Stroud - Philosophers Past and Present - Selected Essays (1)

PHILOSOPHERS PAST AND PRESENT

This page intentionally left blank

Philosophers Past and Present

Selected Essays

BARRY STROUD

CLARENDON PRESS · OXFORD

3 Great Clarendon Street, Oxford ox2 6dp

Oxford University Press is a department of the University of Oxford.It furthers the University’s objective of excellence in research, scholarship,

and education by publishing worldwide inOxford New York

Auckland Cape Town Dar es Salaam Hong Kong Karachi Kuala Lumpur Madrid Melbourne Mexico City Nairobi

New Delhi Shanghai Taipei TorontoWith offi ces in

Argentina Austria Brazil Chile Czech Republic France Greece Guatemala Hungary Italy Japan Poland Portugal Singapore South Korea Switzerland Thailand Turkey Ukraine Vietnam

Oxford is a registered trade mark of Oxford University Press in the UK and in certain other countries

Published in the United Statesby Oxford University Press Inc., New York

© in this volume Barry Stroud 2011

The moral rights of the author have been assertedDatabase right Oxford University Press (maker)

First published 2011

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means,

without the prior permission in writing of Oxford University Press, or as expressly permitted by law, or under terms agreed with the appropriate

reprographics rights organization. Enquiries concerning reproduction outside the scope of the above should be sent to the Rights Department,

Oxford University Press, at the address above

You must not circulate this book in any other binding or cover and you must impose the same condition on any acquirer

British Library Cataloguing in Publication DataData available

Library of Congress Cataloging in Publication DataLibrary of Congress Control Number: 2011929418

Typeset by SPI Publisher Services, Pondicherry, IndiaPrinted in Great Britain

on acid-free paper byMPG Books Group, Bodmin and King’s Lynn

ISBN 978–0–19–960859–1

10 9 8 7 6 5 4 3 2 1

For David

This page intentionally left blank

CONTENTS

Acknowledgements viii

Introduction 1

1. Our Debt to Descartes 11

2. Berkeley v. Locke on Primary Qualities 29

3. Colours and Powers 49

4. The Study of Human Nature and the Subjectivity of Value 65

5. Hume’s an Enquiry Concerning Human Understanding 103

6. Ayer’s Hume 117

7. Hume’s Scepticism: Natural Instincts and Philosophical Refl ection 144

8. “Gilding or Staining” the World with “Sentiments” and “Phantasms” 167

9. The Constraints of Hume’s Naturalism 189

10. Practical Reasoning 207

11. The Charm of Naturalism 223

12. The Transparency of ‘Naturalism’ 240

13. Anti-Individualism and Scepticism 256

14. Sense-Experience and the Grounding of Thought 273

15. The ‘Unity of Cognition’ and the Explanation of Mathematical Knowledge 290

16. Contemporary Pyrrhonism 307

17. Perceptual Knowledge and Epistemological Satisfaction 322

Index 335

ACKNOWLEDGEMENTS

The essays collected in this volume were originally published as follows:

‘Our Debt to Descartes’ in J. Broughton & J. Carriero (eds), A Companion to Descartes , Blackwell, Oxford 2008, pp. 513–525

‘Berkeley v . Locke on Primary Qualities’ Philosophy 1980, pp. 149–166 ‘Colours and Powers’ previously unpublished ‘The Study of Human Nature and the Subjectivity of Value’ in G. Peterson

(ed.), The Tanner Lectures on Human Values vol. X, Salt Lake City 1989, pp. 213–259

‘Hume’s Enquiry Concerning Human Understanding’ Afterword to D. Hume, Una investigación sobre el entendimiento humano (Spanish translation with parallel English text of D. Hume, An Enquiry Concerning Human Understanding ) ed. and tr. V. Sanfélix and C. Ors, Madrid 2004, pp. 383–399

‘Ayer’s Hume’ in L. Hahn (ed.), The Philosophy of A. J. Ayer , (Library of Living Philosophers), Chicago 1992, pp. 609–631

‘Hume’s Scepticism: Natural Instincts and Philosophical Refl ection’ Philosophical Topics 1991, pp. 271–291

‘ “Gilding or Staining” the World with “Sentiments” or “Phantasms” ’ Hume Studies 1993 pp. 253–272

‘The Constraints of Hume’s Naturalism’ Synthèse 2006 pp. 339–351 ‘Practical Reasoning’ in E. Ullmann-Margalit (ed.), Reasoning Practically ,

Oxford University Press, New York 1999 pp. 27–38 ‘The Charm of Naturalism’ Proceedings and Addresses of the American

Philosophical Association , November 1996, pp. 43–55 ‘The Transparency of “Naturalism” ’ The Romanell Lecture, Proceedings

and Addresses of the American Philosophical Association , November 2009, pp. 158–169

‘Anti-Individualism and Scepticism’ in M. Hahn & B. Ramberg (ed.), Refl ections and Replies: Essays on the Philosophy of Tyler Burge , MIT Press, Cambridge, Mass. 2003, pp. 15–27

‘Sense-experience and the Grounding of Thought’ in N. Smith (eds), Reading McDowell: on “Mind and World” , Routledge, London 2002, pp. 79–91

‘The “Unity of Cognition” and the Explanation of Mathematical Knowledge’ Philosophical Topics 2001, pp. 415–428

‘Contemporary Pyrrhonism’ in W. Sinnott-Armstrong (ed.), Pyrrhonian Skepticism , Oxford University Press, New York 2004, pp. 174–187

‘Perceptual Knowledge and Epistemological Satisfaction’ in J. Greco (ed.), Sosa and his Critics , Blackwell, Oxford 2004, pp. 165–173

INTRODUCTION

The essays collected here examine and discuss views and attitudes held by different philosophers at different times. The aim through-out is to make progress on the problems those philosophers were concerned with. Seven of the essays are about Hume or about move-ments of thought in which he played a signifi cant part. Three others are about thinkers who set certain tasks for philosophy before Hume put pen to paper. The remaining seven or eight consider the views of philosophers of the twentieth and twenty-fi rst centuries. Many of the essays confront in one form or another the implications of a certain traditional conception of “immediate experience.” Another theme is the prospects of a defensible form of something called “naturalism.” Some essays explore the sources of confl ict between those two ideas.

All but one of the essays have been printed before; they appear here with the permission of their original publishers. Each was writ-ten for a particular occasion and is meant to stand on its own. There is some overlap between a few of the essays, but I have left them as they were fi rst printed. I collect them together here to make them more conveniently available and to encourage further refl ection on the relations among them and among the different philosophers and issues they are about. Despite their dealing with particular philoso-phers and ranging over a large portion of the history of philosophy, I am reluctant to call them without qualifi cation essays in the history of philosophy. Philosophy certainly has a history, and everything that happens in philosophy is part of that history. In that sense what each of these essays is about is part of the history of philosophy. But in that same sense any essay dealing with problems put forward by philosophers is about something that is part of the history of philosophy. And if such an essay is any good, it will be part of that very history. But it will not for that reason alone be an essay in the history of philosophy.

Some of these essays are historical in the sense of describing move-ments or infl uences or developments of philosophical lines of thought over time. But most of them are meant to contribute in one way or

2 Introduction

another to philosophical understanding, and so to philosophy. That does not mean history does not come into it. History is obviously involved in understanding philosophers of the past. To understand Hume, or Berkeley, or Descartes, you need to know what was really at stake for that philosopher: what issues he thought confronted him, what alternatives he saw as feasible, what outcomes he deemed acceptable, and so on. Those are historical questions, and you need good answers to them to understand the problems those philoso-phers faced and what should be done about them.

But good answers to the same kinds of questions are needed in responding to the philosophical work of one’s contemporaries. It is perhaps easy to think that does not raise historical questions in the same way. That seems to me naive, and to ignore the importance of historical understanding for a proper grasp even of how things really stand right now and how they got that way. I think this is borne out by illuminating recent work in the history of analytic philosophy that has improved our understanding of many of the problems we now face. In any case, careful, informed exposition, criticism, and devel-opment of the problems and views of our contemporaries are essen-tial to any solid contribution to philosophy today. I think the same is true of similarly informed, sympathetic treatments of the problems and views of philosophers of the past. It is in that spirit that I offer these essays about philosophers from different times and places.

“Our Debt to Descartes” was published in A Companion to Descartes , edited by Janet Broughton and John Carriero (Blackwell, Oxford, 2008). A full account of Descartes’ philosophical legacy would touch on almost everything that has happened in philosophy since his time. But no treatment, however selective, could ignore Descartes’ posing the general epistemological problem of our knowledge of the external world. Deeply implicated within that prob-lem is a dualistic conception of the mind in relation to rest of the world. In one form or another dualism remains at the centre of cur-rent controversy in the philosophy of mind. Descartes’ defence of his dualism started from the distinctive character of the thought “Cogito.” This essay draws attention to some applications of that Cartesian idea not explicitly envisaged by Descartes. Some of its implications are at work in the philosophy of Kant, and others in more recent philosophical accounts of direct reference, especially of indexical terms. Both ways of exploiting the special character of ‘Cogito’-like thoughts appear to offer the prospect of thoroughly anti-Cartesian

Introduction 3

pictures of the mind and its relation to the world. But even such an outcome, arrived at in those ways, would be part of our still-growing debt to Descartes.

“Berkeley v. Locke on Primary Qualities,” published in Philosophy in 1980, examines Berkeley’s objections to Locke’s view that our ideas of so-called ‘secondary’ qualities like colour, taste, and smell do not “resemble” anything in the objects that cause them, while our ideas of “primary” qualities like weight, shape, and motion do “resemble” qualities in independent objects. Locke’s support for that view came from the fact that colours, tastes, smells, etc., have no place in the independent world as described by the most successful “corpuscular” scientifi c account of it. Berkeley argued that all the qualities we perceive—shape, size, and motion as well as colour, taste, and smell—depend on the state of the perceiving subject. That would appear to be no objection to Locke if his defence of his view does not rest on the distinctive “relativity” of the perception of “secondary” qualities. But Berkeley’s real opposition to Locke lies in the doctrine that esse is percipi , which he claims to be supported by the facts of perception as he understands them. That metaphysical doctrine implies the impossibility, and for Berkeley even the unintelligibility, of any qualities existing in unthinking, unperceived things. Berkeley was right that there is a real confl ict between his views and Locke’s.

“Colours and Powers” has not been printed before. I presented it at a conference in Oxford in 2003 honouring the work of Michael Ayers. In his masterly book on Locke Ayers criticizes Locke’s con-ception of colour-words as primarily naming simple ideas of colour while “secondarily” naming whatever it is in objects that causes those ideas. That view leaves open the possibility of different perceivers getting ideas of a different colour from the same object while each ‘secondarily’ applying the same colour-word to that object. Ayers thinks something like this possibility, although not Locke’s way of accounting for it, must be preserved. He thinks we must be able to describe the effects of coloured objects on perceivers by applying colour-terms not directly to objects but to those “sensory effects” themselves in order to capture “the phenomenal quality” of our “sensory experience.” I argue that we can describe those effects by using colour-words as true only of the objects we see, with no need for a distinctive application of them to something called “sensations” or “impressions.” I fi nd no case has been made for colour-words as standing only for the powers objects have to produce certain kinds of

4 Introduction

“sensations” or “sensory effects” in perceivers. I go on to sketch what I see to be some of the wider implications of this for the understand-ing of psychological states and attitudes generally.

“The Study of Human Nature and the Subjectivity of Value” com-prises the two Tanner Lectures on Human Values I delivered in Buenos Aires in 1988. They were published in Grethe B. Peterson (ed.), The Tanner Lectures on Human Values X (University of Utah Press, Salt Lake City, Utah, 1989). This long essay describes and explores some of the apparent consequences for the understanding of human evaluative thought of the Enlightenment conception of a “science of human nature” as Hume understands it. I discuss the parallel, explicitly drawn by Hume, between the subjectivity of val-ues and the traditional doctrine of secondary qualities, and suggest in each case how attachment to a fully naturalistic science of human nature can make acceptance of both those views seem unavoidable. That would leave us at best with a conception of a colourless, value-free world on to which we merely “project” whatever colours and feelings we fi nd in our immediate experience as we respond to what that independent world presents us with.

The essay “Hume’s Enquiry Concerning Human Understanding ” appears here for the fi rst time in English. It was published in Spanish in a facing-page English–Spanish edition of that Enquiry : David Hume, Una investigación sobre el entendimiento humano , edited and translated by Vicente Sanfélix Vidarte and Carmen Ors Marqués (Ediciones Istmo, Madrid, 2004). It describes the overall point and structure of Hume’s fi rst Enquiry and how he saw it as achieving more convincingly the goals he had set himself in his earlier and more diffi cult Treatise of Human Nature . Conclusions reached there about the role of reason and of the idea of cause and effect are applied in the fi rst Enquiry not only to our beliefs about the natural world but also to freedom of action and to the grounds for religious belief. The fi rst and the last sections of the Enquiry also explain better than before the distinctive character and importance of “scepticism” as Hume understands it.

“Ayer’s Hume” was my contribution to the Library of Living Philosophers volume The Philosophy of A. J. Ayer , edited by Lewis Edwin Hahn (Open Court, La Salle, Illinois, 1992). Ayer died just as my essay was fi nished, so there is no reply by him in that volume. Hume was a central fi gure in Ayer’s formation. He claimed that the version of logical positivism he defended in Language, Truth, and

Introduction 5

Logic and elsewhere descended directly from Hume. This had a pro-found infl uence on the interpretation of Hume’s philosophy for dec-ades and has not completely disappeared even today. The essay draws attention to some fundamental aspects of Hume’s views—not least the idea of an empirical study of the fundamental principles of human nature as an appropriate goal for philosophy—that that widely shared interpretation either ignores or distorts. There is also the question of the proper goal for philosophy as Ayer sees it quite independently of the interpretation of Hume.

In “Hume’s Scepticism: Natural Instincts and Philosophical Refl ection,” published in Philosophical Topics in 1991, I do my best to bring out the full richness, and the signifi cance both for philoso-phy and for human life, of what Hume means by the “scepticism” he endorses and recommends. It is not simply a philosophical doctrine or theoretical intellectual position, but more a set of attitudes and inclinations, and so potentially a way of life, perhaps arrived at in part by philosophical refl ection. I try to explain why Hume thinks such a stance or set of attitudes would be best for most human beings and also how he can support that endorsement while holding just such “sceptical” attitudes himself. I think the essay makes a case for the indispensability of a proper understanding of the intricate struc-ture and role of the “Conclusion” of Book One of the Treatise for any serious account of Hume’s philosophy.

“ ‘Gilding or Staining’ the World with ‘Sentiments’ and ‘Phantasms’ ” was published in Hume Studies in 1993. It asks how Hume can account for our having the thoughts we do about the beauty, deformity, good-ness, or badness of various things in the world, given that our feelings or “sentiments” play a crucial role in such evaluations. The problem Hume faces extends beyond evaluation to thoughts about the colour and smell and taste of things as well as to causation, enduring objects, and the self. Since we encounter no such facts directly in our experi-ence, our thoughts about them are generated by certain things hap-pening only in our minds. The essay argues that Hume’s view of the mind and thought as nothing more than the comings and goings of “impressions” and “ideas” lacks the resources to explain satisfactorily how we can even so much as think thoughts that in that way go “beyond” the sensory “data” we receive. I believe the diffi culties I identify apply even to “projectivist” accounts of this or that kind of thought that do not explicitly commit themselves to the details of a strictly Humean conception of the mind.

6 Introduction

“The Constraints of Hume’s Naturalism” was fi rst published in Synthèse in 2006. “Naturalism” can perhaps be understood as the project of explaining all aspects of human life, thought, and action on the basis of what can be found to be so in observable nature. There seems nothing essentially problematic in such an enterprise. But in Hume’s hands it reveals that human beings immediately perceive at best only something that falls far short of the world of enduring objects, causal connections, thinking and experiencing persons, and good and bad states of affairs that all human beings believe in. Given those “sceptical” conclusions, explaining how human beings come to have any such beliefs seems to require explaining how what is so in the austere independent world combines with what is true of human beings to produce the richer conception of a world that they all accept. The essay describes how acceptance of those “sceptical” con-clusions restricts what a Humean “naturalist” can appeal to in explaining those ways of thinking in a way that leaves the would-be naturalism forever dissatisfying.

“Practical Reasoning” was fi rst published in Edna Ullmann-Margalit (ed.), Reasoning Practically (Oxford University Press, New York, 2000). It takes a few elementary steps into the huge question of how thinking and deliberating about what to do issues in action. The essay starts from Hume’s claims that “reason alone can never be a motive to any action of the will” and that “it can never oppose pas-sion in the direction of the will.” These claims depend on Hume’s view of the passions, in particular desires, and they result in an essen-tially fi rst-person conception of reasons for action. That idea is present also in Bernard Williams’ defence of exclusively “internal” reasons for action. I argue that questions of what to do or what should be done are not essentially fi rst-personal in any such restricted “internal” sense, even though what is actually done as a result of the deliberation is always up to only one person.

“The Charm of Naturalism” was my Presidential Address to the Pacifi c Division of the American Philosophical Association, pub-lished in Proceedings and Addresses of the American Philosophical Association in 1996. It takes up the question of the nature of natural-ism more broadly, moving beyond Hume to the prospects of philo-sophical illumination from any distinctive enterprise worthy of that name. If naturalism as a method of investigation can appeal only to what is so in “the natural world,” there is a question what that conception of the natural world excludes, and why. If it excludes

Introduction 7

nothing, or only whatever there is no good reason to accept, natural-ism would be indistinguishable from responsible enquiry in general. If what it excludes is thereby rendered unavailable for explaining the ways of thinking and feeling that are to be given a naturalistic explanation, there is a question how irreducible beliefs that are not expressible in fully “naturalistic” terms are to be explained. They apparently include not only beliefs about the colours of things and about enduring objects and causation, but also evaluative beliefs, and even beliefs in the truths of logic and mathematics.

“The Transparency of ‘Naturalism’ ” returns to the same theme, this time in the invited Romanell Lecture on Philosophical Naturalism published in the Proceedings and Addresses of the American Philosophical Association in 2008. This too raises the problem of naturalism’s distinctive character. Because the term “naturalism” takes its meaning in each case from what is conceived as falling within the limits of the “natural” world, it is best under-stood by looking right through the term itself to whatever it is being used to stand for in this or that application. The essay rehearses some of the dissatisfactions of Humean naturalism and explains how parallel diffi culties would undermine any apparently tough-minded naturalism that tries to exclude even intentional attitudes and meaning from the fully “natural” world. That would leave us with no psychological facts, and so not even a conception of the world, to be explained naturalistically or otherwise. In disputes about naturalism, what is in question seems always to be something else, not the merits of naturalism itself.

“Anti-Invidualism and Scepticism” was published in Martin Hahn and Bjorn Ramberg (eds), Refl ections and Replies: Essays on the Philosophy of Tyler Burge (MIT Press, Cambridge, Mass. 2003). I fi rst presented it at a conference on Burge’s work ten years earlier. At that time Burge understood his “anti-individualism” to imply that “error presupposes a background of veridicality.” The essay explores the ways that powerful idea is to be developed into a guarantee against philosophical scepticism about the world. Burge seeks not only what he calls a “general” or “transcendental” guarantee, but something that would justify perceptual knowledge claims in the face of scepticism even in particular cases. He thinks that is achieva-ble even though, in any given case, all of a person’s perceptual capaci-ties could be mistaken about how things are. My question is about that goal, and how it is to be achieved. Is the goal to show that that

8 Introduction

general possibility of error is no threat to a person’s knowing some-thing about the world by perception in a particular case? Or is it rather to show not just that there is no general threat but that philo-sophical scepticism is actually false, since the person in the case con-sidered does know by perception that things are thus and so? And is this stronger conclusion to be reached by philosophical argument?

“Sense-Experience and the Grounding of Thought” was pub-lished in Nicholas H. Smith (ed.), Reading McDowell: On Mind and World (Routledge, London, 2002). What is in question is the proper understanding of perception, or what McDowell calls “experience,” in our thinking and knowing what we do about the world. I am very sympathetic to McDowell’s project, and to his way of carrying it out. I agree that it must be possible to understand ourselves as per-ceiving that things are thus and so, where that says how things are in the world independently of us and our perceptions. The question is how perception is to be understood to secure that result. McDowell’s invocation of what he calls “impressions” seems to me to threaten that possibility. I question whether his legitimate con-cern for the “spontaneity” involved in perception and perceptual knowledge requires anything more than our having the capacity to perceive and thereby to know that p when (but not whenever) it is the case that p . I do not fully understand why, or even whether, McDowell thinks, or thought, that a purely “experiential,” non-judgemental component of perceptual knowledge can somehow be sliced off from the rest.

“The ‘Unity of Cognition’ and the Explanation of Mathematical Knowledge” was published in Philosophical Topics in 2001. I pre-sented it at a conference on the work of Alvin Goldman in that same year. The essay describes some of the main ingredients in the devel-opment of Goldman’s important work in epistemology, focussing especially on the requirement that the reliable procedures essential for gaining knowledge must include facts about the psychology of the believer; relations simply among the propositions believed are never alone enough. “Unity of cognition” is the idea that this require-ment holds for explaining knowledge of all kinds: of mathematical and other necessary truths as well as of contingent facts. I see this conception of reliable procedures as promising for explaining math-ematical knowledge without familiar but ultimately fruitless appeals to “concepts,” “meanings,” or “analyticity.” But Goldman is appar-ently committed to the eventual success of what he calls a “primary

Introduction 9

epistemology” in which what it is for someone to have followed reli-able procedures, and to have done so reliably, will be described in purely “non-epistemic” or “non-normative” terms. I express doubts about that enterprise while encouraging the explanation of mathematical knowledge along the lines Goldman has so ably defended before taking that last reductionist step.

“Contemporary Pyrrhonism” was published in Walter Sinnott-Armstrong (ed.), Pyrrhonian Skepticism (Oxford University Press, New York, 2004). I presented the essay fi rst at a conference on the work of Robert Fogelin at which he was present. Since he is a good-natured friend of mine I tried to encourage discussion by adopting a somewhat incredulous, mock-challenging tone that is perhaps not strictly appropriate for sober philosophical interchange. I hope what remains of it here does not obscure the great extent to which I agree with Fogelin’s account of knowledge and sympathize with his approach to scepticism. I fi nd especially promising his idea that in making a knowledge-claim or in assessing the knowledge-claims of others one takes a non-relativized stand oneself on the adequacy of the reasons for the belief; it involves an evaluative claim of one’s own. I raise the question whether that feature of knowledge might alone put one in a better position, as a knowledge-claimer, to resist the undermining force of the possibilities that in philosophy seem to lead inevitably to scepticism.

“Perceptual Knowledge and Epistemological Satisfaction” appeared in J. Greco (ed.), Ernest Sosa and his Critics (Blackwell, Oxford, 2004). This essay briefl y examines Ernest Sosa’s “externalist” answer to the traditional epistemological question of how we know things about the world around us. I agree with Sosa’s rejection of “internalist” theories for the reasons he gives, but I express doubts about his own positive account of knowledge in terms of its being no accident or coincidence that one’s belief is true. The diffi culty I see does not lie in any circularity in the account but in Sosa’s apparent concession that the most we can strictly speaking know by percep-tion alone is “the character of our experience” and not the way things actually are. Sosa regards that concession as in itself unthreatening if it is not combined with the further assumption made by “internal-ists” that knowledge of the world is arrived at by reasoning from some prior knowledge, and eventually from what we get from perception alone. For Sosa it is enough for knowledge if it is no acci-dent or coincidence that things are the way one takes them to be

10 Introduction

when “the character of one’s experience” is so and so. I try to explain how and why a restricted account of perception that denies all purely perceptual contact with the way things are would leave us always in an unsatisfactory position for understanding our knowledge of the world and explaining it even to ourselves.

1

Our Debt to Descartes

The philosophical, mathematical, and scientifi c legacy of Descartes is by now so deep and so pervasive in our culture that its full extent can no longer be measured with certainty or precision. I make no attempt to do so here. Nor do I try to trace the historical stages of acceptance, absorption, misunderstanding, and transformation by which it has all come down to us. But I want to draw attention to several large ideas or lines of thought that are unmistakably Cartesian in character and probably in origin which seem to me of great inter-est and importance. They have remained conspicuously at or near the centre of philosophy for at least the last one hundred years or so. Their infl uence was perhaps stronger during that period than in the one hundred or so years before that.

Descartes’ philosophy is distinctive in one way in starting from the idea of a secure method of “rightly conducting one’s reason and seeking the truth in the sciences”. 1 That was a perfectly reasonable goal, since Descartes was in fact seeking the truth in the sciences – or anywhere else it could be found – and he wanted to be sure when he had found it and when he had not. He recognized no sharp distinc-tion between philosophy and other ways of seeking the truth about the world; successful enquiry in any of its forms was his goal. But he thought validation or legitimation of one’s apparent results was needed for real progress. Whatever was acquired by a demonstrably secure method would be guaranteed to be true and so could be relied on in future work.

Another distinctive feature of Descartes’ philosophy is his particu-lar way of pursuing this question of method. He found that it was

1 Discourse on the Method of Rightly Conducting One’s Reason and Seeking the Truth in the Sciences , John Cottingham, Robert Stoothoff, Dugald Murdoch (eds), The Philosophical Writings of Descartes , Cambridge Universsity Press, Cambridge, 1985, 2 vols., p. 111 . (Hereafter CSM 1 or 2).

12 Philosophers Past and Present

not enough to study “the great book of the world” as he had done, and acquire a complicated body of information as to what is so. All those beliefs and convictions would themselves stand in need of explicit justifi cation as to their origin or legitimacy. To this end he resolved to undertake studies within himself, “to converse with myself about my own thoughts”, 2 and “to reform my own thoughts and construct them upon a foundation which is all my own”. 3 This was to be the source and basis of all the knowledge Descartes could eventually call his own. “I found myself as it were forced to become my own guide,” 4 he tells us.

The refl ections he recounts are not meant to be only of biographical interest. Descartes presents himself (or somebody) not simply as a certain Frenchman who has certain thoughts about himself at a cer-tain time, but, in Bernard Williams’ words, “as an example – though a genuinely existing, particular example – of the mind being ration-ally directed to the systematic discovery of truth”. 5 “The Meditations are not a description but an enactment of philosophical thought, fol-lowing . . . the fundamental route by which human thought should move from everyday experience to greater philosophical insight.” 6 The “guide” Descartes was forced to follow was to be something each of us can discover and follow for ourselves.

A large philosophical project with these two features – the goal of validating or legitimating what we take to be our knowledge of the world, and each person’s securing the validation for himself, in the fi rst-person singular – was at the centre of philosophy for most of the twentieth century, at least in English-speaking lands. By then philosophy, following Kant, had come to understand itself as some-thing distinct from, but somehow still related to, straightforward investigation of “the great book of the world”. It was to pursue a more detached and more critical and so uniquely philosophical task. The project of justifying or legitimating what we take to be our knowledge, or at least exhibiting a possible justifi cation of it, was the goal of what came to be known as “epistemology”. As the project was conceived in its purest, classical form, it was a question of

2 Discourse (CSM 1 p. 116). 3 Discourse (CSM 1 p. 118). 4 Discourse (CSM 1 p. 119). 5 B. Williams, Descartes: The Project of Pure Enquiry , Penguin, Harmondsworth

1978, p. 19 . 6 B. Williams p. 20.

Our Debt to Descartes 13

showing how each person, proceeding on his own from what is avail-able to him in sense experience, can be justifi ed in believing every-thing we all take to be part of human knowledge of the world. In the casual, offhand way philosophers seem to have with nomenclature, it even came to be called “Cartesian epistemology”.

There are ideas at the heart of this project that are ideas to be found in Descartes. To that extent they are part of Descartes’ legacy. That is not to say that the use to which those ideas have been put, and the consequences to which they were taken to lead, follow from or can reasonably be derived from Descartes’ own understanding of them.

In his attempt to “reform” and “construct” his thoughts on a secure “foundation” Descartes in his Meditations resolved to with-hold his assent from anything in which he fi nds “at least some reason for doubt” and which therefore is not “completely certain and indubitable”. 7 The fi rst step in his application of this method was the denigration or apparent dismissal of “the senses” as a source of knowledge of the physical world. There was nothing new in itself in drawing attention to familiar perceptual illusions or delusions. But Descartes’ imaginative application of his general method of doubt to all knowledge gained through the senses led to something completely new. He exposed what came to be seen as a problem or challenge to human knowledge that had never even been recognized, let alone squarely faced or answered, in philosophy up to that time.

This is what is called “the problem of our knowledge of the exter-nal world”, or simply “the problem of the external world”. And its emergence – even the meanings of the very terms in which it is expressed – can certainly be attributed to the novel kind of refl ection Descartes engaged in in his fi rst Meditation . His way of introducing the possibility of dreaming as a reason to doubt the deliverances of sense-experience on any particular occasion opened the door to an apparently much more devastating and completely general possibil-ity that would seem to threaten the prospect of any perceptual knowledge of the world around us at all. 8

7 Meditations I (CSM 2 p. 12). 8 For a highly illuminating account of what is new in Descartes and completely

unanticipated in antiquity, and what thereby gives rise to the modern problem of the external world, see M. F. Burnyeat, “Idealism and Greek Philosophy: What Descartes Saw and Berkeley Missed”, Philosophical Review 1982 .


If, as Descartes found, “there are never any sure signs by which” a particular sense-experience can be recognized as a perception of how things actually are rather than part of a dream, 9 there is noth-ing to be found in any particular experience itself that enables the perceiver to distinguish the two. Descartes’ method of withholding judgement on any matter “which admits of the slightest doubt” 10 would therefore require at least provisionally withholding judge-ment from the proposition that things actually are as one’s particu-lar sense-experience presents them as being. Even though one has the experience, one could still be wrong in believing that that is how things are. That same method applied with complete generality demands withholding the corresponding judgement with respect to every sense-experience. This is the general possibility of error that Descartes illustrates with the fi ction of the evil demon. And it is the acknowledgement of this general possibility that gives rise to the problem of the external world.

A powerful demon whose sole aim is to deceive everyone on every matter on which they could possibly be fooled could produce sense-experiences that would seem to indicate to perceivers how things are beyond them but would lead them astray if accepted as having that signifi cance. Any step beyond what was strictly speaking perceived would be false. This does not mean that such perceivers could know nothing. What would remain immune to perceptual error under those circumstances would be the distinctive character of the person’s sense-experience of the moment. There would have to be something of that kind that a suffi ciently careful perceiver could not be wrong about, since what he receives in perception is what serves as the stimulus or basis of any mistaken beliefs he falls into. To operate through sense-perception, a deceiver must actually give his victims something in perception that can then lead them astray. To fi nd that kernel in one’s experience, and to withhold judgement about every-thing that goes beyond it, would be to restrict oneself to the deliver-ances of sense-experience alone. That would leave one safe from the scheming of even the most powerful would-be deceiver.

This conception of what perception alone at its best can give us fi nds later expression in Berkeley’s idea that “the senses perceive nothing which they do not perceive immediately: for they make no

9 Meditations I (CSM 2 p. 13). 10 Meditations II (CSM 2 p. 16).


inferences”. 11 Hume for the same reason thought it was not “conceiv-able” that the impressions of our senses, considered on their own, could deceive us. “For since all actions and sensations of the mind are known to us by consciousness, they must necessarily appear in every particular what they are, and be what they appear.” To suppose otherwise would be “to suppose, that even where we are most intimately conscious, we might be mistaken”. 12 The impossibility of purely perceptual error is also at work in H. H. Price’s well-known confrontation with a tomato.

When I see a tomato there is much that I can doubt. I can doubt whether it is a tomato that I am seeing, and not a cleverly painted piece of wax. I can doubt whether there is any material thing there at all. . . . One thing however I cannot doubt: that there exists a red patch of a round and somewhat bulgy shape, standing out from a background of other colour-patches, and having a certain visual depth, and that this whole fi eld of colour is directly present to my consciousness. . . . [By this] I mean that my consciousness of it is not reached by inference, nor by any other intellectual process . . ., nor by any passage from sign to signifi cate. 13

C. I. Lewis thought such a directly presented element can be identi-fi ed in any current sense-experience by applying the following test: “Subtract, in what we say that we see, or hear, or otherwise learn from direct experience, all that conceivably could be mistaken ; the remainder is the given content of the experience inducing this belief.” 14 The implication in each case is that what one strictly speaking perceives in a given experience is only that about which one could not conceivably be wrong, given that experience. In invoking this idea these later philosophers all remain fully in line with Descartes’ own application of his general method of doubt to the senses.

This idea of the restricted range of what the senses alone at best can give us, when conjoined with the assumption that the world can be known at all only somehow on the basis of what we receive through the senses, is the source of the special epistemological problem of the

11 G. Berkeley, Three Dialogues Between Hylas and Philonous , in Berkeley: Philosophical Works (ed. M. R. Ayers), Dent, London 1975, p. 138 .

12 D. Hume, A Treatise of Human Nature (ed. Selby-Bigge), Oxford University Press, Oxford 1958, I, 4, ii (p. 190) .

13 H. H. Price, Perception , Methuen, London 1932, p. 3 . 14 C. I. Lewis, Analysis of Knowledge and Valuation , Open Court, LaSalle, Illinois

1946, pp. 182–183 .


external world. It is a completely general problem, and it brings the whole “external” world into doubt. If the evil demon possibility were realized, all links between what is perceived and what is otherwise so in the world would have been broken. To know on the basis of sense-experience that that possibility is not in fact realized, and that one is perceiving things as they actually are, would require having some sense-experiences. Those sense-experiences could help settle that second-level question only if they themselves indicated that the evil demon possibility is not in fact realized. But if one’s perceiving what-ever one perceives is compatible with that possibility’s being realized (as this view of perception implies), then no particular perception could be known or reasonably believed to indicate that the possibility is not realized. Just as there are no sure signs by which an experience can be recognized as a waking experience and not a dream, so there will be no sure signs by which an experience can be recognized as not produced by an evil demon but produced by things’ being as the per-ception represents them as being.

With possible doubts generalized to all sense-experience in this way, even the existence of one’s own body is cast into doubt, insofar as that body is thought to be known through the senses. A perception of what I take to be my own hand before me, or of what I take to be a rumbling in my stomach, is just as vulnerable to an evil demon’s deceptive machinations as perceptions of a mountain or of a piece of paper. So what become epistemically problematic as a result of these possible doubts are not only bodies spatially distinct from me. Even my own body is cast into what in this special (non-spatial) sense is now to be called the “external” world. That is a world “external” to, or “beyond”, each perceiver, in the sense that it is something that no one can perceive or know about by sense perception alone. Perceivers are never presented in perception with anything they thereby perceive to be true of an “external” world of bodies, including their own bodies.

Myles Burnyeat points out that this thought makes room for the question whether there is any body at all – anything other than mind or thought. He argues that this question is something completely new in the history of philosophy; it arises for the fi rst time out of Descartes’ refl ections on sense perception. An “external” world understood in this way is a world distinct from everything that is true of the “subjec-tive experiences” of perceivers. So by putting “subjective knowledge at the center of epistemology” Descartes “thereby made idealism a


possible position for a modern philosopher to take”. 15 No such pos-sibility was even contemplated in the philosophy of antiquity. Nor was Descartes himself drawn in the direction of idealism by these refl ections. But his attack on the senses as a source of knowledge of the world eventually made idealism a live philosophical option.

Much of the history of epistemology in the twentieth century was an attempt to show in one way or another how we can and do in fact have knowledge of the world around us. This problem took on its especially obstinate character only because of the view of sense per-ception that arises from Descartes’ refl ections. If that view is not accepted, and it is granted that we are capable of perceiving the way things are in the world around us, it would be easy to explain our knowledge of the world. We could know how things are by seeing or otherwise perceiving them to be a certain way. The challenging epis-temological problem of the external world was not to be answered so straightforwardly. It was understood as a question about how we get knowledge of the world around us on the basis of sense-experience even though no particular sense-experience can reach as far as any fact of the world we claim knowledge of.

This way of putting it can make it sound as if there can be no satis-factory solution to the problem at all. I think this way of putting it accurately describes the problem as it was understood by those impres-sive philosophers of the twentieth century who took it most seriously. And I think there can be no satisfactory solution to the problem so understood. With only the restricted resources this view of perception allows to be available, I think it cannot be shown how anyone could know or have reason to believe anything about an “external” world on the basis of sense-experience, or indeed anything beyond the charac-ter of the perceiver’s sense-experience of the moment.

It is sometimes suggested that the trouble lies only in the appar-ently arbitrary “Cartesian” requirement that the experiences adequate for a satisfactory solution must be restricted to the perceptions of each single, individual perceiver. With the shared experiences of mankind to rely on there is thought to be no diffi culty in explaining how we get reason to believe the things we do about the world. But these Cartesian refl ections about perception do not begin from an arbitrary insistence on an individualistic solution to the problem of knowledge. The essentially fi rst-person character of the project is

15 Burnyeat, op. cit. , p. 33.


itself a by-product of this very view of perception and the concep-tion of the “external” world it leads to.

It is not an arbitrary or unjustifi ed assumption to say that each perceiver perceives only what he or she perceives. It is a simple truth. The question is what a perceiver can properly speaking be said to perceive. This view says that everyone perceives only that about which he could not conceivably be mistaken given the perception in question, and that therefore no one can ever perceive that any other perceiver, or indeed any enduring object at all, including his own body, exists. The threat of solipsism is a consequence, not a presup-position, of this view of perception. So the thought of there being other perceivers whose perceptions cohere with and so help support one’s own beliefs is a thought about a world “external” to whatever one perceives. It is therefore not available as contributing to anyone’s perceptual justifi cation of his beliefs. It is part of the problem, not the solution.

The refl ections that seem to lead to this understanding of percep-tion, and so to the general epistemological problem, can be found above all in Descartes. But that is not to say that Descartes himself was in this apparently hopeless plight, or that he thought anyone else was in it. He was concerned at the fi rst step of his method to expose the pretensions of the senses as a source, or perhaps the only source, of knowledge of the world. Having done so, he looked elsewhere for something invulnerable to all possible doubt, and so “completely cer-tain and indubitable”. His strategy was to try then to extend the scope of whatever certainty he could fi nd at fi rst by further refl ection at each point on what must be so if he has been able to reach the certainty he has achieved so far.

The starting-point and foundation of the whole project is the dis-covery of something Descartes saw must be true even if he was deceived as much as a powerful demon could possibly deceive him. No such demon could bring it about that he is nothing while he thinks he is something. To be deceived is to think something that is not so, but he could not be deceived in thinking that he exists. No one who thinks could think falsely that he exists. Descartes sees that “this proposition, I am, I exist , is necessarily true whenever it is put for-ward by me or conceived in my mind”. 16 This is certainly among the most important and longest-lasting ingredients of Descartes’ legacy,

16 Meditations II (CSM 2 p. 17).


and it has borne rich and, in combination with other Cartesian ideas, sometimes apparently unpalatable, fruit. The full signifi cance of the fundamental insight can be seen to extend in several directions, not all of which even now have been exhaustively explored.

There is fi rst a question of the precise character and distinctive status of propositions like ‘I think’ and ‘I exist’. There is a further question of how large and varied the class of propositions with that distinctive character can be seen to be. And there is a question about the source and nature of the certainty that discovering truths with that special status can provide, and how far the range of that cer-tainty can be found to extend. These questions are obviously all con-nected. They are central to Descartes’ philosophical enterprise and, as things turned out, to modern philosophy at large.

A thinker obviously could never be wrong in thinking ‘I think’. That is something that must be true if he thinks anything at all, even if he thinks ‘I am not thinking’. The same is true of ‘I exist’; it cannot possibly be truly denied. Descartes fi nds it “self-evident” that it is “impossible” for someone to think without existing. 17 He “see(s) very clearly that in order to think it is necessary to exist”. 18 So the truth of a thinker’s thought ‘I exist’ is a necessary condition of the thinker’s thinking anything. The same holds for all other necessary conditions of that thinker’s thinking. Everything that must be true if a thinker thinks is something that could never be false if thought by that thinker.

Many propositions are not truly deniable because they are neces-sarily true. They could not be false under any circumstances, so they could not possibly be false if thought by any thinker. Their truth is in that sense a necessary condition of any thinking. But when Descartes says that ‘I exist’ “is necessarily true whenever it is put forward by me or conceived in my mind” he does not mean that ‘I exist’ is a truth of that kind, or that his thinking guarantees that it is necessarily true in that way. He means only that, necessarily, if he thinks then he exists.

He is in fact René Descartes, who existed from 1596 to 1650. So what he said of himself – and could never possibly truly deny – during that period is something that is no longer true of him. If certain things had been different in certain possible ways, what is true of him now (that he does not exist) would even have been true at a time – say, 1638 – when he

17 Replies to Second Set of Objections (CSM 2 p. 100). 18 Discourse (CSM 1 p. 127).


actually (and so truly) thought ‘I exist’. He could have failed to exist in 1638. His existing is therefore a necessary condition of his thinking without itself being something that holds necessarily. It is, in an appro-priately absolute sense, something contingent. But it is a necessary con-dition of his thinking, which itself is also something contingent. That is the distinctive character of propositions such as ‘I think’ and ‘I exist’ that Descartes fi rst draws attention to.

Something’s being a necessary condition of a person’s thinking is not in itself enough to give that person knowledge or certainty of its truth. Although the thinker could not possibly truly deny anything that stands in that relation to his thinking, there might be many things which he does not know or even suspect have that status. If he is unaware that a given proposition falls into that special category for him, he will not recognize that it is a truth he could rely on without any possibility of error in all his further refl ections. For a proposition to take on that kind of role for a thinker he must recognize, with certainty, that its truth does stand in that necessary relation to his thinking.

Descartes holds that in his original certainty of his own existence as a thinking thing he fi nds what is required for his being certain about anything. 19 This mark of certainty is then to be invoked at each step in attempts to extend his certainty further. At the fi rst step he recognizes that his “clear and distinct perception” of his own exist-ence is a perception in which he could not possibly be wrong. What he “clearly and distinctly” perceives to be so in that case could not turn out to be mistaken. But that is because no one could fail to exist if he thinks or perceives anything at all. The further refl ections Descartes engages in at later stages concern not simply the condi-tions of his thinking anything at all, but the conditions of his think-ing in certain specifi c ways or having certain determinate thoughts. His goal is to arrive at further conclusions about how things are or must be from the fact that he has the specifi c thoughts he knows he has.

These further inferences are accordingly more problematic, and have understandably been the focus of most of Descartes’ critics from his own day to ours. There is a recurrent question whether and how the rich conclusions he claims to arrive at can be reached from such beginnings alone. And are the principles invoked in drawing

19 Meditations (CSM 2 p. 24).


those conclusions known with the same kind and degree of certainty – and is the same mark or criterion of certainty invoked – as that with which a person knows that he could not possibly be wrong in thinking that he exists?

For instance, Descartes examines the contents of his own mind and fi nds there an idea of an eternal, infi nite, omniscient, omnipo-tent creator. His refl ections on how he could possibly have such thoughts lead him from the fact of his thinking them to the eventual conclusion that such a being exists with all those attributes. This being is even said to be the guarantor of the truth of everything that Descartes comes to perceive clearly and distinctly to be true. No such outside help was needed to guarantee the truth of his original thought ‘I exist’. That was seen to be true simply because it could not be false if he thought it; and seeing that that is so is all he needed to be certain of his existence.

Having found that he cannot doubt that he himself exists and is something that thinks, Descartes also tries to determine by refl ection exactly what he is. He cannot doubt his own existence, but his attack on the senses shows that he can doubt the existence of all bodies. From his having these two thoughts he appears to conclude that a certain thing is possible: he could exist without a body. The earlier refl ections on God appear to play a role here. So he is something essentially non-bodily. The resulting “dualism” of a non-thinking body distinct from a non-bodily “mind” has done much to give the legacy of Descartes a bad name.

It is not surprising that so much critical attention has been directed to almost every step of these complex lines of thought. If we now have any greater understanding and appreciation of the precise source of the diffi culties they raise, we remain at least in that indirect way indebted to Descartes. But the diffi culty of extending the distinctive status and undeniable certainty of ‘I think’ and ‘I exist’ to other and more substantive propositions does not detract from the fundamental importance of Descartes’ recognition of that distinctive status and of what it promises. Nor is diffi culty alone a deterrent to trying to extend it further.

The revolutionary philosophical signifi cance of exploring the necessary conditions of thinking is the idea so richly exploited by Kant. He was interested not simply in his own thoughts or the thoughts of an individual thinker, but in any thinking at all by anyone. And since for him even experience of a world is impossible without thought, he


focussed on the necessary conditions of any possible thought or expe-rience. He saw clearly that all such conditions must have a very special standing in our thought or knowledge of the world. No one could pos-sibly deny any of the propositions that have that standing and be right. Given that there is thought, such things would simply have to be true. For Kant this was the key to philosophical progress. Metaphysics could be a legitimate intellectual enterprise with secure results only if it restricts itself to the investigation of the general conditions of anyone’s thinking or experiencing anything at all.

For Kant, as for Descartes with ‘I think’ and ‘I exist’, the interest was not primarily in those propositions that must be true if anyone thinks because they are in an absolute sense necessary and so must be true whatever else is so. Kant regarded propositions of that kind, which cannot be denied without contradiction, as ‘analytic’; the concept of the predicate of the judgement is already ‘contained’ in the concept of the subject. But Kant sought non-‘analytic’ or ‘synthetic’ propositions that express necessary conditions of thought and experience in general. That is because he sought signifi cant metaphysical results about the way the world is or must be. ‘Analytic’ truths that reveal only which concepts are ‘contained’ in which other concepts would give us only conditional truths about the way things must be if they are a certain other way. Kant’s conception of metaphysics was to yield substantive, non-conditional truths about the world.

Kant explored the necessary conditions of thinking certain specifi c kinds of thoughts, or thinking in certain determinate ways, as Descartes had also done. But Kant did not simply fi nd certain thoughts in his own mind and look for something that must be so in the world if he has them. Having recognized the distinctive undeniability of all necessary conditions of thinking or experience in general, he thought he could prove that there are certain ways of thinking, or certain determinate thoughts, that any thinking being must have or be capa-ble of. Whatever then could be found to be a condition of someone’s thinking in any of the specifi c ways he identifi ed would also be a necessary condition of thought and experience in general. It would therefore have the same distinctively undeniable status in our thought as Descartes had discovered for ‘I think’ and ‘I exist’.

For Kant, as for Descartes, the signifi cance of there being propositions with that distinctive status was clear; the problem for each of them was to show that the conclusions they most wanted to establish about the world do indeed fall into that privileged class of not-truly-deniable


propositions. Kant’s strategy was to show that having those concepts or a capacity for making those kinds of judgements that he claimed are necessary for any thought or experience at all necessarily involves having a capacity to think in certain other ways, or to deploy certain other concepts, and so on to further and further conditions. It was to be a demonstration of necessary connections between the possession of certain capacities or concepts and the possession of certain others. It was not simply a question of the necessity with which one concept or set of concepts must be true of something if a certain other con-cept or set of concepts is true of it. That would yield only ‘analytic’ truths, and so only conditional necessities. What was needed to yield substantive metaphysical results were the necessary conditions of thinking, and so of thinking certain determinate thoughts about the world, not simply necessary connections between the contents of those thoughts.

Even if Kantian necessities between different ways of thinking could be fi rmly established by this procedure they would seem to fall short of implying that the judgements shown to be required for the ways of thinking in question are actually true. It is one thing to dis-cover that it is not possible to think at all or to think in certain ways without having certain other determinate thoughts or capacities. It looks like a different and stronger claim to hold (as with ‘I think’ and ‘I exist’) that it is not possible to think such thoughts without their being true. Some explanation would seem to be needed of how the actual truth of the thoughts can be inferred from the fact that they are, or even must be, thought.

Kant in effect denied this apparent gap with his doctrine of tran-scendental idealism. According to that view, what we have thoughts and experiences about in fulfi lling the necessary conditions of think-ing and experiencing anything at all is the world in which what we think and experience to be so must in general be true. There can be no completely global gap between how we think and experience things to be and how things are.

This would mean that the apparently hopeless plight that Descartes’ attack on the senses seemed to imply for the human perceiver simply could not arise. A single individual might in very bizarre circum-stances fall out of virtually all sensory contact with the world for a while, but not for long, and not for a perceiver’s whole experiential life. Nor could the human race in general be in such an unfortunate position. So if Kant is right, human beings could not be faced with


what came to be seen as the completely general problem of our knowledge of the external world. Having the thoughts and experi-ences necessary even to seem to face that problem would be enough in itself to guarantee that the world to which one seeks epistemic access is, and in fact must be, experientially available.

This was Kant’s conclusion from his refl ections on the necessary conditions of any thought or experience. Descartes’ own refl ections on the conditions of his having the thoughts he fi nds within himself even after his attack on the senses and its accompanying doubts did not lead him to resolve the apparent epistemic quandary in that way. He was not tempted by, or probably even aware of, the allure of idealism. His hopes for knowledge of the world were to rest ulti-mately on the good will of a benefi cent creator. The Kantian idealist solution invokes no supernatural being, but it is possible to feel that it comes at an equally unacceptable price. Both theories involve dis-covering by refl ection that something that is true of the way things are in the world is a necessary condition of thinkers’ thinking or experiencing the things they do. That would yield indubitability equal to that of ‘I think’ or ‘I exist’ if the necessities in question could be demonstrated with that same kind of certainty. A gap appears to arise at what would be the fi nal step in any demonstration of either of these different positive doctrines about the world. That apparent gap is the site of some of the most intense and most productive phil-osophical refl ection of the last three hundred and fi fty years.

The Kantian strategy, for all its idealist extravagance, draws atten-tion to something that Descartes tended to overlook: the conditions of having the thoughts, beliefs, and perceptions essential to possess-ing any conception of a world at all. That can encourage the more specifi c question of what must be true of thinkers and perceivers for them even to be vulnerable to the kind of wholesale falsity in their conception of the world that the deceiving demon is imagined to pro-duce. Descartes does not take up this question directly. His refl ec-tions on the possible doubts generated by the thought of a deceiving demon lead him to regard the senses as providing only sensations or “images” or “ideas” of “things appearing” to be certain ways. He does not ask what makes it possible for perceivers even to have per-ceptions with such contents. But he holds that any states of affairs in the world that might be responsible for the presence of such percep-tual representations are beyond the reach of perception alone. This is


the view that has dominated the understanding of perception since Descartes’ day and has generated the problem of the external world.

As Descartes imagines the challenge to the senses, if everything in the world was under the control of a deceiving demon, then anyone who had thoughts and perceptions of just the kinds we all now have in everyday life would have almost entirely false beliefs and, beyond the minimal core of perceptual experience produced by the demon, entirely non-veridical perceptions. It is compatible with that possibil-ity as I myself think of it, for instance, that all that exists in the whole universe is just that deceiving demon and me as I am now. That would certainly render false almost all my current beliefs and perceptions.

But although that would leave me almost completely deceived, is it also true that if all that existed in the universe was a would-be deceiv-ing demon and me, then I would have or could be given just the beliefs and perceptions that I have as things are now? That is harder to accept, whatever we try to imagine the demon doing. Of course if we imagine, as I suppose we must, that the demon is a supernatural being, he might seem capable of doing anything that can be consist-ently described. So he might be said to convince me that there are mountains and pieces of paper and such things, or to persuade me to believe that I see a red tomato on a table, or somehow simply to give me such beliefs or perceptions, even though everything I think and believe under those circumstances is false. But attending to the con-ditions of thought and perception raises a question about whether and how he could actually succeed in doing that.

It is not enough simply to stipulate that the demon does it. To be convinced or persuaded of something I must at least understand what I thereby come to accept. And for that I would need the requisite concepts or capacities for judgement of certain kinds. And do we know that someone in a world radically different from the world we believe in and take ourselves to perceive could even have the concepts and capacities involved in understanding what we all now understand and believe to be so in the world as we take it to be? Again, it might be stipulated that the deceiving demon simply gives his victims the requisite concepts or capacities, by magic, as it were. But that is not to explain how he does it, or to show that it is even possible. Can we really make sense of how any agent in the world is able to do such things, or what he could give those thinkers and perceivers that would yield that kind of understanding?


Behind this question lies an issue at the forefront of much recent philosophical discussion. In its most general form it is the question whether or in what ways certain things must be so in the world that thinkers think about and perceive in order for those thinkers even to have the particular thoughts and perceptions they have. Descartes had already found one instance of a positive answer to this general question with his ‘I think’ and ‘I exist’. He knew that such things had to be true in the world if he even so much as has the thought that they are true. Their falsity would have meant that he did not have those thoughts.

Descartes had at his disposal one way of explaining how the exist-ence or identity of a thought can depend on what is so in the world that is thought about. He asked himself what he is, while his attack on the senses was still in effect, and he saw that he could not doubt that he exists although he could doubt that any bodies exist. He saw that it is not by his knowing or believing something about a particu-lar body that he is certain that he exists. Nor could he even be certain at that point that there are such things as spiritual beings or Christian souls in the world, so he is not certain of his own existence by know-ing that he is one of those things either. At that stage of his refl ec-tions he realized that “none of the things that the imagination enables me to grasp is at all relevant to this knowledge of myself that I possess”. 20

This means that he could withhold judgement about all such worldly matters and still have no doubt that he exists. Even if he freed himself from all beliefs about what is so in the world or what or who he is, he would still have the thought that he exists, and it would be true. His having that thought therefore does not require that he know or believe or think of himself in any particular way. What thought Descartes thinks when he thinks ‘I exist’ depends on the identity of the thinker who has that thought. If some other person were to think ‘I exist’ that would be a different thought; what would be so in the world if that person’s thought were true is not the same as what is so if Descartes’ thought ‘I think’ is true. One of those thoughts could exist without the other. The difference between them is a difference in what is so when they are thought.

In this way Descartes’ application of his method of doubt can be seen to contain the seeds of the idea that his ‘I’ refers unfailingly to

20 Meditations II (CSM 2 p. 19).


him quite independently of how he might happen to think of himself. He does not need to think of himself in some particular way, or have certain beliefs about who or what he is, in order for the reference of his ‘I’ to get fi xed on to the right thing. It is true that he recognizes with no possibility of doubt that he is also a thing that thinks. That is something he cannot avoid believing of himself, but his having that belief is not what guarantees for him the secure reference of his ‘I’. For one thing, being a thing that thinks does not suffi ce to secure unique reference. And being the (unique) thing that thinks is not something he can know to be true of himself in his current state of doubt. So it cannot be used to secure unique reference either. He nonetheless succeeds in referring to himself, and himself alone, in his thought ‘I exist’.

Other forms of direct indexical reference are present in thoughts of other kinds whose existence and identity depend in similar ways on what is so in the world the thinker in question thinks about. This can be seen to guarantee the distinctive undeniability of certain other parts of what any thinker needs to think in order to have a function-ing conception of a world in which he exists. Any agent needs a notion of ‘here’ and ‘now’, for instance, to place his thoughts about himself and his possible actions at a particular place and time. And what thoughts he expresses in using those words depends not on what he believes about the world at the place and time he has the thoughts, but on where and when he actually is in the surrounding world that contains him.

The same can be true of certain uses of demonstrative terms like ‘this’ and ‘that’ when used to pick out items available in perceptual experience. An object thought of or even perceptually presented in a certain way can be essential to the identity of a particular thought about the world even if the way that object is thought about or per-ceived is not essential to that very thought about it. In philosophy today we are increasingly familiar with the even broader question of how far something like this indexical or demonstrative aspect of thought can be seen to extend, and whether any cognitive links between thinkers and something they could think about would be possible without it.

These are all further extensions of a line of thinking arguably present in rudimentary form in Descartes’ original insight into his ‘I exist’. Although he did not pursue its implications in this particular way, it is something that perhaps promises eventually to overcome


the almost irresistible idea of a completely global independence of all thought and perception from whatever world there might happen to be that either does or does not match up to those thoughts and perceptions. To get beyond that (Cartesian?) conception altogether would be a formidable advance in human understanding. We are indebted above all to Descartes for his forceful articulation of that fateful idea, and again to Descartes and to anyone else who does anything to explain and help undermine its undeniable appeal.

2

Berkeley v. Locke on Primary Qualities

Locke was once supposed to have argued that since the colours, sounds, odours, and other ‘secondary’ qualities things appear to have can vary greatly according to the state and position of the observer, it follows that our ideas of the ‘secondary’ qualities of things do not ‘resemble’ anything existing in the objects themselves. 1 And Berkeley has been credited with the obvious objection that similar facts about the ‘relativity’ of our perception of ‘primary’ qualities show that they do not ‘resemble’ anything existing in the objects either, so that both ‘primary’ and ‘secondary’ qualities exist only ‘in the mind’. 2 The fal-sity of this view of Locke has been amply demonstrated in recent years, 3 but no corresponding revision has been made in what remains the standard interpretation of Berkeley’s criticisms of Locke. His objections therefore appear to be based on misunderstanding and to

1 Some of the many who attribute this method of argument to Locke are R. Jackson, ‘Locke’s Distinction Between Primary and Secondary Qualities’, in C. B. Martin and D. M. Armstrong (eds), Locke and Berkeley (New York: Doubleday, 1968), p. 72 ; R. I. Aaron, John Locke (Oxford: Oxford University Press, 1937), p. 117 ; G. J. Warnock, Berkeley (London: Pelican, 1953), p. 94 ; P. Cummins, ‘Perceptual Relativity and Ideas in the Mind’, Philosophy and Phenomenological Research 24 (1963), p. 208 ; J. F. Thomson, ‘Berkeley’, in D. J. O’Connor (ed.), A Critical History of Western Philosophy (New York: Free Press, 1964), p. 243 ; J. Bennett, Locke, Berkeley, Hume: Central Themes (Oxford: Oxford University Press, 1971), p. 95, 112 ; I. C. Tipton, Berkeley: The Philosophy of Immaterialism (London: Methuen, 1974), p. 37 .

2 See, e.g., Jackson, op. cit. , p. 72; Aaron, op. cit. , p. 117; R. Popkin, ‘Berkeley and Pyrrhonism’, Review of Metaphysics , 1951–52, p. 234 ; D. J. O’Connor, John Locke (London: Pelican, 1952), p. 65 ; Cummins, op. cit. , p. 209; J. Bennett, op. cit. , p. 95; and ‘Substance, Reality, and Primary Qualities’, in Martin and Armstrong, op. cit. , p. 108–109; J. L. Mackie, Problems from Locke (Oxford: Oxford University Press, 1976), p.12, 24 .

3 The earliest and best account I know of is to be found in M. Mandelbaum, ‘Locke’s Realism’, in his Philosophy, Science, and Sense Perception (Baltimore: Johns Hopkins Press, 1964) . See also E. M. Curley, ‘Locke, Boyle, and the Distinction between Primary and Secondary Qualities’, Philosophical Review 81 (1972) ; P. Alexander, ‘Boyle and Locke on Primary and Secondary Qualities’, Ratio 16 (1974) ; J. L. Mackie, op. cit. , ch. 1. For an account that agrees with many of the main points, often for dif-ferent reasons, see J. Bennett, op. cit.


be irrelevant to what is now seen to be Locke’s actual view and his reasons for holding it. I think this account of Berkeley, like the old view of Locke, is a purely fi ctional chapter in the history of philoso-phy, and in this paper I try to show that Berkeley’s criticisms involve no misunderstanding and amount to a direct denial of the view Locke actually held.

Locke was a follower of the ‘corpuscular philosophy’ of Boyle and others, 4 according to which the physical world is composed of a large number of solid atoms with size, shape, position, and motion or rest, but without colour, sound, odour, hardness, or heat. In saying that our ideas of the primary qualities of things ‘resemble’ qualities possessed by objects in the world but that our ideas of secondary qualities do not he was simply expressing the scientifi c view that everything that hap-pens in the world, including our perceiving the colours, sounds, odours, and so on that we do, is caused by the action of physical particles pos-sessing only primary qualities of the sort listed. We do have ideas of such qualities, and all objects do possess such qualities. But what, in an object, correspond to our ideas of secondary qualities are only the powers the object possesses, in virtue of the primary qualities of its fundamental parts, to produce certain ideas in sentient beings who come into contact with it. In advancing this view Locke was not con-cerned with the problem of how we can reliably tell that things really are as they appear to be, nor did he argue for the asymmetry between our ideas of primary qualities and those of secondary qualities on the ground that familiar facts about the ‘relativity’ of perception hold for the latter but not for the former. He simply supposed, quite reasonably, that only the kinds of qualities referred to in physical explanations ‘are really in them,—whether any one’s senses perceive them or no: and therefore may be called real [or original or primary ] 5 qualities, because they really exist in those bodies’. 6

To understand Berkeley’s response to this aspect of Locke’s philosophy it is not enough to select a few sentences that could be

4 See, e.g., Aaron, op. cit. , pp. 111–113; J. Gibson, Locke’s Theory of Knowledge and its Historical Relations (Cambridge: Cambridge University Press, 1960), pp. 260–265 ; Mandelbaum, op. cit. , pp. 1–15; Curley, op. cit. , pp. 441–450; Alexander, op. cit. , pp. 51–64; Mackie, op. cit. , p. 15. For Boyle’s views see his ‘Experiments and Observations Upon Colours’, in Works , I (London: 1772), and ‘The Origins and Forms of Qualities, According to the Corpuscular Philosophy’, in Works , III.

5 J. Locke, An Essay Concerning Human Understanding , II , viii, A. C. Fraser (ed.) (New York: Dover, 1959), 9 .

6 Locke, op. cit. , II, viii, 17.

Berkeley v. Locke on Primary Qualities 31

taken as an attack on something Locke might be supposed to have said. It is important fi rst to see in general what Berkeley was up to, and then show in some detail how his criticisms of Locke fi t into it.

It has been widely supposed for some reason that Berkeley was pri-marily concerned with what we have come to see as an epistemologi-cal problem of ‘appearance’ and ‘reality’ – with whether and how we can know how things ‘really’ are, as opposed to how they ‘appear’. 7 But in fact he thought the main inadequacy of previous philosophies was their faulty notion of existence, and the fi rst six sections of the Principles , 8 in which the main topics for discussion are introduced, are concerned with the question of what exists and of what it is to exist. It is announced, with little or no direct argument, that unperceived existence is really unintelligible. Some of the illusory attractiveness of rival doctrines is attributed to the infl uence of the pernicious theory of abstract ideas, but Berkeley is most interested simply in laying his view before the reader. Some truths, after all, ‘are so near and obvious to the mind that a man need only open his eyes to see them’ (§6). In those early sections there is no attempt to explain how a person knows, on a particular occasion or in general, that ‘those bodies that compose the mighty frame of the world’ exist, or that they really have the properties they appear to have, or are ordinarily taken to have. There is no mention of such questions at all.

From the fact that the esse of sensible things is percipi Berkeley concludes in §7 that ‘there is not any other substance than spirit , or that which perceives’. And for a fuller proof of this point he argues that sensible qualities cannot exist in an unperceiving thing, since they are really ‘ideas perceived by sense’, and it is impossible for an idea to exist in something that does not perceive. In his statement of this view Berkeley does not distinguish between primary and second-ary qualities. The list he gives of sensible qualities (‘that is, ideas per-ceived by sense’) includes ‘colour, fi gure, motion, smell, taste, and such like’. The point is that all the qualities we perceive, since they are

7 For the idea that some such epistemological problem is at the heart of Berkeley’s concern with primary and secondary qualities see, e.g., Popkin, op. cit. , pp. 238, 246; Warnock, op. cit. , pp. 99–100; Bennett, Locke, Berkeley, Hume , pp. 114–115; and per-haps Mandelbaum, op. cit. , pp. 3–4. For a convincing exception see M. R. Ayers, ‘Substance, Reality, and the Great, Dead Philosophers’, American Philosophical Quarterly 7 (1970), pp. 42–47 .

8 G. Berkeley, Principles, Dialogues, and Correspondence , C. M. Turbayne (ed.) (Indianapolis: Bobbs-Merrill, 1965) . Section numbers in the text refer to the Principles . Page numbers refer to the Dialogues .


‘objects’ of perception, must be ideas, and therefore exist only in so far as they are perceived.

In §8 he entertains a reply to this sweeping conclusion. It is granted that perhaps the ideas themselves do not exist ‘without the mind’, but nevertheless there might well be things ‘like’ our ideas, ‘whereof they are copies or resemblances, which things exist without the mind in an unthinking substance’. Berkeley regards this ‘hypothesis’ or sugges-tion as unintelligible, and most of the rest of the positive part of the Principles , up to the objections he begins to consider in §34, is an attempt to expose in one way or another the absurdities involved in it. Its defects all derive from its central, but incoherent, assertion of unperceived and unperceiving existence.

In §8 his response is that ‘an idea can be like nothing but an idea’, so the appeal to the alleged ‘likeness’ will be to no avail. Either the ‘supposed originals or external things’ are perceivable, and hence are ideas after all, and therefore do not exist in an unthinking substance, or else they are unperceivable and so can stand in no relations of resemblance to our ideas. Does it even make sense to say that a col-our is like something invisible, or something hard or soft is like some-thing intangible?

There are two things to notice about this argument, whatever one thinks about its cogency. 9 First, it says nothing about any problems involved in knowing that our ideas ‘resemble’ something existing in an unperceiving substance; what is in question is whether ideas ‘can’ resemble non-ideas, whether it is possible ‘for us to conceive’, or ‘whether it be sense’ to speak of, a likeness between them. And sec-ondly, the distinction between primary and secondary qualities plays no role. That is not even mentioned until the next section. So Berkeley at this point is certainly not arguing that the reasons Locke (or any-one else) gave for thinking that our ideas of secondary qualities do not resemble anything in the object show just as conclusively that our ideas of primary qualities do not resemble anything in the object either. It is the very idea of resemblance between ideas and non-ideas that is being questioned.

In §9 there is the fi rst mention of those who distinguish between primary and secondary qualities and claim that the ideas of the latter are not ‘resemblances of anything existing without the mind, or

9 Some of the assumptions on which it might be thought to rest are discussed in P. Cummins, ‘Berkeley’s Likeness Principle’, in Martin and Armstrong, op. cit.


unperceived’. The distinction is introduced by Berkeley as a sophisti-cation of, or a restriction on, the ‘likeness’ view expressed in §8. There the suggestion was that our ideas, which are admitted to have no existence ‘without the mind’, are nevertheless ‘resemblances’ of things existing ‘without the mind in an unthinking substance’. In §9 the suggestion is that only some of our ideas resemble things existing ‘without the mind’. Those who distinguish primary from secondary qualities acknowledge that our ideas of secondary qualities are not ‘the resemblances of anything existing without the mind, or unper-ceived, but they will have our ideas of the primary qualities to be patterns or images of things which exist without the mind, in an unthinking substance which they call “matter”.

It is clear, then, that Berkeley understands those who distinguish between primary and secondary qualities to be committed to the existence of things unperceived, or to certain qualities existing ‘in’ something unperceived and unperceiving. That is the sense he usually gives to his much-used expression, ‘without the mind’. The view he is going to reject is one that he takes to imply the denial of his own view that esse is percipi or percipere , and that is what he concentrates on. In his arguments against it in §9 and in §§10–13 he expresses no con-cern about whether we can tell that objects ‘really’ have the primary qualities they ‘appear’ to have, nor does he argue that we are unable to tell whether a particular object before us ‘really’ has the determinate size or shape it appears to have because of certain familiar facts about the ‘relativity’ of perception. There is no suggestion that Berkeley thought Locke had argued that way, or even that Locke was at all concerned with a problem about ‘appearance and reality’.

Berkeley’s argument in §9 merely repeats what he takes himself to have shown already:

But it is evident from what we have already shown that extension, fi gure, and motion are only ideas existing in the mind, and that an idea can be like noth-ing but another idea, and that consequently neither they nor their archetypes can exist in an unperceiving substance.

This is just the same dilemma that was presented in §8, now applied only to primary qualities: either they are perceivable, and therefore are ideas, and their esse is percipi , or else they are unperceivable and therefore cannot resemble our ideas. In neither case can it be said that primary qualities, so understood, exist ‘in’ an unperceiving substance.


The objection in §10 attempts to expose a different absurdity in the view that primary qualities exist in bodies, but colours, sounds, heat, cold, and the like do not. The point is that since it is impossible to con-ceive of a moving, extended body without also giving it some colour and other sensible qualities in addition to the primary ones, the two kinds of qualities are ‘inseparably united’ and are ‘not, even in thought, capable of being abstracted’ from one another. Berkeley might be expected to conclude that the philosophers’ conception of colourless, soundless, etc., bodies is therefore simply incoherent, or unintelligible, or even that it does not represent anything that could possibly exist, but in fact he tries to exploit what he regards as an inconsistency in their view for his own ends. He relies on the fact, granted by his opposition, that sensible qualities like colours, sounds, heat, cold, and so on, do not exist ‘without the mind’; for such qualities esse is percipi . He then argues that if the two kinds of qualities are ‘inseparably united’, and colours, sounds, etc., exist only when perceived, or ‘in the mind’, then the same must hold for the primary qualities also:

Where therefore the other sensible qualities are, there must these be also, to wit, in the mind and nowhere else.

This argument, like its predecessors, attributes no concern with epis-temological issues about appearance and reality to Locke, and it mentions no facts about the ‘relativity’ of perception of either pri-mary or secondary qualities. It is designed to show that the primary qualities cannot be separated off from the others by virtue of existing ‘without the mind’ or unperceived. With respect to existence, all qualities are in the same boat.

§11 is another attempt to show that extension, and derivatively, solidity, which cannot be conceived without extension, ‘exists not in an unthinking substance’. Once again Berkeley tries to force on his opponent a distinc-tion between perceived extension and unperceived extension and to show the former to be an idea and therefore to exist ‘in the mind’, and the latter to be incoherent. Perceived extension and motion allow of relative terms like ‘great’ and ‘small’, ‘swift’ and ‘slow’. And such degrees

are allowed to exist nowhere without the mind, being entirely relative, and changing as the frame or position of the organs of sense varies.

So extension with degrees of largeness, and motion with degrees of swiftness, must exist ‘in the mind’. Berkeley thinks this leaves his opponent in an unsatisfactory position:


The extension, therefore, which exists without the mind is neither great nor small, the motion neither swift nor slow; that is, they are nothing at all.

Only the dreaded doctrine of abstract ideas, which would provide an idea of extension in general, or motion in general, would seem to promise a solution, but for reasons we need not go into Berkeley thinks that leads nowhere.

Again, this argument does not mention secondary qualities, and therefore does not exploit any alleged similarities between our ideas of them and our ideas of primary qualities. It is not an argument from the fact that our judgments of greatness or smallness vary ‘as the frame or position of the organs of sense varies’ to the conclu-sion that things have no size, or that we can never ascertain what their size is. Nor does it suppose that the view against which it is directed is one concerned with determining the ‘real’ extension of things as opposed to the extension they ‘appear’ to have. So I think there is still no evidence that Berkeley misunderstands Locke’s views about primary and secondary qualities and our ideas of them.

§12 offers an argument about number that more or less exactly parallels the argument about extension and motion in §11, and §13 is hardly relevant to the dispute at all, being largely a denial by Berkeley that he possesses an idea of unity.

By the end of §13 of the Principles , then, Berkeley has given a number of arguments against the view that there are certain qualities that exist unperceived, or exist ‘in’ some unperceived and unperceiv-ing substance. But only in §14 do we fi nd the sentences that have seemed most strongly to support the standard interpretation of Berkeley’s views and of his understanding of Locke. It begins with a clear indication that it is introducing additional considerations—something new and different from what has gone before:

I shall further add that, after the same manner as modern philosophers prove certain sensible qualities to have no existence in matter, or without the mind, the same thing may be likewise proved of all other sensible qualities whatsoever.

I think there is some doubt about whether Berkeley includes Locke among the ‘modern philosophers’ here, but if he does have Locke in mind then it looks as if he has misunderstood him on at least one point. The ‘manner’ of proof ascribed to these philosophers is as follows:


Thus, for instance, it is said that heat and cold are affections only of the mind, and not at all patterns of real beings existing in the corporeal sub-stances which excite them, for that the same body which appears cold to one hand seems warm to another.

If, as I think has been shown, Locke never supports his views about secondary qualities in this way, and if Berkeley is asserting here that he did, then he is mistaken. 10

But the point in §14 is that the same kinds of reasons for which ‘it is said’ by ‘modern philosophers’ that heat and cold are ‘not at all patterns of real beings existing in . . . corporeal substances’, could be used to show the same thing about fi gure, extension, and motion:

Now, why may we not as well argue that fi gure and extension are not patterns or resemblances of qualities existing in matter, because to the same eye at different stations, or eyes of a different texture at the same station, they appear various and cannot, therefore, be the images of anything settled and determinate without the mind?

And just as ‘it is proved’ that sweetness is not in the object because the same object, while remaining unaltered, can taste sweet at one time and bitter at another:

Is it not as reasonable to say that motion is not without the mind, since if the succession of ideas in the mind become swifter, the motion, it is acknowledged, shall appear slower without any alteration in any external object?

There is a defi nite ad hominem or conditional nature to this argu-ment, and it is clear that Berkeley is to a considerable extent separat-ing himself off from the ‘modern philosophers’ he has in mind. He makes his point both times rather tentatively, in the form of a ques-tion, rather than asserting confi dently that primary qualities simply do not exist in external objects for the reasons given. He is careful at each point to say that ‘modern philosophers prove . . .’, ‘it is said that heat and cold . . . ’, ‘it is proved that sweetness . . .’. So he is not obvi-ously endorsing those ‘proofs’. He is arguing only that if they are

10 Popkin, op. cit. , claims that Berkeley in §14 is advancing an argument he found in Bayle, and that Bayle in turn got it from Abbé Foucher, a ‘sceptical’ opponent of Descartes and Malebranche. Certainly Foucher did not have Locke’s Essay in mind, and it is not obvious that Bayle did. Locke and Boyle are squarely in the ‘atomist’ tradition and do not use epistemological arguments from the ‘sceptical’ tradition to support their views about colours, odours, etc. Berkeley shows in §15 that he is aware that these ‘sceptical’ arguments do not prove what Boyle wanted them to prove. See below, pp. 37–38.


thought to work for secondary qualities then it is equally reasonable to think that they work for primary qualities as well.

This comes out clearly in §15, where he sums up the point of the previous section:

In short, let anyone consider those arguments which are thought manifestly to prove that colours and tastes exist only in the mind, and he shall fi nd they may with equal force be brought to prove the same thing of extension, fi gure, and motion.

We know that the non-endorsement is explicit here, that the insertion of ‘which are thought manifestly to prove . . . ’ is to be taken seriously as a disclaimer, since Berkeley immediately points out that the ‘proofs’ do not really establish what they have been thought to establish:

Though it must be confessed this method of arguing does not so much prove that there is no extension or colour in an outward object as that we do not know by sense which is the true extension or colour of the object.

This shows fi rst of all that Berkeley did not think that ‘Colour does not exist in outward objects’ or ‘Extension does not exist in outward objects’ is established or proved by the fact that an object appears to have different colours and different sizes in different perceptual situ-ations. I have tried to show that he never argues that way in the Principles prior to §14, and in §14 itself he mentions but does not endorse arguments to that effect given by others.

There is perhaps some reason to believe he mistakenly thought Locke was one of the ‘modern philosophers’ who had argued that way, but even if that is so, the conditional or ad hominem argument of §14 does not provide the only considerations he brings against Locke’s views. What he says prior to that section indicates the kinds of objec-tions he has to any view, including Locke’s, which speaks of things existing unperceived, or existing “in” an unperceiving substance. It is those earlier arguments he is referring to in §15 when he says:

But the arguments foregoing plainly show it to be impossible that any colour or extension at all, or other sensible quality whatsoever, should exist in an unthinking subject without the mind, or, in truth, that there should be any such thing as an outward object.

This is a much stronger conclusion than the claim that we can never know by sense what is the true colour or extension of an outward object. It amounts to the view that for all sensible qualities whatsoever, esse is percipi . An ‘outward’ object as Berkeley understands it here is an


object that exists ‘without the mind’, or in other words, unperceived. Locke’s view is to be rejected because it requires the existence of such objects, or of qualities existing ‘in’ something unperceived. That is the view Berkeley wants to demolish, and he sees perfectly clearly that it is not demolished simply by the fact that apples or pennies look to be different colours and different sizes in different perceptual situations.

Even if Berkeley did mistakenly think that Locke had argued that way about secondary qualities, he nevertheless shows that in another way he understands Locke’s conclusions correctly. In criticizing the ‘modern philosophers’ by pointing out that perceptual ‘relativity’ shows at most that we can never ‘know by sense’ what the true colour or extension of the object is, Berkeley shows that he takes the ‘mod-ern philosophers’ to be interested in something stronger than that epistemic conclusion. His criticism is that their arguments do not succeed in showing that ‘there is no extension or colour in an out-ward object’, and that would be a relevant criticism only if that were the conclusion they wanted their arguments to prove. So even if Berkeley did include Locke among the ‘modern philosophers’ here, it is clear that he did not take him to be primarily concerned to answer only an epistemological question about whether or how we can accurately ascertain the true qualities of an object. 11

No such question is at stake in Berkeley’s discussion of primary and secondary qualities in §§9–15 of the Principles ; those sections amount to an exposition and defence of the view that esse is percipi or percipere . Putative ways around that doctrine are suggested, most of them made from the standpoint of the standard or ‘received’ view of the scientifi c philosophers, and all suggestions are rejected for relying in one way or another on the incoherent notion of unperceived exist-ence, or the existence of something ‘in’ an unperceived, unperceiving thing. The discussion continues in the next two sections with a more detailed examination of what is there called ‘the received opinion’. And once again its real defi ciencies are not hard to discover:

But why should we trouble ourselves any further in discussing this material substratum or support of fi gure and motion and other sensible qualities?

11 If Berkeley admits in the Principles that an object’s feeling cold to one hand and hot to another does not prove that there is no heat or cold in it, then why did he use so many simi-lar arguments from perceptual ‘relativity’ to establish that esse is percipi in the Dialogues three years later? The question is raised and discussed by Tipton, op. cit. , pp. 39–41, 237–240. I try to answer the question below, pp. 42–47, in discussing the Dialogues .


Does it not suppose they have an existence without the mind? And is not this a direct repugnancy and altogether inconceivable? (§17)

This is a frontal attack on the central notion of the scientifi c philoso-phers; it is not merely the complaint that on their view certain things would remain unknowable.

Berkeley does go on, starting in §18, to raise questions about how the existence of such things as material substances could be known. But in doing so he does not mention the distinction between primary and secondary qualities or our ideas of them. And what he does say about the knowability of material substances in §§18–20 is concerned with whether and how things that are ‘without the mind’ could be known. By ‘without the mind’ he means ‘unperceived’; anything that existed ‘without the mind’ would be a counter-example to ‘ Esse is per-cipi or percipere ’. Since he thinks everyone will agree that the senses alone provide knowledge only of things that are immediately perceived and not of things that exist unperceived, the questions he raises about the knowability of things ‘without the mind’ are not about the know-ability of familiar things like tables and trees, or of the way things ‘really’ are, as opposed to how they ‘appear’. 12 What he questions is the ‘connection’ between allegedly unperceived, unperceiving substances and the things that we do perceive. He argues that there is no necessary connection between them, in the sense that we cannot deduce from any fact about an unperceived, unperceiving thing the conclusion that it produces such-and-such ideas in the mind. 13 Nor can there be any intelligible causal connection between them, if that means that the nature or characteristics of the unperceiving substance comprehensibly explain or account for its producing the ideas it does. 14

12 Further evidence that this was not his main concern is offered below, p. 40ff. 13 Locke would agree with this. See, e.g., Locke, op. cit. , II, viii, 13, where he speaks

of God annexing certain ideas to certain motions, and of the possibility of God’s act-ing otherwise.

14 Locke would agree. He thinks the structure and relations of the minute, insensible particles of a thing are causally responsible for the thing’s having the qualities that it does, and for its producing the effects that it does, including those effects that are ideas in the minds of perceivers. But he admits that: Impressions made on the retina by rays of light, I think I understand; and motions from thence continued to the brain may be conceived, and that these produce ideas in our minds, I am persuaded, but in a manner to me incomprehensible. This I can resolve only into the good pleasure of God, whose ways are past fi nding out. ( Locke, ‘An Examination of P. Malebranche’s Opinion of Seeing All Things in God’, in Works , VIII (London: 1794), 217 )


This is not to say that for Berkeley there can be no explanation of how and why we get the ideas we do. His present point is only that no intelligible explanation can be found in an appeal to an unperceiving thing that is said to cause our ideas. There are certain ‘laws of nature’ in accordance with which our ideas come and go, and Berkeley has every reason to promote the discovery of more and more such ‘laws’. 15 But what we learn in those discoveries is simply that ‘such-and-such ideas are attended with such-and-such other ideas in the ordinary course of things’ (§30), and that is not to discover any ‘necessary con-nection between our ideas’ (§31), or to discover any power or agency in them. For Berkeley all the things we perceive are ideas, and ideas themselves are completely ‘inactive’:

A little attention will discover to us that the very being of an idea implies passiveness and inertness in it, insomuch that it is impossible for an idea to do anything or, strictly speaking, to be the cause of anything. (§25)

When we discover correlations between two sorts of perceivable things, then, we are not discovering that things of one sort are the causes of things of the other sort, since all causes are active.

This line of thought is what is responsible for the differences between Berkeley and Locke about the causation of our ideas. It might look as if Berkeley is simply denying Locke’s own scientifi c account, or perhaps even denying that a scientifi c account can be given, when he concludes from the last-quoted passage:

Whence it plainly follows that extension, fi gure, and motion cannot be the cause of our sensations. To say, therefore, that these are the effects of powers resulting from the confi guration, number, motion, and size of corpuscles must certainly be false. (§25)

But that ‘follows’ simply because nothing that is not active can cause anything. Locke’s corpuscularian view, which is quite clearly what Berkeley is referring to here, is to be rejected because it confl icts with the principles that only minds are active and can cause things, and that esse is percipi or percipere . If what Locke thinks of as the causes of our ideas are perceivable then they must themselves be ideas and hence, strictly speaking, not causes; or else they must be unperceivable,

15 He believes that increasing our knowledge of the ‘laws of nature’ is a way of increasing our knowledge and appreciation of the wisdom and benefi cence of the Author of Nature (§66). Even such a committed ‘mechanical philosopher’ as Boyle would agree. See his ‘The Christian Virtuoso’, in Works , V, pp. 514, 515.


and either that is unintelligible or it implies that the causes are active things. 16 But that does not rule out a scientifi c study of perception; it only imposes constraints on how that study is to be understood.

This last alternative—that the causes of our ideas are active—is pretty clearly what Berkeley had in mind in various entries in his Philosophical Commentaries 17 which try to locate the real source of the differences between Locke’s views and his own. For example:

282 Bodies etc. do exist whether we think of ‘em or no, they being taken in a twofold sense—Collections of thoughts & Collections of powers to cause those thoughts. These later exist, tho perhaps a parte rei it may be one simple perfect power.

He had raised the same question earlier:

84 Powers Quaere whether more or one onely?

And he expressed greater confi dence in the answer later:

838 Every sensation of mind w ch happens in Consequence of the general, known Laws of nature & is from without i.e. independent of my Will dem-onstrates the Being of a God, i.e. of an unextended incorporeal Spirit w ch is omniscient, omnipotent, etc.

Similar thoughts are expressed more fully in Principles §§28–33, where Berkeley fi rst advances his views about the causation of our ideas, and are exploited later, especially in §§50–66, when he discusses the relations between his views and the mechanical philosophy. That philosophy is not to be rejected because the ‘laws of nature’ it discovers about perceivable things are false, but because it relies on ‘inactive’ causes or on unperceivable—and hence for Berkeley non-existent—entities or qualities.

After the interlude about the knowability of material substances in §§18–20 Berkeley returns to his main theme, the defence of ‘ Esse is percipi or percipere ’. And in §§22–23 he gives the famous argument that it is impossible even to conceive of anything at all existing

16 Berkeley’s objection is therefore based on something more complicated than a certain simple ‘view of the relation between philosophy and science’ favoured by ‘epis-temologists’ who wish ‘to free philosophic questions from any direct dependence upon science’ (Mandelbaum, op. cit. , pp. 3–4). What is at issue is, at least, the proper concep-tion of causality, and how an ‘inactive’ thing could cause anything.

17 G. Berkeley, Philosophical Commentaries , A. A. Luce (ed.) (London: Nelson & Sons, 1944) . Numerals refer to the entries as numbered in this edition.


unperceived or ‘without the mind’. It is clear why Berkeley is ‘content to put the whole upon this issue’. ‘The whole’ point for which he has been arguing is that unperceived existence is impossible, and if he can demonstrate it ‘with the utmost evidence in a line or two to anyone that is capable of the least refl ection’, he will have succeeded in the single most important aim of his philosophizing. That is the main conclusion he wanted to explain and defend all along, and I have tried to show that it is only in connection with that thesis that Berkeley’s whole discussion of primary and secondary qualities in the Principles is to be understood. Seen in that light, it is clearly and intentionally directed against a view that Locke actually held.

The purpose of the Dialogues is ‘to treat more clearly and fully of cer-tain principles laid down in the [ Principles ] and to place them in a new light’ (p. 106). And it is clear from the beginning that once again the aim is to show that what philosophers call ‘material substance’ does not exist, and that no absurd or sceptical consequences follow from its denial. It is to be rejected because it confl icts with the doctrine that esse is percipi or percipere . Berkeley proceeds more methodically and con-secutively here than he did in the Principles . He does not simply state his view and then explain and defend it; he tries to establish it step by step, gaining concessions on apparently uncontroversial points and then revealing as time goes on that there is nothing left to be conceded, that his whole theory of existence has been established. That procedure fi ts more easily into dialogue form, and it is not surprising that it should be used by someone disappointed by the immediate rejection of his fi rst effort on the subject.

The fi rst general step is to establish that for ‘sensible things’, esse is percipi . That is a thesis about existence, and Philonous undertakes to prove it for each kind of sensible quality in turn. He proceeds system-atically through the different kinds of sensible qualities, heat and cold, tastes, odours, sounds, and colours, arguing in each case that at any rate the perceived heat, taste, colour, etc., exists ‘only in the mind or only in so far as it is perceived. Hylas keeps fi nding himself driven to invoke some mysterious unperceivable quality in order to avoid the conclusion that esse is percipi .

After having conceded that none of the sensible qualities so far considered has any existence ‘without the mind’ Hylas introduces in his defence the distinction between primary and secondary qualities. As in the Principles , it is clear that it is introduced as a sophisticated


or restricted version of the view that all the qualities we perceive have an existence ‘without the mind’. Given the primary-secondary dis-tinction, it can be argued that only some do. And again it is clear that extension, fi gure, etc., are to be thought of as existing ‘without the mind’ in the sense of being ‘inherent in external, unthinking sub-stances’ (p. 127). That is still Berkeley’s target; it is the place at which Locke’s view of primary qualities comes into direct confl ict with the doctrine that esse is percipi or percipere .

When Hylas fi rst introduces the distinction Philonous asks ‘But what if the same arguments which are brought against secondary qualities will hold good against these also?’ (p. 127). Since he goes on actually to present arguments in the case of each kind of primary quality against its existence ‘without the mind’, it cannot be said that Berkeley is arguing simply that if secondary qualities can be shown by certain arguments not to exist ‘without the mind’ then the same kinds of arguments can show that the same holds for the primary qualities. That conditional plays no role in Berkeley’s actual procedure, since he goes on to give direct arguments for each kind of quality. Philonous’s question is nothing more than an announcement of the course the ensuing discussion is to take.

Furthermore, when Philonous refers to ‘the . . . arguments which are brought against secondary qualities’ he is obviously referring to the very arguments he himself has just given to show that sensible tastes, colours, and so on, have no existence ‘without the mind’. He is not simply referring in a sweeping manner to the ‘method of arguing’ of the ‘modern philosophers’, as in the Principles , or to those argu-ments which are thought manifestly to prove that colours and tastes exist only in the mind. Here we know exactly what sorts of argu-ments he has in mind, because he has just given them.

It is true that some (but by no means all) of the arguments given so far in the Dialogues do turn in part on facts about the ‘relativity’ of perception, and that the arguments given later against the primary qualities invoke similar considerations, 18 so it looks as if Philonous is indeed referring to the same sort of arguments as Berkeley attributed to the ‘modern philosophers’ in the Principles . And we saw that

18 Considerations of the ‘relativity’ of perception play no role in the arguments about odours and sounds, and only a minor role at one point in the argument about heat, but they are explicitly invoked in the arguments about extension and fi gure, motion, and solidity.


Berkeley thought such a method of argument ‘does not so much prove that there is no extension or colour in an outward object as that we do not know by sense which is the true extension or colour of the object’. But I think there is no inconsistency here.

Berkeley thinks that facts about the ‘relativity’ of perception can be used to prove that perceived colour or perceived extension do not exist ‘without the mind’, and hence that those facts can be used to establish, as in the Dialogues , the thesis ‘For sensible things, esse is percipi ’. But that alone does not prove quite generally that ‘there is no extension or colour in an outward object’, or the thesis ‘For every-thing that is not a mind, esse is percipi ’. Berkeley of course believes the stronger conclusion, but he does not believe it solely on the ground that all the things we can ever perceive are such that their esse is percipi . He also believes that the notion of unperceived existence is incoherent or unintelligible, and that the only things of whose exist-ence it makes sense to speak are either perceived or perceiving things. That does not follow simply from the fact that the things we do per-ceive are such that their esse is percipi . The ‘relativity’ arguments can therefore be used to establish that for sensible things, esse is percipi , without establishing by themselves the stronger conclusion that no qualities at all exist in ‘outward’ objects. In the step-by-step procedure of the dialogue form Hylas is slowly brought to concede the former while having to adopt more and more desperate measures to avoid the latter.

The subsequent discussion of whether the primary qualities are ‘inherent in external unthinking substances’ parallels the earlier argu-ments about the secondary qualities in almost every respect, except that in this case considerations about the ‘relativity’ of perception play a more prominent role. 19 The aim is always to establish that at least the perceived extension, fi gure, and so on does not and cannot exist in an unthinking substance. That is why Hylas is led to concede that, for example, ‘the very fi gure and extension which you perceive by sense exist in the outward object or material substance’ (p. 127).

19 One reason why Berkeley uses arguments from the ‘relativity’ of perception here, and not the earlier considerations about pleasure and pain, is that he thinks pleasure and pain are not annexed to our ideas of primary qualities as they are to those of secondary qualities. This is offered as an explanation of why people are more willing to grant that secondary qualities exist only ‘in the mind’ than they are to grant it about primary qualities (see p. 131). This would suggest that Berkeley derived little or no pleasure from, for example, viewing sculpture or architecture.


That might look like the incredible view that we always accurately ascertain the true shape and size of the objects we perceive, or that our perception of the primary qualities of things is always completely reliable. 20 But the reliability or accuracy of our perception of primary qualities is not in question, and there is no attempt to refute the ridiculous view that our perception of the sizes or shapes of things is always reliable. What is being refuted is the view that what is perceived exists in an ‘outward object or material substance’, i.e. in something that exists ‘without the mind’ or unperceived. And that refutation holds just as much for colours and sounds as it does for extension and fi gure.

That Berkeley is not concerned here with questions about reliably or accurately ascertaining how things ‘really’ are as opposed to how they ‘appear’ is also clear from an exchange later in the fi rst Dialogue in which Hylas tries to argue that even if one’s ‘sensation’ cannot exist ‘without the mind’, still it does not follow that the ‘object’ can-not. Again Philonous has little diffi culty convincing him that the ‘object’ he has in mind must be the perceived object, the object of the senses. And Philonous, Berkeley’s spokesman, raises no scruples against saying that both primary and secondary qualities are ‘in’ that object:

Phil. That the colours are really in the tulip which I see is manifest. Neither can it be denied that this tulip may exist independent of your mind or mine; but that any immediate object of the senses—that is any idea, or combination of ideas—should exist in an unthinking substance, or exterior to all minds, is in itself an evident contradiction. Nor can I imagine how this follows from what you said just now, to wit, that the red and yellow were on the tulip you saw , since you do not pretend to see that unthinking substance. (p. 135)

He can ‘concede’ that the perceived thing really does have both exten-sion and colour because he has never been trying to establish anything incompatible with that. It is only an ‘outward’ object’s possession of extension or colour, or indeed the very possibility of its existence, that he questions. And ‘outward’ here means ‘unperceived’.

20 The alleged view that we are never wrong in perception, that things are always exactly what they seem, is perhaps the position often described in textbooks and even in serious philosophical discussion as ‘naive realism’. No such view is a serious candi-date as an account of human beings’ conceptions of their relation to the world, and its only connection with naiveté would seem to lie in the naiveté (if that is what it is) involved in thinking that actual human beings ever think that way.


Berkeley insists throughout his philosophizing that he is defending common sense and denying nothing that the plain, non- philosophical man believes. It is no part of his aim to deny that things have colour, shape, size, or sound. In that respect there is no difference between primary and secondary qualities. He thinks qualities of both kinds are possessed by the ordinary sensible things—houses, mountains, rivers, etc.—that we perceive:

222 Men are in the right in judging their simple ideas to be in the things themselves, certainly Heat & colour is as much without the mind as fi gure, motion, time, etc. 21

Heat and colour are said to be as much ‘without the mind’ as fi gure, motion, etc., in a sense of that phrase explained elsewhere in the Philosophical Commentaries :

863 Bodies exist without the Mind, i.e., are not the Mind, but distinct from it. This I allow, the Mind being altogether different therefrom.

882 I will grant you that extension, Colour, etc may be said to be without the Mind in a double respect, i.e. as independent of our Will & as distinct from the Mind.

Clearly ‘without the mind’ is not being used here in the sense of ‘unperceived’, as is Berkeley’s usual practice in the Principles and Dialogues . He uses it here to express, not an unintelligible ‘philo-sophical’ theory of existence outside of all minds whatever, but the perfectly ordinary commonplace that what people perceive is some-thing distinct from (i.e. not identical with) their own minds, and something they do not simply invent.

Berkeley thinks that once we recognize that the things in the world of which we speak and in which we take an interest are the very things we perceive then there is no reason at all to deny that they have both size and colour, shape and smell. That is why he thinks of himself as a defender of common sense against the confusions of the materialists: 22

19 In ye immaterial hypothesis the wall is white, fi re hot, etc.

21 Berkeley, Philosophical Commentaries . 22 This is shown clearly in C. M. Turbayne, ‘Kant’s Relation to Berkeley’, in L. W.

Beck (ed.), Kant Studies Today (La Salle, Ill.: Open Court, 1969) , where Berkeley’s position is compared with Kant’s rejection of ‘transcendental realism’.


His sense of the unnatural, contrived allure of the ‘materialist hypothesis’ is perhaps best expressed in another entry in Philosophical Commentaries :

392 There are men who say there are insensible extensions, there are others who say the Wall is not white, the fi re is not hot &c We Irish men cannot attain to these truths.

Hylas employs various suspect distinctions or notions in the rest of the fi rst Dialogue in an attempt to grant that the esse of sensible things is percipi without having to conclude that in general esse is percipi . He tries to distinguish between what is perceived and what causes the ideas we get, or between sensible extension and ‘absolute’ extension. He tries to distinguish the ‘object’ from the ‘sensation’, or the act of sensation from its object. Philonous shows that none of these attempts succeed in giving a sense to unperceived existence, and Hylas is forced once more to revert to the notion of a material sub-stratum , which he claims is needed because qualities cannot be con-ceived to exist without a ‘support’. It is not diffi cult to show that this is not literally true and that no explanation has been given of what the metaphor is intended to convey.

Once again Berkeley, in the person of Philonous, is ‘content to put the whole upon’ the issue of whether it is possible to conceive ‘any mixture or combination of qualities, or any sensible object whatever, to exist without the mind’ (p. 140). He can rest everything on that question, as in the Principles , because that is what he has been con-cerned to establish all along, viz. that esse is percipi or percipere . That is a stronger conclusion than ‘For sensible things, esse is percipi ’; he thinks he has already established, or has been conceded, that weaker conclusion. The ‘conceivability’ argument, if successful, would estab-lish the stronger conclusion once and for all, and it would doom to failure in one fell swoop all further attempts to grant that sensible qualities exist only ‘in the mind’ while insisting that certain unperceived or unperceivable qualities nevertheless have an independent existence in ‘outward’ objects. The argument plays the same role as in the Principles , and for the same reasons.

I conclude that Berkeley’s discussion of primary and secondary qualities in the Dialogues differs in no essential respects from his treatment in the Principles , and that in neither case does it rest on any serious misunderstanding of Locke’s views on the subject. In


both Berkeley’s works the discussion occurs in the midst of attempts to support or defend the view that esse is percipi or percipere , and Locke’s view of primary qualities is attacked for its commitment to the existence of unperceived things or qualities. The only possible evidence for any misunderstanding comes from Berkeley’s remarks about the ‘modern philosophers’, and that would show at most that he wrongly supposed that Locke had argued in a certain way; it would not amount to a misunderstanding of what Locke’s views about primary and secondary qualities actually were. And even that minor misunderstanding could be attributed to Berkeley only if it could be shown that he was including Locke among the ‘modern philosophers’ who argued from the fact that the same body can feel cold to one hand and hot to another to the conclusion that there is no heat or cold in the object. Berkeley thinks the argument is invalid. Locke never argues in that way, nor does Berkeley, and I think the preponderance of the evidence shows that Berkeley never thought that he did.

3

Colours and Powers

I want to take up what Michael Ayers says in his book on Locke 1 about the relation between the colours of objects and the powers objects have to produce perceptions of certain kinds in perceivers. I very much admire that book, and its author, and I am not sure how or whether I disagree with the main things he says about colours and powers. But there is something he thinks is very important about per-ceptions of colour that I have doubts about. My uncertainty might be connected with something else of wider signifi cance that he also thinks is very important.

At the end of chapter 22, on sensation, Ayers describes a view he calls “the linguistic theory of psychological structure”, which is:

in effect the bold claim that . . . if we abstract away from language, all psychological structure withers away, and, without structure, ‘the mind’ withers away. The mind is just an inwardly cast shadow of the external employment of language, and the traditional enterprise of considering states of consciousness or ‘modes of thinking’ in abstraction from language is entirely misconceived. (p. 206)

Ayers regards this theory as anathema. He rejects it on the grounds that:

Unless there were animal experience and intelligence and emotional life prior to language, there would be nothing to incorporate and comprehend lan-guage, or to be transformed by it. Our power to use language cannot, there-fore, be the explanation of the intentionality of our mental states. (p. 206)

I am uncertain whether the issue about perceptions of colour on which I think I disagree with Ayers is connected with his resistance to this “theory of psychological structure”. That is a question I will return to, when we might be in a better position to see what this

1 Michael Ayers, Locke: Epistemology and Ontology , Routledge, London and New York, 1991 (page numbers in parentheses in the text refer to this book).


“linguistic” theory or conception of the mind really is, and what it does and does not imply.

But fi rst, the question about colours and powers is whether we can understand the colours of objects to be nothing more than powers objects have to produce perceptions of certain kinds in certain kinds of perceivers in certain kinds of circumstances. It might seem that Locke did understand the colours of objects in that way, since it is central to what Ayers calls Locke’s “offi cial” view that the word ‘blue’, for instance, functions in two different ways. In its primary employment the word is the name or sign of a simple idea. A simple idea is said to signify or represent the object that causes it to appear in the mind. But in addition to naming that idea in the mind, the word ‘blue’ also secondarily or derivatively names the object or cause that produces the idea that the word directly or primarily names.

This view as it stands leaves open the possibility of one and the same object causing an idea of a certain kind in one person and an idea of a different kind in another person. Because of a difference in their perceptual mechanisms, two perceivers could get different kinds of ideas from the same object, but they could both still apply the word ‘blue’ correctly to the object that caused those ideas. That would be a secondary employment of that word; it would (secondarily) name the single cause of their different ideas. If the two perceivers get different kinds of ideas from that same object, the word ‘blue’ in its primary employment would name a different kind of idea, and so would mean something different, for the two different speakers.

This is so far no more than an abstract possibility left open by the structure of Locke’s “offi cial” view. But it is a possibility Ayers thinks Locke was right to insist on, since it is important to acknowledge the mechanisms by means of which objects affect a person and give him the ideas he gets. Locke apparently believed that although there is a possibility of this kind of variation in people’s ideas, that is not what actually happens. He thought people’s perceptual mechanisms are in fact probably suffi ciently similar that objects of the same kind always produce ideas of the same kind in everyone under the same conditions.

How Locke could ever know such a thing, or even have any reason to believe it, I must say, is beyond me. But in any case, if he is right about that, the word ‘blue’, as things are, names the same kind of thing for everyone in both its primary and its secondary employment. What it names in its secondary or derivative employment is whatever

Colours and Powers 51

it is in objects that causes the simple ideas that the word names in its primary employment. And, Locke says, that is nothing other than the power such objects have to produce simple ideas of that kind in perceivers. A word like ‘blue’ secondarily or derivatively names that power.

Does this amount to the view that an object’s being blue is its power to produce ideas of a certain kind? Or, generalizing from blue to all colours, that the colours of objects are powers? I think that, so far, it does not amount to such a view. That is because there is nothing in the theory as stated so far to indicate what it is to think of an object’s colour, or how we manage to do it, or even how we can think of an object as having any colour at all. Even if we grant that the word ‘blue’ names a power in an object, that does not show that the colour of an object is a power. That requires, at the very least, that the word ‘blue’ be a colour word. And what makes the word ‘blue’, or any word, a word for a colour? For Locke what determines the meaning of a word is the kind of idea the word stands for. A word is a colour word only if it names or is the sign of an idea of colour. But what makes an idea, and especially a simple idea, an idea of colour ?

This is a question Locke is not in a good position to answer. As Ayers points out, Locke understands simple ideas that are caused in us by the presence of an object as nothing more than “blank sensory effects” of that object. By that I take him to mean that as long as the simple ideas produced by objects of a certain kind have some distinctive feature that links them with their cause, those simple ideas stand for or signify that cause. The origin of simple ideas is what alone determines what simple ideas of distinctive kinds signify or represent. They represent what causes them. In that sense there is no possibility of simple ideas’ being false of the objects that distinctively cause them. Simple ideas are “blank sensory effects” in the sense that what marks them as ideas of this or that kind is whatever distinctive feature the ideas caused by those objects happen to possess. Words for simple ideas therefore name different kinds of “blank sensory effects” and thereby (secondarily or derivatively) name the powers objects have to produce those effects.

This still does not explain what makes a simple idea an idea of colour , and so what makes a word that names that simple idea a col-our word. Ideas that all share a certain distinctive feature are not thereby ideas of that very feature. There are too many different pos-sible candidates to single out one of the shared features in particular


over others. Even if a certain group of ideas all share the same colour (if that makes any sense), that does not make them ideas of that dis-tinctive colour. Thinking of simple ideas as nothing more than “blank sensory effects” of a certain cause tells us nothing about the contents of those ideas, or even whether they have any contents at all. Their unique origins and distinctive properties are not enough to account for them as ideas of blue, or of red, or of anything else.

Taken strictly, then, Locke’s “offi cial” view says only that a word like ‘blue’ directly names a certain kind of sensory effect that is pro-duced by objects of a certain kind, and derivatively names the power objects have to produce those effects. Strictly speaking, all we have so far is a name, the word ‘blue’, and a certain class of sensory effects that the word is said to name in its primary signifi cation, along with the power an object has to produce those effects, which the word names or indicates in its secondary signifi cation. But this alone gives us no way of saying that that power is the colour of an the object, or that the colour of an object is a power. It has not been explained how we are even able to think of the colours of objects at all.

As Ayers points out, what is missing in all this is any account of what an idea is an idea of . As he puts it, to characterize an idea as an idea of, say, blue is to characterize it in terms of an aspect of its intentional content – to specify “the conditions under which it would be veridical” (p. 210). This seems to me exactly right; it brings in the idea of truth as essential to the understanding of intentional content. That is what Locke’s conception of simple ideas as “blank sensory effects” of objects does not capture. The conditions under which an idea of blue would be veridical are conditions under which a thing to which the idea is applied is blue. That is what determines whether or not an idea is the idea of blue; not its distinctive origin or any qualities it might happen to possess as a mere sensory effect. It is because it is an idea of a certain distinctive quality: blue. To have an idea of a physical object in space as having that quality would be to have a thought that is true if and only if that object is blue.

That, I believe, is how colour words actually function. Ayers appears to agree. But he takes this to show, as he puts it, that “predi-cates like ‘blue’ are primarily predicated of objects rather than sensa-tions” (p. 210). That qualifi cation suggests that although predicates like ‘blue’ are primarily predicated of physical objects, he thinks they are secondarily or derivatively predicated of or applied to ideas or sensations. That would be a simple reversal of Locke’s view, with the


order of priority going in the opposite direction. I think that still concedes too much to Locke’s way of thinking of colour words as getting their meanings by standing for ideas or sensations in some way. On the view of the intentional content of ideas that I think Ayers and I agree about, it is enough to say that the word ‘blue’, when predicated of anything at all – whether it be an object or an idea or a sensation or whatever it might be – is veridical or true of that thing if and only if the thing is blue. There is no need to distinguish between primary and secondary applications of the word to different kinds of things.

Ayers wants to leave room for what he calls two different kinds of application (if not predications) of words for colours and for what he calls other secondary qualities. He thinks that in their primary appli-cation such words are true or false of objects in space. But he thinks the words can also have a secondary application to ideas, or to what he sometimes calls “sensations” or “impressions” or “sensory effects”. He says, for instance, that the word ‘blue’ as it appears in the phrase ‘a sensation of blue’ does not indicate that there is a sensation that is blue, but rather that “if the sensation is to be veridical, then some-thing must be causing it which is . . . blue” (p. 210). Here the word ‘blue’ is thought of as primarily predicated of an object in space – some-thing which is or is not blue. But Ayers thinks the phrase ‘a sensation of blue’ also serves to characterize a kind of “sensation”, and it does so by characterizing the cause of that sensation. That is a secondary or derivative application of the word ‘blue’; the word applies to a sen-sation by means of its cause. And that is what seems to bring in the idea of causes or powers as referents of colour words.

Ayers thinks we must make room for these two different applica-tions of colour words in order to acknowledge the effects of the per-ceptual mechanisms through which objects affect us. This is what he thinks Locke was right to recognize, and it is what leaves open the abstract possibility that (as Ayers puts it) “with a different mechanism, the sensory effects of the same intrinsic attributes [of an object] might be different” (p. 210). So we need some way to describe those possibly different “sensory effects” themselves, and not just the object that causes them. For Ayers this means that we need some way of describ-ing “the phenomenal quality” of our sensory experience. And he thinks we can do this by using a colour predicate to connote what he calls a “phenomenal attribute” of an “idea” or a “sensation” or a sensory effect, not simply a property of an object in space.


Given that we need some way to describe our perceptual experi-ences or “sensations”, this raises the question why we cannot do it by using predicates like ‘blue’ that are understood as true or false of objects in space. If we use colour predicates to describe what people see, why would we not thereby describe the kinds of experiences they have? Ayers appears to think that restricting ourselves only to colour predicates thought of as true of objects in space would leave us unable to account for certain facts of perceptual experience. I say that is what he appears to think; he does not directly argue for that claim or specify the facts of experience he thinks could not be accounted for.

What he actually does is to show that a certain view that under-stands colour predicates as used only as applied to objects does not give a satisfactory account of the “phenomenal quality” of percep-tual experience. He thinks Locke’s view, which allows for both ‘pri-mary’ and ‘secondary’ applications of colour words, therefore shows a fi rmer grasp of what Ayers calls “the ontology of percep-tion” (p. 210).

The view Ayers opposes is what he calls “the ‘private language’ argu-ment” or “the Wittgensteinian view”. That view stresses the public use of colour predicates as applied to objects in space, and the dependence of that use on agreement in judgements expressed in a shared language. On that view, speakers who share the same colour predicates will agree among themselves not only in their judgements about which objects are blue or yellow, but also about which objects look or seem blue or yellow to them. They will therefore agree in general in their judgements of ‘the phenomenal quality’ of their experience as well as in their judgements about the colours of objects. So on this “Wittgensteinian view”, what Locke allows as at least a possibility – that different per-ceivers could get experiences with different ‘phenomenal qualities’ from the same objects – would turn out to be incoherent.

If we try to apply the “Wittgensteinian” view to the thought of two groups of perceivers whose perceptions of colour are transposed so that the colour one group sees on looking at violets is the colour the other group sees on looking at marigolds, we seem to get nowhere.

With such community of judgement . . . goes community of meaning. Consequently we must admit that after all violets and marigolds give rise to the same visual sensations, namely sensations of blue and yellow respectively, in both purportedly differentiated sets of observers. No difference in the phe-nomenal quality of the experience of such observers can be described or indicated in language. (p. 209)


Ayers rejects that view on the grounds that it “does not seem in the least senseless” to speculate about “the phenomenal quality of experience which goes beyond mere community of judgement” (p. 211). We can make perfect sense of the same objects’ looking one colour to one set of observers and a different colour to a different set of observers with different perceptual mechanisms.

On this point, it seems to me, Ayers is surely right. We can make sense of different or deviant or defective perceptual mechanisms leading different perceivers to see different colours when looking at the same objects. If we ourselves were among those perceivers, we might conclude that some or all of those people see colours that are not there, or that things look to them to be a colour they are not. We can understand that someone with perceptual mechanisms different from ours might see blue in the presence of something that is yellow. That is possible, and is not in the least senseless.

But does that mean that we must understand colour predicates to be predicated of or applied to something other than objects in space in order to account for all the facts of perceptual experience? Does it mean that we must acknowledge an application of colour words to things like “ideas” or “sensations” or “visual impressions” or “sen-sory effects”? Must some such objects be reckoned as part of the “ontology of perception”? And if so, must we further acknowledge that colour words in such ‘secondary’ applications refer to or some-how indicate the powers objects have to produce such things? I agree with Ayers in rejecting what he calls “the Wittgensteinian view”, but I would answer No to all these questions.

As for the “ontology” of perception, Ayers speaks primarily of “sensations”. But in speaking of what he calls “sensations of blue”, for instance, he seems to have in mind simply seeing blue. And to describe someone as seeing blue, I don’t think we need such things as “sensa-tions” or “impressions” for the colour word to apply to. The only objects we need are the perceiver and, perhaps, the object that is seen, if there is one. Of course there is also the seeing itself, which is an ‘object’ in the sense of being picked out by a singular term. It is some-thing like an event, or an activity, or a state, and it can be described in terms of its content – what it is a seeing of – which does not require referring to another object. So to account for the role of the word ‘blue’ in the phrase ‘a sensation of blue’, or ‘seeing blue’, I don’t see that we need any ‘secondary’ application of the word to something that is not an object in space.


So I do not agree with Ayers’ claim that the word ‘blue’ as it enters into the phrase ‘a sensation of blue’ indicates that if that sensation is veridical then something physical must be causing it which is blue. To say of a sensation or a seeing that it is of blue is only to say that it is a veridical perception of something if and only if what it is a percep-tion of is blue. It is a perception of a property that it would be true to attribute to a thing if and only if the thing is blue. That is the condi-tion of its ‘veridicality’. But a sensation or a seeing of blue could be caused by something that is not blue.

As far as I can see, this view of the conditions of ‘veridicality’ or the intentional content of perceptions of colour is all we need to explain how colour terms are used to describe our experiences. We use them in saying what we see, or how things look to us, or in describ-ing the content of an experience, or what an experience is an experi-ence of . That is to describe our experiences. In using colour words to describe our experiences we use them with the same meanings they have as predicated of objects in space. Those words have the mean-ings they do only because there is a general practice of using such words to say things about the colours of objects around us. Without widely shared judgements as to the colours of things, there would be no colour terms at all. So broad agreement as to the colours of things is a condition of a community’s having any colour terms as we under-stand and make use of them in describing objects and describing our experiences.

We can describe our experiences in those determinate ways only because we know what words like ‘blue’ and ‘yellow’ and so on mean, or what blue and yellow are. We thereby know what it is to see blue and to see yellow. So we know and can recognize the difference in what Ayers calls the “phenomenal quality” of seeing blue and seeing yellow because we know the relevant difference between blue objects and yellow objects. With mastery of that distinction, we can grant the possibility of some people having experiences of seeing blue in the presence of objects that are yellow, and vice versa. I think some-one could even know that he was having such an experience, and would be understood if he reported it to others. “My God,” he might say, “what’s gone wrong? I look at the butter and I see blue!”

What Ayers calls “the ‘private language’ argument” or “the Wittgensteinian view” was wrong to deny the possibility of a differ-ence in “phenomenal quality” of observers’ experiences due to differ-ent perceptual mechanisms. But I think widespread agreement in


judgements as to the colours of things is necessary for intelligible talk about the colours of things. The general agreement needed to secure fi xed meanings for colour terms as applied to objects can allow for the use of those same terms, with those same meanings, to describe experiences the contents of which would deviate wildly from shared judgements of the colours of objects. So I think we cannot conclude from the need for community of judgement about the colours of objects that “no difference in the phenomenal quality of the experi-ence of such observers can be described or indicated in language” (p. 209). We can indicate and describe in language the difference between people who see blue and others who see yellow in the pres-ence of the same objects.

I take it that that “Wittgensteinian view” as Ayers understands it is an instance of the more general view he calls the “linguistic theory of psychological structure”. According to that theory, “the traditional enterprise of considering states of consciousness or ‘modes of think-ing’ in abstraction from language is entirely misconceived” (p. 206). I agree that what Ayers calls “the Wittgensteinian view” must be rejected. I have suggested that a ‘community-based’ conception of the meanings of colour terms as predicated of objects can account for differences in the “phenomenal quality” of perceptions of colour. Does that commit me to some version of the dreaded “linguistic the-ory of psychological structure”? I don’t know. This is the uncertainty I mentioned at the beginning.

As far as I can see, even a creature without language, and with no possibility of ever getting any language, could see blue. We know what blue is, so we have no diffi culty in understanding that. If seeing blue is a “state of consciousness” – as I suppose it is — then this allows for “states of consciousness in abstraction from language”, as the “linguistic theory of psychological structure” apparently does not. But for other “states of consciousness” I am not so sure. Can we say of a non-linguistic creature who sees blue in the presence of vio-lets that those violets thereby look blue to that creature? Do the vio-lets look to that creature to be blue? Does the creature, in normal conditions, take the violets to be blue? I am not sure we can say any of this, even if the creature sees blue when it looks at the violets.

For thinking that the violets look blue, or judging that they look blue, as for judging that they are blue, it is clear that much more than simply seeing blue is needed. I think someone could be in such “states of consciousness” – if that is what they are – only if he has the concept


of blue, and so has concepts of other colours as well. And I don’t think anyone who is not a master of a language with colour terms has those concepts.

Does this mean I am committed to thinking that “the mind is just an inwardly cast shadow of the external employment of language”, as the “linguistic theory” apparently holds? Again, I am not sure. I do think a tremendous range and variety of psychological states are possible for human beings only because they can speak; they could not be in any such states without the capacities that language pro-vides. But I do not disagree with Ayers that:

unless there were animal experience and intelligence and emotional life prior to language, there would be nothing to incorporate and comprehend lan-guage, or be transformed by it. Our power to use language cannot, therefore, be the explanation of the intentionality of our mental states. (p. 206)

But I do think our capacity to speak is a condition of virtually all human intentionality, in the sense that the scope of our intentional-ity would be reduced to a more-or-less sub-human level without the power of language. So perhaps I am acquiescing in a weaker version of the “linguistic theory of psychological structure”. But is that weaker version equally unacceptable?

I want to argue in any case that the conception of intentional con-tent of colour perception that I think Ayers and I agree about repre-sents an obstacle to accepting the idea that the colours of objects are powers. It does not stand alone; other things also count against that view. As Ayers so clearly explains, the power an object has to produce something or other is individuated or distinguished from other powers of the object only in terms of the particular kinds of effects the object is disposed to produce. So if an object’s being blue is to be identifi ed with the object’s having a power to produce certain distinctive effects, those effects must be identifi ed or individuated independently of the power the object has to produce them. The power can be identifi ed only if the distinctive effects are independently specifi ed.

Locke’s view of simple ideas as blank sensory effects of causal interactions with objects can seem at fi rst glance to be congenial to the idea of colours of objects as powers. On that view, it looks as if sensory effects of distinct general kinds could be independently iden-tifi ed, and the powers objects have to produce effects of those kinds in certain circumstances could then be individuated in terms of them. That encourages the thought that objects’ having powers to produce


such effects is enough in itself to explain all our experiences of and thoughts about the colours of objects. But as Ayers explains, to think of perceptions of colour as having intentional contents, and not sim-ply as blank sensory effects, is implicitly to deny the possibility of any independent specifi cation of those perceptions, and hence of any independent identifi cation of the power to produce them.

That is how the idea of specifying the content of a perception in terms of the conditions under which it would be veridical represents an obstacle to the view that colours are powers. In identifying a per-ception as a perception of something blue, for instance, we draw on what we think when we think an object is blue. The thought of an object as having a colour is prior to the identifi cation of the distinctive perceptions that the object has a power to produce. So the thought of an object as having a certain colour is appealed to in identifying the power that objects that have that colour are said to have.

Does this mean that the colours of objects are not the powers objects have to produce perceptions of certain kinds? Does it mean that the proposed identifi cation of colours with powers is wrong? I want to say “Yes, it does mean that”. Ayers says “No, it does not”, or at least “Not exactly”. He is very careful here. He rightly stresses the point just made:

because the relevant effect, in the case of a quality like yellow, is a way in which objects appear or are presented to us, that effect cannot be distin-guished independently of the quality in the object. The effect, the visual impression of yellow, has to be distinguished in terms of the phenomenal attribute, rather than the reverse. (pp. 216–217)

But he immediately follows that sound observation by saying:

Nevertheless, the characterization of the phenomenal attribute as a power is not simply false or inept. (p. 217)

This seems to say that although the phenomenal attribute distinctive of perceptions that are perceptions of yellow cannot be identifi ed independently of what is thought to belong to an object in the thought that it is yellow, there remains something right (or not sim-ply false or inept) in the thought of that attribute as a power of the object. Ayers earlier made the point — which I disagreed with – that the word ‘blue’ in the phrase ‘a sensation of blue’ indicates that if the sensation is to be veridical, something physical must be causing it which is blue. He now says:


That point may not show that the characterization of secondary qualities as powers is false, but it does show that such a power must be an epistemologically very peculiar power. So much, perhaps, is only to be expected of a power which is directly and primarily manifested in a feature of the appearance of its possessor to observers. (p. 210)

This appears to concede that although it is not strictly false that the colour of an object is a power the object has, if it is a power, it is a very peculiar power.

I think there is no question that a blue object – a violet, say – has a certain power. I don’t think it is even a very peculiar power, although it is perhaps not often remarked on. A blue violet has the power to cause, in suitably placed and suitably equipped observers, a seeing of something blue. That power is manifested in that very seeing – in the appearance to observers of something blue, or in something blue’s being visually presented to observers. That is what the presence of a blue violet, in the right circumstances, brings about; the violet’s being there is what makes us see blue. But I cannot see how a violet’s having that power could be the same as its being blue – that its being blue just is its power to produce such a seeing. I do not deny they are the same because I think some blue things do not have that power, or because I think some things that have that power are not blue. It is just that I do not see that being blue and having the power to produce seeings of blue are one and the same thing – the same property, if you like.

A blue violet is a visible object, and its being blue is a visible prop-erty of it. Being visible means being capable of being seen, or having the power or disposition to be seen, under the appropriate condi-tions. Visibility is a power, but not really a very peculiar power. It is a power that a great many objects and a great many properties of objects have. Being visible is a power that is manifested in the seeing, or the being seen, of something with a certain property. A circular object, as long as it is not too big or too small, is also a visible object, and its being circular is a visible property of it. It has the power or disposition to be seen – to produce in appropriately placed observers a seeing of something circular. That is how that power is manifested. But no one, I take it, would suggest that an object’s being circular just is the power it has to produce seeings of something circular. I think it would be false or inept to identify an object’s being circular with the power it has to produce perceptions of something circular. Again, not because some circular objects do not have that power, or some objects that have that power are not circular.


In the case of a circular object it seems right to say that the object has the power to produce seeings of something circular in the appro-priate circumstances precisely because it is circular. Square objects do not have that power. The circular shape of an object is part of the reason we see something circular when we look at it. Similarly, it seems to me, the reason we see something blue when we look at a violet in the appropriate circumstances is that the violet is blue. It has the power to produce such seeings precisely because it is blue. Yellow objects don’t have that power; that is because they are yellow.

So we could perhaps say that the colour, or the shape, of an object is a property of the object that accounts for its having the power to produce perceptions of something with that colour, or with that shape. But that is not to identify an object’s having that colour, or that shape, simply with its having such a power. It is to identify being blue, or being circular, with a property the having of which gives an object the power to produce perceptions of something blue, or of something circular. And that is the property of being blue, or being circular. That makes it no surprise, or nothing peculiar, that the power an object has to produce perceptions of something blue, or something circular, is manifested in a distinctive character of just such perceptions.

There is perhaps another, more Lockean, way of understanding what Ayers calls the “epistemologically very peculiar power” that the colour of an object might be identifi ed with. He says what is peculiar about the power he has in mind is that it is “directly and primarily manifested in a feature of the appearance of its possessor to observers”. This talk of an “appearance” might be taken to refer to something that is produced in perception, and so to a distinctive “feature” of that “appearance” that the object that produces it has the power to pro-duce. This would be a very familiar traditional conception of what we might call the “ontology of perception”. It can make it look as if all that is needed to account for the “phenomenal attribute” present in perceptual experiences of colour is some object somewhere that has the power to produce “appearances” with such distinctive “features”. Nothing more than such a power would have to be ascribed to the object that produces those “appearances”. So what are vulgarly called the colours of objects could be said to be nothing other than the pow-ers objects have to produce such “appearances”.

This is, more or less, Locke’s view. And although it encourages the idea that colours are powers, it faces all the diffi culties we have seen.


I think it can give no account of how it is possible to think of an object as being coloured. So it can give no account of how it is pos-sible for us to think an object’s being coloured is nothing more than its having a certain power. That is because we cannot really account for the intentional content of perceptions of colour. The view treats “appearances” or “ideas” or “sensations” as nothing more than blank sensory effects of perceivers’ interactions with objects. That is what seems to make room for the idea that the colour of an object is noth-ing more than its power to produce just such sensory effects. But that leaves out of the account the intentional content of those sensory effects – what they are experiences or perceptions of – so the power with which the colour of the object is to be identifi ed is left unspecifi ed.

The view of intentional content that I think Ayers and I share holds that sensory effects can be identifi ed as perceptions of a certain colour only in terms of their distinctive intentional content. That content is to be specifi ed by giving the conditions under which it would be true that an object has that particular colour. So the thought of an object’s having a certain colour must make independent sense to us prior to the identifi cation of those sensory effects. That is what I think is incompatible with an object’s having a certain colour being the same as – nothing other than – its having a power to produce those effects.

If we tried to generalize this Lockean conception of the “ontology of perception”, so that what it says about perceptions of colour is meant to apply to all perception of anything, even so-called primary but still visible qualities of objects, I think no visible qualities could intelligibly be predicated of objects at all. Strictly speaking, the only visible qualities we see would be qualities or features of “appearances” or ideas. There would be no room within such a view for thought of the visible qualities of any objects, and so no room for thought of vis-ible objects at all, since they are visible through their qualities. The most that could be said on that view is that we receive “ideas” or “appearances” with certain distinctive features we can notice, and we somehow think of objects that we never perceive as producing those “appearances”, or as having a power to produce them.

On that view, an object might produce an “idea” or “appearance” with a certain “phenomenal attribute” to which we attach the name ‘circular’. That would not so far give us a way of thinking of the object that produced that “appearance” as circular. We could think of the object – if we could think of it at all – only as an unperceived


something that produces that distinctive “phenomenal” effect. But we could not think of our being aware of that “phenomenal” effect as our perceiving an instance of the property of being circular, or of something’s being circular. To think of what we perceive in that way requires a capacity to think of an object as being circular, and so far we would have no way of thinking of objects as having the (purely “phenomenal”) features we can notice in our experience. We would so far have no notion of an object with perceptible qualities at all. 2

If it is granted that the thought of an object’s being circular or hav-ing other visible qualities is already intelligible to us, there is no dif-fi culty in our recognizing that that is what our perceptions are perceptions of, and so no diffi culty in our thinking of objects as hav-ing the visible properties we are aware of in perception. But if the thought of an object’s being circular is intelligible to us, and we can make sense of perceptions as being of something circular, why can it not equally be granted that the thought of an object’s being blue, or yellow, is also antecedently intelligible to us, so we can make sense of our perceptions of the colours of objects in those same terms? If we do think of the colours of objects in that way, as I think we do, I think we cannot understand the colours of objects to be nothing other than powers.

I think this shows at the very least that what the Lockean view says about perceptions of colour cannot be generalized to apply to all perception of all visible qualities of objects. This leaves us with the question at the heart of this whole issue. If not all perceptions of the qualities of objects can be understood in the way the Lockean view would understand the perception of the colours of objects, what is the difference between being circular and being blue that encourages the idea that what causes perceptions of something circular is an object’s being circular, but what causes perceptions of something blue is not an object’s being blue, but only its having a power to pro-duce such perceptions? I will not try to answer the question. Whatever the differences might be, I don’t think they are to be found in any

2 This is a version of what I think Hume struggles with in his reductio ad absurdum of the doctrine of secondary qualities. If we try to think of all our perceptions of objects as that doctrine says we are to understand our perceptions of secondary qualities, Hume found, “we utterly annihilate all these objects”, and “nothing we can conceive is possest of a real, continu’d, and independent existence”. See that section of his Treatise called “Of the modern philosophy”. L.A. Selby-Bigge (ed.), 2nd edition revised by P. H. Nidditch, Oxford, Oxford University Press, 1978. p. 228.


differences between the ways we think or speak about circular things and the ways we think and speak about blue things. Whatever differ-ence is felt – if it is felt – must have deeper metaphysical roots.

I conclude by returning briefl y to “the linguistic theory of psychological structure”. If we understand psychological states as having intentional contents, as we do, and we understand the con-tents of perceptions and thoughts and other such states in terms of the conditions under which the perceptions would be veridical or the thoughts true, then in making sense of whatever structure we think psychological life possesses, we make essential use of the idea of truth, or the idea of something or other’s being so.

It is because we are masters of language that we are able to think in that way. That does not mean that we must attribute language to every creature we can understand to have any kind of psychological life. But it does seem to require that a creature must be capable of thoughts of something or other’s being so, or being true, in order to be capable of any reasonable level of psychological complexity, espe-cially if that complexity is present in the relations among the con-tents of those different psychological states. And how far and in what respects creatures who cannot speak are capable of states with such complex, interrelated contents is a diffi cult question. It seems to me that if we take away language an enormous amount of psychological structure does wither away. To that extent a great deal of what we call ‘the mind’ could perhaps be described as “an inwardly cast shadow of the external employment of language”. But rather than denying something intrinsic to mental life, or restricting its range, that seems on the contrary to indicate the enormous extent to which language makes possible the rich and complex kind of psychological life we all know we have.

4

The Study of Human Nature and the Subjectivity of Value

The Tanner Lectures on Human Values Delivered at The University of Buenos Aires June 7, 1988

I believe that the philosophical study of moral and other values is fi lled with diffi culty. Our best attempts to understand evaluative thought of all kinds — all questions of good and bad, better and worse — seem to me to distort or threaten to obliterate the very phe-nomenon we want to understand. But if we refrain from pressing for a philosophical exposé of values we appear to ourselves to be simply acquiescing in a way of thinking and acting without understanding it. And that leaves us dissatisfi ed. So we persist, and end up misrepre-senting and so still not understanding the phenomenon of value.

I would like to present at least the outlines of the dilemma I see. Having it clearly before us is a necessary step toward fi nding some way out of it. In these two lectures I can explain it only sketchily and at a regrettably high level of generality.

I stress that it is a philosophical way of thinking about values in general that I am interested in, not any particular morality or political arrangement or set of values in itself. It is a very powerful conception of what is really going on when human beings deliberate, evaluate courses of action, and make choices, or assess the choices and actions of others. If some such conception really is at work in our under-standing of ourselves, it can be expected to affect the way people think concretely about what they are doing, and why. And it can come in that way to affect what people actually do. There are perhaps good reasons in general to doubt that such an abstract, purely philo-sophical theory could ever have such palpable effects. But on the other hand it seems hard to deny that many of the ways we think and speak about our current social arrangements, and the justifi cation typically offered for them, do rest on some such conception of value


in general. I will not have time to go into the question of the extent to which that is really so.

I will fi rst try to describe the main outlines of the conception of evaluative thought that I have in mind and then identify the kind of distortion or denial I think it leads to. Then, in the second lecture, I will turn to the question of how and why we are so inevitably driven toward that dead end. I think it comes from nothing more than our desire to understand ourselves in a certain way.

I

Any attempt in philosophy to understand morality or evaluative thought generally leads almost inevitably to what I shall call “subjec-tivism.” It is not always easy to notice this tendency, let alone to lament it, since the kind of view one is led to appears to have gained the status of orthodoxy. There seems to be no other way to think about values. And so, we think, nothing true is being distorted or denied at all.

The idea, in a word, is that values are “subjective,” that questions of value are not questions with “objective” answers, that the good-ness or badness of a thing or a course of action is not something that belongs to the world as it is in itself, independently of us. There are many different versions of this single thought. I will not be concerned with each one of them. It is what they all have in common that leads to the diffi culties I see.

What they all have in common is the thought that there are no evaluative facts. In general, when we say or believe something, if things are the way we think they are, if the world is in fact the way we say it is, then what we say or believe is true. When what we say is false, things are not that way, the world is not in fact as we say it is. In sci-ence and all other forms of inquiry we seek the truth. By that I mean nothing lofty, abstract, or metaphysical. I mean only that in this or that particular way we want to fi nd out what is so, how things are, what the world is like in one or another respect. The question can be quite particular and trivial (e.g., Where is that book I was reading yesterday?) or extremely general and profound (e.g., What, if any, are the fundamental elements of the universe?). It is always a question of what is so — what are the facts. Whether or not something is so, whether there is anything there to be discovered or not, is in general something that holds quite independently of whatever we might

The Study of Human Nature and the Subjectivity of Value 67

happen to think about it, however we might feel about it, or even whether we are at all interested in it or not.

On the “subjectivist” view, matters of value — of the goodness or badness, the beauty or worth, of a thing or action — are not in this way anything to be found among the facts of the world at all. They are therefore not part of anything that scientifi c or any other kind of cognitive investigation could study and try to make progress on. There is in that sense no possibility of moral or aesthetic or, in gen-eral, evaluative knowledge. Not because our faculties are too weak to discern the true value of things, and not even because evaluative mat-ters are so complex that we can never expect universal or even wide-spread agreement about them. It is, rather, that in the realm of values there is simply no “objective” truth to be known. The world in itself is just what it is; it is simply there. It is the totality of facts, and it is value free.

Of course, human beings do take an interest in certain facts. They care about certain things and not about others, they want certain things, they try to bring about certain states of affairs and to prevent others. Those are undeniably facts of the world. Human beings are part of the world, and they do think and feel and act in those ways. In short, human beings value some things or states of affairs more highly than others. That is a fact of the world, but it is not an evalu-ative fact. It is not a matter of one thing’s being better than another. It is simply a matter of human beings’ regarding one thing as better than another. For the “subjectivist,” there are “objective” facts of what humans do, but not of the value of what they do, or of the value of anything else.

What “subjectivism” denies, then, is not that human beings do place value on certain things, but only that there is such a thing as being correct or incorrect in those valuings, as we can be correct or incorrect in our beliefs about the facts. When we say or think that something is good or worthwhile, or evil or ugly, our thinking or say-ing it is certainly something that is so, but either there is nothing at all that makes what we think true or false, or if in some way there is, it is only something about us, something “subjective.” And it is noth-ing evaluative.

Hume put the view this way:

Take any action allow’d to be vicious: Wilful murder, for instance. Examine it in all its lights, and see if you can fi nd that matter of fact, or real exist-ence, which you call vice . In which-ever way you take it, you fi nd only


certain passions, motives, volitions and thoughts. There is no other matter of fact in the case. The vice entirely escapes you, as long as you consider the object. You never can fi nd it, till you turn your refl exion into your own breast, and fi nd a sentiment of disapprobation, which arises in you, towards this action. . . . So that when you pronounce any action or charac-ter to be vicious, you mean nothing, but that from the constitution of your nature you have a feeling or sentiment of blame from the contemplation of it. Vice and virtue, therefore, may be compar’d to sounds, colours, heat and cold, which, according to modern philosophy, are not qualities in objects, but perceptions in the mind. 1

Hume thought this “discovery” about the nature of morality was a great advance in the study of human nature. It was the “discovery” (although Hume, of course, was not the fi rst to make it) of what I am calling the “subjectivity” of value.

Hume thought that not only values and colors are “subjective,” but also, most famously, causality itself. Given the way human minds work, we will inevitably come to believe in necessary causal connections between some of the things we experience. But nothing in the world corresponds to that belief. “Necessity is something, that exists in the mind, not in objects,” he said. 2 We think causal necessity is something “objective,” but it is really nothing but a “subjective” “determination of the mind.” 3 This famous treatment of the idea of causality can still serve today as our best model of “subjectivism.” Other more recent varieties can all be measured against it.

There are many different positive versions of the “subjectivist” idea. For Hume, in speaking of necessity there is really nothing to speak of except what is in your own mind. And, as he says, when you “pronounce” upon the value of something “you mean nothing, but that . . . you have [a certain] feeling or sentiment” toward it. Taken lit-erally, that implies that value judgments are really just statements of feeling. That particular idea is not essential to “subjectivism.” Another version says that if you say that something is vicious you are not stating that you have got a certain feeling but rather are simply expressing or giving vent to a feeling or attitude you have toward the thing. Your remark is like a cheer or a sigh and is therefore neither

1 David Hume, A Treatise of Human Nature , ed. L. A. Selby-Bigge (Oxford: Clarendon Press, 1978), pp. 468–469 .

2 Ibid. , p. 165. 3 Ibid.


true nor false. 4 Or you might be both stating facts about the action and expressing a feeling toward it. Another version says that you are reporting or expressing a feeling and also encouraging others to have that same feeling or attitude. 5 For some “subjectivists” feelings are not involved at all; in “pronouncing” upon the value of something you are recommending or prescribing it, not saying anything that is true or false of it. 6 A quite different kind of theory holds that when you say that something is vicious you are saying only that the thing is such as to produce certain feelings or experiences or desires in human beings of such-and-such kinds. Whether the thing does or would have such effects is a straightforward matter of fact. But for the “sub-jectivist” there is nothing evaluative in the facts such judgments state. There couldn’t be. They speak only of non-evaluative effects to be brought about in human beings by certain “objective,” non- evaluative states of affairs.

“Subjectivism” carries with it a certain view of moral discussion or disagreement. It cannot see it as a dispute as to how things are, or what is so. Those who dispute about whether it is better to do X or to do Y when it is not possible to do both do not dispute about any mat-ter of fact. Of course, they might disagree about certain facts as well, but the purely evaluative dispute is not factual. The disputants’ valu-ings or attitudes or feelings are opposed to each other, so that at most only one of them can prevail, but the one who does prevail cannot be said to be getting things right while the other is getting them wrong. The one who prevails gets, or gets more of, what he values. But their dispute, if it is evaluative, is not a dispute about whether the world is such that X is better than Y or that Y is better than X.

The theory obviously has great appeal. It is extremely widely believed, in one form or another. In fact, it can seem to be the only kind of account there could be, largely because it alone among all theories avoids what would otherwise be an apparently insoluble problem. If values were part of the “objective” world, what sort of thing could a value be? How could there be such a thing as an evalua-tive fact or state of affairs? We know that where a thing is, what

4 See, e.g., A. J. Ayer, Language, Truth, and Logic (New York: Dover Publications, 1946), chap. 6 .

5 See, e.g., C. L. Stevenson, Ethics and Language (New Haven: Yale University Press, 1958), chaps. 2, 4, 9 .

6 See, e.g., R. M. Hare, The Language of Morals (Oxford: Clarendon Press, 1952), chaps. 1, 12 .


shape it is, how much it weighs, even what color it is or how much it costs, even whether human beings want it or get pleasure from it — are all matters of “objective” fact. But how could there be an add-itional fact to the effect that the thing is good or bad, or better than something else? The unintelligibility or “queerness” of what values would apparently have to be if they were “objective” has been one of the strongest arguments for “subjectivism.” As befi ts a metaphysical theory, it is defended on what are really metaphysical grounds.

The theory also has its moral or political appeal. It seems to express something to which we attach positive value — the idea that nothing or nobody can push me around in matters of evaluation. There is no position from which one person’s values can be criticized as incorrect or misguided. Nor is a person’s choice of what to do or the best way to live constrained by some “objective” standard against which it can be measured. The thought that the world cannot force us to accept one set of values rather than another can be liberating. It does not necessarily make life easy. There are great differences and confl icts among people’s valuings, and social and political life is a matter of resolving those confl icts and reconciling opposed interests. But what calls for solution is the question of which is to prevail. Each oppos-ing interest must somehow be accommodated. All are there to be dealt with, and there are none that can be dismissed on the grounds that they are mistaken.

I have called what is common to all forms of “subjectivism” a meta-physical theory. It involves a conception of what the world is really like — a specifi c, determinate idea of the nature of “objective” real-ity. It is a world that lacks some of the things that most people appear to believe it contains. It is in that sense a more restricted world than what we seem to accept in everyday life. For Hume it contained no necessary causal connections between events, and no colors or sounds. No causal sentences or color sentences were true of the world. For the “subjectivist” about values no evaluative sentences are true of the world, even though we appear to say and think that some things are good, or are better than others. Evaluative thoughts or beliefs or attitudes are part of the world, but there is nothing in the world that makes those thoughts true or false. All such evaluative facts have been eliminated from the subjectivist conception of what the world is like.

Eliminating something from our conception of the world is in ordinary circumstances a familiar procedure. We could be said to be


doing it whenever we fi nd out that something we used to believe is not so. This happens every day in small matters and, over longer periods of time, cosmically. Great scientifi c breakthroughs are some-times needed to bring about an altered conception of the world. Other, smaller changes take less. But in every case those old ways of thinking are then abandoned.

With the metaphysical theory of “subjectivism” things are differ-ent. Human beings — even “subjectivists” — continue to talk about and appear to believe in those very things that the theory claims are not really part of the world. There is a sense in which they are not abandoned. We cannot help getting experiences of color and believ-ing on that basis that objects around us are colored. We do inevitably come to value certain things more than others. The “subjectivist” philosopher of human nature says that those things we inevitably perceive and come to believe in are not in fact to be found in the “objective” world. But any such theorist, being human, will inevit-ably get those very perceptions and beliefs that the theory says are only fi ctions and cannot be true. The “subjectivist” will inevitably believe that grass is green, for example, while also holding that no object in the world has any color. And he or she will regard a particu-lar murder as vicious or bad while also insisting that no value state-ments are true, that the viciousness or badness of something is nothing in the world.

This seems to require of “subjectivism” both detachment from and engagement with the very same experiences, ideas, and beliefs. We must stand apart from our color beliefs and our evaluations while also holding onto them. Given the force with which the world inevitably operates on us, this would seem to make refl ection on the austere, restricted reality of “subjectivism” at best unstable — a momentary grasp of what you take to be the way things really are, from which your humanity immediately rescues you, plunging you back into a rich world of colors and vice and virtue which refl ection had appar-ently revealed to be nothing but illusions generated only by your own constitution. No one has given more poignant expression to this plight, while remaining in the grip of both sides of it, than Hume. 7

But it is not just a matter of psychological instability, or oscilla-tion. It is a question of whether that restricted view of the fully

7 See especially Treatise , book 1, part 4, section vii.


“objective” world can even be reached. That is the question I want to ask. Can we coherently think of a world in which all our valuings are exposed as only “subjective”? Could we then continue to understand ourselves to be making any evaluations at all? I think neither defend-ers nor opponents of “subjectivism” have taken this question ser-iously enough.

We say how we think the world is by saying what we believe to be so. But as long as we simply specify how things are, or how we take them to be, we will never arrive at the view that I am calling “subject-ivism.” In fact, if we tried to specify all the things we believe, and we took that list to express our conception of what the world is like, what we believed would be incompatible with “subjectivism.” One thing I believe is that grass is green; another is that some acts are vicious murders, that the deliberate killing of a human being is a very bad thing. If I take these beliefs to express part of my concep-tion of the world, I will have to conclude that it is a fact, or part of the way the world is, that grass is green, that some acts are vicious murders, and that the deliberate killing of a human being is a very bad thing. My conception of the world will not then be “subjectivis-tic” about colors or values. So at the very least the “subjectivist” account of the world must not include the contents of any of those beliefs. In saying how things really are it must not mention the colors of things or their value.

But merely leaving such things out of one’s conception of the world is not enough in itself to express the “subjectivist” conception. To leave certain features out of my conception of the world is not necessarily to conceive of a world which lacks those features. I might concentrate for some reason on only certain aspects of things. For example, I might think only about the size of the objects in my house, without mentioning their location, where I got them, or how much they cost. But that does not mean that I think that only their size is real, that they do not really have any location, any origin, or any cost. Similarly, I might specify a huge number of physical facts about the movements of particles, the presence of certain forces in the world, and so on, without mentioning the colors of anything. But I do not thereby imply that I think things have no color. I simply say nothing about their color one way or the other. And if I say only that certain physical movements occurred and the effect was the death of a human being, I say nothing about the value of what went on, but I do not imply that it was not in fact a vicious murder, or that I believe it was


not. So merely stating some of one’s beliefs about the world without mentioning the colors of things, or their value, does not automat-ically make one a “subjectivist” about colors or about values. Leaving something out is not the same as saying that there is no such thing.

“Subjectivism” clearly needs the thought, then, that colors, or val-ues, or whatever is said to be purely “subjective,” are not part of the world. Rather than merely conceiving of a world without conceiving of colors or values, it must conceive of a world which lacks colors and values. It involves a claim of exclusiveness. The negative claim about what the world does not contain is as essential to “subjectiv-ism” as the positive claim about what the world is really like.

But “subjectivism” also requires the thought that people neverthe-less do have beliefs about, or experiences of, those very features which it holds are not part of reality. The point of calling the source of those beliefs or experiences merely “subjective” is that we only think things are that way, or we have experiences which we wrongly take to represent the way things are. Without that, there would be nothing to be a “subjectivist” about. The theory is a theory about human thoughts or beliefs or experiences. So it cannot deny that we have such thoughts and beliefs and experiences.

If all this is what “subjectivism” requires, how is it to be shown that “subjectivism” is true in a particular domain? How is it to be shown, for example, that there is nothing in the world corresponding to our beliefs about colors or about values — nothing to make them true or false? With a theory like Hume’s it can look easy. He thought that all that was available to us in perception of the world were momentary, independent atoms of sensory information. Anything we ever think about must somehow be constructible out of such meager data. The task of his science of man was to explain how we develop our elaborate conception of the world with so little informa-tion to go on. Given only such restricted data, various features of our own minds will obviously have to play a large role. To the extent that our own mental operations alone can explain the origin of ways of thinking that go beyond what is available in the minimal data, those ways of thinking will be seen to have a wholly “subjective” source. The world would not have to contain anything corresponding to those ways of thinking in order for them to arise quite naturally in us as they do.

This is a strategy that many “subjectivist” philosophers since Hume have made use of, and continue to make use of today. If you can


explain how people come to think or experience something without having to suppose that those thoughts or experiences represent any-thing that is so in the world that gives rise to them, you will have exposed the thoughts or experiences as “fi ctions” with a wholly or partly “subjective” source. “Objective” reality would therefore include no more than what is found to be essential for explaining everything that happens, including human beings’ getting the thoughts and beliefs and experiences we know they get. So to say that colors are not part of the world, or that nothing in the world as it is in itself has any value, would be to say that nothing like “Grass is green” or “That was a vicious murder” has to be taken to be true in order to explain why people come to think that grass is green or that an act was a vicious murder. No colors would need to be ascribed to anything in the world in order to explain people’s color perceptions and beliefs. And no values would need to be ascribed to anything in the world in order to explain why human beings value things as they do.

I call such explanations, if they are successful, “unmasking” explan-ations. They unmask or expose some of our beliefs or experiences as illusory in the sense of not actually representing the way things are in the world, even though it is perfectly understandable why we inevit-ably get such beliefs or experiences, given what we are like and the way the world works on us. Whatever we cannot help regarding as true in order to explain our thinking and experiencing what we do must be reckoned as part of the way the world is. Those indispensable beliefs about the world will not then have been exposed or unmasked by an explanation of their origins. On the contrary, they will have been vindicated. They will have been shown to represent things as they really are. But for all the rest, the world is not really the way they represent it as being.

This might be called an explanatory test or criterion of reality. The world as it is in itself amounts to all, but only all, those truths that are suffi cient to explain what is so. Anything that is not needed for that explanatory purpose is not to be reckoned as part of the way things are.

I have said that this is one possible route to “subjectivism” about values, or about colors. It seems to rely on a certain faith in the sim-plicity of the universe. It sees the world as highly effi cient and eco-nomical, as no richer than it needs to be for the explanatory purposes of science. I do not want to speculate about the origins of such a faith. Nor will I go into the details of any particular attempts to


establish “subjectivism” by an appeal to unmasking explanations. I can only say that it seems to me extremely implausible to think that they alone could do the job. They seem to work best, as in Hume’s case, when you have already arrived at a restricted conception of what the world really contains. But establishing “subjectivism” in a particular area is a matter of arriving at that appropriately restricted conception of the world in the fi rst place.

I want to turn away from all questions about how “subjectivism” about values might be established and look instead at what must be an essential ingredient in any form of the view, however it is arrived at. There must be some way of understanding the presence of what those unmasking explanations, if they were appealed to, would be supposed to explain or unmask. The “subjectivist” view of the world, for all its zeal in eliminating certain features we unrefl ectively seem to think are there, still must acknowledge as part of the world all those perceptions, beliefs, and attitudes of human beings which it claims have only a “subjective” source. And there is a question of how the presence of those perceptions, beliefs, and attitudes is to be understood.

The question arises as much for colors as it does for values, and it will be helpful to look at that case fi rst. To entertain the view that colors do not belong to the “objective” world, but are at best pro-jected onto, or falsely believed or perceived to be present in, a world that does not really contain them, we must ourselves attribute no color to anything (since we say there is none in the world) while never-theless believing that there are many perceptions of and beliefs about the colors of things. The question is whether we can do that. It obvi-ously depends on what perceiving colors or believing that things are colored amounts to, and on what it takes for us to understand that such psychological phenomena occur. If we are “subjectivists,” it cannot depend on our supposing that any of the contents of such perceptions or beliefs are actually true of the world they are about. Can we make sense of the perceptions or beliefs if we no longer make that assumption?

We can, of course, understand people to have many beliefs about, and perhaps even perceptions of, things which we ourselves know do not exist. I understand that people believe in, and sometimes describe themselves as seeing, for example, ghosts or angels. People also think about centaurs and golden mountains. There is no doubt that such psychological phenomena occur. The explanation traditionally


offered for our understanding of such facts relied on a simple com-positional theory of thought. The concept of a ghost or a centaur was said to be a complex idea. It represents nothing that exists in the world, but it is a compound made up of simple elements, some of which do indeed fi nd counterparts in the world. Our attributing thoughts or beliefs about non-existent things to others therefore does not require that we ourselves believe the world to be populated with the things those complex ideas represent. We can see how people come to think that way without our agreeing that the thoughts they have are true.

Even this theory does not completely sever our understanding of the thoughts of others from all our own beliefs about the way things are. The presence in our common world of objects like horses’ bodies or the heads of men (or other even simpler things) is what enables us to think about such things and to attribute thoughts with those contents to the minds of our fellow humans. But even if that theory is perfectly satisfactory for thoughts about centaurs or golden mountains — which I do not believe it is — it would be of little help in explaining how the “subjectivist” can understand the presence of thoughts and perceptions of color. Surely our idea of color cannot be built up out of simpler elements that are not themselves colors at all. Perhaps some particular colors or shades can be understood as mixtures or combinations of other colors or shades, but there are no “elements” which are not colors but which somehow could be com-bined in thought or experience to give us the idea of color in the fi rst place.

Particular shades of color have traditionally been thought to be so simple that we can all understand what it is to have a perception of them simply by having such perceptions. It has been suggested that we each understand in our own case what it is to have a perception of green, say, simply by perceiving green; we know what that is like. If we could each understand it in that way, we could perhaps then say that what others have when they have a perception of green is just the same as what we have. We know in that way what green is, so we know that the feature that others perceive when they perceive green is that same feature. And that is what we all ascribe to objects when we believe that they are green. This seems to involve no ascription of green to anything in the world and yet to acknowledge the presence of perceptions of green and beliefs about green things on the part of human beings constituted more or less as we are.


I do not fi nd this traditional theory plausible, for reasons I will only state and not develop. I believe we could never come to understand in that fi rst-person way what it is to have a perception of green. The theory says that having a perception of green, or perhaps several of them, is enough to teach us what having a perception of green is. But simply having perceptions of green could never be enough. There is no way of being directly acquainted with something, or simply gazing at or experiencing a particular item, and from the mere occurrence or presence of the thing somehow coming to under-stand it as a thing of a certain sort rather than of some other sort. That is what we must do if we are to understand something as a perception of green rather than, say, as a remarkable event, which it might also be. Nor is there any possibility, on the sole basis of “having” it, of under-standing that we have got the same sort of thing this time as we have had before. Every two things are the same in some respect or other, and also different in countless respects, so whether we have got the same kind of thing on a second occasion depends on which respects are relevant and which not. And that cannot be fi xed by an original item about which we understand nothing but which we merely “have.”

Some surroundings are needed to make a thought into the thought of a certain kind of thing, so some surroundings are needed to make a thought about a psychological occurrence into a thought about a perception of green, say, and not something else. But if in trying to supply the surroundings needed to ascribe perceptions of particular colors to perceivers we fi nd that we ourselves must also ascribe colors to some of the things we take them to be perceiving, we will have abandoned the “subjectivist” conception of reality. We will be con-ceiving of the world as containing colored things. The “subjectivist” thought must leave room in the world for perceptions of and beliefs about color, but the price of our understanding such things to be part of the world would be our also taking the world to contain colored things. Color perceptions and beliefs could not then be unmasked as illusory or as having no counterparts in the way things are. So “sub-jectivism” could not be established.

It might seem that that is not so, since there is at least one version of “subjectivism” on which it remains true that objects are colored. It says that what is ascribed to objects when we apply color words to them is a disposition to produce perceptions of color in appropri-ately placed perceivers in certain specifi ed conditions. Objects really do (or do not) have such dispositions. So on that view our beliefs


about the colors of things would indeed describe things as they are in the world. We would not be precluded from truly ascribing colors to objects.

But that dispositional theory does not really avoid the diffi culty. It explains what it is for an object to be colored in terms of perceptions of color, but it says nothing about what a perception of color is, or what it takes for us to understand that there are such things as per-ceptions of color. Nor does it explain what a perception of green in particular is a perception of. The greenness involved in perceptions of green — what makes them perceptions of green — cannot itself be equated with a disposition to produce perceptions of green, even if the greenness of objects can be explained that way.

That dispositional account of the greenness of an object makes essential use of an idea of green that cannot in turn be explained in that same dispositional way. So it must hold that there are percep-tions of green even though no objects in the world possess that fea-ture that they are perceptions of. And that is the same problem that faced other versions of “subjectivism” about colors. It must explain how we can understand particular perceptions to be perceptions of, say, green and not something else, while at the same time we hold that no objects in the world possess that feature that they are percep-tions of.

When we think about what actually happens in everyday life, it seems that we constantly do rely on the public accessibility of such states of affairs as the greenness of grass in ascribing perceptions of greenness to our fellow human beings. We attribute color to objects in the world as a condition of attributing particular contents to per-ceptions. If that is so, and inescapably so, we will not be able to think of the world in the way “subjectivism” requires.

We can now see, I hope, a parallel diffi culty for the “subjectivity” of values. This is where the threat of distortion or denial comes in. Those who think that a particular act was an act of murder and was vicious or wrong seem to have a certain thought about that act: they think it was wrong. Perhaps they think in general that the deliberate killing of a human being is a very bad thing. We can speak of such persons as having certain moral views or beliefs or opinions (in this case not very controversial). “Subjectivism” cannot deny that people have such views. It must insist on the fact. The question is how it can acknowledge and understand that fact while also holding that no such thoughts are ever true or false of the world.


I have said that there are many different positive theories of evalu-ative judgment which are all compatible with the negative “subjectiv-ist” thesis that values are not part of the “objective” world. One view is that the assertion of “That act was wrong” reports the presence of a certain feeling that the speaker has toward that act (for Hume, a “sentiment of disapprobation”). Another view holds that what is being said is that the act is such that all human beings of certain kinds would get a certain feeling toward it if they knew of it. Both these views see the so-called evaluative judgment as a factual asser-tion about actual or hypothetical feelings on the part of certain human beings. In that respect they are like the dispositional analysis of an object’s color.

One merit of all theories of this sort is that they preserve one strik-ing feature of our evaluative thought. They allow that our reactions to the world do involve genuine beliefs about the goodness or bad-ness of things. They see us as asserting what we take to be truths about the world. And there is very good reason for insisting that we think of our moral judgments as either true or false. Not only do we seem to believe them and assert them and try to support them by reasoning. Moral sentences can also be embedded in other sentences in what certainly looks like a purely truth-functional way.

For example, from the sentence “That act was wrong” and the sentence “If that act was wrong then whoever did it deserves to be punished” it follows logically that whoever did that act deserves to be punished. Any view which says that moral or other evaluative judgments are not assertions or are not, strictly speaking, true or false has great diffi culty in accounting for that logical implication. Take the extreme emotivist view which says that in uttering “That was wrong” I am not asserting anything but only expressing my own distaste or my disapprobation of the act in question. That view can really give no account of the validity of the inference at all. In saying “If that act was wrong then ——” my “If” does not signify that I am somehow hypothetically or conditionally expressing a feeling. There is no such thing as hypothetically expressing a feeling. Of course, I can say “If I feel such-and-such then ——” and then refl ect on what follows from my having a certain feeling, or what would be true if I had one. I can also draw conclusions from the supposition not that I have a certain feeling but that a certain feeling has been expressed. But the antecedent in all those reasonings would be a straightforward factual proposition about feelings or the expression


of feelings. They would not be mere expressions of feeling which are neither true nor false.

Other kinds of non-propositional theory are more complicated, but they all face similar diffi culties. Some hold that to make a moral judgment is not to say anything true or false but to prescribe a cer-tain course of action in the way that imperatives order or demand certain courses of action. But still there is a diffi culty about how one can hypothetically prescribe or recommend something. One can cer-tainly prescribe or order something that is hypothetical or condi-tional — “If you go out, shut the door after you.” But that is an order to do something if certain conditions are fulfi lled. The imper-ative does not appear as the antecedent of a conditional propos-ition. Moral judgments like “That was wrong,” it appears, do occur as the antecedents of conditional propositions, and inferences are validly drawn from them. But if they are prescriptions, it would seem that in such positions they serve to issue prescriptions only condi-tionally. And what could that be? It would not be entertaining the hypothesis that a certain prescription has been made or that a cer-tain course of action has been recommended. Those are both straightforward factual propositions which are either true or false. They can easily be embedded in other sentences. But how could a prescription itself be embedded in a conditional sentence? It seems that it would have to be something like a hypothetical issuing of a prescription. But there is no such thing.

Another type of view, perhaps closest to what Hume says about causal necessity, is that in making moral judgments we do take our-selves to be expressing beliefs which are either true or false, but we are deeply confused and mistaken. What we are really doing is pro-jecting something we feel when perceiving or thinking about an action onto that action itself and mistakenly supposing that it “object-ively” resides there. “The mind,” Hume says, “has a great propensity to spread itself on external objects, and to conjoin with them any internal impressions, which they occasion, and which always make their appearance at the same time that these objects discover them-selves to the senses.” 8 In making moral judgments we think we are ascribing moral characteristics to the acts we observe; we treat our moral views as if they were, so to speak, propositional, but in fact they are mere projections. We “gild” or “stain” the facts with our

8 Treatise , p. 167.


feelings, but all that is strictly true in what we say is the purely fac-tual, non-evaluative content to which something in the value-free world could correspond.

This kind of view seems to me to serve the interests of “subjectiv-ism” best. But so far it gives no account of what our making a moral judgment really amounts to. It does not explain what we are saying when we say or believe that a particular act was wrong. We are said to take something we feel and project it onto the world, believing it to be a property belonging to things that exist there. But how do we do that? We do not think that objects and events in the world actually have the very feelings that give rise to our own “pronouncements.” The most that could be said is that we ascribe to things in the world, not the feeling itself, but what the feeling is a feeling of — that very feature that we are aware of “in our own breast.” In the case of caus-ality Hume thinks we get what he calls an impression of necessity, and it is that very feature — necessity — that we ascribe to the con-nections between some of the events we observe. In the case of color it is, say, greenness that we perceive and then project. What is the corresponding feeling or impression in the moral case? What we think or judge is that the act was bad, or vicious, or wrong, so it would seem that it must be a feeling of badness, or vice, or wrongness. But what is such a feeling? The question for “subjectivism” is whether and how we can understand those particular feelings that it says either generate or are referred to in our moral judgments or opinions.

Hume calls the feelings in question sentiments of approbation or disapprobation. But what makes a feeling a feeling of disapprobation or disapproval? Not just any bad or negative feeling will count. To disapprove of something is to think it bad, to make an unfavorable evaluation of it. So a particular feeling will be a feeling of disap-proval only if it is generated by or suffused with the thought that the thing in question is bad. But that is precisely the evaluative thought that the theory is trying to account for. It must explain how we can think something is bad or wrong without itself attributing badness or wrongness or any other evaluative feature to anything.

This same diffi culty faces other versions of “subjectivism” in which feelings are said to play an essential role in moral judgments. If my moral judgment is a report that I have a certain feeling, or that all human beings would get a certain kind of feeling under certain con-ditions, the kind of feeling in question must be identifi ed before we can know what is being said. Not just any feeling will do. Hume says


that the feeling arising from virtue is “agreeable,” and the feeling of vice is “uneasy” or unpleasant, but in saying that an act is wrong, even if I am indeed saying something about how people do or would feel, I am not saying only that they would get unpleasant or disagree-able feelings from the act. They might get unpleasant feelings from something they eat, but that would not make what they eat bad, or vicious, or wrong. So we still need some explanation of what it is to think that something is bad, or vicious, and some account of how we can intelligibly attribute such thoughts or attitudes to people.

I do not mean to suggest that, as things actually are, there is any special diffi culty about our doing that. We often agree in our moral assessments of particular acts and in many of our more general evaluative opinions. We come to share values, when we do, by grow-ing up and living in a culture in which they are endorsed and acted on. We recognize the badness of certain acts, and we recognize that other people have beliefs or reactions that are appropriate to the badness of the acts we all observe. Their responses count as disap-proval because they involve the thought that the acts are bad — a thought which we know to apply truly to just such acts as these. Our ascriptions of evaluative attitudes or feelings to human beings go hand in hand with our ascriptions of value to things and actions in the world.

I need not share all those moral assessments that I can correctly attribute to others. I can recognize that others think that a certain sort of thing is bad even if I do not think the thing is bad, because I too can have that same thought about other things. I do not have to agree in each particular case, any more than I must agree with someone else’s judgment about the color of something in order to attribute a belief about or a perception of color to that person. Knowing that a blue light is shining on a white wall, I will know that a person looking at it sees blue and, if he doesn’t know about the light, that he also believes that the wall is blue. I know the belief is false, but I can attribute that belief to him. I can do that because of my own general competence in the language of color and my know-ledge of what colors things are in the environment. Similarly, if I do not agree with a person’s evaluative judgment, I can still correctly attribute it to him and understand what it is for him to hold that view, because of my own general competence in the language of evaluation and my know ledge of the evaluative features of the environment — what things are good or bad, better or worse than others.


The traditional theory of simple and complex ideas was a way of accounting for the possibility of false belief or of a lack of corres-pondence between people’s ideas and the world. But that theory seems no more plausible here as a way of understanding the possibil-ity of evaluative thought in general than it seemed in the case of color. Perhaps some particular evaluative concepts can be defi ned in terms of others, but surely we cannot expect all evaluative notions to be reduced to terms that are not evaluative at all. There are no simple non-evaluative “elements” which could somehow be combined in thought or experience to give us the idea of value, and hence the pos-sibility of evaluation, in the fi rst place. This irreducibility is one of the few things on which most modern moral philosophers would seem to agree.

There is no question that we do make moral judgments or evaluate things or states of affairs, and that we do attribute such judgments, or reactions involving such judgments, to others. The question is not whether we all do it in real life. The question is whether someone who consistently holds to the “subjectivity” of all values could do it. Could someone make sense of the idea of there being feelings or attitudes of disapproval, say, if that person did not also hold the view that certain kinds of acts are bad, or wrong, or worthy of disap-proval? What made it seem possible in the case of color was the thought that perceptions could somehow be directly recognized as intrinsically of a certain specifi c kind — that we can simply read off from our perceptions themselves what features they are perceptions of — whatever we take the world to be like, whether we think it con-tains any colored things or not. I think there is a tendency to rely on a similar thought in the case of values. We are thought to be able to recognize what we feel simply by feeling it, by being aware of some felt feature in our experience, whatever we take the world to be like, whether we attribute any negative or positive value to anything in the world or not.

I have already suggested why I think that sort of view could not be right even about perceptions of color. I do not think its prospects are any better in the case of values. Even supposing that we could isolate in our experience some feeling or attitude or response which plays an essential role in moral or evaluative judgment, there would still be the problem of what that feeling could be said to be like. It would have to be identifi ed and classifi ed only in terms that are somehow immedi-ately available to consciousness, not in terms of any evaluative


judgments that defi ne or accompany it. It would have to be the kind of feeling or response that a person could have without having any moral or evaluative opinions at all.

This would have the consequence that the only materials available to us for understanding what appears to be evaluative thought and for seeing how it fi gures in human action and human social arrangements would be simple, isolated feelings with no evaluative content. They might be such things as pleasant or unpleasant sensations, feelings of satisfaction or dissatisfaction, or simple likes or dislikes. Or, moving away from feelings, they might be such things as basic unmotivated desires or wants or preferences, or even more indiscriminately, those all-purpose motivators called “pro-attitude” and “con-attitude.”

Even such apparently scaled-down materials are not necessarily on the non-evaluative ground fl oor. One can feel pleased that justice has been done, for example, and if that is a feeling of pleasure it is still not independent of its evaluative content. Someone who did not think that justice had been done could not have such a feeling. And liking something or somebody can be a matter of thinking well of the thing or person, and that again has an essential evaluative component. Even wanting, or preferring, or simply being for or against a certain thing can also be an evaluative attitude or state. Preferring that virtue be rewarded, for example, or being for a just solution, or being against the unjust acts of one’s government — these are moral attitudes and not simple motivating feelings or wants that might rise up in a non-evaluating agent. It is not clear to what extent there could even be such a thing as a non-evaluating agent — at least, a human agent.

I have suggested that it is the thought that values and evaluation would not otherwise be intelligible that can lead to the idea that they must ultimately be explained only in terms of likes and dislikes, pleas-ures and pains, or basic human desires. Perhaps something like that is what fi nds expression in the popular half-thought that morality is after all just a matter of what people want, or what they like or don’t like. Or worse still, the thought that it is just a matter of whose likes and dislikes are going to prevail. And now there is the view, in the United States at least, that morality is itself just one among a great many “special interests” that have to be accommodated in society. People are thought to be just pushing their own personal interests or seeking their own “gratifi cations” in one way or another, and the “morality lobby” is encouraged to fi ght it out with the military, the corporations, the doctors, the judges, and so on.


Whether such views are really derived from the position I am call-ing “subjectivism” I don’t know. But it does seem to me that to hope that the feelings or attitudes essential to evaluative judgment can be identifi ed and understood neutrally, on the basis of some intrin-sically felt quality alone, would be disastrous for making sense of what is, after all, a fundamental aspect of human life. If it is only our very engagement in a set of values that makes it possible for us even to recognize the phenomenon of evaluation, the demand of evaluative neutrality would have the effect of denying or obliterating the very phenomenon we want to understand. If engagement or participation is essential, we can never get ourselves into a position to discover that all values are “subjective,” that the goodness or badness of some-thing is not part of the way things are.

II

I have been trying to identify some of the diffi culties in “subjectiv-ism” and to draw attention to what looks like a serious obstacle to our ever even arriving at that conception of the world. If we tried to adhere strictly to what “subjectivism” requires of us, it seems that we could not consistently or coherently come to think that it is true of our evaluative thought. But even if that is so it might not seem like much of a threat. Why can’t we simply abandon “subjectivism” and look elsewhere for a more satisfactory understanding of evaluative thought and its role in our lives? I now want to suggest why I think that will not be easy.

“Subjectivism” in one form or another appears to be the inevitable result of our trying to understand ourselves in a certain way, and the goal of understanding ourselves that way is not easily abandoned. It seems impossible to conceive of a better way of understanding how human beings work. The model goes back to one of the greatest achievements of the Enlightenment — the idea of a “science of human nature.” It was to be a thorough, systematic investigation of the principles of human nature that would eventually explain every aspect of the personal, social, cultural, and political life of human beings. That the proper aims of individual human beings and the best social and political arrangements among them should somehow be determined in the light of truths about their nature and the world they live in was not in itself a new idea. What was unique to the Enlightenment was an open-ended curiosity about what that human


nature is really like and, most important, an empirical, secular idea of how it is to be known.

If human beings were endowed by a supernatural power with cer-tain capacities and goals and were placed in a specially ordained position in the world, that would obviously limit both the content and the justifi cation they could fi nd for the beliefs they arrived at and the goals they aspired to. The source of any knowledge they acquired or any values they pursued would then lie in something other than those human beings themselves and the familiar world they could know they inhabit. Their behavior in pursuit of that knowledge and those values would therefore not be fully understandable to them by human intellectual means alone. There could be no properly scien-tifi c understanding of human behavior or human life.

The impressive growth in understanding the inanimate physical world, culminating in the work of Isaac Newton, was the decisive step toward the downfall of that traditional picture. The mathematical sci-ence of nature was seen to rest on no supernatural hypotheses and to proceed carefully no further than experience, and solid reasoning based on that experience, could take it. It does not matter now whether that really was a correct perception of the revolution in physical sci-ence or not. In any case it served as the source of the Enlightenment ideal of understanding what we might call the human world through the application of just such a broadly scientifi c enterprise to the study of human nature and human life. It put human beings at the center, and it insisted that the way things should be in any human world must be based only on what human beings can reliably fi nd out about them-selves and the natural observable world they live in.

One of the earliest and still one of the best statements of this goal is Hume’s A Treatise of Human Nature . It was written in the belief that “all the sciences have a relation, greater or less, to human nature. . . . Even Mathematics, Natural Philosophy, and Natural Religion , are in some measure dependent on the science of MAN ; since they lie under the cognizance of men, and are judg’d of by their powers and faculties.” 9 The aim of the book was to lay the foundations of a genuine science of man that was to be like nothing that had gone before.

Hume had found that the “moral Philosophy transmitted to us by Antiquity,” like ancient “Natural Philosophy,” was “entirely

9 Treatise , p. xv.


Hypothetical, & depend[ed] more upon Invention than Experience. Every one consulted his Fancy in erecting Schemes of Virtue & of Happiness, without regarding human Nature, upon which every moral Conclusion must depend.” 10 His plan was to appeal to noth-ing but experience—to “introduce the experimental method of rea-soning into moral subjects.” He would take human beings as they are, he would observe them “in the common course of the world, . . . in company, in affairs, and in their pleasures,” 11 and he would try even-tually to explain all the overt and hidden richness of their behavior, thought, and feeling. That would require the discovery of general principles of human nature, but the only possible access to such principles and explanations would be judicious generalization from whatever we can fi nd out by “the cautious observation of human life.” 12 Where our best experience remains still silent on some ques-tion of human nature or human destiny we must willingly confess our ignorance and perhaps try harder to discover how we work and what the world holds in store for us. But we must not let the natural human desire for some answer or other lead us to invent comforting stories for which we can honestly fi nd no support in experience, or impose such conjectures or hypotheses on the world and then base our beliefs and behavior on such creatures of the imagination.

The positive project of a naturalistic study of human beings is familiar to us today in what we call the social sciences. They are such a pervasive and powerful feature of the modern world that it is some-times diffi cult to remember that they have not always been with us. Ideally, they promise a dispassionate, scientifi c understanding of thought, feeling, and behavior in every area of human life. The infor-mation they would provide is to be the basis of all personal, social, and political organization and improvement, just as the physical sci-ences provide the facts and theories that engineering and technology then make such spectacular use of in the purely inanimate domain. By now we are used to a division of the study of human behavior into different, highly specialized fi elds such as economics, psych-ology, and sociology. Hume probably never envisaged the technical professionalism of today. But the idea that human beings can be

10 The Letters of David Hume , vol. 1, ed. J. Y. T. Grieg (Oxford: Clarendon Press, 1969), p. 16.

11 Treatise , p. xix. 12 Ibid.


studied and understood in this way is an Enlightenment idea. It is just what he had in mind.

Hume’s enthusiasm for the idea went beyond the social benefi ts to what he saw as the directly cognitive or intellectual pay-off of a sci-ence of man. Since all sciences — even mathematics and physics — fall under the cognizance of men, he thought it was “impossible to tell what changes and improvements we might make in these sciences were we thoroughly acquainted with the extent and force of human understanding, and cou’d explain the nature of the ideas we employ, and of the operations we perform in our reasonings.” 13 These hopes have scarcely been vindicated. It is diffi cult to point to concrete advances in the sciences that have been generated by Hume’s own epistemological theory of our ideas and reasonings, and presumably few today would think of looking to psychology, say, or sociology, as a possible source of breakthroughs in mathematics or physics. But even without the hope of such direct scientifi c consequences, the Enlightenment idea of the study of human nature still serves to determine our culture’s conception of understanding ourselves. It remains diffi cult to imagine any other way of getting the kind of understanding of ourselves that we seek.

The goal is an understanding of human nature. The method is to study human beings in interaction with the world and thereby to explain how they come to think, feel, and act as they do. But not just any story — even any true story — about the relation between human beings and the world will give us what we want. For instance, we are interested in how people come to believe and know what they do about the world around them, but we would not be satisfi ed with the obvious truth that they learn to think and speak and they come to know things about the world by seeing and touching things and in other ways perceiving what is true of the things around them. That is all true, but it does not explain what we want to understand. Similarly, we are interested in how people come to value things and why they endorse the particular values that they do, but we would not be satis-fi ed with the obvious truth that they grow up and are socialized into a particular culture and come to accept many of the value beliefs cur-rent in that culture. That again is certainly true. But such general truths about human beings and human life do not explain what we want the philosophical study of human nature to explain.

13 Treatise , p. xv.


It is not just a question of detail. Even a full, detailed story of how a particular person or a particular group comes to know or to value a certain sort of thing would not satisfy us. We want to understand certain pervasive or fundamental aspects of human life in general .

Morality, for example, is a quite general phenomenon which seems distinctive of the human species. No other creatures seem moved by considerations of good and bad, right and wrong. To understand that aspect of human life, then, would be to understand how there comes to be such a thing as morality at all, how it works, and what makes it possible. Our possession of an elaborate conception of the world we live in and its history, how it works, and how it affects us, is also something a science of human nature should be able to explain. We want to know how there comes to be any such thing at all. The same is true of our beliefs about the colors of things in particular, or all our beliefs to the effect that certain things are causally connected with other things. The question in each case is how we come to have any thoughts or beliefs or responses of that general kind at all.

There is implicit in this kind of question a certain idea of how best to answer it. It seems to be the only way in which it could be answered. Obviously, if human beings come to act or think or feel a certain way only after interaction with the world around them, there must be something about human beings — something about what we are like — and something about what the world is like which combine to produce in human beings the way of acting or thinking or feeling in question. Even if facts of the world affected us directly by simply impressing themselves on our minds whenever we opened our eyes or ears, there would still be something true of us that was partly respon-sible for our getting all the beliefs we get. It would be because we are capable of passively receiving information about the world in that way. So it seems that any explanation of distinctive, pervasive fea-tures of human life would have to be a two-part explanation. It would involve an “objective” factor — what the world is like, or how things are independent of us — and a “subjective” factor — what we are like, or how things are with us.

If both factors are always present, if human beings themselves always play some role in acquiring their conception of how things are, then the study of human nature will naturally take the form of asking how much, and what, the human subject does contribute. That would be to isolate and identify those elements of human nature that are responsible for our conceiving of and responding to the


world in the ways we do. How much of what we think or feel about the world is due to us, to the way we are, to the “subjective” factor, and how much is due to the way things are independently of us, to the “objective” factor? The intellectual goal expressed in terms of this bipartite conception serves as our model for understanding our-selves, for seeing how we work, for identifying what is distinctively human.

Hume’s Enlightenment project of a “science of human nature” embodied just this conception of how to understand ourselves. It was the search for, among other things, “principles of human nature” — those features of human minds and sensibilities that are responsible for our thinking, feeling, and acting as we do. Hume sought those “principles” which he called “permanent, irresistable, and universal.” They are the “foundation” of all our thoughts, feelings, and actions in the sense that, as he put it, “upon their removal human nature must immediately perish and go to ruin.” 14 They will be those “prin-ciples” that are at work in anything recognizable as human life as we know it — ways of thinking and acting without which nothing would think or act as humans do. They are to be identifi ed by asking what must be true of human beings as we know them in order for them to think and feel and act in the ways we know they do.

This Enlightenment project of isolating and identifying the “sub-jective” factor in human thought and experience is by no means only a thing of the past. We fi nd it in any philosopher who would distin-guish in general between the “given” and the “interpretation,” between the “data” we receive and the construction we put upon them, or between the “fl ux of experience” and the “conceptual scheme” we impose upon it to make sense of our experience and to learn from it. In our own day, for example, W. V. Quine in his Word and Object has put his task almost mathematically as follows: “we can investigate the world, and man as a part of it, and thus fi nd out what cues he could have of what goes on around him. Subtracting his cues from his world view, we get man’s net contribution as the difference. This difference marks the extent of man’s conceptual sovereignty.” 15

It is not just a matter of how the world affects us. We know that something outside us acts on us; our cognitive and affective responses

14 Treatise , p. 225. 15 W. V. Quine, Word and Object (Cambridge, Mass.: Technology Press of MIT and

John Wiley & Sons, 1960), p. 5 .


are caused by something in the world. So in that sense we know that there is something or other in the world which, in conjunction with facts about us, makes us think and feel as we do. But it is not merely a question of causation. We are interested in a more complex relation between our beliefs and responses and the world. We want to know the nature of whatever causes there are. Is there anything in the world that not only causes but somehow matches up with, or corresponds to, or is adequately represented by, those things that we think and feel? Is the world, or does the world have to be, anything like the way we think it is, in order for it to have given us the thoughts and feelings about it that we have? What is there, if anything, that renders those thoughts or responses true? Or are they merely “subjective” responses with nothing corresponding to them in the “objective” world? And with these questions we have arrived at the inquiry that I have already suggested seems so easily to lead to “subjectivism” with respect to many of the things we believe.

It leads easily to skepticism about the world too. Descartes’ evil demon represented the threat that the way things are independently of me might be extremely different from the way I take them to be. If the demon exists, he alone exists beyond me, and his clever machin-ations make me think that I live in a world of earth and water, trees and buildings, and other people with human bodies like the one I think I’ve got. He gives me such thoughts and beliefs, so they are produced by something “objective” and independent of me, but there would be almost nothing in that world corresponding to any of those thoughts. They would almost all be false. The challenge of skepticism is to show how I know that I do not live in such a world. Once we see human knowledge as a combination of an “objective” and a “subject-ive” factor in this way, and we acknowledge the possibility of a largely or even entirely “subjective” source for most of our beliefs, it seems impossible to explain how those beliefs could ever amount to know-ledge or reasonable belief.

But our concern here is not the epistemological question of know-ledge or the reasonableness of our beliefs. The “subjectivism” I have been describing results from a metaphysical project which relies on some unquestioned knowledge of the world. It subtracts from some of our beliefs, not the causes that produce them, and not necessarily our warrant for accepting them, but rather their correspondence with anything that holds in the way things are independently of us. Whatever support we might have for them therefore cannot be


“objective.” It cannot be based in anything that is to be found in the way things are. The same bipartite conception of the ultimate source of all our beliefs and responses is at work in this project. Everything is to be assigned either to an “objective” or to a “subjective” source.

It is because we want to understand human values or human valu-ing in general in terms of this bipartite picture that I think we are inevitably driven toward “subjectivism.” To understand the general phenomenon of human evaluation we want to understand the con-tent of such evaluative thoughts — what exactly is being said or thought when someone says, “That was a vicious act” or “It is better to comfort someone than to kill him.” To ask what is involved in such a thought is to ask what, if anything, could make it true, or false. If we answer, simply, that it is the goodness or badness of things that makes such remarks true or false, we feel we are not really explaining the content of those evaluative “pronouncements.” In saying that it is true that an act is vicious if and only if the act is indeed vicious we are making use of the idea of viciousness, but we are not explaining it. We feel we will understand evaluation only when we know, as it is often put in philosophy, what it is to be vicious, or bad, or wrong.

When Hume investigated that fundamental feature of human life which he called our “reasoning from causes to effects,” our getting beliefs about the necessarily causal connections between things, he did not restrict himself simply to identifying the circumstances in which we get such beliefs. If he had, he could have said that a belief that two things are causally connected arises in us whenever we are presented with instances in which one thing causes another. Even if that were true, it would not help us understand human thought about causality in general because it does not explain what it is for one thing to cause another. It was because Hume wanted to explain not just the origin but also the contents of our causal beliefs that he was able to fi nd nothing in the world corresponding to their special claim of necessity. Necessity, then, could be only something “subjective,” or nothing at all. It was the very desire for a completely general account that would explain how any thought about necessity is pos-sible at all that led to “subjectivism” about necessity.

If the contents of our evaluative beliefs are going to be explained in terms of what is or could be the case in the world, it does seem that we will have explained evaluative thought in general in that way only if whatever we fi nd in the “objective” world to explain it is something non-evaluative. Otherwise, there will still be some evaluative content


that will not have been explained. And since the non-evaluative always falls short of exhausting the special, apparently evaluative, content of our thought, the source of that special evaluative element will inevitably have to be located in us, not in things as they are independently of us. It will be assigned to the “subjective” and not the “objective” factor. All our value beliefs, then, insofar as they are really evaluative, will be “subjective” and will not assert anything that is or even could be true of the way things are.

In the fi rst lecture I tried to cast doubt on our ability to carry out this metaphysical project of conceiving of a world in which we have genuinely evaluative beliefs or attitudes while holding that none of them is true of the world. It requires detachment from, or non-endorsement of, the contents of all our evaluations. And it requires an acknowledgment of the fact that we nevertheless do make such evaluations. But the fulfi llment of either one of those two requirements threatens the possibility of fulfi lling the other. Without ascribing value to things in the world, and hence holding evaluative beliefs of our own that we take to be true, it is diffi cult to see how we could interpret and hence understand other people, as well as ourselves, as holding any evaluative beliefs at all. We would not have what it takes to see the world as containing genuine evaluations — any thoughts to the effect that something is good or bad, or better or worse than something else. If we do fi nd thoughts in the world which we understand to be genuinely evaluative, then it seems that we must already hold certain evaluative beliefs or opinions of our own. But then our disen gagement from all values would have been abandoned. We would be t aking certain value sentences as true.

I did not try to prove any of that. I said what I could to make it plausible. Without offering further argument for it here, I want to examine some of its consequences. Suppose, as I have been suggest-ing, that there is simply no understanding of evaluation in completely non-evaluative terms. It has long been accepted that there is no hope of strictly defi ning evaluative notions somehow in purely non- evaluative terms. That is, perhaps, the real lesson of G. E. Moore’s misnamed “naturalistic fallacy,” which preoccupied moral philoso-phers for so long. 16 I am suggesting, not an obstacle to defi nition of the contents of our evaluations, but rather an obstacle to our even

16 G. E. Moore, Principia Ethica (Cambridge: Cambridge University Press, 1956), chap. 1 .


understanding or acknowledging the phenomenon of human evaluation at all. Evaluation is a fact of human life — something that human beings do — which it seems we cannot acknowledge without our also engaging in the practice ourselves. And that means that we cannot even understand that it is going on without our being prepared to take certain evaluative beliefs or “pronouncements” to be true.

Just suppose for a moment that that is right. Now if it is also right (as I think it is) that accepting “subjectivism” about values would have to involve our disengaging from all values while still making sense of evaluative thought, then we would not really be able to accept a “subjectivist” picture of human values. We could not coher-ently get ourselves into the position of discovering that none of the evaluative beliefs or attitudes of human beings corresponds to any-thing that is so in the way things are. But then, if “subjectivism” is the inevitable outcome of trying to understand human values in terms of that traditional bipartite conception of human beings and their rela-tion to the world, it would follow that we cannot really carry out the Enlightenment project of determining the “objectivity” or “subject-ivity” of values in general. The detachment or disengagement we would need would rule out the very understanding that we seek. That would be disturbing, and dissatisfying, given the natural appeal of that picture. It seems like a perfectly comprehensible intellectual goal — in fact, the very model of what it would be to understand general aspects of human nature in the right way. But it would be unattain-able. We could never fully understand ourselves in that way.

This is not to say that we could not study the phenomenon of human evaluation, and indeed human values themselves, and learn much more than we now do. There is a great deal that we do not know and should be trying to fi nd out, not only about what things are good and bad, and why, but also about how people acquire the values they do. How does it happen that an infant who comes into the world with needs and impulses and a native set of behavior pat-terns comes by the time it is an adult to possess a complex set of evaluative beliefs and responses? How do adults with fi rmly held evaluative opinions about certain matters come over the course of time to change them? And how can we arrange things in society so that people on the whole make such changes in the direction of more informed, more considerate — in a word better — evaluative atti-tudes? There is no answering such questions in the abstract. It obvi-ously depends on the particular people and the particular culture and


on countless other factors in ways we still do not understand very well. But such questions, however complex, can be answered — or at least progress can be made.

Any study of human socialization or human development along these lines would be a study of how a human being or a group of human beings gets absorbed into a culture whose members already have some values or other, or how the possession of one set of values gets transformed into possession of another. It would explain at most the transmission of values, perhaps even the transmission of the very idea of value, from those who have it to those who do not. But to explain how something is transmitted or changed is not necessarily to explain what it is that is transmitted or changed. It is not necessar-ily to explain what those attitudes are, or what it is to hold them. What is it to think that deliberate killing of a human being is a very bad thing, or to think anything evaluative about anything? This is a philosophical question about human valuing as such.

If that is what we want to understand about ourselves or about human nature in general, then the metaphysical project I have been describing seems inevitably to come into play. This is what leads us to “subjectivism.” We want to understand the nature of any evaluative thoughts or attitudes. We ask what their special content is, what is really being thought. And that fi rst takes the form of asking what would be so if they were true. In trying to answer that question, either we merely repeat the thought — “Killing a human being is bad” is true if and only if killing a human being is bad — and so we do not feel we are explaining it, or we try to express its content in other terms that reveal in some illuminating way what is really being said. If those further terms are still evaluative, we will not feel that we have explained what it is for any evaluative thought or attitude to be true; we will simply have exchanged one such thought for another.

So if we are going to make any progress in explaining the evalu-ative as such, we will either say that having what we call an evaluative attitude or opinion is not really a matter of thinking something to be true — but instead is expressing a feeling or issuing a prescription or making a recommendation or some such thing — or we will say that it is a matter of our thinking true something that is really non- evaluative and so could hold in the “objective” world — perhaps something about non-evaluative feelings that we and others do or would feel under certain conditions. Each of these alternatives is a ver-sion of what I am calling “subjectivism”: there are no “ objective” evaluative


facts or states of affairs. And each of these alternatives appears to deny or obliterate the very phenomenon we set out to understand. That looks like the inevitable outcome of our trying to understand human evaluation in general.

It is this very desire to explain human evaluation in general that seems to preclude us from invoking unexplained evaluative truths or states of affairs in any account of the special contents of human evaluations. Thus do we inevitably banish genuinely evaluative facts from any world in which we can make sense of what seems to be evaluation. But in trying consistently to adhere to nothing more than that shrunken conception of what is really so we would fail to make sense of the idea that human beings have such things as evaluative thoughts, attitudes, or responses. We would lose those very attitudes that “subjectivism” about values claims have nothing corresponding to them in reality but are nothing more than our “subjective” responses to an “objectively” value-free world.

I would draw here on the parallel I see with the case of colors. If we did not make categorical ascriptions of colors to things around us we could not acknowledge the existence of such things as perceptions of colors or beliefs about the colors of things on the part of human beings. We could not conceive of the world as containing those very perceptions and beliefs that “subjectivism” about colors claims have nothing corresponding to them in reality and are nothing more than our “subjective” responses to an “objectively” colorless world.

I want to say more about what this idea amounts to and exactly what it implies about our understanding of values and colors, and what it does not. It says in its strongest form that we cannot think of a world in which people perceive particular colors or believe that things are colored without ourselves being prepared to ascribe color categorically to things in the world. We cannot understand human beings to have evaluative opinions or attitudes to the effect that such-and-such is good or bad without ourselves sometimes recognizing the goodness or badness of certain things. And in making those ascriptions of color, or of value, we are taking certain things to be true. We take it to be part of the way things are, for example, that grass is green, or that the deliberate killing of a human being is a very bad thing. Our engagement with, or endorsement of, aspects of the world of those general types is required for our ascribing to human beings beliefs or attitudes with those types of contents. What we take


to be facts of the world are implicated in our making sense of thoughts of the world. The two cannot be pried apart completely.

I am here endorsing particular instances, having to do with colors and with evaluation, of what appears to be a quite general fact about our understanding one another in the ways we do. There are condi-tions of the successful ascription to human beings of beliefs, atti-tudes, perceptions, or feelings with specifi c contents. This is something that I believe lies at the heart of Wittgenstein’s later work. Donald Davidson has stressed its importance in what he calls “radical interpretation.” 17

If we are to interpret someone as believing or perceiving or feeling some particular thing or as having a certain specifi c attitude, we must somehow connect those specifi c psychological states we are attribut-ing to that person with facts or events or states of affairs in the world that we take them to be about. If we ourselves had no opinions about what is so and what other people are most likely to be attending to in the environment, we would be in no position to attribute any beliefs or perceptions or attitudes to them at all. We interpreters and ascrib-ers of beliefs and other psychological states must therefore be engaged in the world and take certain things to be true of it if we are ever going to attribute psychological states to anyone. And we have no choice but to ascribe to others, at least in general, beliefs in and per-ceptions of and attitudes toward some of the very things we our-selves take to be true of the world. We cannot make sense of other people as believing something we know to be obviously false unless we have some explanation in the particular case of how they come to get it wrong. And that explanation will work only if we understand them to share in common with us other beliefs and attitudes in the midst of which their particular, localized error (as we see it) can be made intelligible.

This still leaves considerable room for difference or disagreement. Whole areas of belief or perception might be found to diverge if there remains enough overlap to serve as the shared base of what the interpreter could then see as the others’ deviance. But it seems to me, as I have been suggesting, that the deviance could not go as far as the interpreter’s fi nding that others had perceptions of color and beliefs

17 See, e.g., Donald Davidson, Inquiries into Truth and Interpretation (Oxford: Clarendon Press, 1984), essays 9–12 .


about the colors of things which he held did not agree with anything at all that he took to be true in the world. Nor could he fi nd that they had evaluative beliefs or attitudes about the goodness or badness of things none of which he shared because he had no evaluative beliefs or attitudes at all.

Davidson has sometimes drawn from his main claim about inter-pretation the conclusion that most of our beliefs must therefore be true. 18 They must be, if we can even understand the fact that we and other people have any beliefs at all. And this seems to imply that the truth of the majority of our beliefs is a necessary condition of our having them — that if we have any beliefs or attitudes at all, the list of sentences which state the contents of those beliefs or attitudes will contain mostly truths. That would connect our beliefs necessarily with the way the world is. If I am right to apply this thesis about interpretation to color beliefs in particular, that would imply that most of our color beliefs are true. It would mean that we are getting the colors of things, on the whole, right. We can’t help it. And applied to our evaluations it would mean that, in general, our beliefs about what is good and bad, better and worse, are true. On the whole, the things that we think are good, or bad, really are good, or bad. Of course, in the case of values there appears to be much less widespread agreement, so it is not easy to speak without further qualifi cation of what “we” believe, what “our” evaluative beliefs are. But despite that apparent lack of agreement, there must still be enough common ground somewhere among all those who hold evaluative beliefs to make possible the ascription of such beliefs to them. That common core, or at least the major part of it, is then said, on Davidson’s view, to be true.

This is not a conclusion I wish to draw — at least not if it is taken as a defense of the “objectivity” of colors or values as opposed to their alleged “subjectivity.” If we knew, by this kind of transcenden-tal argument, that most of our beliefs had to be true, and in particu-lar that most of our color beliefs are true, and that most of our evaluations are true, we could easily be led to ask, in the spirit of the traditional metaphysical project, how that could possibly be so. How could it be that our believing what we do requires that the world should be a certain way? This will make us look once again at the

18 Donald Davidson, “A Coherence Theory of Truth and Knowledge,” in Kant oder Hegel , ed. D. Henrich (Stuttgart: Klett-Cotta, 1983) .


contents of those beliefs. What exactly are we believing when we believe that, say, grass is green, or that the deliberate killing of a human being is a bad thing? And what is it to believe such things? We want to understand such beliefs in general. Just as that quest drove us toward “subjectivism” earlier, so it would drive us toward “subjectiv-ism” again. There seems to be no other way to account for the neces-sary connection that would have been proven to hold between the body of our beliefs as a whole and their truth.

This is just the position of Kant’s Critique of Pure Reason , perhaps the greatest attempt there has ever been to prove that the truth about the way the world is cannot come apart in general from our thinking and perceiving in the ways we do. Kant thought there were necessary conditions of the possibility of all thought and experience, and that not all those necessary conditions are themselves just further thoughts or beliefs. They include as well many non-psychological truths, so not only must we think a certain way, but the world independent of us must be a certain way, in general, if we are even able to think of or perceive anything at all.

Kant saw that some philosophical theory was needed to explain this necessary link between thought and experience and the world, and his explanation was the theory of transcendental idealism. It was the only explanation he thought there could be. We can know that the world in general must conform to our thinking and perceiv-ing in certain ways because our being able to think and perceive what we do actually “constitutes” the world that we perceive and believe in. The price of showing that our thoughts and perceptions and the truth of their contents cannot come apart in general was that the truth of what we believe about the world somehow consists in our having the kinds of thoughts and perceptions that we do. The world turns out to be dependent on our thoughts and perceptions in some way after all. That is a form of idealism, which is one variety of what I am calling “subjectivism.”

If we ask in a similar vein how our color beliefs, or our evaluative beliefs, could not fail to be true, a more particular version of that same idealism or “subjectivism” will seem like the only possible answer. What it is for color judgments to be true, for there to be a world of colored objects, it would say, is just for human beings to agree for the most part in their ascription of colors to things. There might be considerable disagreement in particular cases, but on the whole there would be nothing more to things’ being colored than


human perceivers agreeing in general in the perceptions they have and the judgments they make about the colors of things. Similarly, for things to have value, and to have the particular values they have, would simply be for human beings to agree in general in their ascrip-tions of value to things. Again, there is room for wide disagreement and uncertainty, but on the whole the truth of value judgments would amount to nothing more than agreement, or the possibility of agree-ment, in human beings’ evaluative beliefs.

This is clearly just “subjectivism” approached from a different direction. In terms of the traditional dichotomy, it locates the source of the truth of color judgments and of value judgments on the “sub-jective” side. Or rather, like all forms of idealism, it in effect collapses what was originally thought of as the “objective” into the “subjec-tive.” The facts of the “objective” world that make our color judg-ments or our value judgments true would be facts only about us, about what we say and do, and not about an independent world that we say those things about. The only form of so-called “objectivity” granted to those judgments would be intersubjective agreement. And understood in terms of the traditional dichotomy that is not really “objectivity” at all.

This is precisely the kind of view of the world that I have been sug-gesting we can never reach. Describing it this way perhaps brings out why. We cannot make sense of the idea that the truth of judgments of a certain kind amounts to nothing more than human agreement with respect to the contents of those very judgments. If human beings agree in certain judgments, there must be something they agree about. We must be able to make independent sense of their making such judgments in the fi rst place if we are to fi nd anything for them to be in agreement about. That is the point of Davidson’s requirements on interpretation. The content of the judgment must be identifi able independently of the fact that the judgment is made. And that is why our taking certain things to be true of the world must be involved in interpretation from the beginning.

Davidson himself would make no appeal to idealism or any form of “subjectivism” to explain why he thinks most of our beliefs must be true. It comes from the conditions of interpretation alone. But making sense of what people are saying and doing, and ascribing various psychological states to them, is something that we human beings do. The conditions of interpretation or understanding are conditions of our doing something, or our succeeding in doing it, not


simply conditions of something’s being so independently of our efforts to understand. If our taking certain things to be true is a con-dition of ascribing to people beliefs and perceptions and attitudes with specifi c contents, then there will necessarily be considerable agreement among us. But that does not strictly imply that what we largely agree about must be true. The truth of something does not in general follow from the fact that some or many or even all human beings agree about it. Nothing about the conditions of interpretation can obliterate that fact.

Of course, if we all agree about many things then we will regard them all as true. We will hold, of those things that we all agree about, that they are all true. But it still does not follow from our acknow-ledged agreement that they are all true, even if we insist that agree-ment is indeed necessary for interpretation and mutual understanding. So even if we must all share certain evaluative (or color) judgments if we are to see ourselves as having any evaluative (or color) opinions at all, I do not see that the truth of any evaluative (or color) statements themselves, as opposed to our believing them to be true, would follow from that. Since we do believe them, we will assert them to be true. But no fact of the world would have been shown to be a necessary condition of our believing things about the world. This is still com-patible with our insisting that we must take certain things to be true of the world in order to see ourselves as believing or having any opin-ions or attitudes about anything.

The fact about interpretation and the ascription of belief is the important point. I think it is enough in itself to prevent us from ever arriving at the “subjectivist” picture of the world. By that I mean only that we could never consistently arrive at the “subjectivist” con-ception of values or color, not that that conception is false, or neces-sarily false, or a contradiction, as it would have to be on the stronger conclusion that sees a necessary connection between the body of our beliefs as a whole and their truth. That stronger conclusion would say that there could not possibly be such things as evaluative beliefs, or color beliefs, unless they were on the whole in fact true. And that is the kind of necessary connection that only idealism or some form of “subjectivism” would seem able to explain.

To think that we are forced to that strong conclusion by what I am calling the fact about interpretation would be to take the kind of step Bishop Berkeley took to his form of idealism. Because he thought that we cannot conceive of an object without perceiving it,


he thought that we cannot conceive of an object that remains unper-ceived. He concluded that an object could not possibly exist unper-ceived — it is inconceivable. But that is to start with the fact that we cannot do something, that we cannot perform a certain feat, and to conclude that a certain thing could not possibly be so. If we distin-guish, as we must, between what we cannot do and what cannot be so, between what we cannot consistently think and what is in itself inconsistent, the fact about the conditions of interpretation will not support the idea that our color beliefs or our evaluative beliefs are simply such that most of them must be true. That conclusion any-way would once again encourage idealism or “subjectivism.” But the fact about interpretation would mean that that is no threat. We would be in no position to deny or refute “subjectivism” on the grounds that it is inconsistent, but we would never be able consist-ently to reach the thought of the truth of “subjectivism” either. Denying “subjectivism” is not the only way of avoiding it.

So although I strongly resist the “subjectivity” of value, I do not wish to be understood as defending the idea that on the contrary values are “objective.” The tendency to draw that conclusion directly from the unacceptability of “subjectivism” is good evidence, if more were needed, of the power of that traditional metaphysical dichot-omy. It still represents the structure in terms of which we want to understand things. We feel that colors, or values, or whatever it might be, must be either “subjective” or “objective.” We think that either there is something corresponding to our thoughts about them in the fully “objective” world or there is not. So if we cannot really under-stand values to be “subjective,” we think they must be “objective” after all.

I believe that we cannot get a satisfactory understanding of our-selves in that way. That leaves us dissatisfi ed. It seems as if it couldn’t be simply impossible. So we persist. And once again we apply the traditional dichotomy. This tendency, I believe, is the place to look for the real source of “subjectivism” about values. It would help explain why we can expect that some form of the view will always be with us.

5

Hume’s an Enquiry Concerning Human Understanding

An Enquiry Concerning Human Understanding was fi rst published in London in 1748. In this book a brilliant young philosopher at the height of his powers presents to the eighteenth-century public a com-pletely new and revolutionary conception of philosophy. Philosophy as David Hume understands it is the investigation of human nature, and he here explains how pursuing philosophy in the way he recom-mends will lead to a novel and eventually liberating understanding of human beings and human life.

The outcome would be novel because an accurate investigation of what human beings are actually like can be expected to expose and so discredit a long-standing misconception about the role of reason in human life. It would be liberating because that new understanding of human nature would lead refl ective human beings to greater calm, ease, and balance in their understanding of themselves and the world, and so in their dealings with one another. That would tend to sup-plant the prejudices, factionalism, and superstitions that have so bedevilled social and personal life in the past, and would lead in time to greater human happiness and fulfi llment.

These are very large, ambitious claims to be made for what is no more than a set of philosophical ideas, and Hume was well aware of the diffi culty of persuading anyone of them. But there was special reason for optimism. He does not simply present an appealing pic-ture of human life for the reader’s contemplation. He tells us how this new liberating understanding is to be achieved. And the proced-ures he recommends are themselves based on facts discoverable by that very study of human nature. This means that if human nature and human thinking are in fact as Hume says they can be discovered to be, then all those who observe and investigate human beings in the ways he recommends will come to the same general conclusions about what makes us think and feel and act in all the ways we do. And by arriving in that way at such an understanding of oneself and

virtualur

Resaltado

virtualur

Resaltado

virtualur

Resaltado


one’s fellow human beings, one would fi nd oneself in just the liber-ated frame of mind and spirit that Hume promises.

The huge ambitions behind this provocative book make it one of the best expressions of attitudes characteristic of what later came to be called the Enlightenment. Put in the broadest terms, that was the idea of taking every aspect of human life as much as possible into human beings’ own hands. It was a matter of getting as free as pos-sible of the unreasonable restrictions of the past, of going farther than had even been thought possible before in the direction of exclu-sively human responsibility for the conditions of human life. The goal was to fi nd out as much as could possibly be discovered about what the world is actually like and how and why human beings oper-ate in it in all the ways they do. Human life was to be conducted as much as possible in the light of all that knowledge, so it could be lived in informed acknowledgement of what are found to be truly unavoidable limitations on human liberty and happiness rather than in meek, benighted acquiescence in the imagined requirements of some supposed but unknown power.

This idea of the investigation of the natural and the human world for human purposes is so familiar to us today that it is sometimes diffi cult to remember that it has not always been with us. It was the beginning in Hume’s day of what we now take for granted as the “social” or “human” sciences – psychology, economics, politics, soci-ology, etc. It was inspired by the overwhelming success – one might almost say the birth – of the new mathematical science of physical nature. Here the monumental achievements of Newton stood as an intellectual model for all. The larger idea was to pursue scientifi c understanding of all aspects of the natural world and to apply its results to better the conditions of human life. Nothing was to be taken as sanctioned by time or tradition alone. Only what could be supported by known or discoverable facts of the natural world was to be worthy of endorsement. The idea was to study and rethink and if necessary refashion human life from the ground up.

Hume certainly had such ideas and hopes, and was inspired by them. Although he was only thirty-six years old when this book was published, he had already written a much more elaborate and more comprehensive defence of the same general project earlier, before he was twenty-fi ve. From his youth he had studied intensely the litera-ture of Greek and Roman antiquity. He was also familiar with the works of a few provocative thinkers of his own day who had begun

virtualur

Resaltado

virtualur

Resaltado

virtualur

Resaltado

Hume’s an Enquiry Concerning Human Understanding 105

to put the study of the mind of man on a new footing. He reports that when he was about eighteen “a whole new scene of thought” opened to him and “transported” him “beyond measure”. The next fi ve or six years were spent enthusiastically developing this vision and applying it carefully to virtually every aspect of human life. The result was A Treatise of Human Nature , a huge, wide-ranging work of three separate books, ‘Of the understanding’, ‘Of the passions’, and ‘Of morals’, published in 1739–40.

That book had apparently convinced no one. In fact it had hardly been noticed. In Hume’s own words “It fell dead-born from the press, without reaching such distinction as even to excite a murmur among the zealots”. Disappointed but undeterred, he resolved to try again. He thought the manner rather than the matter of his earlier presentation had been responsible for its failure. This time he began with a much shorter book, dealing with only that part of his general picture of mankind that concerns thought, perception, belief, and reasoning: hence An Enquiry Concerning Human Understanding . Later he took up other central aspects of human life dealt with in the Treatise and published An Enquiry Concerning the Principles of Morals in 1751.

In the years immediately following the failure of the earlier Treatise Hume had gained considerable popular success with the publication of a number of essays on political and moral subjects, so he pre-sented his philosophical ideas in this fi rst new Enquiry as a series of connected essays, each giving only the highlights of parts of the sin-gle comprehensive conception lying beneath. In writing in essay form this second time around Hume was also much more attentive to the diffi culties he knew he faced in explaining to even the most interested and sympathetic general reader the real sources of such a sweeping and revolutionary understanding of human life.

Thorough and accurate exploration of the fundamental principles of human thought and experience demands prolonged, diffi cult thinking on many interrelated and highly abstract questions. The true value of philosophy is simply unavailable without such detailed work. But Hume knew he must write for a general literate public that does not engage in detached, abstruse refl ections of that kind and has little sympathy for them. He had to build a bridge from the deep results he thought could be discovered in abstract philosophy to the general public he knew would be better off if it could understand and accept them. That is why the book begins with an explanation of


what he calls two different “species of philosophy” and of the rela-tion between them.

The “easy and obvious” philosophy found in literature and polite conversation refl ects on what it fi nds to be agreeable and disagreeable aspects of human nature with a view to directing human beings to virtue and happiness. It paints striking pictures of human life to inspire us with certain sentiments and attitudes, and to move us to action. What Hume calls “accurate and abstract philosophy” is more detached and dispassionate than that, and is typically pursued alone in a philo-sophical thinker’s study. It stands in relation to those more popular refl ections as the work of an anatomist stands to that of a painter. Just as the beauty and accuracy of a painter’s fi gures would suffer without the patient, accurate research of the anatomist, so the “easy and humane” philosophy of literate society would carry little conviction without an accurate understanding of the nature and force of the rea-sonings and sentiments that move us.

Not everyone is prepared to engage in the “painful and fatiguing” refl ections necessary for genuine philosophical discoveries, nor should they be. But to ignore or dismiss such diffi cult, abstruse enquiries altogether would leave the “easy and agreeable” philosophy the victim of whatever pleasant pictures happen to capture the imagin ation, whether they represent human beings as they really are or not. So the “abstract and profound” philosophy whose results Hume sketches through the middle parts of the Enquiry are essential to the success of his project.

Those results of what Hume calls “the science of man” are the outcome of a search for the fundamental principles of human thought, belief, and action. And like any science, that search must proceed from the careful observation of its subject matter: human beings and human life. Hume thinks that in the past too much of what had been thought and said about human life and how best to live it had come from arbitrary and unrealistic preconceptions or demands, or from entrenched religious doctrines, and not exclusively from dispassionate observation of the way human beings actually are. Insistence on empirical investigation of the human mind and how it works is fundamental to Hume’s revolutionary account. In this book he presents only enough of its basic results to bring the potentially liberating picture before the reader’s mind.

First, and most basically, there is human thought. To think of anything at all is to have an idea of it in the mind. Ideas are the

virtualur

Resaltado

virtualur

Resaltado


“materials” of thinking, so to understand how thinking is possible we must discover “the origin of ideas”. No ideas are innate to the mind as it is before any experience. Ideas, Hume fi nds, get into the mind only through what we experience in sensation and perception. That puts limits on what we can think about. We can have thoughts of a golden mountain or a virtuous horse which we have never seen, for example, but only because those ideas are made up of ideas we have received in experience and have combined in new ways. The ideas we use in such thoughts come from sense experiences of gold, of a mountain, of a horse, and of virtue, which is something we feel. Although we seem capable of an almost unbounded range of thoughts and imaginings, for Hume “the creative power of the mind amounts to no more than the faculty of compounding, transposing, augmenting, and diminishing the materials afforded us by the senses and experience”.

One of the basic principles of the human mind, then, is that all our ideas are copies of earlier impressions of the senses or are made up of ideas which are in turn copies of earlier impressions. Hume thinks this can be discovered by taking any idea now in a person’s mind and fi nding that its ingredients can be traced back to their origins in the person’s experience. Without the appropriate impressions there would be no idea there to consider; there would be no way for it get into the mind. This is confi rmed, for example, by the blind, who receive no impressions of colour and so have no ideas of colour at all.

The ideas that make up the “materials” of thinking do not come and go in the mind completely randomly or with no pattern. Human thoughts run together in connected sequences; we think of one thing because we have just thought of a certain other thing, and so on. This does not mean that the connections are always easy to identify or understand. Hume thinks there are certain general principles accord-ing to which all thoughts are connected — these are the laws of “the association of ideas”. He sees the search for those laws as parallel to the physical scientist’s search for general laws to account for the inter-actions of all the particles of the material universe.

Perceptions and thoughts do crowd in upon the mind, but what is essential for anything resembling human life are beliefs or expect-ations that go beyond what a person perceives or thinks about at a particular moment. This is what enables a person to anticipate things before they happen, and so to deliberate and direct his actions in the light of what he believes to be so. Intelligent human thought and

virtualur

Resaltado

virtualur

Resaltado

virtualur

Resaltado

virtualur

Resaltado

virtualur

Resaltado

virtualur

Resaltado


action would not be possible without such a capacity. The heart of Hume’s account of this fundamental feature of human life comes in Sections IV and V, with his “sceptical doubts” about the operations of the human understanding and then a “sceptical solution” of those doubts. It is the story of how we learn from experience.

Hume spends considerable time showing fi rst that we get beliefs or expectations of something we are not perceiving at the moment only on the basis of what we have perceived so far. From a perception or thought alone, with no previous experience of the things perceived or thought about, we would come to no particular belief about what is or will be so beyond the present experience. A totally inexperienced Adam, suddenly cast into this world with no previous exposure to what goes on here, would have no more basis for expecting to sink to the bottom of a lake he sees before him than for expecting to walk safely across its surface. Only repeated experience of the way the world has been found to be in the past can lead the mind to one belief rather than another about how the world will be in the future.

What Hume is most eager to show about this movement of the mind from having perceived things to be a certain way to believing that things are or will be a certain way is that the transition is “ not founded on reasoning, or any process of the understanding”. This is the key to his whole new philosophy, and a proper understanding of it is what can lead to the promised liberation.

It might seem that we could reason from our past experience of eating bread and fi nding ourselves nourished by it to the expectation that a piece of bread now before us will also nourish us if we eat it. But no such reasoning could be “demonstrative” in the sense of prov-ing the absolute impossibility of bread’s failing to nourish us in the future although it has done so in the past. When we form an expect-ation on the basis of past experience we always presume that things will continue to be in the future as they have been in the past. Experience would be no use to us if we did not proceed on that sup-position. But it is at least conceivable, and in that sense possible, for the supposition to be false. It is possible for the course of nature to change.

This means that if we have any reason to believe that the bread before us will nourish us it must be because we have some reason – but not a “demonstrative” or conclusive reason – to believe that the course of nature will not change in this case, and that some things in the future will continue to be as they have been in the past. But

virtualur

Resaltado

virtualur

Resaltado

virtualur

Resaltado

virtualur

Resaltado

virtualur

Nota adhesiva

parte clave

virtualur

Resaltado


the only reason we could have for believing that supposition must itself come from what we have found to be so in our past experi-ence. There is no other possible basis for believing something that goes beyond past and present experience, as that supposition does. But we could not reason from the way things have been found to be in the past to the supposition that in the future they will continue to be as they have been in the past. As Hume puts it, that would “be evidently going in a circle, and taking that for granted, which is the very point in question”. No arguments from past experience can support the supposition that the future will resemble the past since all arguments from past experience are founded on that very supposition.

This clever proof shows that in arriving at beliefs on the basis of experience human beings always take a step that is not supported by reasoning or any process of what Hume calls “the understanding”. This would seem to bring even our most familiar practices of com-mon life into doubt or disrepute. For those who insist that human life is to be controlled or guided by reason it could lead to despair. There is no hope of responding to this conclusion by resolving to form beliefs about the future only when they are supported by reason. That would be fatal, if we could do it. Since no such beliefs can be supported by reason, we would then have no expectations at all, and so could not direct our actions towards the ends we seek. But there is no danger of anyone’s fi nding himself in that position. Human beings simply cannot help getting beliefs about the future from what they have found to be so in the past.

The lesson Hume draws from this is “that it is not reasoning which engages us” to expect the future to resemble the past we have observed. So some other principle of human nature must be responsible for this fundamental operation of the mind. Hume fi nds that it is “custom or habit”, not reason, that is “the great guide of human life”. After per-ceiving a constant conjunction of objects of two different kinds human beings simply cannot help coming to expect something of the second kind from the appearance of something of the fi rst kind. That is an “ultimate principle” of the human mind that Hume does not pretend to explain further.

It is an operation of the soul . . . as unavoidable as to feel the passion of love, when we receive benefi ts; or hatred, when we meet with injuries. All these operations are a species of natural instincts, which no reasoning or process of the thought and understanding is able either to produce or to prevent.

virtualur

Resaltado

virtualur

Resaltado

virtualur

Resaltado


This is confi rmed by the fact that animals also learn from experi-ence, but it would be absurd to say that they are guided in these infer-ences by some process of argument or reasoning. Nor do young children form expectations in that way. Hume says we all simply share the same instinctual operation of projecting into the future the regu-larities we have observed to hold in the past. If that can be called a kind of reasoning, we can speak of “the reason of animals” as well as of human beings. But it is not an activity conducted by the faculty of reason or the intellect. It is “nothing but a species of instinct” which human beings possess in common with animals and on which the whole conduct of life depends.

The real “solution” to “sceptical doubts” raised about the power of reason in acquiring beliefs about the world is therefore that those beliefs are not and could not be undermined by any doubts about their basis in reason. Human nature, or our instincts, are too strong for any such intellectual “doubts” to succeed. Hume calls this a “sceptical solution” of the doubts. What he means by that, and how it can lead to the desirable state of mind promised by this new con-ception of philosophy, is best explained in the last Section of the Enquiry , on “the academical or sceptical philosophy”.

Before we get to that Hume uncovers another “ultimate principle” of the human mind that is equally fundamental to our thought. He regards it as one of his most important discoveries, and he is eager to illustrate some of its benefi cial consequences. We make sense of things in the world in terms of cause and effect. When we expect something on the basis of what we have observed so far we believe it will happen because of what has happened so far; given the events of the past, we believe things must happen in a certain way. Since we have this idea of the “necessary connection” between cause and effect, an empirical study of human beings will reveal how we get the idea. It turns out that its source lies not in the world we think about, but in ourselves.

Hume fi nds that in any particular instance we never directly per-ceive any necessity or causal connection between the things we see. We see one billiard ball strike another, and we see the second ball move, but we get no sense impression of any necessity with which the ball must move. We perceive a conjunction, but no connection, between the two events. Nor do we perceive any more than such a conjunction even when we fi nd a particular movement of our own body to follow immediately after our internal “act of will” to move it.

virtualur

Resaltado

virtualur

Resaltado

virtualur

Resaltado

virtualur

Resaltado

virtualur

Resaltado

virtualur

Resaltado


But after repeated observation of similar pairs of events we do come to believe that an event of the one kind is the cause of an event of the other kind. That belief involves an idea of a necessary or causal con-nection between things, and that idea, like all ideas, must be derived from some impression.

We do not get an impression of any necessity with which one thing is connected with another in a particular instance. What is not to be found in one instance is not to be found in a series of instances exactly similar to it. So we do not get the idea of necessity from impressions of some connection we perceive to be present in each of many instances. But we do get the idea of necessary connection after the repeated observation of one kind of event always following another. So that repeated observation itself must be the source of that new idea. The idea is generated in us by something that happens in our minds when we perceive uniformities in the world, not by anything that is present in the world we perceive. What happens is that we come to feel a determination of the mind to pass from an idea of a thing of one kind to an idea of its usual attendant. And that feeling or internal impression is the source of our idea of a necessary or causal connection between things.

Hume thinks it is an empirically discoverable fact that the human mind works this way. It is an operation of the mind that is as “ul timate”, as unavoidable, and as little subject to the dictates of reason as our instinctual tendency always to expect the uniformities we have observed in the past to continue into the future. This means that we will inevitably fi nd causes and effects to be present in the world wherever we fi nd regular correlations between events of differ-ent kinds. This is the key to Hume’s solution to the long-disputed question of “liberty and necessity”: the question of how there can be free, responsible action in a world in which everything that happens is caused by something else that happens.

Hume’s solution, or dissolution, of this problem is to show that it is simply a confusion to see any confl ict between liberty and necessity or to feel threatened in one’s freedom by the doctrine that all human actions are caused by other events that precede them. He thinks every one has always in fact accepted both the liberty and the neces-sity of human actions, despite apparently endless quarreling on the subject. And he thinks his novel account of the source of the idea of necessary connection is what fi nally exposes the whole controversy as idle or merely verbal.

virtualur

Resaltado

virtualur

Resaltado

virtualur

Resaltado

virtualur

Resaltado


Just as we cannot help coming to believe that certain events in inanimate nature are causally connected when we observe one kind of event regularly following another, so when we fi nd that certain motives, beliefs, and desires of human beings are regularly followed by actions of certain kinds we come to believe that those actions are caused by those states of mind. There is as much observable regular-ity in the operations of mind and action as there is in the rest of nature. In fact we draw conclusions about the world by combining reasoning about voluntary actions together with reasoning about inanimate nature with no sense that the necessity is different or weaker in one case than in the other. Wherever we fi nd uniformity of sequence we come to believe in necessity.

It is only because we think in these ways that we can make sense of what happens in the world. We explain the movement of a billiard ball by the fact that it was struck by another ball. We explain why somebody did a certain thing by citing the desires and beliefs and intentions with which he acted. If we did not think that those states of mind actually brought about the action in question they would not help explain why it happened. We would see no connection between them. That is why Hume thinks that anyone who under-stands human actions at all accepts the necessity with which they happen, given their antecedents.

For that same reason he thinks everyone who believes that some actions are freely performed – and so believes in liberty – also accepts the doctrine of necessity. Freedom or liberty is “a power of acting or not acting, according to the determinations of the will”. Someone acts freely if he chooses or decides to do something and does it with-out constraint. And everyone believes that people often act in that way: that a person’s decision or desires are what sometimes actually bring about his actions. If that were not so, the action would not be an action of that agent at all; there would be no connection between what went on in the person’s mind and what then happened. So the idea of liberty or freedom of action of the kind we want to preserve actually requires acceptance of the doctrine of necessity. Only confu-sion could have led to the thought that there is some confl ict between liberty and necessity. Hume briefl y explores several possible sources of the confusion. But his positive account of the idea of necessity is meant to ease our minds once and for all.

Those same results from the “science of man” are also applied in Hume’s infamous essay ‘Of miracles’, which is Section X of the Enquiry . This was written earlier and was to have been included in

virtualur

Resaltado

virtualur

Resaltado


the Treatise , but Hume withdrew it from that book out of fear of causing offence. The response to it in the Enquiry showed that his fears were justifi ed. The argument is simple and clever (‘Jesuitical’ Hume called it), and so almost certain to cause offence in those against whom it is directed. The offence is not lessened by Hume’s advertising his argument as “an everlasting check to all kinds of superstitious delusion”.

Briefl y put, it works like this. Whatever evidence there is in the world for the truth of Christianity comes from human testimony of the miracles that are said to make God’s presence known. The force of evidence from testimony about anything is affected by many fac-tors, including how unusual or unlikely the event in question would be if it had occurred. A miracle is not just an unusual event; it is a violation of a law of nature. But our acceptance of laws of nature is supported by wide experience of exceptionless correlations between events of two kinds. So no belief could get stronger support from any experience than the support our experience of the world gives us for accepting something as a law of nature.

This means that any evidence from testimony in support of an occurrence that violates a law of nature would also have a uniform, exceptionless body of experience against it. But such a body of evi-dence is the strongest evidence we could ever have for anything. So the preponderance of all the evidence available would always be in favour of the exceptionless, unviolated law of nature rather than the truth of what the testimony reports. So no human testimony could ever have suffi cient force to support belief in a miracle. Hume points out, mis-chievously, that it could do so only if the falsity of that very testimony would be an even greater miracle than the event it is testimony for.

This is a subtle, provocative argument. But it is not an argument against miracles as such. It does not say that there are no miracles, or that there could not be. Hume argues only that human testimony can never give us reason to believe in miracles. The distinction is important; it allows Hume to represent himself, more cautiously, as an opponent only of those who seek to defend Christianity on the principles of human reason rather than resting it on faith. His conclusion is only that “mere reason”, even with the help of experience, is not suffi cient to convince us of its truth. Whether this warrants his earlier boast that the argument he has discovered will serve as “an everlasting check to all kinds of superstitious delusion” is a question he does not return to.

His discovery that the idea of necessary connection comes only from our experience of observing regular sequences of events of

virtualur

Resaltado

virtualur

Resaltado

virtualur

Resaltado


different kinds is applied to other questions of religious signifi cance. Since the only causes and effects we can be led to believe in on the basis of observation alone are those we are led to by the correlations we actually observe, any inclination to believe in causes or effects that go beyond that basis must arise from some other source, not simply from what we observe. This sounds unremarkable and uncontrover-sial. But Hume uses it to devastating effect against the prospects of “natural religion”: the project of discovering in the observable world around us reasons for believing certain things about the cause or designer of the world and what he has in store for the well-being or future state of human beings on earth.

The essay “Of a particular providence and of a future state” presents the attack very indirectly. Hume goes to even greater lengths than usual to remove his own personal convictions from the centre of the controversy. Here a narrator (Hume?) reports a conversation in which he persuades a friend to pretend to make a speech before the people of ancient Athens defending Epicurus’ philosophy of nature against the charge that it denies a divine existence and so is danger-ous to society. The speech then given turns out to defend just the position Hume thinks follows from his new understanding of the source of the idea of necessary connection and of how we arrive at beliefs about the world on the basis of experience.

The basic thought is that, even granting that from the order and regularity we observe in the universe we cannot help inferring a cause in the form of an architect or designer of it all, we cannot ascribe to that cause any greater plan or any more qualities than those suffi cient to produce the effects we observe. If we do ascribe added qualities to the cause, it cannot be on the basis of anything we observe in its effects. A weight of ten ounces raised in a scale leads us to believe that the other side of the scale contains more than ten ounces, but it can provide no reason to believe that it con-tains precisely one hundred rather than one thousand ounces. To anyone who imagines that the cause of the world must have the power to produce something much greater and more perfect than the world of misfortune and pain and disorder we see around us, the imaginary Epicurus replies:

You forget, that this superlative intelligence and benevolence are entirely imaginary, or, at least, without any foundation in reason; and that you have no ground to ascribe to him any qualities, but what you see he has actually exerted and displayed in his productions.


Again this does not deny that there is such a “superlative intelli-gence and benevolence”. It denies only that belief in such a being has any “foundation in reason”. No reasons or grounds discoverable in the observable world can give support to specifi c religious doctrines that assert more than what is to be found in the patterns the world presents to us. This theme is developed more fully and with great bril-liance in Hume’s Dialogues Concerning Natural Religion , published posthumously in 1779. It remains one of the greatest works ever writ-ten on the subject.

The benefi cial effects to be expected from absorbing these and other results of the “science of man” as Hume understands it are summed up and defended in the fi nal section of the Enquiry , “Of the academical or sceptical philosophy”. The philosophy recommended there is “sceptical” in a sense akin to that of the sceptics of antiquity, who were said to have achieved a tranquil, trouble-free way of life by suspending judgement on all questions as to how things are and going along with the way things naturally strike them.

Human beings are naturally curious about themselves and their world and cannot altogether resist the urge to refl ect on what they do and how they do it. When pursued with no reliable guide or touch-stone, this can lead eventually to mystery, anxiety, and frustration, and so can leave one open to whatever reassuring or disturbing picture of the human condition the higher fl ights of imagination and superstition manage to construct. The “abstract and profound” refl ections essential to Hume’s “science of man,” when pursued to the limit, lead inevitably to negative and apparently despairing conclusions of the kind he illus-trates in his treatment of beliefs about the future and about causal connections between things. No reassuring foundation in reason can be found for such beliefs, and they gain no support from anything available in experience. But the crucial discovery is that we simply can-not abandon such fundamental beliefs; we will inevitably fi nd ourselves believing them as long as we experience anything at all.

Here lies the “sceptical” point of engaging in the “painful and fatiguing” refl ections that lead to those excessive and unbelievable conclusions. The doubts and scruples we become rightly convinced of concerning those natural beliefs simply cannot be put into practice. It is in fi nding this to be so in our immediate experience that we come fi nally to appreciate the real power of nature and instinct over rea-son. To live in full recognition of that force of nature in human life would be to enjoy the “mitigated scepticism” that Hume recommends


as the best way to live. It is a state of mind and spirit achievable only when the “profound” but excessive and unliveable conclusions of “abstract” philosophy have been tempered or mitigated by our irre-pressible natural instincts and inclinations. But both the “accurate and abstract philosophy” and the inevitable operation of natural instinct are needed for that happy outcome.

Hume believes that appreciating the force of nature over reason in this way would lead human beings to greater “modesty and reserve” in their everyday refl ections and interactions and to greater ease and patience in the face of ignorance and uncertainty. It could lead as well to a greater concern with the complexities of common life and with issues on which progress can be expected than with those “distant and high enquiries” in which uncertainty and frustration seem inevitable.

While we cannot give a satisfactory reason, why we believe, after a thousand experiments, that a stone will fall, or fi re burn; can we ever satisfy ourselves concerning any determination, which we may form, with regard to the origin of worlds, and the situation of nature, from, and to eternity?

It is true that giving free rein to the imagination and superstition on such matters would also be to yield to natural inclinations. But sim-ply following nature in every direction is not necessarily the best way to live. Dogmatism and superstition, perhaps, come naturally to us, but Hume recommends philosophy over superstition. He does so not because he thinks he has shown that philosophy leads to the truth and superstition does not, or that the conclusions of philosophy are supported by better reasons than those of superstition. It is rather that superstition, as he observes, “seizes more strongly on the mind, and is often able to disturb us in the conduct of our lives and actions”. Some people are more disposed than others to indulge in such disturb-ances, but they are the source of faction and discord in the conduct of personal and public life.

No such dangers are to be expected from philosophy. That is why Hume proposes “accurate and just reasoning” as “the only catholic remedy, fi tted for all persons and all dispositions”. It alone, when prop-erly followed, can subvert dogmatism and divisive prejudices and lead to a less disturbed, more fully satisfying, and more balanced way of life. And for Hume the only kinds of consideration that could support any claim about how best to live are the nature and quality of the sat-isfactions to be experienced by the human beings who live that way.

6

Ayer’s Hume

Ayer admires Hume as “the greatest of all British philosophers”. 1 In our own century, he has the highest regard for Russell. 2 He would no doubt look with favour on a history which placed his own work as the next giant step in the march of British empiricist philosophy from Hume through Mill to Russell and beyond. He defended the positiv-ism of his fi rst book as “the logical outcome” of the empiricism of Berkeley and Hume. 3 But Berkeley, for all his cleverness, seems never to have been Ayer’s cup of tea. He certainly was not sound theologic-ally. On that score, and most others, Hume is more reliable—and truly simpatico . Many of the specifi c doctrines of Language, Truth and Logic have been abandoned or revised beyond recognition over the years, but Ayer’s general philosophical outlook has scarcely changed since then. There remains at its centre his endorsement of what he sees as the achievement of Hume.

But even Russell found that Hume, “by making [the empirical phil-osophy of Locke and Berkeley] self-consistent, made it incredible. He represents, in a certain sense, a dead end: in his direction, it is impos-sible to go further”. 4 That did not seem to stop Russell from trying to go further in that same direction. He continued to “hope that something less sceptical than Hume’s system may be discoverable”. 5 It cannot be said that he discovered it.

I think that is because Russell’s verdict was right. It is impossible to go further in the direction he had in mind. But what direction is that? And was it Hume’s direction? Ayer thinks it was. He has remained committed to a certain conception of what the only or at

1 A. J. Ayer, Hume , New York, Hill and Wang, 1980. p. 1 . 2 A. J. Ayer, Wittgenstein , London, Weidenfeld and Nicolson, 1985. p. 145 3 A. J. Ayer, Language, Truth and Logic , 2nd ed., New York, Dover, 1952. p. 31 . 4 B. Russell, A History of Western Philosophy , New York, Simon and Schuster, 1945.

p. 659 . 5 Ibid.


least the proper task for philosophy can be. And that conception, combined with his admiration for Hume, is what I think leads him to his understanding of the aims of Hume’s philosophy. But that seems to me to ignore or at least to distort what is most important and still most fruitful in that philosophy. Ayer also shares many specifi c doc-trines with Hume. But I think those views, given Ayer’s conception of the philosophical enterprise, are the real source of the dead end which Russell encountered and, if I am right, Ayer still faces. I there-fore would like to take up Ayer’s understanding of Hume, his con-ception of philosophy, and the relation between the two.

I

It is not easy to say what Ayer thinks Hume was up to as a philosopher. Or rather, it is not easy to say what he thinks Hume was up to; he seems to have no doubt what it had to be if it can be called philosophy. He invokes Hume’s views often in his philosophical writings, but always in the treatment of some specifi c issue or as the source of this or that particular problem. Even the brief general introduction to Hume in the Past Masters book gives no satisfyingly comprehensive description of what Hume himself was trying to do or saw himself as doing.

Ayer duly reminds us that seventeenth- and early eighteenth-century writers did not distinguish philosophy from the sciences as we do today. He mentions Kemp Smith’s interpretative emphasis on Hume’s attempt to “explain the principles of human nature”. 6 He acknowledges Hume’s wish to develop a “science of human nature” 7 or a “science of the mind”. 8 He quotes Hume’s remark in the Introduction to the Treatise that “ ’Tis impossible to tell what changes and improvements we might make in these sciences [of “ Mathematics, Natural Philosophy, and Natural Religion ”] were we thoroughly acquainted with the extent and force of human understanding, and cou’d explain the nature of the ideas we employ, and of the operations we perform in our reasonings”. 9 But he identifi es Hume’s project with the program of Locke’s Essay Concerning Human Understanding : to answer the questions “What are

6 A. J. Ayer, 1980, op. cit. , p. 18. 7 Op. cit. , p. 24. 8 Op. cit. , p. 25. 9 D. Hume, A Treatise of Human Nature , ed. L. A. Selby-Bigge, 2nd edition revised

by P. H. Nidditch, Oxford, Oxford University Press, 1978. p. xv .

Ayer’s Hume 119

the materials with which the mind is furnished, and what uses can it make of them?”. 10 Perhaps Hume’s Lockean phrases here suggest endorsement of the humble “underlabourer’s” task of “clearing the ground a little, and removing some of the rubbish that lies in the way to knowledge”. 11 But that does not account for the unmistakable opti-mism of Hume’s Introduction—the expectation of untold “improve-ments” in those sciences promised by leaving aside “the tedious lingring method” and marching up “directly to the capital or centre of these sciences, to human nature itself”, from which “we may extend our con-quests over all those sciences, which more intimately concern human life”. 12 We do not have to believe that Hume’s hopes for “the science of MAN” have been realized in order to ask what those hopes might have been, what reasons he had for entertaining them, and how he thought they might be fulfi lled.

The Treatise bears the conspicuous subtitle “Being an Attempt to introduce the experimental Method of Reasoning into MORAL SUBJECTS”. Hume stresses that “as the science of man is the only solid foundation for the other sciences, so the only solid foundation we can give to this science itself must be laid on experience and observation”. 13 He warns that it is “impossible to form any notion of [the mind’s] powers and qualities otherwise than from careful and exact experiments, and the observation of those particular effects, which result from its different circumstances and situations”. 14 Ayer does briefl y explain the division between natural philosophy and moral philosophy, and he notes Hume’s attachment (which he says is shared with Locke) to the idea that the experimental method of rea-soning is applicable to the moral sciences. He concedes that Hume would probably describe his own work as falling within the domain of “experimental reasoning”. 15 But he thinks Hume did not really understand that method as it was employed to such impressive effect in the great work of Newton. And he thinks that it was only to a very limited extent (and then chiefl y in writing about morality) that he

10 Op. cit. , p. 25. 11 J. Locke, An Essay Concerning Human Understanding , ed. A. C. Campbell, New

York, Dover, 1959. p. 14 . 12 D. Hume, 1978, op. cit. , p. xvi. 13 Ibid. 14 Op. cit. , p. xvii. 15 A. J. Ayer, The Central Questions of Philosophy , Harmondsworth, Penguin Books,

1977. p. 23 .


advanced any empirical generalizations which could be tested by experiment. 16

Whatever Hume might say he is doing or wants to do, for Ayer it is clear that “his main business is with concepts”. 17 His philosophical task was to “analyze” them. In this he is apparently not unique. “The majority of those who are commonly supposed to have been great phil-osophers were primarily not metaphysicians but analysts”. 18 Without a sharp distinction between philosophy and science Hume could not be expected to have understood his philosophical work in this way. Ayer sees him as to that extent confused, often expressing his views or his problems in misleading psychological terms, and sometimes even pursuing questions of what we would nowadays call psychology.

But apparently he was not completely confused. There is the famous concluding passage of An Enquiry Concerning Human Understanding which Ayer quotes often as a statement of Hume’s own understanding of his enterprise.

When we run over our libraries, persuaded of these principles, what havoc must we make? 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 reason-ing concerning matter of fact and existence ? No. Commit it then to the fl ames: for it can contain nothing but sophistry and illusion. 19

In Language, Truth and Logic Ayer sees this as a rhetorical version of his own thesis that “a sentence which does not express either a for-mally true proposition or an empirical hypothesis is devoid of literal signifi cance”. 20 In The Central Questions of Philosophy he takes this same passage “implicitly” to express the view that analysis is the only proper activity for philosophy. 21 In Hume he quotes it again, as the best short summary of Hume’s whole “general outlook”, this time linking it most closely with resistance to the “licence” of the imagin-ation which breeds theology and superstition. 22

16 Ibid. 17 Ibid. 18 A. J. Ayer, 1952, op. cit. , p. 52. 19 D. Hume, Enquiries Concerning the Human Understanding and Concerning the Principles

of Morals , ed. L. A. Selby-Bigge, Oxford, Oxford University Press, 1975. p. 165 . 20 Ayer, 1952, op. cit. , p. 54. 21 A. J. Ayer, 1977, op. cit. , p. 23. 22 A. J. Ayer, 1980, op. cit. , p. 96.

Ayer’s Hume 121

These different claims for this stirring passage are not obviously equivalent. Is it really making a claim about the nature of philosophy at all? It says nothing directly about the “literal signifi cance” or insig-nifi cance of any sentences. It is true that it is based on Hume’s div-ision of “all the objects of human reason and enquiry . . . into two kinds, to wit, Relations of Ideas , and Matters of Fact ”. 23 But for Hume a “matter of fact” is something the opposite of which is pos-sible in the sense of implying no contradiction. It is also something that comes to be known or even believed only on the basis of experi-ence. But that does not imply that no “matter of fact” would be intel-ligible, and hence would even be a “matter of fact” at all, unless its holding or not holding were to some degree verifi able in human ex perience. That is what Ayer’s verifi ability criterion of meaningfulness says.

Even if Hume were concerned with meaningfulness, and were expressing Ayer’s verifi ability criterion in this passage, it would not follow that he is here declaring that philosophy, or what he is trying to do in the Treatise and Enquiries , is exclusively an exercise in con-ceptual analysis or refl ection on “relations of ideas”. He could hold the criterion and reject everything that violates it while still seeing his own writings as perfectly meaningful applications of “the experimen-tal method of reasoning” to the study of human nature. That is what he says he is doing.

Ayer thinks the view that philosophy is analysis “can be made to follow” from Hume’s attack on “school metaphysics” in this pas-sage. 24 But what Hume says here about “divinity or school metaphys-ics” is, fi rst, that any book treating of such matters will contain nothing but sophistry and illusion if it does not contain abstract rea-soning concerning quantity or number or experimental reasoning concerning matter of fact and existence. And on Hume’s view that is true of any book on any subject. Second, he suggests, but without documentation, that there is no such reasoning in books of those kinds. But that is the complaint that those books lack certain kinds of reasoning, or that they lack reasoning altogether, not necessarily that they do not contain any sentences of certain legitimate kinds. Again, if Hume thinks their sentences are indeed meaningless, it does not follow that he thinks the proper task for philosophy is analysis.

23 D. Hume, 1975, op. cit. , p. 25. 24 A. J. Ayer, 1977, op. cit. , p. 23.


But even if their sentences are perfectly meaningful those books would still be nothing but sophistry and illusion if they give no rea-sons for believing anything they say. That will be so if, as is likely, they appeal only to the imagination, which is “naturally sublime, delighted with whatever is remote and extraordinary, and running, without control, into . . . distant and high enquiries”. 25

That does seem to be what Hume has in mind. The passage occurs at the end of the last section of the Enquiry where Hume is listing the advantages to mankind of the various forms of mitigated scepticism he has been recommending. One “natural result of the Pyrrhonian doubts and scruples, is the limitation of our enquiries to such sub-jects as are best adapted to the narrow capacity of human understanding”. 26 To see where it is best to concentrate our efforts “it suffi ces to make the slightest examination into the natural powers of the human mind and to . . . fi nd what are the proper subjects of sci-ence and enquiry”. 27 We can then leave “the more sublime topics to the embellishment of poets and orators, or to the arts of priests and politicians”. 28 There the imagination runs free, with its inevitable sophistry and illusion. Ayer’s third reading of the oft-quoted passage therefore seems to me closest to the truth. But it implies nothing about philosophy as exclusively, or even partly, analysis of concepts.

Hume does not think his own Treatise and Enquiries are to be com-mitted to the fl ames. He thinks they contain lots of good reasoning. But not because he thinks they contain only a priori reasoning con-cerning “relations of ideas”. He would exempt them from the charge of sophistry and illusion because he thinks they contain experimental reasoning concerning matter of fact and existence as well, even rea-soning on questions of “the most profound metaphysics”. 29

II

The problem of fi nding a place for a special subject called philosophy on a complete map of human knowledge or enquiry was especially pressing to the logical positivists, in whose name Ayer was writing in

25 D. Hume, 1975, op. cit. , p. 162. 26 Ibid. 27 D. Hume, 1975, op. cit. , p. 163. 28 Op. cit. , p. 162. 29 D. Hume, 1978, op. cit. , p. 189.

Ayer’s Hume 123

1936. It could be said indeed that the nature and possibility of philosophical knowledge was one of the major concerns of that movement. It certainly dominates Language, Truth and Logic ; four—and indirectly a fi fth—of its eight chapters are fully occupied with the nature of philosophy or with the special character of philosoph-ical issues or disputes. The problem was how philosophy as a serious intellectual enterprise was possible.

For Kant philosophical knowledge was synthetic and a priori. Its results were known independently of experience, they were genuinely ampliative—extending beyond the contents of their constituent concepts—and we could be assured that they could not be otherwise. For Wittgenstein in the Tractatus there could be no philosophical propositions and so no genuine philosophical knowledge or statable philosophical results. Philosophy was not a body of doctrine but an activity; its aim was the logical clarifi cation of thoughts. Any neces-sity apparently possessed by certain sentences was due entirely to their empty, tautological character; they say nothing about the world. The positivist account of philosophy was a combination of these views, remaining on the whole much closer to Kant than to Wittgenstein while rejecting the central idea of the Critique of Pure Reason . All genuine knowledge of the world is part of empirical sci-ence, and so not philosophy. All mathematics and logic—in fact, every thing knowable a priori—is analytic, true solely by virtue of the meanings of its constituent terms. There is no synthetic a priori knowledge. Any philosophical knowledge will therefore be independ-ent of experience, it can be knowledge of what could not be other-wise, but it will be only analytic. It can reveal nothing about the way the world is, but only about the concepts in terms of which we under-stand and come to know things about the world.

Ayer sees this conception of philosophy as the unavoidable result of a straightforward process of elimination. 30 Philosophy must be con-ceptual analysis if it is anything respectable because “all the other avenues of knowledge are thought to be already pre-empted”, and the philosopher has “no right to trespass”. 31 The crucial step is the assump-tion that philosophy cannot be empirical because all knowledge of the world is part of empirical science. Kant thought philosophy could not be empirical because its results were known to be necessarily true; and

30 A. J. Ayer, 1952, op. cit. , p. 33. 31 A. J. Ayer, 1977, op. cit. , p. 22.


for him necessity was a “sure criterion” of the a priori. 32 Ayer shares with Kant that assumption about necessity and a priori knowledge, but it alone does not imply that philosophy must be a priori. Even combined with the thesis that all necessity is due to meaning it implies only that all a priori knowledge is only of analytic truths. But that still does not yield the conclusion that the philosopher can legitimately “trespass” only on the analytic or a priori side of the line and not on the synthetic or contingent or empirical side. It does not even imply that the philosopher must be restricted to one side or the other.

With respect to many of the particular philosophical doctrines which were thought to support this conception of philosophy as analysis Hume was of course much closer to Ayer and the positi-vists than he was to Kant. He held that of all those thoughts we can form that are capable of truth or falsity, some are necessarily true in the sense that their negations are contradictory, but they are true solely in virtue of the relations among their component ideas. We can accordingly know such things to be true by the operation of pure thought alone, by refl ection on our ideas. All the rest are such that their negations are not contradictory, so whatever the actual truthvalue of each of them happens to be, it at least could have been otherwise; it is not guaranteed by “relations of ideas” alone. Refl ection on ideas therefore could never in itself produce belief in one such “matter of fact” rather than its opposite, or give us any reason to believe it. Only actual sense-experience could do that. All beliefs in matters of fact are founded on experience.

But this set of largely epistemological views is consistent with a number of different conceptions of philosophy, or with holding no very determinate conception of it as a special discipline at all. The problem for the positivists was largely a problem inherited from Kant, or more precisely from their otherwise accepting the Kantian framework while rejecting the synthetic a priori. The terms in which it had to be solved were Kantian terms. But Hume was a pre-Kantian philosopher.

III

Ayer shares a number of other specifi c theses or doctrines or prob-lems with Hume, but here again I think we cannot assume on that

32 I. Kant, Critique of Pure Reason , tr. N. Kemp Smith, London, Macmillan, 1953. p. 84 .

Ayer’s Hume 125

basis alone that Hume therefore was pursuing what Ayer holds to be the only legitimate goal for philosophy. And when we look at what Hume actually does, I think we fi nd that he was not in fact pursuing that goal. Insofar as he could even formulate it or understand it, he thought it was impossible—and precisely because of his attachment to a number of views which Ayer also holds.

Some of the most important of those shared views could be expressed somewhat loosely as follows. We never directly perceive physical objects or states of affairs in the public world, but only fl eeting and momentary impressions or sense-data which we cannot be wrong about at the time they are present to our minds. Each of us has thoughts of or beliefs about a great many things other than our current impressions or sense-data—past or future impressions, for example, or enduring physical objects and their properties, causal connections between objects or events, laws of nature, the thoughts and feelings of other people, our own past and future selves, the goodness or badness of people’s characters and actions, the exist-ence of a supernatural God, and so on. Any such thoughts we can form, any beliefs we might arrive at about anything, must be con-structed by mental operations working only on materials derived from our immediate sense-experience.

No impression or sense-datum ever provides us with an instance of the identity of a physical or a mental thing (including ourselves) over time, or of the causal connection we believe to hold between two things when we believe that one is the cause of the other. No impres-sion or sense-datum provides us with an instance of the thoughts or feelings of other people, or of what we ascribe to an action or char-acter when we believe it to be good or bad, virtuous or vicious.

What we experience at one moment can give us reason to believe some other matter of fact which we are not experiencing at that moment only if we have reason to believe that it is connected in some way with that absent matter of fact. But the connection in question could not be necessary and so discoverable by the oper-ation of pure thought alone, since for any two distinct matters of fact, there is no contradiction involved in supposing that one of them holds and the other does not. Whether there is a connection between what we experience and what we believe on the basis of it must therefore be a further matter of fact, so any reason we might have for believing in such a connection must also be found in our experience.


These doctrines are baldly stated and not very carefully formu-lated in either Hume’s terms or Ayer’s. I make no claim of complete-ness or of any order of logical priority among them. But even in this summary form I think it would be granted that they capture a great deal that is important and central to each philosopher. Ayer com-bines them with the view that the task of philosophy is analysis. For him that implies that its results are not empirical or contingent. Philosophy is not empirical science.

I have mentioned Hume’s stated attachment to “the experimental method of reasoning”, his devotion to “experience”, and his resolve in the investigation of human nature not to “establish any principles which are not founded on that authority”. 33 But it is not just a matter of programmatic announcements at the beginning. When he gets down to work in detail he seeks “the origin of our ideas” and offers “plain and convincing” “phenomena to prove” that all our simple ideas are preceded into the mind by their corresponding simple impressions. 34 He tries to discover how we come to have the idea of causation, since “ ’tis impossible perfectly to understand any idea, without tracing it up to its origin, and examining that primary impression, from which it arises”. 35 He notes the fact that past experience of a conjunction of things of two kinds and a present impression of something of one of the kinds inevitably leads us to think of something of the other kind, and he asks “whether experience produces the idea by means of the understanding or of the imagination”. 36 He wants to know what “determines” us to make that transition. 37 He asks “what causes induce us to believe in the existence of body” 38 or what “makes us attribute to [our impressions] a distinct and continu’d existence”, 39 and he identi-fi es to his satisfaction the “principles” or “propensities” of “the imagin-ation” that are responsible for it. He asks what “gives us so great a propension to ascribe an identity to [our] successive perceptions, and to suppose ourselves possest of an invariable and uninterrupted exist-ence thro’ the whole course of our lives”. 40 And to the question of how

33 D. Hume, 1978, op. cit. , p. xviii. 34 Op. cit. , pp. 4–5. 35 Op. cit. , pp. 74–75. 36 Op. cit. , p. 88. 37 Op. cit. , pp. 88–89. 38 Op. cit. , p. 187. 39 Op. cit. , p. 194. 40 Op. cit. , p. 253.

Ayer’s Hume 127

that particular thought is “produc’d” he never fi nds an answer that satisfi es him. In moral philosophy as well the problem is to fi nd the impressions from which our judgements of moral good and evil are derived, to discover “after what manner they operate on us”, 41 and thereby to “fi nd those universal principles, from which all censure or approbation is ultimately derived”. 42

These are the words of someone who wants to understand in each case certain very general facts of human nature—how and why human beings come to think or feel or judge in the ways they do. And he wants to answer such questions in the only way he thinks they can be answered, by relying on what can be found out by observing human beings and the world they live in.

As this is a question of fact, not of abstract science, we can only expect suc-cess, by following the experimental method, and deducing general maxims from a comparison of particular instances. 43

Since thought, belief, and attitude are in question, the project might involve an occasional “analysis” or defi nition of one or another of our ideas—or more likely a denial of equivalence or implication between two ideas—but the goal throughout is an empirically-based explanation of highly general ways of thinking, feeling, and acting that are deeply characteristic of human beings as we know them. The facts in question, and the principles introduced to explain them, although extremely general, are still to be understood as contingent. Things could have been otherwise.

Ayer is obliged to fi nd Hume’s way of putting his project unfortunate—a product of confusion or naivete. He is not surprised that Hume’s views have been so often misunderstood; the fault does not all lie with his interpreters. One diffi culty is what he sees as Hume’s “misguided insistence on tracing ideas to their origin”. 44 But if our ideas have an origin, as it would seem they must, what can be wrong with trying to fi nd out what it is? Presumably what is mis-guided about it in Hume’s case is the “confusion of psychological with logical questions of which he is generally guilty”. 45 Ayer praises the Enquiry over the Treatise account of causation on the ground

41 Op. cit. , p. 470. 42 D. Hume, 1975, op. cit. , p. 174. 43 Ibid. 44 A. J. Ayer, 1980, op. cit. , p. 55. 45 Op. cit. , p. 57.


that it is “less encumbered with what would now be reckoned as psychology”. 46 But to have done something in 1740 that would now be reckoned as psychology is not necessarily to have been confused, any more than it is a sign of confusion to have done something in 1740 that would now be reckoned as economics, especially if psych-ology, economics, and other specialized subjects regarded in some quarters as “sciences” today grew directly out of the more encom-passing “science of MAN” of the eighteenth century, as they did. I think the confusion Ayer has in mind is the confusion of thinking that empirical investigation of the origins of human thoughts, feel-ings, and attitudes can answer questions about the meaning, struc-ture, and logical relationships among the contents of those thoughts and attitudes. But it will follow that a philosopher explicitly con-cerned with origins is guilty of that confusion only given a determin-ate conception of philosophy as properly restricted to the a priori analysis of concepts or ideas. And that was not Hume’s conception of what he was doing.

Analysis for Ayer “covers quite a number of activities, which differ from one another either in their methods or their aims or both”. 47 One is giving explicit defi nitions of individual words or ideas by means of synonymous words or equivalent ideas. Another is produc-ing what he called in Language, Truth and Logic “defi nitions in use” 48 which state that, or show how, sentences containing the term in ques-tion can be translated into equivalent sentences which do not contain that term or its synonyms. He saw “the traditional problem of per-ception”, for example, as the problem of showing how sentences about material things can be translated without remainder into sen-tences about “sense-contents”. 49 That would be to provide a phenom-enalistic reduction of material things to “sense-contents”, or to show that material things are “logical constructions” out of “sense- contents”. I think it is in that same sense of “analysis” that he there saw Hume’s treatment of causation as an attempt to answer “the analytic question, What is it that we are asserting when we assert that one event is causally connected with another?”. 50 The task was to reduce the problematic idea to some combination of elements

46 Op. cit. , p. 6. 47 A. J. Ayer, 1977, op. cit. , p. 44. 48 A. J. Ayer, 1952, op. cit. , pp. 59ff. 49 Op. cit. , p. 64. 50 Op. cit. , p. 54.

Ayer’s Hume 129

which are immediately available in sense-experience. The same was true of Hume’s problem of giving “an analysis of the notion of a self” or personal identity; what relations must obtain between sense-experiences for them to belong to the same self ? 51

Hume usually concentrates on our possession of individual ideas or terms. I doubt that he can be credited with the notion of defi ning a problematic term by eliminating it from full sentences in which it can appear in favour of equivalent full sentences which make no use of that term. That is an idea that had to await recognition of the primacy of the judgement or sentence over the idea or term, something con-spicuously lacking in much modern philosophy before Kant. It is the key to the notion of reduction so prominent in analytic philosophy of the twentieth century. Ayer himself has largely abandoned the search for tight reductions, especially phenomenalistic reductions, of the ideas that interest him, and seems accordingly to have become more circumspect in attributing precisely that analytical goal to Hume. But he continues to take it for granted that when Hume considers percep-tion, causation, enduring bodies, personal identity, and moral judge-ment he primarily wants somehow to account for the content of what we think in each case in terms of what is available to us in our immedi-ate experience. The point is to analyze our ideas or concepts or beliefs.

There is one promising exception to this general line of interpret-ation. In his Hume he expresses some to my mind well-justifi ed doubts, but only in the case of morality. Hume famously declares: “when you pronounce any action or character to be vicious, you mean nothing, but that from the constitution of your nature you have a feeling or sentiment of blame from the contemplation of it”. 52 This is just the kind of remark which has seemed to many to support the idea that he is engaged in analyzing the meanings of the things we say or think, but Ayer believes that that is not the way to understand what Hume is doing here. He is not advancing a thesis about the logical equivalence of moral judgements and descriptive statements about someone’s actual or possible feelings, or claiming that in making moral judgements we are covertly asserting something about ourselves or other moral judges. 53

Ayer thinks that what he is doing can still be called “analysis.”

51 Op. cit. , 125. 52 D. Hume, 1978, op. cit. , p. 469 53 A. J. Ayer, 1980, op. cit. , p. 84.


There is indeed a sense in which he is offering an analysis of our moral judge-ments, but the analysis is not intended to supply us with a recipe for translat-ing the sentences which express them. It consists rather in an account of the circumstances in which we are induced to employ moral predicates, and of the purposes which their employment serves. 54

Whatever it is called, it is clear that a study of the circumstances in which we are induced to employ moral predicates, and the purposes served by that employment, could be carried out only by observation of human beings and the circumstances in which they fi nd them-selves. That they think and behave in those ways is a contingent mat-ter of fact. The results of an investigation into those facts would not be true by virtue of the relations of ideas alone. This task would therefore be “analysis” in name only. It would not issue in “analytic” truths, true by virtue of meaning alone. If it is nonetheless a proper task for philosophy, it would leave Hume the philosopher free to develop an empirical “science of human nature” in just the way he so clearly seems to say he wants to. That is just the interpretation of Hume that I am suggesting is to be preferred everywhere, not only in his treatment of morality.

Ayer does not say much to explain why he rejects in this case what he elsewhere has taken to be the proper understanding of Hume’s philosophical task. Perhaps it is because no account of “the circum-stances in which we are induced to employ moral predicates, and of the purposes which their employment serves” would in itself yield an account of the meaning of those predicates or provide translations of their content in terms of those “inducing circumstances” alone. That would be to distinguish questions of meaning from questions of origin. And of course it is right to distinguish them. But that is the very distinction that Ayer usually charges Hume with blurring, or not even seeing. Why then in this case does he think Hume saw it and abided by it? And if he did, what prevented him from doing so in his treatment of the other important ways of thinking he considers?

Perhaps Ayer makes an exception in this case because he thinks the reductive or defi nitional strategy would commit Hume to the mis-taken view that moral or evaluative statements are strictly equivalent to descriptive or factual statements. Ayer seems to agree that that view is mistaken in his discussion of the relation between ‘is’ and ‘ought’ two pages later in Hume . Even in Language, Truth and Logic

54 Ibid.

Ayer’s Hume 131

he rejected any defi nitional equivalence between normative and fac-tual propositions on the ground that whatever factual statement might be true of a thing there is no particular normative statement that it would at the same time be self-contradictory to deny of it. That general failure of implication is what led him to the emotive theory of moral judgements, a theory he says would probably be closest to Hume’s intentions if he were indeed searching for an ana-lysis or reformulation of their content. But Ayer thinks that is not what Hume is doing.

I surmise that Ayer does not attribute to Hume the search for a reduction or defi nitional equivalence in this case because he thinks Hume holds an even more radical view of morals which in broad outline Ayer thinks must be correct. An equivalence in content between moral judgements and statements about actual or possible feelings would imply that something’s goodness or badness consists in the fact that people do or would feel a certain way towards it. Denial of the equivalence could be taken to support the view that moral facts or states of affairs are completely different in kind from all such natural facts. But the more radical view says that there is no such fact as a fact of something’s being good or bad at all. Of course, things have a tendency to make us feel one way or another towards them, and that is a fact. But when we make a moral judgement we are not simply asserting that dispositional statement about the thing in question. Nor are we asserting something “non-natural” about it. We are evaluating the thing, or making a normative judgement about it, but on the radical view there is no fact or state of affairs in the world that could make the normative aspect of the judgment either true or false. The normative judgement in that sense has no factual content, so no defi nition of its normative content can be given. Any attempt to give one would either employ an equally unex-plained evaluative term or, if it did not, would reduce the evaluative judgment to a statement of fact and hence erase its special evalu-ative character. So we can perhaps say what people are doing when they make moral judgments, and we can investigate the conditions under which they make them, and why, but we cannot understand them as evaluative judgments and at the same time give an analysis or defi nition in equivalent factual terms of their content. If Ayer thought that Hume held some such view of morals it would explain why he does not think he tries to provide translations or reductions of the contents of moral judgments.


I think this is just the kind of view of morals Hume does hold. Because he fi nds nothing in the world that could correspond to the special evaluative aspect of our moral judgments, the problem of understanding them is the problem of explaining how we come to make moral verdicts or put forward moral judgments at all, and why we arrive at the particular moral views that we do. That is not a ques-tion of analysis in the sense of being answerable by refl ection on rela-tions of ideas alone or issuing in analytic truths. It is a question of fact, to be answered by empirical observation of human beings and the world they inhabit. But the same is true of each of the other major objects of study in Hume’s pursuit of “the science of MAN”.

In the case of causality Hume insists from the beginning that part of what we ordinarily believe when we think that two things are related as cause and effect is that there is a “necessary connection” between them. That is more than contiguity and temporal priority. “An object may be contiguous and prior to another, without being consider’d as its cause. There is a NECESSARY CONNEXION to be taken into consideration”. 55 Of course, we do not get that idea from a single instance, but when we have found things of one kind constantly conjoined with things of another “we then call the one object, Cause ; the other, Effect . We suppose that there is some con-nexion between them; some power in the one, by which it infallibly produces the other, and operates with the greatest certainty and strongest necessity”. 56 Hume’s question is how we come by this idea of necessity; how that special modal aspect of our thought about cause and effect comes into our minds, and what leads us to apply it in the ways we do.

In the course of his search for the source of that idea he fi nds that it is not derived from any impression of the qualities of objects or of the relations between them. Only the repeated observation of a con-junction between things of two kinds gives rise to it. But mere repeti-tion cannot reveal something new in the single instances which was not there to begin with, or produce something new in them. So the idea of necessary connection must have its source solely in our minds, not in the objects the observation of which eventually produces that idea. In summing up this genetic story Hume allows himself to say such things as:

55 D. Hume, 1978, op. cit. , p. 77. 56 D. Hume, 1975, op. cit. , p. 75.

Ayer’s Hume 133

When we say, therefore, that one object is connected with another, we mean only that they have acquired a connexion in our thought, and give rise to this inference, by which they become proofs of each other’s existence:. 57

Upon the whole, then, either we have no idea at all of force and energy, and these words are altogether insignifi cant, or they can mean nothing but that determination of the thought, acquired by habit, to pass from the cause to its usual effect. 58

I do not think this talk of “meaning” should be taken any more strictly or literally here than Ayer is prepared to take it in Hume’s remarks about moral judgements. The kind of view he is propound-ing is structurally similar to his account of morality. The point is, fi rst, that there is no logical equivalence between causal statements containing the idea of necessity and any factual statements about the qualities or relations of objects which actually exist in the world independently of us. If there were, we could get the idea of neces-sity directly by observing those qualities or relations. But the idea of necessity for Hume has an entirely subjective source. It is an idea to which no fact or state of affairs in the world could correspond; there is nothing objective that could render that special modal aspect of our causal judgements either true or false. We can try to understand how and why we make causal judgements, but any attempt to express in equivalent terms what they mean would either leave the problematic modal idea unanalyzed or would eliminate it in favour of some purely extensional condition to which it is not equivalent.

Hume explicitly draws a parallel between our idea of necessity and our ideas of colours as “the modern philosophy” understands them. In each case we get the ideas only from something that occurs in our minds, and we mistakenly suppose that the qualities the ideas are ideas of are somehow conjoined with objects outside our minds, “tho’ the qualities be of such a nature as to admit of no such con-junction, and really exist nowhere”. 59 There is really nothing in the world corresponding to them. This “propensity” of the mind to project the features of which it gets internal impressions on to objects outside the mind that cause those impressions but do not really possess those features is equally present in the case of morality. There

57 Op. cit. , p. 76. 58 D. Hume, 1978, op. cit., p. 657. 59 Op. cit. , p. 167.


too Hume draws a parallel with our ideas of colours. 60 The same treatment is clearly to be applied both to the idea of goodness or badness and to the idea of necessity, and for the same reason. No equivalents for those ideas can be found in our experience, so the task is to explain what it is about human beings and their interactions with the world that leads them to get them. That is Hume’s main philosophical task, and it is an empirical investigation into a matter of contingent fact.

The same is true of his account of the idea of the continued and distinct existence of objects. There is no question of fully analyzing the content of that idea in terms of elements that are directly available in immediate experience, or giving a phenomenalistic reduction of talk of enduring objects in purely sensory terms. For that we need the idea of the identity of an object over time, and we are never presented with instances of such identity in our experience, nor can we construct the idea solely out of any relations which actually hold among the ingredients of our immediate experience. Hume identifi es those fea-tures of our experience which “give rise to” or “produce” the idea of continued and distinct existence, but they are not proposed as equiva-lent to its content. “Constancy” and “coherence” are certain forms of noticeable resemblance among our impressions which lead us into the thought of something continuing to exist, but “ ’Tis a gross illusion to suppose, that our resembling perceptions are numerically the same”. 61 Nothing in our experience does in fact continue to exist. “The fi ction of a continu’d existence, . . . as well as the identity, is really false”. 62

Hume’s task then is to explain how that false idea arises. There is no doubt that we have it. He fi rst identifi es (in Ayer’s phrase) “the circumstances in which we are induced to employ” it, and then seeks those “principles of the imagination” which are activated by the con-stancy and coherence among our impressions to eventually produce the idea in question. The “imagination” is the place to look because neither the “senses” nor “reason” alone or in combination could give us the idea. As in the case of morality and of causation, there is for Hume no logical equivalence between anything we can fi nd in our experience and the content of the new, richer idea we possess and want to explain. There is in that sense nothing in the world corres-

60 Op. cit. , p. 469. 61 Op. cit. , p. 217. 62 Op. cit. , p. 209.

Ayer’s Hume 135

ponding to our idea of continued and distinct existence—it is a “fi ction”—just as there is nothing corresponding to the idea of good-ness or badness or of causal necessity. The challenge in each case is to explain how we nevertheless come to think in those ways. We would not be human without them.

The same is true of our very thought of ourselves—the idea of personal identity. We think of a person as one thing that remains the same through time. But there is no hope of fi nding some relations which actually hold among perceptions and which constitute their all belonging to the same self or person. We are “nothing but a bundle or collection of different perceptions, which succeed each other with an inconceivable rapidity, and are in perpetual fl ux and movement”. 63 That is all there is. “There is properly no simplicity in [the mind] at one time, nor identity in different; whatever natural propension we may have to imagine that simplicity and identity”. 64 We do have such a “propension”, and Hume is interested in its source and in how it works to give us an idea of ourselves and others. For him that is all there is to be understood about our thoughts of persons.

The factual question about how we come to think in a certain way is all there is to be answered, in this case as in the others, because the world simply does not contain anything that fulfi lls the contents of these thoughts about it. We think of a person or a mind as made up of a series of perceptions, but “identity is nothing really belonging to these different perceptions, and uniting them together; but is merely a quality, which we attribute to them, because of the union of their ideas in the imagination, when we refl ect on them”. 65 In that sense, “the identity, which we ascribe to the mind of man, is only a fi ctitious one” 66 just as the ideas of goodness, of necessity, and of continued and distinct existence are “fi ctions” in relation to what actually holds in the world we apply them to. As in those cases, so with regard to personal identity, “The only question, therefore, which remains, is, by what relations this uninterrupted progress of our thought is produc’d, when we consider the successive existence of a mind or thinking person”. 67 And that is not a question of the analysis of concepts.

63 Op. cit. , p. 252. 64 Op. cit. , p. 253. 65 Op. cit. , p. 260. 66 Op. cit. , p. 259. 67 Op. cit. , p. 260.


Two caveats must be entered about this way of describing Hume’s strategy. First, “the world” which is here said not to contain such items is for Hume the world of immediate experience. What there is and what we can be presented with in experience are taken to be one and the same. That is because for Hume all our ideas must be derived from impressions; there can be no other source of our thoughts and beliefs. But the point about “analysis” remains. For all the import-ant ideas and attitudes Hume considers and spends most time trying to explain, they are interesting and challenging precisely because they go beyond anything that could be found among our impres-sions. They are the only ideas which Hume as a philosopher is inter-ested in. “Analysis” of their contents in terms of those original data is therefore out of the question. So “principles of human nature” or “propensities” of the mind or the imagination must be appealed to to explain how those richer thoughts can make their appearance in the mind.

Because of his offi cial commitment to the principle that every sim-ple idea is preceded into the mind by its corresponding impression, Hume sometimes troubles to fi nd a new impression that is produced by the mind or the imagination in the process of generating one of those rich, important ideas. That is true of his account of the idea of necessary connection, for example. An impression of refl ection, i.e., an impression generated by goings-on in the mind, is said to be the source of that idea. So in that case there is a sense in which the idea in question cannot be said to go beyond anything that is found among our impressions. The idea of necessity is produced directly by some-thing among our impressions, viz., an impression of necessity. But how that impression gets that particular content—what makes it an impression of necessity —is not something that Hume ever manages to explain. He knows what goes on in the mind when the idea of necessity makes its appearance, so what he calls the impression is just something that has to be there to produce the idea of necessity, given the principle that all simple ideas are derived from impressions and the further assumption that the idea of necessity is a simple idea.

When much more complicated processes of the imagination have to be appealed to, as in the case of continued and distinct existence, or personal identity, Hume does not bother trying to fi nd a single impression from which the problematic idea might be said to be derived. The highly elaborate “suppositions”, “feignings”, and “fi c-tions” that the mind has to indulge in are quite enough to have to

Ayer’s Hume 137

invoke, without their also having to produce individual impressions which then give rise to the idea. He does not acknowledge that the explanations he offers in these cases in terms of operations of the mind alone violate his principle of impressions as the source of all ideas. That could be because in these cases he regards the ideas as complex, not simple, so the principle does not apply.

The second warning about the claim that there is nothing in the world corresponding to these important ideas is that of course it does not represent what those who already have those ideas ordinarily believe. We make moral judgments, we believe causal statements, we think of objects and persons as enduring over time. That is the way we take the world to be, once we have those ideas. And our thinking of the world in those ways is just what Hume wants to understand and explain. He does not suppose that his philosophical demonstration of the “fi ctitious” character of all those ideas could ever persuade people not to employ them. We cannot help thinking in those ways, whatever the relation between those ideas and reality might be.

This is the source of the “melancholy” plight Hume fi nds himself in. His philosophical refl ections reveal to him the gap between the restricted data available to us in experience and the richness of the ideas we all somehow arrive at on the basis of them. Obviously only the operations or products of our own minds could be responsible. But then it seems impossible to “conceive how such trivial qualities of the fancy, conducted by such false suppositions, can ever lead to any solid and rational system”. 68 And “this sceptical doubt . . . is a malady, which can never be radically cur’d”. 69 It can never be cured by more or deeper thought because it is an unavoidable result of refl ecting on the relation between our important ideas and their source in our experi-ence; the ideas just are much richer in their content than anything available to us, so they could only have a subjective source. Those who never refl ect philosophically are spared Hume’s malady. They con-tinue to believe in a world of causal connections, enduring objects and persons, and good and bad actions and characters without ever real-izing the “illusion” involved in the possession of all those ideas. Hume’s “science of human nature” is a study of the position they are actually in, and how they get to be that way.

68 Op. cit. , p. 217. 69 Op. cit. , p. 218.


Another form philosophical “analysis” can take for Ayer is what he calls “the theory of knowledge”. Its aim is “to arrive at a satisfac-tory defi nition of knowledge; to determine what sorts of propos-itions can be known to be true; and to explain how these propositions can be known to be true”. 70 The fi rst is relatively unimportant and the second and third combine into the task of showing how we are justifi ed in accepting the propositions we accept. So far I have been discussing Hume’s account of our ideas or thoughts, but when we move to the question of our beliefs or knowledge it remains true that his main task is not one of analysis in this epistemological sense either. He demonstrates that none of our beliefs about matters of fact we are not observing at the moment is to any degree justifi ed by any present and past experience we might have, however rich and varied it might be. His conclusion is:

That there is nothing in any object, consider’d in itself, which can afford us a reason for drawing a conclusion beyond it; and, That even after the observation of the frequent or constant conjunction of objects, we have no reason to draw any inference concerning any object beyond those of which we have had experience . . . . 71

This is perhaps the most famous and certainly the most devastating element of Hume’s sceptical legacy. It represents a complete dead end for the project of showing our beliefs to be empirically justifi ed. If it is right, it really is impossible to go further in that direction.

Of course, Hume’s reaching that conclusion does not in itself show that his main task was not to answer the question of how our beliefs are empirically justifi ed. It might be said that he just gave a sweep-ingly negative answer to his main philosophical question. But that sceptical answer, for all its importance to subsequent epistemology, is only a small part of Hume’s philosophical interest in our beliefs in unobserved matters of fact. He wants to know how they arise, and why we believe the particular things we do, and he looks for answers among the conditions which “cause” or “produce” those beliefs in us. No such beliefs arise without some present experience, and that ex perience will lead to a particular belief only with the help of some relevant past experience. But the question is “Whether experience produces the idea [of something currently unobserved] by means of

70 A. J. Ayer, 1977, op. cit. , p. 58. 71 D. Hume, 1978, op. cit. , p. 139.

Ayer’s Hume 139

the understanding or of the imagination; whether we are determin’d by reason to make the transition, or by a certain association and rela-tion of perceptions”. 72 That is an empirical, causal question about how the mind works.

Hume eliminates “the understanding” as the source of the “transi-tion” and establishes that “reason”, even “reason” combined with past experience, is not what “determines” us, by showing that past experience, however rich, can never give us any reason for believing what we do about what we are not observing at the moment. And that clears the fi eld for his positive project of discovering what does in fact “determine” us to believe what we do, and how. The “imagin-ation” is the only possible source; nothing more than “a certain asso-ciation and relation of perceptions” can be appealed to. And the task then is to describe those connections among perceptions, to identify the “principles of the imagination” that are actually at work in the genesis of our beliefs. That is not a question of “analysis” in Ayer’s sense of “theory of knowledge”.

Again, the idea that no beliefs are ever justifi ed by past experience is not something that those who never refl ect philosophically on their position believe. In everyday life we cannot help regarding some beliefs as more reasonable than others and forming and acting on one set of expectations rather than another. Hume denies none of that. His question is how we do it; what leads us to think and act in those ways. His answer confl icts with what we all believe in everyday life. It does not just analyze everyday or scientifi c concepts; it reveals what there really is in the world that they purport to represent, and why and how we continue to employ them despite our having nothing we could acknowledge as a defence of our thinking, believing, and acting as we do.

Hume would not suggest that he had done all there is to be done towards understanding human thought, belief, and action, even at the very high level of generality on which he operates. His project is in that sense still wide open; no dead end threatens. But he pursues that project with a number of specifi c views about the nature of our thoughts, beliefs, and feelings—indeed, about what it is to have a cer-tain thought, belief, or feeling—which can seem too obscure or restrictive to allow even for fruitful description of the kinds of human phenomena he wants to explain. It would still be fully within the

72 Op. cit. , pp. 88–89.


spirit of Hume’s enterprise to jettison the rigid theory of impressions and ideas in terms of which he conceives of the mind, and try for a richer and therefore more realistic description of the actual position ordinary human beings naturally fi nd themselves in. It is an illusion to suppose that at present we have a proper understanding even of the nature of the very phenomena that we would expect a science of the human to explain—such things as our holding causal beliefs and making inferences from past experience, our possessing a conception of an independent world, our thinking of ourselves as persons who endure through time, our evaluating actions and characters, even our perceiving something that is right before our eyes. It will be time to search for explanations of such phenomena—if that is what will then be wanted—once we have described them more accurately and so have a better understanding than we do at present of just what those phenomena amount to. That is only one of the ways in which Hume’s philosophical project could be extended.

There is no reason to suppose that even that job would be simple, or one-dimensional. And whether any of it, or which parts, could properly be given the name ‘philosophy’, is perhaps of little concern. But it is continuous with what has been called philosophy. And it remains within what Hume calls “the science of human nature”.

IV

I have tried to explain why I think Ayer’s understanding of Hume tends to obscure or distort what Hume is really up to and what is most characteristic and most fruitful in his philosophy. But it is not just a question of the interpretation of a long-dead philosopher. Ayer’s philosophical aims, along with his closeness to Hume and his conception of the British tradition in which he sees himself carrying on Hume’s work, seem to me to leave him too close for comfort to the philosophical dead end that so clearly threatens in that direction.

The threat comes from what Ayer shares with Hume. They both are impressed by the poverty of the data available to us in immediate experience in relation to the richness of the ideas or concepts we come to form on that basis. That gap is the source of Hume’s explanatory challenge: how do we actually get across it? Ayer at fi rst aspired to reduce those thoughts to their sensory bases, so no gap would be left. That was the point of the thesis of Language, Truth and Logic that all empirical statements are equivalent to statements about sense-data.

Ayer’s Hume 141

He abandoned that phenomenalism when the promised equivalences were not forthcoming; no conditions expressed in purely sense-datum terms could plausibly be said to be necessary for the existence of a physical object, nor could any be said to be suffi cient. 73

But even if phenomenalism or other forms of reductionism had worked they would not have accounted fully for our understanding of the concepts they served to reduce. Talk of physical objects was to be equivalent to talk of sense-data which would be had under certain conditions, not just of a collection of actual sense-data alone. And that involves the idea of natural necessity or law-like connection, or at any rate something non-extensional typically implied in subjunct-ive conditionals. Without that special modal feature reductionism would never have got off the ground. Something like Berkeley’s God would have been needed to make it plausible. But that necessity or nomic connection is not something we can ever be aware of in an individual sense-datum or in any collection of them. That is a point on which Ayer has always agreed with Hume. But it represents an obstacle for the purely “analytic” project of specifying in experiential terms what that necessity amounts to—what it means. Phenomenalistic reductionism offers no answer; it makes essential use of the idea without explaining its content. To try to reduce necessity or law- likeness in turn to discoverable features of experience such as “regu-larity” or the “constant conjunction” of properties would in effect obliterate the very distinction between extensional generalizations of fact and nomic or law-like generalizations that is needed to leave any hope for plausible analytic reductions of any interestingly problem-atic notions. No such obstacle arises for Hume. He insists on the impossibility of defi ning necessity in terms of what is available in immediate experience, and concentrates instead on explaining how in fact we come to think that way, despite the obvious gap.

Ayer has suggested avoiding the defi nitional obstacle in the case of natural necessity by concentrating instead on the different attitudes people might take to generalizations they accept. 74 No explanation of the notion of natural law along these lines would yield an analytic equivalence: human attitudes are neither necessary nor suffi cient for

73 A. J. Ayer, 1977, op. cit. , pp. 106–107. 74 A. J. Ayer, “What is a Law of Nature?”, in his The Concept of a Person and Other

Essays , London, Macmillan, 1963. pp. 230–234 .


laws of nature to hold. So whatever illumination might be gained would not help further the explicit reductionism of Language, Truth and Logic . Ayer has in any case moved away from that project in general in favour of a looser relation between experience and the pos-sibility of thought and belief which nevertheless preserves the spirit of empiricism he has always defended. That has meant in effect a shift away from questions of meaning and understanding, strictly speaking, to questions of belief and justifi cation. The focus is more and more on “theory of knowledge”: how our beliefs or theories about the world are sustained or supported by our experience.

Here too there remains a huge but familiar obstacle. Ayer now sees all those beliefs which go beyond the strict contents of immediate experience as something like hypotheses or postulates which gain their empirical support from their explanatory function. Our view of the world is to be vindicated in more or less the same way as a scien-tifi c theory is thought to be supported by its success. All forms of this widely held type of view still face the Humean challenge of showing how actual experience can support any inference that goes beyond it. Most philosophers simply ignore it. Not Ayer. But he is too close to Hume to meet the challenge, or even to defl ect it.

He accepts, albeit in slightly qualifi ed form, Hume’s principle of the “atomicity” of observed matters of fact: the inference from one matter of fact to another is never demonstrative. 75 And he agrees that we have reason to believe a matter of fact we are not observing at the moment only if it is connected in some way with what we are observ-ing now or have observed in the past. But he also accepts what he has always rightly seen to be the further Humean requirement that not only must there be such a connection or similarity between the observed and the unobserved, we must in addition have reason to believe that there is. That is the fatal step that leads to the general sceptical conclusion. The only reason we could have for believing in a connection or similarity between the observed and the unobserved must be found in what has already been observed, and so to rely on it at any given time would be, in Hume’s words, “evidently going in a circle, and taking that for granted, which is the very point in question”. 76 The conditions necessary for justifi ed belief in unob-served matters of fact can therefore never be fulfi lled. There is no

75 A. J. Ayer, Probability and Evidence , London, Macmillan, 1973. pp. 6–10 . 76 D. Hume, 1975, op. cit. , p. 36.

Ayer’s Hume 143

evading this conclusion on Hume’s, and Ayer’s, premisses. Given that starting point, it is impossible to go further in the direction of justify-ing any beliefs or theories about the world. Ayer seems to endorse Russell’s gloomy verdict when he writes:

What we want and cannot obtain, except by circular argument, is a justifi ca-tion for our actual interpretation of the lessons of the past; a justifi cation for adhering to a special corpus of beliefs. That we cannot obtain it is an insight which we owe to Hume. 77

77 A. J. Ayer, 1980, op. cit. , p. 74.

7

Hume’s Scepticism:

Natural Instincts and Philosophical Refl ection

Philosophy for the Greeks was not confi ned to abstract theory but was also meant as a guide to the living of a good human life. Hume was steeped in the literature of antiquity. I think there is a close kinship between his conception of philosophy and that ancient con-ception. It is something we tend to miss when we look back at Hume for our own purposes from here and now. I want to try to bring out the connection by identifying what Hume thought philosophical refl ection could reveal about human nature, and what he, therefore, thought the point, or the human good, of philosophical refl ection can be. His own direction in philosophy took him closest to that way of life said to have been achieved by certain ancient sceptics.

Some of the most personal, and the most moving, passages in all of Hume’s philosophical writings appear in that puzzling, confessional last section of Book One of the Treatise prosaically entitled “Conclusion of this Book.” Earlier in the Book he has presented his accounts of the origins of some of the most fundamental modes of human thought — causality, identity, enduring objects, the self — and here he steps back to ponder what he calls the “leaky, weather-beaten vessel” in which he is about to launch into the “immense depths of philosophy” 1 still before him in Books Two and Three. Given the discouraging conclusions he has reached earlier about the human understanding, he “despairs” at the “wretched condition, weakness, and disorder of the faculties” 2 which he and all other humans have to rely on. He is stricken with “melancholy,” he fancies himself “some strange, uncouth, monster,” he fi nds nothing but “doubt and ignor-ance” in his mind: 3 “Every step I take is with hesitation, and every new

1 D. Hume, A Treatise of Human Nature , ed. L. A. Selby-Bigge (Oxford: Oxford University Press, 1978), p. 263 .

2 Treatise , p. 264. 3 Ibid.

Hume’s Scepticism 145

refl ection makes me dread an error and absurdity in my reasoning.” 4 Even after “the most accurate and exact” enquiry, he can give no rea-son why he should assent to any particular conclusion; he simply feels a strong “propensity” to consider objects “ strongly in that view, under which they appear” to him at the moment. 5

When we trace up the human understanding to its fi rst principles, we fi nd it to lead us into such sentiments, as seem to turn into ridicule all our past pains and industry, and to discourage us from further enquiries. 6

But by the end of that same section, less than ten pages later, he nev-ertheless fi nds his spirits raised. His renewed hope of bringing the neglected science of man “a little more into fashion” has somehow served to “compose” his “temper from that spleen, and invigorate it from that indolence, which sometimes prevail” 7 upon him in intense philosophical refl ection. He is once again in an “easy disposition,” and feels it proper after all to indulge his “inclination in the most elaborate philosophical researches.” 8 He continues to philosophize.

What is responsible for such a quick and complete reversal? And why does Hume even mention the matter in what otherwise looks like a purely theoretical treatise on philosophy? I think trying to answer these questions is no mere biographical exercise. I suggest that the interlude is not to be understood as simply an embarrassing personal confession of the youthful author’s loss of nerve in the face of his negative conclusions. I think it should be taken seriously as an expres-sion of Hume’s philosophy, not just a description of his odd state of mind while writing it. And although the personal, confessional voice is absent from the more polished pages of the later Enquiry Concerning Human Understanding , I think the same theme is taken up there and developed more thoroughly and more systematically. If we can understand what goes on in that section of the Treatise and in the corresponding parts of the fi rst Enquiry we will understand a great deal about what Hume thought philosophy could be, and do.

We must start, as Hume does, with the plight he fi nds himself in at the end of Book One of the Treatise . His examination of the nature of belief and of the role of reason in the genesis of beliefs has shown

4 Treatise , p. 265. 5 Ibid. 6 Treatise , p. 266. 7 Treatise , p. 273. 8 Ibid.


that even when he is most careful and cautious he will have no more reason to believe any particular conclusion than to disbelieve it; he will simply feel a strong propensity to view things in one way rather than another. His experience presents him with certain regularities, and habit leads him to expect them to continue in the future. That alone is the source of all his beliefs about matters of fact. If certain ideas did not get “enlivened” by the imagination, and thereby trans-formed into beliefs, he would never give his assent to anything and never be able to extend his view beyond what is immediately present to his consciousness at the moment. His ability to do so is “founded on” nothing more than the operations of the imagination which serve to make some ideas more “lively” or “vivacious” than some others.

This is in part the discovery that certain operations of the mind must be present if we are able to think and believe and act in the ways we do. In itself that is not something which should throw us into despair. Surely we must acknowledge that some features or oper-ations of the mind must be at work if our minds are functioning at all. What specifi cally troubles Hume is that it is “the imagination” that is at work, and that the quality of the imagination by which “the mind enlivens some ideas beyond others” is “seemingly . . . so trivial, and so little founded on reason.” 9

In calling them “trivial” Hume does not mean that the properties of the imagination are trivial in their effects. Without the operations of the imagination which he is interested in, we could not think at all. Some of those principles are “changeable, weak, and irregular,” but there are others which are “permanent, irresistable, and universal,” and which serve as “the foundation of all our thoughts and actions, so that upon their removal human nature must immediately perish and go to ruin.” 10 So the principles of the imagination are not trivial for human nature; they make it what it is.

In saying that the quality of the imagination by which “the mind enlivens some ideas beyond others” is “so trivial, and so little founded on reason” Hume does not mean that it leaves us uncertain or might lead us astray. He does worry in the Enquiry that “custom . . . like other instincts, may be fallacious and deceitful,” 11 but that would be

9 Treatise , p. 265. 10 Treatise , p. 225. 11 D. Hume, Enquiries Concerning Human Understanding and Concerning the

Principles of Morals , ed. L. A. Selby-Bigge (Oxford: Oxford University Press, 1975), p. 159 (my italics) .


cause for vigilance, not despair. Nor does he fi nd himself uncertain whether there are causal connections, enduring objects, and persons. He is not doubt-ridden; he cannot help believing in them. Even his philosophical accounts of belief, reason, causality, and the existence of objects remain convincing to him. That is itself part of his diffi culty.

Hume thinks he has discovered that the imagination is a principle “so inconstant and fallacious” that it will inevitably “lead us into errors, when implicitely followed (as it must be) in all its variations.” 12 For example, the power of the imagination is what makes us reason causally, and also what makes us believe in external objects that are not perceived. But those two operations can sometimes confl ict. He thinks it is not “possible for us to reason justly and regularly from causes and effects, and at the same time believe the continu’d exist-ence of matter.” 13 That is not all. It is the discovery that “the memory, senses, and understanding are . . . all of them founded on the imagination” 14 that would “seem to turn into ridicule all our past pains and industry.” 15 In our attempts to understand the world we push on to discover what really makes things happen as they do. We seek “that energy in the cause, by which it operates on its effect; that tie, which connects them together.” 16 But Hume’s theory of man has shown that “this connexion, tie, or energy lies merely in ourselves” and is only a determination of the mind acquired by custom. 17 It is “an illusion of the imagination” to think that we have any insight into the connection even in the most familiar, everyday cases, let alone in more esoteric matters at the frontiers of science. 18 Causality and most of our other important ideas have been exposed as mere “fi ctions” or “illusions.” That is one source of Hume’s despair.

The predicament causes him despair because it presents him with the problem, as he puts it, of “how far we ought to yield to these illusions”; 19 and he sees no way to answer the question. To assent to “every trivial suggestion of the fancy,” would lead to so many “errors, absurdities, and obscurities” that we would be “asham’d of our

12 Treatise , pp. 265–266. 13 Treatise , p. 266. 14 Treatise , p. 265. 15 Treatise , p. 266. 16 Treatise , p. 267. 17 Treatise , p. 266. 18 Treatise , p. 267. 19 Ibid.


credulity”; but to try to reject the imagination and “adhere to the understanding” alone would lead nowhere, since “the understanding, when it acts alone, and according to its most general principles, entirely subverts itself, and leaves not the lowest degree of evidence in any proposition.” 20 Hume thinks he has already shown that if we believed only what we have good reason to believe we would believe nothing. So we have “no choice left but betwixt a false reason and none at all.” 21

I think what Hume has in mind in speaking of the seemingly “triv-ial” qualities or operations of the imagination is that those oper-ations are found to be only trivially or accidentally connected with the truth of the beliefs which are their effects. Neither the word “triv-ial” nor the word “accidental” is quite right, but his point is expressed most clearly at the end of that section of the Treatise called “Of scep-ticism with regard to the senses” (I,iv.2). Having explained how vari-ous operations of the imagination lead us from our sense impressions to a belief in the continued and distinct existence of objects, Hume confesses that he is at the moment inclined to place little or no faith in his senses or imagination. He says he “cannot conceive how such trivial qualities of the fancy, conducted by such false suppositions, can ever lead to any solid and rational system.” 22 That is because those properties of perceptions which combine to “produce the opin-ion” of continued existence – namely, constancy and coherence – “have no perceivable connexion with such an existence.” 23 What produces our beliefs has no “perceivable connexion” with their truth. I think that is the best description of the unsatisfactory position Hume fi nds himself in.

It means that, given the ways our minds work, and given what our experience presents us with, we will inevitably come to believe what we do whether that belief is true or not. There is no connection between our believing in the things we believe in – causal connec-tions, enduring objects, and so on – on the one hand, and the exist-ence of such things, on the other. All our beliefs in such things could be false or “illusory” even though it remains perfectly intelligible how we come to acquire them. That is the disturbing feature of Hume’s explanations of the origins of our beliefs in terms of the operations

20 Treatise , p. 267. 21 Treatise , p. 268. 22 Treatise , p. 217. 23 Ibid.


of the imagination. He fi nds that our most important beliefs have a “trivial” or “accidental” origin in the sense that our having those beliefs bears no relation to their being true or to our having any rea-son to believe them. Their truth or reasonableness does not fi gure in the explanation of their origin. To say that their origin is “acciden-tal” is of course not to say that the beliefs have no causes at all. It is to say that, given that we have the beliefs in question, it is at best an accident if the beliefs happen to be true; their being true, or their being false, makes no difference one way or the other to our having the beliefs. We would have had them in either case. And that is a dis-turbing position to fi nd oneself in when refl ecting on one’s beliefs.

Probably no one is in a position to say with full confi dence that none of his beliefs is “accidentally” produced in that way. Perhaps each of us could go so far as to say that we know it must be true of some of our beliefs. But Hume actually specifi es a number of very important beliefs and shows of each of them that they only have an “accidental” origin in the sense in question. He shows that that is so for all our beliefs in causal connections, the independent existence of objects, and the enduring self. His theory of their origins explains how we get those beliefs without implying that they are true, or even reasonable.

In fact it is worse than that. Hume does not just fail to assert or to imply the truth of those fundamental beliefs in explaining their ori-gins; he explicitly denies it. Causal necessity, he says, does not exist in objects. 24 It is a “false opinion” or a “gross illusion” to suppose that objects remain numerically the same after an interruption in our per-ceptions. 25 And “there is properly no simplicity in [a mind] at one time, nor identity in different.” 26 If we became convinced of these negative conclusions, as Hume’s philosophy is meant to convince us, but we nevertheless continued to believe in causal connections, inde-pendent objects, and the self, we might well become, with Hume, “asham’d of our credulity,” and we might resolve to bring our beliefs more into line with reason and with what we have come to see is the way things really are.

But Hume’s “despairing” predicament is that no such resolution could have any effect. We will remain “asham’d of our credulity” if we

24 Treatise , p. 165. 25 Treatise , pp. 209, 217. 26 Treatise , p. 253.


submit to every “trivial suggestion” of the imagination, but it would be literally fatal if we could somehow avoid the imagination and per-form the impossible feat of believing only what is based on reason or solid evidence. Given Hume’s negative conclusions about reason, that would mean that we would have no beliefs at all, except perhaps about what is immediately present in our experience at the moment. And that would mean that life would be impossible. Without beliefs:

All discourse, all action would immediately cease; and men remain in a total lethargy, till the necessities of nature, unsatisfi ed, put an end to their miser-able existence. 27

There seems to be nothing that can be done. Neither side is tolerable; and yet there seems to be no escape. Despair arises not just from discovering the “illusory” or “fi ctional” character of our most import ant beliefs, but from the recognition that we simply cannot avoid indulging in such “fi ctions” if we are to have any beliefs at all. Hume’s theory taken as a whole shows that most of our beliefs must be wrong or unreasonable; given the way we are, we could not have those beliefs unless that were true.

But still, the recognition of this depressing state of affairs is only part of what is needed to understand what Hume is most concerned to show in his philosophical works. The plight or dilemma alone is not enough. What is even more important is the way the dilemma is resolved. He confesses that he simply does not know “what ought to be done” in choosing between “a false reason and none at all.” But he does tell us “what is commonly done.” 28 As a matter of fact, he says, “this diffi culty is seldom or never thought of; and even where it has once been present to the mind, is quickly forgot, and leaves but a small impression behind it.” 29

What kind of solution is that? How does it help to resolve the pre-dicament for us to be told that most people never recognize the plight they are in, or if they do, that they soon forget about it? This might look like a comment on the idleness or irrelevance of philosophy, of its lack of impact on what human beings actually do. Hume some-times says things which give that impression: for example, his wry observation that “errors in religion are dangerous; those in philosophy only ridiculous.” 30 But I think Hume’s point here is no passing jibe at

27 Enquiry , p. 160. 28 Treatise , p. 268 (my italics). 29 Ibid. 30 Treatise , p. 272.


philosophy. It is a very important observation; and, for Hume, a philo sophical observation.

For Hume it is essential to one’s understanding of human nature, and to one’s life – and therefore philosophically important – to rec-ognize the force of natural instinct over the deliverances of reason. It is important to see what actually happens to someone who is rightly convinced of Hume’s negative conclusions and is thereby thrown into the plight he describes. What always in fact happens is that “nature” quickly dispels the clouds that “reason” is incapable of dispelling. 31 We fi nd ourselves “absolutely and necessarily determin’d to live, and talk, and act like other people in the common affairs of life,” 32 whatever our philosophical conclusions might have been and whatever doubts or despair we might have arrived at in our philo-sophical studies. Hume is right to emphasize that philosophers too eat and drink and converse and sometimes even play backgammon as other mortals do. A thoughtful person can perhaps be brought by philosophical refl ection temporarily to “renounce all belief and opinion,” 33 but his resolution, however strong, cannot last very long.

External objects press in upon him: Passions solicit him: His philosophical melancholy dissipates; . . . 34

The bent of his mind relaxes, and cannot be recalled at pleasure: Avocations lead him astray: Misfortunes attack him unawares: And the philosopher sinks by degrees into the plebeian . 35

Section IV of An Enquiry Concerning Human Understanding is called “Sceptical Doubts Concerning the Operations of the Under-standing.” That is where Hume establishes the negative conclusions parallel to those reached in Book One of the Treatise : in particular, “that it is not reasoning which engages us to suppose the past resem-bling the future, and to expect similar effects from causes which are, to appearance, similar.” 36 The next section is surprisingly called “Sceptical Solution of These Doubts.” What is the solution Hume offers? Why

31 Treatise , p. 269. 32 Ibid. 33 D. Hume, Dialogues Concerning Natural Religion , ed. N. Kemp Smith (Indianapolis:

Bobbs-Merril, 1947), p. 132 . 34 Ibid. 35 Dialogues , p. 133. 36 Enquiry , p. 39.


does he call it a “sceptical” solution? I will return to this second ques-tion after we see what the “solution” is.

The “solution” is the one I have already mentioned from the Treatise – that the negative philosophical conclusion about our beliefs and our reasoning will never in fact undermine the beliefs and reasonings of everyday life. The inferences we actually make from our experience will not be affected one way or the other by the true discovery that in all such so-called reasonings there is always a step which is not supported by argument or by any process of the under-standing. What that discovery shows is that, since the mind is not engaged by reason, it “must be induced by some other principle of equal weight and authority; and that principle will preserve its infl u-ence as long as human nature remains the same.” 37 That principle is “Custom or Habit”; 38 it is a fundamental principle of human nature which we cannot pretend to explain further.

Hume therefore sums up the fi rst part of the section called “Sceptical Solution of These Doubts” by declaring that “the conclu-sion of the whole matter” is quite simple. It is that “all belief of mat-ter of fact or real existence” is simply the “necessary result” of a receptive mind being placed in certain circumstances; coming to believe something after having observed a constant conjunction of objects of two kinds is “an operation of the soul, when we are so situ-ated, as unavoidable as to feel the passion of love, when we receive benefi ts; or hatred, when we meet with injuries.” 39

All these operations are a species of natural instincts, which no reasoning or process of the thought and understanding is able either to produce or to prevent. 40

And that is really the end of the matter. That is the “solution.” “At this point,” Hume says, “it would be very allowable for us to stop our philosophical researches.” 41 The most important general point about human nature has been made.

The point is that “nature breaks the force of all sceptical argu-ments in time,” 42 or that “nature is always too strong for principle.” 43

37 Enquiry , pp. 41–42. 38 Enquiry , p. 43. 39 Enquiry , p. 46. 40 Enquiry , pp. 46–47. 41 Enquiry , p. 47. 42 Treatise , p. 187. 43 Enquiry , p. 160.


No philosophy committed only to Hume’s negative conclusions could possibly be put into practice as it stands. But this does not imply that we should therefore pay no attention to those negative conclusions or to the philosophical arguments which lead to them. It does not follow that any consideration of such reasonings must be completely idle and without effect.

For Hume we must see and appreciate both the doubts and nega-tive conclusions and the so-called “solution” if we are to discover the important truth about human nature. We must fi rst fi nd the negative “philosophical” or “sceptical” view completely convinc-ing – indeed, unanswerable – in order to perceive and acknowledge the sheer force of custom, habit, or instinct which can submerge it with hardly a trace. If we never philosophized and reached the “sceptical” conclusion, that discovery would be lost to us. We might fi nd out somehow that there are certain things which we cannot help believing, but we would never understand why, or how. And if Hume is right about the overwhelming force of instinct, if we tried to accept the negative “philosophical” view by itself, we would fi nd it intolerable. Even in our purely philosophical moments, we will at best fi nd ourselves in Hume’s desperate plight; and even then those moments and that plight cannot last very long. Any doubts we arrive at will be unstable and will produce only “moment-ary amazement and confusion.” 44 But if we never arrived at any of those doubts in the fi rst place, the “solution” would be no solution at all. We must be “once thoroughly convinced of the Pyrrhonian doubt, and of the impossibility, that anything, but the strong power of natural instinct, could free us from it.” 45 We must get both. So it is essential for Hume to present both sides: both the “doubts” and the “sceptical solution” of those doubts. That is what he does explicitly in Sections IV and V of An Enquiry Concerning Human Understanding . And that is what I think he does less explicitly but more personally and more dramatically in the “Conclusion” of Book One of the Treatise .

I turn now to the point, or the pay-off, of “resolving” the confl ict in Hume’s way. It reveals something important about his conception of his philosophical task, and so brings him closer to the ancient sceptics.

44 Ibid. 45 Enquiry , p. 162.


Hume adopts and recommends what he calls “the sceptical phil-osophy,” but he uses the words “sceptical” and “scepticism” in a variety of ways. He speaks of the “sceptical reasonings” 46 or “scep-tical arguments” 47 which lead him to what he calls “total scepticism” 48 or “sceptical” 49 or “Pyrrhonian doubt.” 50 I think there is no doubt that Hume as a philosopher believes that those negative conclusions are correct; they represent a signifi cant part of his contribution to philosophy. But his recommendation of what he calls “scepticism” is not simply a recommendation of the acceptance of those sceptical conclusions. He knows that those conclusions are not believable in everyday life. Nature will always submerge them in time. So he does not recommend them as a set of principles to be adopted and used to guide our thought and conduct.

When he acknowledges in the “Conclusion” of Book One of the Treatise that he must inevitably “yield to the current of nature” he goes on to remark that “in this blind submission I shew most perfectly my sceptical disposition and principles.” 51 And in An Enquiry Concerning Human Understanding the “solution” provided by our natural instincts to the philosophical doubts or denials is described as a “sceptical solu-tion.” This brings me to the second of the two questions I raised earl-ier: why does Hume call the “solution” which consists in yielding to the inevitable force of nature a “sceptical solution”?

I believe it is because submission to the forces of nature under certain conditions can be a form of “scepticism” in the sense of the ancient sceptics who reputedly found a trouble-free way of life in fol-lowing their natural inclinations. Some of them appear to have thought that they could achieve that blessed state only if they had no conviction or beliefs as to how things are. Hume thought no one could live without convictions or beliefs, but he saw himself in the old sceptical tradition at least in his recommendation of acquiescence in the face of what is most fully “natural.” Richard Popkin has for this reason called Hume a “Pyrrhonist.” 52 Even if human belief and

46 Treatise , p. 186. 47 Treatise , p. 187. 48 Treatise , p. 268. 49 Treatise , p. 218; Enquiry , Section IV. 50 Enquiry , p. 162. 51 Treatise , p. 269. 52 Richard E. Popkin, “David Hume: His Pyrrhonism And His Critique of Pyrrhonism,”

The Philosophical Quarterly 1 (1951) . Reprinted in V. C. Chappell, ed. Hume (Garden City, NY: Doubleday, 1966) . Page references here are to the Chappell volume.


reasoning cannot be avoided, if they are as natural and predictable as any other events in the world, Popkin thinks there is no need to try to avoid them on sceptical principles. Accepting the inevitability of beliefs and convictions would be “merely a legitimate extension of the Pyrrhonian principle of living according to nature.” 53

But it is equally important in Hume’s view to acknowledge the naturalness and virtual inevitability of refl ecting philosophically on the human condition. At the end of Book One of the Treatise he confesses that he fi nds himself “naturally inclin’d” towards philo-sophical refl ection; he “cannot forbear” indulging in it; he is “uneasy” to fi nd he does not understand certain things. 54 Even if he could man-age to ignore it, he says, “I feel I shou’d be a loser in point of pleas-ure; and this is the origin of my philosophy.” 55 Human beings are easily led to philosphize.

It is true that “profound philosophical researches” will always leave us dissatisfi ed. We will inevitably be led to the “temporary mel-ancholy and delirium,” 56 of what Hume calls “Pyrrhonism, or exces-sive scepticism.” 57 But arriving even for a moment at such unstable and incredible results can nevertheless be a good thing. We must fi rst see and accept the truth of that “excessive scepticism” in order fully to appreciate the real force of nature or the imagination over reason. And given our undeniable natural instincts, the process of following out the reasoning to that excessive scepticism can itself have good effects which cannot be achieved in any other way. It can lead to what Hume calls “a more mitigated scepticism or academical philosophy, which may be both durable and useful” when the excessive scepticism is “in some measure, corrected by common sense and refl ection.” 58 So the “excessive” position which Hume calls “Pyrrhonism” can have good and lasting effects even if it cannot be permanently believed or followed. To ask what those effects are, and why they are good, is to ask what Hume thinks “ mitigated scepticism or academical philoso-phy” is, and why he recommends it.

53 Popkin, p. 89. 54 Treatise , pp. 270–271. 55 Treatise , p. 271. 56 Treatise , p. 269. 57 Enquiry , p. 161. Whether Hume is historically correct in calling the view he has in

mind “Pyrrhonism,” and whether Pyrrho or any of his followers were in fact “exces-sive” sceptics in this sense, are questions I do not try to answer here.

58 Enquiry , p. 161.


First, it is important to see that “mitigated scepticism” as Hume understands it is not a set of doctrines or truths. It is something we can fi nd ourselves with, or a state we can fi nd ourselves in, when the refl ections leading to excessive scepticism, have been tempered or mitigated by our natural inclinations. So mitigated scepticism is not just a qualifi ed or watered-down version of the complete or excessive scepticism which Hume arrives at in his uncompromis-ing, negative philosophizing. In particular, it is not the thesis that we can never be absolutely certain of anything but can at most have beliefs which are only probable. 59

Popkin calls Hume a Pyrrhonist, but Hume speaks of Pyrrhonism as an unacceptable “excessive” form of scepticism. When he is in his study and follows those “profound reasonings” that lead to it he fi nds himself fully convinced, so he does not regard Pyrrhonism as “exces-sive” in the sense of going beyond the truth in what it says. It is rather because it cannot be put into practice; the “doubts or scruples” it rightly arrives at cannot in fact have “any constant infl uence on the mind.” 60 A Pyrrhonist “must acknowledge, if he will acknowledge anything, that all human life must perish, were his principles univer-sally and steadily to prevail.” 61 Of course, there is no danger of that. The inevitable force of nature is always too strong for the Pyrrhonian conclusions to be put into effect. So in Hume’s view, no one could possibly be a Pyrrhonist. In particular, Hume himself could not. He too, like everyone else, must follow nature.

Popkin acknowledges that what he calls “epistemological Pyrrho-nism” is the only possible outcome of philosophizing in Hume’s way, and he agrees with Hume that it can never cause us to adopt what he calls “a practical Pyrrhonian attitude.” 62 But he nevertheless regards Hume as a Pyrrhonist, in fact as a “consistent Pyrrhonist,” because he simply does what comes naturally and so follows the ancient sceptical or Pyrrhonian tradition of living under “the guidance of Nature” alone. For Popkin, “the true Pyrrhonist is both a dogmatist and a sceptic,” 63 but of course not at the same time. It is a matter of alternating “moods,”

59 Popkin describes Hume’s mitigated scepticism this way in his entry “Scepticism” in The Encyclopedia of Philosophy , ed. P. Edwards (New York: Collier-Macmillan, 1967), vol. 7, p. 455 .

60 Enquiry , p. 160. 61 Ibid. 62 Popkin, p. 94. 63 Popkin, p. 95.


even of “split personality.” 64 Since he is led to believe things by the force of nature, the Pyrrhonist will be “as dogmatic and as opinionated as one is naturally inclined to be.” 65 In one mood, the necessities of nature overcome him, and he is “dogmatic”; he asserts and believes things. In another mood, the obstacles to reliable knowledge impress him, and he is “sceptical”; he sees there is little or no reason to assert or believe any-thing. But both moods are produced by natural forces. “In being entirely the product of nature he welds his schizophrenic personality and phil-osophy together. He believes whatever nature leads him to believe, no more and no less.” 66 This is said to yield a “quietude” or peace of mind that is not open to what Popkin calls the “orthodox Pyrrhonist.” Such a person would want to be undogmatic and to suspend judgement about everything, but that would put him into continual confl ict with nature, and he could not succeed. Nature would inevitably force on to him the very beliefs he offi cially doesn’t want. And he would be unhappy at his failure. But the Humean Pyrrhonist as Popkin sees him will have a “peaceful attitude” towards any “dogmatic view” he adopts, “since how he feels about it will be natural, and there will be no attempt to combat his inclinations.” 67

I think this picture of an easy-going, peaceful way of life does not really capture everything that Hume is getting at in his recommenda-tion of “the sceptical philosophy.” Following nature is certainly involved in being a mitigated sceptic, but that is not all there is to it. Someone who was “sceptical” only some of the time, and “dogmatic” the rest of the time, would not really be following a sceptical way of life. He would simply be a creature of nature. And every creature is a creature of nature. But not everyone leads a sceptical way of life. What is required for the kind of scepticism Hume recommends is not just following nature, but following nature while at the same time acknowledging or realizing nature’s inescapable force. We can achieve that realization only by fi rst becoming convinced of the “profound reasonings” of the philosophers. We see that we can have no reason to believe any of the things we so naturally believe, and then we fi nd, in our very thoughts and actions as we leave our studies, that those

64 Popkin, p. 98. 65 Popkin, p. 95. 66 Ibid. 67 Popkin, p. 94.


undeniable negative conclusions immediately give way in the face of the overwhelming power of nature. We cannot continue to endorse or express those conclusions in the ways we live our lives, but they nevertheless continue to have certain effects.

This living in the acknowledgement of, or acquiescing in, both the profound philosophical “doubts” and the natural “solution” of those doubts is what Hume calls “mitigated scepticism.” It is something that is “ consequent to science and enquiry.” 68 It is a “natural result of the Pyrrhonian doubts and scruples,” 69 and it could not have been achieved without them. It is a state which arises when “excessive” Pyrrhonism is tempered or mitigated by our natural inclinations, as it inevitably will be. And it is therefore a state we can fi nd ourselves to be in all the time. It is not just a passing mood; it can be a way of life. But it would not be possible without both of the ingredients Hume stresses. What he calls Pyrrhonism alone is impossible; it can-not be lived. But without the startling effects of that excessive Pyrrohonism our natural inclinations operating on their own would not lead to a truly sceptical way of life either. They would not give us anything like the tranquillity or quietude or peace of mind sought and perhaps even found by the sceptics of antiquity.

Hume in effect explains how his mitigated scepticism might lead to a kind of tranquillity. He distinguishes two different sorts of durable and useful effects which the pursuit of excessive scepticism and the inevitable force of our natural inclinations can combine to produce. The fi rst is a greater “modesty and reserve” in all our thoughts and reasonings. Simply following the urges of nature is not best for “the greater part of mankind”; they “are naturally apt to be affi rmative and dogmatical in their opinions.” 70 They are uneasy and impatient with indecision, hesitation, or balance of opinions. They “throw themselves precipitately into the principles, to which they are inclined.” 71 But Hume thinks that “a small tincture of Pyrrhonism might abate their pride” and help such people achieve or at least approach that “degree of doubt, and caution, and modesty, which, in all kinds of scrutiny and decision, ought for ever to accompany a just reasoner.” 72 Following the “profound researches” of the Pyrrhonist can lead to greater easiness

68 Enquiry , p. 150. 69 Enquiry , p. 162. 70 Enquiry , p. 161. 71 Ibid. 72 Enquiry , pp. 161–162.


in the face of ignorance and uncertainty, less precipitousness in adopt-ing beliefs simply in order to free oneself from indecision, and less obstinacy in holding on to the beliefs one has.

That is only one kind of benefi cial effect.

Another species of mitigated scepticism which may be of advantage to man-kind, and which may be the natural result of the Pyrrhonian doubts and scruples, is the limitation of our enquiries to such subjects as are best adapted to the narrow capacity of human understanding. 73

The humbling experience of becoming convinced of the imperfec-tions of our faculties brought out by the Pyrrhonian reasoning will tend to confi ne our refl ections more modestly “to common life, and to such subjects as fall under daily practice and experience,” and to lead us away from “all distant and high enquiries.” 74

While we cannot give a satisfactory reason, why we believe, after a thousand experiments, that a stone will fall, or fi re burn; can we ever satisfy ourselves concerning any determination which we may form, with regard to the origin of worlds, and the situation of nature, from, and to eternity? 75

In emphasizing the indispensability of philosophical refl ection to the truly sceptical way of life Hume seems to me closer to the ancient sceptical conception of the quest for ataraxia or tranquillity than is Popkin’s fully natural, but possibly unrefl ective, way of life. Pyrrhonism as Sextus Empiricus describes it is not just any natural way of life. It involves a certain mode of enquiry which, if successful, can lead to tranquillity. The beginnings of the sceptical way of life lie

73 Enquiry , p. 162. 74 Ibid. 75 Ibid. Similar words are put into the mouth of Philo in Dialogues Concerning

Natural Religion , pp. 131–132: “Let us become thoroughly sensible of the weakness, blindness, and narrow limits of human reason: Let us duly consider its uncertainty and needless contrarieties, even in subjects of common life and practice: Let the errors and deceits of our very senses be set before us; the insuperable diffi culties, which attend fi rst principles in all systems; the contradictions, which adhere to the very ideas of matter, cause and effect, extension, space, time, motion; and in a word, quantity of all kinds, the object of the only science, that can fairly pretend to any certainty or evi-dence. When these topics are displayed in their full light, as they are by some philoso-phers and almost all divines; who can retain such confi dence in this frail faculty of reason as to pay any regard to its determinations in points so sublime, so abstruse, so remote from common life and experience? When the coherence of the parts of a stone, or even that composition of parts, which renders it extended; when these familiar objects, I say, are so inexplicable, and contain circumstances so repugnant and contra-dictory; with what assurance can we decide concerning the origin of worlds, or trace their history from eternity to eternity”?


in the perplexity which enquiring minds naturally get into when they seek the truth. They begin with the idea that fi nding and grasping the truth will give them tranquillity, but they quickly fi nd themselves torn between confl icting opinions and are unable to determine which of them are true. The sceptical strategy is to exploit those very contra-dictions or confl icts, to oppose every argument with an opposite one of equal weight. This “main basic principle of the Sceptic system” is what eventually leads enquirers to tranquillity by encouraging them to suspend judgement and to cease dogmatizing. 76

It is true that the sceptic simply follows nature or goes along with appearances. But that natural form of life arises only for someone who has already followed the sceptical enquiry and has subjected himself to the sceptical tropes that are said to produce the required suspension, so it is available only to those who begin with refl ection. A blissful peasant who never refl ected or who never felt or was moved by anxiety about his lack of understanding of the way things are would not lead a sceptical life, however blindly and calmly he was carried along by his natural instincts.

It is the sceptical, and not simply the natural, way of life that Hume recommends. The life of a blissful peasant is not something most of us can even aspire to, let alone achieve. Most of us, like Hume him-self, are “naturally inclin’d to carry [our] view into all those subjects, about which [we] have met with so many disputes in the course of [our] reading and conversation. [We] cannot forbear having a curios-ity” about the sources of our beliefs and actions. 77 We are “uneasy” to think of what we do “without knowing upon what principles [we] proceed.” 78 For almost all of us, then, Hume thinks that what he calls the sceptical or philosophical way of life would be best. 79

76 Sextus Empiricus, Outlines of Pyrrhonism , trans, R. G. Bury (Cambridge, MA: Harvard University Press, 1976), I, 6, 9 .

77 Treatise , p. 270. 78 Treatise , p. 271. 79 He mentions only one exception: “I am sensible . . . that there are in England , in

particular, many honest gentlemen, who being always employ’d in their domestic affairs, or amusing themselves in common recreations, have carried their thoughts very little beyond those objects, which are every day expos’d to their senses. And indeed, of such as these I pretend not to make philosophers, nor do I expect them either to be associates in these researches or auditors of these discoveries. They do well to keep themselves in their present situation; and instead of refi ning them into philosophers, I wish we cou’d communicate to our founders of systems, a share of this gross earthy mixture, as an ingredient, which they commonly stand much in need of, and which cou’d serve to temper those fi ery particles, of which they are compos’d.” ( Treatise , p. 272)


Why would that way of life be best? Even if Hume is right that refl ection on the “profound” Pyrrhonian reasoning would in fact result in what he calls mitigated scepticism, why is that a reason for us to engage in that kind of sceptical reasoning? And is it really true, as Hume says, that the doubt, caution, and modesty which he thinks we would achieve is something that ought to be found in every “just reasoner”? Why is that so? Is there any good reason for us to resist our natural temptation to go beyond common life and experience into “distant and high enquiries” about creation and eternity and fate? Why would that be the best way to live? And how can Hume, of all people, presume to tell us what is the best way of life for human beings?

Hume has answers to all these questions. I think they reveal even more clearly his affi nity with the sceptics of antiquity. The answers all rest on human nature. He knows that “ ’tis almost impossible for the mind of man to rest, like those of the beasts, 80 in that narrow circle of objects, which are the subject of daily conversation and action.” 81 We will inevitably venture out beyond them into “distant and high enquiries” some of the time. Since we know that for most of us such journeys are inevitable:

we ought only to deliberate concerning the choice of our guide, and ought to prefer that which is safest and most agreeable. And in this respect I make bold to recommend philosophy, and shall not scruple to give it the preference to superstition of every kind or denomination. 82

We know that “the imagination of man is naturally sublime, delighted with whatever is remote and extraordinary.” 83 Given free rein, it cre-ates a fertile fi eld for the “embellishment of poets and orators” or “the arts of priests and politicians.” 84 Such manipulators thrive on superstition, which exploits our fears and our ignorance, and pushes the mind further in a direction in which its natural bias or propensity already leads it. 85 New worlds of undreamed-of and inaccessible beings are invented to answer our anxious questions and allay our fears. To follow the high-fl ying imagination in this way and to go

80 And perhaps those gentlemen in England. 81 Treatise , p. 271. 82 Ibid. 83 Enquiry , p. 162. 84 Ibid. 85 Enquiry , p. 40.


along with superstition would be to yield to some of our natural inclinations, but Hume fi nds it the wrong thing to do simply because “superstition seizes more strongly on the mind, and is often able to disturb us in the conduct of our lives and actions.” 86 Philosophy can-not do that. That is why he recommends philosophy over superstition.

There are admittedly certain kinds of philosophy which “may only serve, by imprudent management, to foster a predominant inclin-ation, and push the mind, with more determined resolution, towards that side which already draws too much, by the bias and propensity of the natural temper.” 87 So not just any kind of philosophy will be a suffi cient antidote to superstition. Hume regards “the Academic or Sceptical philosophy” as most harmless and innocent in this respect. 88 It fl atters no natural passion but the love of truth, and it encourages modesty, doubt, and suspense of judgement in matters too large for our easy comprehension.

Nothing, therefore, can be more contrary than such a philosophy to the supine indolence of the mind, its rash arrogance, its lofty pretensions, and its superstitious credulity. 89

Hume recommends, not philosophy in general, or any old philosophy, but non-pretentious, non-superstitious philosophy: “the Academic or Sceptical philosophy.”

What is important in understanding Hume’s defense of this form of scepticism is not just the recommendation he makes of a philoso-phy and a way of life, but also the basis on which the recommenda-tion is made. He does not justify it on the grounds that by giving completely free rein to the imagination we would be led into beliefs that are false, to a tissue of errors and illusions. He is in no position to say that profound, careful philosophy is superior in that respect. Nor does he suggest that the conclusions of the sceptical philosophy are supported by good reasons while elaborate superstitious pictures of the world are not. He says simply that superstition “seizes more strongly on the mind, and is often able to disturb us in the conduct of our lives and actions.” 90

86 Treatise , pp. 272–272. 87 Enquiry , p. 40. 88 Enquiry , p. 41. 89 Ibid. 90 Treatise , pp. 271–272.


Philosophy on the contrary, if just, can present us only with mild and moder-ate sentiments; and if false and extravagant, its opinions are merely the objects of a cold and general speculation, and seldom go so far as to inter-rupt the course of our natural propensities. 91

That is why errors in philosophy are only ridiculous while those in religion are dangerous, and that is why philosophy is always to be preferred. 92 We will be better off with it: less disturbed, more content, and more balanced.

I suggest that much of An Enquiry Concerning Human Understanding , and certainly its opening and its closing sections, represents a defense of and a plea for the sceptical philosophy on just these grounds. Different species of philosophy are distinguished in Section I, and “many positive advantages” of “an accurate scru-tiny into the powers and faculties of human nature” are listed. 93 And Section XII makes a case for the superiority of mitigated scep-ticism in securing those advantages. In defending the study of phil-osophy, the Enquiry can also be seen as a justifi cation of Hume’s more youthful Treatise which had fallen “ deadborn from the Press: without reaching such distinction as even to excite a Murmur among the Zealots.” 94 “Profound and abstract philosophy” of the kind he pursued there, he admits, is “painful and fatiguing,” and it is often obscure, 95 but it should not be rejected entirely on that account alone. There is indeed a plausible objection against a great portion of “profound reasonings, or what is commonly called metaphysics ”: 96

that they are not properly a science; but arise either from the fruitless efforts of human vanity, which would penetrate into subjects utterly inaccessible to the understanding, or from the craft of popular superstitions, which, being

91 Ibid. 92 See Philo’s question in Part XII of Dialogues Concerning Natural Religion (p. 220):

“How happens it, then . . . if vulgar superstition be so salutary to society, that all history abounds so much with accounts of its pernicious consequences on public affairs? Factions, civil wars, persecutions, subversions of government, oppression, slavery; these are the dismal consequences which always attend its prevalency over the minds of men. If the religious spirit be ever mentioned in any historical narration, we are sure to meet afterwards with a detail of the miseries which attend it. And no period of time can be happier or more prosperous, than those in which it is never regarded, or heard of”.

93 Enquiry , p. 13. 94 “My Own Life,” in The Letters of David Hume , ed. J. Y. T. Grieg, 2 vols. (Oxford:

Oxford University Press, 1932), vol. 1, 2. 95 Enquiry , p. 11. 96 Enquiry , p. 9.


unable to defend themselves on fair ground, raise these entangling brambles to cover and protect their weakness. 97

But even if that is true of most philosophy, especially of lofty, preten-tious, system-building philosophy, it is no reason for philosophers to abandon the fi eld and “leave superstition still in possession of her retreat,” 98 heaping “religious fears and prejudices” 99 on defenseless minds. It is all the more reason to enquire carefully into the nature and extent of the human understanding, to pursue as profound and abstract an investigation as is needed to “discover the proper prov-ince of human reason.” 100

We must submit to this fatigue, in order to live at ease ever after. And must cultivate true metaphysics with some care, in order to destroy the false and adulterate. 101

In saying what we must do Hume does not argue here that careful, profound reasonings are likely to lead to the truth, while superstition and prejudice are not. He says only that the proper study of the human faculties will enable us to live at ease in a way that supersti-tion and prejudice will not. Sheer “indolence” or lack of interest might protect a few people from the effects of a “deceitful philosophy,” 102 but for most of the rest of us it will be “overbalanced by curiosity.” 103 And although the philosophical refl ection we cannot avoid might well lead at fi rst to “despair,” it “may give place after-wards to sanguine hopes and expectations.” 104 The point is that we will sometimes raise our minds above the matters of common life and experience, and in doing so we need a guide.

Accurate and just reasoning is the only catholic remedy, fi tted for all persons and all dispositions; and is alone able to subvert that abstruse philosophy and metaphysical jargon, which, being mixed up with popular superstition, renders it in a manner impenetrable to careless reasoners, and gives it the air of science and wisdom. 105

97 Enquiry , p. 11. 98 Enquiry , p. 12. 99 Enquiry , p. 11.

100 Enquiry , p. 12. 101 Ibid. 102 Perhaps those gentlemen in England are safe. 103 Enquiry , p. 12. 104 Ibid. This is precisely the course of Hume’s own sentiments as reported so dra-

matically in the fi rst person in the “Conclusion” of Book One of the Treatise . 105 Enquiry , pp. 12–13.


These are large claims to be made in the name of philosophy as Hume understands it. His support for them comes from his view of human nature and of what is necessary for a good human life. He knows that “the mere philosopher” is widely believed “to contribute nothing either to the advantage or pleasure of society.” 106 But on the other hand “the mere ignorant is still more despised.” 107 Each repre-sents only one extreme of what is essential to a balanced human life. Human beings are reasonable, and so need science and learning as part of their “proper food and nourishment.” 108 But they are also sociable beings, as well as active beings, and they need business and occupation. They need relaxation and enjoyment as well.

It seems, then, that nature has pointed out a mixed kind of life as most suit-able to the human race, and secretly admonished them to allow none of these biasses to draw too much, so as to incapacitate them for other occupations and entertainments. Indulge your passion for science, says she, but let your science be human, and such as may have a direct reference to action and society. Abstruse thought and profound researches I prohibit, and will se verly punish, by the pensive melancholy which they introduce, by the endless uncertainty in which they involve you, and by the cold reception which your pretended discoveries shall meet with, when communicated. Be a philoso-pher; but, amidst all your philosophy, be still a man. 109

Hume recommends the pursuit of the sceptical or academical philos-ophy as the best or perhaps the only way to achieve this most natural and therefore most blissful human condition. It represents not only the best way to be a philosopher, given that one is human, but also the best way of being human, given that one will inevitably try to under-stand oneself and the world. The pursuit of the sceptical philosophy is the best way of giving adequate expression to all the tendencies or propensities which constitute human nature. It will lead temporarily to “Pyrrhonism, or excessive scepticism,” but, given our natural instincts, there is no danger that that will paralyze us. And our recog-nition of the superior force of those natural instincts will provide a somewhat bemused detachment from the empty pretensions of rea-son alone. Unlike the beasts, we will raise our minds from time to time beyond “that narrow circle of objects, which are the subject of daily con-versation and action,” and so to some extent we will satisfy our curiosity

106 Enquiry , p. 8. 107 Ibid. 108 Ibid. 109 Enquiry , p. 9. Nature here appears to be speaking directly to the youthful author

of A Treatise of Human Nature .


about “high and distant” matters. But we will do so without giving the imagination completely free rein alone, or simply indulging a predom-inant bias or inclination in our natural temper. With philosophy as our guide, we will be immune to the disturbances of religion and other forms of superstition. We will have achieved a happy “determination” in which no side of our nature “ draws too much.”

What is behind Hume’s defense of the sceptical philosophy is noth-ing more than his views about the nature of human beings. Doubt, caution, and modesty – the effects of the sceptical enquiry – ought to be found in every “just reasoner,” not because that is our God-given duty, but because we will be better off that way. We will be less disturbed, more completely satisfi ed, more balanced. For the same reason we ought to restrain the infl amed imagination and overcome superstition. Not because there is some a priori imperative for us to do so, but because more potential sources of turmoil and distress in human life will be avoided that way. For Hume those are the only sorts of facts which could ever lie behind any just claim about how people ought to live. In that respect too he is fully in accord with the sceptics of antiquity.

Unlike those ancient sceptics who tried to avoid all convictions or beliefs, Hume is not reluctant to believe or even to state the facts of human nature on which his conception of the best kind of human life depends. His acceptance of these facts is not inconsistent with the kind of scepticism he advocates. In fact, he cannot help believing them. Or if he does occasionally feel a certain doubt or reluctance, it is only during the “momentary amazement and confusion” brought on by intense philosophical refl ection. And for him, as for the rest of us, that “delirium” soon passes away. 110

110 In the end perhaps the best recommendation of scepticism or of any other way of life is to be found not in confi rmation of the general facts of human nature on which it is thought to be based, but on the actual lives of its practitioners. If we take Hume’s own life as an expression of his scepticism, Adam Smith’s description of that life is a strong recommendation indeed: “His temper, indeed, seemed to be more happily bal-anced, if I may be allowed such an expression, than that perhaps of any other man I have ever known. Even in the lowest state of his fortune, his great and necessary fru-gality never hindered him from exercising . . . acts both of charity and generosity. . . . The extreme gentleness of his nature never weakened either the fi rmness of his mind, or the steadiness of his resolutions. His constant pleasantry was the genuine effusion of good-nature and good-humor, tempered with delicacy and modesty, and without even the slightest tincture of malignity. . . . And the gaiety of temper, so agreeable in society . . . was in him certainly attended with the most severe application, the most extensive learning, the greatest depth of thought, and a capacity in every respect the most comprehensive. Upon the whole, I have always considered him, both in his lifetime and since his death, as approaching as nearly to the idea of a perfectly wise and virtuous man, as perhaps the nature of human frailty will permit.” ( Dialogues , pp. 247–248)

8

“Gilding or Staining” the World with “Sentiments” and “Phantasms”

Hume’s “science of human nature” is meant to explain, in theory, how human beings come to have all the ideas, thoughts, and beliefs that we know they have. All such mental items are to fi nd their source, one way or another, in experience. But given Hume’s conception of perception and feeling, and his understanding of the relation between perception or feeling and the rest of our mental life, there is an import-ant class of thoughts which present a special problem for him.

The question is whether Hume’s theory can really explain how we get those thoughts and whether, if the kind of explanation he offers does not succeed as it stands, it could ever be improved on while remaining faithful to the general structure of his conception of the mind and its relation to the world. Many who philosophize today in the spirit of Hume while rejecting what they see as unacceptable but dispensable details of his way of thinking would appear to hold that it can. I think no satisfactory explanation along the right lines has yet been given, or even suggested.

The thoughts I am concerned with are primarily thoughts of some-thing or other’s being so. I do not mean only beliefs or judgments that something or other is so; there is also the contemplation or entertaining of something as being so, whether it is actually believed or judged to be so or not. For example, looking at the billiard table, I come to believe that the white ball’s hitting the red ball will cause the red ball to move in a certain direction. I also think that if the white ball causes the red ball to move in that direction, the red ball will go into the corner pocket. In this second, conditional, thought, I think of the white ball’s causing the red ball to move, but I do not then express the belief that it will, which I have in the fi rst thought.

This is one example of the kind of thought I have in mind. It involves what is for Hume the problematic idea of one thing’s causing another. Another example, from a seemingly very different area, is the thought of an action’s being evil, or vicious, or blameworthy.


I might observe someone doing something and immediately come to think that it is bad or vicious. Or I might think, purely hypothetically, that if any person were to commit a suffi ciently vicious or evil act, he should be executed—or perhaps, more humanely, that he should not, even if what he did is vicious or blameworthy. These thoughts involve what is for Hume the problematic idea of vice, or moral evil, or blameworthiness.

A third example involves the idea of beauty. I can fi nd a particular object beautiful when looking at it, or, with no particular object in mind, I might seek something beautiful. And I might think that if I had something beautiful, I would be fortunate or happy.

There seems to me no doubt that we all have thoughts like this. What binds these apparently different examples together is that the ideas involved in each case are special or problematic for Hume in the same way. “Take any action allow’d to be vicious,” he says, “Wilful murder, for instance.”

Examine it in all its lights, and see if you can fi nd that matter of fact, or real existence, which you call vice . In which-ever way you take it, you fi nd only certain passions, motives, volitions and thoughts. There is no other matter of fact in the case. The vice entirely escapes you, as long as you consider the object. 1

The idea of vice or viciousness does not denote anything in the “object” to which it is applied. What you think to be true of the “object” simply isn’t there.

The “object” is also said to be the wrong place to look in the case of beauty.

EUCLID has fully explained every quality of the circle, but has not, in any proposition, said a word of its beauty. The reason is evident. Beauty is not a quality of the circle. It lies not in any part of the line whose parts are all equally distant from a common center. . . . In vain would you look for it in the circle, or seek it, either by your senses, or by mathematical reasonings, in all the properties of that fi gure. 2

There is nothing in any object which can properly be called its beauty. For Hume this is not a matter of controversy.

1 David Hume, A Treatise of Human Nature (T ), edited by L. A. Selby-Bigge revised by P. H. Nidditch (London: Oxford University Press, 1978), p. 468 .

2 David Hume, “The Sceptic,” in Essays, Moral, Political and Literary, edited by E. F. Miller (Indianapolis: Liberty Classics, 1985), p. 165 .

“Gilding or Staining” the World 169

If we can depend upon any principle, which we learn from philosophy, this, I think, may be considered as certain and undoubted, that there is nothing, in itself, valuable or despicable, desirable or hateful, beautiful or deformed. 3

Considering only the objects in question in themselves, there is no vice or evil or beauty or ugliness to be found.

Something parallel is true of causation. However closely we scru-tinize a single instance of one billiard ball’s causing another to move,

we fi nd only that the one body approaches the other; and that the motion of it precedes that of the other, but without any sensible interval. ’Tis in vain to rack ourselves with farther thought and refl exion upon this subject. We can go no farther in considering this particular instance. (T 77)

But we cannot say that contiguity and succession alone give us a “compleat idea of causation.” “There is a NECESSARY CONNEX-ION to be taken into consideration.” But, Hume says:

Here again I turn the object on all sides, in order to discover the nature of this necessary connexion, and fi nd the impression, or impressions, from which its idea may be deriv’d. When I cast my eye on the known qualities of objects, I immediately discover that the relation of cause and effect depends not in the least on them . When I consider their relations , I can fi nd none but those of contiguity and succession; which I have already regarded as imper-fect and unsatisfactory. (T 77)

There simply is no such connection to be perceived in any particular case. After we have observed a series of several resembling instances of contiguity and succession, we do in fact come to think of two sorts of things as causally connected. But the repetition alone does not reveal something in the current instance that was not to be found in any of the earlier and exactly resembling instances; nor does it produce something new in the later resembling instances, each of which is independent of all the rest. Hume concludes:

There is, then, nothing new either discover’d or produc’d in any objects by their constant conjunction, and by the uninterrupted resemblance of their relations of succession and contiguity. But ’tis from this resemblance, that the ideas of necessity, of power, and of effi cacy, are deriv’d. These ideas, therefore, represent not any thing, that does or can belong to the objects, which are constantly conjoin’d. (T 164)

3 “The Sceptic,” Essays, p. 162.


This is perhaps the best description of what is special or problematic about the ideas centrally involved in each of the kinds of thoughts I want to consider. The idea in question does not represent anything “that does or can belong to the objects” which we think of by means of that idea. We think of those objects as being a certain way, but they are not and cannot be that way. There is nothing in, or perceivable in, an act of willful murder that is its vice or its being vicious; beauty is not a quality of any object; there is nothing in, or discernible in, any two objects or the relations between them that is the necessary or causal connection between them. But it appears that we can and do think of some actions as being vicious, of some objects as being beautiful, and of one thing’s causing another. We appear to have thoughts in which we predicate those very qualities of certain objects or relations.

The problem then is to explain how we come to have such thoughts. It is not just a matter of identifying the occasions on which thoughts like that fi rst come into our minds. It is also a question of what hap-pens to us on those occasions, and of exactly how whatever happens brings it about that the thought we eventually get is the thought of an act as being vicious or of an object as being beautiful or of one event as being the cause of another.

What Hume thinks happens to produce a thought of an action’s being vicious or of an object’s being beautiful is that in each case we feel a certain “sentiment.” If you are looking for the vice or vicious-ness of a certain action:

You never can fi nd it, till you turn your refl exion into you own breast, and fi nd a sentiment of disapprobation, which arises in you, towards this action. Here is a matter of fact; but ’tis the object of feeling, not of reason. It lies in yourself, not in the object. (T 468–9)

Similarly, the beauty of a circle is not a quality of the circle.

It is only the effect, which that fi gure produces upon a mind, whose particu-lar fabric or structure renders it susceptible of such sentiments. 4

If we never got such “sentiments” we would never “pronounce” any-thing to be “valuable or despicable, desirable or hateful, beautiful or deformed”:

these attributes arise from the particular constitution and fabric of human sentiment and affection. 5

4 “The Sceptic,” Essays, p. 165. 5 “The Sceptic,” Essays, p. 162.


That is not to say that our getting the relevant “sentiments” is always completely independent of all reason or judgment or thought. Discernment of beauty can be improved; with practice and learning, “the organ acquires greater perfection in its operation.” 6 And not just any “sentiment” of pleasure or pain derived from a person’s action or character makes us praise or condemn it.

’Tis only when a character is consider’d in general, without reference to our particular interest, that it causes such a feeling or sentiment, as denominates it morally good or evil. (T 472)

Experience and informed refl ection might well be necessary to arrive at such “ steady and general points of view” (T 581–2). But even when thought or refl ection is needed, some actual “sentiments” or feelings are needed as well. Without them, we would never “pronounce” on the moral qualities of actions or characters, or on the beauty or ugli-ness of objects around us.

The “sentiment” that always arises in such cases is something new, something beyond or at least different from any thought or belief produced by reason or the understanding.

All the circumstances of the case are supposed to be laid before us, ere we can fi x any sentence of blame or approbation. . . . But after every circum-stance, every relation is known, the understanding has no further room to operate, nor any object on which it could employ itself. The approbation or blame which then ensues, cannot be the work of the judgment, but of the heart; and is not a speculative proposition or affi rmation, but an active feel-ing or sentiment. 7

The distinction Hume draws here marks the difference in general between the distinct faculties which he calls “reason” and “taste.”

The former conveys the knowledge of truth and falsehood: the latter gives the sentiment of beauty and deformity, vice and virtue. The one discovers objects as they really stand in nature, without addition or diminution: the other has a productive faculty. . . . From circumstances and relations, known or supposed, the former leads us to the discovery of the concealed and unknown: after all circumstances and relations are laid before us, the latter makes us feel from the whole a new sentiment of blame or approbation. (EPM 294)

6 “Of the Standard of Taste,” Essays, p. 237. 7 David Hume, Enquiries Concerning Human Understanding and Concerning the

Principles of Morals (EHU and EPM), edited by L. A. Selby-Bigge, revised by P. H. Nidditch (Oxford: Oxford University Press, 1975), p. 290 .


Something “new” is also produced in the case of causation, and never by reason or the understanding. There is nothing in a series of resembling pairs of objects which answers to the idea of a necessary connection between their members:

yet the observation of this resemblance produces a new impression in the mind, which is its real model. For after we have observ’d the resemblance in a suffi cient number of instances, we immediately feel a determination of the mind to pass from one object to its usual attendant, and to conceive it in a stronger light upon account of that relation. This determination is the only effect of the resemblance; and therefore must be the same with power or effi cacy, whose idea is deriv’d from the resemblance. (T 165)

The “new” or “added” ingredient is something in the mind. The inde-pendent but resembling instances of contiguity and succession therefore “have no union but in the mind, which observes them,” as Hume puts it (T 165). I take this to mean that there is no necessary connection between the objects; we only think that there is. That is also the way to take his famous (and otherwise disastrous) pro-nouncement that “Upon the whole, necessity is something, that exists in the mind, not in objects” (T 165): we think things are necessarily connected, but they really are not. That would make the remark about necessity parallel to the even more famous (and almost equally disastrous) adage that “Beauty is in the eye of the beholder.” This is not to be taken to mean that beholders have beautiful eyes.

What is important for Hume is that it is what he calls “the imagin-ation,” not reason or the understanding, that is the source of the “new” or “additional” item which must make its appearance in the mind if we are to be led to “pronounce” any “sentence of blame or approbation,” or of beauty or deformity, or of causal or necessary connection. In all these cases the new item is an impression—a “sen-timent” or feeling or an impression of refl ection. 8 How does the

8 There are other thoughts which are problematic for Hume in this same way but which do not arise from particular impressions or sentiments at all. We think of objects as continuing to exist unperceived, although we never encounter such things in our experience, and “the fi ction of a continu’d existence . . . , as well as the identity, is really false” (T 209). We think of minds as existing through time, although “the iden-tity, which we ascribe to the mind of man, is only a fi ctitious one” (T 259); “there is properly no simplicity in it at one time, nor identity in different; whatever natural propension we may have to imagine that simplicity and identity” (T 253). In these cases, because of certain general “principles of the imagination,” the mind is equally naturally led in confl icting directions, and we are said to “feign” or “imagine” certain things as a way of resolving the confl ict. The resolution in each case appears to be


appearance of one of those things in the mind have the effect of giv-ing us thoughts (or “ideas”) of vice, of beauty, of causation, or of any other qualities or relations we ascribe to objects, when according to Hume those qualities and relations do not and cannot actually belong to “objects as they really stand in nature” (EPM 294)?

He is aware that the idea that objects do not really stand in causal relations or necessary connections to one another in nature will be greeted as an astonishing and violent “paradox.” He thinks there is a deep “biass of the mind” against it (T 166–7). But he thinks that the source of that very “biass” also provides the explanation he is look-ing for. In the case of causation, for example, we know that a certain “internal impression” arises in the mind after the observation of a constant conjunction between objects of two kinds. And:

’Tis a common observation, that the mind has a great propensity to spread itself on external objects, and to conjoin with them any internal impressions, which they occasion, and which always make their appearance at the same time that these objects discover themselves to the senses. Thus as certain sounds and smells are always found to attend certain visible objects, we nat-urally imagine a conjunction, even in place, betwixt the objects and the qual-ities, tho’ the qualities be of such a nature as to admit of no such conjunction, and really exist no where . . . the same propensity is the reason, why we sup-pose necessity and power to lie in the objects we consider, not in our mind, that considers them. (T 167)

In the fi rst Enquiry he describes that same “spreading” or “conjoin-ing” operation this way:

as we feel a customary connexion between the ideas, we transfer that feeling to the objects; as nothing is more usual than to apply to external bodies every internal sensation, which they occasion. (EHU 78n)

We “feel a determination of the mind” to “pass from one object to the idea of its usual attendant” (T 165), and it is that impression, or what it is an impression of, that we somehow “spread” on or “transfer” to or “conjoin” with the objects now before us, and so come to “im agine” or “suppose” that they are causually or necessarily connected.

strictly cognitive, or intellectual. We introduce a new thought or way of thinking into our repertoire; no feeling or “sentiment” works on us in addition to the ideas we pos-ses. I will not enter further into Hume’s explanations of “fi ctions” or “illusions” which arise in this way without a feeling or sentiment. I concentrate here on the relation between feeling or perception and thought, and how the one is supposed to lead to the other in these problematic cases.


In the case of morals, the understanding fi rst discovers and judges the relevant matters of fact in the case, and then “the mind, from the contemplation of the whole, feels some new impression of affection or disgust, esteem or contempt, approbation or blame” (EPM 290). There is again a certain “propensity” at work which starts from that impression or sentiment and somehow takes us beyond the deliver-ances of observation and the understanding alone. It is the imagin-ation which in all these cases exhibits:

a productive faculty, and gilding or staining all natural objects with the col-ours, borrowed from internal sentiment, raises in a manner a new creation. (EPM 294)

The “new creation” is eventually a conception of a world containing good and evil actions, admirable and contemptible characters, and beautiful and ugly objects. It is only because we naturally get certain feelings or impressions, and, even more importantly, only because of the mind’s “productive faculty” in “gilding or staining” the world with what those feelings give us, that we ever come to think in those ways at all. Our moral and aesthetic judgments do not report the presence in objects of qualities which really belong to those objects to which we appear to ascribe them.

Hume draws the same parallel to explain the formation of moral and aesthetic judgments as he drew earlier with thoughts of neces-sary connection. There he compared necessity to sounds and smells; here he adds colors and heat and cold, and invokes the “modern phi-losophy’s” doctrine of “secondary” qualities. The mind operates in moral thinking just as that view says it does with respect to colors, sounds, heat and cold.

Vice and virtue, therefore, may be compar’d to sounds, colours, heat and cold, which according to modern philosophy, are not qualities in objects, but perceptions in the mind . . . (T 469)

In each case the mind “transfers” features of its internal workings or contents to an external world which does not really contain them.

The question is how this “gilding or staining” is supposed to work. What is involved in the mind’s “spreading” itself on to exter-nal objects and “conjoining” with them, or “transferring” to them, something “borrowed” from internal impressions or sentiments? In making the transition, Hume says, the mind “raises in a manner a new creation.” I take him to mean that in our thoughts we


somehow come to endow objects with something “new,” with cer-tain qualities or relations which they do not possess “as they really stand in nature, without addition or diminution” by us. This mental operation I have called “projection,” 9 no doubt more in the dic-tionary than in the psychoanalytic sense. We put on to objects in our thoughts about them certain features that they do not really possess. We take something mental and see it as external. That is how I take Hume’s metaphors of “gilding or staining,” or “spread-ing” something on to, a neutral and unsuspecting world. How does it work?

There is a real problem here for Hume. To put it another way, there is a real problem here for anyone who would interpret Hume as hold-ing that we do really think of objects as causally or necessarily con-nected, or as evil or vicious, or as beautiful. I do want to interpret Hume that way. I think human beings do have such thoughts, so it would be a good thing if Hume’s science of human nature could acknowledge that we have them. That theory is meant to explain every thought and feeling human beings have. But the problem for Hume is that if we do have thoughts of causation, or of the vice or beauty of things, they are thoughts which do not represent “any thing, that does or can belong” to external objects “as they really stand in nature.” He has a view of the world or of “nature” according to which no such qualities or relations could belong to or hold between the objects that make up that world. That is one of the things that make it so hard for him to explain how such thoughts or beliefs are possible. What the thoughts are about is never to be found in the world. It seems then that we could arrive at them only by “adding” something to our conception of the world, by “gilding or staining” it with something that does not really belong to it. The source of that extra “stain” or “gilding” could only be the mind itself, or its con-tents, so it is from there that we must “borrow” whatever materials are used in the “spreading” or “transferring” operation. Anyone who thinks that we do have such thoughts, and who shares Hume’s restricted conception of what the world or “nature” can contain—as

9 See Stroud, Hume (London: Routledge and Kegan Paul, 1977) , e.g., pp. 86, 87, 185, 186. I fi nd now that the term was used to refer to this operation in Hume by Paul Grice in the early 1970s. See Grice, “Method in Philosophical Psychology: From the Banal to the Bizarre” [Presidential Address to the Pacifi c Division of The American Philosophical Association, March 1975], The Conception of Value (Oxford: Oxford University Press, 1992), p. 146 .


many philosophers apparently still do—would seem forced into an account along some such lines.

The questions any such account must answer are: what do we “borrow” from our internal impressions, and what do we ascribe to the external objects we “gild or stain”? We presumably do not “bor-row” the internal impression itself and ascribe it in thought to an external object. We do not think that the sequence of events on the billiards table—the one ball’s striking the other and the second ball’s moving—itself has a feeling or impression like the feeling Hume says we humans get when we observe it. Nor do we think that when the second ball is struck it moves off with a feeling like that. 10 We do not think that an act of willful murder itself has a feeling of disgust or disapprobation, any more than we think that a painting on a wall has a sentiment of pleasure or awe. That is nonsense in each case. It is not the internal impression itself that we ascribe to the external object. Rather, it seems that it should be what the impression is an impres-sion of that we so predicate.

But Hume’s view of impressions—or at least of those impres-sions he seems to have principally in mind in his “gilding or stain-ing” metaphor—makes it diffi cult for him to appeal in the right way to what impressions are impression of. He thinks primarily of colors, sounds, smells, heat, and cold, and all of them he says are “nothing but impressions in the mind” (T 226). The point is not that impressions of colors and sounds are impressions in the mind, but that colors and sounds are impressions in the mind. That is the view he attributes to “modern philosophy” (T 469). It is because such impressions always “attend” the perception of certain external objects that:

we naturally imagine a conjunction, even in place, betwixt the objects and qualities, tho’ the qualities be of such a nature to admit of no such conjunc-tion, and really exist no where. (T 167)

10 It must be admitted that Hume sometimes suggests that we do suppose something like that.

No animal can put external bodies in motion without the sentiment of a nisus or endeav-our; and every animal has a sentiment or feeling from the stroke or blow of an external object, that is in motion. These sensations, which are merely animal, and from which we can a priori draw no inference, we are apt to transfer to inanimate objects, and to suppose, that they have some such feelings, whenever they transfer or receive motion. (EHU 78n)


This suggests that Hume endorses “modern philosophy’s” view that the redness we see is nothing more than a feature of our impres-sions. In his essay “Of the Standard of Taste” he says he wants to explain how one color can be “denominated” the “true and real” color of an object, “even while color is allowed to be merely a phan-tasm of the senses.” 11 If that were what color is, it would not be some-thing that could ever be in the same “place” with an apple. In that respect it would be like pain; the pain we feel is not something that could intelligibly be located in, or belong to, or be predicated of, an external object that causes it. It does not exist in, or belong to, the world of external objects at all. In that sense, the felt quality of a painful sensation could be said to exist “no where,” i.e., in no place. But presumably in that case no one thinks that it does. There is no “spreading” or “gilding” the objects of the world with pain when we have sensations of pain. What is perceived or felt when a painful impression is present is not something that coherently admits of attribution to an inanimate external object.

The same would be true of the disgust or displeasure we might ex perience when observing an act of willful murder, or the pleasure we might get from seeing a great painting, if they too are on Hume’s view just impressions or feelings of certain distinctive kinds. To try to predi-cate them of the objects that cause them would be to ascribe a feeling or impression to an act of murder or to a painting. And that is absurd. The impression or feeling that Hume says comes into the mind when we see objects of one kind constantly followed by objects of another kind would also on that view be yet another distinctive impression. Like a pain, it would be simply an impression or feeling of a certain kind which differs in directly perceivable ways from impressions of other kinds. What distinguishes them in each case would be perceivable or felt qualities of the impressions themselves. Those same qualities which serve to distinguish one kind of impression or feeling from another therefore could not also be thought to be qualities of external objects, any more than the pain we feel or the painfulness of a painful sensation is something that could be a quality of an external object. If impressions of something are understood in that way—as we speak of a “sensation of pain”—then what they are impressions of is not some-thing that could also be thought to be a quality of an object.

11 “Of the Standard of Taste,” Essays, p. 234.


To understand the operation of “gilding or staining” the world with something “borrowed from internal sentiment” in that way, then, would mean that that operation could never really succeed in producing an intelligible thought which attributes certain “added” features to external objects or to the relations between them. At best it would produce a kind of confusion or nonsense on our part, per-haps with an accompanying illusion of having coherent thoughts of that kind when we really do not. There certainly are suggestions that Hume sometimes thinks of it that way, especially in what he says about the idea of necessary connection. Necessity, he says, “is noth-ing but an internal impression of the mind, or a determination to carry our thoughts from one object to another” (T 165). If that is what necessity is, then it would seem that any thoughts about neces-sity would be thoughts either about an impression or about a deter-mination or transition of the mind. But then we could not intelligibly think that necessity, so understood, is a feature of the relation between two external objects or events—that the two are necessarily con-nected. We could not think that one thing must or had to happen, given that something else had happened earlier.

Hume appears to endorse that conclusion in his gloss on the formula that “necessity is something, that exists in the mind, not in objects”:

nor is it possible for us to form the most distant idea of it, consider’d as a quality in bodies. Either we have no idea of necessity, or necessity is nothing but that determination of the thought to pass from causes to effects and from effects to causes, according to their experienc’d union. (T 165–6)

This seems to say that we can think intelligibly about the passage of our thought from one thing to another, and we can think intelligibly about the impression or feeling of determination which accompanies that transition, but that is really all there is to think about in connec-tion with necessity. We cannot intelligibly think that something has to happen, or happens of necessity, or that one thing is necessarily connected with another.

When Hume says that we nevertheless do

. . . suppose necessity and power to lie in the objects we consider, not in our mind, that considers them; notwithstanding it is not possible for us to form the most distant idea of that quality, when it is not taken for the determin-ation of the mind . . . (T 167)

he implies that we are at best confused in our attempts to think of things as causally or necessarily connected. What he says we “suppose”


(“Necessity lies in the objects we consider”) is not really something we could ever have “the most distant idea of.” On this reading, the only idea we could have of the necessity involved in causation is apparently not an idea of any quality which we could intelligibly think belongs to, or could be predicated of, the relation between two objects. It could only be an idea of something (an impression or feeling) which always accompanies the observation of certain pairs of objects, and that is something in the mind, not a quality of the objects or of the relations between them.

There is no question that we can think clearly and without confu-sion about the passage of our thought from one thing to another, or about impressions which appear in our minds on certain occasions. But such thoughts do not involve “gilding or staining” anything in the world with qualities it does not really possess. If Hume is right, we do in fact feel or experience something when the mind passes from one idea to its usual attendant, and in announcing the presence of such an impression we would be stating no more than a straightfor-ward autobiographical fact. Or, in the moral case, I might say, as Hume suggests, “I feel a sentiment of disgust or disapprobation when I consider that act of wilful murder” (T 469); and if I do, what I say will be no more than the ungilded truth. I would not be “adding” or “spreading” any extra quality on to that act or on to anything else.

We can think, equally clearly, and truly, not only about ourselves and the goings-on in our own minds, but also about external objects. If we think of objects of a certain kind, that observation of them is always accompanied by an impression of a certain kind—a feeling of pleasure, or disgust, or perhaps a “feeling of determination”—then again, in having such thoughts we are not “spreading” a “new cre-ation” on to those objects, or “adding” something to them which does not belong to them “as they really stand” in the world. We merely think, without projection or confusion, about a relation which actu-ally holds in nature between certain objects and certain human feel-ings. So although we can and do have perfectly intelligible thoughts of these two kinds, they involve no “gilding or staining.” Nor do they involve ascribing beauty or viciousness or a causal connection to any objects either. But those are the kinds of thoughts which need to be explained. If we have them, Hume must acknowledge that we have them, and his theory of the mind must eventually account for them.

To have only thoughts about our impressions, or dispositional thoughts about the natural tendency of external objects to produce


such impressions in human minds, we perhaps do not need to think of the impressions involved as anything other than impressions or feelings with their own distinctive and directly perceivable characteristics—on the model of sensations of pain. But that is what we found stands in the way of the apparently most straight-forward understanding of the operation of “spreading” or projec-tion. We could not then take the quality which distinguishes impressions of one kind (e.g., disgust) from impressions of another kind (e.g., pain) and somehow predicate that very quality of any external object. The feature of the impression which in that sense makes it the kind of impression it is cannot also intelligibly be thought to be a quality of an external object. But although Hume does often appear to think of impressions of colors, sounds, smells, and heat and cold in this way, and perhaps also feelings of various kinds of pleasure and displeasure as well—along with sensations of pain—it is clear that that view of impressions cannot be accepted in general. Not all impressions “of F” can be understood on the model of sensations “of pain.” There must be another way of distinguish-ing impressions from one another, another way of understanding what it is for an impression to be an impression of such-and-such, if we are ever able to think of perceived qualities as belonging to objects in the world.

If, as Hume holds, every case of perceiving something is a matter of our having an impression of something, then if every impression were just an impression with a certain distinctive felt or perceived character, we could never come to think of external objects as having any of those very qualities that we can perceive. An impression of a round ball, for example, or of the roundness of the ball, would then also be just an impression with a certain distinctive character perceiv-ably different from other impressions like pain or disgust or pleasure. The quality that we are aware of in having such an impression could not coherently be thought also to be a quality of an external object. And if that were true of all impressions, and so of all perceivable qualities, then either impressions or feelings or things in the mind would be the only things that we could think of as having any qual-ities at all—as Berkeley held—or none of the qualities that we could think of an external object as having could be qualities which we could also perceive anything to have. Thought of objects which are not in the mind, if it were possible at all, would be in that way com-pletely cut off from perception or feeling. The objects of thought and


the objects of perception would never be the same. What we can per-ceive and what we can think would not even overlap.

To avoid that unacceptable dilemma, at least some impressions must be understood “intentionally,” as being “of” something that could be so, or of something that could be thought to be true of external objects. Hume apparently fi nds no diffi culty in thinking in this way about an impression of a round ball, for example, or of the roundness of a ball, or an impression of one round ball’s striking another. It seems that we can and do have such impressions, and when we think of one ball striking another, our thought has the very same content; the very qualities and relations that we sometimes perceive—roundness and striking—we also think are qualities or relations of the balls we think about. What we can fi nd in perception is in that case reproduced in thought. We attribute some of the very qualities and relations we perceive to the objects we think about.

It must be said that it is diffi cult to understand how we could ever have an impression of one round ball’s striking another if the thought of two such objects standing in that relation to each other made no sense to us, or was something that we did not think could be so. Our being capable of a perception with just that content would seem to require our fi nding intelligible the thought of one ball’s striking another. For Hume, it is the other way around. Our getting an impres-sion of something is what makes it possible for us to have an idea of that same something. 12 So he needs an independent specifi cation of what we can and cannot, strictly speaking, get impressions of. But the special problem which arises for the problematic thoughts we are interested in is that the impressions which are said to produce them cannot in that sense be impressions of “anything, that does or can belong” to external objects. They are not “of” anything that can be so, or that we can perceive to be so, in the world.

This is sometimes obscured by the way Hume occasionally describes those impressions or feelings. In the case of causation, for example, he speaks of “this connexion . . . which we feel in the mind” (EHU 75). He says that after having observed two kinds of events in constant conjunction the observer “now feels these events to be con-nected in his imagination” (EHU 75–6). But of course on Hume’s

12 Here I ignore for the moment the distinction between simple and complex impres-sions and ideas. The sweeping generalization strictly holds only for simple perceptions.


view of the world there can be no such thing as a necessary connec-tion between two events, and no such state of affairs as two things’ being connected in the mind. Nothing in the world is actually con-nected with anything else, anywhere. So we can never perceive a con-nection which holds between two things, and if we can nevertheless be said to “feel” them to be connected, it must be because the idea of two things’ being causally connected already makes sense to us. If we really did have such a feeling, there would presumably be no diffi culty in “transferring” the content of that feeling to objects in the world and thinking that it is true of the relation between them. We could reproduce in thought exactly what we had found in feeling or percep-tion. But if we must possess the idea of necessary connection in order to “feel” that two things are connected, even in the mind, then it would seem that we could have such a “feeling” only if we had already performed the operation of “gilding or staining.” Hume thinks that that operation is the only way we come to think of two things as causally connected in the fi rst place.

Hume is more careful in describing the experience of exercising the will. He easily resists the suggestion that we get the idea of cause or power from the way in which parts of our bodies and many of our thoughts can be seen to obey the will. He does not deny that we observe the motion of the body to follow upon a “volition” to move it, but we are never able, he says, “to observe or conceive the tie which binds together the motion and volition, or the energy by which the mind produces this effect” (EHU 74).

the will being here consider’d as a cause, has no more a discoverable connex-ion with its effects, than any material cause has with its proper effect. (T 632)

There is no impression of the will’s effi cacy or power; all we are aware of in action is at fi rst a felt “volition,” and then an impression of what happens next.

It would be no better to appeal not to the power of the will but to its powerlessness. 13 If we have experienced a correlation between things of two kinds in the past, and an idea of a thing of the fi rst kind appears in the mind, then whatever we happen to will or not to will at that time, an idea of a thing of the second kind will inevitably present itself. That is one of Hume’s fundamental “principles of the imagination.” If in those

13 This has been suggested by David Pears, Hume’s System (New York: Oxford University Press, 1990), pp. 110–115 .


circumstances we were to get an impression of the inevitability with which that idea appears in the mind, or of our powerlessness to resist its appearing there, we could presumably then ascribe that very feature that we get an impression of—viz., inevitability or powerlessness—to the happenings on the billiards table and elsewhere in the world of objects. We could reproduce in thought exactly what we can fi nd in perception or feeling. But again it is Hume’s view that we could get no such impression or feeling. We could feel a certain desire or “volition”—for example, we will the appearance of an idea other than the idea of the second ball’s moving, or perhaps we decide to will nothing at all—and then we immediately fi nd that the idea of the second ball’s moving nevertheless appears. Repeated experiments show that that same idea always appears in the appropriate circumstances whatever “volitions” are present in those circumstances. But that discovery of the goings-on in our minds involves no impression of the ineffi cacy or powerlessness of the will. It involves only the awareness of many very different “vol-itions,” which according to Hume are themselves just different impressions, 14 followed always by the appearance in the mind of one and the same kind of idea. There is and can be no impression of the inevitability with which things happen, even in the mind; there are impressions only of what happens, or of its happening. If we did have an impression or feeling of the inevitability of something’s happening, or of our powerlessness to prevent it, it could only be because we had already acquired the idea of power and were able to recognize its absence. But such an idea is for Hume the product of the operation of “gilding or staining”; if it were required for the very impression from which that operation is supposed to start (as it, in fact, seems to be), it could not also be the product of that very operation.

In explaining his view of morals Hume is careful to point out that:

We do not infer a character to be virtuous, because it pleases: But in feeling that it pleases after such a particular manner, we in effect feel that it is virtu-ous. (T 471)

But again, that a given character is virtuous is on Hume’s view not something that is or could be so as things “really stand in nature.” If

14 I desire it may be observ’d, that by the will, I mean nothing but the internal impres-sion we feel and are conscious of, when we knowingly give rise to any new motion of our body, or new perception of our mind. (T 399)


we could have a feeling that a certain character is virtuous, it would have to be because we are already capable of intelligibly predicating virtuousness of some of the actions or characters we observe or think about. Simply feeling or thinking that an action pleases us in a cer-tain way does not involve projecting or “spreading” anything on to the action. But feeling or thinking that the action is virtuous does. The “gilding or staining” operation which is supposed to lead to such thoughts could not therefore start from just such a feeling or impres-sion. It must start from a feeling or impression which is “of” something, or has an object, in the “intentional” sense; but it cannot be “of” any object or quality or relation which could be part of the way things “really stand in nature.” If it were, no “gilding or stain-ing” would be necessary.

We can of course have many false thoughts about the world, and even impressions of things that do not really exist. Hume’s view is that in the normal case that is because there are combined in our thoughts or impressions ingredients which we can and do fi nd in our experi-ence; it is only the complex combination which happens to fi nd no counterpart in the world. If I believe that there are unicorns I am wrong about the way things are, and if I open my eyes and get an impression of a unicorn I am not perceiving anything that actually exists. But in each case the “intentional” object of my thought or per-ception is something which in the widest sense could be so in the world; it just happens not to be. I can perceive and think what I do in that case because I have perceived both horses and horns in the past, and the thought of a creature with a horse’s body and a horn between its eyes is perfectly intelligible to me, even if that idea applies to no actual thing. What is especially problematic for Hume is not this ordin ary kind of contingent falsehood or delusion. The world as he conceives of it does not just happen to lack causal connections, virtu-ous characters, and beautiful objects. He does not just think that if things had been different in certain intelligible ways, those qualities and relations would have been there. There is no coherent place for them in any world which he conceives of. What is problematic is there-fore to explain how we can have intelligible thoughts or perceptions which do not represent “any thing, that does or can belong” to the way things “really stand in nature,” if we take the ways things could “really stand in nature” to exhaust the range of what could be so.

That is the problem I fi nd at the center of Hume’s philosophy. It is not unique to Hume. There is admittedly a completely general


problem of intentionality which he faces because of his own special conception of the mind and its contents. Strictly speaking, I believe he is not really in a position to explain how we could ever have any thoughts at all of something’s being so. That is largely because of that “theory of ideas” he inherited from Locke and Berkeley—a way of thinking about the mind which he seems to have imbibed without question just as he unhesitatingly took in the air he breathed. On that conception, the “objects” of the different senses—the only things sensed—are in each case strictly speaking only qualities. For Berkeley, for example, the only or proper objects of sight are colors and shapes, of touch, certain textures and degrees of hardness, and so on. The theory really leaves no room for the intelligible predication of those or any other perceived qualities to an enduring object, despite those philosophers’ understandable tendency to continue to speak as if it did.

In the face of this diffi culty Berkeley held that an object is really nothing more than a collection or combination of sensible qualities, or what he equivalently called “ideas.” But he never explained what a “combination of sensible qualities” amounts to. He was right to fi nd no help in what he thought was Locke’s idea of a “substratum”—a je ne sais quoi which somehow “supports” the qualities—but he was in no better position than Locke to explain how we can think, of an apple, that it is red, and round, and on the table. What looks like predication of such qualities to an object can be for him nothing more than a thought of a number of qualities somehow being “present” together. I think Hume makes no advance on Berkeley or Locke on this crucial matter.

I believe the diffi culty is connected with something deeper: the absence from this theory of any adequate notion of judgment, or assertion, or putting something forward as true. With no account of judgment it would be hard to fi nd a place for predication; predication yields a thought that is capable of truth or falsity. That is why I think Hume ultimately cannot even explain the possibility of our thinking of a particular ball as round, or as striking another. Thinking for him is too much like being presented with pictures. But even to see some-thing as a picture of a round ball, or of one ball’s striking another, one must be able to think of a ball as round, or as striking another, and that involves the ability to predicate a quality of an object, and to think of one object as related to another. Without an explanation of how we can make sense of such thoughts there can be no account


of how we could even have such a thing as an impression of one round ball’s striking another. Thought of an object, and of its having qual-ities and relations, must be possible for us in order to have such ex periences, and that requires in turn the possibility of judgment or assertion. This is obviously the kind of objection which Kant, for one, would be eager to press against Hume.

There is, then, a completely general problem of intentionality for Hume which should not be overlooked, or minimized. But even if we drop the restrictions imposed by the theory of ideas, as most philosophers nowadays would claim to have done, the most troubling aspect of the problem I have been drawing attention to seems to me to remain. It has to do with that notion of “the world” or of the way things “really stand in nature” which is supposed to exclude beauty and causal connections and the virtuousness or viciousness of actions and characters. A ball’s being round, or its striking another ball, is allowed to be part of that “world,” and so unproblematically available as an object of an impression. It might even be said that such things’ happening in the world is what explains why we get impressions of a ball’s being round, or of one round ball’s striking another; what fi xes the content of the perceptions we get on certain occasions is precisely what is so or what is going on in plain view on the occasion in question. If we followed Hume in supposing that thought is ultimately derived from perception, we might then be able to explain how it is pos-sible to think of one round ball’s striking another. We reproduce in thought what we have found in perception. But even on that view of thought we could not be said to have found the source of any thoughts we might have of the causal connection between the movements of those balls, or of the beauty of any objects, or of the virtuousness or viciousness of any actions or characters. On the view of “the world” shared by Hume and his many followers, there can be no such things in the world for us ever to get impres-sions of. That is precisely why the mental operation of “gilding or staining” the world is needed; it alone is supposed to produce the thought of something that is not really there.

Of course, one could simply deny that we ever do get any thoughts which in their content go beyond the way the world could really be: we think only either about happenings in our own minds or about the dispositions of objects to produce effects in minds that observe them. There is no doubt that we can and do think of the world in those


ways, but I have been considering the view that we also believe more; that we predicate moral and aesthetic qualities of objects and attribute necessity to some of the relations between them. But if we do eventually come to think coherently of beautiful objects, of virtu-ous and vicious actions and characters, or of causal connections between things—however we manage to do it—how can we then hold that the world does not and cannot include such things? How can we make sense of the idea that we do indeed think things are that way, if we also think that they really are not? And if things in the world are not really that way, how can we explain the fact that we nevertheless think that they are? The Humean suggestion I have been considering is that our thoughts are generated by a creative or productive process which “takes” something or other from our impressions or feelings and leads us somehow to “spread” what it takes on to objects which we unproblematically believe to populate the world. But that is only a prejudice or a hope or a fairy-story without a convincing account of exactly what we “take,” and exactly how it is turned into some-thing which it becomes intelligible to predicate of objects or the rela-tions between them.

The feelings or impressions from which the “gilding” story is sup-posed to start cannot be described from the outset as impressions of something in which the very feature that we are said to “spread” appears already in intelligible predicative position, applied to an object. No “new creation” would then be needed to give us the cor-responding ideas or thoughts about objects in the world. The thoughts we eventually get are said to predicate something “new” or “added” to objects “as they really stand in nature.” So something must hap-pen between the original impression or feeling and the subsequent idea to generate what to us will be a newly intelligible predicative thought.

I think we do have intelligible thoughts about beautiful objects, the virtuousness and viciousness of actions and characters, and causal connections between things that happen. But if we do, how can we avoid regarding such thoughts as representing the way things are, or the way things are in the world? What is the notion of “nature” or “the world” employed by Hume and all those phil osophers who hold that only some, but not all, of the things we seriously and unavoida-bly believe represent things “as they really stand in nature” or in “the world?” For them it is apparently not true that the world is everything that is the case, or that the world we believe in is everything we believe


to be the case. They draw an invidious distinction within all the things we believe; the “world” they think we believe in amounts to some-thing less than the truth of everything we believe. But on that view the very possibility of our having and making sense of those “extra” thoughts has still to be accounted for. 15

15 I am grateful to Janet Broughton and Hannah Ginsborg for very helpful critical comments on earlier versions of this essay.

9

The Constraints of Hume’s Naturalism

It was exactly one hundred years ago that Norman Kemp Smith (or Norman Smith as he then called himself) published a two-part article in Mind called ‘The Naturalism of Hume’. 1 It gave the outlines of a new understanding of what Smith saw as the enduring signifi cance of Hume’s philosophy. And it amounted to a defence of Hume against what was then the dominant and largely unsympathetic inter-pretation. That view was perhaps best represented by the detailed exposition and unrelenting attack of T. H. Green.

Green had called Hume many things, 2 but never a naturalist. According to Green, Hume’s philosophical achievement was com-pletely negative: he explicitly reduced the already unbelievable con-ceptions of thought and experience to be found in Locke and Berkeley to their absurd and paradoxical consequences. He showed only that on that basis we can know nothing of any external world of con-nected, enduring bodies or events or processes, and nothing of any thinking, active beings, even including ourselves. We can have no rea-son to believe in any such things; there is nothing to be found in any world we can make sense of beyond the momentary impressions and ideas that pass fl eetingly into and out of our awareness.

Green was one of the two editors of what for many years was the standard edition of Hume’s collected philosophical works. In the long introduction he wrote to those works Green launched an uncom-promising diatribe against everything he took Hume to believe. Why Green thought anyone would ever want to read the volumes he had so laboriously edited if he was right about what they contain is a question he did not try to answer. An answer of sorts was given sixty

1 Norman Smith, ‘The Naturalism of Hume’, Mind 54, 1905, pp. 149–173; pp. 335–347 . 2 Green speaks of Hume’s “inconsistency”, his occasional “charlatanry”, and his

“undoubted vanity”. See T. H. Green, Hume and Locke , Thomas Crowell, New York, 1968 .


years after Green, still very much in the same spirit, by H. A. Prichard of Oxford, who said:

Hume’s reputation as the arch sceptic leads all of us, I think, to approach the Treatise for the fi rst time with a certain trepidation - . . . Yet I think that close consideration of the argument will convince anyone that our fears are groundless. In fact to my mind the Treatise is one of the most tedious of books, and close examination of it renders me not sceptical but angry. Of course, there is a great deal of cleverness in it, but the cleverness is only that of extreme ingenuity or perversity, and the ingenuity is only exceeded by the perversity.

It could be wished that the student of philosophy could be spared all contact with Hume, and thereby the trouble of rooting out some of the more gratuit-ous forms of confusion common in philosophy. 3

Norman Smith’s aim in his original article was to demonstrate the presence in Hume of an original and positive philosophy of great interest and importance. In Smith’s words:

the establishment of a purely naturalistic conception of human nature by the thorough subordination of reason to feeling and instinct is the determining factor in Hume’s philosophy. 4

This promised a constructive way of accounting for what is undeni-ably a negative or sceptical side of Hume’s philosophy, but without supposing that the main goal of his investigations was to reach such destructive and potentially dispiriting conclusions. Like Descartes and Boyle and other philosophers before him, Hume too was also to be seen as making use of the lessons of philosophical scepticism rather than simply accepting its apparent conclusions as they stand. A new and positive understanding of human nature was to be achieved by appreciating what those unavoidably sceptical lines of thought really reveal. A ‘sceptical’ enquiry into human nature will be an essential part of any positive attempt “to introduce the experimental method of reasoning into moral subjects” – the overall purpose of A Treatise of Human Nature , as Hume announced in its subtitle.

On Smith’s view, Hume was fully aware of the failure of the Lockean theory of ideas to account for all our experience and know-ledge. But for Hume the failure was due to the faulty conception of

3 H. A. Prichard, Knowledge and Perception , Oxford University Press, Oxford, 1950, pp. 174–175 .

4 Smith, op. cit. , p. 150.

The Constraints of Hume’s Naturalism 191

the function of reason that was implicitly combined with that theory. Hume’s conclusion (in Smith’s words) was that:

We cannot by means of our reason explain any of the ultimate characteris-tics of our experience – the origin of our sensations, the true ‘secret’ nature of causal connexion, apprehension of external reality, appreciation of beauty, judgment of an action as good or bad. And the alternative is not scepticism, but the practical test of human validity. Certain beliefs or judg-ments . . . can be shown to be ‘natural’, ‘inevitable’, ‘indispensable’, and are thus removed beyond the reach of our sceptical doubts. 5

With all such judgements, as Hume himself puts it in a well-known passage, “Nature has not left this to [our] choice, and has doubtless esteem’d it an affair of too great importance to be trusted to our uncertain reasonings and speculation” ( T I, 4, ii; 187). 6 Or again, “Nature, by an absolute and uncontroulable necessity, has determin’d us to judge as well as to breathe and feel” ( T I, 4, i; 183).

Here “nature” is emphasized as the dominant force. And Smith speaks accordingly of Hume’s “naturalistic view of reason” 7 and of his insistence on “purely naturalistic explanations” 8 of all aspects of human behaviour and human life. Why this insistence on the force of nature and on the need for purely naturalistic explanations of human thought and behaviour? There would be no need to insist that the study of ani-mal behaviour must proceed “naturalistically”. That would go without saying. How else could we fi nd out about animals except by observing them as they actually are in nature and seeing how they “naturally” or “by nature” come to respond and act in all the ways they do? Why then does a special case need to be made for a fully “naturalistic” study of human nature or human life? Why are we not equally ready to attribute all human responses, thoughts, and actions to basic propensities of human beings to respond “by nature” to the world that is presented to them?

Asking this question helps to identify the special place and the special need for the negative or sceptical philosophical reasonings in

5 Smith, op. cit. , p. 152. 6 References to Hume’s works in the text are to Book, Part, Section, and Selby-Bigge

pagination of Hume, A Treatise of Human Nature , ed. L. A. Selby-Bigge, Oxford University Press, Oxford, 1978 (as T ) and to Section, Part, and Selby-Bigge pagination of Hume, Enquiries Concerning the Human Understanding and Concerning the Principles of Morals , ed. L. A. Selby-Bigge, Oxford University Press, Oxford, 1975 (as E ).

7 Smith, op. cit. , p. 158. 8 Smith, op. cit. , p. 166n.


Hume’s overall program. They are needed because human thoughts, judgements, and actions have been widely thought to be arrived at by reason or reasoning in a way that cannot be captured in a purely naturalistic study. On that conception of reason and of its role in human life, belief or judgement or action is thought to be the out-come of independent refl ection or consideration or inference that is meant to provide justifi cation or rational support for the conclusions that we are thereby led to accept. Reason on this view involves a grasp of one thing’s being reason to believe another, and acceptance of the idea that assent to a conclusion is recommended or warranted only if all the considerations that lead to it give one reason to believe it. Human thought and action, then, would be seen as distinctively human only to the extent to which it can be understood in terms of its reasonableness, or its conformity to reason, not simply as a phe-nomenon explicable by the blind forces of nature.

But this view of reason, as Smith points out, “is at the mercy of the philosophical sceptic”. 9 Pushed to the limit, as it is in Hume’s hands, and as the view itself seems to require, it can be exploited to bring into question any conclusion that anyone thinks is warranted or any of the principles thought to warrant any conclusion. This view of reason opens up the possibility of a never-ending process of doubt. Smith sees Hume’s use of sceptical reasonings as intended to show that, if we take this view of human reason seriously as the source or explanation of human belief, we would have to admit that human beings would believe nothing at all. The point of Hume’s sceptical considerations, then, is precisely to discredit this concep-tion of reason and to displace it at the centre of human life with whatever can be found to be true of human beings by the cautious observation of the natural processes by which they arrive at the responses, judgements, and actions by which they actually live. No such negative, sceptical step would be needed as part of a natural-istic study of non-human animals. The preconception that human beings are different from the other animals in just this respect is what must be independently exposed and abandoned before the positive naturalistic study of human beings can be assessed and accepted on its merits alone.

Norman Smith’s goal, then, was to explain how the two differ-ent aspects of Hume’s philosophy – the “sceptical” side and the

9 Smith, op. cit. , p. 165.


“naturalistic” side – can be understood to fit together. They must both be seen as essential to what Hume was trying to bring out about human beings. This I believe to have been a real interpret-ative breakthrough, and a positive and lasting contribution to our understanding of Hume’s philosophy. It did not win immedi-ate or universal acceptance. Hume continued to be regarded as little more than a sceptic in philosophy through the early and middle years of the twentieth century, as the remarks of Prichard and many others amply show.

Smith developed his original idea much more fully and in more convincing detail thirty-fi ve years later in his book, The Philosophy of David Hume of 1941. 10 In the interval, as he explains in that book, he had come to the conclusion that Hume “entered into his philosophy” “through the gateway of morals”, 11 and that he was heavily infl uenced by Francis Hutcheson’s writings, especially on the relative roles of reason and feeling or passion. Smith’s idea was that what Hutcheson had held to be true about the domin-ance of feeling or sentiment in judgements of beauty or goodness was generalized by Hume to all judgements whatever, culminating in the view (in Hume’s words) that “belief is more properly an act of the sensitive, than of the cogitative part of our natures” ( T I, 4, i; 183). This was the revolutionary Humean turn, and it took both the ‘sceptical’ and the ‘naturalistic’ strands of his philosophy to establish it.

The philosophical world (especially in Britain) had other things on its mind in 1941 than the scepticism and naturalism of Hume. Again there was little or no immediate acceptance of Kemp Smith’s interpretation. And by the time the war ended and philosophy came fully alive again Hume was being enthusiastically co-opted as an enlightened early precursor of the then-popular logical posi-tivism or logical empiricism. He was even said to have devised an “empiricist criterion of meaning” that would condemn to the fl ames all volumes “of divinity or school metaphysics, for instance” on the grounds that they are “cognitively meaningless”. 12

10 Norman Kemp Smith, The Philosophy of David Hume , Macmillan, London, 1941 . 11 Smith, The Philosophy of David Hume , p. vi. 12 A. J. Ayer, Language, Truth and Logic , Oxford University Press, New York, 1946, p. 54 .


On that roughly positivistic understanding of Hume’s philosoph-ical importance there was no room for any naturalistic investigation of how human beings come to think and feel and act in all the ways they do. That would be an empirical study, and there is certainly nothing wrong with such a thing in itself. But on the positivistic conception, no such study could be part of philosophy, however interesting it might be on other grounds. Philosophy could be at best purely a priori , and so exclusively “analytic”. This was another obstacle to ready acceptance of Kemp Smith’s emphasis on Hume’s “naturalism”.

What has subsequently come to be called “the Kemp Smith inter-pretation of Hume” has been disputed in recent decades and in some cases rejected on quite different grounds. Much of what has been in question are largely historical claims about the extent to which the works of Hutcheson actually had the effects on the ori-gins of Hume’s philosophy that Kemp Smith claimed for them. There has also been the question of whether Kemp Smith got Hutcheson himself right. On these matters I take no detailed stand – although I must say that Kemp Smith’s account has always looked extremely plausible to me, especially when you see those sentences in Hume’s moral philosophy that are more or less directly lifted from Hutcheson, and then you see the same terms and the same distinctions employed to similar effect in Book I of the Treatise , where morality is not in the picture at all. But what I think can no longer be disputed, and what I regard as a secure and permanent contribution to Hume scholarship, is Kemp Smith’s account of the relative roles of the negative and the positive, the “sceptical” and the “naturalistic”, aspects of Hume’s philosophy, and his explan-ation of how the two were seen by Hume to fi t together.

That both these elements are essential to the combination as Hume understands it makes it wrong, I believe – or at least it would serve no useful purpose – to call Hume, simply, only a sceptic, or only a nat-uralist. Applying only one or the other of the two labels to Hume would be an especially serious distortion of his thought. And I think we owe to Kemp Smith the proper understanding of how and why that is so.

Naturalism by now is widely believed to be a very good thing in philosophy – perhaps the only thing. So there is perhaps no longer resistance to Kemp Smith’s account of Hume’s naturalism on purely philosophical grounds. But that account says that both the so-called ‘sceptical’ and the so-called ‘naturalist’ sides are essential to a proper


understanding of Hume’s philosophy. And that combination is what gives that philosophy its distinctive character and its distinctly pro-vocative air. It is also what makes that philosophy so hard to accept, or makes the attempt to accept it so disturbing. I think this is not always appreciated, especially by advocates of naturalism. The spe-cial embarrassments and dissatisfactions of Hume’s naturalism in particular are what I want to try to bring out more clearly. They per-haps suggest a more general lesson about naturalism, at least when it is indulged in for philosophical purposes.

Naturalism alone as an attitude or a programme of study is not the source of the diffi culty. There is nothing in itself troubling in study-ing human beings as part of the observable world and trying to explain how they come to think and feel and act in all the ways they do. I don’t mean that it is easy to explain everything – or even any-thing – in that way. I mean only that there is a reasonably straightfor-ward enterprise of treating the human beings you are studying as part of the very nature you know they operate in and trying to dis-cover the ways they are naturally disposed to behave in such a world, and why. That is what we do in the study of other animals. And Hume always wanted to keep the comparison with other animals clearly in mind in what he called his “science of human nature”.

The naturalistic study of human beings or of anything else in nature requires that the person carrying out the study have a concep-tion of what that nature in which those beings operate is actually like. He must have some view or other of the natural world. In the study of human beings in particular it is also necessary to have some con-ception of what those human beings do and think and feel in response to that world of nature. That is what is to be explained. Hume wanted to discover and formulate what he called general “principles of the mind” or “principles of the imagination” that would explain how what human beings receive from the world they live in produces the elaborate conception of the world and their place in it that in one form or another all human beings eventually have. To put it in crude terms, it is a question of how human beings get, or what takes them, from input to output. This requires a detailed conception of what that “output” actually is. What is it about human thought, feeling, and action that the “science of human nature” is supposed to explain? Whatever it is, it is something that the student of human nature must be able to recognize and understand and describe. That is what gives him his subject matter. In the study of the other animals, for all their


complexity, what is to be explained seems by comparison much less elaborate.

The negative or sceptical side of Hume’s enterprise is meant to show that the processes by which human beings get from their input to their output is “ not founded on reasoning, or any process of the understanding” ( E IV, 2; 32). The transition must be understood in some other way. One reason for that is what Hume sees as the poverty of the input. Here the theory of ideas from Locke and Berkeley is strongly at work. All anyone is given in experience are momentary, fl eeting impressions of this or that sensory quality. And from those beginnings alone no thought could take anyone by reasonable, defen-sible steps to the elaborate view that human beings in fact all have of a world of enduring independent bodies in constant interaction with one another and with the human beings that observe them.

Even the belief that there are such things as active, thinking human beings with experiences and thoughts and feelings of their own goes far beyond anything that anyone could be led to believe by any kind of reasoning from the meagre resources human beings are restricted to in sensory experience. And that is not simply because of the richness or complexity of what human beings come to believe about the world. Even if two sensory qualities have always appeared one after the other in a person’s experience in the past, the simple expectation of something of the second kind, given an experience of something of the fi rst, is “ not founded on reasoning, or any process of the understanding” either. Human beings always do in fact get such expectations under such conditions, Hume says, but nowhere in “all the branches of human knowledge” is there anything that could supply an “argument” from the way things have been found to be in the past to a conclusion about how they will be in the future ( E IV, 2; 35). So much the worse for other more complex beliefs that go beyond what can ever be presented in any particular experience or set of experiences, such as beliefs in enduring physical objects, in causal connections between them, or in the effects of such objects on the human beings who take themselves to perceive them and believe in them.

The wish to see human beings as generically distinct from the rest of the animal world in being uniquely rational, then, leads to the conclusion that human beings in fact have no reasons for believing or doing what they do. There are reasons why they think and behave as they do. It is the task of the “science of human nature” to discover


such reasons. But human beings are not given anything in their ex perience which is, or which they can rightly see to be, reason to believe and do what they do. They might believe that they do have good reasons for some of their beliefs, but that is only because they have not pushed far enough in satisfying the demands of reason. ‘Sceptical’ considerations, relentlessly applied, show that they would be wrong in their beliefs in their own rationality. Given the way human minds work, human beings who receive the kinds of experi-ences they get from the world cannot help coming to believe the things they do. But no one can defend those beliefs with reason while accepting the demands that that conception of unfettered reason seems to impose. The inevitability of belief, and the impossibility of defending it by reasons seen to be sound even in the face of ‘sceptical’ attack, reveals what Hume calls:

the whimsical condition of mankind, who must act and reason and believe; though they are not able, by their most diligent enquiry, to satisfy themselves concerning the foundation of these operations, or to remove the objections, which may be raised against them. ( E XII, 2; 160)

This “whimsical” state of unresolved dissatisfaction is not a condi-tion that a naturalistic student of human nature will actually fi nd mankind to be in. It is not a state that any human being can naturally fi nd himself to be in for very long, as Hume points out:

Most fortunately it happens, that since reason is incapable of dispelling these clouds [of dissatisfaction], nature herself suffi ces to that purpose. ( T I, 4, vii; 269)

Nature breaks the force of all sceptical arguments in time, and keeps them from having any considerable infl uence on the understanding. ( T I, 4, i: 187)

The force of nature here, and how it does its work, is what the nat-uralistic student of human nature seeks to discover and understand. That would be to explain how in fact human beings get all the beliefs and other attitudes that they do.

But there is a certain constraint on such a naturalistic study as Hume envisages it. The investigation must proceed in full recognition of what I have called the poverty of the input — the restricted char-acter of the impressions that originally give rise to the whole concep-tion of the world and their place in it that human beings end up with. This means that with only fl eeting and momentary impressions to start with, the task for the “scientist of human nature” is almost


exclusively that of discovering some general “principles of the imagin ation” in accordance with which human beings with such lim-ited experiences will universally and inevitably come to believe and act in all the ways they do. This is what imposes a troubling constraint on Hume’s naturalism. It is not a constraint on naturalism as such, or on naturalism in every form, but on Hume’s naturalism in particular.

What I have called the poverty of the input has the consequence that, for instance, the public world of independently existing phys-ical objects that all human beings inevitably come to believe in plays no role at all in the naturalistic explanation of how human beings come to believe in such a world. In that section of the Treatise called “Of Scepticism with regard to the Senses” Hume explains the origin of the belief in what he calls “the continued and distinct existence of objects”. His explanation is intricate and diffi cult to follow in detail, but one thing that is clear about it is that it makes no appeal at any point to the existence of any such independently existing objects. The story unfolds completely in terms of various principles accord-ing to which perceptions come and go in the mind, and a kind of mental “inertia” by which the mind slides easily along certain series of perceptions despite the obvious dissimilarities among some of its members. These tendencies would lead to intolerable confl icts or tensions in the mind’s operations in general were it not for the fact that the mind naturally indulges in what Hume calls a certain “feign-ing” ( T I, 4, ii; 208) or “supposing” ( T I, 4, ii; 199) that has the func-tion of removing the potential confl ict and thereby putting the mind at ease.

As a result of these elaborate operations of the “principles of the imagination” all human beings acquire a belief in the continued and distinct existence of objects. Hume describes that belief as a kind of “fi ction” ( T I, 4, ii; 205, 209, 215) – something that the mind is naturally but inevitably led to invent and then to hold on to for its own ease or equilibrium. Nothing in nature is called on to explain that belief. Or rather nothing in anything other than psychological or mental nature is appealed to; nothing beyond momentary and fl eet-ing perceptions and those principles of the imagination that hold in the mental realm of the comings and goings of perceptions. For all that Hume says in his naturalistic explanation of the origin of the belief in independently existing objects, human beings would be nat-urally brought to believe in such things whether there really were any such enduring objects in the natural world or not.


The same is true of the belief in a continuing personal self or sub-ject that every human being cannot help acquiring. That idea too is a “fi ction” that naturally arises in any series of perceptions that ex hibits certain recurrent characteristics. As Hume explains it:

However at one instant we may consider the related succession [of percep-tions] as variable and interrupted, we are sure the next to ascribe to it a per-fect identity, and regard it as invariable and uninterrupted. Our propensity to this mistake is so great . . . that we fall into it before we are aware;

[We] yield to it, and boldly assert that these different related objects are in effect the same . . . In order to justify to ourselves this absurdity, we often feign some new and unintelligible principle . . . Thus we feign the continu’d existence of the perceptions of our senses, to remove the interruption; and run into the notion of a soul , and self , and substance , to disguise the vari-ation. ( T I, 4, vi; 254)

This means, as Hume puts it, that “The identity, which we ascribe to the mind of man, is only a fi ctitious one” ( T I, 4, vi; 259); “there is properly no simplicity in it at one time, nor identity in different; what-ever natural propensity we may have to imagine that simplicity and identity” ( T I, 4, vi; 253). These “mistakes” or “suppositions” or “imaginings” are discoveries of Hume’s naturalistic “science of human nature”. They account for the origin of beliefs that all human beings have. But they are discoveries of nothing in nature except cer-tain series of perceptions and certain natural propensities for percep-tions of certain kinds to occur in series of those kinds. Flesh and blood human beings as they walk the earth do not come into the story at all. Nor does the earth, for that matter, or any of the objects that naturally exist on it.

All such objects are further “fi ctions”, and there are other Humean “fi ctions” or “suppositions” that human beings also inevitably fall into. Perhaps the best known of all is the belief in causal or necessary connections between the objects and events that human beings take themselves to observe in the world. Having seen things of one kind invariably follow things of another kind, no one can help expecting something of the second kind given an experience of something of the fi rst kind. Not only that; everyone will believe that something of the second kind must happen. “There is a NECESSARY CONNEXION to be taken into consideration” ( T I, 3, ii; 77). We “immediately conceive a connexion between” what we have observed and what we expect ( T I, 3, xiv; 163). But repeated observation of things of two kinds produces


something new only in the mind, not in the objects themselves. So the source of the idea of necessity is only something that happens in the mind.

Upon the whole, necessity is something, that exists in the mind, not in objects; nor is it possible for us ever to form the most distant idea of it, consider’d as a quality in bodies. ( T I, 3, xiv; 165–166)

In human beings’ thoughts of causation, then, just as in their thoughts of the continued and distinct existence of objects and their thoughts of human subjects or thinkers or agents, they take something that arises in the mind solely as a result of the comings and goings of perceptions there, and they form a new idea that they then project outwards in their thoughts and beliefs and come to regard it as rep-resenting a feature of the way things actually are in the world. That explains how and why human beings come to believe certain things about the world they live in. But their actually living in such a world and fi nding such things within their experience of it is no part of that naturalistic explanation.

The same is true of judgements of taste and of what Hume calls “moral distinctions”. Even in something you regard as “vicious” – an act of “Wilful murder, for instance” – “You never can fi nd” the vice “as long as you consider the object”. You must “turn your refl exion into your own breast, and fi nd a sentiment of disapprobation, which arises in you, towards this action”. That is something that “lies in yourself, not in the object” ( T III, 1, i; 468–469). We make what look like judgements ascribing certain properties to objects or actions independent of us, but our doing so is to be explained exclusively in terms of certain “sentiments” or “passions” in our minds. There is nothing in any of the objects to which human beings seem to ascribe such properties that could explain their having the responses or mak-ing the judgements about them that they do.

Kemp Smith’s hypothesis was that Hutcheson’s use of this idea in the case of beauty and goodness is what served as Hume’s model for all the other more elaborate “fi ctions” or “feignings” or “suppos-itions” he eventually introduced in explaining what seem to be the cooler and more detached operations of “the understanding”. Those explanations are much more elaborate than an apparently simple story of projecting our feelings or sentiments on to an otherwise neu-tral world. But for all the differences among the different “fi ctions” that Hume invokes, Kemp Smith was defi nitely on to something that


they all have in common. What is true of the story that we “spread” or project our feelings on to the world and of all the other more elaborate “fi ctions” Hume introduces is that in no case is anything that is actually so in the public, observable world brought in to explain why human beings come to believe that the world is this or that way. That is at least one thing that is meant by calling the result-ing belief a “fi ction”.

Perhaps Hume thinks something even stronger is common to all these “fi ctions”. In the course of explaining the origins of all these beliefs he says, for instance, that “’tis a false opinion that any of our objects, or perceptions, are identically the same after an interruption” ( T I, 4, ii; 209). “’Tis a gross illusion to suppose, that our resembling perceptions are numerically the same” ( T I, 4, ii; 217), even though “the imagination is seduc’d into” believing in such an identity ( T I, 4, ii; 209). As we saw, he also says “there is properly no simplicity in [the mind of man] at one time, nor identity in different”, although we have a “natural propensity” to believe in that simplicity and identity ( T I, 4, vi; 253). And causal necessity, he says, “is something that exists in the mind, not in objects” ( T I, 3, xiv; 165). There is no such connec-tion between things we come to think of as causes and effects. And in the case of moral judgements he says “vice and virtue . . . are not qual-ities in objects, but perceptions in the mind” ( T III, 1, i; 469). Taken strictly, the “fi ction” in each of these cases appears to be something that Hume as a “scientist of human nature” regards as false. He explains the “fi ctions” or “mistakes” or “illusions” or “feignings” or “suppositions” that he fi nds human beings inevitably fall into. And on this reading the beliefs they come to hold in those ways would be one and all false.

It is this attribution of “fi ctions” to all human beings that I think gives Hume’s version of naturalism its peculiar character and its dis-tinctly provocative air. And whatever exactly he means by “fi ctions” – whether or not he means that they are all strictly false – that is what makes it so hard to accept his naturalistic explanations. It is not hard to accept that human beings, for various intelligible reasons, some-times believe things without having suffi cient reason to believe them. It is not hard to accept that human beings sometimes even come to believe things that are actually false. A naturalistic student of human nature could take such things in his stride. He could come to one or the other of those conclusions by explaining how and why the human beings he studies nonetheless come to believe the things they do.


Their having no reason to believe them, or even their being false, would be no impediment to explaining in particular cases why people believe such things.

Even if Hume does not explicitly regard the so-called “fi ctions” he attributes to human beings as not supported by reason or as strictly false, his explanations fi nd those beliefs to be solely the products of the mind’s own self-contained activities. In calling them “fi ctions” he would at least be pointing out that the beliefs in question do not have to be true in order for human beings inevitably to come to believe them. Just as we do not have to credit animals with beliefs about, say, the causal connections between things in order to explain their behav-iour in seeking food or in avoiding danger when it threatens them, so the Humean naturalist does not have to attribute to human beings true beliefs in the continued and distinct existence of objects or in human agents or in causal connections or in the moral qualities of actions in order to explain people’s believing such things and direct-ing their behaviour accordingly. Exactly what Hume had in mind in calling such beliefs “fi ctions” is perhaps open to question. Does he mean that the beliefs are strictly speaking false, or not?

Whatever he means, there are two distinctive features of the Humean naturalistic study that make it more diffi cult to accept its conclusions than those of more familiar and less constrained nat-uralistic investigation. For one thing, Hume does not simply fi nd occasional absence of good reasons or even occasional outright error somewhere within the rich body of human belief. The “fi ctional” or purely subjectively-generated character of human belief is said to extend to all beliefs in any enduring bodies or in any active, thinking subjects or in any causal connections between things or in any moral qualities of any actions at all. The whole conception from one end to the other is seen to be an elaborate put-up job that the human mind cannot help indulging in as long as it receives the kinds of impres-sions that set its “principles of the imagination” in action. The truth of any of the beliefs it acquires either is denied or at least not affi rmed in the explanation of the mind’s coming to believe it. This is a com-pletely general account of virtually all human belief in anything at all. And that makes it much harder to accept its results than those of accounts that reveal only occasional error or illusion or “fi ction”.

What makes it doubly diffi cult to accept these results is another obvious but signifi cant feature of naturalistic study. What is true of the naturalistic study of both animals and human beings is that the


study is carried out in each case by human beings. That presents no diffi culty for our understanding whatever we fi nd to be true of ani-mals. But when human beings are both the agents and the objects of naturalistic study, what is found to be true of the objects studied must somehow also be understood to be true of those of us who conduct the study. And that is what makes it hard to accept Hume’s results. We can perhaps be led, at least for a moment or two, to sup-pose that those conclusions might be true of others. But it is hard for any of us to accept the results of the Humean study as true of our-selves. It is diffi cult if not impossible for us to accept that we too inevitably indulge in what we regard as nothing more than “fi ctions”. That does not give us the kind of understanding or illumination of ourselves that we expect from the study of other parts of nature.

I have said that the naturalistic student of human beings or of anything else in nature must have some conception of what that nature in which those beings operate is actually like. Hume speaks of “the cautious observation of human life” and of fi nding our experi-ments “as they appear in the common course of the world, by men’s behaviour in company, in affairs, and in their pleasures” ( T Intro.; xix). Those are the kinds of things that we all believe are happening in the everyday world we take ourselves to believe in. But if that whole world of enduring, interacting bodies and human persons act-ing and communicating with one another is not part of the Humean naturalistic explanation at all, and is itself a “fi ction” and is under-stood as such by us, where does that leave us?

If we take it that the belief in such a world is a “fi ction” in the strong sense of being strictly speaking false, we are left with no world at all in which the alleged objects of our naturalistic study can be believed to reside. Even if we cannot help believing that there is such a world, and it is populated with beings that are of great interest to us, we cannot acknowledge that our belief in such a world is false and continue to fi nd the same kind of interest in its illusory inhabitants. There would be no believed-in world of nature left to us in which to conduct our naturalistic study of what goes on there. We could not consistently believe both that the world is the way the Humean explanations show that we cannot help believing it is and that our beliefs in such a world are one and all “fi ctions” in the sense of being false.

Even if that world we all believe in is simply not appealed to any-where in the Humean naturalistic explanation of how we come to


believe in it, it will be a “fi ction” in a weaker but still disturbing sense. It will be something we fi nd we cannot help believing in. Even if we are not obliged to explicitly regard it as false, we will see it as some-thing we would believe in, just as we do, whether there actually were any world like that or not. We would see that we would inevitably have come to believe in it whatever the world around us, if any, actu-ally happens to be like. I can perhaps understand how I could see others’ beliefs about the world as “fi ctions” in that sense; I would explain how they come to believe what they do whatever the world around them happens to be like. But can I see my own beliefs as “fi c-tions” in that way?

What could I appeal to to explain them? And if I accept that what I too believe is nothing more than a “fi ction” in that sense, can I con-tinue to regard that conception of mine as something I believe to be true? Can I hold that conception of the world to be true when I con-cede that I would inevitably believe exactly what I now believe even if the world were not the way my conception of the world says it is? Could I see myself as anything other than a credulous victim of the blind workings of certain “principles of the imagination”? Is that a satisfactory outcome of the naturalistic study of human beings, including myself ?

Hume was well aware of the disturbing plight the consistent pur-suit of his naturalistic science of human nature seemed to leave him in. He gives poignant expression to it in the concluding section of Book I of the Treatise . He fi nds himself almost in “despair” at the end of what he calls “the most accurate and exact” of his naturalistic “reasonings” about the human understanding ( T I, 4, vii; 265). He realizes that he can say nothing more than that he has “a strong pro-pensity to consider objects strongly in that view, under which they appear” to him ( T I, 4, vii; 265). That is all he can say on his own behalf and in support of the view of the world that he fi nds he can-not help holding. He acknowledges that “habit” alone determines him to “form certain ideas in a more intense and lively manner, than others” ( T I, 4, vii; 265). But that seems to him such a “trivial” qual-ity of the mind, “and so little founded on reason” ( T I, 4, vii; 265). And he laments that “if we assent to every trivial suggestion of the fancy . . . we must at last become asham’d of our credulity” ( T I, 4, vii; 267). But there is no alternative. He sees that he will inevitably believe just what he does, and as a result of the operations of “principles of the imagination” that he cannot help believing in. That discovery, he


says, is what “cuts off all hope of ever attaining satisfaction” ( T I, 4; 267). And of course here Hume is speaking personally. He tries to apply the results of his naturalistic investigations to his own, fi rst-person case.

Now we know, as Hume points out, that this “despair” cannot last. “Nature herself suffi ces” to “dispel these clouds” of “philosophical melancholy and delirium” ( T I, 4, vii; 269). Hume dines, he plays a game of backgammon, and converses and is merry with his friends, and all is well. All is well psychologically, that is, at least for a while, when the Humean scientist of human nature “cannot fi nd it in [his] heart to enter into [his naturalistic investigations] any farther” ( T I, 4, vii; 269). But however agreeable that state might be, it is not a satis-factory outcome of the naturalistic study of human beings. We can have agreeable conversation and backgammon with our friends with-out giving any thought to the naturalistic study of human nature. But the fact that nature can be relied on to deliver us into such a happy state whenever we despair of understanding human nature in the right way does not itself yield a satisfactory understanding of human thought and belief and action. That is what we seek. The unsatisfac-tory and unacceptable picture we come up with on Hume’s view is simply submerged by the force of nature; it is not made more intelli-gible or more palatable to us.

Hume fi nds himself in this plight precisely because he thinks he has explained how human beings come to believe all the things they do. He thinks he has got the correct explanation, but he fi nds it dis-satisfying or unacceptable when he tries to apply it to his own case. But has he really explained any such thing, naturalistically or not? That is the question. Are “principles of the imagination” according to which mental particulars like impressions and ideas come into and go out of existence ever enough in themselves to explain how human beings come to believe certain things and have thoughts about the world that are either true or false?

I would say that he does not succeed in that explanatory task. One of the problems that creates a big obstacle is the nature of belief or judgement , as opposed to having an impression or an idea. Hume thought he was one of the fi rst philosophers to recognize that prob-lem. And I think he knew he did not really solve it. If that is right, the explanations he offers of our believing in objects and persons and causation and morality and so on do not actually succeed in explain-ing what they are meant to explain. The only materials Hume thinks


are available – impressions and ideas coming and going in the mind – are not enough in themselves, however elaborately they might be com-bined, to yield judgements or beliefs which we hold to be true or false. To explain how human beings come to have such attitudes, and so to act and interact in the world in the light of them, I think you need a conception of nature that includes a whole world of enduring bodies in space and human bodies and human actions in interaction with them and with one another. Only with such a relatively rich concep-tion of nature is there any hope of explaining what we want the sci-ence of human nature to account for.

But then the conception of the world of nature that any naturalis-tic student brings to his study of human beings will be the very same, at least in general terms, as the conception of the world he fi nds his human beings to hold and which he explains their possession of. The question for such a naturalist will be, in effect, how those human beings he studies come to believe that the world is more or less just the way he already takes it to be. At its most general, this will be the task of general learning theory, or the theory of human development and socialization.

Hume saw his enterprise as more challenging and potentially more illuminating. He wanted to explain naturalistically how human beings arrive at something that is completely “new” and almost infi n-itely richer than the meagre set of mental materials from which he thought they had to begin. Such an explanation, if it worked, would be truly illuminating. It would be impressive in direct proportion to the size of the gap that it posits between the impoverished non- propositional input human beings receive and the profuse and highly articulated output of beliefs and other attitudes they end up with. But it is just that huge gap that I think left Hume in the unsatisfac-tory position he knew he had reached. I don’t think he succeeded in crossing that explanatory gap. He thought he had crossed it, but when he tried to apply to himself and his own thoughts the kinds of explanations that seemed to work so well for others, it still left him forlorn.

10

Practical Reasoning

I want to explore some basic questions about what practical reason-ing is, or is supposed to be, and what is special about it. Unfortunately, I do not get beyond a few fundamental issues. Even so, I get far enough to fi nd myself at odds with what has often been said about them. Any wider implications for the proper understanding of action, reasons for action, and even morality, must be left for other occasions.

Each of us faces practical problems. What shall I do? What shall I do now? What shall I do tomorrow? What is the best thing for me to do about this? What should I do? (understood simply as a prac-tical question for me and not necessarily what society or morality or honor or some other such system requires). I just want to know what to do. Questions like this are asked and answered by and about par-ticular people in determinate circumstances, at particular times and places, every day.

We settle such questions in different ways. Or perhaps there are different sorts of question. For example, I decide now to go to the movies tomorrow. After some thought, that is what I conclude is the thing to do tomorrow, and I now decide and so intend to do it. For the moment that question is settled. I can always re-open the ques-tion between now and tomorrow and settle it in a different way. And even if I don’t, I still might not do tomorrow what I settled today is the best thing to do. As for what to do in the here and now, that ques-tion is settled by doing something at the time, even if that means, as we say, doing nothing. If I carry out my present decision to go to the movies at 6:00 PM tomorrow, then that is what I will do then. But if at that time tomorrow, despite my decision, I am sitting motionless nowhere near a movie theater, then that will be the way I settle the question of what to do at 6:00 PM tomorrow. Setting a practical ques-tion of what to do in the here and now is acting. Whatever an agent does at the time is what settles the question.


A philosophical interest in practical reasoning is not simply an interest in how agents come to form intentions and decisions and act on them when the time comes. It appears to be more a concern with the role of something called “reason” or reasoning in settling prac-tical questions. But does this mean “settle” in the sense of arriving at a conclusion about what to do, or in the sense of deciding or intending to do something, or in the sense of actually doing something? Perhaps the issue is whether practical questions can be settled in any of those ways by “reason” alone. Is “reason” on its own ever practical?

That seems to be the question Hume answers negatively when he says that “reason alone can never be a motive to any action of the will.” 1 He thinks “the greatest part of moral philosophy, ancient and modern,” has been “founded” on the idea of a “combat” of “reason” with “passion,” and on the idea that “every rational creature . . . is oblig’d to regulate his actions by reason” and to oppose any other “motive or principle” until it has been “brought to a conformity with that superior principle.” 2 Hume means to reject that whole tradition of moral philosophy. He thinks reason cannot be a motive to any action, so it cannot even possibly oppose passion in what he calls “the direction of the will.” 3 That is why, in his famous words, “we speak not strictly and philosophically when we talk of the combat of passion and of reason. Reason is, and ought only to be the slave of the passions, and can never pretend to any other offi ce than to serve and obey them.” 4

As Hume thinks of “reason,” it seems right to say that it alone could never produce action. What settles or resolves practical ques-tions in the here and now, at the time of acting, is the agent’s actually doing something. No action would occur at all if the putative agent were indifferent to all potential alternative courses of action. Hume thinks “reason” or “the understanding” is “cool,” “indifferent,” unengaged, and in itself “begets no desire or aversion.” 5 It can dis-cover relations among ideas and—in the expansive way Hume allows himself to think of it in his moral philosophy—it can discover what is so and what are the causes and effects of events. But “it can never

1 D. Hume, A Treatise of Human Nature (ed. L. A. Selby-Bigge), Oxford, 1978, p. 413 . 2 Treatise , p. 413. 3 Treatise , p. 413. 4 Treatise , p. 415. 5 D. Hume, Enquiries Concerning the Human Understanding and Concerning the

Principles of Morals (ed. L. A. Selby-Bigge), Oxford, 1975, p. 172 .

Practical Reasoning 209

in the least concern us to know, that such objects are causes, and such others effects, if both the causes and effects be indifferent to us.” 6 If the operations of reason alone leave us always indifferent as to what we discover is so, then no operations of reason could on their own produce action. Reason alone cannot settle a practical question in the here and now.

Given Hume’s understanding of the two separate faculties of “rea-son” and “passion,” it seems equally true that “passion” alone can-not produce action either. At least not intentional action. If bodily movement occurs, something must have been suffi cient to cause it. But for Hume what is at stake is not just movement but what he calls “the direction of the will.” Preference, desire, or inclination is what leads to, and so is present in, intentional action: what is done is done as leading to the satisfaction of a certain preference or desire. And that requires that the agent have certain thoughts, in the sense that certain intentional descriptions are true of him. It must be true of an agent in an intentional sense that he is reaching out for a drink, for example, or choosing thirst-quenching over motionlessness, in order for his “will” to be involved in his action of taking a drink, and so for his desire for a drink to be relevant to the “direction” of his will. And that is a matter of how he regards what he is doing, what he sees or believes or otherwise takes to be so, at least in the sense of having certain intentional attitudes. And that involves “reason” in Hume’s expansive sense: the “discovery of truth or falshood,” 7 or the taking of certain things to be true.

If we take seriously the idea of reason and passion as two distinct faculties or departments of the mind, then, it seems that neither alone can cause intentional action; both are needed. If reason’s practical insuffi ciency on its own were enough to render it a “slave” to passion, passion’s practical insuffi ciency on its own would be enough to render it a slave to reason. It is probably best to drop the slavery metaphor altogether.

In any case, all this has to do with acting in the here and now, with what settles a practical question at the time of action. Hume makes action sound like an outcome of certain mental dynamics: a combin-ation of psychic forces issuing in an action. From his observations on the impotence of reason as a force in generating action, he draws

6 Treatise , p. 414. 7 Treatise , p. 458.


the conclusion that an action cannot strictly speaking be “contrary to reason.” 8 Much of modern moral philosophy has been based in one way or another on this Humean conclusion. I think the conclu-sion as it is usually understood does not follow.

Hume’s claim that “reason alone can never be a motive to any action of the will” is a thesis about the impotence of reason in the production of action at the time of action. So when he puts the point by saying that it is “impossible, that reason and passion can ever . . . dis-pute for the government of the will and actions,” 9 it must be the execu-tive branch of government that he has in mind. To infer that therefore the will and its actions cannot be contrary to reason would be like inferring that since only the executive and not the legislative branch of government can carry out policy, what the executive branch does cannot be contrary to what the legislature does. It is true that a legis-lature cannot oppose or prevent executive action by executing an action in the opposite direction, as it were. It cannot “dispute” or be in “combat” with the executive in that sense. But what the executive does can nonetheless confl ict with, or fail to accord with, what the legislature legislates. Its actions can be contrary to the legislature’s deliverances. And a person’s action can be contrary to the deliver-ances of reason if it is contrary to what can be discovered by reason-ing to be the thing, or the best thing, for the person to do.

This is true of reasoning in its role of settling practical questions in advance. If I come to a fi rm conclusion about what there is most reason, or what is the best thing, for me to do, my reasoning to that conclusion does not alone determine whether the action actually takes place or not. In that sense Hume was right. Not even my rea-soning combined with my deciding or forming the intention to carry it out is enough alone to produce the action. Of course, we engage in the reasoning for the sake of action, and most of us aspire to be the sort of person who carries out what he has decided to do. Some people are better at that than others—or at least quicker. If the con-nection broke down in general, we would have no grounds for ascrib-ing decisions or intentions in advance to agents. But that we have that aspiration, and that people in general live up to it, is not itself a prod-uct of reason or reasoning. If the deliverances of a legislature were never carried out, we would have no grounds for ascribing to it any

8 Treatise , p. 416. 9 Treatise , p. 416 (my italics).


legislative function; it would just be going through the motions. But the fact that its decisions are carried out is not a product of its legisla-tive activity alone.

I do not want to press the analogy. I do not suggest that in thinking or reasoning in advance about what to do we are passing laws for ourselves, let alone for other people. The point is only a reminder of the difference between drawing a conclusion about what to do, or the best thing to do, and carrying out one’s conclusions in action.

The fact that reasoning to a practical conclusion does not neces-sarily settle the executive question of action is revealed in another way by the fact that I can reach a practical conclusion about what is the best thing for someone else to do in a certain situation, just as others can about me. Drawing such a conclusion obviously does not lead to the action in question. Even if I continue to believe my con-clusion about what someone else should do right up until the time for action, it does not lead to a decision or intention or action on my part. My thought and reasoning about other people’s actions, and theirs about mine, is not deliberation . One can deliberate only about one’s own actions, and only about one’s future actions at that.

That is perhaps true of what is properly called deliberation. But we can and do engage in thought directed toward what to do, where what is in question are the actions of someone other than the person engaged in that thinking and reasoning. We do call in others to aid us in deliberating on important questions, as Aristotle pointed out. 10 There would be little point in my doing that if others could not ask and express views about what is the thing, or the best thing, for me to do. The fact that they are not themselves deliberating does not mean that they are not trying to answer the same practical question I am trying to answer.

Of course, if they are helping me, their way of putting the question is not the same as mine. My question contains the indexical expres-sions “I” or “me”; theirs has “you” or “he.” But in the reasoning involved in answering the question, the different indexicals are in a way inessential. All of us take into account whatever we know about me and the circumstances and the alternative actions we contemplate. To draw the best practical conclusion, we have to be sure not only that we are getting things right, but that we are getting all the relevant facts. That poses a different kind of problem for me than it does for

10 The Nicomachean Ethics of Aristotle , (tr. W. D. Ross), Oxford, 1954, III, 3 (p. 56).


others thinking about me, especially with regard to what I want, what I would like, and what I value. But given all the facts we can discover, the question for all of us is the same: what is the thing for me to do? The fact that only I will act does not mean that in the thought and reasoning stage of the proceedings I am asking a question different from the one my advisors are asking. And for all of us, it is a practical question.

This might seem to be contradicted by Bernard Williams, who says in Ethics and the Limits of Philosophy that “practical thought is rad-ically fi rst-personal. It must ask and answer the question ‘what shall I do?’ ” 11 If this means that others cannot ask and answer questions about what is the thing, or the best thing, for me to do, it doesn’t seem right. They can. If it means only that the questions others can ask and answer about me are not practical questions, we have seen one way in which that is right. They are not practical questions for them: other people are not going to have to act on the answer to the question what I should do. The reasoning others engage in about me is not for them deliberation. Williams also puts his point that way: “Practical deliberation is in every case fi rst-personal, and the fi rst person is not derivative or naturally replaced by anyone .” 12

I think it is right that the fi rst-person pronoun in practical reason-ing is not simply replaceable by the variable “anyone.” But I do not think that means that practical thought is essentially fi rst-personal in the sense of requiring the indexical “I.” Some name or personal pro-noun or other singular term is essential in the formulation of the practical conclusion; it must indicate for whom the thing to do is such and such. But it does not have to contain a pronoun, even a fi rst-person pronoun. It is also true that in order to put a practical conclu-sion into effect, an agent must have a belief expressed in fi rst-personal terms. He must believe that he is the one for whom the thing to do is such and such. But that is not required—although it might well be present—in the reasoning involved in answering the practical ques-tion in advance of action.

I can answer a question about what someone with such-and-such beliefs, desires, values, and goals should do in a certain situation, and then fi nd that the general description is in fact true of me. My practical

11 Bernard Williams, Ethics and the Limits of Philosophy , Cambridge, Mass., 1985, p. 21 .

12 Williams, Ethics and the Limits of Philosophy , p. 68.


thought or reasoning before I make that discovery was not fi rst- personal, but now I am in a position to draw a fi rst-personal conclu-sion from it, by instantiation, as it were. I might conclude that the thing for me to do is just what I said such a person should do. I might draw that conclusion, but even without repudiating my earlier reason-ing, I might not. To put it in the unfortunate phrase that can so bedevil this subject, “reason” does not “force” me to draw that conclusion about myself.

I may not draw it, although I fi nd nothing wrong with the answer I gave to the completely general and impersonal practical question, because what seemed all right to say is the best thing for someone of a certain description to do does not look so good when it is applied to me, even though I fi t that description. That need not be simply the arrogant insistence that I am special or above it all or not subject to the reasons that apply to everyone else. It can be the result of my knowing or believing or feeling something else about myself that was not mentioned in the original question or reasoning. Although the original description is true of me, it is not everything that is true of me. And the additional considerations can count against the reason-ableness of detaching the practical conclusion of that earlier reason-ing and applying it directly to me.

This is exactly parallel to “factual” or “theoretical” thought or rea-soning about what to believe, or what is so. I can answer an imper-sonal and completely general question about what a person who has such-and-such beliefs, and is confronted with such-and-such evi-dence, should believe, or what such a person has most reason to believe, and then discover that that description in fact applies to me. If my answer was that the person so described should believe p , I might then draw that conclusion about myself, and accordingly come to believe p . If I do, I transmit my acceptance of the original beliefs and evidence through to the recommended proposition p , and so accept it too. But I need not do that, even if I continue to accept those original beliefs and evidence. I might know or believe other things that, when put together with those original beliefs and evi-dence, make it unreasonable for me to believe p , and more reasonable to believe something else. In that case, I will not conclude by direct instantiation that I should now believe p . What I fi nd it right to say about what a person should believe, given that he has the original beliefs and evidence, does not look so good when applied to me, even though I still accept those beliefs and evidence. That is not all that is


true of me. So I still face the epistemic question of what is best, or is most reasonable, for me to believe.

The question of what I should now believe is therefore not the same as the general impersonal question of what anyone who has such-and-such beliefs and evidence should believe, even if I have those beliefs and evidence. To answer the question about myself, I will need some general knowledge of what sorts of considerations are good reasons to believe certain things, but I cannot take the answer to any such question and automatically apply the answer to me as I am. The same holds for practical questions of what is the thing for me to do. I need some idea of what anyone who has such-and-such beliefs, desires, values, and goals should do in such-and-such circumstances, but I cannot proceed immediately and uncritically from an answer to such a question to a conclusion about what I should do, even if I have those beliefs, desires, and values. In this respect, questions of what to do and of what to believe are parallel.

So I think Williams is right that the question of what I should do is not simply the question of what anyone of such-and-such a descrip-tion should do. But that is not because the question is essentially fi rst-personal. There is nothing essentially fi rst-personal in the recog-nition that a general description true of me might not be all that is true of me. Others who know additional considerations in my case could see the unreasonableness of drawing the practical conclusion about me that it would otherwise be most reasonable to draw from the general description alone. Williams grants that the question of what I should believe is not essentially fi rst-personal. But that is not simply the question of what anyone of such-and-such a description should believe either. It is a question of what the particular, specifi ed person should believe in the actual situation that obtains. The prac-tical question similarly asks what the specifi ed person should do, as things actually are. In each case it is a question that anyone can ask and try to answer.

The fact that no general, impersonal answer is directly applicable to an actual case is a consequence of an important general fact about reasons or reasoning. If P is good reason to do or believe something, it does not follow that the conjunction of P and anything else consist-ent with it is also good reason to do or to believe that thing. Even though certain considerations make it reasonable to do or believe something, the addition of further considerations can make it less reasonable or even unreasonable to do or to believe it. In that sense,


both the practical and the epistemic questions are always open-ended, and the fi nal verdict is always less than logically inescapable, even given extensive general knowledge of what considerations are good reasons for acting or believing in certain ways.

My question of what I should do is open-ended in that way, and so are other people’s questions about what I should do in that situation. In giving an answer about ourselves or about some other person, we say more than just what is best for anyone of a certain description to do or to believe. We conclude “ a should do F” or “the best thing for a to do is F,” or we use the fi rst-person pronoun to refer to a : “I should do F.” In making that assertion we detach a categorical or unconditional conclusion about a particular person from reasoning involving facts about that person and the situation. But in detaching and stating that conclusion, we no longer mention those facts from which it is derived. If we did, and so stated only a relation between the accepted facts and a certain practical or epistemic conclusion, we would still be one step short of drawing a practical or epistemic con-clusion. The question of what that particular person in that particu-lar situation should do or believe would not have been settled. At most we would have answered a general and impersonal question of what someone of such-and-such a description should do or believe in a situation of a certain kind. Detaching a non-conditional practical or epistemic conclusion is needed to conclude the reasoning about what an actual person should do or believe, as things are.

In the epistemic case, not only might a set of beliefs and evidence support the conclusion p , while additional facts about the believer make it unreasonable for him to believe p , but those additional facts can even make it unreasonable for him to continue to accept all those original beliefs and evidence. This holds even in the rare case in which the original beliefs logically imply the conclusion p . Even then, reason does not force me to believe p , even if I believe the premises and I know that p follows from them. I might see right before my eyes that p is false, or at least highly doubtful. So I should abandon or re-examine my acceptance of those original beliefs. This is equally obvious in the more typical case in which my earlier beliefs do not logically imply the conclusion p . Even if they only strongly support it, I can still fi nd that additional considerations give me more reason to reject some of those earlier beliefs than I have to accept that conclusion.

This shows, I think, that no fi xed general description of a person of a certain kind can capture the full question of what a particular


person should (or has most reason to) believe or do. The question of what I should believe in the situation in which I fi nd myself is not the same as the question of what anyone who fi ts a certain general description should believe, even if I fi t that description and know that I do. In asking the question about myself I am in effect asking whether that description is all that is relevantly true of me and my situation or, if it is, whether it should continue to be true of me. Even if I grant that I now have the beliefs in question, the question is whether I should continue to believe everything I now believe. The question of what to believe is in that sense always open, whatever I believe at the moment. I do not mean that it cannot be answered, but only that the answer is never simply dictated by what I already believe. To detach the best conclusion, I have to be ready to subject even my earlier beliefs to further scrutiny or criticism.

The same is true of the practical question of what to do. Even if I have such-and-such beliefs, desires, goals, and values, and I know what it would be best for anyone who has those particular attitudes to do, I cannot conclude directly that that is what it would be best for me to do. There might be other things true of me that make it more reasonable for me to abandon or re-examine some of those beliefs or values or goals than it is to accept that practical conclusion as applied to me. Here too the question I face is in effect whether that original description is all that is relevantly true of me and my situation or, if it is, whether it should continue to be true of me. Should I continue to want, value, and aspire to all the things I now do? As in the epi-stemic case, I have to be ready to re-examine and re-evaluate all the considerations and circumstances from which the best practical con-clusion is to be drawn. Nothing can be accepted as fi xed and beyond scrutiny or criticism.

The epistemic question of what to believe presents itself to the believer himself as the question of what is so, or is most likely to be so. The reasons that I canvass for and against my believing p are other things that I believe, such as q and r and s , not the fact that I believe q or believe r or believe s . If I take the fact that it has always rained in Berkeley in January as good reason to believe that it will rain there next January, it is the past rain that I take as the reason for expecting rain, not the fact that I believe it has always rained there in the past. My believing that it has always rained does not amount to much of a reason for expecting rain in Berkeley in January. In trying to determine what to believe, or what there is most reason to believe,


my focus is on the world, or on what is so, not on my attitudes toward what is so. I have many beliefs about how things are, and I want to fi nd out something. And the question is settled for me, for the time being, by my coming to believe something. But my coming to believe what I do is to be based on what I already believe or can fi nd out, not on my believing it or fi nding it out.

Another person who asks what I have most reason to believe about a particular matter must ascertain what I already believe about it, so to that extent he will attend to or take notice of the fact that I believe certain things. But again what he is interested in is the relation between what I now believe and what it is best for me to believe. He asks what those things that I believe do in fact establish or give me most reason to believe, not what my believing them establishes or gives me most reason to believe. If he wants to explain why I believe something that I already believe, he will perhaps cite my holding cer-tain attitudes. But explaining why someone believes what he does is not the same as determining what he has most reason to believe.

The question of what it is best for a particular agent to do is paral-lel in this respect. It presents itself to the agent himself as the ques-tion of what to do, or what is best. The considerations he canvasses for and against his doing a certain thing are what he expects, wants, and values, not the fact that he expects or wants or values those things. If I take the pleasure of the sweet taste I expect to get from eating a chocolate as a reason for me to eat it right now, the expected pleasure of the sweet taste is what I take to be the reason for me to eat it, not the fact that I expect or want the pleasant sweet taste. In that sense, I consider or attend to what I want and expect, not to my own attitudes or to the fact that I have them. Of course, it is only because I have the attitudes I do that the practical problem takes the particular form for me that it does. If I didn’t want the pleasure of a sweet taste, the fact that such a taste is to be expected from eating the chocolate would not for me be a good reason to eat it right now. But that does not mean that what I take to be a reason to eat it is my wanting the pleasure of a sweet taste. My having that want is what is expressed or exhibited in my taking the expected sweet taste as a rea-son to eat the chocolate. It is not what I take to be a reason to eat the chocolate.

Another person who asks what I have most reason to do must ascertain what I want or value, so to that extent he will attend to the fact that I want or value certain things. But again what he is interested


in is the relation between what I want and value and the properties of certain proposed actions. His assessment rests on the relation that he thinks holds between getting a sweet taste (which is something I want) and the action of eating the chocolate, not on the relation between my wanting to get a sweet taste and the action of eating the choc-olate. This latter relation is what he might appeal to in order to explain my action of eating the chocolate, if I do. Perhaps I ate it because I expected it would give me a sweet taste, and I wanted the pleasure of a sweet taste. Here, my holding the attitudes I do is essen-tial to the explanation. But explaining an action that has occurred is not the same as determining what an agent faced with a practical question has most reason to do.

Another person’s assessment of my position might differ from mine, so we come to different conclusions about what I should believe or do. This can happen in different ways. In the epistemic case, we might differ about what conclusion it is most reasonable to draw from what we agree are all the relevant considerations. Or we might differ in what relevant considerations we take into account. Someone who believes that a weather pattern approaching the Pacifi c coast in late December will bring more than a month of drought will con-clude that there is not good reason to expect rain in Berkeley next January even though it has always rained there then in the past. If I have heard nothing of that weather pattern, I will continue to believe that it is going to rain then. He will perhaps grant that, given only what I believe, I have most reason to believe that it will rain, but he will think I should take into account the available information about the drought. He will think that to believe that it will rain would be contrary to what there is most reason to believe. If I have heard the story of the drought but think there is no good reason to believe it, I will disagree. There is disagreement between us, not only about the weather in Berkeley next January but also about whether the past rain we agree about is in fact good reason to expect rain next January. This looks like disagreement about what is so. If it matters, it is some-thing we can discuss and try to settle. The same dialectical process can go on in me. If I learn of an impending drought, and so realize that there is not in fact most reason for me to expect rain, I will see that I had been wrong earlier, although I did not think so at the time. Within me, or between me and another, there can be disagreement, and then agreement, about the answer to the same question: what is there most reason for me to believe?


The practical case is parallel. You and I might agree about all the relevant considerations yet differ in what we think it is most reason-able to conclude from them. More typically, we might differ about what they are and so draw different conclusions. Someone who believes that I am allergic to chocolate will think there is not most reason for me to eat the chocolate even though it is true that I will get a sweet taste from it, which is something I want. He will think that my eating the chocolate would be contrary to what there is most rea-son for me to do. Someone who has heard nothing of my allergy will disagree. This again looks like disagreement as to what is so. Different judges give confl icting answers to the question “What is there most reason for me to do?”

Given that not even the person’s original beliefs, desires, and val-ues can be taken as fi xed and beyond question in trying to answer an epistemic or practical question, different people’s assessments of what the person should believe or do can differ in more radical and apparently more intractable ways. One person can fi nd that, given the person’s beliefs, it is reasonable for him to believe p , but that there is more reason for him to reject some of those beliefs and so not to believe p at all. Or that, given the person’s present wants and values, it is reasonable for him to do so-and-so, but that there is more reason for him not to want or value all the things he now does, and so not to do so-and-so. What would be best is for him to change some of his already-accepted beliefs or evaluations. This conclusion is something with which the agent in question might disagree. If it matters, it is something they can discuss and try to settle. They might come to agree, or they might not.

Confl icting verdicts about what to believe or to do are not restricted to interpersonal disagreement. A person can reach a verdict that con-fl icts with an earlier conclusion he drew about himself. He will then think that if he had believed or acted in accord with that earlier con-clusion it would have been contrary to what is most reasonable for him to believe or to do. That is just what another person thinks of a belief or action that goes against the conclusion he drew. Both are attempting to answer the same question: what should that person believe or do?

Bernard Williams resists the idea that what he calls “practical deliberation” and “factual deliberation” are symmetrical in these ways. Although in each case I can stand back from my current beliefs, experiences, desires, and values and refl ect on them with an eye to


what to do or what to believe, he thinks what he calls factual deliber-ation is “impersonal” or only “derivatively” fi rst-personal, since it seeks the truth. He says the question “What should I think about this question?” could as well be “What should anyone think about this question? . . . Refl ective deliberation about the truth indeed brings in a standpoint that is impartial and seeks harmony, but this is because it seeks the truth, not because it is refl ective deliberation.” 13 But he thinks that, in practical deliberation, I do not completely detach myself from my beliefs and desires, and I am not required to take the result of anyone else’s properly conducted deliberation into account.

The action I decide on will be mine, and . . . its being mine means not just that it will be arrived at by this deliberation, but that it will involve changes in the world of which I shall be empirically the cause, and of which these desires and this deliberation itself will be, in some part, the cause. It is true that I can stand back from my desires and refl ect on them. . . . [But] The I that stands back in rational refl ection from my desires is still the I that has those desires and will, empirically and concretely, act. 14

It is true that the person who asks “What should I do about this?” is the person who will act and so will settle the question of what to do. But it is equally true that the person who asks “What should I believe about this?” is the person who, if he settles the question, will believe something. The question is not simply “What should anyone who believes, desires, and values what I do believe or do?” The refl ective question in each case is what that particular person should now believe or do, where the answer might involve rejecting or ignoring or changing some of the beliefs, desires, or values that he already has. To that extent, the agent must stand back from all his current beliefs, desires, and values with an eye to altering them in the right or best way. He alone is in a position to do that precisely because they are his beliefs and attitudes. But the epistemic or practical refl ection is directed always toward what the particular person should believe or do as things actually are. It presents itself to every inquirer, including the agent, as a question of what is so.

Aristotle held that deliberation is investigation or search—trying to fi nd out something—even if not all investigation or search is delib-eration. 15 The fact that others can ask and try to answer the same

13 Williams, Ethics and the Limits of Philosophy , p. 69. 14 Williams, Ethics and the Limits of Philosophy , pp. 68–69. 15 Aristotle, p. 57.


question that I ask about the best thing for me to do suggests that there is something to be found out, something for all of us to be right or wrong about, and that we are all trying to fi nd out what it is. Even if others who try to determine what I should do are not deliberating, they could be said to be investigating or searching, just as I am. In this important respect, there is no difference between so-called prac-tical and so-called theoretical reasoning.

That there is something to be found out is perhaps further sup-ported by the fact that I can refuse or be reluctant to go along with a practical conclusion I have reached, even though I have found it to be the best thing, or the thing there is most reason, for me to do. Refusal or inability to do what one acknowledges one should do is often regarded in discussions of practical reason as unreasonableness, or irrationality, or “weakness of will.” But one form such a refusal or reluctance can take can be a very good thing in an agent. It can be a response to feeling that the recommended action is somehow just not right, or not the thing for me, or maybe even simply out of the ques-tion for me, even though I cannot quite put my fi nger on why. If I could, it would be by fi nding some defect in what led me to that practical conclusion, or some additional considerations that point in a different direction. Identifying those additional factors would en able me to articulate my reluctance or to defend my new conclusion. But even without fi nding identifi able weaknesses in his earlier reasoning, or specifi able considerations in favor of a new conclusion, a sensitive, careful agent will continue to take seriously whatever reluctance he feels in the face of a practical conclusion he has reached. A more refl ective agent will try to get to the bottom of it. Being capable of both reactions can be a virtue, just as a good nose or a good feel for the case can be an epistemic virtue in other kinds of investigation. An experienced detective or experimental scientist or other investigator learns to respect and pursue the feeling that some-thing is just not right, even though all the reasons he can specify point unmistakably in a certain direction. He feels that there is defi nitely something more to be found out, even though he knows that so far he has not found it. The same seems true in deliberation and action.

Many people resist the idea that one person can answer the ques-tion of what another person should do. They fi nd it offensive. They think nobody can tell anyone else what to do; it threatens or denies a person’s autonomy. But it is just not true that nobody can tell anyone else what to do. They can, and at times they should. It is often just


what is called for. Furthermore, someone who reaches a fi rm conclu-sion about what I should do in a particular situation is not thereby telling me what to do. For one thing, he needn’t tell me anything. And even if he tells me what conclusion he has reached, he is not thereby telling me what to do in any way that threatens my autonomy. He is saying what I should do, or what is the thing, or the best thing, for me to do. Even if he is right about that, my autonomy is not thereby reduced.

I can reach a verdict myself about what to do and then act on it. If that counts as autonomy, then my not acting on it and doing some-thing else instead can be acting autonomously too. So the fact that someone else reaches a verdict about what I should do does not reduce my autonomy either, even if I know about it. I might act on it and I might not. I do not mean that my autonomy cannot be reduced or threatened in any way. Of course it can, in many familiar ways, and by people with power over me who tell me what to do. But my autonomy is not threatened by other people reaching practical con-clusions about what I should do. What we want in the way of auton-omy, if we want it, is autonomy of action, including the actions of thinking and reasoning. The idea that our actions or thoughts accord with what we take ourselves to have the most reason to do or think is in itself no threat to autonomy. So it is no threat to autonomy if our actions accord with what someone else takes us to have most reason to do or think. Reason or reasoning, whoever engages in it, is not a force, and so not a force that restricts or constrains action. 16

16 I am grateful to Samuel Scheffl er for helpful discussions of earlier drafts of this essay.

11

The Charm of Naturalism

Presidential Address delivered before the Seventieth Annual Pacifi c Division Meeting

of The American Philosophical Association in Seattle, Washington, April 5, 1996

I want to make some very general observations on what many see and applaud as a broadly “naturalistic” turn in recent philosophy. There seems little doubt at fi rst glance that there is such a thing, at least judging from what many now call what they are doing. Something known as “epistemology naturalized” has been with us for some time. Or at least a recommendation to that effect was made some time ago. 1 More recently we have been encouraged in such enterprises as “nat-uralized semantics,” “naturalizing belief,” and, even more generally, “intentionality naturalized.” And now there is the even more general project (why not go all the way?) of “naturalizing the mind” (the title of a delightful recent book) 2 . I have even seen something called “nat-uralizing responsibility.” And there are no doubt many other efforts at “naturalization.”

Is there more to all this than just a trendy label? What, if anything, is behind it? Is it something distinctive, and new? And if so, is it a good thing?

These questions are not easy to answer. The idea of “nature,” or “natural” objects or relations, or modes of investigation that are “naturalistic,” has been applied more widely, at more different times and places, and for more different purposes, than probably any other notion in the whole history of human thought. The earl-iest turn towards naturalism that I have heard of was in the fi fth

1 See W. V. Quine, “Epistemology Naturalized,” in his Ontological Relativity and Other Essays , Columbia University Press, New York, 1968 .

2 Fred Dretske, Naturalizing the Mind , MIT Press, Cambridge, Mass., 1996 .


century BC And they seem to have been happening every so often ever since.

When we look at this most recent enthusiasm for what its propon-ents call “naturalism,” I think we fi nd that, whatever they are excited and optimistic about, it is not naturalism as such. With two excep-tions that I will mention in a moment, I think there is nothing in naturalism alone that is suffi ciently substantive to be philosophically controversial. What is usually at issue is not whether to be “natural-istic” or not, but rather what is and what is not to be included in one’s conception of “nature.” That is the real question, and that is what leads to deep disagreements. And as far as I can see, those disagree-ments are not themselves to be settled by what can be recognized as straightforwardly “naturalistic” means. So one thing that seems not to have been “naturalized” is naturalism itself. If it were, the resulting naturalistic view of the world might be impressively comprehensive and illuminating, and superior to views of other kinds, but if it had those virtues it would have them on its own merits, not simply because it is an instance of something called “naturalism.”

“Naturalism” seems to me in this and other respects rather like “World Peace.” Almost everyone swears allegiance to it, and is will-ing to march under its banner. But disputes can still break out about what it is appropriate or acceptable to do in the name of that slogan. And like world peace, once you start specifying concretely exactly what it involves and how to achieve it, it becomes increasingly diffi -cult to reach and to sustain a consistent and exclusive “naturalism.” There is pressure on the one hand to include more and more within your conception of “nature,” so it loses its defi niteness and restrict-iveness. Or, if the conception is kept fi xed and restrictive, there is pressure on the other hand to distort or even to deny the very phe-nomena that a naturalistic study—and especially a naturalistic study of human beings—is supposed to explain. The source of these two confl icting movements of thought is what I want to illustrate.

But the fi rst thing to do with naturalism, as with any philosophical doctrine or “ism,” is to ask what it is against . What does the so-called “naturalistic turn” turn away from, or deny?

Here we have to distinguish two aspects of naturalism. There is naturalism as a view of what is so, or the way things are, or what there is in the world. And there is naturalism as a way of studying or investigating what is so in the world. A naturalistic study of human beings would study and understand them in relation to the rest of

The Charm of Naturalism 225

nature. Obviously, what you think the natural world is like will have an effect on how you investigate the things in it, and what you think is the best way to understand them. The two aspects of naturalism are connected.

Under the fi rst aspect, as a doctrine about what is so, or what there is, naturalism says that there is nothing, or that nothing is so, except what holds in nature, in the natural world. That is not very inform-ative so far, but even without specifying it more precisely it already seems to exclude some things that many people have apparently believed in. Naturalism on any reading is opposed to supernatural-ism. Here we have what looks like a substantive issue, or at any rate something controversial. Not everyone regards exclusive naturalism as beyond question or as an unqualifi ed good thing.

This is the fi rst of the two exceptions I mentioned. By “supernat-uralism” I mean the invocation of an agent or force which somehow stands outside the familiar natural world and so whose doings cannot be understood as part of it. Most metaphysical systems of the past included some such agent. A naturalistic conception of the world would be opposed to all of them.

Supernaturalism as a doctrine about what is so can have conse-quences for the study of human beings—in particular, how they believe and come to know things. In epistemology there have been many supernaturalists. Descartes thought that human knowledge cannot be accounted for without a benevolent, omniscient, and omnipotent God who guarantees the truth of what human beings clearly and distinctly perceive to be true. For Berkeley, God’s agency is the only active force there is in the world of things we perceive and know about. Without him there would be nothing for us to know. Even Locke relied on a benevolent agent as the ultimate source of those cognitive faculties which are all that human beings need to get along in the world they fi nd themselves in. These are not fully nat-uralistic accounts of human knowledge. They appeal to something beyond the natural world. In going against this supernatural consen-sus, Hume is almost alone among the greats. His credentials as a fully naturalized—or at least as a non-supernaturalized—metaphysician and epistemologist are impeccable. The same is probably true of John Stuart Mill, if he counts as one of the greats. But there have not been many.

In the sense in which naturalism is opposed to supernatural-ism, there has been no recent naturalistic turn in philosophy. Most


philosophers for at least one hundred years have been naturalists in the non-supernaturalist sense. They have taken it for granted that any satisfactory account of how human belief and knowledge in general are possible will involve only processes and events of the intelligible natural world, without the intervention or reassurance of any super-natural agent. Many people regard that as on the whole a good thing. But it is nothing new.

In fact, the long-standing naturalistic consensus is being chal-lenged more directly now, when the virtues of naturalism are being so loudly proclaimed, than it was during the long period when they went more or less without saying. Alvin Plantinga, for example, argues that no satisfactory general explanation of human knowledge can be given on a naturalistic basis. He thinks that justifi cation or warrant, which is essential to knowledge, can be understood only in terms of the proper function of human cognitive capacities. And that in turn, he thinks, requires a divine designer of those capacities. Successful epistemology therefore “requires supernaturalism,” 3 in particular, what he calls “theism.” 4 He is apparently not alone in that belief. If Plantinga and his friends convince others, there will be a general turn away from naturalism. That shows that it is naturalism that is now old hat. It is not something towards which there has been a recent, glorious turning.

Even supernaturalists like Plantinga and Descartes, Locke, Berkeley, and others would still count as “naturalized epistemolo-gists” in at least one current sense that has been given to that phrase. Epistemology has been said to be “naturalized or naturalistic” as long as it tries to explain only how human beings do in fact arrive at their beliefs rather than how they ought to arrive at them. 5 If that is enough to make an epistemology naturalistic, then virtually every philosopher in history has been a naturalized epistemologist. They have all been concerned to describe and understand the human con-dition as it is, to see and to explain how we actually get all the know-ledge we’ve obviously got. If God plays a role in human beings’ coming to know things, that will be part of the answer to the purely “descriptive” question of how human beings in fact arrive at their

3 A. Plantinga, Warrant and Proper Function , Oxford University Press, New York, 1993, p. 46 .

4 Ibid. p. 237. 5 For this description of naturalized epistemology, see the Introduction to H.

Kornblith (ed.), Naturalizing Epistemology , MIT Press, Cambridge, Mass., 1985 .


knowledge. Even supernaturalism as a view of what is so is not incompatible with naturalized epistemology in this curiously weak, so-called purely “descriptive” sense.

This shows that the fi rst aspect of naturalism dominates over the sec-ond. If you do not start out with any restrictions at all on what the world you are studying contains, studying things only as part of the natural world does not amount to anything very defi nite. Some determinate conception of what the natural world is like is needed to give substance to the claim that one’s epistemology, or one’s study of any other aspect of the world, is naturalistic.

The second exception to the idea that there is no real dispute about naturalism is perhaps best illustrated (at least in epistemology) by Quine, who after all, as far as I know, is the person who coined the phrase “epistemology naturalized.” He was responding to Carnap, with whom he had a real dispute. Carnap sought a reduction of all talk of external bodies to talk only about possible sense-experiences. It was intended as what came to be called a “rational reconstruction” of our science or knowledge. It would have shown how our concep-tion of the world could be supported solely by materials to be found in immediate sense-experience. But no satisfactory translation or reduction was found. The idea of “rational reconstruction” does not in itself require that statements about the external world must be translatable into statements about immediate sense-experiences. It requires only that it be shown how our beliefs about the world could be justifi ed by information that we could get through experience or observation. That general task of what might be called “hypothetical (or reconstructed) justifi cation” was pretty much the task of analytic epistemology through the middle fi fty years or so of the twentieth cen-tury. One form it took, and perhaps still takes, is confi rmation theory.

Quine’s so-called “naturalistic” turn was to say “why all this cre-ative reconstruction, all this make believe?” 6 Why ask how statements of the kinds human beings believe could be confi rmed by sense- experiences they could conceivably have? As Quine put it: “Better to discover how science is in fact developed and learned than to fabri-cate a fi ctitious structure to a similar effect.” 7 The question is how science is “developed and learned.” It is not just a question of the logical relations among the propositions human beings believe.

6 W. V. Quine, op. cit. , p. 75. 7 Quine, op. cit. , p. 78.


Something is at stake here between Quine and Carnap, but it is not the merits of naturalism. It is really a dispute about what phil-osophy is or ought to be doing. Quine obviously has no quarrel with the idea of reducing one domain of discourse to another, if you can do it. Carnap and the positivists obviously have no quarrel with the idea of natural scientifi c studies of human belief and knowledge, or even of institutions like science. Naturalism as a way of investigat-ing the world is thought by all to be nothing but a good thing. But for logical positivism no such studies could be part of philosophy. Philosophy could be only a priori . Its only subject matter could therefore be the “concepts,” or the logical relations among the “principles,” employed in the sciences. Its only task could be “anal-ysis.” It could not pronounce professionally on the actual acquisi-tion and development of science, but only on what it called its “logic.”

Quine’s rejection of the very notion of the a priori left him with no such constraints. Study human knowledge in the same way you would study anything else in nature, he says, and don’t worry much about what label you attach to what you are doing. That meant that a task continuous with what epistemologists had attempted in the past could now proceed scientifi cally. Epistemology would in that sense be part of natural science, and it would study the acquisition, transmis-sion, and growth of natural science. The idea is, in Quine’s words, “that knowledge, mind, and meaning are part of the same world that they have to do with, and that they are to be studied in the same empirical spirit that animates natural science. There is no place for a prior philosophy.” 8

That same empirical spirit is present in the study of the history of science, which could also be described as a form of naturalism in the investigation of human knowledge. The history of science has of course been with us almost as long as science has, but its fl ourishing in the 1960s was in part also a reaction against the abstractions of logical positivism. The positivists focussed on what is known, or on the form of what is known, rather than on the knowing of it, or on the processes of fi nding it out. They did not study science as a human enterprise that develops in different ways at different times as a result of different sorts of forces. That could not be part of philosophy for them. The growth of the history of science in the last thirty years has

8 Quine, op. cit. , p. 26.


changed and enormously enriched the picture. And it certainly has been a very good thing.

Quine himself at one time seemed not so sure. He thought the historically oriented work of people like Kuhn, Polanyi, and Hanson had “loosed a wave … of epistemological nihilism” (as he put it) and tended to “discredit the idea of observation,” to “belit-tle the role of evidence and to accentuate cultural relativism.” 9 These are curious complaints for a fully naturalized epistemologist to make. Scientifi c epistemology must be prepared to accept what-ever the empirical study of human beings actually reveals. If it turns out that human knowledge is acquired without there being a fi rm, fi xed line between so-called “observational” and “non-obser-vational” terms, or if what a philosophical “theory of evidence” calls “evidence” is never actually appealed to in the acceptance and rejection of scientifi c hypotheses, then so be it. That will have to be accepted as the way knowledge is in fact acquired. If cultural rela-tivism turns out to be the best way to account for what happens in human life, the committed naturalist has to accept cultural relativ-ism. (What he should do fi rst of all, of course, is try to fi gure out what the term ‘cultural relativism’ actually means. But that is another story.)

The point is that conclusions of naturalized epistemology can be drawn only from the study of what actually goes on with human beings. If it turns out that women’s knowledge differs in certain ways from men’s, for instance, or poor southern blacks’ knowledge from that of affl uent urban whites, that is something that a naturalized epistemologist should welcome, or at any rate should not resist. Studies in the sociology, economics, and politics of knowledge could also be called “naturalistic epistemology” too. The lively interest in such matters these days is certainly on the whole a good thing. Not because naturalism is a good thing, but because coming to see more and more differences among things in the world—if they are actually there—is almost always a good thing.

I want to draw attention to a confl ict or tension that I think is present in a commitment to naturalism. It arises most clearly when we move beyond questions about this or that culture or this or that institution within a culture to that more general level at which phil-osophers typically ask about apparently universal features of human

9 Quine, op. cit. , p. 87.


life. Now I mean naturalism in every area of philosophy, not just epistemology.

Naturalism as a view of what is so, or what the world is like, must be given some determinate and restricted content. That means that anything that human beings think about, believe in, care about, or value that lies outside that restricted conception cannot really be seen as part of the natural world in which they live. But since it cannot be denied that people do have the very thoughts, beliefs, values, and concerns in question, the contents of those attitudes will have to be understood and accounted for in terms of something less than their possible truth. What human beings think, feel, and care about must be fully expressible somehow within the restricted resources available in the naturalist’s world. And that can lead to distortion. If, to accom-modate psychological phenomena and their contents in all their complexity, the restrictions are lifted, naturalism to that extent loses its bite. This is the basic dilemma I want to bring out.

I can illustrate it by starting with an extreme naturalist view. I would say that it is a ridiculously extreme position, were it not for the fact that many philosophers I respect appear to hold it. It says that the natural world is exhausted by all the physical facts. That is all and only what the natural world amounts to on this view; there is nothing else in nature. First of all, this view is probably not itself reached by purely naturalistic means. It not only states all the physical facts, which presumably can be determined by broadly naturalistic means. But it goes on to say that those are all the facts there are—that they are the whole truth about the world. And that claim is more than the conjunction of all the physical facts. It excludes everything else from being true, as they alone do not. Is the exhaustiveness that is essential to physicalism something that is naturalistically or physicalistically arrived at? That is one question.

Second, a natural world conceived of only as the totality of all the physical facts obviously does not contain any psychological facts. There are no truths to the effect that someone believes, knows, feels, wants, prefers, or values anything. Of course, anyone who holds that the physical is all there is might hold that everything we think along those lines is really just physical facts in disguise. In any case, that would leave no psychological facts for a naturalistic theory of the world to explain. The study of human beings on such a restricted physicalist conception would be just a study of physical goings-on, including some that happen to go on in human organisms.


The case is extreme because it does not include very much for a study of human beings to explain. Without at least biological facts in your naturalistic conception of the world you will not have much to investigate that is distinctively or interestingly human. But if the physicalist conception is expanded to include biological facts as well, what exactly are such facts thought to add? Do biological facts include the “intentional” facts of human beings believing, knowing, feeling, wanting, preferring, and valuing certain things? Some would say not, since these are just “folk” ways of speaking. Organisms inhabiting the natural world are not be thought of as having any such attitudes, or as acting from them, on that view. That would mean that naturalism could never be faced with the problem of explaining how and why human beings come to believe and feel and want the things they do. There would be no such facts. Naturalism as to what is so would be so restrictive as to leave naturalism as a method of investi-gation with much less to do.

There is an embarrassing absurdity in this position which is revealed as soon as the naturalist refl ects and acknowledges that he believes his naturalistic theory of the world. If persons with attitudes like belief and knowledge are not really part of nature, he cannot consistently say that about himself. I mean he cannot say it and con-sistently regard what he is saying as true. In fact he cannot say any-thing and regard it as true, or think of himself as saying it, if he holds such a restricted naturalistic conception.

It looks as if any sensible naturalism will have to acknowledge that human beings do in fact have a complex set of attitudes, feelings, evaluations, institutions, and so on. If it is going to explain what is so, it will have to explain how and why human beings think and feel and act in all the ways they do. It will offer those explanations by appealing not only to the ways human beings are, but also to facts of the natural world surrounding and affecting those human beings. To explain why people believe that there are such things as rectangular tables, for example, or red apples, it will trace the connections between human beings who perceive things and a world that contains rectan-gular tables and red apples. It is because things are as they are in the natural world, and because humans are as they are, and interact as they do with their surroundings, that they get the beliefs they do and are on the whole right about the natural world.

Even this simple general picture leaves room for human attitudes directed towards objects or states of affairs which restrictive forms of


naturalism can fi nd no room for within their conception of the world. For example, many philosophers now hold that things as they are in the world of nature are not really colored. There are rectangular tables in the natural world, perhaps, and there are apples in the natural world, but no red apples (and no yellow or green ones, either). This view appears to be held largely on the grounds that colors are not part of “the causal order of the world” or do not fi gure essentially in any purely scientifi c account of what is so. Scientifi c naturalism accord-ingly excludes them.

But even on this view those false beliefs and illusory perceptions of the colors of things must themselves be acknowledged as part of nature. A naturalistic investigator must somehow make sense of them as the psychological phenomena they are. Since he holds that there is no such fact as an object’s being colored, he cannot specify the con-tents of those perceptions and beliefs in terms of any conditions that he believes actually hold in the world. If he could, that would amount to believing that there are colored things in the world after all. Scientifi c naturalism denies that. But still, the beliefs and perceptions with those particular contents must be accounted for.

An easy way around this diffi culty has suggested itself to many phil-osophers, at least in this case. They take the apparently more sensible scientifi c naturalist view that there really is no systematic error in our beliefs about the colors of things. The beliefs are not in general false, since there is something in the restricted naturalist’s world to give con-tent to them after all. Beliefs about the colors of objects, it is said, are really beliefs about certain dispositions which those objects have to pro-duce perceptions of certain kinds in certain kinds of perceivers in cer-tain kinds of circumstances. Objects in nature really do have those dispositions. So the beliefs are preserved as largely true. The color of an object depends on what kinds of perceptions it is disposed to produce.

A dispositionalist theory of this kind can succeed only if it can specify the contents of the perceptions of color which it says physical objects have dispositions to produce. They cannot be identifi ed as perceptions of an object’s having some quality that objects actually have in that restricted naturalist’s world. They cannot be identifi ed simply as perceptions of an object’s having a disposition to produce just these perceptions under certain circumstances. The question is: which perceptions? There must be some way of identifying the per-ceptions independently of the object’s disposition to produce them. So it looks as if they must be identifi ed only in terms of some so-called


“intrinsic” quality that they have. Not a quality that the perception is a perception of , but simply a quality of the perception itself.

I doubt that we can make the right kind of sense of perceptions of color in this way. So I doubt that any dispositional theory can give a correct account of the contents of our beliefs about the colors of things. The way we do it in real life, I believe, is to identify the con-tents of perceptions of color by means of the colors of the objects they are typically perceptions of. It is only because we can make intel-ligible non-dispositional ascriptions of colors to objects that we can acknowledge and identify perceptions as perceptions of this or that color. But if that is so, it requires our accepting the fact that objects in the world are colored, and that is what the restrictive naturalist who denies the reality or the objectivity of color cannot do.

None of this is something I can hope to establish here. The point is only to draw attention to what I see as a general problem of restrict-ive naturalism. Exclude colored objects in general from the world, and you are in danger of losing the capacity to recognize perceptions of and beliefs about the colors of things. Include colored objects, and the contents of those perceptions and beliefs no longer go beyond what is so in the natural world.

I have found in my experience that this tension is not widely felt or acknowledged. Most philosophers regard it as so obvious and uncontroversial that colors are not real, or are in some way only “subjective,” that they simply do not recognize what I think is the distortion or incoherence they are committed to. That is something I continue to ponder, and try to get to the bottom of. But a problem of this same form is at least sometimes recognized elsewhere. Two large areas of philosophy are problem areas precisely because some form of restrictive naturalism looks like the only possibility in those cases. I have in mind the areas of mathematics and morality, or evaluation generally.

Human beings have evaluative beliefs and attitudes; they regard some things as better than others, they think that a certain thing is the thing to do on a certain occasion, and so on. To understand and acknowledge the presence of these human attitudes in the world, the naturalist must understand their contents—what those human beings actually think or believe. Naturalism is widely understood to imply that no evaluative states of affairs or properties are part of the world of nature. On that assumption, either evaluative thoughts and beliefs take as their “objects” something that is not to be found in the


natural world at all, or their contents are equivalent to something that is true in that world, so they are not really evaluative.

One way to embrace the fi rst option would be to say with G. E. Moore that evaluative statements are assertions about a “non-natural” world, or that they ascribe “non-natural” properties to objects in the natural world. We might then wonder what that “non-natural” domain is like, and how it is related to what goes on before our eyes. And what-ever it is, we might wonder why we should ever take any interest in it. Values might then be other-worldly, and have nothing to do with us. If all that is just too mysterious, we could keep to this fi rst option by saying instead that evaluative attitudes do not have contents that are true or false at all. In evaluating something we are prescribing, recom-mending, approving, or encouraging something, but not ascribing any properties to it or saying anything true of it beyond the “natural” properties we think it has got.

This last idea, I believe, distorts our actual thought and practice. It cannot give the right kind of sense to the evaluative thoughts we have or the inferences we regard as valid when combining evaluative and non-evaluative propositions. Again, that is not something I am going to try to prove here. I simply draw attention to the source of the pres-sure towards some such emotivist or expressivist theory. It comes from a restricted naturalistic conception of what the world contains. Nature itself, it is said, is value-free. So evaluations cannot be strictly speaking either true or false. That is one alternative. It is not an inevit able consequence of a restricted naturalistic view of the world. The same restrictive view of nature is what leads a non-expressivist like Moore to the idea of values as “non-natural” or in some sense “other-worldly.”

Dissatisfaction with both non-naturalism and expressivism leads the restrictive naturalist to the second option, and so to some form of reductionism. Human attitudes that appear to be evaluative are to be seen as attitudes with contents which can and do hold in the restricted natural world after all. They can be true or false, but the conditions of their truth are purely natural and so non-evaluative.

If such a reduction is expressed in terms of the dispositions nat-ural objects or states of affairs have to produce certain reactions in human beings, it faces the same kind of problem as the dispositional-ist view of colors. Those reactions themselves must somehow be identifi ed, and if they are left as reactions with evaluative contents no naturalistic progress will have been made. Reductionism threatens to


take away the evaluative aspect of the attitudes, feelings, and reac-tions that objects are said to produce, just as I think it cannot make the appropriate identifi cations in the case of perceptions of color. It cannot get the contents of our beliefs or attitudes right. To insist that evaluative attitudes simply must be so reducible, and to restrict one-self to reduced or non-evaluative terms alone, would be in effect to eliminate the evaluative vocabulary altogether. Everything we say or think that is intelligible and either true or false would have to be said or thought without it. Here again it is the restrictive naturalism that produces the pressure.

The same pattern is present in the philosophy of mathematics, where the quandary is perhaps most obvious, and has certainly been widely acknowledged. There is no question that we have mathemat-ical and logical knowledge. Could there be an explanation exclusively in restricted naturalistic terms of how we come to have that know-ledge? It would have to make sense of what we believe in mathematics and logic, and could it do so by giving an account of the conditions under which such things are true or false? If so, would that mean that mathematical and logical facts are to be understood as part of nature? Many would insist that even if in some sense or other it is true that seven plus fi ve is twelve, it is not a natural fact, not a fact of the nat-ural world. But we do all believe it, even know it to be true. A restrict-ive naturalist who holds that what mathematical statements assert is not part of the natural world he believes in would have to explain our knowledge of logic and mathematics without himself appealing to any mathematical or logical facts at all.

This has been tried, or at least proposed. But when we look at what has been the most widely-canvassed strategy for carrying it out I think everyone has to confess to a certain dissatisfaction. The main idea has been to locate the source of mathematical and logical truth somehow “in us,” and not in the world independent of us. All such truths have been said to be “analytic” or “true solely in virtue of the meanings of their constituent terms,” something which “we” are in some sense solely responsible for. Since words mean only what we determine or “decide” they are to mean, logical and mathematical truths are said to be true, if at all, only “by convention.” These are all attempts to make sense of mathematical and logical knowledge on the assumption that all of it is “empty” or, in the positivists’ phrase, “devoid of factual content,” and says nothing about the way the world is. Anyone who holds such a view would have to account for human beings’ believing


certain things which he himself does not acknowledge to be states of affairs that hold in the world as he conceives of it.

There is good reason to think that no such theorist would be equipped even to identify, let alone explain, the mathematical know-ledge he admits we all have. For one thing, no naturalistic reduction looks even remotely plausible in this case. Facts about what human beings do, how they think or speak, even how they decide to think and speak, or what conventions or rules they have adopted—all this seems in principle insuffi cient to express the contents of the things we believe when we believe that seven plus fi ve equals twelve or that every thing that is both red and round is red. All of human beings’ doing or deciding or intending whatever they do is contingent, some-thing that could have been otherwise. But it could not have been oth-erwise than that seven plus fi ve is twelve or that everything that is both red and round is red. No contingent truths, however important, could be adequate to express such necessities.

What is more, any naturalism that takes a specifi cally scientifi c form, and says that the natural world is the world described exclu-sively in the terms of the natural sciences, would seem forced to accept truths of logic and mathematics anyway. They are needed in the formulation of physical, chemical, and biological theories. And in any case, it is completely unrealistic to expect a naturalistic theor-ist of any persuasion to get along without any mathematical and logical beliefs of his own. The acceptance of some such truths might even be essential to coherent thought; we could not think without them. If that is so, is that a natural fact, a fact of the natural world? If that meant that it had to be contingent, it is hard to see how it could be. But if for whatever reason we grant the indispensability of logical truths for the possibility of thinking at all, then we have to face the consequences of our really accepting it. That is, we must acknowledge that we do in fact think in those ways, that we do believe that everything that is both red and round is red, that seven plus fi ve equals twelve, and so on. We thereby acknowledge that those and other such demonstrable or undeniable propositions are true.

If the naturalist does or must accept logical and mathematical truths in order to have a determinate conception of the world at all, what becomes of the idea that those propositions do not state any-thing that holds in the natural world? What is the conception of nature that is said to exclude them? It can no longer be identifi ed as simply the world that a scientifi c naturalist believes in, since if he now accepts logical


and mathematical propositions, they are not excluded from what he believes. If this still counts as naturalism, it will be a more open-minded or more expansive naturalism. It does not insist on, or limit itself to, a boundary fi xed in advance. It will have expanded to include whatever has been found to be needed in order to make sense of eve-rything that is so in the natural world. What cannot be avoided is to be accepted. To say that not everything that is accepted is accepted as part of nature raises the question of how the naturalist distinguishes what he thinks of as the natural world from all the rest of what he takes to be the case. And more importantly, what, if anything, now turns on making that distinction?

The same question arises in the case of evaluation. If the goodness or other evaluative aspect of something is not a “natural” quality of it, what exactly is a natural quality? After years of effort G. E. Moore admitted that the best he could come up with was that a natural prop-erty is a property “with which it is the business of the natural sciences or of psychology to deal.” 10 But if that is what a natural property is, then the famous “naturalistic fallacy”—the mistake of giving a “nat-uralistic” defi nition of ‘good’—would be simply the attempt to replace ethics by one of the natural sciences. “Non-naturalism” in ethics would then be nothing more than the view that ethics is not one of the natural sciences. There would be nothing other-worldly or mysterious about that kind of non-naturalism. Who would not want to be an ethical non-naturalist on that defi nition?

To agree that ethics is not one of the natural sciences, or that goodness or badness is not a scientifi c matter, is not to concede that nothing is better than anything else, or that no evaluations are true or false. Not everything that is so is the subject matter of some natu-ral science. If it is true that evaluations cannot be reduced in general to non- evaluative propositions, then our understanding of evalua-tions cannot be seen as built up out of non-evaluative ingredients alone. Anyone who could identify the presence of evaluative atti-tudes in the human beings he observes must understand what evalu-ative attitudes are, even if he does not agree with those he discerns in others. That suggests that he must have some evaluative atti-tudes of his own, on pain of his not being able to recognize them

10 See C. Lewy, “G. E. Moore on the Naturalistic Fallacy”, in P. F. Strawson (ed.), Studies in the Philosophy of Thought and Action , Oxford University Press, Oxford, 1968, p. 137 .


in others. If he acknowledges those attitudes of his, his total view of what is so will contain evaluative states of affairs. He will hold that certain things are better than others, that a certain thing is the thing to do on a certain occasion, and so on. His conception of what is so will have been forced to expand, just as I think it must expand in order to recognize beliefs in logical and mathematical truths, and perceptions and beliefs concerning the colors of things. It expands in each case into a more open-minded or less restricted naturalism.

What I am calling more open-minded or expansive naturalism says we must accept everything we fi nd ourselves committed to in account-ing for everything that we agree is so and want to explain. We want to explain the thoughts, beliefs, knowledge, and evaluative attitudes that we think people have got. If mathematical and logical truths have to be accepted in order to make sense of those attitudes, then they must be accepted, however in some sense “non-natural” they might seem. If some evaluative propositions must be endorsed in order even to recognize the evaluative attitudes of others, then evalu-ative states of affairs must be included too, however diffi cult it might be to decide which particular evaluations are correct. If we have to hold that objects are colored in order to specify and acknowledge all the perceptions and beliefs that we know people have, then the colors of things must be allowed into the picture, and not in reductionist form.

Those who remain committed to a determinate and restricted con-ception of the natural world will have to locate the contents of all those attitudes somehow within that restricted world. If that leads to a distorted conception of the attitudes that people on earth have actu-ally got, as I think it does, the determinate and restricted naturalism is what is responsible for the distortion. A more open-minded or expan-sive naturalism will admit states of affairs and psychological phenom-ena that are found problematic from a more restricted nat uralistic point of view. With no restrictive commitment in advance, a more open-minded naturalism will feel no pressure to exclude from the pic-ture anything that is needed.

By now it should begin to look as if this expandable or more open-minded form of naturalism does not amount to anything very substantive or controversial. It is “open” because it is not committed in advance to any determinate and therefore potentially restrictive conception of what is so. Rather than calling it open-minded nat-uralism we could just as well drop the term ‘naturalism’ and call it


open-mindedness. It says that we must accept as true everything we fi nd we have to accept in order to make sense of everything that we think is part of the world. If that is still called “naturalism,” the term by now is little more than a slogan on a banner raised to attract the admiration of those who agree that no supernatural agents are at work in the world.

12

The Transparency of ‘Naturalism’

I am not sure I can live up to the terms of the invitation to give the Romanell Lecture on Philosophical Naturalism. I fi nd on refl ection that I cannot say what naturalism is. Maybe that is the question: what is it? Of course I am not the fi rst person to notice the indeterminate-ness of naturalism. But that has not always prevented others from going on about it in the past.

The question of what exactly naturalism is, what it implies and what it is meant to exclude, cannot be avoided in any assessment of the merits of naturalism. That is true whether naturalism is under-stood as a doctrine or theory about what is so, what the world is like, or as a method or policy for investigating what the world is like. In each case there is a real question of what naturalism is: what does the doctrine or theory say about what is or is not so? What is or is not a naturalist method or policy? Very few determinate answers have ever been offered to these unavoidable questions.

This is an instance of a diffi culty I often fi nd in philosophy. I would make a modest recommendation to help overcome it. If you hold or endorse or look with favour on a certain doctrine or theory or method or policy in philosophy, I think you should state that doctrine or policy, say what it is or what it says, and not just name it. Once you state it, the name you give it doesn’t matter, and the merits of the doctrine or policy can be assessed on their own.

For that reason I call my talk “The Transparency of [the term] ‘Naturalism’ ”. I think the term on its own does not carry a very determinate meaning. And when you look through it to see what it is meant to stand for in this or that application, it turns out to differ from case to case. So rather than starting with the question of what naturalism is, which in any case will not go away, I want to start with a different question.

What is the point of being or not being a naturalist? What is at stake? Why would any serious, sensible person be concerned to

The Transparency of ‘Naturalism’ 241

promote or defend naturalism? I do not mean this as a rhetorical question. I do not suggest that naturalism is so obviously good that it needs no defence, or that it is so bad that it cannot be given any. I want to ask what is in it for the self-styled naturalist? Why is it felt that naturalism needs to be defended? Or attacked, for that matter? This is a question not about naturalism itself, but about naturalists, or anti-naturalists: those who take a stand one way or the other on the question of naturalism. It is to that extent a diagnostic or histori-cal question. How do things stand, or how has it come to be, that it is felt to be important or even interesting today to promote or attack naturalism?

One thing that has certainly been at stake is the fate of supernatu-ralism. If there is anything naturalism is opposed to, it must be super-naturalism. But that does not take us very far, since to say what supernaturalism is you need some conception of the natural that supernaturalism is said to go beyond. And that is just the question. Nor does the contrast with supernaturalism take us very far on the question of what is at stake, since it does not explain why it is impor-tant, if it is, to oppose supernaturalism, or what the dispute with naturalism really turns on. Is supernaturalism to be rejected simply because it confl icts with naturalism, or for some other reason?

Here one thinks of Hume. He never called himself a naturalist, although he has been called that by others. The label is perhaps appropriate since Hume wanted to understand everything about human nature by appealing only to what human beings can fi nd by experience to be true of the world, including how human beings fi nd out about the world and act in it. His “attempt to introduce the experimental method of reasoning into moral subjects” was to pro-ceed without appeal to anything beyond that. That was something his great predecessors – Descartes, Berkeley, Malebranche, even Locke – had not managed to do. They all brought into their accounts of human perception, thought, belief, and action some divine or “higher” being or power.

Hume’s account of human nature did without any such “higher” being or force. If that is an expression of naturalism, and so a rejection of supernaturalism, it does not oppose supernaturalism on the grounds that it confl icts with something called ‘naturalism’. Hume thought belief in the kind of being or force those other philosophers believed in cannot be supported by anything to be found in the world by employ-ing the capacities and procedures human beings are naturally endowed


with. “Natural religion” was to proceed only from the world as human beings can observe it to be, and Hume thought no support could be found for believing in anything very determinate beyond that. Belief in something supernatural in that sense could be explained by natural “principles of the imagination” operating on human beings in the nat-ural world as it is, without the beliefs’ being true.

Hume’s naturalism, then, if it can be called that, is not an attachment to some doctrine or policy he called ‘naturalism’. It is rather an expres-sion of his idea of what he called a fully “experimental philosophy”. That is the policy of starting where we are, with the experiences and capacities we naturally fi nd ourselves with, and trying to understand and explain the actual mechanisms by which we come to think and believe and feel and act in all the ways we do. We can fi nd out about the world only by making use of the capacities and procedures we are endowed with for fi nding out about it. And we can discover what those capacities and procedures are, and how we come to learn things by using them, only by making use of those very procedures and capacities in our observation of human beings in the observable world. There is no higher authority for our believing anything about the world or about ourselves. This amounts in effect to resistance to supernaturalism, but not on the basis of prior commitment to something called ‘naturalism’. It is simply acceptance of, and acquiescence in, the most Hume thinks can be discovered by human means about the human capacity to know and believe things.

But in Hume’s hands this completely general “experimental” or “naturalistic” explanatory project led to a deeply unsatisfactory out-come. Supernatural beings or forces were not the only things that had to be left out of one’s account of the world. Hume thought his inves-tigation equally showed that human beings’ beliefs in causal connec-tions between things that happen in the world are also without foundation. Careful study shows that there really are no such connec-tions anywhere; belief in them can be explained without supposing there are any. Hume also found that the universally-shared human belief in objects existing independently of us and continuing to exist when they are not perceived is to be explained by nothing more than the presence of fl eeting, momentary perceptions that come and go in people’s minds in accordance with general “principles of the imagination”. And although every human being inevitably believes in his or her own identity as the same person through time, this belief too is revealed as nothing more than a “fi ction” generated by the operation


of those same “principles” of the human imagination. There does not have to be any such identity in anything in the world in order to explain why human beings all believe there is.

All this is the outcome of Hume’s completely general “experimen-tal” study of human beings. It is a paradoxical and unacceptable result. It is unacceptable because we human beings are not only the objects of the study said to yield these results; we are also the agents who carry out the very investigation that is said to lead to them. So we must be able to see the results we arrive at as true of ourselves. We start out believing we are surrounded by a world of enduring, well-behaved objects and fellow human beings causally interacting with one another, and we ask how we come to believe that the world is that way. That is part of the task of “the science of human nature”. On Hume’s account we discover by this kind of investigation of ourselves and our capacities that even those general beliefs that express the con-ception of the world we start from are no more than “fi ctions” that human beings simply cannot help indulging in and accepting.

Hume sometimes describes these “fi ctions” as beliefs that are actu-ally false. He says causal necessity is “something that exists in the mind, not in objects”, 1 for instance, or “There is properly no simplicity in [the mind] at one time, nor identity in different”, 2 although we can-not help believing that there is. Sometimes he does not actually deny that the beliefs are true, but says only that we can see we have no reason to believe them. We are not able to satisfy ourselves “concern-ing the foundation” of these ways of thinking or to remove the “objections that can be raised against them”. 3 This makes us “asham’d of our credulity” 4 while continuing to believe the things we cannot help believing. On either way of understanding the plight, it is not a satisfactory position for the “experimental” study of human nature to have left us in. We are unable to see ourselves as intelligibly related to the natural world we take ourselves to be part of.

What leads to this disaster is not Hume’s commitment to “experi-mental philosophy”. Nor is it his wanting to discover and explain the actual mechanisms of human thought, belief, and action as they are

1 D. Hume, A Treatise of Human Nature , (ed. Selby-Bigge), Oxford University Press, Oxford, 1978, p. 165 .

2 Hume, op. cit., p. 253. 3 D. Hume, Enquiries Concerning the Human Understanding and Concerning the

Principles of Morals (ed. Selby-Bigge), Oxford University Press, Oxford, 1975, p. 160 . 4 Treatise , p. 267.


to be found in the world. The trouble lies in the details of what Hume found to be true about human beings. What prevented him from giv-ing a satisfactory explanation of how human beings believe and come to know the sorts of things they do was the limited resources he thought inquiring human beings are actually restricted to. And he thought it can be discovered that we are so restricted by applying “the experimental method of reasoning” to ourselves. Human beings are restricted to nothing more than “momentary and fl eeting” per-ceptions that come and go according to general “principles of the imagination”. But on that basis alone human beings cannot both explain and at the same time satisfactorily endorse the beliefs whose possession they would explain in that way.

I think the situation Hume’s account would leave us in is even worse than he fully acknowledges. He sees human beings as presented in per-ception with nothing more than fl eeting impressions that leave imprints in the mind in the form of resembling ideas. And he thought there is no way to go from beginnings like that to any beliefs we have reason to hold about the way things are in the wider world. But with only Humean impressions and ideas to work with, I think there is no explaining how anyone could even so much as think or believe anything at all. Hume did recognize the special diffi culty of the nature of belief. He thought he was one of the fi rst philosophers even to notice the problem. But he never found a satisfactory solution, and he knew he had not. The intentionality of thought, and the corresponding possi-bility of both true and false thoughts, is a challenge for any theory with only the limited non-propositional resources Hume had to work with. The thought of something that is either true or false cannot be built up out of such entities alone, whatever they might be. And there is no accounting for human thought and belief without acknowledging that people believe things that are true or false, and do so for reasons that are to be found in the world they observe and investigate.

By the middle of the twentieth century philosophy had become much more self-conscious about its own standing than it had been in the eighteenth century. This was largely the legacy of Kant. Philosophy was seen to occupy a distinctive position in human culture. Its methods, and its results, were to be understood as independent of anything discover-able by sense-experience or empirical observation. Despite the strong and continuing infl uence of Hume, at least in English-speaking regions, the “experimental” study of human nature as he had understood it came eventually to be regarded as not part of philosophy at all.


The logical positivists eliminated the synthetic a priori from the otherwise Kantian conception of human knowledge they inherited. That left them with more or less the same general theory of human knowledge that Hume had defended two hundred years earlier. But now there was a new conception of philosophy to go along with it. Philosophy, it was said, could proceed only a priori , independently of all experience. And since the only things knowable a priori were “ana-lytic” propositions or mere “relations of ideas”, that was the most philosophy itself could discover or establish.

On this view, any distinctively philosophical interest in human thought, belief, and experience would have to confi ne itself to what could be established a priori . It would provide at best an “analysis” of the concept of knowledge or belief or action or whatever it might be, or at best an account of the logical relations among propositions, abstractly considered. Central to philosophy on this conception was what came to be called “the logic of science”, part of which was a “theory of confi rmation” that would exhibit in purely formal terms the relations between the “data” or “evidence” observers can be aware of and what they can thereby know about the world. It was a ques-tion of the logical or evidential relations among certain propositions. That anyone does in fact know such things, and if so how they know them, was to be no part of the strictly philosophical enterprise. Hume’s “science of human nature” as he understood it was thereby disqualifi ed from the domain of philosophy.

All this is what Quine challenged in his “Epistemology Naturalized” in 1968. That essay has been hailed as a decisive turn – or return – towards naturalism in philosophy. For twenty years before that Quine had been challenging the whole positivistic conception of human knowledge embodied in the distinction between analytic and synthetic truths or Hume’s “relations of ideas” and “matters of fact”. In “Epistemology Naturalized” he questioned the attempt to account for our knowledge of the external world as “a logical construct of sense data” as he thought Russell had attempted to do. 5 There had been a

5 W. V. Quine, “Epistemology Naturalized”, in his Ontological Relativity and Other Essays , Columbia University Press, New York, 1969, p. 74 . Quine thought this project had come closest to completion in Carnap’s Der Logische Aufbau der Welt . There is a real question whether that was Carnap’s aim in that book. For a defence of the idea that it was not, see, e.g., Michael Friedman, “Carnap’s Aufbau Reconsidered” and “Epistemology in the Aufbau ”, in his Reconsidering Logical Positivism , Cambridge University Press, Cambridge 1999 .


thriving project from the 1930s onwards of giving a “rational recon-struction” of human knowledge of the world on the basis of whatever human beings were thought to perceive “directly”. The task was to identify some assumptions on the basis of which sensory “data” pre-sumed to be available to human beings could provide good reasons to believe the sorts of things we all believe about the world. The question was hypothetical and broadly logical; any principles that could be found to do the job would answer the epistemological question.

Quine’s response to this philosophical enterprise in “Epistemology Naturalized” was to ask “Why all this creative reconstruction, all this make believe? . . . Why not just see how this construction really proceeds?”. 6 This is very much in the spirit of what Hume had in mind in those philosophically innocent days before Kant. Hume wanted to discover and explain the mechanisms of human thought, belief, and feeling by nothing more than “the cautious observation of human life”. Quine put it more succinctly and more scientifi cally: “Why not settle for psychology?”. 7 This was to be an “experimental” or observational study of actual human beings. Quine’s endorsement of this kind of epistemology was part of his admirable resistance to the idea that philosophy must be exclusively a priori , or perhaps, even better, his resistance to the idea that there is any a priori knowledge at all. If Hume’s “experimental” project of the “science of man” can be called ‘naturalistic’, then so too could Quine’s conception of epis-temology, for the same reason.

In fact Quine was more open-minded than his mention of psychology alone suggests. “If we are out simply to understand the link between observation and science, we are well advised to use any available information, including that provided by the very science whose link with observation we are seeking to understand”. 8 This too is in the spirit, if not exactly the letter, of the Humean enterprise.

The theory Hume relied on to explain the link between what Quine calls “observation and science” was the theory of impressions and ideas and the “principles of the imagination” according to which such things come and go in the mind. Hume thought that theory was borne out by the observation of human beings. But that way of trying

6 Op. cit. , p. 75. 7 Ibid. 8 Op. cit. , p. 76.


to understand ourselves made it impossible to account for our thought about the world in a way we could fi nd satisfactory. What Quine thought can be discovered by the study of human beings is not the same as what Hume thought. But following Quine’s recommen-dation as he understands it would leave us in a position structurally similar to the position Hume’s account left us in. We would be no better off.

The “psychology” Quine is willing to settle for is a chapter of what he calls “natural science”.

It studies a natural phenomenon, viz., a physical human subject. This human subject is accorded a certain experimentally controlled input – certain pat-terns of irradiation in assorted frequencies, for instance – and in the fullness of time the subject delivers as output a description of the three-dimensional world and its history. 9

In studying the transition from those human sensory stimulations to that “torrential” scientifi c output there is no question of our seeking to validate or confi rm or support the conception of the world human beings end up with. We simply want to understand how they get it. And we can make use of anything we happen to accept about the world from any scientifi c source, including the very science whose acceptance we are trying to explain.

This leaves us with the same kind of diffi culty that confronted Hume. We as “scientists of human nature” are both the objects and the agents of the study, so we must fi nd a way to see the results we arrive at about human beings as true of ourselves. On Quine’s account, we “scientists” of human nature do that by studying:

how the human subject . . . posits bodies and projects his physics from his data, and we appreciate that our position in the world is just like his. Our very epistemological enterprise, therefore, and the psychology wherein it is a component chapter, and the whole of natural science wherein psychology is a component book – all this is our own construction or projection from stimulations like those we were meting out to our epistemological subject”. 10

When this is taken strictly and literally, I think it is as unsatisfactory as the position Hume left us in, and for the same kind of reason. To “appreciate” that we are in the position Quine describes is to see even

9 Op. cit. , pp. 82–83. 10 Op. cit. , p. 83.


our granting that there are human subjects receiving stimulations in a world we believe in as itself part of a “torrential output” generated in us by certain patterns of irradiation we receive in assorted frequencies at our sensory surfaces. Even our taking it that there are such patterns of irradiation, and that we receive some of them, or even that we have sensory surfaces at all, is to be seen in turn as part of that “description of the three-dimensional world” that irradiations apparently produce in us. Our having such a view of the world is not itself simply an irradiation of a surface or a piece of “data” we receive from any such irradiation. We must “appreciate” that the whole of what we call “nat-ural science” (including all that talk of irradiations) is itself nothing more than “our own construction or projection” from some “meager data” we have received from somewhere.

If this is an unsatisfactory position to be left in, as I think it is, it is not because what it says is not true. Even if we do receive such stimu-lations at our sensory surfaces, as I am sure we do, and even if they do lead us to believe all the things we come to believe about the world, we are still left at best, as Hume left us, with no way of seeing that the world we believe in gives us reason to believe that it is the way we think it is. In that respect we would remain at best, in Hume’s words, “asham’d of our credulity”.

That is one kind of dissatisfaction. But I think the position Quine would leave us in is worse than that. Being “asham’d of our credulity” at least involves our having beliefs . And on Quine’s conception of “sci-ence” it is not strictly speaking a “scientifi c” fact that someone believes something. It could not be discovered by “science” so understood that human subjects do “deliver as output a description of the three-dimen-sional world and its history”. The very phenomenon “naturalized epis-temology” is meant to explain – human beings believing and knowing things – therefore could not show up as part of “natural science”. The most that can be discovered by that chapter of natural science called psychology as Quine understands it is that human organisms receive stimulations of various kinds at their sensory surfaces and issue out-puts in the form of utterances or assents to utterances. Even the term ‘assent’ is dubious. If to assent is to agree or accept something as true, it can never be “scientifi cally” discovered that someone assents to some-thing. The most that can be “scientifi cally” observed is an organism’s issuing or responding to an utterance. And to fi nd that someone utters or has a disposition to respond to an utterance in a certain way is not to discover what the person means by it or thereby assents to or accepts.


I suggested that Hume does not have the resources to account for the presence of thought and belief in the world. I think the same is true, strictly speaking, of Quine. Given his conception of science, which is also a conception of what there is in the world for any natural science to discover and explain, a purely “scientifi c” epistemology meant to explain human belief and knowledge could not even fi nd its potential subject matter.

If that is so, it is not because of Quine’s “experimental” and accordingly “naturalistic” approach to the study of human beings. Nor is it because of his resistance to purely a priori philosophizing. That was not the source of Hume’s diffi culty either. The limitation comes from Quine’s conception of nature or the natural world, of what there is in the world for any “natural science” to study and understand. It is his commitment to a determinate conception of the limits of nature or the natural. It is a doctrinal “naturalism” according to which everything that is so in the natural world, and therefore everything there is for science to discover, is expressible in exclusively extensional terms. Quine thinks that is the only way things really are. He does not tell us how that conclusion was reached. He does not tell us by what “experimental” or “naturalis-tic” investigation of the world it has been shown that that is how things are. But that doctrine, with its accompanying restriction on what there is for science to fi nd out, is what stands in the way of our fi nding his “scientifi c” version of “epistemology naturalized” a sat-isfactory explanation of how we come to think and believe and know the sorts of things we do.

This will seem uncharitable to Quine, in taking him too strictly. Quine would not deny that we think and believe and know many things, just as none of the rest of us would deny it. That is true. Hume would not deny it either. But the question is not whether David Hume or W. V. Quine say and believe what all the rest of us say and believe about ourselves. The question is whether the kind of “experimental” or “naturalistic” philosophies Hume and Quine defend can actually acknowledge and explain our believing all those things that no one would deny. If they cannot, as I suggest they cannot, it is not because those philosophies are “empirical” or fol-low only “humanly natural” ways of trying to discover things about the observable world. It is not because they are “naturalistic” in that sense. It is because, in Hume’s case, there is error or distortion in how he thinks human beings actually work. And in Quine’s case


there is a restricted conception of what there is in the world for any “scientifi c” study of the world to discover.

Quine’s commitment to an exclusively extensionalist conception of the facts of nature is embodied in his doctrine of “the indeterminacy of translation”. 11 All human dispositions to utter or respond to the utterances of others are compatible with countless different and mutually incompatible translations of those utterances into other utterances and so compatible with the person’s meaning any one of many different things by a particular utterance or response. Utterances and responses to utterances can themselves all be described in fully extensionalist terms, but no attributions of meaning or understand-ing or intention to something is equivalent to any such purely exten-sionally-stated matters of fact. The same point is brought out on essentially the same basis by the argument for Kripke’s “sceptical paradox” that “there can be no such thing as meaning anything by any word”. 12 Whatever is described in purely extensional terms, and so not described as meaning anything or being understood in some particular way rather than others, can never fi x or determine or con-stitute something’s meaning one thing rather than another.

As it stands, this is no more than a denial of the equivalence of intentional and purely extensional language. The non-equivalence is expressed in the so-called Brentano thesis, roughly stated by Quine as the idea that “there is no breaking out of the intentional vocabulary by explaining its members in other terms”. 13 Quine observes that:

One may accept the Brentano thesis either as showing the indispensability of intentional idioms and the importance of an autonomous science of inten-tion, or as showing the baselessness of intentional idioms and the emptiness of a science of intention. My attitude, unlike Brentano’s, is the second. 14

This attitude, I believe, is what is responsible for the unsatisfactory position Quine’s “scientifi c” epistemology would leave us in. What is that uncompromising extensionalist attitude itself derived from? Is it the result of some “experimental” or “naturalized” study of the world and the human beings who inhabit it? Whatever its source, Quine remained unrepentant. In a late piece of reminiscence called

11 See W. V. Quine, Word and Object , MIT Press, Cambridge, Mass., 1960, chapter 2 . 12 Saul A. Kripke, Wittgenstein on Rules and Private Language , Harvard University

Press, Cambridge, Mass., 1982, p. 55 . 13 Quine, op. cit. , p. 220. 14 Op. cit. , p. 221.


“Confessions of a Confi rmed Extensionalist” he said “I doubt that I have ever fully understood anything that I could not explain in exten-sional language”. 15 If he is right about that, it means this very remark of his is something he does not fully understand.

There certainly has been a tendency in philosophy to think of uncompromising “naturalism” as carrying with it suspicion of or resistance to intentional vocabulary and the intentional phenomena it gives expression to. That has encouraged the apparently futile effort to explain intentional attitudes in exclusively non-intentional terms. But there is no human thought, knowledge, or action without human intentional attitudes. So there is no acknowledging and understand-ing thought, knowledge, and action without acknowledging and understanding those attitudes. If human beings are part of the “nat-ural” world, our having intentional attitudes must be understood as part of the “natural” world. We are the kinds of beings who “ naturally” have such attitudes.

Quine in his response to the Brentano thesis is dismissive of what he calls an “autonomous science of intention”. He appears to regard belief in the possibility of such a science as the price of granting the indispensability of intentional idioms. We know why Quine thinks no genuine “science” is in the offi ng. But insisting on the indispensabil-ity of the intentional and its irreducibility to the extensional does not imply acceptance of an “autonomous” domain of intentionally described facts that can be understood in isolation from the rest of what is so in the world.

There was a time when there were no beings with intentional atti-tudes at all. Now they are all over the place. One might take an inter-est in how that transition occurred; how there came to be subjects with intentional attitudes in a world that originally did not contain them. Explaining how that came to be so in the “natural” world could perhaps be called “naturalizing the mind” (to coin a phrase). 16 It is by no means clear how such an explanation would proceed, or how sat-isfying we could fi nd it. Could we expect to explain from the ground up how there came to be such a thing as language and meaning in the world, and so how human beings have come to acquire the range of

15 J. Floyd & S. Shieh (ed.), Future Pasts: The Analytic Tradition in Twentieth-Century Philosophy , Oxford University Press, New York, 2001, p. 217 .

16 See Fred Dretske’s Naturalizing the Mind , MIT Press, Cambridge, Mass., 1995 , where he pursues a somewhat different project.


intricate capacities involved in their knowing all the sorts of things they now know? And would such an explanation invoke only facts expressed in exclusively non-intentional terms?

To understand anyone as having determinate intentional attitudes we must see both the person and his attitudes in relation to the world he is in contact with and thinking about. We must make use of what-ever we know or believe about the world and the person that helps us make the right kind of sense of what he does. Any attitudes a person can be understood to have must be seen as intelligibly related to other attitudes the person can be understood to hold. The world the person is in touch with and thinking about must therefore be understood as one in which something’s being so can be reason to think that some-thing else is so, and believers must be understood as capable of rec-ognizing such reasons. Those who attribute intentional attitudes must also believe certain things about the world and recognize some things as reason to believe certain other things about it. We are unavoidably immersed in the very world we credit ourselves and oth-ers with believing and knowing things about. This makes it diffi cult to envisage a successful explanation of intentional attitudes in gen-eral, from scratch.

Even if we had an explanation of that kind that satisfi ed us, it would not imply that someone’s having an intentional attitude is equivalent to some combination of extensionally-describable states of affairs. There was a time when there was no mathematical knowledge in the world either. Now there is a lot of it. But a success-ful “naturalistic” explanation of how mathematical knowledge came to be, if we had one, would not imply that mathematical propositions are equivalent to some combination of non-mathematical facts of the kind that played a role in the emergence of that very knowledge. Knowing how mathematical knowledge came to be is not itself math-ematical knowledge, but it is diffi cult to see how mathematical knowledge in general could be explained without some mathematical knowledge on the part of the explainer. In the same way, it is diffi cult to see how intentional attitudes in general could be explained with-out some intentional attitudes on the part of the explainer and so an understanding of the world as already containing, for those who can notice them, reasons to hold certain attitudes towards it.

If we are to make any sense of human thought, belief, knowledge, and action there is no abandoning the appeal to intentional attitudes. The practice of making sense of ourselves and others in that way is


certainly something we could understand much better than we now do. The search for a better understanding of ourselves as thinkers and believers could be called a “naturalistic” study of human beings as part of nature, or perhaps even a “science of human nature”. But it does not really matter what you call it. What matters is a clear understanding of the kinds of questions you are pursuing and an unprejudiced and undoctrinal openness to what it would take to answer them.

A more open-minded and truly experimental approach along these lines is vigorously defended by Penelope Maddy in her book, Second Philosophy . 17 Maddy promotes what she calls “a naturalistic method”, and she uses the word ‘naturalism’ throughout, but she admits that the word is really not much help for her purposes. She thinks that by now ‘naturalism’ has come to mean little more than “a vague science-friendliness”, 18 or else it means so many different things in different people’s hands that it is better to drop the term altogether and simply describe the various projects in question. I applaud this recommen-dation. Maddy calls the kind of investigation she has in mind ‘Second Philosophy’. I fi nd this congenial too, since that name means nothing at all until Maddy specifi es what she means by it. To understand her, we have to look right through that label to see what she means to draw our attention to, just as I think we have to do with the term ‘naturalism’ itself.

Maddy nonetheless sees herself as squarely in what she thinks of as the “naturalistic” tradition of Hume and Quine. They are her heroes – especially Quine – but she does not follow them into the dead-ends that I think undermine their otherwise enlightened enterprises. Hume, she says, “loses his naturalistic faith” because he is somehow led “down his skeptical course”. 19 That is what she thinks must be resisted. She sees her Second Philosophy as “a variety of post-Quinean naturalism”, 20 but without what she calls “the less naturalist-friendly strands of Quinean thought”, such as the indeterminacy of translation, the inscrutability of reference, and ontological relativity. 21 They too must be resisted.

17 Penelope Maddy, Second Philosophy: A Naturalistic Method , Oxford University Press, New York, 2007 .

18 Op. cit. , p. 1. 19 Op. cit. , p. 46. 20 Op. cit. , p. 96. 21 Op. cit. , p. 84n.


Maddy is more open-minded and undoctrinal in not committing herself in advance to any defi nition of “science” or any limits to the kinds of things science can fi nd out. Conclusions reached in Second Philosophy are not to be defended on the authority of something called “science”, or even something called “Second Philosophy” or “naturalism” either, but only on the basis of the evidence actually available or discoverable. Second Philosophy is accordingly not a set of doctrines or a specifi ed method. It is at best a kind of inquiry “without strict disciplinary allegiances”. We can understand it only by understanding in each case the questions it considers and the ways it tries to answer them.

One thing that links this enterprise most strongly with the “ naturalism” of Hume and Quine is Maddy’s rejection of the idea that questions in philosophy are to be answered by exclusively a pri-ori means. With those pretensions exposed and set aside, Maddy sees no danger of “common sense and its scientifi c refi nements” being “convicted of undercutting the reasonableness of their own methods”. 22 That is what I have suggested happens in the philoso-phies of Hume and Quine.

Traditional classifi cation of questions or methods appropriate to philosophy or to this or that science play no role in the kind of inquiry Maddy recommends. For instance, she thinks our knowledge of rudi-mentary logic is best explained as knowledge of the general structure of the world around us. Drawing on psychological research, some of it on infants, she thinks human beings acquire that knowledge because their “most primitive cognitive mechanisms” allow them to detect that structure in the world, and their “cognitive mechanisms” are that way because human beings live in a world that has just that structure. This account, of course, is open to empirical refutation. Nothing we believe about anything is immune to that risk. The only human route we have to any understanding of what is so is through careful use of our natural capacities and everything we think we know about the world, including what we know about the capacities we make use of in trying to understand it.

By this point in the gradual paring-away of naturalism, I think we have reached the level of platitude, perhaps enlightened platitude. If this is naturalism, there is no longer anything very determinate that the term can be said to mean. To my original question, “Why would

22 Op. cit. , p. 45.


any serious, sensible person be concerned to defend naturalism?”, the answer seems to be, at least with respect to philosophy, that there is good reason to resist a priori philosophical restrictions that would limit what a study of something or other can legitimately do. We should just try to fi nd the best ways we can to understand the things that puzzle us, and proceed carefully. If this is naturalism, it sounds like a good idea. But it is not a defi nable method, and it implies noth-ing determinate about what is so in the world.

I think this shows what I call the transparency of ‘naturalism’. In philosophy you always have to look right through the term to see in each case what it is meant to stand for. It has some determinate but still very general meaning as opposition to certain a priori philosoph-ical doctrines. But otherwise we are left with nothing much more determinate than the idea of careful, informed investigation, or per-haps just responsible inquiry. That is not nothing to be left with. It is all we’ve got. But there is no need to decide in advance how inquiry into whatever interests us must, or must not, proceed.

13

Anti-Individualism and Scepticism

I take anti-individualism in this context to be the view that what thought a person is thinking at a particular time depends in part on that person’s present and past relations to what is so in the world around him. I think that idea is immensely important in understand-ing without confusion or paradox how human thought and experi-ence as we know them are possible. There is no telling, so far, just how rich the consequences of that idea will turn out to be. In this essay I consider some of its apparent implications. What the view implies obviously depends on what it does and does not say, and that in turn depends on exactly how it is formulated and defended. Here I will enter into those crucial questions of formulation only as far as I think I need in order to examine some of the very general epistemo-logical implications of the kind of view of perception, thought, and belief defended by Tyler Burge. I will have in mind as well the very closely related ideas of Donald Davidson.

The view that certain relations must hold between a thinker and the world around him in order for anything he does or whatever happens to him to count as his thinking a particular thought fi nds a parallel in the perhaps more familiar point that certain relations must hold between a perceiver and the world around her in order for what hap-pens to her on a particular occasion to count as her perceiving a par-ticular thing, or her having a perception of a particular kind. In the basic cases—those without which we could not understand people to be having perceptions of certain kinds at all—the perceptions that are attributed to perceivers are perceptions of something that is so in their perceptible environment. This does not mean, of course, that every-one always perceives what is there and only what is there. But even when we misperceive or perceive some object or state of affairs that is not there, our having a perception of just the type we are having on that occasion (which is in fact false or illusory) also depends on our past history and the world in which the very perceptual capacities we

Anti-Individualism and Scepticism 257

are exercising at the moment have been developed and exercised. A different world, or a different past for us, could have meant that we were having different perceptions from what we are having right now, even if all the stimulations at our sensory surfaces remained just what they are at the moment. As Burge puts it, “most perceptual represen-tations . . . represent what, in some complex sense of ‘normally’, they normally stem from and are applied to”. 1

It is not easy to state in simple and adequate terms exactly what this implies about the reliability of human perception. But at the very least I think we can see that there is, and is meant to be, a reassuring point here. It is not that none of us is ever wrong or ever misperceives anything. It is not even that no one person could be more wrong than right in perception over a long period of time. But the view does require that for someone to misperceive, or to perceive things as they are not, it must be possible for him or others to perceive things as they are. Things’ being the way they are perceived as being is in that sense a condition of there being perceptions of things being one way rather than another at all. Without enough “normal,” that is, veridical, perceptions of certain kinds, no one could be understood to have any perceptions of those or any other kinds, and so no one could even be understood to have non-veridical or illusory percep-tions. There is a certain reassurance in the thought that if we ask “Could all or most of the perceptions we actually have misrepresent or mislead us about what they in fact stem from?” we must on anti-individualist grounds answer “No, they could not.”

Burge puts this reassurance at one point by saying that in the case of perception “we are nearly immune from error in asserting the existence of our perceptual kinds, and of other kinds that are taught by more or less immediate association with perceptually based applications”. 2 There is a question here of how “nearly immune from error” we are, and even what different degrees of immunity amount to. These are some of the questions of adequate formulation I men-tioned at the beginning. The precise character and measure of this immunity do not matter at this point. I think we get the idea, and it is enough to be going on with. Let us say that, given the conditions

1 Burge, Tyler. 1986. Cartesian Error and the Objectivity of Perception. In Subject, Thought, and Context. McDowell and Pettit (eds.). Oxford: Oxford University Press. p. 131.

2 Ibid.


of identifi cation of perceptual experiences as being of this or that type, or as having this or that content, we could not be largely wrong in perception. Our perceptions are in large part perceptions of things as they are. Even with the scope of possible error left vague in this way, we can go on to ask about the status of this conclusion and about its implications for certain varieties of scepticism about the senses and about the empirical world.

Is the reassurance this view provides an adequate reassurance against scepticism about perception? That will depend on what scep-ticism is, and on what it takes to oppose it or to provide reassurance against it. But if this kind of view does have anti-sceptical conse-quences in the case of perception, whatever they might be, can those consequences be carried over directly, or at least developed along similar lines, to apply to belief and knowledge in general, and so to oppose epistemological scepticism everywhere?

I am not sure what Burge’s answers to these questions would be. He certainly seems to accept what I have just said about perception. He thinks the anti-individualist view in that case provides what he calls a “qualifi ed basis for the oft-repeated slogan that error presup-poses a background of veridicality”. 3 He also draws a parallel between the external conditions of perceptual experiences of certain kinds and the conditions of thoughts and beliefs more generally. In fact he thinks the point about perception can help explain and make plausible his anti-individualist theory of thought; it can bring out the way in which entertaining a certain thought also has its conditions in the wider world. He holds that in both cases those external condi-tions’ holding is not something the perceiver or the thinker himself must know (or at least must know empirically) in order to have the perception or thought in question, or even to know that he has it. This last point is the key to his account of self-knowledge in general, or knowledge of one’s own psychological states.

Despite this parallel, and the fact that he thinks we are “nearly immune from error” in asserting the existence of the things we per-ceive, Burge thinks, as he puts it, that “we are not immune from fairly dramatic and wholesale error in characterizing the nature of the empirical world”. 4 He also thinks the “oft-repeated slogan that

3 Burge, Tyler. 1986. Cartesian Error and the Objectivity of Perception. In Subject, Thought, and Context. McDowell and Pettit (eds.). Oxford: Oxford University Press. , pp. 130–131.

4 Ibid. , p. 131.


error presupposes a background of veridicality” is “sometimes mis-used”. 5 Since he appears to contrast the scope of the possibility of “fairly dramatic and wholesale error” about the empirical world with our near immunity from error in the case of perception, this might suggest that he thinks that slogan is “misused” in trying to show that we cannot be largely wrong in our beliefs about the world in general. He is inclined to believe that Quine and Davidson in par-ticular “sometimes use this important idea [as expressed in the slo-gan] with insuffi cient discrimination”. 6 Burge uses that same idea himself to show that we are “nearly immune from error” in the case of perception. Does he think that Quine and Davidson are indis-criminate in applying it more generally to all our beliefs about eve-rything, even well beyond the limits of perception? Or is there some other way in which they are not suffi ciently discriminating in their use of it?

Davidson does conclude from what looks like the same general conception of the identifi cation of thoughts and beliefs, as well as perceptions, that, as he puts it, “most of our beliefs are true,” 7 even that “most of our beliefs must be true”. 8 He thinks the anti- individualism that guarantees the truth of most of our beliefs “serves to rescue us from a standard form of skepticism” 9 or, more cau-tiously, that it has within it “the makings of a cogent argument against some forms of skepticism”. 10 Perhaps Burge thinks that draw-ing that optimistic-sounding conclusion directly in that way is an insuffi ciently discriminating use of the important idea. He himself holds, to the contrary, that “there is no easy argument against skepti-cism from anti-individualism and authoritative self-knowledge”. 11 What, then, are we to make of Burge’s own reassuring-sounding con-clusion about perception—that we are “nearly immune from error” in that case? If that counts as an appropriately careful and discrim-ating use of the idea that error presupposes veridicality, does it not

5 Ibid. 6 Ibid. , p. 131n. 7 Davidson, Donald. 1983. A Coherence Theory of Truth and Knowledge. In Kant

oder Hegel? D. Henrich (ed.) . Stuttgart: Khitt-Cotta. p. 435. Also in 1986, Truth and Interpretation, Lepore (ed.), Oxford: Blackwell.

8 Ibid. , p. 431. 9 Ibid. , p. 438. 10 Davidson, Donald. 1988. Reply to Burge. Journal of Philosophy 8511: 664–665 . 11 Burge, Tyler. 1988. Individualism and Self-Knowledge. Journal of Philosophy

85(11): 655 .


rescue us or save us from some forms of scepticism, at least about the senses or the things we perceive? If it is not possible for all or most of the perceptions we understand ourselves to have to misrepresent or mislead us about what they in fact stem from, is that not an argument against scepticism?

I think (I’m not sure about this) that Burge’s answer would be “No”—that the general consideration, as it stands, is not an argu-ment, or at least not an easy argument, against scepticism. Why not? I do not think he means that although scepticism about our percep-tual beliefs has been ruled out, scepticism about beliefs that go beyond what we can perceive remains a possibility. For one thing, that would seem to require some sharp and lasting and therefore dubious distinction between what we can perceive to be so and what we believe or hypothesize or infer to be so on the basis of what we perceive. It is dubious because such a distinction seems on the con-trary precisely to encourage scepticism. It is accordingly more realis-tic to allow that the limits of what we can perceive can be gradually moved further and further outward as the sophistication of our knowledge and of our perceptual capacities increases, so that there would eventually remain virtually nothing that a suffi ciently informed and perceptive person could not be said to perceive to be so under some circumstances or other. That would mean that whatever immunity from error we enjoy in perception is at least theoretically extendable to everything, or virtually everything, we believe about the world. Second, the traditional source of scepticism about our knowledge of the world in general, at least in the so-called modern era, has been the threat of scepticism about perception. If scepticism had somehow been blocked there, or had not even got off the ground, it is not easy to see what would be left of scepticism about our knowl-edge of the world in general.

I think Burge must mean that the reassuring conclusion about the reliability of perception does not itself amount to an argument, or at least an easy argument, against scepticism, even with respect to our perceptual beliefs. Scepticism in this context is presumably the view that we do not know anything about the world around us, or that we have no reasons to believe the things we believe about it. If an argu-ment “against scepticism” is an argument for the negation of that doctrine, it would have to show that we do know or have reason to believe the things we believe. If that is what Burge means by an argu-ment against scepticism, then I agree that what I have been calling the


reassurance provided by anti-individualism does not amount to such an argument. For one thing, even if it does imply that most of our beliefs about the world are true, or by and large true, it does not fol-low that those true beliefs amount to knowledge. That is a point Davidson concedes, although he continues to hold that anti- individualism nevertheless “serves to rescue us from a standard form of scepticism”. 12 I think the distinction is important. Providing the negation of scepticism and so knowing that it is false is not the only way to be saved from it.

When Burge says “there is no easy argument against skepticism from anti-individualism and authoritative self-knowledge”, he is speaking in part about what he calls “transcendental” responses, in particular, Hilary Putnam’s attempt in the fi rst chapter of Reason, Truth, and History to show that, given an anti-individualist account of meaning and belief, even a brain in a vat could not be largely wrong about what is happening in its environment. Burge thinks those considerations “do not do much to undermine skepticism”. 13 If he means that they do not show that we know that we are not brains in a vat and so know what is happening now in our environment, I agree. I do not think a valid argument can be found which a person might deploy, starting from anti-individualist premises, to take him to the conclusion that he knows he is not a brain in a vat, or that he knows most of the things he knows about the world, in a way that refutes philosophical scepticism. But again the question is whether anti-individualism must be shown to have that implication in order to have anti-sceptical consequences.

Burge thinks that what he calls “transcendental” responses would at most provide only “‘general’ guarantees against skepticism”. 14 Perhaps that is why he thinks they do not amount to a good argu-ment against it. He does not say what a “general guarantee against skepticism” is or would be, or why it would not be just what we want. It is pretty clearly not what he wants. He contrasts transcendental responses with what he regards as a more promising line, which he says would actually “justify particular perceptual knowledge claims in the face of skepticism”. 15 This suggests that a completely general

12 D. Davidson, 1983, op. cit. , p. 438. 13 T. Burge, 1988, op. cit. , p. 655n. 14 Ibid. , p. 655. 15 Ibid. , pp. 655–656.


anti-individualist reassurance about the reliability of perception would not be enough to counter scepticism as Burge wants to coun-ter it. I have agreed that the reassuring general point does not imply that scepticism is false—that we do know or have reason to believe all or most of the things we think we know. But it appears that Burge would like to prove, for some of the particular things that philosoph-ical scepticism would say we don’t know, that we do or can know them after all. The merely general reassuring-sounding conclusion implies nothing about any particular case. If Burge does want to prove, in particular cases, that we know—if that is for him a condi-tion of success for an argument against scepticism—that would explain why he thinks “there is no easy argument against skepticism from anti-individualism and authoritative self-knowledge”. There is no easy argument from those premises that can be used to prove, in a particular case, that we know in that case.

He hints that a non-easy, or at least a more complicated, argu-ment to that effect could eventually be given. He thinks it would show that we can know, presumably in a particular case, that no demon is fooling us, or that we are not brains in a vat. We could know that, he suggests, “by inferring it from our perceptual knowledge”. 16 This suggestion is made briefl y, and in a footnote, and is not developed further. He says there that “this is a complicated matter best reserved for other occasions”. 17 Well, this is another occasion.

As I understand it, the argument would show, of particular per-ceptual knowledge claims made on particular occasions, that they are true or justifi ed. That is what would make it more than a “gen-eral guarantee”. But it would have to reach that conclusion while remaining consistent with another feature of perception that Burge stresses, namely, that “in any given case, all of a person’s perceptual capacities . . . could in principle be mistaken about the empirically perceivable property (object, relation) being perceptually presented”. 18 This, he holds, is a consequence of the fact that our perceptual experience as we understand it is of things or relations or properties that are “ objective ”—they are as they are independ-ently of anyone’s thoughts or experiences of them, or even of

16 T. Burge, 1988, op. cit. , p. 655n. 17 Ibid. 18 T. Burge, 1986, op. cit. , p. 125.


whether there are any mental phenomena at all. An anti-individualist or externalist theory of perception would presumably imply that, even in the face of that general possibility, if things are in fact a certain way on a particular occasion, and the person’s perceiving them to be that way is connected in the right way with their being that way, then the person is perceiving things as they are on that occasion, and in that way thereby knows that they are that way. If all of that were so, it seems, then given anti-individualism, scepti-cism would be false. The person would know. But that is still a com-pletely general, or only conditional, claim. The stronger or non-easy complex argument that Burge envisages would presumably enable the person to prove, at that time, that what she perceives to be so at that time is so, and that she therefore does know the truth of what she perceives to be so. And she would have to prove it in that par-ticular case even though in any particular case a perceiver can be wrong about how things are even though she perceives them to be that way. I agree that that would be no easy argument. This is a fascinating suggestion that I hope Burge will pursue.

I am still far from certain that it is really what he has in mind. Rather than speculate about it further, I would like to raise a prior question. Do we have to prove such a thing, or in any other way prove that scepticism is false, in order to oppose it on anti-individualist grounds? Is the general reassurance provided by anti-individualism not enough in itself to block any serious sceptical threat? At the risk of being thought too friendly to so-called “transcendental” responses, this is the question I would like to explore.

Davidson holds on anti-individualist grounds that most of our beliefs are true, or that they are by and large or for the most part true. The anti-individualist grounds he relies on are to be found in the conditions of what he calls interpretation—one person’s understand-ing and communicating with another—which involves the attribu-tion to that other of thoughts, perceptions, beliefs, and other propositional attitudes. We can do that only by connecting the atti-tudes we attribute to people with circumstances or states of affairs in the world in which we as interpreters and they as interpreted interact. This inevitably produces a large measure of agreement among those who can understand and communicate with one another, according to Davidson. Because an interpreter relies on what he himself believes to be true of the world, he will fi nd those he interprets to have largely true beliefs.


Burge’s anti-individualism, even about perception alone, is also a theory of our practices of attributing psychological states and attitudes with determinate contents to people in the world we inhabit. The thought experiments he appeals to to support his anti- individualism turn on and exploit our capacities of belief- and attitude-attribution as they actually are. We are convinced by his “experiments”, if we are, because we fi nd that in the situations he describes his subjects to be in we think it correct to attribute to them the attitudes that his anti-individualist theory implies that they have. What we are told is so in the world surrounding the subject helps deter-mine what attitude we correctly ascribe to him, whether he knows it is so or not. If the thought experiments succeed, they show us that our attitude-ascribing practices, and so our understanding of per-ception, thought, and other mental attitudes, are anti-individualistic. The world and our thoughts about the world cannot come completely apart. In that sense we understand that we are “nearly immune from error” in our perceptual beliefs.

To declare, as a consequence of this, that most of our beliefs must be true, could be just a way of announcing, on anti-individualist grounds, what one holds to be true of one’s fellow human beings. That is one way of understanding Davidson’s conclusion. “Given that they have got the particular beliefs I have ascribed to them”, he could be saying, “those beliefs must be by and large or for the most part true”. That too could be the kind of “immunity from error” that Burge thinks we must enjoy in perception. We could not under-stand people to perceive this, that, and the other as we do without acknowledging that those perceptions we ascribe to them are by and large veridical. That is a condition of perception-attribution on the anti-individualist theory. The people we understand to have perceptions of this or that kind do not just happen to have avoided error for the most part in their perceptual beliefs; they are largely “immune from error”. We fi nd that they could not be largely wrong, given that they have got the kinds of perceptions we ascribe to them.

But saying that most of our beliefs must be true, or that people couldn’t be largely in error in their perceptual beliefs, can also sound like a stronger thesis to the effect that beliefs and percep-tions simply must be largely true or veridical—that that is a condi-tion of anyone’s having beliefs and perceptions at all. This is certainly suggested by Davidson’s saying “belief is in its nature


veridical”. 19 That makes it sound as if beliefs simply could not be false, that no reasonably rich set of beliefs could fail to be largely true, that if there are beliefs at all, they must be for the most part true. That thesis, it seems to me, would threaten the objectivity of what we believe to be so—the idea that the truth or falsity of what is believed to be so is in general 20 independent of its being believed to be so by this or that person or group, or even by all human beings universally. The stronger thesis could allow that such inde-pendence holds for each particular belief, or perhaps even for each relatively small subclass of beliefs, taken on its own. But it appears to rule out any such independence for all or most of a large set of beliefs taken as a whole. The stronger view seems to imply that the world would have to be by and large the way that any group of believers with a suffi ciently rich and comprehensive set of beliefs believed it to be. Not just because human beings can reasonably be expected to get things more or less right, but simply because that is what beliefs are. They are, in their “nature”, veridical.

That stronger thesis, if it were correct, would be, or would pro-vide, a foolproof antidote to scepticism, it seems to me, even though it would remain, in Burge’s terms, completely “general”. It does not directly imply the negation of scepticism—that we know or have reason to believe the truth of those things we believe—but anyone who came to realize that “belief is in its nature veridical” in the strong sense could easily conclude that he, or at any rate human beings generally, could not go for the most part wrong in believing whatever they believe. This kind of reassurance is what I think Davidson had in mind when he argued in “A Coherence Theory of Truth and Knowledge” that anyone who came on refl ection to see that “belief is in its nature veridical” would have a reason to believe that most of our beliefs are true, and thus the threat of pervasive failure or defect in the grounds of our beliefs would then have been defused. We could see by refl ection that if what we seek is the truth, we cannot fail, at least for the most part. This is very reassuring news.

19 D. Davidson,1983, op. cit. , p. 432. 20 I say “in general” because there are some things that must be so if anyone believes

anything, e.g., that someone believes something. But for what does not fall into that special class of beliefs, the truth or falsity of what is believed is independent of its being believed.


But like much good news, this strategy also has its negative side. As the independence or objectivity of what is believed fades, believing what is for the most part true becomes less and less of an achieve-ment. It is something in which, on this stronger view, we could not possibly fail. Such are the embarrassments of a coherence theory of truth, and the perennial disappointments of idealism.

I do not think the stronger thesis that “belief is in its nature veridical” follows from anti-individualism about thought, perception, and belief. Perhaps the suggestions in some of Davidson’s formulations that he is taking the phrase in that stronger sense are what inclines Burge to believe that Davidson sometimes uses the slogan that “error presup-poses a background of veridicality” with “insuffi cient discrimination”. I do not know whether that is what he has in mind, but I do think it would be a misuse or an insuffi ciently discriminating use of that idea to derive from the slogan what I have been calling the stronger thesis. 21

Anti-individualism is supported by our practices of attributing thoughts and attitudes with determinate contents to people in our world. It is a condition of our ascribing the thoughts we do that we understand them to be related in appropriate ways to the objects or states of affairs in the surrounding world we take them to be about. Without our own knowledge of and engagement with the world we could make no sense of anyone’s having thoughts or experiences of this or that kind at all. Human beings understand the mind anti-individualistically. They can identify the contents of minds only in terms of what they also take to be true of the independent world. But none of that implies that what people believe about the world is in fact true, or that the beliefs they inevitably ascribe to their fellow human beings are in fact for the most part true either.

This is not a sceptical point. I do not mean to be casting doubt on any of our beliefs about the world. The point is only that, however overwhelmingly unlikely, even in some sense unthinkable, it might be that we are largely or even slightly wrong in our beliefs about the world, our not being wrong, or our having for the most part beliefs that are true, does not strictly follow from the conjunction of the

21 I have argued this at greater length in application to Davidson’s theory in Hahn, Lewis E. (ed.). 1999 . The Philosophy of Donald Davidson . Chicago: Open Court. For its connection with the “transcendental” strategy once pursued by P. F. Strawson, see my “Kantian Argument, Conceptual Capacities, and Invulnerability”, in Parrini, P. (ed.). 1994. Kant and Contemporary Epistemology . Dordrecht: Kluwer.


fact, fi rst, that we have a great many beliefs about the world, second, that we understand our fellow human beings to have a great many determinate beliefs with this or that content, and third, that we do or even must ascribe those beliefs on anti-individualist grounds. If those three things are true, no one could consistently fi nd that his fellow human beings believe this, that, and the other about the world, but that those beliefs are all or for the most part false. The conditions of his ascribing those beliefs to them rules out that possibility. But it does not follow that the beliefs in question, or therefore in general any beliefs that any people could be understood as having, are in fact largely true. The stronger thesis about belief and truth is not sup-ported by what supports anti-individualism.

Anyone who thinks that people do have certain determinate beliefs, and who thinks so on anti-individualist grounds, will of course regard those beliefs as for the most part true. She will not see the people she interprets as holding the beliefs she says they have got unless she also sees them as having mostly true beliefs. But that is not surprising, if she sees those others as sharing her beliefs to a large extent. Every believer regards her own beliefs as true. That they are true, even if they are, is not something that follows from that alone. It does follow from what such a person believes—not from her believing it—that the beliefs she shares with those others are true. She believes many things about the world, and she believes that others also believe many of those things she believes. So from the totality of everything she believes it follows that most of those others’ beliefs are true. But that is not a special consequence of anti-individualism. It is a consequence of the fact that from any proposition p, and the proposition that someone believes that p, it follows that that person has a true belief. If you write down everything a belief-attributor believes, it will fol-low from everything you write down that the beliefs she attributes are for the most part true.

There is then a weaker sense in which “most of our beliefs are true” and “error presupposes a background of veridicality” are legitimately supported by what supports anti-individualism. It says: we can under-stand ourselves to have false beliefs or illusory perceptions only if we understand ourselves to have a set of beliefs or perceptions that are largely or for the most part true or veridical. This still leaves unspec-ifi ed the exact scope and quality of possible error in ways I men-tioned earlier. But it retains the idea of a kind of reassurance provided by anti-individualism. To say not merely that we are for the


most part not in error in our perceptual beliefs, but that we are “nearly immune from error” in those beliefs, would then be to say that we cannot consistently fi nd ourselves to have certain determinate perceptual beliefs without also fi nding ourselves not to be largely in error about them. Finding them largely erroneous would be suffi cient for fi nding those beliefs not to be present after all. The idea that “belief is in its nature veridical” can accordingly be taken in the sense that, as we might put it, belief attribution is in its nature truth-ascribing (for the most part), or that attributed beliefs are necessarily regarded as, for the most part, true.

Scepticism is the outcome of an effort to examine our beliefs about the world all at once. The line of thinking that eventually leads to it starts from what looks like the uncontroversial observation that all our beliefs could be false, consistently with our holding them. This so far is only what might be called a logical point. Much more is needed to get to a sceptical conclusion, but this is where the reasoning that typically leads to it starts. We are then challenged to say how we know that that admitted possibility is not actual—that it is not the case that our beliefs are not true even though we all believe that they are. The rest of the reasoning attempts to show, in one way or another depending on the case, why the challenge cannot be met. What I have called the stronger thesis about belief and truth would imply that this very fi rst step is wrong. On that view, there is simply no possibility that a large and reasonably comprehensive set of beliefs is entirely or for the most part false. What the sceptical reasoning would start from is therefore on this view actually a contradiction. There could be no serious question of how we know that that alleged possibility is not actual. It is not a possibility at all, so there is nothing that we have to show is not actual.

I have said that this view, if it were true, would still not refute scep-ticism in the sense of implying its negation. But it would decisively stop the typical sceptical challenge from getting off the ground. But it seems to me to deny too much, and so to be too strong, and too quick. I do not think it follows from anti-individualist premises about belief attribution. I think we must grant the completely general and abstract point that the truth of all or most or even any of a set of beliefs does not follow simply from their being held. Given that they are all believed, it is still in that sense possible for them to be false. That is the logical point. But what the slogan that “error presupposes a background of veridicality,” or the idea that “we are nearly immune


from error” in our perceptual beliefs, do imply, in the sense in which I think they are supported by anti-individualism, is that anyone who understands people to have a set of determinate beliefs must take those beliefs to be for the most part true. She could not say “these are the beliefs they hold, and they are completely or for the most part in error”. So the abstract possibility from which the sceptical reasoning typically starts could pose no serious threat as applied to any particu-lar set of beliefs anyone might consider. Not because it is in no sense a possibility at all that those beliefs are false, but because, even though it is in the weakest sense a possibility, it is not a possibility anyone could ever consistently fi nd to be actual when specifi c, determinate beliefs are under consideration.

The epistemological investigation of human knowledge, of which scepticism is one possible outcome, involves scrutiny and assessment of the particular beliefs we human beings have actually got. The pos-sibility we are asked to consider at the beginning is a possibility with two parts, or conjuncts. It is the possibility that we have all the beliefs we now have in this, that, and the other aspect of the world, and that those beliefs are all or for the most part false. Anti-individualism with respect to our thoughts about and attributions of beliefs means that we could not consistently fi nd the fi rst half of the possibility to be realized without fi nding its second half not to be so. And if we found the second conjunct to be true, we could not consistently fi nd the fi rst half of the possibility to be realized. Attribution of the beliefs we attribute requires fi nding them for the most part true; fi nd-ing a certain set of propositions for the most part false rules out assigning them as contents of the beliefs of people with whom one shares a common world.

Our position with respect to that original possibility is therefore similar to a person’s relation to the apparently paradoxical sentence ‘I believe that it is raining, and it is not raining’. That is not some-thing one could consistently believe or assert—but not because what it says is something that could not possibly be true. It is possible that I believe that it is raining when it is not raining. That is a possibility with two parts, or conjuncts. The fi rst does not imply that the second is false, and the second does not imply that the fi rst is false. It is in that sense a genuine possibility. But no one can consistently hold in his own case that that possibility is actual, that both conjuncts are true. This brings out the difference between something’s being simply inconsistent or impossible (which my believing that it’s raining, and


its not raining, is not) and something’s being impossible for anyone consistently to believe or discover. I am suggesting that this distinc-tion applies to the possibility that human beings have all the beliefs they now have, and that those beliefs are all or for the most part false. It is the impossibility of consistently fi nding or believing that possi-bility to be actual that I think follows from anti-individualism about our thoughts, perceptions, and beliefs.

If that is so, then we cannot take seriously the possibility that the beliefs we take ourselves to have are for the most part not true. There is no serious challenge to our beliefs about the world expressed in the question “Given that the truth of your beliefs and the veridicality of your perceptions do not follow simply from your having them, how do you know that, taken all together, they are not all or for the most part false or illusory?”. That is the question that the traditional epis-temological challenge presses. Failure to answer it satisfactorily at any point leaves us eventually with scepticism. Applied to a particu-lar case, we fi nd in various ways that we cannot appeal to anything we already believe or to any perceptions we might get or seek in order to answer the question, since at every step all of our beliefs and per-ceptions of the world are in question all at once. But given an anti-individualist understanding of beliefs and perceptions, for any specifi c set of beliefs and perceptions we take ourselves to have, the question cannot really present the serious challenge it might other-wise appear to lead to. Not because the challenge can be met by establishing independently—somehow without relying on any per-ceptions or beliefs—that our perceptions and beliefs are veridical or true and amount to knowledge of the world. Not because we could not possibly fail to be largely right in whatever set of perceptions and beliefs we happened to have. But because our considering the par-ticular set of perceptions and beliefs that we are asking about in itself guarantees that we fi nd them to be for the most part true.

Does it follow from that that we know those beliefs are true? No, it does not. Does it follow from that that the beliefs in question are even largely true, whether we know it or not? Again, I think it does not. We must always admit that our beliefs could be false. Asked right now about the things we believe, we will of course say they are true. Asked whether we know them to be true, we will say, of many of them, per-haps on refl ection, that Yes, we do know them to be true. But we are fallible human beings with beliefs about an independent world, and it is always an open question whether what we think we know is really


something that we know, or is even something true. What I have described as the reassurance derived from anti-individualism does not confl ict with that. Although I cannot consistently believe both that I believe that it is raining, and that it is not raining, I can still ask myself whether it is raining or not, or whether I am right in believing that it is raining. So, in the more general case, I can still ask, “Are those beliefs that I think people have true, or for the most part true, or not?”. And just as I answer the fi rst question by going outside and fi nding out whether it is raining, so the way to answer the more gen-eral question is to put myself in the best position I can for fi nding out whether the things I and others believe about the world are true. If I have not already done that, or even if I have, I might fi nd on further investigation that some or even many of the things I believe are not true. I will then abandon those beliefs. But I might equally fi nd that all or a great many of them are indeed true, and in fact so well sup-ported that I do not hesitate to say that I know they are true. The open question has been, to the best of my knowledge and at least for now, answered.

If I answered the question in that way, would philosophical scepti-cism about our knowledge of the world have been refuted? I would say “No”. Would what I say nevertheless confl ict with scepticism, even if it does not refute it? This question is harder to answer. Would we still face a challenge to our alleged knowledge starting from the thought that all or most of our beliefs and perceptions could be false, and so demanding some reason to think that that possibility is not actual? Here I am inclined to answer, on anti-individualist grounds, “No, that is not a possibility we can take seriously; it is not a way we could fi nd things to be.” Even if that is right, and reassuring, it does not amount to the negation of philosophical scepticism. But still I say, for the very best reasons we can fi nd, that we know a great many things about the world around us.

Philosophical scepticism is a negative outcome of an effort to assess all of our knowledge of the world all at once, and perforce from a position in some sense outside it. The denial or negation of that assessment, although perhaps reassuring if reachable, would presumably have to be made from that same outside-of-all-of-our-knowledge-of-the-world position. That is why I think we should not aspire to deny the thesis of scepticism, any more than we want to accept it. The great promise of anti-individualism as I understand it is that it would reveal how and why it is impossible for us even to get


into that position with respect to any comprehensive set of beliefs or experiences we can recognize ourselves to have. We understand our-selves to have beliefs only by for the most part endorsing them. We could then consistently arrive at neither the assertion nor the denial of philosophical scepticism. If we could accept the fact that our beliefs about the world are none the worse for all that, perhaps fi nal satisfaction would be at hand.

14

Sense-Experience and the Grounding of Thought

John McDowell’s Mind and World seeks to expose and thereby to exorcise a line of thinking that leads to a philosophical challenge to the possibility of thought. Theories meant to answer the challenge cannot succeed; the defects of each drive us to an opposed theory, and then back, in endless oscillation. The key to the exorcism is to see how and why we are not forced to choose among the unsatisfactory theories. To be liberated in this way would not be to answer the philo-sophical question. Ideally, the challenge will simply disappear, or will lose whatever point it had seemed to have.

The troubling refl ections start from the idea that thought has a sub-ject matter, or is about something, only if certain conditions are ful-fi lled. One such requirement is found in the idea of truth. A belief or judgement to the effect that things are thus and so is correctly or incor-rectly held according to whether or not things are thus and so. Thinking is in that way “answerable” to the way things are – and so to something independent of thought – for its being correctly or incorrectly execut-ed. 1 But it is the “cognitive predicament” 2 of human beings to confront the world by means of sense-experience, so McDowell holds that refl ection on thought’s answerability to the world must begin for us with an understanding of our access to “the empirical world”. 3 This is where we can easily fall into the quandary he wants to show how to avoid. For “how can we understand the idea that our thinking is answerable to the empirical world, if not by way of the idea that our thinking is answerable to experience”? 4 But to be answerable to experi-ence our thought must be subject to a “verdict” from experience; some-thing that indicates whether our beliefs or judgements have been

1 Mind and World , Cambridge, Mass., 1996, p. xii. 2 Mind and World , p. xii. 3 Mind and World , p. xii. 4 Mind and World , p. xii.


adopted correctly or incorrectly, or at least reasonably or unreaso nably. And there are obstacles in the way of seeing how such a verdict from “the tribunal of experience” 5 is even possible.

Human sense-experience is a matter of impingements by the world on beings who possess certain sensory capacities. Such impingements are events in nature; something that is impinged upon is affected in a certain way. That is a causal connection, describable in terms of sci-entifi c laws of nature. But to understand two things as connected in accordance with a law of nature is not to understand one of those things as making the other reasonable in any way, or as justifying or supporting it. It is to see that one thing happens because something else happens, but not to see that one thing is correctly or incorrectly executed, or is warranted or justifi ed in the light of something else. Understanding sensory impacts in this way makes it impossible to see them as constituting a tribunal “standing in judgement over our beliefs”. 6 The idea that mere impingements on our sensory surfaces can “ground” our beliefs and make them reasonable is nothing more than a myth: The Myth of the Given. But with no experiential “grounding”, thought will not be answerable to anything independ-ent of thought, and so will degenerate into a series of “moves in a self-contained game”, 7 “a frictionless spinning in a void”. 8 We are left with a familiar kind of philosophical anxiety: how is thought about the world possible at all?

McDowell’s way of avoiding this quandary is to explain how and why there is no tension between sense impressions’ being occurrences in nature and their grounding one’s beliefs. The liberation comes from seeing how certain happenings in nature can nonetheless sup-port or make reasonable beliefs about the way things are. Mature human beings possess concepts which they employ in thinking about the world. Impacts by the world on the receptive capacities of such beings are not mere impacts describable only in terms of scientifi c laws of nature. They are typically cases of a person’s perceiving that things are thus and so in the world. A suitably equipped subject “takes in, for instance, sees, that things are thus and so ”. 9 That things are thus and so is also something one can think, or judge, or believe.

5 Mind and World , p. xii. 6 Mind and World , p. xiii. 7 Mind and World , p. 5. 8 Mind and World , p. 11. 9 Mind and World , p. 9.

Sense-Experience and the Grounding of Thought 275

Having an experience of seeing that things are thus and so can there-fore make it highly reasonable (to put it mildly) to believe that things are thus and so. That belief gets a very favourable verdict from that experience.

Of course, one can be misled into thinking that one takes in that things are thus and so when they are not. Human beings are not infallible, even in perception. But the key to McDowell’s exorcism is to reveal how, when one is not misled, one perceives how things are. That is why, as he points out, it does not matter much that one can be misled. Actually being misled can matter a great deal; it can even be fatal. But the general possibility of being misled is no obstacle to an understanding of the possibility of thought. The perceptual access one has to the way things are when one is not misled is what supports or warrants one’s belief to the effect that things are that way. Not only has one’s thought been correctly executed; one also has good reason to believe that it has. When sense-experience is understood as “openness to the layout of reality” 10 in this way we can see how an independent reality can “exert a rational infl uence on what a subject thinks”. 11

This insistence that in sense-experience we can take in the way things are, or that things are thus and so, is obviously crucial to McDowell’s way of bypassing the philosophical quandary. He is admirably clear and forthright about its importance, and explictly sets himself against the widespread tendency “to conclude that even a non-misleading experience cannot genuinely be a case of openness to reality”. 12 If, even when things go as well as they can, there is a dif-ference, and hence a gap, between what we perceive to be so and what is so, we could never have “the fact itself impressing itself on a perceiver”. 13 And for McDowell that “strains our hold on the very idea of a glimpse of reality”. 14 It would leave inexplicable how think-ing can be answerable to anything independent of itself.

McDowell is surely right to insist that in experience we can see that things are thus and so. The history of philosophy (if nothing else) should have convinced us by now that no conception of perception which offers us at best only something less than what is so in the

10 Mind and World , p. 26. 11 Mind and World , p. 26. 12 Mind and World , p. 111. 13 Mind and World , pp. 112–113. 14 Mind and World , p. 112.


independent world could explain how we can have any reason to believe anything about such a world. Most of Mind and World is taken up with the diagnostic task of explaining how and why such a conception of experience has come to seem unavailable, and with doing what is needed to bring it clearly and convincingly into view. It is not easy, as McDowell is fully aware; a trouble-free conception of sense-experience has certainly eluded a great many impressive philosophers.

I will not take up McDowell’s rich exploration of the ground that must be cleared in order to put the otherwise inevitable quandary behind us. I fi nd his account extremely illuminating, and promising of even greater riches. Rather I want to ruminate on the conception of sense-experience that will ensure what he is right to insist is needed – perceptual access to the way things are. Does even McDowell avoid all danger in this traditional minefi eld?

To suppose that mere impacts from the world, conceived as such, could justify or make reasonable any beliefs they might happen to generate is to fall victim to the Myth of the Given. McDowell accord-ingly stresses that we must start from the idea that “the world’s impressions on our senses are already possessed of conceptual content”. 15 So what is needed is “a notion of experience as an actual-ization of conceptual capacities in sensory consciousness”. 16 This means that anyone who receives sense impressions has certain con-ceptual capacities. Where there are no such capacities, there is no experience of the way things are. Creatures with sensory capacities but without conceptual capacities of the kinds human beings so con-spicuously possess do not really have experience as McDowell con-ceives of it. They might be perceptually sensitive to various features of their environment, but they do not have experience of the world.

To have the requisite conceptual capacities is to be capable of activities of thinking which are “responsibly undertaken by a subject who is in control of the activity”, 17 and so “is responsive to rational relations, which link the contents of judgements of experience with other judgeable contents”. 18 In perception, what one perceives to be

15 Mind and World , p. 18. 16 “Precis of Mind and World ”, Philosophy and Phenomenological Research 1998,

p. 367. 17 “Precis of Mind and World ”, Philosophy and Phenomenological Research 1998,

p. 367. 18 Mind and World , p. 12.


so is, after a certain point, not directly under one’s control. But the capacities brought into play in passive perceptual experience must be available to the subject in fully active, and so non-perceptual, think-ing. Impacts by the world on the sensory capacities of a person with such conceptual capacities are sense impressions with conceptual content.

That things are thus and so is the conceptual content of an experience, but if the subject of the experience is not misled, that very same thing, that things are thus and so , is also a perceptible fact, an aspect of the perceptible world. 19

This looks like just the conclusion we want. That conclusion would seem to be secured by the idea that having

a sense-experience or impression, at least when things go well, is sim-ply a matter of seeing or otherwise perceiving that things are thus and so. The experience of seeing that something or other is so cer-tainly involves receiving an impact from the world on the senses. And no one could see that p without understanding the thought that p , and so having the conceptual capacities required for understanding that thought. So seeing that p is “an actualization of conceptual capacities in sensory consciousness”. Impressions for McDowell are also what can give us reasons to believe something about the empi-rical world. And seeing that p can certainly give a person reason to believe that p . The justifi cation of the person’s belief can be traced back to that person’s experience of seeing that p . If that experience is an impression, our beliefs would be answerable to our impressions, and so answerable to experience, by our sometimes seeing or other-wise perceiving that things are thus and so.

This is what McDowell wants his conception of experience to deliver, but he says many things about impressions which suggest that he does not regard them as simply cases of seeing that things are thus and so in the independent world. Or if that is how he thinks of them, he does not think it is enough to explain in the right way how they ground empirical beliefs. I would like to understand why not.

One indication that he thinks something is missing is his dissatis-faction with the views of Donald Davidson, whom he describes as not merely committed to the idea that “experience cannot count as a reason for holding a belief”, 20 but even as suffering from “a block

19 Mind and World , p. 26. 20 Mind and World , p. 14.


that prevents [him] from seeing any possibilities in this direction”. 21 But there seems to be nothing to prevent Davidson from accepting a view of impressions as cases of seeing that p . He has famously argued that a certain physical movement’s occurring in a suitably-equipped person can be a case of that person’s acting with a certain intention 22 , and that the utterance of a certain sound by a person equipped with certain conceptual capacities can be a case of that person’s express-ing a certain belief. 23 It is hard to see why he could not also accept the idea that the world’s having a certain impact on a person adept in the use of the relevant concepts can be a case of that person’s seeing that it is raining. And how could Davidson, or anyone, deny that seeing that it is raining can give one good reason to believe that it is raining? Looking and seeing what is going on is the best way to get a reason-able belief about the weather.

It is true that some things Davidson says appear to confl ict with the idea that sense-experience can give one reason to believe something. He speaks of sense impressions or sensations as causes of certain effects in perceivers, and denies that as such they can pro-vide reasons for believing something. In a sentence McDowell puts a lot of weight on, Davidson says “Nothing can count as a reason for holding a belief except another belief”. 24 The sentence is unfor-tunate, fi rst, in its apparent implication that one can have a reason for holding a belief only if one also holds a different belief. That would mean that seeing that it is raining could not be one’s reason for believing that it is raining, since there is not “another belief” involved—only the proposition that it is raining. But since what matters for the reasonableness of a belief is not the proposition believed but the reasonableness of the person’s believing it, it is pos-sible for one’s having a certain attitude towards a proposition (e.g., seeing that p ) to justify or give one reason to hold a different attitude towards that same proposition (e.g., believing that p ).

Putting too much weight on the unfortunate sentence also obscures Davidson’s reason for denying that sense impressions or sensations can give one reasons to believe something: sense impressions understood as mere impacts from the world have no content. That is

21 Mind and World , p. 18. 22 See, e.g., “Agency”, in his Essays on Actions and Events , Oxford, 1980. 23 See, e.g, “Radical Interpretation”, in his Inquiries into Truth and Interpretation ,

Oxford, 1984. 24 Quoted by McDowell in Mind and World , p. 14.


something McDowell himself insists on. He says he has no objection to what Davidson expresses in the unfortunate sentence if it means “nothing can count as a reason for holding a belief except some-thing else that is also in the space of concepts”. 25 But if that is what it means, it does not confl ict with the view that seeing that p , which is receiving an impact from the world by a person possessing the relevant concepts, can give one reason to believe that p .

McDowell nonetheless thinks Davidson is not in a position to explain in the right way how experience can give one reason to believe things about the world. One thing he fi nds lacking is any account of that “external constraint” that is essential for thought’s having any bearing on an independent reality. We must acknowledge a role in perception for what McDowell calls “receptivity”, as well as for those essential conceptual capacities that come from the side of “spontan-eity”. But in understanding that “external constraint” “we must not suppose that receptivity makes an even notionally separable contri-bution to its co-operation with spontaneity”. 26 To suppose that it does, or could, leads directly to the Myth of the Given, and hence to the fatal oscillation.

The idea of an impression as a seeing that p appears to fulfi ll this condition. Seeing that p occurs only when there is an impact by the world on a perceiver equipped with conceptual capacities suffi cient for grasping the thought that p . Only then is there “receptivity” in perception. An impact on an organism that lacks those capacities, or in whom they are inoperative, is not a case of seeing that p ; it is in that respect a different kind of impact. But the difference between the two kinds of impact lies only in the absence or presence of those conceptual capacities that are essential to seeing that p . So receptivity does not make a contribution to seeing that p that is “even notionally separable” from the contribution of the conceptual capacities char-acteristic of spontaneity. It is nonetheless required for seeing that p , and so required for such an experience to give one reason to believe that p .

As McDowell thinks of reasons to believe, an impression gives one reason to believe something only if it is, or provides, the person’s reason for believing what he does. It is not enough for an experience simply to render a person’s belief reasonable, even to a very high

25 Mind and World , pp. 140, 143. 26 Mind and World , p. 51.


degree. This is the point behind his criticism of Gareth Evans’ and Christopher Peacocke’s appeal to states or experiences with what they call “non-conceptual content” that is not “available” to the subject at the time. 27 Even if something square is present, for exam-ple, whenever a person is in a certain sensory state and believes that something square is present, it does not follow that that sensory state is the person’s reason for holding that belief. 28 Since the belief would be true whenever it was held in those circumstances, it could perhaps be said to be highly reasonable for anyone in that state to hold it then. But for McDowell that belief would not be “based on” the sensory state in the required way. The so-called sensory experi-ence that gave rise to it would fall beyond the reach of the concep-tual capacities involved in spontaneity, and in that sense would be “blind”. 29 It would not even count as an experience. Experience involves awareness, and that requires possession and deployment of conceptual capacities that can be exercised in “active, self-conscious thinking”. 30

The idea of impressions as seeings that p also appears to fulfi ll this requirement. No one could have an experience of seeing that p without the conceptual resources for understanding the thought that p . The experience therefore has “conceptual content”; it is not “blind”. Someone who sees that p is aware that p , or “takes in” that p . The content of that experience is therefore “available” to the person to be rationally linked with other “judgeable contents”. Having that experi-ence could be said to be the person’s reason, or what gives the person reason, to believe what he does.

But I think that for McDowell this still does not explain how what he calls an impression can count as a person’s reason for holding a belief about the world. Appealing only to seeings that p , even if we call them experiences, does not do justice to that cooperation of receptivity and spontaneity that he sees as essential to experience. A person who sees that it is raining judges or believes or accepts or otherwise puts it forward as true that it is raining. That judgement or belief or assertion is a judgement about the independent world, so for McDowell its justifi cation must eventually be traced back to an

27 Mind and World , pp. 50–55, 162–166. 28 Mind and World , pp. 163–164. 29 Mind and World , p. 54. 30 Mind and World , p. 55.


impression that grounds it. “When we trace the ground for an empir-ical judgement, the last step takes us to experiences.” 31 But tracing the justifi cation back only to what I have called the experience of see-ing that it is raining would trace it back only to something that still involves judgement or belief about the independent world. To see that p is to judge that p . And for McDowell that judgement, like all empirical judgements, will be justifi ed only if it is grounded in an experience which is an impression.

In receiving a sense impression, the perceiver is passive. It is true that one receives an impression with content only if the conceptual capacities that are present and operative in it are “integrated into a rationally organized network of capacities for active adjustment of one’s thinking to the deliverances of experience”. 32 But those “deliv-erances of experience” are not themselves results of the “active, self-conscious thinking” characteristic of exercises of what McDowell calls spontaneity. That is a form of “freedom”, and requires assess-ment, decision, and judgement. But “in experience, one fi nds oneself saddled with content. One’s conceptual capacities have already been brought into play, in the content’s being available to one, before one has any choice in the matter”. 33

This is perhaps what leads McDowell to speak of impressions as “appearances” or “appearings”. Having a sense-experience is “hav-ing things appear to one in a certain way”. 34 Those appearances are “rationally linked” to the rest of our thought through “the way appearances can constitute reasons for judgements about objective reality – indeed, do constitute reasons for judgements in suitable circumstances”. 35 In such circumstances, an empirical judgement is “satisfactorily grounded in how things appear to one”. 36

Speaking of impressions as “appearances”, and distinguishing them from the judgements one makes on the basis of them, is per-haps a way of drawing attention to the passivity of perception as contrasted with the “freedom” that is essential to “active think-ing”. But how is that contrast to be drawn? McDowell expresses it this way:

31 Mind and World , p. 10. 32 Mind and World , p. 29. 33 Mind and World , p. 10. 34 Mind and World , p. 62. 35 Mind and World , p. 62. 36 Mind and World , p. 62.


Minimally, it must be possible to decide whether or not to judge that things are as one’s experience represents them to be. How one’s experience repre-sents things to be is not under one’s control, but it is up to one whether one accepts the appearance or rejects it. 37

This strongly suggests that whenever one receives an impression one is presented with an “appearance” that “represents” things to be a certain way, but one can decide not to accept or judge that things are in fact that way. Withholding one’s assent from the “appearance” would not be to see that things are that way. Seeing that p involves judging that p , but having an impression that merely “represents” things to be such that p does not. So although in seeing that p there would be an experience that grounds the judgement that p , seeing that p would not itself be that grounding experience. On this view, having an impression would be part of, but it would not be the same as, seeing that p . Impressions would be restricted to “appearances” which involve acceptance or endorsement of nothing beyond “how one’s experience represents things to be”, and so nothing about the independent world.

Do we always have an impression of something less than that things are thus and so every time we see that things are thus and so? One reason to think so might be the possibility of the empirical judgement’s being false even when made on the basis of an impres-sion. An object can fail to be square even though a very sharp-sighted person who looks at it carefully judges that it is square. In such a case, the person does not see that the object is square, since it is not square, but she certainly sees something, or has some kind of impres-sion. We know from the tortuous history of the philosophy of per-ception that it is very tempting to conclude that a perceiver must have the same kind of impression when the object is in fact square and when it is not but only looks that way. The content of the impression in either case implies nothing about the independent world. The per-ceiver judges on the basis of that impression, either truly or falsely, that the object is square.

This cannot be how McDowell thinks of impressions as “appear-ances”. This view really would rule out the possibility of “the fact itself impressing itself on a perceiver”. 38 I think it is fatal even to the prospect of any reasonable belief about the world. But McDowell

37 Mind and World , p. 11. 38 Mind and World , pp. 112–113.


has convincingly shown that it can be avoided. The general possibil-ity of an empirical judgement’s being false does not imply that when-ever a person has an impression that grounds that judgement it is an impression that is non-committal about the judgement’s truth. 39 So there is no reason here for thinking that impressions are always of something less than the truth of the empirical judgements we make on the basis of them.

Another reason for thinking of “appearances” as non-committal in that way might be the thought that in the passive receptivity of perception alone we merely “receive”, and do not make or accept any judgement at all. As an illustration of the way an experience can ground a belief McDowell says that an ordinary person asked why she holds a belief to the effect that an object within her fi eld of view is square could give the unsurprising reply “Because it looks that way”. He says that gives a reason for holding the belief. 40 Now when she says “Because it looks square” she is presumably express-ing the belief that it looks square (to her). And the content of that belief, together with an assumption about conditions’ being normal – or at least the absence of information to the contrary – could be given as a reason for believing that the object (probably) is square. But that traces the justifi cation of the original belief back only to a belief or judgement about how things look. For McDowell there must be an experience which is the believer’s reason, not just a judgement or belief.

Perhaps in saying “Because it looks square” the person is refer-ring to such an experience. Its looking that way to her is certainly an experience she might have. But if she has such an experience, it must be true that that object looks square to her, and if that is the way it looks to her, she must be aware that that object looks square to her. Experience involves awareness. But for her to be aware that that object looks square to her is for her to accept or endorse as true the proposition ‘That object looks square (to me)’. She is aware that that proposition is true. So it cannot be that in receiving an impression that is an appearance there is no acceptance or judge-ment at all.

39 See e.g., Mind and World , pp. 112–113 and references there to his “Criteria, Defeasibility, and Knowledge” and “Singular Thought and the Extent of Inner Space”.

40 Mind and World , p. 165.


McDowell might seem to be denying this when he says “Its appear-ing to me that things are thus and so is not obviously to be equated with my believing something”. 41 But that denies only that sensory appearings and believings are the same thing. He is right that they are not; it is possible to believe something without anything appear-ing to me in perception at all. The question is whether things could perceptually appear to me to be a certain way without my believing anything at all. McDowell says that in sense-experience “one fi nds oneself saddled with content . . . [which is] available to one, before one has any choice in the matter”. 42 But to be “saddled” with a certain content in perception is not simply for that content to be “available” to be entertained or contemplated, as it is in the unasserted ante-cedent of a conditional proposition, for instance. To take in some content in perception is to have accepted or endorsed that content, or to fi nd oneself accepting or endorsing it.

‘Believe’ is perhaps not the best word to capture the attitude of acceptance or endorsement involved in perception, especially if it suggests actively making up one’s mind. 43 The word ‘judgement’ seems even less apt, with its connotation of deliberation and deci-sion. But there is awareness or acceptance of some kind in perceptual experience, not mere entertainment of content. Even to be aware that things appear to be thus and so is to take in or acknowledge that they appear to be that way. It is not to remain neutral on the question of how things appear to you to be.

So the passivity and absence of choice in perception, which is the mark of receptivity and “external constraint”, cannot be equated with the absence of all judgement or assent. In being “saddled” with content one is “saddled” with assent to or affi rmation of that con-tent, or at least of some content or other. When the content of one’s impression is things’ appearing to be thus and so, it looks as if the impression that grounds the judgement involved in accepting that content must be the very impression in the receiving of which that judgement is made. There is nothing else that could serve as its ground. To insist that its ground must be found elsewhere, because a judgement can be adequately grounded only if it is based on an impression that does not involve making that very judgement, would

41 Mind and World , p. 140. 42 Mind and World , p. 10. 43 See Mind and World , p. 60.


lead to a sceptical regress; no judgement would be grounded. But if an experience involving a judgement can ground that very judgement in the case of minimally-committal “appearances”, what reason is there to deny that an empirical judgement about the independent world can be grounded by an experience that involves making that very judgement?

Someone who is asked why she believes that a certain object in her view is square might also give the unsurprising reply “Because I see that it is square”. That is an even better reason for her to believe that it is square than that it looks square to her. It involves judging or asserting that the object is square. That judgement is part of the experience of seeing that the object is square. Must its justifi cation be traced back to an “appearance” whose content implies nothing about the way things are in the independent world?

There is no question that we sometimes — even fairly often – have impressions that are minimally committal in that way. If I know or believe that two lines in a familiar confi guration are the same length, although they now appear to me not to be, I will not judge or believe that those lines are unequal. I do not go beyond the appearance in what I accept. It is in that sense “up to me” to judge that the lines are unequal when I see what I see, and in such a case I do not make that judgement. But when I do judge that two lines that I see are unequal, that judgement is also “up to me” in the same sense, even if it is something I cannot help believing at the time, and something I never reject. Sticking by it or not is “up to me”; the judgement is revisable, even if never revised. If I come to see or suspect that it is not true, or that I do not see that it is true, I can settle for its appearing to me that the lines are unequal. Or if I stick by the original judgement and it is, unknown to me, false, then I did not see that the lines are unequal, even though I thought I did. Since in that case I had an impression with some content or other – but not that the lines are unequal – it could again be said that I had an experience of the lines appearing to me to be unequal. But none of this shows that I have such an experi-ence every time I see that two lines are unequal in length.

The fact that judgements made in perceptual experience are revis-able in the light of refl ection or further experience can make it seem that continued acceptance of a perceptual judgement about the inde-pendent world cannot be based solely on the original experience that gave rise to it; further refl ection or experience serve to support it as well. And this can make it tempting to suppose that the content of


that original experience in itself must have implied nothing about the independent world, and its acceptance was direct and unrevisable. But yielding to this temptation is another way of falling into that “highest common factor” view of perception and reasonable belief that McDowell has rightly rejected.

That view also assumes that what is accepted in “appearances” is known or endorsed in some direct way, and is unrevisable. But that is not so. One’s acceptance of how things appear to be is also revisable, just as one’s beliefs about what is so are. Closer scrutiny or further refl ection can reveal to me that two lines that I see do not really appear to be the same length, although I thought earlier that they did. Finding them to appear equal or not is “up to me” in the same sense as fi nding them to be equal or not is. As far as I can see, to acknowledge this revisability of perceptual judgements is not to deny the presence of an “external constraint” on thought or the receptivity involved in perception.

So although I fi nd McDowell certainly right to stress that sense impressions have content, because of the conceptual capacities of the perceiver, and right to see those impressions as giving the perceiver reason to believe something, I am less than certain of how he thinks of those impressions, and of how he thinks they provide such rea-sons. If all impressions are “appearances”, but even seeing that things are thus and so in the independent world counts as receving an “appearance”, there is no diffi culty. We can get reason to believe that p by seeing that p . And if all impressions are “appearances” whose contents always fall short of implying anything about the independ-ent world, the “openeness to reality” that McDowell rightly demands of a satisfactory account of experience is lost. No one could ever see that things are thus and so; the most anyone could be aware of in perception alone would be things’ appearing to be thus and so. “The fact itself” would never be impressed on a perceiver. That threatens even the possibility of reasonable belief about an independent world. How could “appearances”, so understood, ever give one reason to believe anything about what is not an “appearance”?

McDowell nonetheless puts his problem of how experience can give rational support to empirical judgements as that of accounting for “the way appearances can constitute reasons for judgements about objective reality”. 44 His answer appeals to “rational relations”

44 Mind and World , p. 62.


between experiences and the judgements about the world that we make on the basis of them, 45 and he appears to think of those rela-tions as holding between the contents of the experiences and the beliefs. 46 “Because it looks square”, he says, is “easily recognized” as giving a reason to believe “It is square”; 47 it is the person’s reason for believing that the object is square. But is it recognized as a reason because we discern a certain “rational linkage” between the content of the experience and the content of the belief ?

I think the content of an experience alone cannot give a person reason, or be a person’s reason, to believe something. The content of an experience is typically expressed in a proposition, and propos-itions are not reasons, nor do they make other propositions reason-able. Propositions are true or false, not reasonable or unreasonable, justifi ed or unjustifi ed. Even if one proposition implies another, it does not justify, support, warrant, or make reasonable that other proposition. What is justifi ed or reasonable or supported or war-ranted is a person’s accepting a certain proposition, or rejecting it, or taking some other attitude towards its truth.

It is true that when asked for his reason for believing something a person can simply cite a proposition. A detective who is asked “Why do you believe this suspect didn’t do it?” could answer “Because he was in Cleveland that night”. But in giving that as her reason the speaker indicates that the proposition she cites is something she believes. Her believing that proposition – in fact, her believing it for good reason – is required for her having that reason to believe that the suspect didn’t do it. If she had no idea that the suspect was in Cleveland, even if it were true, she would not have the reason she has for believing that he didn’t do it. And even if it were not true, she would still have good reason for believing he didn’t do it if she had good reason to believe that the suspect was in Cleveland.

The same holds for experiences. It is not simply the content of a person’s experience that gives him reason to believe something; it is the person’s experiencing, or being aware of, or accepting, or somehow “taking in” that content. Taking in that a certain object looks square can give you reason to believe that the object is square. That cannot be because “It looks square” is always reason to believe “It is square”.

45 Mind and World , p. 52. 46 See, e.g., Mind and World , p. 166. 47 Mind and World , p. 165.


If you already know that an object is not square, the experience of its looking square does not give you reason to believe that it is. In the right circumstances, an object’s looking square could even be excel-lent reason to believe that it is spherical. Whether the way something looks is a reason to believe that it is that way or not depends on what else is true in the situation, and on what else you have reason to believe.

Can we say, then, that there is a “rational relation” holding for the most part, or other things being equal, between “It looks square” and “It is square”? It is probably true that, for the most part, or normally, anyone to whom an object looks square has reason to believe that the object is square. That is an observation about how often, or under what circumstances, having a certain perceptual experience gives a person reason to hold a certain belief, or under what circumstances having a certain attitude gives one reason to hold a different attitude. Some such connections hold in all circumstances. Seeing that an object is square always gives one reason to believe that it is square.

For present purposes the crucial question about any “rational rela-tions” between “It looks F” and “It is F”, however they are under-stood, is whether they also run in the opposite direction. Does seeing that an object is square require having an impression of its looking square? Must the justifi cation of all empirical judgements be traced back to “appearances” understood as minimally committal in that way? I fi nd I do not know McDowell’s answer to this question. I can-not see how or why he could possibly say “Yes”. Yet he seems reluc-tant to hold that seeing that p , which involves judging that p , is the kind of experience that could count in the right way as grounding a person’s belief that p . He insists that thought will have a bearing on reality only if it is ultimately grounded in something “outside thought”, in the sense of outside thinking or judging. But why can seeing that p not serve as that ground? It is something “outside”, or beyond, think-ing. It is seeing , and not merely thinking or judging, that p .

More recently McDowell has invoked the chicken sexers of philo-sophical legend to illustrate what he regards as a condition any satis-factory account of perceptual experience must meet. 48 As he describes them, those experts declare a chick with which they are presented ‘Male’ or ‘Female’ without knowing how they do it, and even without

48 See “In Conversation: Donald Davidson: The McDowell Discussion”, Philosophy International, London, 1997 (videotape).


claiming that the two sexes look different to them. The best chicken sexers get it right most of the time; truth is what counts. What he fi nds unacceptable is the idea that this could be the way things are in general, for all perceptual experience and all concepts. Even those chicken sexers, he thinks, will presumably acknowledge that they rec-ognize something as a chick because it looks like a chick.

It is no doubt implausible to suppose that speakers know nothing about what features they rely on in applying any of their concepts to anything. The richness and variety of the concepts one would have to have in order to have any concepts at all are probably enough to rule that out. But McDowell appears to require that a satisfactory account of visual perceptual experience should show how the way things look to us is what we rely on in judging, on the basis of experience, how things are. Taken with complete generality, for all concepts applied in experience, that would lead to a regress; no concepts at all could be applied in experience. But if there is no regress because it is conceded that the concept “looks F” can be directly recognized to apply, and not on the basis of its looking as if it looks F, why must the concept “is F” be thought to be justifi ably applicable only on some such indi-rect basis? A perceiver often has no reason for applying a certain concept beyond the fact that he sees that it applies to the item to which he applies it. “How do I know that this colour is red? – It would be an answer to say: “I have learnt English”.” 49

49 L. Wittgenstein, Philosophical Investigations , (tr. G. E. M. Anscombe), Oxford, 1953, §381 .

15

The ‘Unity of Cognition’ and the Explanation of Mathematical Knowledge

Alvin Goldman’s work is central to epistemology as we know it. It is more than central. It has helped determine the course the subject has taken for the last thirty-fi ve years or so. And epistemology has been central to philosophy during that period. It is true that epis-temology has broadened out a little over the years—as have many epistemologists. From a more-or-less exclusive concern with per-ception and knowledge in the middle years of the twentieth century it has turned more recently to wider issues of thought, belief, cogni-tion, and experience. It could even be said to have changed its name, being referred to nowadays more often as something like “philoso-phy of mind.” But what remains at the heart of the subject, what-ever it is called, is the attempt to understand fundamental aspects of any possible human life. And Alvin Goldman’s work is at the center of that subject. It is partly formative of the philosophy we have all been doing.

Questions of how we know, perhaps even how we know anything, become crucial at times of innovation or revolution within human knowledge itself. In periods of apparent stability and consensus phil-osophical interests turn elsewhere. When epistemology was central to philosophy in the twentieth century it took over more or less without question the elaborate conception of ideas, impressions, or the imme-diate given of sense-experience that earlier philosophers had resorted to in response to the new science of the seventeenth century.

Historically speaking, this is a very odd feature of twentieth-century epistemology. Given the great innovations or revolutions in human knowledge in the early years of the century, it is not surprising that epistemology once again became central. But those breakthroughs were largely in mathematical physics and in the nature of mathematics itself. The theory of relativity, the development of mathematical logic, quantum theory, the incompleteness of arithmetic—all this can hardly

The ‘Unity of Cognition’ 291

be said to have given special urgency to the question whether sense-data exist when they are not perceived, or whether they have backsides as well as the sides we see.

What is striking about twentieth-century epistemology is not only its concentration on problems of perception. One can also be struck by what I think is the poverty of its treatment of the question one would have thought would get the most attention: knowledge of mathematics and of necessary truth in general.

Throughout the century the dominant idea was that mathematical or necessary knowledge is knowledge only of what is somehow implicit in our ways of understanding the world, and not of the world itself. There was a long-standing, vaguely expressed feeling that it is all to be explained somehow as a matter of “meaning” or “concepts” or “analyticity.” But it can hardly be said that that consensus about the “emptiness” or “lack of factual content” of mathematical or ne cessary truth ever added up to a satisfactory explanation of how knowledge of such truths is possible.

That was partly because of a related consensus that the only proper task of philosophy is something called “analysis.” The idea was that philosophers interested in knowledge should be interested only in the analysis or defi nition of the concept of knowledge and related con-cepts. But the search for the necessary and suffi cient conditions of knowledge tended to focus on simple cases of perceptual knowledge, where the question was thought to be easiest, and so could be answered fi rst. I think that assumption was unwarranted. It has been one of the main obstacles to progress in understanding mathematical knowledge. What were taken to be the simplest cases are much more complicated than they looked.

It would seem that any successful philosophical theory of human knowledge should give a satisfactory account of mathematical knowledge and of how we know so many things to be necessarily true. Mathematics, after all, is one of the greatest achievements of the human mind.

Goldman has not explicitly put forward a positive theory of math-ematical knowledge. But when we look at his epistemological work with this question in mind I think the course of his refl ections over the years shows real promise. He seems to me to be in a much better position to explain human beings’ knowledge of mathematical or necessary truths than many of those philosophers whose views he has been right to repudiate.


Whether certain other aspects of his philosophy prevent him from fully exploiting what I see to be the strengths of his position is some-thing I will only touch on at the end. It is a question for him alone to decide.

In “A Causal Theory of Knowing” 1 in 1967 he made an important advance. What exactly that advance was, or what its real importance was, were perhaps not immediately clear. The announced goal of that paper was to answer Gettier’s challenge by adding to the traditional analysis of “S knows that p ” the further condition that there be a causal connection between the fact that p and S’s believing that p . The account was explicitly restricted in scope. It was not intended to apply to knowledge of what Goldman then called “nonempirical truths.” But he observed in passing that “certain elements” in his theory might well prove to be “relevant to the analysis of knowledge of nonempirical truths.” 2

I think that was correct and prophetic. But it must be said that it did not seem so right away. In fact, for Paul Benacerraf in 1973 3 a generally causal account of knowing was seen as the greatest obs-tacle to a satisfactory unifi ed account of both mathematical knowl-edge and mathematical truth. He thought the best account of the semantics of mathematical sentences represents them as quantifi ying over abstract objects like numbers or sets, and such objects stand in no causal relations to anybody or to anything else. How then could any mathematical truths be known? The challenge came directly from the idea that for knowledge that p there must be a causal connection between the grounds of the person’s belief and the subject matter of p . Far from being potentially “relevant” to mathematical knowledge, this made it look as if the causal theory ruled out the possibility of such knowledge altogether.

Drawing on Goldman, Benacerraf thought that some causal connection must be part of knowledge because we can deny that a person knows a certain thing if we know that there was no such

1 Alvin Goldman, “A Causal Theory of Knowing,” Journal of Philosophy (1967) . 2 I fi nd to my surprise that this remark does not appear in the original version of the

paper in the Journal of Philosophy in 1967; there Goldman says that he thinks “the traditional analysis is adequate for knowledge of nonempirical truths” (p. 357). The earliest appearance I have found of this newer observation is in a reprinting of the paper in Michael D. Roth and Leon Galis (eds), Knowing: Essays in the Analysis of Knowledge (New York: Random House, 1970) . It appears in some, but not all, later anthologies in which the paper is reprinted.

3 Paul Benacerraf, “Mathematical Truth,” Journal of Philosophy (1973) .


connection, that the person could not in fact have got the evidence needed for knowing that thing. That is true, and it is an important observation about knowledge. But it does not show that knowledge that p requires that the belief that p be caused by the fact that p . The most it shows is that whether somebody knows a certain thing or not—and especially whether or not he has come to know something he didn’t know before (which is primarily what Goldman and Benacerraf were talking about)—depends on how the person comes to believe the thing. What matters for the question of knowledge is the origin of the belief. And knowing that a person’s belief did not have a certain kind of origin can be enough to show that it does not count as knowledge.

That was the important point behind Goldman’s “A Causal Theory of Knowing.” A full understanding of knowledge must include some fact about the person’s acquisition of the belief or about his continu-ing to believe it. And behind that point lies what I think is an even more important point. The relevant fact about the person’s acquisi-tion or retention of the belief is not itself just a matter of what prop-ositions the person believes or what relations hold between those propositions. The necessary constraint on his acquiring or holding the belief in a certain way cannot be expressed solely in terms of rela-tions among propositions the person believes. Something else must be true of the believer. That something else is, broadly speaking, something psychological.

That this was the real importance of Goldman’s causal theory of knowing became clearer when he went on to refi ne the theory and explicitly abandoned the requirement that the belief that p must be caused by the fact that p . In his paper “Discrimination and Perceptual Knowledge” 4 in 1976 he concentrated instead on why that alleged condition is not suffi cient for knowledge either. It is possible for someone to be caused by the fact that p to believe that p without that person’s knowing that p . Even if there is such a causal connection—a barn’s being there is what causes a person to believe that there is a barn there, for example—whether the person knows or not depends on the process by which the belief is acquired on that occasion. This focusses attention on the right thing: the believer, and not just what he believes and the fact that causes that belief. Knowledge cannot be

4 Alvin Goldman, “Discrimination and Perceptual Knowledge,” Journal of Philosophy (1976) .


understood in isolation from further psychological facts about the believer.

Justifi ed belief—which is essential to knowledge—was then to be defi ned partly in terms of the kinds of processes by which the person acquires or sustains the belief. But the truth of the belief itself does not have to enter into the conditions required for a process to be of the type that produces a justifi ably acquired belief. This is what leaves open the possibility of knowledge of facts or states of affairs with which the person who comes to believe in them is not in direct causal contact. And that is what allows the theory to be generalized beyond the case of direct perception—perhaps even to knowledge of facts with which it is not possible for anyone ever to be in causal contact.

The processes that produce justifi ed belief, and hence are essential for knowing, or coming to know, are for Goldman reliable processes. His main epistemological task has therefore been to say which pro-cesses are reliable and how and why they are reliable. Rather than recounting the details of those attempts at defi nition, I want instead to draw attention to several important aspects of the general enter-prise as Goldman conceives of it, whatever the detailed defi nitions turn out to be. The features I have in mind seem to me both congenial in themselves and congenial to an eventual understanding of math-ematical and necessary knowledge.

First, as I have already mentioned, the project allows for the pos-sibility of acquiring a belief that p by a reliable causal process with-out the fact that p being itself part of that causal process. That is surely something that must be insisted on for all mathematical know-ledge. That does not mean that causation is not involved in math-ematical knowledge. Coming to know something you didn’t know before—even in mathematics—involves change, and so, presumably, it involves some kind of causal connection between what was true of you earlier and what is true of you later. It is hard to see how that transition could happen without some causal processes going on. But that does not require that what is said to be so by the mathematical truth in question—or by any mathematical truth—is what caused the change in belief. Goldman’s theory of knowledge, as far as I can see, does not require that it should.

Second, a true belief counts as knowledge by being acquired by a reliable process or method. And for Goldman the reliability of a process or method in producing knowledge can be a matter of its pedigree or ancestry. It is not simply a question of its being a process


or method of a certain kind, abstractly described. The procedures a person employs must in fact be reliable to give the person knowledge, but they must have been acquired by reliable means as well. 5 This too is congenial to an understanding of mathematical knowledge. As things are, we go to considerable effort to make sure that the ways human beings acquire their basic mathematical skills are highly reli-able. Only those who come by their reliable mathematical procedures in such ways can be said to know the truths that they can see such procedures to generate.

But even following reliably acquired reliable procedures will not produce knowledge if what they start from is not itself justifi ably acquired. This third aspect adds a historical dimension to the under-standing of knowledge. 6 The justifi cational status of a belief—and so its status as knowledge—depends on its prior history. This too is something I think we should insist on for mathematical knowledge, both individually and collectively.

Individually, a person knows some new (to him) mathematical fact only by building on what he (and ultimately others) knew already. It is only because what a pupil is taught in school is something that is already known that her acquiring the beliefs and capacities she does in that way counts as knowing. And collectively, current mathematics as a subject could get nowhere without the mathematics of the past on which it builds. It is probably impossible to understand how math-ematical knowledge in the present is possible without taking into account the history of mathematics. But what has a history is not the sheer truth of those mathematical results. It is their being known , and their being available as known and understood. That has a history, and that is what is relevant to the question of current mathematical knowledge. Goldman’s epistemological views can be seen as an encouragement to the study of that history. According to many other philosophical theories it would be at best an irrelevant digression.

Another, fourth, aspect of Goldman’s picture of knowledge and belief acquisition seems to me especially encouraging to an under-standing of mathematical knowledge. That is his forthright rejection of something I think continues to lead so many others astray. A

5 Alvin Goldman, Epistemology and Cognition (Cambridge, Mass.: Harvard University Press, 1986), pp. 51–53 .

6 Alvin Goldman, “What Is Justifi ed Belief ?,” in George Pappas (ed.), Justifi cation and Knowledge (Dordrecht: Reidel, 1979), pp. 14–15 .


person has a justifi ed belief for Goldman only if his belief has been reached by transitions from earlier states in accordance with certain “epistemic principles” or “rules.” The problem of defi ning justifi ed belief is therefore the problem of identifying and formulating such correct principles or rules of transition and explaining why conform-ing to them and not to others is required for justifi ed belief or know-ledge. 7 Which “J-rules,” as he calls them, will do this job?

It has apparently proved tempting to think of necessary truths or truths of logic and mathematics as themselves providing just such inferential epistemic principles or rules. It has even been supposed that this is the only role of such truths. This is perhaps what gives rise to the widespread but scarcely coherent thought that necessary and mathematical truths are in some distinctive way purely “formal,” or express only requirements on the “form” or “structure” of thinking, and so say nothing categorical about the way things are.

If anything can be said for that idea—that mathematical and neces-sary truths put constraints on the “form” of our thought—then the same can be said of all truths equally. Thought, in order to be correct, must conform to the truth, whatever kind of truth that might be. Goldman admirably resists the temptation others have felt to put ne cessary or logical or mathematical truths in a special epistemic posi-tion because of this. He is clear that such truths do not themselves give us, or even directly support, the kinds of “epistemic principles” he seeks. 8 He seeks “epistemic principles” or “rules” which govern psychological states and transitions between them—psychological processes of belief acquisition or revision. But necessary truths of logic and mathematics say or imply nothing about such states or tran-sitions, and nothing about whether any of them are justifi ed.

If it is necessarily true that if p then q , it does not follow that any-one who believes that p should believe that q , or even that such a person is justifi ed in believing that q . Nor does it follow that acquir-ing a true belief in accordance with that transition will always yield knowledge or justifi ed belief. If the person already knows (or even suspects) that it is false that q , she should not believe that q , even if she believes that p . And if she accepts the necessary truth linking the two she will have excellent reason to re-examine her reasons for believing that p .

7 See, e.g., Goldman, Epistemology and Cognition , ch. 5. 8 See op. cit. , p. 81–89.


This is a very important point—both in general and in application to mathematical or necessary knowledge. 9 What a person should believe at a certain time cannot be simply read off facts about the logical relations among the propositions the person believes at that time. There is in that sense no such thing even as deductive inference, if that means an inference whose conclusion is somehow purely deductively justifi ed. There obviously is such a thing as coming as a result of inference to believe a proposition which stands in the rela-tion of deductive consequence to propositions you believed earlier. But what makes the acceptance of that proposition justifi ed—and so potentially knowledge—is not simply the fact that it stands in that relation to those propositions. A proposition can stand in that rela-tion to what you believed earlier without your being justifi ed in accepting it.

This is not only a point about deduction. It holds as well for non-deductive relations among propositions. If there is such a thing as a non-deductive relation between propositions which is defi nable in completely non-epistemic terms (something many philosophers appear to have dreamed of under the name of “confi rmation”), then even if proposition p stands in that relation to another proposition q, and you believe that p , it does not follow that you should, or that you have good reason to, believe that q .

Goldman sees clearly that the notion of justifi cation, or proper “epistemic principles,” or appropriate procedures for gaining justi-fi ed belief and knowledge, cannot be defi ned in terms of logical or mathematical truths or relations among propositions alone. That is what reinforces the relevance of what he calls the psychology of the believer or knower for questions of knowledge and belief. How the person comes to believe is always relevant—not just what he believes, however rich the set of propositions he believes might be. And this holds just as much for mathematical and necessary knowledge as it does for other kinds of knowledge. The fact that in mathematics and logic there are deductive relations holding between the propositions believed makes no difference. The point about justifi cation is fully general. This is what Goldman calls “the unity of cognition.” 10 He endorses that idea; and I applaud his endorsement of it.

9 The point has been clearly made and repeatedly stressed by Gilbert Harman. See, originally, his “Induction” in M. Swain (ed.), Induction, Acceptance and Rational Belief (Dordrecht: Reidel, 1970) . 10 Goldman, Epistemology and Cognition , pp. 299–304.


A distinction had traditionally been drawn between “formal” and “material” truths, or “truths of reason” and “truths of fact.” It sounds like a distinction between different kinds of truth, but it was more than that. It was meant to imply that the human ways of know-ing truths of those kinds are fundamentally different as well. Acceptance of some such epistemic distinction is one thing that has made it so diffi cult to fi nd a satisfactory explanation of mathematical and necessary knowledge. Goldman’s “unity of cognition” opposes that basic epistemic bifurcation, and I think he is right to oppose it. He holds, and I think rightly, that “the cognitive ingredients” of belief and knowledge acquisition “are fundamentally similar across different domains,” and that “the classical attempt to dichotomize epistemology is fundamentally misguided.” 11

The idea is that there is a single thing or attitude in question—knowledge, or justifi ed belief—which we can get or try to get toward many different kinds of truths. But to reject the exclusive epistemo-logical dualism of the tradition is not to deny that there are many differences among the ways we know things. Ways of knowing can be classifi ed in countless and even overlapping ways. How many ways of knowing things are there? Let me not count the ways. But still we can say that how we come to know what is so right before our open eyes in good light is obviously different in important respects from how we come to know things in set theory or in sociology. The point is that they are not different as knowledge or justifi ed belief.

I hinted earlier that simplistic assumptions about how we know what is so right before our open eyes in good light have stood in the way of understanding mathematical knowledge. Perception is a causal process, and it seemed to be assumed that when we open our eyes the world simply impresses on us certain facts that hold in our vicinity—as if that is all it takes for us to know that there is a barn before us, for example. With that simple conception as your model of what it is like to get the most direct and so most secure knowledge of how things are, mathematical and logical knowledge are bound to seem hard to account for.

It seems to lead almost inevitably in mathematics and logic to talk of something called the “pure light of reason,” or “non-empirical intuition.” Some such faculty will seem to give us the only kind of direct contact with something somehow analogous to sense

11 Op.cit , p. 304.


perception. But of course no such faculty is or could be like sense perception in precisely the respects that matter most. There is, liter-ally, no contact with anything at all in awareness of mathematical truth; the alleged objects of awareness are not the sorts of things that can cause anything. Terms like “intuition” or “rational insight” then come to seem like nothing more than names for a certain kind of belief or conviction , perhaps strong or even unshakable conviction. But conviction, even unshakable conviction, is not suffi cient for knowledge. Knowledge in mathematics then appears unexplained, perhaps even mysterious.

Accepting the parallel with a simplistic conception of perceptual knowledge is what I think leads to the impasse. Almost from the beginning Goldman drew attention to the fact that the requirements for knowledge even in cases of apparently straightforward percep-tion are more complicated than that simple picture allows. Standing with eyes open in good light in the presence of a barn, and even being thereby caused to believe that there is a barn there, is still not suffi -cent for knowing. I think refl ection on other necessary conditions of knowledge can further soften what is felt to be the sharp contrast between perceptual knowledge and knowledge of mathematical or necessary truth.

Not every creature that opens its eyes in good light in the presence of a barn, or a human hand, thereby comes to know that a barn, or a human hand, is in its vicinity. In order to gain knowledge of such things it must be possible for the person to have thoughts of such things as barns and hands. The physical causal factors that are at work on a human being in the presence of a barn are more or less the same as those at work on a cow, at least for a large part of the story. But those infl uences lead to knowledge that a barn is present only in the human being who is suitably equipped conceptually, or who has certain capacities or abilities which I take it cows don’t have. I don’t mean cows can’t know anything. I mean a cow cannot know that there is a barn in front of her. What a creature can know, even when undergoing certain effects in a given situation, depends partly on what the creature is able to think, and is otherwise equipped to do. A cow cannot judge that here is a barn. A suitably equipped human subject can. That capacity is essential to knowledge that here is a barn, even for a creature face to face with a barn.

A subject who is suitably equipped can come to know a certain thing simply by fi nding herself in a certain situation and undergoing


by means of a reliable process a transition from one cognitive state to another. Understanding how that happens is needed for understand-ing how people come to know things. That involves at the very least understanding how the person was originally equipped. How a new belief, perhaps even justifi ed belief or knowledge, is acquired depends not only on what happens to the person but also on what state the person was originally in, and on what it took for the person to be in it. That must be understood in order to understand perceptual know-ledge. And it must be understood in order to understand mathemat-ical knowledge as well, even if the original states in each case are very different.

This precondition is what I think we should look at more closely to begin to understand mathematical knowledge. We need a more his-torical understanding of how a person, or a culture, comes to know what we know in mathematics. The same is true of music or painting. How music or painting come to be as they are at a certain period can-not be understood without seeing them as the outcome of the music or painting of the past. That is just as true of an individual person’s music or painting as it is of a whole culture.

Understanding mathematical knowledge involves understanding fi rst how a person gets into the position even of having mathematical thoughts. We must be capable of having thoughts of barns, or human hands, in order to get knowledge of the presence of such things by perception. So we must understand the capacity for mathematical thoughts, and how we get it, in order to understand how we acquire and extend mathematical knowledge. No doubt it is easier to understand how an individual person comes to have thoughts of barns or hands in a culture in which such thoughts are already present than it is to explain how there come to be any thoughts about barns or hands at all. It is easier to explain how a person comes to produce the music or painting she does in a culture that already has a trad-ition of music or painting than it is to explain how there comes to be such a thing as music or painting in the fi rst place. With a musical or painting tradition already in place, it is a matter of explaining how an individual gets socialized into the culture and comes to understand things that others already understand, and composes or paints things that are intelligible in that culture.

I think it is the same with mathematics. There are well-known fi xed procedures, and standards relentlessly insisted on, for a person’s suc-cessful socialization into recognized mathematical practice. Basic


mathematical capacities can be developed, and greatly expanded, so that a talented individual can eventually come, at least for a little while, to know something that no one else knows at all. One of the beauties of mathematics is that it can even be mathematically proved, and so known, that a proposed extension of knowledge is really sound—that a certain transition from what was known earlier to a later conclusion actually yields knowledge.

But such relatively intelligible developments within mathematical knowledge, either for an individual or for a culture, start with some-thing that is already mathematical knowledge and build on it. How a person goes from having no mathematical thoughts or knowledge at all to understanding some mathematical concepts and coming to know even the basic facts of addition and multiplication is more dif-fi cult to explain. In order to result in knowledge there must be a pro-cess of reliable transitions from non-mathematical beginnings. But those transitions will not be reliable in the sense of being provably sound, as extensions within already existing mathematical knowledge can be shown to be. The problem is obviously compounded when it is not a matter of the origins of mathematical thoughts in an indi-vidual person, but in a culture at large, or the whole human race. By what transitions from more primitive, non-mathematical coping with the world do there arise thoughts with purely mathematical contents, and so eventually mathematical belief and knowledge? That is part of what we need to know to understand fully how mathematical knowledge is possible, just as understanding how thoughts of endur-ing physical objects are possible is part of understanding perceptual knowledge of barns or of human hands.

Goldman has every reason to encourage such investigations as part of his very general epistemological project. In a recent article he touches briefl y on the genesis of numerical concepts from everyday transactions with objects. 12 That seems to me just the sort of thing to concentrate on: the conditions even for having the mathematical thoughts that are essential to the beliefs and eventually the justifi ed beliefs that constitute mathematical knowledge.

But now a question arises about how Goldman sees his more gen-eral epistemological project, and whether we can reasonably expect the mathematical knowledge that we all possess to be explained in

12 See, e.g., Alvin Goldman, “A Priori Warrant,” Philosophical Perspectives 13 (1999).


the way he appears to have in mind. I will end by simply raising this question.

In his theory of knowledge he says he wants to identify what he calls “substantive” conditions of justifi ed belief and knowledge. He wants to be able to say when a belief is justifi ed—and so when know-ledge is present—but he apparently wants to do so without using any “epistemic” terms like “justifi ed”, “good grounds”, “reason to believe,” and so on. 13 He takes truth to be a “nonepistemic” notion, and he wants to restrict himself to that, along with “probability” or “ratio,” as well as “belief” and other appropriate “nonepistemic” terms for describing the relevant processes.

I am not sure why he insists on such a requirement for a successful theory of knowledge, or even the extent to which he does insist on it. Of course he agrees that there are rules of thumb, strategies, heuris-tics, even sound procedures and demonstrably reliable methods that we can learn and try to follow in moving from one cognitive state to another. These are only what Goldman calls “methods,” and we might use the notion of one thing’s being good reason to believe another, or of a certain conclusion’s being justifi ed , and so on, in describing such reliable methods or rules. But “methods” are to be distinguished from what he calls “processes.” “Methods” have to be reliably acquired in order to give us justifi ed belief. And the “processes” by which they are acquired must be reliable. So “processes” are said to be primary, and “processes are just executed [or undergone], not ‘applied’ in the way that a rule or an algebraic or statistical technique is ‘applied.’” 14 What Goldman calls “psychologically basic processes” are part of “our native cognitive apparatus” or of “the architecure of cognition.” 15 They are what he thinks we really need to identify and understand in order to understand human knowledge. And they are to be described in non-epistemic terms. That is the domain of what he calls “primary epistemology,” 16 the basis of all other procedures or techniques. It is this task, I believe, that he thinks “cognitive science” contributes to.

I am not sure about this. I would like to know whether Goldman does understand his fundamental epistemological task in this way, and if so, why he imposes this requirement. I don’t mean why he

13 Goldman, “What Is Justifi ed Belief ?,” pp. 1–2. 14 Goldman, Epistemology and Cognition , p. 92. 15 Op. cit. , p. 92. 16 Op. cit. , p. 93.


thinks that there are processes describable in “nonepistemic” terms, without which we would not know anything. Obviously, there are physiological processes going on all the time—and a good thing too! But it seems to me that appeal to physiological processes alone could never make intelligible to us how human beings come to know the things they do, or in particular how they achieve justifi ed belief or knowledge, in mathematics or anywhere else. If we really were strict about appealing to nothing more than physiological processes we could not even identify people as being in psychological states like belief or justifi ed belief. For that we need psychological, and not only physiological, terms.

Goldman describes the basic processes he has in mind as “ psychological processes”; they are presumably processes by which thinkers move from one psychological (and not solely physiological) state to another. They are processes involving the acquisition or retention of a belief. He takes “belief” itself to be a non-epistemic term. 17 And he wants the processes which lead to justifi ed belief acquisition or retention to be specifi ed in completely non-epistemic terms as well. Not non-psychological terms, but non-epistemic terms. My question is why he thinks, if he does, that a satisfactory philosophical theory of knowledge employing psychological terms must not use or rely on any “epistemic” psychological terms in explaining how people get into the psychological states that the theory is meant to explain.

I think I understand why some people insist on such a strict require-ment in the theory of knowledge. It is a holdover from the old idea that the proper task of philosophy is analysis, combined with the idea that a satisfactory analysis must be expressed in terms of a dif-ferent kind from the term to be analyzed. But after his early “A Causal Theory of Knowing,” in which Goldman described himself as trying to answer Gettier’s purely defi nitional question, he seemed to repudiate any such purely “analytic” conception as his task. He wants to explain how we know the things we do. He wants to identify processes of belief acquisition which we can see to be the kinds of processes that produce knowledge or justifi ed belief, and then explain how people come to know mathematical truths by undergoing pro-cesses of those kinds. Is it reasonable to think we will be able to do that without using any “epistemic” terms?

17 Goldman, “What Is Justifi ed Belief ?,” p. 2.


For some other philosophers the demand for a completely “non-epistemic” understanding of knowledge comes out of their devotion to something they call “naturalism.” They think epistemic terms are evaluative or normative, and so beyond any respectable naturalistic domain. Goldman says he too wants his epistemology to be “natural-istic,” but he espouses only what he calls a “moderate naturalism” which appears to carry no such restrictive implication. His moderate naturalism says only that epistemic warrant is a function of the psychological processes that produce or preserve belief, and that epistemology needs appropriate help from science. 18 That would seem to leave a “moderate naturalist” free to use anything he can get his hands on to help explain how human beings come to know things, as long as it is something he himself thinks he knows. And a great many of the things we think we know are expressed in “epistemic” or nor-mative terms.

Goldman agrees that epistemic terms are “evaluative,” or terms of appraisal, and in explaining why he does not want them to appear in his “primary epistemology” he draws an analogy with what he calls “normative ethics.” That is the attempt to specify “substantive” “non-ethical conditions that determine when an action is right.” 19 So “primary epistemology,” by analogy, would specify “nonepistemic” conditions that determine when someone has a justifi ed belief or knowledge.

“Normative ethics,” so described, sounds to me like non-normative non-ethics. It sounds like a meta-ethical program driven by a dubious defi nitional or reductionistic demand. Is there perhaps a lingering defi nitional or reductionistic aspiration in Goldman’s epistemology? And would it leave him with a correspondingly non-normative non-epistemology? If so, I think it comes into confl ict with what I fi nd to be his much more encouraging explanatory aims.

When I say there is confl ict I mean that by not permitting himself the use of any “epistemic” or normative terms in trying to explain human knowledge—if that is what he is doing—Goldman seems to me to be putting himself under a terrifi c handicap. Could you really hope to explain how people come to know the things they do in mathematics without describing them as taking or seeing one thing to be good reason to believe something else, or one thing to be a

18 Goldman, “A Priori Warrant,” p. 3. 19 Goldman, “What Is Justifi ed Belief ?,” p. 1.


proof of another? A person’s taking something to be good reason is not suffi cient for its being a good reason, just as someone’s taking something to be a proof does not alone imply that it is a proof. So it looks as if a theorist of knowledge, in assessing the state of a person’s belief or knowledge, would also have to take certain things as proofs, or as good reason to believe something else, in order to explain the person’s mathematical knowledge. The theorist himself will employ “epistemic” terms in identifying and explaining the knowledge he observes others to possess.

This raises the question of what is involved in the very idea of something as a psychological, and not simply a physical or physio-logical, process or state. We presumably cannot do without some-thing like the concept or attitude of belief or acceptance in epistemology. Is even that concept really “nonepistemic”? Can we make sense of it, and apply it correctly, without making use of any “epistemic” notions? Could we ascribe beliefs to our fellow human beings—as we must, in order to see them as knowers—without also ascribing to them some understanding of what is or would be a rea-son to believe a certain thing? It looks as if we ourselves must at least take certain things to be good reason to believe something in order to ascribe knowledge to others and so to explain how they come to have the knowledge we recognize them to have. In ascribing beliefs and knowledge we operate perforce from within our developed capacity to understand the world in “epistemic” terms.

To insist that in all strictness we must understand and explain the place of cognition in the world without the use of any epistemic views of our own seems to demand that we must understand the phe-nomenon of human knowledge without ourselves assuming that anything is known or is reason to believe anything else. That would be to demand an external, alienated, and epistemically non-committal perspective for explaining human knowledge: an understanding of knowledge from somewhere completely outside it.

Something like this was tried, or demanded, before in philosophy, in the case of perceptual knowledge. Perception was to be under-stood on the assumption that nothing we immediately or directly perceive was to imply anything about how things actually are, or any-thing about anyone’s knowing or having reason to believe anything about how things are in the wider world. This left the staggering problem of how perception could ever give us any knowledge of an independent world. The situation looked, and still looks, and I believe


is, hopeless. There must be a way of overcoming or setting aside the restriction and explaning perceptual knowledge from within what we already know of the independent world.

Given the “unity of cognition” that I think Goldman is right to insist on, and so the unity of our ways of understanding human knowledge in general, I think the restriction to “primary epistemol-ogy” and so to “nonepistemic” terms alone must also be set aside. It threatens to rule out the possibility of explaining the mathematical knowledge of fl esh and blood human beings as the observable nat-ural phenomenon we can all understand it to be.

16

Contemporary Pyrrhonism

Writing under the title ‘Contemporary Pyrrhonism’ turns out to be more diffi cult than it might seem. How many contemporary Pyrrhonists do you know? What do they believe?

I think we all might have said that we know at least one contemporary Pyrrhonist, and that we all know the same one. He is the author of Pyrrhonian Refl ections on Knowledge and Justifi cation . 1 That is a book I greatly admire; I agree with most of it, I have learned a great deal from it, and I am very much in sympathy with the spirit and point of it. 2 The book sets out to explain and defend Pyrrhonism as a philosophy, and that philosophy seems to me exactly the kind of response I think we should make to what we all know by now as traditional epistemology. So I too, on those grounds, would identify Bob Fogelin as a contemporary Pyrrhonist.

But in a few other parts of that book, at what one hopes are only brief moments of weakness, the author appears to lose courage and to abandon what on his own account Pyrrhonism really amounts to. The weakness – if that is what it is – is understandable. It is probably not easy being a Pyrrhonist. But maybe it is not really weakness. Maybe he is willing to defend those puzzling parts of the book as well. Then is he really a Pyrrhonist? Or is Pyrrhonism something more than what he says it is in the parts of the book that I most admire? Then what is it? This begins to identify the quandary I fi nd myself in. I think the best I can do is to explain it more fully.

Put bluntly, what I don’t understand is why, having shown such insight and such good sense in responding in his Pyrrhonian way to the whole enterprise of explaining and justifying our knowledge of the world in general without bringing any of our ordinary or scientifi c

1 Oxford University Press, New York, 1994. (Page numbers alone in parentheses in the text refer to this book.)

2 See my review in The Journal of Philosophy 1995, pp. 662–665.


beliefs into doubt, and having given along the way an account of knowledge that is admirably free of all traces of relativism or contextualism, Fogelin nonetheless slips (or maybe even leaps) right back into endorsing a way of thinking that leads him to obviously false conclusions about what he and the rest of us know. And he apparently does so in the name of Pyrrhonism. This is what anyone concerned with contemporary Pyrrhonism must try to understand better.

Pyrrhonism as Fogelin fi rst describes it – what he sometimes calls “updated Pyrrhonism” – is a form of philosophical scepticism. That is, it is a sceptical or negative response to something that arises in philosophy . What arises there is a concern with the possibility of human knowledge in general. That is the subject matter of “philo-sophical epistemology”. It tries to account for human knowledge of the world in general, or at least for as much of it as can be accounted for in completely general terms. It is the attempt to explain how we know or have good reason to believe all or most of the things we think we know; to show that and how our beliefs about the world are in general justifi ed or warranted or well-supported on the basis of the grounds we have for holding them.

The updated Pyrrhonist holds that that attempt can never succeed. That is the sceptical or negative verdict. No arguments starting from the grounds that the philosophical epistemologist thinks we have for our beliefs can provide support for those beliefs without either circularity or infi nite regress or arbitrary assumption. All proposed justifi cations will fall prey to one or the other of these ancient modes of Agrippa, and so will fail. That is what the Pyrrhonist philosopher argues. He does not mean that no reasonings at all can avoid those pitfalls. He simply invokes the modes of circularity, infi nite regress, and the rest against the episte-mological enterprise of showing how our beliefs in general are justifi ed on the assumed basis. Avoidance of those pitfalls is a condition of suc-cess of that enterprise, and the Pyrrhonist argues only that that condi-tion cannot be fulfi lled within the terms of that enterprise.

This is just how the sceptics of antiquity argued against the Stoic theory of knowledge. Stoics held that knowledge is possible only because we sometimes have “kataleptic” or cognitive perceptions, which could not possibly be wrong. Sceptics argued that, given what a “kataleptic” perception is supposed to be, or what it would have to be to give knowledge in the way Stoics claimed, there are and could

Contemporary Pyrrhonism 309

be no such perceptions. So on the Stoics’ own grounds, knowledge is impossible. This is a conditional claim; it is made only about or from within the Stoics’ conception of knowledge. It does not imply that nobody knows anything. It does not imply that people should not believe the things they now believe. It does not say anything about the knowledge or beliefs of any actual human beings on earth, except this: if people know things only if they have “kataleptic” perceptions, then nobody knows anything.

The updated Pyrrhonist accordingly says that the philosophical problem of the justifi cation of our beliefs in general cannot be solved. On the standards implicit in that project, everyone should strictly speaking withhold judgement on everything. “We know nothing (or almost nothing)” is the only reasonable conclusion from the tradi-tional justifi cational project of philosophical epistemology.

That, I believe, is Pyrrhonian scepticism as Fogelin understands it. Now I, for one, believe that conditional proposition about the traditional epistemological project. Understood correctly, the project cannot succeed. So if that is Pyrrhonism, maybe there is at least one contemporary Pyrrhonist after all. If so, then I think there are two of us, because that is the proposition Fogelin argues for and accepts in the second half of Pyrrhonian Refl ections on Knowledge and Justifi cation . It is one of the many things in that book that I agree with him about. But maybe that belief alone is not enough to make one a Pyrrhonist.

Someone who arrives at that Pyrrhonist verdict about the familiar enterprise of traditional epistemology might be happy to see the end of that project. He could then turn with relief to other things; per-haps even to the quite distinct diagnostic question of how it could seem that there is such an epistemological problem, or what can make it look so pressing. That in itself could prove deeply interesting and illuminating, especially if, as I believe, there is a very strong tendency to continue to think in ways that generate that problem even among those who say they have no interest in epistemology at all. Only when someone came along with what looked like yet another way of show-ing how all our beliefs about the world in general really are justifi ed after all would the Pyrrhonist have to go back to pointing out once again how the whole thing, even in this new form, cannot really get off the ground.

One thing a Pyrrhonist who has turned away from the traditional philosophical project could do is know things. Or at least nothing he


had shown in his dialectical engagement with that project would imply that he cannot. So he could often say that he knows such-and-such and be right in what he says, just as he could think many other people are right. He could also take an interest in what he is saying when he and others say such things. He could try to say what the word “know” means, or what people mean when they say something of the form “S knows that p ”. He could also try to describe the con-ditions under which people typically say such things. In fact, it is hard to see how he could say what it means without paying attention to the conditions under which people say it.

Discoveries of this kind about his linguistic community would not themselves actually imply that anybody knows anything – that state-ments of the form “S knows that p ” are sometimes true. But that does not matter. A Pyrrhonist with semantic interests need not be trying to answer that question. Nor will what he discovers about how he and his fellow human beings behave give a satisfactory answer to the question raised by “philosophical epistemology”. But he already believes that there is no satisfactorily positive answer to that question anyway. Although what he fi nds out about how the word “know” is used and what it means does not answer those questions at all, what he fi nds out could still be true, and he could know it to be true. Or at least there is nothing about Pyrrhonism or about being a Pyrrhonist that implies that he could not.

Bob Fogelin as a Pyrrhonist says what knowledge is. “ ‘S knows that p ’ means ‘S justifi ably believes that p on grounds that establish the truth of p ’ ” (p. 94). That is what someone says of a person in say-ing that the person knows that p , whether the person in question is the speaker or someone else. Saying that someone knows something involves taking a stand oneself on the adequacy of the person’s grounds for the belief. A claim to knowledge might be denied or shown to be false because the grounds do not really establish the truth of the belief, or because the belief was not justifi ably arrived at, even though the person who makes the claim thinks both those con-ditions hold. Someone who has or later gets information that was not available to the person making the claim might know that for one or both of those reasons the knowledge claim is false. This is what Fogelin thinks happens when we hear of cases like those familiar from Gettier in which we agree that the person in question does not know, even though he has a true belief which is supported in a cer-tain way. Those examples do not count against Fogelin’s defi nition of


knowledge, since we also fi nd either that the person did not justifi ably arrive at the belief, or that his grounds do not in fact establish the truth of what he believes. The person does not know, but the defi nition of knowledge is so far sustained.

To say that the grounds of a person’s belief establish the truth of that belief is not to say that the grounds imply that the belief is true. A speaker does not have to think that his grounds make it absolutely impossible for his belief to be false in order to think that they estab-lish its truth. In saying that they establish its truth he is making an epistemic claim or commitment, not a point about implication. It is a claim that, under certain circumstances, could be, and could be shown to be, wrong.

We all make knowledge-claims on such grounds without having considered each and every possibility which, if realized, would mean that we do not know what we claim to know. The same is true of assertion in general, not only of assertions of knowledge. We put forward something as true without considering every one of the pos-sible ways in which it could have been false if it were false. This fact about the conditions under which people assert or claim to know things is also something that a Pyrrhonist can come to know by observing his fellow human beings. We say things in full recognition that we can be wrong in the things we say. Of course, we don’t think we are wrong at the time we say them. But we say things, and believe them to be true, while acknowledging that we are fallible human beings. As far as I can see, an “updated Pyrrhonist” can know that we behave in these ways. And he, like the rest of us, engages in these very practices.

It is somewhere within this area that we now approach the point — the top of the slope, as it were — at which what looks like Fogelin’s backsliding away from his Pyrrhonism begins. The Pyrrhonist leaves behind the philosophical problem of accounting for the possibility of our knowledge of the world in general. He sees that its failure presents no obstacle to the knowledge we all claim and possess in everyday life. But Fogelin thinks that if we refl ect on how and under what circumstances we all make the knowledge-claims we do in everyday life, we will begin to get cold feet about our putative know-ledge. He, any rate, gets cold feet. He fi nds himself inclined to say that he doesn’t know some very ordinary, apparently unquestionable things that he would have said he was certain he knew before he began refl ecting on what he calls our “epistemic practices”. And he


thinks all the rest of us can be brought to agree that, strictly speak-ing, we don’t know such things either.

Now why does he think that? And is he being a Pyrrhonist in think-ing it? Or does he abandon his Pyrrhonism to the extent to which he follows that inclination? These are the questions I fi nd it diffi cult to answer.

One reason I fi nd it hard to see why he thinks refl ection on this aspect of our “epistemic practices” will tend to have this undermin-ing effect is that I fi nd it hard to determine what he thinks the refl ec-tion he has in mind actually amounts to. He thinks it can start from what looks like a “legitimate complaint” or question that someone he calls “the philosophical critic of our common ways of making epi-stemic judgements” could raise. Now fi rst it is not clear who he thinks this character is. Is he the traditional epistemologist whose justifi ca-tory project the Pyrrhonian believes can be shown to be doomed? Conceding that we cannot know on those traditional standards pre-sumably leaves everyday knowledge and our “epistemic practices” untouched. Or is this philosophical observer of “our common ways of making epistemic judgements” simply the curious Pyrrhonian, musing on the human scene? If so, what is his worry? Or is this “phil-osophical critic” somebody else altogether? Is it Bob Fogelin? And if so, what is his worry?

The worry is said to start from the fact that we claim to know things without explicitly considering every one of the ways in which what we say could be false. But, and this is how Fogelin fi rst puts the refl ection, “How can we say that grounds establish the truth of a proposition while at the same time admitting that these grounds do not completely exclude the possibility that the proposition in ques-tion is false?”. After all, the refl ection continues, “If we recognize that a proposition might be false, don’t we have grounds for doubting that proposition, and isn’t having grounds for doubting some prop-osition incompatible with knowing it to be true?” (p. 89).

Now let us concede, without going into it more carefully, that if we recognize that something we believe might be false, then we have some grounds for doubting that it is true. And let us concede that having grounds for doubting that a certain thing is true is incompat-ible with knowing it to be true. How does that amount to a diffi culty for our saying that a person’s grounds establish the truth of his belief without logically implying that it is true? There might be a diffi culty if, in saying that the grounds for a belief establish its truth without


completely excluding the possibility that the proposition is false, we were conceding that even given those grounds, the proposition might be false. But surely that is not correct. Someone who fi nds the truth of a belief to be established by its grounds – even if those grounds don’t imply it – holds that the belief is true. He could not then hold that the belief might be false.

It certainly seems that Bob Fogelin is not making any such assumption. He carefully explains and defends the importance, and the importance of the special epistemic character, of the ver-dict that a person’s grounds establish the truth of his belief, even though they don’t imply it. That is the key to his whole account of knowledge. We can explain in a particular case how and why we think a specifi ed set of grounds establishes the truth of something a person believes. We then think the truth of that belief has been established. Of course, we cannot explain the idea of “establishes the truth of ” in non-epistemic terms which mention only the rela-tions among the propositions believed; there is no defi nition or reduction of the idea in neutral, non-warrant vocabulary. To sup-pose that there must be would be to fall prey to something akin to G. E. Moore’s naturalistic fallacy. That is what Fogelin thinks most ‘defi nition-of-knowledge epistemologists’ have been suffer-ing from since Gettier’s challenge in 1963, if not earlier. In saying that the grounds establish the truth of a belief one is taking a stand oneself on the adequacy of epistemic support for the belief one regards as true.

So I don’t see how Fogelin could be tempted by the kind of refl ec-tion I have just considered. But I must say he seems to be tempted by it. Here is what he says:

Normally we ignore these [remote] possibilities [which would render our assertion false], but if we dwell on them, our level of scrutiny will rise, and we will fi nd ourselves unwilling to claim to know many things that we usually accept as items of knowledge. Do I, for example, know my own name? This seems to me to be as sure a piece of knowledge as I possess. But perhaps, through a mix-up at the hospital, I am a changeling. I’m really Herbert Ortcutt, and the person who is called “Ortcutt” is actually RJF. These things, after all, do happen. Given this possibility, do I know my own name? I’m inclined to say that I do not. . . . [And he thinks he is not alone] . . . When pressed in this way people . . . will acknowledge that strictly speaking – if you are going to be picky – given that they do not know they are not changelings, they do not know their own names. (pp. 93–94)


He says, “Given the possibility that there was a mix-up at the hospi-tal, I am inclined to say that I do not know my own name”. Now in what sense is that possibility “given”? Is it that his grounds for believ-ing that his name is Bob Fogelin do not logically imply that there was no mix-up at the hospital, and so do not imply that that is his name? Well, fi rst, I wonder whether that is really true. Is there nothing in his grounds that implies that there was no mix-up at the hospital? Isn’t it likely that part of his grounds for believing that his name is Bob Fogelin is that he was given that name in the hospital and that he still had it (so to speak) when he got home? And doesn’t that imply that there was no mix-up at the hospital? Or is it that his grounds for believing those grounds in turn do not imply them? Is even that true? Or is it rather that his having those grounds does not imply that there was no mix-up? But can we be sure that even that is so?

In any case, let us grant that there is nothing in his grounds that implies that there was no mix-up at the hospital. Then the possibility in question is a failure of implication of one thing by another. Given the truth of his grounds, it is still possible that there was a mix-up at the hospital. But when he said he knew his name is Bob Fogelin he was saying that his grounds for that belief establish its truth, so how does the mere failure of implication work to undermine that epi-stemic judgement of his? Why is he inclined to say “Given that pos-sibility, I don’t know”?

This is not to challenge what he actually says about the rest of us. He says that, when pressed, people will acknowledge that, “given that they do not know that they are not changelings, they do not know their own names”. And that is probably true. If they don’t know that they are not changelings, then they don’t know who they are, and they would probably admit it. But how is it to be shown that people do not know that they are not changelings? The fact that the truth of their belief that they are not changelings is not logically implied by their grounds for believing it does not show that they do not know it. On the Fogelin account of knowledge, when someone says that she knows that she is not a changeling, and knows what her name is, she says that her grounds establish the truth of what she believes. So what leads Bob Fogelin to withhold such an epistemic claim in his own case?

Now he could be relying on something here that nobody else knows, and that he is reluctant to reveal. He says that perhaps he is Herbert Ortcutt, and the man known all these years as “Ortcutt” is


really Bob Fogelin. This name “Ortcutt” is immediately suspicious. We know that at least one of the Ortcutts is believed to go in for spying. That is Bernard J. Ortcutt, grey-haired pillar of the com-munity sometimes seen in suspicious circumstances wearing a brown hat. 3 But Bernard J.’s subversive activities, however impres-sive, would be as nothing compared to the deception brother Herbert would have pulled off if Fogelin’s, or rather Herbert Ortcutt’s, speculation is right.

But no, I don’t think Bob has some secret information that he is not at liberty to reveal. He simply refl ects on the possibility of a mix-up. Even given all his grounds for believing that his name is Bob Fogelin, that possibility is apparently what inclines him to say that he doesn’t know that that is his name.

Another way he puts the refl ection is to say “it seems entirely nat-ural to ask how grounds can establish the truth of something when at the same time there are undercutting possibilities that have not been eliminated” (p. 94). This is meant to be a description of what we do – of our “epistemic practices”. When we claim to know some-thing “we assert something, thus committing ourselves to it without reservation, while at the same time leaving eliminable refuting pos-sibilities uneliminated” (p. 94).

But if I say that I know that the name of this man before us is Bob Fogelin – something I certainly do say, and without reservation – do I leave “uneliminated” the possibility that there was a mix-up at the hospital and it is really Herbert Ortcutt instead? I think I do not. I say that I know the name of this man is Bob Fogelin, son of the parents of Bob Fogelin, so what I say is inconsistent with, and in that sense eliminates or rules out, a mix-up at the hospital. What I believe is established by my grounds eliminates that possibility as actual. “But what reason do you have to eliminate that possibility?”, some-one might ask. I say I have all the reasons that I have for believing that his name is Bob Fogelin. And those reasons, I judge, are enough to establish the truth of that. That is what I commit myself to in say-ing that I know his name is Bob Fogelin.

This might seem presumptuous or arrogant on my part. How can I claim to know that this man’s name is Bob Fogelin when the man himself is inclined to say he doesn’t know it? Don’t I have to admit

3 W. V. Quine, “Quantifi ers and Propositional Attitudes”, in his The Ways of Paradox , Random House, New York, 1966, p. 185 .


that I could be wrong about there having been no mix-up at the hos-pital? These things, after all, do happen, as Bob says. Well, yes, they do, but in saying I know his name I am saying or implying that no such thing happened in this case. I admit that I am not infallible. Someone who knows or has reason to believe something that I am not aware of might reasonably conclude that I am wrong – that my grounds do not establish that his name is Bob Fogelin after all. That is a possible development. The matter can be settled only by looking at that person’s reasons for doubt. Similarly with my view about Bob Fogelin (or this man now before us). I think he is wrong to say that his grounds do not establish the truth that his name is Bob Fogelin. I think he does know. He is inclined to say he does not. The matter can be settled only by looking at his reasons for doubt. But that is just my problem: what are his reasons for doubt?

He fi nds himself unwilling to claim that he knows his own name when he “dwells” on possibilities like a mix-up at the hospital. What he thinks happens when we dwell on such possibilities is that “our level of scrutiny” of our claim to know “will rise, and we will fi nd ourselves unwilling to claim to know many things that we usually accept as items of knowledge” (p. 93). “Refl ection on remote possi-bilities”, he says, “can raise the level of scrutiny and thus lead us to withdraw epistemic commitment in a wholesale way” (p. 94).

It is hard to see how refl ection on a possibility can have this effect on our knowledge-claims, especially given Fogelin’s own conception of knowledge. The possibility involved in this case is that of his name’s not being Bob Fogelin (because of a mix-up at the hospital) even though our grounds for believing that that is his name are true. To dwell on that possibility would therefore be to dwell on a failure of implication. Is that enough to raise the “level of scrutiny” of our knowledge-claim? I don’t think so if, while we are dwelling on the fact that our grounds do not imply what we claim to know, we can also at the same time refl ect on what our grounds establish . We will then fi nd that the possibility we were dwelling on is not actual. What we think is established and so known is something that implies that there was no mix-up at the hospital. Of course, that is not all it implies. This man’s name being Bob Fogelin also implies that even after the hospital those sneaky Ortcutts did not get into the Fogelin home and substitute their own baby Herbert, who was such a dead ringer for baby Bob that the parents never noticed. And of course it implies the non-actuality of many other such possibilities.


So I still don’t know what refl ections on possibilities can have such devastating undermining effects on our knowledge. But let me turn now to the question whether those refl ections, whatever they are, are “Pyrrhonian” refl ections. Fogelin says the refl ections can “lead us to withdraw epistemic commitment in a wholesale way”. So “the recog-nition that we make knowledge claims without [eliminating these defeators]”, he says, “gives one as robust a skeptical challenge as one could like” (p. 193). Is that challenge a “Pyrrhonian” sceptical chal-lenge? It is said to reveal “the fragility of our common epistemic practices” (p. 193). So I ask: is someone who notices that “fragility”, and dwells on it, and so falls prey to a wholesale withdrawl of epi-stemic commitment from things he thought he knew in everyday life, really exhibiting the true spirit of Pyrrhonism?

It seems a far cry from the untroubled Pyrrhonist described at the beginning. That was someone who, having shown the impossibility of any positive answer to the traditional philosopher’s question about knowledge in general, nonetheless remains unperturbed and calmly goes along claiming to know many things, and usually being right about it, and assessing both positively and negatively the everyday knowledge-claims of others in his community. But by the end of the book the Pyrrhonist is described as one who notices and refl ects on the “fragility” of the practices he engages in, and as a “natural con-sequence”, fi nds that he has “unleashed what amounts to an unmiti-gated skepticism” (p. 195) about all of knowledge. The dilemma he is led to is simply “incapable of resolution” (p. 203).

I think something has gone wrong here. What is at stake is not simply the question of a label – is this Pyrrhonism or not? What is at stake is whether refl ection on human knowledge does or must leave us vulnerable to this kind of collapse. And if so, what kind of refl ec-tion does it take? Here is something on which I think I really do dis-agree with Fogelin.

He says that refl ection on the conditions under which we make everyday knowledge-claims raises “the level of scrutiny” so that we are led to abandon those claims, and to deny the claims of others. In order to be led into such doubts he thinks we do not need to invoke what he calls “skeptical scenarios”. Those are “radical” or “globally dislocating” (p. 193) or “systematically uneliminable” (p. 91) possi-bilities, such as dreaming or total hallucination or perhaps a brain in a vat. Such possibilities are not needed because refl ection on “the fact that our empirical claims are made in the face of unchecked, though


checkable, defeators” (p. 193) is enough to raise “as robust a skeptical challenge as one would like”.

This is what I would deny. At least, I have been unable to identify any refl ections on such possibilities that I think would have that sweeping undermining effect. Fogelin holds that “the theory of knowledge, in its traditional form, has been an attempt to fi nd ways of establishing knowledge claims from a perspective where the level of scrutiny has been heightened by refl ection alone” (p. 99). I don’t disagree with that, as it stands. Refl ection alone can present us with a question about knowledge of the world in general. And I do think that by refl ecting on our knowledge from within the traditional justi-fi catory project, we do and must end up with total scepticism. That is what “Pyrrhonian scepticism” says, and that is what Fogelin argues for in the second half of his book. We all agree about that.

But I think the refl ections that have that negative outcome in that project are not just refl ections on the fact that we make knowledge-claims without checking every one of the possible ways in which what we say could be wrong. I think the reason “Pyrrhonian scepticism” is correct in its response to the traditional epistemological project is precisely because the threatening possibilities in that case are system-atically or globally ineliminable. The refl ections leading to that trad-itional problem rest on the idea that all knowledge of the world around us comes ultimately from perception, and that what we receive in perception can be seen to be limited in a certain systematic way. Once it is shown or granted that we could perceive everything we do even though the world around us were very different from what we believe it to be on the basis of all our perceptions, there is no way for anyone to get any reason to believe one thing rather than another about the world around us. This is where Pyrrhonian deployment of the modes of Agrippa comes into play.

Any attempt to go from what we get in perception to anything beyond it in the world around us will fall into circularity or regress or unjustifi ed assumption. It will be circular because you have to appeal to something or other beyond what is perceived in order to get any reason to believe something that goes beyond what is perceived. It will be regressive because whatever could be appealed to (something perceived) could do what is demanded of it only with the help of something else of the same kind (something else perceived), and so on without end. Or it will rest on an unjustifi ed assumption if you just help yourself to something beyond the perceived data to support


a conclusion about the world beyond. With only what falls on the side of perception to appeal to, there is no legitimate way to get beyond it. The systematic failure of all such attempts to transcend the available data is what the Pyrrhonian refl ections reveal. On that point, as I said, I think the Pyrrhonist is completely right, and for the reasons he gives.

With knowledge-claims as understood in everyday life things are not the same. Take Fogelin’s everyday claim to know his own name – before he began to get any doubts. Nothing the Pyrrhonian invokes to show that knowledge as the traditional epistemologist tries to explain it is impossible can be shown to stand in the way of that everyday knowledge. Fogelin mentions the possibility of a mix-up at the hospital, but that possibility can be shown not to be actual, and so eliminated. And that can be done without falling into any circu-larity, infi nite regress, or unjustifi ed assumption. Suppose that after all these years it is discovered that there was a security camera at work in the hospital, and that it was trained continuously on the young Fogelin from the time of birth until his discharge from the hospital. We can play the fi lm and see that there was no switch. That would settle the question without circularity or regress: there was no mix-up at the hospital. But nothing could settle the traditional philosopher’s question of which of several competing possibilities holds in the world around us, if it can be settled only by perception, and whatever anyone could perceive always falls short of any states of affairs of the world.

That is one difference between knowledge in everyday life and what the traditional epistemological project requires. Another differ-ence is that if I claim here and now to know that this man’s name is Bob Fogelin – as I do – and I am asked what reason I have for elim-inating the possibility that there was a mix-up at the hospital, I can reply – as I did – that I have all the reasons that I have for believing that his name is Bob Fogelin. Those grounds, I believe, establish that that is his name, and that in turn implies that there was no switch. The possibility is eliminated as inconsistent with something I know. That is how I know there was no switch.

But what looks like this same kind of move does not work within the traditional epistemological project. I do not successfully elim-inate the possibility that I am dreaming that there is a tomato before me by claiming that I know that the tomato I see is really there, so I know that I am not dreaming. I can say that that is something I know,


and what I say implies that the dream possibility is not actual, but once I acknowledge that all my perceptions are restricted in the way the traditional problem depends on, I see that I have to take it back. The position I understand myself to be in gives me no more reason to believe that there is a tomato there than that there is not. But in the everyday case, if I think all my grounds establish the truth that this man’s name is Bob Fogelin, and nothing has come along to reveal that those grounds are weaker than I thought, I can with the same reasons continue to claim to know that that is his name, and also claim to know anything else that I know follows from that.

That philosophical view of the limited resources of perception is at the heart of the traditional project. Any view that systematically restricts the kind of data available to us as grounds for knowledge car-ries with it that disastrous sceptical conclusion. That is what the Pyrrhonian use of the modes of Agrippa reveals, just as sceptics of antiquity exposed the disastrous consequences of the Stoic conception of perception and knowledge. But those consequences are disastrous only for those philosophies, only within a certain philosophical enter-prise, and only with those restrictive views of perception. And restrictive views of what is available in any possible perception involve the use of what Fogelin calls “sceptical scenarios” – “globally” and “systemati-cally ineliminable” general possibilities. The negative Pyrrhonian ver-dict can be a correct conditional proposition about any such epistemological enterprise without having any implications one way or the other for what we do, or should do, in everyday or scientifi c life.

So that is the problem I have found myself faced with. Is Bob Fogelin a contemporary Pyrrhonist or not? We know he is contem-porary, but is he a Pyrrhonist? The question is whether, in those refl ections that produce his inclination to say he doesn’t know his own name, he is following only Pyrrhonian refl ections. Or has he been seduced away from Pyrrhonism through not having rid himself completely of the kind of corrupting thoughts that keep alive the traditional concern with the possibility of human knowledge in gen-eral? There are distressing signs that that this might be so. He says he fi nds that “demands for philosophical modes of justifi cation can spring quite naturally” from “refl ection on our ordinary modes of justifi cation”. And this leaves him feeling “the need for something more” (p. 203).

But where could “something more” be found? Observation and description of what we actually do, he thinks, would never be enough.


“It is possible to describe those circumstances under which we employ epistemic claims in a nontentative way”, he says, but “this . . . does not show that our epistemic practices are legitimate” (pp. 199–200). That, apparently, is the worry he is left with. It is not a worry about some particular claim to know, or even about claims of a certain kind. It is a demand for justifi cation of our “epistemic practices” in general.

To get to the bottom of Fogelin’s dissatisfaction here we would need to understand better how someone with such an admirable non-reductive conception of knowledge, and someone with his philo-sophical feet always apparently so fi rmly planted on the ground, can nonetheless be lured away from the comforts of an unthreatening Pyrrhonism by some so-far-unexplained longing for legitimacy. It is a tribute to the work of Bob Fogelin that to make progress on this question about Bob Fogelin would be to understand better the source of the disturbing and still-not-fully-understood appeal of traditional epistemology itself.

17

Perceptual Knowledge and Epistemological Satisfaction

Ernest Sosa has written so well about so many different topics that there is no hope even of expounding, let alone examining, the full range of his contributions in a single essay. Even within the theory of knowledge he has directed his careful scrutiny to such issues as the defi nition of knowledge, contextualism, foundationalism, coherence theories, epistemic virtues, and different pictures of knowledge behind apparently intractable epistemological controversies. Recently he has been attending to a general question about the very possibility of a philosophical theory of knowledge. He fi nds a form of subjectivism or relativism or skepticism rampant both in the general culture and within the apparently stricter precincts of philosophy that would bring the whole enterprise to ruin. It is based on recognition of an inevitable circularity, perhaps in all intellectual endeavor, but certainly in the search for a philosophical theory of human knowledge.

Sosa wants to defend the traditional epistemological enterprise against such charges. A successful philosophical theory as he under-stands it would show that, and how, we know the sorts of things we think we know, and so would provide a general explanation of human knowledge. It would be established by careful study of what human beings are like and how they use the abilities they are endowed with to come to know the things they know. He insists that there is no diffi -culty in principle with human beings coming in this way to know such things about human knowledge. “We can legitimately and with rational justifi cation arrive at a belief that a certain set of faculties or doxastic practices are those that we employ and are reliable,” he says. 1

I want to take up Sosa’s defense of the traditional epistemological enterprise. I think there remains a question about its philosophical

1 Ernest Sosa, “Philosophical Scepticism and Epistemic Circularity,” Proceedings of the Aristotelian Society supp. vol. 68 (1994) (cited as PSEC), p. 285 .

Perceptual Knowledge and Epistemological Satisfaction 323

prospects that he does not consider or put to rest. In fact, I believe the kind of theory he favors is vulnerable to the diffi culty I see. But I do not think it is a question of circularity. As a general objection to the possibility of understanding knowledge, I think the charge of circularity is without force.

Sosa’s optimism about the epistemological project is expressed in what he appears to regard as a rhetorical question. If we have “legit-imately and with rational justifi cation” arrived at an explanation of human knowledge, he asks, “Why could we not conceivably attain thereby a general understanding of how we know whatever we do know?” (PSEC, 285). I think this raises a real question which can be given a good answer. Even if we “legitimately and with rational jus-tifi cation” arrived at a theory of knowledge of the kind Sosa has in mind, I think there is a way in which we would not thereby attain a satisfactory general understanding of how we know what we know. I will try to explain why not. This is something I have tried to bring out before, apparently with little success. 2 It is worth trying again; I think something important about the philosophical understanding of human knowledge is at stake.

Sosa starts with a question of the kind he thinks a successful theory of knowledge should answer: “Is the existence of external things just an article of faith?” (RK, 410). He takes the question from G. E. Moore, who understands external things as things to be met with in space. Tables, trees, socks, mountains, even soap bubbles, are things of a kind to be met with in space. And they occupy their positions in space, and so exist, if they do, whether anyone ever encounters them or perceives them or thinks about them or not. The question is to be understood so that if any such things exist, then there are external things. That, of course, would not settle the epistemological ques-tion. There could be such things even if no one knows that there are. Nor would everyone’s believing that there are external things settle

2 See my “Understanding Human Knowledge in General,” in M. Clay and K. Lehrer, eds., Knowledge and Skepticism (Boulder, CO: Westview Press, 1989) . Sosa found the alleged shortcoming illusory in PSEC. I replied in my “Scepticism, ‘Externalism’, and the Goal of Epistemology,” Proceedings of the Aristotelian Society supp. vol. 64 (1994) that he had not correctly identifi ed the diffi culty I have in mind. In his “Refl ective Knowledge in the Best Circles,” The Journal of Philosophy 94 (1997): 410–430 (cited as RK), he appears to understand it as a charge of circularity. Both papers by me are now reprinted in my Understanding Human Knowledge. Philosophical Essays (Oxford: Oxford University Press, 2000).


the question of knowledge; belief can be a matter of faith. But if anyone knows or has good reason to believe that tables, trees, socks, or mountains exist, then that person knows or has good reason to believe that external things exist; he is not just taking it as an article of faith. That would be a positive answer to the question about human knowledge of external things.

If many people knew that there are external things, but no one knew that anyone had that knowledge, then no one would know the answer to the epistemological question. There would be a positive answer — the existence of external things would be something known and not just an article of faith – but no one would know it. If many people after epistemological investigation came to believe that human beings know that there are external things, that would not be enough for a satisfactory outcome of their epistemological investigations, just as everyone’s believing that there are external things was not enough to settle the question of their knowledge. Even if the investigators’ belief were in fact true, that still would not be enough. The truth of the answer they accept would not give them the understanding they seek unless they could recognize that they know or have good reason to believe that answer. Rightly fi nding themselves with knowledge or good reason to believe that the answer they accept is true would give them a satisfactory understanding of human knowledge.

One way in which the existence of external things would be some-thing that we know, and not just an article of faith for us, is if we could often see that there is a table in the room, a tree in the garden, socks on our feet, and so on. That would be knowledge by perception of the existence of external things. I think we can and do see such things to be true, and so thereby know that there are external things. That is one answer to a question about our knowledge of external things; what might be called the most straightforward answer.

If perception is indeed a way of coming to know something about external things, then I can also know by perception that that answer to that epistemological question is correct. I can often see that some-one right in front of me sees that there is a table in the room and thereby comes to know that there is a table in the room. I see what he does and I see that he knows. So I can see, and thereby know, that that is how people come to know that there are external things. The truth of that straightforward epistemological explanation is some-thing I can know to be true by perceiving that it is true, just as I can know that there are external things by perceiving that there are.


I think there is no suspicious circularity in this way of coming to know that there are external things. Circularity can enter the picture only where there is a chain of inference or a course of reasoning by which a conclusion is reached. But the straightforward answer says nothing of a chain of inference or reasoning. It says that one sees that there is a table in the room, not that one infers that there is a table in the room from something else. And to see that p is to know that p . It is not that one infers that one knows that p from the fact that one sees that p . Whoever sees that p thereby knows that p . Whoever sees that there is a table in the room knows that there are external things.

Just as there is no circularity in coming to know in that way that there are external things, so there is no suspicious circularity in com-ing to know in that same way that this answer to the epistemological question is true. This straightforward answer does not say that one reaches the “conclusion” that people know by perception that there are external things, and so perception is reliable, by inferring it from something else one sees or knows. It says simply that I can see and thereby know that people see and know in that way that there are external things. There is no inference or chain of reasoning, and so no room for circularity. It is true that I use my eyes in fi nding out how people fi nd out about the things around them, and I fi nd that they use their eyes, but there is nothing circular or illegitimate about that. There is no illegitimacy or paradox in using our eyes to fi nd out how the human eye works, or using the larynx to lecture on the workings of the larynx. It is no different in this case, given that we can see that there are external things.

So the straightforward explanation involves no circularity, and it is something I can know to be true by seeing that it is true. If that is the answer Sosa would give to the question about our knowledge of external things from which he begins, there is no disagreement between us, either about circularity or about our knowledge of exter-nal things. We can “legitimately and with rational justifi cation” arrive at a belief that knowledge of external things is acquired by percep-tion, and we can “attain thereby a general understanding” of how we know many of the things we know.

Many philosophers will grant that perhaps there is nothing wrong with that answer as far as it goes, but will feel that there is a deeper and more challenging question about perceptual knowledge to which it is not a satisfactory answer. One source of this feeling is the demand


that it should somehow be established that human beings ever do see that there is a table in the room, a tree in the garden, or any other fact involving the existence of an external thing. The straightforward answer, it is thought, simply assumes or takes it for granted that that is so. On that assumption, the answer is perhaps unobjectionable. But for a serious and satisfying theory of human knowledge, it is felt to be something that must be shown to be true, not simply assumed.

I think this familiar reaction, as stated, is based on a misunder-standing of the straightforward answer. That answer does not simply assume or take it for granted that people can see that there is a table in the room or that other external things exist. It says that is some-thing that almost anyone can see to be true right before his eyes; it is not something that is or must be assumed. So it can be established that people can see that there are external things. The expressed dis-satisfaction with the answer appears to rely on a lingering suspicion of circularity. But we saw that there is no room for circularity when that answer is rightly understood.

The straightforward answer would be found dissatisfying by some philosophers because they believe that no one ever does see that p , where what takes the place of ‘ p ’ implies the existence of an external thing like a table or a tree or a sock. They think certain general refl ec-tions about sense perception should be enough to convince anyone of that fact. So they think the straightforward answer is not, strictly speaking, true. For such philosophers, this presents a deep and chal-lenging epistemological question: how do we get knowledge by per-ception of the existence of such external things, given that we never, strictly speaking, see or otherwise perceive that they exist?

This is perhaps what has come to be called the philosophical prob-lem of our knowledge of the external world. But with this under-standing of the restricted deliverances of unaided perception, the word ‘external’ takes on new signifi cance. It no longer just denotes things to be met with in space, like tables and trees. It now applies to everything that is not, strictly speaking, perceived to be so; and what is perceived turns out to be much less than might originally have been thought. The problem then is how we can come to know or have rea-son to believe anything about what is “external” to, or beyond, the limited domain of what we strictly speaking perceive.

It is to this kind of question that Sosa appears to think the theory of knowledge he has in mind can give a satisfactory answer. The question he addresses is one to which he thinks it is at least possible


to give an answer that fails through circularity or regress. That is what he says is true of all “internalist” theories of knowledge. In attempting to show how we have perceptual knowledge of something beyond what is perceived, either they rely at some point on know-ledge of something beyond what is perceived, and so are “epistem-ically circular,” or they appeal only to what is perceived and so to a regress of perceptions which establish nothing beyond themselves (PSEC, 267–8, 286). Sosa accordingly favors a form of “externalism” that he thinks is invulnerable to that charge. But he appears to regard his “externalism” as providing a satisfactory non-circular answer to the same question that “internalism” fails to answer.

Sosa thinks the reason “internalist” theories fail is that they try to explain knowledge of facts which we do not, strictly speaking, per-ceive as knowledge arrived at by inference or reasoning from some-thing we do perceive or are aware of in experience. Accepting the restrictions on unaided perception which create the epistemological challenge, the most we can strictly perceive or be aware of is what Sosa calls “the character of our experience” (RK, 412), or perhaps its “qualitative character” (RK, 413), not facts which hold independ-ently of our experiencing them. “Experience as if there is a fi re before us does not entail that there is a fi re there, experience as if here is a hand does not entail that here is a hand, and so on” (RK, 413), he says. And that is the most that experience alone provides us with. But he regards it as “doubtful” that “any allowable form of infer-ence . . . will take us from the character of our experience to the sort of knowledge of our surroundings we ordinarily claim” (RK, 412).

On this too I agree completely with Sosa, demurring only at his understated “doubtful.” I think it is not just doubtful that such infer-ences could support our beliefs about the independent world, but impossible. I think Sosa would not disagree. That is precisely his case against “internalism.” Some philosophers would draw from this dead end the skeptical conclusion that therefore perceptual knowledge of an independent world is impossible. Sosa concludes only that percep-tual knowledge of facts which go beyond what is perceived cannot be understood as inferential knowledge. The appeal of his “external-ism” is that it offers “a way to explain how we can know that p with-out reasoning from prior knowledge” (RK, 418).

We do not have to enter into the precise details of what Sosa thinks is the most promising formulation of this “externalist” theory in order to assess it as an answer to his epistemological question. “The


key idea exploited is this: you can know something non-inferentially so long as it is no accident or coincidence that you are right” (RK, 418). Applied to perception, this means that one will have perceptual knowledge of certain facts if there is a non-accidental or reliable con-nection between one’s perceptual experiences and the facts that one believes in as a result of them. The reliability of perception for Sosa is a matter of there being “experience/belief connections” of that kind. “Good perception is in part constituted by certain transitions from experiences to corresponding beliefs – as is the transition from the visual experience characteristic of a tomato seen in good light to belief in the tomato” (RK, 421). In order for those who enjoy that characteristic visual experience to know that there is a tomato there, and so for perception to be a reliable source of knowledge of the independent world, perceivers do not need to know that the “transi-tion” is reliable; it is enough for it to be reliable (RK, 426).

This conception of knowledge is what Sosa sees as the key to his answer to the question how we get perceptual knowledge of the exist-ence of things like tomatoes, given that we never, strictly speaking, perceive them; their existence is never entailed by our having the per-ceptual experiences we have.

That presents no obstacle to knowledge if there are reliable con-nections between the perceptual experiences we enjoy and the inde-pendent world we believe in. This view says that human beings then know by perception that there are tomatoes and other external things. That human beings get such knowledge in that way is also something that we can come to know by observing human beings. If there are reliable connections between the perceptual experiences we enjoy while engaging in such epistemological investigations and the world of people and other external things that we are studying, and we come to believe under those circumstances that human beings have perceptual knowledge of external things, then we know (on this con-ception of knowledge) that Sosa’s answer to his epistemological question is true. We have “legitimately and with rational justifi ca-tion” arrived at “a belief that a certain set of faculties or doxastic practices are those that we enjoy and are reliable.”

Having fulfi lled in this way what the “externalist” view says are suffi cient conditions for knowing how human beings get perceptual knowledge of external things, have we thereby attained a satisfactory general understanding of how we know what we know about the independent world? I think there is still a way in which we have not.


Sosa’s account leaves us in what I think is still an unsatisfactory pos-ition for understanding whatever knowledge we have. But it is not unsatisfactory because it suffers from some kind of circularity. There is no circularity involved in fulfi lling the conditions Sosa says are suf-fi cient for perceptual knowledge. Nor do I protest that those condi-tions are not suffi cient for knowledge. That is a complex question which turns on the precise formulation of the “externalist” defi nition of knowledge. But the diffi culty I think we are left in would remain even if we grant that people know just the things Sosa’s theory implies that they know.

Someone who accepts Sosa’s theory knows (on that conception of knowledge) that the human beings he observes know by perception that there are external things. He knows that because he knows that there are reliable connections between the perceptual experiences those people receive and the external things they believe in. It is not simply that he believes that there are such connections. The theorist would concede that his believing alone would not be enough for him to know how people come to know the things they do by perception. He thinks he has a satisfactory explanation of their knowledge because he thinks that explanation is something he knows to be true. And if it is true – if there are reliable connections between people’s perceptual experiences and the facts they come to believe in as a result of them – the theory implies that the theorist who accepts that explanation does know that those people know in that way that there are external things.

But that theorist, in light of his own theory, must acknowledge that he does not, strictly speaking, ever see or otherwise perceive that those human beings and other external things that he is interested in are there. Nor does he ever perceive the reliability of the connections that he believes hold between them. The most he is perceptually aware of or presented with in experience are the qualities or charac-ter of his perceptual experiences. Of course, he believes in those human beings and other external things, and in the reliable connec-tions between them, even if he never sees that they are there. He comes to believe in them as a result of undergoing certain perceptual experiences. And his theory says that he thereby comes to know of them by perception if there are reliable connections between those perceptual experiences he has and the facts he comes to believe in as a result of them. Of course, he believes that theory of knowledge, and he believes that there are such reliable connections, so he will


confi dently assert that he knows what he thinks he knows about the world, and in particular that he knows that human beings get know-ledge of the world by perception in the way he thinks they do.

Any theorist of this kind who refl ects on his position will concede, as before, that his merely believing is not alone enough for him to know how people know what they do. He would not thereby achieve a satisfactory explanation of human perceptual knowledge. And I think he must also concede that even knowing that people know things in that way would not be enough, if knowing is simply a mat-ter of fulfi lling the conditions Sosa’s theory says are suffi cient for knowledge. All the theorist can appeal to in accounting for his own knowledge as more than confi dent belief are the perceptual experi-ences he knows he has had, the beliefs he holds, which he believes to be the result of those experiences, and the theory of knowledge that he also believes. That theory says that if one further condition holds, then he does know what he thinks he knows. And he believes that that further condition holds. But still he remains in no better position for understanding himself as knowing what he thinks he knows than someone who refl ects on his knowledge with equal confi dence and in an equally satisfactory way and yet knows nothing at all.

Sosa in his concern for circularity imagines beings who do not resort to ordinary sense perception but consult a crystal ball to fi nd out about the world. He wonders whether they could show without circularity that that practice is reliable, and if not, whether ordinary sense perception is any better off in that respect. If the idea is to establish the reliability of sense perception as a source of knowledge of the world by appeal only to what is here taken to be, strictly speak-ing, perceived, then the answer seems to me “no,” just as the reliabil-ity of crystal-ball gazing as a way of knowing cannot be established by appeal only to what is seen in crystal balls. But again, I do not think circularity is the issue in the plight of Sosa’s “externalist” epis-temological theorist. What matters is the poverty of the resources available to him for understanding his own knowledge. Crystal-ball gazers help make the point.

A committed crystal-ball gazer could refl ect on what he takes to be his crystal-ball gazing knowledge of the world and claim to understand it in a way parallel to Sosa’s imagined “externalist” the-orist. He believes many things about the independent world, but he has never seen or otherwise perceived anything except what he fi nds on gazing into his crystal ball. But he believes a theory of knowledge


to the effect that if there are reliable connections between his seeing what he does in the ball and facts in the wider world that he believes in, then he knows what is so in the wider world. Often, when he gazes into the ball, he sees certain things and then fi nds himself believing that many other crystal-ball gazers know things about the world around them by gazing into their own crystal balls. He believes that because he believes that there are reliable connections between those people’s seeing what they do and the world they come to believe in as a result of it. So, given the theory he also believes in, he takes himself to know that crystal-ball gazers know things about the world in that way.

In refl ecting on his knowledge and explaining to himself and others how he knows what he knows, he will concede that his merely believing what he does about his knowledge is not alone enough for him to have a satisfactory explanation of it. He thinks he has a satisfactory explanation because he thinks he knows what he claims to know about crystal-ball gazing knowledge. He recognizes that he has certain experiences, and certain beliefs which he believes to be the result of them, and he believes a theory which says that if those experiences and beliefs are connected in a certain way with facts in the wider world, then he has crystal-ball gazing knowledge of that world.

The difference between the positions of the two theorists lies only in the believed-in connections between the relevant experiences and the wider world. The theory says in each case that if such connec-tions hold, that theorist knows. Each theorist, confi dently sticking to his own story, believes that they hold in his case and not the other. Each might even try to settle the matter by consulting his own ex perience and his own theory, and fi nd himself content with the discovered result. In that respect, the two positions are equally satis-factory, or unsatisfactory.

It is perhaps tempting to say that what distinguishes them is only something that lies beyond the knowledge of either theorist; it is a matter only of what is actually so. But on Sosa’s “externalist” con-ception of knowledge, that is not right. If reliable connections hold in one case, then according to the theory that theorist knows; the dif-ference is not beyond his knowledge. But the question is not whether one of those theorists knows. The question is whether holding such a theory leaves anyone in a position to gain a satisfactory understand-ing of knowledge of the world, even if he fulfi lls the conditions Sosa’s


theory says are suffi cient for knowledge. Could someone in such a position come to recognize himself as knowing, and not merely con-fi dently believing, perhaps even truly, that sense perception is a way of getting knowledge of the world and crystal-ball gazing is not?

I think that, on the understanding of perception that appears to be involved in Sosa’s question about perceptual knowledge, the answer is “no.” On that view, what we are aware of in perception is restricted to features of our perceptual experiences. The external facts we know as a result of those experiences are nothing we ever perceive to be so. What we get in sense perception therefore bears the same relation to the world we think we know by that means as what is seen in crystal-ball gazing bears to the world the gazers think it gives them know-ledge of.

What we believe about the world goes well beyond the restricted domain of experience, and, if the connections are right, it is some-thing we know. But to know of the reliability of the connections which must hold in order for us to know, we can do no more than attend to what then comes within the restricted domain of our ex perience and, if we believe it, and if the connections are right, thereby come to know that those connections are reliable. So we can understand the position we are in only as follows: if there are reliable connections between our perceptual experiences and what we think we know, then we know by perception that there are external things. And we believe that there are such connections. But anyone who thinks that all it takes to have a satisfactory understanding of percep-tual knowledge is to conclude by modus ponens that we know by per-ception that there are external things would have to concede that the crystal-ball gazers have a satisfactory understanding of crystal-ball gazing knowledge. They could draw the corresponding conclusion equally confi dently from what they believe about themselves.

No comparable doubts affect the straightforward answer to the epistemological question we considered earlier. It says that we know that people know that there are external things by perceiving that there are external things. That is something we can see, and thereby know, to be true of human beings and other external things. So we have no trouble seeing and so knowing that there are no reliable con-nections between what people see in crystal balls and what goes on in the world beyond them. If the gazers could raise their eyes from their crystal balls and see what is so in the world around them, they could see that too. It might seem that Sosa’s “externalist” theorist is in a


superior position in this respect, because if he could see what is so beyond his limited perceptual experience he would see that what he believes to be so is reliably connected with his getting the experiences he gets. But no such vindication is available to him. It would be pos-sible only if he could see or otherwise perceive what is so in the world around him, not only the character of his perceptual experiences. And anyone who can do that thereby knows what is so in the world; he has no need for an “externalist” theory of knowledge to explain how his experiences enable him to know what he knows.

The conclusion I would draw from all this is that in order to achieve a satisfactory understanding of our knowledge of the world we must set aside or overcome the idea that the deliverances of perception even at its best are limited to the character of one’s perceptual experi-ences alone. Sosa appears to regard that idea as unthreatening as long as knowledge of the wider world is not thought of as a result of reasoning from that prior knowledge. But even without invoking rea-soning his “externalist” account still leaves us with something that is epistemically prior to any knowledge of an independent world. If there are no reliable connections between the perceptual experiences we receive and the world we believe in as a result of them, we know nothing of the wider world even though we know what experiences we are having. Perceptual knowledge of external things is seen as a combination of some prior knowledge which is not knowledge of external things plus something else. That is what I think leaves us in the plight I have described.

Perhaps I am wrong to attribute to Sosa without qualifi cation the view that we perceive at best only the character of our perceptual experiences. That seemed necessary to make sense of him as trying to answer the kind of question his “externalist” theory is meant to answer. He insists that knowledge requires “that one be adequately related, causally or counterfactually, to the objects of one’s know-ledge” (RK, 430). He thinks that would be so if there were reliable connections between one’s perceptual experiences and the wider world, so he appears content with a severely restricted view of the objects of perception. But one is also “adequately related . . . to the objects of one’s knowledge” if one sees, and in that way knows, that there is a table in the room, or a tree in the garden. One could not see that such a thing is so unless it were so; there being a table, or a tree, is a condition of anyone’s seeing that there is a table, or a tree, there. This is a form of “externalism” too, but not in the sense of


Sosa’s epistemological theory. Of course the table, or tree, alone is not suffi cient. As Sosa points out, “we must be both in good internal order and in appropriate relation to the external world” (RK, 430) in order to have knowledge of it. But seeing that there is a table, or a tree, is a highly “appropriate” relation in which to stand to the world in order to know that there is a table, or a tree. It is suffi cient for knowing such facts.

It would be no simple matter to say what “good internal order” a person must be in in order to see that there is a table in the room, even when a table is right before him in good light. It is a question of what it takes for him to be capable of having the thought, and so being able to recognize, that there is a table in the room. Exploration of those conditions would contribute to an explanation of how per-ceptual knowledge of external things is possible. But it would not explain it as a combination of some knowledge that is prior to any knowledge of external things plus something else. It would leave us in a position to say: “The existence of external things is not just an art-icle of faith; it is something we can see and thereby know to be true.” If Sosa would give that answer to his question, then again there is no disagreement between us. But that is just the straightforward answer; something everybody knows. It does not look like an answer to a deep and challenging question that we need an “externalist” or any other kind of philosophical theory of knowledge to answer.

INDEX

Aaron, R. I. 29 nn. 1 and 2 , 30 n. 4 Alexander, P. 29 n. 3 , 30 n. 4 Aristotle 211 , 220 Ayer, A. J. 4–5 , 69 n. 4 , 117–43 , 193

n. 12 on Hume on nature of

philosophy 118–24 ‘Humean’ aspects of his

views 124–40 prospects of his ‘Humean’

project 140–3 Ayers, Michael 3 , 31 n. 7 , 49–64

on colours as powers 58–64 on ‘linguistic theory of

psychological structure’ 49–50 , 57–8 , 64

on ‘phenomenal quality of experience’ 53–7 , 59 , 62–3

on ‘primary’ and ‘secondary’ applications of colour-terms 53–6

on the ‘private language’, ‘Wittgensteinian’ view 54–5 , 56–8

Bayle, Pierre 36 n. 10 , 190 Benacerraf, Paul 292 Bennett, Jonathan 29 nn. 1 and 2 ,

31 n. 7 Berkeley, George 2 , 3 , 13 n. 8 , 14 ,

29–48 , 101 , 117 , 180 , 185 , 225 , 226 , 241

on ‘ esse is percipi ’ 33 , 38–9 , 41–3 , 47 on ‘relativity of perception’ 33–4 ,

37–8 , 43–4 Boyle, Robert 29 n. 3 , 30 , 30 n. 4 ,

36 n. 10 , 40 n. 15

Brentano, Franz 250 , 251 Broughton, Janet 188 n. 15 Burge, Tyler 7 , 256–72

on knowledge of the world 256–66

on perception 256–8 and scepticism 260–72

Burnyeat, Myles 13 n. 8 , 16 , 17 n. 15

Carnap, Rudolf 227 , 228 , 245 n. 5

colours 3 , 49–64 , 176–7 perception of 50 , 75–6 as powers of objects 3 , 49 , 51 ,

58–63 , 77–8 predicates of 51–7 , 60–3

‘primary’ and ‘secondary’ applications of 52–4

‘subjectivity’ of 3 , 4 , 5 , 29 , 68 , 72 , 75–8 , 96–7 , 232

Cummins, P. 29 n. 2 , 30 n. 9 Curley, E. M. 29 n. 3 , 30 n. 4

Davidson, Donald 97–8 , 100 , 256 , 259 , 261 , 263–6 , 277–9 , 288 n. 48

Descartes, René 2 , 11–28 , 91 , 190 , 225 , 241

on ‘Cogito’ 2 , 18–22 , 26–8 and direct reference 26–8 on knowledge of the world

12–18 on method 11–12

Dretske, Fred 223 n. 2 , 251 n. 16

Epicurus 114 Evans, Gareth 280

336 Index

external world, knowledge of 2 , 7 , 8 , 12–18 , 307–21 , 322–34

Descartes on 12–18 ‘externalist’ accounts of 9 , 256–66 ,

299–300 , 327–9 , 331–4 dissatisfactions of 329–34 Fogelin on 310–11 , 323–5 perception and 7 , 8 , 13–18 , 256–8 ,

280–9 , 318–20 , 324–5 , 326–8 scepticism about 7 , 9 , 260–72 ,

308–18 , 326–8 Sosa on 323–5 , 327–34 Stoics on 308–9

Fogelin, Robert 9 , 307–21 on knowledge 9 , 310–11 on ‘possibilities of error’ 312–16 as Pyrrhonian sceptic 307–9 ,

311–15 Friedman, Michael 245 n. 5

Gettier, Edmund 313 Gibson, J. J. 30 n. 4 Ginsborg, Hannah 188 n. 15 Goldman, Alvin 8 , 290–306

on accounting for mathematical knowledge 292–6 , 303

on defi ning knowledge 292–6 , 303 on ‘primary epistemology’ 302–6

Green, T. H. 189–90 Grice, H. P. 175 n. 9

Hanson, Norwood Russell 229 Hare, R. M. 69 n. 6 Harman, Gilbert 297 n. 9 Hume, David 1 , 2 , 4–7 , 15 , 67–8 ,

70–3 , 79 , 80–1 , 86–93 , 167–88 , 208–10 , 225 , 253 , 254

on belief and judgement 185–7 , 205–6

on causation 68 , 110–11 , 132–3 , 169–72 , 181–3

Hume, David (cont’d ) Enquiry Concerning Human

Understanding 4 , 103–16 on evidence for miracles 112–13 on ‘fi ctions of the

imagination’ 198–202 , 242–3 on inference from observed to

unobserved 107–10 , 138–9 on judgements of value 67–8 ,

129–32 , 170–1 , 200 on liberty and necessity 111–12 and logical positivism 4–5 , 117–43 on natural religion 114–15 ‘naturalism’ in 87 , 189–206 ,

241–7 , 249 on ‘projecting’ sentiments 167–88 ,

200–2 scepticism in 4–6 , 115–16 ,

144–66 , 192–4 , 196–7 on ‘the science of man’ 86–93 ,

104 , 106–7 on ‘secondary’ qualities 63 n. 2 ,

174 Hutcheson, Francis 193 , 194 , 200

Jackson, R. 29 nn. 1 and 2

Kant, Immanuel 2 , 21–4 , 46 n. 22 , 123–4 , 244–6

on conditions of thought and experience 21–4

on thought and the world 99–102 transcendental idealism of

99–102 Kemp Smith, Norman 118 , 189–94 ,

200 Kornblith, Hilary 226 n. 5 Kripke, Saul 250 Kuhn, Thomas 229

Lewis, C. I. 15 Lewy, Casimir 257 n. 10

Index 337

Locke, John 3 , 29–39 , 49–64 , 117 , 119 , 185 , 190 , 225 , 226

on ‘primary’ and ‘secondary’ application of colour-terms 52–4

on ‘primary’ and ‘secondary’ qualities 29–30

on simple ideas of sensation 50–2

McDowell, John 8 , 273–89 vs Davidson 277–9 on the myth of the given 274 , 276 on perception 8 , 274–7 , 280–8 on the possibility of thought 273

Mackie, J. L. 29 nn. 2 and 3 , 30 n. 4 Maddy, Penelope 253–4 Malebranche, Nicolas 39 n. 14 , 241 Mandelbaum, M. 29 n. 3 , 30 n. 4 ,

31 n. 2 , 41 n. 16 mathematical knowledge 8 , 235–7 ,

252 , 291–2 , 295–301 , 306 Mill, J. S. 117 , 225 Moore, G. E. 93 , 234 , 237 , 313 , 323 naturalism 1 , 6–7 , 223–39 , 240–55

in Goldman 302–6 in Hume 87 , 189–206 , 241–7 , 249 in Maddy 253–4 in Quine 227–9 , 245–51

Newton, Isaac 86 , 119

O’Connor, D. J. 29 n. 2

Peacocke, Christopher 280 Pears, David 182 n. 13 perception:

knowledge by 13–18 , 256–72 , 280–9 , 299–300 , 305–6

objects of 14–16 , 262–3 , 274–7 , 280–9 , 328–9 , 333–4

reliability of 257–63 , 328–9 Plantinga, Alvin 226

Polanyi, Michael 229 Popkin, R. 29 n. 2 , 36 n. 10 , 154–7 ,

159 Price, H. H. 15 Prichard, H. A. 190 , 193 Putnam, Hilary 261

qualities, ‘primary’ and ‘ secondary’ 3 , 4 , 29–48 , 63 , 174

Quine, W. V. 90 n. 15 , 223 n. 1 , 227–9 , 253 , 254 , 259 , 315 n. 3

naturalism in 227–9 , 245–51

reasoning: from observed to

unobserved 107–10 , 138–9 practical 6 , 207–22

Russell, Bertrand 117 , 143

scepticism: Burge and 260–72 Fogelin and Pyrrhonian 307–9 ,

311–15 in Hume 4–6 , 115–16 , 144–66 ,

192–4 , 196–7 as not consistently

acceptable 269–70 and problem of external world 7 ,

9 , 258–72 Sosa vs 322–5 , 327

Scheffl er, Samuel 222 n. 16 Sextus Empiricus 160 n. 76 Smith, Adam 166 n. 110 Sosa, Ernest 9 , 322–34

on knowledge of external world 9 , 323–5

‘externalist’ theory of 327–9 , 331–4

dissatisfactions of 329–34 on objects of perception 328–9 ,

333–4

338 Index

Stevenson, C. L. 69 n. 5 Stroud, Barry 175 n. 9 , 266 n. 21 ,

307 n. 2 , 323 n. 2

Thomson, J. F. 29 n. 1 Tipton, I. C. 29 n. 1 , 38 n. 11 Turbayne, C. 31 n. 8 , 46 n. 22

value, judgements of 67–9 , 79–84 , 93–8 , 129–32 , 170–1 , 200 , 233–5

Hume on 67–8 , 129–32 , 170–1 , 200

‘subjectivity’ of 4 , 66–85 , 92–102 supported by ‘unmasking

explanation’ 73–5

Warnock, G. J. 29 n. 1 , 31 n. 7 Williams, Bernard 6 , 12 , 212 , 214 ,

219–20 Wittgenstein, Ludwig 97 , 123 ,

289 n. 49

Documents

Stroud - Philosophers Past and Present - Selected Essays (1)