause form. The causes of
form are in between the forms, in the emptiness that separates them, a fertile emptiness not
unlike the quantum void, surging with formless virtuality.

5 The Zhuangzi
I pass briefly to a second Daoist classic, Zhuangzi. A passage reads:

258

Naturalism with Chinese characteristics

You hide your boat in a gully or your net in a swamp and call them secure. But in the
middle of the night a strong man could still take them on his back and leave, and you
would be asleep and not know. But if you hid the world in the world, you would
have nothing to lose. This is the essence of what lasts.13
We think we are clever and know how to use one thing to master another: use our boat to
master the water, our net to master the fish, and our foresight to master the thieves. But for
everything we use we are used in turn by something else. We need the boat and the fish because
we are used by the body that afflicts us with hunger, and we need to hide the boat because the
instruments of our livelihood are a resource thieves prey on. No use is secure. For each thing
there is a stronger, and nothing escapes its nemesis. To exist is to be a thing among things, and
that is to be menaced and ultimately overcome by them.
When everything is ceaselessly changing, everything becoming something else, the very idea
of “thing” becomes questionable. What we call things are artifacts of our perspective, ways we
pattern and punctuate nature’s formless, nameless eventuality. The whole idea of a thing whose
nature you might know is a misapprehension. If you want knowledge that does not falter, that is
effective and nourishes life, do not look for it in things separate and distinct. Learn instead to see
the whole world immanent in each seemingly separate thing. What makes any one thing one
makes the world, the whole world process being incipient in each single thing. If everything is
immanent in everything, then nothing can be lost or pass away and there is nowhere one is not
at home.

6 Neo-Confucianism
In the third century BCE, Zou Yan, whom Joseph Needham calls China’s first man of science,
introduced a new discourse on qi. Causes and conditions formerly attributed to local, ancestral,
and heavenly spirits begin to be explained in terms of the harmony, balance, and resonance of
qi. Music becomes a favored metaphor of harmonious process.14
Qi is a word for any gaseous substance – steam, clouds, smoke, the air, or breath – and the
modern Chinese word for weather. Traditional natural philosophy ascribes to it two characteristics, the first being to come in gradations of subtlety, for instance, now ice, now steam;
and second, dynamism, constantly moving and changing, like flowing water or rising steam. Qi
is material, provided that matter is understood dynamically, as interchangeable with energy and
never at rest. It has spatial extension and is in constant transformation independently of
awareness. As the original material of all things, it penetrates everywhere and makes all things
resonant.
Qi itself emerges from the more metaphysical interaction of the empty and the full,
which I understand in terms of actual and virtual.15 The full is actuality, full form fully
present; the empty is virtual potentiality, the fertile emptiness of the Daodejing in a new
implementation. Ceaselessly fluctuating tendencies of each to become the other create
energetic flows laminar and turbulent, which keep the qi far from equilibrium. This
movement is expressed in the world as it comes under our senses by the alternations of yin
and yang and the cycles of the Five Phases.16 Yin-yang is not a diametrical opposition but
implies nonantagonistic gradients on a scale. Phenomena analyse into infinitely integrated
proportions: a predominantly yin phenomenon always has a germ of yang, and in that germ,
a germ of yin, bearing yet another germ of yang and so on ad infinitum. Everything is made
of qi, everything bears a distinctive yin-yang signature, and everything interacts, becomes
correlated, and resonates.
259

Barry Allen

These ideas were quickly assimilated into Daoist natural philosophy. It took longer for them
to find an expression among the Confucians, but by the Song dynasty (960–1279)
Confucianism had taken up these ideas and reinvented itself as a philosophy of nature. A
leader in this movement was the early Song-dynasty Confucian Zhang Zai. He nicely recapitulates the dynamic that immanently energises nature’s cycles. “The vast emptiness cannot
be devoid of qi, qi cannot but condense and become the myriad things, and the myriad things
cannot but dissolve and become the vast emptiness”.17
Neo-Confucian qi looks very like Stoic pneuma. Both philosophies advance immanent,
background-independent ideas of nature. On the Stoic theory, original matter is passive and
formless, without cohesive force. Natural forms appear when a fiery rational pneuma mixes
into everything, producing a tension that makes distant parts continuous. Pneuma is corporeal
because it acts and is acted on, which is the Stoic criterion of body, yet they regard pneuma
more as force than matter, a continuous field spreading through space, interpenetrating every
body, and generating all the phenomena. Stoics also divinised this pneuma, identifying it with
the intelligent source of order in nature, which is where the comparison with China breaks
down.18
Zhang’s cosmos is an ateleological pattern that never began and therefore has no creator
or intelligent purpose. It resembles a machine in being an interacting assemblage but differs
in having no function other than its own operation. There is no creationism in traditional
Chinese thought, according to which nature is the outcome of a transcendent purpose. This
is humanity’s most flattering cosmology, suffused with intentionality, logos, rationality,
presence and purpose, all qualities consciousness recognises as its own. The Chinese cosmos
is autistic by comparison. There is no intentionality behind natural change, which is
spontaneous, from itself, an unregulated regularity, immanently orderly but transcendently
lawless.
Nature’s dynamism is sublime but godless. Human virtue depends on the sagacity to find the
way it functions and make that our principle in life. The difference between Confucian and
Daoist philosophy is a difference in how to do that. Zhu Xi, another Song-dynasty Confucian,
expresses his conviction that such a principle inheres in nature. “The blue sky is called heaven;
it revolves continuously and spreads out in all directions. It is sometimes said that there is up
there a person who judges all evil actions; this assuredly is wrong. But to say there is no ordering
principle would be equally wrong”.19
Zhu promoted a particular explanation of that “ordering principle”, in terms of what he calls
li, translated “principle”. The word originally designated lines that run through jade, which
became an image for the continuity or interpenetration of nature. The li are lines, imparting
consistency and interpenetration; more than that, however, they are lines to follow, being keenly
discerned by sages. We appreciate the principle of a thing when we know how to use it with
propriety, how to handle it well, do right by it, giving it a righteous place in the ceremonial
economy of the cosmos.
Zhu’s principal source for this concept was an earlier Song Confucian, Cheng Yi (a cousin of
Zhang Zai), according to whom
there is a difference between acting after having grasped [a principle] and acting
after having taken thought. If you have grasped [a principle] in yourself, the action
will be as simple as using your hand to lift a thing; but if you have to think, it is not
yet within yourself, and action is like holding one thing in your hand to take
another.

260

Naturalism with Chinese characteristics

For Cheng, Zhu, and their influential stream of Confucianism, the many li of the many
things ultimately form one cosmic Principle. Cheng writes, “There is one principle in all
the innumerable things, and even a single thing or activity, however small, has this principle”. It became a slogan in his school to say, “Principle is one, its particularizations are
diverse”.20
This Principle immanently regulates the economy of the world, sustaining its unity from
within rather than beyond. Principle is an endlessly reticulated network, whose resonant ripples
reach everywhere, like veins coursing through jade. There is nowhere they do not lead,
nothing they do not weave into an economy with others and ultimately with everything. This
Principle makes things, not what they are (metaphysical essence), but what they are good for, what
they can do, which establishes the norm of their adaptation in the fabric of the world, and the
way they change to avoid untimely destruction.
Notice that the conditions that engender the norms of nature are immanent, but also
processual. I point this out because naturalism is typically an idea of immanence without much
attention to process and the question of the temporality of immanence.

7 The comparison with liberal naturalism
One cannot say what “naturalism” is without using the word “science”, meaning modern
science, which did not exist in traditional China. Even “liberal naturalism”, which aspires to
take seriously people’s everyday experience and accept the unreduced reality of ordinary midsized objects, takes pains to reassure us that science and its exemplary objectivity are not denied
but only qualified as partial.21 The liberality of this naturalism consists in a refusal to limit
philosophical methods to those of the natural sciences, and a recognition of legitimate forms of
non-scientific explanation, for instance in art criticism and the explanation of action by reasons.
That is not, of course, to say traditional China had no science, but none that could play the
role required for the “naturalism” polemic in any of its current forms, liberalor physicalist. As
their scientific understanding of nature grew, traditional assumptions were never threatened.
Their advances in astronomy, biochemistry, medicine and other sciences never made intellectuals think their deepest beliefs were being challenged. Buddhism did affect them that way;
they felt a challenge and an opponent and did not restrain their polemic. But nothing coming
from their many lines of natural research was alienating. It would have been practically impossible for any new scientific insight to present a conundrum for their ethical and political
thought.
They do not need naturalism because they were never supernatural, and they do not need
liberal naturalism because they were never scientistic or positivistic. The point of “liberal
naturalism” is to loosen bonds that never bound China. If, however, we set the polemic with
positivism aside and consider the meaning of “nature” in liberal naturalism, we find an idea of
immanence that is not entirely dissimilar to ideas in traditional Chinese thought. When,
thinking comparatively of our naturalism, we canvass the theme of immanence in Chinese
philosophy, we find appreciation of something not usually thematic in naturalism, namely,
considerations of genesis and movement that take time more seriously than European philosophies typically do. Other philosophers who advance an idea of immanence appreciate the
need for a consistent philosophical position on time, the outstanding example being Gilles
Deleuze (2001). Liberal naturalism may still require liberation from the residual positivism that
assumes time is an essentially spatial fourth dimension. A really liberal naturalism would think of
time as we experience it, not as it is represented in physics.

261

Barry Allen

Notes
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21

On the difficulty of using “naturalism” as a category in the history of philosophy, see LoLordo (2011).
Pines (2002).
Confucius (2001, 3.13, 6.28, 9.5).
Ibid., 11.9, 12.5.
Ibid., 6.22.
Ibid., 1.12. All the Chinese thinkers discussed in this chapter are treated in more detail in Allen (2015).
Xunzi, “Discourse on Heaven”, in Ivanhoe and Norden (2001, 269–274).
Mencius, in Ivanhoe and Norden (2001, 148).
Johnston and Ping (2012, 481)
The idea of “schools” of Chinese philosophy is not satisfactory; see Smith (2003) and Sivin (1978).
Daodejing, in Ivanhoe and Norden (2001, 163).
Ibid., 167–168.
Zhuangzi, in Ivanhoe and Norden (2001, 236).
Needham (1978) and Brindley (2012).
Allen (2015).
Major (1993).
Zhang Zai, in Kim (2010, 114). On Zhang, see also Kasoff (1984), and on Song-dynasty neoConfucianism, Angle and Tiwald (2017).
On pneuma and the Stoic philosophy of nature, see Sambursky (1959) and Hahm (1977).
Zhu Xi, in Needham (1956, 492). It would be wrong to think Chinese cannot distinguish “is” and
“ought” or that they were unaware of this distinction, Graham (1990, 430).
Graham (1958).
De Caro and Macarthur (2004, 2010).

References
Allen, B. Vanishing into Things: Knowledge in Chinese Tradition. Harvard University Press. (2015).
Angle, S. and Tiwald, J. Neo-Confucianism: A Philosophical Interpretation. Polity Press. (2017).
Brindley, E. Music, Cosmology, and the Politics of Harmony in Early China. State University of New York
Press. (2012).
Confucius. Analects. In Readings in Classic Chinese Philosophy, 2nd edition. Trans. P.J. Ivanhoe and B.W.
van Norden. Hackett. (2001).
Deleuze, G. Pure Immanence. Trans. Anne Boyman. Zone Books. (2001).
De Caro, M. and Macarthur, D. Naturalism in Question. Harvard University Press. (2004).
De Caro, M. and Macarthur, D. Naturalism and Normativity. Columbia University Press. (2010).
Graham, A. Two Chinese Philosophers: Ch’eng Ming-Tao and Ch’eng Yi-Ch’uan. Lund Humphries. (1958).
Graham, A. Studies in Chinese Philosophy and Philosophical Literature. State University of New York
Press. (1990).
Hahm, D. The Origins of Stoic Cosmology. Ohio State University Press. (1977).
Ivanhoe, P.J. and van Norden, B. (eds.) Readings in Classical Chinese Philosophy. Hackett. (2001).
Johnston, I. and Ping, W. (eds. and trans.) Daxue and Zhongyong. Chinese University of Hong Kong
Press. (2012).
Kasoff, I. The Thought of Chang Tsai (1020–1077). Cambridge University Press. (1984).
Kim, J-Y. A Revisionist Understanding of Zhang Zai’s Development of Qi. Asian Philosophy, 20, (2010).
111–126.
LoLordo, A. Epicureanism and Early Modern Naturalism. British Journal of the History of Philosophy, 19(4),
(2011). 647–664.
Major, J. Heaven and Earth in Early Han Thought. State University of New York Press. (1993).
Needham, J. Science and Civilization in China, vol. 2: History of Scientific Thought. Cambridge University
Press. (1956).
Needham, J. The Shorter Science and Civilization in China. Cambridge University Press. (1978).

262

Naturalism with Chinese characteristics
Pines, Y. Foundations of Confucian Thought: Intellectual Life in the Chunqiu Period 722–453 BCE. University
of Hawaii Press. (2002).
Sambursky, S. Physics of the Stoics. Routledge and Kegan Paul. (1959).
Sivin, N. On the Word ‘Taoist’ as a Source of Perplexity. Religious Studies, 17, (1978). 303–330.
Smith, K. Sima Tan and the Invention of Daoism. Journal of Asian Studies, 62(1), (2003). 129–156.

263

PART IV

Applications of liberal naturalism

23
LIBERAL NATURALISM AND
AESTHETICS: ART UP CLOSE
AND PERSONAL1
David Macarthur

1 Introduction
In this article, I shall consider the relationship between naturalism and aesthetics. I will not
attempt to define at the outset what kind of naturalism best accommodates aesthetic phenomena
by which I shall mean, for present purposes, our experience of artworks.2 What an appropriate
naturalism looks like is precisely what is in question. In order to address this question, I want to
critically discuss Murray Smith’s Film, Art and the Third Culture (2017) because it has the virtue
of putting this question at the center of its account of the aesthetics of film.3 It also has the aim
of articulating a naturalism that makes available a “cooperation” or “integration” (2) of scientific
explanation and humanistic interpretation or understanding.4 Of course what cooperation or
integration means in this context is part of what is in question.
Smith provides the outlines of what he calls “a third cultural approach” to film art which
draws upon discoveries or insights of phenomenology, psychology and neuroscience – thus,
potentially overcoming the sense of a deep rift between the sciences and the humanities in
Western society famously articulated by C.P. Snow in his book, The Two Cultures (1959). It is
perhaps no surprise, given these ambitions, that Smith prefers a “nonreductive cooperative
naturalism” whose job is to reconcile the manifest and scientific images of the world without
any wholesale replacement of the former by the latter (4).
At this general level of description, I am sympathetic to Smith’s reconciliation strategy and
his stated aims in attempting to articulate a naturalised aesthetics suitable for art in general, and
film in particular. The idea of a non-reductive cooperative naturalism suggests what in previous
work I have called a liberal naturalism, a philosophical orientation that contrasts with orthodox
scientific naturalism in being non-reductive; and from the point of view of which, unlike that
of Sellars and a whole generation of scientistic philosophers contemporaneous with him (most
notably Quine), there is no irredeemable general clash between the manifest and scientific
images of the world.5 Liberal naturalism, in contrast to orthodox scientific naturalism, takes the
manifest image seriously as the home of human intelligibility and responsiveness which is largely
taken for granted by the development of various forms of scientific intelligibility.6
But, as always in philosophy, there is the question whether Smith has earnt the right to his
claims – in this case, of cooperation, integration and reconciliation. I want to take issue specifically
with Smith’s idea that he can combine nonreductive explanatory ambitions whilst endorsing the
DOI: 10.4324/9781351209472-28

267

David Macarthur

orthodox form of contemporary naturalism, namely, scientific naturalism. I shall argue for two main
ideas: first, that earning the label “non-reductive” in the setting of a scientific naturalism is a lot
more problematic than Smith supposes and once one sees the difficulties that arise then one’s sense
of the philosophical landscape changes dramatically; and, second, whereas Smith thinks the way to
thicken explanations in (film) aesthetics is to go sub-personal by invoking sub-personal neural
mechanisms in the explanation of aesthetic phenomena, I suggest that such a move is optional and,
in any case, philosophically problematic. The most accommodating form of naturalism from the
point of view of aesthetics is not a scientific naturalism concerned with the objective causes of
natural phenomena, but a liberal naturalism which takes seriously persons and their experiences,
both of which are fundamental to the manifest image. The manifest image is manifest to persons;
and manifestness is a function of its being experienced by persons. Aesthetics is, first and foremost, a
person-level phenomena – concerning one’s experience of artworks and one’s capacity to elaborate
their human significance and value in everyday psychological or social terms – even if the complexities and depths of these phenomena imply that they are anything but straightforward.

2 Can scientific naturalism be nonreductive?
Let me begin with Smith’s characterisation of naturalism,
Typically naturalism in the sense at stake here is held to have two aspects: a substantive
(metaphysical) commitment to the study of all phenomena, including human behavior, as a part of the physically constituted, biologically evolved world; and a methodological commitment to the methods and standards of the natural sciences.
To approach some phenomenon naturalistically is to seek to place and explain it
within the natural order, while the rational-empirical methods of science—framing
hypotheses, seeking evidence, considering alternative hypotheses and countervailing
evidence—are, on this view, our best bet in realizing this aim. (22)
From such remarks, we gather that Smith’s naturalism is a form of scientific naturalism. The
world is, metaphysically speaking, nothing other than the “natural order”, that is, the scientific
image of the world provided by the natural sciences: primarily, the image provided by physics,
chemistry and biology.7 And our best bet as to what is genuine knowledge or understanding is
arrived at by applying the experimental methods of natural scientific inquiry. This definition of
naturalism, which gives metaphysical and methodological weight to the natural sciences, is, in
fact, the orthodox form of scientific naturalism in contemporary philosophy that we associate
with philosophers such as Andy Clarke, Fred Dretske and W.V. Quine – all of whom Smith
invokes as examples of the kind of naturalism he espouses (22).
Scientific naturalism, so understood, is in the business of naturalising various phenomena, that
is, as Smith explains, “to seek to place and explain [the problematic phenomena] within the
natural order”. Such naturalisation projects are attempts to answer various “placement problems”
that arise in the context of the claim that the world is nothing other than the natural order
discovered by the natural sciences. How are we to find a “place” for those aspects of human
experience, thought or talk – say, mind, morality and mathematics to give three well-known
examples – that, on the face of it, do not fit into the natural order, so conceived? The theoretical
options are few and familiar: eliminate the putative entities from one’s ontology; reduce them to
naturalistically acceptable entities; or reconstrue the discourse that putatively describes or refers to
the problematic phenomena as nondescriptive, nonfactual or nonrepresentational. But since
268

Liberal naturalism and aesthetics

Smith accepts a “representational theory of mind” (222) and a “space of aesthetic reasons” (29)
which engages with “aesthetic properties” (53), then one might reasonably suppose that reduction
is the only available option that his naturalism allows him.
This poses a serious conceptual problem for Smith in so far as he claims to defend a nonreductive naturalism that is not in the business of eliminating or reducing the aesthetic phenomena it intends to study. But one cannot simply add the label “nonreductive” to scientific
naturalism as if it were a matter of taste whether one is reductive or not. Scientific naturalism is
inherently reductive if one is not prepared to dispense with the phenomena one is attempting to
explain. If a scientific naturalist is not to ontologically reduce some phenomena that does not fit
into the natural order and she refuses to radically revise how we talk, then the only alternative is
to take the drastic step of denying any representational account of mind and language with
regard to the phenomena in question – in this case aesthetics. But that would be to deny that we
enjoy aesthetic experiences of aesthetic properties which we represent in language using aesthetic predicates. That is, reductionism can only be avoided at the cost of elimination of the
phenomena in question. Rule out elimination and reductionism is the default position.
The alternatives for a scientific naturalist who embraces representationalism are stark: ontological
reduction or elimination of the putative entities. In either case, the scientific naturalist is committed to
the denial of any distinctive subject-matter that is not a part of the scientific image of the world.
Regarding aesthetics, since the natural order does not, prima facie, contain aesthetic properties then one
has no option but to treat aesthetic properties as naturalistic properties; or, in so far as aesthetic properties
are nonnatural (i.e. nonscientific: not posits of successful scientific explanations) to deny that there are
any aesthetic properties. It may seem easy to avoid the harsh implications of scientific naturalism if one
focuses on the Davidsonian point that there may be no meaning-preserving translational reductions of
aesthetic predicates to naturalistic predicates.8 But Davidson’s larger point is that a reductive ontology of
mind does not require a translational reduction of psychological predicates.9 And the same is true for
aesthetics: a reductive account of aesthetic properties is compatible with denying that there is any
translational reduction of aesthetic predicates. But then an inexplicable gap opens up between the
ordinary meaning of aesthetic language and what it is taken to be really about – on Smith’s account,
unobvious (perhaps unobservable) physical features of the artwork or subpersonal (hence unobservable)
neurophysiological features of the viewer – in either case, items of which we are not aware. We are
thus led to an incoherent conclusion that what aesthetics is really about is unobservable since what
we are left with does not rightly count as aesthetics at all.
The problem for Smith is that if scientific naturalism denies that there is any distinctive
aesthetic subject-matter, then there is no such project as a nonreductive integrative study of
aesthetic experience or properties in relation to the study of natural-scientific facts and properties. Since Smith in no way wants to eliminate aesthetic subject-matter from the universe, the
only option left to him, within his naturalist framework, is reductive integration: studying aesthetic subject matter as a part of the larger study of the natural order. The distinctiveness of
aesthetic language (i.e. its untranslatability) is accorded no ontological significance or weight
within a scientific naturalist outlook. Indeed, it is made mysterious by grounding aesthetic
significance in, for example, unobservable properties and relations recognised by the natural
sciences. But why must we accept this reductive and obfuscating way of looking at things?

3 Triangulation and “thick” explanations
Let us return to the issue of naturalistic explanation in aesthetics. Smith’s naturalistic program
replaces conceptual analysis with a naturalistic explanatory strategy he calls “triangulation”
which he elaborates as
269

David Macarthur

the principle that at the outset of enquiry we take seriously not two but three factors –
three levels of analysis with their attendant types of evidence – that we have at our
disposal with respect to mental phenomena: the phenomenological level…; the
psychological level…; and the neurophysiological level (60).
Although Smith claims that “the three levels…are interdependent”, his conception of the
relations between these levels assumes an important explanatory asymmetry between them. For
the most part, the phenomenological level only supplies data for the explanatory ambitions
of the other two levels, the psychological (by which he means the level of scientific psychology)
and the neurophysiological.10
A central feature of Smith’s naturalistic aesthetics is its focus on sub-personal mechanisms in
order to generate what he calls “thick explanations” (9) of film spectatorship such as suspense,
or empathy with the characters observed on film, or the perception of color and sound in film,
as well as our emotional engagement with the worlds of film.11 Smith remarks:
Thick explanation might be thought of as a model or ideal of the naturalistic explanation of artistic and cultural phenomena (51).
The term “thick” is here being used as a term of praise; “thin”, in contrast, is derogratory. On
this account, “thick” naturalistic explanations are regarded as superior to the allegedly “thin”
explanations of ordinary engagement and art criticism which rely on a nontheoretical practitioner point of view and “folk” (or everyday) psychology:
If I restrict my analysis of a film sequence to folk psychology and the concepts of the
practitioner, I can offer an explanation, but a relatively ‘thin’ one. Reference to subpersonal phenomena enables a ‘thicker’ explanation, by revealing the mechanisms that
filmmakers intuitively exploit (52).
Elsewhere, Smith adopts the even stronger thesis that if one wants a naturalistic account of film
spectatorship, then one is “required” to invoke subpersonal (presumably neurophysiological)
mechanisms:
a naturalistic approach to film spectatorship requires that we do not stop with the
experience and explanatory concepts related to the level of personhood (beliefs, desires, intentions, feelings), but urges that we thicken our account by engaging with the
sub-personal mechanisms underpinning personhood (48). [Emphasis added].
On Smith’s naturalistic aesthetics, then, appealing to scientific psychology and, especially brain
science (e.g. neurophysiology), are the only hope for generating substantial explanations of our
aesthetic experience of film – a surprising result given Smith’s initial claims of cooperation,
integration and reconciliation of the sciences and the humanities.

4 The case of empathy
Smith cites empathy as a prime example of a mental state involved in our aesthetic experience
of film that neurophysiology can shed important light upon – again given the assumption that
“we thicken our account by engaging with the sub-personal mechanisms underpinning personhood” (48). I shall take it as a test case of what a thick explanation of an important aesthetic
270

Liberal naturalism and aesthetics

phenomenon looks like in Smith naturalistic aesthetics. It will, therefore, be worth examining
the case in some detail to ask whether thick explanations can really carry the weight that Smith’s
naturalistic explanatory program puts on them.
Smith formulates empathy thus,
In imagining how some other specified agent sees the world and in imagining how they
think and feel, I empathize with the agent. Call this other focused personal imagining –
imagining the experience of others ‘from the inside’ in the words of Kenneth Walton.
Such imagining allows us not merely to recognize or understand but to grasp directly…the emotional frames of minds of others (179).
This is a more sophisticated and specialised notion than the one we find in a standard dictionary:
“Empathy = def. the ability to understand and share the feelings of another”.12 As Smith
astutely notes, one can understand an emotion without feeling it, at least there and then; and
one can share it in the sense of having the same type of emotion (say, joy or grief) without
sharing it from the other subject’s own point of view. To empathise, as Smith sees it, requires that the
relevant imagining must be focused on what it is like “from the inside” for the other. Let us
examine this conception further.
Smith claims that the neurophysiological discovery of mirror neurons – “neurons that are
both active when a subject performs and observes an action” (64) – provides a good example of
the power of subpersonal mechanisms in the explanation of person-level aesthetic experiences
in film and the arts generally. The main explanatory advantages of the appeal to mirror neurons
for the aesthetic phenomenon of empathy are claimed to be these:
1

3
4

Empathy is a genuine mental state: Mirror neurons provide indirect scientific evidence for the
existence of empathy, a contested mental state, by providing “a neural mechanism that may
underpin the experience of empathy and the psychological function it performs” (73).
A simulation account of empathy is most likely correct: Mirror neurons provide indirect scientific
evidence for simulation theory in the debate between those who take our ability to ascribe
mental states to others to depend on an implicit “theory of mind” (so-called theory theory
accounts) and those who take this ability to depend on an empathic capacity to simulate the
mental states of others (simulation accounts).
Explanation of functional nuances of empathy: “Neural evidence sheds light on the functional
nuances of the phenomena that elude ordinary experience and reflection” (103).
Explanation of the variability of empathy between people: “In the case of mirror responses, brain
research is gradually giving us a picture of how the targets of empathic response (sensations,
emotions, actions) may vary across primate species and, among humans, may vary according to background and experience” (104).

Let us consider each of these claims in turn. The first two are really responses to skepticism
about empathy as Smith understands it, since according to the theory theory account all we
have at our disposal is a theoretical inference which, if all goes well, produces understanding of
the other’s feeling (and perhaps also sympathy as a result of that understanding) but not empathy, that is, feeling for the other but not with them. Smith takes it as obvious that the scientific
discovery of mirror neurons provides evidence, albeit not decisive evidence, that we can feel
the very same emotion as the screen figure we are watching; that is, an imaginative feeling “from
the inside” of that particular other in their particular circumstance.

271

David Macarthur

But there are serious problems with Smith’s demanding conception of empathy. To empathise with another person is “to stand in their shoes”, as we say. How are we to unpack the
metaphor of standing in another’s shoes? One option would be to simply imagine oneself in the
same physical situation as the other person, say, imagining what one’s sensations would be like if
one stubbed one’s toe on a rock or cut one’s finger with a kitchen knife supposing this is what
happened to the person we are empathising with. But empathy is not supposed to be simply
having the same experience in the restricted sense of having the same physical sensations –
something that seems relatively easy to imagine and understand at least if we assume, reasonably
enough, that we have the same human physiology. On the contrary, on Smith’s account, one is
supposed to imagine what it feels like for them emotionally “from the inside”, as we say, employing a metaphor of inner and outer spaces. The suggestion is that we are attempting to know
what emotions the other is going through in a rich sense which typically requires imaging being
in certain complex affective and cognitive states, given that many emotional states presuppose,
or are at least partly caused by, beliefs – something that is obvious in the case of so-called
“cognitive emotions” such as shame, pride and guilt.13
So let us consider a more interesting case. How is one to empathise with another person –
call him John – going through a divorce whether in actual fact or on the screen? What is
involved in the imaginative exercise of standing in John’s shoes? Supposing one is unmarried or
a woman then one will have no past experience of one’s own of the relevant kind to draw
upon. And even if one is a married man, how would one know what to imagine? Divorces, like
marriages, are as different from each other as people or artworks or trees or flowers or beetles
are. As we say, when it comes to psychological matters like this, the Devil is in the details. And
that is just the problem. The more we try to imagine the specific psychological details of what it
is like for some particular other to be going through their divorce, or a particular episode within
it, the more formidable the problems we face become, both epistemological and conceptual.
For example, how are we to imagine being unhappily married to John’s wife? Do we (can
we) really know what that is like? What is required to imagine that? Are we to imagine how his
particular character traits, his system of beliefs, his cares and concerns and so on, mesh or fail to
mesh with her particular character traits, her system of beliefs, her cares and concerns? And since
we presumably have a partial and imperfect grasp of both his and her past and current mental
life, attempting this complicated imaginative project seems to confront us with major epistemic
shortcomings. It is hard to avoid the thought that we do not know nearly enough about the
psychologies of each of these two people to engage in the relevant form of imagining.
Furthermore, since their past life together is relevant to imagining what it is like to be him now,
our epistemic shortcoming begins to seem overwhelming. How can we possibly imagine feeling
as he does without inventing a great deal to fill in all the gaps in our knowledge of his past?14
And, then, why would that be empathising with him rather than a creative imaginative construction of our own occasioned by his outward circumstances but without really being a way
of knowing his particular feelings at all? Our lack of knowledge of John’s mental life (not to
mention his wife’s mental life and her past experiences and the ways they interacted over a
significant period of time living together) and his past experience of the world and its vast and
minutely complex consequences for his current feelings surely makes this sort of counterfactual
thinking a piece of imaginative fiction rather than genuine access to psychological facts about
him. As Karsten Stueber perceptively remarks, like many explanations of empathy, Smith’s
“fail(s) to explain why…our ‘feeling into’ the other’s mind is more than mere projection”.15
Empathy is supposed to be a route of access to the feelings of another – in ordinary life or on
screen – but the more we attempt to unpack Smith’s demanding conception of standing in
another’s shoes at a certain time and place the more we rely on already knowing a great deal about
272

Liberal naturalism and aesthetics

their minds – including how they thought and felt about all sorts of things in the past. But
knowing how they think and feel is the very problem that the appeal to empathy was invoked to
solve in the first place! And the problem of vicious circularity is not the only problem. The closer
we want to come to the goal by imaginatively filling out the other’s psychology and its embeddedness in the world including the complex interplay between psychology and environment,
natural and social, past and present and so on, it seems that what we are really attempting to
imagine is being the other person. And how am I supposed to do that and at the same time keep a
grip on my own different identity as apparently required by the logic of my empathising with them?
The problem here is one of conceptual incoherence, not merely an epistemic lack. I’m not John
and I can’t imagine being him in the rich sense required and at the same time retaining a sense of
myself as myself. Smith’s conception of empathy thus runs aground.

5 Must we go subpersonal in aesthetics?
Let us now consider Smith’s third explanatory claim that “neural evidence sheds light on the
functional nuances that elude ordinary [aesthetic] experience and reflection” (103). Smith
provides the following example:
It is plausible to suppose that film-makers may exploit our empathetic capacities in the
way that they represent action and emotional expression. Hitchcock, in particular,
sought empathetic effects through precisely wrought renderings of gesture and facial
expression. In a famous sequence in Strangers on a Train (1951), for example,
Hitchcock’s meticulous crafting of the action, sustaining our attention to the grasping
motions and expressive exertions of Bruno (Robert Walker) as he reaches for a cigarette lighter stuck in a drain, invite us to ‘feel into’ his emotions” (77).
This seems to me a plausible-enough description of the way Hitchcock invites an empathic
reaction in the audience; but it does nothing to establish the explanatory significance of neural
mechanisms (say, involving mirror neurons) that, by hypothesis, play a causal role in making
our aesthetic experience possible. The entire description is at the person level and makes no
reference to subpersonal brain processes. No thick explanation of person-level phenomena is
offered on the basis of neural evidence. Smith goes on to remark,
many films create ‘scenes of empathy’ – scenes in which the viewers are exposed to
sustained close-ups of emotive facial expressions, often combined with affective aural
cues, apt to elicit empathic feelings via the mechanism of mirror neurons. (77)
This passage gives the game away. The last clause invokes mirror neurons only as the neural
state that implements empathic feelings. It does not provide any of the advertised “functional
nuances” nor “further fine-grained level of detail” (103) about the person-level phenomenon of
empathy in the cinema that Smith claimed as the explanatory advantages of his naturalistic
triangulation approach. One might recall Austin’s remark about the pretensions of philosophical
explanation: “There’s the bit where [they] say it and the bit where [they] take it back”.16
What we can acknowledge is Smith’s basic claim that “cognitive architecture is shaped by
neural architecture” (103). But in the context of his nonreductive account of the mind this
dependency claim is of no explanatory help at all in understanding our experience of film (or art
in general) since we do not understand the relevant notion of being “shaped” in this context.
Smith rekindles the old physicalist problem of invoking supervenience as if it were an
273

David Macarthur

explanation. To say (irreducible) aesthetic experience supervenes on physical properties (say,
body plus environment) is useless for explanatory purposes if we do not understand the nature
of this dependence.17 In general, if we cannot explain the way in which the personal depends
on, or is shaped by, the subpersonal then we cannot countenance Smith’s claim that understanding the neural facts explains empathy in some robust or substantial “thick” sense.
Smith final point about the empirically discovered variability of empathic response in humans would appear to undermine, rather than support, his explanatory ambitions. As his explanation of empathy in Stranger’s on a Train makes clear, the thick explanations that Smith
presumes – but, I have argued, fails – to establish are based on the scientific discovery of general
causal patterns such as what causes, or is caused by, mirror neuron excitation. Science aims to
discover general truths: the existence of certain universal causal laws; or, at least, certain more or
less localised causal patterns. But there is no such thing as a science of aesthetics because, as Kant
famously observes, “there can be no rule in accordance with which someone could be compelled to acknowledge something as beautiful”.18 Put otherwise, there is no rule or law connecting one or more of the myriad features of a particular artwork with a particular judgment of
taste, say, a judgment of beauty. Artworks, like persons, are unrepeatable unique entities (or
haecceities, to borrow a term from metaphysics) whose useful criticism cannot rest content with
generalisations, whether wide or narrow. For this reason, science at the personal or sub-personal
level is of only limited usefulness in the appreciation of an artwork. Fruitful criticism is concerned
with the singularity or haecceity of aesthetic objects at the person level; science is concerned to
explain by appeal to general causal patterns or laws, often regarding unobservables.
Admittedly, there are some features of our aesthetic experience that can be explained by
science. Smith remarks,
Attention to the sub-personal level allows us to explain such basic constitutive features
of cinema as seeing depth and motion, and integrating sound and image, about which
analysis at a personal level has little to say (49).
But depth and motion perception in the cinema are features of film as such, not any film in
particular. If we are interested in our aesthetic experience of a specific film and what sets it apart
from other films, explanations of depth and motion perception are largely beside the point.
We have seen that Smith has not provided any compelling case for going subpersonal in
(film) aesthetics. The questions posed by film, for example, involve descriptions of people in
psychological and social terms that demand answers concerning motivations, intentions, wishes
and wants, hopes and fears, loyalties and betrayals, good and evil and so on. In the context of
scientific naturalism, such psychological and moral categories are left unexplained by the subpersonal mechanisms that they depend upon. To say that mentality depends upon a functioning
brain is true but explanatorily empty at least when it comes to explaining our mentality as we
experience it. I have also argued that scientific naturalism is inherently reductive – although
Smith attempts (unsuccessfully) to sidestep its reductive implications. Indeed, it typically issues
promissory notes for future explanations involving hoped-for reductions. But it is a naturalist
insight that we must respect empirical reality, how things actually are. Promissory notes are
wishful thinking, and not to be taken seriously in philosophy.

6 A liberal naturalist aesthetics: a brief sketch
In view of the problems facing scientific (hence, reductive) naturalism, let us adopt in its place a
“liberal naturalism”, a genuinely nonreductive naturalism that is much better suited to Smith’s
274

Liberal naturalism and aesthetics

cooperative reconciliatory ambitions (–which is not to prejudge the question of how far these
ambitions can be realised). Liberal naturalism, as a form of naturalism, accepts that the world is
nothing more than the natural order. But notice that what is meant by “the natural order” on
this reconception is not limited to the scientific image of the world, whether restricted to the
natural sciences alone or expanded to include the social sciences. The natural order is simply the
non-supernatural order; a conception which makes conceptual space for the recognition of
nonscientific non-supernatural phenomena such as persons, ordinary artifacts (e.g. chairs) and
artworks (e.g. novels).19
When I speak of artworks as nonscientific, I am not denying that they can be objects of
scientific study. What I am denying is that their existence depends on whether or not they are
“posits” of a successful scientific theory, to use Quinean terminology. In other words, they have
a life science knows nothing about outside of the context of scientific inquiry. If we understand
“science”, plausibly enough, as an objective empirical study of the causal structures of the
world, then there can be no science of art, as Kant saw, even if we can (in a sense) study
artworks and our experience of them scientifically.20 The concepts of artwork and of the categories of artwork are not fitting concepts for scientific study since, for one thing, as Quine
saw, their individuation conditions (as well as the identity conditions of their features) are too
messy and indeterminate to allow for reproducible identification, repeated verification and
objective measurement – all requirements of being a fitting object of scientific study. But such
shortcomings need not apply to the concepts of the chemical compounds that make up paint or
plaster (etc.) or to the concepts of various neurophysiological reactions people undergo when
observing an artwork – for example, the activation of mirror neurons. Put otherwise, the
concept of art, or of specific categories of art, are not suitable scientific concepts since they are
too subjective to count as such. But, of course, artworks can be studied as physical things or
chemical things or economic things or in many other ways within various scientific
subdisciplines.
The concept of art does not figure in successful scientific explanations any more than the
concepts of chair, table or person do. Since there is no science of such things, no objective
causal patterns in which they are caught up qua chair, table or person, they count as “nonscientific” as I am using that term. But that does not make them nonnatural. In saying this, I am
refusing a physicalist (or narrow scientific naturalist) conception of artworks as nothing more
than a class of physical objects with certain interesting perceptual features. What is art then?
From within a liberal naturalist perspective, let us say that art is something which invites, sustains
and resists a certain sort of critical attention and engagement.21 Since art bears an internal relation to
our experiences of art, it is dependent upon human subjectivity, and in particular, the way art
engages with the human significance and value of things – aspects of the world that pose serious
(if not insurmountable) problems for objective scientific study. The scientific naturalist typically
looks for hidden causal processes to explain the appearances; the liberal naturalist thinks that
there are depths in the appearances themselves that we may be unaware of which careful
criticism can bring to light. As Oscar Wilde observed, “It is only shallows people who do not
judge the world by appearances. The true mystery of the world is the visible not the
invisible”.22

7 Thick and thin explanations of art
Smith claims that the person-level explanations of film and other arts that we tend to offer in
aesthetic reflections and art criticism are “thin”. But as we have seen Smith has not made good
on his claim that going subpersonal is able to provide “thick” explanations of artworks or their
275

David Macarthur

aesthetic features. On the contrary, it seems to me that what Smith has put his finger on is not
the poverty of person-level explanations of art but the poverty of the psychological concepts and
explanations that are typically deployed within philosophical aesthetics. Anscombe famously complained that moral philosophy “should be laid aside…until we have an adequate philosophy of
psychology in which we are consciously lacking”.23 I take it much the same can be said of
aesthetics, especially the philosophical criticism of art that puts weight on psychology rather
than politics.24 One way of responding to this lack of an adequate philosophy of psychology
within aesthetics is to adopt a liberal naturalism which makes allowance for a deeper more
complex conception of the person, and of aesthetic experience, and of human psychology in
general – understood as manifest phenomena in the manifest image – without needing to
venture beyond the person-level. To provide an example of an explanatorily thick aesthetics
along these lines, I want to briefly discuss Stanley Cavell’s philosophical aesthetics.25 I will then
give an example of the fruitfulness of the resources of Cavell’s approach in making available a
person-level explanation of the condition of film spectatorship.
On Cavell’s conception, one’s experience of an artwork provides hunches, associations,
intuitions, glimmers, prods, beckonings and so on, that one is invited to take an interest in and
regard as a talisman capable of guiding one’s thinking and further aesthetic explorations.26 The
task that confronts one, in so far as one takes a reflective interest in one’s own experience, is to
make these inklings articulate, to put them into words. Cavellian aesthetics thus calls upon us to
take an active interest in the task of describing or articulating our own experience, to notice
differences and similarities and their nuances, however small, seemingly insignificant and undeclarative – in order to allow them to become significant, perhaps momentous, even lifechanging. Cavell explains,
My manner, namely to introduce a remark in a guise (calling attention to itself) meant to
mark an intuition that I find guiding, or whose obscurity or incompleteness is meant to
be undisguised, intended to remind myself in public, as it were, that I find significance
here that I have not earned, to which accordingly I know I owe a return.27
Rather than appealing to a philosophical or scientific theory as a source of (additional) explanatory weight, Cavell teaches us to look deeper into our own experience itself, to take it
seriously and to plumb its depths and conditions – for the results of this investigation may
challenge all existing theories.28 Cavell speaks of a method of “finding that an object of interpretation has become a means of interpretation”29; or, more fully and specifically, letting our
experience of the object of interpretation [an artwork] become a means of interpretation [of
that work]. A powerful or penetrating interpretation of a singular artwork is based in our
experience of that work.
To attempt to explain an artwork by seeing the work as an instantiation of a general theory is
to explain the work in terms of its being a kind or type – precisely a denial of its particularity.
And, anyway, why should we assume a given theory, presumably accepted on prior grounds, is
appropriate for, or adequate to the explanatory challenges of, the work? General theories are in
the business of providing general explanations based on causal laws or patterns or general rules.
The problems Smith’s naturalism faces include the limits of any theoretical understanding in not
having the resources to do justice to the particularity of the artwork;30 as well as the justification
for the general strategy of using neurophysiology as a key to understanding art of no matter
what category or quality or power.31 We might say Cavell wants to get us to observe at art from
close up, without theoretical preconceptions; Smith from the detached stance of scientific
theory.
276

Liberal naturalism and aesthetics

8 Cavell’s problem of the other and film spectatorship
Ironically, Smith himself suggests an explanatory strategy that can deepen person-level explanations of art, namely, to “subject our assumptions and beliefs to skeptical pressure” (52).
Cavell, once again, provides a useful framework to develop this insight – so long as we are
prepared to understand skepticism in unorthodox (not strictly epistemological) terms. For
Cavell, skepticism is an existential threat to mutual intelligibility and (because of that) selfhood,
which manifests itself in the human drive to a special form of self-defeat – the perversity of using
language to undermine the work we want language to do for us in everyday communicative
settings. For example, in Descartes’s Meditations, “metaphysical” doubts such as that we might
be dreaming now – a doubt that would be absurd to raise in ordinary contexts such as a court of
law – are employed to undermine all knowledge of the external world. Cavell supposes that in
order to be a self, one must express oneself in one’s home language in ways that can be acknowledged by others (fellow-speakers). Cavell then asks us to take seriously the question
whether we know each other, that is, whether we acknowledge (say, know and appropriately
respond to) the humanity in one another as expressed in word, gesture and action. In our
ordinary lives, we bear the responsibility for our knowledge of others and for responding appropriately towards them (attentively, sympathetically, justly, etc.) in light of this knowledge.
What Cavell calls “the problem of the other” is that we find this responsibility to know others,
or the demand to acknowledge them, burdensome;32 and so, we catastrophically disown our
knowledge, which leads us to wound others or into tragic failures of responsiveness. Cavell
draws the startling conclusion that “we live of skepticism” of others in our everyday lives
together.33
Suppose we accept this Cavellian way of seeing things. Then we might explain the power of
film in terms of allowing us to see others unseen, from a position that frees us of the responsibility to know or acknowledge them. Film spectatorship, from this perspective, provides a
respite from the ordinary condition of living together in community with its endless requirement to acknowledge others; where what is required to acknowledge others fully is
without any well-defined standard or clear limit. Film thus allows us a unique vantage from
which to explore our everyday condition of intersubjectivity, and the all too familiar slights,
forgettings, inattendings and all the other small and large fallings away from (full or adequate)
acknowledgement of others.34 To think of film in these person-level terms makes available a
thick explanation of it and of its importance in our lives. Yet I do not see how an explanation
offered in Cavellian terms of notions like skepticism, other minds, acknowledgement and the
like would be illuminated at all by going subpersonal. Nor do I see that any such move is
warranted or required. Being open to art and philosophy of art is being open to one’s
experience.

Notes
1 Wittgenstein writes, “In order to see more clearly, here and in countless similar cases, we must look at
what really happens in detail, as it were from close up” (2009, 51).
2 Most discussions of this issue that I am aware of assume a more or less reductive scientific naturalism.
So, for example, David Fenner (1992) identifies naturalism with materialism in articulating what he
calls a “modest aesthetic naturalism”.
3 Page numbers in the text refer to Smith (2017) throughout.
4 Smith speaks of his naturalism as a “cooperative naturalism” (3).
5 See De Caro and Macarthur (2004, 2010) for articulations and explorations of liberal naturalism; and
Macarthur (2017) for an argument that challenges the idea that there is a serious clash between the
scientific and manifest images of the world.

277

David Macarthur
6 Here I mean to echo Wittgenstein’s (2009, §116) methodological remark, “What we do is bring words
home from their metaphysical to their everyday use”, the context of which makes clear that this is
equivalent to bringing words “home”.
7 Smith is representative in only mentioning physics and biology. It is curious that chemistry is hardly
ever mentioned in the context of scientific naturalism. One wonders why.
8 A similar obfuscation is possible in the case of eliminativism since eliminativist ontology is compatible
with the retention of aesthetic discourse – and so, need not eliminate the surface linguistic phenomena
– on the instrumentalist grounds that it serves some non-descriptive function.
9 Davidson (2001).
10 Smith speaks of the “aesthetic properties and experiences that we are interested in explaining” (53).
The idea that aesthetic experiences might themselves be explanatory goes missing.
11 It is also important to consider what is meant by “explanation” in these contexts. For example, is it the
causal explanation of the perception of color or sound? Or is it aesthetic explanation of the meaning or
significance or value of the colors and sounds we perceive in film? And why speak in these abstract
terms? Isn’t it people and their actions or social interactions; or mountains, stormy skies, bridges,
buildings and music that we perceive in film? Admittedly we have to be able to perceive colors and
sounds to perceive these things but that kind of dependence does not really illuminate perceiving such
phenomena. It is a condition of such perception.
12 Oxford American English Dictionary.
13 Note that the existence of these complex emotions does not show that emotions are cognitive or reasonsensitive. Plausibly, basic emotions (e.g. anger, fear, joy, love, anxiety) are noncognitive and nonrational – an
insight found in Plato, Hume, James and Freud. We might further suppose that terms like “pride” pick out a
compound mental state made up of [belief + basic emotion] where the embedded non-cognitive nonrational
basic emotion is typically caused by the embedded belief. So pride might be the result of a belief that one has
achieved something notable or praiseworthy combined with a feeling of joy which it helps to bring about. If
that is so, then that the grammar of emotion terms is misleading in so far as it suggests that there are cognitive
emotions. See Talia Morag, “Skepticism about the ‘Cognitive Emotions’” unpublished MS.
14 Murray is aware of the issue but he skates over the difficulty it creates. As he puts it, empathy must be
distinguished from imagining myself in the situation of another (“in his shoes but possessed of my own
rather than his traits, states, and history” (183). He does not ask the question that inevitably arises here:
how do we know enough about the traits, states and history of the other to engage in the appropriate
imagining that constitutes empathy?
15 Stueber (2019, online).
16 Austin (1962, 2).
17 Jaegwon Kim (1993, 165–166) remarks:
The thesis that mental properties supervene on the physical turns out to be a conjunction of the
following two claims: the covariance claim, that there is a certain specified property of covariation
between the mental and the physical; and the dependence claim, that the mental depends on the
physical. But the thesis itself says nothing about the nature of what kind of dependency
it is; nor how the dependency grounds or explains the property covariation.
[Emphasis added].

In other words, a supervenience relation (which implies a dependency relation), by itself, is nonexplanatory!
18 Kant (2001, 5:215–216).
19 The “relaxed [aesthetic] naturalism” envisaged by Anthony Savile (2000, 63) may count as a form of liberal
naturalism. He remarks, “in the aesthetic arena [naturalism] does not find itself endorsing the existence of
supernatural entities and properties, nor does it endorse existences other than those we find cognitively
accessible” (61). This impression is strengthened by two further considerations: (1) his acknowledgement that
the aesthetic properties of things are normative (fn. 20) and (2) his aesthetic naturalism explicitly contests “a
scientism fostered by the spectacular successes of the natural sciences in the late nineteenth century” (63).
20 Science in this sense is empirical science. It does not presume to capture everything that people ordinarily call “science” since, I take it, there is a lot of confusion about what science is in popular
culture. I am defining “science” in a way that captures a lot of core things said about science by

278

Liberal naturalism and aesthetics

21
22
23
24
25
26

27
28
29
30

32
33
34

scientists; as well as connecting up with a plausible account of the rise, since the seventeenth century,
of a form of scientific intelligibility that depends on explaining the behavior of natural phenomena by
bringing them under some relevant causal law or pattern.
I cannot defend this Kant-inspired conception of art in the present context, but it has the important
virtue of making sense of the “timelessness” of art.
Wilde (2007, 19).
Anscombe (1958, 1).
The political criticism of art has its own problems one being that it tends to reduce art to a political
message and thus fails to account for the “timelessness” of art.
See, especially, the essays collected in Cavell (2002[1969]) and Cavell (1979b) which is devoted to his
philosophy of film.
Other strategies are available such as: (1) taking seriously a depth psychology such as that of Freud
which invokes the concept of an unconscious thus making available an expansion of one’s critical
vocabulary to include paradoxical states such as, for example, self-deception, rationalisation, projection
and objectification (2) unsettling the idea of a unified rationally ordered mind by following Plato,
Hume, James and Freud, in regarding emotions as disruptive and non-rational parts of the soul, capable
of dividing the mind against itself.
Cavell (2002, xxiv).
This is what Cavell (2004) thinks Wittgenstein’s Philosophical Investigations (2009) does: it explodes all
past theories of aesthetics as inadequate to its own aesthetics.
Ibid., xxii.
Smith is aware that “the particularity of art” poses an “important challenge” (16) to his naturalistic
approach but he fails to explain how the general explanatory strategies of the latter – he mentions Paul
Ekman’s concept of the blending of emotions – provide an adequate response to it. His discussion of
“nuance ineffability” in music might explain how “we can discern but not memorize very subtle variations in pitch” (16) but it does nothing to explain the significance and value of particular musical works.
To adopt the scientific point of view in aesthetics has the following (disastrous) consequences:
(1) objectification of the artwork and of our response to it; (2) a focus on the public, verifiable and
reproducible – training one away from one’s experience of the work and what it may reveal;
(3) abstraction of the artwork from its relations to other artworks and the history of art and (4) a requirement to explain by appeal to the universal rather than the particular. Smith wants to avoid these
implications by attempting to take seriously aesthetic experience and the aesthetic properties it reveals
but his commitment to scientific naturalism undermines this attempt.
Cavell (1979a, 430).
Ibid., 440.
It is worth noting that Cavell (1979b, 26) resists this line of thought on the somewhat questionable
grounds that viewers of film are not, on his view, required to acknowledge the human characters on
film. Cavell defends this idea on the basis that characters on screen are ontologically displaced from
being present to the audience by means of the mechanism of film projection. For more discussion of
this Cavell-inspired but non-Cavellian reading of film spectatorship, see Macarthur (2016).

References
Anscombe, G.E.M. Modern Moral Philosophy. The Journal of Philosophy, 33(124), (1958). 1–19.
Austin, J.L. Sense and Sensibilia. Oxford University Press. (1962).
Cavell, S. The Claim of Reason. Oxford University Press. (1979a).
Cavell, S. The World Viewed: Reflections on the Ontology of Film: Expanded Edition. Harvard University Press
([1971]/1979b].
Cavell, S. Must We Mean What We Say? Cambridge University Press. ([1969]/2002).
Cavell, S. The Investigations’ Everyday Aesthetics of Itself. In J. Gibson and W. Huemer (eds.), The
Literary Wittgenstein, pp. 21–33. Routledge. (2004).
Davidson, D. Mental Events. In Essays on Action and Events. Clarendon Press (2001).
De Caro, M. and Macarthur, D. (eds.) Naturalism In Question. Harvard University Press. (2004).
De Caro, M. and Macarthur, D. (eds.) Naturalism and Normativity. Columbia University Press. (2010).
Fenner, D.E. Modest Aesthetic Naturalism. The Journal of Aesthetics and Art Criticism, 50(4), (1992).
283–289.

279

David Macarthur
Kant, I. Critique of the Power of Judgment. Trans. Paul Guyer and Eric Matthews. Cambridge University
Press. (2001).
Kim, J. Supervenience and Mind. Cambridge University Press. (1993).
Macarthur, D. Living Our Skepticism of Others Through Film: Remarks in Light of Cavell. Sub-Stance,
45(3), (2016). 120–136.
Macarthur, D. Liberal Naturalism and the Philosophy of the Manifest Image. In A. Gare and W. Hudson
(eds.), For a New Naturalism. Telos. (2017). Chap. 4.
Morag, T. Skepticism About the Cognitive Emotions. Unpublished ms.
Smith M. Film, Art and the Third Culture: A Naturalized Aesthetics of Film. Oxford University Press. (2017).
Savile, A. Naturalism and the Aesthetic. British Journal of Aesthetics, 40(1), (2000). 46–63.
Snow, C.P. The Two Cultures: Enlarged Edition. Cambridge University Press. ([1959]/1998).
Steuber, K. Empathy. In E.N. Zalta (ed.), The Stanford Encyclopedia of Philosophy, Fall edition. (2019).
https://plato.stanford.edu/archives/fall2019/entries/empathy/
Wilde, O. The Collected Works of Oscar Wilde. Wordsworth. (2007).
Wittgenstein, L. Philosophical Investigations, 4th edition. Trans. G.E.M. Anscombe, P. Hacker, and J.
Schulte. Blackwell. ([1953]/2009).

280

24
LIBERAL NATURALISM,
AESTHETIC REFLECTION, AND
THE SUBLIME
Jennifer McMahon

1 Introduction: liberal naturalism and a perceptual affordance
The standard scientific naturalism and much empirical analytical philosophy modelled upon it,
would reduce normative entities like value and meaning to physical laws governing bodies and
the brain. This picture of the world acknowledges the laws and entities of the hard sciences such
as physics as the most objective, true and actual image we have of the world.1 According to this
view, other levels of explanation such as those whose terms are conceived to interpret and
explain phenomena at the level of mind, meaning, and morals can only be explained as pseudoimages of the world. Philosophers who accept this hierarchy refer to theories of mind, meaning,
and morals as error-theories whose objects of focus can be referred to as queer objects, and the
processes by which these objects engage us as non-cognitive.
Liberal naturalism, in contrast, is a philosophical position which places such normative
entities within the realm of nature. The methodology is largely analytic-empirical involving
conceptual analysis and explanatory theory compatible with, rather than reduced to, contemporary evolutionary theory and cognitive science; hence the emphasis on experience.2 R. F.
Crespo puts this in terms of the first person perspective, and argues,
the ‘first person perspective’ cannot be grasped by a restricted naturalism, and it
justifies the use of a liberal naturalism standpoint. Typically, liberal naturalism upholds
that some things are evidently shown by our everyday experience, from a first person
or ‘common sense’ perspective, and cannot be reached by the methods of natural
sciences, a ‘third-person’ perspective.3
A philosophy of experience treats seriously how we feel, emote, and what motivates us, and the
contingencies involved in what we know and consider right and good.
On the face of it, this may sound like business as usual for aestheticians. Afterall, isn’t this
what most analytic philosophers of art assume: that there are different levels of explanation even
though they each defer to compatibility with the scientific world-view. So how does enunciating the tenets of liberal naturalism change any of that? In fact, the reductive impulse of
scientific naturalism has so consumed and dominated philosophical thinking over the past

DOI: 10.4324/9781351209472-29

281

Jennifer McMahon

century that the very terms we use to think about such aspects of experience as meaning, value
and pleasure are depleted of much of their content. Philosophical aesthetics is a case in point.
Metaphysicians and epistemologists grounded in scientific naturalism have arguably dominated
the philosophical scene for the last century. This has had consequences for the shape of topics and
debates that have characterized philosophy and in particular, philosophical aesthetics. Philosophical
aesthetics has been forced to operate as a kind of error theory, its objects treated as non-existent or
superfluous, and the defining mental processes engaged by aesthetic objects as non-cognitivist.
Aestheticians have either adopted terms more or less defined by such classifications, or alternatively,
introduced convoluted strategies to reduce relevant items to scientific entities. The interpretations
developed or facilitated through a more liberal naturalism must of course be compatible with
scientific enquiry into mind and brain, but the level of explanation appropriate to understanding the
objects in question is the level of intentional objects rather than physical structure or computation.
To say for example, that particular neurons, neural structures or even certain computational
functions are activated while enjoying the view or while engrossed in a fiction, does not contribute
anything to the kind of understanding with which philosophical aesthetics is concerned. Pleasure,
for example, is assumed by the more restricted naturalism to be a basic sensation; the kind of
sensation we have in common with other animals. The extent of its efficacy is to instinctively draw
us toward its brute source and away from its opposite. Treating pleasure as an irreducible aspect of
experience is taken straight from scientific naturalism. However, it does not take particularly keen
observational skills to notice that different culturally based behaviors suggest that we learn to take
pleasure in certain things, and our pleasures can be composed of various associations, attitudes and
ideas surrounding their objects. In other words, pleasure can be intentional. In philosophical aesthetics this is an idea that fails to gain much traction largely because the domain is dominated by
attempts at analytical respectability. The approach taken is to toe the line of thought which reduces
pleasure to a non-cognitively generated sensation and hence avoid any conception whose terms
might sound in error or queer to the scientific naturalist’s ear. For such an aesthetician, mental states
such as “intentional pleasure” that are not limited to the terms of scientific explanation simply do
not exist. In contrast, liberal naturalism according to John McDowell who coined the term, does
not require the integration of all our capacities into a narrow scientific framework.4 Instead, in
defence of trusting our observations, David Macarthur writes:
A more plausible naturalism [than the standard scientific naturalism] would not begin
by dogmatically assuming that nature is exhausted by a restrictive conception of scientific nature – posited by successful scientific theorizing – if for no other reason than
that naturalism rightly aspires to rid philosophy (so, of course, itself) of a priori
dogmatism – what Quine rightly castigates as ‘first philosophy’ – in favor of more
practice-based empirically sensitive conceptual thinking.5
I will argue that once aesthetic experience is clearly analysed in terms of a more liberal naturalism, the result is that the conceptual shackles imposed by notions of error theory, queer
objects and non-cognitivism (the philosophical implications of scientism) are removed to reveal
new understandings of key concepts in aesthetics.
We can begin by drawing an analogy between an ethical liberal naturalism and an aesthetic
liberal naturalism. According to Hans Fink, there are several forms of naturalism, and he follows
McDowell in arguing that “one specific form of naturalism which is very common in modern
philosophy, … [is] based on an unduly restricted conception of nature and bound to misrepresent the ethical” and we might add, bound to misrepresent the aesthetic. And this is
scientific naturalism. Fink writes that:
282

Liberal Naturalism

An ethical naturalist is someone who insists on a fundamental continuity between the
ethical and the natural. Ethical values or norms can and should be accounted for
within the realm of nature and in terms of or based on ordinary natural facts.”6
The next line should not be indented to mark a new paragraph. Please check my
original document and resist arbitrarily starting new paragraphs.
The same could be claimed for aesthetic values. However, as Hilary Putnam has argued, scientific naturalism finds no place for such unobservable objects as values and mind.7 In this
chapter, I will refer to the narrow and restricted kind of naturalism, as the scientific image.
The scientific image treats aesthetic experience as a matter of private reverie or
mindless gratification of some kind. The aesthetic experience by this image is just a
feel-good state understood in terms of individual brains. An interesting upshot of this
is the chasm that grew up between the twentieth century formulations of aesthetic
terms compared with earlier understandings. A key aspect of aesthetic experience
which interested those eighteenth-century philosophers responsible for setting an
agenda for philosophical aesthetics, and carried through by later philosophers and
poets who wrote essays on their practices in the nineteenth century, was that aesthetic
experience involves a process like perception or a judgment which structured the
subjective in terms of certain objective constraints. The details as theorized varied
between writers but generally entailed comparisons and concepts of enculturated
persons. Without taking on the various metaphysical commitments of their eighteenth
and nineteenth-century authors, we can still recognize the cognitive and intentional
aspect of their construal of aesthetic experience.8
The above indented text should not be indented but only the first line to mark a new
paragrpah. And the sentence below this - beginning with “For example, “should
continue on, in the same paragraph.
For example, according to Immanuel Kant, the pleasure that accompanies the representations of
aesthetic reflective judgment is a kind of cognition (as opposed to a sensation).9 As argued
above, an aesthetician shaped by the philosophical assumptions of reductive scientific naturalism
simply cannot understand this. In contrast, a liberal naturalism demonstrates much greater
explanatory power for how we engage with the world and develop cultural objects. Once
intentionality is woven through our responses, aesthetic experience can be understood in
contemporary terms, as cognitive.
The reductive tendencies which took hold in twentieth century philosophy led to aesthetic
experience being treated as non-cognitive. Ironically, the conditions were then set for the rise
of “cognitive aesthetics” which attempts to gain some respectability for aesthetic engagement
by changing the object of aesthetic experience to literal or scientifically reducible content. But
this approach distorts the objects and processes of aesthetics by mistakenly accepting the way
feelings and attitudes are caricatured as almost epiphenomenal by the scientific image. An
aesthetics which understands value and meaning as foundational rather than secondary does not
need a theory of “cognitive aesthetics” because all aesthetic experience can be understood
accurately to be intentional.
Liberal naturalism is not a new approach; one could cite a number of recent philosophers
who work in this vein such as Stanley Cavell, McDowell the later Putnam. They all recognize
that their approach introduces an alternative type of naturalism to scientific naturalism.10 And
283

Jennifer McMahon

there is a case to mount for interpreting many historical philosophers in terms of liberal naturalism as for example regarding the most influential aspects of Kant’s aesthetics.11 Nonetheless,
it took David Macarthur and Mario De Caro to bring its tenets into focus as a distinct metaphysical position and to articulate its philosophical relation to the scientific image. Macarthur
argues that “getting clear about the 2nd person stance and what it makes available provides a
way of exploring the realm of natural (i.e. non-supernatural) non-scientific items that liberal
naturalism makes newly available.”12 By second person, Macarthur aims to draw out “an intersubjective form of intelligibility”13 according to which the agent’s perspective is as explanatory of our reality as the explanatory concepts of science.
Arguably an exemplary concept conceived in terms of a more liberal naturalism is the
concept of perceptual affordance. This has a long history, but can reasonably be understood to have
taken its place in mainstream vision-theory through James J. Gibson’s 1966 book The Senses
Considered as Perceptual Systems. It is the idea that an object is not just perceived in terms of
sensuous properties like shape and form, but also in terms of the possibilities it offers for the
perceiver. I will refer to the latter as involving a blueprint for action. So according to the theory
of perceptual affordance, object recognition does not simply entail labelling an object with a
correct name but in addition to this, responding to the blueprint of action that it affords. Ripe
fruit suggests eating, a baby kitten warm regard, an injured person sympathy or assistance and so
on. The situation may prevent actual active response – the ripe fruit may be on the desk of your
bank manager - but the affordance, as Gibson explained it, entailed this blueprint. As such, a
concept of any object includes the actions it motivates or the orientation it prompts.
The aesthetic concept which is the topic here is arguably the most foundational of perceptual
affordances. The kind of action and orientation that it indirectly motivates is required for the kind
of complex sociability required of human cultures. Perceiving an object as sublime involves
existential themes. Under liberal naturalism, certain key aesthetic concepts such as the sublime
take on a new significance, because they represent the process whereby meaning and significance
are ascribed to objects.14 Gibson’s affordance was something that evolved in an enculturated
person over-time as they interacted with the environment under adaptive pressures (those who
perceived fruit as occasions for sustenance lived long enough to reproduce etc.) but importantly
also incorporated cultural norms into the blueprint for action that the object afforded. In the case
of aesthetic experience of the sublime, the affordance is an underlying attitude to the relation
between the world, ourselves, and others. As such, analyzing the sublime in terms of a perceptual
affordance reveals that our experience of the sublime like other idea-generating aesthetic experiences expresses our rationally normative relation to the world.
The experience of the sublime motivates a feeling of respect and responsibility through a
particular kind of pleasure. This is an intentional pleasure; a pleasure understood as such by
liberating naturalism from a dogma that would reduce aesthetic pleasure to instinct. It was the
normative aspect in the way in which our pleasures could be cultivated as demonstrated by the
sublime, which would seem to have interested Immanuel Kant and which continues today to
motivate philosophical commentary on the sublime by many Kant scholars. However, many
contemporary debates are hampered by unacknowledged variations between the relevant
background theories held by the protagonists, on topics like pleasure, imagination, and perception. By treating the sublime as a perceptual affordance I avoid the contortions which arise
from attempting to reconcile perception with the ascription of values and ideas; that is, I avoid
the theoretical difficulty of explaining how the world can be perceived as amenable to our
hopes and ideals.
In this chapter we will consider how treating the sublime as a perceptual affordance clarifies the
historical approaches which treat the sublime as an experience of our agency and avoids the
284

Liberal Naturalism

reductive thrust of many influential twentieth century theories in the philosophical literature
which are hampered by the reductive notions of pleasure and reflection. The benefit of accounts which recognize their foundations according to liberal naturalism, is that the feeling
involved in such experiences can be conceived as interwoven and shaped by cognition. This
provides a basis for remapping certain key terms in philosophical aesthetics. As such, the theory
of the sublime provided here, demonstrates the greater explanatory power of liberal naturalism
in contrast to a bald scientific naturalism.
The arguments of the following sections will include a reframing of reflective judgment in
terms of perceptual affordance (section 2); intentional pleasure (section 3); non-perceptually
represented perceptual properties (section 4); and intersubjectivity (section 5). Finally an account of the way this re-envisaging of the relevant terms results in an economy of means for
linking the sublime and moral motivation is provided (section 6) before drawing together the
strands of each section to reveal a liberal naturalist account of the sublime.

2 Reflective judgment
The way we currently understand the nature of the sublime is attributed in broad outline to
Kant’s account in the late eighteenth century. The focus is typically on the non-cognitive aspect
included in the standard interpretation of Kant’s aesthetic theory but the interpretation provided here shifts this emphasis somewhat by noticing how the relation Kant articulated between
various kinds of judgments includes a type of judgment which arguably translates as a perceptual
affordance; and so in our current framework, a cognitive judgment. While Kant adopted the
descriptions of the sublime from other philosophers such as Edmund Burke15, he reasoned what
would need to be the case regarding the system of the mind in order for such experiences to be
possible. In doing so, he contributed a refinement to what was noticed about the sublime, and
what was considered meaningful and significant about such experiences.
Kant saw in the sublime the human capacity for awe and wonder but with an interesting
twist. He thought the awe and wonder while prompted by a natural object initially had a
sobering quality to it, which we nonetheless enjoyed on some level due to the particular type of
narratives we typically conjured in response. As such the sublime was not a sensation but a
judgment. Kant uses the term “judgment” not to denote a measure, grading or final evaluation,
but rather to denote a particular kind of intellectual operation that is conditioned upon having
had prior cognitively shaped cultural experiences.
Kant distinguishes between two kinds of judgment, one determinate, and the other reflective.
A determinate judgment involves identifying something in the world; picking something out.
This is possible according to Kant in virtue of a concept we hold of the thing picked out. The
process of perceiving the thing involves subsuming the visual array under this concept. It is the
concept which both selects or registers the relevant details and presents the item to us in a way by
which it acquires meaning for us.
In reflective judgment, on the other hand, one looks but without the kind of focus which
predetermines what one perceives. So while in determinate judgment it is as if one simply
registers what is out there; in contrast, in reflective judgment it is as if the meaning and significance of the object is ascribed by us, providing the object with its significance for us. Most
importantly, the experience though engaging a heightened form of subjectivity, feels as though
it is objectively significant, that is, almost as though the object were expressive of the ideas we
summon forth.
Among reflective judgments Kant distinguishes three kinds: judgments of beauty, judgments
of the sublime and judgments of purposiveness. Another way of thinking about these reflective
285

Jennifer McMahon

kinds of judgment is that they involve ascriptions which fall into general categories that are
applied universally. “Ascriptions” are applications of meaning and significance, as opposed to
straightforward descriptions or referents. That is, we should think of them as resulting not in
perceptual givens as we do in the case of objects like tables and chairs (albeit with their particular “affordances”), but rather more fundamental perceptual conditions that we apply universally such as finding the world conducive to our interests (beauty); finding that our humanity
allows us a freedom not available to non-human life (the sublime); or that the world is
knowable and we have a desire and means to know it (purposiveness). In other words “ascriptions” involve evaluative content: the attitudes and orientations we bring with us that constitute
in part the meaning we give to the object. An “ascription” is not a label, nor is it simply a
blueprint for action if this is conceived as formulaic; but rather, an orientation without which
any blueprint for action would not take hold. In cases of the sublime, it is not simply that the
object is a functional object like a chair (whose affordance would include that it take our
weight), but that in addition, it would need to occasion reflection on our human status in the
natural world. Clearly, chairs are not typically the kind of thing we find sublime.
So here I use “ascription” to emphasize the way “affordances” are comprised by learning
from experience including culturally specific learning, even if the content is typically constrained by universals. In the case of aesthetic concepts like the sublime, the ascriptions reflect
the need for beings with life-questioning capacities to develop constructive narratives around
conceptions of self, given the prima facie dispassionate and alienating face of nature. The need for
this kind of elaboration has been recognized in various philosophical contexts.16 But none of
these debates and discussions acknowledges that much of the theoretical work has already been
done by the concept of perceptual affordance.17
Understanding reflective judgments in a contemporary theoretical context proves quite
helpful in distinguishing between ungrounded belief and what Kant might have described as
thinking “of something supersensible in a way which is serviceable to the experiential use of our
reason” ([AK 8: 136-37] 1996: 10).18 My aim is to present the sublime as an experience the
nature of which demonstrates a case where perceptual experience involves affordances which
invite certain kinds of interactions including the identification of certain ends. In short, the
condition for the sublime is having the ability to think, and the structures of mind and experience that this entails. The sublime involves finding the very perception of certain phenomena sufficient to prompt a sense of ourselves as more than the sum of our instincts, and this
is both sobering and liberating.

3 Intentional pleasure
The experience of the sublime is usually discussed in terms of natural objects or events which
are a potential threat to us. They might be threatening by their monolithic size or by their
potentially destructive power. On the face of it, such objects or events do not put us in a
pleasant state of mind. And yet, we seem to enjoy them for this very unpleasantness as long as
we are actually out of immediate danger.
All philosophical accounts of the sublime include terms which reflect elements of pain and
pleasure. Such accounts vary on whether these are successive components or combine in a
sobering kind of pleasure. Most of the debates around the sublime begin by attempting to justify
a position on this question. But the way pain and pleasure themselves are understood impacts
upon how this question is answered. Consider that the feeling of the sublime seems to precede
any thoughts we may have about it. Yet a certain enculturation into ideas of nature as
“landscape” or nature as a metaphor for emotional states, condition experiences of the sublime.
286

Liberal Naturalism

Experientially, the sublime engages us because of the objective stimulus it represents. It might
just be a towering edifice or a stormy sky. To be sublime though it must engage us subjectively;
for example, we feel the sublime object’s magnificence in its size or power.
Burke explained the initial feeling of the sublime rather reductively as fear, while Kant
thought it was a feeling of being inadequate to the task of grasping the form of the object’s
immense size or power.19 According to Kant, the structure of the experience prompts an
analogy between physical size and power in nature on the one hand, and the immense scope
and freedom of the human mind on the other. Most significantly for Kant, the sublime prompts
an appreciation of the contrast between the agent-less forces of nature compared to our own
agency. As Kant explains:
[I]n our aesthetic judgment nature is judged as sublime not insofar as it arouses fear,
but rather because it calls forth our power (which is not part of nature) to regard those
things about which we are concerned (goods, health and life) as trivial, and hence to
regard its power (to which we are, to be sure, subjected in regard to these things) as
not the sort of dominion over ourselves and our authority to which we would have to
bow if it came down to our highest principles and their affirmation or abandonment.
… To be sure, this principle seems far-fetched and subtle, hence excessive for an
aesthetic judgment; but the observation of human beings shows the opposite, that it
can be the principle for the most common judgings even though one is not always
conscious of it. ([AK 5: §28, ‘262] 2000: 145).
After Kant, the sublime is formulated as a perceptual experience involving an inhibition of our powers or emotions, followed by their release, by which one experiences a relief-type pleasure. But their release requires some intellectual input
which would seem to presuppose a facilitating kind of pleasure. As Kant puts it, the
sublime involves an “[E]motion, a sensation in which agreeableness is produced only
by means of a momentary inhibition followed by a stronger outpouring of the vital
force” ([AK 5: §14, ‘226] 2000: 111). It might be worth mentioning here that while
Kant does analyse the sublime in terms of universal principles, grounded by the moral
law and the idea of freedom within us ([AK 5: §29, ‘275] 2000: 156), it is consistent
with Kant’s account to envisage the manifestation of such principles as quite varied
across different cultures and sub-cultures, hence compatible with the normative relation between us and the way we represent objects.20
Paul Guyer argues that Kant equivocates on the role of propositional content when discussing
the dynamical sublime (an experience of power in nature rather than size, the latter the
mathematical sublime according to Kant). Guyer thinks the text may suggest that a judgment or
propositional attitude occurs in response to the inhibition, the ameliorating ideas of which give
rise to the experience of release (1993: 213-14 [AK 5: §28, ‘260, ‘264]). Kant thought the
feeling of being inadequate to the task of imaging the object, automatically gives way to a sense
of our superiority over the determinism of nature and this is felt as a release. It seems reasonable
to propose, and Guyer thinks there is textual evidence to suggest, that Kant thought the sense of
release was prompted by a certain narrative generated by the perceiver (Kant [AK 5: ‘257]).
Both Burke and Kant, along with other eighteenth century authors on the sublime, thought
that the response to the sublime was universal, though in Kant’s case, given certain conditions.
This was explained by Burke in terms of survival instincts but in Kant’s case by the structure of
the mind. Kant postulated a mind structured in such a way that our rational selves could be seen
287

Jennifer McMahon

to be freely oriented even if embedded in a larger system of physically determined laws.
Nonetheless Kant thought that though everyone had the potential to experience the sublime,
unlike beauty the sublime required enculturation. For Kant, only those capable of responding
to reasons and meanings, in other words, a rational normativity, could experience the sublime
and he associated this with enculturation: that is, growing into one’s society and away from
being determined by appetite and self-interest.21 Kant wrote:
[W]e cannot with the same readiness count on others to accept our judgment about
the sublime in nature [as for the beautiful]. For it seems that, if we are to pass
judgment on that superiority of [such]natural objects, not only must our aesthetic
power of judgment be far more cultivated, but also … [i]n order for the mind to be
attuned to the feeling of the sublime, it must be receptive to ideas ([AK 5: ‘265]
1987: 124).
There is a difference between concepts and ideas in Kant’s system of the mind; suffice to say
here that the ideas he had in mind were of a particular kind. In the next paragraph he writes:
But the fact that a judgment about the sublime in nature requires culture … still in no
way implies that it was initially produced by culture and then introduced to society by
way of (say) mere convention. Rather, it has its foundation in human nature: in
something that, along with common sense, we may require and demand of everyone,
namely, the predisposition to the feeling for (practical) ideas, i.e. to moral feeling ([AK
5: ‘265] 1987: 125).22
The issue of the ideas involved in the sublime is a knotty one. While commentators accept the
role of ideas in the sublime, their discussions and analysis do not always make it entirely clear
what role the ideas play, at what point they come into the experience, and of what they actually
consist. There is general consensus that one does not need to contemplate any particular set of
ideas in order to prompt an experience of the sublime. The perceptual object itself prompts the
relevant experience through the mechanism of a perceptual affordance as explained in the previous
section. But it does seem that the experience simply manifests as the kind of experience that
would be explained if we were entertaining certain kinds of ideas. I will elaborate further in the
next section.
For now consider that, in the context of the tenets of liberal naturalism, the sublime is of
interest because the assumption of philosophical accounts of aesthetic concepts as construed by
Kant, is that perceptual experience is itself theory laden or at least is in part constituted by
affordances, and these affordances, where aesthetic experience is concerned, involve prescriptions to value and act in certain ways. These affordances are culturally based to varying degrees,
even when they answer to desires which could be deemed universal in beings like us with
reasoning powers. If it is true that everyone needs to feel their reasons for living reflected in
nature and society, the narratives that serve this end might vary between cultures or traditions.
But this narrative is the affordance that characterizes the sublime. As such the sublime demonstrates that to explain the pain and pleasure of the sublime requires the explanatory power
of the first and second person perspectives available through a liberal naturalism. What is also
clear, is that after Kant, the pleasure involved is not instinctual as Burke may have postulated,
but instead a pleasure that is taken in the object in virtue of enculturation. Regarding the
pleasure taken in beauty which also applies to the sublime, Kant writes:

288

Liberal Naturalism

only that of the taste for the beautiful is a disinterested and free satisfaction; … [in
contrast] [a]n object of inclination and one that is imposed upon us by a law of reason
for the sake of desire leaves us no freedom to make anything into an object of pleasure
ourselves ([AK 5: 210] Guyer 2000: 95).

4 Non-perceptually represented perceptual properties
The explanation given for the tenor of the experience of the sublime is that we actually have
certain ideas in front of mind that become so interwoven into the perceptual experience of the
object that we feel that the object is expressive of these very same ideas. This is what is meant by
“the ascription” of certain narratives to an object. There is a certain corroboration between
philosophers, poets and artists regarding the kind of ideas involved. Even before Kant, an
appreciation of landscape as sublime was understood to be an indication of an enlightened
mental state and elevated moral disposition.23 And by the time of the early nineteenth century,
after Kant, the influential and much acclaimed British poet William Wordsworth thought that
while poetry must necessarily be pleasurable, it was a pleasure we took in the way poetry
heightened our interest in moral relations.24
As discussed earlier, philosophers after Kant have postulated that the experience involves a
sense of some kind of constraint (a sense of our frailty or physical limitations) evoked by the size
or might of an object. In response our way of representing the object mitigates these feelings
with ideas of some kind. The thing is though, our awareness is of the object rather than ideas
peculiar to us. That is, the ideas summoned up to mitigate the negative aspect are experienced
as if the sublime object is expressive of them. How is this possible? According to Bence Nanay:
A major question in philosophy of perception is about which properties are perceived
and which ones are inferred or non-perceptually represented. Beliefs can represent
their objects as having any property. In the case of perceptual states, in contrast, the set
of properties they represent their objects as having is limited. The question is how
limited this set of properties is. Color is a good candidate for a perceived property,
whereas being made in Australia is a good candidate for a non-perceptually represented one. But there are many kinds of properties in between that are more
difficult to categorise (2018: 53).
The sublime could reasonably be considered as “in between”. Like color, the subjective basis is
sub-personal, but unlike color, it seems to come with a more fleshed out narrative even if this
narrative is brought to bear unconsciously, with the effects of the narrative felt, rather than read
off, so to speak. However, it is the case that in certain cultures, the perception of a certain color
can afford certain values such as luck, spirituality, purity and so on. This is closer to the sublime
than if we simply perceived color as purely descriptive, a means to object recognition. But the
structure of the sublime as theorized suggests a more palliative narrative. As such, rather than
posited in terms of a property, perhaps it is more accurately conceived as a propositional attitude
or value. Nanay continues with some examples of what he has in mind concerning the
properties which fall in between perceptual and non-perceptually represented properties:
[I]t has been argued that we perceive objects …, as being causally efficacious …, or in
terms of the way [they] ... function for us so for example, edible, climbable or Q-able
in general …, or as having some kind of normative character or value …, as having
dispositional properties …, and as having moral value …. (Nanay 2018: 53-4).
289

Jennifer McMahon

The notion that value is a component of our descriptions is given more and more attention
across all areas of philosophy. Or put another way, attitudes saturate our perceptions. In this
vein, Stephen White argues that “our capacity for action presupposes that affordances are a part
of our perceptual experience” (2004: 218); and he argues this based on the inability to find the
basis of motivation in objective, descriptive beliefs when understood as untainted by human
interests. That is, if we did perceive things only in terms of objective properties isolated from
any affordances or ascriptions which imbue the object with their purposes, meaning, significance and value for us, it is difficult to envisage how we motivate our actions at all.
Unless perception involves the ascription of meaning and significance to particular objects,
the world would indeed seem an alienating place. But the sublime demonstrates the way
narratives become embedded within our worldviews and shape our perceptions and experience.
As we have seen, the structure of the sublime prompts the ascription of a certain narrative to the
object in perceiving it; or at least our response to the sublime would be explained if this were
the case.
While Kant refers at certain points in his Critique of the Power of Judgment to the super-sensible
substrate of humanity which he understands as the ground for human agency, it does not distort
his analysis of aesthetic concepts such as the sublime by instead grounding this capacity in the
higher cognitive powers of human beings. In fact, when he introduces the notion of the Sensus
Communis in the “Critique of Aesthetic Judgment”, he provides the means whereby the natural
tendency to sociability in us leads to reflective judgment: a judgment that one compares with
reason in general (what it is assumed others would judge) ([AK 5: ‘293] 2000: 173). This kind of
judgment draws us toward the kind of exchanges that foster the conditions required for calibration of values and meaning. A liberal naturalism is “on offer” in Kant’s aesthetic theory as he
treats aesthetic experience in terms of reflective judgments that exhibit intentionality and hence
are normative rather than lawful by our contemporary theoretical framework.
In order to consider the explanatory power of the above account of the sublime, imagine
you are visiting some extraordinary natural site such as the Grand Canyon in Arizona. Imagine
yourself standing there, looking directly at it. You would be expected to have certain feelings in
response to its sheer size and the various details of the terrain. This would be expected to hold
your attention and keep you riveted to the spot. You would also be expected to have certain
kinds of thoughts that could be understood to be tied to the feelings aroused by the object.
What kind of thoughts and feelings would you consider apt given the view?
Imagine that you are quietly gazing out over the Grand Canyon and you explain your
thoughts as taken up with an impending tennis match you have been anticipating. Most people
would take this as a sign that you were distracted from the view even if you were looking in its
direction. Or instead, imagine you merely described the shapes and colors, lines, form and
texture, without any evaluative language. Many would regard this kind of response lacking in
some respect, as if you were missing something. In contrast, if you explained your thoughts in
terms of the grandiosity of the big themes of life and suggested the view simply made day-today concerns seem trivial, most people would understand you as having been deeply moved by
the view. It is what makes the latter content apt that is of interest to us here.
Conceiving of the sublime as a perceptual affordance provides a contemporary version of Kant’s
account. The experience is perceptual and entails ascribing meaning and significance to the
object as though it were simply perceived in it. Kant wrote:
[t]hat which, without any rationalizing, merely in apprehension, excites in us the
feeling of the sublime, may to be sure appear in its form to be contrapurposive for our
290

Liberal Naturalism

power of judgment, unsuitable for our faculty of the presentation, and as it were doing
violence to our imagination, but is nevertheless judged all the more sublime for that.
([AK 5: ‘245] 2000: 129).

5 Intersubjectivity
Above we saw that the structure of the sublime involves a negative and positive aspect,
characterized broadly as having a sobering quality which is either enjoyed itself or at least
prompts a more enjoyable response. In some accounts it is as though the sobering quality is
converted into pleasure. This might seem to be Burke’s idea that we feel fear but on realizing
our safety from any actual threat, we enjoy the stimulation of fear. But on closer scrutiny it
would seem fear must be held in mind in order for this stimulus to be enjoyed even if it is
accompanied with safety-securing narratives.25
The positive aspect to the experience of the sublime relates to a unique feature of humanity
which is being able to choose one’s goals and intentions. This is both exhilarating and sobering,
according to the Kantian sublime. How we explain this and even the content of our reflection is
a matter of debate among commentators. However, that the sublime is complex but pleasurable, and is the kind of experience we find deeply moving, is beyond dispute among those who
write on the sublime.
Katerina Deligiorgi accepts the Kantian conception of the two aspect nature of the sublime
where the positive response is primed by some sense of our physical limitations to which we
respond by thinking of our own cognitive efficacy. She argues that “the mere thought that we
have the capacity for agency” (2014: 32) evokes the pleasure of the sublime. Paul Crowther also
adopting the Kantian view, argues that to ameliorate the negative aspect of the sublime experience, a sense of our freedom from nature’s determination is evoked, and this facilitates
respect for persons, as we reflect upon this aspect of humanity: that human beings can
“comprehend things which far exceed their sensible capacities” from the inside as it were (1989:
173). For Sandra Shapshay (2013) in a similar vein, the experience of the sublime focuses our
attention upon the object and our relation to it and by analogy our relation to and position
within the world. Deligiorgi, Shapshay, and Crowther, all have accounts which incorporate
reflection as part of the sobering aspect which is furthered by the pleasure of heightened
subjectivity. The sources of pain and pleasure in the sublime involve objective and subjective
components, both instinctual and intellectual. However, the ameliorating ideas projected onto
the object rely on cultural context for their content (ideas internalized from our interactions
within our communities), and so are more accurately understood as inter-subjective.
Malcolm Budd (2002) furthers his own earlier account of the sublime (1999) which had
been very influential in later twentieth century Anglo-American aesthetics. He sets aside the
more architectonic aspects of Kant’s account and adopts the reductive approach of Burke’s
account. However unlike Burke for whom the dangers of nature were the main objective
feature, Budd sides with Kant in as much as he pinpoints the monolithic or powerful in nature
as the prompt, but deviates from both Burke and Kant in characterizing the negative aspect as a
feeling of insignificance in the face of nature’s power and might. Rather than feel pleasure in
being safe from the dangers of nature as Burke suggests, Budd argues that we feel a certain relief
in the feeling of insignificance prompted by sublime objects in nature. So like Burke, no special
gravitas just a relief-pleasure which is necessarily characterized by an absence – in this case not
Burke’s absence of threat or fear but the absence of responsibility or performance anxiety – and
so the pleasure is somewhat laced through with a sense of our limitations which in this case is
meant to be reassuring. It is as if we can snuggle into ourselves knowing that we cannot be
291

Jennifer McMahon

expected to achieve the power and might of such extra-ordinary natural things. But this account does not do so well in accounting for the kind of reflection associated with the sublime.
A relief-pleasure ends reflection rather than prompts it.26 We could extend this objection to
Burke’s account also.
Ronald Hepburn (1996) who was considered something of a guru on natural aesthetics last
century, treated the subjective component along Kantian lines involving a sense of our superiority over nature. But in a significant break from Kantian tradition, he treated this reflection
as a kind of private reverie and as such the content of the reverie could well be idiosyncratic. It
was typical during this period in Anglo-American aesthetics to treat aesthetic experience as
bordering on a kind of sensuous gratification with the individual treated as an isolated primary
unit. Aesthetic reflection was often treated as private reverie like day-dreaming and Hepburn’s
conception is a case in point. But if this were the case, we would not have responded with
perplexity when the peruser of the Grand Canyon reflected upon an upcoming tennis match or
focussed their full attention upon line, shape and texture. In contrast to Hepburn’s view, the
content of the sublime experience exhibits inter-subjective characteristics and so we do expect a
certain kind of shared mental content to accompany the experience. And Liberal Naturalism
provides greater explanatory power in accounting for such observations.
It is of interest here to compare Budd and Hepburn’s accounts with those of Delgiorgi,
Crowther and Shapshay’s. Budd and Hepburn’s views are neatly contained by the kind of
scientism it is argued in the first section of this chapter, characterized and limited philosophical
aesthetics last century. In contrast Delgiorgi, Crowther and Shapshay put the tenor of the
experience first and develop theories to account for it, rather than allow metaphysical dogma to
limit what they recognize in the experience. In this they exercise a more liberal naturalism, and
it must be said, achieve greater explanatory power by their accounts.
Nonetheless there is another question which separates the more liberal naturalist accounts.
The question of the nature of the relevant mental content has led to debates around whether
the experience is inwardly or outwardly focussed (Deligiordi 2014: 30). The problem is theorized in terms of how the subjective component could be objectively focussed. However, such
problems arise from other background assumptions, in this case by omitting perceptual properties which are inferred or non-perceptually represented, as discussed by Nanay in the previous
section above. If we think of the mechanism of the sublime as a perceptual affordance, then the
way cultural attitudes and dispositions add content to and in turn shape our experience is
explained (Kant introduced the Sensus Communis to account for this as discussed in the previous
section).27 Reflection which is part of the defining aspect of the sublime does typically manifest
in ways which are influenced by our personal predicament. But nonetheless, that the content of
the reflection tends to exhibit existential themes relative to one’s time in history and culture, is
understood to be universal by these writers on the sublime. A perceptual affordance is necessarily
object centred but hooks our subjectivity into the object.
When considered as a perceptual affordance, the distinction between object-centredness and
subjectivity becomes less sharp. The non-perceptual nature of the ideas we associate with the
sublime are experienced as though they are expressed by the object, and orientate us in a certain
life-reaffirming way. Kant avoided collapsing the perceptual experience into inward reverie even
though he argued that the sublime was of the human mind rather than an objective property (AK
5: ‘245). The point of the sublime for Kant was that unless we hold certain assumptions about the
world and society, we would not orientate ourselves in the appropriate way to want to know the
world or cooperate with other people. For Kant this orientation is given a supersensible ground
through the moral law within us, even though this law can be manifested through aesthetic
reflective judgments prompted by certain natural objects. But moving away from Kant’s
292

Liberal Naturalism

metaphysics, we can see that certain orientations to the world must be assumed in order for us to
function effectively. Through a liberal naturalism, we can explain how we orientate ourselves as
we do by assuming that we acquire certain attitudes or what we would now call affordances
through certain experiences. Aesthetic experiences like the sublime are bound to be had as they
promote an elevated sense of humanity and the pleasure we take in this is assumed to be shared,
and so the opportunity to have them will be facilitated in human culture. Scenic routes, look-outs
and the cultivation of natural landscape destinations are relevant examples.
In sum, so far we have seen that the idea of the sublime is that feelings aroused in
us by certain natural objects are feelings that orientate us to the world in ways conducive
to our flourishing. Furthermore I have argued that questions regarding the structure of the
sublime – the relation of pain to pleasure, the role of ideas, and the relation of objectivity and
subjectivity – can be answered if the sublime is understood as a perceptual affordance.

6 The Sublime and morality
As we have seen, while there is broad agreement among those who accept a Kantian theory of
the sublime concerning the tenor and significance including that the pleasurable aspect involves
some kind of ameliorating narrative generated by the perceiver, there is some variation concerning the details of this narrative. While most Kantian influenced accounts understand the
content to orientate us toward but above nature in some respect, Shapshay (2013), Emily Brady
(2013) and Christine Battersby (2007) argue in different ways that the sublime actually primes us
to feel incorporated into nature. They treat this response as priming a moral obligation in us to
respect nature, and for both Brady and Battersby this specifically includes respect for diversity in
nature. However, at times one might be forgiven for concluding they had collapsed the sublime
into the experience of beauty. Traditionally, beauty is about finding the world a perfect fit for
the kind of beings that we are. The sublime in contrast, is thought to set us above the rest of
nature as beings capable of taking responsibility for our actions. And in this way, the sublime is
thought to orientate us to the world and each other as moral beings.
Francois Lyotard (2011 [1984]) adapts the Kantian notion of the sublime to explain the
possibility of cultural renewal. He argues that the sublime can be understood as an experience of
the as yet unsayable; “as yet” because Lyotard (after Adorno, and in turn arguably after Kant)
assumes our concepts evolve in response to cultural renewal, and as such their evolution outruns
our linguistic terms. The experience of the sublime captures just that aspect of our conceptual
framework that is exhausted by new ways of construing experience. Not being able to say what
one means is frustrating, but at the same time it is liberating to find that we are not constrained
by literal preordained assemblages of terms and expressions. The idea is that art is the vehicle for
such expression. As such Lyotard removes any direct link between the sublime and morality,
but it can be argued that he nonetheless maintains an indirect link.
Lyotard’s account can be understood as Kantian in that for Kant, judgments of the sublime
do not lay claim to “any cognition of the object” but like all reflective judgments they
“nevertheless [are] still related to concepts” ([AK 5: ‘244] 2000: 128). For Kant, in the sublime
“the mind is incited to abandon sensibility and to occupy itself with ideas that contain a higher
purposiveness” ([AK 5: ‘246] 2000: 129). Kant distinguishes between “our cognition of natural
objects” and “our concept of nature”. The experience of the sublime expands the latter, not the
former ([AK 5: ‘246] 2000: 130). This means that reflective judgments involve how we
construe nature rather than how nature presents to us. In line with this, Lyotard identifies the
sense in which the sublime expands our understanding. And he does this in a way which shows
how human artefacts might exploit this capacity. Lyotard argued that the sublime grounded the
293

Jennifer McMahon

possibility of avant-garde art which in his account (after Adorno) is a mechanism for cultural
renewal. Cultural renewal demonstrates the exercise of agency and the way in which communities of people are the authors of their own destinies. In this sense the sublime is a sign of
our capacity for morality.
The various Kantian conceptions including Lyotard’s account, provide a link either directly
or indirectly to morality. We have considered the way this link is thought to occur by various
writers whether as respect for humanity, pleasure in our agency, enabling cultural renewal or
respect for nature. Writers like Melissa Merritt (2012), and Robert Clewis (2009) conceive the
sublime entirely in terms of alerting us to capacities within us that are conditions for morality.
Presumably Kant was interested in the sublime for this very reason. The conception of the
sublime developed here as based within the contemporary terms of a more liberal naturalism,
draws upon the idea of a perceptual affordance to explain the intersubjective nature of the sublime
rather than Kant’s reliance on the moral law within. We could also draw upon research which
shows the way imagery to which we are previously exposed, can shape and influence what we
subsequently perceive (what we notice, foreground, and the significance it holds for us)
(Fazekas and Nanay 2017) as a way of demonstrating the way affordances work. The structure
of the sublime experience as outlined above, the particular way pain and pleasure are evoked
and combined, and the inclination for narrative that characterizes the human mind (Currie
1995), in addition to the influence of imagery on perception, all support a conception of the
sublime as indicative of our agency and hence our moral capacity. In contrast, it is clear that
scientistic accounts, which only recognize visual elements, representational perceptual properties, and instinctual responses or private reverie in place of ideas, cannot do justice to the
experience of the sublime which I have argued includes a sense of our morality.

7 Conclusion
It is clear that the experience of the sublime is an aesthetic experience which occasions and
promotes ideas which further our efficacy in the world as individuals and communities. The
writers on the sublime who prioritize their experiences and observations over metaphysical
dogma such as narrow scientific naturalism, find that those aspects of us which are a condition
of our moral capacity are raised to consciousness in experiences of the sublime. That this is a
universal response is suggested by the cross-cultural capacity for awe and wonder in the face of
some phenomena that without our intellectual intervention would be de-motivating for our
human endeavors and schemes. But it also shows how we can exercise agency in how we
orientate ourselves to the world and each other which speaks to our moral capacity.
We have seen that our starting point was the structure of the sublime as identified and
described by Kant. According to Kant, the sublime involves an “agreeableness … produced
only by means of a momentary inhibition followed by a stronger outpouring of the vital force”
([AK 5: § 14, ‘226] 2000: 111). However, as we have seen there are a range of positions taken
by commentators on the sublime regarding the relation of pain and pleasure, and the role of
ideas and their relevance to moral conceptions of self. Many disagreements over the relation
between pain and pleasure follow from the account of pleasure, imagination, and perception
that the commentator implicitly holds. When such processes or faculties are envisaged within
the scientific image, much is omitted, and the result is a notion of the sublime which seems little
more than a feeling of relief from danger or perhaps a stimulus to day-dreaming. In contrast, our
own observations, other writing on the sublime, in addition to expressions of the sublime in
certain artistic and literary works, would suggest otherwise.

294

Liberal Naturalism

In this chapter I show how treating the sublime as a perceptual affordance explains the link
between pain and pleasure, and the inter-subjective and objective focus engaged by the sublime. My approach demonstrates the greater explanatory power in reconstruing reflective
judgment in terms of perceptual affordance which in turn explains the concepts of intentional
pleasure, intersubjectivity and moral awareness for a fuller and more experientially complete
account of the sublime.
Kant’s conception of the sublime is quite compatible with the possibility that the narrative or
ideas through which we experience the sublime might change and vary from age to age and
between cultures. In many respects the theoretical commitments of the Kantian sublime nicely
foreshadow certain tenets of liberal naturalism. The tenor of the ideas posited to explain the
experience of the sublime are devised to show how a perceptual experience, while subjective,
can make a claim of normative validity, or in other words, a claim on everyone’s assent.
Construed in this way, this account of the sublime avoids many of the distortions which arise by
understanding reflective judgment, pleasure, perception, intersubjectivity and morality in an
overly atomistic manner. It also explains why Kant was interested in reflective judgments of the
sublime in the first place. Kant thought that the sublime was not so much a symbol of morality
like beauty ([AK 5: ‘353] 2000: 227) but that the sublime raises ideas within us like the idea of
freedom which is conducive to morality.28 We can reject his understanding of the lawful
quality of morality and give instead a greater role to the pressures of inter-subjectivity which
arguably help shape perceptual affordances. In this chapter, the explanatory power of liberal
naturalism for the sublime is revealed through the concept of perceptual affordance.

Notes
1 For a detailed account of the contrast between liberal naturalism and scientific naturalism see
Macarthur (2019, p. 572).
2 For a detailed account of the aims and objectives of Liberal Naturalism, see: De Caro and Macarthur
(2004, 2010).
3 Crespo (2019, p. 256).
4 McDowell (2004, p. 95).
5 Macarthur (2019, p. 573).
6 Fink (2006, p. 203).
7 Putnam (2002, p. 15).
8 See McMahon (2017) for a detailed account of the textual evidence behind a reading of Kant’s aesthetic theory according to which it entails cognitive engagement. Also see McMahon (2020) in the
section called “The Modern Origins of Beauty Theory”.
9 Kant ([CJ AK: 5: 305, §44] 2000: 184-85. For a discussion of this, see McMahon (2020) in the section
entitled “The Modern Origins of Beauty Theory”. When certain terms used by Kant in his “Analytic
of the Beautiful” are put into the context of the metaphysical commitments and conceptual framework
of his time, and understood in light of what he writes in the “Deduction of Pure Aesthetic Judgments”,
an interpretation of his aesthetic theory becomes clearer and arguably more accurately reflects Kant’s
philosophical objectives.
10 See McDowell (2002; and 2004).
11 It is not only Kant for whom such a defence is possible. See Giladi (2014) who interprets Hegel’s
idealism in terms of liberal naturalism; and Zuckert (2015) who reconciles culturalism and naturalism in
Herder’s aesthetics..
12 Macarthur (2019, p. 577).
13 Macarthur (2019, p. 578).
14 In 1979 Paul Guyer wrote that the sublime “will not be of much interest to modern sensibilities”
(1979, 400, fn 2). In 1993 he revised this, writing in response to the feedback he had received to his
1979 book that: “[n]o statement in that book has come in for more criticism than this remark” (1993,

295

Jennifer McMahon

15
16

19
20
21

22
23
24
25

26
27
28

p.187). And see an example of renewed interest in the sublime, in Arcangeli and Dokic (2020) who
present another contemporary reading of the Kantian structure of the sublime.
For a detailed discussion of the various influences on Kant’s conception of the sublime, see
Guyer (2014).
See Nichols and Stich (2000) for their concept of the “script elaborator” which they use to explain the
interface between imaginings and perception. But see Langland-Hassan (2016) who argues that their
reliance on the “script elaborator” simply boxes the hard problem of imagination (which is the way he
categorizes the problem of ascription) and gives it a name, rather than solving the problem. In contrast,
I argue that the concept of perceptual affordance addresses the problem of ascription without proposing
expensive mental hardware.
More work on making this link could be done by drawing upon Fazekas and Nanay (2017) which
provides an account showing how imagery (as top down) is the mechanism by which the cognitive
penetration of perception is mediated. Fazekas and Nanay’s research, though focussed on a different
problem, further corroborates the active role the experience of the sublime can play in shaping behavior. See also Nanay (2010).
This issue is also relevant to Kant’s distinction between private and public reason. This is discussed in
Kant ([AK 5: §40, ‘293-296] 2000: 173-76). See also O’Neill (2011) where she argues that the Sensus
Communis as represented in the Critique of Aesthetic Judgment is Kant’s best demonstration of public
reason.
Kant discusses the difference between his own account and that of Burke’s in ([AK 5: 277-278] 2000,
pp. 158-59).
Kant himself expresses this acknowledgment of the cultural variations in the manifestation of such rules
and principles more generally, for example, in his study of culturally based norms in his Anthropology
from a Pragmatic point of View.
For a discussion of Kant’s notion of “freedom to think” see the discussion in McMahon (2014a: 109110). For the essay in which Kant discusses “freedom to think” see “What Does It Mean To Orient
Oneself in Thinking” ([AK 8: 144] 1996: 16). Also for the difference between private and public
reason see O’Neill 2011.
I quote from Pluhar’s translation for these extracts on the sublime for ease of understanding. However,
the translation closer to the original, even though some revision occurs there also to assist understanding, can be found in Guyer and Matthews 2000: 148-149.
Hirschfeld ([1779-85] trans. Parshall 2001). For a discussion, see McMahon (2014a, p.136-39).
Wordsworth and Coleridge ([1800, 1802] 1963: pp. 250-51, fn.50).)
Deligiorgi argues that the experience of the sublime orientates us to the objective world in certain ways
and this involves our attention be directed to the object rather than lost in personal reverie (2014: 30).
She argues that this suggests that pain is not converted to pleasure but remains as part of the experience
in some sense.
For an account of the difference between relief-pleasure and facilitating-pleasure which would suggest
that the pleasure of the sublime is the latter, see Mohan Matthen (2018, 2017).
A discussion of how exchanges between varying perspectives provide occasions for a calibration of
terms within a community can be found in McMahon (2014b).
I would like to thank Talia Morag for inviting me to present at the Liberal Naturalism conference in
Melbourne in November 2017; and for the feedback she and the other members of the audience
provided at that time. I have written on naturalizing Kant’s conception of beauty (2007) and the role of
the imagination in the sublime (2014a ch.6) but Morag’s conference prompted me to articulate more
explicitly the metaphysical commitments of my position. I am also grateful for the invitation from
Mario De Caro and David Macarthur to contribute a chapter to this anthology. And to the support
provided by the Australian Research Council Discovery Grant DP150103143.

References
Arcangeli, M. and Dokic, J. At the Limits: What Drives Experiences of the Sublime. British Journal of
Aesthetics. (2020). 10.1093/aesthj/ayaa030.
Battersby, C. The Sublime, Terror and Human Difference. Routledge. (2007).
Brady, E. The Sublime in Modern Philosophy: Aesthetics, Ethics, and Nature. Cambridge University Press. (2013).

296

Liberal Naturalism
Budd, M. Delight in the Natural World: Kant on the Aesthetic Appreciation of Nature Part III: The
Sublime in Nature. British Journal of Aesthetics 38(3). (1999): 233–250.
Budd, M. The Aesthetic Appreciation of Nature. Oxford University Press. (2002).
Burke, E. A Philosophical Enquiry into the Origin of Our Ideas of the Sublime and the Beautiful, ed. Paul Guyer.
Oxford University Press. ([1757])/2015).
Clewis, R. The Kantian Sublime and the Revelation of Freedom. Cambridge University Press. (2009).
Crowther, P. The Kantian Sublime: From Morality to Art. Clarendon Press. (1989).
Currie, G. Image and Mind: Film, Philosophy and Cognitive Science. Cambridge University Press. (1995).
De Caro, M. and Macarthur, D. Introduction: Science, Naturalism, and the Problem of Normativity. In
M. De Caro and D. Macarthur (ed.) Naturalism and Normativity, pp. 1–19. Columbia University
Press. (2010).
De Caro, M. and Macarthur, D. Introduction: The Nature of Naturalism. In M. De Caro and D.
Macarthur (ed.) Naturalism in Question, pp. 1–17. Harvard University Press. (2004).
Deligiorgi, K. The Pleasures of Contra-Purposiveness: Kant, the Sublime, and Being Human. The Journal
of Aesthetics and Art Criticism 72(1), (2014) 25–35.
Dimitrakos, T. Integrating First and Second Nature: Rethinking John McDowell’s Liberal Naturalism.
Philosophical Inquiries 8(1), (2020).
Fazekas, P. and Bence, N. “Pre-cueing Effects: Attention or Mental Imagery?” Frontiers in Psychology.
(2017). DOI: 10.3389/fpsyg.2017.00222.
Fink, H. Three Sorts of Naturalism. European Journal of Philosophy 14(2). (2006). 202–221.
Gibson, J.J. The Senses Considered as Perceptual Systems. Houghton Mifflin. (1966).
Giladi, P. Liberal Naturalism: The Curious Case of Hegel. International Journal of Philosophical Studies 22(2).
(2014). 248–270. DOI: 10.1080/09672559.2014.886280
Guyer, P. A History of Modern Aesthetics Vol. 1: The Eighteenth Century. Cambridge University Press. (2014).
Guyer, P. Kant and the Claims of Taste. Harvard University Press. (1979).
Guyer, P. Kant and the Experience of Freedom. Cambridge University Press. (1993).
Hepburn, R. Landscape and the Metaphysical Imagination. Environmental Values 5. (1996). 191–204.
Hirschfeld, C.C.L. Theory of Garden Design. Ed. and trans. Linda B. Parshall. University of Pennsylvania
Press. ([1779-1785]/2001].
Kant, I. Anthropology from a Pragmatic Point of View [1798]. Trans. Mary J. Gregor. Nijoff. (1974).
Kant, I. Critique of Judgment [1790]. Trans. W. S. Pluhar. Hackett. (1987). [AK 5: 167–356].
Kant, I. Critique of Practical Reason [1788]. In Immanuel Kant: Practical Philosophy, pp. 135–271. Trans. and
ed. M.J. Gregor, intro. A.W. Cambridge University Press. (1996). [AK 5: 4–163].
Kant, I. Critique of the Power of Judgment [1790]. Trans. P. Guyer and E. Matthews, ed. Paul Guyer,
pp. 1–230. Cambridge University Press. (2000). [AK 5: 167–356].
Kant, I. What Does It Mean to Orient Oneself in Thinking? [1786]. In Immanuel Kant: Religion and
Rational Theology, trans. and ed. Allen W. Wood and George Di Giovanni, pp. 3–18. Cambridge
University Press. (1996) . [AK 8: 134–146].
Langland-Hassan, P. On Choosing What to Imagine. In A. Kind and P. Kung (ed.) Knowledge Through
Imagination, pp. 61–84. Oxford University Press. (2016).
Lyotard, J.F. The Sublime and the Avant-garde. In The Continental Aesthetics Reader, ed. Clive Cazeaux,
pp. 586–595 (second edition). Routledge. ([1984]/2011).
Macarthur, D. Liberal Naturalism and the Philosophy of the Manifest Image. In A. Gare and W. Hudson
(ed.) For a New Naturalism, pp. 50–65. Telos. (2017).
Macarthur, D. Liberal Naturalism and the Scientific Image of the World. Inquiry: An Interdisciplinary Journal
of Philosophy 62(5), (2019). 565–585. DOI: 10.1080/0020174X.2018.1484006
Matthen, M. New Prospects for Aesthetic Hedonism. In J.A. McMahon (ed.) Social Aesthetics and Moral
Judgment: Pleasure, Reflection, and Accountability, pp. 13–33. Routledge. (2018).
Matthen, M. The Pleasure of Art. Australasian Philosophical Review 1(1), (2017). 6–28.
McDowell, J. Two Sorts of Naturalism. In J. McDowell (ed.) Reason, Value, and Reality. Harvard
University Press. (2002).
McDowell, J. Naturalism in the Philosophy of Mind. In M. De Caro & D. Macarthur (eds.) Naturalism in
Question. Harvard University Press. (2004).
McMahon, J.A. Aesthetics and Material Beauty: Aesthetics Naturalized. Routledge. (2007).
McMahon, J.A. Art and Ethics in a Material World: Kant’s Pragmatist Legacy. Routledge. (2014a).
McMahon, J.A. Beauty. In Oxford Research Encyclopedia of Literature. Oxford University Press. Article
published September 28, 2020. doi: 10.1093/acrefore/9780190201098.013.1050

297

Jennifer McMahon
McMahon, J.A. Immediate Judgment and Non-Cognitive Ideas: The Pervasive and Persistent in the
Misreading of Kant’s Aesthetic Formalism. In M.C. Altman (ed.), The Palgrave Kant Handbook,
pp. 425–446. Palgrave Macmillan. (2017).
Merritt, M. The Moral Source of the Kantian Sublime. In T. Costelloe (ed.), The Sublime: From Antiquity
to the Present, pp. 37–49. Cambridge University Press. (2012).
McMahon, J.A. “The Sense of Community in Cavell’ Conception of Aesthetic and Moral Judgment”.
Conversations: The Journal of Cavellian Studies 2. (2014b). 35–53. https://uottawa.scholarsportal.info/
ojs/index.php/conversations/issue/view/222
Nanay, B. “Against Aesthetic Judgments”. In J.A. McMahon (ed.) Social Aesthetics and Moral Judgment:
Pleasure, Reflection, and Accountability, pp. 52–65. Routledge. (2018).
Nanay, B. Attention and Perceptual Content. Analysis 70, (2010). 263–270, DOI: 10.1093/analys/anp165
Nichols, S. and Stich, S. A Cognitive Theory of Pretense. Cognition 74, (2000). 115–147.
O’Neill, O. Kant’s Conception of Public Reason. In C. Payne and L. Thorpe (eds.) Kant and the Concept of
Community, pp. 138–149. University of Rochester Press. (2011).
Putnam, H. The Collapse of the Fact/Value Dichotomy and Other Essays. The MIT Press. (2002).
Shapshay, S. Contemporary Environmental Aesthetics and the Neglect of the Sublime. British Journal of
Aesthetics 53(2), (2013). 181–198.
White, S.L. Subjectivity and the Agential Perspective. In M. De Caro and D. Macarthur (eds.) Naturalism
in Question, pp. 201–227. Harvard University Press. (2004).
Wordsworth, W., and Coleridge, S.. Lyrical Ballads. Edited by R. L. Brett and A. R. Jones. Methuen.
([1800, 1802]/1963).
Zuckert, R. Adaptive Naturalism in Herder’s Aesthetics: An Interpretation of Shakespeare. Graduate
Faculty Philosophy Journal 36(2), (2015). 269–293.

298

25
PHILOSOPHY OF PERCEPTION
AND LIBERAL NATURALISM
Thomas Raleigh

1 Introduction
Perceptual experience is the moment where a subject’s conscious mind “makes contact” with
her mind-independent world. The central, traditional problems for philosophers of perception
have been to give an account of this conscious perceptual “contact” – that is, to say both what
its metaphysical nature or structure is, but also how it manages to perform the epistemic role of
providing the subject with information about her surroundings, allowing her to form accurate
beliefs, to become knowledgeable about her environment.
This chapter will explore the connections between these traditional problems of perception
and the recent opposition between “Strictly Scientific”, reductive forms of Naturalism vs.
“Liberal”, nonreductive forms of Naturalism. It is a striking fact about the recent history of
philosophy that two of the most prominent figures associated with Liberal Naturalism – John
McDowell and Hilary Putnam – were also both extremely influential on the philosophy of
perception and in particular on the rise of relational and disjunctivist theories of perceptual
experience. It is therefore worth considering what exactly the relationship is between Liberal
Naturalism and theories of perceptual experience. In particular, we will consider whether
holding a specific theory about the nature of perceptual experience should incline one toward
or away from Liberal Naturalism. And conversely whether a commitment to some form of
Liberal Naturalism should incline one toward or away from one or more of the standard views
about perceptual experience.
Liberal Naturalism is not a precisely defined philosophical thesis. Rather, the label “Liberal
Naturalism” refers to a cluster or family of views about what it takes for something to count as a
real, genuine part of the natural world.1 Liberal Naturalism is standardly characterised in opposition to Strictly Scientific Naturalism, which seeks to limit what counts as a real or genuine part
of the natural world to whatever entities and features appear in our best/(final/completed?)
scientific theories – or which can be explained by or reduced to the entities and features
admitted by science.2 In contrast, Liberal Naturalists insist that there is some range of phenomena that cannot be reduced to scientific entities/features (nor can be properly investigated
and theorised using scientific methods) but which are nonetheless perfectly real and genuine
features of the natural world and so which need not be thought of as somehow supernatural or
“spooky”.3 In order to get a more precise grip on the commitments of a distinctively Liberal
DOI: 10.4324/9781351209472-30

299

Thomas Raleigh

Naturalist take on some topic/phenomenon, one would want to say whether and how it differs
from some kind of nonreductive materialism4 and also how exactly it differs from nonnaturalism. However, given that these sorts of questions are addressed extensively by other
chapters in this volume, I will not attempt to say anything further here to pin down the notion
of Liberal Naturalism more precisely.
The plan for this chapter then is as follows: Section 2, below, briefly outlines the traditional
problem of perception and the main standard views in current philosophy of perception.
Section 3 considers how Liberal Naturalism interacts with the philosophy of perception.
Section 4 focuses on the work of McDowell and Putnam. Section 5 discusses the topic of color.
Finally, Section 6 provides a very brief conclusion.

2 The traditional problems of perception
The traditional philosophical problem of perception can be stated briefly as follows: our ordinary
(intuitive, pretheoretical, “naïve”) conception of perceptual experience is that we are simply and
directly conscious of the external, mind-independent features in our immediate vicinity. This is
sometimes put in terms of our minds being “open” to the external world, or that we are simply
“presented” with or “acquainted” with external objects in our surroundings. But then reflection
on the possibility of perceptual illusions and hallucinations – and perhaps also phenomena of
perceptual relativity and also the time-lag between light reflecting from a distal stimulus and our
experience of it – can seem to undermine this ordinary intuitive conception of experience and to
suggest that our perception of our external surroundings is importantly indirect, involving conscious awareness of some kind of mental intermediary or representation. This general dialectic can
be pursued with an emphasis on the metaphysics of experience – that is, as a question about the
metaphysical relation between certain kinds of conscious experience and the external environment that these experiences are experiences of. And it can also be pursued with a more epistemological emphasis via the question of whether/how perceptual experiences can be provide us
with knowledge and justified beliefs about our surroundings. Perhaps the simplest form of argument that is meant to undermine our ordinary conception of perceptual experience is the
argument from hallucination (similar, though somewhat more complicated, arguments for the
same conclusion can be mounted based on the possibility of illusions, or on the relativity of
perceptual appearances to facts about the perceiving subject):
1
2
3

In a (totally) hallucinatory experience, we are not in direct conscious contact with our
external surroundings.
Common Kind Assumption: Hallucinations and normal perceptual experiences are
fundamentally the same kind of conscious state.
SO: in normal perceptual experiences, we are not in direct conscious contact with our
external surroundings.

The first premise here is rarely contested – perhaps it is thought to simply fall out of the
meaning of “hallucination”.5 So the main philosophical action has concerned the second
premise – the Common Kind assumption. There are two main lines of thought that are
standardly offered in support of this Common Kind Assumption:
(i) For any normal perceptual experience, there could be a hallucinatory experience that is
subjectively indistinguishable. And if two experiences are subjectively indistinguishable,
then they are fundamentally the same kind of conscious state.
300

Philosophy of perception

(ii) For any normal perceptual experience, there could be a hallucinatory experience that
involves exactly the same neural processes in the brain. And if two experiences involve
exactly the same neural processes, then they are fundamentally the same kind of conscious
state.
Metaphysical Disjunctivism6 is the denial of the Common Kind Assumption – even if the
perceptual experience and the hallucinatory experience may be subjectively indistinguishable
and neurally identical, they are fundamentally different kinds of conscious states. (What constitutes the phenomenal character in the good, perceptual case is fundamentally different from
what constitutes the phenomenal character in the bad, hallucinatory case.) Disjunctivists have
thus argued against both (i) and (ii), denying that either subjective indistinguishability or neural
similarity would entail that there is the same fundamental kind of conscious state in both cases.
If metaphysical disjunctivism can be made plausible then it would allow us to endorse both the
intuitive, “naïve realist” idea that in normal successful perception we are directly conscious of
our external surroundings and also the intuitive idea that there can be hallucinatory situations
where we merely seem to be perceiving our surroundings though we are not in fact in conscious contact with the external environment.
The conclusion of the argument from Hallucination is explicitly about the metaphysics of
experience: it denies the claim that normal “veridical” experiences can consist in a direct relation of conscious acquaintance with the external environment. But if we are thinking about
the epistemology of perception, the idea that what we are directly given in experience, as a basis
for forming perceptual beliefs, is just a subjective appearance of the external environment –
something that is epistemically in common to both normal perceptions and hallucinations
(or other skeptical scenarios involving demons, Brains-in-Vats, etc.) – raises the following,
traditional skeptical line of thought:
1

All that we are directly given in any perceptual experience as a basis to form perceptual
beliefs about the external world are mere perceptual appearances of the external world –
something which may or may not turn out to be genuine perceptions.
Such a perceptual appearance that p cannot justify a belief that p (where p is some claim
about the external world) unless we have some basis for supposing that the perceptual
appearance really is a genuine perception – that is, that experiences are mostly accurate/
reliable.
There is no nonquestion-begging basis for supposing that perceptual appearances are
(mostly) accurate/reliable.

SO: (4) Perceptual experiences cannot justify beliefs about the external world.
Notice that the first premise here is plausibly consistent with accepting some form of Naïve
Realist or “Relational” view concerning the metaphysics of experience. For even if we are
sometimes directly acquainted with external items and features, one might still hold that we
cannot tell just by introspecting such an external-world-involving experience whether it really
is a genuine, world-involving perception or whether it is a hallucination (or dream). And so
epistemically speaking all one is “given” to work with is the perceptual appearance of the
external world (even if in fact that appearance is partially constituted by the external world.)7
Epistemological disjunctivism denies the first premise: in good cases we are presented with
more than just a mere appearance that p, we can be presented with the very fact that p. Even if
we can be deceived by hallucinations/illusions into thinking that we are in the good case when
we are actually in the bad case, still the epistemic position of the subject who is actually in the
301

Thomas Raleigh

good, perceptual case is fundamentally different to the subject who has a mere, illusory perceptual appearance that p.8 See, for example, McDowell (1994, 2008) and Pritchard (2008,
2012) for influential statements of the view. In contrast, Macarthur (2003) and Wright (2008)
are both dubious of the supposed antiskeptical power of epistemological disjunctivism.
Jim Pryor (2000) influentially argued against premise (2), insisting that a perceptual experience as of p can justify a belief that p even in the absence of any positive belief or evidence
that one’s experience is generally reliable – so long as there is no positive belief or evidence that
one’s experience is unreliable. This position has come to be known, following Pryor, as
“Dogmatism” about perceptual justification.
Some traditional foundationalists accept the first two premises but deny the third, arguing
that we can have a legitimate basis for supposing that our perceptual appearances are generally a
reliable guide to the external world. For example, Russell (1912) and Bonjour (2003) claim that
we have a good abductive basis for taking our experiences to be mostly reliable. Whereas
Descartes (in)famously appealed to the perfection of God as an a priori basis for supposing that
perceptual appearances are generally an accurate guide to our surroundings.
Some traditional coherentists can be understood to simply accept this whole argument, but
without thereby embracing skepticism concerning knowledge of the external world, since they
deny that perceptual experiences play the role of justifying beliefs in the first place, insisting that
“only a belief can justify another belief” (Davidson [1983]/2001). Instead, a coherentist claims
that our beliefs about the external world can be justified on the basis of their coherence with all
the other beliefs in our total set of beliefs.9
The traditional problem of perception has given rise to a range of standard positions. A now
somewhat old-fashioned kind of taxonomy divided theories of perception into three groups:
•
•
•

DIRECT REALIST theories – which claim that we are (at least sometimes) directly
conscious of our mind-independent surroundings (in some sense of “direct”).
INDIRECT REALIST theories – which accept that we perceive the mind-independent
world via a “veil of appearances”.
PHENOMENALIST/IDEALIST theories – which deny that there is a mind-independent
world distinct from all conscious appearances in the first place.

These days, a standard sort of taxonomy would identify the following four major positions:
•

•

Sense-Data views – in perceiving our environments, we are directly conscious of
nonenvironmental particulars – “Sense-Data” – whose properties constitute the phenomenal character of our experiences. We are thus only indirectly conscious of the
physical objects in our environment.10 See, for example, Russell (1912, 1927), Moore
(1910) Price (1932), Broad (1925), Jackson (1977) , Foster (2000), Robinson (1994) and
Bermudez (2000).
Adverbial views – perceptual experiences are conscious mental states which do not have a
binary, act-object structure. Instead, the perceiving subject is understood to be conscious in
a specific way. So, for example, rather than standing in a relation of (direct) conscious
awareness to an external instance of redness, or to a red sense-datum, the subject is said to
experience “redly” – that is, in a specific phenomenally red manner. See, for example,
Ducasse (1942), Chisholm (1957), Cornman (1971), Sellars (1975) andTye (1975, 1984).
Representational/Intentionalist views – perceptual experiences are conscious mental
states with a representational content, which content at least partially determines the
phenomenal character of perception. The perceptual experience, like a belief or a desire, is
302

Philosophy of perception

•

intentionally directed towards some object or objects, but this object need not actually
exist. The general idea that perceptual experiences are representational is these days the
orthodoxy in philosophy. There is then is a huge variety of views within the broad representational framework, which disagree over the precise nature of the content of the
representational state and it’s relation to phenomenal character. See, for example, Harman
(1990), Tye (2000), Byrne (2001), Dretske (2003), Chalmers (2004) and Siegel (2006).
Naïve-Realist/Relational views – in perceiving our environments, we have direct
conscious awareness of physical objects and features in our environment, which at least
partially constitute the phenomenal character of the experience.11 See, for example,
Campbell (2002), Brewer (2011), Martin (1998, 2002, 2004, 2006), Travis (2004) and
Johnston (2006).

Sense-Data and Adverbial views would both naturally be counted as avowedly indirect.12
Whereas, Naïve-Realism aims to capture notion of “direct contact” between mind and world
as literally as possible with its claim that external features can partially constitute a perceptual
experience. Representational views are these days generally presented as being a direct view
(we are directly aware of external features themselves by being in a conscious state which
accurately represents them). However, most Representational views typically accept the
Common Kind assumption and so it not entirely clear whether they really capture the intuitive
notion of direct open-ness to the world. On the other hand, there are some representational
theorists who are happy to accept that perceptual experience is metaphysically indirect (involves
awareness of intermediaries). And of course, there are also various possible “hybrid” positions
which combine elements from two or more of the above positions.13
To be clear, there is much more to the philosophy of perception than just the traditional
problems outlined above! Just to give a few examples of some much-discussed questions: how
to differentiate the different senses, what the causal conditions on perception are, whether and
to what extent perceptual experience can be “cognitively penetrated” by our beliefs and biases.
One topic that is especially relevant to Liberal Naturalism is the opposition between “Thick”
vs. “Thin” accounts of which properties we can (directly, genuinely) perceive. According to
Thin views, the phenomenal character of perceptual experience is exhausted by the experiential
presentation of “low-level” properties such as shapes and colors and textures.14 According to
Thick views, the phenomenal character of perceptual experience can also sometimes (often)
involve the experiencing of “high-level” properties such as: natural kinds, causal relations,
meanings, functions, affordances and perhaps also normative and value properties (see
Section 3.3).15

3 How does philosophy of perception interact with questions
about naturalism?
Philosophical views on what is most fundamental or real in nature have long influenced and
been influenced by philosophical accounts of perceptual experience. Perhaps the clearest historical example is how the mechanical philosophies of the early modern period gave rise to the
distinction between primary vs. secondary qualities and to theories about how we perceive and
what we perceive. Various important early modern thinkers – including Galileo, Descartes,
Hobbes, Locke and Boyle – held that mechanical properties, such as size and shape, motion and
rest, were the fundamental properties of matter. Whereas other nonmechanical properties –
including not only colors, smells and tastes, but also electrical and magnetic properties – were
held to be nonfundamental and to be explained by the fundamental mechanical properties. This
303

Thomas Raleigh

is a general metaphysical picture of the material world that can be thought of as a species of strict
scientific naturalism; the mechanistic philosophers sought to explain everything physical in
terms of a relatively small set of (what they took to be) scientifically respectable properties.
Those features that were not thus explicable – such as the qualitative natures of the colors –
were held to be mind-dependent, features of our ideas rather than being genuine, mindindependent features of the physical world. This general mechanistic world-view was reflected
in, for example, Locke’s account of how we perceive the world. According to Locke we only
have direct conscious access to our own ideas – that is, mental entities – which mediate our
perceptual access to material objects and features in our surroundings. Material objects and ideas
can “resemble” each other in some respects but not others. Both ideas and material objects can
have sizes and shapes and motions, the “primary qualities”. But our ideas of colors and smells
and tastes are supposed to have no resemblance in the material objects that normally cause these
ideas. Thus secondary qualities as attributed to physical objects are held to be “in truth nothing
in the Objects themselves, but Powers to produce various Sensations in us by their primary
Qualities”– that is, colors, smells and tastes, qua properties of external objects, are observerdependent, dispositional properties, that depend on the more fundamental, observerindependent primary qualities.16 And the qualitative natures of colors, smells and tastes are held
not to be features of the external environment at all, but rather features of our own sensations/
ideas that we mistakenly “project” onto the external, material bodies. These early modern
debates show the interplay between a general metaphysical picture of what counts as a genuine/
real feature of the (mind-independent) natural world and an account of how we make conscious perceptual contact with that world. (Though of course in this period a mechanistic
account of material objects and their properties would be standardly combined with a dualist
metaphysics when it comes to mental states and properties.)
But leaving aside the historical interplay between these topics, what general theoretical
connections are there between Liberal Naturalism and the standard positions in philosophy of
perception? More specifically: whilst there is in fact a notable overlap between philosophers
who endorse Liberal Naturalism and those who endorse some species of naïve-realism and/or
disjunctivism, is there any substantial theoretical connection between these positions? And
conversely, is there any genuine tension or inconsistency between Liberal Naturalism or Strictly
Scientific Naturalism and any of the other standard views on perception?
One obvious way that there could be a flat-out inconsistency is if one held some strict/
reductive version of naturalism according to which there really are no such things as X’s. Then
one would, of course, be committed to rejecting any view of perception that claimed that we
can perceive X’s or that X’s are somehow essentially involved in perception. So, for example,
suppose X is: external, physical colors – any strict scientific naturalist who denied that such
color properties exist in the physical environment would also clearly have to reject any Naïve
Realist theory which held that external colors are a constitutive element in visual experience.
Or, suppose X is: private sense-data or qualia – any strict scientific naturalist who denied the
existence of these private items would obviously have to reject any theory which was committed to such entities.

3.1 Naïve realism and liberal naturalism
It is sometimes claimed that Naïve Realism enjoys some kind of default or presumptive status,
so that the view should only reluctantly be given up in the face of persuasive arguments.17 As
discussed above, whilst arguments from hallucination and illusion have been thought persuasive
by many philosophers, disjunctivism has recently emerged as a possible strategy for resisting
304

Philosophy of perception

these arguments. However, another reason that might persuade one to abandon NR is the idea
that modern science has shown us that the world is not really populated by objects composed of
a homogenous substance with sharp boundaries and colored surfaces – though it may appear so
in perceptual experience. Rather, it turns out that the objects in our environment are actually
composed of scattered clouds of micro-particles that have neither familiar macroscopic shapes
nor colors. So we could not, as Naïve Realism claims, be directly perceptually acquainted with
external instances of such familiar shapes and colors.18 Here, it seems that one way for a Naïve
Realist to resist this latter, “scientistic” line of thought would be to embrace something like
Liberal Naturalism and insist that there can be more in our natural environments for us to
perceive than is described by fundamental physics. Thus Naïve Realism might incline one
toward Liberal Naturalism at least with respect to color (qua intrinsic sensible quality of external
objects) and macroscopic shape. However, it is also possible to combine Naïve Realism with a
reductively physicalist account of color19 – for further discussion of color see Section 5.
Another connection between Naïve Realism and Liberal Naturalism emerges via the topic
of “structural realism”.20 Epistemic structural realism is the idea that our scientific theories can
at best provide Knowledge of the causal-dispositional structures and mathematical relationships
that obtain, but not the intrinsic/categorical nature of the relata that stand in these relationships.21 Ontic structural realism, in contrast, is the thesis that there is, fundamentally, nothing
more to nature/reality than these structural relations. Notice that epistemic structural realism
holds that there is more to the natural world than can be captured by scientific investigation –
since the categorical nature of the world is supposed to be something that is over and above the
structural relationships between properties and magnitudes that science discovers and describes.
Thus, it should count as a form of Liberal Naturalism (even if the idea is that the categorical
nature of reality will remain forever unknowable to us). Now, one motivation that has been
advanced in favor of Naïve Realism is that by positing a relation of direct conscious acquaintance with external qualities it allows us to gain Knowledge of (some of) the categorical
nature of the external world – see Brewer (2011), Campbell (2009), Johnston (2006) and Logue
(2012). Any such form of Naïve Realism, which holds that perceptual experience reveals to us
categorical features/properties of the natural world that would not otherwise be knowable to
us, will thus be committed to a form of Liberal Naturalism.

3.2 “Indirect” views of perception and liberal naturalism
It is clearly possible to endorse some kind of “indirect” theory about perceptual experience
(sense-data) and yet consistently be a liberal naturalist. If you think that perception is mediated
by some kind of veil of appearances, this is obviously compatible with endorsing a liberal,
nonreductive ontology. But in fact the connection between Liberal Naturalism and some indirect views is plausibly stronger than mere compatibility. For, as mentioned above, it would
seem that most sense-data theories could be understood to entail a kind of Liberal Naturalism.
According to Russell22 and Moore, sense-data were supposed to be real, mind-independent
entities, but which were not part of the shared physical environment and so which were not
open to scientific investigation. According to more recent Sense-data theories of Jackson (1977)
or Bermudez (2000), sense-data are mind-dependent but are located out in the one same
physical space that we occupy. All of these sense-data theorists would seem to be immediately
committed to a liberal ontology, which accepts the existence of irreducible entities over and
above anything posited by the physical sciences.23
Compare: Chalmers (1996) has influentially argued that qualia (phenomenal properties)
cannot be reduced to or fully explained by any physical property and so must be recognised as a
305

Thomas Raleigh

further fundamental feature of the natural world, alongside fundamental particles and forces.
(Chalmers explicitly compares this to Maxwell’s realisation that electromagnetism cannot be
explained mechanically and so must be treated as a brute, sui generis phenomenon.) This kind
of view is explicitly committed to the existence of properties that are additional to anything that
is mentioned in our best current scientific theories – so in one sense this could be seen as
endorsing a kind of Liberal ontology in opposition to Strictly Scientific Naturalism. On the
other hand, the motivation for Chalmers’s kind of view is quite different to the motivations of
most theorists commonly thought of as Liberal Naturalists. For Chalmers’ accepts the cogency
of the location problem challenge for phenomenal properties. And so then when he find that
phenomenal properties cannot be reduced to or fully explained by any of the scientifically
kosher elements we currently recognise, he draws the radical conclusion that phenomenal
properties are a new kind of fundamental property in the universe, on a par with fundamental
forces like electromagnetism. More generally, many indirect theories of perception have been
motivated by the idea that perceptual appearances cannot be properly explained in terms of the
external physical items we perceive and so they locate the sensible properties we are aware as
aspects of our own conscious minds, or as instantiated by nonphysical sense-data. Whereas, in
contrast, most Liberal Naturalists start by rejecting the whole idea that every genuine property
or phenomenon must somehow be “located” within the domain described by our scientific
theories. Thus phenomenal properties – and likewise intentional properties or rational properties or moral properties – need not be thought of as posing any special problem for a naturalistic world view in the first place.
The moral then: theories that are standardly thought of as “indirect”, accepting mental or
non-environmental intermediaries that we are directly aware of, are not only compatible with a
Liberal Naturalistic ontology, many such theories entail an ontology that is more liberal than
Strictly Scientific Naturalism would countenance. However, some of the standard motivations
that lie behind Sense-Data or pure Qualia views are quite opposite to the motivations that have
generally driven most Liberal Naturalists.

3.3 Representational views and liberal naturalism
There are many different kinds of representational/intentional accounts of perceptual experience. But as a broad, general framework the idea that perceptual experiences are fundamentally
representational mental states is pretty neutral between Liberal Naturalism and Strictly Scientific
Naturalism. The general representational framework allows for the possibility that perceptual
contents are “thick” (or “rich”). And so properties and features that a Liberal Naturalist might
want to insist are genuine and natural could then be included in the content of perception – for
example, moral, normative and evaluative properties, artefactual/social properties, primitive
color properties. Both McDowell (1985) and Stephen White (2010) are Liberal Naturalists who
argue that we can directly and noninferentially perceive objects as morally valuable.24
Whereas, a “thinner” representational theorist could restrict the range of properties represented in perception to only those that a strict scientific naturalist would accept. Notice
however that the distinction between thick vs. thin accounts of the representational content of
experience does not always line up neatly with opposition between Liberal vs. Strictly scientific
naturalism. For example, the idea that perceptual experience represents causal relations between
objects or events is standardly counted as a thick view of perceptual content – but of course
causal relations are presumably perfectly acceptable within a strictly scientific ontology.
Alternatively, one could allow that some properties that would not be recognised by Strictly
Scientific Naturalism are standardly represented in perceptual experience, but claim these
306

Philosophy of perception

properties are not actually instantiated in the environment, so such perceptual representation is
systematically illusory – for example, this is Chalmers’s position on primitive color properties
(see error-theories in Section 5 on color).
However, it should be noted that for some Liberal Naturalism theorists – notably McDowell
and Putnam – it is important that experience somehow involves representational content (or at
least draws on conceptual capacities) in order for it to be able to play the epistemic role of
allowing us to make rationally evaluable judgments about the external, natural world. In the
work of McDowell at least, this is bound up with large, Liberal Naturalist theses that the natural
world is constitutively such as to be conceptualisable and that the “rational realm” is part of the
natural order. We turn to the specific views of McDowell and of Putnam in the next section.

4 McDowell and Putnam
In recent Anglophone philosophy, two of the most prominent advocates for Liberal Naturalism
have been John McDowell and Hilary Putnam, both of whom were also seminally important
advocates for disjunctivism about perceptual experience (though as we shall see, both of these
philosophers’ views on perception evolved over time). In the first of his 1994 Dewey Lectures
(later reprinted as “The Threefold Cord”), Putnam wrote:
Shortly after I began writing these lectures in the Berkshires, I went to the excellent
bookstore on Water Street in Williamstown and looked at the collection of philosophy books on display there. What I wanted to see was what was available about
perception. Although Wittgenstein was well represented…there was not a single book
by John Austin – the philosopher whose Sense and Sensibilia represents the most
powerful defense of what I am calling ‘natural realism’ in the history of philosophy.
Nor…was there any other work by a natural realist, nor, for that matter, was there a
single book devoted to perception as a topic…25
Twenty-five years later, nobody could complain that perception is a neglected topic in present
day “analytic” philosophy! This is in large part due to the influence of Putnam and McDowell.
Given their influence on contemporary philosophy, it will be instructive to briefly consider
how their views on Naturalism and on Perception interact.
John McDowell’s interests in the philosophy of perception are primarily epistemological. He
summarises his own disjunctivist position in the following passage:
In perceptual experience a subject has it appear to her that things are a certain way in
her environment. In some cases (‘good’ cases), the experience makes it manifest to the
subject that things are that way; in others (‘bad’ cases), the appearance that things are
that way is a mere appearance. An experience in which the appearance that things are a
certain way is a mere appearance can be such that the subject cannot tell that it Is not
an experience of the first kind. The point of my epistemological disjunctivism is that
this indiscriminability does not imply a match in epistemic significance.26
McDowell takes the main import of arguments from illusion and hallucination to be the following supposed epistemological moral: that when we see our surroundings we have to make a
move from a “lesser informational state” about how things look/appear to us, to a stronger
claim about how our external surroundings actually are.

307

Thomas Raleigh

Consider the Argument from Illusion. Seeing, or perhaps having seen, that things are
thus and so would be an epistemically satisfactory standing in the space of reasons. But
when I see that things are thus and so, I take it that things are thus and so on the basis
of having it look to me as if things are thus and so. And it can look to me as if things
are thus and so when they are not; appearances do not give me resources to ensure that
I take things to be thus and so, on the basis of appearances, only when things are
indeed thus and so. If things are indeed thus and so when they seem to be, the world is
doing me a favor. So if I want to restrict myself to standings in the space of reasons
whose flawlessness I can ensure without external help, I must go no further than
taking it looks to me as if things are thus and so.27
Rejecting this “Cartesian” picture is McDowell’s main concern in the philosophy of perception. He often draws a comparison with the case of testimonial knowledge, where he would
also reject the idea that we always have to make an inferential move from a lesser informational
state about the fact that a testifier, S, asserted that p, in order to come to know that p by
testimony. Rather, McDowell would insist, we can often simply be told that p and thus come
to know that p. Likewise, one can often simply see that p and thus know that p without making
any kind of inference from how things look or appear to one.
McDowell then is certainly an epistemological disjunctivist. It is less obvious whether he
should also be counted as a metaphysical disjunctivist. There are passages where he seems to
claim that in the “Good Case” of successful perception, external facts/circumstances are essentially involved in subjective appearances, whereas they are not in the “Bad Case” of experiencing a “mere appearance”:
Of facts to the effect that things seem thus and so to one, we might say, some are cases
of things being thus and so within the reach of one’s subjective access to the external
world, whereas others are mere appearances. In a given case the answer to the
question ‘Which?’ would state a further fact about the disposition of things in the
inner realm (a disposition less specifically mapped by saying merely that things seem to
one to be thus and so); since this further fact is not independent of the outer realm, we
are compelled to picture the inner and outer realms as interpenetrating, not
separated from one another by the characteristically Cartesian divide.28
This talk of the “interpenetration” of inner and outer is most naturally understood, I suggest, as
making a metaphysical claim about what constitutes a perceptual experience – that is, that external circumstances can partially constitute a subjective perceptual perspective on the world.
Likewise in a later paper responding to Tyler Burge he states that: “a subject’s inner world does
not have the characteristic Cartesian independence from the outer world” (2010, 245).
However, in response to Burge, he makes clear that he is happy to accept that there can
presumably be a neuro-computational state in common to both the Good and Bad cases, and that
all he really wants to insist on is that there is a fundamental epistemological difference between
these two situations. It is also worth emphasising that McDowell, despite his claims that in
perception our minds are open to the external world, is not comfortably categorised as a being
simply a Naïve-Realist given the large representational component in his views. In Mind and
World, he insisted that perceptual experiences have propositional contents and moreover that
these contents are conceptual– that is, in having such a representational perceptual state we
necessarily deploy conceptual capacities. However, partly in response to the work of Charles
Travis (2004), McDowell later recanted the claim that experiences have propositional content:
308

Philosophy of perception

I used to assume that to conceive experiences as actualizations of conceptual capacities,
we would need to credit experiences with propositional content, the sort of content
judgments have. And I used to assume that the content of an experience would need
to include everything the experience enables its subject to know non-inferentially.
But both of these assumptions now strike me as wrong.29
McDowell’s rejection of “Cartesian” accounts of perceptual experience is connected with his
wider rejection of strict, “scientistic” forms of naturalism. According to McDowell, it is the scientistic assumption that everything – including the mind and conscious experience – can be
explained in the standard causal-physical terms employed by the natural sciences which underlies
the appeal of the Cartesian idea that conscious perceptual appearances cannot be intrinsically
world-involving. Causal relations are external and nonessential to the physical items that stand in
these causal relations to each other. And once we think of a conscious perceptual experience as
being just the endpoint of a causal chain, running from the light being reflected from a physical
surface, through stimulation of the retina, to some pattern of neural activity, then the distal stimulus
(i.e. the perceived object) is bound to be treated as external and non-essential to the experience.
Thus we arrive at the idea that what perception delivers are mere appearances, which may or may
not have been caused in a reliable way by the external world. In rejecting this Cartesian picture,
McDowell does not, of course, deny that there is some causal chain of events running from the
light reflecting from the perceived object through to various events in the retina and visual cortex
and so on, which is necessary in order to perceive. What is being denied is the idea that a perceptual
experience can be identified with some pattern of neural activity that can be intrinsically specified in
isolation from the rest of that causal chain – or more generally that the mind can be thought of as a
separate, self-standing realm that is affected only causally by the external world.
Hilary Putnam credited McDowell (along with William James and John Dewey) for awakening his own philosophical interest in perception.30 Having been deeply impressed and influenced by McDowell’s “Mind and World”, Putnam used the occasion of his 1994 Dewey
Lectures to defend what he called “natural realism”:
A natural realist, in my sense, does hold that the objects of (normal ‘veridical’) perception are ‘external’ things, and more generally aspects of ‘external’ reality. But the
philosopher whose direct realism is just the old causal theory of perception with a bit
of linguistic cover-up can easily go along. “We perceive external things – that is to
say, we are caused to have certain subjective experiences in the appropriate way by
those external things,’ such a philosopher can say. The natural realist, in William
James’s sense, holds, in contrast, that successful perception is a sensing of aspects of the
reality ‘out there’ and not a mere affectation of a person’s subjectivity by those aspects.
I agree with James, as well as with McDowell, that the false belief that perception must
be so analysed is at the root of all the problems with the view of perception that, in on
form or another, has dominated Western philosophy since the seventeenth century.31
He also characterises perceptual experience as “transactional”:
As Dewey might have put it, perception is transactional. We are aware of ourselves in
interaction with our perceptual objects. I am aware of a series of visual, tactile, etc.,
perspectives on the chair without ceasing to perceive the chair as an object that does not
change as those perspectives change.32

309

Thomas Raleigh

These passages are naturally read as expressing a commitment to something like a Naïve realist
or relational view concerning the metaphysics of perceptual experience – on which external,
environmental features constitutively feature in experience. And indeed at this point in his
philosophical development, Putnam denied that we can even make sense of qualia as they are
often conceived – internalist, non-representational and nonconceptual conscious properties.
However, Putnam’s views on perceptual experience, as with so many topics, changed very
significantly over time. Due mainly to the influence of his former student Ned Block,33 Putnam
moved away from the McDowellian disjunctivism of his Dewey Lectures and came to accept
that there are after all something like “Qualia” – though understood in what he takes to be an
empirically respectable sense, identified as features of brain states.
The particular form of disjunctivism I sympathized with in the Dewey Lectures is due
to John McDowell (although…I refused to endorse the claim that all experiences are
conceptualized…). I now believe…that the best “mesh” of phenomenological considerations and neuroscientific ones will involve recognizing that many of our experiences are not conceptualized in any of the senses that McDowell has proposed,
and will involve rejecting the idea (advocated by Wittgenstein and later by Hinton)
that there is no possibility of a scientific criterion for sameness and differences of, say
color experience of different subjects in which intersubjective spectrum inversion is an
intelligible possibility. In sum…there is a qualitative and nonconceptual dimension of
experience that can be scientifically investigated.34
My present view is almost the complete opposite of McDowell’s.35
But whilst Putnam toward the end of his life came to reject many aspects of McDowell’s views
and to accept the existence of (something like) qualia, he remained interested in developing a
viable form of naïve-realism – for example, a late paper from 2011 is entitled “How to be a
sophisticated ‘Naïve-Realist’”. Putnam also suggested that a psychologically plausible form of
naïve-realism could be supported by appeal to the sort of Enactivist, sensorimotor theories of
perception advanced by J.J. Gibson and later by Alva Noe (another pupil of Putnam). These
theories treat perception as an active process of perceptual exploration by an embodied subject,
embedded in an environment. And so perceptual experiences are understood as (partially)
constituted (as opposed to merely caused) by the interaction – or “transaction” to use Putnam’s
preferred term – between the subject and her environment. This embrace of Enactivism about
perception illustrates Putnam’s late return to functionalism. Having been the leading proponent
of functionalism in the 1960s, in the late 1980s Putnam famously came to reject functionalism as
being incompatible with externalism about mental content. But by this late stage in his career,
Putnam thought that there can be a nonobjectionable form of liberal functionalism, which defines
functional states in wide/broad terms that allow for externalism about content. For example,
Putnam states that a liberal functionalist
will follow Gibson’s pioneering lead in seeking for a description of the complex relations between the properties of the object in question, the organism’s repertoire of
actual and possible sensory-motor transactions with the object, and the ways in which
the object is perceived.36
Alas, Putnam did not live long enough to fully flesh out this late, “sophisticated” form of naïve
realism. But he did insist on the epistemological point that the qualia which he now accepted
310

Philosophy of perception

are (somehow!) involved in perceptual experience cannot be what justify our perceptual beliefs.
Rather, Putnam thought that there must be something else, which he called “apperceptions”,
which play this justificatory role and which are, unlike qualia, conceptually saturated and externalistically individuated. Putnam was clear that there is no rational or epistemic relation
between qualia and apperceptions – we do not “construct” apperceptions from qualia.
However, it is perhaps not entirely clear from these late writings what the positive relationship
between qualia and apperceptions – that is, between mere brain states and fully-fledged perceptual experiences of the external world – is supposed to be exactly.

5 Color
Color is an especially important and contested topic both for debates about naturalism and for
the metaphysics of perceptual experience. It is very common to find scientists claiming that
physical objects are not really colored – or at least not in the way that we naively take them to
be – and that the qualitative natures of colors as we experience them are really just features of
our subjective human experiences, which we naively and mistakenly “project” onto external
physical items. This was the view of Galileo, Boyle, Newton, Maxwell and Helmholtz. And
more recently, such a view has been endorsed by leading vision scientists such as Edwin Land
(1983), Semir Zeki (1983) and Stephen Palmer (1999).37 For example:
People universally believe that objects look colored because they are colored, just as
we experience them. The sky looks blue because it is blue, grass looks green because it
is green, and blood looks red because it is red. As surprising as it may seem, these
beliefs are fundamentally mistaken. Neither objects nor lights are actually “colored” in
anything like the way we experience them. Rather, color is a psychological property
of our visual experiences when we look at objects and lights, not a physical property of
those objects or lights. The colors we see are based on physical properties of objects
and lights that cause us to see them as colored, to be sure, but these physical properties
are different in important ways from the colors we perceive.38
This view about the metaphysics of color is very often presented as a scientific result – established by such phenomena as subjective variation amongst human observers, metamers, the
contribution of opponent processing, the different color spaces for different species (see Hardin
1988 for a philosophically sophisticated presentation of this scientific case for “subjectivism”
about color.) Of course if this eliminativist view about the nature of color were correct, it
would immediately rule out Naïve-Realist/Relational accounts of visual experience – at least
insofar as there would be no external colors out there for us to be acquainted with that could
explain our “colorful” visual phenomenology. However many philosophers – and not just
Naïve Realists – have denied that vision science really does establish any such result about the
nature of color. The metaphysics of color has given rise to a very large literature, which could
be taxonomised in various ways and in more or less detail. But for present purposes we can
briefly identify four main positions:
1

Reductive physicalism
Colors are objective, physical properties of objects (and light sources). Contra the claims of
many vision scientists, colors are real, mind-independent properties located out in the
external environment, just as we naively suppose them to be, though they turn out to have
complex physical natures that goes beyond their manifest appearance and which is open to
311

Thomas Raleigh

empirical investigation. For example, Byrne and Hilbert (2003) and also Tye (2000) argue
that colors are (roughly) spectral reflectance properties – that is, a complex tendency to
reflect light only of certain wavelengths – which are ultimately grounded in the case of
surface colors in the microscopic texture of the object’s surface. This sort of view is an
avowedly reductive account of color that seeks to locate color as part of the natural, external
world by identifying it with a complex, scientifically respectable property. It is thus in line
with Strictly Scientific Naturalism.39
2

Primitivism
Primitivists also take colors to be real, mind-independent properties, located out in the
external environment. However unlike reductive physicalists, primitivists deny that these
properties can be reduced to some complex physical property of surfaces (or volumes, or
light sources etc.), or indeed to any other physical property. Colors are held to be sui
generis, simple, irreducible properties of physical items whose essential qualitative nature is
revealed in how they look – though it is generally allowed that these sui generis color
properties do at least supervene on some or other physical properties of the object (e.g.
perhaps the surface reflectance properties mentioned above). In denying that colors need to
be reduced to scientifically acceptable physical properties in order to be a genuine/real part
of the natural world, primitivism effectively amounts to a (limited/specific) form of Liberal
Naturalism.40

Dispositional and relational views
Dispositional theories of color, qua properties of external objects, treat them as dispositions
to cause certain kinds of experiences in certain kinds of conscious observers. One version
of this idea takes these to be dispositions to cause the experiences in “normal” observers in
“normal” viewing conditions – which then requires the dispositional theorist to give a
principled, non-circular account of what counts as normal. (See Peacocke 1984 for an
attempt to provide such a non-circular account. See Levin (2000) for arguments that no
such account is possible.) A different version of dispositionalism avoids this requirement by
relativising these dispositions to different perceivers and different viewing conditions –
hence the label “relational”. On this latter view objects do not, strictly speaking, possess such
colors as red or green or blue simpliciter. Rather they possess such relational color properties as:
red for perceiver x in viewing condition 1; green for perceiver y in viewing condition 2, etc.41
Dispositional views rely on distinguishing the dispositional property, or power, of an external object to cause certain experiences in certain observers (in certain conditions), from the
phenomenal properties of the experiences thus caused – that is, from color as experienced.
There is then a choice between treating this dispositional story as revisionary with respect to our
naïve color beliefs and color concepts, or treating it as vindicating or everyday beliefs and
concepts. The revisionary line was adopted by many of the great scientists mentioned above –
for example, Galileo, Boyle and Newton – as well as many early modern philosophers, such as
Locke and Descartes. They held that the color properties of physical objects are secondary
qualities, powers to cause certain experiences in us, but this is not how we naively, pretheoretically think of colors. This revisionary version of dispositionalism would also seem to be
in line with a Strictly Scientific form of naturalism insofar as it treats our manifest image of
colors as shown to be erroneous by higher scientific standards.
In contrast, a number of somewhat more recent dispositionalists have held that the dispositional view is in perfect accord with our naïve, common-sense conception of colors –
and indeed that colors as we experience them appear to be dispositional.42 This version of
312

Philosophy of perception

dispositionalism, if it could be made plausible, would avoid any conflict between science and
our everyday conception and experience of color. It is thus more amenable to Liberal Naturalists
– though it certainly does not require that one accept anything like Liberal Naturalism.
4

Error theoretic views
Error-theoretic or eliminativist views can be thought of as more extreme versions of the
revisionary kind of dispositionalism sketched above. For whereas the revisionary dispositionalist still allows that it is true to say that there really are colored objects (though these
external color properties are not how we naively think of or experience them), the error
theorist insists that (strictly speaking) there simply are no colors at all. Our experiences as of
colored objects are thus held to be systematically inaccurate and illusory. And so the only way
to make sense of our practice of attributing color properties to external objects is as a useful
fiction. One can also understand this sort of eliminativist position as agreeing with the
Primitivist that color properties are simple sui generis intrinsic properties, whose essential
qualitative nature is revealed in how they look – the eliminitavist simply goes on to deny that
these properties are actually instantiated anywhere (in our actual universe).43

As mentioned above, of these four main positions it is Primitivism which amounts in itself to a
form of Liberal Naturalism. Whereas Reductive Physicalism, revisionary dispositionalism and
Error theories are often motivated by something like Strict Scientific Naturalism. However, it is
clearly possible to be a Liberal Naturalist in general whilst holding any position, or none at all,
concerning the specific topic of colors. As a case in point: John McDowell endorses the nonrevisionary form of dispostionalism about colors, allowing that their nature is essentially tied to
our specific human kind of sensory experiences – though he does not think of such experience
as something “inner” or “private”:
It is one thing to gloss being red in terms of being such as to look red, and quite
another to gloss it in terms of being such as to induce a certain ‘inner experience’ in
us. Note that ‘red’ in ‘looking red’ expresses a concept of ‘outer experience’ no less
than does ‘red’ in ‘being red’, in fact the very same concept.44
Values are not brutely there – not there independently of our sensibility – any more
than colors are: though, as with colors, this does not prevent us from supposing that
they are there independently of any apparent experience of them.45
One particularly important discussion of color by a prominent Liberal Naturalist is Barry
Stroud’s “The Quest for Reality” (1999). Stroud considers the long history of scientists and
philosophers who claim to “unmask” colors as really existing only in the mind. His argument is
subtle and complex, but it hinges on considering whether these kind of skeptical conclusions
about color can even be coherently believed. The color skeptic wants to suggest that reality “in
itself” does not correspond to our thoughts and experiences of color. But Stroud claims that to
make sense of this skeptical conclusion we need to be able to understand which properties are
being (allegedly) misattributed to physical, external objects – that is, we need to be able to say
what the supposedly mistaken content is of our everyday color thoughts and experiences. And
he argues that the color skeptic cannot do this without relying on our everyday concept of
colors as qualities of external items. We need, according to Stroud, to think of the external
environment as genuinely colored in order to ascribe each other thoughts and experiences of
color. So in order to formulate her skeptical claim – that our thoughts and experiences of color
313

Thomas Raleigh

are systematically misleading – the skeptic has to draw on and employ the very conception of
colors that she is seeking to undermine. However, Stroud does not go on to draw the positive
conclusion that colors therefore must really be genuine qualitative features of our external
environments. Repeating his earlier general criticism of transcendental arguments (see Stroud
1984), Stroud points out that from the fact that: in order to understand our everyday color
thoughts and experiences, we have to think of the environment as genuinely containing colors,
it does not follow that the external environment really is that way. It remains a possibility that
our color thoughts and experiences are radically mistaken even if we cannot coherently disavow
such thoughts or fully understand what it would be for them to be so radically mistaken.

6 Conclusion
In summary then, it seems that Liberal Naturalism is at least logically compatible with any of the
main philosophical theories concerning perceptual experience. That there are no strict relations of
inconsistency or entailment here should be unsurprising given that (as mentioned back in
Section 1) Liberal Naturalism is not a precisely defined thesis, but rather a loose family of views
and kindred philosophical approaches. But there is plausibly still a significant negative connection
between Liberal Naturalism and the philosophy of perception insofar as Strictly Scientific
Naturalism would rule out Primitivism about color and so seem to be less congenial to NaïveRealism about perceptual experience. More generally, insofar as Liberal Naturalism holds that we
need feel no philosophical anxiety about the ontological status or scientific respectability of
various common-sense or “manifest image” phenomena – not only color but also everyday
objects and artifacts, mental states, morality, normative properties and so on – it will be a natural
ally to any theory of perception that holds we really can perceive such phenomena.46

Notes
1 Some prominent advocates of Liberal Naturalism include: P.F. Strawson, Hilary Putnam, John
McDowell, Huw Price, Richard Rorty, Jennifer Hornsby and Barry Stroud. Some would also interpret Wittgenstein as espousing a form of Liberal Naturalism – see Cahill and Raleigh (2018) for a
collection of essays focused on Wittgenstein’s Naturalism.
2 Some prominent advocates of Strictly Scientific Naturalism include: W.V.O. Quine, David
Armstrong, David Lewis, Frank Jackson, David Papineau and Alex Rosenberg.
3 Two particularly important and influential collections of essays devoted to the topic of Liberal
Naturalism are De Caro and Macarthur (2004), De Caro and Macarthur (2010).
4 It is plausible that at least most Liberal Naturalists would accept a weak supervenience thesis to the effect
that everything somehow or other supervenes on the total, universe-wide state of the fundamental
particles and forces that are investigated by physics. This sort of supervenience is widely accepted to be
necessary, though not sufficient, for some form of materialism or physicalism, since some nonnaturalist
or dualist positions could also accept such supervenience.
5 However, see Raleigh (2014), Ali (2016), Masrour (forthcoming) for papers which explore denying
the first premise. See also Stoneham (forthcoming) for similar exploration of this sort of move in the
context of dreams.
6 Metaphysical disjunctivism is standardly credited first to Hinton (1967, 1973). It was then developed
by Snowdon (1981) and McDowell (1982). Metaphysical disjunctivism is expounded and defended
most extensively in the work of M.G.F. Martin (2004, 2006).
7 Many thanks to Craig French for helpful correspondence.
8 The precise relationship between epistemological disjunctivism and metaphysical disjunctivism is a
substantive question. But prima facie, it seems that the two theses are independent – one could endorse
either one without endorsing the other. Thus an epistemological disjunctivist could allow that, or at
least be agnostic whether, there is a common kind of conscious phenomenal character in both good
and bad cases. And conversely, as mentioned in the main text, above, it seems that a metaphysical

314

Philosophy of perception

9
10

11
12

13
14
15
16
17
18
19
20
21
22
23

24
25
26
27
28
29
30
31
32

disjunctivist could accept that we are in a common epistemic position in both good and bad cases. See
French (2016, 2019) for further discussion.
However, some coherentists do allow that experiences play some kind of important epistemic role –
see, for example, Rescher (1973) and Bonjour (1985).
Note: whilst more recent sense-data theorists such as Robinson (1994) and Foster (2000) take Sense-Data
to be mental particulars, Russell (1912) and Moore (1910) insisted that are not mental or minddependent, though they are supposed distinct from anything in our physical environments. A further
complication: Bermudez (2000) and Jackson (1977) both take sense-data to be located out in normal
physical space, though they are still mind-dependent. Note also: as Russell and Moore originally introduce the term, “sense-data” is just meant to refer to whatever it is that we are directly conscious of – so
it is not part of the definition of “sense-data” that they are nonenvironmental or nonphysical entities,
though this is something we may discover by reflection, introspection and philosophical argument.
However, see Beck (2019) for a version of the Relational view which denies that phenomenal
character is constituted by external items/features.
Though it is typically denied that we perceive Sense-Data, perceptual contact with the external world is
indirect insofar as we perceive our physical objects and features in virtue of being directly conscious of
something else, the Sense-Data, distinct from those physical features. In contrast, Adverbial theorists
deny that experience has an “act-object” structure and so they are not obviously committed to the idea
of a mental intermediary. Nevertheless, they would deny that in perceptual experience we are “open to
the world” and would accept that in both the metaphysical and epistemic senses, what we are “given”
is fundamentally the same in both perceptual and hallucinatory cases. Thanks once more to Craig
French for helpful discussion.
For example, McDowell, at least at the time of Mind and World (1994), espoused a hybrid position
combining elements of Naïve Realism and Representationalism.
For examples of “sparse” accounts (see Brogaard 2013; Byrne 2009; Pautz 2008; Prinz 2013; Sosa
2015; Tye 1995, 2000).
For examples of “rich” accounts (see Bayne 2009; Butterfill 2009; Scholl and Tremoulet 2000; Siegel
2006, 2009; Nanay 2011).
Locke ([1690], 1975), II. iii.9, 135).
See Allen (2020) for an extended argument that Naïve-Realism enjoys a “special status” amongst
philosophical theories of perception.
See Raleigh (2020) for discussion of this line of thought with respect to macroscopic shape.
See Raleigh (2018) for discussion of how Naïve-Realism might accommodate a physicalist, nonprimitivist account of color.
The label “structural realism” is due originally to Maxwell (1970).
Roughly the same position has also been given the rather more pessimistic/skeptical labels “Kantian
Humility” (Langton 1998) and “Ramseyan Humility” (Jackson 1998).
At least, this was Russell’s initial view at the time of Problems of Philosophy. By the time we get to
“Mysticism and Logic and other Essays” (1918), Russell had become a neutral monist and so held that
sense-data were ultimately physical.
It is perhaps worth noting that there may be possible sense-data theories in logical space on which
sense-data are simply identical to some physical feature that is acceptable to Strictly Scientific Naturalism
– for example, if sense-data are simply identical to physical processes in the brain. Indeed this may be
true of Russell’s later neutral monist account of sense-data – see fn. 20 above. Thanks to both Craig
French and Ori Beck for helpful discussion on this point.
For example, McDowell writes: “The idea of value experience involves taking admiration to represent
an object as having a property that…is conceived to be not merely such as to elicit [admiration]…but
rather such as to merit it”. (1985, 118).
Putnam (1999, 11).
McDowell (2013, 259/260).
McDowell (1995, 877–878).
McDowell (1986, 150, emphasis added).
McDowell (2009, 258).
See fn. 1 in Putnam (1999).
Putnam (1999, 11).
Ibid., 159.

315

Thomas Raleigh
33 Putnam wrote of Block’s influence on him:
Two papers by Ned Block – a lecture titled ‘Wittgenstein and Qualia’, which I heard Block
deliver at the ‘Putnam Fest’ conference in my honor in Dublin in 2007, and a paper of his titled
‘Consciousness, Accessibility, and the Mesh between Psychology and Neuroscience’, which
appeared later in the same year…have had an impact on my thinking about the phenomenology
of perception comparable to the impact on my later philosophy of mathematics that reading
Quine’s ‘On What There Is’ (1948) and ‘Two Dogmas of Empiricism’ (1951) in my twenties
turned out to have. (2016, 169–170).
34
35
36
37
38
39
40

41
42
43
44
45
46

Putnam (2016, 156, [emphasis added]).
Ibid., 184.
Ibid., 167.
Byrne and Hilbert (2003) comment: “In fact, the most popular opinion, at any rate among color scientists, may
well be the view that nothing is colored – at least not physical objects in the perceiver’s environment…” (3)
Palmer (1999, 95).
See, for example, Armstrong (1969), Hilbert (1987), Byrne and Hilbert (2003), Tye (2000).
See, for example, Hacker (1987), Campbell (1994), McGinn (1996), Gert (2008), Allen (2016). Strictly
speaking, some kind of non-naturalism about colors might also allow that colors supervene on features
described by the physical sciences and that colors are “genuine” or “real”, though without being part
of the “natural” realm. Many thanks to Anders Nes for helpful correspondence on this point.
See Cohen (2009) for excellent defense of this relational view.
See, for example, Dummett (1979), McDowell (1985), Peacocke (1984).
See, for example, Hardin (1988), Boghossian and Velleman (1989), Averill (2005) andMaund (2006).
McDowell (1994, 31 fn 7).
McDowell (1998b, 146).
I am extremely grateful to Ori Beck, Craig French and Anders Nes, all of whom generously provided
me with superbly helpful written comments on an earlier draft of this chapter. Needless to say, they are
in no way responsible for any errors or infelicities that remain and should not be assumed to endorse
any of the opinions I express in this article. My thanks also to the editors, David Macarthur and Mario
De Caro, for all their hard work and patience in putting this volume together.

References
Ali, R. Does Hallucinating Involve Perceiving? Philosophical Studies, 175(3), (2016). 601–627.
Allen, K. A Naïve-Realist Theory of Colour. Oxford University Press. (2016).
Allen, K. The Value of Perception. Philosophy and Phenomenological Research, 100(3), (2020). 633–656.
Armstrong, D. Color-Realism and the Argument from Microscopes. In Brown and Rollins (eds.),
Contemporary Philosophy in Australia. Allen and Unwin (1969).
Averill, E. Toward a Projectivist Account of Color. The Journal of Philosophy, 102(5), (2005). 217–234.
Bayne, T. Experience and Content. Philosophical Quarterly, 59(236), (2009). 429–451.
Beck, O. Rethinking Naive Realism. Philosophical Studies, 176(3), (2019). 607–633.
Bermudez, J. Naturalized Sense Data. Philosophy and Phenomenological Research, 61(2) (2000). 353 – 374.
Boghossian, P. and Velleman, D. Colour as a Secondary Quality. Mind, 98(389). (1989). 81–103.
Bonjour, L. The Structure of Empirical Knowledge. Harvard University Press. (1985).
Bonjour, L. A Version of Internalist Foundationalism. In L. Bonjour and E. Sosa (eds.), Epistemic
Justification: Internalism vs. Externalism, Foundations vs. Virtues. Blackwell. (2003).
Brewer, B. Perception and its Objects. Oxford University Press. (2011).
Broad, C.D. The Mind and Its Place in Nature. Routledge & Kegan Paul. (1925).
Brogaard, B. Do We Perceive Natural Kind Properties? Philosophical Studies, 162(1), (2013). 35–42.
Butterfill, S. Seeing Causings and Hearing Gestures. The Philosophical Quarterly, 59(236), (2009). 405–428.
Byrne, A. Intentionalism Defended. Philosophical Review, 110(2), (2001). 199–240.
Byrne, A. Experience and Content. Philosophical Quarterly, 59(236), (2009). 429–451
Byrne, A. and Hilbert, D. Color Realism and Color Science. Behavioral and Brain Sciences, 26(1),
(2003). 3–21.

316

Philosophy of perception
Cahill, K. and Raleigh, T. (eds.) Wittgenstein & Naturalism. Routledge. (2018).
Campbell, J. A Simple View of Color. In John Haldane and Crispin Wright (eds.), Reality, Representation
and Projection, pp. 257–269. Clarendon Press. (1994).
Campbell, J. Reference and Consciousness. Oxford University Press. (2002).
Campbell, J. Consciousness and Reference. In B. McLaughlin, A. Beckermann, and S. Walter (eds). Oxford
Handbook of Philosophy of Mind Oxford. Oxford University Press. (2009).
Chalmers, D. The Conscious Mind. Oxford University Press. (1996).
Chalmers, D. The Representational Character of Experience. In B. Leiter (ed.), The Future for Philosophy.
Oxford University Press. (2004).
Chisholm, R. Perceiving: A Philosophical Study. Cornell University Press. (1957).
Cohen, J. The Red and the Real: An Essay on Color Ontology. Oxford University Press. (2009).
Cornman, J. Materialism and Sensations. Yale University Press. (1971).
Davidson, D. A Coherence Theory of Truth and Knowledge. In Subjective, Intersubjective, Objective.
Oxford University Press. ([1983]/2001).
De Caro, M. and Macarthur, D. (eds.) Naturalism in Question. Harvard University Press. (2004).
De Caro, M. and Macarthur, D. (eds.) Naturalism and Normativity. Columbia University Press. (2010).
Dretske, F. Experience as Representation. Philosophical Issues, 13(1), (2003). 67–82.
Ducasse, C.J. Moore’s Refutation of Idealism. In P. Schilpp (ed.), The Philosophy of G. E. Moore,
pp. 223–252. Northwestern University Press. (1942).
Dummett, M. Common Sense and Physics. In G.F. MacDonald (ed.), Perception and Identity.
Macmillan. (1979).
Foster, J. The Nature of Perception. Oxford University Press. (2000).
French, C. The Formulation of Epistemological Disjunctivism. Philosophy & Phenomenological Research,
92(1), (2016). 86–104.
French, C. Epistemological Disjunctivism and its Representational Commitments. In Duncan Pritchard,
Casey Doyle, and Joe Milburn (eds.), New Issues in Epistemological Disjunctivism. Routledge. (2019).
Gert, J. What Colors Could Not Be: An Argument for Color Primitivism. Journal of Philosophy, 105,
(2008). 128–157.
Hacker, P. Appearance and Reality: A Philosophical Investigation into Perception and Perceptual Qualities.
Blackwell. (1987).
Hardin, C. Color for Philosophers: Unweaving the Rainbow. Hackett. (1988).
Harman, G. The Intrinsic Quality of Experience. In J. Tomberlin (ed.), Philosophical Perspectives, Vol. 4,
pp. 31–52. Ridgeview. (1990).
Hilbert, D.R. Color and Color Perception: A Study in Anthropocentric Realism. CSLI Publications. (1987).
Hinton, J. Visual Experiences. Mind, 76, (1967). 217–227.
Hinton, J. Experiences: An Inquiry into Some Ambiguities. Clarendon Press. (1973).
Jackson, F. Perception: A Representative Theory. Cambridge University Press. (1977).
Jackson, F. From Metaphysics to Ethics: A Defence of Conceptual Analysis. Oxford University Press. (1998).
Johnston M. Better Than Mere Knowledge? The Function of Sensory Awareness. In T. Gendler and J.
Hawthorne (eds.), Perceptual Experience. Oxford University Press. (2006).
Land, E.H. Recent Advances in Retinex Theory and Some Implications for Cortical Computations:
Color Vision and the Natural Image. Proceedings of the National Academy of Sciences, 80(16), (1983).
5163–5169.
Langton, R. Kantian Humility. Oxford University Press. (1998).
Levin, J. Dispositional Theories of Color and the Claims of Commonsense. Philosophical Studies, 100(2),
(2000). 151–174.
Logue, H. Why Naïve Realism?. Proceedings of the Aristotelian Society, 112, (2012). 211–237.
Locke, J. An Essay Concerning Human Understanding. Ed. P. Nidditch. Clarendon Press. ([1690]/1975).
Macarthur, D. McDowell, Scepticism, and the ‘Veil of Perception’. Australasian Journal of Philosophy, 81(2),
(2003). 175–190.
Martin, M.G.F. Setting Things Before the Mind. In A. O’Hear (ed.), Contemporary Issues in the Philosophy
of Mind, pp. 157–180. Cambridge University Press. (1998).
Martin, M.G.F. The Transparency of Experience. Mind and Language, 17, (2002). 376–425.
Martin, M.G.F. The Limits of Self-Awareness. Philosophical Studies, 120, (2004). 37–89.
Martin, M.G.F. On Being Alienated. In T.S. Gendler and J. Hawthorne (eds.), Perceptual Experience.
Oxford University Press. (2006).

317

Thomas Raleigh
Masrour, F. On the Possibility of Hallucinations. Mind, 129(515), (2020). 737–768.
Maund, B. The Illusion Theory of Colours: An Anti-Realist Theory. Dialectica, 60(3), (2006). 245–268.
Maxwell, G. Structural Realism and the Meaning of Theoretical Terms. Minnesota Studies in the Philosophy
of Science, 4, (1970). 181–192
McDowell, J. Criteria, Defeasibility and Knowledge. Proceedings of the British Academy, 68, (1982).
455–479.
McDowell, J. Values and Secondary Qualities. In Ted Honderich (ed.), Morality and Objectivity: A Tribute
to J. L. Mackie, pp. 110–127. Routledge and Kegan Paul. (1985).
McDowell, J. Singular Thought and the Extent of ‘inner space’. In John McDowell and Philip Pettit
(eds.), Subject, Thought, and Context. Clarendon Press. (1986).
McDowell, J. Mind and World. Harvard University Press. (1994).
McDowell, J. Knowledge and the Internal. Philosophy and Phenomenological Research, 55(4), (1995).
877–893.
McDowell, J. The Disjunctive Conception of Experience as Material for a Transcendental Argument. In
A. Haddock and F. Macpherson (eds.), Disjunctivism: Perception, Action, Knowledge. Oxford University
Press. (2008).
McDowell, J. Avoiding the Myth of the Given. In The Engaged Intellect. Harvard University Press. (2009).
McDowell, J. Perceptual Experience: Both Relational and Contentful. European Journal of Philosophy,
21(1), (2013). 144–157.
McGinn, C. Another Look at Color. Journal of Philosophy, 93(11), (1996). 537–553.
Moore, G.E. Sense-Data. Reprinted in Some Main Problems in Philosophy. Allen & Unwin. ([1910]/1953).
Pautz, A. What Are the Contents of Experiences? The Philosophical Quarterly, 59(236), (2008). 483–507.
Nanay, B. Do We See Apples as Edible? Pacific Philosophical Quarterly, 92(3), (2011). 305–322.
Palmer, S. Vision Science: Photons to Phenomenology. MIT Press. (1999).
Peacocke, C. Colour Concepts and Colour Experience. Synthese, 58(3), (1984). 365–381.
Price, H.H. Perception. Methuen. (1932).
Prinz, J. Siegel’s Get Rich Quick Scheme. Philosophical Studies, 163(3), (2013). 827–835.
Pritchard, D. McDowellian Neo-Mooreanism. In F. Macpherson and A. Haddock (eds.), Disjunctivism:
Perception, Action, Knowledge. Oxford University Press. (2008).
Pritchard, D. Epistemological Disjunctivism. Oxford University Press. (2012).
Pryor, J. The Skeptic and the Dogmatist. Noûs,34(4), (2000). 517–549.
Putnam, H. The Threefold Cord: Mind, Body and World. Columbia University Press. (1999).
Putnam, H. Naturalism, Realism and Normativity. Ed. M. De Caro. Harvard University Press. (2016).
Raleigh, T. A New Approach to ‘Perfect’ Hallucinations. Journal of Consciousness Studies, 21(11–12),
(2014). 81–110
Raleigh, T. Visual Acquaintance, Action & The Explanatory Gap. Synthese, (2018). doi: 10.1007/s11229018-01907-x].
Raleigh, T. Science, Substance and Spatial Appearances. Philosophical Studies, 177(8), (2020). 2097–2114.
Rescher, N. The Coherence Theory of Truth. Oxford University Press. (1973).
Robinson, H. Perception. Routledge. (1994).
Russell, B. The Problems of Philosophy. Oxford University Press. (1912).
Russell, B. Mysticism and Logic and Other Essays. Longmans, Green & Co. (1918).
Scholl, B.J. and Tremoulet, P.D. Perceptual Causality and Animacy. Trends in Cognitive Sciences, 4(8),
(2000). 299–309.
Sellars, W. The Structure of Knowledge. In H-N. Castañeda (ed.), Action, Knowledge, and Reality,
pp. 295–347. Bobbs-Merrill. (1975).
Siegel, S. Which Properties Are Represented in Perception? In T.S. Gendler and J. Hawthorne (eds.),
Perceptual Experience. Oxford University Press. (2006).
Siegel, S. The Visual Experience of Causation. Philosophical Quarterly, 59(236), (2009). 519–540.
Snowdon, P. Perception, Vision and Causation. Proceedings of the Aristotelian Society, 81, (1981). 175–192.
Sosa, D. What Does it Matter What it’s Like?. Philosophical Issues, 25(1), (2015). 224–242.
Stoneham, T. Dreaming, Phenomenal Character, and Acquaintance. In J. Knowles and T. Raleigh (eds.),
Acquaintance: New Essays. Oxford University Press. (2019).
Stroud, B. The Significance of Philosophical Skepticism. Oxford University Press. (1984).
Stroud, B. The Quest for Reality: Subjectivism and the Metaphysics of Colour. Oxford University Press. (1999).
Travis, C. The Silence of the Senses. Mind, 113(449), (2004). 57–94.

318

Philosophy of perception
Tye, M. The Adverbial Theory: A Defence of Sellars Against Jackson. Metaphilosophy, 6(April), (1975).
136–143.
Tye, M. The Adverbial Theory of Visual Experience. Philosophical Review, 93, (1984). 195–225.
Tye, M. Ten Problems of Consciousness. MIT Press. (1995).
Tye, M. Consciousness, Color, and Content. MIT Press. (2000).
White, S. Phenomenology and the Normativity of Practical Reason. In De Caro M. and D. Macarthur
(eds.), Naturalism and Normativity. Columbia University Press. (2010).
Wright, C. Comment on John McDowell’s ‘The Disjunctive Conception of Experience as Material for a
Transcendental Argument’. In A. Haddock and F. Macpherson (eds.), Disjunctivism: Perception, Action,
Knowledge. Oxford University Press. (2008).
Zeki, S. Colour Coding in the Cerebral Cortex: The Reaction of Cells in Monkey Visual Cortex to
Wavelengths and Colours. Neuroscience, 9, (1983). 741–765.

319

26
ETHICS AND LIBERAL
NATURALISM
Hans Fink

1 Introduction
During the first half of the twentieth century, it was the orthodoxy within analytical philosophy
that ethical naturalism was simply a mistake. To be a naturalist in ethics was to commit the
naturalistic fallacy as it was first exposed by G.E. Moore with his “open question argument” in
Principia Ethica (1903) or by later reformulations in A.J. Ayer’s Language, Truth and Logic (1936)
or R.M. Hare’s The Language of Morals (1952). It was widely accepted that contrary to the
claims of earlier naturalists, there was an ontological and epistemic gap between empirical facts
on the one hand and ethical values and norms on the other, or a logical gap between ordinary
descriptions on the one hand and moral evaluations and prescriptions on the other.
Moore had accused Bentham, Mill and Spencer of committing the fallacy of not noticing
this gap simply by offering definitions in natural, empirical terms of the indefinable and nonnatural property which he took the adjective “good” to be standing for. He did this at the point
in time when psychology and sociology were newly established university departments eager to
show their credentials as morally neutral, empirical sciences and when philosophy was redefining itself in logistic and antipsychologistic terms. Here was common ground between the
analytical philosophy originating in Cambridge under the influence of Russell and Moore and
the phenomenological movement beginning in Göttingen under the influence of Husserl. The
situation was somewhat different in the USA due to the influence of pragmatism, but even here
he value-freedom of science and the autonomy of ethics became broadly accepted ideas.
In the decade after WWII both logical positivism, existentialism and Hare’s prescriptivism
retained the idea of an insurmountable gap between facts and values, but moral values and
norms were no longer seen as somehow more real and rational than the merely empirical, but
rather as less real and rational because basically a matter of emotional or volitional attitudes and
not a matter of true or false beliefs. Ethics was necessarily based on noncognitive personal
commitments alien to the objective stance of the natural and social sciences. No normative
“ought” could be derived from a descriptive “is”. Ethical naturalists were philosophers who had
failed to understand this.
In the late 1950s and early 1960s, some moral philosophers began losing their fear of the
naturalistic fallacy and challenging or softening the sharp distinctions standardly taken to separate facts from values and norms, descriptions from evaluations and prescriptions. They
320

DOI: 10.4324/9781351209472-31

Ethics and liberal naturalism

reacted against such antinaturalistic noncognivism by showing that moral evaluations need not
and should not be taken to be free-floating in relation to descriptions of the facts of the situation
in question.1
In the 1970s influential works by among others, John Rawls, Jürgen Habermas and P.F.
Strawson opened new ways of understanding ethics in cognitive terms. At the same time, new
forms of scientific naturalism following Quine’s call for “naturalisation” in philosophy became
highly influential and gave rise to forms of ethics relying on developments in evolutionary
biology or experimental psychology often combined with forms of utilitarianism. Even more
moral philosophers, however, reacted against this tendency not by defending some form of
antinaturalism, but rather by presenting liberal forms of naturalism that did not prioritise the
natural sciences while claiming that the norms and values we live by should nevertheless be
understood as ordinary and necessary aspects of the natural world that we live in and that
scientists study certain, but far from all aspects of. Moral discourse should not be viewed as
clearly separable from but rather seen in continuity with factual discourse. The last decades have
thus seen much discussion between two ways of understanding what naturalism both in general
and in ethics would amount to. On the one hand there is what Strawson called “hard naturalism”, “strict naturalism” or “reductive naturalism”, on the other hand, what he called
“liberal naturalism”, “soft naturalism”, “nonreductive naturalism”, “social naturalism” or
“catholic naturalism”.2
I have been a participant observer to these developments, especially in ethics. I began
studying philosophy in the 1960s at Aarhus University, Denmark in a department dominated by
ordinary language philosophy as practiced in Oxford. I specialised in ethics which at that time
just meant metaethics, and I wrote my very first paper defending the idea of a logical gap
between “is” and “ought”, but soon began questioning that idea and defending a form of
ethical naturalism that could be called liberal naturalism though that term was not available at
the time. To distinguish my view from many other ethical theories that have an equally good
claim to the term liberal naturalism, I often call it “straightforward naturalism” or “absolute
naturalism”. In what follows, I shall explain what this amounts to, allowing myself to take my
own work as the point of departure for saying something more general about the enormous and
extraordinarily complex topic of ethics and liberal naturalism. Here, I can do little more than
give a rough and sketchy presentation of the version I favor.

2 No naturalistic fallacy
I wrote my M.A. thesis (in Danish) in 1968 on what was then the most recent discussion of the
naturalistic fallacy. I argued that both Moore in his original characterisation of the fallacy and
Hare in his reformulation of it took conceptions of ethics and conceptions of naturalism for
granted that no one, and certainly no naturalist, should feel forced to accept. Moore simply
assumed that the moral value referred to by “good” is a unique and indefinable object of
intuitive experience, and that any form of naturalism (or, indeed, supernaturalism) involves
defining what is indefinable or mistaking a synthetic proposition about goodness for an analytic.
But naturalists could just deny that this is how moral goodness should be understood, and that
they need to commit such fallacies, even if some naturalists might have done so. They should
not see themselves as trying to define the ethical in nonethical terms, but rather to show that
ethical terms need not be understood as super- or subnatural.
Hare similarly took it for granted that words like “good” and “ought” have a commending
and prescriptive force that descriptive words lack, and that naturalists are engaged in attempts to
understand the ethical using only terms that lack the force necessary for being ethical. But again,
321

Hans Fink

naturalists could and should just deny that there is a clear and logically relevant difference
between descriptive terms on the one hand and commending and prescriptive terms on the
other. They should not see themselves as trying to replace commending and prescriptive terms
with terms that are not commending or prescriptive, but rather to explain how value terms can
have commending force while being at the same time wholly descriptive, and to explain how
descriptions and prescriptions may well fit seamlessly together in arguments. There is no way of
knowing in advance that this attempt is bound to fail.
My conclusion was that naturalists may commit all kinds of mistakes, but that there is no fair
and neutral ground on which it has been or can be shown that there is one fallacy that they
necessarily commit.3 Hare was therefore wrong when he had claimed that naturalism in ethics
was a on a par with attempts to square the circle and to justify induction (92). This could be the
case only if he in all seriousness relied on the glaringly question begging definition of naturalism, he had given at 82: “The term [“naturalism”] has, unfortunately, since Moore’s introduction of it, been used very loosely. It is best to confine it to those theories against which
Moore’s refutation (or a recognizable version of it) is valid”.
I further discussed Hare’s position in relation to recent papers by Philippa Foot and John Searle
and I sided with the latter in their rather different attacks on the metaethical orthodoxy that Hare
represented. I also briefly discussed how the distinction between metaethics and normative ethics
may itself install a sharp fact-value distinction that fits better with antinaturalism than with naturalism. In arriving at my conclusions, I was much influenced by John Dewey. He was one of the
few major philosophers who had consistently resisted all forms of antinaturalism without ever
coming close to a restrictive scientism. His naturalism was both a naturalistic humanism seeing
everything human as natural, and a humanistic naturalism seeing science as a human endeavor to
understand our own nature in its continuity with the nature of everything else in nature.4 I had
been translating some of his works into Danish and was directly inspired by his paper
“Antinaturalism in Extremis” in the collection of essays Naturalism and the Human Spirit, edited by
Y.H. Krikorian (1944). He there complained against his opponents: “The very fruit of antinaturalism is made the ground of attack upon naturalism” (10) and that seemed a precise characterisation of what Moore and Hare had been doing. They both took an antinaturalistic moral
exceptionalism for granted; Moore by seeing what he called the science of ethics as somehow
ontologically elevated above and epistemologically more certain than the merely factual and
fallible empirical sciences, Hare by seeing moral language as involving a personal commitment
making it in principle less answerable to facts than descriptive and scientific language.
I followed Dewey in seeing naturalism in ethics as the philosophical attempt to avoid moral
exceptionalism. Human life is life lived in the light of ideals and standards for individual and
collective behavior, and their ontological and epistemological status has been discussed and
revised throughout the ages. They have been seen as divine, absolute, eternal, universal, purely
rational and thus somehow more rigid and objective than anything in the natural human life
that they regulate, or they have been seen as human, all too human, in a way that makes them
purely emotional, merely subjective or conventional and relative to time and place, and thus
action guiding for those who accept them but less rigid and objective than factual knowledge.
Ethical naturalism can be understood to be the attempt to avoid both such inflation and such
deflation of the ethical. It is not so much a unified doctrine or a research program as it is a
consistent anti-antinaturalism. If there were no influential antinaturalistic views around there
would be little point in claiming to be a naturalist in ethics, one could just proceed with trying
to understand the ethical in human life better, but as things are, it remains important to stand up
for ethical naturalism against widespread explicitly antinaturalistic assumptions, and equally
important to avoid simplistic, reductive and naïve naturalistic accounts of values and norms.
322

Ethics and liberal naturalism

3 A naturalistic understanding of nature
From Dewey, I also learned that naturalism is not a self-explaining term and that a lot hinges on
which conception of nature a naturalist (or an antinaturalist) takes for granted or argue for.
“There is no word in the history of thought which carries more varied meanings than ‘nature’;
naturalism shares in its diverse significations” as he put it.5 The disagreement between ethical
naturalists and ethical antinaturalists is as often due to differences in their conceptions of nature
and the natural as it is due to differences in their conceptions of ethics and the ethical. On one
possible conception of nature everything is natural, and nature is all there is, so, of course,
everything ethical is natural and is to be understood as such. On other possible conceptions of
nature there are alternatives to being natural, and nature is understood as a realm contrasting
with other realms regarded as somehow supernatural, unnatural, or nonnatural. On such restricted conceptions of nature, it may be equally obvious that the ethical cannot be understood
without reference to something regarded as somehow beyond nature, be it divine or purely
rational; be it personal decisions or cultural traditions.
There is no neutral ground on which to decide between comprehensive or restricted
conceptions of nature. Intellectual history has left us with many different and mutually incompatible conceptions, and our concept of nature is not just complex but hypercomplex.
Nevertheless, it seemed to me then, as it does now, that a consistent naturalist must be prepared
to argue for an all-inclusive, comprehensive understanding of nature. There would be no point
in calling a philosophical position “naturalism” if it did not involve the metaphysical (or, if you
like, antimetaphysical) claim that everything is natural, and nature is all there is. We are natural
beings through and through and our culture shows how nature can also be. Ethical naturalists
should defend naturalism as the default position and claim that to think otherwise is to mystify
or dilute the ethical while at the same time trivialising nature and the natural. The burden of
proof should be on ethical antinaturalists to come up with a clear account of that which they
regard as other than nature and yet ethically relevant for natural human beings in nature. They
should not be allowed to take a restricted conception of nature for granted without argument.
And ethical naturalists should themselves be careful not to equivocate between comprehensive
and restricted meanings of the word “nature” the way scientistic naturalist do.
That nature is best understood in such a comprehensive way is defended by John Stuart Mill
in his posthumously published essay “Nature”. He took the nature of something in particular to
be all that determines how it may influence or be influenced by everything else, and he took
nature in general to be the totality of the systematic interconnection of all particular natures. He
even claimed that this is the correct and truly scientific meaning of the word “nature”.6 I don’t
think that there is any way in which this could be established, but I do think that Mill is right in
singling out such a comprehensive and radically antidualistic conception of nature as somehow
privileged over the many different contrasting conceptions, in that all the different contrasts that
these conceptions appeal to can, and in a way must, be seen as contrasts within nature rather
than contrasts with nature. To contrast with something else, it is necessary to have a nature in
the sense given by Mill and thus automatically to belong to nature in the comprehensive sense.
The naturalistic point is that nature should not be regarded as merely the uncultivated or less
cultivated parts of the surface of the earth, or merely the external world, or merely our physical
and biological origins. We ourselves and all our culture, all our knowledge, all our misunderstandings, all our dreams, all our fictions, all our phantasies, and, indeed, everything
ethical have their own nature and belong to nature and are to be understood for what they are
in their relations to everything else within that all-encompassing context. This is straightforward
naturalism because it does not await the success of any particular program of naturalisation.
323

Hans Fink

And it is absolute naturalism in the Spinozean sense that nature is completely independent of
everything else just because there is nothing else it could be dependent on. “Supernatural” is
neither available as a category insulating beliefs from critique nor as a term of abuse, and “the
unnatural” or “the nonnatural” should not be taken to belong to separate metaphysical realms
of their own.
Hard naturalists who take the results of the natural sciences as we now know them as
defining of nature are passing a restricted conception of nature off as if it was comprehensive.
Scientists who stick to their science will not equivocate here. An initial abstraction from everything human in nature was crucial for the establishment of branches of science that could
exorcise anthropomorphism and teleology from the study of lifeless nature. This undoubtedly
allowed great advances which, however, threaten to function ideologically the moment a
metaphysical claim to universality is based on this initial abstraction, placing everything that has
been abstracted from in a metaphysical limbo. The purposiveness of human activities, including
the activities of the scientists themselves, needs to be fully accepted and explained, and not
explained away as somehow less real or less natural than the interactions of fermions and bosons.
Appearances must be saved, explained and accepted as real appearances for natural human
beings in nature. Anthropomorphic terms need not be exorcised from the scientific study of
human beings. Moral values and norms should not first be excluded from nature and then
included again only on the condition that they can be fully interpreted in terms of what was left
as nature after their exclusion.
It thus seems clear to me that liberal naturalists are right to accuse reductive naturalists of
equivocating or flip-flopping between a restricted and a comprehensive understanding of
nature, and of having an unduly restricted sense of science and scientific method which is,
according to Dewey’s more liberal conception, “after all but systematic, extensive and carefully
controlled use of alert and unprejudiced observation and experimentation in collecting, arranging, and testing evidence…”7 But it seems to me that many forms of liberal naturalism are
inconsistent in their understanding of nature, too. McDowell, for one, retains a sharp distinction between the first and the second nature of human beings, where our second nature is
explicitly regarded as sui generis relative to our first nature. A dialectical relation is set up between on the one hand the first-person and normative perspectives characteristic of our second
nature and of ethics and on the other the third-person and causal law perspective characteristic
of our first nature and of the natural sciences. This ignores the fact that first nature is the result
of an abstraction rather than a separate, existing realm and it leaves open the question of the
exact character of the relation between first and second nature, between first and third person
perspectives, and between the realm of law and the normative space of reasons. The gap between first and second nature is presented as categorical and therefore threatens to become a
dualism in spite of the supposed naturalness of our second nature; but here the hard naturalist’s
claim that there is one nature only is surely right.8

4 A naturalistic understanding of value in general
I did my D.Phil. at Oxford University in the early 1970s having both Foot and Hare as supervisors. My thesis was called The Analysis of Goodness and was an attempt to give a naturalistic
and descriptivist account of our use of the word “good” in all its varieties. I was much inspired
by The Varieties of Goodness by G.H. von Wright, though trying to establish a kind of unity
underlying the varieties in a way he explicitly didn’t. I was also influenced by the account of the
semantics of “good” given by Paul Ziff in his Semantic Analysis. Rather than challenging the
fact-value distinction by discussing so called “thick” ethical terms like “rude” “honest”, or
324

Ethics and liberal naturalism

“courageous”, I tried to establish that the “thinnest of thin” value terms, “good”, in all its uses,
including uses in moral and aesthetic contexts, is an incomplete term always working in tandem
with other terms. There is no goodness simpliciter or tout court, just as there is no bigness or
fastness as such. To be good is to be better than certain others or something else in one or more
respects that can be explicated but are often left to be implicitly given by the context. We could
not know what a good X is without knowing what type X belongs to and in what context it is
seen, and here it is no use to say that to call X good just is to commend it or to say that it is to be
preferred, because we no more know what commending or preferring would amount to
without knowing something about what X is and what context it is seen in. A comparison with
others in a universe of comparison and along a dimension of comparison that is more or less
precisely specified in the context is always involved.
A key move was then to argue that the specification of the dimension of comparison is
always a matter of specifying a process that can be performed more or less well. The use of the
positive form of the adjective “good” is always somehow dependent on the comparative,
adverbial form “better than” as used of some process more or less directly associated with the
item evaluated.9 There could be no goodness in a static universe. Things or persons can be
more or less good at doing something. This is what von Wright calls technical and instrumental
goodness. States of affairs, events, or actions can be more or less good according to how they
contribute to the degree of success of some process taken as given. This is utilitarian or auxiliary
goodness. Comparative, adverbial evaluation is somehow presupposed in all positive, adjectival
evaluation.
But what is it for a process to go better than another similarly specified process? That
depends entirely on the verb in question. Some types of processes are simply defined in terms of
the achievement of a result that allow of degrees. To clean is to make cleaner, and to clean
better just is to achieve a higher degree of cleanness than is achieved in another attempt at
cleaning. Other processes need a supplementation of the meaning of the verb for them to be
qualified by the adverb “better than”. To run better is normally implicitly taken to be to run
faster, but if explicitly stated it could also be to run for a longer time, or to run more gracefully.
The world of sports offers examples of how performances are meticulously defined so that it can
be decided as objectively as possible who is the most successful.
Comparative, adverbial evaluation can thus be a completely straightforward factual matter if
the process in question is one-dimensional and well specified, and when it is not well specified,
this is the direct reason for the lack of precision of the evaluation, not something inherent in the
value predicate as such. Often we use the word “good” quite loosely, but we can always make
our ideas more precise by specifying a universe of comparison and a processual dimension of
comparison. “Good” is never used as a mere projection of warm feelings on an otherwise
neutral world. No matter how much I like the blackbird on my lawn this is not going to
make it a good blackbird. We actually don’t know what a good blackbird would be supposed to
be because no process is directly associated with being a blackbird the way certain processes are
being associated with being a carrier pigeon or a laying hen. We can, however, rank blackbirds
the moment a relevant process is taken as base. “Good” is often said by liberal naturalists to be
supervenient on other properties. The logical structure of that relation, often left somewhat
mysterious, can be explicated in great detail if the logically adverbial character of “good” is
taken into account.
The real problem about evaluation is thus not that there can be no true value judgments, but
that there are all too many true value judgments, and that we can decide objectively between
them only by taking the issue to be one-dimensional. A higher degree of cleanness may be
achieved by spending time and energy that could have been used for other purposes or by using
325

Hans Fink

chemicals that may be harmful in other respects. So whether the better cleaning is the best
cleaning all things considered is an open question. This does not take anything away from the
objectivity of any of the evaluations involved, but it shows practical decisions to be a matter of
balancing different concerns. This can be done objectively if a well specified process is agreed
upon, but otherwise not. For completely naturalistic reasons Moore was therefore right that it is
always an open question whether something that is called good for some reason is “really” good
because there is always other respects that may also be relevant. But if something is good or bad
in one respect it remains good or bad in that respect no matter how bad or good it is in other
respects. And moral goodness is not a special, overriding, form of goodness exempt from this
logic. The word “good” is used in contexts we call moral, but there is no special moral sense of
the word.

5 A naturalistic understanding of ethical normativity
Back in Aarhus in the 1970s and teaching courses in ethics and political philosophy, I was much
influenced by the intellectual climate at the time and the way in which Alasdair MacIntyre in
his A Short History of Ethics insisted on seeing work in moral philosophy in its social and historical context. A study of Hegel’s critique of Kant and work on the young Marx and the early
Habermas gave a needed expansion to my earlier rather narrow analytical work. I wrote a small
book for my elementary courses, later translated as Social Philosophy, in which I presented the
accounts of ethics and politics that philosophers had been giving in the past and were giving in
the present as deeply integrated in the historical development of the societies in which they
lived. In a way I saw this as being in continuity rather than discontinuity with ordinary language
philosophy. Wittgensteinian language games are played and Austinian speech acts performed in
specific societies under historical change. Ethical naturalism must be psychologically, sociologically and historically well-informed as Dewey would be the first to agree.
I also began teaching courses on The Ethical Demand the main work from 1956 by the
theologian and phenomenological philosopher K.E. Løgstrup, who had been my teacher, and
who is clearly the most original and influential moral philosopher in Denmark in the twentieth
century, and who is now beginning to be considered by moral philosophers in the Anglophone
world as well.10 He insisted that the central ethical content of Christianity, the demand for love
of neighbor, could and should be understood and argued for philosophically without appeal to
anything regarded as being either supernatural or a question of personal commitment. He could
thus be called a liberal naturalist in ethics (and perhaps even in theology). What I find most
important in the present context is the way his ethical naturalism refers to a normativity implicit
in the factual that is prior to and independent of all explicit normativity be it in the form of
divine, social or personal demands.
Human life is social life, but that means life lived together with quite particular others. My
life is what it is, because of what my relations to certain others are and have been. Moral
philosophers often make the mistake of defending concern for the other by offering selfish
reasons for unselfish behavior, as if we were basically isolated individuals with primarily selfish
concerns only. But this is a naïve abstraction overlooking the way our life is always already
completely entangled with the life of others. We are deeply interdependent, and we cannot
help having to rely on the good will of others while running the risk of being exposed to their
ill will. This is a completely reciprocal relation. But when another has shown her dependency
on me, it is up to me what to do, but it is not up to me whether what I do to the other is good
or bad for her. This is a factual matter, often complicated and involving the weighing of
different concerns in the short and the long run, but it may be entirely clear that what I have
326

Ethics and liberal naturalism

done is bad or evil for the other in that it neglects or exploits her dependency. Here there is an
asymmetry between good and bad. If nothing bad was done everything would be good, but we
would have no occasion for noticing this. The good is ontologically prior, but the bad is
epistemically prior. (Health and peace are similarly situated in relation to illness and war).
It is ethically demanded in the situation that you do what you sincerely and imaginatively
think is best for the other. This may, however, very well be in conflict with what the other
explicitly wishes from you or with other legal or moral demands that you are equally under in
the situation. You are then in a real conflict. Moral philosophy may help you describe this and
offer reminders of what is generally at stake in such conflicts without, however, relieving you of
the burden of having to act with no certainty that your action is the right or best all things
considered. Even if you are convinced that it is, this may still involve doing harm that is not
nullified by the good that you have also achieved. Every choice may leave you with remainders
that it is your responsibility to take care of and not just neglect or deny the relevance of.
Ethical naturalism may take many forms, but on my understanding inspired by Løgstrup (and
by Cavell, too) it should first of all be an ethics of responsibility. Utility, duty, virtue and rights
all come into the picture, but your responsibility for what you do to another is not to be fully
subsumed under any rule or principle other than that it shall be done for the best of the other
and for the other’s sake – and then not even for the sake of following this as a principle.
Morality is not to be seen as an ideal, unified set of reasons offering solutions to our dilemmas.
What matters in human life is manifold and it is a naturalistic claim that nothing that we should
do our best not to overlook anything that matters but to take it all into account for what it is, no
more, no less. We have not been promised that it should be easy.

Notes
1 Geach (1956), Foot (1958) and (1958–1959), MacIntyre (1958), von Wright (1963) and Searle (1964)
were pioneerworks in this movement. State of the art in the early 1960s are well represented in
Foot (1967).
2 Strawson (1985). De Caro and MacArthur (2004, 2010) give an excellent overview of the following
discussion.
3 In a way, this was just an update of W.K. Frankena’s conclusion in his discussion of Moore’s arguments
in Frankena (1939).
4 The former phrase is used in Dewey (1944, 15), the latter in Dewey (1981, 10).
5 Dewey (1984, 74).
6 Mill (1969, 374–375).
7 Dewey (1944, 12).
8 I have argued this point against McDowell in Fink (2006). In his reply, he insists that a vague notion of
nature is sufficient for his purposes.
9 I was here inspired by Dewey’s adverbial account of truth.
10 His work has so far been discussed in a monography: Stern (2019), and two collections of papers:
Andersen and Niekerk (2007) and Fink and Stern (2017). I have contributions in both. Alasdair
MacIntyre, Zygmunt Bauman, Simon Critchley and Stephen Darwall are among the moral philosophers who have also written about Løgstrup.

References
Andersen, S. and Niekerk, K.V.K. (eds.) Concern for the Other. Perspectives on the Ethics of K.E. Løgstrup.
Notre Dame University Press. (2007).
Ayer, A.J. Language, Truth and Logic. Victor Gollancz. ([1963]/1962).
De Caro, M. and Macarthur, D. (eds.) Nikaturalism in Question. Harvard University Press. (2004).
De Caro, M. and Macarthur, D. (eds.) Naturalism and Normativity. Columbia University Press. (2010).

327

Hans Fink
Dewey, J. Antinaturalism in Extremis. In Y.H. Krikorian (ed.), Naturalism and the Human Spirit. Columbia
University Press. (1944). Reprinted in The Later Works of John Dewey, Vol. 15. Southern Illinois
University Press (1989).
Dewey, J. Experience and Nature In The Later Works of John Dewey, Vol. 1. Southern Illinois University
Press. ([1925]/1981).
Dewey, J. Half-Hearted Naturalism. In The Later Works of John Dewey, Vol. 3. Southern Illinois University
Press. ([1925]/1984).
Fink, H. Social Philosophy. Methuen. (1981). (Later editions: Routledge).
Fink, H. Three Sorts of Naturalism. European Journal of Philosophy, 14(2), (2006). Reprinted with a reply
from McDowell in J. Lindgaard (ed.) John McDowell. Experience, Norm and Nature. Blackwell.
Fink, H. and Stern, R. (eds.) What is Ethically Demanded. K.E. Løgstrup’s Philosophy of Moral Life. Notre
Dame University Press. (2017).
Foot, P. Moral Arguments. Mind, 67(268), (1958). 502–513.
Foot, P. Moral Beliefs. Proceedings of the Aristotelian Society, 59(1), (1958–1959). 83–104. (Reprinted in
Foot 1967).
Foot, P. (ed.) Theories of Ethics. Oxford University Press. (1967).
Frankena, W.K. The Naturalistic Fallacy. Mind, 48, (1939). (Reprinted in Foot 1967).
Geach, P. Good and Evil. Analysis, 17, (1956). (Reprinted in Foot 1967).
Hare, R.M. The Language of Morals. Oxford University Press. (1952).
Krikorian, Y. H. (ed.), Natuaralism and the Human Spirit. Columbia University Press. (1944).
Løgstrup, K.E. The Ethical Demand. Intro. Fink H. and MacIntyre A. Notre Dame University Press.
(1997). (A new translation with an introduction by B. Rabjerg and R. Stern has been published by
Oxford University Press, 2020).
MacIntyre, A. Hume on ‘Is’ and ‘Ought’. Philosophical Review, 68, (1958).
MacIntyre, A. A Short History of Ethics. MacMillan. (1966).
Mill, J.S. The Collected Works of John Stuart Mill, Vol. X. University of Toronto Press/Routledge and
Kegan Paul. (1969).
Moore, G.E. Principia Ethica. Cambridge University Press. (1903).
Searle, J.R. How to Derive ‘Ought’ from ‘Is’. Philosophical Review, 73, (1964). (Reprinted in Foot 1967).
Stern, R. The Radical Demand in Løgstrup’s Ethics. Oxford University Press. (2019).
Strawson, P.F. Scepticism and Naturalism: Some Varieties. Methuen. (1985).
von Wright, G.H. The Varieties of Goodness. Routledge and Kegan Paul. (1963).
Ziff, P. Semantic Analysis. Cornell University Press. (1960).

328

27
KANTIAN CONSTITUTIVISM
AND THE NATURALISTIC
CHALLENGE
Carla Bagnoli

1 Introduction
One of the most fundamental problems in moral philosophy is to account for the objectivity of
ethical knowledge. Broadly speaking, ethical knowledge is the sort of knowledge implicated in
doing the right thing. It is the sort of knowledge that enables agents to tell right from wrong and
which guides (rational) agents in acting as they should. The challenge behind the attempt to
offer a satisfactory account of ethical knowledge has been accounted in various manners. One
way to frame the challenge is to say that ethical judgments appear to aspire to objectivity and
practicality, but these aspirations push in opposite directions, and it is an open question whether
and how to reconcile them. For many, the main challenge in defending ethical objectivity is
that it seems to commit one to a queer ontology, insofar as it entails knowledge of properties of
the world that are “intrinsically motivating” (Mackie 1977; Smith 1994).
Underlying these debates about the ontological commitments of ethical objectivity, there are
deeper divisions concerning not only the very standard of ethical objectivity in ethics but also
the standards of practical significance. The first issue divides those adopting an ontological
approach to objectivity and those proposing a practical approach to it. The second issue divides
those concerned with the practical powers of reason, and those investigating the capacity to
motivate action through rational argumentation, rhetorical persuasion, or psychological inducement. The prevailing view, most often combined with the rejection of the ontological
approach, is that ethical judgments guide action through conative states and emotions, rather
than via reasoning and rational argumentation. Thus, the practical guidance of ethical judgment
is supposed to have little to do with the normative authority of reason. This is a move that has
been recurrently associated with naturalism. Broadly speaking, ethical naturalism holds that the
semantic features of ethical judgments require no special ontology. Moral facts – if there are any
– are natural facts. Moral properties are nothing but natural properties. In this broad sense, any
meta-ethical account of ethics that is consistent with naturalistic position in general philosophical inquiry counts as naturalistic. But this general claim can be phrased in more or less
stringent terms (Lutz and Lenman 2018). From a metaphysical point of view, naturalism admits
to existence only natural things; and, from an epistemological point of view, it holds that we
know such things via empirical methods. When this view is combined with realism, it holds
that there are objective, mind-independent moral facts, and we know about them using
DOI: 10.4324/9781351209472-32

329

Carla Bagnoli

empirical methods. This is the view I call “reductive naturalism”, which stands in contrast to
antirealist theories of various sorts (e.g. error theory, Humean constructivism, relativism, expressivism), and Kantian constructivism.
A seeming advantage of reductive naturalism is that it solves all these debates with a simple
move, which admits of a naturalistic ontology to which all special ontologies belong and
eliminates the problem of normativity as a pseudo-problem. According to its critics, the disadvantage of this move is that it does not account for the subjective experience of moral life,
understood as the life of mutually dependent and vulnerable subjects, governed by and accountable to principles of rationality, and participating in a community regulated by norms.
Practical subjects represent themselves and others as governed by (ethical) norms, which can be
defended, challenged, observed and violated. To be guided by norms requires an understanding
of the authority of such norms, which eludes reductive naturalism. Reductive naturalism seems
right in pressing a unified naturalistic ontology, but its reductivist methodology bears epistemic
costs that seem intolerable exactly because it imposes a loss of understanding of the subjective
experience of ordinary normative practices.
Liberal naturalism offers a very plausible way to avoid this problem in that it acknowledges
that riddles and frictions underlying the dynamics of moral life are genuine phenomena which
call for philosophical explanation, and yet admit no redundant or extravagant ontology (De
Caro and Macarthur 2010, 9). Philosophical explanation aims to elucidate and further our
understanding of paradigmatic ethical phenomena: it is pluralistic in its methodology and does
not stand in need of any special ontology.
Within these methodological coordinates, Kantian constructivism represents a promising
approach to ethical objectivity, which shares some fundamental tenets of liberal naturalism
(Macarthur 2019). While Kantian constructivism agrees with moral realism that there are
objective distinctions, such distinctions do not rest on any separate ontology. Kantian constructivism adopts the naturalistic constraint on ontology, and is parsimonious in admitting
moral facts and moral properties. There are no moral facts prior to and independently of the
very activity of practical reasoning (Rawls 1980; Korsgaard 1996, 38–40). While admitting a
naturalistic ontology, Kantian constructivism claims that the focus of the discussion of ethical
objectivity is practical rather than ontological, and proposes that we understand the practical
significance of ethical judgments in terms of rational authority. In this chapter, I shall discuss the
constructivist account of rational authority, and address some worries raised by reductivists.

2 Wronging and punishing
How and whether ethical judgments are authoritative is a contested matter in meta-ethics.
Relatedly, it is a matter of contention in ethics what the source of moral authority is. The issue is
relevant for determining the source and power of moral obligations, but also the legitimacy of
punishment. Indeed, a central problem for an account of ethical objectivity concerns the responses
appropriate to wrongdoing and the legitimacy of punishment. This is a problem that sets ethics
apart from other informal domains of inquiry about rational choice and rational thought.
Wronging someone is not just a matter of judging that a (wrong) action is a course of action open
to us; nor is it just choosing a wrong course of action over others. Wronging someone is wronging
someone. It is a kind of conduct that is subject to blame and other forms of punishment. The
legitimacy of blame and punishment depends on firm distinctions between right and wrong, and it
is a matter of dispute whether such distinctions are objective and how to draw them.
Grounding ethical objectivity on wrong-making and right-making properties has proven to
be an unsuccessful strategy. One problem is that such a strategy seems to invoke a special sort of
330

Kantian constitutivism

moral ontology, and violates the requirement of ontological parsimony (Mackie 1977).
However, ethical objectivity could be defended in ways that do not violate such requirement
because they do not assume anything special about moral ontology. In this vein, ethical naturalism holds that right and wrong are concepts that do not represent special properties or
clusters of properties. The function of ethical judgments of right and wrong is not to describe a
special sector of reality but to guide conduct. But how to explain the action-guiding features of
ethical judgments? How are agents endowed with the concept of right and wrong driven to act
morally? Can they act morally when moral action is not in their own interest, or does not
conform their own particular convictions?
Such questions relate the issue of ethical objectivity to the issue of normativity, that is, the
issue explaining how the concepts and judgments of wrong can guide rational agents in action.
Ethical naturalism holds that such questions admit of empirical answers which require nothing
exoteric. There is nothing special about the ontology of right and wrong, nor is there anything
mysterious or spooky about the bindingness of right and wrong. In fact, one way to explain
away the queerness associated with the claims of ethical objectivity is to focus on punishment
first. As John Stuart Mill (1861) writes in Chapter 2 of Utilitarianism, when considering whence
the sense of obligation comes from we find discussing emotional responses to wrongdoing, such
as guilt or remorse. Such emotional reactions are appropriate in the aftermath of wrongdoing,
and count as “internal sanctions”, that is, sanctions that have been internalised and are part of
the moral psychology associated with wrongdoing. They explain how and why the agent
endowed with a normal moral cast of mind paradigmatically feels bound by the moral obligation and refrains from wrongdoing. This is a promising line of investigation because it
purports to account for the objectivity and authority of moral distinctions regarding right and
wrong without committing to any redundant or suspicious ontology. Furthermore, this account
coheres with a naturalistic understanding of ethics as a cooperative enterprise marked by a sense
of human fellowship, and it does not assume any foundational value prior to and independently
of such cooperative interactions. It does lend itself nicely to evolutionary interpretations of the
ethics of right and wrong (Gibbard 1990). To this extent, the naturalistic defense of ethical
objectivity resists the sort of debunking arguments that seems fatal to nonnaturalistic absolutism.

3 Ethical naturalism and the paradox of punishment
In the Sources of Normativity, Christine M. Korsgaard argues that the naturalistic proposal of
accounting for the distinctions between right and wrong is not satisfactory. Her argument is that
the naturalistic strategy of explaining the normative force of ethical judgments of right and
wrong in terms of punishment – for example, in terms the emotional reactions that are appropriate to it – fails. It fails in a particular way: it is a self-defeating account, which dissolves the
very phenomenon that it sets out to investigate. Interestingly, in her view, this is the same
problem that affects other forms of objectivism, which takes punishment to explain the authority of moral obligations.1 It is reductive insofar as it explains the alleged authority of moral
obligations in terms of the manner of its social and psychological enforcement.
According to Korsgaard, Thomas Hobbes’ claim that God’s authority rests on his “irrestistible power” is a paradigmatic example of this reductivist strategy (Hobbes, Leviathan II.31,
246). On Hobbes’ account, the agent is obliged to fulfill his duty forced by the prospect of
punishment. This construal reveals that the source of obligation is rather mundane: it is a
psychological mechanism that is both external to the subject (i.e. the agent bound by the
obligation) and to the sovereign (i.e. the alleged source of moral obligation). This way, it is
shown that the sovereign’s authority is merely executive: “The sovereign’s authority now
331

Carla Bagnoli

consists entirely in his ability to punish us. Although sanctions are not our motive for obedience, they are the source of the sovereign’s authority and so of our obligations” (Korsgaard
1996, 29). This treatment of the authority of obligation leads to the paradoxical result that
wrongs are always punished, or they are not cases of wrongdoing. If one commits a moral crime
and gets away with it, i.e. escaping the vigilance of the sovereign and his ability to punish, the
Hobbesian accounts cannot explain on which grounds what the agent did was wrong and how
his wrongdoing counts as a moral transgression. Which norm did he violate? Whose authority
has he trespassed? Moral transgression is possible because the sovereign enforces moral obligations and has unlimited ability to punish it. But if “irresistible power” just means “power
unsuccessfully resisted”, then moral authority amounts to the successful exercise of power, and
“things always turn out right. For no one can ever do what he lacks the power to do”
(Korsgaard 1996, 29). Korsgaard’s argument aims to establish that there is a distinction between
power and authority. To confound the two is to misunderstand the whole issue about the
normativity of moral obligations. The naturalistic strategy, which reduces the authority of moral
obligation to its enforcement, is self-defeating.
The point against Hobbes’ ethical naturalism can be generalised. Korsgaard extracts a form of
the naturalistic reduction of normativity and formulates a methodological dilemma that binds all
naturalistic ethical theories:
If we try to derive the authority of morality from some natural source or power it will
evaporate in our hands. If we try to derive it from some supposedly normative
consideration such as gratitude or contract, we must in turn explain why that consideration is normative or where its authority comes from (Korsgaard 1996, 30).
Korsgaard’s critical move is avowedly analogous to G.E. Moore’s open question argument
against ethical naturalism (Korsgaard 1996, 43, 124, 215, 218). For Moore, the naturalistic
impasse should be avoided by arguing that normativity cannot be explained because it is a
fundamental notion that stands for an irreducible nonnatural property. For Korsgaard, however,
the nonnaturalist strategy fails for similar reasons as naturalism fails, that is, because they do not
recognise reason as the source of normativity and of objectivity. These opposing views share a
realist framework, which displaces the normativity of ethical judgments. By contrast, Kantian
constructivism proposes that ethical objectivity be interpreted in terms of rational justification,
which counts as the sole genuine source of authority.
The promise of the constructivist view can be illustrated by focusing on the case of radical
disagreement. In such a case, each of the conflicting parties may (legitimately) claim that they
are absolutely right, and the opposing parties absolutely wrong. Rational argumentation is
bound to fail in this sort of disagreement if the opposing parties take the disagreement to be
about fundamental ethical properties (natural for ethical naturalism, nonnatural for intuitionism). They each expect others to converge onto an (ethical) feature of the world that the
other cannot possibly recognise as such (Korsgaard 1996, 68ff; Engstrom 2013, 137). The loss
implicated in this strategy is not only epistemological. By taking normativity as a primitive
aspect of ethics which resists philosophical explanation, moral realism gives up ethical argumentation as a way of inducing moral change and progress. One is left with the view that
ethical truths are out there, whether or not people recognise them. The naturalist’s selfdefeating account for normativity is an especially significant loss. And the nonnaturalist approach, which substitutes authority with un-argued confidence in the objective moral domain,
is no less disappointing: “the normative question arises when our confidence has been shaken,
whether by philosophy or by the exigencies of life” (Korsgaard 1996, 40).
332

Kantian constitutivism

4 The claim of universal authority
Kantian constructivists pursue the project of vindicating the objectivity of ethical judgments in
terms of rational justification and their authority in terms of the authority of reason. In contrast to
“dogmatic” forms of rationalism, they argue that the function of reason is constructive or generative. Korsgaard phrases the contrast between standard realism and constructivism in terms of
rational procedures (1996, 38–40). This phrasing captures an important aspect of constructivism as
an account of ethical objectivity. While Kantian constructivism agrees with moral realism that
there are correct ways of drawing the distinctions between right and wrong, it disagrees that such
distinctions depend on the ontological domain of ethics. There is no moral domain prior to and
independently of practical reasoning. Thus, the objectivity of ethical distinctions does not rest on
moral ontology, but on the proper activity of reason, which is principled (Rawls 1980, Lecture III;
O’Neill 2002; Bagnoli, 2015). The realist thinks that reasoning delivers ethical truths because
it tracks them down, and thus assumes that there is some distinctive moral domain prior to and
independent of our thinking of it. Cognitive activity amounts to “letting one’s thinking be
controlled by how things are” (Mackie 1977, 22). By contrast, Kantian constructivists focus on
practical reasoning as a shared enterprise which produces ethical agreement. This is not to say
that all kinds of ethical disagreements are eliminable or should be eliminable by reasoning. On
the contrary, constructivists may recognise significant forms of ethical disagreements about
morally decent forms of life and morally permissible courses of action. Yet, engaging in practical
reasoning enables practical subjects to build practical reasons that others could share, and thus
provides a solid basis for cooperative interactions.2 Practical subjects are subjects capable of
acting on reasons, or more precisely on principled reasons, that is, reasons that we could share.
The core of constructivism is that practical reasons are not fixed by an order of moral values
“apart and distinct from how we conceive of ourselves”. Thus, solving ethical disagreements
does not require us to discover the truth of the matter but construct reasons we could share
starting with the reasonable grounds for agreement “rooted in our conception of ourselves”.
This is the meaning of John Rawls claim that “moral objectivity is to be understood in terms of
a suitably constructed social point of view that all can accept” (Rawls 1980, 519).
One significant advantage of the requirement that practical reasons are principled is intelligibility (O’Neill, 1996, 57). Universal reasons make acting at least in principle intelligible to
others. Others can challenge, support or obstruct what we do insofar as they understand what
we are doing as something we do for a reason. But the main advantage of producing universal
reasons is practical. In contrast to foundationalist (naturalist or anti-naturalist) approaches,
Kantian constructivism does not search for proofs. Rather, it is an informal account of practical
reasoning that aims at an audience, which is universal rather than parochial. To address a
universal audience, reasoning must be structured by the requirement of universality. Reasoning
that starts from particular substantive standpoints is conditional on such specific standpoints, and
thus claim only restricted authority. Such reasoning is “private” in that it can serve only limited
communities that share a common conception of the world, and thus it does not help constructing the sort of objectivity required for ethics. Ethical objectivity demands universal authority. And only universalist reasoning can respond to this demand. When the authority of
reason is not premised on special membership in a concrete community it is unconditional,
binding for all possible agents endowed with rationality. This claim makes constructivism better
suited than communitarian reasoning to account for contexts marked by value pluralism. In
such contexts, the appeal to specific interests, identities, and plans can reach out only to a
limited audience, and thus creates incommunicable differences or reinforces divisive confrontations (O’Neill 1989; Rawls 1993).
333

Carla Bagnoli

The defense of universality as the form of reasoning is a defining feature of Kantian constructivism, and hence not a dispensable feature of it.3 But it is debatable whether it is consistent
with the anti-realist and naturalist aspirations of constructivism. The first worry is that Kantian
constructivism betrays its antirealist commitments insofar as and because it takes universality as
the constitutive norm of reason. Along these lines, Sharon Street complains that “Korsgaard’s
argument does not ultimately remain true enough to the central constructivist point that there
are no facts about normative reasons apart from the standpoint of an agent who is already taking
things to be reasons”, (Street 2012, 48). This is because of the pivotal role that Korsgaard
ascribes to the moral value of humanity in structuring practical reflection.4 A second worry is
that to conceive of practical reason as constructive, that is, generative and productive, is problematic. In this vein, Nadeem Hussain and Nishitan Shah argue that Korsgaard’s constructivism ends up with the same mysterious account of normative truths as anti-naturalism
moral realism (Hussain and Shah 2006, 292). They take the constructions of reason to be “acts
of ex-nihilo creation”, which make constructivism vulnerable to the same objection of ontological queerness as the standard forms of moral realism.5 They conclude that Korsgaard fails
to establish Kantian constructivism as a self-standing meta-ethical alternative to moral realism
(Hussain and Shah 2006, Hussain and Shah 2013). Their argument is based on several claims,
but ultimately, their point is that she fails to distinguish questions about practical reason from
meta-ethical questions.6 This point is not theoretically innocent, since an integral part of the
constructivist argument is that meta-ethical questions such as ethical objectivity cannot be
adequately dwelt but calling into play the practical and constructive function of reason (O’Neill
1989; Bagnoli 2015; Engstrom 2013). In fact, the Kantian core of the constructivist position is
that the objectivity of moral obligation can be vindicated only by vindicating the constructive
or generative function of reason (Rawls 1980, 2000; O’Neill 1989).
The reply to these reductivist objections revolves around the explication of the distinctive
role of universal principles in practical reasoning as well as in the constitution of rational agency.
On the constructivist view, the requirement of universality is formal in the sense that it is
structural of reasoning as such, but also in the sense that it is constitutive of rational agency. The
claim that universality is a formal requirement puts constructivism at odds with relativist
conceptions of ethics and practical reasoning, and with standard realism. Substantive moral
principles are not the object of rational intuition. It pertains to the original self-conscious
activity of practical judgment, through which one constitutes oneself as a rational agent capable
of choice. This self-representation is the condition of possibility of rational choice, but it is not a
theoretical cognition (Korsgaard 1996, 14). Rather, it is a matter of awareness, totally internal to
the subjective dimension. That is, it pertains to the agent’s own self-conception of oneself as a
practical subject operating together with other practical subjects. This sort of awareness is
present in any act of practical judgment. Does this claim infringes naturalism in that it presumes
a too strong sense of freedom?
In fact, constructivists such as Korsgaard have argued that the ethics of autonomy is the only
ethics appropriate to the world conceived under the naturalistic constraint (Korsgaard 1996).
This is not to say that this view of moral action is merely compatible with naturalism.
Reductive naturalism leaves out an important aspect of moral and rational agency, which
Kantian constructivism sets out to explain. And yet the explanation that Kantian constructivism
affords is bound by the naturalistic constraint discussed at the outset, in that it does not admit of
any specious or redundant ontology.
Agents are “practical subjects” insofar as they capable of governing their own lives and
answerable to the demands of practical reason. They understand and accept their finitude and,
consequently, their exposure to contingency as a constitutive condition of acting. But they are
334

Kantian constitutivism

also subjectively aware of their susceptibility to practical reason. As rational agents, they act and
think under the representation of themselves and others as free. Is this compatible with massive
and systematic error about the world and ethical distinctions, and also about taking one another
responsible judges of what they think and do? It is logically possible that we are massively and
systematically deluded about the objectivity of ethical distinctions and reciprocal attributions of
epistemic and ethical responsibility. But to entertain and sustain this logical possibility is selfdefeating, as it undermines every belief, action and attitude (Bagnoli 2015).

5 The constitutivist claim
Critics might insist that there still is a sense of circularity in the constructivist argument for the
rich practical understanding of objectivity, which is disguised when the theory is formulated in
terms of correct decision procedures (Timmons 2003; Scanlon 2014). The hidden point is that
the procedure is constitutive of rationality, and thus it is not up for individual choice. To be a
rational agent requires much more than avail oneself to and managing a procedure of reasoning.
Thus, the hidden point is not a substantive claim that specific principles and skills are built into
rational justification. Rather, the hidden point is that rational justification necessarily takes a
distinctive form, namely, universality. The suspicion is that there cannot be any vindication of
reason by means of reason alone. In other words, reflexivity of reason does not bring the sort of
unconditional authority which is the intended target at which Kantians aim. To say that the
point about the form of rational justification is “hidden” is to suggest that Kantian constructivism relies tacitly on realist foundations. But this is an uncharitable reading that mischaracterises the constructivist project, which openly defends the claim that reasoning itself is
structured by the requirement of universality.
Constructivists concede that the general Kantian argument for the vindication of reason by
reason alone sounds circular (O’Neill, 2004). Indeed, it is avowedly circular, in the innocuous
sense that it does not rest on external foundation. If it did, its authority would be derivative and
parasitic on the authority of its fundaments. The Kantian claim is that the normative authority
must be underivative to be genuine. Thus, the authority of reason is authentic because it does
not derive from elsewhere, that is, from fixed starting points external to the domain of reason
itself. Reason is reflexive and thus applies to itself its scrutiny. There is no vicious circularity
affecting this claim. The point is simply that reasoning is the only activity that can authorise and
justify one’s claims and attitudes. To do so, reasoning must be governed by an internal norm,
not imported from outside its domain. This is why universality must be the norm constitutive
of reasoning (O’Neill, 1989, 108; Korsgaard, 2009). Such norm does not have any specific
content; it is merely the principle that there must be a principle.7 The argument is support of
universality is not ontological or theoretical but practical, and tied to the claim that ethical
objectivity demands universal authority. The appeal to universality as a norm constitutive of
reasoning is not sufficient to determine what to so. Thus, it would be a mistake to conceive of it
as a complete decision procedure.

6 Conclusion
Kantian constructivism enters the debate about ethical objectivity with two related proposals.
The first is that ethical objectivity comprises a claim to universal authority. The second is that
only reason provides universal authority, through rational justification. The practical significance of ethical judgments is an integral part of their objectivity. There is no tension between the practicality and the objectivity. The tension depends on misconceptions of the
335

Carla Bagnoli

standards of objectivity and of practicality. All opposing theories in metaethical debates conceive of objectivity as an ontological claim about what there is independently of our manner of
knowing, understanding and representing. Standard forms of realism and antirealism at the
opposite spectrum take reason to be a theoretical capacity, that is, the capacity to acquire
knowledge of external objects. They also conceive of practicality as the capacity to motivate
agents to action, quite independently of their capacity for reasoning. They deny that reason has
any power besides representing and tracking objects and properties out there. Its theoretical
function is to achieve a true representation of reality.
By contrast, constructivism holds that the practical significance of ethical judgments depends
on their capacity to guide rational agents in action. The norm the attains at the formation of
ethical judgments is also constitutive of rational agency. The significance of reason is both
ontological and practical. From an ontological point of view, reason is generative and productive. Reasons are not out there prior to and independently of reasoning; they are the
products of the activity of reason. Such activity is correct not because and insofar as it matches
an outside reality, but because and insofar as it is governed by the appropriate norms. The norm
is appropriate because and insofar it expresses persons’ basic self-conception (Rawls 1980,
559–560, cf. 518–519, 571). It is the norm that governs self-reflection and that guides rational
agents in action (Korsgaard 1996, Chap. 4). It is because such a norm expresses our persons’
basic self-conception that can determine them to act accordingly. This is how constructivism
reconciles the objectivity and the practicality of ethical judgments or, more precisely, how it
explains away their apparent tension. If ethical judgments have cognitive import it is because
they convey knowledge of oneself as a practical subject, capable of acting on a reason, and of
interacting with others on the basis of reasons they could share.

Notes
1
2
3

4
5

6
7

“The problem here is a general one, which applies to any attempt to derive normativity from a natural
source of power”, Korsgaard (1996, 29, 30, 145–46, 160–161).
I use Korsgaard’s phrase, but constructivists have offered different ways of cashing out the claim that
reasons are universal, see, for example, Scanlon (1998) and O’Neill (1989).
Humean constructivists argue that constructivism can do without the constitutivist norm of universality, and indeed it should in order to keep faith to its anti-realist commitment, see Street (2008,
2012). As I have argued elsewhere, to release the requirement of universality as a constitutive norm
amounts to giving up the constructivist project of a distinctive meta-ethics, different than expressivism
and realism, see Bagnoli (2019).
A related concern is that Korsgaard fails to ground moral obligation on rationality alone. The deduction of moral obligations from reason either fails or ultimately rests on substantive grounds, see, for
example, Cohen (1996) and Gibbard (1999).
It is not clear how Kantian constructivism is exposed to the objections of epistemological implausibility
and arbitrariness that undermine standard moral realism, according the constructivist. It is also unclear
why constructed moral facts would be more problematic from an ontological point of view than
projections in error theory or quasi-realism. It seems that for the objection to work it has to target also
non-standard forms of antirealism. But the result is that the only two alternatives in meta-ethics are
standard moral realism (reductive naturalism or intuitionism) and standard antirealism (emotivism and
expressivism), see, for example, Darwall et al. (1992). Compare Schafer (2014).
They object that Kantian constructivism builds upon a critique of nonreductive moral realism which
confounds two different questions about normativity (Hussain and Shah 2006, 269).
This is the formal warrant of the autonomy of reason, which requires independence from “contingent,
subjective conditions that distinguish one rational being from another” (Kant Ak 5, 21), and also “lawgiving of its own on the part of pure and, as such, practical reason [which] is freedom in the positive
sense”, (Kant Ak 5, 33).

336

Kantian constitutivism

References
Bagnoli, C. Authority as a Contingency Plan. Philosophical Explorations, 22(2), (2019). 130–145.
Bagnoli, C. (ed.), Constructivism in Ethics. Cambridge University Press. (2013).
Bagnoli, C. Moral Objectivity: A Kantian Illusion? The Journal of Value Inquiry, 49, (2015). 31–45.
Copp, D. Is Constructivism and Alternative to Moral Realism? In Bagnoli, (2013). 108–132.
Cohen, G.A. Reason, humanity, and the moral law. In G. A. Cohen , R. Guess , T. Nagel , and B.
Williams (eds.), The Sources of Normativity, pp. 167–188. Cambridge University Press. (1996).
Darwall, S., Gibbard, A., and Railton, P. Toward Fin de Siecle Ethics: Some Trends. The Philosophical
Review, 101, (1992). 115–189.
De Caro, M. and Macarthur D. (eds.) Naturalism and Normativity. Columbia University Press. (2010).
Enoch, D. Taking Morality Seriously. Oxford University Press. (2011).
Engstrom, S. Constructivism and Practical Knowledge. In Bagnoli (ed.), Constructivism in Ethics.
pp. 133–152. Cambridge University Press. (2013).
Fitzpatrick, W. How Not to be an Ethical Constructivist: A Critique of Korsgaard’s Neo-Kantian
Constitutivism. In Bagnoli, 2013, (2013). 41–62.
Gibbard, A. Morality as consistency in living. Ethics, 100, (1999). 140–164.
Gibbard, A. Wise Choices, Apt Feelings: A Theory of Normative Judgement. Harvard University Press. (1990).
Hussain, N. and Shah, N. Misunderstanding Metaethics: Korsgaard’s Rejection of Realism. Oxford Studies
in Metaethics, 1, (2006). 265–294.
Hussain, N. and Shah, N. Meta-Ethics and its Discontents: A Case Study of Korsgaard. In Bagnoli, 2013,
(2013). 82–107.
Kant, I. Kants gesammelte Schriften. Preussische Akademie der Wissenschaften. (1907). (abbreviated Ak).
Korsgaard, Christine The Sources of Normativity. Foreword O’Neill. Cambridge University Press. (1996).
Korsgaard, C.M.. Self‐Constitution. Action, Identity and Integrity. Oxford University Press. (2009).
Korsgaard, C. M. Valuing Our Humanity. In R. Dean and O. Sensen (eds.), Respect, pp. 171–191.
Oxford University Press. (2021).
Lenman, J. and Y. Shemmer (eds.) Constructivism in Practical Philosophy. Oxford University Press. (2012).
Lenman, J. Expressivism and Constructivism. In Lenman and Shemmer, 2012, (2012). 213–225.
Lutz, M. and Lenman, J. Moral Naturalism. In E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy, Fall
2018 edition. (2018). https://plato.stanford.edu/archives/fall2018/entries/naturalism-moral/.
Mackie, J. L. Ethics: Inventing Right and Wrong. Penguin. (1977).
Macarthur, D. Liberal Naturalism and the Scientific Image of the World. Inquiry, 62(5), (2019). 565–585.
O’Neill, O. Vindicating Reason. In P. Guyer (ed.), The Cambridge Companion to Kant, pp. 280–308.
Cambridge University Press. (1992).
O’Neill, O. Constructions of Reason: Exploration of Kant’s Practical Philosophy. Cambridge University
Press. (1989).
O’Neill, O. Toward Justice and Virtue: A Constructivist Account of Practical Reasoning. Cambridge University
Press. (1996).
O’Neill, O. Autonomy and the Fact of Reason. In O. Hoffe (ed.), Kritik der praktischeVernunft
(7–8, 30–41). pp. 81–97. Akademie Verlag. (2002).
O’Neill, O. Self‐Legislation, Autonomy, and the Form of Law. In H. Nagl‐Docekal and R. Langthaler
(eds.), Recht, Geschichte, Religion: die Bedeutung Kants fur die Gegenwart, pp. 13–26. Akademie
Verlag. (2004).
Rawls, J. Kantian Constructivism in Moral Theory. Journal of Philosophy, 77, (1980). 515–572.
Rawls, J. Political Liberalism. Columbia University Press. (1993).
Rawls, J. Lectures on the History of Moral Philosophy. Harvard University Press. (2000).
Scanlon, T. M. What We Owe To Each Other. Belknap. (1998).
Scanlon, T. M. Being Realistic About Reasons. Oxford University Press. (2014).
Schafer, K. Constructivism and Three Forms of Perspective‐Dependence in Metaethics. Philosophy and
Phenomenological Research, 89(1), (2014). 68–101.
Shafer-Landau, R. Moral Realism: A Defense. Oxford University Press. (2003).
Smith, M. The Moral Problem. Oxford University Press. (1994).
Street, S. What Is Constructivism in Ethics and Metaethics. Philosophy Compass, 5, (2010). 363–384.
Street, S. Coming to Terms with Contingency: Humean Constructivism about Practical Reason. In
Lenman and Shemmer, 2012, (2012). 40–59.
Timmons, M. The Limits of Moral Constructivism. Ratio, 16(4), (2003). 391–423.

337

Carla Bagnoli
Wallace, R.J. Constructivism About Normativity: Some Pitfalls. In Lenman and Shemmer, 2012,
(2012). 18–39.
Wiggins, D. A Neglected Position? In J. Haldane and C. Wright (eds.) Reality, Representation, and
Projection, pp. 329–336. Oxford University Press. (1993).

338

28
THE RATIONAL WOLF.
MORAL PHILOSOPHY AS KEY TO
JOHN MCDOWELL’S LIBERAL
NATURALISM
Sofia Miguens

1 Our conceptual capacities and their place in nature
McDowell’s philosophy is an investigation of our conceptual capacities and their place in
nature.1 In moral philosophy and in general, his liberal naturalist view is put forward against
universalist and Platonistic views on the one hand and scientistic naturalism on the other. Here,
I will try to work out the idea that McDowell’s moral philosophy provides a useful key to
understand the shape of his liberal naturalism. But first I would like to point out something
about the idea of “being natural” and philosophical naturalism. Naturalism is currently a very
popular position in many domains of philosophy, from epistemology to ethics to philosophy of
mathematics. Being a naturalist might simply mean that one regard humans, as rational and
moral beings, as part of nature, and not otherworldly. This seems uncontroversial enough. Yet
if we ask ourselves who, among contemporary philosophers, is a self-professed naturalist, or gets
classified by others as such, the resulting list might strike us as surprisingly heterogenous, since it
ranges from John Dewey to Ludwig Wittgenstein, from David Papineau to David Chalmers to
John McDowell. One doubts that there is any common denominator to all these naturalists.
Actually one could think of a test, thinking, for example, of philosophy of mind. Let us say that,
looking at the field of philosophy of mind, we compare the stand of naturalism and of physicalism. These are sometimes (tendentiously) identified yet whereas there are well-known
arguments in favor of physicalism (e.g. an Argument from Causal Closure), a similarly clear case
of an argument for naturalism is simply absent. What we find are rather arguments from naturalism, more often than not leading to a certain sort of project of reduction (of mind, meaning,
morality or modality). All we are left with are cases of being a naturalist and how the argumentation goes in each case. I want to look at McDowell’s moral philosophy as one such case.2

2 Animality and rationality: the rational wolf thought experiment
McDowell’s moral philosophy often takes the form of a discussion of the classics, in particular of
Aristotle, Hume and Kant. The story is long and full of controversial exegesis. Since I need
something quite simple which might be illuminating for his moral philosophy as a whole, I will
DOI: 10.4324/9781351209472-33

339

Sofia Miguens

consider the fable (or thought experiment) of the rational wolf from the article Two Sorts of
Naturalism.3 The thought experiment is about what being rational is for a natural being. Talking
about being natural here is not a prolegomenon to talk of reduction but rather to considering
(our) rationality against the background of (our) animality. In McDowell’s terms, the task is to
conceive of the authority of reason in the face of the authority of our first nature (“first nature”
is McDowell’s term for the picture of humans provided by natural science). The very way of
setting the problem (in particular the focus on reason, or rationality) makes clear that McDowell’s
naturalism, often called second nature naturalism,4 by itself involves the rejection of what
Macarthur and De Caro call the ontological and methodological doctrines of scientific naturalism.5 In fact, it is such rejection that opens space for what we might call philosophical
anthropology, which includes the challenge of conceiving of our conceptual capacities (moral
capacities included) as part of the our nature.
The rational wolf thought experiment illuminates such space. It asks us to suppose that an
animal other than a human becomes rational. In McDowell’s terms, it becomes capable of
giving expression to conceptual capacities and of asking for reasons of its own behavior. He
would then ask himself: should I do as all wolves do? Should I hunt with the pack? Should I
cooperate? Need I do it? Why should I do it? Notice that being rational for the rational wolf is
not conceiving its own behavior as just another phenomenon in the world, which it then
conceptualises. Being rational involves being able to step rationally out of oneself and ask: why
should I do as other wolves (or humans) do? It amounts to, as McDowell puts it, letting (one’s)
mind roam over possibilities of behavior other that what comes naturally to wolves (or humans). This,
in fact, is one central point of Two Sorts of Naturalism which comes out as part of the criticism of
Philippa Foot’s version of (Aristotelian) naturalism. McDowell believes Foot’s is a less satisfying
sort of naturalism – not because it is Aristotelian (there is much of that in his own position) but
because it does not pay too much attention to the concept of nature in its own right, something
which McDowell sees as crucial for naturalism. Not paying sufficient attention to the concept
of nature brings the risk of appealing, as Foot does, in a less than reflective way, to (supposedly)
natural facts as underlying, for example, what it is for a human life to flourish. McDowell’s own
attention to the concept of nature in relation to that of reason is at work in his analyses of our
responsiveness to reasons and his view of value.

3 McDowell’s case for liberal naturalism (I): responsiveness
to (moral) reasons
What then does the authority of reason amount to when it comes to responsiveness to reasons
in rational animals such as ourselves? This is one question McDowell deals with in “Are Moral
Requirements Hypothetical Imperatives?”6 At the background lies Philippa Foot’s critique of
Kant. In face of Kant’s views on the inescapable character of moral reasons, Foot asks: What
does it mean to say that we ought to do x? (e.g. what does it mean to say that I ought to pay my
100.000 euro debt to my friend Susana?) We inherited from Kant, and very often accept, the
idea that there is a distance between hypothetical and categorical imperatives: moral imperatives
are categorical. That I ought to pay my debt to Susana is categorical. But what is really at stake
when we say that moral imperatives are categorical, or that we have moral reasons to do X?
According to the orthodoxy, hypothetical imperatives are conditional in that they state that I
should do X if I want Y. In contrast, categorical imperatives are unconditional in the sense that
the recommended action imposes itself as an end in itself. But this is precisely what Foot rejects:
if that were the only difference then we should take social rules (e.g. rules of etiquette) to be
categorical. After all their use is also clearly non-hypothetical – it is not dependent on any
340

The rational wolf

further ends. (Think of “One ought to eat with knife and fork”; this is not dependent on any
further ends).
If we are not willing to think of moral imperatives as etiquette imperatives, then we should
think of another way of supporting the idea that moral imperatives are categorical. One such
way is to claim that what Kant means is that in acting morally we do as reason dictates. But this
is precisely what Foot thinks is ungrounded. She thinks it is perfectly rational that someone asks
himself or herself Why should I be moral? Why be moral? The immoral – he or she who sees no
reason to obey moral precepts – cannot be accused of being irrational. Many things may be said
about a person not willing to do x, when one thinks he, or she, ought to do it. We may say that
she is cruel, selfish, rash or imprudent. Yet someone who thinks that moral imperatives are
categorical would have to claim that she is irrational. But that we cannot do. Moral requirements
do not per se give us a reason for acting. Moral reasons are available only for she who cares about
moral good – such is Foot’s claim. It is in this sense that moral imperatives are conditional, or
hypothetical.
McDowell does not agree: for him moral imperatives are indeed categorical. He does agree
with Foot though in that there is no irrationality in not conforming to them. His strategy for spelling
this out is to focus on the virtuous person. Of course, he is never fully explicit about what a
virtuous person is, nor need he be, he thinks. That there are virtuous persons is simply the
starting point for understanding morality as an exercise of reason. If one is a virtuous person the
thing to do in a particular situation is objective, it is simply there to be seen. The virtuous
person is distinguished by the way she reads events: the way she perceives things to be gives her
the reasons to do x. She recognises requirements: no further desire needs to be added.
In the arena of moral philosophy, it matters to put things in such terms against the Humean.
The Humean sees reason (the representing of circumstances) as motivationally inert – it is my
having a brute desire for E which constitutes my having a reason to pursue E. For the Humean,
there is thus a gap between representing things as thus and so and being moved to act; such gap
can only be filled by a desire.
Naturally, a whole view of perception goes with McDowell’s criticism of the Humean
belief-desire model. My main point here is just that McDowell thinks that there is no neutral
perception of facts which is shared by every perceiver, to which a motivational layer is then
added. Beliefs themselves are motivating. If in a given situation someone does not see X as the
thing to do (e.g. I do not see that I ought to my debt to Susana) this happens not because she
lacks the desire a virtuous person has but because she does not see reality the same way a
virtuous person does.
So Foot thinks that moral imperatives are hypothetical and thus Kant was wrong whereas
McDowell thinks moral imperatives are categorical and Kant was right. They do give us the
thing to do unconditionally in that there is no supplementary desire needed for being motivated
to do the thing to do. Hume was wrong there. Yet the fact that one does not see certain traits of
moral reality is no sign of irrationality. For McDowell, moral requirements are categorical not
because they are recognisable by every rational being but because once they are recognised they
necessarily motivate those who see them. In McDowell’s expression moral reasons silence every
other reason present.
Notice that McDowell does not share Kant’s rationality-based universalism. Notice also that
since the question here is seeing (i.e. conceiving things a certain way) and not reasoning, no
rational argument can bring an agent to see a situation a certain way. In order to see things as
being a certain way, one has to be the right sort of person, a virtuous person. That involves
Bildung, that is, education and custom, and thus second nature. A proper account of responsiveness to moral reasons thus involves the agent’s capacities to see moral saliences as well as the
341

Sofia Miguens

circumstances the agent is in. That considered, he claims that the virtuous person has an understanding of the situation which involves not only having the belief that there is something to
do but also being motivated to do it. She sees the thing to do; for her the thing to do is
objective. Truth and objectivity are present in the account, but not an external standpoint from
which to recognise them.
The important point about rationality here is that being virtuous is not something which
separates rational from irrational people. It is rather a matter of Bildung, of training of the
practical intellect so that certain reasons to act become visible for an agent. In other words, the
exercise of our moral capacities is dependent on the tuning and shapening of (moral) perception
by education. It is always in particular circumstances that virtue becomes an habit, and thus
“second nature” to a human.
Notice that if things are so with our moral capacities, then not only ethics cannot be formalised into a set rules to be applied in similar cases (knowing the thing to do in each case
cannot be deduced from general principles, since it requires judgment in context) but also
matters ethical will turn out to be quite different from what one might have expected. They
will turn out on matters of perceiving things in certain ways and in that sense close to perceptual
judgements. Imagine that I am watching a woman being stoned to death for adultery and I see it
is as unimaginably cruel – still the person next to me sees it as rightful. There will never be
anything like a rational proof of a judgements such as “To stone this adulterous woman to death
is the thing to do” or “To stone this adulterous woman to death is not the thing to do”. But
McDowell’s point is that this does not per se mean that reason is not involved in the agent’s
thinking in any particular circumstances what the thing to do is.
Is this not clear relativism? It is certainly a picture in which recognising the objectivity of
moral reasons is dependent on capacities which are admittedly parochial. The problem is, how
do we know that it is our eyes are opened to the right reasons and not somebody else’s?
McDowell is aiming for a very difficult balance: he does not defend that moral reasons are
universal, but rather that they necessarily appeal to rational beings for whom virtue has become
second nature. There is thus no Kantian universality of rationality. But neither is there anything
like a natural foundation for morality. Our moral capacities are not part of our first nature: they are
rational capacities, they are conceptual and contextually tuned, and as such they are not merely
a rechanneling of natural impulses. For a natural creature being rational is being able to step
behind natural impulses and ask for reasons. There is no fixed way of doing that. That is the
point of freedom in the thought experiment of the rational wolf. That is what he means by
letting (one’s) mind roam over possibilities of acting other that what comes naturally to wolves (or
humans). This is what McDowell believes being rational is for a human animal. That is one
important point of the thought experiment.
It is maybe not too surprising that the question of the relation between reason and nature
brings McDowell close to Kant. At the end of “Two Sorts of Naturalism”, he observes that
there is a Kantian flavor in his naturalism “though it is entirely free of formalism and supernaturalism”. He goes as far as saying that
It takes reflection on Kant, of all people, to show us the way to an acceptable picture
of the relation of reason to nature.7
The Kantian traits of McDowell’s position are the connection between reason and freedom and
the idea that freedom is not a channeling of our first nature. Remember that for McDowell there
can be no account of rational moral capacities without an account of appeal of virtue to reason and
such appeal simply cannot be grounded on first nature. This is one point of the rational wolf
342

The rational wolf

thought experiment. The problem the rational wolf poses himself is meant to show that for a
rational creature its own behavior is not, for itself, just one more phenomenon in the world to be
conceived as something that happens. It is a question of agency (Kant would speak of spontaneity).
This is the problem of freedom, not the free-will problem, conceived as a question regarding the
relations between free will and determinism. It is not about determinism. It is about Why should I
be moral? why should I hunt with the pack?. In other words, McDowell is saying that we need
something more than the free will problem to talk about freedom: we need a view of the relations
between reason and freedom. This is not what usual approaches to free-will give us.8

4 McDowell’s case for liberal naturalism (II): the place of value
in the natural world
A view of value is one further element of McDowell’s case for liberal naturalist. How are we to
understand our capacity to see value? Can it fit the viewpoint of natural science upon nature? This is
the question famously posed by J.L. Mackie, whose error-theory answers it negatively. Spelling
out what is wrong with such answer is instrumental for McDowell case for liberal naturalism.
The error-theorist grants that our evaluative discourse has cognitive content: it does seem to
us that there is value in the world. Yet there is not. It is a massive error, an illusion: values are
not part of the fabric of the world, Mackie claims.
In “Values and Secondary Qualities”9 McDowell takes the steps he believes are needed to
counter this conviction: basically he proceeds to bring apart the conception of nature and value
it is built on.
Ultimately, he proposes that in order to be moral realists we do not have to think of moral
properties as primary properties, or assume that the world as it is in itself can only be described
in terms of primary properties. In fact, it is this idea of the world as it is in itself that is a fantasy.
An analogy with color does work here. Colors are not less real because they are to be understood in terms of the object’s disposition to present a certain kind of perceptive appearance
and thus in terms of how they appear to a subject. The property of an object “being red” is to
be understood as “being in certain way such that under certain circumstances it appears red” to
a mind. Moral properties, like color properties, depend on being perceived by subjects with the
appropriate sensibility in certain circumstances. There is no such thing as “being red” which is
not appearing red to some mind. Yet this does not per se mean that such qualities are not there
to be perceived independent of that particular appearing to a particular mind. Such properties
are not subjective in the sense of being illusory. They are not illusory – in fact they are there to
be experienced (by many minds).
McDowell’s main point here is that a conception of the world should include room for
experience of the world and for what there is from the viewpoint of such experience – and this
is not the case with Mackie’s conception. His conception of the world, or nature, is simply too
thin; it identifies nature with the content of a view from nowhere, where there are only
primary properties. Such view of nature should, according to McDowell, be rejected.
Of course, there is another alternative here. We might regard value not as illusion but simply
as spreading (ourselves) unto the world. This is what Hume did, and this is the point in
contemporary moral philosophy of, for example, Simon Blackburn’s projectivist quasi-realism,
a Humean, that is, a non-cognitivist, view of our moral capacities.10
The neo-Humean sees our moral judgements not as descriptions of reality but as expressions
of our attitudes before it. According to Humean projectivism, properties that seem to genuinely
belong to objects are just a projection or reflex of our subjective responses to a world which
in fact does not contain such properties. Blackburn’s account of this situation, that is, his
343

Sofia Miguens

quasi-realism, is McDowell’s target in “Projection and Truth in Ethics” and “Non-cognitivism
and Rule-Following”.
For McDowell, the core mistake of the projectivist is that she explains genuine traits of
reality as reflexes of our subjective responses. In other words, according to the projectivist, there
is priority and explanatory independence of subjective responses in relation to the aspects to be
explained. McDowell’s claim is that there is no such priority: our feelings and the traits of reality
at stake (moral properties) are paired as siblings, and not as parents and children. Neither are
moral properties prior and independent to our subjective responses as the realist would have it,
nor are our subjective responses prior to moral qualities as the projectivist thinks.
The opponent can obviously point out that McDowell’s no-priority view is circular, and
needs to appeal to something like the default human sensibility and doubt that there could be such
thing. The view is also obviously conservative: Blackburn accuses McDowell of merely citing or
postulating the ethical verdicts of our own concepts and practices. These are the accusations
McDowell is defending himself from in “Non-cognitivism and Rule-Following”. His claim is
that when we say that acting virtuously is what the virtuous person does that is simply the end of
the line. We cannot go any further; there is no way out of this circle. The situation simply reflects
the fact that we cannot think of value from without our evaluative experience. That does not
mean, though, that there is no value or that value is merely a reflex of our responses. Only if we
believe that it is possible to step back from our ongoing practices to ground them will we believe
that a realism dependent on human sensibility is not sufficient. But for McDowell there is no
sideways-on view available here to allow us to pursue such grounding; we cannot transcend our
practices nor our parochial viewpoint towards a supposed “reality such as it is in itself”. When it
comes to accounting for our moral capacities there is nothing to look for beyond that which we
have learned when we were introduced into a practice. In McDowell’s famous simile there is
nothing to keep us on the rails,11 except for the shared practices themselves. Practices are all we
have, even to account for the rightness of practices.

5 Some conclusions regarding liberal naturalism: the fight
against an ungrounded attraction and a transformative (and pragmatist)
view of rationality
I said at the beginning that we should not expect to find an argument for naturalism, or for
liberal naturalism, in McDowell’s work. McDowell has a case, a case for liberal naturalism
which is built by fighting, in several fronts, what he sees as the ungrounded attraction of a bad
metaphysics.12 He believes there is a spell, which should be broken, hanging over the discussions of the nature of reasons and value above, as it does over others (e.g. discussions of
meaning, perception or other minds). The spell is the picture of “shallow empiricist naturalism”. It has us accept that what science aims to discover is nature of reality in so far as it can
be characterised in absolute terms, as a view from nowhere. It leads to forms of naturalism based
on a concept of nature according to which meaning and value are, in McDowell’s expression,
“injected from the outside”. In contrast, as we saw in the cases of responsiveness to reasons and
of value, the liberal naturalist has come to accept that she is dealing with the interior of nature.
She sees responsiveness to reasons as not attached to anything like universal rationality yet still
objective. She sees moral values as attuned to the sensibility of particular agents and moral
properties as anthropocentric yet real. She believes that we examine our moral practices (and
our conceptual practices in general) from within, and that in doing that there is no possibility of
dissociating descriptive from normative elements. She thinks that there is no such thing as an
evaluatively neutral reality onto which moral judgment projects, or injects, our values.
344

The rational wolf

I started out suggesting that we regard McDowell’s moral philosophy as key to his global
philosophical outlook. This is relevant also because McDowell’s philosophy is a central case of
liberal naturalism in the current philosophical landscape. So what we now have is the following.
McDowell’s whole philosophy, his moral philosophy included, is a view of reason, or rationality, a
framework in which to think about the capacities of a rational animal. It is in relation to rationality,
the rationality of a rational animal, that the question of what natural is should be posed. According
to McDowell our rationality is natural – it is part of our nature, second nature to our first nature –
our animality. For us, thus, there is no getting outside our rationality as there is no stepping outside
it in order to account for it. There is no external standpoint. This leads to what Matt Boyle calls a
transformative view of rationality,13 a view that clashes head-on with the additive views characteristic of most current naturalist approaches. This is how Boyle sees the contrast:
Additive theories of rationality, as I use the term, are theories that hold that an
account of our capacity to reflect on perceptually-given reasons for belief and desirebased reasons for action can begin with an account of what it is to perceive and desire,
in terms that do not presuppose any connection to the capacity to reflect on reasons,
and then can add an account of the capacity for rational reflection, conceived as an
independent capacity to ‘monitor’ and ‘regulate’ our believing-on-the-basis-ofperception and our acting-on-the-basis-of-desire.
There is no such thing for McDowell, as there is not for Boyle.14 We cannot have an account of
what it is to perceive or desire, in terms that do not presuppose any connection to the capacity
to reflect on reasons and then add an account of the capacity for rational reflection. This is
crucial when we are considering first nature facts (e.g. neuroscience facts) about ourselves and it
amounts to a critical stance on many current naturalisms. Most current naturalists do not pay
sufficient attention to the concept of nature and thus unreflectively pursue additive theories.15
One last point should be considered in McDowell’s second nature naturalism regarding our
moral capacities. We might accept that there is no view from nowhere, no sideways on view on
ourselves and on our conceptual capacities and their exercises – the account of moral capacities
illustrates that. But of course, in the case of moral capacities there is still another problem. The
problem is that we can very well conceive of different ways of applying ethical concepts. We
just have to look around us and we do see alternative applications of ethical concepts – what
reasons do we then have to think that our way of applying concepts is the right one? How can
we ever argue that one particular way of applying moral concepts is superior to another?
This is one major objection to McDowell’s view and it comes neither from the reducionist
viewpoint of the scientific naturalist, nor from the Kantian universalist, nor the Platonic realist.
We find it, for example, in Bernard Williams in Ethics and the Limits of Philosophy (1987) and
Williams’ student Alan Thomas sums it up like this in his 2006 book Value and Context (2006):
The proposed form of cognitivism [i.e. McDowell’s view] can give an excellent account of
particular forms of ethical reasonings and practices as they arise in a given historical community, with its culturally specific concepts and practices. However, it fails to allow for the
possibility of a certain kind of radical, distinctively modern form of reflection in which we
take a critical stance towards the pratices of our own historic community or are challenged
by practices and ideals of other communities (our way of going on is just local).16
In other words, McDowell’s account lacks the capacity to account for any kind of (critical)
moral insight into our own ongoing moral practices. If our own purchase of moral properties
345

Sofia Miguens

takes place strictly from within one particular ongoing form of life, our own, it seems we are
helpless in responding to the fact that there seems to be more than one such form of life, each
giving access to its own range of “moral properties”. We are left empty handed to cope with
radical pluralism, that is, with the existence of competing ethical claims within coexistent moral
practices. In Thomas’ terms such form of pluralism simply goes beyond “a reasonable pluralism
within morality” which is a “welcome complement of our idea of autonomy”. It presents a
further challenge, it is something different. Again in Thomas’ terms, such radical pluralism can
be perceived as an avenue not to freedom but to nihilism.
Thomas is right that McDowell’s proposal as it stands is derailed by Williams’ critique. Yet
he himself suggests a way out. I believe his suggestion is a good suggestion and converges with
an idea regarding the nature of rationality which find articulated by some pragmatists, for
example, Stephen Stich.17
Thomas’ suggestion is that McDowell’s proposal can be put back on track if one isolates and
criticises one key assumption: we have to distinguish two ideas concerning rationality and moral
capacities (and the way we conceive of our moral capacities as rational). We can indeed form a
conception of what it would be for an alternative framework of ethical judgment to be superior
to another. What we cannot do (without a grand, teleological, Hegelian historical metanarrative) is to iterate such conception in order to yield the idea that that framework cannot be
surpassed by another. This is the distinction we need if we are to be able to defend our own
moral capacities (as rational capacities) across entire frameworks of moral belief, across cultures
and historical differences. This is how one might allow for the bearing of the kind of moral
insight into our own ongoing moral practices that Williams urged was lacking in McDowell’s
cognitivism. We may very well claim that A is superior to B – what we are not in a position to
claim is that say, Z, is the ultimate framework of ethical judgment.
Thomas’s suggestion concerns rationality and moral capacities, and is formulated within a
discussion of moral and political philosophy but in broad strokes it converges, as I said, with the
pragmatist view of rationality put forward by Stephen Stich. The basic elements of such theory
are the following.18 From a pragmatist viewpoint being rational for an agent is doing well in
pursuing goals. Such goals are not true beliefs, or valid reasonings, or being rational per se; they
are whichever goals the (real, physical, biological) agent has. Cognitive processes are agents’
tools for reaching such goals. The important point is that the evaluation of cognitive tools is
possible but only in a comparative, rather than absolute, way. There is no God’s eye view on
doing well, no doing well an sich. Yet there is such thing, for an agent, as doing better and doing
worse (in a world). In other words, it is possible to say whether a cognitive system is doing
better or worse than another system, for a particular task, in a particular situation or context,
with a particular cognitive equipment. But there is no such thing as the doing well of the ideal
rational agent independent of a world and of situations therein. In fact, there is no such thing as
an ideal rational agent. The comparison that matters for cognitive evaluation is the comparison
between actual alternatives, that is, alternatives which are equally available for a particular agent
in particular circumstances. That may very well be all that is needed for a liberal naturalist
account such as McDowell’s to stand.

Notes
1 The exercise of our conceptual capacities in perception is crucial for McDowell’s whole view; the
classic formulation of McDowell’s position on perception can be found in Mind and World (McDowell
1994). According to McDowell perceptual experiences represent things as being a certain way. A
perceiving is a taking-to-be, which then is (or not) endorsed by judgement. In more recent

346

The rational wolf

2
3
4
5
6
7
8
9
10
11
12
13
14

15
16
17
18

formulations (see, for example, Avoiding the Myth of the Given, McDowell 2009, 267), McDowell
speaks of perceivings as seemings and thus claim-like. “Claim” is a term of Wilfrid Sellars, a term which
McDowell thinks is “wrong in the letter but right in spirit”. For McDowell’s conceptualist view of
perception and what it stands opposed to in current philosophy of perception, see Miguens (2019).
My main references are the articles collected in Part II (Reason, Value and Reality) of
McDowell (1998).
McDowell (1998).
This is a reference to Aristotle, for whom a man’s character is second nature to him.
De Caro and Macarthur (2004, 2010).
McDowell (1998).
McDowell (1994, 1998).
This is fact is one source of a current attraction within analytic philosophy to German Idealism (and to
Kant). The other is that normativity is a problem for naturalism.
McDowell (1998).
Blackburn (1984).
McDowell Virtue and Reason (91).
Naturally, McDowell’s Wittgensteinian conception of philosophy informs the whole approach. See
Miguens (2019).
Boyle (2016).
Boyle identifies two difficulties for the additive approach, each analogous to a classic problem for
Cartesian dualism. The interaction problem concerns how capacities conceived as intrinsically independent of the power of reason can interact with this power in what is intuitively the right way. The
unity problem concerns how an additive theorist can explain a rational subject’s entitlement to
conceive of the animal whose perceptual and desiderative life he or she oversees as “I” rather than “it”.
He argues that such difficulties motivate a general skepticism about the additive approach to rationality.
Appealing to first nature facts, as many contemporary naturalists do comes too close to a single view of
what natural is for a rational animal. McDowell thinks such claims are undue.
Thomas (2006, 2–3).
Stich (1990).
Miguens and Pinto (2018).

References
Blackburn, S. Spreading the Word. Clarendon Press. (1984).
Boyle, M. Additive Theories of Rationality: A Critique. European Journal of Philosophy, 24(3), (2016).
527–555.
De Caro, M. and Macarthur, D. Naturalism in Question. Harvard University Press. (2004).
De Caro, M. and Macarthur, D. Naturalism and Normativity. Columbia University Press. (2010).
Foot, P. Morality as a System of Hypothetical Imperatives. In P. Foot (ed.), Virtues and Vices, pp. 157–173.
Oxford University Press. (1978).
Mackie, J. Ethics – Inventing Right and Wrong. Penguin. (1977).
McDowell, J. Mind and World. Harvard University Press. (1994).
McDowell, J. Mind Value and Reality. Harvard University Press. (1998).
McDowell, J. Values and Secondary Qualities. In J. McDowell (ed.), Mind Value and Reality, pp. 131–150.
Harvard University Press. (1998a).
McDowell, J. Projection and Truth in Ethics. In J. McDowell (ed.), Mind Value and Reality, pp. 151–166.
Harvard University Press. (1998b).
McDowell, J. Non-Cognitivism and Rule-Following. In J. McDowell (ed.), Mind Value and Reality,
pp. 98–218. Harvard University Press. (1998c).
McDowell, J. Are Moral Requirements Hypothetical Imperatives? In J. McDowell (ed.), Mind Value and
Reality, pp. 77–94. Harvard University Press. (1998d).
McDowell, J. Might There Be External Reasons? In J. McDowell (ed.), Mind Value and Reality,
pp. 95–111. Harvard University Press. (1998).
McDowell, J. Two Sorts of Naturalism. In J. McDowell (ed.), Mind Value and Reality, pp. 167–197.
Harvard University Press. (1998f).
McDowell, J. Virtue and Reason. In J. McDowell (ed.), Mind Value and Reality, pp. 50–73. (1998g).

347

Sofia Miguens
McDowell, J. Avoiding the Myth of the Given. In J. McDowell (eds.), Having the World in View,
pp. 256–272. Harvard University Press. (2009).
Miguens, S. and Cadilha, S. John McDowell – uma análise a partir da filosofia moral. Colibri. (2014).
Miguens, S. Is There a Single Way for All Humans to Be Human? Some Problems for Aristotelian
Naturalism in Contemporary Moral Philosophy. Filosofia, Revista da Faculdade de Letras da Universidade
do Porto, 34, (2018). 167–186.
Miguens, S. and Pinto, J.A. Seeing What a ‘Science of Rationality’ Founders on (With a Little Help From
D. Davidson). Poznań Studies in the Philosophy of the Sciences and the Humanities, 111, (2018). 71–92.
Miguens, S. Is Seeing Judging? Radical Contextualism and the Problem of Perception. In E. Marchesan
and D. Zapero (eds.), Truth, Context and Objectivity. Routledge. (2019).
Miguens, S. Temptation and Therapy – Wittgensteinian Responses to Other Minds Skepticism.
Wittgenstein Studien, 10, (2019).
Stich, S. The Fragmentation of Reason – Preface to a Pragmatic Theory of Rationality. MIT Press. (1990).
Thomas, A. Value and Context – The Nature of Moral and Political Knowledge. Oxford University
Press. (2006).

348

29
RAWLS AND LIBERAL
NATURALISM
Paul Patton

Given that Rawls’s primary concern is to lay out a liberal conception of a just society, it would
be easy to think that if he offered a naturalistic conception of justice, it would have to be a
liberal naturalistic conception. However, there are two senses of “liberal” at work here: one
that countenances a plurality of ontological and methodological approaches in philosophy and
one that countenances a plurality of comprehensive philosophical, moral and religious points of
view capable of forming an overlapping consensus on a liberal conception or family of conceptions of justice. In his publications from the 1980s onwards, Rawls was interested in a
specific form of liberalism that he called political liberalism, by which he meant a political
theory that accounts for the possibility of a sustainable, stable political community in which
there is freedom and diversity of opinion and judgment grounded in different comprehensive
points of view. In contrast to liberal naturalism that countenances a certain kind of diversity of
ontological and methodological approaches to philosophical questions, Rawls’s political liberalism countenances a diversity of moral outlooks that hangs together by means of an overlapping consensus.
The question at the heart of this chapter is the nature of the relationship between liberal
naturalism and Rawls’s account of the conditions of such a consensus. The question is not
whether Rawls himself ever advocated a view about naturalism: he did not. Indeed, in his later
work he advocated a position of neutrality in relation to controversial philosophical, religious
and other views that appears to make it difficult to determine any precise relationship between
his political philosophy and naturalism, liberal or otherwise. However, I argue below that this
“method of avoidance” as he calls it is beside the point in relation to the naturalism or otherwise
of his theory of justice. The real issue concerns the ontological and methodological commitments endorsed by Rawls’s political liberalism: do these imply some form of non-naturalism or
might they be consistent with a liberal rather than a stricter kind of naturalism?
The differences between the Rawls of A Theory of Justice (1971) and the Rawls of Political
Liberalism (1993) complicate the question of his relation to liberal naturalism. For example, A
Theory of Justice made no clear distinction between moral and political philosophy. By contrast,
beginning with “Justice as Fairness: Political Not Metaphysical”, Rawls began to develop the
idea that a liberal theory of justice should be understood as political in a specific sense.1 In
Political Liberalism, he spells out the sense in which the theory of justice is political by reference
to its sphere of application, its relationship to comprehensive moral, philosophical or religious
DOI: 10.4324/9781351209472-34

349

Paul Patton

doctrines and its sources. These conditions are summed up in the idea that a conception of
justice is political in that:
it is framed to apply solely to the basic structure of society;…it is presented independently of any wider comprehensive religious or philosophical doctrine; and that
it is elaborated in terms of fundamental political ideas viewed as implicit in the public
culture of a democratic society.2
The last condition is key to the question of the naturalism of this political philosophy. It raises
the question whether we can make sense of political ideas “implicit in the public culture” of a
democratic society without recourse to any unacceptably nonnatural properties or connections.
The successive versions of the Rawls’s conception of justice rely on a consistent set of such
political ideas. All rely on the fundamental normative idea of society as a fair system of cooperation. All rely on the idea that a conception of justice should be the outcome of “reflective
equilibrium” between the principles proposed and the considered intuitions or judgments of
the people concerned. The status of these “considered intuitions” is crucial to the question of
compatibility with naturalism. We return to this issue below. Continuities aside, there are
significant differences between the early and late accounts of how such a conception might
serve as the basis for a “well-ordered” society. For Rawls, a society is well ordered when it is
effectively regulated by a public conception of justice. This implies three things: first, that
everyone accepts and knows that everyone accepts the same political conception of justice or
some recognisably liberal variant; second, that the basic structure satisfies and is known to satisfy
one or other set of liberal principles of justice; and third, that citizens have an effective sense of
justice, meaning that they understand and can apply the principles, especially in public debate
over claims of political right.
Perhaps the most significant difference between Rawls’s earlier and later work emerges in
the successive accounts of how such a public conception of justice could form the basis of a
stable political order over time. In the preface to Political Liberalism, he devoted particular attention to the concept of stability. A Theory of Justice sought to show that the conception of
justice as fairness was more stable than competing utilitarian conceptions. Unlike principles of
utility, Rawls’s principles of justice did not allow for a situation in which some might be
expected to accept a lesser share of primary social goods for the sake of others.3 He argued that,
for a well-ordered society to be stable, “Its principles should be such that when they are
embodied in the basic structure of society men [sic] tend to acquire the corresponding sense of
justice and develop a desire to act in accordance with its principles”.4 His argument that the
principles of justice as fairness could serve as the basis of stability relied on further claims about
the good of a well-ordered society and the “congruence” of this public good with the individual conceptions of the good present in the society. In turn, this argument for congruence
relied on assumptions about the moral character of citizens and the good of autonomy that
presupposed a particular moral point of view. These assumptions could not be sustained once
the fact of reasonable pluralism in relation to individual conceptions of the good was taken into
account.
This pluralism is not simply a matter of fact that results, for example, from the increasingly
multicultural character of liberal societies. It is rather a permanent feature of liberal democratic
public culture that results from “the work of free practical reason within the framework of free
institutions”.5 What Rawls calls the “burdens of judgment” will ensure that different opinions
will arise in the course of reasoning about political matters. These burdens include the complexity of evidence, the different weights attached to different kinds of consideration, the
350

Rawls and liberal naturalism

vagueness and indeterminacy of political concepts such as equality, fairness, freedom or power
and the effects of different life experience on the ways in which individuals assess evidence and
weigh moral and political values.
These inescapable and perfectly legitimate sources of disagreement in a liberal democratic
society lead to the idea that part of being reasonable is recognising and being willing to bear the
burdens of judgment in our interactions with others. Another part of being reasonable is
willingness to propose principles as fair terms of cooperation and to abide by them, given the
assurance that others will do likewise.6 Rawls’s political liberalism builds on this conception of
reasonableness to argue that liberal principles of justice can only serve to maintain a wellordered society when they are the objects of an overlapping consensus among citizens with
different comprehensive, moral, religious and other views. Political Liberalism is devoted to the
following problem: “How is it possible that there may exist over time a stable and just society of
free and equal citizens profoundly divided by reasonable though incompatible religious, philosophical and moral doctrines?”7 Rawls’s solution relies on the idea that all reasonable citizens,
whatever their comprehensive moral and other doctrines, should be able to agree on a particular
conception or family of conceptions of justice. This is what he means by an “overlapping
consensus”. If the society is well-ordered, the resultant conceptions of justice will be supposed
to regulate the terms of social cooperation, even though they are not coextensive with the
broader philosophical, religious or other moral doctrines to which individuals might subscribe.

1 Realism, constructivism and the method of avoidance
In Lecture Three of Political Liberalism, Rawls argues that he puts forward a constructivist view
about the structure and content of a political conception of justice. By this he means that, once
reflective equilibrium is reached, “the principles of political justice may be represented as the
outcome of a certain procedure of construction”.8 The procedure envisaged is that of the
original position in which “rational agents, as representatives of citizens and subject to reasonable conditions, select the public principles of justice to regulate the basic structure of
society”.9
He specifies that the procedure of construction is based not on theoretical but on practical
reason, which he defines following Kant as that form of reason concerned to produce objects
according to a conception of those objects. In this case, the relevant object is that of a just
constitutional regime given plurality of comprehensive moral views. Theoretical reason, which
is concerned with the knowledge of given objects, can and does still play a role in the process of
devising and selecting the principles of justice. In addition, Rawls notes that the reasoning that
produces a political conception of justice relies on a complex conception of persons as possessing certain relevant moral powers, namely a capacity for a sense of justice and a capacity for a
conception of the good. It also relies on certain fundamental political ideas such as a conception
of society as a fair system of cooperation and an idea of the reasonable that applies to persons,
principles and institutions. As noted above, the idea of the reasonable involves the willingness of
persons to propose and abide by fair terms of cooperation and their willingness to accept the
consequences of the burdens of judgment. This is a political rather than a metaphysical conception of persons. Importantly, political constructivism does not rely on a commitment to the
truth of these fundamental political ideas, only to the truth of the claim that they are ideas
implicit in the public political culture of a democratic regime. On that basis, it suggests that a
conception of justice built upon those ideas is acceptable when it can be endorsed on due
reflection, that is with the achievement of wide reflective equilibrium.

351

Paul Patton

Rawls explicitly contrasts his political constructivism with moral realism in the form of the
rational intuitionist approach proposed by in the eighteenth century by Samuel Clarke and
Richard Price, in the nineteenth by Henry Sidgwick and in the twentieth century by W.D.
Ross. This rational intuitionism is characterised by four features: first, the idea that moral first
principles and judgments, when correct, are true statements about an independent order of
moral values; second, the idea that moral first principles are known by theoretical reason, along
the lines of mathematical knowledge, that is by a kind of perception or intuition along with
reflection; third, whereas political constructivism relies on a conception of persons as rational
and reasonable, rational intuitionism requires only a conception of the person as knower endowed with the capacity to be moved to act in accordance with the principles of morality for
their own sake; finally, rational intuitionism relies on a correspondence theory of truth according to which moral judgments are true when they accurately represent the independent
order of moral values. This contrast between Rawls’s political constructivism and rational intuitionism makes it clear that he cannot be construed as a moral realist in a manner that would
be incompatible with naturalism.
At his point, however, he introduces a further clarification that makes it difficult to fully
align his political constructivism with naturalism, liberal or otherwise. He insists that political
liberalism does not contradict rational intuitionism.10 Nor does it deny that there is an independent order of moral values. It claims only that the values expressed by justice as fairness
follow by means of the original position from the principles of practical reason and the relevant
conceptions of society and the person. It also asserts that these are the most appropriate political
values for a democratic society characterised by reasonable pluralism. Rational intuitionists can
affirm the outcome from within their own comprehensive view and thereby participate in an
overlapping consensus.11
This careful refusal to take a position with regard to the truth or falsity of particular moral,
philosophical or religious views amounts to what Rawls referred to in some articles prior to
Political Liberalism as the “method of avoidance”. So, for example, in “Justice as Fairness:
Political not Metaphysical”, he presented this method as essential to the practical aim of political
liberalism, namely achieving a secure overlapping consensus among citizens with a plurality of
reasonable views on a range of fundamental questions.12 Political liberalism sidesteps divisive
questions about moral truth or the nature of the self by characterising objectivity in terms of a
constructivism that neither asserts nor denies particular theories about the status of moral and
political values. Rawls’s hope is that “by this method of avoidance, as we might call it”, the
differences between moral and political views can be moderated, if not eliminated, “so that
social cooperation on the basis of mutual respect can be maintained”.13 Similarly, in Political
Liberalism, he suggests that it is important that his political constructivism neither affirms nor
denies the truth of any particular reasonable view. Nor should it be indifferent or skeptical
about such truths:
Such skepticism or indifference would put political philosophy in opposition to numerous comprehensive doctrines, and thus defeat from the outset its aim of achieving
consensus.14
Political liberalism simply avoids taking a position on the truth or falsity of religious, moral
or philosophical doctrines. As noted above, political constructivism avoids claiming that the
principles of justice it recommends are true. It bases itself on the shared fundamental ideas of
the public political culture in attempting to develop a conception of justice that accords with
the considered convictions of people. The question of the truth or falsity of particular doctrines
352

Rawls and liberal naturalism

is put aside in favor of overlapping consensus on a conception or conceptions of justice. This
does not require judgment about the truth or falsity of all things that might be claimed from
within a given comprehensive view, although it does require judgment about the truth of some
claims, such as those about the fact of reasonable pluralism or the burdens of judgment. Political
Liberalism leaves such judgments to individual citizens.15
Although Rawls used the phrase “method of avoidance” to describe this strategy in “Justice
as Fairness: Political not Metaphysical” and in “The Idea of an Overlapping Consensus”, the
passage in Political Liberalism corresponding to the one in the latter text removes this phrase.16
Henry Richardson notes that, although the strategy of refusing to take sides is used pervasively
throughout Political Liberalism, there is only one reference in a footnote to the “precept of
avoidance”.17 Moreover, it is noteworthy that, although the strategy of avoidance is employed
throughout Political Liberalism, the phrase is avoided in the passage on page 150 that refers to this
strategy. Richardson offers a possible explanation for this: since the idea that the method of
avoidance is appropriate for a political theory of justice is controversial, “the method of
avoidance counsels avoiding any assertion that the method of avoidance should be followed”.18
In effect, the same higher order strategy of avoidance counsels denying that it should be followed. Instead, Richardson suggests, “one should simply follow it, as Political Liberalism did,
without trumpeting that fact”.19
The question that arises in relation to naturalism is whether this method of avoidance makes
it difficult to impute to Rawls a commitment to naturalism rather than non-naturalism, and to a
weaker form of liberal naturalism rather than a stronger version such as physicalism or? On the
one hand, there does appear to be something odd about suggesting that a political philosophy
that was committed to such a strategy of avoidance could be shown to imply any particular form
of naturalism. On the other hand, the fact that it does not affirm any commitment to naturalism
does not mean that political liberalism is not naturalistic. There is an important difference
between the principles of political liberalism, including the principle of avoiding taking sides on
controversial philosophical issues, and the presuppositions of the theory of justice as fairness
that, Rawls argues, would be compatible with those principles. As noted earlier in relation to
the original position and reflective equilibrium, the theory of justice as fairness relies on certain
claims about the political ideas implicit in public political culture and the bases of consensus on a
restricted range of political issues and principles. Political Liberalism argues that consensus is
possible to the extent that the principles worked up on the basis of those ideas accord with the
considered opinions of a people about what is fair or just. The naturalism or otherwise of this
political philosophy then depends on our understanding of the nature of such considered
opinions or intuitions. If they are compatible with a scientific view of the world, liberally
understood so that it includes the social sciences, then we can align Rawls’s political philosophy
with a liberal naturalism.

2 Naturalism, the original position and the evidential basis of
the theory of justice
There is a growing body of historical literature on Rawls that approaches his work in the
context of the development of twentieth century analytic philosophy. A recent contribution to
this literature by Nikhil Krishnan places A Theory of Justice squarely in the tradition of a revived
naturalist approach to ethics at Oxford initiated by Philippa Foot, Elizabeth Anscombe and Iris
Murdoch.20 In explicit reaction to the nonnaturalism of G.E. Moore, the emotivism of logical
positivist approaches to ethics, and the neo-Kantian formalism of R.M. Hare, these thinkers
shared a methodological commitment towards seeing human ethical practices “in the context of
353

Paul Patton

a more general ‘anthropology’ of human life”.21 Krishnan argues that, partly under the indirect
influence of late Wittgensteinian philosophy but also in the tradition of nineteenth century
British utilitarianism, Rawls thought that the ultimate bases of moral reasoning should be
sought in criteria of ethical judgment actually employed by people. He argues that the novelty
of A Theory of Justice lay not only in its analytical approach to substantive questions in political
philosophy but also in the fact that Rawls
shared the methodological “naturalism” of [analytic philosophy’s] orientation in that
he had a conception of what ethical, and normative political, questions were about,
i.e. their subject matter, that offended against nothing in a fully scientific view of both
human beings and their world.22
Despite such arguments based in intellectual history and Rawls’s biography, there is a way of
understanding the early Rawls that appears to commit him to a nonnaturalist conception of the
argument in support of his conception of justice as fairness. This approach is based on an
understanding of his task as the construction of a conception of justice that would be endorsed
by rational agents under the hypothetical conditions of the original position. Anthony Laden
argues that much of the reception of A Theory of Justice, positive as well as negative, relied on a
“blueprint” of this kind that took Rawls to be arguing from “a conception of human nature or
human rationality to the validity of the two principles of justice”.23
On this view, both the agents or parties to the original position and the conditions under
which they deliberated behind a veil of ignorance are highly idealised constructs. It follows that
the reasons appealed to in justifying the resultant principles of justice will be similarly ideal.
They may be supposed to exist whether or not any actual citizens endorse them. Understood in
this way, the argument of A Theory of Justice appears both be to be deductive, on the basis of a
minimal set of assumptions about human nature, and to take place in a space of reasons that is
independent of the natural social and political world. Actual societies may instantiate the resultant idea of a just society to greater or lesser extent, but the conception of justice is justified
by reason regardless of whether or not it is actualised in any existing society. This approach leads
to a kind of Platonism about political norms that, as De Caro and Macarthur point out, would
be incompatible with any form of naturalism, liberal or otherwise. According to such a view,
the principles of justice would express certain normative facts that exist “independently of
human practices (say, of reason giving) and are, as it were, simply there anyway waiting to be
discovered”.24
Laden argues that this way of reading Rawls is mistaken, both about the nature of his
ambition and about the arguments deployed. It led many commentators to suppose that the idea
of the original position was a key means to make the argument for a particular conception of
justice as fairness. “Making the argument” here can mean one of two things. It can mean that
the original position was a mechanism of theory construction, whereby the circumstances of the
original position, including the veil of ignorance, are employed as a procedure designed to
produce a particular set of principles that define a just basic structure of society. Alternatively,
the device of the original position can be supposed to “make the argument” for a particular
conception of justice in the sense of “arguing for” rather than producing the principles of
justice. Understood in this way, the original position is not so much a means of theory construction as a representational device.
This is how Rawls characterises it in Political Liberalism. In these terms, the procedure involving choice by idealised agents who represent citizens under the idealised conditions expressed by the veil of ignorance “exhibits reasonable conditions to impose on the parties, who
354

Rawls and liberal naturalism

as rational representatives are to select public principles of justice for the basic structure of such a
society”.25 The conception of citizens as rational and reasonable is distributively modeled by
this procedure. The citizen’s capacity for a conception of the good is modeled by the rationality
of the parties in the original position, while their capacity for agreement on principles of justice
is modeled by features of the situation such as the equality of the representatives and the limits
on information available to them as represented by the veil of ignorance. There is in effect a
division of labor between the parties to the original position, who are merely rational, and the
procedure itself that models the reasonableness of citizens who are prepared to agree on
principles that may not directly serve their own interests when they can see that others will do
likewise.
The details of the original position were always supposed to reflect pretheoretical judgments
about what kind of reasons were appropriate to deciding on principles of justice. The exclusion
of knowledge about the social position, moral views or conception of the good of the parties
reflects the idea that these are irrelevant to determining the principles of a just basic social
structure. Rawls did not include knowledge of one’s gender as one of the items excluded by the
veil of ignorance in A Theory of Justice. His later addition of this item to the veil of ignorance
reflects the conviction that a just basic structure should be neutral in this regard. In this manner,
the details of the original position are intimately bound up with conceptions of fairness in a
given society. Since it models
What we regard as acceptable restrictions on reasons available to the parties for favoring one agreement rather than another, the conception of justice the parties would
adopt identifies the conception we regard – here and now – as fair and supported by the
best reasons.26
Understood in this manner, the purpose of the original position is not to generate principles of
justice but to provide a way of justifying them to fellow citizens who might not agree. Laden
argues that the difference between these two ways of interpreting Rawls corresponds to a
difference between two ways of understanding justification. He uses the example of a teacher
who justifies a course of study to her students by appealing to the curriculum mandated by the
state or to her own authority as an experienced teacher. Justification of this kind may be
understood as an “impersonal” practice that works by grounding a given course of action in “a
system of thought, rules or authority that she takes to hold, regardless of whether the person to
whom she is justifying her actions also accepts these premises”.27 However, justification may be
understood quite differently as an intersubjective activity in which success depends on uptake
from those to whom it is offered. Rawls understands justification in the latter sense, as he makes
clear in “Justice as Fairness: Political Not Metaphysical”:
justification is not regarded simply as valid argument from listed premises, even should
these premises be true. Rather, justification is addressed to others who disagree with
us, and therefore it must always proceed from some consensus, that is, from premises
that we and others publicly recognize as true; or better, publicly recognize as acceptable to us for the purpose of establishing a working agreement on the fundamental
questions of political justice.28
Laden relies on this understanding of justification in his account of the role of the original
position. The point of the hypothesis is not to provide a timeless and context independent basis
for the principles of justice but rather to provide justification to fellow citizens for the principles
355

Paul Patton

of justice as fairness. Understood in this way, the original position is directly bound up with the
practice of reason giving in a particular society and, to that extent, appropriately construed in
liberal naturalist terms.
In whatever manner we understand the precise function of the original position in Rawls’s
account, it is clear that the procedure was never wholly independent of human practices and
institutions. The ultimate test of an acceptable conception of justice was always the achievement
of reflective equilibrium between the proposed principles of justice and the intuitions or
considered convictions of individual members of the society. The proposed principles of justice
were supposed to generate judgments on particular issues that more or less coincided with the
considered judgments of members of the society in question about what was fair and just. The
precise status of those judgments, intuitions or considered convictions is open to dispute in
Rawls scholarship. Occasional remarks in A Theory of Justice appear to support a thoroughly
subjective interpretation, for example when Rawls says that it is the views of the readers and the
author that count.29 On that basis, it becomes possible to argue that, despite the “naturalistic
flavor” of Rawls’s method, he is insufficiently committed to naturalism. More careful attention
to empirical evidence about what are the intuitions might produce results incompatible with his
theory of justice as fairness. For example, surveys have shown a higher regard for principles of
desert than Rawls’s theory allows.30 Further experiments designed to test the actual choices
people would make behind a veil of ignorance show a preference for non-Rawlsian principles
in some communities and a degree of variation among culturally different populations.31 These
results do not support the claim that Rawls’s theory of justice would be chosen in many
societies if the benchmark is supposed to be the actual opinions of people.
Against this interpretation of the empirical benchmark of the theory of justice, it can be
argued that the considered judgments or convictions about what is fair or just were never
supposed to be a matter of purely subjective opinion. Rather, they were conceived as views
about right ways of acting that were expressed in the institutions, constitutional settlements,
legal decisions and traditions of interpretation of the society in question. As Rawls wrote in
“Justice as Fairness: Political Not Metaphysical”, his conception of justice “tries to draw solely
upon basic intuitive ideas that are embedded in the political institutions of a constitutional
democratic regime and the public traditions of their interpretation”.32 In this sense, the evidential base for the conception of justice lies not in disembodied judgments nor in the subjective beliefs of individuals but in judgments that are embedded in the history, institutions,
material culture and practices of the society in question. Such “objective” intuitions are not
assessed by surveys but rather by historical, institutional and conceptual analysis. To take one of
Rawls’s examples, the claim that slavery is impermissible in a modern democratic society is
justified by reference to legal doctrines and practices, economic institutions and practices as well
as the opinions expressed in scholarship and reflective public opinion. These are commonplace
elements of our everyday social world. They are comparable to the institutional or interpretative entities such as works of art that liberal naturalism wants to countenance as natural. In
this sense, the evidential baseline for theories of justice is composed of natural rather than
supernatural entities. The reasons that inform the theory of justice can be considered part of
nature in the expanded sense of the term, even though they may not be reducible to or explicable in physicalist terms.

Notes
1 Rawls (1985).
2 Rawls (2005, 11–14, 223).

356

Rawls and liberal naturalism
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32

Rawls (1999, 155).
Ibid., 119.
Rawls (2005, 37).
Rawls (2005, 49).
Rawls (2005, xviii).
Rawls (2005, 89–90, [emphasis added]).
Rawls (2005, 90).
Rawls (2005, 95).
Ibid.
Rawls (1985).
Ibid., 231.
Rawls (2005, 150).
Rawls (2005, 153).
Rawls (1987, 12; 2005, 150).
Rawls (2005, 29, n. 31).
Richardson (2014, 41).
Ibid.
Krishnan (2021).
Ibid., 11.
Ibid., 6.
Laden (2003, 373).
De Caro and Macarthur (2010, 3).
Rawls (2005, 103).
Rawls (1985, 238).
Laden (2014, 107).
Rawls (1985, 229); see also Rawls (1971, 508)
Rawls (1971, 50).
Chan (2005, 455).
Ibid.: 457–563.
Rawls (1985, 225).

References
Chan, H.M. Rawls’ Theory of Justice: A Naturalistic Evaluation. Journal of Medicine and Philosophy, 30,
(2005). 449–465.
De Caro, M. and Macarthur, D. Naturalism and Normativity. Columbia University Press. (2010).
Krishnan, N. John Rawls and Oxford Philosophy. Modern Intellectual History. (2021). doi: 10.1017/S14
79244320000451.
Laden, A. The House That Jack Built: Thirty Years of Reading Rawls. Ethics, 113, (2003). 367–390.
Laden, A. The Practice of Equality. In R. Forst (ed.), Justice, Democracy and the Right to Justification,
Bloomsbury. (2014). pp.103–126.
Rawls, J. A Theory of Justice. Harvard University Press. (1971).
Rawls, J. Justice as Fairness: Political not Metaphysical. Philosophy and Public Affairs, 14)(3), (1985).
223–251.
Rawls, J. On the Idea of an Overlapping Consensus. Oxford Journal of Legal Studies, 7(1), (1987). 1–25.
Rawls, J. Political Liberalism. Columbia University Press. (1993).
Rawls, J. A Theory of Justice: Revised Edition. Harvard University Press. (1999).
Rawls, J. Justice as Fairness: A Restatement. Harvard University Press. (2001).
Rawls, J. Political Liberalism: Expanded Edition. Columbia University Press. (2005).
Richardson, H.S. Avoidance, Method of. In J. Mandle and D.A. Reidy (eds.), The Cambridge Rawls
Lexicon. Cambridge University Press. (2014).

357

30
SCIENTIFIC NATURALISM
AND NORMATIVE EXPLANATION
Robert Audi

1 The historical and philosophical context
By the early Twentieth Century, naturalism had strong proponents on both sides of the
Atlantic. Roy Wood Sellars wrote, “We are all naturalists now”, though he quickly added that
naturalism “is less a philosophical system than a recognition of the impressive implications of the
physical and biological sciences”.1 Wilfrid Sellars, writing three decades later, went further.
Parodying Protagoras, he said, “[S]cience is the measure of all things, of what is that it is and of
what is not that it is not”.2 This extraordinary pronouncement represents the kind of view some
have called scientism, on which knowledge as well as what counts as real are to be determined by
scientific method understood quite narrowly.
Twentieth-Century naturalism may have arisen initially and certainly in part as a reaction
against supernaturalism. Naturalism stresses a nontheistic approach to understanding the world
and, typically, scientific ways of explaining events. Modern empiricism is also a factor: its
adherents see both meaning and knowledge as arising from our experience of the natural world
in a way that makes it easy to conclude that the truths about natural phenomena are the only
basic truths there are. Twentieth-Century positivism, though it develops this view, retains its
naturalism: for positivism, observation and scientific method are the only grounds of substantive
knowledge (knowledge of logical and analytic truths is not considered substantive on this
approach).
Early in the Twentieth Century, naturalism was often clarified by contrast with two major
philosophical views. First, naturalists opposed idealism – roughly, the theory that reality depends on, or is indeed constituted by, minds and their ideas. Second, and perhaps even more
important, naturalists opposed ethical views that posited intrinsic value as an irreducibly nonnatural phenomenon. The rightness of an action, on these views – Moore’s, for instance –
presupposes an irreducibly nonnatural property: goodness. On Moore’s view, pleasure has
intrinsic value, but is not equivalent to the intrinsically good.3
More recently, naturalism has been contrasted with Cartesian dualism and interpreted in
relation to the possibility of an ultimately physicalistic understanding of human behavior. We
are flesh and blood, neurons and synapses; and if the life of the mind is more than a powerless
shadow, it must occur in this biological network. Since at least the 1960s, naturalism has also
become a powerful force in epistemology. Knowers are seen not as meeting evaluative standards
358

DOI: 10.4324/9781351209472-35

Scientific naturalism

that enable us to justify beliefs, but as reliable instruments responding to experience of the
world: they reliably register truth rather as a good thermometer registers temperature.
In metaphysics, then, we have naturalism both as a reaction to supernaturalism and as a rejection of the kind of mind-body dualism we find in Descartes; in ethics, naturalism – at least in
realist, as opposed to noncognitivist versions of ethical naturalism – appears as a rejection of
irreducible notions of value and rightness; and in epistemology naturalism accounts for knowledge
and justification in terms of notions amenable to scientific treatment, particularly the concepts
central in physics and psychology. If there is a unifying conception of naturalism in all these
domains, it is at least not commonly made clear. In broad terms, we might conceive philosophical
naturalism as the view that nature is all there is and the only basic truths are truths of nature.4

2 Scientific naturalism
Many philosophers of science hold or presuppose some form of philosophical naturalism. But
are scientists as such committed to naturalism? Here we should scrupulously distinguish between philosophical naturalism and methodological naturalism. Methodological naturalism –
which, unlike the former, is not a metaphysical position – is, roughly, the view that causes and
explanations of natural phenomena lie in – or at least should be sought in – the natural world,
paradigmatically in terms of what meets at least three commonly endorsed criteria for a scientific
theory – testability, publicity and empiricality (these are taken to be important necessary
conditions, not sufficient conditions, though a case can be made that there is an interpretation
of them on which anything satisfying all three and deserving the name “theory” would be
scientific). Methodological naturalism is a commitment of many scientists. It requires conducting scientific inquiry in what might be called descriptive natural categories and with full
respect for testability and the experimental, public, predictive character of science; but it does
not imply a comprehensive ontology nor even deny the existence supernatural entities.

2.1 Naturalism and the drive for intellectual economy
It must be granted that naturalism as I have characterised it (and as often described) is no clearer
than the notion of the natural, but a standard assumption is that the physical is paradigmatically
natural and, to take the strongest contrast, that God, as widely understood, is supernatural. The
elements of scientific method I am emphasising are methodological, not ontological. As such,
they are neutral with respect to the possibility of supernatural agency, and methodological
naturalism neither affirms nor denies theism. Neither the testability nor the publicity nor the
empirical character of scientific hypotheses entails any substantive ontological conclusion, say
that only natural events can be causes or that there cannot be an infinite chain of causes.
It is, however, commonly and plausibly held that commitment to scientific method presupposes commitment to some form of Ockham’s Razor, understood as a principle calling for
intellectual economy. Does this principle favor philosophical naturalism over methodological
naturalism? The question is not easy to answer. On any interpretation, the Razor implies that
other things equal, the simpler of two competing views is preferable. Isn’t philosophical naturalism a simpler worldview than, say, a theistic one that is otherwise as much as possible like,
say, the scientific realism that a philosophical naturalist can countenance? Answering this requires some account of simplicity, an elusive notion I cannot explicate here.5 Still, if we assume
that the number of basic assumptions and irreducible kinds of entities posited by a worldview is
most important for determining the degree of its simplicity, then we might suppose that there is
some reason to think that the naturalistic worldview is simpler than a theistic view.
359

Robert Audi

2.2 Two conceptions of the Razor as requiring preference for the simpler
So far, I have spoken as if it were clear what the status of the Razor is.6 It is not. Suppose we
take the principle to state that, other things, the simpler of two competing hypotheses is
preferable. The principle is most widely considered epistemic, in the broad sense that it provides
an evidential standard relevant to assessing the candidacy of a belief, hypothesis or theory for
justification or knowledge or both. It also applies to acceptance as a propositional attitude or to
hypotheses as possible objects of acceptance.
May we (other things equal) take the simpler view to be better justified, on evidential
grounds? This is not obvious. What may be obvious is something easily conflated with the
principle conceived epistemically: namely, the quite different principle that, other things equal,
it is preferable to work with a simpler view in constructing a theory or in framing or retaining a
worldview. There is less to deal with – to explain, to connect with other elements and so forth.
That interpretation, however, yields a practical principle that could be true even if simplicity is
irrelevant to evidential status. Call the practical principle in question the principle of least effort.
It might also be called the principle of optimal efficiency to emphasise that it opposes expending
more effort than is needed for one’s governing purpose.
This principle of least effort is plausible on the a priori basis of understanding its content and
may even be self-evident. But a hypothesis that is easier to work with is not thereby better
evidenced, say in terms of confirmatory experiences or predictive power, than an otherwise
equally well confirmed alternative. Is there reason to think that from an external perspective,
simpler views are more likely to be true? Perhaps, for some plausible notion of simplicity, there
is, given common experience and the track record of scientific hypotheses. The intuitively
simpler hypotheses have apparently tended to be better confirmed (though this might be at best
hard to show, because of the difficulty of separating considerations of simplicity from those of
degree of confirmation). I do not see how to demonstrate this, though doubtless we might
entertain an evolutionary explanation. The preference for simplicity could have fitness value:
perhaps our general preference for simpler hypotheses has contributed to our survival as a
species, and perhaps we would not have survived, or at least survived and reached the level of
efficiency we have achieved, if the ways of nature were not better evidenced by simpler
hypotheses.

3 Ontological pluralism
So far, we have seen no decisive reason to think that a scientific habit of mind requires
commitment to philosophical naturalism, as opposed to an ontologically more pluralistic, and in
that way more liberal, worldview. For some highly scientific people, however, including those
who are theists, methodological naturalism is a possible position. We have also seen no decisive
reason to believe that, assuming Ockham’s Razor, theism is clearly less plausible than atheism
given. Even if it is less simple in some way, it does not follow that it is less well evidenced.
Ontological pluralism, at least of a kind affirming the reality of both natural and nonnatural
entities, seems a plausible position.

3.1 Abstract entities
To be maximally simple, a full-scale materialist ontology must apparently discountenance abstract entities. This is a prospect I cannot pursue in detail, but philosophical naturalists are in any
case more concerned with naturalising any reality with causal power (as supernatural beings are
360

Scientific naturalism

thought to have) than with naturalising the abstract, which is typically considered outside the
causal order. Abstract entities are a kind of annoyance to some naturalistic thinkers, since they
must be accounted for in a comprehensive ontology; but if they have no causal power, they at
least do not do anything in the natural world and cannot compete with scientific explanations
or undermine scientific predictions.
Monistic naturalism – which countenances only one basic ontological category, such as the
physical – seems quite implausible. Suppose all natural entities are physical. Can this view even
account for itself, as a bearer of truth value? If, for instance, what is true or false is a proposition
(as the view apparently is), or even a sentence type as opposed to a sentence token, then there
are some abstract entities.7 Universals and numbers present other challenges, but there, too, I
must be content simply to point out obstacles to full-scale materialism rather than argue for
commitment to pluralism that includes the abstract. Suppose, however, that there are properties
that are not even descriptive – and in that minimal sense natural – to begin with. Then there
would seem to be a twofold problem for a thoroughgoing naturalism: first, to eliminate the
normative; secondly, to eliminate the abstract. Here, I will concentrate only on the first problem and then only in the general way appropriate to a broad essay of this kind.

3.2 The normative domain
Supposing there are normative properties, can reductive naturalists accommodate them? The
Cornell Realists have attempted to give a kind of reduction here, but I have elsewhere argued
that although we might regard them as providing reasons to think moral explanations can be
naturalised, they have not shown that moral properties are naturalisable.8 Briefly, the idea is that
when a moral property, such as the injustice of a regime, is cited to explain behavior, say a
revolt against that regime, the (empirical) explanatory work is done by some grounding
property that is natural. For instance, wherever injustice can be justifiedly and correctly cited as
explaining a revolt, the person giving the explanation must be justified and correct in taking to
be present such grounds of injustice as police brutality, arbitrary curfews and poverty-inducing
taxation; and these are natural phenomena that can explain a revolt in terms of social-scientific
generalities that do not require normative concepts in their content. Moreover, if we did not
take injustice to imply such plainly causative elements, we would not see how it explains the
revolt.
There is, however, another route to possible naturalisation of the normative. Suppose for the
sake of argument that (1) we can formulate a disjunction of all the grounds of moral obligation
and (2) these grounds are natural facts (as is plausible). If W.D. Ross’s famous list is as comprehensive as he hoped, we could say that an act is obligatory if and only if (in rough terms) it is
either a doing of justice (in the descriptive sense of treating persons equally), an avoidance of
injury, a promise-keeping, an avoidance of lying, a beneficent deed, an act of selfimprovement, a reparation for injury, or an expression of gratitude. One could do the same
kind of thing in epistemology, for instance holding that a belief is justified if and only if is
grounded in either perception, or memory, or introspection, or reason, or testimony. Even if
these biconditionals are not analytic or any kind of conceptual truth,9 might they suffice for
reduction of the normative properties in question to descriptive properties?

3.3 Property identity as an element in ontological reduction
To answer this, we must at least assume that the equivalence is necessary. But must necessarily
equivalent predicates express the same properties? This is doubtful. Consider the mathematical
361

Robert Audi

property of being even. Is this the same property as that of being a member of the series 2, 4, 6,
8, etc.? The property of being even (being an even integer) is equivalent to the property of
being divisible by 2 without remainder; but is that property identical with, or even equivalent
to, the property of being a member of the set 2, 4, 6…? The members of the set in question are
thus divisible; but is being a member of it the same property as being divisible by 2 without
remainder? The former is a set-theoretical property, the latter a numerical (arithmetical) one,
and they seem different in kind.
This example and others, such as those involving purported reductions of normative concepts or properties, bring out an important point. The relevant sense of “reduction” is no
clearer than our understanding of when two predicates express the same property (or, given any
properties, F and G, under what conditions F = G). This remains a vexed question. Do two
predicates express the same property provided they are (1) synonymous, (2) analytically
equivalent, (3) conceptually equivalent (if that is not equivalent to one of the other categories),
(4) logically equivalent, (5) metaphysically equivalent, (6) synthetically a priori equivalent, (7)
explanationally equivalent, (8) nomically equivalent, (9) causally equivalent or (10) something
else again? Even if we have an answer to this question, the ten notions in question remain
controversial, and they all need analysis.
Whatever our criteria for property identity, it is far from clear that there are disjunctive
properties at all, as opposed to disjunctions of predicates that each express a property. If there
are disjunctive properties, we need an explanation of why specifying that a shape is either
circular or elliptical is not an answer to the question “What are its shape properties?” or,
especially, “What is its shape?” One could argue that the needed explanation is simply pragmatic, say that in asking for a shape we presuppose in normal contexts that a non-disjunctive
specification will be given.10
None of these points detracts from the significance of the disjunctively specified grounds of
obligation cited above. They yield multiple paths from certain descriptive notions to the crucial
notion of obligatory action. Similarly for the grounds of justification, which might be argued to
be disjunctively specifiable in terms of, say, the classically recognised basic sources of justification: perception, memory, introspective consciousness and intuitive reason. But this point
about the grounding of the normative shows – assuming the grounds in question are genuinely
constitutive – only that the normative properties (and the concepts of them) are anchored in the
natural world, not that those properties are genuinely natural or even equivalent to disjunctive
properties, if there are such properties.
This is a good place to note that what has been said of properties may well not apply to
concepts. If C1 and C2 are concepts expressed in different terms, the two conceptual description are not taken to express the same concept unless they are synonymous or have means
so closely interconnected that it is plausible to construe the thought that C1 as equivalent to the
thought that C2. This does not require indicating something about what the relevant things are
“like”, especially where this is related to explanation, prediction, connections to other things,
or all of these (and perhaps other elements). In this very fine-grained way of individuating
them, concepts are more like propositions than like properties.

4 Nonreductive naturalism
The view that nonnatural properties are strongly supervenient on natural ones is sometimes
called nonreductive naturalism, and the name is better understandable if the supervenience is
paired with a grounding relation. It normally is paired with one, though in the mental-physical
case the grounding relation is empirical and presumably nomic, whereas in the normative case it
362

Scientific naturalism

is (on my view and for many others) a priori and metaphysically rather than nomically necessary. Such naturalism is, however, clearly a liberal naturalism: although it may embody a denial
of substance dualism – a view that liberal naturalists reject – it is consistent with property
dualism, since even the strong grounding relation obtaining in the normative cases does not
imply the possibility (or impossibility) of reduction of the normative to the natural.
It should be added that so-called non-reductive materialism as so far described is even
consistent with a version of Cartesian dualism. If mental properties are grounded in (but not
reducible to) physical ones – the kind of natural property important for naturalism in the
philosophy of mind – then the mental properties can be in a different category, perhaps even
being properties of a Cartesian mental substance, though how properties of that kind might be
grounded – metaphysically as opposed to causally – in physical properties is at best puzzling.
There need be no mystery about pain’s being caused by tissue damage, but it is unclear how
pain can be what it essentially is in virtue of facts about tissue damage.
To be sure, from the perspective of non-reductive materialism, mental-physical causal relations may be less easy to understand given Cartesian substance dualism than on property
dualism, but nothing in the concept of causation precludes cross-categorial causal relations. I
would stress, however, that nonreductive naturalists would generally restrict mental properties
to properties of embodied beings and would not countenance Cartesian mental substances. One
metaphysical problem, of course, is how to understand mental properties as at once not reducible to physical ones yet necessarily belonging to physically embodied beings.
Even if mental properties are not themselves physical, psychological science can in principle
proceed as it would if they were physical. If, for instance, there are lawlike connections between, on the one hand, sets of desires and beliefs and, on the other hand, actions, these can be
studied either (1) psychologically in relation to objective criteria for applying the relevant
psychological and behavioral predicates or (2) neurophysiologically in terms of the grounding
properties and their role in producing either the psychological states or the physical movements
those states explain, or both. For the normative counterpart of non-reductive physicalism, the
point is different, since some of the generalisations in question, say that failure of the prosperous
to do any beneficent deeds is wrong, are not empirical. But some of the generalisations are
empirical – say that injustice breeds anger – even if the empirical work, and hence the task for
social-scientific appraisal, is at the level of the grounding properties, say of unequal treatment
and angry behavior. If ethics is not itself a science, there is certainly a host of scientific questions
about the conditions for its realisation in human conduct, about the risks to its grip on human
motivation, and about the identification of those for whom it yields guiding standards – moral
individuals – and those for whom it does not.
It is the metaphysician, not the scientist, who tends to feel a need to reduce non-natural
properties to natural ones. There is no reason to object to the quest and much to be learned
from appraising its degree of success. But, as a philosophical framework appropriate to a scientific habit of mind and supportive of the progress of science, methodological naturalism is as
strong a naturalistic position as required.

5 Moral realism in the intuitionist tradition
My strategy in metaethics – now called “the New Intuitionism”11 – is to develop a moderately
rationalist intuitionism which anchors moral properties in natural (descriptive) ones, as with the
list of grounding properties from Ross (to which I have added two others12). This is a realist
view, but does not require reducing normative properties to descriptive ones. Here I want
simply to indicate that even if one could show that moral properties are natural, the major
363

Robert Audi

elements in my theory would be sustainable: the theory accords well with, but does not depend
on, a nonnaturalist account of normative properties.
As is well known, G.E. Moore held that moral properties are non-natural. My intuitionism
does not presuppose this. The most important point here is that the epistemic status I attribute
to intuitive prima facie duty principles – being necessary and a priori – does not require that the
properties figuring in those principles be nonnatural.13 Nor is that incompatible with or entailed
by the thesis that moral properties are, as Moore, Ross, and other intuitionists also hold,
consequential on natural ones. This neutrality toward Moorean non-naturalism is an advantage
in making a version of intuitionism eligible for endorsement by many contemporary moderate
naturalists, certainly including liberal naturalists.
If it turns out, then, that moral and other normative properties are natural, my view does not
require either changing my position on what properties ground them – certain natural properties – or moving to anti-realism, or altering the relation I have described above between ethics
and science. Nor need the moral epistemology be revised. The relation between moral and
non-moral properties and concepts can have the character I have maintained whether or not
those properties are natural. The metaphysical change from neutrality toward naturalism to
endorsing some version of it would leave both the normative content and the epistemology of
the view essentially unchanged.

6 Normative explanation
The notion of the normative is often taken to be clear, but I do not believe that it is.14 Here I
will say only that by “normative explanation” I mean roughly presumptive explanation by appeal to
a normative concept, fact, or property, where paradigms are moral or epistemic in the way justification is. That an act is (morally) wrong is a clear case or a normative proposition expressing, if
true, a normative fact; so is the proposition that there is reason not to kill persons, and the
related proposition that there is justification to believe global warming has occurred. I have
already cited attempts to explain such phenomena as revolts by appeal to the normative notion
of injustice. Let me go on to cite a presumptive case of moral perception.
Can we see wrongdoing? You see a shooting. The man shot is fleeing and you see and hear
several shots in succession with more than one apparently hitting him in the back. Do you simply
see a physical act and infer rather than see that a (prima facie) wrong was done? How can you see
an instantiation of a moral property if it is not itself in the causal order, much less in the physical
order? Consider injustice as a moral phenomenon. Is it ever observable, in the most basic sense, a
sense that goes with perceptual properties, such as movement and shape? Is seeing injustice, for
example, observational in the sense corresponding to the perceptual properties of color and shape?
Or is such moral perception equivalent to seeing – in a distinctive way that is at least not narrowly
observational – a set of base properties for injustice, such as a patently unequal distribution of
needed food to starving children, where these properties are seen in a way that makes it obvious,
upon seeing them, that an injustice is done? The second alternative points in the right direction.

6.1 Normative properties as consequential on natural ones
In asking about the relation between a moral perception of injustice and seeing the relevant base
properties, that is, the properties on which injustice is consequential, I assume something widely
held: that actions and other bearers of moral properties do not have those properties brutely, but
on the basis of (consequentially on) having “descriptive” properties. Consequential properties
may also be called grounded or resultant, terms that also indicate that a thing possesses the
364

Scientific naturalism

“higher-level” properties because it possesses the base properties. Acts are not simply wrong, in
the way they can be simply hand-raisings (though in certain underlying ways even such basic
acts are not simple). It is essential to the wrongness of a wrong act that it be wrong on the basis of
being a lie, or because it is a promise-breaking, or as a stabbing, and so forth. Similarly, a person
is not simply good, but good on the basis of, or because of, having good governing motives
together with beliefs appropriate to guide one toward constructive ends.15
If, however, we see moral properties on the basis of our seeing nonmoral properties, philosophers will ask whether one ever really sees a moral phenomenon, such as an injustice. One
might think that the phenomenal elements in perception properly so called must be sensory in
the representational way that characterises paradigms of seeing and some of the exercises of the
other senses among the five ordinary senses. But why should we expect perceiving injustice,
which is not a basic kind of perception for us and has a normative, non-sensory, moral phenomenon as object, to be just like perceptions of color, shape, flavor, or sound, which are
physical or in any case sensory, non-normative, and, in typical cases, basic for us? Why should
there not be, for instance, a phenomenal sense of injustice that is not “pictorial” in the way
exemplified by visual impressions of trees or paintings? Here we might consider non-visual
perception. Where a moral perception is auditory, as with hearing a lie, or tactual, as with
feeling one’s face slapped, we are not tempted to expect it to be pictorial, at least in the way
visual experience of many kinds of things may be taken to be.

6.2 The causal element in moral perception
These points do not answer another question raised by positing moral perception: if moral
properties are not in the causal order (even in the limited sense in which, say chemical
properties are), how is moral perception causal? The short answer is that the perceptual experience in, for example, seeing a wrong as such – seeing morally, if you like – is caused by the
instantiation of the base properties, just as the experience in seeing a round disk is caused by the
instantiation of some determinate of roundness; and the phenomenology of the moral experience is an integration between the sensory representation of the base properties and the
perceiver’s moral sensibility. That, of course, is connected with past experience and, normally,
some kind of moral education. Recall the case of injustice as a presumptive cause. Whereas
Cornell realists would account for moral perception here by taking injustice to be a natural
property, on my view it need not be natural but plays a role in the causal order by virtue of
causal work done by its grounding properties in relation to the sensibility of the perceiver.
More specifically, moral perception centers on a phenomenal representation constituted by a
(richer) perceptual response to injustice. The sense of injustice, then, a kind of impression, as
based on, and as phenomenally integrated with, a suitable ordinary perception of the properties
on which injustice is consequential – grounded, in a main use of that term – might serve as the
experiential element in moral perception. Call this an integration theory of moral perception.
One might wonder at this point if there are any natural properties that, like moral properties,
have their causal power only consequentially upon that of related natural properties. I suggest
that shape is one. It is clearly a natural property, yet it is not shape but what grounds a particular
shape, such as roundness, that causally explains such phenomena as its rolling behavior. The
moral property of wronging is similarly related to stabbing, which grounds it where the
wronging perceived is consequential on stabbing. Real properties need not, then, be in the
same metaphysical category as those that ground them.
What kind of naturalism does this picture allow? Reduction of moral to natural properties is
not required. For thoroughgoing philosophical naturalists, although it is not an a priori truth
365

Robert Audi

that mental properties depend on physical ones, the physical has a kind of metaphysical sovereignty. But what of scientific naturalism? Does the scientific study of morality require
reduction? No; reliable connections between natural phenomena can provide necessary or
sufficient conditions for the instantiation of moral properties, and that can suffice for much
theorising, say about moral education or, on the negative side, causes of dishonesty. What, then,
of the unity of science, which apart from reduction must countenance unreduced properties
(these need not in the end be non-natural)? Even if philosophical naturalism represents a
metaphysical idea of unity, it is only a kind of explanatory unity that is needed for a scientific
worldview on which empirical phenomena are explainable by facts in the causal order.

7 Normativity and the scientific habit of mind
It has been widely believed or at least presupposed that if normative properties are not reducible, then normative judgments cannot be objective. We have seen reasons to doubt reducibility but, more important, we may deny the conditional. Let me explain.

7.1 Objectivity
Objectivity comes in many forms. Ontologically, the objective is (very roughly) the real as
opposed to the imaginary. Here I will simply suggest that an epistemic conception is more
useful for explaining how a scientific habit of mind may be sustained without philosophical, as
opposed to scientific, naturalism. We might begin by saying that the objectivity of judgments
may be taken as basic, in such a way that a theory is objective if and only if its truth is appraisable
in terms of objective judgments concerning its truth or success and that a method is objective if
and only if it is intersubjectively applicable, in the sense that its steps are describable empirically,
their accomplishment is appraisable in terms of objective judgment, and their completion and
results within the end aimed at is likewise appraisable. This leaves open how objective judgment
figures in characterising an objective person, but clearly such a person is guided by such
judgments in making decisions in empirical and moral matters. As all this suggests, objectivity
admits of degrees.
If we do take objective judgments as a fruitful starting point for understanding objectivity in
general, it is important to indicate how we might understand those. Intersubjectivity seems
crucial: they are a kind such that any person competent in the relevant subject can appraise
them. To be sure, the very notion of competence is difficult to capture. As with many notions,
we must appeal to intuition and to publicly available information, some of it sociological, as
with identifying competent scientists partly by institutional criteria. Here the demarcation
problem – how to distinguish science from other cognitive disciplines – rears its head, but in the
space I have I can say only that we have such paradigms as Galileo, Newton and Einstein to cite,
and we can appeal to publicity, testability and (for specifically scientific objectivity) empiricality
as necessary conditions for objective judgment.
If we now ask whether normative judgments – and here it suffices to consider just moral cases
– may be seen as objective even if normative properties are irreducible – this essay supports a
positive answer. I have shown in effect how moral properties, because they are grounded in natural
ones, (1) are under the control of natural properties and (2) knowable by appeal to them. (1) is
ontological and (2) epistemological. Unfortunately, neither point yields the conclusion that every
moral question is decidable by an objective method, in the sense that a correct answer is guaranteed
by a finite number of objectively descriptive steps. But this does not apply in scientific cases either.
There too, amenability to an objective method does not entail decidability by it.
366

Scientific naturalism

I am not arguing here that ethics is a science, but only that it is not unscientific in the way
subjective judgments may be, and that it can be objective. As in the scientific case, this allows for
epistemic indeterminacy. That, unfortunately, makes room for subjectivity to intrude, as, for
instance, with questions that suffer from ineliminable vagueness, for example the question when,
in normal human development, personhood, as opposed to mere membership in the human
species – hence “human life” – comes. There is room, then, for a position that represents scientific naturalism as characterised here, and leaves open (1) realism and even (2) rationalism is
ethics, as well as, in philosophy of science, (3) scientific realism as opposed to instrumentalism.
Moral skeptics will object that in the scientific case, epistemic indeterminacy does not
undermine objectivity but does in the moral case. It may force us to suspend judgment about
which of two competing theories is preferable but will leave open the possibility that future
inquiry will settle the matter. Thus, those with a scientific habit of mind must suspend judgment. By contrast, in ethics, there are irresoluble deadlocks besetting such questions as personhood – in the sense entailing full human rights – and where incommensurable evidences
clash, say when we have reasons of protection of the innocent for toppling a dictator and,
opposing them, reasons of non-injury to others – including some innocent noncombatants.
And sometimes practical considerations force moral decisions whereas, in the scientific case,
practical decisions based on one or another contested theory can be delayed. The response to
this apparent asymmetry is that further inquiry may be in practice impossible in some scientific
cases; and in the ethical case as well, irresolubility may not be inferred at any point in a dispute,
and inquiry may go on indefinitely, as it does concerning personhood.

7.2 Religion, politics, and scientific objectivity
If ethics can be squared with the scientific habit of mind and even with scientific naturalism,
what of its relation to theology? In theology we find neither a priori grounds for major principles
parallel to those in ethics nor the publicity, testability and empiricality crucial for objective
judgments in the scientific case. I doubt that the latter holds for all theological claims,16 but I
believe that the point of view of practical reason and indeed of political philosophy, supports a
principle that is consistent with the scientific habit of mind and can also protect the religious as
well as the nonreligious from coercion on the basis of religion or theological claims. It what I
call the principle of secular rationale:
The principle of secular rationale (or natural reason): Citizens in a democracy have a prima
facie obligation not to advocate or support any law or public policy that restricts
human conduct, unless they have, and are willing to offer, adequate secular reason17
for this advocacy or support (e.g. for a vote).18
One might think that those with a scientific habit of mind – particularly if they have a welldeveloped ethical position – can never appeal to religious considerations as important in lawmaking. I doubt that. But they may certainly appeal to them in explaining some aspects of their
views and of their own conduct, such as religious observances in their own tradition or beneficent deeds that many would consider unwisely self-sacrificial. Returning to the difficult matter
of personhood, some scientifically oriented people might forthrightly admit that, on theological
grounds, they accept an ensoulment view of personhood; but they will very likely not think
that there is a scientific, or even an objective, method for determining the truth of this view.
Moreover, any reasonable person will see that people with a different theology could, given
sufficient political power, make laws on the basis of that theology.
367

Robert Audi

It is clearly possible, then, that the secular rationale principle yields a way to avoid the kinds
of often domineering conduct whose spectre is a major source of motivation for philosophical
naturalism. It leaves freedom to follow nonobjective criteria for decisions in one’s personal life,
but calls for objective ways of dealing with issues concerning law-making for the interpersonal
behavior of citizens whose radically different views are not objectively appraisable views. This is
not to say that a liberal naturalism is compatible with any particular theology; but if it can
countenance a conception of human persons who, though embodied, may have some nonphysical properties, it leaves room for a related conception of divinity.
____
Philosophical naturalism can take many forms. In its domain-specific forms, it is more plausible
in some realms of inquiry than in others. Naturalisation in the philosophy of mind, for instance,
seems less difficult to defend than full-scale philosophical naturalism requiring a nominalistic
account of all apparent presuppositions of the existence of abstract entities. Moreover, even
within a single field, such as epistemology, some core notions, say knowledge, seem more
amenable to naturalisation than others, such as justification. In ethics, the project of naturalising
moral properties does not preclude naturalising moral explanations or countenancing moral
perception as a genuinely causal response to objective grounds for their attribution. In its strong
forms, naturalisation can be rationally resisted without extravagant metaphysical or epistemological claims. Moreover – and this is a more important conclusion for this paper – scientific
rigor does not require a commitment to comprehensive naturalism, nor is any stronger naturalism than a methodological version required for cultivation and maintenance of a scientific
habit of mind. It appears, then, that neither science nor ethics nor logic and mathematics, nor
philosophy itself, requires endorsing a full-scale philosophical naturalism. A liberal naturalism,
by contrast, may embrace methodological naturalism in science while countenancing both
abstract entities and also properties that, though mental or normative or both, may be grounded
in the perceptible realms of experience that are universally agreed to be natural.19

Notes
1 Quoted by Hilary Kornblith in Kornblith (1994, 50). Sellars was of course overlooking G.E. Moore
and other nonnaturalists in early analytic philosophy. Cf. Barry Stroud’s comment,
‘Naturalism’ seems to me…rather like ‘World Peace’. Almost everyone swears allegiance to it…
But disputes can still break out…And, like world peace, once you start specifying concretely
exactly what it involves…it becomes increasingly difficult to reach and to sustain a consistent
and exclusive ‘naturalism’. See “The Charm of Naturalism”, in De Caro and Macarthur
(2010, 22).
2 Wilfrid Sellars (1963, 173).
3 For a detailed examination of Moore’s metaphysics of value, see Panayot Butchvarov (1989). Arguably,
though a kind of physicalism often associated with scientific realism is a pervasive element in
twentieth-century naturalism and at least equally at present, anti-realism in metaphysics, unlikely the
thesis of the irreducibility of the normative, is compatible with naturalism.
4 For a broader view that constitutes a philosophical naturalism, see the “liberal naturalism” set forth by
Mario De Caro in De Caro and Macarthur (2004) and in De Caro and Voltolini (2010).
5 Richard Swinburne argues in detail that the theistic view is simpler. For an indication of his view of
simplicity (see Swinburne 2004, 53–66, 70–71 and 179–179; Swinburne 2001).

368

Scientific naturalism
6 For a brief discussion of how to interpret the razor principle and its possible role in evaluating design
arguments, see Robin Collins (Collins, 2006).
7 For an excellent but too little studied paper on this issue, see Richard Cartwright (1962). Among its
points is the important one that sentences and propositions (or any bearer of truth value) differ in their
arithmetic. The number of things said is determined differently than the number of sentences in which
it may be said. If two people who, respectively, speak only English and Polish, say in their own
languages that 7 + 5 = 12, there is one thing said and one truth, but there are two sentences.
8 I appraise one version of Cornell Realism, and sketch some conceptions of moral explanation in
Audi (1993).
9 As suggested in the text, if we had a complete list of the things that constitute a person’s obligation, we
might say that to be obligated to A is equivalent to satisfying one or more of the descriptions.
Depending on the character of the list, we might consider this a (disjunctive) analysis. But one might
also insist on some unifying element, such as the notion of respect for persons, and go on to argue that
here such an element must be normative. More is said below about the possibility of a disjunctive
analysis.
10 To be sure, there is a sense in which for any true proposition, there is a possible question to which it is
a correct answer. My point here is that if the question is what properties a thing has or what properties
of it explain something, it is at best atypical for a plausible answer to cite something that seems a
genuinely disjunctive property. The issue is large, however, and what is said here is meant only to
suggest problems that prima facie favor denying that predicative disjunctions of the kind in question
express disjunctive properties.
11 See Hernandez (2011) for papers articulating and critically appraising this view.
12 See Chapter 5 of Audi (2004) for a description of the added grounding properties as well as refinements
and corrections of Ross’s list of such properties.
13 There is even a way for my view to be adapted to noncognitivism, as I noted in Audi (2004, 151). But
noncognitivism is not, to be sure, a natural option for an intuitionist.
14 This is argued in Audi (2013).
15 That moral properties are consequential is a view articulated in Moore (1903) and Ross (1930, esp. Chap.
2). I develop it further in Chapter 2 of Audi (2004). I here presuppose that certain properties, such as, on
the negative side, killing and, on the positive side, promising are a priori grounds of moral properties, but
my theory of the nature of moral properties does not require a particular list of such grounds.
16 This matter is discussed in detail in Audi (2009).
17 Judgment is required both to determine when a reason is secular – roughly, not evidentially dependent
on theology or religion – and to ascertain when a reason is adequate.
18 See Audi (2011, 65–66).
19 An earlier version of part of this paper was given at the Jagiellonian University, Roma III, and the
University of Pittsburgh. I benefited much from the three sets of discussions. The paper also reflects much
that I have learned from discussions with colleagues, friends and audiences, and it draws on previous work,
most Audi (2000), Audi (2012), and Chapter 10 of Audi (2011). Thanks to Nicholas Rescher for helpful
critical responses and to Mario De Caro for helpful comments on the penultimate draft.

References
Audi, R. Ethical Naturalism and the Explanatory Power of Moral Concepts. In S. Wagner and R. Warner
(eds.), Naturalism: A Critical Appraisal, pp. 95–115. University of Notre Dame Press. (1993).
Audi, R. Philosophical Naturalism at the Turn of the Century. Journal of Philosophical Research, 25,
(2000). 27–45.
Audi, R. The Good in the Right: A Theory of Intuition and Intrinsic Value. Princeton University Press. (2004).
Audi, R. Religion and the Politics of Science: Can Evolutionary Biology Be Religiously Neutral?
Philosophy and Social Criticism, 35(1–2), (2009). 23–50.
Audi, R. Rationality and Religious Commitment. Clarendon Press. (2011).
Audi, R. Can Normativity Be Naturalized? In S. Nuccetelli and G. Shea (eds.), Ethical Naturalism: Current
Debates, pp. 169–193. Cambridge University Press. (2012).
Audi, R. Knowledge, Justification, and the Normativity of Epistemology. Res Philosophica, 90(2), (2013).
125–145.
Butchvarov, P. Skepticism in Ethics. Indiana University Press. (1989).

369

Robert Audi
Cartwright, R. Propositions. In R.J. Butler (ed.), Analytical Philosophy, pp. 81–103. Blackwell. (1962).
Collins, R. The Design Argument. In A. Corradini, S. Galvan, and E. Jonathan Lowe (eds.), Analytic
Philosophy Without Naturalism, pp. 140–152. Routledge. (2006).
De Caro, M. and Macarthur, D. (eds.) Naturalism in Question. Harvard University Press. (2004)
De Caro, M. and Macarthur, D. (eds.) Naturalism and Normativity. Columbia University Press. (2010)
De Caro, M. and Voltolini, A. Is Liberal Naturalism Possible? In M. De Caro and D. Macarthur (eds.),
Naturalism and Normativity, pp. 69–86. Columbia University Press. (2010).
Hernandez, J.G. (ed.) The New Intuitionism. Continuum. (2011).
Kornblith, H. Naturalism: Both Metaphysical and Epistemological. Midwest Studies in Philosophy, XIX(4),
(1994). 50.
Moore, G.E. Principia Ethica. Cambridge University Press. (1903).
Ross, W.D. The Right and the Good. Oxford University Press. (1930).
Sellars, W. Empiricism and the Philosophy of Mind. In Science, Perception and Reality, p. 173. Routledge
and Kegan Paul. (1963). (originally published in Minnesota Studies in the Philosophy of Science
Vol 1, 1956).
Swinburne, R. The Existence of God, 2nd edition. Oxford University Press. (2004).
Swinburne, R. Epistemic Justification. Oxford University Press. (2001).

370

31
SCIENTISM AND
LIBERAL NATURALISM
Massimo Pigliucci

1 Introduction
Liberal naturalism as understood within the context of this volume is an approach that attempts
to strike a reasonable balance between a philosophy that is simply a handmaiden to the natural
sciences and one that rejects them.1 While relatively new on the modern philosophical stage,
liberal naturalism builds on a long tradition in philosophy, one that loosely connects Aristotle,
Spinoza, Hume, Kant and Quine, among others, without necessarily agreeing in full with any
of the views put forth by those thinkers.2 As such, liberal naturalism squarely puts itself at odds
with the emerging phenomenon of scientism, to which this chapter is dedicated.
Here I will (1) discuss what scientism is, (2) provide a few examples of it, (3) explore some
excesses on the other side of the debate, where “scientism” is used as a generic (and unwarranted) trump card to defend irrational or antirational views; (4) connect scientism to our
conceptions of what science itself is and then (5) adopt a more organic view of the relationship
between science and the humanities, with particular reference to philosophy, based on the
much underappreciated framework proposed in mid-twentieth century by Wilfrid Sellars and
his “stereoscopic” view of what he called the scientific and the manifest images of the world. I
will suggest that, just as Sellars himself envisioned (and contra some of his own disciples, both
those of the so-called “right wing” and those of the so-called “left-wing”), it is a major and
crucial task of philosophy to continually monitor and negotiate our conceptualisation of the
relationship between the two images.
In the end, a basic position underlying my approach is that science is not in the business of
arriving at universal truths from a god-eye perspective – because that is simply not possible for
the limited abilities of brains that evolved by natural selection – but rather to aid human beings
in formulating better and better understanding of the world they live in and have to navigate.
Such understanding, while crucially dependent on the natural sciences, cannot and should not
be limited to their particular methods and perspective.

2 What is this scientism thing, anyway?
The quintessential bad boy of philosophy of science famously wrote:

DOI: 10.4324/9781351209472-36

371

Massimo Pigliucci

Science can stand on its own feet and does not need any help from rationalists, secular
humanists, Marxists and similar religious movements…Science must be protected
from ideologies; and societies, especially democratic societies, must be protected from
science…In a democracy scientific institutions, research programmes, and suggestions
must therefore be subjected to public control, there must be a separation of state and
science just as there is a separation between state and religious institutions, and science
should be taught as one view among many and not as the one and only road to truth
and reality.3
When I first encountered this quote from Feyerabend, I was a young scientist with little or no
appreciation of philosophy in general, and philosophy of science in particular. My reaction was of
disgust, anger, even. It was also highly unwise. While still today I am skeptical of the more radical
views expressed by Feyerabend, I have come to appreciate that he had a point. A major one, in fact.4
Even though I didn’t realise it at the time, Feyerabend was one of the first and most vocal
critics of scientism, though so far as I know he did not actually use the term, and he was actually
preceded by people who would have otherwise not agreed with him on a number of other
issues, like philosopher Karl Popper and economist Friedrich Hayek.5
Feyerabend is often interpreted, especially by scientists, as being antiscience, but this is a
gross simplification that is betrayed by a less than casual reading of what he actually wrote.6
Despite his notoriously provocative language, his concern wasn’t with science itself (“science
can stand on its own feet”), but rather with more or less socially nefarious ideological uses of
science (“does not need any help from…religious movements”). While I wouldn’t go so far as
advocating a separation between state and science analogous to that between science and religion (because then science funding would be entirely in private hands, and likely not for the
benefit of the public), the dangers that Feyerabend envisioned are real and have manifested
themselves plenty of time (e.g. in the infamous eugenic movement in the United States during
the early part of the twentieth century, as well as in more recent incarnations of the same
pernicious ideological attitude).7
But what is scientism, in the first place? A number of definitions have been put forth, for
instance:
A slavish imitation of the method and language of science.8
The aping of what is widely mistaken for the method of science.9
[The belief that] science, and only science, describes the world as it is in itself, independent of perspective.10
Scientism is a matter of putting too high a value on natural science in comparison with
other branches of learning or culture.11
The view that the characteristic inductive methods of the natural sciences are the only
source of genuine factual knowledge and, in particular, that they alone can yield true
knowledge about man and society.12
The best way to understand the charge of scientism is as a kind of logical fallacy
involving improper usage of science or scientific claims.13
The belief that the methods of natural science, or the categories and things recognized in
natural science, form the only proper elements in any philosophical or other inquiry.14
While no complex concept – including not just scientism, but in fact science itself, as we shall
see shortly – admits of clear-cut definitions based on a small set of necessary and jointly sufficient conditions, the above statements clearly identify a kind of Wittgensteinian cluster or
372

Scientism and liberal naturalism

family resemblance concept. What these authors have in mind is some sort of trespassing, where
“science” is said to be competent to generate knowledge and adjudicate questions, essentially
no matter what the specific field of scholarship. Which is why I consider discussions about
scientism as a particular instantiation of Popper’s famous demarcation problem, the quest for
what separates science from pseudoscience and nonscience.15
While Larry Laudan famously argued that the demarcation problem is a waste of philosophical energy, a number of authors have successfully overcome his objections and put back the
issue on the radar of philosophers of science.16 The bottom line is that Laudan was looking for a
standard definition in terms of necessary and sufficient conditions, which he correctly concluded cannot be arrived at. But fuzzy boundaries are all we need, and we better be able to
discuss them, on penalty of the very word “science” (not just scientism and pseudoscience)
becoming vacuous and useless. Yet, per Wittgenstein we do play a language game in which
“science”, “pseudoscience” and “scientism” refer to identifiable, if imprecise, notions and our
conceptual arsenal would be severely depleted if we attempted to do without them.17 This is
why I argue in this essay that one cannot meaningfully talk about scientism without also
shedding some light on what science itself is, a topic to which I will return later on.
Despite the overwhelmingly negative characterisation of scientism, there are scientists and
philosophers who have, on the contrary, taken it up as a badge of honor, openly and brazenly
defending scientism as a viable philosophical position. Alex Rosenberg, for instance, forcefully
argues in favor of an extreme version of materialistic reductionism according to which science is
the only human endeavor capable of ascertaining facts about the world.18 And he defines “facts”
in such a broad way as to include, for instance, mathematical statements, not to mention
anything and everything that is the object of study of not just the natural sciences, but also the
social sciences (e.g. economics) and the humanities (including, first and foremost, history and
philosophy). This strategy of colonising other fields by simply expanding the definition of terms
like “science” and “facts” is typical, and indeed a dead giveaway of, the scientistic attitude.
Before we move to briefly examining some illustrative examples of scientism, in order to
better fix our ideas and ground them in practical applications, let me point out that of course the
scientistic attitude is really nothing new under the sun. While this is not the place for a detailed
history of the concept, I would be remiss in not mentioning two of the obvious forebears:
David Hume (rather ironically, since he is claimed also as part of the intellectual lineage leading
to liberal naturalism) and, of course, logical positivism. Hume famously wrote:
If we take in our hand any volume; of divinity or school metaphysics, for instance; let
us ask, Does it contain any abstract reasoning concerning quantity or number? No.
Does it contain any experimental reasoning concerning matter of fact and existence?
No. Commit it then to the flames: for it can contain nothing but sophistry and
illusion.19
This is also known as Hume’s fork, and we ought to be grateful that nobody applied it to
Hume’s Enquiry itself, since it contains very little “experimental reasoning” and pretty much no
“abstract reasoning concerning quantity or number”. While Hume’s rather radical statement
can certainly be understood within the context of what he was reacting against (Scholastic
metaphysics), Rosenberg & Co. have no such excuse 270 years later.
As for logical positivism, here is A.J. Ayer:
The criterion which we use to test the genuineness of apparent statements of fact is the
criterion of verifyability. We say that a sentence is factually significant to any given
373

Massimo Pigliucci

person, if, and only if, he knows how to verify the proposition which it purports to
express – that is, if he knows what observations would lead him, under certain
conditions, to accept the proposition as true, or reject it as being false.20
And we all know that in philosophy of science logical positivism, despite its historical importance, went the way of the dodo. So why resurrect it now?

3 A taste of scientism
Who, exactly, engages in scientism? This is an important question to address, because defenders
of the ideology often claim that it is a straw man, and that nobody actually holds the opinions
that critics of scientism rail against. The following, then, is a minimal gallery of scientistic
rogues, to unequivocally show that scientism is real and widespread, especially among scientists
and science popularisers.
In The Moral Landscape, author Sam Harris argues that science can determine human values,
with no help from moral philosophy, which he dismisses in an endnote as “directly increas[ing]
the amount of boredom in the universe”.21 This sort of thing has become a small cottage
industry, as in the case of Harris’ colleague and supporter, Michael Shermer, who contributed
to the genre with his The Moral Arc, in which he argues along lines that are difficult to distinguish from Harris’s own.22 The reason I find these books scientistic in nature is because of
their intellectual totalitarianism: they do not seem to suggest that science can aid moral discourse (which would be reasonable, and also trivially true and thus not worth writing a book
about), but rather that the scientific approach is the only one that can profitably do so.
Another class of scientistic examples comes from what has now become yet another cottage
industry of articles and books that purport to provide the answer to a number of complex
questions based on fMRI scans (or similar techniques) of the brain. This approach has been
debated and debunked a number of times, but it persists in the media, and is in fact more
popular than ever.23 Just like the case of the relationship between moral philosophy and science,
the issue isn’t that neuroimaging doesn’t tell us anything of value about, say, gender differences,
moral decision making, or a number of other interesting questions. The issue is that what
neuroimaging tells us is always going to be only part of a complex story, and often not a
particularly compelling part at that. It may be interesting, for instance, to pinpoint the neural
correlates of moral decision-making in humans, but to say that (indirectly estimated) activity in
neuroanatomical structure X or Y correlates (statistically) with this or that type of moral
thinking doesn’t tell us even close to what we want to know about why people engage in
certain types of moral thinking (let alone whether they are doing it right!), which is presumably
the result of genetic, developmental, environmental and social factors, inextricably interlaced
with one another.24 Indeed, if taken at face value, such research may even harm our understanding of whatever social-psychological phenomenon we are studying, by providing the illusion that we are nearing the crucial locus of explanation for it, while we are doing no such
thing.
More direct examples of scientism can easily be found in the pronouncements of a number
of public figures in science, for instance: “Philosophy is dead. Philosophy has not kept up with
modern developments in science, particularly physics”, wrote eminent theoretical physicist
Stephen Hawking, right at the beginning of a book best characterised as a popular treatise on
the philosophy of cosmology.25 Or consider what science populariser Neil deGrasse Tyson said
when he appeared on the Nerdist show back in 2014:

374

Scientism and liberal naturalism

My concern here is that the philosophers believe they are actually asking deep
questions about nature. And to the scientist it’s, what are you doing? Why are you
concerning yourself with the meaning of meaning?…The scientist knows when the
question ‘What is the sound of one hand clapping?’ is a pointless delay in our
progress.26
Along similar lines, the cosmologist Lawrence Krauss, in an interview with the Atlantic said:
Every time there’s a leap in physics, it encroaches on these areas that philosophers have
carefully sequestered away to themselves, and so then you have this natural resentment
on the part of philosophers…Philosophy is a field that, unfortunately, reminds me of
that old Woody Allen joke, ‘Those that can’t do, teach, and those that can’t teach,
teach gym.’ And the worst part of philosophy is the philosophy of science; the only
people, as far as I can tell, that read work by philosophers of science are other philosophers of science.27
Finally, Nobel physicist Steven Weinberg wrote:
Philosophy of science, at its best seems to me a pleasing gloss on the history and
discoveries of science…some of [philosophy is] written in a jargon so impenetrable
that I can only think that it aimed at impressing those who confound obscurity with
profundity…The insights of philosophers have occasionally benefited physicists, but
generally in a negative fashion – by protecting them from the preconceptions of other
philosophers.28
Now indulge me for a moment and just imagine if, in all seriousness, I were to write something
along the following lines:
Physics is dead because it has not answered any question raised by philosophers. And
the worst is quantum mechanics. The only people, as far as I can tell, that read works
in quantum mechanics are other people who work in quantum mechanics. Some
physics is written in a jargon so impenetrable that I can only think that it aimed at
impressing those who confound obscurity with profundity.
I hope no further commentary is necessary to drive home the point that these are abysmally
ignorant comments born of precisely the sort of intellectually arrogant attitude that characterises
scientism.

3.1 Scientism as an unwarranted trump card against
embarrassing questions
The debate about scientism is, however, more complex than the picture I have depicted so far
would lead one to believe. Accusations of scientism have been hurled, for instance, by people
who wish to shield metaphysical, theological or even downright pseudoscientific positions from
critical and particularly scientific analysis.
For instance, theologian John Haught has criticised philosopher Daniel Dennett for his belief
in scientific naturalism, which leads Dennett to hold a position according to which “only
nature, including humans and our creations, is real: that God does not exist; and that science
375

Massimo Pigliucci

alone can give us complete and reliable knowledge of reality”.29 Dennett has strongly reacted to
this kind of criticism, labeling them “an all-purpose, wild-card smear…When someone puts
forward a scientific theory that [religious critics] really don’t like, they just try to discredit it as
‘scientism.’ But when it comes to facts, and explanations of facts, science is the only game in
town”.30
It seems to me that both Haught and Dennett are off the mark here. Haught is right that
Dennett and other so-called New Atheists hold a position similar to the one he describes, but he
still needs to marshal out arguments for why such position is philosophically untenable. Hurling
around epithets will hardly do. Moreover, there are other kinds of naturalism – such as the
“liberal” one advocated in this volume – that still present serious problems for any theologian, since
the supernatural, whatever that is, cannot be simply assumed. Convincing arguments and evidence
have to be presented, and the track record of theology is not exactly inspiring in that respect.
That said, Dennett himself cannot dismiss the charge of scientism in a blank statement,
simply because it is inconvenient for his cherished positions. He owes us a good discussion of
what he means by “facts”, for instance. Does he – like Alex Rosenberg and Lawrence Krauss –
include mathematical and logical statements among “facts”? I happen to know for certain that
he doesn’t, but he needs to square that position with fellow atheists. And wasn’t Dennett
himself who memorably stated that “There is no such thing as philosophy-free science; there is
only science whose philosophical baggage is taken on board without examination”? Does this
“baggage” fall into the category of facts? Is it amenable to scientific inquiry?31
An example from pseudoscience will help. Author Deepak Chopra is notorious for making
nonsensical statements about something he calls “quantum mysticism”, a word salad that includes
occasional references to concepts in quantum mechanics in order to peddle his “idealist” view of the
world. When called to task, Chopra deploys the scientistic trump card, as in the following tweet:
Proud to be a ratbag idealist. Realism of any kind – naive realism, representational
realism, or scientific realism, or any theory positing an observer independent reality
can neither be falsified nor validated notwithstanding the claims of apostles of
#Scientism,” which he published in response to a comment by theoretical physicist
Sean Carroll: “Belot, following David Lewis, makes fun of ‘ratbag idealists,’ who
imagine that we can change the laws of nature just by changing how we think about
them. (Apparently Kant was a ratbag idealist.)”32
Examples could be easily multiplied, but if theologians and purveyors of pseudoscience can
deploy “scientism” as a weapon to defend the indefensible, then we do need to be careful how
we ourselves deploy the term. Why not eliminate it altogether, then? Because it serves a positive function (that of highlighting legitimate instances of misuses of terms like “science” and
“scientific”) and because plenty of other perfectly viable words – including both pseudoscience
and science itself (ever heard of “scientific” creationism?) – have been and are being abused,
without for that reason prompting philosophers to drop them from their vocabulary.

4 What is this thing called science?
If we want to understand the abuses of science, both in the sense of applying the concept where
it doesn’t belong (i.e. scientism) and in the sense of pretending that something is scientific while
it isn’t (i.e. pseudoscience), we better have some conception of what science is. Unfortunately,
there is no agreed upon clear cut definition of science, and for good reasons. The Oxford
Dictionary’s stab at it is as good as any:
376

Scientism and liberal naturalism

The intellectual and practical activity encompassing the systematic study of the structure
and behaviour of the physical and natural world through observation and experiment.
This and similar definitions run into immediate problems once we begin to unpack it. How is
this systematic study of the natural world actually carried out? That is, is there such a thing as a
distinct scientific method, different from the methods of, say, mathematics and logic, or history
and literary criticism, or even everyday knowledge? Do all sciences rely on systematic observations and experiments? And so forth.
Even a cursory look at standard textbooks in philosophy of science will quickly disabuse the
non-naive reader of the notion that there is a unified scientific method of any kind. While we
don’t need to go full Feyerabend and embrace methodological anarchy, it is sensible to realise that
science is conducted by using a dynamic set of tools that vary over time and from discipline to
discipline. Alan Chalmers, in his What is This Thing Called Science? deftly explores a number of
philosophical accounts of the workings of science, including induction, falsificationism, Kuhnstyle paradigms and their shifts, Lakatos’ research programs, Feyerabend’s anarchism and
Bayesianism, finding them all wanting as general theories, and yet all useful as shedding partial
light on the inner workings of science.33 The book should be mandated reading for scientists in
training, and certainly for anyone taking the scientistic position seriously.
Or take the role of observations and experiments in science. Carol Cleland has published a
series of landmark papers exploring what she calls the “asymmetry of overdetermination”, a
crucial factor distinguishing a-historical, largely experimental sciences like physics and chemistry, from historical, largely observational ones like astronomy and paleontology – with yet
more disciplines, for instance evolutionary biology, being characterised by a complex mix of the
two general categories.34 The asymmetry of overdetermination lies in the fact that we make
scientific predictions about the future in a very different way from how we develop “postdictions” about the past.
Take one of the most famous cases in paleontology: the discovery of the crater, off the
Yucatan peninsula in Mexico, that represents the “smoking gun” for the long-held theory that a
major cause contributing to the episode of mass extinction at the end of the Mesozoic was a
devastating extraterrestrial impact on our planet. While paleontologists have pretty much no
clue as to how to predict the next impact of that magnitude, relatively few historical traces (a
worldwide layer of iridium at the K/T boundary, the position and size of the crater) were
enough to spectacularly confirm what initially was mere speculation. That’s because historical
traces overdetermine past events: we don’t need all the pieces of a historical puzzle in order to
reliably deduce what happened.
By contrast, modern fundamental physics is extremely good at predicting the outcomes of
experiments, but that is predicated on the practice of carrying out such experiments in a highly
controlled fashion, under strict laboratory conditions. That’s because current events underdetermine future outcomes: minor variations on the initial conditions, or a plurality of interacting
causal factors, may result in quickly diverging predictions about the future (that is one reason why
predictions based on models of climate change come with fairly large confidence intervals).
Cleland’s point is that scientific research is conducted differently as a function of the level of
historicity of a given discipline, with observations and experiments (or a mix of the two) playing
significantly distinct roles in any particular case. This implies that when we call both paleontology and particle physics “sciences” we are using a very broad umbrella term for the sake
of convenience, not because it indicates a lot of actual similarities between the two activities.
(Cleland also rightly argues, as a side product of her analysis, that to insist in thinking that
physics is the “queen”, and hence model of reference, for science in general is misguided.)
377

Massimo Pigliucci

None of this variability and fuzziness of the concept of “science” should be surprising,
because science is a socially constructed human activity. Before anyone jumps on me thinking
that I have just endorsed some kind of extreme sociological-postmodern conception of science,
let me hasten to add that I am referring to a brilliant analysis by Helen Longino, in which she
attempts to stake a sensible middle ground between the sort of indefensible sociological postmodernism a la David Bloor and hard scientistic positions such as those implicit in the words of
the various physicists mentioned in Section 2 above.35
Longino’s basic point is that of course science is a social construction, in the specific sense that
it is a historically and culturally situated type of human enterprise. The way Aristotle was doing
science while studying mollusk shells on the island of Lesbo was very different from what a
modern biologist, informed by the Darwinian theory of evolution, would do it. Bacon’s analysis
of induction in the Novum Organum set the standard for the new mechanism of Galileo and
Newton, but is hardly relevant to contemporary physics. And so forth.36
But, Longino importantly adds, the fact that scientific knowledge is social knowledge does not
license the inference that it is therefore entirely subjective. Quasi-objectivity in science emerges
from two factors:37 (1) there is, presumably, a real, mind-independent world out there, so that
when scientists insist on theories that are not, in fact, empirically adequate, eventually they have to
relent in the face of the realities of the natural world (see, for instance, the Lysenko disaster in
mid-twentieth century Russia)38; and (2) the more science becomes pluralistic in terms of
welcoming perspectives from people of different genders, nationalities and ethnicities, the more
the once dominant biases of a particular group are corrected and counterbalanced (e.g. medical
research is nowadays carried out on populations that are heterogeneous in terms of gender and
ethnicity, unlike earlier research conducted mostly on middle aged white men, the subpopulation
that, at the time, just happen to constitute the near totality of the medical profession).
So “science” turns out to be a term that refers to a broad, inherently fuzzy, historically and
culturally dynamic type of human activity. In a very real sense, science is what scientists do –
within constraints imposed by a mind-independent reality and, hopefully, by an evolving set of
best epistemic practices (the latter being precisely the domain of philosophy of science). This
means that, for instance, thinking of parapsychology, or astrology, or ufology as pseudoscience
is warranted by the fact that: (1) claims made by practitioners of these approaches clash with our
best current understanding of reality; (2) such claims are not the result of best epistemic practices
and (3) scientists, by and large, don’t accept these fields as scientific. In terms of scientism similar
considerations apply, mutatis mutandis and “science” therefore does not provide answers to
moral questions, of to many questions in literary criticism, history and philosophy. That’s
because these disciplines, although they do benefit from constructive interactions with the
sciences, are not themselves sciences. The question, then, is what kind of healthy relationship
can we imagine among natural science, social science and humanities, without incurring in the
scientistic fallacy of sweeping every type of human quest for understanding and knowledge
under a too general, and therefore ultimately meaningless, label of “science”.

5 A better approach: a stereoscopic view of the manifest and scientific images
of the world
Wilfrid Sellars is arguably one of the most underappreciated philosophers of the twentieth
century. In writings like “Philosophy and the Scientific Image of Man”, he articulated nothing
less than a new program for philosophy itself, one that has influenced some of the major
philosophers of the next generation, including but not limited to Paul and Patricia Churchland,
Ruth Millikan, Jay Rosenberg, Richard Rorty, John McDowell and Robert Brandom.39
378

Scientism and liberal naturalism

Here I wish to focus on Sellars’ well known distinction between what he called the scientific
and the manifest images of the world. The notion is radically different from that of some of
Sellars’ own students (like the Churchlands) who compare the view of the world we get from
science with what they rather dismissively label the “folk” understanding of things. Sellars was
on board – as any serious philosopher ought to be nowadays – with the notion that science is by
far our best (really, our only) approach to understanding the natural world. Nothing else tried
by human beings has come even close to revealing the intricacies of the biological and physical
worlds, and to assent to this statement of fact is not an instance of scientism, as should be clear
by now.
However, Sellars (unlike the Churchlands) also understood that the human world of reasons,
values and other first-person attributes is here to stay, and that it cannot be reduced to a
scientific approach for the simple reason that science has no place for that sort of vocabulary so
essential to understanding (and living) the human condition. As deVries aptly summarises it:
Sellars thinks that science’s aperspectival description of the empirical world is the
measure of reality, but he is also committed to the indispensability of the concepts
built into the first-person perspective that makes agency possible. If those concepts
were not involved in the regulation of our behavior, we would not be persons and
could not engage in such activities as moral behavior or scientific research. This is why
Sellars calls for a stereoscopic vision in which the descriptive resources of the sciences
are united with the language of individual and community intentions and the dualism
of the manifest and scientific images is transcended.40
Some of Sellars’ students (the so-called “left wing Sellarsians”), like Rorty, veered toward a
rejection of the scientific image as irrelevant, or even pernicious to, the manifest image. Others
(the “right-wing Sellarsians”), like the above mentioned Churchlands, took the opposite path
and attempted to eliminate the manifest image, reducing it to the scientific one. Both approaches are, in my mind, profoundly misguided, for opposite reasons that Sellars so brilliantly
understood, and to which we should return. Science itself is a human activity, “socially constructed” in the sense elucidated by Helen Longino, and thus irreducibly influenced by the
manifest image. Similarly, there is no understanding of the manifest image that completely
excludes science, since human beings, their emotions and their cognitive abilities are, after all,
the result of a process of evolution by natural selection and other mechanisms.
Hence the notion of a “stereoscopic” vision in which it is the principal task of a new
philosophy, and in particular of a philosophy informed by liberal naturalism, to constantly
negotiate the proper balance between the two images of the world, making sure that neither
ignores or overtakes the other. Philosophers are particularly – I would say uniquely – well
positioned to carry out Sellars’ program, because the nature and history of philosophy has
always included both images of the world. In modern academia philosophy is considered a
humanistic field, but it is by far the closest to science, and has, in fact, historically spun the
sciences, one by one, beginning with mathematics in ancient times and continuing with
physics (Galileo, Newton), chemistry (Boyle), biology (Darwin), psychology (James) and so
forth.
Philosophy, in a very important and very Sellarsian sense, is not a discipline “about”
something in particular, like biology is about living organisms and history is about human
events. It is, rather, a connective tissue that runs across all disciplines, both scientific and humanistic, allowing us to make sense of what both the sciences and the humanities tell us about
the world and ourselves.
379

Massimo Pigliucci

Notes
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40

De Caro and Macarthur (2004, 1). Rosenberg (2004, 1).
Aristotle, Metaphysics; Spinoza (1677), Hume (1748), Hanna (2006, 16), Quine (1981).
Feyerabend (1975, viii).
Pigliucci (2016, 1–6).
Hacohen (2002), Hayek (1980).
Theocharis and Psimopoulos (1987), Horgan (2016).
Ordover (2003), Lombardo (2011).
Hayek (1942).
Popper (1979, 185).
Putnam (1992, x).
Sorell (1994, 1ff).
Bullock and Trombley (1999, 775).
Peterson (2003).
Blackburn (2005, 331–332).
Boudry and Pigliucci (2013, 2017).
Laudan (1983), Boudry (2013), Hansson (2013), Ladyman (2013), Mahner (2013), Pigliucci (2013).
Wittgenstein (2001[1951]).
Rosenberg (2011).
Hume (1748, 166).
Ayer (1936, 16).
Harris (2010, 211).
Shermer (2015).
Fine (2010), Lilienfeld and Satel (2013).
Greene and Haidt (2002).
Hawking and Mlodinow (2010).
Nerdist.com podcast (http://goo.gl/A7DDjN, the relevant bit starts at 2019).
Quoted in Anderson (2012).
Weinberg (1994).
Haught (2008, x).
Quoted in Byrnes (2006).
Dennett (1985, 21).
See https://www.trendsmap.com/twitter/tweet/1056211565978910720 (accessed on 17 November 2018).
Chalmers (2013).
Cleland (2001, 2011).
Longino (1990, 2001), Barnes et al. (1996).
Bacon (1620).
Giere (2010).
Birstein (2009).
Sellars (1962).
A good overview, with references, of Sellars’s work can be found in deVries (2016).

References
Anderson, R. Has Physics Made Philosophy and Religion Obsolete? Atlantic, 23, (April 2012). http://
goo.gl/pOjrSJ.
Aristotle, Metaphysics. http://classics.mit.edu/Aristotle/metaphysics.html (accessed on 17 November 2018).
Ayer, A.J. Language, Truth and Logic. Penguin. (1936).
Bacon, F. Novum Organum. (1620). http://www.earlymoderntexts.com/assets/pdfs/bacon1620.pdf (accessed on 22 December 2018).
Barnes, B., Bloor, D., and Henry, J. Scientific Knowledge: A Sociological Analysis. University of Chicago
Press. (1996).
Birstein, V.J. The Perversion of Knowledge: The True Story of Soviet Science. Basic Books. (2009).
Blackburn, S. Scientism. In The Oxford Dictionary of Philosophy, pp. 331–332. Oxford University
Press. (2005).

380

Scientism and liberal naturalism
Boudry, M. Loki’s Wager and Laudan’s Error: On Genuine and Territorial Demarcation. In M. Pigliucci
and M. Boudry (eds.), Philosophy of Pseudoscience, pp. 85–104. Univerity of Chicago Press. (2013).
Boudry, M. and Pigliucci, M. (eds.) Philosophy of Science: Reconsidering the Demarcation Problem. University
of Chicago Press. (2013).
Boudry, M. and Pigliucci, M. Science Unlimited? The Challenges of Scientism. University of Chicago
Press. (2017).
Bullock, A. and Trombley, S. (eds.) The New Fontana Dictionary of Modern Thought. Harper Collins. (1999).
Byrnes, S. (2006). When it Comes to Facts, and Explanations of Facts, Science Is the Only Game in Town.
New Statesman. https://web.archive.org/web/20111016131703/http://www.newstatesman.com/
200604100019 (accessed on 17 November 2018).
Chalmers, A. What is This Thing Called Science? Hackett. (2013).
Cleland, C.E. Historical Science, Experimental Science, and the Scientific Method. Geology, 29(11),
(2001). 987–990.
Cleland, C.E. Prediction and Explanation in Historical Natural Science. The British Journal for the Philosophy
of Science, 62(3), (2011). 551–582.
De Caro, M. and Macarthur, D. Naturalism in Question. Harvard University Press. (2004).
deVries, W. Wilfrid Sellars. In Stanford Encyclopedia of Philosophy (2016). https://plato.stanford.edu/
entries/sellars/ (accessed on 30 December 2018).
Feyerabend, P. Against Method. Verso. (1975).
Fine, C. Delusions of Gender: How Our Minds, Society, and Neurosexism Create Difference. W.W.
Norton. (2010).
Giere, R.N. Scientific Perspectivism. University of Chicago Press. (2010).
Greene, J. and Haidt, J. How (and Where) Does Moral Judgment Work? Trends in Cognitive Sciences, 6
(2002), 517–523.
Hacohen, M.H. Karl Popper: The Formative Years, 1902–1945: Politics and Philosophy in Interwar Vienna.
Cambridge University Press. (2002).
Hayek, F.A. Scientism and the Study of Society: Part I. Economica 9(35), (1942). 267–291.
Hayek, F.A. The Counter Revolution of Science: Studies on the Abuse of Reason. Liberty Fund Inc. (1980).
Hanna, R. Kant, Science, and Human Nature. Clarendon Press. (2006).
Hansson, S.O. Defining Pseudoscience and Science. In M. Pigliucci and M. Boudry (eds.), Philosophy of
Pseudoscience, pp. 68–84. Univerity of Chicago Press. (2013).
Harris, S. The Moral Landscape: How Science Can Determine Human Values. Free Press. (2010).
Haught, J. God and the New Atheism: A Critical Response to Dawkins, Harris, and Hitchens. Westminster John
Knox Press. (2008).
Hawking, S. and Mlodinow, L. The Grand Design. Bantam. (2010).
Horgan, J. (2016) Was Philosopher Paul Feyerabend Really Science’s ‘Worst Enemy’?. Scientific American
Blogs, 24 October. https://blogs.scientificamerican.com/cross-check/was-philosopher-paul-feyerabendreally-science-s-worst-enemy/ (accessed on 17 November 2018).
Hume, D. 1748. An Enquiry Concerning Human Understanding. https://en.m.wikisource.org/wiki/An_
Enquiry_Concerning_Human_Understanding (accessed on 17 November 2018).
Ladyman, J. Toward a Demarcation of Science from Pseudoscience. In M. Pigliucci and M. Boudry (eds.),
Philosophy of Pseudoscience, pp. 52–67. Univerity of Chicago Press. (2013).
Laudan, L. The Demise of the Demarcation Problem. In R.S. Cohen and L. Laudan (eds.), Physics,
Philosophy and Psychoanalysis, pp. 111–127. D. Reidel (1983).
Lilienfeld, S. O. and Satel, S. Brainwashed: The Seductive Appeal of Mindless Neuroscience. Basic
Books. (2013).
Lombardo, P.A. A Century of Eugenics in America: From the Indiana Experiment to the Human Genome Era.
Indiana University Press. (2011).
Longino, H.E. Science as Social Knowledge: Values and Objectivity in Scientific Inquiry. Princeton University
Press. (1990).
Longino, H. E. The Fate of Knowledge. Princeton University Press. (2001).
Mahner, M. How to Demarcate After the (Alleged) Demise of the Demarcation Problem. In M. Pigliucci
and M. Boudry (eds.), Philosophy of Pseudoscience, pp. 36–51. Univerity of Chicago Press. (2013).
Ordover, N. American Eugenics: Race, Queer Anatomy, and the Science of Nationalism. University of
Minnesota Press. (2003).
Pigliucci, M. The Demarcation Problem. A (Belated) Response to Laudan. In M. Pigliucci and M. Boudry
(eds.), Philosophy of Pseudoscience, pp. 16–35. University of Chicago Press. (2013).

381

Massimo Pigliucci
Pigliucci, M. Was Feyerabend Right in Defending Astrology? A Commentary on Kidd. Social
Epistemology, 5(5), (2016). 1–6.
Peterson, G.R. Demarcation and the Scientistic Fallacy. Zygon, 38(4), (2003). 751–761.
Popper, K.R. Objective Knowledge: An Evolutionary Approach. Clarendon Press. (1979).
Putnam, H. Renewing Philosophy. Harvard University Press. (1992).
Quine, W.V.O. Theories and Things. Harvard University Press. (1981).
Rosenberg, G. A Place for Consciousness: Probing the Deep Structure of the Natural World. Oxford University
Press. (2004).
Rosenberg, G. The Atheist’s Guide to Reality. W.W. Norton. (2011).
Sellars, W. Philosophy and the Scientific Image of Man. In R. Colodny (ed.), Frontiers of Science and
Philosophy, pp. 35–78. University of Pittsburgh Press. (1962).
Shermer, M. The Moral Arc: How Science Makes Us Better People. Henry Holt. (2015).
Sorell, T. Scientism: Philosophy and the Infatuation with Science. Routledge. (1994).
Spinoza, B. Ethics, Demonstrated in Geometrical Order. (1677). http://www.gutenberg.org/ebooks/3800
(accessed on 17 November 2018).
Theocharis, T. and Psimopoulos, M. Where Science Has Gone Wrong. Nature, 329 (21 October), (1987).
595–598.
Tyson, N. deG. Neil deGrasse Tyson Returns Again. (2014). Nerdist.com podcast audio posted by Katie
Levine on March 7, 2014, http://goo.gl/A7DDjN.
Weinberg, S. Against Philosophy. In Dreams of a Final Theory: The Scientist’s Search for the Ultimate Laws of
Nature, pp. 166–190. Vintage. (1994).
Wittgenstein, L. Philosophical Investigations. Trans. G.E.M. Anscombe. Blackwell. ([1951]/2001).

Online References
https://en.m.wikipedia.org/wiki/Scientism
https://en.wikipedia.org/wiki/Liberal_naturalism
https://en.m.wikipedia.org/wiki/Wilfrid_Sellars

382

32
THE FOUNDATIONS OF
PSYCHOANALYSIS AND LIBERAL
NATURALISM: THE FREUDIAN
UNCONSCIOUS AND THE
MANIFEST IMAGE1
Talia Morag

1 Introduction
Freud famously had scientific ambitions for psychoanalysis, which he saw as a developing
natural science of the mind, with the aim to explain a variety of mental phenomena including
psychological symptoms (e.g. psychosomatic symptoms, obsessional behaviors, phobias, anxiety,
melancholic depression, paranoia), and also jokes, slips of the tongue, and the content of
dreams. He says, for example: “the intellect and the mind are objects for scientific research in
exactly the same way as any nonhuman things. Psycho-analysis has a special right to speak for
the scientific Weltanschauung”.2 Perhaps the most well-known criticism against psychoanalysis,
the one that has gripped not only philosophers but also the popular press and the social imagination, consists in the claim that psychoanalysis is a failed science.3
In this paper, I want to focus on just one of those criticisms, which I find the most compelling, that of Adolf Grünbaum, as it appears in his book The Foundations of Psychoanalysis.4 The
core of his criticism, grounded in a detailed engagement with Freud’s writings, comprises a
burning challenge for psychoanalysis, especially in light of the striking absence of a response
from psychoanalytic thinkers. Meeting this challenge, I suggest, involves embracing the criticism and agreeing with Grünbaum, that psychoanalytic explanations are indeed not scientific,
that is, that they do not track objective (reliable, reproducible) causal connections between
objects that can be individuated by clear identity conditions.5 Such a defence, consists in
earning the right to say (and not merely saying) that psychoanalysis is not, nor need it be, a
science. For that purpose, I argue, we need a new understanding of what Grünbaum correctly
identifies as the two main foundations of psychoanalysis, namely the Freudian unconscious and
the therapeutic outcome. Consequently, my main aim in this paper is to provide a new notion
of the Freudian unconscious that can adequately address (by partly accepting and partly diffusing) Grünbaum’s critique.
I turn here to liberal naturalism – not as a doctrine, but as a method of inquiry. I do not see
liberal naturalism as a mere reaction to scientific naturalism, widening or liberalising the
DOI: 10.4324/9781351209472-37

383

Talia Morag

scientific image of nature. That is to say, liberal naturalism is not merely an inventory list of our
ontological commitments that accepts all objects and causes that scientific naturalists do (i.e. the
posits of successful science), but further admits norms, persons, responsibility, first-person experience – all things that are embedded in our everyday practices and of which we have firsthand experience, even if those experiences are not fitting topics for scientific inquiry.6 I turn to
the writings of David Macarthur who conceives liberal naturalism in positive terms as a philosophy of “the manifest image”,7 which effectively comprises what is available to us in perceptual and introspective experience. Liberal naturalism, as the philosophy of the manifest
image, I attempt to show, gives us the resources to answer Grünbaum’s critique of psychoanalysis, and to help provide a new notion of the Freudian unconscious as manifest.

2 The foundations of psychoanalysis
By the time Grünbaum wrote his book Foundations of Psychoanalysis, criticisms of the scientific
status of psychoanalysis were well-known.8 Defenders of psychoanalysis such as Habermas and
Ricoeur accept that psychoanalysis is not a science, nor should it be a science, and claim that the
criticism against it can be neutralised simply by correcting Freud’s self-misidentification of
himself as a scientist.9 Grünbaum has a powerful reply. What is at stake, he effectively claims, is
not Freud’s self-labeling but the commitments of his theory.
Through a close reading of Freud’s writings Grünbaum argues that at the heart of Freud’s
psychoanalysis is the causal notion of the Freudian unconscious. Setting aside for the moment
the problem of there being various versions of the notion of the unconscious in Freud’s writing,
the basic causal claim of psychoanalysis is that certain identifiable behaviors are caused by a
person’s unconscious motivations – be those understood in term of emotions, desires, fantasies
or memories. As Grünbaum puts it, certain hypothesised unconscious motivations X lead to the
occurrence of behavior Y, and are therefore “causes of human conduct Y”.10 That causal claim
figures in psychoanalytic explanations of certain behaviors, ranging from psychopathologies,
moods and laughter to dreams and slips of the tongue.
It is also the causal claim that lies at the foundation of psychoanalytic therapy, namely the
bringing of the unconscious to consciousness. If the repression, namely, the rendering unconscious of X, causes behavior Y about which the patient complains, then the lifting of that
repression should cure the patient; or at least significantly ameliorate their condition.
Grünbaum here effectively invokes a basic interventionist notion of causation as a difference
maker: If X causes Y then the removal of X should make a difference to, and perhaps entirely
eliminating, Y.
As Grünbaum says, this principal causal claim of psychoanalysis is a scientific claim. The
clinical practice and its therapeutic results are purported to be objectively identifiable, consisting
in the long-term elimination of certain behavioral patterns which undermine the patient’s wellbeing in the beginning of treatment. The therapeutic results, in turn, are supposed to qualify as a
cure. And that cure confirms that certain specifiable motivational states of mind, of which the
patient was hitherto unaware, were unconscious “pathogenic” mental states, to use an expression of Freud and Breuer from The Studies on Hysteria.11 Let us call the causal claim so
formulated “the scientific causal claim”, and its corresponding therapeutic claim “the scientific
therapeutic claim” of psychoanalysis.
As Grünbaum notes, Freud is entirely aware of this commitment:
In 1926 he wrote that evidence of therapeutic gain was from the start intrinsic to the
clinical validation of the psychoanalytic theory of the unconscious: “In psycho-analysis
384

The foundations of psychoanalysis

there has existed from the very first an inseparable bond between cure and research”
(S.E. 1926, 20:256). […] Thus, psychoanalytic causal hypotheses that do not themselves pertain either to the dynamics or the outcome of analytic therapy (e.g. the
trauma theory of hysteria) have nonetheless been parasitic on therapeutic results, since
Freud used such results partly to validate making inferences from clinical data at all,
and partly to guide and support the specific inferences themselves.12
Unlike Popper’s famous argument that psychoanalysis is not falsifiable and therefore not a
science,13 Grünbaum shows that Freud’s psychoanalysis is falsifiable. If a cure does not result
from the clinical experience, then the central causal hypothesis of psychoanalysis must be
deemed false. It is also worthwhile noting here that much like Freud, Grünbaum, too, thinks
that psychoanalysis purports to be a natural science. Given we are in the realm of causes,
Grünbaum asserts, then these are ultimately physical causes14 – a claim that presupposes what
has come to be known as “causal fundamentalism”, namely the belief that there is only one
causal order, the physical one, properly articulated by a future finalised science of physics.15
However, Grünbaum’s claim that psychoanalysis entails scientific commitments, need not
rest on causal fundamentalism. Even what has been called broad scientific naturalism accepts the
existence of several levels of causation.16 The psychoanalytic causal hypothesis can still count as
scientific, even if it does not relate to physical causation and even if it turns out not to be
reducible to some future brain science. The main point is that the causal hypothesis commits
psychoanalysis to being an empirical science that stands or falls by way of “the tribunal of
experience”.17
Grünbaum notes Freud’s own acknowledgement of two phenomena that falsify this causal
hypothesis: the recurrence of the psychopathology after analysis has ended; and the known
phenomenon of spontaneous periods of remission.18 These should be enough to declare the
scientific causal claim of psychoanalysis false and psychoanalysis a failed science. But
Grünbaum’s main charge, made repeatedly throughout the Foundations, complains that even
when therapy is successful, it is plausible that it is the result of suggestion by the psychoanalyst
rather than being caused by the removal of repression. I am less impressed by this option, even if
it is a genuine potential problem, since it seems very difficult to prove, and one could claim that
this is a risk run by any and every psychotherapy. I therefore set this charge aside; and in any
case, it is not required by Grünbaum’s overall argument.
In the rest of this section, I suggest a way to diffuse Grünbaum’s criticism, by modifying the
understanding of the causal hypothesis of psychoanalysis. But first I want to extract what I see as
Grünbaum’s main insight, which will serve as the basis for this paper, namely, his identification
of these two foundations of psychoanalysis:
1
2

A causal notion of the unconscious, used for explanations of certain manifest behavioral
patterns.
A notion of well-being in reference to which we regard the removal of repression as
therapeutic.

Grünbaum understands these two notions in a manner that gives rise to the scientific causal
hypothesis of psychoanalysis, which admittedly Freud endorses in his writings.19 In particular,
the notion of well-being that Grünbaum implicitly invokes is the one with which therapy
begins – the patient complains about certain behavioral patterns, which s/he judges to be bad
for her/his well-being, and will be cured when those are eliminated in the long term.

385

Talia Morag

In order to bypass Grünbaum’s critique I make the two following moves. First, I claim that
what explains certain identifiable behavioral patterns (S1, S2,…SN), is not a specific unconscious
motive U1, nor a limited specifiable set of interrelated unconscious motives (U1, U2…UM), but
rather a whole interconnected network of many unconscious emotions, desires, fantasies and
memories, a network whose boundaries are changeable and vague. I base this claim partly on
Freud’s case studies, in particular the longer ones,20 where it is virtually never the case that Freud
identifies only one unconscious mental state as the cause of the relevant behavior.
Second, I claim what today many contemporary psychoanalytic circles implicitly accept,
namely that there is no predetermined notion of well-being that is presupposed in psychoanalysis. It is quite possible that a patient will not eliminate their initial symptom yet stop seeing
it as a problem.21 Or a patient may alleviate their initial symptoms but come to see other
behaviors as problematic. As Cavell once put it, at some point through the psychoanalytic
treatment, “what you recognize as problems are different”.22 The patient’s notion of well-being
changes as part of psychoanalysis – but not based on any notion that we can articulate in
advance. In fact, contemporary psychoanalysts insist that the analyst must never impose a notion
of well-being on the patient.23
The causal claim of psychoanalysis now becomes:
A network of motivational mental states is repressed, and that repression causes behaviors, some
of which may be the topic of complaint (i.e. symptoms).
The complementary therapeutic claim now becomes:
Removal of repression of enough of that network causes the patient’s behavioral patterns to
change, a change that is partial and unpredictable, and that change may be for the better, but in
an indeterminate sense of “better.”
I have elsewhere argued that psychoanalysis nevertheless owes us some kind of notion of wellbeing, as flexible as it may be.24 Whatever that notion of well-being ends up being, it is going to
involve a certain self-understanding or self-knowledge. Although I am merely providing a
sketch of such a notion of well-being, it is clear that it is far removed from the early Freudian
idea of psychoanalysis providing a “cure” to a “pathogen”, expressed most clearly and directly
throughout the Studies on Hysteria, which Grünbaum articulates and criticises.
The causal claim of psychoanalysis, as I now formulated it, begins to sound like an extension
of common-sense psychology.25 Awareness or self-understanding are ordinary terms that we
use in our everyday lives. And positing a network of unconscious emotions and desires,
memories and fantasies for explanatory purposes seems analogous to cases where we try and
understand or explain people’s behavior by ascribing to them a vaguely defined host of beliefs
and desires. It is vaguely defined, because if I ascribe to someone who is walking to the supermarket the belief that the supermarket is open, I may as well also ascribe to that person the
belief that the pavement is solid, and that cars don’t fly, and that it won’t start snowing suddenly
in summer and so on.26
These are cases where people do not express their beliefs and desires, and yet we still allow
ourselves to assume they have beliefs and desires that motivate them. We might be wrong –
they might be just following a habit, or maybe there is another explanation for their behavior.
But at least in some of our everyday explanations we allow ourselves to posit the existence of
mental states in other people. Positing unconscious motivations seems a very similar explanatory
or interpretative exercise.
386

The foundations of psychoanalysis

Some beliefs and desires may be expressed in a person’s behavior and “readable” as such, with no
need to posit anything imperceptible. People can usually feel when someone believes or doubts them
– just by reading their facial expression. At least some of us can feel if someone is sexually attracted to
us – without the need to posit anything, but simply by being aware of their “body language”, the
way they smile, the looks they give. I shall discuss in further detail this kind of “reading” in the next
section, where I shall claim that unconscious motivations can also be “read” in this manner.
Whether posited or “read”, unconscious motivations as they figure in this revised causal
claim, comprise an extension of common-sense psychology. The modified causal claim of
psychoanalysis cannot be considered scientific, if only because of the difficulty to specify objective identity conditions for the causes in question, the unconscious network. According to
the revised causal and therapeutic claims, then, psychoanalysis is not a science, but not because
it is a failed science as Grünbaum argues. If psychoanalysis is an extension of common-sense
psychology, what is, then, precisely, its empirical status? And what route can we take to defend
its theory and practice? This depends, as I show in the next section, whether one endorses a
scientific or a liberal naturalism.

3 Naturalism and the foundations of psychoanalysis
I have argued in the previous section the foundations of psychoanalysis consist in a causal and a
therapeutic claim that are not amenable to scientific study and confirmation. If we wish to
defend psychoanalysis – despite it being nonscientific – from the charges of mysticism and
obscurantism, then it makes most sense to think of it as an experience-based theory and
practice. But how shall we understand the empirical status of psychoanalysis? If psychoanalysis
comprises, as I have claimed in the previous section, an extension of common-sense psychology, then the answer to our question depends on our views on common-sense psychology.
And those, in turn, as I show in this section, depend on whether one endorses a scientific or a
liberal naturalism. In this section, then, I shall examine how scientific and liberal naturalism bear
on the foundations of psychoanalysis, including the notion of awareness presupposed by the
therapeutic claim and, more importantly, the notion of the unconscious presupposed by the
causal claim.
Awareness can be seen as a straightforward phenomenon that all of us experience directly
and without complication. Indeed, this is how both Freud and Hume saw it – as simply what I
have on my mind right now.27 It belongs to what Wilfried Sellars calls “the manifest image” of
the world.28 The manifest image is the world as we experience it through perception and
introspection and that is describable in ordinary language. Or, as David Macarthur says: “The
manifest image [includes] observable and introspectable things, and cognitive extensions of our
awareness of such things based on ‘correlational techniques’ [as Sellars calls them] (e.g. induction, statistical inference)”.29
When Sellars introduces the notion of the manifest image, he contrasts it with “the scientific
image” of the world, “which postulates imperceptible objects and events for the purpose of
explaining correlations among perceptibles”.30 The scientific image, in other words, “involves
acceptance of all the posits of all successful scientific explanations of phenomena”.31 As
Macarthur explains:
[T]wo ideas are […] central to Sellars’s approach: (1) that there is a clash between
scientific and manifest images (call this the clash); and (2) that in adjudicating the clash,
the scientific image has priority (call this scientific priority). Scientific naturalism accepts
the clash and scientific priority.32
387

Talia Morag

But there is no science of awareness. No wonder, then, that the scientific naturalist will see
awareness (or what is more commonly called consciousness) – a phenomenon whose existence
would be quite strange to deny and that clearly belongs to the manifest image – as a problem.
How does awareness fit into the scientific image?33 Insofar as the nonscientific therapeutic claim
of psychoanalysis goes, then, the scientific naturalist who wishes to defend psychoanalysis as a
science, can only do so by waiting for a solution of the “hard problem” of consciousness.
Consider next unconscious mental states that partake in the revised nonscientific psychoanalytic causal hypothesis, taken to be an extension of common-sense psychology. What is the
scientific status of the mental states that partake in common-sense psychology? Scientific naturalists call it a “folk science” or a “folk psychology”.34 As Macarthur explains:
The almost universal assumption that our so-called “folk psychology” is a protoscientific theory, which aims to explain and predict human behavior by hypothesizing
“law-like [causal] relations holding among external circumstances, internal states and
overt behaviors.”35
In other words, scientific naturalists look at our ordinary psychological talk about mental states,
such as beliefs, desires, intentions, emotions, as imperceptible explanatory posits. They are not
scientific posits, since they are too vaguely defined to have the objective identity conditions
required by science36 and since their contents may be partly supernatural.37 The scientific
naturalist can then but hope for a future maturation of “folk psychology” into a science.
From a scientific naturalist perspective, then, the defender of psychoanalysis should be busy
showing why psychoanalysis is indeed an extension of this “folk psychology”, and have faith
that the latter will one day be properly regimented into a science. Without the salvation of a
future science, psychoanalysis will be deemed a meaningless discourse. It is not clear, however,
how long one is supposed to wait before giving up on that prospect. And the problem is not
merely a question of time. This hope faces a conceptual difficulty, namely that it is doubtful
what part of our ordinary psychological talk is covered by proto-scientific “folk psychology”.
As Macarthur explains, our ordinary discourse is imbued with normativity, which is not a fitting
subject for scientific inquiry.38 This kind of scientific-naturalist defence of psychoanalysis as an
extension of common-sense psychology, presupposed by Wollheim and Hopkins, has not done
much for the status of psychoanalysis in the social imagination. At least in that respect, it seems
to have failed. Maybe that is because, if the hope is realised, it is not at all clear what would
remain of psychoanalytic terminology in particular or of common-sense psychology in general.
Let us consider next the nonscientific causal and therapeutic claims from a liberal naturalist
perspective. As Macarthur explains, liberal naturalism, in opposition to scientific naturalism,
rejects both the clash between the manifest and the scientific image and the scientific priority of the
scientific image over the manifest image.39 Due to the scope of this paper, I cannot review here
Macarthur’s persuasive arguments against the clash and scientific priority and in favor of accepting
the manifest image for what it is. I simply proceed by describing the manifest image from the
liberal naturalist perspective, which I adopt.
The manifest image includes some objects whose identity conditions are objective and
which display a measurable regularity (e.g. stars, elevators). Some of their workings can be
explained through imperceptible posits, which are in turn a part of the scientific image, where
their regularities appear in the form of causal laws or patterns.40 But the manifest image also
contains objects that do not have a double life, so to speak, in the scientific image and which are
not supernatural either, such as: introspectable subjective feelings of pain or of sexual attraction,
art and persons.
388

The foundations of psychoanalysis

The manifest image is “experience-based”, as Macarthur says.41 In other words, it is empirical, it includes not just objects but occurrences and processes, both in the world and in our
psychology. The manifest image, then, includes causal connections. For example, I see a puddle
of water on the floor, and I ask: “how did this happen?” and my preschooler provides the
answer: “My water cup fell down”. This causal explanation stays at the level of the manifest
image: the two perceptible situations (puddle, a water cup falling) are connected through a
causal relation.
Let us not worry that I am invoking some kind of an invisible metaphysical glue among
correlated situations. I am not correlating anything to anything else just yet. I am just asking my
preschooler to explain this particular puddle. That is, I am asking for what Woodward calls a
“singular” causal explanation for an individual occurrence.42 The notion of cause I am invoking
is the ordinary notion of a difference maker, which allows for an interventionist question –
“what if my son did not knock over the cup? How would things be different then?” If I can ask
that question in a way that makes sense, then I am asking a question about a causal relation. As
Woodward explains:
Singular causal explanations wear the source of their explanatory efficacy on their face –
they explain not because they tacitly invoke a “hidden” law or statement of sufficient
conditions, but because they identify conditions such that changes in these conditions
would make a difference for whether the explanandum-phenomenon or some specified
alternatives to it would ensue.43
Now the causal connection in this example is actually correlated with many other spilling
occasions, and it has a double life, so to speak, in the scientific image, as an instance of certain
causal laws.44 But at the manifest level, there is no need to invoke scientific notions about
gravity or liquid state or any causal laws. It is only if one is a causal fundamentalist and believes
that causation belongs only to the scientific image that one would object to what I shall now
call “manifest causes”. But one need not be a causal fundamentalist. As Macarthur explains:
Some of the most exciting recent work on causation suggests that we see causation as
having close conceptual ties to explanation and explanatory contexts. Such an account
of causation poses a significant challenge to causal fundamentalism by lending strong
support to the doctrine of causal pluralism. If causal talk is always understood relative
to background explanatory context the fact that there are different levels of explanation implies that there can be a plurality of complete causal explanations for the
same event.45
Some manifest causes do not have a double life in the scientific image. Say I see my husband at
home in the early afternoon. And I ask: “how come you are at home?” and he answers: “the
reading group was cancelled”. These are two perceptible and therefore manifest situations
connected with a causal connection. And there is no need to invoke any causal law here. In fact,
my husband’s action need not be a part of any habit. It could very well be that this occasion will
only rarely occur, or maybe it is a one-off occurrence. Maybe on other occasions when the
reading group is cancelled or when other work activities are cancelled, my husband will go to
meet a friend, or go to work at a café, or take a walk. Maybe on that specific day he suddenly
thinks that life would be better if he could sing and spontaneously joins a choir. I have known
him for many years, but that does not mean I can predict his every step or that I know every
aspect of his personality. Our intimates often surprise us – that is how they can make us laugh,
389

Talia Morag

and move us, impress or disappoint us, challenge us and “keep us on our toes”, as we say. Just
the other day I suddenly heard my husband sing for the first time in ten years of marriage!
Nancy Cartwright, in her criticism of Grünbaum, warns us that people might be very
different to one another, and there might be no point sometimes to look for correlational
patterns over populations. “We cannot rule out the possibility of intrinsic variability”, when it
comes to humans, she says.46 I want to make a stronger claim here, that we cannot rule out that
even one person cannot be summed up in terms of reliable behavioral patterns or very stable
habits. In fact, this claim seems reasonably justified in the manifest image, where people are
sometimes said to behave “out of character”, where people are said to “have changed”, where
they go in and out of addictions, make huge career changes that surprise everyone, or identify
with literary characters in suspense novels or films whose behaviors are unpredictable, and
where great artists produce great works of art which are a “one-off”, and so forth. In other
words, I suggest that singular causal relations in the manifest image, or manifest causes, as I
prefer to call them, can be a part of a flexible habit that comes and goes or changes over time.
Perhaps some manifest causes can even be a “one-off”, not correlated with anything else, that is,
not a part of habit.
I said above that I look at a cause as a “difference maker”. But it is not clear that we must
know what “specified alternatives to it would ensue” without that cause, as Woodward says in
the above quote. Say I come home and am surprised my husband is not at home. And he
explains they called him to go to work. But what if they hadn’t called him? Maybe he would
work from home as I expected. But maybe he would have gone to a café and work there. Or
maybe he would have decided to go see a friend for lunch? Or maybe he would have decided to
take the day off and go to the museum, and so on. In fact, what if he simply decided to go to
work anyway, without being called to go there?47
What is important, I suggest, is not that a difference is actually made, nor that we have a clear
idea of what such a difference would be. What is important, for this phone call from work to
count as a cause, or for their being a causal relation between this phone call and my husband’s
getting out of the house, is that the question “how would things be different if my husband
were not called from work?” makes sense, and that at least some possibilities can be hypothesised as I just did above.
The manifest image thus has an aspect that is nonscientific in the strong sense that it has no
double life in the scientific image. This is the realm of the nonscientific empirical, where persons
with relatively unreliable behavioral patterns exist (amongst other things). We can examine the
manifest image and see what belongs to it by using nonscientific empirical methods. We can
examine our own use of ordinary language, we can take recourse to common knowledge that
we presuppose in our relationships and social practices; we can help ourselves to centuries of art,
poetry and literature and to what they teach us about ourselves and others; and we can refer to
our own everyday experience.48
So where does liberal naturalism stand insofar as common-sense psychology is concerned?
The question now becomes whether, or to what extent, we can identify common-sense
psychology with the manifest image. Awareness, for example, is part of the manifest image.
That is why from a liberal naturalist perspective, the so-called hard problem of consciousness
does not even arise. It is an “experience-based” nonscientific phenomenon that includes everything that is straightforwardly introspectable. It seems completely unmotivated, from a
liberal naturalist perspective, to even discuss the possibility that one day awareness may be
understood in scientific terms.49
From a liberal naturalist perspective, then, the nonscientific therapeutic claim of psychoanalysis, which rests on a notion of well-being that is tied to awareness, is not even a candidate
390

The foundations of psychoanalysis

for science. Yet, being manifest, it does not invoke any supernatural or mystical entities. We still
need a better account of well-being in psychoanalysis and to clarify its dependence on such
awareness of what was hitherto unconscious, but that is now a philosophical (not a scientific)
question that is legitimised by liberal naturalism.
What about the rest of our ordinary psychological talk, such as invoking beliefs, desires,
emotions and intentions, then? The first thing to note, is that it is not at all clear that those
mental states are “internal” or imperceptible. As Macarthur says:
In everyday life it is a familiar fact that we can see a loving expression in a face, we
can hear doubt or excitement or pity or fear (etc.) in a voice, we can understand a
gesture of friendship just by looking at a friendly wave or a gesture of ignorance […]50
In other words, mental states can be and often are perfectly visible, apparent and manifest.
Emotions provide an obvious case for manifest mentality. If I see anyone on the street shouting,
clenching her teeth and her fists, and becoming red in the face, I am not postulating an imperceptible anger to explain these manifest phenomena. I am just describing my experience of
this angry person.
The existence of such prototypical expressive behaviors may lead us into thinking that we
could have a science of emotional expressions, whereby we divide emotions into kinds that are
characterised by these prototypical expressions.51 If we could have such a science, then we
would be entitled to regard physiological phenomena such facial expressions and certain action
tendencies as separate from the emotion itself, namely as phenomena that are causally correlated
with an inner, hidden, mental state that is the emotion. In other words, if emotional expressions
were indeed reliably identifiable as members of kinds, then we would be justified in presupposing that they can be explained by imperceptible causes in the scientific image.
But emotions like anger need not be expressed in these obvious or prototypical ways that are
so straightforwardly interpretable. If this person is my friend whom I have known for a long
time I can tell that she is angry by the way she holds her teacup, and/or by the way she slightly
lifts her shoulders in a tense gesture, and/or by the way she becomes quieter than normally,
and/or by her seemingly peaceful “ah right” utterance. Even if she tries to conceal her anger, I
know her so well, that I am familiar with her “tells”, as we call them. Emotional expressions, in
other words, can be and often are quite idiosyncratic.
Even our intimates, however, are not exactly an “open book”. Some “tells” may come and
go or change over time, and some may be “one off’s”. People, even those we know very well,
maintain a level of mystery and unfamiliarity, so often used in fiction and suspense TV series
(e.g. the “twists” where the audience finds out that their favorite character, to which they are
attached and with whom they feel familiar, is motivated by completely different reasons to
those they assumed). If emotional expressions are not so reliable, then we need not (though we
could) presuppose that they are reliably caused by an inner, hidden, mental state. That is, we
need not think that manifest emotional expressions have a double-life in the scientific image.
And if emotional expressions belong only to the manifest image, then what we do when we
describe someone as angry.... circumstances. Rather, we “read” their anger as it is manifested in
their behavior, gestures, their so-called “body language” and their utterances. “Reading”
people, as we call it, requires sensitivity to perceptible nuances in behavior, which we engage in
on a daily basis as a matter of an unreflective skill or habit. Macarthur calls this “reading” a case
of “seeing in” and emphasises that some people, who Wittgenstein calls “Menschenkenners”, are
better than most of us in seeing mental states in the way people talk and move and behave. I
want to emphasise here that we often are, or at least can be, very good readers of our intimates.
391

Talia Morag

As I have argued elsewhere,52 this kind of reading is nonscientific. And as Macarthur emphasises, reading the expressive mentality of others is not taught or learned as part of a theory.53
Although the ineffability of this capacity together with the fact that it comes in degrees might
cause us some anxiety or uncertainty about our knowledge of others,54 we nevertheless rely on
it in our daily lives together. We all know that we have this capacity as part of our common
knowledge about emotions.
Through this short example for the case of emotions, I want to claim that inasmuch as
common-sense psychology refers to readable expressive behaviors, then it is not a folk science
of imperceptible mental states as explanatory posits, but rather belongs to the manifest image. In
other words, common-sense psychology has a manifest aspect. And insofar as it is manifest, then
from a liberal naturalist perspective, such ordinary psychological talk does not raise a problem
about how it fits into the scientific image nor does it require any special legitimation from the
sciences. If we want to defend psychoanalysis as a nonscientific practice, we need to show that
its foundations belong to the manifest imagine. My task then, in the rest of this paper, is to
provide a notion of the Freudian unconscious as manifest, or rather as belonging to the manifest
image.

4 A liberal naturalist notion of the Freudian unconscious
and the method to its discovery
Saying that the unconscious is manifest not only sounds oxymoronic but further might seem
like an endeavor that is very far removed from Freud and his notion of the unconscious, which
is infamous as being mysterious or even mystical in its invisibility. But I shall offer a notion of a
manifest unconscious that is Freudian in the following senses:
1

It is a causal notion with an explanatory role when it comes to explaining puzzling behaviors
that seem intentional but over which we do not have rational control (e.g. phobias,
compulsive behaviors).
It is a person-level notion. The Freudian unconscious, as is well-known, is supposed to, at
least partly, become conscious as a part of psychoanalytic therapy, which thereby also
includes a certain kind of acknowledgement of responsibility for what was hitherto unconscious.55 If we are responsible, to some degree or other, for our unconscious mental
states, then those cannot be merely sub-personal. I further claim that if we can become
aware of our Freudian unconscious, at least partly, then this possibility makes it a personlevel phenomenon. As we shall see, it is a bit like the pain of an injury during a football
game – players can continue playing while ignoring their pain, but then when they stop,
the pain can be overwhelming. That pain was there to become aware of, and in that sense
is a person-level phenomenon even prior to the attention given to it when the game
is over.
Freud aimed at a unified notion of the unconscious, even if he admitted failure in this
respect.56 The notion suggested here attempts at a unified picture:
A It will show how we can remain unaware of certain mental states for a long time – but
without supposing some mysterious storage room of the mind, known as the “topographical” model.57
B It will characterise the unconscious processes proposed by Freud: condensation and
displacement.58

392

The foundations of psychoanalysis

C It will show how we may behave or speak or think in ways that seem meaningful even if we
do not intend to speak or think or behave in this way. The psychoanalytic way of accounting for this phenomenon consists in positing an unconscious that is like a little
person within us.59 It is beyond the scope of this paper to fully criticise this notion. I
shall suffice in just mentioning in passing the well-known issues with homuncular
views of the mind, mainly their failure in explaining puzzling behavior merely by
pushing the puzzle to another level, namely that of the homunculi. In other words,
the notion I shall offer does not amount to a hidden intentionality.
D It will show how our unconscious has a social psychological aspect. This aspect appears,
for example, in Freud’s discussion of typical dreams in The Interpretation of Dreams, and
plays a large role in Jung’s theory of the “collective unconscious”,60 Lacan’s theory,61
and in the school of relationist psychoanalysis.62
The notion of the unconscious as manifest is not so outrageous as one might initially think.
Consider the following very famous quote from Freud:
When I set myself the task of bringing to light what human beings keep hidden within
them…I thought the task was a harder one than it really is. He that has eyes to see and
ears to hear may convince himself that no mortal can keep a secret. If his lips are silent,
he chatters with his finger-tips; betrayal oozes out of him at every pore.63
It is challenging to try and discover anything manifest in a theory that is ridden with obscure
terminology, much of which seems to postulate hidden entities or processes. In order to unravel
what is manifest in psychoanalysis I shall set all theory aside, and “go back to the rough
ground”, to borrow a phrase of Wittgenstein’s.64 That is, I shall investigate the actual practice of
psychoanalysis and examine the way in which it relates to ordinary language and ordinary
experience, to our common knowledge and to various nonscientific modes of understanding
persons such as novels, old adages, proverbs, and aphorisms. In other words, I shall employ the
nonscientific empirical methodology of investigation of the manifest described in Section 2.
I cannot fully defend this method of investigation here, though it is implicit in much of our
philosophical investigations into ethics and aesthetics. I will suffice on this occasion with a quote
by J.L. Austin:
[O]ur common stock of [psychological] words [and sayings and adages] embodies all
the distinctions men have found worth drawing, and the connexions they have found
worth making, in the lifetimes of many generations: these surely are more numerous,
more sound, since they have stood up to the long test of the survival of the fittest, and
more subtle, at least in all ordinary or reasonably practical matters, than any you or I
are likely to think up in our armchairs of an afternoon.65
In what follows I rely on my own experience of psychoanalytic psychotherapy. It is just my
experience, and yet if psychoanalysis is not a science, then there are no relevant statistics to be
taken account of over the population of analysands. The kind of self-knowledge that psychoanalysis can achieve comes from the analysand’s own experience of the analysis,66 and so I
take myself to be entitled to speak about that, even if I am not a psychoanalyst.

393

Talia Morag

Now one might worry that whatever it is that I am defending depends on the kind of
training my psychoanalytic psychotherapist has undergone and the theories she endorses. My
sincere, albeit provocative, answer to this concern is that I do not believe this detail matters.
Different analysts may have different interpretations of the same situations due to their terminological and theoretical differences. And that would matter if the self-knowledge gained in
analysis – the coming to awareness of what was hitherto unconscious – had to do with such
interpretations or insights.
My claim, then, is that those insights, whether they originate with the analyst or the analysand, sometimes referred to as “pennies dropping”, can feel very meaningful and arouse
various emotional reactions, and yet they do not comprise the main work of psychoanalysis. I
suggest we understand such moments as summary judgments, namely after-the-fact realisations
and rationalisations of what has actually already happened over a long period of time during the
analysis. Such summary judgments belong to the register of awareness. They comprise a way of
reorganising or narrating our self-understanding of our past behavioral patterns or of our relationships, and thereby have a role to play in our self-image. If we want to understand what the
practice of psychoanalysis consists in, we need to understand what precedes these interpretative
insights.
To say it differently, and perhaps in a less provocative manner, I want to suggest that there is
something about psychoanalytic practice which is not theoretically driven and that is common
to all schools of psychoanalysis. And I further claim that this common core is the practical
foundation of psychoanalysis of whatever school, and what makes these schools a group that is
different in kind to, say, cognitive behavioral therapy.67 Something happens in the analytic
situation in-between interpretations, whether the analysand is sitting or lying down, whether
they meet once a week or five times a week. It is a peculiar situation, where one person talks
and another listens. The one who talks is supposed to say whatever comes into their mind, no
matter how strange or embarrassing or irrelevant it might seem. As Freud famously says:
[T]o put himself into a state of quiet, unreflecting self-observation, and to report to
[the analyst] feelings, thoughts, memories – in the order in which they occur to him.
At the same time [the analyst] warn[s] him expressly against giving way to any motive
which would lead him to make a selection among these associations or to exclude any
of them, whether on the ground that is too disagreeable or too indiscreet to say, or
that it is too unimportant or irrelevant, or that it is nonsensical and need not be said.
[The analyst] urge[s] him always to follow only the surface of his consciousness and to
leave aside any criticism of what he finds, whatever shape that criticism may take.68
This rule is extremely hard to follow, actually, and one finds oneself not saying what is on one’s
mind quite frequently. As Jonathan Lear says:
There will be some kind of disruption to the free flow of speaking one’s mind: a pause
or silence, a sudden change of subject, intense fatigue, the eruption of a somatic issue
like coughing, stomach ache, headache, bowel troubles, and so on.69
This inability to follow through and say out loud whatever comes to one’s mind is part of
psychoanalytic practice too, and I will discuss it in the next section. But the very basic activity –
the attempt to say whatever comes to one’s mind – the attempt at “free association”, letting
words and images and memories follow one another as they do and giving them voice, is, I take
it, the core of psychoanalytic practice and is linked with its primary explanatory goal. The aim is
394

The foundations of psychoanalysis

to uncover unconscious emotions, desires, fantasies and memories, which in turn cause various
emotional and behavioral patterns (including psychosomatic symptoms) that impede the analysand’s well-being.
The implicit claim of psychoanalysis, as is well-known, is that free association is the road –
the long road – to the unconscious, whatever that may be.70 In the next section, I will try to
show and explain why free association is – itself – the unconscious revealed.

5 The unconscious is manifest in free association
A The flow of associations in everyday life as the positive aspect of the unconscious
The basic experience of free association, without vocalising it, is readily available to our ordinary introspective experience. Just try and think of nothing and see what happens. This
attempt to think of nothing invariably fails unless one hums or focuses on bodily sensations or
attends to whatever else meditation practices recommend we attend in order to block out
words and images – I use the term “images” broadly to denote any sense modality, such as
imagined or remembered sights, sounds or smells or feelings that “pop”, as we say, into one’s
mind, one after the other. This ordinary experience of a wandering mind, passively aware of
whatever strikes it, as it were, is precisely what we are meant to voice in psychoanalysis.
It seems uncontroversial to claim that the relations between two consecutive words and/or
images that passively strike our attention are not inferential or rather not rational relations –
where by “rationality” I refer to the tradition that includes Kant and Frege, which regard
rationality as irreducibly normative. The words that strike one in attentive reflection on oneself
are often not part of sentence, and they are often followed by words or images that are quite
remote in meaning.
The claim for the nonrationality of associations is familiar from Hume,71 but I base this claim
on the ordinary experience, familiar to many, of what it is like to focus on a task that is
characterised by practical or technical reasoning, or of what it is like to listen to or articulate an
opinion or an argument. Such focus demands ignoring the words and images that appear in
one’s mind, precisely because they are not rationally connected to the practical or verbal or
interpretative task at hand.
If you cannot recall such attempts to focus, recall instead how in ordinary practice we urge
others, in particular children or students, to focus when performing either a physical or an
intellectual task, to not let their minds wander, to listen “carefully”, to pay attention, to “think
about” what they are going to say or do. Frege famously articulates and argues for this demand
to set aside associations that come to our mind in order to focus on inferential or rational
relations.72 The fundamental rule of psychoanalysis comprises the inverse demand.
I take these introspective and linguistic experiences to show both that associations are not
rational and that the flow of associations is effectively constant. Although we ignore associations
most of the time, they become immediately apparent as soon as we quiet down the rational
demands of our everyday engagements.
The very thin theoretical claim of psychoanalysis, as I see it, consists in supposing what that
these associative sequences of words and images do not appear randomly to one’s mind, but
cause one another in a nonarbitrary yet not rational manner. To put it positively, in a way that
echoes Hume,73 one word or image that strikes one’s mind causes the following one through an
imaginative relation.

395

Talia Morag

Each associationist in the British empiricist tradition suggests a different list of associations.74
My inspiration for the list that I propose here comes from my interpretation of Freud’s case
studies. But effectively, these are ordinary relations that are at the basis of poetic or literary
connections with which we are all familiar. A partial list of these imaginative relations includes:
(1) similarity – not just a straightforward similarity in many respects (e.g. chair and couch), but a
similarity in very few respects, the kind of similarity that is at the basis of metaphors, similes and
symbols (e.g. blue eyes, the ocean); (2) inversion – the kind of relation that is at the basis of
oppositions75 (e.g. white-black, tall-short); (3) part-whole relations, which are at the basis of
metonyms and synecdoche (e.g. a cat and its smile in Lewis Carroll’s Alice in Wonderland) and
(4) part-part relations, which can also found metonyms (parts of the same whole, e.g. cup, tea, as
in “cuppa”).
I want to draw attention to the part-part association, because arguably it could be identified
with the association of contiguity. Salt and pepper are two parts of the same whole, whether
that whole is the dinner table or a salt and pepper set, but they are also a frequent example of
contiguity.76 There is, however, a fundamental difference between these two kinds of
association.
Contiguity in space and/or time is an association that is supposed to be formed in one’s mind
as a result of a repeated co-appearance of the two associated items, A and B. Even if A and B are
not always experienced together, they need to be experienced together enough (plausibly
beyond a certain threshold) for the contiguity association to be formed. Once it is formed, then
even if reality does not reliably provide the experience of A and B together, the respective
contiguity association is reliable. That is, in our psychology, whenever we experience A,
whether in perception or as an association that strikes our mind, we expect B to co-appear
with it.
Part-part relations, conversely, do not entail such a built-in reliability. A and B might be
parts of a whole C, but this is a contingent matter. A may also be a part of D without B being a
part of D as well. So, when I experience A, either in perception or as an association that strikes
my mind, B may or may not strike my mind as a result. The part-part association, in other
words, may or may not be activated when I see or imagine either A or B. So, whereas contiguity associations come with an assumption of reliability, part-part associations do not.
I claim in this section that this ordinary associative activity, to which we can attend by
pausing our everyday tasks and conversations, comprises the Freudian unconscious, or rather its
positive aspect – as I shall soon explain. That is, this associative activity is what I claim to cause
or to be expressed in the various behavioral patterns psychoanalysis aims to explain. These
patterns are not reliable. They come and go with periods of remission and aggravation, and
change over time, as can be seen in Freud’s case studies.77 I therefore prefer to limit the role of
contiguities in the proposed list of associations that are relevant for the explanatory aims of
psychoanalysis.
The claim that words and images that strike our minds are connected through imaginative
associations is supported by several ordinary experiences with which we are all familiar. When
we talk to someone and suddenly, she says something that seems out of context, we often say:
“what made you think of that?” That is, we assume that although what she said is not rationally
connected to what was said a moment ago – it is nevertheless causally connected in a nonarbitrary way to the conversation or to certain features or objects in the immediate
environment.
Another common phrase that attests to our familiarity with imaginative associations is “that
reminds me of”. “That reminds me of” is not a “however” or a “moreover” or an “and so it
follows that” or a whole range of phrases which serve as rational connectives among sentences
396

The foundations of psychoanalysis

or topics of conversation. Neither does it designate an arbitrary relation, whereby speakers have
no idea how some word or image came to their minds unbidden. This “reminding” relation is
in fact an imaginative relation, such as a salient similarity or inversion. In fact, many of us are
puzzled when ideas or feelings merely come to us without any way of making sense of their
arrival, as can be attested by our noting that very fact: “I don’t know why, I’m just in a bad
mood today”, or “I haven’t thought about this in years! I have no idea how it suddenly occurred to me now”.
To give a famous example of the nonarbitrary nature of associations, which is also familiar
from everyday life, let us consider “Freudian” slips of the tongue. Somehow a word that was
obviously going to appear in one’s speech, since it is rationally related to the conversation is
replaced by another word that was clearly not intended to be uttered, and yet it is clear to both
interlocutors that the word slipped into speech not merely by chance.
Say someone passes on to her colleague a social media contact of a friend, for the purpose of
inviting her to his party. And he says: “ah great, I’m always happy to broaden my sexual
network”. He clearly meant to say: “social network”, but the word “sexual”, similar in sound
to “social” came out instead. This is an embarrassing moment, because both interlocutors
understand that this mistake is not random but indicative of this young man’s desires. The word
“sexual”, in other words, that struck our young man’s mind, was caused through imaginative
connections to what preceded it. It seems straightforward to suggest a similarity association of
sound, as well as, perhaps, a part-whole between social contacts and sexual contacts, and
perhaps another association between his party and some fantasies he may have about how it
should end, fantasies on which he may enjoy lingering or alternatively that he ignores whenever
they “pop” into his mind. One word here condenses, (to use a Freudian term that I shall later
explain), at least two, if not more, associations.
A psychoanalyst might be tempted to posit here a little person within our young man that
has this sexual desire (the man’s “Id” or “It”) and declare that “It speaks!”78 But there is no need
to posit this hidden intentionality.79 That temptation, I suggest, comes from the imaginative
nature of associations. Imaginative associations lend themselves to meaning giving. Once we
describe the associations in question and find various respects in which they hold they can seem
to be delivering a message. An association, however, is a merely causal happening that can be
described in words. These words are after-the-fact descriptions, not “subtitles” of the
association.
In other words, according to the notion of the unconscious proposed here, it is not the case
that our young man has a hidden desire, locked somewhere in a secret room or belonging to his
secret inner twin, which is causing him to utter the word “sexual”. Rather, this young man has
associated his colleague’s passing on her friend’s contact and the words “party” and “social” to
the word “sexual” – which he voiced. And that associative sequence plausibly went through
another fleeting association to a sexual fantasy. I take this point to show how this notion of the
unconscious can deliver on criterion 3.C from the previous section.
Now this particular example happens to be a true story. The young man was indeed embarrassed and was able, on reflection, to articulate these associations that I listed here. But even
if this was an entirely made-up example, it would still serve my purposes in this paper. It is a
believable example, and if a fictional character made that slip in a sitcom, at least some of us
would laugh. The associative explanation I provide is meant only as a possible explanation, as a
demonstration of what an explanation of this sort would look like. Its plausibility depends on
common-sensical psychological insights, which I take seriously as my method of investigation
of the manifest image (see Sections 2 and 3). In the case of Freudian slips, the insight is our

397

Talia Morag

natural reaction of embarrassment or laughter when they occur, namely our capacity to read
that the utterance is meaningful, and not just a random mistake.
The unconscious as I describe it in this sub-section, to sum-up, consists in the constant flow
of associations, which we are normally unaware of or ignore. It is a person-level phenomenon
because we can become aware of them by paying attention to them through various techniques
(e.g. meditation, listening to oneself). Furthermore, although it is passive and although we do
not choose our associations, we take them to show something about us as persons or about our
social relationships, as is evident from the phenomenon of slips of the tongue. The “what made
you think of that?” question, or “that reminds me of” locution, also demonstrate that we implicitly acknowledge that our associations are our own making, even if we do not perform them
intentionally. I take this point to show how this notion of the unconscious can deliver on
criterion 2 from the Section 3.
B Associative patterns of inattention as the negative aspect of the unconscious
The psychoanalytic situation imposes the strange condition whereby one is meant to give voice
to the everyday flow of imaginative associations. As we are encouraged to pay attention to the
associations that we normally ignore, we discover, as said in the previous section, that it is
actually a difficult task, constantly interrupted. Lear reflects on his practice:
These disruptions are not merely accidental, but are motivated in various ways. They
tend to function as inhibitions; sometimes under the guidance of self-conscious will,
often bypassing the will, often just outside conscious awareness, though it is relatively
easy to draw a person’s attention to them.80
It is easy to understand what a wilful inhibition is – it is simply the desire not to linger on a
word or an image that strikes our mind, because it is unpleasant or embarrassing. One could
report on this manifest introspective awareness. I am not sure what Lear means by “bypassing
the will”. I suggest we understand this metaphor as a behavior that is not intentional, that comes
over us, like an emotion. I suggest that some associations arouse aversive reactions since they are
embarrassing or guilt- or shame-provoking so we automatically ignore them or distract ourselves from them by thinking about something else, e.g. via “a sudden change of subject”.81
And what is an “inhibition” that is “just outside of conscious awareness”? In a footnote
Lear says:
Analysands have reported to me that, if they pay attention, they can feel thoughts
escaping their consciousness; with effort they can draw them back. They can see for
themselves that the thought they were about to lose was not an indifferent thought,
but an unwanted thought, one which was about to lead in uncomfortable directions.82
I suggest that what Lear’s patients are describing here are effectively associations that trigger
longstanding habits of inattention, which I shall describe in this section.
These two claims – about aversive reactions and habits of inattention – might sound like
empirical claims with very little empirical support. So here too I turn to our ordinary experience. The first claim, regarding inattention of an unpleasant association, is effectively a
specific case of the familiar and ordinary phenomenon of self-deception, or of “turning a blind
eye”, as we say. We all accept that self-deception is possible and that individuals are capable of
ignoring their own thoughts and feelings. This is a part of our common-sense psychology in its
398

The foundations of psychoanalysis

manifest aspect. Although none of us can identify self-deception as it occurs to us, we can and
do describe certain behaviors or locutions of our intimates as cases of self-deception.
We read their manifest behavior and their manifest denial of it. There are occasions, for
example, that my husband says repeatedly and emphatically that he doesn’t care about something, but I can tell that effectively he is expressing his care about that very issue through this
denial, and at the same time, I can tell he is not lying to me, but believes, or rather tries to
convince himself that he does not care about that issue (a case of “[Thou] doth protest too
much”, to quote Hamlet’s articulation of this common wisdom).83
Now it is far more difficult, albeit not impossible, to read someone’s ignoring their own
private thoughts or fantasies. When my interlocutor is pausing unexpectedly their speech,
looking for a moment at the corner of the room, and/or having a strange new expression on
their face, and then comes back to our conversation as if nothing happened, I may make the
plausible assumption that he just brushed off an association.84 Nonetheless, even if adding the
ignoring of unpleasant or immoral associations to our repertoire of self-deception is highly
plausible, saying that such moments of self-deception are manifest might seem far-fetched.
Sustained habits of inattention, however, can more readily count, I shall now argue, as
manifest. And if those are manifest, then a singular occasion of ignoring a specific association,
can also count as part of the manifest image. We might not be able to identify each such
moment of self-deception, but we shall be entitled to suppose that such moments exist by a
“correlational technique”, namely that if a pattern exists, then so do its instances.
Consider a woman who got a divorce in the most unpleasant circumstances. Her younger
and arguably more good-looking husband, a journalist who often works overseas, and who does
not earn much money in comparison to her, cheated on her with a younger woman. Then
when she found out and decided to leave him, he got half of her assets in a divorce settlement.
Now she meets a new person, and they form a committed love relationship.
Her close friends can tell that she is deeply humiliated by the nature of this divorce, which is
quite similar to her previous divorce from another “player”. But our woman has not admitted
any such vulnerability. After expressing some contempt and mockery of her ex-husband
throughout the divorce, she simply stops talking about him. Whenever the topic comes up in
conversation with her new love interest or with friends, she creates some distraction and
changes the topic. “Deflection”, as we might call it, is as manifest as can be. She becomes so
trained in deflecting the conversation that she does not wait for the issue of her ex-husband to
explicitly come up. When the conversation reaches the topic of ex-es, of journalism, or of
working overseas, or of places that they have travelled together – she deflects the conversation.
Her sensitive close friends may be able to notice such a pattern of distraction, and read our lady
as pre-empting the need to deflect the conversation from the more specific topic, namely her
ex-husband. In other words, those habits of deflection are manifest since they are readable.
One can imagine that this deflection can be quite draining or difficult, but soon enough it
can become an unreflective habit. If someone mentions a story about a journalist caught up in a
scandal, she will redirect the conversation to a scandal involving some other celebrity. She
might feel a fleeting uncomfortable feeling, but will not reflect on it or give it any attention,
since she is already distracted thinking about something else. In such a case, although a close and
sensitive friend may see a manifest deflection, the person deflecting might just be performing a
habit. This is now a case of a manifest pattern of self-deception.
At some point the topic (i.e. this woman’s feelings about her ex-husband) simply drops out
of the conversation. This woman has not repressed her divorce or its circumstances but has
managed to avoid the acknowledgement of her feelings of humiliation. A few years later, at her
birthday party, her partner and friends make a short video exhibiting some photos from her life.
399

Talia Morag

But the years she spent with her ex-husband are simply not represented. Not even one photo of
her ex-husband of almost a decade is displayed. No comment or joke is being made about her
ex-husband or his betrayal, and yet – it is “the elephant in the room”, as we say.
“The elephant” is precisely what everyone “in the room” (except our lady, that is) can see
without difficulty. It is very big, as it were, that is, obvious, but nobody talks about it or
acknowledges its presence. If a cruel person who knows this woman well yells out something
about this ex-husband and/or how humiliated our lady feels in this regard, people would feel
most uncomfortable precisely because they know that this comment is correct. Everyone would
know that such a comment is ruthlessly true, even if our proud lady laughs it off and denies her
feelings or alternatively becomes angry with the person who made that comment or feels
humiliated by him, without acknowledging the humiliation regarding her ex-husband. This
“elephant” is, therefore, manifest. And if the “elephant” is manifest then it is equally manifest
that our lady has had some practice in avoiding the “touchy” topic.
This story is not meant to provide evidence of a theory about inattention to associations. In
fact, it is more or less an imaginary example. Parts of it are true, parts of it are made-up, and
parts of it are borrowed from my own first-hand familiarity with deflection. I give this example
in order to elucidate my suggestion, namely that we have a capacity to form longstanding habits
of inattention to associations, what Sartre calls “habits of distraction” in his discussion on bad
faith.85 That is, we exercise the capacity to ignore not only words and images that we positively
prefer to ignore, but also words and images that would remind us of what we prefer to ignore.
And – given the imaginative associative nature of this “reminding” – we also have the capacity
to ignore or deflect from or distract ourselves from words and images that would remind us of
other words and images that would remind us of what we prefer not to experience – and so on.
In this manner, certain feelings and emotionally laden memories or desires can remain associatively isolated for a long time. These feelings and these habits of inattention are nevertheless
manifest since they are readable by a sensitive intimate or analyst.
Although this example is partly imaginary, I nevertheless hope that reading it resonates or
reminds at least some readers of their own experience of deflection. The way this example
makes sense of the familiar ordinary phenomena of an “elephant in the room”, may further
increase confidence in the reader that people indeed have this capacity to form associative habits
of inattention. My example is hardly literature, but fictional characters and the descriptions of
what goes on in their minds routinely serve as fountains of psychological and ethical insight in
philosophy.86
We are now very close to understanding “repression” as an associative habit of inattention
that belongs to the manifest image. In fact, the discussion thus far fulfills criterion 3. A from
Section 3, regarding the capacity to ignore and in that way just not experience certain mental
states for a long time. What we need is not just an aggregate of “touchy” issues that may be
avoided for a long time. We need those “touchy” issues, which are associatively isolated from a
person’s awareness, to be connected, to form a network, that together is supposed to be explanatory of various behavioral patterns. as explained in Section 1.
So let’s tell the story of our humiliated lady in a different way. This time the story begins at
the birthday party with the carefully edited video showing. It is strange to see a documentary of
someone’s adult life with 10 years missing from it. So we may ask, what explains this odd
missing chapter? And the answer is available to her intimates who are very aware of the elephant
in the room – well, this chapter is too humiliating for our lady. She can’t deal with it, she more
or less erased it from her everyday trains of thought, and her loved-ones show grace and do not
force the issue. Had this chapter not contained her unfaithful and possibly “gold-digging” ex,
we would not have this gap.
400

The foundations of psychoanalysis

Looking more closely into this explanation, we shall discuss the imaginative-associative patterns
of inattention that enable this strange situation. In fact, we could tell the story I described above as a
series of explanatory questions and answers in the form of an ordinary causal explanation. For
example, I can ask one of her friends, “why did our lady deflect the conversation from the topic of
this scandal about this journalist?” Her friend will answer: “You’ve noticed, ha? Yes… she’s very
good at deflecting the conversation away from anything remotely related to her ex”. And then we
can understand the term “remotely related” as describing an imaginative association.
Now at the birthday party, after hearing about this lady’s incapacity to deal with her humiliation, we can ask: “but how come our lady is so humiliated and cannot cope with this
emotion?” Arguably this situation, albeit a kind of cliché, is not so catastrophic. Other people
go through this, talk it out, and move on, without having to employ all these avoidances. In
fact, our lady has gone through arguably worse situations (e.g. death of a loved-one, a long
period of illness), without taking recourse to these avoidance strategies. Why is our lady repressing this humiliation specifically? Why has our lady cultivated this specific pattern of inattention? It is, in fact, quite puzzling.
This puzzlement is effectively expressed in the familiar term “touchy”. It is a term that
acknowledges that someone seems to be overly emotional about a certain issue, and at the time
provides absolutely no explanation for it. The word “touch” suggests a causal and not a rational
explanation, but on each occasion where we say “touchy” we basically admit we do not know
what causes this oversensitivity.
Here is where we can hypothesise an answer, by re-iterating the same explanatory strategy we
used to explain all the other patterns of inattention to do with the manifest “elephant”: Our lady
is inattending to this humiliation, because this emotion, or the situation that gave rise to it, is
imaginatively connected to some other pattern of inattention. In other words, this divorce situation “reminds” her of, or is imaginatively associated with, another emotion or another desire
or situation from her past or some kind of fantasy that our lady has been repressing for years. She is
repressing this specific humiliating situation as part of some other pattern of inattention. Maybe it
is a previous romantic humiliation, or a series of such romantic humiliations, or an occasion
where she humiliated someone else, a sibling perhaps, or something to do with her feelings about
her parents’ relationship. In fact, given the way patterns of inattention work, namely through
imaginative associations, it is probably connected to several such emotions or desires, some of
which could have been readable to our lady’s intimates in the past.87
This is not a theoretical hypothesis. On the one hand it is too general, because it is pointing
to patterns of inattention without specifying them. On the other hand, it is too specific, because
if we could specify the other patterns of inattention that causally contribute to this “elephant”,
there is no way to generalise such a causal connection over many subjects (e.g. Whenever you
meet a lady that represses her humiliation by her ex you can safely assume that it is because she is
repressing so and so).88
It is nevertheless an empirical hypothesis, because it makes a claim about something that is
not immediately apparent. But in the manifest image an empirical hypothesis is a call to “look
and see”, to borrow another phrase of Wittgenstein’s89 – a method pointing at a road one can
take in the manifest image to see where it leads. This is the method of free association: encourage our lady to speak about her feelings about her divorce, encourage her to associate freely
rather than deflect, and see where her associations take her. They are going to take her to all
kinds of issues in her life that are imaginatively related – in her mind – to this divorce. But since,
as psychoanalytic practice teaches us, “no one can follow the fundamental rule” of free association,90 then her associative path is bound to encounter further occasions where habits of
inattention kick in. What we know is that the associations leading to these other problematic
401

Talia Morag

emotions, fantasies, desires or memories – call them “issues” for brevity – connect her divorce
with these other issues both imaginatively and causally. We know this just by understanding the
nature of associations of the sorts I have been discussing. In our lady’s mind then, each of her
various issues, the ones she has problems acknowledging, the “touchy” ones to which she has
been inattending, the ones that interrupt her associative train of thought in analysis – imaginatively and causally connects to one or more other issues. Together, these issues form a
gradually revealed messy network, her “baggage”, as we call it.
Although I talk of this network as being gradually revealed, and in that sense as still largely
hidden, even its hidden parts are a part of the manifest image. To reappropriate a well-known
metaphor, what is seen – the readable moments or patterns of self-deception and specific issues
that come up in free association – is just the tip of the iceberg. How do we know there is an
iceberg underneath? Although in the case of the iceberg scientific instruments can be used for
detection, at least in principle all we need to do is look and see. We can just dive deeper and
deeper into the ocean and see more or if. Just as an iceberg hidden in the ocean is a part of the
manifest image, so is the messy network of unconscious motivations.
If psychoanalysis is the attempt to associate freely, and voice our associations without deflection or distraction, then it can be seen as the gradual breaking of our habits of inattention.
And it is no wonder that the more we do this, the more potentially upsetting, and the deeper in
the past or in fantasy the associations become – since they have been ignored for the longest,
and anything that reminded us of them has been ignored for a long time as well. This is how I
suggest we read Freud when he says:
Forgetting impressions, scenes or experiences nearly always reduces itself to shutting
them off. When the patient talks about these ‘forgotten’ things he seldom fails to add:
‘As a matter of fact I’ve always known it; only I’ve never thought of it.’91
This interconnected network of unconscious mental states, and nothing short of it, is what explains the lady’s extra “touchiness” about this humiliation. It is therefore unlikely that our lady
will ever expose all of it in psychoanalysis; and certainly not over a short period of time. The claim
is, rather, that she could over time expose enough of it to make a difference and to lessen its
burden, somehow – in a way that is yet to be clarified, and that exceeds the scope of this paper.
These are, I suggest, the two aspects of the unconscious – positively it consists of the images
and words and feelings that strike us, that passively come to our attention, and negatively it
consists in our associative habits of inattending to these associations. This is it – this is IT. Free
association is the unconscious revealed.
As we can see throughout this section, I have developed (from Freudian materials) a personlevel unconscious, which shows how we can be unaware of certain mental states for a long
time. I will call those simply unconscious mental states in the next section. What we need is to
see how this unconscious can explain the kind of behaviors Freud wanted to explain – besides
slips of the tongue, the explanation for which I briefly sketched in this section. I indicated in
this section how “condensation” is a process that easily comes together with this notion. I shall
elaborate on this in the next section, where I also explicate the process of “displacement”.

6 Ticking the Freudian boxes
Why would our lady even begin psychoanalysis? As the philosopher and psychoanalyst Russell
Grigg says, it is a sense of a riddle that motivates people to seek psychoanalysis.92 Our lady

402

The foundations of psychoanalysis

suffers from occasional inexplicable anxiety and too many people, so she says, make her feel
inadequate and insecure, and she does not understand why. Her birthday party was supposed to
be a lovely event, and yet even there she experienced these emotions. In this section, I want to
show what would count as an explanation of such psychological symptoms.
I reiterate that this example is partly imaginary and I do not even know if this lady ever went
to psychoanalysis. My use of this example and the explanation I offer is only to show what an
explanation of puzzling behaviors would look like in an account like this one that relies on
insights of common-sense psychology.93 Any actual explanation of an actual behavior that a
person complains about in the beginning of treatment, would have to follow the “look and see”
method of free association described in Section 4.
The video at the party contains an image of our lady’s trip to Turkey. Turkey was also where
her ex-husband was when she learned about his affair. Our lady normally avoids the subject
matter of Turkey. But it is unavoidable, being right there on the screen to which she is attending. This confrontation affectively perturbs our lady. She has several images of Turkey in
her memories, images that are affectively laden with anger and shock, and that affect is now
enlivened. She quickly focuses on the next image presented on the video, but the affect, say a
faster heartbeat, is already present. This affect may appear mysterious and puzzling, and she
might wonder whether she is anxious or excited or upset and why. This account of the unconscious as imaginative associations, then, can be used to explain inexplicable anxiety or
moods.
At that party, our lady’s gaze lands on a younger woman who is the new wife of one of her
friends. This is an old friend who was her lover in her youth. She always secretly wanted him to
pursue her seriously, and though he gently flirted with her from time to time, he never reciprocated her love. This friend and his new wife are associated through similarity to her
divorce situation, and an image of her ex-husband with his new girlfriend fleetingly strikes her
awareness. She quickly ignores this memory, but the effect of humiliation is already enlivened.
She suddenly feels humiliated by the presence of her old friend and his new wife. Being around
them makes her feel inadequate and insecure. She feels as though they are mocking her even if
she knows it could not be true. She tries to avoid them, yet she is not sure why. She is behaving
in a way that seems meaningful, but she cannot rationalise it in the manner she can rationalise
her intentional actions. But our lady does not have a secret small little lady “inside” her that
causes her to avoid this couple. Rather, our lady is seeing this couple in terms of that image of
her ex-husband with his new girlfriend, and the enlivened affect colors her relation to this
couple at the party.
It is similar to liking or disliking someone we just met without any apparent reason – until
we suddenly realise that we see this person’s face in terms of another similar face of someone for
whom we have these positive or negative feelings. Prior to realising we are seeing one face in
terms of another, what we might call an imagistic “seeing as” experience,94 we are just aware of
the result – namely our inexplicable liking or disliking of the new person. In other words, we
are aware of the “condensation” of the two faces, precisely because we are unaware of the
associations that compose it. The dislike of the new face, caused by this condensation, is what
Freud calls “displacement”.
When we become aware of what causes our dislike of this new person, we may be able to
stop disliking them, or we may still dislike them, but at least we understand the origins of this
dislike. In other words, the similarity associations between the two faces cause the condensation
and, in turn, the displacement. And becoming aware of an association unpacks the condensation. That is, we now perceive one face and at the same time vividly remember the other.
We are aware of the similarity between them, but we no longer focus on just the perceived
403

Talia Morag

face, or rather on the aspect of that perceived face that is similar to the one we remember.
Awareness is the most we can do to remove this kind of associative cause for displaced affect.
Our lady’s feelings about her old friend actually comprise another “touchy” issue, a readable
unconscious emotion. The humiliation by her ex-husband and her unreciprocated love of her
old friend are imaginatively connected. Possibly, her feelings toward this old friend causally
contribute to the feelings of humiliation by her ex-husband in the first place. It is quite likely,
that other unconscious emotions also imaginatively connect with these ones (e.g. her anger
regarding her parents’ divorce, feeling of jealousy toward her younger sister, etc.), as discussed
in the previous section. Some of those may be causally efficacious at this party inasmuch as our
lady is imaginatively reminded of them by this video about her life.
To sum-up this causal explanation for our lady’s avoidance of her old friend and his new
wife: associations connect various people and things from the party (an image from the video,
certain guests) with certain affectively laden remembered or fantasised images that thereby strike
our lady’s mind. These associations condense so that our lady sees certain people from the hereand-now situation in terms of people and things from her unconscious memories and fantasies.
Although these unconscious emotions trigger habits of inattention, their affect is already enlivened and cannot be ignored. If indeed several unconscious mental states are involved in such
a train of associations, the affect – which has a duration – can accumulate. And the associations
may further cause our lady to focus on a certain person or thing at the party, and regard them as
the cause of the aroused affect.
Now we can ask: What if our lady had not seen the young new wife that night? What if that
new wife had not come to the party? Maybe seeing her old friend/lover would have been
enough, especially in the context of the video down memory lane, to enliven an image of her
ex and feel a humiliation that seems caused by her old friend. Or maybe seeing that old friend
without the new wife would cause her to be reminded of some other image that she normally
inattends to and our lady would have just experienced some kind of unfocused anxiety. Or
maybe she wouldn’t feel anything at all. Or maybe seeing the video was too much no matter
what was on or not on it, and she would feel anxious anyway.95 There is no way to know what
would have happened, because these associations do not follow any lawlike pattern – they are
imaginative, contingent and even capricious.96 But it is still a singular causal explanation (as per
Section 2), because the question “how would things be different if…?” makes sense.
Finally, I want to sketch how this notion of the unconscious can also tick the socialpsychological box (3.D). I have so far emphasised the idiosyncratic aspect of imaginative associations, that is, the images that are connected can come from one’s own memories and
fantasies, and in that sense, they are particular to one’s own psychology. But some associations
can be shared among many people who share certain experiences, share a language, share a host
of cultural references. It could be, for example, that our lady’s inattention to her humiliation by
her ex-husband is partly motivated by her aversion to stereotypical gender roles. The familiar
image from movies and novels and from the documented life of celebrities of the insulted older
woman left by her husband for a younger woman is just intolerable to our lady, as it is a social
symbol of female weakness and dependency that she despises. Her way of ignoring that humiliation, by focusing on her contempt to her ex-husband’s financial dependence on her, also
depends on the cultural cliché of gender role reversal. An important part of unraveling our
lady’s motivations in analysis would be to see how she has been partly trapped by these cultural
associations.
The scope of this paper does not allow me to show how fruitful this notion of the unconscious as manifest may be in re-articulating further psychoanalytic terms in a way that reveals
their reliance on ordinary experience. I leave that for future work. When I say “ordinary”, I do
404

The foundations of psychoanalysis

not mean that it is easy to be a psychoanalyst or to that it is easy to “look and see” people’s issues
and sensitivities in free association or to read them in behavior.
The unconscious is part of the manifest image and is there to be seen in much of our
behavior and our introspective experience and once we attend to it we say that “[we] have
always known it”, as Freud’s patients say. Even repression is always there on the surface of our
experience, since it is a continuous effort, as Freud says, which on this account is understood as
the constant maintenance that habits of inattention require.97 It is precisely due to its persistent
appearance in the manifest image, continuously ignored, that the unconscious is the aspect of
the manifest that is very difficult to notice. Wittgenstein’s comments about philosophy fit this
psychoanalytic endeavor of paying attention:
The aspects of things that are most important for us are hidden because of their
simplicity and familiarity. (One is unable to notice something—because it is always
before one’s eyes.) The real foundations of his enquiry do not strike a man at all.
Unless that fact has at some time struck him.—And this means: we fail to be struck by
what, once seen, is most striking and most powerful.98
I tried to show in this paper, that Grünbaum indeed identified the foundations of psychoanalysis –
a notion of the unconscious and a notion of well-being – but that his understanding of these
notions as belonging to the scientific image requires revision. And this revision, in turn, needs a
positive vision for a different form of inquiry into our behavioral and introspective experience. I
hope to have shown that liberal naturalism, conceived as the philosophy of the manifest image,
can be used to provide the real foundations of this inquiry, which are on the one hand simple and
familiar and on the other hand explanatorily powerful.

Notes
1 Many thanks to Justin Clemens, John Monteleone, Kathryn Tabb, Michael Jacovides, Cameron
Buckner, Mike Dacey, Noa Salamon, and especially David Macarthur for their comments.
2 Freud ([1932], 1995, S.E. 22, 159). S.E. refers throughout the footnotes to The Standard Edition of the
Complete Psychological Works of Sigmund Freud, translated by James Strachey, which was originally
published in 1966. This version is from 1995.
3 “Freud bashing”, as it is called, is a fashionable and accepted practice, see for example Crews (1995)
and Cioffi (1998).
4 Grünbaum (1984).
5 Quine (1969, 23).
6 Macarthur (2019).
7 A term coined by Sellars in (1962, 2007).
8 For example, Popper 1962.
9 Habermas (1971, 246–252); Ricoeur (1981[1977], 259).
10 Grünbaum (1986, 219).
11 The term is used throughout Studies on Hysteria, for example, Breuer and Freud (1995[1983], S.E. 2, 12).
12 Grünbaum (1986, 223).
13 Popper (1962, 37–38).
14 Grünbaum (1986, 219).
15 Macarthur (2015, 567).
16 Macarthur (2010).
17 Quine (1961, 41).
18 Grünbaum (1986, 222).
19 This is especially evident in the case studies of Studies on Hysteria. In the Interpretations of Dreams, Freud
describes how free association leads to the exposure of unconscious mental states in both dream analysis
and the revelation of unconscious motives that cause hysteria: “[…] the correctness of our method is

405

Talia Morag

20
21
22
23
24
25

26
27

warranted by the coincident emergence and disappearance of the symptoms” Freud (1995[1900], S. E.
5, 528). See also Freud’s summary in his Encyclopedia entry on psychoanalysis. He there emphasises
repressed sexual fantasies, that causally explain the hysterical symptom. Freud (1995, S.E., 18, 243).
For example, the case known as “The Rat Man”, Freud (1995[1909], S.E. 10, 155–249).
Lear (2011, 70–71) and (2006, 489–542).
Cavell ([1969]/1998), 86).
Symington (2005). See also Clemens’s provocative claim that psychoanalysis does not involve any
notion of well-being: (2013).
Morag (2017).
Some philosophers have seen psychoanalysis as an extension of common-sense psychology: Richard
Wollheim (1993), Jim Hopkins (1996), Sebastian Gardner (2006) and Thomas Nagel (1994). But they
did not make any modifications to the causal claim that Grünbaum identifies and so they are vulnerable
to his critique.
Wittgenstein discusses similar background certainties or “hinge propositions” in for example
(1969, 341).
In The Ego and the Id Freud writes: “Being conscious is in the first place a purely descriptive term,
resting on perception of the most immediate and certain character”. (1995[1923]: S.E., 19, 13–14).
This is also how Hume conceives of awareness in the Treatise, he writes:
We may observe, that it is universally allowed by philosophers, and is besides pretty obvious of
itself, that nothing is ever really present with the mind but its perceptions or impressions and
ideas, and that external objects become known to us only by those perceptions they occasion.
To hate, to love, to think, to feel, to see; all this is nothing but to perceive [to perceive in
Hume’s sense is to be aware of impressions and ideas] (1738: Bk 1, Part 2, Sect. vi).

28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44

45
46
47

Sellars (1962[2007]).
Macarthur (2017).
Quoted in Macarthur (2017, 387).
Macarthur (2017)
Macarthur (2017).
Chalmers (1995) calls this “the hard problem”.
Pettit and Jackson (1990).
Quote from Paul Churchland in Macarthur (2022, forthcoming).
Chomsky (2000).
Macarthur (2017).
Macarthur (2022).
Macarthur (2017).
Note that not all the workings of stars and elevators are amenable to scientific explanation, for example
their aesthetic features, their economic aspects, the way elevators are tailored to the human body, the
way “stars” have been used in poetry. See Macarthur’s discussion of novels in Ibid.
Ibid.
Woodward (1984).
Ibid., 237.
Given the size and speed in question we can invoke Newtonian mechanics to account for the fall of the
water, as well as hydrodynamics, arguably, to discuss the spread of the puddle (spreading of puddles is
actually a challenging question in physics, for a relatively recent theoretical solution, see Huh and
Scriven (1971).
Macarthur (2010, 130).
Cartwright ([2002]/2009).
Nancy Cartwright, in her criticism of Grünbaum, notes that it is not necessary that causes must make a
difference at all, even if they contribute to the effect (Ibid.). Cartwright is talking about cases of
overdertermination, which Freud thought to be the rule, rather than the exception, in psychoanalysis
([1901]/1905), S.E. 7, 60). It is beyond the scope of this paper to discuss this matter further. I have
elsewhere argued that there is still a point to talking about causes as difference makers when it comes to
overdetermination in psychoanalysis, but the difference may be very small.
I further describe this method and show that it is not new to philosophy in Morag (2017).

406

The foundations of psychoanalysis
49 In this paper, I approach liberal naturalism as a positive philosophy of the manifest image. But if one
treats it as a contrast to scientific naturalism, one could remind us that it is physicalism that puts pressure
on qualia and other aspects of consciousness to be explained in physical terms. Liberal naturalism is not
committed to physicalism and thereby avoids the “hard problem” of consciousness.
50 Macarthur (2022); [emphasis added].
51 See Paul Griffith’s discussion, which relies on the work of Paul Ekman, in (1998, 77).
52 Morag (2017, 354–355).
53 Macarthur (2022).
54 Ibid.
55 Lear (1990, 66).
56 Freud (1995[1923], S.E. 19, 18).
57 For example, Freud (1995[1915], S.E. 14, 161–215). This notion is criticised by Sartre, who remarks
that a storage room requires a censorship function that determines which mental states reside there.
This censorship function, in turn, presupposes a homuncular view of the mind, which is famously
nonexplanatory. See Sartre’s discussion on bad faith in (1943, Chap. 1). For an attempt to answer this
criticism see Neu (1988).
58 Freud elaborates on these “primary processes” in his early writings, in particular in The Interpretation of
Dreams (1995[1900], S.E., 4 & 5); in The Psychopathology of Everyday Life (1995[1901], S.E. 6); and in
Jokes and their Relation to the Unconscious (1905, S.E., 8). These processes and their operation are further
exemplified in his case studies. Although the processes do not receive much attention in his later
writings, they are nonetheless implicit in his ongoing analogy between dreams and symptoms. It is also
worthwhile noting that contemporary psychoanalysts hardly engage with discussion on these processes,
with the important exception of Lacan (2006, 412–467). They are nevertheless presupposed in the
psychoanalytic theories of no matter what school.
59 An Id, or maybe also aspects of the super-ego, what is known as the “structural” model of the unconscious: Freud (1995[1923], S.E. 19).
60 See for example Jung’s essays on archetypes (1959, 1980).
61 See Lacan’s discussion of the four discourses as a fundamental social bond in (1969–1970, 2005).
62 Layton (2006).
63 Freud ([1901]/1995), S.E. 7, 77–78).
64 Wittgenstein (1969, 107).
65 Austin (1970, 182).
66 See for example Freud’s discussion on how the knowledge gained in analysis cannot be merely given
by the analyst but “must rest on an internal change in the patient”. Freud ([1932]/1995, S.E. 16,
280–281). See also Symington (2012).
67 For a brief review of cognitive behavioral and its rivalry with psychoanalysis in the popular press, see
Burkeman (2016).
68 Freud ([1932]/1995], S.E. 16, 287).
69 Lear (2014, 83).
70 Grünbaum (1984, 174).
71 Hume (1738).
72 See in particular Frege ([1884]/1980), 37, fn 1) where Frege discusses the meaning of concepts in
general, and criticises Kant for confusing pictorial associations with logical ideas. See also my discussion
of Frege’s principles of inquiry and the nature of associations in my (2016, chap. 5).
73 Hume (1738, bk 1, part1, sect. 4).
74 Dacey (2021).
75 These are like the “contrast” associations suggested by Thomas Brown and Alexander Bain, as explained in Dacey (2021).
76 Mandelbaum (2015).
77 For example, the case known as Rat Man, Freud 1995[1909], S.E. 10, 155–249).
78 Lacan (2006, 341).
79 For a contemporary account of slips as meaningful but not intentional see Monteleone, 2020.
80 Lear (2014, 83).
81 Ibid.
82 Ibid.
83 For an alternative account of such occasions as cases of inference rather than of reading emotional
expression, see Monteleone (2021)

407

Talia Morag
84 This seems to be Lear’s assumption in such cases (2014, 83).
85 Sartre (1943)
86 One might not want to go as far as making an identity claim between literature and philosophy as
Martha Nussbaum does for the case of Henry James (1992, Chap. 4), but my method here, of using a
made-up example in order to elucidate, discuss and make plausible a psychological idea, is not new in
philosophy.
87 This is how I interpret what Freud calls “secondary repression” in his “Repression” (1995[1915], S.E.,
14, 143–158).
88 See Freud’s comment in the Dora case:
Am I now going on to assert that in every instance in which there are periodical attacks of
aphonia we are to diagnose the existence of a loved person who is at times away from the
patient? Nothing could be further from my intention. The determination of Dora’s symptoms is
far too specific for it to be possible to expect a frequent recurrence of the same accidental
aetiology. (1905[1901], S.E. 7, 40)
89
90
91
92
93
94
95
96
97
98

Wittgenstein (1969, 66).
Lear (2014, 83).
Freud (1995[1914], S.E. 12, 148).
Russell Grigg, in conversation.
This kind of use of imaginary examples is not new in philosophical psychology, for example, A.
Rorty (1980).
Morag (2016, 2019).
This is in alignment with Freud’s discussion of his “botanical monograph” dream in The Interpretation of
Dreams, Freud (1995[1900], S.E. 4, 176).
Hegel famously criticised Hume’s associations for their “caprice and contingency” in ([1830], 1971,
Sect. 455).
Freud (1995[1915], S.E., 14, 151); (1995[1926], S.E. 20, 157).
Wittgenstein (1969, 129).

References
Austin, J.L. Philosophical Papers. Eds. J.O. Urmson and G.J. Warnock. Clarendon Press. (1970).
Burkeman, O. Therapy Wars: The Revenge of Freud. The Guardian. (2016, January 7). https://www.
theguardian.com/science/2016/jan/07/therapy-wars-revenge-of-freud-cognitive-behaviouraltherapy
Cartwright, N. How Can We Know What Made the Ratman Sick? Singular Causes and Population
Probabilities. In A. Jokić (ed.), Philosophy of Religion, Physics, and Psychology: Essays in Honor of Adolf
Grünbaum. Prometheus Books. ([2002]/2009)
Cavell, S. Must We Mean What We Say? Cambridge University Press. ([1969]/1998).
Chalmers, D. Facing up to the Problem of Consciousness. Journal of Consciousness Studies, 2, (1995).
200–219.
Chomsky, N. Explaining Language Use. In New Horizons in the Study of Language and Mind, pp. 19–45.
Cambridge University Press. (2000).
Cioffi, F. Freud and the Question of Pseudoscience. Open Court. (1998).
Clemens, J. Psychoanalysis is an Antiphilosophy. Edinburgh University Press. (2013).
Crews, F. (ed.), The Memory Wars: Freud’s Legacy in Dispute. Granta Books. ([1995]/1997).
Dacey, M. (2021). Associationism in the Philosophy of Mind. Internet Encyclopedia of Philosophy. https://
iep.utm.edu/associat/.
Frege, G. The Foundations of Arithmetic: The Logico-Mathematical Enquiry into the Concept of Number [1884].
Trans. J.L. Austin. Oxford: Basil Blackwell. ([1950]/1980)
Freud, S. The Standard Edition of the Complete Psychological Works of Sigmund Freud. Trans. James Strachey.
The Hogarth Press. ([1966]/1995).
Gardner, S. Irrationality and the Philosophy of Psychoanalysis. Cambridge University Press. ([1993]/2006).

408

The foundations of psychoanalysis
Griffiths, P.E. What Emotions Really Are: The Problem of Psychological Categories. Chicago University
Press. (1998).
Grünbaum, A. The Foundations of Psychoanalysis: A Philosophical Critique. University of California
Press. (1984).
Grünbaum, A. Precis of the Foundations of Psychoanalysis: A Philosophical Critique. The Behavioural and
Brain Sciences, 9, (1986). 217–284.
Habermas, J., Knowledge and Human Interests. Trans. J.J. Shapiro. Beacon. 1971 [1968].
Hegel, G.W.F. Hegel’s Philosophy of Mind: Part Three of the Encyclopaedia of the Philosophical
Sciences. Trans. W. Wallace. Clarendon Press. ([1830]/1971)
Hopkins, J. Psychoanalytic and Scientific Reasoning. British Journal of Psychotherapy, 13(1), (1996). 86–105.
Huw, C. and Scriven, L. Hydrodynamic Model of Steady Movement of a Solid/Liquid/Fluid Contact
Line. Journal of Colloid and Interface Science, 35(1), (1971). 85–101.
Hume, D. A Treatise of Human Nature. Everyman’s Library Press. ([1738]/1911).
Jung, C.G.I. The Archetypes and the Collective Unconscious. Collected Works of Carl Jung, vol. 9, part
I, Trans.R.F.C. Hull, Princeton: Princeton University Press. ([1959]/1980).
Lacan, J. The Other Side of Psychoanalysis: The Seminar of Jacques Lacan Book XVII. Trans. R. Grigg. Norton.
([1969-79]/2005).
Lacan, J. Écrits. Trans. B. Fink, H. Fink, and R. Grigg. Norton. ([1966]/2006).
Layton, L. Racial Identities, Racial Enactments, and Normative Unconscious Processes. Psychoanalytic
Quarterly, 75, (2006). 237–269.
Lear, J. Love and its Place in Nature. Farrar, Straus & Giroux. (1990).
Lear, J. A Case for Irony. Harvard University Press. (2011).
Lear, J. Integrating the Non-Rational Soul. Proceedings of the Aristotelian Society, 114(Part 1), (2014). 73–99.
Macarthur, D. Taking the Human Sciences Seriously. In M. De Caro and D. Macarthur (eds.), Naturalism
and Normativity, pp. 123–141. Columbia University Press. (2010).
Macarthur, D. Liberal Naturalism and Second Personal Space: A Neo-Pragmatist Response to ‘The
Natural Origins of Content’. Philosophia, 43,(2015). 565–578.
Macarthur, D. Liberal Naturalism and the Philosophy of the Manifest Image. In A. Gare and W. Hudson
(eds.), For a New Naturalism. Telos. (2017).
Macarthur, D. Liberal Naturalism and the Scientific Image of the World. Inquiry, 62(5), (2019). 565–585.
Macarthur, D. Naturalism and Skepticism about Other Minds: Remarks on the (In)visibility of Other
Minds. In S. Hetherington and D. Macarthur (eds.), Living Skepticism. Brill. (2022).
Mandelbaum, E. (2015). Associationist Theories of Thought. Stanford Encyclopedia of Philosophy. https://
stanford.library.sydney.edu.au/archives/spr2015/entries/associationist-thought/
Monteleone, J.M. Verbal Slips and the Intentionality of Skills. Synthese, 199(1-2), (2020). 1521–1537.
Monteleone, J. M. Free Association as an Epistemic Method. International Journal of Applied Psychoanalytic
Studies, (2021). 1–18.
Morag, T. Emotion, Imagination,and the Limits of Reason. Routledge. (2016).
Morag, T. Wellbeing, Morality, and the Aim of Psychoanalysis. Parrhesia, 28, (2017). 76–86.
Morag, T. Comparison or Seeing-As? The Holocaust and Factory Farming. In A. Gleeson and C. Taylor
(eds.), Morality in a Realistic Spirit: Essays for Cora Diamond. Routledge. (2019).
Nagel, T. Freud’s Permanent Revolution. The New York Review of Books, 41(9), (1994). 34–35.
Neu, J. Divided Minds: Sartre’s “Bad Faith” Critique of Freud. Review of Metaphysics, 42, (1988). 79–101.
Nussbaum, M. Love’s Knowledge: Essays on Philosophy and Literature. Oxford University Press. (1992).
Pettit, P. and Jackson, J. In Defence of Folk Psychology. Philosophical Studies, 59(1), (1990). 31–54.
Popper, C. Conjectures and Refutations: The Growth of Scientific Knowledge. Routledge & Kegan Paul. (1962).
Ricoeur, P. The Question of Proof in Freud’s Psychoanalytic Writings. In Hermeneutics and the Human
Sciences: Essays on Language, Action, and Interpretation. Trans. John B. Thompson. Cambridge University
Press. (1981/1977).
Rorty, A. Explaining Emotions. In Explaining Emotions. University of California Press. ([1978]/1980).
103–126.
Sartre, J.P. Being and Nothingness: An Essay on Phenomenological Ontology. Trans. H.E. Barnes. Philosophical
Library. ([1943]/1956).
Sellars, W. Philosophy and the Scientific Image of Man. In K. Sharp and R. Brandom (eds.), In the Space of
Reasons: Selected Essays of Wilfrid Sellars. Harvard University Press. ([1962]/2007).
Symington, N. Where Is the Medicine of Healing in Psychoanalysis? Australasian Journal of Psychotherapy,
25(2), (2005). 6–22.

409

Talia Morag
Symington, N. The Essence of Psycho-Analysis as Opposed to What Is Secondary. Psychoanalytic Dialogues:
The International Journal of Relational Perspectives, 22(4), (2012). 395–409.
Quine, W.V.O. Two Dogmas of Empiricism. In From a Logical Point of View, pp. 20–46. Harvard
University Press. ([1951]/1961).
Quine, W.V.O. Ontological Relativity and Other Essays. Harvard University Press. (1969).
Wittgenstein, L. Philosophical Investigations, 4th edition. Trans. G.E.M. Anscombe, P. Hacker, and
J. Schulte. Blackwell. ([1953]/2009)
Wittgenstein, L. On Certainty. Trans. G.E.M. Anscombe and D. Paul. Blackwell. (1969).
Wollheim, R. The Mind and Its Depth. Harvard University Press. (1993).
Woodward, J. A Theory of Singular Causal Explanation. Erkenntnis, 21, (1984). 231–262.

410

33
ACTUALISM AS A FORM OF
LIBERAL NATURALISM
Paul Redding

1 What’s in a name?
The liberal naturalist, we are told, seeks an inclusive philosophical position beyond scientific
naturalism, and in this respect, the adjective “liberal” seems particularly apt. Thus, among the
various meanings for the term “liberal” to be found in the Oxford English Dictionary are: “directed to a general intellectual enlargement and refinement”, “not narrowly restricted to the
refinements of technical or professional training” and “free from narrow prejudice; openminded, candid” (Simpson and Weiner 2004, vol. 8, 881–882).
The other concept central to the dictionary meaning of “liberal” is that of freedom, as is found
in the original sense of “those ‘arts’ or ‘sciences’ that were ‘worthy of a free man’; opposed to
servile or mechanical” (Ibid.). Now, abstracting away from the class- and gender-ridden idea of
a liberal culture deemed the appropriate domain of the leisured gentleman, I take it that such a
link to freedom also recommends the adjective “liberal” to the liberal naturalist as well. This is
because the ontology favored by the rival scientific naturalist can be seen as incapable of finding
a place for us human beings in terms resembling those in which we otherwise understand
ourselves, for example, as beings capable of something like free agency or self-determination.
Now traditionally, idealism has been thought to offer some kind of a counter to scientific
naturalism along these lines, but I suspect that the liberal naturalist might assume that the idealist
is committed to the type of supernaturalism to which she is equally opposed. Thus, liberal
naturalism is, we are told, “a philosophical outlook lying between scientific naturalism and supernaturalism” (De Caro and Macarthur, 2015). The liberal naturalist wants to account for
intentional human beings without reference to “the supernatural, whether in the form of
entities (such as God, spirits, entelechies, or Cartesian minds), events (such as miracles or
magic), or epistemic faculties (such as mystical insight or spiritual intuition)” (De Caro and
Macarthur 2010, 3). I don’t believe that idealism – at least the type of idealism espoused by
Hegel – was in fact a supernaturalist doctrine, but I won’t be arguing for that thesis here. The
account of actualism I will present here derives from quite different sources, drawing on work
within recent metaphysical debates generated from developments in modal logic.1
This form of actualism, I will suggest, neatly provides for those “liberal” features of naturalism that the liberal naturalist desires – basically, a place for mind in the otherwise natural
world – but it does so while circumventing the difficult questions that arise when, starting from
DOI: 10.4324/9781351209472-38

411

Paul Redding

the natural world, one attempts to find in it a place for mind. The advantage of thinking of the
“actual world” as the fundamental domain of philosophical enquiry, I’ll suggest, is that it accommodates mind from the outset. Moreover, as the actualist’s position in modal metaphysics is
self-consciously shaped by a need to avoid supernaturalistic assumptions that the actualist claims
to find in rival views, it is both self-conscious and thorough-going in its anti-supernaturalism.
Nevertheless, the distinction between the “actual” and the “real” is far from clear and easily
misunderstood, and so is in need of some initial clarification.

2 The real and the actual
Realists sometimes respond in a somewhat exasperated manner on being asked what exactly
they mean by “reality”. “Reality is…”, they may say, with some associated hand-waving –
“what’s…out there!”. However, in relation to this it is instructive to be reminded that the last
half century has witnessed considerable debate over the viability of this assumption. David
Lewis had famously used the word “reality” in such a way that there was supposed to be more to
reality than what was “out there”, at least if that phrase is meant to capture our world, the world
that “consists of us and all our surroundings, however remote in time and space” (Lewis 1986,
2). Lewis thought of that latter realm as the actual world, and his reality included the actual as
well as many other alternate possible worlds.
Lewis’s extremely counterintuitive metaphysical claims about the reality of alternate possible
worlds arose when the type of modal logic introduced by C.I. Lewis early in the twentieth
century had received been transformed in the late 1950s and early 1960s when technical innovations by Saul Kripke and others allowed the development of so-called “possible-world
semantics”. Kripke had been able to effectively extend existing “model theory” to C.I. Lewis’s
propositional modal logic by offering completeness proofs for Lewis’s various modal systems. The
basic intuition behind Kripke’s formalisation was to qualify the way that the quantifiers of
classical logic worked in modal contexts. In short, the quantifiers of modal sentences were
thought to quantify not over the range of objects specified in the structures of standard modeltheoretic semantics – that is, objects assumed to exist in the actual world – but over a plurality of
“possible worlds”. In this way, possible-world semantics could treat necessary truths in a
Leibnizian manner as truths holding in all possible worlds, and possible truths as ones holding in
some possible worlds, and, not necessarily ours. And while this was seen by many as surely a good
thing from the point of view of theories of logic and language, it raised the type of metaphysical
questions that analytic philosophers had thought they had put behind them when taking the
analytic turn, such as needing to ask about the “reality” of things that didn’t exist but might have.
In his 1959 paper, Kripke had written that “it is not necessary for our present purposes to
analyze the concept of a ‘possible world’ any further” (Kripke 1959, 2), but others clearly felt
the need to scratch this itch. Lewis’s theory of the reality of merely possible worlds was one such
response (e.g. Lewis 1986).
Lewis’s “possibilism”, the view of reality as made up of many possible worlds,2 was
countered by advocates of various versions of “actualism”, including the form of actualism I
want to discuss here – that found in the work of Robert Stalnaker.3 In contrast to Lewis, an
actualist is, like Stalnaker, more likely to simply identify the real with the actual, and so identify
the real with what is “out there”, apparently in conformity with the realist’s intuition.
However, we should pause to consider the full ramifications of this. Thinking of the actual
world as what is “out there” appears to offer support to Lewis’s “indexical” treatment of the
actual, which probably runs contrary to what the realist had intended. Designated in this
gesturing way, actuality is being characterised subjectively in the way that a particular time, say,
412

Actualism

may be subjectively characterised by being referred to as “now”, or as a time that gains its
identity from the now, as with “yesterday” or “tomorrow” for example. This characterisation is
subjective in the sense that “now” is always somebody’s now. But the actualist, or “modal
actualist”, wants, like Lewis, to treat modal talk as meaningful, and so treat alternate possibilities
as in some sense real.4 And so, foregoing the idea of other possible worlds, the actualist must find a
place “within” the actual world for nonactual possibilities, replacing talk of “possible worlds”
with that of “possible alternate states of the world”. This, however, may put pressure on any
easy identification of the real with what is “out there”. The handwaving gesture of the realist
may have been made in the spirit of a “bricks and mortar” style realism, a little like Samuel
Johnson’s famous “refutation” of Berkeley’s immaterialism. But for the actualist, “out there”
would now seem not only to be related to some subjective point of view, it now also has to
“contain” nonexisting alternate states of things. Being a realist about nonactual possibilities is no
simple task. Aristotle had attempted to capture the idea of a potential (dynamis) but from the
point of view that made the actual (energia) primary. This seemed, however, to create an internal distinction within the actual between something not-yet-actual (a potential) and its actuality.5 The messiness of modal notions such as this had led a generation of analytic
philosophers to deny the meaningfulness of modal talk, but the development of possible-world
semantics had put this issue back on the agenda. Besides this, intuitions about realisable but
nonactual possibilities seem part of everyday consciousness, and it is hard to think of ourselves as
agents without some distinction like this one.
Both of these features of the position that we see shaping up as an actualist alternative to
Lewisian possibilism – the idea of the indexicality of the actual, and the need to accommodate
alternatives to the actual states to be found in the actual world – push this alternative toward
embracing a feature that has been traditionally seen as an aspect of idealism. This is the idea that
there is mind in the world. Lewis’s idea that the actual world “consists of us and all our surroundings” effectively defines actuality as containing us, and so only the relatively uncontentious further claim that we have minds is needed to construe actuality as thereby
containing minds. Idealism is usually associated with a stronger claim – that the world necessarily
contains us minded creatures and that our existence is the outcome of some divine plan. But the
actualist as described here has no need to go down that path. There is a sense in which minds
are necessarily in the actual world, but this is the trivial sense deriving from the decision to
characterise the actual indexically like Lewis. Being realists about alternate possible states of the
world, we can postulate states in which minded creatures had not evolved, our existence as
minded creatures being seen as therefore contingent. But just as a world without minded
creatures would be a world without “nows” or “tomorrows”, a world without minds of any
kind in it would lack those features that make it this world, the world alluded to by our wouldbe realist. In this sense, actualism might be seen as a metaphysically inexpensive version of
idealism, achieving what the liberal naturalist wants from nature without some of the surrounding problems. Nature is easily conceivable without minds, and so starting with nature one
has to find a place in it for minds. Start with the actual world, however, and minds are already in
it. We get them for free.
There is another dimension to this view of actualism with minds-built-into-it-from-the-start
that offers a way beyond a problem expressed in relation to the “bricks and mortar” realist’s
conception of how what’s “out there” might accommodate nonactual possibilities. Simply, if
what is “out there” includes other persons conceived as “intentional” beings, then the existence
of alternate possibilities to the actual becomes less obscure. It is part of our general conception
of what it is to have a mind that minds confer rationality and freedom on their bearers, and the
idea of freedom comes along with the idea of the capacity of an agent to in some way bring
413

Paul Redding

about states of the world that would not have otherwise come about by themselves. That is, we
can conceive of these alternatives to the actual as existing in the intentional states of beings
capable of coming up with plans and acting in ways to help realise them. In this sense, the
existence of minds in the world and the existence of unrealised possibilities or potentialities go
together. This now brings into focus a further but related sense in which the existence of other
minds in the actual world brings with it the existence of alternate possibilities for this world.
Other individuals have beliefs about how the world is that can differ from one’s own, and we
can use the idea of “possible worlds” to capture the content and the logic of those belief states.
Just as necessity is truth in all possible worlds, so knowledge is truth in all epistemically
possible worlds. The assumption is that to have knowledge is to have a capacity to
locate the actual world in logical space, to exclude certain possibilities from the
candidates for actuality (Stalnaker 2019, 12). This was the basic drift of the approach to
the modal logics of knowledge and belief that was started by Jaakko Hintikka in the
early days of post-Kripkean modal logic (Hintikka 1962).
Hintikka had taken this application to the intentional states of others as the natural way to
understand modal logic, rejecting the primacy given to “aletheic” modal readings with their
problem-ridden metaphysical dimensions (Hintikka, 2006, 20–23). Starting from C.I. Lewis’s
propositional approach to modal logic, Hintikka attributed propositional contents to the belief
states of others by treating the attribution of those states in a way that others treated necessity
and possibility as propositional operators. Later, this possible-worlds approach to intentional
states would be given a twist by Lewis, and a further twist by Stalnaker to compensate for the
fact that Hintikka’s approach “did not give explicit attention to the interaction of different
knowers” (Stalnaker 2019, 14).

3 The indexicality of the actual world and locating minds from within it
Actualists are likely to conceive of the relation of modal to nonmodal logics differently to the way
that this has come to be conceived from the dominant point of view. I have suggested that any
view of the world indexed to certain cognisers will be that of a world which must, in some sense,
be understood as containing those cognisers, but this might be met with an opposing view with a
long and prestigious pedigree. For example, an opponent might appeal to the type of view expressed by Wittgenstein in the Tractatus Logico-Philosophicus (Wittgenstein 1981), in which the
world seems indexed to a thinker but in a way that avoids the conception of containing that thinker:
5.631

5.632

The thinking, presenting subject; there is no such thing.
If I wrote a book “The world as I found it”, I should also have therein to
report on my body and say which members obey my will and which do not,
etc. This then would be a method of isolating the subject or rather of
showing that in an important sense there is no subject: that is to say, of it
alone in this book mention could not be made.
The subject does not belong to the world but it is a limit of the world.

This idea that “the subject” is not a thing and so does not belong to the world presupposes a
view of logic and the content of judgments that, I will be arguing, the actualist opposes. This
idea of a Tractarian subject that “shrinks to an extensionless point” and that does not belong to
the world but is rather “co-ordinated with it” (Ibid., 5.64) has a long philosophical pedigree
414

Actualism

that includes Kant’s “transcendental subject of apperception”, an “I think” that while able to
accompany all representations does not belong to the world as represented (Kant, 1998, B131). But
the reason that the Tractarian or transcendental subject here does not belong to its world seems
to be a product of how “the world” to which it is coordinated is conceived. In such views, the
world is conceived as a centerless structure that is isomorphic with a set of consistent centerless
propositions, none of which contain indexicals that would tie subjects in determinate ways to
locations in that world. In the Tractarian world there could be no “heres” or “nows” that the
subject in question would grasp as its own, because the presence of indexicals would upset the
needed pattern of consistency given the fact that the truth or falsity of those indexical judgments
vary with time.6 The actualist thinks of indexicality as ineliminable from any subject’s set of
beliefs, however, and so the view of any particular subject’s set of beliefs as captured by a set of
consistent propositions with absolute, timeless truth values must be given up.
The idea of the “ineliminability of the indexical”, call it the “indexicality thesis”, has appeared in various forms in analytic philosophy,7 but I want to focus here on the types of reasons
why an actualist might want to affirm this position. In an example of an indexical judgment from
Perry’s classic paper, an academic in her office intends to go to a meeting that is to start at noon,
and at noon, she leaves the office to do as she had intended. Among the various beliefs needed
to act in this way is the indexical belief she had at noon that “it’s now noon” (Perry 1979, 4).
Beliefs such as this have come to be known as the paradigms of “self-locating” beliefs because
they locate the agent in the world as they conceive it to be. As Stalnaker puts it, “the agent’s
particular location in the world provides the anchor that determines the way all of her thoughts
get their content” (Stalnaker 2019, 73). An agent can think about nonactual possible situations,
but to do so must draw upon semantic resources provided by their actual world. The basic
strategy at work here is to avoid attributing to agents any “magical” conceptions of reference, of
the sort that Hilary Putnam had referred to as requiring “metaphysical glue” to establish
connections between an agent’s thoughts and what they are about (Putnam 1980).
To incorporate indexical beliefs into a system like Hintikka’s epistemic logic David Lewis
had treated the contents of self-locating beliefs in terms of possible worlds which themselves had
a “center” (Lewis 1979), but a problem with Lewis’s approach, Stalnaker thinks, is that it
obscures the distinction between a simple difference of perspective between individual intentional
agents on the one hand and real disagreement between them on the other (Stalnaker 2008,
49–51). In our attempts to understand the world we draw upon the achievements of others and
we need to be able to calibrate our beliefs with those of different and differently located agents.
If I believe that climate change is real, and Tony tells me that it is a hoax, I don’t want to
construe Tony’s beliefs as merely about the world as Tony believes it to be or how it “looks”
from his point of view. His beliefs are also about my world. Reflecting on the disagreement I
will need to locate Tony in my world, and think about how his “location” in the world – the
newspapers he reads, perhaps – might bear on the views he has about the world.
Stalnaker sums up the situation in this way:
the lesson we should learn from the phenomenon of self-locating belief is that we
cannot give an adequate representation of a state of belief without connecting the
world as the subject takes it to be with the subject who has the beliefs…When we
represent the way this individual locates himself in the world as he takes it to be, we
need to include the information about who it is who is locating himself there, and we
need to link the world as the believer takes it to be to the world in which the believer
takes the world to be that way (Stalnaker 2008, 53).

415

Paul Redding

The problem with Lewis’s way of picturing possible worlds was that according to it
belief about what possible world you are in is like belief about what country you are
in, while beliefs about where in the world you are is like a more specific belief about
where, in the country you are (what village, street corner, or mountain top).
But Stalnaker counters this with the idea that
ordinary belief about where you are in the world is always also belief about what
possible world you are in (what possible state of the world is actual). If I am not sure,
as I drive along the highway toward New York, whether I am still in Massachusetts,
then I am not sure whether I am in a possible world in which this stretch of highway is
located in Massachusetts (Stalnaker 2008, 51).
This significance of this is that it provides a way of translating between self-locating and
ordinary beliefs for single, isolated beliefs. Consider the case where Stalnaker asks himself
while driving, “Is this stretch of highway in Massachusetts?” This question could be answered
by a subject-locating belief about that particular stretch of highway picked out by the demonstrative. But the question can be put in a type of inverted way to ask not about this
stretch of highway (a de-re question, as it were) but about the truth or falsity of the belief involved
(a de-dicto question): “Is the belief that this stretch of highway is in Massachusetts, true?”.
Unlike the de-re question, the de-dicto one is not being made exclusively from the perspective of the driver. Anyone could ask a similar question such as “Is the stretch of highway
that Prof. Stalnaker is now driving along in Massachusetts?” – that is, anyone, including
Stalnaker himself.
Any normal speaker is capable of doing this – it is just a variant on the type of translation we
use when, instead of saying “Yesterday I went to the library”, one says that one went to the
library on a certain date. Stalnaker’s “actualist” point, however, is that to make such generalised
“decontextualised” assertions capable of being understood across contexts, a speaker must
employ semantic resources found in the actual world, because “all of an agent’s representations
get their content from that agent’s relation to the things those representations are about”
(Stalnaker 2019, 4) – things to which they can be related in the actual world.
For the actualist, the picture of human subjectivity portrayed in a Tractarian or transcendental
way cannot give an account of how our intentional states can be directed to the world. It is the
assumption that they are that relies on Putnam’s “metaphysical glue”. We must grasp that it is
only because we are, qua minded beings, located in the world in particular places that we can
have intentionally directed thoughts and beliefs, thoughts and beliefs that are in the first instance
from that particular point of view. For this reason, we must rely in the existence of other
differently located intentional beings who express thoughts and beliefs indexed to those different locations. And so,
clarifying a conception of this kind will involve considering different perspectives on
the world, and relations between those perspectives. In a sense, we are looking at
ourselves from the outside, as agents whose interactions with nature and with other
agents are part of an objective world to be described and explained. But we also
recognize that we aren’t really outside: Our third-person view of ourselves is developed and refined within the world, from perspectives within it (Stalnaker 2019, 3).

416

Actualism

4 Redefining metaphysics with an actualist’s account of modality
I’ve been sketching in some of the features of the modal landscape of the contemporary “actualist”, but as mentioned at the outset, my aim is to attempt to use the basic features of the
actualist position developed within modal metaphysics to argue for a broader redefinition of the
very project of metaphysics itself, one that is in the spirit of “liberal naturalism”.
“Actualism” is often discussed as a metaphysical theory in contemporary analytic metaphysics, in which it is set in opposition to its antagonist, “possibilism”, in much the same way
that traditionally opposed metaphysical theories such as realism and nominalism, materialism
and idealism and so forth, are opposed. A recent volume devoted to analytic metaphysics notes
how, having gone out of fashion during much of the first half of the twentieth century within
analytic circles, metaphysics then came back into fashion with the sorts of modal issues discussed
here playing a significant part in this revival (Loux and Zimmerman 2003). The general impression, however, is that of a restitution of metaphysics in the traditional sense of aiming at
some comprehensive account of the “ontological structures” or “ultimate features” of reality.
Such talk, however, strongly conveys the idea of a subject matter, “reality”, as “what’s there
anyway” – a reality as it exists independently of any mind that might attempt to come to terms
with it. But thinking of metaphysics in this way, while it might make sense from the possibilist
side of the debate, does not, I’ve suggested, make sense from the actualist side. The idea of the
restitution of metaphysics via clashes of metaphysical theories fails to take into account how
such debates might contribute to the redefinition of metaphysics itself.
The version of modal actualism that I have sketched here fits poorly into a position within
the metaphysical debate as traditionally understood: it is probably better understood as a “metametaphysical” stance that foregoes the intelligibility of the traditional project of metaphysics as
out of step with the finitistic, antisupernatural view that has become predominant within
“enlightened” and generally liberal thought in the modern West. In this sense, it is a form of
skepticism, but it still sees a significant place for the sort of work typically done in the name of
metaphysics. On an individual level this type of winding back of the ambitions of metaphysics is
neatly expressed by Stalnaker when he writes that while originally motivated to “get to the
bottom of things”, and later realising that “the bottom was further down than I thought”, he
has finally come to the view that “there is no bottom: the best we can do in philosophy is to
chip away at bits and pieces of the problems” (Stalnaker 2019, 1). This is an attitude he
identifies as broadly naturalistic and Humean, and as at one with the idea that
we can’t separate the task of developing and justifying rules for finding out about the
world from the substantive task of developing a view about what the world is like. We
approach both tasks from within, criticizing and refining the methods and beliefs that
we find ourselves with (Ibid., 3).
If naturalism in its liberal form is motivated by the recognition that we belong to the world, then
the other side of this recognition is that the world should be regarded from the start, as containing
the type of beings that we are.

Notes
1

Nevertheless, I have been drawn to such issues in modal metaphysics via relations, both historical and
substantive, to aspects of Hegel’s idealism. See especially the work of John N. Findlay in relation to the
actualism of Prior (1967) on the one hand and nontraditional readings of Hegel on the other
(Redding 2017).

417

Paul Redding
2

3
4
5
6
7

Here I will use the phrase “possibilism” to refer to Lewis’ claim of the existence of possible worlds of
which ours is just one. As has often been pointed out, the term “modal realism” that Lewis used of his
own philosophy is misleading, as Lewis thought of modal issues as being reductively explained in terms
of the plurality of possible amodal worlds.
That is not to say that all such “actualisms” are without supernaturalistic assumptions, that of Alvin
Plantinga (Plantinga 2003) being a case in point.
Kit Fine defines modal actualism as the conjunction of “modalism”, taking modality seriously, and
“actualism”, taking actuality seriously, in this way (Fine 2005, 2). Fine treats Arthur Prior effectively as
the prototype of the modal actualist (Ibid., 133).
On some of the internal dilemmas of Aristotle’s treatment of the potential–actual distinction, see
Bechler (1995).
That is, while the set of propositions containing some indexical ones [p1, p2…pn] might be consistent at
t1, it might be inconsistent at t2.
The classic statement is Perry (1979), but see also Lewis (1979).

References
Bechler, Z. Aristotle’s Theory of Actuality. State University of New York Press. (1995).
De Caro, M. and Macarthur, D. (eds.). Naturalism in Question. Harvard University Press. (2004).
De Caro, M. and Macarthur, D. (eds). Naturalism and Normativity. Columbia University Press. (2015).
De Caro, M. and Macarthur, D. Liberal Naturalism. In R. Audi (ed.), The Cambridge Dictionary of
Philosophy, 3rd edition. Cambridge University Press. (2015).
Fine, K. Modality and Tense: Philosophical Papers. Clarendon Press. (2005).
Hintikka, J. Knowledge and Belief: An Introduction to the Logic of Two Concepts. Cornell University
Press. (1962).
Hintikka, J. Intellectual Autobiography. In R. E. Auxier and L. E. Hahn (eds)., The Philosophy of Jaakko
Hintikka. Open Court. (2006).
Kant, I. Critique of Pure Reason. Trans. Ed. Guyer P. and Wood A.W. Cambridge University
Press. (1998).
Kripke, S.A. A Completeness Theory in Modal Logic. Journal of Symbolic Logic, 24(1), (1959). 1–14.
Lewis, D.K. Attitudes De Dicto and De Se. Philosophical Review, 87, (1979). 513–545.
Lewis, D.K. On the Plurality of Worlds. Blackwell. (1986).
Loux, M.J. and Zimmerman, D.W. Introduction. In M.J. Loux and D.W. Zimmerman (eds.), The Oxford
Handbook of Metaphysics. Oxford University Press. (2003).
Perry, J. The Problem of the Essential Indexical. Nous, 13(1), (1979). 3–21.
Plantinga, A. Essays in the Metaphysics of Modality. Ed. M. Davidson. Oxford University Press. (2003).
Prior, A.N. Past, Present, and Future. Clarendon Press. (1967).
Putnam, H. Models and Reality. Journal of Symbolic Logic, 45, (1980). 464–482.
Redding, P. Findlay’s Hegel: Idealism as Modal Actualism. Critical Horizons, 18(4), (2017). 359–377.
Simpson, J.A. and Weiner, E.S.C. The Oxford English Dictionary, 2nd edition. Clarendon Press. (2004).
Stalnaker, R.C. Our Knowledge of the Internal World. Oxford University Press. (2008).
Stalnaker, R.C. Knowledge and Conditionals: Essays on the Structure of Inquiry. Oxford University
Press. (2019).
Wittgenstein, L. Tractatus Logico-Philosophicus. Trans. (from German) C.K. Ogden, with an Intro. by
Bertrand Russell. Routledge and Kegan Paul. (1981)

418

34
CRITICAL NATURALISM FOR
THE HUMAN SCIENCES
Daniel Andler

1 Naturalism descends on the human sciences
Naturalism is a central issue in philosophy, and has been for the longest time, albeit with widely
different construals, linked to widely different contexts. But it has come to the fore in the last
half-century or so for a reason: the heretofore nonnaturalist majority opinion came under attack
from two sides. On one side were philosophers armed with conceptual analyses purporting to
show that what came to be known as scientific naturalism is the correct position, whether as an
ontological or an epistemological doctrine. On the other were formal and empirical scientists
with a scientific research agenda aiming at uncovering natural processes and states of affairs to
which, one by one, the tenets of nonnatural understanding, whether lay or scientific, could be
identified, on pain of being altogether eliminated.
The two campaigns were roughly convergent, if not fully coordinated, and provided each
other mutual support. They shared a common spirit as well as certain intellectual ancestors. It
could almost seem, from a distance, as if philosophy was the legislative body and this scienza
nuova its executive branch. But as some reflection, in the light of some experience, has shown,
this picture was flawed. The conceptual plea for naturalism, in itself, was – and remains – freefloating, leaving so many options undecided that a plethora of world-views and conversations
proliferate under the banner of scientific naturalism: as a whole, this philosophical movement is
what Putnam calls a I-know-not-what. By contrast, the scientific research program, which I
will henceforth identify, at the cost of some simplification, with cognitive science, is anchored
in a set of structuring hypotheses regarding the natural realm and the sense in which recalcitrant
entities can be repatriated to – or “placed” in – that realm. As these hypotheses evolve, and
placement results begin to accumulate, a clearer picture of what scientific naturalism amounts to
emerges, as a (provisional and partial) end-result, rather than an a priori standard to which
science should abide.1
The human sciences (both those centered on the individual and those tending to society)
have been at the forefront of the naturalism debate. As is well known, they have been seen, by a
majority of their practitioners and philosophical defenders, as fundamentally differing from the
natural sciences, and thus at once challenging ontological naturalism and constituting a living
rebuttal of epistemological naturalism. No wonder then that the recent naturalistic wave has
descended on the human sciences with a vengeance. Like waves breaking on the shore, this one
DOI: 10.4324/9781351209472-39

419

Daniel Andler

upsets its target’s balance in two ways: it first hits it on the head, and then tows it from under.
First, naturalism attempts to delegitimise the human sciences by questioning their basic ontological tenets: on what grounds can they help themselves to such concepts as mind, thought,
meaning, consciousness, self, abstract objects, norms, culture, practices, language, symbol,
history, class, representation, order, power and so on, which are not observable and don’t seem
to exert any well-identified causal effect on the world, notwithstanding what our commonsense
parlance, contaminated by these human sciences, may suggest? Second, cognitive science
(again, broadly construed), claims to provide bona fide scientific accounts that supersede the
received human-scientific ones. This destabilising strategy is all the more effective as naturalistic
methods in the human sciences can prevail themselves of the proven success of the natural
sciences. To be sure, they are deployed on a novel terrain, and thus stretched beyond their
original perimeter of competence. But the same happened when the natural science of living
matter developed, gradually, into present-day biology: the case of a Grand Division between
the sciences of inert matter – physics and inorganic chemistry – and the sciences of living
organisms did not hold for very long.2 Whenever a new method was proposed, it was critically
examined and tested, and only after it had been shown to be sufficiently in line with existing
proven methods, and able to provide reliable, cumulative knowledge that meshed with previously acquired results, was it admitted into the club of natural scientific methods.
The same is happening under our own eyes with cognitive science qua naturalistic human
science. So that, instead of just claiming that human entities must be naturalised, on conceptual
grounds, the new science is actually attempting to show how it can be done. Henri Poincaré,
who was similarly concerned with the unity of nature, put it succinctly: “The question we must
ask is not whether nature is one, but how it is one”.3 In other words, the question for physicists
in the nineteenth century who worried about the “imponderable fluids” such as caloric,
phlogiston and so on, was not, Can we reduce or eliminate them? but rather, How do we do it?
In the case of human phenomena, the ongoing efforts split into two projects. One is to
naturalise the mind: to show that it is part of nature, that all things mental are natural states,
processes or events. The second project aims at producing a naturalistic social science. While
these are distinct projects for the obvious reason that their targets are distinct, they are related in
two ways. First, although not analytically wedded to methodological individualism, cognitive
science requires at least some principle of individual transmission4: social forces, whatever they
are, require the vehicles of individual minds – not necessarily in conscious, explicit mode – in
order to impinge on the world. Second, a crucial recent development in cognitive science
proceeds from the insight that biological evolution has provided the human mind with very
specific social skills: ours is a “social brain”, so that human society and human cognitive capacities end up like lock and key as a result of a co-evolutionary process.
Let us then first remind ourselves of how cognitive science approaches the individual mind.
The theoretical background is provided by the extraordinarily novel configuration of biology
fashioned by the formal sciences in the course of the twentieth century. It made not only
intelligible but highly productive, from the 1960s onward, the concept of the mind as a
physically instantiated computational device operating on representations. This model of the
fully formed mind typical of the adult Homo sapiens, was considerably enriched, both conceptually and empirically, by developmental psychology, as well as by the study of non-normal
or non-typical minds and non-human cognition. Then came the enlargement of the initial
agenda, consisting of such functions as reasoning, memory, language, perception, motor
control, to emotions, consciousness, self, character (along with interindividual differences,
heretofore deliberately ignored), social skills. Finally (in a sense both logical and, to some extent
at least, chronological), biology staged a spectacular return, wielding three distinct though
420

Critical naturalism for the human sciences

intimately related research programs: neuroscience, buttressed by functional neuroimaging and
computational modeling, genetics, evolutionary theory. The long-term outcome of this
monumental undertaking is a three-fold account of the mind as a highly complex informationprocessing system, of the brain as the equally complex biological system that undergirds or
enables the mind, and of the cognitive organ shaped by evolution, both as a system of mental
functions (the “evolved mind”) and as a system of neural functions (the “evolved brain”).
Filling out this very rough sketch cannot be attempted here, and a critical outlook will be
deployed shortly. But it may be worth emphasising at this point that underestimating the force
and impact of these developments would be a mistake at least as serious, and probably more, as
overestimating them. The plausibility of any form of naturalism today cannot be gauged in the
absence of an informed perspective on the ongoing naturalisation programs, which includes
taking them seriously. Liberal naturalists who are content with acknowledging their existence
without reflecting on their output may be naturalists only in name.
Let us now turn to the social sphere. At the outset a misunderstanding threatens, which must
be defused: in some traditions, any phenomenon that involves social processes is by definition not
natural. This is unhelpful: there are non-human social species giving rise to social phenomena
that are natural in the most widely accepted sense of the word (nonhuman animals are part of
nature through and through, even if we wish to distinguish different strata within the nonhuman animal world, accepting for example the notion of animal culture). We must ask
therefore, regardless of our stand on naturalism, which among the social processes in human
societies are natural in a sense shared with nonhuman societies, and which are not, and of these,
what, if anything, makes them nonnatural. In fact, the current naturalistic approaches support
the idea that the natural constitution of human individuals is intimately linked to the specific
character of human society. At any rate, it is advisable not to let a premature terminological
choice obscure the inquiry.
How does the naturalisation of the social sphere proceed? Several approaches are explored, in
interconnected ways. The first is the most abstract and removed from mainstream empirical work
in the social sciences: society is viewed as a complex network of interacting agents who are, for all
practical purposes, information processors that harbor beliefs and preferences under uncertainty.
Powerful analytic tools – sophisticated versions of decision theory, game theory and dynamical
systems – provide elaborate and enlightening accounts of social, political and economic phenomena. Of particular interest in the present context is the treatment of norms, which are not
limited to the maximisation of individual utility and include a plurality of “other-regarding” norms.
They are shown to emerge, first, as a result of gene-culture co-evolutionary processes leading to
Homo sapiens, and subsequently, in social and cultural evolution, as coordinating devices.5
The second approach is closer to a lot of present-day social science, to which it aims at
contributing a better – more detailed and realistic – model of the individual agent. The simple
model of the utility-maximising agent in economics, or the more elaborate models used in
social psychology or political science, have turned out to be deficient. The strategy of abstracting away from individual differences and imperfect performance, in the hope that deviations cancel out, ceases to be an acceptable wager in a naturalistic perspective, both because it
leads to poor results and because better models have become available. Cognitive science
purports to substitute genuine empirical knowledge for fanciful, armchair or introspective
speculations regarding such essential processes and faculties as memory, learning, reasoning,
linguistic communication, and above all decision-making. Behavior that could only be characterised in the traditional framework as irrational, can now be accounted for and seen as
rational in its own right. Predictions improve, and in some cases policy can be oriented towards
better overall outcomes.
421

Daniel Andler

The third approach concerns the genesis of social phenomena, which are purported to result
from two distinct processes: biological evolution and cultural evolution. The first throws light on
the gene-based endowment of modern humans, with the all-important feature, mentioned
earlier, that an essential part of that endowment determines basic features of human natural sociality. Cultural evolution is concerned with the emergence of particular social patterns, a highly
complex process involving a combination of construction, through learning and imitation, of
environmental constraints, and of selection according to a generalised Darwinian dynamics.
Notable examples of what this approach can bring include accounts of two of the most important
and seemingly paradoxical features of “Homo socialis”: humans are essentially co-operative, despite
the advantages of free-riding; and they can collectively adapt to an extremely wide variety of
environments, despite the advantages of being optimally adapted to a given niche.
The preceding remarks constitute no more than an exceedingly schematic and partial sketch.
They are not meant as a defense of naturalistic approaches in the human sciences. They merely
aim at conveying a sense of the field and the reasons for taking it seriously, Verstehen and other
traditional worries notwithstanding.

2 Fluctuat Nec Mergitur6: The human sciences still stand
Indeed, it may be asked, why should anyone want to resist the naturalistic pull? Naturalism
appears to be progressive and fecund, it seems to be the only truly scientific game in town, and
it does not obviously require giving up all of one’s former theoretical commitments. The first
task, though, is to ascertain whether this is indeed the case. The second is to question the
unstated assumption that naturalism itself, as deployed in the abovementioned research programs, meets with uniform success. The third is to challenge the claim to exclusive scientific
privilege. It will then be possible to strike a reasonable, albeit provisional, balance between
acceptance and rejection.
To start with, little if anything has been said of how naturalistic approaches propose to dispose
of the theoretical posits and the empirical results of the human sciences – a key question raised by
every episode in the age-old history of naturalism, but one whose stakes have risen considerably,
in proportion with the growth of science on both sides: the human sciences and the naturalistic
program centered on cognitive science. Today’s scholars differ widely on the issue. At stake are
different kinds of things: entities, like belief, social class, the market, public opinion, incentives…;
properties, like mental capacities such as linguistic competence or consciousness, or social phenomena such as institutions or inflation; purported or real facts, such as young children’s capacity
to master any first language, or liberal economies’ ability to fix prices optimally. Among them,
many appear as composites of more basic components, where the composition seems tractable in a
naturalistic setting. What the naturalist is left with are some hard cases.
Her options are: eliminate, reduce, identify. Elimination is rarely final, reduction, in practice,
rarely perfect, though often informative. Identification is left as the strategy of choice, but runs
into problems as well. In its strong form, which bears on kinds (as when water is identified with
H2O, temperature with mean kinetic energy, lightning with electric discharge in a certain setting), it often – though not always – runs into a form of the open question argument: one is left
wondering whether the target kind is truly the same as the candidate natural kind. The weak form
of identification, supervenience, may offer some ontological comfort but little scientific guidance.
So what is the naturalist to do with recalcitrant entities? What it not opened to her, as stated
earlier, is to accept anything in her ontology without ascertaining, at least in principle, its
natural status, as granted by the best current natural-scientific theories. By the same token, any
theoretical account, explanation or prediction must be shown to be at least compatible with
422

Critical naturalism for the human sciences

naturalistic constraints. However, she may also hold off, like a customs officer awaiting instructions. She need not respond here and now to the pressure that the nonnaturalist social
scientist exerts by merely resorting to one or another concept.
This prudent attitude is a first step that the naturalist can take: she shows her interlocutor in
the social-scientific mainstream some tolerance, asking only some tolerance in return. More
importantly, she makes way for a somewhat liberal understanding of her own commitment. But
more is needed to persuade her to actually broaden her perspective and instead of merely
withholding judgment, acquiesce to something a strict naturalist would reject. Three considerations may nudge her in that direction.
First, she may choose to take stock of the lack of progress in the naturalisation of a number of
key entities, such as, among the more prominent, intentionality, consciousness, normativity
(the last of which I will return to in the final section). Lack of progress is of course a matter of
judgment. Yet one cannot help but notice the sharp contrast between these recalcitrant entities
and others that give rise to concomitant conceptual and empirical fairly rapid progress.
Philosophical reflection on consciousness, to take one example, becomes increasingly arcane,
while science produces non-cumulative, nonconvergent bodies of evidence only weakly
supporting a plurality of theories. The situation for visual perception is diametrically opposite.
This leads no naturalist of any persuasion to conclude that consciousness partakes of the supernatural; but some may want to question the ability of natural science as we know it to
provide a satisfactory account of consciousness, or even to articulate a coherent research program aiming at covering it fully.
That naturalism in the “strong” sense of cognitive science and allied disciplines is “the only
(natural-scientific) game in town” is the second assumption which our naturalist may want to
reconsider, especially with respect to social science. After all, formal and quantitative methods
have been part and parcel of economics, sociology, political science, demography, history,
geography, ethnography… for decades, and have co-existed, more or less peacefully, with
descriptive, idiothetic, narrative, interpretative branches of the same disciplines. And these
formal and quantitative methods are not naturalistic in the strong sense; in fact, they can be and
often are precisely agnostic about the mechanisms that realise the behaviors they measure and
model, and arose in a distinctly anti-naturalistic moment in philosophy and human science. At
the same time, they are compatible with strong naturalism, as the first direction outlined at the
end of the preceding section (the “Gintis” orientation) shows. And they share with the natural
sciences the essential features that are supposed to set these apart from the human sciences.
Yet a third consideration might be that the “strong” naturalistic programs are themselves
hybrid affairs, from which all non-natural concepts have not been purged: cognitive psychology
and neuroscience, evolutionary anthropology, generative linguistics cannot forego intentionality, whose full naturalisation remains on the agenda, as we just reminded ourselves.
Our open-minded naturalist has reached the shores of relaxed naturalism, the mildest form
of liberal naturalism. She remains convinced that every step towards some form or other of
naturalisation contributes to our understanding, not least because it often tells us how to naturalise this or that, and sometimes shows that is can be done in several ways rather than just one.
Insofar as her training permits it, she is happy to devote herself fully to the task. She no longer
believes that science is neatly divided into two separate branches, she clearly sees that the
distance between two areas within any one of them can be greater than that between areas in
different branches. She is excited to see different perspectives brought in harmonious articulation as a result of a discovery or a leap of scientific imagination; but she is not worried about
the persistence of heterogeneous views on the same set of phenomena: she has come to espouse
scientific pluralism. Nor does she feel that the human sciences must either disappear or
423

Daniel Andler

renounce non-naturalistic concepts and methods altogether. Naturalism comes in degrees, and
the human sciences, shaken by the wave, are holding up.7

3 Can we make liberal naturalism more productive?
In fact, the human sciences are becoming stronger and more relevant to social, economic,
political issues, and naturalism has a lot to do with it. It serves them in two distinct ways. The
first is delivered, so to speak, by its “short arm”, one that provides, in a constructive manner,
ideas, concepts, methods, sometimes important results. This contribution is all the more
consequential as it is offered in a liberal spirit: a naturalist who has and shows due respect for
what the human sciences have accumulated over the years in terms of insights, observations,
conceptual frameworks, will have a better chance of being heard by scholars not or not yet
inclined to take the naturalistic turn, and of presenting her own ideas in a way intelligible
to them.
The second impact, delivered by naturalism’s “long arm”, is not so gentle. It challenges the
human sciences’ accounts, either by questioning their conceptual foundations, or by proposing
alternatives, or again by demanding evidence – more precisely, by raising the relative weight of
evidence among the criteria for assessing a given theory or framework. This critical function can
nudge the human sciences towards a greater concern for accountability, and in particular
overcome their perennial “toothbrush problem”: major theorists will only use their own
theory, and use it as the measure of all things, making cumulative knowledge difficult if not
impossible to achieve.
But there is an equally important target for the naturalist’s critical proclivity, that is, naturalism itself. One thing is to broaden the scope of naturalism so as to make it hospitable to a
wider set of entities and methods, at least on a provisional basis. Another is to tighten the criteria
for accepting a theory hatched within naturalism, and resist the temptation of judging it true or
plausible just because, if true, it would vindicate naturalism, or because it appears as the only
naturalistic theory on the market, and thus, however dubious, preferable to any non-naturalist
account. One might think this an unnecessary, and indeed somewhat insulting counsel: why
suspect naturalism of susceptibility to such biases? One reason is the confirmation bias to which
we are all prone. But another is the existence of a militant streak within scientific naturalism,
one which takes the vindication of naturalism as a theory to be the naturalist’s moral duty.8
However it be, naturalism earns its right to criticise non-naturalist approaches in the human
sciences, and more generally to take part in the conversation, only insofar as it applies to itself
the full force of critical thinking, as deployed in the sciences and in philosophy, and not restricting its analytic means to those already certified by strict naturalism. By allowing naturalism
to step, for these methodological reasons, outside of its strict, programmatic boundaries, it
becomes more than just liberal, it becomes critical and self-critical. It also coheres better to the
fundamental inspiration of naturalism in its widest construal, viz. to scrutinise with the utmost
care what is – in this case, to examine the various research programs in the relevant sciences
with sufficient attention so as to discern how they really operate, what they take for granted,
what they postulate, what they offer, and in the light of these observations to evaluate their
claim to make a genuine contribution to understanding, and to do so by naturalistic means.
At this point, then, liberal naturalism has opened up to mainstream human sciences – either
by relaxing the rules of admission into the realm of natural entities and methods, or by giving
them credit for their proven capacity to throw some light on the human sphere, or both. It has
also resolved to apply the full force of impartial rationality to naturalistic programs. While
facilitating collaboration between naturalistic and mainstream social science, these moves by
424

Critical naturalism for the human sciences

themselves cannot bring about substantial synergies. The main reason is that naturalistic research, at the present stage, is essentially theoretical and laboratory-centered, while most
mainstream human science is field-based, in the sense where it relies on real cases, be they
drawn from historical archives, ethnographic or sociological observations, statistics on real
populations or economies, accounts of individual or collective behaviors in actual situations,
descriptions of past and present political and legal systems and so on. And there generally is no
direct route from one area to the other: no question or puzzle raised by the field studies (in the
extended sense just proposed) to which naturalistic theory can directly respond; and conversely,
no issue in the field that naturalistic theory can be directly applied to. This disconnect is reminiscent of a well-identified situation regarding medicine in relation to human biology: to go
from “bench” (the biology lab) to “bedside” (the patient) requires an intermediate phase – socalled translational research – during which vocabularies become partially aligned, questions
reformulated, methods adjusted so as to ensure a fruitful two-way commerce whereby biology
becomes responsive to questions arising in medicine it hadn’t envisaged before, and conversely
medicine begins to look in directions suggested by biology that it has left unexplored. A similar
process, I submit, will allow liberal naturalism to yield its full potential in the human sciences.9

4 Why the human sciences cannot be fully naturalised
The brand of liberal naturalism that I advocate is critical in the everyday sense just made clear. It
is also critical in the Kantian sense10: it embraces the thought that its empire is not boundless. It
does so not from an abstract standpoint, or for the sake of saving some sacred value (such as
resisting the disenchantment brought about, according to people on both sides, by naturalism),
but by taking a hard look at the concrete level of scientific achievements and reflecting on their
resources and their limitations.
The critical naturalist is not alone: non- and anti-naturalists are on the same page, or so it
seems. The difference is that for the critical naturalist, the existence of limits does not discredit
naturalism; to the contrary, it brings into focus what it can achieve. Moreover, the critical
naturalist, instead of rejecting wholesale, from first principles, the naturalist perspective on
recalcitrant phenomena, follows it as far as it goes. On the other hand, unlike the radical
naturalist, he does not assume that all obstacles will eventually be cleared, just because, as the
slogan goes, “there are no miracles”. Indeed: science has no miracle solution for every problem
on the horizon, as the most cursory look at history shows again and again. Realism about
science, in the everyday sense of the word, is part of the naturalistic stance: our scientific
knowledge remains radically incomplete.
I will not pursue this train of thought, and instead will focus in the present (and last) section
on an issue of direct relevance for the human sciences – at both the individual and collective
levels, viz. behavior. Behavior may well be the most important, most salient target of the
naturalistic human sciences. Despite the (exaggerated) reports of behaviorism’s demise, an
overwhelming portion of cognitive science is couched, or can easily be translated in terms of
behavior. Of particular relevance here is the large body of research combining the tools of
decision and game theory with both evolutionary and neuro-psychological approaches of
reasoning in order to construct models of agents’ behaviors in a large set of cases. These models
operate either at the so-called “personal” level, where the agent consciously follows certain
norms and applies certain rules to the situation at hand; or at the “subpersonal” level, where the
agent’s decision is the outcome of neural processes operating according to certain constraints on
informations gathered by the agent from various sources. These neural processes implement
algorithms that lead to the actual decision, which on average leads in turn, thanks to
425

Daniel Andler

phylogenetic adaptation and ontogenetic learning, to the desired outcome – the agent’s mental
flux, during the process, is explanatorily idle.
There is much to admire in, and learn from this line of work. It does however raise two
crucial questions.
The first concerns normativity. The naturalist is correct in assuming that insofar as norms
impinge on behavior, they do so because agents are disposed to abide by them: the disposition
to respect a given norm is thus assumed to be on par with other natural dispositions, such as the
dispositions to speak and understand one’s native language, to respond to threats in such and
such a way, to imitate certain people, to go for sweet foods, to avoid stimuli that we find
disgusting, etc. Norms however don’t impinge on behavior at all in the same way as these
natural dispositions. They bind the agent while leaving it up to her to accept or ignore their
injunctions. The normativity of norms is a puzzle for the naturalist, one which he is tempted to
dissolve by denying its very existence. But as is invariably the case, elimination by fiat isn’t a
stable solution. This is made clear in cases where the agent has a reason not to abide by a given
norm, for example because another norm orients her towards a different decision: her deliberation is not about which of her dispositions will eventually win out, but about how she should
adjudicate their divergent claims. As has been often remarked, there is no way of eliminating
deontic concepts while describing normative events. The naturalist who insists that the agent is
deluding herself, the victim of a massive error, does not have the last word.
This well-known debate, to which I have given short shrift, is however of limited import for
the present discussion. It concerns an ontological issue, which leaves essentially intact a large
part of the agenda of scientific naturalism. After all, if the present uncertainties regarding intentionality, consciousness or free will have not stopped it in its tracks, normativity shouldn’t
either. The second issue is far more consequential, as it impacts the epistemological wing of
naturalism.
Models of decision-making perforce envisage types of situations. They take as input the
parameters of the situation, the goals of the agent, her subjective estimates of the outcomes of
her various possible moves. The output is the decision that the agent should take in order to
optimise her gain (prescriptive mode) or actually does take (descriptive mode). This very broad
framework is easily adjusted so as to accommodate a large variety of decision procedures,
whether automatic and subpersonal, such as those involved in linguistic understanding, or
deliberative, such as those recruited in deliberate problem solving. Early artificial intelligence
and cognitive science aimed for models for the latter kind: they meant to provide the answer to
questions of the form “What is the intelligent agent to do in this situation?”, an answer in the
form of an algorithm. The project foundered when it hit the obstacle of context: however
sophisticated, an algorithm can take into account no more than a predefined set of parameters,
one attached to the type of situation it is built to deal with. Outside the laboratory, people deal
not with a type of situation, nor even with a typical situation, but with a concrete, singular
situation. Rules provided by the algorithm are not sufficient to dictate the proper response,
which requires identifying and weighing the relevant features of this particular situation among
the myriad potential ones. In fact, before the agent can apply the rules relative to the type of
situation of which the situation at hand is a token, she must spot the relevant features and
correlatively identify the situation at hand as one of a certain type.
There are two key points to consider at the present juncture. The first is that in many cases,
there exists no context-identifying algorithm, yet people somehow manage: a lot of the time,
they do “the right thing”. The second is that there is no verification procedure, no supreme
umpire, for what makes their decision the “right” one: it is right insofar as it has survived a
comparative cross-examination by the agent and her real or notional ideal witnesses, who put
426

Critical naturalism for the human sciences

forth arguments and take into account in the fairest possible way all the objective elements in
play, including the eventual success at solving the initial problem, if there is one in a sufficiently
well-defined sense. This judgment is never final and can always be reversed.
The question then arises, of how does anyone do the right thing, absent an algorithm – a
system of rules – that delivers it? The answer cannot lie in some yet unidentified procedure that
works in all or most cases, to which the objection just raised against a system of rules would
apply. Nor can the answer lie in the miracle of a supernatural intuitive faculty. The point is that
in each case there is a perfectly natural process at work, although there is no known natural
process that works in each or even most cases: the “miracle” boils down to an inversion of
quantifiers! This is no more of a mystery than the fact that, as Neurath remarked a long time
ago,11 we have no complete physical theories of falling leaves or of forest fires, yet every episode
of a leaf falling off a tree and every forest fire fall under the jurisdiction of the physical sciences,
and moreover we can identify constraints on the possible trajectories of falling leaves and the
possible unfolding of forest fires. What runs dry in such cases is not ontological naturalism, but
epistemic naturalism.
It may still be objected that quotidian experience, as well as the partial success of formal
decision methods, whether abstract or deployed in artificial systems, provide empirical evidence
for the existence of situations to which all this does not seem to apply. These are situations in
which the context is neutralised, in the sense where it can be fully determined in advance, that
is, in fact eliminated qua context, absorbed in the set of parameters to be included in the
assessment of the situation. In such cases the agent proceeds without hesitation, and there is no
debate about the way to go. This is the limit case on a continuum stretching between full
tractability and deep intractability. But though conceptually extreme, it is pragmatically central:
thanks in part to biological evolution, and to a larger part to cultural evolution, institutional
niche construction and social learning, people go about their daily labors without having to
stop, again and again, to ponder about the situation at hand. Most of the time, the context
problem is presolved.
However, equally confirmed by quotidian experience is the recurrence of situations in which
the problem is humanly intractable, in the precise sense where there is no known tractable algorithmic procedure that determines the relevant features. In such cases, deliberation is a necessary
but insufficient first step, and is followed by, yet doesn’t simply entail, choice: the Aristotelian dyad
(bouleusis/proairesis) is at work in epistemic reasoning as it is in ethical reasoning.12
How does the existence of these situations impact the human sciences? The answer is
straightforward, once it is granted that agents’ reasons for choosing a course of action partially
determine that course of action, which in turn affect the individuals’ behavior and performances
in the long run. Whether singly or cumulatively, individual decisions then affect the trajectories
of and within societies. As in the case of falling leaves and forest fires, there are constraints on
the paths that individuals and societies can take, which naturalistic methods are adept at uncovering. But within the cone of possible paths, the actual path followed and its differences
from other possible paths are of the utmost importance to us. The human sciences thus
comprise an essential idiothetic, descriptive component that is connected, but only partially, to
the network of regularities and constraints brought to light by their naturalistic subdisciplines.
What the liberal naturalist maintains, unlike the anti-naturalist, is that both components can and
must co-exist, and that there is no fixed and fast boundary between them.
This endpoint is hardly novel, in contrast with the sea change envisaged by strict naturalism.
On large issues concerning the human sphere, calls for sea changes are a sign of philosophical
hubris. The liberal naturalist, while welcoming the contribution of his radical colleagues, parts
way with them at this juncture.
427

Daniel Andler

Notes
1 A philosopher who defends scientific naturalism can do so while heavily leaning on, and even contributing to the naturalisation project. David Papineau is a case in point: see Papineau (1993, 2002).
2 This is admittedly a crude oversimplification. Chemistry itself was once divided into two branches,
inorganic and organic, that were deemed essentially different, until the synthesis of urea by Friedrich
Wöhler in 1828 established an ontological bridge between them. Biology did not appear as clearly
continuous with the physical sciences before the mid twentieth century.
3 Poincaré (1902), Chap. IX; my italics.
4 See, for example, Nelson (1990).
5 See, for example, Gintis (2009).
6 “Tossed by the waves, she doesn’t sink: the motto of the city of Paris.
7 For a systematic exposition of this perspective, see Macarthur (2010).
8 See Andler (2010), Williamson (2011).
9 I have refrained up until now from providing examples, due to space constraints and the risk of
courting controversy. Here I make an exception, to deflect the charge of empty talk. In linguistics,
Smolensky et al.’s theory of harmonic grammar is the outcome of in-depth translational research,
allowing for the application of powerful computational ideas to real-life linguistic phenomena
(Smolensky and Legendre 2005). In paleoanthropology, Sterelny’s theory of the emergence of social/
cultural forms and advanced skills (Sterelny 2012) brings together rich field data and evolutionary
cognitive science. In social psychology, Daniel Nettle and others have similarly brought to bear life
history theory on phenomena identified by sociology (Nettle et al. 2013). These are examples, all very
different, of programs moving back and forth between the laboratory and the field.
10 The “critical” label is used, here and elsewhere, as an ordinary epithet qualifying naturalism; I also use
it for the brand of naturalism that I develop at length in Andler (2016).
11 Neurath ([1931]/1983). (esp. chap. 6: Sociology in the framework of physicalism (1931)).
12 The epistemic undecidability that I have been stressing is compounded by the ethical dimension of
most actual situations to which no routine applies; how these two sources of indeterminacy combine
and compound one another is a matter of the utmost importance that cannot be examined here. The
central point is that indeterminacy arises, or would arise, even in the absence of ethical concerns.

References
Andler, D. Context: The Case for a Principled Epistemic Particularism. Journal of Pragmatics, 35(3), (2003).
349–371.
Andler, D. Is Naturalism the Unsurpassable Philosophy for the Sciences of Man in the 21st Century? In F.
Stadler, S. Hartmann, D. Dieks, W.J. Gonzalez, T. Uebel, and M. Weber (eds.), The Present Situation in
the Philosophy of Science, pp. 283–303. Springer. (2010).
Andler, D. La silhouette de l’humain. Quelle place pour le naturalisme dans le monde d’aujourd’hui?
Gallimard. (2016).
Gintis, H. The Bounds of Reason. Game Theory and the Unification of the Behavioral Sciences. Princeton
University Press. (2009).
Macarthur, D. Taking Social Science Seriously. In M. De Caro and D. Macarthur (eds.), Naturalism and
Normativity, pp. 123–141. Columbia University Press. (2010).
Nelson, A.J. Social Science and the Mental. Midwest Studies in Philosophy, 15, (1990). 194–209.
Nettle, D., Frankenhuis, W.E., and Rickard, I.J. The Evolution of Predictive Adaptive Responses in
Human Life History. Proceedings of the Royal Society of London. Series B. Biological Sciences, 280(1766),
(2013). 1343–1352.
Neurath, O. Philosophical Papers 1913–1946. D. Reidel. (1983).
Papineau, D. Philosophical Naturalism. Blackwell. (1993).
Papineau, D. Thinking About Consciousness. Clarendon Press. (2003).
Poincaré, H. La science et l’hypothèse, Flammarion. (1902).
Smolensky, P. and Legendre, G. The Harmonic Mind: From Neural Computation to Optimality- Theoretic
Grammar. MIT Press. (2005).
Sterelny, K. The Evolved Apprentice. MIT Press. (2012).
Williamson, T. What Is Naturalism? The Stone, 5 September (2011).

428

35
HABERMAS AND
LIBERAL NATURALISM
Paul Giladi

1 Introduction
To some extent, the project of articulating Habermas’s complex understanding of the relationship between philosophy and the natural sciences as well as his critique of scientism may be
construed as problematic from the very outset. First, from a purely technical perspective, in
Knowledge and Human Interests, Habermas does not appear to make any ostensible references to
the burgeoning tradition of liberal naturalism. Second, from a purely philosophical perspective,
Habermas’s transcendental pragmatist critique of scientism does not in and of itself commit him
to liberal naturalism, so much so that Habermas himself may even resist being labeled any kind
of “naturalist” even if he tries to fuse Kant’s transcendentalism and Darwin’s naturalism. In these
respects, rather than structuring this chapter around the question, “Is Habermas a liberal naturalist?”, I would like to pose and answer the following alternative question, “In what ways can
liberal naturalism benefit from Habermasian resources?”. While Knowledge and Human Interests is
the principal site of Habermas’s reflections on philosophy and the Naturwissenschaften, I also
want to draw readers to how important aspects of his mature critical social theory play significant roles in the liberal vs. scientific naturalism debate.

2 Vivat imperator scientia naturalis
For Occidental cultural theory, the age of modernity is the totalising prioritisation of and
confidence in the normative authority of reason. Quoting Habermas, the guiding principle of
the Enlightenment was the expectation that the emancipation of natural and normative sciences
from religion “would not merely promote the control of the forces of nature, but also further
the understanding of self and world, the progress of morality, justice in social institutions, and
even human happiness”.1 Given the macrosociological dimension and scope of the project of
the Enlightenment, Weber famously argued that modernity involves the interrelation of rationalisation and disenchantment. The process of rationalisation involves humanity’s attempt to
make all features of reality intelligible, so much so that the cognitive desire to make sense of
things invariably morphs into the “desire to increase mastery, control, over every aspect of the
world”.2 However, the general process of the rationalisation of the world crucially involves
increased exercises of discursive sub-processes: developments in “substantive rationality”
DOI: 10.4324/9781351209472-40

429

Paul Giladi

involve rendering values traditionally associated with religious forms of life and value-systems
more coherent; and developments in “formal rationality” involve methods and practices that
increasingly codify and quantify attitudes and institutions.
The relationship between drives for substantive rational action and drives for instrumental/
formal action invariably causes friction within reason, to the extent that the question for
modernity would not be whether or not reason will be sovereign but which pattern of rationality
and action would emerge hegemonic in the general process of rationalisation. For Weber, the task of
sociology is to explore why instrumental/formal rationality has come to dominate in modern
Western society, and why, by consequence, has nature been disenchanted (entzaubert) and
culture faces “extirpation”.3 Construing modernity as eventually culminating in a state of
“mechanised petrification”4 in an “iron cage” (stahlhartes Gehäuse),5 Weber articulates the
connection between rationalisation and disenchantment in terms of a tragic dialectic of religion.
While, of course, providing a complete explication of Weber’s social anthropology of religion
goes far beyond the scope of this chapter, I think it would be helpful to briefly sketch his central
narrative: in an effort to satisfy nonbiological means of self-preservation, prehistorical human
beings developed fetishist religious practices. Primitive societies often tended to imbue ordinary
objects with magical significance under a form of polytheism, so much so that nature was enchanted as the living embodiment of divine beings. However, over the course of the development of
societal psychology and the ways in which human societies considered how to satisfy their cura
animarum, the fetishist framework gradually gave way to intellectualised monotheistic religions
underpinning the Abrahamic faiths. “Judaism, Christianity and Islam…sought to render suffering comprehensible”.6 This turn to rationalisation, as Weber put it, was motivated by
showing that the world “in its totality is, could, and should somehow be a meaningful
‘cosmos’”.7
With Christianity at least, the kind of systematisation of doctrine and the challenges raised to
Catholicism by the Protestant movement during and since the Reformation establish the ironic
grounds for the progressive secularisation of modern Western society. Although the rise of
institutionalised religion and its correlative theological schema led to the abandonment of
primitive fetishism, the “disenchantment of the world”8 is effected by the power of
Protestantism and Puritanism, which did not simply wish to reject papal authority and revise
Christian theology by rejecting divine mysteries. These movements also wished to construe
religion as allied with formal reason and not merely residing within the bounds of sense:
explanation-bearers were no longer an esoteric group of priests endowed with magical capacities for disclosing “mysterious, unpredictable incalculable forces”.9 If advances in physics and
chemistry since the Renaissance had not already proclaimed scientific nomothetic rationalisation as imperator, the revolutionary impact of Darwinism consolidated natural science’s suzerainty through guaranteeing the disenchantment of the world by the construal of humanity in
purely naturalistic vocabulary: “if only we wished to understand [the conditions under which we
live] we could do so at any time. It means that…we can in principle control everything by means of
calculation”.10

3 Habermas’s transcendental pragmatism in “knowledge
and human interests”
Habermas begins Knowledge and Human Interests with an excoriation of scientism, understood as
the widespread neo-positivist trend across the Western world making philosophy the vassal of
the Naturwissenschaften:

430

Habermas and liberal naturalism

Science can only be comprehended epistemologically, which means as one category of
possible knowledge, as long as knowledge is not equated either effusively with the
absolute knowledge of a great philosophy or blindly with the scientistic selfunderstanding of the actual business of research. Both equations close off the dimension in which an epistemological concept of science can be formed—in which
therefore, science can be made comprehensible within the horizon of possible
knowledge and legitimated. Compared with “absolute knowledge” scientific
knowledge necessarily appears narrow-minded, and the only task remaining is then
the critical dissolution of the boundaries of positive knowledge. On the other hand,
where a concept of knowing that transcends the prevailing sciences is totally lacking,
the critique of knowledge resigns itself to the function of a philosophy of science,
which restricts itself to the pseudo-normative regulation of established research.
Philosophy’s position with regard to science, which at one time could be designated
with the name “theory of knowledge,” has been undermined by the movement of
philosophical thought itself. Philosophy was dislodged from this position by philosophy. From then on, the theory of knowledge had to be replaced by a methodology
emptied of philosophical thought. For the philosophy of science that has emerged
since the mid-nineteenth century as the heir of the theory of knowledge is methodology pursued with a scientistic self-understanding of the sciences. “Scientism”
means science’s belief in itself: that is, the conviction that we can no longer understand
science as one form of possible knowledge, but rather must identify knowledge with
science. (K&HI: 4)
Rather than understand Habermas as a reactionary to the development of natural science over the
years, to the extent that one interprets Habermas as dismissive of natural science as a source of
knowledge and authority of justification, I think one should read this passage in the following
way: as revealing nuanced Kantian (and Hegelian) concerns about philosophy’s self-inflicted
self-renunciation.
For Habermas, the central lesson of Kantianism is that philosophy is a second-order critical
discipline. Contra positivism and scientism, philosophical engagement with the natural sciences
does not amount to adopting the methodologies and basic results of natural scientific inquiry.
Or, to put this another way, the legacy of Kantian critique is not in making metaphysics the
vassal of physics. Rather, under Kantianism, the task of philosophical science is to engage in the
positive practice of critique, namely explicating the logic and conditions of different kinds of
inquiries. Crucially, this “quasi-transcendental”11 notion of critique involves disclosing the
historically evolving presuppositions of first-order disciplines, thereby revealing the necessary
conditions for the possibility of forms of knowledge. As Habermas writes:
These systems of reference have a transcendental function, but they determine the
architectonic of processes of inquiry and not that of transcendental consciousness as
such. Unlike transcendental logic, the logic of the natural and cultural sciences deals
not with the properties of pure theoretical reason but with methodological rules for
the organisation of processes of inquiry. These rules no longer possess the status of
pure transcendental rules. They have a transcendental function but arise from actual
structures of human life: from structures of a species that reproduces its life both
through learning processes of socially organised labour and processes of mutual understanding in interactions mediated in ordinary language. These basic conditions of
life have an interest structure. (K&HI: 194)
431

Paul Giladi

Central to Habermas’s argument is his rejection of a purely representational model of inquiry –
that is, a view of inquiry as disinterested from cultural or historical situatedness. In addition to
claiming that a “view from nowhere” is incoherent given how human cognition is both
embedded and embodied through-and-through, Habermas goes further by claiming that the
transcendental rules governing the practices of inquiry are born out of our evolutionary tale,
specifically through the development of our anthropologicoepistemic interests:
Since it is posited with the behavioural system of instrumental action, this framework
cannot be conceived as the determination of a transcendental consciousness as such.
Rather, it is dependent on the organic constitution of a species that is compelled to
reproduce its life through purposive-rational action. (K&HI: 133–134)
The concept of “interest” is not meant to imply a naturalistic reduction of
transcendental-logical properties to empirical ones. Indeed, it is meant to prevent just
such a reduction. Knowledge-constitutive interests mediate the natural history of the
human species with the logic of its self-formative process (which at this point I can
only assert and not demonstrate). But they cannot be employed to reduce this logic to
any sort of natural basis. I term interests the basic orientations rooted in specific fundamental conditions of the possible reproduction and self-constitution of the human
species, namely work and interaction…Knowledge-constitutive interests can be defined
exclusively as a function of the objectively constituted problems of the preservation of
life that have been solved by the cultural form of existence as such. (K&HI: 196)
These interests both govern and are receptive to our cognitive architecture, psychological
orientations and linguistic practices: different epistemic practices, vocabularies and forms of
action have developed out of different logics of inquiry, because each particular practice works under
a particular cognitive interest: (1) an interest in instrumental control; (2) an interest in communication and (3) an interest in emancipation.12
Instrumental action involves using reason for the sake of technical achievement, specifically
achieving some particular goal of control. In this respect, instrumental reason aims at controlling/dominating the objects of one’s concern. The natural sciences are instantiations of instrumental reason, insofar as they are typified by systematic practices of nomothetic reason
aimed at subsuming phenomena under general laws. Physics, chemistry and biology are bound
together by exemplifying the way in which “the technical cognitive interest defines the framework of the empirical-analytic sciences”.13 As Michael Friedman writes, “…the possibility
of reducing all of the appearances of nature to this basis, in accordance with the law of causality,
is then ‘the condition for the complete conceptualisability of nature’”.14
However, by contrast, communicative action is not modeled on any kind of subject-object
relationship and means-end framework. This is because communicative action is the variety of
activity constituted by communicative interests, since the function of communicative action is
to interpret and to bring about the intelligibility of normative concepts such as justice and
goodness under public reason. Communicative action, therefore, is directed at ends-inthemselves and to realising an intersubjective relationship between agents as much as possible.
In this respect, the human sciences – particularly hermeneutic and historical disciplines defined
eo ipso by their concern for how best to interpret events and intentional agents – exemplify pragmatic
interests. Understood in such a way, Habermas can find kinship here with another Kantian
pragmatist, namely Wilfrid Sellars:
432

Habermas and liberal naturalism

Now, the fundamental principles of a community, which define what is ‘correct’ or
‘incorrect’, ‘right’ or ‘wrong’, ‘done’ or ‘not done’, are the most general common
intentions of that community with respect to the behaviour of the members of the
group. It follows that to recognise a featherless biped or dolphin or Martian as a person
requires that one think thoughts of the form ‘We (one) shall do (or abstain from
doing) actions of kind A in circumstances of kind C’. To think thoughts of this kind is
not to classify or explain, but to rehearse an intention.15
Before moving on to the final human interest that Habermas details, namely emancipatory interest, I would like to establish what I take to be the first of two substantive connections
between Habermas’s transcendental pragmatism and liberal naturalism. Knowledge-constitutive
interests are part of a wissenschaftlich understanding of humanity, since they are the forms in
which human cultural life is produced and reproduced – especially in modern society. However,
there is nothing reducible or eliminable about knowledge-constitutive interests. In other words,
human interests – what one may legitimately deem the steering drives of culture – are not the
kind of phenomena that are candidates for re-description and translation into the vocabulary
and grammar of the natural sciences. Crucially, recognising the irreducibility and ineliminability
of knowledge-constitutive interests does not means there is any ineffable mysteriousness or
queerness to these “quasi-transcendental” phenomena. Much in the same way that John
McDowell’s variety of liberal naturalism has argued there is no inherently anathematic connection between “first nature” (natural scientific discourse) and “second nature” (development
of moral, socio-cultural, aesthetic sensibilities),16 Habermas should not be read as claiming that
there is no room for thinking the heterogeneity of knowledge-constitutive interests is in square
conflict with the claims of natural science. Insisting that knowledge-constitutive interests are
conceptually irreducible to purely causal and descriptive kinds in no way disqualifies oneself from
being scientific or from regarding the natural sciences as authoritative ways of making sense of
things. This reveals Habermas’s nuanced critique of scientism in a way that is of particular use to
liberal naturalism:
Because science must secure the objectivity of its statements against the pressure and
seduction of particular interests, it deludes itself about the fundamental interests to
which it owes not only its impetus but the conditions of possible objectivity themselves. (K&
HI: 311)
Scientism, construed as the tendency to establish instrumental technical interests as hegemonic
over communicative interests, necessarily presupposes the grammar of the manifest image in an
effort to excise it in favor of the pure scientific image. As Robert Hanna writes,
the basic natural sciences, as rational human cognitive achievements, and also natural
scientists themselves, as fully engaging in pre-exact-scientific and trans-exact-scientific
human rationality at every moment of their conscious and self-conscious lives, are
necessarily irreducible to the physical facts known by those very sciences and those very
scientists.17
To those more rehearsed in critical theory terminology, Habermas’s transcendental claim about
the logic of inquiry amounts to an immanent critique of scientism. On this subject, I previously
alluded to Habermas’s rejection of a purely third-personal conception of inquiry, disinterested
from cultural or historical situatedness. One can now see the full force of Habermas’s critique
433

Paul Giladi

here: the harm of a purely third-personal conception of inquiry involves a radical form of
dehumanisation. Our default self-conception as geistig is erased and replaced with a hermeneutically dissonant view of human beings as unmittelbar natürliche. For, what distinguishes Geist
from mere Natur is self-consciousness and the ways in which intentional and communicative action
renders human beings as thoroughgoingly active in the world. Such a position would be illustrative of Kant’s notion of pragmatic anthropology, which crucially draws a distinction between
die Welt kennen and Welt haben: “the expressions ‘to know the world’ and ‘to have the world’ are
rather far from each other in their meaning, since one only understands the play that one has
watched, while the other has participated in it”.18 Trying to establish a “view from nowhere” as
the desideratum of inquiry goes against the very logic of inquiry, a logic which reveals just how
inquiry is mediated through-and-through by a plurality of interests.

4 The emancipatory interest: liberal naturalism and critical theory
The second substantive connection between liberal naturalism and Habermas can be explicated
by looking at some detail at the third and final cognitive interest: the emancipatory interest. For
Habermas, the emancipatory interest “aims at the pursuit of reflection”,19 conceived as the
paradigmatic act of autonomy:
I mean the experience of the emancipatory power of reflection, which the subject
experiences in itself to the extent that it becomes transparent to itself in the history of
its genesis. The experience of reflection articulates itself substantially in the concept of
a self-formative process. Methodically it leads to a standpoint from which the identity
of reason with the will to reason freely arises. In self-reflection, knowledge for the sake
of knowledge comes to coincide with the interest in autonomy…The emancipatory
cognitive interest in the undoing of repression and false consciousness corresponds to
this self-reflective learning process. (K&HI: 197, 347)
Though Habermas broke away from the first-generation critical social theory of Adorno and
Horkheimer, this passage helps indicate the extent to which Habermas retains a significant
commitment to the Frankfurt School’s Western Marxism. Like Horkheimer before him,
Habermas insists that social inquiry ought to combine rather than separate the poles of philosophy and the social sciences to explain and transform all the oppressive, repressive and marginalising circumstances that enslave human beings. The critical social theorist is, therefore, a
“phenomenologically oriented sociologist”,20 tasked with determining and revealing processes
of social development that can be viewed as misdevelopments, disorders or “social pathologies”.21 With this in mind, I now would like to develop the ways in which important features
of Habermas nuanced Marxist-Meadian-Deweyan critical social theory can be deployed in
conjunction with his transcendental pragmatism in Knowledge and Human Interests in relation to
liberal naturalism’s critical metaphysical orientation.
For Habermas, social conflict in late modernity is rooted in the struggle to resist the colonisation of the lifeworld by systems. Conflict is not now principally resulting from dissatisfaction with the material distribution of goods and services in a given society, but rather
resulting from dissatisfaction with the encroachment by systems on the lifeworld’s territory.
Between The Theory of Communicative Action and Between Facts and Norms, the language
Habermas uses to articulate how the lifeworld can and should resist the pathological effects of
juridification is primarily defensive. As he writes:

434

Habermas and liberal naturalism

The goal is no longer to supersede an economic system having a capitalist life of its
own and a system of domination having a bureaucratic life of its own but to erect a
democratic dam against the colonising encroachment of system imperatives on areas of
the lifeworld22…it is a question of building up restraining barriers for the exchanges
between system and lifeworld.23
As I understand Habermas, the way in which one can effect resistance to the system’s colonial
oppression is to act as a border-patroller and to maintain a protective barrier. Crucially, though,
Habermas appears to be committed to the claim that instrumental capitalist structures must be
accepted as having primacy over communicative ones and that the best one can hope for is to
maintain the integrity of the democratic dam.
By analogy, the way in which one can effect resistance to scientism’s colonisation of the
normative space of reasons is to act as a conceptual border-patroller and to maintain a protective
hermeneutic barrier. Recognising the hegemony of the natural sciences and nomothetic rationality’s dominance requires those wishing to resist the totalising encroachment of the space
of reasons by purely naturalistic vocabulary to erect a hermeneutic dam and maintain its
structural integrity as best as one reasonably can. Crucially, though, this defensive strategy
appears to be committed to the claim that nomothetic rationality must be accepted as having
primacy. This insight into the socio-cultural problems of late modernity, typified by the inherent
tension between capitalist drives and democratic impulses, translates rather elegantly to the
debate concerning the Placement Problem.
As Huw Price (2004) suggests, the Placement Problem can be expressed in the following way24:
1
2
3
4

All reality is ultimately natural reality.
Whatever one wishes to admit into natural reality must be placed in natural reality.
Modality, meaning, universals, norms, intentionality and so on do not seem admissible into
natural reality.
Therefore, if they are to be placed in nature, they must be forced into a category that does
not seem appropriate for their specific characters; and if they cannot be placed in nature,
then they must be either dismissed as non-genuine phenomena or at best regarded as
parasitic second-rate phenomena.25

The Placement Problem problematises where “odd” phenomena, such as norms and intentionality, might “fit” in the world described by physics, chemistry and biology. Natural
science and fundamental physics in particular, therefore, are hegemonic, to the extent that they
have become the focal points of dialectic. Why these “odd” phenomena are viewed as problematic is principally because their status as central concepts of the manifest image’s web of belief
means there is invariably foundational friction between them and the mathematisable and
quantifiable features of the scientific image. Such is the latter’s epistemic authority that the
Placement Problem, from the very outset, aims to level out the idiosyncratic dimensions of the
manifest image and the space of reasons in order for them to be deemed legitimate. This is because both
the structure and the discursive grammar of the Placement Problem frame the legitimacy of
modality, meaning, universals, norms and intentionality in terms of whether or not they can be
placed/located in the world described by the natural sciences. Notions of finding a place for mind in the
natural world and making elbow room for intentionality in the world described by physics both seem to
presuppose that one ought to accept from the very outset the vocabulary and general
Weltanschauung of thenatural sciences, and then find some meaningful and coherent way of fitting
in phenomena such as intentionality and normativity into that nomothetic picture.
435

Paul Giladi

Such a way of thinking about reality goes some way to explain why exactly the Placement
Problem grips the philosophic imagination with such force: rational activity is exclusively articulated in terms of the kind of inferential patterns definitive of purely analytical thinking,
namely the kind of thinking symptomatic of instrumental reason. This, in turn, leads to conceiving of the space of reasons and the space of nature as fundamentally in tension with another,
and to regarding the manifest image and scientific image as metaphilosophical antagonists.26
Conceived in this way, the vocabulary of the ideal scientific image becomes epistemically
authoritarian and imperialistic by forcing other forms of enquiry to adopt the discursive recourses and grammars of formal disciplines that are different in various ways to the manifest
image’s web of meanings. Scientism, therefore, as a particular mode of disciplinary ideology,27 is
diagnosed as having made the following error described by Windelband:
[T]he failure to recognise the autonomy of individual provinces of knowledge.28
Recognising the autonomy and heterogeneity of the normative space of reasons in no way
entails conceiving of intentionality, et al. as “imaginary skyhooks”.29 On the contrary, it deepens
our way of viewing reality as intelligible by doing justice to our geistige Einstellung, our status as
self-interpreting amphibians engaging in multifaceted modes of sense-making.30
Habermas himself recognised the deficiency of overly defensive attitudes to systemencroachment, and the way he shifts to a far more positive and ambitious model of resistance in
Between Facts and Norms is one which can and should be extended to the goal of decolonizing the
normative space of reasons from scientistic encroachment. For Habermas in Between Facts and Norms, if
one is to resist and eventually overcome juridification, one must develop deliberative democracy, in which legal power can be rooted in the communicative power of the lifeworld,
especially a well-functioning public sphere and civil society.31 Traversing “the long march
through the institutions”32 is progressively transformative, because debunking the legal positivist framework in favor of a discourse theory of law involves combatting and reversing the
unofficial circulation of power in constitutional democracies. The official circulation of power
in a constitutional democracy involves the public voting and providing input to legislative
assemblies; legislative assemblies then makes laws; the executive enacts these laws; and the
judiciary reflects on these laws in cases of conflict. The unofficial circulation of power, by
contrast, involves political parties, etc. manipulating the public. For Habermas, “in a perceived
crisis situation”,33 the flow of power can be reversed to its official state once the public become
actively aware of its unofficial circulation. This form of social consciousness reveals how one no
longer deems current frameworks as rationally satisfying, thereby compelling agents to radically
revise their socio-political sense-making practices.
By analogy, to resist and eventually get over scientism involves combatting and reversing the
circulation of epistemic power. Paraphrasing Walter Mignolo, this quasi-decolonial way of
thinking is “nothing more than a relentless analytic effort to understand, in order to overcome the
logic of [epistemic] coloniality underneath the rhetoric of modernity, the structure of management and control that emerged out of the transformation of the [epistemic] economy”.34 From
this perspective, scientific naturalism is guilty of a cognitive variety of imperialism, one which is the
theoretical equivalent of Iris Marion Young’s concept of cultural imperialism:
In societies stamped with cultural imperialism, groups suffering from this form of
oppression stand in a paradoxical position. They are understood in terms of crude
stereotypes that do not accurately portray individual group members but also assume a
mask of invisibility; they are both badly misrepresented and robbed of the means by
436

Habermas and liberal naturalism

which to express their perspective. Groups who live with cultural imperialism find
themselves defined externally, positioned by a web of meanings that arise elsewhere.
These meanings and definitions have been imposed on them by people who cannot
identify with them and with whom they cannot identify.35
In the 1990s, the politics of difference focused on questions concerning nationality, ethnicity
and religion. Under this approach, the value of cultural distinctness is essential to individuals and
not something accidental to them: their personal autonomy depends in part on being able to
engage in specific cultural practices with others who identify with one another as in the same
cultural group. For Young, most modern societies contain multiple cultural groups some of
which unjustly dominate the state or other important social institutions, thus inhibiting the
ability of minority cultures to live fully meaningful lives in their own terms. The dominant
group in society can limit the ability of one or more of the cultural minorities to live out their
forms of expression. In other words, the dominant culture threatens to swamp the minority
culture to the extent that particular cultural practices and different hermeneutic spheres – ways
in which members of cultures interpret their experiences – are crowded out or erased. Under
this analogy, the concern about scientific naturalism is that the vocabulary of the ideal scientific
image becomes epistemically authoritarian and imperialistic by forcing other forms of inquiry to
adopt the discursive recourses and grammars of formal disciplines that are fundamentally different in
various ways to the manifest image’s “web of meanings”.
For Habermas, the lack of respect for the peculiarity and sui generis features of the normative
space of reasons is a crisis of communication: the scientistic model remains locked in the viewpoint
of instrumental reason and is therefore inhibited from radically revising the very notion of how sense-making
ought to be constituted and practiced. Given the difference between natural science and philosophy in
terms of how they respectively make sense of things, it would be incorrect to suppose that natural
science and philosophy should be understood in terms of a Geistig hierarchy. This is because the way
in which natural science makes sense of things is so different to the way in which philosophy makes
sense of things: conceived in this way, one ought not to regard natural science and philosophy as
rival forms of intelligibility competing with one another to best satisfy our desire for understanding
our world. On the contrary, they should be seen as complementary reflective practices, practices
which are jointly indispensable for adequately and holistically engaging with our environment:
Reason’s interest in emancipation, which is invested in the self-formative process of
the species and permeates the movement of reflection, aims at realising these conditions of symbolic interaction and instrumental action. (K&HI: 210–211)
Reflection on our discursivity illuminates the particular kind of amphibian epistemic architecture we have for experiencing the world from our Geistig perspective. To quote Bernard
Williams here, who elegantly expresses a similar claim: “…I take philosophy to be, part of a
more general attempt to make the best sense of our life, and so of our intellectual activities, in
the situation in which we find ourselves”.36 For all of the indisputably important and impressive
noetic achievements of the natural sciences, the march to scientism constitutes a type of “selfrenunciation” and a failure of rationality.

Notes
1 Habermas (1997, 45).
2 Breen (2016, 9).

437

Paul Giladi
3 Karlberg (1980, 1176).
4 PE: 124.
5 PE: 123. “Iron cage” is Talcott Parsons’s translation of Weber’s term. However, arguments have been
made that “a shell as hard as steel” is in fact a better rendering. For further on this debate, see Baehr
(2001) and Chalcraft (1994).
6 Breen (2016, 10–11).
7 SPWR: 281.
8 SV: 13.
9 Ibid.
10 (Ibid., 12–13).
11 K&HI: 194.
12 Viz:
There are three categories of processes of inquiry for which a specific connection between
logical-methodological rules and knowledge-constitutive interests can be demonstrated. This
demonstration is the task of a critical philosophy of science that escapes the snares of positivism.
The approach of the empirical-analytic sciences incorporates a technical cognitive interest; that
of the historical-hermeneutic sciences incorporates a practical one; and the approach of critically
oriented sciences incorporates the emancipatory cognitive interest. (K&HI: 308).
13
14
15
16
17
18
19
20
21
22
23
24

K&HI: 178.
Friedman (2013, 82).
Sellars (1991, 39–40).
Viz. McDowell (1997).
Hanna (2014, 756).
APPV: [120], 4.
K&HI: 198.
T&CT: 192.
Viz. Honneth (2007, 4).
TCA II: 364.
Habermas (1992, 444).
Price’s Placement Problem owes much to Jackson (1998), where it is dubbed “The Location
Problem”, although Price endeavours to distinguish them (Price 2013, 27n).
25 I have chosen to use “phenomena” here, as I am not keen on using expressions such as “things” or
“entities”, since they risk reifying normativity et al.
26 Price also argues that a key presupposition of The Placement Problem is what he calls “the Semantic
Ladder”, which derives from representationalism. This is “the assumption that the linguistic items in
question ‘stand for’ or ‘represent’ something non-linguistic…This assumption grounds our shift in
focus from the term ‘X’ or concept X, to its assumed object, X” (Price 2013, 9). This assumption is
required for the Placement Problem, for without it we cannot transition from merely studying human
linguistic practices (such as ethics talk), to considering their implications for the world. So, one may
evade the Problem by refusing to climb the Ladder.
27 See Giladi (Forthcoming 2022).
I think it is important to note here that construing scientism as ideology ought not to use science
as a scapegoat for the pathological features of capitalism. Once one sees that pragmatic realism in
philosophy of science does not entail – and in fact, strictly speaking, undermines a crass, nonNeurathian unity of science thesis, “scientism” just becomes a chimera. Given this, the following pertinent question arises: “why, from a diagnostic perspective, does scientism still persist?”
Scientism is, therefore, peculiar, because it persists despite resting on implausible grounds, since
“the omnipresent neo-Pythagoreanism of contemporary science is surely not adequately justified by its empirical successes” (Dupré 1995, 224). I think a particularly compelling answer

438

Habermas and liberal naturalism
to this question involves explaining scientism’s persistence in terms of scientism’s status – not
science’s status – as the theoretical concomitant of the kind of social pathologies caused by the
ideological exercise of formal reason in capitalist modes of production.
28 Windelband (1980, 171).
29 Baker (2013, xxii).
30 The use of “self-interpreting” is a nod to Charles Taylor; the use of “amphibian” is a nod to Hegel; and
the use of “sense-making” is a nod to Adrian Moore.
31 The Democratic Principle: “Only those statutes may claim legitimacy that can meet with the assent of
all citizens in a discursive process of legislation that in turn has been legally constituted” (BF&N: 110).
32 TSTPS: 210.
33 BF&N: 380.
34 Mignolo (2011, 10).
35 Young (1990, 59).
36 Williams (2006, 182).

References
Baehr, P. The “Iron Cage” and the “Shell as Hard as Steel”: Parsons, Weber, and the stahlhartes Gehäuse
Metaphor in ‘The Protestant Ethic and the Spirit of Capitalism’. History and Theory, 40, (2001).
153–169.
Baker, L.R. Naturalism and the First-Person Perspective. Oxford University Press. (2013).
Breen, K. Under Weber’s Shadow: Modernity, Subjectivity and Politics in Habermas, Arendt and MacIntyre.
Routledge. (2013).
Chalcraft, D. Bringing the Text Back In: On Ways of Reading the Iron Cage Metaphor in the Two
Editions of The Protestant Ethic. In L.J. Ray and M. Reed (eds.), Organizing Modernity: New Weberian
Perspectives on Work, Organisation and Society, pp. 16–45. Routledge. (1994).
Dupré, J. The Disorder of Things: Metaphysical Foundations of the Disunity of Science. Harvard University
Press. (1995).
Friedman, M. Philosophy of Natural Science in Idealism and Neo-Kantianism. In K. Ameriks, N. Boyle,
and L. Disley (eds.), The Impact of Idealism: The Legacy of Post-Kantian German Thought. Volume 1:
Philosophy and Natural Sciences, pp. 72–104. Cambridge University Press. (2013).
Giladi, P. Scientism as Ideology; Speculative Naturalism as Qualified-Decoloniality. In L. Corti and J-G.
Schülein (eds.), Life and Cognition: Understanding Nature between Classical German Philosophy and
Contemporary Debates. Springer. (Forthcoming 2022).
Habermas, J. Knowledge and Human Interests. Trans. J.J. Shapiro. Beacon. (1971).
Habermas, J. The Theory of Communicative Action. Vol. 2: Lifeworld and System. Trans. T. McCarthy.
Beacon. (1987).
Habermas, J. The Structural Transformation of the Public Sphere: An Inquiry into a Category of Bourgeois Society.
Trans. T. Burger and F. Lawrence. MIT Press. (1989).
Habermas, J. Postmetaphysical Thinking. Trans. W.M. Hohengarten. MIT Press. (1992).
Habermas, J. Between Facts and Norms: Contributions to a Discourse Theory of Law and Democracy. Trans. W.
Rehg. MIT Press. (1996).
Habermas, J. Modernity: An Unfinished Project. In M.P. d’Entrèves and S. Benhabib (eds.), Habermas and
the Unfinished Project of Modernity, pp. 38–58. MIT Press. (1997).
Hanna, R. Husserl’s Crisis and Our Crisis. International Journal of Philosophical Studies, 22, (2014). 752–770.
Horkheimer, M. Traditional and Critical Theory. In M. O’Connell (ed. and trans.), Critical Theory: Selected
Essays by Max Horkheimer, pp. 188–243. Continuum Press. (1999).
Jackson, F.C. From Metaphysics to Ethics: A Defense of Conceptual Analysis. Oxford University Press. (1998).
Kant, I. Anthropology from a Pragmatic Point of View. Ed. and Trans. R.B. Louden. Cambridge University
Press. (2006).
Karlberg, S. Max Weber’s Types of Rationality: Cornerstones for the Analysis of Rationalisation Processes
in History. American Journal of Sociology, 85, (1980). 1145–1179.
McDowell, J. Mind and World, 2nd edition. Harvard University Press. (1997).
Mignolo, W.D. The Darker Side of Western Modernity: Global Futures, Decolonial Options. Duke University
Press. (2011).

439

Paul Giladi
Price, H. Naturalism without Representationalism. In M. De Caro and D. Macarthur (eds.), Naturalism in
Question, pp. 71–88. Harvard University Press. (2004).
Price, H. Expressivism, Pragmatism, and Representationalism. Cambridge University Press. (2013).
Sellars, W. Science, Perception and Reality. Ridgeview Publishing. (1991).
Weber, M. The Social Psychology of the World Religions. In H.H. Gerth and C.W. Mills (eds.), From
Max Weber: Essays in Sociology, pp. 279–313. Routledge. (1948).
Weber, M. The Protestant Ethic and the Spirit of Capitalism. Trans. T. Parsons. Routledge. (1992).
Weber, M. Science as a Vocation. In D. Owen and T.B. Strong (eds.), and R. Livingstone (trans.), The
Vocation Lectures, pp. 1–31. Hackett Publishing. (2004).
Windelband, W. Rectorial Address: History and Natural Science. History and Theory, 19, (1980). 169–185.
Young, I.M. Justice and the Politics of Difference. Princeton University Press. (1990).

440

36
STRAWSON AND NONREVISIONARY NATURALISM
Hans-Johan Glock

I am of conciliatory temper – which sometimes extends,
in philosophy, into an attempt to reconcile views which appear to be
sharply opposed to each other.
(Strawson 2011 [1998a], 246–247)

1 Introduction
In 1922, R.W. Sellars proclaimed: “We are all naturalists now” (1922, i). On the other hand,
concerning the American philosophical scene in the aftermath of WWII Wilfrid Sellars observed: “As for Naturalism. That, too, had negative overtones at home. It was as wishy-washy
and ambiguous as Pragmatism. One could believe almost everything about the world and even
some things about God, and yet be a Naturalist” (1979, 2). In a similar vein, Ernest Nagel
cautioned: “the number of distinguishable doctrines for which the word ‘naturalism’ has been
counted in the history of philosophy is notorious” (1954, 3). This diagnosis is even more
apposite today (see also Keil 2006). Naturalism has become the biggest “research program” in
town. In the wake of Quine, many contemporary philosophers feel downright contractually
obliged to profess allegiance to some version of it. Alas, there are almost as many conceptions of
naturalism as there are proponents. In the first instance, one can distinguish three types of
naturalism.
•
•
•

Ontological naturalism equates what exists or is real with the spatiotemporal world;
Epistemological naturalism insists that any genuine knowledge must be capable of being
formulated through theories of (natural) science;
Metaphilosophical naturalism claims that philosophy is a branch of or continuous with
(natural) science.

On the assumption that the spatiotemporal order consists exclusively of physical phenomena
like matter and energy which are subject to the laws of physics, ontological naturalism turns into
a type of physicalism. But there is also an epistemological/methodological variant of ontological
naturalism-cum-physicalism. Instead of specifying directly what categories of things qualify as
DOI: 10.4324/9781351209472-41

441

Hans-Johan Glock

natural/physical, it delegates that task: only those entities exist that feature in the explanations of
either natural science in general or of physics as the most fundamental natural science.
Ontological, epistemological and metaphilosophical naturalism can all take on either an
eliminativist or a reductionist form. Faced with apparent counterexamples – entities beyond the
natural world described by physics, knowledge claims of a non-scientific kind or philosophical
methods that do not rely on science – a naturalist has two options. She can either dismiss them as
spurious; or she can try to show that on closer scrutiny they boil down to naturalistically respectable phenomena – to empirical methods, scientific findings and spatio-temporal entities,
respectively. It is exclusively this reductionist option, however, which fuels the ubiquitous
projects of naturalising phenomena such as intentionality, meaning, consciousness or morality. The
aim of such an enterprise is to demonstrate that an apparently non-natural phenomenon is real
only because it is really something else (Fodor 1987, 98), namely something which is part of the
natural order and can therefore be accommodated within science. By the same token, the discipline dealing with the phenomenon – for example, logic, epistemology, metaphysics – will be
transformed into a branch of empirical science – for example, psychology, biology or physics.
In Scepticism and Naturalism Strawson characterised his position as a form of naturalism. Not
even in that work, however, did he subscribe to any of the views that my fairly standard taxonomy recognises. That at any rate is what I shall argue in the sequel. But I shall also explore
whether there are further types of naturalism which are more congenial to Strawson’s philosophical outlook, such as “liberal naturalism”. My verdict will be qualified: Strawson’s naturalism, as
far as it goes, is anthropological instead of scientific, and descriptive rather than revisionary,
shunning both eliminativism and reductionism. But, I shall urge, it is none the worse for that. For
Strawson’s approach is more sober and realistic than these mainstream naturalisms.

2 Strawson and naturalism: an initial checkup
A thumbnail sketch of Strawson’s achievements could run as follows.
Strawson, P.F. (1919-2006) British philosopher, the leading member of the later
phase of post-war linguistic philosophy at Oxford. Strawson made seminal contributions to philosophical logic, metaphysics, epistemology and meta-philosophy. He
developed austere and abstract arguments in a lucid and elegant fashion. His early
writings form part of so-called ‘ordinary language philosophy’ in so far as they criticize
orthodoxies of logical analysis by reference to ordinary use. His later writings are
constructive, and have led linguistic philosophy back to metaphysics along Kantian
lines. During his final years, he tried to combine this ‘descriptive metaphysics’ in the
spirit of Aristotle and Kant with a ‘naturalism’ inspired by Hume and Wittgenstein. In
spite of these transformations, there are two abiding concerns in his work. One is
describing the most general and pervasive features of human thought about the world,
in particular the operations of reference and predication. The other is promoting a
non-reductive, ‘connective’ type of conceptual analysis. Both come together in
Strawson’s defence of the existence of universals and intensions (meanings) against
nominalism and extensionalism (see Glock 1999; see also Snowdon 1998; Sen and
Verma 1995; Brown 2006).
Before turning to Scepticism and Naturalism, we should note that in his other writings Strawson
repudiates all the aforementioned forms of naturalism. The context is his abiding dialogue with
Quine, the undisputed prophet of contemporary naturalism (see Glock, 2003a, 23–30).
442

Strawson and non-revisionary naturalism

Metaphilosophical Naturalism
On the basis of rejecting the distinction between analytic (conceptual) and synthetic (factual)
propositions, Quine claimed that proper philosophy is “part of”, or “continuous” with science.
Together with Grice, Strawson wrote an influential defence of the analytic/synthetic distinction
(2011 [1956]). In keeping with this semantic stance, he propagated conceptual analysis and its
extension – descriptive metaphysics – as the proper method of philosophy. And he explicitly
contrasted that conception with Quine’s “scientific philosophy” (e.g. 2011 [1990]). Connective
analysis and descriptive metaphysics do not emulate natural science by trying to describe and
explain reality on the basis of experience. Nor do they purport to capture essential features of
reality independently of thought and experience, by contrast to revisionary metaphysics. Instead,
they reflect on and explicate our extant conceptual scheme – they are an exercise in “conceptual
self-consciousness” or “self-understanding” (1966, 44; 2011, 225; see Heyndels 2020: 3–4).

Epistemological naturalism
Philosophy as propagated and practiced by Strawson aspired to a non-scientific understanding of
our conceptual scheme. For this reason alone, he had no truck with the scientistic slogan that
there is no knowledge other than that of natural science. But his animadversions to epistemic
naturalism reached deeper. He harbored a life-long devotion to the arts, especially literature. In
his youth he was especially interested in literary studies and dreamt of becoming a poet. He
continued to be fastidious about style and to hold that beauty matters in its own right (see 2011
[1998], 228; Ferdinand 2008, 2, 4, 11). As a philosopher, he defended the idea that the humanities
as well as common sense and experience vécu provide us with insight and knowledge (2011 [1990],
175–177). The suggestion that talk about architectural and literary styles is dispensable because it is
only required for purposes of “criticism” provoked him into abandoning his customary mild
manner: “From such philistinism as this we can only avert our eyes” (1997, 35).

Ontological naturalism
On the one hand, Strawson’s seminal work on reference incorporates Wittgenstein’s insight that
the meaning of an expression must not be confused with the object, if any, it stands for. This
militates against postulating supernatural referents for meaningful expressions that lack a physical
referent. Similarly, Strawson’s Kant-inspired attack on the idea of a soul-substance (1974, Chap.
8) remains one of the most astute critiques of Cartesian dualism. His criticisms of Kantian
phantasies of a noumenal realm (1966) is equally compelling. On the other hand, Strawson
explicitly defended reference to abstract entities, including entities of an “intensional” nature,
such as universals and meanings. In later years he had no compunctions about subscribing to a
form of “Platonism”, though a “demythologised” one. This kind of Platonism recognises intensions as objects in a thin logico-grammatical sense, as subjects of meaningful predication, while
repudiating the idea that they are entities inhabiting a realm beyond space, time an causation just
as material objects inhabit the spatiotemporal world (see 2011 [1990], 175; 1997, 2–3, 78, 83–84).
On all three scores, Strawson directly opposed both eliminativist and reductionist naturalism.
Philosophical reflection on our conceptual scheme is both legitimate and sui generis; common sense,
historical and humanist disciplines provide genuine knowledge that is not available through natural
science; the existence of intensions can neither be denied nor can it be analysed in terms of spatiotemporal phenomena recognised by physics. But in order to understand Strawson’s attitude toward
naturalism adequately, we need to turn to his most explicit and sustained treatment of the matter.
443

Hans-Johan Glock

3 “Reductive” vs. “non-reductive” naturalism
Skepticism and Naturalism revolves around a distinction between “strict”, “reductive” or “hard”
naturalism on the one hand, “catholic”, “liberal”, “non-reductive” or “soft” naturalism on the
other (1985, 2; unless otherwise specified, all references are to this text). For the most part,
Strawson refers, to, respectively, “reductive” and “non-reductive naturalism”. He condones
non-reductive naturalism and defends it in areas such as morality, secondary qualities, the mindbody problem and “intensional entities” (69), that is, universals (properties) and meanings.
What precisely does Strawson mean by non-reductive naturalism? Answering that question
is complicated by the fact that he opposes it to a daunting variety of positions, concerning
different topics, and in several ways, as he later admitted (2011 [1998a], 242–243). As we shall
see, it is even moot whether his position is a bona fide version of naturalism. The common
denominator of the positions he opposes to non-redcutive naturalism is “skepticism” in the
sense invoked in the book’s title. Traditional skepticism denies the possibility of knowledge. By
contrast, this more general skeptical point of view challenges fundamental aspects of our common
sense worldview.1
…as against all these reductivist (or skeptical) stances, as also against skepticism more
traditionally and properly so-called, I have tried to set up another kind of
Naturalism—a non-reductive variety—which recognizes the human inescapability
and metaphysical acceptability of those various types of conception of reality which
are challenged or put in doubt by reductive or traditionally sceptical arguments (68,
see also 40)
We need to distinguish the contrast between non-reductive naturalism and reductive naturalism
from that between non-reductive naturalism and skepticism. Regarding morality, secondary
qualities, mental phenomena and intensions, skepticism can be a straightforward manifestation of
naturalism. But it is not reductionism. What Strawson combats is primarily a version of eliminative
naturalism. At stake is the very existence of phenomena recognised by common sense: moral responsibility, secondary qualities, conscious experience, properties and meanings. One might
defend the epithet “reductivist” on the grounds that reductive naturalism finds these phenomena
wanting precisely on the grounds that they are not reducible to a more scientifically acceptable,
physicalist base. However, this exculpation ignores an important point. With the partial exception
of the mind-brain identity thesis, what Strawson’s text confronts are not ingenious attempts to
reduce the contested phenomena to a natural base, but their outright “skeptical” denial.
Strawson’s response to that denial revolves around a distinction between “two standpoints” –
the “participant” or “involved” perspective of everyday life vs. the “objective” or “detached” one
of science which abstracts from the experience of human individuals. Reductive naturalism insists
that only adopting a “purely objective” or “purely naturalistic” standpoint reveals the true nature
of reality. Non-reductive naturalism does not dispute that we sometimes can, or even have to,
adopt an objective standpoint. Yet it denies that this entails that the phenomena that we acknowledge and naturally respond to from the participant standpoint are illusory. Strawson tries to
steer a course between the Scylla of reductive naturalism and the Charybdis of super-natural
“myths” or “fictions”, through what he calls a “relativizing move” (45–53). He rejects the apparently mandatory choice between these two perspectives. The idea of a “metaphysically absolute…superior standpoint” from which the true nature of the world reveals itself is an “illusion”
(38, 52–53). Neither the participant nor the detached perspective provide such a stand-point, yet
both have a role to play.
444

Strawson and non-revisionary naturalism

Strawson’s relativising move plays out differently in various areas. As regards morality, he is
concerned with our reactive attitudes toward the good or ill will of fellow humans and with our
practice of holding them responsible for their intentional actions. By “phenomenal qualities” he
means the sensory qualities of the objects of perception, such as their colors and smells. In both
cases, we normally adopt the participant perspective of everyday experience; yet we can or even
must adopt the objective perspective of the natural, psychological and social sciences (42),
though only temporarily and on special occasions (Chap. 2). As regards the mind, Strawson
rejects a straightforward double-aspect version of the “[mind-brain] identity theory”, according
to which “one and the same event” can be characterised both in “biographical” and in
“physiological” terms. But he accepts that
what happens to the subject is, from the subjective or biographical point of view, his
undergoing the conscious experience and, from the physical scientist’s or physiological
point of view, the occurrence in him of the physiological event. Nevertheless, the two
are not identical. Rather, the physiological event is the ‘physical basis or realization’ of
the conscious experience (66–67; [my emph]).2
Finally, as regards “intensional entities” (meanings, universals, rules) we can embrace either
“realism”, the “inside view” of “speakers and thinkers”, or “nominalism”, the “externalist or
strictly or reductively naturalist standpoint”. According to the latter, nothing exists outside of
the spatiotemporal world; our impression that we are contemplating abstract entities or being
compelled by rules are nothing but the result of behavioral dispositions (see 68, 78–79, 93).
Note, however, that Strawson doubts that the kind of “reconciliation” of these standpoints that
he propagates for morality and perceptible qualities can be achieved with respect to intensions
(94–95; see also 1997, 62–63).
Concerning knowledge, skepticism takes the form of scepticism. The latter is in no way a
reductive project; it bluntly disputes the existence and possibility of knowledge.
Correspondingly, as regards the possibility of knowledge, non-reductive naturalism contrasts
not with reductive naturalism, but with other strategies for responding to scepticism. Such antisceptical strategies can be divided into direct and indirect ones (Glock 2016, 292–293). The
former are foundationalist. They try to meet the sceptical challenge through identifying beliefs
that are proof against skeptical doubts. The latter reject the skeptical challenge according to
which we need to demonstrate that our purported knowledge is irrefutable or rests on indubitable foundations. To proponents of the indirect strategy, the “scandal of philosophy” is
not that a proof of the existence of the external world is yet to be given – as Kant felt – but that
“such proofs are expected and attempted again and again”, as Heidegger remarked (see 24).
Indirect responses in turn divide into several types. Two of them make an appearance in
Strawson: “non-reductive” naturalism on the one hand, transcendental arguments on the other.
Strawson had revived transcendental arguments in Individuals and Bounds of Sense. One type of
transcendental argument is elenctic. It aims to show that skeptical doubts are incoherent or selfrefuting, because the skeptic questions the validity of conceptual scheme which is needed to
state the skeptical problem. He himself employs concepts which make sense only on the assumption of conceptual connections he rejects. Therefore, the skeptical position could not be
stated unless it were unfounded (1959, 35, 105–109). For instance, scepticism about other
minds founders because the possibility of ascribing experiences to oneself presupposes the
possibility of ascribing them to others.
In Scepticism and Naturalism, Strawson conceded that transcendental arguments only establish
connections within our conceptual scheme, not anti-skeptical conclusions to the effect that we
445

Hans-Johan Glock

possess secure knowledge of a mind-independent reality.3 A naturalism inspired by combining
Hume and Wittgenstein is supposed to fill the gap (10). While skeptical arguments are irrefutable, they are also “idle” (19, 28). They cannot persuade us; for due to our natural dispositions, we cannot help believing, for example, in the existence of material bodies and other
minds. Strawson’s naturalism is precisely “the rejection of the invitation” to supply “the reasoned rebuttal that the skeptic perversely invites” (12). The obvious common denominator in
non-reductive naturalism as opposed to reductive naturalism and non-reductive naturalism as
opposed to scepticism is non-revisionism: Human nature prevents us from abandoning the
convictions, attitudes and practices opposed by skepticism. There is also an important difference. The reality of moral and phenomenal properties can legitimately be viewed from two
standpoints – objective and participant; in more complex ways, this holds for the conscious
experiences and intensions as well. By contrast, there is no analogous liberty concerning our
commitments to the existence of material bodies and to the canons of induction. There simply
is no standpoint, Strawson insists, from which we could view the external world as illusory or
completely chaotic. Hence, the relativising move is not available for our most basic assumptions
about the existence and regularity of the external world.

4 A naturalistic turn?
Scepticism and Naturalism marks a transition in Strawson’s oeuvre. Some commentators have
gone further and diagnosed a “naturalistic turn” (Stern 2003), in line with the general turn
analytic philosophy has taken following Quine. However, there are two reasons for correcting
that diagnosis.
The first is straightforward. The label “naturalism” only makes an appearance in Scepticism
and Naturalism. But the Humean idea that certain convictions and attitudes attacked by skepticism are inexorably entrenched is far from novel. It was anticipated by Strawson’s influential
discussion of the inescapability of reactive attitudes in “Freedom and Resentment” (1974
[1962]). Going back further, a little-noticed debate took place between Strawson and Wesley
Salmon in 1957–1958 (see Heyndels 2020, 134). Scepticism about induction questions the
rationality of predicting the future on the basis of the past. In Introduction to Logical Theory (1952,
chap. 9) Strawson had argued that inductive reasoning is not a method for predicting the future
which can be assessed for its rationality, since it defines what counts as rational prediction: we
call a prediction reasonable precisely if it is supported by previous experience.
According to Salmon (1957), this implies the inacceptable thesis that inductive beliefs are
merely conventional and, by implication, a matter of arbitrary choice. Referring to Hume,
Strawson (1958) pleaded not guilty: insisting on the impossibility of justifying our canons of
induction is compatible with acknowledging that we are naturally, prerationally committed to
them and therefore have no choice as to whether or not to “adopt” them. Strawson’s initial
treatment of induction qualifies as an indirect response to scepticism, and his defence against
Salmon is non-reductive naturalism avant la lettre.
The second reason to demur from the diagnosis of a naturalistic turn in Strawson’s oeuvre is
more complicated. Strawson summarises the move from “transcendental” to “naturalist” strategies against scepticism as follows:
Having given up the project of wholesale validation, the naturalist philosopher will
embrace the real project of investigating the connections between the major structural
elements of our conceptual scheme (1985, 19).

446

Strawson and non-revisionary naturalism

But naturalism, as Strawson here describes it, does not differ from the non-reductive, connective
analysis that always propelled his descriptive metaphysics. Furthermore, the move from transcendental validation to naturalistic description will simply change the topic, unless the latter also
offers a response to the skeptical challenge. And so it does, as Strawson emphasises. This naturalistic response even shares an important feature with the transcendental one, namely being
indirect. Skeptical doubts are not refuted by reference to allegedly indubitable beliefs; instead,
they are repudiated on the grounds that they imply abandoning categories that are indispensable
to human thought (see below).
Such continuities militate against the diagnosis of a turn, a decisive change of direction. They
leave open the question of whether Strawson’s general approach, including his trademark
project of descriptive metaphysics, had been naturalistic all along. However, the label notwithstanding, descriptive metaphysics as presented in Individuals does not just provide an inventory of our actual conceptual scheme. It also adopts a validatory stance. Strawson writes of
commonsensical beliefs in the primacy of material bodies and persons:
It is difficult to see how such beliefs could be argued for except by showing their
consonance with the conceptual scheme which we operate, by showing how they
reflect the structure of that scheme. So if metaphysics is the finding of reasons, good,
bad or indifferent, for what we believe on instinct, then this has been metaphysics.
(1959, 247)
Why can one argue for these beliefs in this fashion? Because certain features of our conceptual
scheme are indispensable, and hence immune to the doubts of skeptics and the reforms of
revisionary metaphysicians. Description yields validation, since some concepts do not just in fact
play a central role in our conceptual scheme, they must play such a role in any conceptual
scheme we are capable of understanding.
Prima facie these observations rehabilitate the idea of a naturalistic turn. For in Scepticism and
Naturalism Strawson explicitly contrasted “validatory or revisionary” with descriptive metaphysics.
to establish the connections between the major structural features or elements of our
conceptual scheme--…may well seem to our naturalist the proper, or at least the
major task of analytic philosophy. As it does to me. (Whence the phrase ‘descriptive
[as opposed to validatory or revisionary] metaphysics’) (23).
The bracket notwithstanding, however, validatory and revisionary metaphysics are not
equivalent, by Strawson’s own lights. Au contraire! Validatory metaphysics seeks to show against
the skeptic that our conceptual scheme is legitimate and hence need not be revised.
Furthermore, Strawson’s brave new naturalism is not all that remote from the old fashioned
transcendental arguments it is supposed to displace. Heyndels maintains that Strawson never
seriously harbored the ambition to vindicate our conceptual scheme against skeptical challenges.
But his own painstaking research points in a different direction. Heyndels (2020, 101–105)
illuminatingly distinguishes three themes in non-reductive naturalism.
[Inescapable] It is “in vain” or “useless” to doubt the external world, moral, phenomenal and
mental properties. For we simply “cannot help believing” that material bodies, moral and
phenomenal properties exist (11, 39).
[Natural Framework] (1) Such “inescapable” beliefs delineate the framework within which
alone questions of justification are possible. (2) Therefore these beliefs themselves are not even
candidates for either doubt or justification in the first place (41; see also 1974 [1962], 18, 23).
447

Hans-Johan Glock

[Not A Belief] Labels like “belief” and “proposition” are inadequate to characterise these
convictions (1998b, 370).
[Inescapable] marks the Humean side of non-reductive naturalism, which has been the main
focus of interpreters. [Not A Belief] strikes a note familiar from Wittgenstein’s On Certainty
(1969 [1951]): the real foundations of our beliefs are (social) practices and attitudes rather than
beliefs (of individuals). That note is pragmatist in a lose sense of the term, and compatible with
non-reductive naturalism. But [Natural Framework] threatens to upset this picture. For it
encapsulates the guiding idea of elenctic transcendental arguments: both doubt and justification
require a framework; and skeptical doubt transgresses the limitations which this framework
imposes, by removing its supporting beams.
The analogy to transcendental arguments is obvious. The skeptic is charged with sawing off
the branch on which he himself is sitting, by questioning the preconditions of human thought.
But this leaves open the nature of this self-stultification. Is it of a merely psychological kind, as a
naturalist gloss would have it? Or does it have an argumentative, rational side, as the transcendental strategy demands? Is a skeptic merely paralysed in his thinking and acting? Or is there an
intellectual flaw in his position? These questions have both an exegetical and a substantive dimension. Concerning the former, my answer is as follows. In Scepticism and Naturalism as well as in
subsequent writings, Strawson vacillated; in doing so he is often swayed by On Certainty, which
similarly combines pragmatist and transcendental ideas. On the one hand, Strawson insists that
the point has been, not to offer a rational justification of the belief in external objects
or other minds, or of the practice of induction, but to represent sceptical arguments
and rational counter-arguments as equally idle—not senseless, but idle—since what we
have here are original, natural, inescapable commitments which we neither choose
nor could give up (28).
On the other hand, he characterises his approach as follows:
It is not merely a matter of dismissing the demand for a justification of one’s belief in a
proposition on a ground that one can’t help believing it. That would be weak indeed.
The position is, rather that the demand for justification is really senseless. Indeed both
Hume and Wittgenstein, in their several and distinct ways, go further, questioning, in
effect, the appropriateness of the ordinary concepts of ‘belief’ and ‘proposition’ in this
connection (1998b, 370)
The two passages can be rendered consistent if one interprets “senseless” as signifying bereft of
linguistic meaning in the first, pointless in the second. But the difference between these two senses of
“senseless” was obvious to Strawson. Furthermore, in writings prior to Scepticism and Naturalism
he clearly opted for the linguistic sense. He maintained that the skeptic himself employs concepts
which are intelligible only on the tacit assumption of conceptual connections he explicitly rejects.
Therefore, the sceptical position could not be stated unless it were unfounded.
He [sc. the skeptic] pretends to accept a conceptual scheme, but at the same time
quietly rejects one of the conditions of its employment. Thus his doubts are unreal,
not simply because they are logically irresoluble doubts, but because they amount to
the rejection of the whole conceptual scheme within which alone such doubts make
sense (1959, 35; see 106, 109).

448

Strawson and non-revisionary naturalism

Similarly, both before and after Scepticism and Naturalism Strawson insisted that the concepts on
which descriptive metaphysics focuses are non-contingent in that they are “limiting” or “necessary features in any conception of experience which we can make intelligible to ourselves”
(1966, 24, 44, 68; 1992, 26).
The charge of vacillation stands, therefore. But this opens up the more important substantive
issue. Does either claim about scepticism provide an intellectually sustainable resting place? In
my view, the transcendental one does (Glock 2012, 413–417; 2017, 224–226). If the skeptic
abandons preconditions of human thought and rational discourse, as Strawson contends, he
suffers the kind of self-refutation elenctic transcendental arguments try to unmask. It would
therefore be precipitate to rest content with a flat-footed Humean naturalism, according to
which the sceptical doubt is correct though impotent. Even if transcendental arguments cannot
establish any “ontological” conclusions about reality, they may be able to silence the skeptic. If
one can show that the skeptic employs concepts that are incompatible with his own doubts,
then it prevents him from making a coherent contribution to the debate. That is not the same as
proving that we have knowledge. But to silence sceptical doubt by means of argument is to
resolve the philosophical problem that it poses.

5 Anthropological naturalism
So far Strawson’s “non-reductive naturalism” comes across as primarily anti-eliminativist rather
than non-reductivist, and as not all that naturalistic. It does not fit any of the standard categories
of naturalism; and its response to scepticism is elenctic in character: jettisoning central parts of
our commonsensical conceptual scheme, the argument goes, is self-defeating, since it transgresses the bounds of linguistic sense. Nevertheless, the undeniable Humean streak in Scepticism
and Naturalism also manifests a type of naturalism that is not catered for by received taxonomies.
It is “simply not in our nature” (41) to abandon the participant perspective, either completely
(morality, perceptible qualities, intensions) or even momentarily (external objects, causation)
concerning morality and perceptible qualities. Strawson’s is an anthropological naturalism that
revolves around the idea of human nature.
In an excellent survey of naturalisms, Keil (2006, 292–294) protests that Strawson’s claim to
the title is a terminological sleight of hand which trades on the ambiguity of the word “nature”.
It can mean the totality of the natural world; it can also mean the “essence” or “real character”
of a thing. Keil rightly observes: “not any kind of reference to human nature indicates a naturalistic position”. Obviously, some views of the essence of humans – notably those of Plato
and Descartes – fail to qualify because they portray humans as being otherworldly at least in
part, for example, due to their possession of an immaterial soul. Nevertheless, Keil is mistaken
to suggest that it is therefore an “absurdity” to think that some appeals to human nature qualify
as naturalist. For it is equally obvious that an account of the nature of human beings, as of the
nature of morality, perceptible qualities, the mind, meaning and so on – can be naturalistic.
Strawson’s position falls into this category, for several reasons. First, he incisively criticises
the aforementioned dualist picture: human beings are flesh-and-blood animals. Now, it is
perfectly true that Strawson does not reduce human beings to solitary organisms whose essence
lies in their bio-chemical make-up. He favors a “social naturalism” whose progenitor is
Wittgenstein rather than Hume (see 24, 39). His account stresses the unique social, linguistic
and cultural dimensions of human life. Pace Keil, however, this does not disqualify Strawson
from being an anthropological naturalist. For those dimensions form part of our heritage as
denizens of the natural realm. It is part of our nature – that is, our real character – to be part of
nature – that is, the natural order. That is why the scientific perspective on human beings is not
449

Hans-Johan Glock

just legitimate but immensely fertile. At the same time, it is part of our real character to
constitute a distinctive part of the natural order – to be primates that are not just intelligent and
social, but also endowed with language, dependent on cooperation and the division of labor,
subject to norms and capable of progressive cultural development (see Glock 2012). That is
why the participant perspective is inescapable. But this very inescapability is due the kind of
animal we are – that is why it pertains au fond to practices and attitudes that antecede intellectual
beliefs – as of [Not A Belief] above.
The spirit of such an anthropological naturalism is evident in Wittgenstein’s remark:
Commanding, questioning, recounting, chatting, are as much a part of our natural
history as walking, eating, drinking, playing (1967 [1953], Sect. 25).
It is epitomised succinctly in Herder’s revolutionary riposte to the supernaturalist dogma that
language must have divine origins: “Already as an animal, man [der Mensch] possesses language”
(2002 [1772], 65). The capacity for language, norms, culture, etc. is part of a specifically human
yet nonetheless completely natural and unmysterious form of life.
In one passage Strawson briefly defends both his resistance to reductive naturalism and his
laying claim to being a naturalist. The forms of action, thought and speech that skepticism
repudiates by appeal to the hard sciences are simply not up for grabs. We “cannot help playing”
these language-games. Resigning oneself to this fact is therefore “more thoroughgoing naturalist”, whereas reductive naturalism is “merely naturalistic” (41). The anthropological naturalism of Scepticism and Naturalism implements a project that Strawson had sketched in earlier
writings – “explanatory metaphysics”. The latter “explains not just how our concepts and types
of discourse operate, but why it is that we have such concepts and types of discourse as we do;
and what alternatives there might be”. Explanatory metaphysics investigates the “natural
foundations” of our “conceptual apparatus in the way things happen in the world, and in our
own natures”. And it considers conditionals about how our conceptual scheme might change
given different empirical conditions (1956, 108; 1962, 317). In fact, however, this enterprise is
anthropological rather than metaphysical. It flies the flag of a realistic anthropology, one that
acknowledges the more or less stable parameters set by human nature.
Alas, Strawson remains general and vague about how the aspects of human nature that he
invokes against reductive naturalism fit into the natural order. His hints have been picked up,
for example, by McDowell’s (1996, 86–94) “naturalism of second nature”. But all of these
philosophical endeavors require elaboration through an empirical anthropology which continues the unduly neglected tradition of natural history (e.g. Tomasello 2014). This kind of
anthropology is both a biological and a social-cum-human science; it thereby straddles the
divide diagnosed by proponents of liberal naturalism (e.g. MacArthur 2010).

6 Relaxed realism and liberal naturalism
Strawson grants a point which threatens to unsettle the reconciliatory programme of anthropological naturalism. Concerning some of the phenomena disputed between skepticism and its
opponents, our natural inclinations point in opposite directions. The existence of universals,
intensions and rules is the paradigmatic case. On the one hand, there is “a strong natural
inclination to believe that whatever exists at all exists in nature”. On the other hand, we
naturally claim to recognise conceptual relations that obtain between concepts and are guided
by rules (71 vs. 91). Both the external standpoint of the reductive naturalist and the internal
standpoint of an individual speaker and thinker appear to be valid (93).
450

Strawson and non-revisionary naturalism

A resolution to this conflict is not to be found in contemplating the conflicting intellectual
and emotional pressures weighing in, respectively, on the side of nominalism and realism. It is
to be found in philosophical argument. Strawson was alive to this challenge. He convincingly
defended the existence of abstract and intensional entities like properties, kinds and propositions
against nominalism and extensionalism. The extension of an expression is what the expression
stands for or what it applies to. By contrast, the intension of an expression is an aspect of what
the expression means, standardly that aspect which determines its extension.
Quine grudgingly accepts the existence of abstract entities of an extensional kind, notably of
sets, because mathematics and hence the natural sciences are ontologically committed to their
existence.4 At the same time, he unconditionally rejects abstract entities of an intensional kind,
such as properties or propositions. His rationale is that there are no criteria of identity for
“intensions”: while we can specify conditions under which two expressions have the same
extension, we cannot specify conditions under which they are synonymous, that is, have the
same intension or meaning (see Glock 2003a, chaps. 2–3, 6–7).
Strawson counters these qualms with three moves. He accepts Quine’s general idea of ontological commitment: a statement is committed to those objects which must exist if it is true
(see 83; 1959, 15; 1992, 4–50). But he resists Quine’s specific criterion of ontological commitment:
“To be is to be the value of a bound variable” (Quine 1953, 14–15). This oracular dictum is
designed to ensure that only singular but not general terms carry ontological responsibilities, on
the grounds that only they are subject to quantification. Accordingly, abstract singular terms like
“redness” but not concrete general terms like “red” commit us to properties.
Yet prima facie, at least, general terms in predicative position are also accessible to quantifiers
and hence commit their users to the existence of properties (Strawson 1971, 65–66;
1992, 48–49).
1
2
3

Betty is witty entails not just
Someone is witty but equally
There is something Betty is (viz. witty)

Fa
∃xFx
∃ΦΦa

In statements like (1) Strawson distinguishes between the copula “is”, which stands for
nothing, and the adjective, which stands for a property. In consequence, he accepts that (3) is an
instance of what Quine calls objectual quantification (1997, 5, Chaps. 3–4). This is based on the
conviction that all genuine existence statements amount to “There is an object x such that” and
invites trouble. Unfortunately, on this assumption, (3) is indeed infelicitous. For it would have
to be rendered as
(3’)There is an object Φ such that a is Φ
Yet (3) does not claim that there is an object – the property of being witty – Betty is, but
merely that there is something she is, namely witty. “There is something which” has a wider
range than “there is an object which”. “Something” is syntactically transcategorial: it can
quantify into the positions of both singular terms, as in (2), and of predicates, as in (3). How
precisely such quantification is to be formalised is a moot issue (see Künne 2003, 6.2.3). The
crucial point, however, is that there are perfectly acceptable idioms of ordinary language which
capture quantification into predicate position (Glock, 2003a, 56–58). And Strawson is surely
right that formalisation must pay tribute to the inferences recognised in non-regimented discourse rather than the other way around. By acknowledging non-objectual quantification into
predicate position, we can therefore strengthen Strawson’s case against extensionalism.
451

Hans-Johan Glock

Strawson’s second move is directed against Quine’s programme of avoiding commitment to
intensions through “critical paraphrase” into canonical notation. Contrary to Quine, the
transition from
(4)Betty possesses wit
to (1) does not reduce our ontological commitments. For it is necessarily the case that (1) is
true under precisely the same conditions as (4). And whether someone is committed to the
existence of certain things depends on what she says, not on how she says it. The extensionalist
nicely fits Strawson’s description “Committed in thought to what we shun in speech, we
should then seem like people seeking euphemisms in order to avoid explicit mention of distasteful realities” (1997, 58; see 1992, 48–49).
Strawson’s final counter-argument is directed against Quine’s standard of ontological admissibility – “No entity without identity”. This standard demands that the existence of a type of
entity (events, for example) can only be accepted if we can provide “criteria” or “conditions” of
identity for entities of that type. Strawson likewise insists on criteria of identity. “There is
nothing you can sensible talk about without knowing, at least in principle, how it might be
identified” (1997, 22). But this is not to underwrite Quine’s demand for formal criteria of
identity which apply to all objects of that kind in a rigid and clear-cut fashion. It is not even
required that the objects we talk about be enumerable. There are no cast-iron ways of establishing how many colors, intellectual qualities, character traits, literary styles, affective attitudes, smells and ways of walking or talking there are. Nevertheless, these things are “as
distinctive, as easily recognizable, as anything in human experience”, and they are the subjects
of “pointful predications” (1997, 23, 35). Indeed, enumerability is a question not of the degree
to which certain types of entities are real, but of the degree to which the corresponding sortals
are regimented. There are ways of counting the properties Bohr’s 1913 theory ascribes to
electrons, simply because that theory draws precise distinctions in these respects. Conversely,
although mountains are supremely real, there is no cast-iron method of counting them.
All we need is a way of making clear what we are talking about. In this sense, however, we
do possess criteria of identity for colors or character traits, for although we cannot count them,
we can distinguish different colors and character traits. That is to say, we can tell that an object
is red rather than green, or that a person is arrogant rather than timid. Universals carry their
own criteria of identity. In fact, this is a precondition for the possibility of (re-)identifying
particulars. Such identification is not possible through mere ostension alone (“this?”), it also
requires that particulars can be subsumed under sortal terms (“this? F”), that is, assigned to
general kinds (82–83, 1959, Chap. 1).
In this respect, Strawson’s non-reductive naturalism is perhaps the most impressive implementation of the idea of “liberal naturalism” debated in this collection (see also De Caro and
Voltolini 2010). It repudiates not just epistemological but also ontological naturalism, yet without
committing itself to the existence of entities which are part of a quasi-causal supernatural order.
Abstract entities are not beyond space and time, they are simply aspatial and atemporal (1997,
60–61). Their existence is implied by statements we know to be true. To be sure, we need to spell
out what it means for them to exist. In the case of abstract entities this is done by referring to
cognitive capacities of the kind highlighted by anthropological naturalism – capacities for identifying and classifying particulars in the case of universals, capacities for linguistic understanding in
the case of meanings. There is an obligation of conceptual clarification here, yet no license for
suspecting ab initio a metaphysical mystery.
Snowdon and Gomes (2019, Sect. 7) call Strawson’s approach to ontology “relaxed realism”;
and they rightly note that it has been quietly influential among Oxford philosophers. We are
entitled to be realists about aspects of the world that are suspect to reductive naturalism,
452

Strawson and non-revisionary naturalism

including color, mentality, meaning and perhaps value, without reducing them to physical
phenomena. Yet this neither questions the claims of natural science nor lapses into metaphysical
mystery-mongering.
However, Strawson and his followers have perhaps been slightly too relaxed. The course
between the Scylla of reductive naturalism and the Charybdis of supernaturalism is more
precarious than it appears in Strawson’s inimitably elegant and relaxed prose. First, it is unclear
in what sense, if any, the existence of a type of entity can be relative to a conceptual framework.
Second, acknowledging entities that are atemporal and aspatial seems to require rejecting an
assumption that has been profoundly entrenched in modern philosophy. It is the conviction
that only the ultimate mechanical causes, those which explain the behavior of all observable
phenomena, exist properly speaking and are genuinely real, whereas “causally idle” entities
must be reduced or eliminated. If liberal naturalism à la Strawson is tenable, this claim cannot be
accommodated by the latter’s conciliatory spirit. It must be straightforwardly wrong rather than
one-sided. I am far from relaxed about the prospects of overcoming this identification of
genuine existents with ultimate efficient causes. But one point is clear: it is neither a finding
of empirical science, nor a result of sober analysis. Instead, it is a metaphysical thesis, a dogma of
ontological naturalism.

Notes
1
2

3
4

I shall use “skepticism” for the general attack on common sense, “scepticism” for the denial of
knowledge.
By contrast “Where we may freely operate with the notion of identity is with respect to the subject
himself, a being both conscious and corporeal, one and the same subject of both the personal and the
physiological histories” (66). The bearers of mental properties are flesh-and-blood subjects rather than
separate soul-substances.
He capitulated to Stroud’s (1982 [1968]) influential criticism. Prematurely, I argue in Glock 2003b:
33-42.
He is therefore an ontological naturalist only in the methodological sense mentioned at the start of this
essay.

References
(i) Works by Strawson
Strawson, P.F. On Referring. Mind, 59, (1950). 320–344.
Strawson, P.F. Introduction to Logical Theory. Methuen. (1952).
Strawson, P.F. Construction and Analysis. In A.J. Ayer and W.C. Kneale et al.. (eds.), The Revolution in
Philosophy. Macmillan. (2011).
Strawson, P.F. On Justifying Induction. Philosophical Studies, 9, (1958). 20–21.
Strawson, P.F. Individuals: An Essay in Descriptive Metaphysics. Methuen. (1959).
Strawson, P.F. Analysis, Science and Metaphysics. In Philosophical Writings ([1962]/2011). Oxford
University Press.
Strawson, P.F. The Bounds of Sense: An Essay on Kant’s Critique of Pure Reason. Methuen. (1966).
Strawson, P.F. Logico-Linguistic Papers. Methuen. (1971).
Strawson, P.F. Subject and Predicate in Logic and Grammar. Methuen. (1974).
Strawson, P.F. Freedom and Resentment. In Freedom and Resentment: And Other Essays, pp. 1–28.
Methuen. ([1962]/1974).
Strawson, P.F. Skepticism and Naturalism: Some Varieties. Methuen. (1985).
Strawson, P.F. Analysis and Metaphysics: An Introduction to Philosophy. Oxford University Press. (1992).
Strawson, P.F. Entity and Identity. Clarendon. (1997).
Strawson, P.F. Reply to Ernest Sosa. In L.E. Hahn (ed.), The Philosophy of P.F. Strawson, pp. 288–292.
Open Court. (1998b).

453

Hans-Johan Glock
Strawson, P.F. and Grice, H.P. In Defense of a Dogma. In Philosophical Writings, pp. 166–177. Oxford
University Press. ([1956]/2011).
Strawson, P.F. Two Conceptions of Philosophy. InPhilosophical Writings, pp. 166–177. Oxford University
Press. ([1990]/2011).
Strawson, P.F. Intellectual Autobiography. In Philosophical Writings, pp. 227–247. Oxford University Press.
([1998a]/2011).

(ii) Works by others
Arrington, R.L. (ed.) A Companion to the Philosophers, pp. 524–528. Blackwell. (1999).
Brown, C. Peter Strawson. Acumen. (2006).
De Caro, M. and Voltolini, A. Is Liberal Naturalism Possible? In M. De Caro and D. Macarthur (eds.),
pp. 69–86. Columbia University Press. (2010).
Dummett, M. The Place of Philosophy in European Culture. European Journal of Analytic Philosophy, 8(1),
(2012). 14–23.
Ferdinand, C.Y. (ed.). A Tribute to Peter Strawson. Oxford. (2008).
Fodor, J. Psychosemantics. MIT Press. (1987).
Glock, H.-J. Quine and Davidson on Language, Thought and Reality. Cambridge University Press. (2003a).
Glock, H.-J. Strawson and Analytic Kantianism. In H.J. Glock (ed.), Strawson and Kant. Oxford University
Press. (2003b).
Glock, H.-J. What is Analytic Philosophy. Cambridge: Cambridge University Press. (2008).
Glock, H.-J. Strawson’s Descriptive Metaphysics. In L. Haaparanta and H. Koskinnen (eds.),Categories of
Being, pp. 391–419. Oxford University Press. (2012).
Glock, H.J. Philosophy Rehinged? International Journal for the Study of Scepticism, 6, (2016). 274–308.
Glock, H.-J. Ordinary Language Philosophy and Descriptive Metaphysics. In A. Preston (ed.), Analytic
Philosophy: An Interpretive History, pp. 214–228. Routledge. (2017).
Herder, J.G. Philosophical Writings. Cambridge University Press. ([1772]/2002).
Heyndels, S. Philosophy as an Exercise in Conceptual Self-consciousness: an Explication of P.F. Strawson’s philosophical Methodology. Doctoral Dissertation. Leuven. (2020).
Keil, G. Naturalism. In D. Moran (ed.), A Companion to Twentieth Century Philosophy, pp. 254–307.
Routledge. (2006).
Künne, W. Conceptions of Truth. Oxford University Press. (2003).
LePore, E. and Smith, B. The Oxford Handbook of the Philosophy of Language. Oxford University
Press. (2006).
McDowell, John Mind and World. Harvard University Press. (1996).
Nagel, E. Sovereign Reason. Free Press. (1954).
Quine, W.V.O. From a Logical Point of View. Logico-Philosophical Essays. Harper & Row. (1953).
Quine, W.V.O. Word and Object. MIT Press. (1960).
Salmon, W. Should We Attempt to Justify Induction? Philosophical Studies, 8, (1957). 33–48.
Sellars, R.W. Evolutionary Naturalism. Open Court. (1922).
Sellars, W.F. Naturalism and Ontology. Ridgeview. (1979).
Sen, P.K. and Verma. R.R. The Philosophy of P.F. Strawson. Indian Council of Philosophical
Research. (1995).
Snowdon, P. The Routledge Encycplopedia of Philosophy. Ed. E. Craig. Routledge. (1998).
Snowdon, P. and Gomes, A. Peter Frederick Strawson. In Edward N. Zalta (ed.), The Stanford Encyclopedia
of Philosophy. (2019). https://plato.stanford.edu/archives/spr2019/entries/strawson/.
Stern, R. On Strawson’s Naturalistic Turn. In H.-J. Glock (ed.), Strawson and Kant. Oxford University
Press. (2003).
Stroud, B. [1968]. Transcendental Arguments. In R.C.S. Walker (ed.), Kant on Pure Reason, pp. 117–131.
Oxford University Press. (1982).
Tomasello, M . A Natural History of Human Thinking, Harvard University Press. (2014).
Wittgenstein, L. Philosophical Investigations. Trans. G.E.M. Anscombe. Blackwell. ([1953]/1967)
Wittgenstein, L. On Certainty [German-English Parallel text]. Ed. G.E.M. Anscombe and G.H. von
Wright, Trans D. Paul and G.E.M. Anscombe. Blackwell ([1951]/1969).

454

37
PUTNAM AND LIBERAL
NATURALISM
Massimo Dell’Utri

1 Introduction
Being a naturalist was for Hilary Putnam at the same time unavoidable and difficult. Given his
recognition of the ineliminable role the natural sciences play in the understanding of human
beings, resorting to their perspective in epistemological and metaphysical issues was for him
virtually mandatory. On the other hand, Putnam was a religious person who deeply felt the
need for a spiritual dimension to life, and subscribed to a kind of “naturalistic theology, that best
reconciles my own spiritual needs with my critical intelligence” (Putnam 2016c, 163).
Combining these two dimensions of his personality proved to be one of the most difficult tasks
of his intellectual life.
Moreover, for Putnam the difficulty of being a naturalist is increased by the vagueness of the
term “naturalism” itself. It is a very popular term in papers by philosophers of different varieties,
he remarks, but it hardly ever gets defined in these papers, and the upshot is that most of the
people working within a naturalist framework often have a different understanding of what
being a naturalist amounts to. To give an example of the vagueness, emptiness and sometime
implausibility, of the term, he cites the definition to the effect that to subscribe to naturalism is
to believe that “all phenomena are subject to natural laws”. But “if all that is involved in being a
naturalist is thinking that there aren’t phenomena that actually violate natural laws, then who
isn’t a ‘naturalist’?” (Putnam 2004b, 60). And if the definition a naturalist is working from is one
according to which naturalism implies believing that the methods of the natural sciences are
applicable in every area of inquiry, where this is supposed to mean that “there is a science that
resembles physics (complete with laws, theories, experiments and so on) of every single thing
[…], then why wouldn’t it be perfectly respectable not to hold such an implausible
view?” (Ibid.).
The two aforementioned definitions are instances of what he has termed “scientistic” or
“hard naturalism”, which represents a reductionist stance to the effect that only the natural
sciences are qualified to make the right epistemological and metaphysical claims, and aims
therefore at bringing all the claims made within our conceptual system by other disciplines back
to the framework of the natural sciences (hard naturalists of different varieties differ on what the
best ways to accomplish this may be). Putnam was a fierce opponent of this kind of naturalism
(cf. Putnam 1983c, 211: “scientism is, in my opinion, one of the most dangerous contemporary
DOI: 10.4324/9781351209472-42

455

Massimo Dell’Utri

intellectual tendencies”), and tried to remedy the vagueness he detected in it by presenting a
view characterised by the rejection of the items in the following list:
all appeals to supernatural entities in philosophy […], the positivist demand that
aesthetic and ethical concepts be reduced to the concepts of the natural sciences […],
the positivist view that history will eventually become a “science”, [the idea that]
normative utterances [are] somehow “second grade” or merely “expressive”, […] a
Platonic realm of normative facts independent of human practices and needs, [and]
Moorean quasi-mystical faculties of moral intuition
(Putnam 2016b, 22).
This is the view called liberal naturalism – a phrase coined by Mario De Caro and David
Macarthur that Putnam happily took up. However, thus conceived liberal naturalism does not
yet fully represent Putnam’s position, given that the denial of the items in the list above encompasses a wide spectrum of positions some of which Putnam wanted to distance himself
from. To that denial one should add the endorsement of “an unreduced notion of representation” – where representation is “a relation between organisms (and states of organisms
and, derivatively, bits of language) and real things, properties, and events” (Putnam 2016b, 40).
This means both that there is something “out there” to be represented by our mental and
linguistic activity, and that the relation of representation in force cannot be confined to a
material (or any other) dimension. Moreover, since the very activity of representation presupposes a number of values, the irreducibility of representation carries with it the irreducibility
of the normative. In brief, liberal naturalism of a Putnamian variety is characterised by a realistic
attitude both in metaphysics and toward the normative. I will try to clarify these two features.

2 Realism in metaphysics
Putnam’s realistic attitude is characterised by the conviction that there are facts of the matter that
objectively fix what is correctly said in a given region of discourse. His realism passed through
various stages – at least three (cf. Dell’Utri 2016) – all with a different name. Indeed, in
chronological order we can distinguish between metaphysical realism, internal realism and natural
realism. The latter is a view he began to develop from 1990 onward and has a feature in
common with the earlier metaphysical realism, that is, the idea that the world is independent
from our cognitive abilities. This is a feature that was lost in the second stage, where one of the
central claims is to the effect that “the mind and the world jointly make up the mind and the
world. (Or […] the Universe makes up the Universe – with minds – collectively – playing a
special role in the making up)”: Putnam 1981, xi). The main difference between the first and
the third stage is that, according to the former, the world has a fixed structure of objects and
properties (it’s a so-called ready-made world), whereas according to the latter the world possesses
no predetermined totality of objects and their properties. Moreover, according to metaphysical
realism there is a sharp line between properties we discover in the world and properties we
project onto the world, whereas in the third (as well in the second) stage this conviction is lost
in favor of the idea of a deep interpenetration of fact and convention.
Within each stage Putnam displayed a specific interpretation of the concept of truth.
Roughly, metaphysical realism combined with a correspondentist interpretation to the effect
that there is a fixed relation between portions of reality and statements (whatever their content)
that, when it is present, decrees the truth of a given statement. Internal realism gave a pragmatist
interpretation to truth, in that it is seen as what can be asserted in epistemically-good-enough456

Putnam and liberal naturalism

conditions – an idea reminiscent of Charles Sanders Peirce’s account of truth, although different
in one important respect: there is no reference to a purported ideal limit of inquiry. Finally,
natural realism chimed in with a pluralist interpretation of truth according to which truth
amounts to many different things – as many different things as there are kinds of true propositions, and as there are domains in which a proposition can be true: empirical, mathematical,
logical, ethical, juridical, religious and so on. In brief, truth is not one, but many. Notice that
this kind of alethic pluralism does allow talk of correspondence, provided that you are not
presupposing the existence of a metaphysically strong and well-determined relation that connects every statement to the portion of the world it relates to. The only philosophically legitimate possibility of talking of correspondence is to give an actual semantical interpretation to a
Tarskian biconditional of the form “It is true that p if, and only if, p” (for any statement p), that
is, an interpretation that sees any instance of the biconditional as an actual words-world relation.
This allows us to point to the right-hand side of a Tarskian biconditional and identify the bit of
reality (the fact of the matter) that is responsible for the truth of p (cf. Putnam 1994, 253). Thus, in
this metaphysically light sense that denies the existence of a predetermined strong and unique link
between language and the world, all true statements can be said to correspond to reality, regardless of their content, where this correspondence can boil down to coherence, warranted
assertibility, correspondence in the strict sense, utility and so on, depending on the statement
at hand.
Yet this is still too poor a description of Putnam’s own versions of realism, given that many
philosophers might be seen to subscribe to one or the other version and, nonetheless, have
substantial differences with Putnam’s position. To get a picture of what Putnam was up to we
have to introduce two features that are shared by all three stages and have a direct bearing on his
own kind of naturalism (features that, I should add, only the last stage manages to account for
properly). These are conceptual pluralism, on the one hand, and reference understood as “a relation
between linguistic items and worldly objects” (Putnam 2016b, 24), on the other. (The latter is
part of the “unreduced notion of representation” mentioned above.)

2.1 Conceptual pluralism
Conceptual pluralism is the idea according to which “it is an illusion that there could be just
one sort of language game which could be sufficient for the description of all of reality”
(Putnam 2004a, 22). On the contrary, more than one theory, more than one kind of statement,
more than one vocabulary, more than one conceptual scheme, more than one language game is
of service to our cognitive purposes, even though they are subject to different standards, possess
different sorts of application, and have different logical and grammatical features. But, again, a
further (and very important) element is to be added if one wants to gain a complete grasp of the
Putnamian notion of conceptual pluralism. And this is the claim that all conceptual schemes that
turn out to be a useful guide to reality are of equal validity, so that, for example, the areas of
intellectual analysis that a positivist-minded philosopher would deem devoid of epistemological
value are in fact bona fide attempts at knowledge – ethics and aesthetics being cases in point.
With this clarification in mind we have a point of view according to which those conceptual
schemes are all genuine ways of getting in touch with reality, and all of them represent accounts
of the world that have to be taken seriously and not as a mere façon de parler, since all of them are
forms of reflection “fully governed by norms of truth and validity”, as Putnam put it borrowing
a phrase from James Conant (cf. Conant 1997, 202).
To have an instance of conceptual pluralism, consider an account of the contents of a room
in the vocabulary of fundamental physical theory and an account in the vocabulary of tables and
457

Massimo Dell’Utri

chairs. They are seen by Putnam as having the same epistemological dignity: both can genuinely
aspire to represent a description of what is out there. There is no question of one account being
privileged in respect to the other, no question of reducing the conceptual framework of
common sense to the framework of physics: both are on a par when it comes to gaining an
understanding of the world in which we live. As I mentioned, the same goes for ethical and
aesthetical statements: they are involved in the practice of inquiry like any other statement
traditionally so involved, and equally subject to testing for their truth or falsity. Only, in cases
like these the kind of test and the nature of truth differ from the ones that are in place in
empirical situations.
Notice that Putnam subscribed to conceptual pluralism because for him it is reality itself that
asks us to do so: conceptual pluralism is a consequence of his picture of reality. According to the
latter, the world is no monolithic unchanging block but a multilayered ductile and constantly
changing complex that “has many levels of form, including the level of morally significant
human action, [so that] the idea that all of these can be reduced to the level of physics [is] a
fantasy” (Putnam 2008, 5–6). Each level is singled out thanks to the role it plays in our lives and
keeps its ontological dignity insofar as it plays this role. It follows that, besides the previously
mentioned moral level, also the material, aesthetical, religious, economical, juridical and
mathematical levels belong to the world, as do all the aspects of reality that elicit in us a
perceptual or intellectual response and constitute our experience – in the broadest possible sense
of the term.1
Within each level one can speak of correspondence to the facts in the light metaphysical sense I
was hinting at above: each one of them possesses features that warrant the objective validity of
the statements made in it. Therefore, when a statement is objectively valid, this happens because
there is a fact of the matter that makes it so – some aspect of reality the statement refers to that is
responsible for that validity. Reality impacts on us in many ways, perceptual and nonperceptual: it demands something from us, it challenges us by presenting problematic situations
and pushing us to respond to requests that don’t just exist at a physical level, but which involve
the whole complexity of our human nature. This is why, according to Putnam, reality isn’t
morally, juridically, aesthetically or religiously inert. That things are so appears for him clear if
we reflect on the fact that, even though the values guiding our verbal and non-verbal behavior
are of our own making, we make them in response to needs we didn’t create – needs that then
guide us in the evaluation of the correctness or incorrectness of those values as responses to
those very needs (cf. Putnam 2002, 97, 108; 2008, 6). It is a virtually endless process that testifies
to the magmatic, kinetic and multifaceted character of reality:
An essential part of the “language games” that we play in science, in morals, and in the
law is the invention of new concepts, and their introduction into general use; new
concepts carry in their wake the possibility of formulating new truths
(Putnam 2002, 109).
New truths that unlock new portions of the world, and new portions of the world that allow
for the formulation of new truths: that is why “we endlessly renegotiate – and are forced to
renegotiate – our notion of reality as our language and our life develop” (Putnam 1999, 9).
This very endless renegotiation of the notion of reality shows how the idea that the
notions of object and existence are given once and for all and possess a steadiness of their
own is unjustified. On the contrary, according to Putnam the many uses of language
that we display in the many contexts of our life reveal that these notions have a
458

Putnam and liberal naturalism

distinctive openness, by virtue of which what counts as an object and what counts as
existent isn’t univocally predetermined from the start. Granted, we use the words
“object” and “existence” in order to refer to the middle-sized physical objects we
come across in our daily lives, and also to the many objects we can perceive by means
of instruments. We can thus take this as the initial use of these words. However, the
fact that we then speak of things such as World War II, the color of the sky, mirror
images, objects of desire, the number two, the ideal of freedom, the value of coherence, the main character in the novel we are reading and so forth, is a pretty strong
indicator that “object” and “existence” are used in a multiplicity of ways that have
hardly anything to do with their initial use.
So, according to Putnam the reality we are embedded in is multilayered and plastic, and
conceptual pluralism is the best epistemological stance we can adopt to describe it. It should be
noted in passing that conceptual pluralism is different from what Putnam has called conceptual
relativity, but the two combine nicely together, as far as the plasticity of the world is concerned,
and therefore conceptual relativity too helps us to understand the character of Putnam’s liberal
naturalism. Conceptual relativity is the phenomenon given by the cognitive equivalence of
sentences, theories, conceptual systems that, taken at face value, are incompatible: for example,
two sentences that say strongly different things about the very same portion of reality and that,
nevertheless, are both true given that all the available evidence supports both sentences equally
well. Take for instance that situation in quantum mechanics where an electron passing through
a certain spot in spacetime is described by one theory as a particle with a well-defined momentum and an undetermined position and by another theory as a wave with a particular
length. Both theories are equally justified by the facts, since the facts do not exclude one of the
two to the advantage of the other – yet the two theories say incompatible things.
Equivalent descriptions of this kind are no exclusive phenomenon of physics. On closer
inspection, they are quite pervasive: they are in place in all the cases in which it is possible to
choose between equivalent alternatives. For instance,
(1) the choice, in different contexts, of including or not including mereological sums
in one’s ontology; (2) the choice, in formalized geometry, of taking points to be
individuals or taking them to be convergent sequences of spheres (or of other solids—
not to mention the various different ways in which the notion of a “convergent
sequence” has been formalized!); (3) the choice, in a certain portion of classical
electrodynamics between taking the action between charged particles to be mediated
by “fields” or by “point-source retarded potentials”; and (4) the choice, in mathematical logic, between taking sets to be characteristic functions or taking them to be
primitive objects and taking functions to be sets of ordered pairs
(Putnam 2001, 432–423).
In all these cases nothing forces us to choose one alternative over the other, because there is no
fact of the matter that would adjudicate the question in one way over the other: in the last case,
“no mathematician would suppose that this ‘ontological difference’ had the slightest mathematical or metaphysical significance” (Ibid.). Each theory and each sentence do justice to the
phenomena just like their alternative – the adoption of the one or the other being just a matter
of convention.
Careful though. The incompatibility between two (or more) equivalent descriptions is such
only when considered at face value. When so taken, it is obvious that, say, points and convergent
459

Massimo Dell’Utri

sequences of spheres are mutually incompatible, but the incompatibility vanishes if we take
another important feature of equivalent descriptions into account, that is, that there is a way of
translating each sentence of the one description into a sentence of the other “which ‘translates’
scientific explanations into equally acceptable scientific explanations” (Putnam 2001, 436–437;
for details, cf. Putnam 1983b). Thanks to this kind of translation we can see that the two
equivalent descriptions merely look incompatible if considered, as it were, at first blush. They
simply embed different conventions. Which ones? The different conventional choices consisting
in taking individuals as points or as convergent sequences of spheres – choices that in their turn
reveal what we saw in relation to conceptual pluralism, that is, a different use of the relevant
metaphysical words such as “object”, “individual”, “substance”, “fact”, “state of affairs”,
“exists”. As in the case of conceptual pluralism, we have that the world does not impose a
unique and univocal use of these words, but their use gets extended as soon as we adopt a
conceptual scheme that proves to be useful in the description of a given situation. Whoever
insists upon refusing to acknowledge this, Putnam claims, “tries to preserve […] the naïve idea
that at least one Category – the ancient category of Object or Substance – has an absolute
interpretation” (Putnam 1987, 36).
In a nutshell, even though “conceptual relativity implies [conceptual] pluralism [and] the
reverse is not the case” (Putnam 2004a, 49), both cooperate in making it impossible for there to
exist a privileged vocabulary to account for reality. Still, the difference between them lies in the
fact that conceptual relativity applies (when it applies) within one layer of reality individually
considered, whereas conceptual pluralism has to do with the very existence of a plurality of
layers of reality – the ones we refer to in thinking and talking.

2.2 Reference
Putnam thought that reference is an irreducible concept. Why did he think so? The answer is
connected to the impossibility of having a fixed relation of correspondence between language
and the world. Indeed, if we are looking for the link that unites language to reality and consider
the thing from an abstract logical point of view, then
there are many different ways of putting the signs of a language and the things in a set
S in correspondence with one another, in fact infinitely many if the set S is infinite
(and a very large finite number if S is a large finite set). Even if the ‘correspondence’
has to be a reference relation and we specify which sentences are to correspond to
states of affairs which actually obtain, it follows from theorems of model theory that
there are still infinitely many ways of specifying such a correspondence
(Putnam 1983c, 206–207; for technical details, cf. Putnam 1980).
It follows that, from a purely logical perspective, the attempt to reduce reference to a specific
relation collapses owing to the impossibility of singling out a distinct relation as the intended one –
let alone a relation that conforms to a hard naturalist’s taste. Likewise, if we consider the matter
from the concrete point of view of daily life and try to highlight the relation of correspondence
that links true empirical statements to reality, no matter what the empirical statement is and no
matter what the occasion of its utterance may be, again we find ourselves facing a plurality of
kinds of correspondence:
“This piece of beef weighs one pound” may “correspond to reality” by the standard of
correspondence appropriate to a butcher’s shop, but be extremely wrong by the
460

Putnam and liberal naturalism

standard of laboratory science. And the kind and degree of correspondence changes
again when the statement is “John is very neurotic,” or …
(Putnam 2013, 98).
Worse still, any strategy aimed at showing that, despite all this, there are routes through which
reference can be made attuned to the demands of the natural sciences won’t work, according to
Putnam.
One of these strategies consists in trying to naturalise semantics by claiming that referring to an
object o is permitted by a causal interaction with o, where the causal relation that makes this
interaction possible is built-in into the very material structure of reality. But, Putnam rhetorically
asks, is causation really something physical that implies an actual and necessary relation between
objects or events? He remarks that if this were so, then a cause would be “a sufficient condition for
its effect; whenever the cause occurs, the effect must follow (at least in a deterministic world)”
(Putnam 1983c, 212). However, “cause” virtually never means this. For example, when we try to
find the cause of a fire in the forest, we suddenly find that “cause” has many elements to it: the
dryness of the leaves, their proximity to the campfire, the temperature of the day and so on, and
only some of them are picked up in our search for the responsible event – where we choose this or
that element depending on its advantages given the aims and interests motivating our explanation:
what we call the “cause” of what depends on more than the intrinsic physical
properties of the objects and events concerned; speaking of causality is a way of introducing an explanatory structure, and […] the same events get structured differently
when different interests are brought to them (which is not to say that any event can be
structured in just any way, of course)
(Putnam 1992, 378).
“Causes” can therefore very often be paraphrased as “explains”, so that “What is and what is not
a ‘cause’ or an ‘explanation’ depends on background knowledge and our reason for asking the
question” (Putnam 1983c, 214). And all this is not written into the physical system itself, given
that “salience and relevance are attributes of thought and reasoning, not of nature” (Ibid., 215).
Another strategy which attempts to make the relation of reference palatable to a naturalist
perspective consists in defining the relation disquotationally – that is, non-semantically –
thereby neutralising the problem of reference from the start. This boils down to the claim that
all there is to reference is that a word a refers to an object x if it satisfies the following schema:
(x)(“a” refers to x if, and only if, a = x) (cf. Horwich 1998, 112). But as an account of what we
humans do when we use a word in our daily practice this claim is rather “empty”: it is a purely
logicolinguistic remark quite irrespective of any actual “word-worldly object” relation. And the
error in this account of our ability to refer to something is, according to Putnam, crystal clear:
If there is no determinate correspondence between any of our terms and nonlinguistic
objects […] then I can utter […] as many […] instances of the ‘disquotation’ scheme
“T” refers to Ts as I want, and no relation between the term […] and any particular
object external to my words is thereby defined
(Putnam 2004b, 63).
Even the account of truth that in the same vein the disquotationalist provides is wanting. They
claim that all there is to truth are instances of the disquotation scheme: “s” is true if, and only if,
s (for every sentence s), and emphasise that, whenever s is uttered in a language that is different
461

Massimo Dell’Utri

from the one in which the attribution of truth to it is made, the two occurrences of s must
translate one another. Disquotationalists therefore recognise that the notion of translation is
needed if their strategy is to work. But “translating sentences presupposes knowing what their
descriptive constituents refer to” (Putnam 2016b, 38), which makes it clear that, paradoxically
enough, disquotation (and alethic deflationism for that matter) presupposes an actual and fullblown relation of reference – that is, that very relation to empirical particulars and properties
that is acquired via “information-carrying causal interaction with those particulars and properties” (Putnam 2004b, 63) and allows us to speak of the many things and events obtaining in
the many levels of reality. This paves the way for the realistic attitude toward the normative that
is a fundamental feature of Putnam’s liberal naturalism.

3 The pervasiveness of the normative
What Putnam stressed more and more is that “describing language use involves normative concepts: things we say are variously true and false, justified and unjustified, succeed in referring and fail to
refer, appropriate and inappropriate, and so on” (Putnam 2004b, 70). And, as conceptual pluralism
shows, this applies to all the various sorts of statements concerned with the many levels of reality:
they “are bona fide statements, [i.e.] fully governed by norms of truth and validity” (Ibid., 61). In
all levels of reality, therefore, truth is relevant, showing that the normative is everywhere:
to regard an assertion or a belief or a thought as true or false is to regard it as being
right or wrong, [where] what sort of rightness or wrongness is in question varies
enormously with the sort of discourse
(Putnam 1999, 69).
This claim has two consequences. First, that since we can formulate objectively valid and true
claims also in ethics, aesthetics, politics, sociology, psychology, mathematics and so on without
making any reference whatsoever to purported objects, “an indicative sentence can be a bona
fide statement without being a ‘description of reality’” (Putnam 2004b, 70). You can have
objectivity without necessarily referring to objects. The only requirement is the adoption of a
fallibilistic outlook, which is of inestimable help in distancing oneself from what is deemed true on
the basis of the best theories available in a given historical context. Second, the very practice of
the natural sciences presupposes epistemic values such as instrumental efficacy, plausibility,
reasonableness, simplicity, naturalness, conservatism, and even the aesthetical value of beauty
(cf. Putnam 2002, Chap. 2), so that “science depends on what is not fully scientific at every
point” (Putnam 2016b, 43). Discarding any Platonic realm of normative facts independent of
human practices and needs, and any Moorean quasi-mystical faculties of moral intuition,
Putnam thus shows that the normative proves to be virtually ineliminable.
But then,
if there is no eliminating the normative, and no possibility of reducing the normative
to our favorite science, be it biology, anthropology, neurology, physics, or whatever,
then where are we? We might try for a grand theory of the normative in its own
terms, a formal epistemology, but that project seems decidedly overambitious. In the
meantime, there is a great deal of philosophical work to be done, and it will be done
with fewer errors if we free ourselves of the reductionist and historicist hang-ups that
have marred so much recent philosophy
(Putnam 1983d, 247).
462

Putnam and liberal naturalism

Note
1 For a distinction between a narrow and a broad sense of experience, cf. Putnam (2012b).

References
Baghramian, M. (ed). Reading Putnam. Routledge. (2013).
Clark, P. and Hale, B. (eds.). Reading Putnam. Basil Blackwell. (1994).
Conant, J. Wittgenstein’s Philosophy of Mathematics. Proceedings of the Aristotelian Society, 97, (1997),
195–222.
De Caro, M. and Macarthur, D. (eds.). Naturalism in Question. Harvard University Press. (2004).
Dell’Utri, M. Putnam’s Conception of Truth. European Journal of Analytic Philosophy, 12, (2016). 5–22.
Hill, C.S. (ed.). The Philosophy of Hilary Putnam (Philosophical Topics), Vol. 20, pp. 1–408. University of
Arkansas Press. (1992).
Horwich, P. Truth, 2nd edition. Clarendon Press. (1998).
Putnam, H. Models and Reality. Journal of Symbolic Logic, 45, (1980). 464–482.
Putnam, H. Reason, Truth and History. Cambridge University Press. (1981).
Putnam, H. Realism and Reason: Philosophical Papers, Vol. 3. Cambridge University Press. (1983a).
Putnam, H. Equivalence. In Realism and Reason: Philosophical Papers, pp. 26–45. Cambridge University
Press. (1983b).
Putnam, H. Why There Isn’t a Ready-Made World. In Realism and Reason: Philosophical Papers,
pp. 205–228. Cambridge University Press. (1983c).
Putnam, H. Why Reason Can’t Be Naturalized. In Realism and Reason: Philosophical Papers, pp. 229–247.
Cambridge University Press. (1983d).
Putnam, H. The Many Faces of Realism. Open Court. (1987).
Putnam, H. Reply to Richard Healey. In The Philosophy of Hilary Putnam (Philosophical Topics),
pp. 377–379. University of Arkansas Press. (1992).
Putnam, H. Simon Blackburn on Internal Realism. In Reading Putnam. Basil Blackwell.
pp. 242–254. (1994).
Putnam, H. The Threefold Cord. Mind, Body, and World. Columbia University Press. (1999).
Putnam, H. Reply to Jennifer Case. Revue internationale de philosophie, 218, (2001). 431–438.
Putnam, H. The Collapse of the Fact/Value Dichotomy and Other Essays. Harvard University Press. (2002).
Putnam, H. Ethics Without Ontology. Harvard University Press. (2004a).
Putnam, H. The Content and Appeal of ‘Naturalism’. In M. De Caro and D. Macarthur (Eds.), Naturalism
in Question, pp. 59–70. Harvard University Press. (2004b).
Putnam, H. Jewish Philosophy as a Guide to Life. Rosenzweig, Buber, Levinas, Wittgenstein. Indiana University
Press. (2008).
Putnam, H. Philosophy in an Age of Science. In M. De Caro and D. Macarthur (Eds.), Physics, Mathematics
and Skepticism. Harvard University Press. (2012a).
Putnam, H. The Depths and Shallows of Experience. In Philosophy in an Age of Science. Physics, Mathematics
and Skepticism, pp. 567–583. Harvard University Press. (2012b).
Putnam, H. Comments on Richard Boyd. In Reading Putnam, pp. 95–100. Routledge. (2013).
Putnam, H. Naturalism, Realism, and Normativity, in M. De Caro (Eds.). Harvard University Press. (2016a).
Putnam, H. Naturalism, Realism, and Normativity. In Naturalism, Realism, and Normativity, pp. 21–43.
Harvard University Press (2016b).
Putnam, H. Reading Rosenzweig’s Little Book. Argumenta, 2, (2016c). 161–168.

463

INDEX

Abbau 62, 64
abstract entities 360–361
abstractive impoverishment 62, 64
action 39
actualism 411–418; indexicality 414–416;
metaphysics and 417; realism and 412–414
adverbial views 302
aesthetics 267–279; cognitive 283; empathy
270–273; film spectatorship 277; liberalist
naturalist 274–275; nonreductive scientific
naturalism and 268–269; subpersonal in
273–274; thick explanations 270, 275–276; thin
explanations 270, 275–276
Agrippa trilemma 73
Akaike Information criterion (AIC) 197–198
Allen, B. 255–263
analogy 41–42
The Analysis of Goodness (Fink) 324
analytic metaphysics 193–199
Andler, D. 419–428
Anscombe, G.E.M. 7, 18, 122, 124, 125, 138,
276, 353
anthropological naturalism 449–450
Anthropology for Physicians and Philosophers
(Platner) 42
anthropomorphism 42, 324
The Anti-Christ (Nietzsche) 49
antifoundationalism 193
antimony 207–208
anti-naturalism 247
anti-supernaturalism 177
Aquinas, T. 124, 131
Aristotelian naturlism 68n2
Aristotelianism 206, 215n2
Aristotle 1, 7–15, 15n7, 113, 124, 126, 131;
Categories 7, 10; on character and nature 13–15;
ethical naturalism 7–8; Eudemian Ethics 8–9; on

habit (hexis) 10–11; Metaphysics 8, 11;
Nicomachean Ethics 7, 12, 14; Physics 8, 11;
unidirectionality of habits 12–13
Articulating the World (Rouse) 179, 180
ascriptions 286
astrology 378
asymmetry of overdetermination 377
Audi, R. 358–369
Austin, J. L. 307, 393
autonomy 39, 49, 73, 179, 184, 232, 251, 334,
346, 434, 436
avant la lettre 70
Avela, F. 78
avoidance 351–353
Ayer, A. J. 320, 373
Bacon, F. 378
Baggini, J. 172, 175n7
Bagnoli, C. 329–335
Baier, A. 28
Baker, L. 210, 219, 223–224, 237
base rates 196–197
Baumgarten, A. G. 45
Bayes’ rule 196–197
Bayesian Information Criterion 198
Beck, L. 23–24
Being and Nothingness (Sartre) 74
Being and Time (Heidegger) 136
beliefs 40
Benovsky, J. 210
Bentham, J. 320
Berghofer, P. 75
Berkeley, G. 2
Bermudez, J. 305
Between Facts and Norms (Habermas) 434, 436
Beyond Good and Evil (Nietzsche) 49
Biddle, J. 185

464

Index
Big Bang theory 249, 253n10
Bildung 342
Bilgrami, A 27, 187n11
Blackburn, S. 142, 146, 343
Block, N. 310, 316n33
Bloor, D. 378
Bobzien, S. 15n8
body language 391
Boghossian, P. 214
Bonhoeffer, D. 222
Bonjour, L. 302
Book of Songs 257
Bounds of Sense (Strawson) 445
Boyd, R. 122
Boyle, M. 345
Braddon-Mitchell, D. 175n6
Brandom, R. 156–157, 378
Brick, D. 122
Budd, M. 291–292, 292
Burge, T. 188n13, 308
Burgess, J. 101
Burke, E. 291–292
Butchvarov, P. 368n3
Byrne, A. 312
Canberra Plan 169, 175n6
Cantor, G. 99
Carnap, R. 98
Carroll, L. 395
Carroll, S. 376
Cartesian dualism 347n14, 363
Cartwright, N. 390, 406n47
Cartwright, R. 369n7
Caruso, G. 210
Casebeer, W. 122
Categories (Aristotle) 7, 10
catholic naturalism 321
Catholicism 430
causal fundamentalism 385
Cavell, S. 147, 276–277, 279n34, 386
Chalmers, A. 377
Chalmers, D. 172, 305–306, 339
Chappell, S. G 134
character 13–15
chemistry 428n2
Cheng Yi 260–261
Chiaradonna, R. 7–15
Chinese naturalism 255–263; beginnings of
255–256; Confucianism 256–257; Daodejing
257–258; liberal naturalism, compared with 261;
neo-Confucianism 259–260; Tian 255–256;
Zhuangzi 258–259
Chomsky, N. 210
Chopra, D. 376
Christianity 230, 430
Churchland, P. 117, 134, 210, 215, 378
Clarke, S. 352

classical pragmatism 83–91; evolutionary theory
and 83–87; scientific naturalism and 87–89
Clauberg, J. 21–23
Cleland, C. 377
Clewis, R. 294
Cogito 50
cognitive aesthetics 283
cognitive emotions 272
color 310–314; dispositional and relational views
312–313; error theoretic views 313; primitivism
and 312; reductive physicalism and 311–312
Comments on a Certain Broadsheet (Descartes) 21–22
common life 33
common sense 1, 2–3, 47, 117, 206, 228–229, 233,
246, 250–252, 281, 288, 312, 386–388, 390,
392, 403, 406n25
complementarity of opposites 258
Comte, A. 60–61
Conant, J. 457
conceptual capacities 308
conceptual pluralism 457–460
Confucianism 256–257
constructivism: Kantian 329–335; political
351–353
contra positivism 431
Cornell Realists 122
cosmic exile 100
cosmos 39
Cottingham, J. 218–219, 223–224
counterfeit service of God 42
covariance claim 278n17
cowardice 50–51
Crane, T. 215
Crespo, R. F. 281
The Crisis of the European Sciences and Transcendental
Phenomenology (Husserl) 65
critical theory 434–437
Critique of Practical Reason (Kant) 41
Critique of Pure Reason (Kant) 38
Critique of the Power of Judgment (Kant) 290
Crowell, S. 57
cultural anthropology 88
cultural evolution 422
cultural imperialism 436–437
curve fitting 197–198
D’Oro. G. 175n3
Dainton, B. 240, 241
Dancy, J. 135
Daodejing (Classic of the Dao and Its Power)
257–258
Darwin, C. 24, 28, 85–87
Darwinism 122, 430
Davidson, D. 179, 181
Daybreak (Nietzsche) 49, 50–51
De Anna, G. 122–132

465

Index
De Caro, M. 1–3, 20, 71, 127, 166, 177, 182, 190,
198–199, 205–216, 250, 284, 340, 354, 456
De Igne (Kant) 36–46
Deism 230
Deleuze, G. 35n3, 261
Deligiorgi, K. 291
Dell’Utri, M. 455–463
Democritus 207
Dennett, D. 375–376
dependence claim 278n17
Derrida, J. 47
Descartes, R. 19, 193, 207, 235, 277, 302
Descartes, Rene 21–24
deVries, W. 110–120, 378
Dewey, J. 1, 83, 84, 86, 88, 90, 91n6, 192, 211,
322, 339
direct realist theories 302
disappearing agent 209
disenchantment 429
disenchantment of the world 62
displacement 403
Doctrine of Types 48
dogmatism 302
dokei 15n7
Douglas, A. 17–25
doxa 72
Drummond, J. 59–60
dualism 79
Dubouclez, O. 24
Dummett, M. 19, 23
Dupré, J. 77, 187n1, 210, 438n27
earth sciences 36
Ebbs, G. 97–106
The Ego and the Id 406n27
Einstein, A. 77, 195
eliminitavist naturalists 134, 313
Ellis, F. 128–129, 218–224
emancipatory interest 434–437
embodiment 73–78
emotional coloring 65, 66
emotivism 123–124, 336n5
empathy 270–273
empiricism 84
Empiricism and Subjectivity (Deleuze) 35n3
emptiness 258
Enlightenment 429
Enlightenment naturalism 230
An Enquiry Concerning Human Understanding
(Hume) 26
An Enquiry Concerning the Principles of Morals
(Hume) 26
ens realissimum 38, 41
episteme 72
epistemological disjunctivism 301
epistemological naturalism 31, 441–442, 443
epistemology 102–105

error theoretic views 313
The Essayer (Galilei) 206–207
Esthétique du Mal (Stevens) 137
The Ethical Demand (Løgstrup) 326
ethical naturalism 7–8, 331–332
ethical realism, optimism about 139
ethics 320–327
Ethics (Spinoza) 17–18
Ethics and the Limits of Philosophy (Williams) 137
ethnical naturalism 148n21
ethnical normativity 326–327
Eudemian Ethics (Aristotle) 8–9
evolutionary pluralism 84
evolutionary theory 249; practical consequences of
85–87; theoretical consequences of 83–85
evolved brain 421
evolved mind 421
e-world 160–161
expanded naturalism 128–129
expansive naturalism 129
explication 99–101
expressivism 156, 162n9, 330, 336n3, 336n5
faith 40
fallacy 321–322
fallacy of subtraction 106
Farina, Fl. 7–15
fecundity of emptiness 258
Ferguson, A. 43
fetishism 42
Feyerabend, P. 372, 377
Fiddes, P. 222
Film, Art and the Third Culture (Smith) 267
Fine, A. 77
Fine, K. 418n3
Fink, H. 68n2, 282, 320–327
first conjecture 104
“The Fixation of Belief ” (Peirce), Peirce, C. S.
87–88
Fodor, J. 50
folk physics 3
folk psychology 3, 134, 137, 270, 388
folk science 388
Foot, P. 7, 122–132, 322, 340–341, 353; defence
of ethical naturalism 123–127; liberal naturalism
and normativity 130–132; “Morality as a System
of Hypothetical Imperatives” 125; Natural
Goodness 123, 124, 130; naturalism 127–130;
position in ethical naturalism 122–127
Forman, D. 8; on nature (physis) 8–10
formlessness 258
Foucault, M. 28
Frankfurt School 434
free association 395–402
freedom 411
freedom of the will 39–40
French Revolution 138–139

466

Index
Freud, S. 28, 44, 383–405, 406n1; Dora case
408n88; The Ego and the Id 406n27, 407n58;
Interpretation of Dreams 405n1, 407n58; Jokes and
their Relation to the Unconscious 407n58; The
Psychopathology of Everyday Life 407n58
Friedman, M. 432
functional pluralism 156–157
Gadamer, H. G. 227, 229
Galilei, G. 206–207
Gallagher, S. 167, 175
Gardner, S. 406n25
Geist 60–61, 65, 68n3, 73
Geistigkeit 62
Geneaology (Nietzsche) 49
German Materialists 48
Gibson, J. J. 284, 310
Giladi, P. 429–439
Glock, H.-J 172, 441–453
God 19–21, 41, 91n6; conception as a supernatural
person 221; liberal naturalism and 218–224;
naturalism and 218; rethinking 221–223;
supernaturalism and 218–219
God, Value, and Nature (Ellis) 218–224
Goodman, N. 142
Grelling’s paradox 99
Griffin, J. 224
Grigg, R. 402–403
Grünbaum, A. 383–384, 384–385, 390
Gustafsson, M. 107n20
Guyer, P. 287
Habermas, J. 321, 384, 429–439; on emancipatory
interest 434–437; Between Facts and Norms 434,
436; Knowledge and Human Interests 429,
430–434; on substantive rationality 430; Theory
of Communicative Action 434; transcendental
pragmatism of 430–434
Habermas, Jürgen 434
habit (hexis) 10–11; unidirectionality of 12–13
Hacker-Wright, J. 129
Hacking, I. 187n5
Haidt, J. 52
hallucination 300
hallucinations 300
Halvorson, H. 249
Hanna, R. 433
Haraway, D. 185
hard naturalism 321, 324, 455
Hare, R. M. 138, 320, 353
Harris, S. 119n4
Haugeland, J. 180
Haught, J. 375
Hawthorne, J. 193
Hayek, F. 372
Heidegger, M. 47, 56, 136, 445

Hepburn, R. 292
Herder, J. G. 50
hexis 7, 10–11
Heyndels, S. 447
highest good 41
Hilbert, D. 312
Hintikka, J. 414, 415
Hobbes, T. 207, 331–332
Homo sapiens 420, 421
Homo socialis 422
Honest to God (Robinson) 222
Hook, S. 192
Hookway, C. 150n44
Hopkins, R. 406n25
Horkheimer, M. 434
human nature 9, 27–28, 30, 32, 34, 42–45, 48,
123, 160, 354, 446, 449–450
human sciences 419–428; liberal naturalism and
424–425; naturalisation of 425–427; naturalism
and 419–422
Hume, D. 1, 26–35, 103, 235, 373, 406n27, 446;
An Enquiry Concerning Human Understanding
(Hume) 26; An Enquiry Concerning the Principles
of Morals (Hume) 26; as liberal naturalist 30–32;
as scientific naturalist 27–30; therapeutic liberal
naturalism 32–34; A Treatise of Human Nature
(Hume) 26, 27, 29, 33
Hume’s fork 373
humility 50–51, 220
Hursthouse, R. 7, 125, 130
Hussain, N. 334
Husserl, E. 1, 3, 55–68, 68n4, 73, 240; The Crisis of
the European Sciences and Transcendental
Phenomenology (Husserl) 65; critique of
naturalism 55–57; Ideas II 60, 63; Introduction to
Ethics 63; Introduction to Philosophy 60; on nature
63–67; Nature and Spirit 60; Philosophy as a
Rigorous Science 56; principle of all principles 75
Husserl, Edmund, on natural teleology 59–63
Husserliana (Husserl) 63
Hutto, D. 165–175
Huxley, A. 122
hypertraditional societies 136
idealism 411, 413
Ideas II (Husserl) 60, 63
identity theory 445
identity-preserving properties 66
idolatry 42
immanent naturalism 258
immaterialism 413
immortality 41
indeterminacy 105
indexicality 414–416
indirect realist theories 302
indispensability arguments 90
Individuals (Strawson) 445

467

Index
injustice 141–142
integral theory of moral perception 365
intellectual economy 359
intentional pleasure 286–289
intentional state 115
interests 432
The Interpretation of Dreams (Freud) 393
Interpretation of Dreams (Freud) 405n1, 407n58
intersubjectivity 291–293
Introduction to Ethics (Husserl) 63
Introduction to Logical Theory (Strawson) 446
Introduction to Philosophy (Husserl) 60
intuitionism 124, 332, 336n5; moral realism and
363–364; new 363; rational 352, 363; social 52
inversion 395
Ionian School 1
irreducibility 247–248

Krishnan, N. 353–356
Kukla, R. 185

Jackson, F. 190, 305
James, W. 28, 83, 85, 91n6
Jardine, N. 150n44
Jefferson, T. 141–142
Johnston, M. 241–242
Johnston, S. 413
Jokes and their Relation to the Unconscious (Freud)
407n58
Kant, I. 36–46, 41, 125, 135, 193, 207–208, 211,
274, 275, 283, 285, 287–290; on action 39; on
belief 40; body 39; cosmos 39; Critique of
Practical Reason 41; Critique of Pure Reason 38; De
Igne 36; freedom 39; freedom of the will 39–40;
on highest good 41; human nature 42–45; mind
39; on religion 41–42; on sensibility and
understanding 37; on the sublime 293–295; on
symbolism 41–42; on theoretical reason 37–39;
on transcendental conditions 37; transcendental
unity of apperception 37
Kantian constructivism 329–335; constitutivist
claim 335; ethical naturalism and 331–332;
punishment and 330–331; universal authority
and 333–335; wronging and punishing 330–331
Kantianism 431
Keil, P. 449
Kelly, K. 195
Kemp Smith, N. 29
Khlentzos, D. 142
Kim, J. 278n17
Knowledge and Human Interests (Habermas) 429,
430–434
Knowles, J. 157
Kornblith, H. 73, 368n1
Korsgaard, C. 61, 237, 333
Krauss, L. 375
Krikorian, Y. H. 322
Kripke, S. 142, 145, 147, 412

La Rochefoucauld, F. 53
Laden, A. 354
Ladyman, J. 192, 245
Land, E. 311
Lange 180
The Language of Morals (Hare) 320
Language, Truth and Logic (Ayer) 320
Larrabee, H. 192
Laudan, L. 373
lawful patterns 180–181
Lear, J. 394, 398
Lebenswelt (life-world) 71
left wing 371
Leibniz, G. 19–20
Leiter, B. 47–53
Levin, S. 83–91
Lewis, C. I. 412, 414
Lewis, D. 27, 194, 376, 412, 415
liberal functionalism 310
liberal naturalism: actualism as 411–418; aesthetics
and 267–277; agenda and strategy of 57;
Chinese naturalism, compared with 261;
classical pragmatism and 83–91; core
commitments of 166; critical theory and
434–437; epistemology of 71; ethics and
320–327; Foot 122–132; God and 218–224;
Hume 26–35, 30–32; Husserlian
phenomenology and 55–67; Merleau-Ponty
70–73; moral reasons 340–343; naturalistic
stance 198–199; ontological commitments
245–453; ontology of 71; origins and prospects
of 205–216; orthodox naturalism and 178–179;
overview 1–3; perceptual affordance 281–285;
philosophy of perception and 299–315; as place
of value in the natural world 343–344;
placement problems 205–206; pragmatic
naturalism and 89–91; psychoanalysis and
383–405; Putnam 455–463; radical naturalism
and 177–187; reconciliation problem and
210–215; relaxed naturalism and 165–174;
relaxed realism and 450–453; scientism and
371–379; Sellars 110–119; Spinoza 17–25;
subject naturalism and 152–162; Taylor
227–234; version 2.0 252; Williams 134–148
liberal, defined 411
liberalism: moderate naturalism and 250–251;
ontological 161; political 349–354
life-world 1
likelihoods 196–197
location problems 154, 248
Locke, J. 235–236, 304
Løgstrup, K. E. 326
Longino, H. 378, 379
Lyotard, F. 293–294

468

Index
Macarthur, D. 1–3, 127, 160, 166, 177, 179, 182,
190, 198–199, 236, 242, 251, 267–279, 282,
284, 302, 340, 354, 387–389, 391–392, 456
Machery, E. 169, 170–171, 173, 210
MacIntyre, A. 122, 129, 326
Mackie, J.L. 137, 215n1, 329, 343
Maddy, P. 192
Making Sense of Humanity (Williams) 137
manifest causes 389
manifest image 1, 113–114, 245, 388–389
Marxism 434
materialism, non-reductive 363
Materialists 48
Mazzoni, J. 206, 215n2
McCabe, H. 223
McDowell, J. 2, 8, 27, 30–31, 53, 57, 58–59, 60,
68n2, 70, 89, 90, 122, 129, 130, 152, 160–162,
167, 179, 181, 211, 213, 218–220, 224, 227,
232, 282, 299, 302, 306, 307–311, 324, 339,
378, 433; “Are Moral Requirements
Hypothetical Imperatives?” 340; cases for liberal
naturalism 340–344; Mind and World 308;
naturalism of second nature 450; “Noncognitivism and Rule-Following” 344;
“Projection and Truth in Ethics” 344; “Two
Sorts of Naturalism” 342; “Values and
Secondary Qualities” 343
McGinn, M. 131, 210
McMahon, J. A. 281–296
McTaggart, J. E. 238–240
meaningfulness 106
mechanical properties 303–304
Meditations (Descartes) 277
Meehl, P. 120n14
Mellor, D. H. 215
Mencius 257
Menschenkenners 391
Merleau-Ponty, M. 56, 70–81, 231; liberal
naturalism and 70–73; as liberal naturalist 78–80;
on perception and embodiment 73–78;
Phenomenology of Perception 74, 80n3; Structure of
Behavior 73; The Visible and The Invisible
76–77, 79
Merritt, M. 293–294
metaphilosophical naturalism 441–442, 443
metaphysical glue 416
metaphysical naturalism 31, 253n8
metaphysical realism 147, 456–462
metaphysics 417; base rates 196–197; curve fitting
197–198; likelihoods 196–197; model selection
197–198; moderate naturalism and 250–251;
realism in 456–462; simplicity 195; simplicity in
metaphysical reasoning 198–199; of value 368n3
Metaphysics (Aristotle) 8, 11
methodological doctrine 190–191
methodological naturalism 78, 87–88, 178–179,
253n8, 359, 368

methodological separatism 72, 78
Mignolo, W. 436
Miguens, S. 339–347
Mill, J. S. 97, 320, 323, 331
Millikan, R. 378
Millum, J. 129
Mind and Cosmos (Nagel) 58
Mind and World (McDowell) 58, 308
mind-brain theory 445
mirror neurons 272
M-Naturalism 48, 53
model selection 197–198
model-theoretic argument 142
moderate naturalism 246, 250–252
monism 215
monistic naturalism 361
Moore, A. 145, 150n44
Moore, G. E. 1, 3, 122–123, 213, 305, 320, 332,
353, 358, 364
Morag, T. 383–405
The Moral Arc (Shermer) 374
moral judgments 48
The Moral Landscape (Harris) 374
moral perception 365–366
moral properties 329
moral realism 50; anti-naturalism 334; in the
intuitionist tradition 363–364; Kantian
constructivism and 330, 333–334, 336n5; liberal
naturalism and 219; normative explanation
364–366; rational intuitionism and 352
moral reasons 340–343
moral sources 233–234
morality 48, 293–294
Morganti, M. 245–253, 251
Mounce, H. O. 31–32
Murdoch, I. 218–219, 220–221, 223, 353
mysterianism 210
Nagel, E. 441
Nagel, T. 23, 58–59, 208, 237–238, 406n25
naive empiricism 43
naïve-realism 304–305, 310
naïve-realist/relational views 303
Nanay, B. 289
Natali, C. 15n7
natural attitude 73
Natural Goodness (Foot) 123, 124, 130
natural objects 39
natural phenomena 161
natural realism 309
natural selves 242–244
natural teleology 59–63
naturalised conceptual analysis 169–174
naturalism 2, 7, 19–20, 127–130, 134–135,
449–450; characterisation of 268; Chinese
255–263; epistemological 441–442, 443;
historicising 229–231; for human sciences

469

Index
419–428; Husserl’s critique of 55–57;
intellectual economy and 359;
metaphilosophical 441–442, 443; monistic 361;
Nietzsche 47–53; nonreductive 269, 360–362;
non-revisionary 441–453; ontological 441–442,
443; perception and 303–307; phenomenology
and 56; pragmatic 87–89; psychoanalysis and
387–392; scientific 87–89; of second nature
160–162; soft 56; as a stance 192–193; theory of
justice and 353–356; Twentieth-Century 358;
types of 441–442
Naturalism and Ontology (Sellars) 110–111
Naturalism and the Human Spirit 192
naturalistic attitude 66
naturalness 247–248
nature 8–10; character as second 13–15; culture
185; naturalistic understanding of 323–324
Nature and Spirit (Husserl) 60
Naturwissenschaften 429, 430
Nehamas, A. 47
neo-Aristotelian eudaimonism 137–138
neo-Confucianism 259–260
neo-Darwinist evolutionary theory 182–184
neo-Kantianism 45
Neoplatonism 20, 21
Neo-Positivism 127
Neta, R. 247
Neta’s dilemma 247, 250–251
Nettle, D. 428n9
Neurath, O. 193, 427
neuroaesthetics 209
neuro-computational state 308
neurophysiology 270
New Atheists 376
New Intuitionism 363–364
Newton, I. 195
Nicholls, S. 296n16
Nicomachean Ethics (Aristotle) 7, 12, 14
Nietzsche, F. W. 43, 47–53; The Anti-Christ 49;
Beyond Good and Evil 49; Daybreak 49, 50–51;
Geneaology 49; ressentiment 48; Twilight of the Idols
49, 51–52
Noe, A. 310
Nola, R. 175n6
nonnaturalism 79–80, 135, 213, 300, 332,
353–354, 368n1, 419, 423
nonperceptually represented perceptual properties
289–291
non-reductive materialism 111, 300, 363
non-reductive naturalism 270, 321, 360–362,
444–446, 449, 452
nonrevisionary naturalism 441–453
nonsupernatural order 275
noologists 37
normative domain 361
normative explanation 364–366
normative properties 364–366

normativity 130–132, 366–368; two-dimensional
biological 182–184
noumena 208
Novum Organum (Bacon) 378
Nussbaum, M. 408n86
O’Shea, J. 111
object naturalism 153, 253n4; assertions 156–157;
functional uniformity 156; Price’s rejection of
157–158; rejection of 154–157;
representationalism and 155–156
objective being 21
objective views 207–208
objectivity 366–367
observation sentence 103
Ockham’s razor 75, 359–360
On Certainty (Wittgenstein) 448
O’Neill, O. 333
ontological commitments 245–453; criteria for
247–248; downward path 249–250;
irreducibility 247–248; liberal naturalism and
246–247; naturalness 247–248; Neta’s dilemma
247; presuppositions 247–248; supernaturalism
and 249–250; upward path 247–248
ontological doctrine 190–191
ontological liberalism 161
ontological naturalism 441–442, 443
ontological pluralism 360–362; abstract entities
360–361; normative domain 361; property
identity and 361–362
ontological reduction 361–362
ontological structures 417
ontology 101–102
original image 120n6
originary givenness 75
originary presentive intuition 76
orthodox naturalism 70, 72, 79, 81n7,
178–179, 181
other focused personal imagining 271
Overgaard, S. 175n3
Palmer, S. 311
Papineau, D. 77, 339, 428n1
parapsychology 378
Parfit, D. 138, 238–240
Parmenides 20
particularism 135
part-part relations 394
part-whole relations 395
Patton, P. 349–356
Peirce, C. S. 83, 457; “The Fixation of Belief ”
87–88
perception 73–78, 299–315; adverbial views 302;
color and 310–314; common kind of
assumption 301–302; direct realist theories 302;
hallucinations and 300; indirect realist theories

470

Index
302; indirect views of 305; naïve-realist/
relational views 303; naturalism and 303–307;
phenomenalist/idealist theories 302;
representational/intentionalist views 302–303;
sense-data views 302; traditional problems of
300–301; transactional 309
perceptual affordance 281–285, 286, 288, 290,
292–295, 296n16
perceptual properties 364
Pereboom, D. 209, 210
Perfect Model Model 178
perfectability of human nature 44
persons: concept of 242–243; selves and 240–242
persons as subjects 236–238
perspectivism 123
phenomenal qualities 445
phenomenalist/idealist theories 302
phenomenology 56
Phenomenology of Perception (Merleau-Ponty) 70, 74,
80n3
philo-physics 80
philosophical anthropology 231–233
Philosophical Investigations (Wittgenstein)
130–131, 147
philosophical naturalism 26, 28, 153, 177, 184,
192–193, 193, 251, 339, 359–360, 368
philosophy 45, 431; scandal of 445; as a secondorder critical discipline 431; therapeutic liberal
naturalism and 32–34
Philosophy as a Rigorous Science (Husserl) 56
Philosophy in its Proper Bounds (Machery) 170
philosophy of nature 165–167
physical 120n14
Physics (Aristotle) 8, 11
physis 7, 8–10
Pigliucci, M. 371–379
placement problems 84, 154, 208–211, 237,
238–240, 248, 435–436, 438n26
Platner, E. 42
Platonism 21, 23, 128, 215n2, 218, 220, 250, 443
Platonists 206–207, 220, 339, 354
pluralism 84; causal 389; conceptual 457–460;
evolutionary 84; functional 156–157; monism
and 215; ontological 360–362; radical 346;
scientific 181, 423
A Pluralistic Universe (James) 91n6
pneuma 260
Poincaré, H. 420
political liberalism 349–354
Political Liberalism (Rawls) 349–353, 354–355
Popper, K. 195, 372, 385
possibilism 412–413, 417, 418n2
postulational methods 114
pragmatic anthropology 36, 42–43, 434
pragmatic belief 40
pragmatic naturalism 84, 87–91

pragmatism 441; classical 83–91; transcendental
430–434
Pragmatism (James) 85
prescriptivism 123–124
presuppositions 247–248
Price, H. 70, 152–163, 179, 236, 253n4, 435; on
e-world 160–161; on philosophical naturalism
153; on placement problems 205; rejection of
object naturalism 157–158
Price, R. 352
primacy of practice 90
primary qualities 304
primitivism 312
Principia Ethica (Moore) 320
principle of all principles 75
principle of continuity 84
Principles of Philosophy (Descartes) 19
Pritchard 302
property identity 361–362
propositional approach 414
Protestantism 430
proto-science 3
Pryor, J. 302
pseudoscience 373, 376, 378
psychoanalysis 383–405; causal claim of 386;
complementary therapeutic claim of 386;
foundations of 384–387; free association and
395–402; naturalism and 387–392; socialpsychological box 402–405; unconscious and
392–402
The Psychopathology of Everyday Life (Freud) 407n58
punishment 330–331
Puritanism 430
purposiveness 285–286
Putnam, H. 2, 70, 90, 142, 147, 188n13, 283, 299,
307–311; on liberal functionalism 310; on liberal
naturalism 455–463; metaphysical glue 416;
natural realism 309; on realism in metaphysics
457–460; on scientific naturalism 455; “The
Threefold Cord” lectures 307
qualia 248, 310
queer 215n1
The Quest for Reality (Stroud) 313
Quine, W. V. O. 1, 28, 48, 77, 87, 97–106, 104,
107n20, 191, 247, 250, 275, 321; on abstract
entities 451; epistemology 102–105; fallacy of
subtraction 106; indeterminacy 105; logistical
99; on mathematics 107n40; meaningfulness
106; ontology 101–102; regimentation and
explication 99–101; second conjecture 104;
working from within 98
racism 142
Radical Enlightenment 187n11
radical immanentism 219, 223, 378

471

Index
radical naturalism 177–188
radical pluralism 346
Rahner, K. 222
Raleigh, T. 299–315
Randall, J. 192, 219
ratcher effect 187n8
rational intuitionism 352, 363
rational wolf 339–347; animality 339–340; liberal
naturalism and 340–343; rationality 339–340;
thought experiment 339–340
rationalisation 429
rationalism 84, 333, 367
rationalists 37
Rawls, J. 70, 321, 349–356; “Justice as Fairness:
Political not Metaphysical” 353, 355–356; “The
Idea of an Overlapping Consensus” 353; Political
Liberalism 349–353, 354–355; A Theory of Justice
349–350, 353–354, 355, 356
ready-made world 456
realism 351–353, 412–414; in metaphysics
456–462
The Reality-Structure of the Life-World: Nature as
Abstract Core-Layer of the World 63
reason 37–39
reconciliation problem 210–215
Redding, P. 160, 411–418
reductive naturalism 100, 321, 330, 334, 336n5,
444–446, 450, 453
reductive physicalism 311–312
re-enchantment 60
reflective judgment 285–286
regimentation 99–101
Reichenbach, H. 193
reidentifiability 238–240
reidentification 236
relata 155
relaxed naturalism 68n2, 165–175, 278n17;
contrasting visions of nature and 167;
philosophy of nature 165–167; scientific
naturalism and 165, 168–169
relaxed realism: liberal naturalism and 450–453;
naturalised conceptual analysis 169–174
religion 41–42
representation relation 155–156
representational/intentionalist views 302–303,
306–307
representationalism 155–156, 269, 438n26
ressentiment 48
restrictive naturalism 30–32, 152
Reynolds, J. 70–81
Reynolds, Jack 56
Richardson, H. 353
Ricoeur, P. 384
right wing 371
Ritchie, J. 190–202, 192
Robinson, J. 222
Romantics 233–234

Rorty, R. 47, 117, 139, 147, 230, 232, 378
Rosenberg, A. 27, 167, 168, 173, 209, 210, 216n6,
373, 376
Rosenberg, G. 205
Rosenberg, J. 378
Ross, W. D. 192, 245, 352, 361, 364
Rouse, J. 177–188; Articulating the World 179, 180;
liberal and radical conceptions of 184–187;
liberal and radical conceptions of normativity as
natural 184–187; orthodox naturalism 178–179;
scientific understanding 179–182; on temporal
externalism 186; two-dimensional biological
normativity 182–184
Russell, B. 302, 305
Salmon, W. 446
Sartre, J. P 74, 407n57
Savile, A. 278n17
Savini, M. 22
Scepticism and Naturalism (Strawson) 442, 444–446,
447, 448, 449–450
Schechtman, M. 237
schemata 41
science 376–378; common sense and 206–207;
empirical 278n17
Scientia Intuitiva 23
scientia mensura principle 72
scientific creationism 376
scientific image 113, 114–117, 118–120, 180
scientific naturalism 27–30, 87–89, 100, 165,
168–169, 228–231, 245, 246, 358–369, 455;
definition of 208–209; historicising naturalism
229–231; intellectual economy and 359;
methodological doctrine 190–191; moral
perception and 365–366; moral realism and
363–364; nonreductive 268–269; nonreductive
naturalism and 362–363; normativity 366–368;
objectivity 366–367; Ockham’s razor and
359–360; ontological doctrine 190–191;
ontological pluralism 360–362; placement
problems and 208–211; politics and 367–368;
religion and 367–368; as a troubled doctrine
190–191
scientific realism 78, 111–112, 117, 119, 139, 229,
359, 367, 376
scientific understanding 179–182
scientism 216n6, 219, 371–379, 431, 433; debates
about 375–376; definition of 371–374; science
and 376–378
scientistic naturalism 1, 2
script elaborator 296n16
Searle, J. 206, 322
second conjecture 104
Second Philosophy 192
Self-Constitution (Korsgaard) 61
self-givenness 75
Sellars, R. W. 358, 441

472

Index
Sellars, W. 30, 31, 72, 73, 79, 110–120, 154, 158,
245, 267, 378–379, 387, 433–434, 441;
“Counterfactuals, Dispositions, and the Causal
Modalities” 112–113; manifest image 113–114;
Naturalism and Ontology 110–111; “Philosophy
and the Scientific Image of Man” 113–114, 118;
scientific image 114–117, 118–120; scientific
realism 117; scientism and 371–379; Sellarsian
naturalism 110–113
selves 235–255; naturalism 242–244; persons and
240–242; persons as subjects 236–238;
reidentifiability 238–240; temporal perspective
238–240
Semantic Analysis (Ziff) 324
semantic deflationism 155, 156
semantic ladder 438n26
semantic temporal externalism 188n13
sense-data 302, 305, 315n10, 315n12
The Senses Considered as Perceptual Systems
(Gibson) 284
Sensus Communis 290, 292
Shah, N. 334
Shapiro, L. 152–163
Shapshay, S. 291
Shermer, M. 374
A Short History of Ethics (MacIntyre) 326
Short Treatise on God, Man, and His Wellbeing
(Spinoza) 17
Sider, T. 193
Sidgwick, H. 352
similarity 395
Singer, P. 138
skepticism 32–34, 48, 77, 103, 142–143, 271, 277,
303, 352, 417, 444–446
slavery 141
smallism 79
Smith, A. 43
Smith, B. 26–35
Smith, M. 267–270, 273–274
Smith, N. 227–234
Smolensky 428n9
Snow, C. P. 267
social brain 420
Social Darwinism 122, 127
social intuitionism 52
social naturalism 321, 449
Social Philosophy (Fink) 326
social world 184
Sociobiology: The New Synthesis (Wilson) 122
soft naturalism 56, 321
Song dynasty 260
Sources of Normativity (Korsgaard) 331
Sources of the Self (Taylor) 229, 230
space of reasons 89–90, 129
Spencer, A 85, 122, 320
Spinoza, B. 1, 17–25, 207; Cartesian philosophy
and 21–24; Ethics 17–19; on God 19–21;

naturalism 19–20; order of nature 19; Short
Treatise on God, Man, and His Wellbeing 17
Staiti, Andrea 55–68
Stalnaker, R. 412, 414, 415–416
Sterelny, K. 428n9
Stevens, W. 137
Stich, S. 210, 296n16, 346
Stokes, P. 235–244
Strangers on a Train (film) 273–274
Strawson, G. 235, 240, 241
Strawson, P. F. 321
Strawson, P.F 30, 70, 100–101, 113, 207, 441–453;
achievements 442; Bounds of Sense 445; on
criteria of identity 452; Individuals 445;
Introduction to Logical Theory 446; on naturalism
442–443; on non-reductive naturalism 449; on
relaxed realism 450–453; Scepticism and
Naturalism 442, 444–446, 447, 448, 449–450
Street, S. 334
Streumer, B. 137
strict naturalism 321
Strictly Scientific Naturalism 299
Stroud, B. 28, 31, 34, 103, 313, 368n1
structural realism 305
Structure of Behavior (Merleau-Ponty) 73
The Studies on Hysteria (Breuer and Freud) 384, 386
Stueber, K. 272
Sturgeon, N. 122, 139
subject naturalism 152–163, 177, 253n4; definition
of 153–154; liberal naturalism and 159–160;
naturalism of second nature and 160–162;
rejection of object naturalism 154–157
subjective views 207–208
sublime: intentional pleasure 286–289;
intersubjectivity 291–293; morality and
293–294; non-perceptually represented
perceptual properties 289–291; reflective
judgment 285–286
substantive rationality 429
supernatural entities 17
supernaturalism 1, 218, 218–219, 228–229, 247,
249–250, 323, 411; moral sources and 233–234
superstition 42, 140, 219, 228–229
supervenience 212–213, 273, 278n17, 314n4,
362, 422
Sydney Plan 175n6
symbolism 41–42
Takeuchi Information Criteria (TIC) 198
Tappan Zee Bridge 172
Taylor, Charles 227–234; criticisms of scientific
naturalism 228–229; Enlightenment naturalism
230; philosophical anthropology 231–233;
Sources of the Self 229, 230
teleology 20, 55, 59–63, 80, 131, 228–229, 324
Teller, P. 178–179
temporal externalism 186, 188n13

473

Index
temporal perspective 238–240
Teske, J.G. 36
The Foundations of Psychoanalysis (Grünbaum)
383–384
The Varieties of Goodness (von Wright) 324
theoretical reason 37–39
The Theory of Communicative Action (Habermas) 434
A Theory of Justice (Rawls) 349–350, 353–354,
355, 356
theory of mind 272
theory theory accounts 272
therapeutic liberal naturalism 32–34
third cultural approach 267
Thomas, A. 345–346
Thompson, M. 125–126, 129
Thucydides 53
Tian 255–256
Tillich, P. 222
Timmons, M. 335
Tomasello, M. 187n8
Tractatus Logico-Philosophicus (Wittgenstein)
414–415
transactional perception 310
transcendental conditions 37
transcendental ego 74
transcendental phenomenology 56, 73
transcendental pragmatism 430–434
transcendental unity of apperception 37
Travis, C. 308–309
A Treatise of Human Nature (Hume) 26, 27, 29, 33
true naturalism 218, 221, 223
Truth and Method (Gadamer) 227
Truth and Truthfulness (Williams) 137
Twilight of the Idols (Nietzsche) 49, 51–52
The Two Cultures (Snow) 267
Two Sorts of Naturalism (McDowell) 339
Tye, M. 312
type-facts 48
type-identity theory 209, 445
Tyson, N. 374–375
ufology 378
unconscious 392–402
universal authority 333–335
Ur-naturalism 193
user illusion 78
Utilitarianism (Mill) 331
Value and Context (Thomas) 345
van Frassen, B. 192
van Inwagen, P> 208
A View from Nowhere (Nagel) 208
The Visible and The Invisible (Merleau-Ponty)
76–77, 79

Voltolini, A. 20
von Wright, G. H. 324–325
Wallace, M. 210
Walton, K. 271
Weber, M. 62, 429–430
Weinberg, S. 375
What is This Thing Called Science (Chalmers) 377
White, S. 290, 306
Whyman, T. 235
Wilde. O. 275
The Will to Believe and Other Essays in Popular
Philosophy (James) 91n6
Williams, B. 134–150, 168, 173; Ethics and the
Limits of Philosophy 137, 345; on French
Revolution 138–139; Making Sense of Humanity
137; on naturalism 135; neo-Aristotelian
eudaimonism 137–138; rejection of ethical
naturalism 148n21; on slavery 141; Truth and
Truthfulness 137
Wilson, E. O. 122
Wilson, M. 187n10, 190–202
Wiltsche, H. 75
Windelband, W. 436
Wisssenschaften (sciences) 47, 48, 52
Wittgenstein, L. 31, 130–131, 174, 211, 339, 373,
391, 405, 414, 446, 448, 450; On Certainty 448
Wittich, C. 24
Wolffian philosophy 36
Wollheim, R. 406n25
Wood, A. 36–46
Woodsworth, W. 289
Woodward, J. A. 389
Wright, C. 302
wronging 330–331
wssenschaftlich 433
Xunzi 256
yin-yang 259
Young, I. M. 436–437
Zahavi, D. 78, 235
Zarahustra 51–52
Zeki, S. 311
Zhang Zai 260
Zhongyong 257
Zhou dynasty 255
Zhu Xi 260
Zhuangzi 258–259
Ziff, P. 324
Zimmerman, D. 193

474