Seeing, Knowing, Understanding: Philosophical Essays 0198809751, 9780198809753

Citation preview

OUP CORRECTED PROOF – FINAL, 26/4/2018, SPi

Seeing, Knowing, Understanding

OUP CORRECTED PROOF – FINAL, 26/4/2018, SPi

OUP CORRECTED PROOF – FINAL, 26/4/2018, SPi

Seeing, Knowing, Understanding Philosophical Essays

Barry Stroud

1

OUP CORRECTED PROOF – FINAL, 26/4/2018, SPi

3

Great Clarendon Street, Oxford, OX2 6DP, United Kingdom 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. Oxford is a registered trade mark of Oxford University Press in the UK and in certain other countries © in this volume Barry Stroud 2018 The moral rights of the author have been asserted First Edition published in 2018 Impression: 1 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, by licence 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 work in any other form and you must impose this same condition on any acquirer Published in the United States of America by Oxford University Press 198 Madison Avenue, New York, NY 10016, United States of America British Library Cataloguing in Publication Data Data available Library of Congress Control Number: 2017961723 ISBN 978–0–19–880975–3 Printed and bound by CPI Group (UK) Ltd, Croydon, CR0 4YY

OUP CORRECTED PROOF – FINAL, 26/4/2018, SPi

Contents Acknowledgements Introduction 1. What is Philosophy?

vii 1 7

2. The Pursuit of Philosophy: The Dewey Lecture

24

3. The Epistemological Promise of Externalism

40

4. Epistemological Self-Profile

55

5. Explaining Perceptual Knowledge: Reply to Quassim Cassam

62

6. Scepticism and the Senses

71

7. Seeing What is So

86

8. Perceptual Knowledge and the Primacy of Judgement

100

9. Doing Something Intentionally and Knowing that You are Doing It

114

10. Judgement, Self-Consciousness, Idealism

128

11. Feelings and the Ascription of Feelings

141

12. Kant’s ‘Transcendental Deduction’

151

13. Unmasking and Dispositionalism: Reply to Mark Johnston

167

14. Are the Colours of Things Secondary Qualities?

179

15. Concepts of Colour and Limits of Understanding

193

16. Logical Aliens and the ‘Ground’ of Logical Necessity

205

17. Ways of Meaning and Knowing Moral Realities

218

18. Meaning and Understanding

233

19. Davidson and Wittgenstein on Meaning and Understanding

255

Index

275

OUP CORRECTED PROOF – FINAL, 26/4/2018, SPi

OUP CORRECTED PROOF – FINAL, 26/4/2018, SPi

Acknowledgements The essays in this volume were originally published as follows: ‘What is Philosophy?’ in C. P. Ragland and S. Heidt (eds), What Is Philosophy? (Yale University Press 2001), pp. 25–46. ‘The Pursuit of Philosophy’, The Dewey Lecture, Proceedings and Addresses of the American Philosophical Association (2008), pp. 117–29. ‘The Epistemological Promise of Externalism’ in R. Schantz (ed.), The Externalist Challenge: New Studies on Cognition and Intentionality (De Gruyter 2005), pp. 181–91. ‘Epistemological Self-Profile’ in J. Dancy, E. Sosa, M. Steup (eds), A Companion to Epistemology, second edition (Blackwell 2010), pp. 190–4. ‘Explaining Perceptual Knowledge: Reply to Quassim Cassam’, European Journal of Philosophy (2009), 590–6. ‘Scepticism and the Senses’, European Journal of Philosophy (2009), pp. 559–70. ‘Seeing What is So’ in J. Roessler, N. Eilan, H. Lerman (eds), Perception, Causation, and Objectivity: Issues in Philosophy and Psychology (Oxford University Press 2011), pp. 92–102. ‘Perceptual Knowledge and the Primacy of Judgment’, Journal of the American Philosophical Association (2015), pp. 385–95. ‘Judgement, Self-Consciousness, Idealism’ in D. Emundts (ed.), Self, World, and Art: Metaphysical Topics in Kant and Hegel (De Gruyter 2013), pp. 37–46. ‘Doing Something Intentionally and Knowing That You Are Doing It’, Canadian Journal of Philosophy (March 2013), pp. 1–12. ‘Feelings and the Ascription of Feelings’, Teorema (January 2011) pp. 25–33. ‘Kant’s “Transcendental Deduction” ’ in J. O’Shea (ed.), Kant’s Critique of Pure Reason: A Critical Guide (Cambridge University Press 2017), pp. 106–19. ‘Unmasking and Dispositionalism’, Philosophy and Phenomenological Research (2004), pp. 202–12.

OUP CORRECTED PROOF – FINAL, 26/4/2018, SPi

viii

ACKNOWLEDGEMENTS

‘Concepts of Colour and Limits of Understanding’ in F. Gierlinger & S. Riegelnik (eds), Wittgenstein on Color (De Gruyter 2014), pp. 105–13. ‘Ways of Meaning and Knowing Moral Realities’ in G. Ortiz Millán y J. Cruz Parcero (eds), Lenguaje, Mente y Moralidad: Ensayos en homenaje a Mark Platts (Instituto de Investigaciones Filosóficas, Mexico 2015). ‘Meaning and Understanding’ in O. Kuusela and M. McGinn (eds), The Oxford Handbook to Wittgenstein (Oxford University Press 2011), pp. 294–310. ‘Davidson and Wittgenstein on Meaning and Understanding’ in C. Verheggen (ed.), Wittgenstein and Davidson on Thought, Language, and Action (Cambridge University Press 2017), pp. 123–38.

OUP CORRECTED PROOF – FINAL, 26/4/2018, SPi

Introduction I assemble here nineteen recent essays of mine on distinct but interconnected subjects. All but two of the essays have been published before; they appear here with the permission of their original publishers. They remain in the form in which they were published, despite some overlaps between them, and despite the urge to improve them in ways that now seem obvious. The first two essays stand somewhat apart from the others, at least in genre and scope. “What is Philosophy?” was presented at a conference on that question at Yale University and was published in C. P. Ragland and S. Heidt (ed.), What Is Philosophy?, Yale University Press 2001. Some contributors to that conference tried to answer that impossible question; others more or less ignored it. I remained within sight of the question, trying to explain what makes it so unmanageable, probably satisfying no one, including myself. “The Pursuit of Philosophy” is The Dewey Lecture I was invited to give to the Pacific Division of the American Philosophical Association in 2008. It was published in the Proceedings and Addresses of the American Philosophical Association in that year. The Dewey Lecture is meant to be autobiographical, sketching the significant outlines of a philosophical career. Although I believe the essay is a true account of the life it describes, I have found no academic person born after about 1950 who can believe for a moment that philosophical life was ever really like that. Essays three through nine pursue questions about the possibility and explanation of human knowledge. In The Significance of Philosophical Scepticism of 1984 and epistemological essays of mine in the 1980s and 1990s, I concentrated on the special character, and the depth, of the philosophical problem of our knowledge of the external world. I think that is the problem to be understood and fully appreciated first. In these six essays I go on to suggest a way of understanding how perceptual

OUP CORRECTED PROOF – FINAL, 26/4/2018, SPi



INTRODUCTION

knowledge of the world is possible without purporting to answer that traditional question. We must reject the assumption about perceptual experience that has made the problem so intractable. “The Epistemological Promise of Externalism” was published in R. Schantz (ed.), The Externalist Challenge: New Studies on Cognition and Intentionality, De Gruyter, Berlin & New York 2005. While granting the inevitable dissatisfactions of “externalist” definitions of the concept of knowledge, the essay begins to explore the promise of “externalist” or “anti-individualist” accounts of the contents of the thoughts and beliefs one must have even to be faced with that challenging problem. That strategy, and its role in overcoming the traditional problem, is further explained in “Epistemological Self-Profile”, a description of my work I was invited to contribute to the second edition of A Companion to Epistemology, edited by J. Dancy, E. Sosa, and M. Steup for Oxford University Press in 2010. “Explaining Perceptual Knowledge: Reply to Quassim Cassam”, is my reply to Cassam’s “Seeing and Knowing” which challenges some of the conditions Cassam thinks I have imposed on a satisfactory explanation of our knowledge of the external world. He argues that the conditions he specifies can be fulfilled in ways that explain how the knowledge is possible. I first heard this paper presented at a conference on scepticism in Edinburgh. The organizers of the conference then invited me to write a reply, and both papers were published in the European Journal of Philosophy in 2009. What is at stake is the conception of what is perceived to be so that is needed to account for the kind of perceptual knowledge we all know we have. That is what must be in question in any promising move away from the overly restrictive conception of perceptual experience that gives rise to the hopelessness of the traditional problem. I presented “Scepticism and the Senses” to that same conference in Edinburgh and it was published in the European Journal of Philosophy in 2009. It takes up directly that impossibly restrictive conception of perceptual experience and begins to outline what is needed for a more accurate and so more trouble-free account of what we can and do in fact perceive. What is needed is a distinction between perceiving an object— something referred to with a noun or a singular term—and perceiving that such-and-such is so—something expressed by a sentence. We can perceive particular objects without believing or knowing anything about them. It is only with such “propositional” objects of perception that

OUP CORRECTED PROOF – FINAL, 26/4/2018, SPi

INTRODUCTION



directly perceptual knowledge of the world is possible, since knowledge is knowledge of what is so. The importance of this distinction, and of the possibility of our having perceptions of the two different kinds, is central as well to “Seeing What is So”, in J. Roessler, N. Eilan, H. Lerman (ed.), Perception, Causation, and Objectivity: Issues in Philosophy and Psychology, Oxford University Press 2011, and “Perceptual Knowledge and the Primacy of Judgment”, in the Journal of the American Philosophical Association 2015. These essays elaborate and defend a conception of perceptual knowledge that draws on our capacity to apply predicates we understand to objects we encounter in perception. Exercising that capacity competently is a way of knowing by perception that such-and-such is so, where its being so or not is not necessarily dependent on its being perceived or even thought to be so by anyone. “Doing Something Intentionally and Knowing That You Are Doing It” was presented at a conference in London on the philosophy of action and published in the Canadian Journal of Philosophy in 2013. It draws attention to the parallel between the competent exercise of conceptual capacities involved in perceptual knowledge of the world and those competences essential to intentionally doing something and knowing that you are doing it. The ways of knowing what is known are different in the two cases, but the role of the conceptual capacities is strikingly similar in each. And in both cases, competent exercise of those capacities amounts to knowledge. “Judgement, Self-Consciousness, Idealism”, was my contribution to a conference in Berlin in honour of the work of Rolf-Peter Horstmann. It was published in D. Emundts (ed.), Self, World, and Art: Metaphysical Topics in Kant and Hegel, De Gruyter, Berlin 2013. The essay takes some steps into the question of the kind of self-consciousness that must also be present among the capacities required for propositional thought. Does it involve some kind of activity, or some awareness of something, on the part of the thinker, and if so, what activity, or awareness of what? This raises the spectre of idealism. Can the role of self-consciousness in thought be understood without implying that the truth of propositional thoughts is dependent in some way on the activities or experiences of the thinkers of them? “Feelings and the Ascription of Feelings” was my contribution to a workshop in Oviedo on David Finkelstein’s Experience and the Inner and

OUP CORRECTED PROOF – FINAL, 26/4/2018, SPi



INTRODUCTION

was published in Teorema January 2011. Here I take up Finkelstein’s sensitive treatment of some remarks by Wittgenstein about the conception of oneself, or the kind of self-consciousness, involved in ascribing feelings and sensations rather than thoughts or beliefs to ourselves. We express our feelings and sensations in various ways in the things we do, including saying or thinking of ourselves that we feel so-and-so. Do we refer to ourselves in such first-person remarks, and if so is what we say about ourselves something we can know to be true? And if so, how? “Kant’s ‘Transcendental Deduction’ was published in J. O’Shea (ed.), Kant’s Critique of Pure Reason: A Critical Guide, Cambridge University Press 2017. This essay is meant to present a straightforward structural description of Kant’s conception of what the transcendental deduction is supposed to do, and how it is supposed to do it. It explains the point and strategy of the “deduction” as Kant understands it, and the demanding conditions of its success, without entering into complexities of interpretation or critical assessment of the degree of success actually achieved. “Unmasking and Dispositionalism: Reply to Mark Johnston” is my response to Johnston’s “Subjectivism and Unmasking”, which was directed towards my The Quest for Reality. His essay and this one were both published in Philosophy and Phenomenological Research 2004. Johnston defends an ontological account of what colours are and explains how, on that view, it could be true that no colours belong to the everyday objects we perceive in the world. My resistance to the subjectivity of the colours of things turns rather on the proper understanding of colour terms as predicates ascribing colours to objects, not as names or terms referring to the colours. What is in question is whether Johnston’s conception of colours can make sense of our understanding people as perceiving and believing what they do about the colours of things. I have argued that we could not consistently recognize people as perceiving the colours of things they do while holding that the objects they perceive do not really have colours independently of perceivers’ responding to them in the ways they do. “Are the Colours of Things Secondary Qualities?” has not been published before. I include it as part of what I hope is still a fruitful exchange with John McDowell about the status of the colours of things. I now think the differences between us turn on our different understandings of perceptual experience, and on what each of us thinks is required for facts of the world to be directly available to us in sense-perception.

OUP CORRECTED PROOF – FINAL, 26/4/2018, SPi

INTRODUCTION



I argue that acknowledging that the colours of objects can be known to us in that direct way is incompatible with holding that the colours of things are secondary qualities. I regard the accounts of perceptual knowledge that I explain and defend in earlier essays in this volume as support for that conclusion. “Concepts of Colour and Limits of Understanding” was presented at a conference in Vienna on Wittgenstein on colour and was published in F. Gierlinger and S. Riegelnik (ed.), Wittgenstein on Color, De Gruyter, Berlin 2014. This takes up some puzzling reflections by Wittgenstein on the possibility of understanding concepts of the colours of things different from those already familiar to us. If we cannot understand them, does this show that, with colour, things must be this way; that we have the “correct” concepts? This encourages further questions about the “source” of whatever necessities we recognize among colour concepts. Wittgenstein raises similar difficulties about apparent “necessities” among other kinds of concepts. Is there an intelligible question here about necessity that we could find a way to answer? “Logical Aliens and the ‘Ground’ of Logical Necessity”, was presented at a conference in Porto on James Conant’s “The Search for Logically Alien Thought” twenty years after its publication. This essay is to be published in C. Travis & S. Miguens (ed.), Logically Alien Thought, Harvard University Press. This focusses mainly on Conant’s account of Frege’s conception of the special character of the laws of logic, and on whether or how that conception figures in Frege’s opposition to “psychologistic logicians” who apparently envisage the possibility of laws of logic contrary to our own. I question the “difficulty” Conant finds Frege involved in in arguing against that view, as well as his attributing to Frege some kind of explanation of the “source” or “ground” of the special character of logical laws. The challenge of explaining the necessity of what we regard as necessary remains. “Ways of Meaning and Knowing Moral Realities” was presented at a conference in Mexico in honour of the work of Mark Platts. It was published in Spanish in G. Ortiz Millán y J. Cruz Parcero (ed.), Lenguaje, Mente y Moralidad: Ensayos en homenaje a Mark Platts, Instituto de Investigaciones Filosóficas, Mexico 2015. This general outline of the structure of Platts’ “realist” or “objectivist” account of evaluative thought stresses its direct connection with the non-“epistemic” conception of meaning and understanding from which it is derived. This can explain

OUP CORRECTED PROOF – FINAL, 26/4/2018, SPi



INTRODUCTION

how exercise of the conceptual capacities involved in evaluative attitudes essential to intentional action can yield thoughts we regard as true or false independently of anyone’s actually taking those attitudes. This mirrors the conceptions of perceptual knowledge of the world and of one’s knowledge of one’s own intentional actions explained and defended in other essays in this volume. “Meaning and Understanding” was presented at a conference on Wittgenstein in Santa Cruz, California and published both in O. Kuusela and M. McGinn (ed.), The Oxford Handbook to Wittgenstein, Oxford University Press 2011 and in Jonathan Ellis and Daniel Guevara (ed), Wittgenstein and the Philosophy of Mind, Oxford University Press 2012. This essay tries to identify and explain an important source of the distinctive character of Wittgenstein’s philosophy: the idea that there is no understanding language, or the thoughts and attitudes we hold about the world, as he puts it, “from outside”. The general idea is present in Wittgenstein’s philosophy from the beginning, but it takes different forms at different times. What does this apparently uncontroversial remark even mean, and what does it imply about the kind of understanding we can expect to achieve of ourselves? An answer is exhibited in detail in Wittgenstein’s later writings, and in this essay I trace its development there and try to explain its significance while drawing out some of its implications. The importance of what looks like that same idea can also be recognized from different directions. The essay “Davidson and Wittgenstein on Meaning and Understanding” was published in C. Verheggen (ed.), Wittgenstein and Davidson on Thought, Language, and Action, Cambridge University Press 2017. Here I draw attention to what I see as a deep accord between Davidson and Wittgenstein on this matter, despite the different starting-points from which they reached it. This second essay focusses on Davidson’s account of meaning and its formal parallels with Wittgenstein’s conception of meaning as use. The first essay concentrates on the centrality in Wittgenstein of the fundamental idea of working “from within”. The two essays are closely connected, although written for different purposes, with different aims. There is inevitably considerable overlap between them.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

1 What is Philosophy? Faced, as I now am, with the question “What is philosophy?”, my first reaction is that the question is absurd. On further reflection, I find that it is not so much the question that is absurd as the attempt to answer it. Then, getting more personal, I realize that what is even more absurd is for me to try to answer it. But now, having reached this more personal perspective, I can shift the question a little, so it becomes not “What is philosophy?” but only “How do I see philosophy?” That is still daunting— how do I see philosophy?—but it feels more manageable. But then any answer I might give is not likely to be of wide interest. It will be like the message of a flea, working away at his own square inch of flesh, reporting on “How I see the elephant.” Trying to answer the large, impersonal question is absurd or discouraging in the way that it is absurd or discouraging to try to answer questions like “What is painting?” or “What is music?” It is not that nothing can be said, even something that is true. But what can be said that would be helpful to someone who wants an answer? For one thing, painting and music and philosophy are human activities with a history. They are as they are here and now because of the ways they were elsewhere and in earlier times. A large part of understanding them lies in understanding how they got to be the way they are from what they were before. But the completely general questions seem to ask simply “What are they?” And someone who asks that question out of complete ignorance—someone from Mars, say—is not likely to be satisfied with a story of how one kind of painting or music or philosophy led to another, and eventually to what we have here and now. This brings out part of the absurdity in trying to answer the completely general question: it is left indeterminate what someone who asks the question wants to know. How much is such a person supposed to know already, or not to know, about human life? Where are we allowed to start in trying to tell someone what painting or music or philosophy is?

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



WHAT IS PHILOSOPHY ?

Someone from Mars, say, who has never visited this planet, might be satisfied with being told that here on earth some of us paint, and that painting is marking a surface with colour, or that we enjoy music, which is sound and silence ordered by a composer in certain ways. That is true, as far as it goes, but it is probably not going to satisfy anyone a little closer to home who asks what painting, or music, is. And the same is true of “What is philosophy?” or even of my more personal question: “How do you see philosophy?” Another part of the absurdity in trying to answer the completely general question is that philosophy is many different things, and it has been conducted in many different ways. There is no reason to insist that it be only one thing, or that it can be summed up or defined in some illuminating way even for those who know a great deal about human life. That is why I restrict myself to saying how I see philosophy. But there is another disclaimer to make even about that. I am suspicious of, and not very interested in, what painters say about their own paintings, for example. Technical details of execution are fine, materials and procedures fascinating, but when it comes to what the painting is really “about” or what the painter was really doing—not to mention what painting is in general—I don’t think the painter is in any special position. The best thing to do is to look closely at the painting itself. That is what shows what it is and what it does, and what the painter was doing in painting it (if anything shows it). I have the same feeling about philosophy. It is much better—more reliable—to look at what a philosopher does in a particular case than to listen to what he or she says about what is being done. It is in the actual work that a philosopher’s conception of philosophy, or his or her way of doing it, is really to be found. So I am suspicious of pronouncements, even by philosophers, about what philosophy is, or how it ought to be done. It is a real philosophical problem, I think, to say what is going on in a piece of philosophizing. The philosopher who is doing it is not necessarily in a privileged position about it. With all these reservations in mind, I can nonetheless try to say something about what philosophy is—or how I see it. Just as it is true, but not very enlightening, to say that painting is marking a surface with colour, so, at that same level of banal truth, we can say that philosophy is thought. It is reflection on very general aspects of the world, and especially those aspects that involve or impinge on the lives of human beings. Human beings are those who carry out philosophical reflection, so it can

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

WHAT IS PHILOSOPHY ?



be said to be a form of self-reflection. But not just any thought—even any self-reflective thought about very general aspects of the world that involve or impinge on the lives of human beings—is philosophy. Nor is all marking a surface with colour painting, or all ordered combinations of sound and silence, music. This does not take us very far toward an answer to the question of what is special or unique about philosophy, but it does give us something. It points to something important, whatever the special or unique characteristics of philosophical thinking turn out to be. The important, but not always acknowledged, fact is that philosophy does not necessarily exist, and certainly does not thrive, in every society or culture. But every society or culture surely has some set of ideas about the way things are in general, and how they do and do not impinge on the lives of human beings. Some such general conception of the world, and some conception of themselves on the part of human beings, seems essential to, or at least universal in, all human life. All human beings have certain interests and concerns, and certain fears and aspirations, which in one form or another are present in human life wherever it appears. And something institutional or social is needed for human beings to cope with or pursue those natural concerns, and to provide ways for groups of people to live together in some kind of unity. That is a large part of what a culture is, or does. But I do not regard the contents of all such cultural formations as philosophy, or each one of them as a philosophy. Every culture has thoughts and attitudes about fundamental features of a human being’s relation to the rest of the universe, about death and human finitude, and about the appropriate ways of interacting with other human beings. Those are the kinds of things philosophy is about. But not every way of coming to terms with them, even every way of thinking about them, amounts to philosophy. This is obvious in one way simply from the fact that there are a great many societies, in both the past and the present, for whom there are more important things than philosophy. Concerns about the life and death of human beings, about their relations to the world and to other human beings, must be attended to before the possibility, or even the conception, of philosophy could begin to arise. Philosophy exists, or flourishes, only under certain social conditions, not all. But if societies in even the most desperate circumstances, where there is no place for philosophy, nonetheless need certain shared understandings or ideas of

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



WHAT IS PHILOSOPHY ?

themselves and their relation to the rest of the world in order to cohere and function at all, then the presence of such ideas does not necessarily amount to philosophy. Obviously, a society must be pretty advanced, compared to the minimum conditions needed to function or even to flourish, in order for that society to support or make room for philosophy. Philosophy needs reasonably benign social conditions, and a relatively high level of economic development, at least for some of the people. But that alone is no guarantee that philosophy will have an official place in such a culture. Of course, philosophy can be done outside, or even against, an official culture. But that happens only where a tradition of previous philosophizing has not been completely extinguished. And even then it is difficult, and in the end does not lead to much in philosophical terms, except possibly keeping the tradition somehow available until more promising times. Those repressive states that have attacked or tried to suppress philosophy by turning their unemployed and harassed philosophers into street sweepers or window washers do not lack a set of ideas that they expect their citizens to share and to guide their lives by. The doctrines they promulgate are general ideas and principles about the ends and value of human life and how best to achieve them. But they are not for that reason alone what I would regard as philosophy. They might even be called “philosophy,” at least by the state that is trying to enforce them. But that doesn’t mean that such states, or even their subjects who accept the ideas and live by them, are supporting or engaging in philosophy. A society must be not only fairly advanced, but fairly comfortable with itself, to provide an official place for the best that philosophy can offer. By now, for us, it is hard to think of any place other than a university where that might be done—at least systematically and communally. That was not always so. The university has not been with us as long as philosophy has. But the age of royal or aristocratic patronage—or any form of private patronage—is behind us. Much work now being done in philosophy is supported by forms of social patronage other than the university. Resources that could obviously be put to pressing social uses sometimes go to philosophers. This is one mark of the advanced state, even the civility, of our Western culture. It is also perhaps a mark of our society’s attitudes toward some of those pressing social issues. But the university now is the place where philosophy thrives.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

WHAT IS PHILOSOPHY ?



This strikes me as both a good and a bad thing. It is bad because, as the university becomes increasingly professionalized, it has increasingly professionalized philosophy. This, in my opinion, has rendered much more of philosophy sterile, empty, and boring. What institutions demand from individuals, institutions get. And what universities, even the best universities, now demand from individual professors, on the whole, is quantity of publication, frequency of citation in the professional literature, widely certified distinction in the profession, and other quantifiable measures of an impressive resumé. And that is what professors of philosophy are on the whole providing—to the detriment of philosophy as I see it. But the connection between philosophy and the university is a good thing, too. It provides a place in which philosophy can be done, and with wise leadership can be done in conditions in which it can thrive. It requires teaching: the need to say clear things in public, and to make them accessible to others and subject to critical assessment. And teaching is what passes on the subject, or the tradition, if only in the form in which it is received by those now entering into it. The university also provides the freedom to think about something for a long time, and in new ways that might show no beneficial social effects. Philosophical work (even work that seems to go nowhere) is officially supported by the university, and so by the society at large, and not for any of the specific conclusions it is expected to reach. It is regarded as more important that the activity should go on than that it should have this or that specified outcome. Results, in the form of conclusions reached, or propositions established, are not what matters. This is a good thing as I see it because I do not regard philosophy as a set of results or doctrines, in the sense of conclusions reached, or propositions established. It may well be that every society or culture needs ideas or beliefs or doctrines to give sense and direction to the lives of its members. And accepting a belief or doctrine, or espousing an ideology, is a form of thought, of thinking. But it is not the kind of thought that philosophy is. Or rather, that kind of thought is compatible with the absence of philosophy, or of what is valuable about philosophy at its best. Religion, too, is a form of thinking, and is certainly about matters of great concern to human beings. But I think religious belief or religious thought is not the same as philosophy, or philosophical thought. As I put

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



WHAT IS PHILOSOPHY ?

it a moment ago, it is compatible with the absence of philosophy, even though it includes beliefs, attitudes, and precepts that are based on general ideas about human beings and the world. Such thoughts and beliefs express a conception of the world, or a creed, but philosophy as I see it is not a creed. Nor is it the effort to discover or produce a creed. I think the attitudes or motives out of which philosophy is pursued are not the same as a typically religious attitude or orientation. A philosopher is not fundamentally at peace with the world. He does not, as a philosopher, regard himself as somehow ultimately in good hands, and safe, and so willing to acquiesce in something that can or must be accepted without being understood. Of course, a philosopher, like everyone else, must acquiesce in the face of the facts—or at least of the facts that cannot be made otherwise. But philosophy depends on undying curiosity, and the pursuit of limitless enquiry. It arises out of a wish, or an attempt, to grasp the world as it is, as it is open to our view. The difficulty is to get the right, or the intellectually satisfying, view of it. Philosophy is interesting and challenging to me personally only insofar as there is no other-worldly theological account of things. If I really thought there were, I believe I would not find serious thought about such matters interesting or rewarding in the way philosophy can be. The very effort would seem to founder on the epistemic difficulty of how far we could hope to penetrate the ultimately inscrutable ways of God or whatever else was thought to be in charge. But even if I went in for such thoughts and beliefs, and they were important for the way I lived my life, I do not think that in accepting them I would be engaged in what I regard as philosophy. Not because philosophy as I think of it is idle and can have no effect on one’s attitudes and one’s relations to the world and to life. I mean only that philosophy as I understand it is different, and that it has its effects, if it does, in other ways. It is not a matter of arriving at conclusions that are applied to or used to guide or order one’s life. Philosophy is thought, or reflection, that is done purely for the sake of understanding something, solely to find out what is so with respect to those aspects of the world that puzzle us. The activity is in a certain sense endless, even if it ends for each human being who engages in it. But that does not mean that it does not issue in anything, or that it has no effects. It is just that the effects, if they come, do not take the form of discoveries of conclusions or doctrines which serve to direct or guide one’s life.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

WHAT IS PHILOSOPHY ?



When I say that philosophy as I see it is not a creed, and not the search for a creed or set of doctrines—and so not a social or political creed or ideology, and not a religious creed or set of attitudes either—I also mean that it is not the search for what might be called a philosophical creed or a set of philosophical doctrines or theses either. Philosophy as I see it is an activity, not a set of doctrines or truths at all. Nor is its point or goal to discover philosophical theses or doctrines. There is in that sense no such thing as, for example, twentieth-century philosophy or eighteenth-century philosophy, in the way there is such a thing as twentieth-century physics or present-day chemistry or molecular biology or seventeenth-century physics. Of course, there is lots of disagreement and uncertainty in those fields, and lots of activity leading off in different directions. And that has been true at every stage of their development. But there is such a thing as what all current physicists or chemists know: a body of doctrine that can be called truths of physics or chemistry as those subjects now stand. I don’t mean that none of it will ever be changed, but only that there is a body of accepted truth that physicists or chemists share. It is the core, for now, of what physics, or chemistry, says: what is so, physically or chemically speaking. It is in that sense that I think there is no such thing as a body of philosophical doctrine, or truth: what is so, philosophically speaking. There is no such thing now, and there was no such thing in the past. Many people will agree because, as they are happy to point out, philosophy makes no progress, and philosophers do not agree about anything—or at least not much. That is said to be because philosophers are inveterately, perhaps even inherently, disputatious, and because their problems are unreal, since there are no facts in philosophy, and no way to verify or falsify its claims. According to this idea, even when many philosophers do agree about something, that is all it is—widespread agreement, rather than the discovery of philosophical facts or truths. In some circles now the view of philosophy I have just described is said to be true of all intellectual efforts to understand anything. There are held to be no facts or truths of any kind, even in physics or chemistry. There is just more widespread agreement in certain areas than in others, which leads us to speak of “facts” or “discoveries” in physics and chemistry and such fields, but not in politics or morals or philosophy. That usage itself is seen as just something else that most people agree about.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



WHAT IS PHILOSOPHY ?

This more general view, if it can be called that, can perhaps make philosophy look not so bad after all, at least comparatively speaking. It turns out to differ from other intellectual pursuits only in the amount of agreement to be found in it. This could even make philosophers look good: maybe they are more courageous and more independently minded than those flocks of physicists and chemists and mathematicians who (on this view) are sheepishly going along with whatever consensus seems to be growing. But I will not dwell further on the sad story of this fashionable conception of human thought. I think there is a different explanation for why there is no such thing as a body of doctrine, or a core of discovered truth, that represents the results of twentieth-century philosophy, or even eighteenth-century philosophy. I think it is because philosophy is just not the same kind of enterprise as physics or chemistry or other ways of coming to know about the world. It is true, of course, that certain topics or questions dominate philosophy in certain places and times, and certain ways of dealing with them prevail and then give way to other topics and methods. But those are interests and approaches and procedures, not results. They do not include a basic core which anyone doing philosophy at a certain time and place knows or accepts, and which represents philosophy’s achievement up to that point. There is of course a tradition of philosophical works, just as there is in painting and music. It is the works, and not something recognizable as their results, that constitute the tradition. That is why philosophy, like painting and music, must be understood at any given time at least in part historically. There is no way of understanding the issues and what is at stake other than seeing where they come from and why they present themselves at that time in the ways they do. The need for a past, for a tradition to start from, is a commonplace of the history of painting. Painters paint as they do in response to the painting that precedes them. André Malraux dramatized the familiar point this way: It is a revealing fact that, when explaining how his vocation came to him, every great artist traces it back to the emotion he experienced at his contact with some specific work of art. . . . Never do we hear of someone who, out of the blue so to speak, feels a compulsion to “express” some scene or startling incident. . . . An old story goes

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

WHAT IS PHILOSOPHY ?



that Cimabue was struck with admiration when he saw the shepherd-boy Giotto, sketching sheep. But, in the true biographies, it is never the sheep that inspire a Giotto with the love of painting; but, rather, his first sight of the paintings of a man like Cimabue. What makes the artist is that in his youth he was more deeply moved by his visual experience of works of art than by that of the things they represent.¹

Something like this is no less true, I believe, of philosophy. G. E. Moore reports in his autobiography that he thinks “the main stimulus to philosophize” for him was “certain philosophical statements which [he] heard made in conversation.” He says, “I do not think that the world or the sciences would ever have suggested to me any philosophical problems. What has suggested philosophical problems to me is things which other philosophers have said about the world or the sciences.”² This is often cited as a defect or limitation of G. E. Moore and his philosophy, and indeed, sometimes, of so-called analytic or linguistic philosophy in general. It is thought to reveal the superficial and derivative, even parasitic, character of that philosophy, its dry academic or professorial sources, and its distance from the “real” problems that the world presents to any thoughtful, sensitive human being who is directly and passionately engaged with it. I think rather that the observation shows Moore’s acuteness, and his honesty. Whatever one might think of Moore’s limitations, perhaps even his blindnesses, as a philosopher, he was not blind or limited in this case. What he reports about himself is something I believe to be true of philosophy in general. I find it interesting that many philosophers would deny it, and would regard it as demeaning or indicative of shallowness to acknowledge that it is true of them. Of course, those philosophers will concede that it is necessary to become acquainted with philosophy, to learn something about it, if you want to do something in it. But many would deny that what they do is derived from previous philosophy in a way they might grant that present-day painting or music are derived from their pasts, and not fully intelligible apart from them. Many philosophers say they are just ¹ André Malraux, The Voices of Silence (tr. Stuart Gilbert), Princeton University Press 1978, p. 281. ² G. E. Moore, “Autobiography,” in The Philosophy of G. E. Moore, ed. Paul A. Schilpp, Open Court 1968, p. 14.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



WHAT IS PHILOSOPHY ?

interested in solving certain problems that present themselves, or in understanding certain phenomena, or putting forward correct theories of this or that aspect of the world. They are not concerned with what was said about such matters in the past, although they grant that others, even others in philosophy, might be interested in such questions. W. V. Quine, for example, has remarked that people go into philosophy for different reasons: some are interested in the history of philosophy, and some are interested in philosophy. He would put himself in the second category. But the fact that Quine is not primarily interested in studying the history of philosophy does not mean that he philosophizes in ignorance of or in isolation from that tradition. In fact, Quine knows a great deal about the history of philosophy. It is just that the history of philosophy for him was Carnap, and Principia Mathematica. He knows both of them very well. His philosophy is in large part a response to them. There is a lot more of C. I. Lewis in it than he would care to admit, as well. The same sort of thing is true of any theorists who claim to be doing nothing more than directing their philosophically unfettered minds onto problematic aspects of a more or less neutrally describable public world. There are many philosophical theorists, and philosophical theories and doctrines. That is to say, many philosophers describe what they are doing in those ways. What puzzled G. E. Moore was the relation between the kinds of things such philosophers say and facts of the everyday world that we all know. That seems to me to be the right kind of question to ask about those things called “philosophical theses” or “doctrines” or “results.” What do they imply about what we all know, or about what is so, outside philosophy? Does “Time is unreal,” for example, imply that I did not have my breakfast after I got out of bed, or not? And if it does, isn’t it simply false, since I did get out of bed first this morning? I find it exhilarating when philosophers raise, and press, questions like this, even about so-called philosophical theories themselves. I say this is the right kind of question to ask, or at least to start from, since it leads to questioning the very nature of a “philosophical” thesis or doctrine or theory. It is a way of pursuing our question “What is philosophy?”, but from down on the ground, as it were, as it presents itself in particular real cases. This is not something that much concerns many of those who are busy producing, or at least seeking, theses or theories in philosophy. It seems to be very important to many

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

WHAT IS PHILOSOPHY ?



philosophers to have some position or doctrine or theory with which to identify themselves. This now seems to dominate at least that part of philosophy that is conducted in English. Theories are constructed, or proposed, and philosophical discussion amounts to pitting these positions or theories against one another. It is apparently thought to be a virtue to stick by your theory as long as you can, making adjustments under pressure only if opposed theories seem to be gaining an advantage, and to keep trying to give your own position the best run for its money in the competitive marketplace. This serves a professional enterprise very well. Each operator has his or her own niche in the subject, and so a certain professional identity. And the relative fortunes of each player’s project can be charted to some extent. The whole picture appears to be modeled on a certain conception—no doubt a mythical conception—of how real science proceeds, with a further economic metaphor now added to the myth. One hears a lot in philosophy these days of “buying into” this or that theory or “ism” or “research program.” When I first went to California in the flourishing 1960s, graduate students who found a new idea or suggestion promising might say “I’ll buy that.” Now the most they can afford is to buy “into” something. They are prepared to put what money they have into “anti-realism” versus “realism,” say, or “externalism” versus “internalism,” and then try to defend their investment against all comers. This can sometimes lead to something new and worthwhile. Philosophy throughout its history has thrown off many questions and programs that have become self-sufficient subjects or even sciences of their own. A chair in physics in Oxford today is still called the chair of “Natural Philosophy.” But it is not within the Sub-Faculty of Philosophy. This might be one thing that accounts for the fact that there are no results in philosophy: as soon as there are real results, it no longer counts as philosophy. But I think the professionalized, scientistic conception that many people now have of how to proceed in philosophy is unfortunate. Despite many impressive achievements of sheer brain power, it remains, to my mind, unsatisfactory and, if not unphilosophical, at least not sufficiently philosophical. In the terms I used earlier, I would say that it is compatible at a certain point with the absence of philosophy. It has led to what I think is a certain complacency, even a certain blindness, in the face of what remains philosophically important. I don’t mean that the best constructive theoretical philosophers are complacent in defence of their

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



WHAT IS PHILOSOPHY ?

theories or blind to possibly damaging objections to them. But I think that in philosophical enterprise or penetration they tend to stop one step too soon. They are insufficiently critical of what a philosophical thesis or theory amounts to, and what it means to adopt or accept such a thing. That makes philosophical theorizing look too much like accepting an ideology or a creed or a religion. It does not take into philosophical consideration the very attitude of so-called acceptance itself, or the very status of the doctrines or theory accepted. It takes for granted that we understand what sort of thing a philosophical doctrine or theory is, and asks only which ones to accept and which to reject. I find this unsatisfactory, or not sufficiently philosophical, because I think that what is then taken for granted is something we do not understand: what is this enterprise we philosophers are engaged in, and what can be expected from it? That is a form of the question “What is philosophy?” and it is a question for philosophy itself. Who else will address it? Philosophical theories about this or that aspect of the world do not address it. They are instances of the very thing it asks about. That is what we want to understand better than we do. What philosophical theories or theses are, or what the answers to philosophical problems say or mean, can only be as clear and as well understood as the questions they answer or the phenomena they are meant to account for. But those problems, or the conceptions that give rise to them, are themselves the results of thought—of thinking of things in certain ways. And to understand those problems, even before trying to answer them, we have to identify and understand those ways of thinking, and assess them. But those ways of thinking are also at least in part the product of previous philosophizing, and that too we have to identify and understand. Only in this way, I think, can we come to know what we are doing in philosophy. We have to get at the sources of the so-called problems, to see where they come from and why they take the forms that they do. This is what I think is left out or simply taken as known by those who are busy answering the questions, solving the problems, or producing theories of heretofore problematic phenomena. This is something I think we can no longer take for granted in philosophy. It is something that many theories since the time of Kant have tried to account for. Those theories by now have been repudiated. And rightly so. But putting scientistic slogans in their place is no advance. What we need now is a much closer look at what actually

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

WHAT IS PHILOSOPHY ?



goes on. But not in general—in detail. We need to see where philosophical problems come from in order to understand the special character of what is said in attempts to solve them. One word sometimes applied to the kind of philosophical attitude or curiosity I am recommending is therapeutic. The term is unfortunate, and does not capture what I have in mind. Wittgenstein does say that there are different methods in philosophy, “like different therapies.”³ But that might mean, allowing for the bad grammar, only that there are different methods in philosophy, just as there are different therapies in psychotherapy. That does not imply that the different ways of doing things in philosophy are therapies. On that reading, the remark means only the equivalent of “There is more than one way to skin a cat.” With a cat there is no obscurity or uncertainty about the goal: you either end up with a fully skinned cat or you don’t. But what is the analogue, even in psychotherapy, of the skinned cat? And if you think in that case that you can identify in advance a clear goal that could be reached by several different therapeutic means, what is the parallel in philosophy? What goal are different methods in philosophy all designed to achieve? Therapeutic as a term puts too much weight on an identifiable outcome. I prefer the word diagnostic. What philosophers need now is a diagnosis or uncovering of what they regard as their problems or their questions, and some understanding of the nature and sources of the kinds of things they think philosophy should account for. Wittgenstein also says, as translated by Elizabeth Anscombe: “The philosopher’s treatment of a question is like the treatment of an illness.”⁴ This suggests that the philosopher, or the philosopher Wittgenstein approves of, treats a philosophical question like an illness—or even that he thinks a philosophical question is an illness. I think that is an unfortunate suggestion. The thing to do with an illness, after all, is to get rid of it. Hence the idea of therapy or cure. But, as we have just seen, Wittgenstein also says there are many different therapies, and you can get rid of something in lots of different ways. Why would “philosophical methods,” if there are any, be any better than anything else that works? Best of all would be not to get the thing in the first place. And with philosophical ³ Ludwig Wittgenstein, Philosophical Investigations (tr. G. E. M. Anscombe) Macmillan 1953, §133. ⁴ Ibid §255.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



WHAT IS PHILOSOPHY ?

questions, that can be arranged. So on this reading, Wittgenstein would be suggesting that nothing would be lost without philosophy. I think Wittgenstein’s remark should be taken in another way, which puts it much closer to the conception of philosophy or philosophical activity I think we now need—or need to resuscitate. It makes a place for philosophy. What Wittgenstein writes, in the German, is: “Der Philosoph behandelt eine Frage; wie eine Krankheit,” which could be put something like this: “The philosopher treats a question; as an illness (is treated).” The stress is on the verb. The philosopher treats a question; the doctor treats an illness. The parallel is with what is done, not necessarily with what it is done to. Well, how is an illness treated? First of all, and crucially, it has to be identified. “What have we got here, exactly,” we ask, “and how does it differ from other things that are very similar but different?” These symptoms must then be diagnosed. What are they indications of? What lies behind them? How did things develop so that these symptoms show up in this form here and now? The time for therapy and cure can come only after these questions have been answered. The point is that treatment begins with identification and understanding, and an illness to be treated is understood in terms of its origins or causes. We understand what it is in terms of how it came to be. Treating a question is not the same as answering it. Answering it might be the worst thing to do with it. I believe that happens in philosophy. Answering a question before understanding what it is and where it comes from is like applying a cure for an illness without having identified what illness it is. It can make things worse, and certainly harder to identify and understand. To understand what philosophical theses or theories are, we have to understand the nature and sources of the problems to which they are answers. And to understand those problems we have to identify and understand their sources. I do not mean that we must always seek their temporal or historical sources. We need to identify the assumptions, the demands, the preconceptions, and the aspirations that lead to a question’s having the particular significance it now has for us, but that need not mean going back to earlier stages of the philosophical tradition. The particular frame of mind that is responsible for the question is something we are in right now. That does not mean it is therefore easy to identify. But given that philosophy is always in part a response to previous philosophy, we can be

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

WHAT IS PHILOSOPHY ?



pretty sure that the sources of questions that now lie within us, and to that extent seem uncontroversial, are the products of earlier philosophizing or earlier ways of thinking about the world. That is why I think philosophy is inseparable from the history of philosophy. But not every particular attempt to plumb the sources of a philosophical problem must take us backward in time. The important thing is to gain some understanding of the origin and special character of the problem or issue right now, whatever that takes. So philosophical investigation as I think of it should extend in this way to the process or activity of philosophy itself. An unrelenting selfconsciousness is essential to the task. I think only the application of philosophical self-reflection to the very procedures and products of selfreflection can reveal what philosophy is, or what it can offer. But there always has to be something more to reflect on, or to start from, than just the activity of philosophizing itself. There has to be something we think, something we are trying to understand, some puzzling phenomenon or aspect of the world. There must be some “data,” so to speak, to reflect on, or to come to terms with, even if they are only the way a problem presents itself to us, or a felt need to understand things in a certain way. There must be some things we know, or things that are so, or things that we think and cannot deny, that are firmer than anything that philosophical theorizing can undermine. I don’t mean anything here like “foundations” of knowledge, or “selfevident truths.” I mean facts, convictions, or attitudes that we are engaged with as mature thinkers and agents in a public world—something we are involved in, or subject to, as active, feeling, human beings, and which philosophy is presumably in some way meant to help illuminate. There must be some things which, in Thomas Nagel’s phrase, we “think straight.” There must be something we are involved in that is not philosophy. In reflecting on such things philosophically, we have to have the strength to recognize and hold on to such things, and not to distort or deny them, in the face of philosophical reflection. Otherwise, whatever importance or interest philosophical thought can have is lost. This is what I find admirable about G. E. Moore, for all his limitations in recognizing what a philosophical attack on such things might be up to. There is virtually no philosophical diagnosis in Moore; there is just the steadfast insistence that this is true, that is false, and that we all know such-and-such. On its own, as it appears in Moore, that can be

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



WHAT IS PHILOSOPHY ?

philosophically dissatisfying; it can seem shallow and uncomprehending. But without some such engagement and steadfastness about something, without something we unshakably think to be so, at least at the time, philosophical self-reflection on its own becomes vulnerable to complete self-absorption—what Christopher Ricks has called “narcissistic regression.”⁵ If we focus only on the things we are inclined to say now, and why we are inclined to say them, and why we are inclined to say that about our present inclination, and so on, we are left with mere play. We have nothing to think about except the rhetorical effects of the rhetorical devices we have employed in previous responses. We would then have lost the connection between our thinking and anything else in the world, especially anything that matters. That really would be, again in Ricks’s words, a kind of “kamikaze of the intellect.”⁶ So I think that in philosophy we have to look at each problem or issue or phenomenon, and respond to it authentically as it presents itself to us at the time, without denying or dismissing what we honestly cannot deny or dismiss. We cannot simply accept it without scrutiny, and we cannot simply declare the problem bogus or the alleged phenomenon illusory on the basis of some quite general theoretical belief about the contrived or contingent sources of philosophical problems. We have to take the particular issue as seriously as we can bring ourselves to feel it. We have to participate in, not just comment on, philosophical thought. If we do that, what do we get? Or what can we expect? Again, I think nothing illuminating can be said in general terms. If we hold on to our attachment to the world, what we can hope for in a particular case is at least a sharper awareness, and so a fuller appreciation, of the detail and complexity of those aspects of the everyday world that gave rise to philosophical reflection about them. By seeing the distortions, or even perhaps the impossibility, of a detached, theoretical account of something, we might come to appreciate better what the philosophical enterprise aspires to. To see how and why we can never achieve such a point of view on the world might provide its own kind of human or self-understanding—perhaps the best or the only kind we can really hope for. But for any of this to be possible—and this is my main point—we have to participate in the activity of philosophy. There is no shortcut, and ⁵ Christopher Ricks, T. S. Eliot and Prejudice, Faber and Faber 1994, p. 21. ⁶ Ibid., p. 90.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

WHAT IS PHILOSOPHY ?



no formula. So to the completely general, detached question “What is philosophy?” I say “Don’t ask; don’t tell.” It is a question to be treated, not answered. Trying to answer it in that general form gets you nowhere. You have to look at some particular bit of philosophy, or better still do some philosophy—think about something, and try to get to the bottom of it—and then ask yourself what is going on. “What is really being said and done here,” we have to ask, “and what does it imply about this or that undeniable fact of the world?” I think philosophy promises most only if you do this each time the philosophical urge arises, or each time a philosophical issue comes along. This is especially true when it comes along in your own thoughts, and in your own voice.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

2 The Pursuit of Philosophy The Dewey Lecture

It is a great honour to be invited to give The Dewey Lecture. I am thrilled by the opportunity, and I am grateful to the Pacific Division of the APA for appointing me. I hope I can live up to what is expected. I learned, mostly after I had agreed to do it, that it is an especially daunting task (at least for me) for a number of different reasons. The Dewey Lecture at the APA is something relatively new. It is not meant to be a more-or-less standard philosophical talk. Rather, the Dewey Lecturer is expected to reflect on his life and work, on his education, his teachers and mentors, on the origins and development of ideas and approaches that have been influential, and in general on a career in philosophy as a discipline or profession. What is wanted is a brief but revealing philosophical or intellectual autobiography that gives a sense of the sweep of a philosophical career. It is both the word ‘revealing’ and the word ‘career’ here that give me pause. As for ‘revealing’, there is a standing danger that must be acknowledged and faced up to right from the beginning by anyone who engages in elderly, avuncular reflections of this kind. The danger has been identified in perhaps its most frightening form, characteristically, by Nietzsche. It quite often happens that the old man is subject to the delusion of a great moral renewal and rebirth, and from this experience he passes judgements on the work and course of his life, as if he had only now become clear-sighted; and yet the inspiration behind this feeling of well-being and these confident judgements is not wisdom, but weariness.¹ ¹ F. Nietzsche, Daybreak, sec. 542 (quoted in Bernard Williams, Philosophy as a Humanistic Discipline, Princeton University Press 2006, p. 198).

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

THE PURSUIT OF PHILOSOPHY



This raises questions about who might be interested in hearing such a lecture and, more disturbingly, why they would be interested. Maybe most of you are here out of a more-or-less clinical interest. In any case, you are here. Who knows why? Everyone has his reasons. As for the word ‘career’, one idea behind the Dewey Lecture is that it could be of interest to younger philosophers who are still finding their way, to see how a senior philosopher has developed and reached whatever place and whatever points of view he now occupies. Of course, that doesn’t mean it would have to serve as a positive model. It could go either way. But I think my own history cannot really offer very much useful guidance of that kind one way or the other, for reasons I will try to explain. One reason is that I didn’t think of what I was doing in terms of seeking a ‘career’ or a ‘profession’ at all. What attracted me to philosophy was in part precisely the idea that it wasn’t like getting a job or following a professional career. It seemed to me nothing like studying to become an engineer or a doctor or a lawyer. I thought of philosophy as something you studied just for its own sake. That was its charm; it wasn’t for anything. When I was at university and told people I was studying philosophy and they asked, as they inevitably did, “What are you going to do with that?”, I would never have answered that I was going to get a Ph.D. and become a professor. I did get a Ph.D. and I did become a professor, but at that time I had no such aspiration. That wasn’t the point. And even when I did get a Ph.D., and took a job teaching philosophy, I don’t think I saw myself even then as pursuing that professional goal. Like most American philosophers more or less exactly contemporary with me, I did not actually apply for the job I got when I left graduate school. I was offered a job, and I took it. It seemed like a good thing to do then. And Berkeley turned out to be such a good place that it is the only job I have ever had. In that particular respect my case is somewhat unusual. Not unique; there have been a few other complete academic stay-at-homes in philosophy besides me, but not many. So like others who got their Ph.D.s around the time I did, I have never prepared an academic dossier with a description of my thesis and a writing sample, or asked people to write letters of recommendation to be included in it and sent around the world. I have never been to a professional convention seeking a job or been interviewed in a hotel room or anywhere else, or ever had a “fly-back” or given a “job talk”. I have been involved over the

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



THE PURSUIT OF PHILOSOPHY

years in what has come to be called “placement”, but only in helping as a would-be “placer” and never as a “placee”. I don’t want to give the impression that I was alone in all of this, or that those of us who were so fortunate at that time fell into some special, elevated, elite category of job-seekers. Far from it: things were simply done differently in those days. It is not much of an exaggeration to say that then, in the early 1960s, the scene was dominated by those who were seeking employees, not those who were seeking employment. They were after us, not we after them. That was all part of the good fortune of those years for most of us. No such good fortune had been available to the generation slightly older than we were; for them, jobs had been very difficult to find, or to keep. I see the good timing of our cohort as just one part of the great good fortune that I find I have been surrounded with my whole life, often without even knowing it, at least at the time. It is not that I have had no active role in what has come to be—obviously not—but looking back on those early years I am not at all sure that I then had the will and the determination and the focus on professional achievement and advancement that for many years now has seemed to be required of all aspiring young philosophers. It felt different then. So in that respect I have been doubly fortunate. All this makes it difficult for me to be much help in giving support or advice to others on the basis of my own past experience. In my experience there seems to have been so little strategy or planning or professional decision-making; it didn’t seem called for. That is why I resolved, when speaking to younger philosophers about jobs and careers and professional advancement, never even to mention how things were, or how things were done, long ago when I was roughly their age. I still think it is a bad idea. It is unseemly to speak openly about the whole thing in public, especially with younger people present (I mean people under about 65). As with efforts towards truth and reconciliation, it is sometimes best for all, and best for the future, that some things about the past be left unsaid. But here I am before you, charged with the task of speaking openly about the past! As for my own past, I started out with the good fortune of having parents who were completely supportive of my doing anything I wanted to do. They never raised the question “What are you going to do in life?”, or “Why don’t you do so-and-so or become such-and-such?”, or “Why on earth are you doing that? What are you going to get out of that?”.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

THE PURSUIT OF PHILOSOPHY



Never. And not because they were not interested or didn’t care what I did. It was because they did care, and they knew how to let me do it. They trusted me and believed in me without imposing conditions or demands of their own (or were so good at it that I didn’t feel them). In any case, I felt no pressure, and there was no rush. When I finished high school I didn’t even apply to the university until a month or so before the academic year started. You could do that then, as long as you had achieved the right honours in enough of the high school subjects required for university. Universities in Canada were virtually free; there was only an insignificant registration fee. And I lived in Toronto, which happened to be the home of what was then Canada’s best university. So I lived at home and went to that university. When I first went there I must have had very little idea of what philosophy is. Maybe I thought it was something you did or talked about in smoky cafés, wearing a black turtleneck sweater. But I must have had or got a better idea than that because during my first year of studying it I chose it as my major or “Honours” subject. There were very good teachers at the University of Toronto, and the instruction in philosophy was almost completely historical. We read lots of things very carefully and tried to figure out what they said, and why they said it. Roughly speaking, we started with Thales and ended with Quine (actually we started with Plato and then went back to Thales). All that was also part of my good fortune, as I now see it. I never really felt very secure in my understanding of what was going on in philosophy. I did well enough in my studies as an undergraduate. I got good grades, but I knew that there was a lot I really wasn’t getting. There were enjoyable and fruitful discussions on almost everything with one outstanding fellow student, John Woods. That meant a lot to me, and it helped me. But I continued to feel the shallowness or fragility of my understanding in a way I thought more sophisticated students didn’t feel. Despite the doubts and the naiveté and the deeper puzzlement, I think I did come to learn a fair amount about a number of philosophers of the past, if not about the history of philosophy itself, and that turned out to stand me in good stead when I got to graduate school. But that was no part of any plan of mine then. Although by the end of my undergraduate years I had done quite well in philosophy, I never would even have thought about going to graduate school if it had not been for the encouragement of a couple of my

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



THE PURSUIT OF PHILOSOPHY

teachers. Here, I can now see, is a good lesson for all of us: to recognize, and encourage, our most interested, promising students. Of course, in one respect things then were the same as they are now: going to graduate school costs money. And I didn’t have any money. But that was the late 1950s, and it was known that a huge increase in university attendance was just over the horizon as the postwar baby boom generation was to reach college age. This was the source of the good fortune we were all about to experience. So there was an increasing amount of financial support for the higher education of prospective university teachers. I was encouraged to apply for what were then called Woodrow Wilson Fellowships, initiated for just that purpose. I learned a few years ago that 200 of those fellowships had been awarded each year in all of North America, but just at the time I reached my senior year at the university the number was increased five-fold, to 1000. Another stroke of luck or good fortune. I got one of those fellowships, and I know that without it I would not have gone to graduate school in philosophy. Or if I had gone, I would not have done in philosophy what I have been able to do. So I owe a great deal to the Woodrow Wilson Foundation—as they keep reminding me every year in their regular fund drives. That kind of financial support is part of the good fortune all of us have been enjoying for some years. By now virtually everyone is more-or-less fully supported in one way or another through graduate school in a great many of the best Ph.D. programs. That is not true even now of most other countries in the world, and it has not been true for most of the history of philosophy. We are all fortunate to live today in a place, and an age, in which the advanced study of philosophy is in effect subsidized for those who can do it best. I went with that Woodrow Wilson fellowship to Harvard, and that, eventually, changed everything. I found myself among a large group of fellow students who were extremely good, and had had training and experience in good philosophy departments in discussing philosophical questions and trying to get somewhere in thinking about this or that. And the ‘this’ or ‘that’ they were good at thinking about did not have to be identified by philosopher’s proper names, such as Locke’s view of substance or Kant’s conception of the transcendental unity of apperception. I thought I knew something about some of those things from a kind of spectator’s point of view, but I saw that the best of my fellow students

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

THE PURSUIT OF PHILOSOPHY



had a much better idea of how to do the subject of philosophy—how to pursue it, and advance it, not just to learn about it. This is what I too began to develop a sense of, and to appreciate, and to try to imbibe. Harvard was a serious intellectual place, where philosophy was strenuously pursued at the highest level. So my arriving there, and thriving, was another stroke of great good fortune. I had had a year-long course on the Critique of Pure Reason as an undergraduate and I was much exercised by the whole question of a priori knowledge. That was one of the reasons I wanted to go to Harvard in particular. I knew Quine was there, taking on that question in a new way that excited me, and in roaming the stacks in the University Library in Toronto I had found Morton White’s newly-published Toward Reunion in Philosophy on the same themes, and he too was at Harvard. He eventually became director of my Ph.D. thesis, in which he was extremely helpful and supportive, a gentle critic, and a valued advisor, both about the thesis and about other things. I don’t think he ever said it to me in so many words, but one thing I learned from my experience with him is that the most valuable quality a Ph.D. thesis can have is that it be finished. In the Harvard philosophy department I also found, among others, Burton Dreben, who at that time had a big effect on my philosophical development. He was a very active, engaged teacher, and you could discuss things with him almost endlessly. The subject gradually began to open up for me, along with possible ways of getting somewhere in it. Dreben stressed always the historical aspect of the problem, which was congenial to me, and he was always insistent on the distinction between purely technical or formal results or issues on the one hand and whatever philosophical constructions or readings might be imposed on such results on the other. This is one of the most valuable lessons I think one can learn as a fledgling philosopher. Not everyone seems to learn it equally well. It is something that was always stressed at Harvard in those years. In this, I think, we all saw Quine as the model. There were Quine’s writings, of course, which I studied carefully. Quine himself was helpful to me. He was very conscientious. When I gave him something of mine to read he read it right away, and we discussed it. He told me what he thought was right and what was not, he raised questions I sometimes couldn’t answer, and he made suggestions about where to go next. I attended his seminars and his mathematical

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



THE PURSUIT OF PHILOSOPHY

logic course, and I was in his lectures on the philosophy of language at the very time Word and Object came out. For the second half of that semester we all could sit in the class and turn the clean new pages of the book before us as Quine went from one point to the next by shifting a card from the top to the bottom of the well-thumbed deck of three-by-five cards he lectured from. One completely unexpected find for me at Harvard was Rogers Albritton. He turned out, over the years, to be simply the best philosophical discussant and conversationalist I have ever seen or talked with. He was at Harvard all the time I was there, he came as a visitor to Berkeley in later years, I twice taught as a visitor at UCLA when he was there, and I kept up with him at other places and times in between. I learned something, and was both challenged and enlightened, by virtually every philosophical encounter I had with him. And he was always very supportive of me, and encouraging. He was a major figure in my philosophical life, and not just at the beginning. It was in those early years at Harvard with Albritton and, in a different way, with Dreben, that I began the serious study of Wittgenstein that I have continued ever since. At Harvard, the first year I was there, Gwil Owen was visiting from Oxford, and the next year David Pears, also from Oxford. I eagerly took their classes and learned a great deal from each of them about different ways of doing philosophy. The depth and complexity and philosophical richness that Owen drew attention to in those works of antiquity that he knew so well was both a revelation and an inspiration to me. I continued to attend his classes and discussion groups in Oxford in later years whenever I could. And with all the subtlety and finesse of David Pears, it was somehow still possible to feel that even without his brilliance you could do the subject that way too! But we do not profit most from our teachers by trying to emulate them. For one thing, it doesn’t work. We gain most from our teachers by responding to what we find best in them and then trying to incorporate those virtues into a developing philosophical identity of our own. I have continued to learn from David Pears; we have written on some of the same subjects and responded to each other’s work for decades. But without doubt the greatest influence and the most beneficial effects in my graduate education came from my fellow graduate students. I believe that is quite generally true, for almost every student of philosophy.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

THE PURSUIT OF PHILOSOPHY



That is primarily why it matters for your future success where you go to graduate school. The faculty at the place matters; I mean the quality of their minds and their teaching, not their reputation or fame. But it is most important to study at a place where you are surrounded by good, engaged, active philosophers of your own age. I had the good fortune at Harvard to be surrounded by and in touch with many of those who became leading figures in the subject in later years. And you could tell at the time that that was going to be so. Of course, there were many fewer Ph.D. programs in those years, so the concentration of outstanding students and potentially leading philosophers in one place was perhaps higher in a few places then than it is in any one place now. Now there are so many good philosophers, and so many good places to study, that the best seem to be coming from all over. I was a beneficiary of this concentration at Harvard. With the faculty and the fellow graduate students I was surrounded by there, I couldn’t help but benefit. It is a well-known principle in sports, and ought to be as widely proclaimed in philosophy, that if you have a certain ability, and you play with those who are better than you are, especially if you play with the very best there are, you get better. Whatever progress I have made in philosophy I attribute to the truth of that principle and to my good fortune in having available to me some of the best philosophers there are (or were). I knew very little about Berkeley before I went there in 1961, but when I got there I found an intense group of younger philosophers, each one different in interesting ways from each of the others. That has remained a general feature of Berkeley over the years, even with different people. It is one of the things that makes the place so rich, and so stimulating. When I first arrived in Berkeley Stanley Cavell was still there, writing and giving a seminar with Thompson Clarke on chapters of what was to be his Ph.D. thesis and eventually The Claim of Reason. His explanations of the detail and the significance of Wittgenstein’s later philosophy both in those pages and in person were to me a complete breakthrough to something new and powerful, both in understanding Wittgenstein and in understanding philosophy. It was a great disappointment to me, and to all of us in Berkeley, when he left for Harvard the next year. Someone who was in Berkeley when I arrived, and has never left, was John Searle, by now the only person who has been in the Berkeley philosophy department longer than I have. He had come straight from

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



THE PURSUIT OF PHILOSOPHY

Oxford, which was then the centre of English-language philosophy, and he was spreading the word, while wiping the floor with all comers then as he has continued to do ever since. David Shwayder, another Oxford product with long Berkeley connections, was also part of that early 1960s department. Soon there was also Thomas Nagel, who had been a pupil of Austin at Oxford; he and I had been fellow graduate students at Harvard. Then George Myro came, another of our Harvard contemporaries. He too had studied at Oxford. And then Bert Dreyfus, who had been many places, but definitely not Oxford. It is difficult to recapture, or even to remember, the excitement, the variety, and the energy of all those different forces pushing and pulling in Berkeley at that time. Later in the 1960s Paul Grice arrived, and then soon Hans Sluga and then Michael Frede. These colleagues remained a continuing source of knowledge and stimulus to me to move in new, both historical and philosophical, directions. Grice’s arrival meant a lot to us in Berkeley; he formed a kind of centre around which many of us revolved, some more than others. He explored almost every area of philosophy at one time or another. He was an incredibly clever and resourceful and strategic philosopher, and a tremendously effective teacher, especially one-on-one. When he retired, as was compulsory in those years, Donald Davidson arrived as his successor. But Grice was still there too. Berkeley as a philosophical community now had them both. And when mandatory retirement also eventually claimed Donald Davidson, he simply changed his official relation to the university and carried on just as before. Bernard Williams arrived as his successor, and they both remained active colleagues of each other and of all of us for the rest of their lives. And through that whole period, roughly from the arrival of Davidson, Richard Wollheim also was in Berkeley for part or all of every academic year. He made a wonderful, and of course uniquely Wollheimian, contribution to philosophical life in Berkeley, eventually even serving as chair of the department. All these senior colleagues were good friends of mine and outstanding philosophical companions. I was one of the main beneficiaries of their combined philosophical and personal richness. I found great support and encouragement in their relations to me as a philosopher. They were all older and wiser than I was, and that was an important factor. I was devastated when they all died within less than six months of one another.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

THE PURSUIT OF PHILOSOPHY



But from the 70s on outstanding, congenial younger colleagues had been arriving who came to mean just as much to me and to my interests and development in philosophy. Sam Scheffler and Alan Code, in particular, and then Janet Broughton, and then Hannah Ginsborg, and now, much more recently, a group of powerful younger philosophers, among them, I am happy to say, the Chair of this session, John Campbell. This is beginning to sound like a who’s who of the Berkeley philosophy department over the years, but I am only trying to mention some aspects of the good fortune that has surrounded me and some of the people who have made such a difference to whatever I have been able to do. I have mentioned only teachers and colleagues. To give a full accounting I would have to mention all those philosophers whose works I read and ponder and respond to. And then there are all those who have been students at Berkeley at one time or another. But this is the kind of thing that is true of all of us here. We all live in a wider and flourishing philosophical culture of which we are all beneficiaries. That wider culture includes not only individuals, but communities or places, other places. I have mentioned Harvard. Oxford, too, has a very special role in my life. After Berkeley, where I have lived the longest, and Toronto, where I grew up and went to university, Oxford is the place I have lived next longest. I never was a student there, but I have spent more time there than I did in graduate school in Cambridge, Mass. I have lived four different sabbatical years in Oxford from the 60s through to the 80s, and many other long or short stays for philosophical or other purposes. I certainly would not have done what I have done in philosophy without the intellectual sustenance and challenge I have always found there. I have also had the good fortune to spend two long happy periods teaching as a visitor at UCLA, a philosophical community and a group of philosophers I am very fond of and feel strangely part of and continue to learn from. These have been communal and supportive philosophical places for me; and they have meant a lot. But it is also good to have places of philosophical solitude and reflection where I can be, at least philosophically, completely on my own. Venice, and Rome, are special places of that kind for me. Within our philosophical culture, whether widely or more narrowly conceived, I must say that of all the philosophers I have ever interacted with, I owe most to my erstwhile colleague Thompson Clarke. He was in Berkeley when I first got there, and it has turned out that in decades of

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



THE PURSUIT OF PHILOSOPHY

continual discussion, joint classes, and other forms of shared exploration of different kinds of philosophical questions with him, I have learned more than from any other source about the special character of philosophy and about the level of depth that needs to be reached to make real progress in it. My whole way of thinking about philosophy has been affected by him in ways that I could no longer identify, and I am happy to acknowledge it here. I see it as a huge part of that same good fortune I have been talking about. When I arrived in Berkeley, fresh from graduate school, I did not have any conception that such a development or transformation of my philosophical thinking was even possible, let alone any expectation that such a thing would happen. It certainly was not the result of any “career decision” or “professional strategy” I was wise enough to adopt and follow. It was the result of my finding in Thompson Clarke what I recognized to be real philosophical thinking, and coming to participate in it in my own way. What is real philosophical thinking? It can be many different things. Wittgenstein observes that “a man who is not used to searching in the forest for flowers, berries, or plants will not find any because his eyes are not trained to see them and he does not know where you have to be particularly on the lookout for them”. Similarly [he says], “someone unpractised in philosophy passes by all the spots where difficulties are hidden in the grass, whereas someone who has had practice will pause and sense that there is a difficulty close by even though he cannot see it yet”. But not seeing the difficulty clearly at first should not be surprising if you know “how long even the man with practice, who realizes there is a difficulty, will have to search before he finds it”.² The philosophers I admire most possess just that kind of acute sensitivity to philosophical difficulties. They are open to potential philosophical riches, and they find them, in what look to most of the rest of us like very unpromising places. And, what is equally important, those philosophers I admire most know how to keep searching when they know they haven’t really found the right thing yet. This is not the only kind of philosophical ability there is. Sheer intellectual firepower combined with prodigious energy can produce impressive results which must be taken account of one way or the other. But for me, those I most admire have a

² L. Wittgenstein, Culture and Value (ed. G. H. von Wright). Blackwell 1980, p. 29.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

THE PURSUIT OF PHILOSOPHY



firm foothold in reality and a ‘nose’ or feel for real problems, along with the patience to unfold the detail of what has to be overcome to achieve the kind of understanding that can mean the most to us. I am happy to know or to have known some philosophers like that. They have meant the most to me. Where does an interest in or a passion for philosophy come from? What kind of mentality seeks it out, or goes for it when it comes across it? There is obviously no simple answer; there are many different philosophical mentalities. There is a certain kind of thought or feeling that occurs to many people at some time during their youth or adolescence that is regarded as evidence of at least an incipient philosophical sensibility. I have in mind the thought, as it is sometimes put, that “the world I inhabit is my world” or “my world is the only world there is” or (even more grandiosely) that “everything there is depends on me for its existence”. I don’t mean that people seriously put forward these views and act on them—or at least not much. But they are thoughts that have passed through young people’s minds and have led them to think about themselves in the world in what we have come to think of as philosophical ways. I have even heard it said of one very distinguished philosopher that at a remarkably early age he asked his mother “Mommy, does God exist even when I am not thinking of him?”. These are thoughts people can have; some of my best friends report having had them. I have to admit that I have never had any thoughts of this kind. I draw no inference from this; I simply report it as a fact. It could be connected with my feelings about idealism in philosophy. I just cannot see how idealism could be true, or even really thought to be true. Maybe I exaggerate a little. I confess that there was a brief time when I thought what I wanted to say was that transcendental idealism is true, you just can’t say it. But then I saw that what I had wanted to say is not something that can be said. So I didn’t say it. As with idealism, so with verificationism, if there is any difference between them. I mean the idea that something can mean or say only as much as we can find out or have reason to believe. I just don’t see how anything like that could be so. This is probably another instance of my not getting the idea of things’ being so or not as somehow depending on me, or on us. It is not that nothing depends on me, or on us, of course. But it seems to me that most of what is so is not like that.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



THE PURSUIT OF PHILOSOPHY

Maybe my blindspot for idealism or solipsism can be explained genetically. I once was driving with my aged mother—she was then at least ninety-five—through some dense woods in Canada that looked as if they had been the way they were for a very long time. We had not had anything even remotely resembling a philosophical conversation— ever, really. So I was completely staggered when my mother suddenly announced in the middle of that silent forest, “You know, I never got that thing about a tree falling in the forest; why wouldn’t it make a noise?” I am not sure we really have reliable access to the origins of our interests and attitudes and motivations in philosophy. I believe what I have now been saying about how it was for me. But in all this I could be revealing not just weariness, as Nietzsche would have it, but illusion or self-deception as well. One thing I know I have always thought about philosophy is that it is especially difficult. And for a long time now I have been struck by one distinctive feature of philosophy and the urge to engage in it that I think is partly responsible for that special difficulty. In philosophy we take a very general interest in those pervasive, fundamental features and capacities by which human beings differ from everything else in the universe: language, thought, perception, knowledge, science, evaluation, intention, action, morality. And in our attempt to understand these things we find ourselves having to make use of the very capacities or ways of thinking and responding that we are hoping to get a proper philosophical understanding of. This can leave us dissatisfied. We don’t want simply to take our current beliefs and attitudes for granted without further understanding or backing. Like good “anthropologists” of the human species, we want to make fully intelligible to ourselves the relation that all those distinctively human responses stand in to what is really so in the world. But we investigators are human beings with just the attitudes and responses we want to understand. We are both the objects and the agents of the very study we are engaged in. This can lead us to wonder whether we could ever get a fully satisfactory grasp of what we want to understand about ourselves in philosophy. It can come to seem that full philosophical satisfaction would require an understanding of ourselves from a position or a set of beliefs or attitudes other than or somehow “outside” our current conception of how things are. That current conception, and how it has been possible for us to arrive at it, after all, is just what we want to understand in relation to the

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

THE PURSUIT OF PHILOSOPHY



surrounding world. And that doesn’t seem possible by simply taking that conception for granted from the beginning. Sometimes, when we seek such an “elevated” or “external” point of view on ourselves, and think we have found one, the news is not reassuring, or satisfying. I think this happens, for instance, when we seek an understanding of human knowledge in general. We then seem forced to the conclusion either that we don’t really know anything at all about the world, or at least that we have no satisfactory understanding of how we do know the things we think we know. Similar conclusions can be reached in the same way about meaning or thought or value. We seem to reach a position from which it looks as if nobody ever really means any particular thing by any utterance he makes or has one determinate thought rather than another, or that nothing really has any value or worth on its own, unless we “invest” it with some value by our feeling or desire or concern for it. In each case there seems to be a point of view from which conclusions like these can be found unavoidable. But they cannot be fully satisfying. They do not give us what we want. Or they do not give back to us what we think we already have. This makes it overwhelmingly attractive to try to find some way to deny or at least to avoid such discouraging general conclusions. But it seems to me that many philosophical attempts to give positive answers to the completely general questions that concern us, and so to return us to the comfort of our ordinary ways of thinking and justifying what we think and do, end up giving a distorted account of what the thoughts and attitudes and feelings that we want to understand actually are. Whatever satisfaction we achieve is gained only at the price of misunderstanding ourselves. And that provides no lasting satisfaction either, at least to those who can recognize the distortion as distortion. If I am right about this back and forth movement of thought, I think the best we can do in the face of it is to recognize the apparently undying urge for the kind of “outside” or “elevated” position that gives the investigation its special philosophical character. And while recognizing that urge we can nonetheless keep pointing out how and why the goal we feel we seek has not yet and perhaps never can be reached. We can continue to draw attention to the distortions or misunderstandings or failed aspirations that seem endemic to the apparently reassuring answers to the completely general “anthropological” questions that concern us.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



THE PURSUIT OF PHILOSOPHY

That can lead us closer to the actual thoughts and attitudes we want to understand, but it must be granted that this activity can also be found dissatisfying, if not worse. It can be felt at best only to identify or describe those thoughts and attitudes, without offering any satisfyingly positive explanation of what we feel we want to understand. It promises no final philosophical theory of knowledge or language or value or whatever it might be, and so no prospect of a satisfying resting-place from which the human phenomena that puzzle us can be viewed once and for all with the illumination we seek. I think this is true. I think there is no such satisfying resting-place on offer to the philosophically curious human investigator who takes seriously the status of his own philosophical investigation. And it can be profoundly, one might say metaphysically, dissatisfying to acknowledge this. But I think this recognition of the inevitability of metaphysical dissatisfaction can itself reveal something important about the human condition or human aspiration. It can bring home to us the irrepressible human effort to have something which at the same time we can come to see we cannot really have (at least in the form in which we most desire it). In that way it brings into question the very possibility of the kind of understanding of ourselves that seems to be embodied in the idea of a philosophical theory. Exposing that idea for the illusion or will-o’-thewisp that it is represents perhaps a different goal, or different hopes, for the pursuit of philosophy. But it does not mean abandoning the enterprise of trying to understand ourselves in the ways we have always sought in philosophy. In fact, it is a different way of contributing to that kind of understanding. This enterprise of understanding ourselves in general is what we are all engaged in in one way or another. And the fact that we are doing it, or the doing of it, is what counts, not some final prize that is supposed eventually to make the whole effort worthwhile. All this philosophical activity is what I have come to feel part of and to identify with. I have not thought of myself primarily as absorbed by or identified with a certain distinctive career or profession. It has certainly happened that I have a career and a profession; I cannot deny it. It is something I am especially aware of when I am overcome with the administrative duties that make such demands on all of us. I feel then, in the University of California, like a rather low-level state employee: just the kind of thing, or the kind of job, I thought I was avoiding by studying philosophy.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

THE PURSUIT OF PHILOSOPHY



But beyond those bureaucratic demands, I have increasingly felt that what I most belong to and so most identify with is that huge human enterprise called philosophy, something that has been with us for a very long time. But it is not just a matter of how long it has been going on. I don’t feel related to philosophy in just the way I think a doctor or a lawyer might feel about being part of something that has been going on for a long time. I feel that what I have been doing, what all of us here are doing now, is recognizable as carrying on the same kind of investigation that Plato and Aristotle and Descartes and Hume and Kant and Hegel and Russell and all the other greats of our subject were involved in. I don’t mean that anyone among us now is in that stellar company. I mean only that I think of what I am trying to do, and what all of us are doing, as continuing the same enterprise those giants of thought were engaged in, but in the form it has now reached for us. And that is the identity that I find to be the greatest source of satisfaction.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

3 The Epistemological Promise of Externalism There have been two distinct but related tendencies in the philosophical theory of knowledge. One, going back to Socrates as represented by Plato, is to ask what knowledge is: to seek to define it and to distinguish it from truth, belief, certainty, reasonable belief, and other closely related notions. The other approach focusses on human beings as they find themselves in the natural world and asks how they come to know the sorts of things they know. Questions of this second kind tend to become central to philosophy at times of upheaval, innovation, or even revolution within human knowledge itself. This was especially true of Europe in the seventeenth century. The rise of the ‘new science’ eventually overturned deeply entrenched ways of thinking of the world and our relation to it. Philosophical concerns with knowledge in our own day are still dominated by descendants of problems and theories introduced at that time. A central problem for philosophy since that time has been to understand the role of sense-experience in human knowledge, and to see whether or how we can know what we do about the world on the basis of what we perceive to be so. When that question is understood as completely general—about any perceptual knowledge of the world at all—it has proved difficult to avoid the conclusion that we perceive much less than we might unreflectively think we do. A familiar line of reflection seems to imply that, strictly speaking, sense perception alone is no reliable guide to the way things are in the world that produces it. Unaided by knowledge from other sources, it seems incapable of giving us access to an independent public world at all. This raises the question how we know any of the things we think we know about the world. This question is potentially disturbing or challenging only to those who have found the reach of sense perception to be severely limited

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

THE EPISTEMOLOGICAL PROMISE OF EXTERNALISM



compared to the richness of what they think we know on that basis. If the world around us were available to perception alone—if we could sometimes perceive the objects around us and perceive that many of the things we believe are true of them—there would be no general problem of explaining our knowledge. We would know what much of the world is like simply by perceiving it. So if an urgent challenge to our unreflective confidence about the world in general can be seen to arise at all, it is because of the force of those philosophical reflections that seem to shrink what we get through the senses alone to some meagre and apparently inadequate basis. The threatening reflections about sense perception start from the undeniable fact of perceptual illusions or mistakes. It sometimes happens that a person perceives something that is not there, or misperceives what is there, and so on the basis of perception is wrong about what is so. This obvious fact of life in itself implies no general scepticism. If there are also occasions on which we succeed in seeing or otherwise perceiving how things are in the independent world, then we do often know how things are, and we know it by perception. But if reflection on the nature of perception leads to the conclusion that what we perceive even when we happen to get true beliefs about the world extends no further than what we also perceive in illusions or when our beliefs about the world are false, then perceptual knowledge of the world in general is threatened.¹ What is true of perceptual illusions comes to be taken as representative of what perception even at its best can provide us with. That generalizing move from ‘some’ to ‘all’ is what leads to the conclusion that we never strictly speaking perceive what is so in the independent world. And that in turn seems to make it impossible to understand how we get knowledge of the world by perception. The philosophical study of knowledge in the early and middle years of the twentieth century was dominated by just such a conception of sense perception. It did not yield satisfactory answers to the question of how knowledge of the world is possible on that basis. One strategy was to try to bring the world we believe in closer to what we can perceive; every meaningful

¹ For influential examples of reflection along these lines see, e.g., G. E. Moore, “Sense-Data”, ch. 2 of his Some Main Problems of Philosophy, Allen & Unwin, London 1953; H. H. Price, Perception, Methuen 1932, ch. 1; C. I. Lewis, Analysis of Knowledge and Valuation, Open Court, ch. 7.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



THE EPISTEMOLOGICAL PROMISE OF EXTERNALISM

and potentially knowable proposition about the world was to be expressed exclusively in terms of possible sense experiences.² No plausible equivalences were found. Another idea was that everything we believe about the world beyond the strict limits of possible perception is a ‘posit’ we introduce, or a huge ‘theory’ we construct, to make the best overall sense of the limited sensory data we receive.³ But indefinitely many different ‘posits’ or ‘theories’ can be constructed, all consistent with the same perceptual data but incompatible with one another, so perception alone could provide no distinctive reassurance about the particular ‘theory’ we actually hold about the world. Taking fully seriously the idea of strict limits to what human beings perceive makes it seem impossible to account for the knowledge of the world that we all think we have. Scepticism seems the only strictly correct answer. The project of defining knowledge resurfaced in that same period in the twentieth century, partly under the influence of the idea that the proper or perhaps only method of philosophy is ‘analysis’—the definition or clarification of those fundamental concepts we use to make sense of the world in the ways we do. Not only were relentless efforts made to define the concept of knowledge,⁴ with more and more elaborate conditions being added to eliminate yet another clever counter-example to the necessity or sufficiency of what had been offered so far. It was also widely believed that success in that task would overcome the unsatisfactory sceptical plight that had haunted epistemology. The idea was that the restricted deliverances of sense perception seem insufficient for knowledge or reasonable belief about the world only because of a mistaken or distorted conception of knowledge. The promise was that by getting the definition of knowledge right we would see that the apparent threat of scepticism vanishes.

² See, e.g., C. I. Lewis, Analysis of Knowledge and Valuation ch. 8; A. J. Ayer, “Phenomenalism”, in his Philosophical Essays, Macmillan 1959. For a clear demonstration that the sought-for equivalences do not hold see, e.g., R. Chisholm “The Problem of Empiricism”, The Journal of Philosophy 1948. ³ See, e.g., the last section of W. V. Quine, “Two Dogmas of Empiricism” in his From a Logical Point of View, Harvard University Press 1953 or his Word and Object MIT Press 1960) ch. 1. ⁴ The response to the challenge of E. Gettier, “Is Justified True Belief Knowledge?”, Analysis 1963, eventually amounted to hundreds of articles or books up to at least the middle 1980s.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

THE EPISTEMOLOGICAL PROMISE OF EXTERNALISM



This strategy had much to recommend it. Anyone who feels forced to the conclusion that what is available in sense perception could never explain how knowledge of the world is possible must be thinking of knowledge in some way or other. If the conception of knowledge at work in such thoughts could be shown to be faulty, there would be no good reason to draw that conclusion. And if on a different and more accurate conception of knowledge, strict limitations on sense perception represent no threat to knowledge of the world, the definitional or ‘analytic’ project would be vindicated as providing the solution to the philosophical problem of knowledge. This strategy seemed to gain further encouragement from the line of reasoning by which many philosophers explained the source of the traditional epistemological quandary. Restricted as we are to limited perceptual resources, it was argued, we could have reason to believe something about the world beyond them only if we could discover some connection between our limited sensory data and what is so in that wider world. No such connection could be established a priori, independently of experience; it could be discovered only on the basis of experience. But no appeal to sense experience alone could ever take us beyond the limited sensory data. So nothing available within sense experience alone could give us any reason to believe anything about what is so beyond it. That is why sense experience was said to be incapable of supporting any belief about a wider world. This reasoning appears to assume that in order to know or have reason to believe something, I must know or have reason to believe that I know or have reason to believe it. Sensory data were held to give us no reason to believe things about the wider world because we could discover nothing within those data that would give us reason to believe in any connection they might bear to something outside them. Given only our restricted perceptual data, we could find nothing we could recognize to be such a reason. An externalist definition or conception of knowledge seems to promise a way out of this quandary. It implies that we do not have to meet that requirement in order to know or have good reason to believe something. According to externalism, the conditions of knowledge or reasonable belief are not conditions we must know or even believe to be fulfilled in order to fulfill them. To know or reasonably believe what is so it is sufficient if the conditions of knowledge or reasonable belief are in fact fulfilled.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



THE EPISTEMOLOGICAL PROMISE OF EXTERNALISM

The view is externalist in holding that some of the conditions of knowledge or reasonable belief can be fulfilled independently of, or ‘outside’ or ‘externally’ to, what the subject believes about the fulfillment of those conditions. Just as one can know something without knowing that one knows it—even while believing that one does not know it—so one can have reason to believe something without knowing or reasonably believing that one does. If that is correct, and if the traditional problem arises only from insisting on stronger conditions for knowledge or reasonable belief, then the problem arises only from demanding more than we actually require for knowledge. An epistemological challenge raised on that basis would be empty; it would rest on a misunderstanding or distortion of the meaning of ‘know’ or ‘reasonably believe’ as they are actually used. This is the substance of an epistemological externalist diagnosis of the traditional problem. To make good on this claim and on the consequent dissolution of the traditional sceptical threat, a correct definition of knowledge must be found, and it must be externalist in the required sense. Many philosophers have been working on that definitional project. The task has not been completed to everyone’s satisfaction. Some insist on defining knowledge without using ‘epistemic’ terms like ‘justified’ or ‘reason to believe’. Knowledge is accordingly to be understood as something like a true belief that is acquired or sustained in a certain way. One ‘reliabilist’ form of externalism says that the true belief must be acquired or sustained by a properly-functioning, reliable belief-forming mechanism, where reliability is defined in terms of something like the frequency with which it produces true beliefs.⁵ Another externalist definition says that a knows that p if and only if a knows something r that is a conclusive reason for p in the sense that it would not have been true unless it was true that p.⁶ Another version, related to this, says that a knows that p if and only if a’s belief that p ‘tracks’ the truth.⁷ Less ambitious externalist definitions of knowledge do not require the elimination of all ‘epistemic’ terms. But a correct definition even containing such terms must remain

⁵ See Alvin Goldman, “What is Justified Belief?”, in G. Pappas (ed.), Justification and Knowledge, Reidel 1979. ⁶ F. Dretske, “Conclusive Reasons”, Australasian Journal of Philosophy 1971. ⁷ R. Nozick, Philosophical Explanations, Harvard University Press 1981, ch. 3.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

THE EPISTEMOLOGICAL PROMISE OF EXTERNALISM



externalist in the required sense in order to undercut the sceptical quandary in this way. I do not want to take up the question of the precise formulation, and so the definitional adequacy, of epistemological externalism. Rather I would like to assess the role that any such definition might play in the diagnosis and dissolution of the traditional epistemological problem. The quandary I have described arises from the realization that the most we can perceive is severely restricted compared to the richness of what we claim to know on the basis of it. Epistemological externalism promises to remove the difficulty by supplying a correct understanding of the idea of knowledge or reasonable belief. I believe that no one who sees our perceptual position as restricted in the way the traditional problem implies could get from an externalist definition of knowledge the understanding and reassurance he seeks about knowledge of the world. If all the limited perceptual information we receive is in fact reliably produced by just those unperceived states of affairs which we all believe to hold in the wider world, then according to the externalist definition we know that the world is as we believe it to be. But that would not give a reflecting philosopher who arrives at that restrictive view of perception a satisfactory understanding of how knowledge of the world is based on sense perception. The deficiency I have in mind would not be immediately discernible to a philosopher who reflects only on other people’s sense perception and knowledge, not his own. Even if he believes that other people’s perceptions alone give them no information as to what is so in the world, he might still try to explain their beliefs by appealing to what he takes himself to know about the world. He might find that the beliefs those people get on the basis of their restricted perceptions are in fact true and are produced by the world by processes that are reliable. That would be to find that the conditions expressed in an externalist definition of knowledge are fulfilled by those people. He could conclude that therefore they know what they claim to know. He would not have to regard what he takes to be the limited scope of their strictly perceptual data as a threat to their knowing what they think they know or to his own understanding of their knowledge. But someone who reflects on the role of sense perception in human knowledge in general must also acknowledge that, being human, he is one of the people to whom his conclusions apply. If he holds as a general

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



THE EPISTEMOLOGICAL PROMISE OF EXTERNALISM

thesis that perception even at its best gives no information about the independent world, then he must concede that his own sensory data are restricted in that same way. But that is to concede that whatever is so beyond what he perceives could be any one of indefinitely many different possible ways compatibly with the limited sensory data he receives. Nothing he could ever strictly speaking perceive to be so could discriminate among those possibilities. Of course, he might believe that the world is in fact only one of those possible ways, and he might think he knows which way that is. Perhaps he also believes, and claims to know, that the world’s being that particular way is precisely what produces in him the limited sensory data he receives. He might also believe that it does so by processes that produce mostly true beliefs. If this last belief of his is in fact true, then he fulfills the conditions expressed in a reliabilist externalist definition of knowledge. If that definition is correct, it follows that he knows what he thinks he knows about the world, despite the poverty of his purely perceptual information. The question is not whether that is something the person knows; an externalist definition of knowledge implies that it is. The question now is whether a person’s fulfilling the conditions of that definition of knowledge would put him in a satisfactory position for understanding himself as knowing by means of perception the things he thinks he knows about the world. I think it would not, as long as he holds that perception alone gives no information about what is so in the independent world.⁸ The philosopher who reflects along those familiar lines could insist that if the world as he believes it to be does produce his perceptions and beliefs by processes that are reliable, then (according to externalism) he knows what he thinks he knows about the world. But he would also have to acknowledge that if his perceptions and beliefs are produced by a world that is different from the way he believes it to be, then he does not know what he thinks he knows, despite his believing it. And he would have to concede that all the strictly perceptual information he could ever get is compatible with each of those

⁸ I argued this in “Understanding Human Knowledge in General” in M. Clay and K. Lehrer (ed.), Knowledge and Skepticism, Westview 1989. I tried again in “Scepticism, ‘Externalism’, and the Goal of Epistemology”, Proceedings of the Aristotelian Society: Supplementary Volume LXVIII 1994. Both essays are available in my Understanding Human Knowledge, Oxford University Press 2000.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

THE EPISTEMOLOGICAL PROMISE OF EXTERNALISM



different possibilities. Nothing he could ever perceive to be so could distinguish between them. On the assumption that knowledge of the world is possible at all only on the basis of what is available in perception, this would leave the theorist unable to understand himself as knowing or having any reason to believe anything about what is so beyond what he perceives to be so. He might believe that he knows or has reason to believe such things, but he would see that he cannot legitimately draw any stronger conclusion on the basis of anything he could ever perceive. In his attempt to understand his own knowledge, he would find himself in the quandary that severely restricted perceptual data have always seemed to lead to: scepticism about any matters of fact beyond the limited sensory data. If this is right, then the diagnosis of the traditional sceptical reasoning offered by proponents of an externalist definition of knowledge was off the mark. It is not acquiescence in a faulty non-externalist conception of knowledge that leads to the epistemological quandary. The severely restricted scope of perception is what creates the threat. A philosophical theorist who accepts strict limitations on perceptual access to the world can never achieve a satisfactory understanding of any knowledge he thinks he has of what lies beyond those limits, even if he accepts an externalist definition of knowledge. And as long as such a restricted view of the scope of perception seems inevitable, or even plausible, we will not have a philosophically satisfying explanation of how perceptual knowledge of the world is possible. I think the greatest promise of epistemological advance in the face of this quandary lies not in an externalist definition of knowledge or even in a positive externalist answer to the epistemological question. Promise of more secure progress lies rather in externalism’s potential for exposing and thereby undermining the assumptions needed for anyone to be faced with that apparently insoluble problem in the first place. What generates the completely general problem is the idea that no one ever perceives what is so in the independent world. The problem then is to explain how any beliefs whose contents go beyond that restricted domain of perception can gain any support from what is perceived. To be faced with that problem we must understand ourselves as never perceiving what is so in the world beyond or independent of what is perceived while nonetheless holding a great many determinate beliefs about that world. But could anyone coherently understand himself and others in that way? What

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



THE EPISTEMOLOGICAL PROMISE OF EXTERNALISM

must be true of a person who has perceptions with determinate contents, and beliefs about an independent world, but who never perceives what is so in the world that he believes in? Is that even possible? Could anyone consistently find that that is the position human beings are in? These are questions about the conditions of being in certain psychological states, and of understanding oneself and others as being in such states. To explore the promise of externalism on these matters would be to move away from a focus on knowledge in particular towards an understanding of psychological states and their contents in general. There is a conception of the mind that is broadly-speaking externalist in holding that the contents of a person’s perceptions, thoughts, beliefs, and other attitudes are fixed in part by what is true in the world in which the person has those attitudes. This is not simply the truism that people get into the psychological states they do because of what happens in the world around them. Rather the idea is that the very contents of a person’s perceptions or thoughts or beliefs on a particular occasion—what they are perceptions of or thoughts or beliefs about—are determined in part by what is or has been so in the world in which the person acquires those attitudes. Attitudes have those contents the possession of which best explains a person’s acting and feeling and responding as he does. This is the basic externalist idea that seems to me to have the richest epistemological potential. Taken in one way, it appears to have the consequence that our perceptions and beliefs about the world are in general correct.⁹ That is because whatever states of the world are responsible for our having the perceptions and beliefs we do are what, in general, those perceptions or beliefs are about, or are directed towards. For any particular perception or belief, or even for a sizeable number of them, it is possible for a person to be wrong, even wildly wrong, in what he perceives or thinks. But the externalist idea is that perceptions and beliefs in general must in large part represent the world as it in fact is; they would not be the very perceptions and beliefs they are if they did not. That is the sense in which what is so in the world in which people hold certain psychological attitudes is what determines the very contents of those attitudes. For ⁹ For an emphasis on this implication of externalist views see, e.g., Donald Davidson, “A Coherence Theory of Truth and Knowledge” in his Subjective, Intersubjective, Objective, Oxford University Press 2001.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

THE EPISTEMOLOGICAL PROMISE OF EXTERNALISM



the whole body of perceptions and beliefs taken all together, then, most of them, or most of the most fundamental of them, must be true. This has a very reassuring ring, and it might be taken as enough in itself to solve the problem of knowledge of the world once and for all. It means that the beliefs and other propositional attitudes that any philosopher would attribute to himself and others in the course of investigating human knowledge would be largely true. We could not turn out to be completely or mostly wrong about the world. If such an externalist conception of thought and experience could be established, it looks as if we would have a general guarantee of the truth of most of the beliefs anyone could ever have. Even if such an externalist view of the contents of beliefs could be established, I do not think it would be enough to give us a satisfactory explanation of how human knowledge of the world is possible. For one thing, true belief is not the same as knowledge or true reasonable belief, so an epistemological question can remain open even if we are assured that beliefs are mostly true. What needs to be accounted for is not just the truth of our beliefs, but human access to that truth. How are our beliefs known to be true, or how do we have good reason to believe them? These questions are not answered by the truth of our beliefs alone. An externalist conception of the mind and its contents in general would seem to provide answers to these questions since it implies that what we perceive to be so must also be true, at least for the most part. And if the things we perceive to be so are included among the very things we also believe to be so in the independent world, then not only the truth but also our knowledge of the truth of many beliefs about the world would be guaranteed along externalist lines. Anyone who understands perception in this externalist way could see that what we perceive gives us good reason to believe many of the things we believe about the world. There can be no better reason for believing something, and so no better way of coming to know it, than seeing or otherwise perceiving that it is true. Such an externalist view of perception would therefore account not only for the truth of our beliefs, but also for our access to those truths, and so for our knowledge of the world. This understanding of perception would provide a satisfactory account of our knowledge precisely because it accords us unproblematic perceptual access to the way things are in the independent world. We would not have to infer how things are from something we perceive;

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



THE EPISTEMOLOGICAL PROMISE OF EXTERNALISM

we would often see or otherwise perceive that things are a certain way. That would give us knowledge of many things that are so in the world, and that knowledge would give us good reason to believe many other things about the world that we do not perceive to be so. But if further reflection on the nature of perception led persuasively to the conclusion that we actually have no such perceptual access to the independent world, the satisfactory understanding of our position promised by an externalist conception of the mind and its contents would no longer be available. In yielding to those threatening reflections about sense perception we could no longer understand ourselves as having the reasons to believe things that we thought we had. This apparent threat to our understanding of human knowledge would be overcome once and for all if it could be shown that perception simply cannot be understood in that way. One way to do that would be to show that the restrictive view of perception is wrong and that an externalist conception of perception and its contents is correct. That would require establishing the truth of the externalist view that what we perceive to be so is included among the very things we believe to be so in the independent world, and that what we perceive and believe to be so is, for the most part, in fact true. But there is a question whether that view can be established in a way that provides the reassurance it seems to promise. The view cannot be shown to be true by first identifying all of a person’s or a community’s perceptions and beliefs and then discovering independently that all or most of them are true. That could indeed give us great reassurance, but the idea behind the externalist view is that the contents of perceptions and beliefs cannot be identified independently of identifying what is so in the world in which they are held. That means that anyone who attributes perceptions or beliefs to anyone must hold some determinate beliefs himself about the world. But fulfilling that condition of the successful attribution of perceptions and beliefs does not give an attributor any greater reassurance of the truth of the beliefs he attributes than the assurance he has of the truth of the beliefs he relies on in attributing those beliefs to others. An attributor will inevitably regard as true most of the beliefs he attributes. But if his grounds for his own beliefs could be brought into question by reflections on the nature of perception, he could be faced with a general problem of knowledge even if he began by accepting an externalist view of mental contents.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

THE EPISTEMOLOGICAL PROMISE OF EXTERNALISM



Support for the externalist view comes from the evident impossibility of anyone’s identifying or attributing beliefs or perceptions or other attitudes without holding some beliefs himself about what is so in the world. There appears to be no way for one person to come to any verdict about what another person perceives or believes except by finding some intelligible connection between that person’s responses and what is so in the world around him. That requires beliefs about the world on the part of the attributor, and beliefs about one thing’s being reason to believe another. He must appeal to the truth of such beliefs in making sense of what the other person does. That means that in general, or in the normal case, what one person comes to believe another person perceives or believes will be something that the attributor himself also perceives or believes to be so. Very general agreement between them is a consequence of fulfilling the conditions of successful attribution of any perceptions or beliefs; disagreement is intelligible only against a larger background of mostly shared attitudes. The externalist view in its reassuring-sounding form says that the contents of the perceptions and beliefs that people attribute to one another are for the most part true. But I think it must be conceded that the truth, in general, of the perceptions and beliefs that people all agree in attributing to one another does not follow from their agreeing in attributing them. Widespread agreement in perceptions and beliefs—even inevitable agreement—does not imply that what everyone agrees in perceiving or believing to be so is in fact true, even for the most part.¹⁰ So it does not imply that the externalist view of perception and belief in its reassuring-sounding form is correct. And it therefore does not imply that a restrictive view of the scope of perception is wrong. Even if this failure of implication is conceded, I think it represents no obstacle to externalism’s potential for showing that perception cannot be understood in the restrictive way that generates the apparently insoluble problem of knowledge of the world. To avoid the consequences of that view of perception it is not necessary to establish the positive doctrine that what we perceive to be so is in fact, for the most part, just the way

¹⁰ I have argued for this in “Radical Interpretation and Philosophical Scepticism” in L. E. Hahn (ed.), The Philosophy of Donald Davidson, Library of Living Philosophers, Open Court 1999 and in “Anti-Individualism and Scepticism” in M. Hahn and B. Ramberg (ed.), Reflections and Replies: Essays on the Philosophy of Tyler Burge, MIT Press 2002.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



THE EPISTEMOLOGICAL PROMISE OF EXTERNALISM

things are in the world. It would be enough to show that on that restrictive view of perception no one could find himself presented with the completely general epistemological problem in the first place. That would not amount to a reassuring positive answer to the epistemological problem. It would not show that the restrictive view of perception is false, and that we do perceive and so know how things are in the independent world. But it would completely eliminate any danger of arriving at a negative sceptical answer to the epistemological problem from general reflections about what human beings perceive. One way of working towards such an ambitious result would be to show that if the restrictive view of the scope of perception were true, no one could consistently attribute any perceptions or beliefs to anyone. That would mean that the general epistemological problem could not then arise. Someone who does not attribute perceptions or beliefs to anyone cannot be faced with the problem of explaining how what people perceive gives them reason to believe what they believe about the world. Contrapositively, anyone who found himself faced with such a problem concerning perceptions and beliefs he thinks people actually have could not consistently hold that the contents of all those perceptions reach no further than what the restrictive view of perception implies. And if some of the perceptions he attributes are conceded to be perceptions of what is so in the independent world, once again the general epistemological problem does not arise. Either way, no one could be faced with a pressing general problem of how what is perceived can give us reason to believe what we believe about the world. I think the ideas behind an externalist understanding of the mind and its contents strongly encourage pursuit of such an outcome. It is not a project that anyone can be said to have explicitly carried out in convincing detail. Nor is it a goal to be reached overnight. It would require a thorough investigation in an externalist spirit of what a person must believe about the world in order to recognize the presence in it of people with perceptions and beliefs with determinate contents that he understands. Without being able to think of human beings in that way, no one could be presented with a pressing epistemological problem. And on an externalist understanding of the conditions for the attribution of such perceptions and beliefs, there is good reason to think that an attributor must take himself and others sometimes to perceive what is so in the independent world.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

THE EPISTEMOLOGICAL PROMISE OF EXTERNALISM



To ascribe beliefs about the world to anyone we must find some way to connect something we take to be so in the world with the believer’s believing what we think he believes. It is not enough simply to find something that we believe to be so in the person’s vicinity, for instance, or something of which he has no inkling. Ascribing beliefs requires finding some reason the believer actually has to believe what we take him to believe: either something that is a reason or at least something he takes to be such a reason. People’s reasons for believing what they do typically lie in other things they believe. An assumption essential to the traditional epistemological problem is that our reasons for believing what we do about the world ultimately depend at least in part on what we perceive to be so. That means that anyone who attributes beliefs must eventually attribute to the believer perceptions of something which is, or which is taken to be, reason to believe what he is thought to believe. The contents of the beliefs we attribute are to be identified somehow in terms of what the perceptions we attribute are thought to give the believer reason to believe. This applies both to attributions of beliefs and perceptions to ourselves in the first person and to the attribution of such attitudes to others. As things are, we typically ascribe perceptions with determinate contents by finding that the perceiver perceives something that we perceive or know or believe to be so in the world we all inhabit. On the view that no one ever strictly speaking perceives what is so in the independent world, we could not attribute perceptions to ourselves or others in this familiar way. Even if we had many beliefs about the way the world is, and we invoked the truth of those beliefs in attributing beliefs to ourselves or others, we could not find that anyone ever perceives the truth of any of those things we believe to be so. To do so would be to grant that some people do perceive what is so in the world after all, and that is what the restrictive view of perception denies. So the only perceptions we could ascribe to anyone on that view would be perceptions whose contents imply nothing about how things are in the independent world. But we would nonetheless have to regard perceptions of that kind as providing reasons to believe things about the independent world if we are to attribute any such beliefs at all. The epistemological promise I see in an externalist conception of the mind and its contents in general lies in its potential for showing how and why no one could do that. Focus on the conditions of the attribution of

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



THE EPISTEMOLOGICAL PROMISE OF EXTERNALISM

perceptions and beliefs raises the question of what contents, if any, we could ascribe to any perceptions at all if they had to be understood only in that restrictive way. And even if we could recognize some such perceptions, that raises in turn the further question of how and whether we could regard them as giving anyone reason to believe anything that goes beyond those restricted contents. If we could not, we could attribute no beliefs about the world to anyone. Pursuing these questions would lead to greater appreciation of what I think externalism shows to be a necessary connection between what a person can be understood to think or believe to be so and what he can intelligibly be thought to perceive to be so.¹¹ The restrictive view of perception, and so the traditional problem of perceptual knowledge of the world, thrive on denying or ignoring the connection. Exploration of that important link is what I think promises the richest epistemological rewards.

¹¹ I begin to explore this connection and its anti-subjectivist implications as applied to perceptions of and beliefs about the colours of things in my The Quest for Reality: Subjectivism and the Metaphysics of Colour, Oxford University Press 2000, chs. 5–7.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

4 Epistemological Self-Profile My interests in epistemology were generated at first more by questions about necessary truth and our knowledge of it than by “empirical” knowledge of the world around us. A widely-shared conception of perception and of the methods of confirmation thought to be sufficient to yield knowledge of the world made it look as if necessary truths could be known only in some other “non-empirical” or “a priori” way, and perhaps were not really true of the world at all. Despite disagreements of detail about exactly what can be perceived and exactly how those perceptual data support the conclusions drawn from them, this generally hierarchical or foundational picture of the basic structure of “empirical” knowledge dominated twentieth-century epistemology and gave determinate shape to the philosophical problem of our knowledge of the external world. The problem was to explain how knowledge of the world is possible in general, not simply how human beings come to know certain things about it on the basis of certain other things they already know about it. The goal was to explain, at least in outline, how anyone ever comes to know anything at all about a public objective world on the basis of perceiving the sorts of things human beings are capable of perceiving. On the prevailing conception, what was “directly” available to unaided perception alone was not part of that “external” world at all. The idea was that we could perceive everything we strictly speaking perceive without knowing anything about any such world, or even without the world’s being anything like what we believe it to be. This represented a truly formidable challenge. It seemed to me that some proposed solutions that accepted the challenge on its own terms (e.g., phenomenalism) simply did not work, and that others did not strictly fulfill the conditions that any successful solution would have to meet. At one point or another they all surreptitiously ascribed to the

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



EPISTEMOLOGICAL SELF - PROFILE

human subjects they were studying some knowledge of just the kind they were supposed to be accounting for. Paying close attention to the distinctive character of the completely general question of knowledge, and to the demanding requirements for a fully successful answer to it, made it look more and more as if no satisfactory account could be found. A negative or sceptical answer to the question of knowledge seemed to me the only acceptable outcome. Not only necessary truths, but no truths about the world at all seemed knowable on that conception of perception and knowledge. One response to this predicament, if it was recognized at all, was to reject the general epistemological problem as somehow illegitimate. Since a negative, sceptical answer to the question of human knowledge in general is absurd, paradoxical, or even in a sense unthinkable, it was tempting to conclude that there really must be no serious or perhaps even fully intelligible epistemological problem of that form. One strong version of this strategy was the roughly Kantian idea that the very formulation of the question carries within it a guarantee of an unthreatening positive solution to it. Our having any conception of ourselves as knowers of a world at all was thought to be enough to imply the possibility of human knowledge of such a world. This encouraging conclusion could perhaps look reachable on Kant’s own view of the world as “empirically real” but “transcendentally ideal”. But on any more reasonable conception of a world independent of us and our responses it was difficult to explain how a necessary connection could be found, or forged, between our thinking of the world in certain ways and the world’s actually being, and being knowable as being, any of those ways. This was the problem I drew attention to in “Transcendental Arguments” (Journal of Philosophy 1968). Without transcendental or any other form of idealism to appeal to, it looked as if only something like an implausible verifiability account of meaning could connect intelligible thoughts and beliefs about the world with the guaranteed possibility of knowledge of the world we thereby think about. What seemed to me to be needed for any real progress was greater understanding of the sources of the general epistemological problem. In the early chapters of The Significance of Philosophical Scepticism (Oxford University Press 1984) I tried to present the problem clearly from the ground up, hoping to identify the sources of those demanding requirements on the understanding of knowledge that seemed so difficult to

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

EPISTEMOLOGICAL SELF - PROFILE



fulfill. Later chapters investigated different attempts to dismiss or expose the philosophical question, or to sidestep or overcome it in some way. Something is to be learned from each, but I think none of them gets to the heart of the matter. Explicit presentation of the sceptical reasoning has been followed by a variety of fruitful responses to the problem since about the middle 1980s. One reaction has been not to challenge directly the reasoning by which the sceptical philosophical conclusion is reached, but to question the very significance or import of that conclusion itself. By drawing attention to the relativity to context or to current interests and inquiries that is characteristic of claims involving ‘know’ and other epistemic terms in everyday life, this brings into question the implications of conclusions reached in “philosophical” reflection for anything we ordinarily think or claim to know when not philosophizing. It remains unclear to me whether a “contextualist” strategy along such lines even countenances the possibility of a positive answer, or at least acknowledges the intelligibility of an all-encompassing question, about human knowledge. If philosophical scepticism is the best answer to a question that arises only in that special context called “philosophy”, is there nonetheless a question about the possibility and explanation of human knowledge in general? And is there a “contextualist” form of answer to that question about knowledge? Or can we at best simply describe the many different contexts in which we say ‘Human beings know many things’ and ask, ‘How do they know them’? Another kind of reaction has taken the form of careful responses to one or another step of the sceptical reasoning. This is not to treat the question as unreal or unintelligible, but rather to attempt to disarm the threat of a sceptical answer to it by exposing and rejecting one or another of the assumptions on which it rests. It can leave room for a positive answer to the general epistemological problem once the alleged fallacies of scepticism have been set aside. In at least some quarters these matters are still in dispute. One suggestion was to deny the necessity of an assumption apparently central to the sceptical case—that if one knows that p and knows that ‘It is true that p’ implies ‘It is true that q’ then one knows that q. But blanket rejection of that principle would seem to leave a great deal of our otherwise unproblematic knowledge unaccounted for. A more plausible strategy was to deny that one’s knowing that p implies one’s knowing

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



EPISTEMOLOGICAL SELF - PROFILE

that one knows that p. This has encouraged a variety of “externalist” or “reliabilist” conceptions of knowledge according to which conditions sufficient for one’s knowing that p can be fulfilled even if one does not know or have reason to believe that they are fulfilled. The success of any such theory depends first on finding a set of conditions that correctly captures all possible positive and negative instances of knowledge— a continuation of the Gettier “definitional” project that I have not entered into. But even a “definition” that is correct in that sense will give us what we want only if a non-“externalist” conception of knowledge actually plays an essential role in the sceptical reasoning and nothing else within that way of thinking still represents an obstacle to the kind of understanding of knowledge that we seek. I think the major obstacle in the traditional epistemological project is its commitment to the restrictive view of the proper objects of perception that has been with us in one form or another since at least the seventeenth century. That is what gives sense to the very idea of the world as “external” to or “beyond” everything we could ever directly perceive to be so. In “Understanding Human Knowledge in General” (1989), “Scepticism, ‘Externalism’, and Goal of Epistemology” (1994),¹ and “Perceptual Knowledge and Epistemological Satisfaction” (2004)² I tried to illustrate the inevitable dissatisfaction that I think even acceptance of an enlightened “externalist” conception of knowledge would leave us in as long as we continue to hold any such restrictive view of perception. The idea is that even if, according to an “externalist” conception of knowledge that we accept, we do know some particular thing about the world around us, we still will not satisfactorily understand our possession of that knowledge if we grant that we can never actually perceive that things are a certain way in the world we claim to know about. What I think is needed to undermine the appeal of any such view of perception is more than an “externalist” conception of knowledge. We need a better understanding of the conditions of our even thinking of ourselves as having the thoughts, beliefs, and perceptions involved in making sense of the philosophical question in a way that could seem to present a completely general challenge. The full Kantian aspiration of deriving substantive conclusions about the world and our knowledge of ¹ Both reprinted in my Understanding Human Knowledge, Oxford University Press 2000. ² Reprinted in my Philosophers Past and Present, Oxford University Press 2011.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

EPISTEMOLOGICAL SELF - PROFILE



it from no more than the conditions of our thinking of the world in certain ways still seems to me too demanding to promise results we could find palatable. But deeper understanding of the conditions of our even thinking and believing what we do about the world might serve to undermine the sceptical threat without going so far as to imply that the world is a certain way or that we do in fact know the truth of those things philosophical scepticism says we do not know. Establishing the denial of scepticism is not the only possible way to overcome what looks like a sceptical threat. It would be enough to find that we cannot consistently get into the position of understanding and recognizing the force of any such general threat to whatever knowledge we have. I have outlined in very programmatic form one version of a strategy along these lines and begun to explore some of its epistemological implications in “Kantian Argument, Conceptual Capacities, and Invulnerability” (1994), “Radical Interpretation and Philosophical Scepticism” (1994), and “The Goal of Transcendental Arguments” (1999)³. The idea is that in order to investigate human knowledge in general we must at least attribute to the human beings we would study, including ourselves, certain thoughts and beliefs and experiences concerning the world in which we take those people to live. There is a question of the conditions of anyone’s having the particular thoughts, beliefs, and experiences that we attribute, and there is a question of the conditions of our being able to ascribe such thoughts, beliefs, and experiences to anyone. The thought behind this strategy is that to understand people as having certain thoughts and experiences, and to be in a position to attribute such attitudes to them (which is required for thinking of them as faced by the general epistemological problem), we must understand those psychological states as connected in some ways with the surrounding non-psychological world that we take the thoughts and experiences in question to be about. In “The Epistemological Promise of Externalism” (2005)⁴ I tried to sketch how we must move beyond “externalism” about knowledge to a broader “externalist” or “anti-individualist” understanding of all mental contents. That would mean that we can make the attributions needed to formulate the general question of knowledge only because we ourselves are engaged in and ³ All reprinted in my Understanding Human Knowledge, Oxford University Press 2000. ⁴ In this volume.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



EPISTEMOLOGICAL SELF - PROFILE

have experiences and beliefs about the very non-psychological world that those attitudes are directed towards. This obviously does not imply that those beliefs of ours are true or that the world is just as we experience it to be. But it does mean that anything we must believe about the world in order to attribute to ourselves or others certain beliefs and experiences cannot consistently be regarded as false while we at the same time hold that we or others do have the beliefs and experiences we attribute to them. Whatever is indispensable to making any such attributions cannot itself be considered false consistently with our making the attributions in question. The inconsistency involved here would be structurally similar to that of the paradoxical sentence ‘I believe that it is raining, and it is not raining’. Although it is possible for both conjuncts of that sentence to be true together, no one could consistently hold that what the whole sentence says is true. Applied to the general problem of our belief in a world independent of us and our responses, this would mean that if there are certain things that must be believed by anyone who thinks of himself and others as believing in such a world, then the (admitted) possibility that we believe in such a world while those certain things are false is not a possibility we could consistently believe to be actual. Realizing that that is so, we would recognize no pressing need to explain how we know that that conjunctive possibility is not actual. Whatever is indispensable in this way to pursuit of a traditional epistemological inquiry into the credentials of our beliefs could be said to enjoy a certain kind of epistemological invulnerability. It is not something that could consistently be brought into question by any epistemological assessment that requires acceptance of that very belief as true. This does not imply that the indispensable belief in question is true, or known to be true, or that it could not be abandoned in the face of weighty considerations against it. It would not have been shown to be invulnerable to all possible attack. The invulnerability in question, given the indispensability, would be only invulnerability against a certain kind of negative epistemological assessment. But that is just the kind of assessment to which the traditional epistemological project would subject our beliefs about the world in general. So far this is no more than a sketch of the skeleton of what I see as a promising line of investigation into the sources of the traditional epistemological enterprise and what has looked like its inevitably sceptical

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

EPISTEMOLOGICAL SELF - PROFILE



outcome. It remains to be seen how far it can take us. But even a fully successful investigation along these lines would not yield a reassuringly positive anti-sceptical answer to a general question about human knowledge. It would at best expose the traditional epistemological challenge as something we cannot get into a position to be faced with. And that can leave us philosophically dissatisfied—as if there is something we still lack. We will have put the completely general epistemological problem forever behind us only when we have managed to overcome that apparently undying desire to explain human knowledge as a single, comprehensible whole, and from a position somehow recognizably outside it. Such a view from nowhere can appear to reveal the truth of a completely general philosophical scepticism. I think it is better to try to explain why there can be no such thing as a view from nowhere than to argue that with greater care we can see from there that scepticism is not really true.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

5 Explaining Perceptual Knowledge Reply to Quassim Cassam This is a response to Quassim Cassam’s “Seeing and Knowing”.¹ Cassam wants to overcome or undermine what he thinks is a potentially devastating dilemma standing in the way of any philosophically satisfactory explanation of knowledge of the world around us by means of the senses. He thinks that dilemma is embodied in certain requirements or constraints that I have laid down or insisted on for any satisfactory explanation of human knowledge of the world. He rightly asks why we should accept those requirements, or why we should think they cannot be met. He concludes that the alleged dilemma is not exhaustive and so presents no obstacle to a satisfactory philosophical understanding of knowledge, and he offers a positive explanation of his own. I think I can see how it might look as if my own response to the philosophical problem of our knowledge of the external world is what Cassam takes it to be, but I must say that it has never seemed that way to me. I do not think of myself as approaching the problem with a set of “requirements” to be met or with any demands about what our epistemological aims “should” be. I find myself presented with a line of thinking that I believe any interested, informed person will find challenging when he turns his attention sympathetically to what has come to be called the philosophical problem of our knowledge of the external world. My concern throughout has been to see whether we could learn something interesting about human knowledge, and about a philosophical ¹ European Journal of Philosophy 2009 (page numbers in parentheses in the text refer to this essay).

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

EXPLAINING PERCEPTUAL KNOWLEDGE



explanation of human knowledge, by understanding better than we do what is really at stake in that traditional problem. Of course, that might be a hypothesis it is not very profitable to pursue. Many philosophers appear to think it has already turned out that way. But if there are things to be learned from a study of that problem it is best to consider the problem in its strongest, most challenging form. And that means we must get as good an understanding of the source of the problem as we can, and then respond to it in the light of that enriched understanding. I think determining what is required for a satisfactory answer is a matter of discovery and careful description based on understanding, not a matter of simply laying down requirements. That does not mean it is easy to say exactly what a satisfactory solution to the problem aspires to, or what kind of understanding of our knowledge it would give us. The task I think of myself as engaging in is trying to discover and describe what the most sympathetic understanding of the problem can be found to reveal about what a satisfactory explanation of human knowledge really demands of us, and why. In that task, I think there are some pretty reliable “data” that would be recognized by everyone. But it is difficult to state in more general and illuminating terms what exactly those “data” can be taken to show about the epistemological challenge as it is best understood. The “data” I have in mind are what I think anyone would recognize as appropriate and inappropriate responses to attempts to meet the challenge given the understanding of the special character of the problem. For instance, when Descartes thinks he knows he is sitting by the fire but then realizes that he could be dreaming, he thinks he could know by means of the senses that he is sitting by the fire only if he knows that he is not dreaming at that moment. I think we can all see that it would be no solution to the problem Descartes faces at that point for him to ask a helpful friend whether he really is sitting there or only dreaming that he is. Asking a friend is sometimes a good way of finding out whether you are dreaming or not. But I think anyone who understands the philosophical problem as it is meant to be understood will see that it is not a satisfactory response to the challenge Descartes takes himself to be presented with there. That is what I am calling a “datum” in our understanding of the problem. Given that that response is inappropriate, the more difficult question is why. What can be said about what is wrong with that kind of answer at

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



EXPLAINING PERCEPTUAL KNOWLEDGE

that point? What do we understood about the problem when we see that that kind of suggestion is no good? This is a question about what is required for a philosophically satisfactory solution to the epistemological problem. And the answer will come out of one’s understanding of the problem. We might say that asking a friend whether you are dreaming is no good because it would be appealing to something you already take yourself to know about the world as a way of explaining how you know anything at all about the world in the face of the possibility that you could be dreaming. This would be to recognize that the problem is completely general—is as concerned with any knowledge of the external world at all. Any putative knowledge of the world that was appealed to would therefore be just as much in question as the original claim to know that one is sitting by the fire. These are attempts to identify what is required, or what falls short, of a fully satisfactory solution to this philosophical problem. Other requirements might be drawn from considering other ways in which apparently natural responses to the challenge can be recognized to fail. If all knowledge of the world on the basis of the senses is what is in question, and nothing that is already taken to be known about the external world can be appealed to, further requirements appear to follow. It looks as if knowledge of the world could be explained as coming somehow from perception only if what we get through the senses alone is always something less than knowledge of anything that is so in the external world. So even if perception gives us knowledge of something, it is tempting to say that the general epistemological problem, properly understood, requires “explaining knowledge of the world on the basis of another, prior kind of knowledge that does not imply or presuppose any of the knowledge we are trying to explain”. We could even say, more non-committally, that the problem requires “showing how our knowledge of the world comes to be out of something that is not itself knowledge of the world”. These attempts at more general diagnosis are of course all phrases I have used in trying to say what is required for a philosophically satisfactory explanation of knowledge of the world on the basis of the senses. They are derived from reflection on the ways in which various proposed responses to the epistemological challenge can be recognized to fail, and on what would accordingly be needed for them to succeed. Cassam quotes several of these more general remarks and takes them to

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

EXPLAINING PERCEPTUAL KNOWLEDGE



express requirements or constraints that I impose or insist on for any satisfactory solution to the problem. Probably none of these statements is fully satisfactory in general terms, independently of the particular kind of failed response it is drawn from. Not all of the descriptions are even correct in their own terms, as he shows, and they are not all equivalent. That would be unfortunate if these statements about what the problem demands were meant to formulate a completely general and inescapable dilemma—a kind of impossibility proof. But I have tried to explain why that is not the way I think of them. I think our understanding of the problem and our sense of the inadequacy of certain responses to it is the source of our recognition that various purported explanations of human knowledge of the world fail to give us what a satisfactory solution to the problem demands. Cassam thinks what he takes to be the conditions of success can be met, and he proposes a kind of explanation he thinks meets them. Given our understanding of the general philosophical problem as we have come to understand it, is the explanation he proposes satisfactory? What he offers is not just what he calls a “modest” explanation that explains a person’s knowing something by citing some other state the person is in that implies that he knows the thing in question. He proposes what he calls a “full blooded” explanation that explains “the transition from not knowing to knowing” (27), or shows “how our knowledge of the world could come to be out of something that is not knowledge of the world” (20). It is meant to explain how we get knowledge of the world by perception, but the kind of perception in question does not amount to knowing or even believing anything about the world. What Cassam thinks can explain our knowledge of the world is seeing or otherwise perceiving something, where the seeing is what Dretske called “simple seeing”: seeing some object x, not seeing something or other to be so.² From perceptions like that, which need not involve belief or knowledge about x, Cassam claims we can explain how people know things about the world. I think he is right about that. Someone standing in front of a barn and looking at it who is asked how he knows there is a barn in front of him can say ‘I see it’. That can be a satisfactory explanation of how he knows

² F. Dretske, Seeing and Knowing, Routledge 1969, ch. 2.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



EXPLAINING PERCEPTUAL KNOWLEDGE

there is a barn in front of him. It is a case, as Cassam puts it, of “something epistemic [his knowing] coming to be out of something non-epistemic [his simply seeing a barn]” (23). I agree that in the circumstances Cassam imagines that can be a perfectly satisfactory explanation of the person’s knowing what he does. As with all explanations, other things must also be true for that explanation to succeed. Those “background conditions” that need not be mentioned in the explanation itself typically involve facts of the surroundings in which what is to be explained takes place, including in this case facts of the orientation and capacities of the person involved. But in the conditions imagined, appeal to the ‘non-epistemic’ fact that the person sees a barn can explain the ‘epistemic’ fact that he knows there is a barn there. In other circumstances, a person whose well-functioning eyes are open and facing a nearby barn would not know in that way that there is a barn in front of him. Someone carefully rehearsing his part in a movie in dense fog might know from the screenplay he holds close to his eyes that there is now a barn in front of him. His seeing what is written on the screenplay appears to require knowledge or understanding on his part, but we could explain how he knows by saying simply ‘He sees the screenplay’. If that is something perceptual but ‘non-epistemic’, it could nonetheless explain his knowing what he does. A person’s knowing or coming to know something can be explained in many different possible ways, by appealing to many different kinds of factors, even different ‘non-epistemic’ factors. Austin reminded us long ago that someone can explain how he knows that there is a bittern at the bottom of the garden by saying ‘I was brought up in the fens’.³ The great variety of different possible factors to explain even one and the same event or state of affairs is due partly to the great variety of different things we can be interested in or can fail to know in wondering why something or other happened or is so. We might wonder how a certain person came to be in a position to recognize a bird as a bittern when he sees it, or how someone who cannot see five feet in front of him found out by sight about a barn ten feet away. Or we might not know by which sense a person in otherwise normal circumstances found out about a nearby barn: did he see it, or touch it, or even, perhaps, smell it? In answering any of ³ J. L. Austin, “Other Minds” in his Philosophical Papers, Oxford University Press 1961, p. 47.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

EXPLAINING PERCEPTUAL KNOWLEDGE



these and countless other similarly diverse questions we could explain a person’s knowing something, and in many cases appeal to ‘nonepistemic’ factors would be enough. But the philosophical problem of human knowledge by means of the senses is about human knowledge in general. And I think we can recognize that simply generalizing explanations of particular items of knowledge like this to human knowledge in general would not give us the kind of explanation we seek. From the person who sees the barn, or the screenplay, for instance, we could reach the generalization that human beings know the kinds of things they do about the world because they see things. I think we recognize that that does not give us a satisfactory answer to the philosophical problem we are interested in. But if that is not a satisfactory explanation, it is not because it is not true. It is true, after all, that people know the kinds of things they do about the world because they see things. It is also true, in Austin’s example, that the person knows the bird is a bittern because he was brought up in the fens. That case is not promisingly generalizable as it stands, since not everyone is brought up in the fens. But even to say that people know the kinds of things they do about the world because they are brought up in certain surroundings is, after all, true. But despite its being true, I think it does not give us the kind of explanation of human knowledge in general that the philosophical question asks for. What is lacking is at least some account of how seeing things, or being brought up in certain surroundings, or any other ‘non-epistemic’ state of affairs, leads to knowing something about the world. That requires some understanding of the capacities a person is master of and exercises on particular occasions in getting knowledge of the world in that way. The knowledge of the presence of a barn by the person who sees a barn and by the person who sees a screenplay can be explained in each case by their ‘non-epistemically’ seeing something. But they come by that knowledge in different ways. What else is true of them that accounts for that difference? What must be true of someone brought up in the fens if he knows simply by seeing a bird that it is a bittern? Even if it is true in general that only those who see (or otherwise perceive) objects in the simple, ‘non-epistemic’ sense ever know anything about the world, we do not understand that knowledge if we do not know anything else about perceivers that helps explain how they know the kinds of things we all know about the world on the basis of the senses.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



EXPLAINING PERCEPTUAL KNOWLEDGE

Finding the appeal to simple ‘non-epistemic’ seeing of objects insufficient on its own for the kind of understanding we seek is not to deny that that kind of perception is fundamental in the sense of being indispensable to the acquisition and possession of any knowledge of the world at all.⁴ It can be fundamental or basic in that sense without being fundamental in the sense of providing ‘foundations’ or ‘basic premisses’ from which conclusions about the world are to be inferred and thereby known. The indispensability of ‘non-epistemic’ perception gives no support to epistemological “foundationalism” in that sense. Cassam does not claim that it does, although he says he takes up the epistemological challenge by “focusing on the non-epistemic foundations of our knowledge” (28). The kind of explanation he offers implies only that ‘non-epistemic’ seeing can “result” (26) in knowing something about the world, or that such knowledge “originates” (27) in ‘non-epistemic’ perceptual encounters. That does not make simple ‘nonepistemic’ perception any more “foundational” or “basic” than any other factors that are equally indispensable for knowledge of the world. There are suggestions that Cassam perhaps thinks ‘non-epistemic’ perception has a more distinctively “foundational” role. He says “Seeing a barn in front of me can give me a reason to believe that there is a barn in front of me” (27), and even that “It is the seeing of a barn that tells one that there is a barn there” (25). This makes it sound as if simple ‘nonepistemic’ seeing has a greater role than I think Cassam should allow. If I am otherwise properly equipped, simply seeing a barn can take me from not knowing there is a barn there to knowing it. But I don’t think my seeing a barn tells me that it is there—or tells me anything, except perhaps that my eyes are in good working order. And seeing a barn can bring it about that I have a reason to believe something I did not have reason to believe before. In that sense it can give me a reason to believe that there is a barn there. But I don’t think my seeing a barn is my reason to believe it is there. Seeing a barn can be the reason why I believe it is there; but if I believe it is there because I see it, it is the barn, or the presence of the barn, that is my reason, not my seeing the barn.

⁴ Tyler Burge has stressed the indispensability of such de re perceptions for the possibility of any states or attitudes with propositional contents at all. See, e.g., “Postscript to ‘Belief De Re’ ” (esp. pp. 77–80) in his Foundations of Mind, Oxford University Press 2007.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

EXPLAINING PERCEPTUAL KNOWLEDGE



I think it is better to say, as Cassam usually does, that seeing a barn can “amount to” knowing that there is a barn there. ‘Non-epistemic’ seeing is not itself a case of knowing, but it “amounts to” knowing if certain other conditions are fulfilled, just as writing your name on a piece of paper in certain conditions can “amount to” your buying a house. It is only in the exploration of what these further conditions might be, and what is essential to their fulfillment, that I think we can hope to learn the best lessons from consideration of the philosophical problem of the external world. I think seeing a barn can “amount to” knowing that there is a barn in front of you only if you are a competent thinker with a capacity to understand and make judgements involving concepts that you know how to apply to things around you like barns. We are all so equipped that when we ‘non-epistemically’ see a barn in the appropriate circumstances we know that there is a barn there. I have suggested in my “Scepticism and the Senses”⁵ that the only satisfactory explanation of how that kind of knowledge is possible is that we see that there is a barn there. The matter of fact that we believe or know to be so in the world is the very fact that we see to be so. I think that kind of ‘propositional perception’ of the way things are must be acknowledged as part of the human repertoire in order to account for the knowledge of the world that we know we have. The promise I see in exploring the conditions of successful ‘propositional perception’ of the way things are is that propositional thought requires the understanding and possession of concepts that apply to whatever we think about. That understanding is embodied in a capacity for judgement. So, someone in full possession of a set of concepts must be capable of judging, and so accepting, certain things about the world to which his concepts apply. A capacity for perceptual judgement, in particular, is a capacity to perceive and so to judge, under the appropriate conditions, that such-and-such is so. That would mean that anyone who comes to know something about the world by ‘non-epistemically’ perceiving some object or other must already know or be capable of knowing something about how things are in the world. So someone could be understood to know something about the world on the basis

⁵ In this volume.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



EXPLAINING PERCEPTUAL KNOWLEDGE

of the senses only if he is also understood to know or be capable of knowing things about the world. Cassam points out that this does not show that the knowledge one gains on a particular occasion of ‘non-epistemic’ perception is arrived at on the basis of the knowledge one has in virtue of possessing the concepts needed for acquiring that perceptual knowledge. That ‘background’ knowledge need not be seen as “epistemically prior” in that sense to the knowledge gained by simple ‘non-epistemic’ perception; it can be no more than what he calls an ‘enabling’ condition for the acquisition of that perceptual knowledge. But that does not affect the implications I would draw from a perceiver’s having to know something or other about the world in order for his ‘non-epistemically’ seeing something to “amount to” his knowing something about the world on a particular occasion. The notion of “epistemic priority” need not come into it. If it is impossible for anyone to possess the conceptual and judgemental capacities essential to gaining perceptual knowledge by ‘nonepistemic’ perception without also knowing or being capable of knowing things about the independent world, then appealing to simple ‘nonepistemic’ perception alone does not explain in general “how our knowledge of the world could come to be out of something that is not knowledge of the world”. It does not explain in completely general terms “the transition from not knowing to knowing”. Those are conditions of success that Cassam sets for the kind of explanation of knowledge he offers. But I do not think those conditions are met when an explanation of human knowledge in general is in question. That does not mean that ‘He sees it’ (understood ‘non-epistemically’) cannot be a perfectly satisfactory explanation of how someone knows that there is a barn in front of him. The promise I see in exploring the conditions of successful ‘propositional’ perception of the way things are is not that it might eventually give us what I think Cassam’s suggestion on its own does not give us: a satisfying answer to the completely general philosophical question. Rather, if those conditions of ‘propositional’ perception are as I think they are, and we can understand clearly why they must be that way, we could not intelligibly understand ourselves to be faced with a completely general problem of how we can know anything at all about the world on the basis of the senses. That is an outcome to be hoped for from a better understanding of perceptual knowledge of the world.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

6 Scepticism and the Senses I will speak in very general terms about some lessons I think can be learned from philosophical scepticism and some strategies to be pursued in understanding human knowledge in the right way. The best route through messy or dangerous terrain can sometimes be found by surveying it from a great height. There is a certain kind of philosophical reflection on human knowledge in general that can lead to the conclusion that no one can know anything at all about the world around us. I think the considerations that can make that conclusion look plausible, perhaps even unavoidable, always depend on certain ideas about sense-perception and its role in knowledge of the world. I doubt that anything about knowledge itself— any distinctive features or requirements of knowledge in general, considered independently of what we can and cannot perceive—could make a general denial of the possibility of knowledge look plausible. Identifying the assumptions about sense-perception that are responsible for the ease with which a sceptical conclusion can be reached, and learning why those assumptions are unacceptable and what should be put in their place, would be a rich reward for taking the traditional sceptical reasoning seriously and trying to get to the bottom of it. We could perhaps even discover how knowledge by perception is to be correctly understood. That would be a substantial, positive payoff: much better than what is offered by those who see in the sceptical reasoning nothing more than arbitrary insistence on impossibly high standards for knowledge or one or another form of confusion, equivocation, or legerdemain. When Descartes in his first Meditation¹ began to investigate what he called the “foundations” or “basic principles” on which he thought all his ¹ Descartes, Meditation I, in J. Cottingham, R. Stoothoff, D. Murdoch (ed. and trans.), The Philosophical Writings of Descartes, Cambridge University Press 1984, vol. 2 (page numbers in parentheses in the text refer to this work).

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



SCEPTICISM AND THE SENSES

former beliefs had rested, he went straight for “the senses”. “Whatever I have up till now accepted as most true (or true par excellence) I have acquired either from the senses or through the senses”, he says (12). It seems to him at first that many beliefs derived from the senses about one’s immediate environment are simply impossible to doubt: for instance, his perceptions of the room he is in at the moment. But then he reminds himself of dreams, and he acknowledges that on “other occasions” he has been “tricked by exactly similar thoughts while asleep” (13). That would represent no threat to his knowledge if he could tell right now that he is not asleep and dreaming; he would know that he is not being tricked in that way this time. But that is what he finds he cannot know even in the present case. He says that careful thought about it leads him to “see plainly that there are never any sure signs by means of which being awake can be distinguished from being asleep” (13). I think Descartes is persuaded of this because at this point in his first Meditation he is considering only knowledge based on the senses alone. That he is thinking of it in this way at the beginning is borne out by what he says later in the sixth Meditation, when he looks back and recalls the reasons he says had originally undermined his faith in the senses. There were the familiar mistaken judgements about things that are too small or too far away to be seen properly, for instance. But he also mentions two other “very general” reasons he had for doubting the deliverances of the senses at that earlier point. The first was that every sensory experience I have ever thought I was having while awake I can also think of myself as sometimes having while asleep; and since I do not believe that what I seem to perceive in sleep comes from things located outside me, I did not see why I should be any more inclined to believe this of what I think I perceive while awake (53).

But that fact—that each of his sensory experiences, as far as he could tell, could be a dream—was good reason for doubt about “things located outside” him only because at that early point he was also entertaining a “second general reason” for doubt. As he puts it: since I did not know the author of my being (or at least was pretending not to), I saw nothing to rule out the possibility that my natural constitution made me prone to error even in matters which seemed to me most true (53).

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

SCEPTICISM AND THE SENSES



As long as both those “general reasons” for doubt were in play, he says he “had no trouble in refuting” the “reasons for my previous confident belief in the truth of the things perceived by the senses” (53). By the time he reaches the sixth Meditation, Descartes no longer entertains those two earlier general reasons for doubt. He now thinks he knows the author of his being, and he knows that that author guarantees that everything he, Descartes, clearly and distinctly perceives to be true is true. His “natural constitution” does not make him “prone to error even in matters which seem to him most true”. That is why he can say in the sixth Meditation that there is a perfectly good way of distinguishing between dreaming and waking. He can sometimes tell that he is not dreaming on the basis, roughly speaking, of the greater “coherence” of his waking experience. This does not directly contradict what he said in his first Meditation.² There, the question was only whether he can know he is not dreaming on the basis of what can be known from the senses alone. By the sixth Meditation Descartes thinks he knows things from sources other than the senses. And he thinks the additional non-sensory knowledge he has uncovered can be combined with knowledge from the senses alone to give him knowledge that could not be got from the senses alone. With the help of the author of his being he can use what he gets through the senses to know that he is not dreaming at a particular time. Descartes originally regarded the possibility of dreaming as a general reason to doubt because, as he put it, “every sensory experience I have ever thought I was having while awake I can also think of myself as sometimes having while asleep”. When only knowledge based on the senses alone is in question, a plausible line of reasoning can lead from that premiss to the conclusion that “there are never any sure signs by means of which being awake can be distinguished from being asleep”. Not only is the reasoning plausible; when properly understood, I think it is valid. If each sensory experience I have, as far as I can tell from that experience alone at the time, could be a dream, I could know of “things located outside me” by means of sensory experience alone only if I know or have reason to believe that the sensory experience I am having is a real perception and not a dream. Nothing in that sensory experience itself ² For the idea that there is a conflict between the claims about dreams in the first and the sixth Meditation see, e.g., Bernard Williams, Descartes: The Project of Pure Enquiry, Harmondsworth 1978, pp. 309–10.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



SCEPTICISM AND THE SENSES

can reliably indicate to me that it is not a dream. So to know on the basis of sensory experience that it is a real perception and not a dream I would have to have some other sensory experience or perform some sensory test that would indicate to me that the experience in question is real and not a dream. But whatever that sensory experience or the sensory outcome of that putative test happened to be, it would not in itself reliably indicate to me that the first experience I was testing for was real and not a dream unless I knew or had reason to believe that the new sensory experience of testing was itself real and not a dream. No sensorily distinctive mark or feature of that new experience could indicate that. So some further experience or test would be needed to determine whether it was a dream or not. And so on. Under those conditions it would be true, as Descartes says, that “there are never any sure signs by means of which being awake can be distinguished from being asleep”. No sensory experience or any set of sensory experiences alone could ever suffice. If I have to know that I am not dreaming in order to know anything about “things located outside me”, this means I could never know any such thing on the basis of the senses alone. This conclusion, reached in this way, does not rest simply on the requirement that knowing that you are not dreaming is a necessary condition of knowing something about “things located outside” you. It could be said that if that is simply a general necessary condition of knowing such things, the condition will be fulfilled whenever you do know something about “things located outside” you. So in itself it would be no threat.³ But this argument demands more than that. This argument accepts that knowing that you are not dreaming is a requirement on knowing that there are “things located outside” you, but it also insists that that condition is to be fulfilled on the basis of sensory experience alone. That is what represents a threat to knowledge, given the conception of sensory experience that is at work in the argument. On that conception, the requirement of knowing that you are not dreaming, and knowing it on the basis of sensory experience alone, introduces a regress that precludes fulfillment of what it says is a necessary condition ³ This is in effect the response of James Pryor to the general sceptical argument he takes me to reconstruct from Descartes’ Meditations. See his “The Skeptic and the Dogmatist”, Nous 2000.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

SCEPTICISM AND THE SENSES



of knowing anything about “things located outside” us. For each putative case of knowing such a thing, a necessary condition of knowing it can be fulfilled only by knowing something else, and so by fulfilling a different necessary condition of knowing that different thing. Each step in the attempt to fulfill a necessary condition of the truth of the original knowledge-claim introduces a different and so far unfulfilled requirement. On these assumptions no one could know anything about “things located outside” us. This conclusion, taken on its own, seems so outrageous that it can look as if Descartes must have something other than ordinary, familiar, everyday (or every night) dreaming in mind. Don’t we all often and easily know on particular occasions that we are not dreaming? If Descartes thinks we do not know such a thing, or even cannot know it, it is easy to think he must have in mind some very special or extraordinary kind of state, not dreaming “in the ordinary sense”. This is the response of Penelope Maddy.⁴ She points out that the dream possibility as Descartes understands it seems to operate as “the functional equivalent of the Evil Demon hypothesis”. That hypothesis implies that anything you could appeal to to show that the Demon is not fooling you at any given moment could be produced by the Demon himself, as far as you could tell, and so you could never know. If that is true of this “dream hypothesis” as well, you could never know that you are not dreaming. Maddy takes this to suggest that Descartes cannot have “ordinary” dreaming in mind, since we can often tell by ordinary procedures that we are not dreaming. I think that at the point we are now considering in Descartes’s reasoning the dream possibility is equivalent in its force to the Evil Demon hypothesis. The Evil Demon hypothesis is stronger in also threatening knowledge of mathematics and other things “related to the intellect” (121) but with respect to sensory knowledge of the contingent world around us, the two “hypotheses” or possibilities have the same effect. Any sensory experience you might appeal to to show that a particular experience is not a dream could, as far as you could tell, be a dream. But the dream possibility has this implication at this point not because something other than ordinary, familiar dreaming is in question. It is because of the way the requirement of knowing that you are not

⁴ Penelope Maddy, Second Philosophy, Oxford University Press 2007, p. 23.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



SCEPTICISM AND THE SENSES

dreaming on a particular occasion must be fulfilled, as Descartes here thinks of it. It is not that the requirement cannot be fulfilled because some “extraordinary” kind of “dreaming” is in question. It is rather that the requirement cannot be fulfilled when the only resources available are limited to what can be known on the basis of the senses alone. What is a reasonable requirement on knowing something about “things located outside” you—that you know you are not dreaming in the ordinary, familiar sense—becomes an “extraordinary” and apparently unfulfillable requirement on knowing such a thing under the restrictive conditions imagined to be required for knowledge in the argument of the first Meditation. Of course, Descartes does not believe this gloomy sceptical conclusion about our knowledge of the world by the time he reaches his sixth Meditation. That is because by then he no longer thinks of the possibility of knowledge of the world as restricted to what is available from the senses alone. The requirement that you have to know that you are not dreaming in order to know any such thing then has a different significance; it does not represent the same kind of obstacle to knowledge of the world as it had seemed to earlier. I think this way of understanding Descartes’s first Meditation reasoning can reveal something important about how perceptual knowledge must be understood. Or at least the unacceptability of its conclusion can reveal how perception must not be understood if it is to be the basis of our knowledge of the world. The senses are obviously important for knowing anything about the world. That is the undeniable point from which Descartes starts. On his view we cannot know anything about “things located outside” us on the basis of perceptual knowledge alone. But Descartes appears to assume that we can know something on that purely sensory basis alone. And that purely sensory knowledge, he thinks, although not itself knowledge of “things located outside” us, can be combined with knowledge available from some non-sensory source to give us knowledge of the familiar world. So our knowledge of the world is a combination of knowledge from two distinct sources, not from sense-perception alone. The additional non-sensory knowledge Descartes himself draws on is knowledge of a beneficent God who guarantees that the highly restricted material we get through the senses, when used carefully, can serve as a reliable indication of how things are in the wider world.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

SCEPTICISM AND THE SENSES



I think few of us will be tempted by that particular Cartesian route to knowledge of the world. But it can still look, and it has looked to many, as if the only way to explain perceptual knowledge of the world is to find some other non-sensory source of knowledge to supplement what we get from the senses alone. Bertrand Russell, in his book Human Knowledge: Its Scope and Limits,⁵ for instance, thought the best we could do was to identify certain “nonempirical” “principles” or assumptions that we can see must be accepted if what we get from the senses alone is to support or validate the kinds of things we all believe about the world. Others have sought other kinds of non-sensory or a priori guarantees that we are getting things right on the basis of what we get from the senses. This is not the lesson I think we should draw from Descartes’s first Meditation argument. I do not think anything along these lines can give us the kind of understanding of human knowledge that we seek. For one thing, the demand for non-sensory knowledge to supplement what we get from the senses rests on the assumption that we never actually perceive anything to be so in the world around us. If that view of perceptual knowledge is accepted, how could we find something from some completely non-sensory source that would secure the right kind of connection between what we can perceive and what is so in a world beyond all possible perception? And whatever looked as if it would do the job, what reasons could we have to accept it? It seems that the most we could know is that if we are to have any reason to believe what we do about the world there must be some such connections between what we can perceive and what is unperceivably so, and we must have some reason to believe in some such connections. But that is only a conditional statement. It says what we have to have reason to believe, but it gives no independent reason to believe anything of that kind. It just shows how desperately some such additional reasons would be needed, on this understanding of the limits of what we can know by perception alone. This might be thought too quick. It could be said that we ourselves do not have to recognize that certain “non-empirical principles” linking our perceptions with the world give us reason to believe what we do; it would be enough if some such “principles” were in fact true. There would then

⁵ B. Russell, Human Knowledge: Its Scope and Limits, Allen & Unwin 1948.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



SCEPTICISM AND THE SENSES

be reliable connections between our perceiving what we do and the way things are beyond it, whether we knew it or took ourselves to have reason to believe it or not. What we believed about the wider world would then be largely true, and we would believe it because of what we find in our perceptual experience. I think this kind of explanation would still leave us in an unsatisfactory position for understanding ourselves as knowing things about the world. Whether we know or not would depend only on the connection or lack of connection between our perceiving what we do and the truth of what we believe about the wider world, whether we know that such connections hold or not. But this view still accepts the idea that we never perceive that such-and-such is so in the independent world. Nothing that we believe to be so in that world is ever perceived to be so; it is something we know or have reason to believe only if the relevant connections do in fact hold. Of course, we believe that the world is a certain way, and we believe that such connections hold between what we perceive and what we believe, but on this view we cannot perceive that they hold, or perceive anything that implies that they hold. On this view, the whole conception of the enduring world we believe in is nothing more than a huge hypothesis, no part of which can ever be perceived to be so by anyone. The most we would be in a position to say about ourselves is that we believe that the world is the way our perceptions lead us to believe it is, but without our ever actually perceiving that the world is that way. Understanding ourselves as at best in this kind of position is unsatisfactory because whatever we can say about ourselves could with equal conviction and support be said by someone or some community for whom the relevant connections between what they perceive and the wider world they believe in do not in fact hold. If such benighted people could nonetheless believe certain things about the world, and could also believe that they know many of those things to be true, they would not know what they think they know, since the relevant connections do not in fact hold. If such connections do hold in our case, as we think they do, then we do know, according to that “externalist” conception of knowledge. But there is nothing we can say to express our understanding of our own epistemic position that those others could not equally say about their position, even though they do not in fact know. That is the unsatisfactory position I think this kind of explanation would leave us

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

SCEPTICISM AND THE SENSES



in. We cannot understand ourselves as knowing how things are in a world that we can never perceive any part of.⁶ For this and other reasons, I think that to understand properly how our knowledge of the world is possible we cannot settle for less than our being able sometimes simply to see or in general to perceive that suchand-such is so in the world we take ourselves to have knowledge about. That is the moral I would draw from Descartes’s sceptical argument. We should get as far away as possible from the conception of perceptual knowledge to be found in Descartes’s first Meditation. We should abandon the idea of purely sensory knowledge that is neutral with respect to how things are in an independent world. I don’t mean that we cannot sometimes know simply how things look or appear or otherwise seem to us at a given moment without committing ourselves as to how things actually are. I mean we should question the whole idea of purely sensory knowledge understood as knowledge that someone could have without knowing or having the resources for knowing anything else—in particular, without knowing anything about “things located outside” us. That is the assumption about perception that appears to have completely general sceptical consequences. That is the assumption that must be rejected. It raises many large and difficult questions that I cannot settle here. Is it really possible to have sensory or perceptual knowledge of many things without having any purely perceptual knowledge of the wider world? Could someone know how things are within his perceptual experience— how things are perceived to be—without knowing or being capable of knowing by perception alone anything that is and would be so whether it is perceived or known by him or anyone else or not? If what a person perceives and thereby knows to be so is something that would be so whether anyone perceived it or not, that would be perceptual knowledge of how things are in the independent world—a world that does not depend for its existence on being perceived. The conception of perceptual knowledge that lies behind the sceptical reasoning denies that that is possible. So the question it raises is whether everything anyone could

⁶ I have tried to explain and defend this response in “Understanding Human Knowledge in General” and “Scepticism, Externalism, and the Goal of Epistemology” (both reprinted in my Understanding Human Knowledge, Oxford University Press 2000) and in “Perceptual Knowledge and Epistemological Satisfaction” reprinted in my Philosophers Past and Present, Oxford University Press 2011.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



SCEPTICISM AND THE SENSES

ever perceive and know to be so by perception alone could be limited to things that are so only if they are perceived and thereby known to be so. Perceptual knowledge understood in that way would not be knowledge of any truths about “things located outside us”, in Descartes’s phrase. I think the sceptical reasoning shows that we could have no satisfactory understanding of our knowledge of the world on that conception of perceptual knowledge. That is the moral I would draw from Descartes’s first Meditation reasoning. That is why I think that in understanding how our knowledge of the world is possible we cannot settle for less than our sometimes actually perceiving that such-and-such is so in the world. By this I mean more than that we must be understood to perceive objects. Perceiving objects that exist in the independent world, objects that would be there whether anyone perceives them or not, is not enough to explain how we know that those objects are there, or how we know anything about them. You can perceive an object without knowing anything about it; without even having any beliefs about it. The kind of perception I think is needed to account for knowledge of the world is therefore what might be called propositional perception: not just perceiving an object x, but perceiving that such-and-such is so. That kind of perception does typically involve perceiving one or more objects. But knowledge is a matter of knowing that something or other is so. So for the explanation of knowledge based on perception it must be explained how our perceiving what we do contributes to our knowing that such-and-such is so. Our perceiving that such-and-such is so would explain our knowing that such-and-such is so. There is no better reason for believing and claiming to know that p than seeing or otherwise perceiving that p. Since this kind of perceiving is a way of knowing, one strategy for coming to understand this kind of perceptual knowledge is to start with what it takes to know something. Knowing that something or other is so requires some way of thinking of or understanding what you claim to know. It involves possession of some concepts that the knower is master of; what he claims to know is expressed in terms of those concepts. Concepts are “predicates of possible judgements”, as Kant rightly insisted and many philosophers since his time have ignored or denied, to their peril. To possess a concept, where that is relevant to knowledge, is more than simply being able to respond in a distinctive way to objects of a certain kind, or to respond in different ways to objects that differ from

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

SCEPTICISM AND THE SENSES



one another in certain ways. The electric eye at the supermarket door can do that much. To possess and understand a concept is to have a capacity to employ that concept in making judgements, in saying or thinking something that is true or false. That does not mean that one does or must actually judge that way. But what can be thought or supposed or entertained without accepting it is something that could be accepted or judged true if the circumstances were right. So someone who possesses concepts and thereby has a capacity to make or accept judgements of a certain kind when appropriate must have some sense of what those appropriate conditions are and a capacity to recognize the presence of those conditions some of the time, on pain of not being intelligible as possessing the concepts in question. To recognize certain conditions as making it appropriate to judge or accept or endorse such-and-such as true is to take certain considerations as reasons in favour of making such a judgement or believing such-andsuch. A capacity to recognize or acknowledge something as reason to believe or accept something is essential to a capacity to judge or believe that such-and-such. Each exercise of that capacity is itself a matter of judgement—of knowing when and how to respond to certain accepted considerations in the appropriate ways. No one could know anything without such capacities. Perceptual knowledge requires a capacity for judging or accepting something that the person can perceive to be so. So such a perceiver must have concepts that could express something he can see or otherwise perceive to be so. And he must know or otherwise recognize when such a thing is so, or when such a judgement is true. That would be to perceive that p, and so to know that p. If a person’s purely perceptual knowledge did not extend to anything that is or would be so independently of its being perceived, what would such a person be capable of perceiving and thereby knowing to be so? In particular, what kinds of concepts could a perceiver with such a limited range be said to be master of? It is difficult to see how such a perceiver could possess and apply concepts that can also be truly applied to perceived but independentlyholding states of affairs. But then concepts of ordinary, everyday objects in familiar surroundings, like tomatoes on a plate or trees on a hillside or waves at a beach, would seem to be disqualified; anyone who possesses such concepts and is not blind could sometimes see that that is how

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



SCEPTICISM AND THE SENSES

things are on a plate or a hillside or a beach. It would not help to describe the person’s perceptual knowledge as limited to its being for him only as if there are tomatoes on a plate, or his having a perception that is said (in the annoyingly evasive phrase) to be “as of ” there being tomatoes on a plate. That a person’s perceptions are as of one state of affairs rather than as of some different state of affairs is something the person must be able to recognize and therefore to understand if he can be said to know by perception that it is as if, or that his perception is as of, there being tomatoes on a plate before him. This is not to say that anyone who has such a perception must believe that there are tomatoes on a plate before him. The question is only whether someone could even make sense of his having a perception as of such-and-such without understanding what would be so if such-andsuch were so, and therefore without a capacity to perceive it to be so if it is the kind of matter of fact that can be perceived. Could someone even be said to understand what tomatoes and plates are, and what it is for tomatoes to be on a plate, if he does not have the capacity ever to see that such a state of affairs holds right before his eyes? But there must be some determinate states of affairs he has a capacity to perceive to hold if he is capable of perceptual knowledge at all. Perhaps a person of such restricted perceptual knowledge could be said to perceive and thereby to know only that there is something red and something white. But even that is questionable as something he can be said to know to be so. Does he understand the words ‘red’ and ‘white’ to be predicates, for instance? And if so, what kinds of objects does he understand them to be predicable of? If they are predicates of such things as tomatoes and plates, we are back with the question whether he could understand such predicates at all if he cannot ever perceive and thereby know that there are red tomatoes on a white plate. Or perhaps he could be said only to perceive red, or redness, and white, or whiteness. But that is to say only that he sees certain objects or properties; an x that he is aware of in some way. That says nothing about what he makes of the object he sees, or what he knows about it, or what he knows about anything. In seeing only such an object a perceiver might be below the minimum conditions for understanding what he is perceiving in one way rather than another, and so for knowing by perception that anything at all is so. Simply to perceive or be aware of something is not so far to be

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

SCEPTICISM AND THE SENSES



intelligibly connected with knowing anything of the kind human beings know about the world. These are the kinds of questions I think we must pursue in order to expose and eventually undermine the ways of thinking about senseperception that have seemed to imply that we can know nothing by perception about the world around us. They are questions about what it takes for someone to be in the kinds of psychological states that are even so much as relevant to questions about perception and knowledge of the world. And they are questions of what we can and cannot find possible, or intelligible. It is we, after all, who confront or are challenged by the problem of our perceptual knowledge of the external world. But we can understand ourselves as perceiving and believing and knowing things only from within whatever position we already occupy in understanding the world, and so only as part of understanding everything else that is so in the very world we all perceive and believe in. We are all perforce “externalists” in that sense about what perceptions, thoughts, and beliefs we can understand human beings to have, to express, and to act on. This is no mere “externalism” about the concept of knowledge in particular. It is a conception of the mind and its contents, or an understanding of psychological facts in general, as being all of a piece with nonpsychological facts of many kinds. To think of psychological facts in this way is to see them as not intelligible on their own, completely divorced from facts of the accompanying non-psychological world in which we understand them. If this is how we do and must make sense of things, we can understand ourselves and others as having certain kinds of perceptions and thereby believing and knowing certain things about the world only because we accept and are immersed in the very general conception of the way things are that we in large part share with those we can understand. We find without difficulty that we can sometimes see and thereby know such things as that there are red tomatoes on a white plate, for instance, or even that we are sitting by a fire in a dressing gown with a piece of paper in our hand. We can make no sense of the alleged possibility of all human perceivers finding themselves only in the general plight that would be described by generalizing directly from what someone perceives and thereby knows to be so on a particular occasion when he cannot tell by perception alone whether he is dreaming or not.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



SCEPTICISM AND THE SENSES

I have given no argument for most of these strong and far-reaching claims. I have been trying to identify something essential to the traditional problem that I think we cannot find intelligible and possible. But what follows about epistemological scepticism? My suggestion is that, given the conditions of our acknowledging any psychological states, and in particular any determinate perceptions, beliefs, and knowledge, as part of the world, we cannot make full sense of perceptual knowledge as restricted only to a limited domain that includes no facts of the wider world. Does that imply that the answer to the traditional problem of our knowledge of the external world is ‘Yes, we do know that things are in general as we perceive them to be’? I want to say ‘No, that does not follow’. I think our not finding such restricted perceptual knowledge intelligible or possible on its own shows rather that we cannot really get ourselves into a position from which a completely general sceptical problem of our perceptual knowledge of the world presents itself to us. But from that it does not follow that we do have the knowledge of the world that a sceptical answer to the traditional problem denies that we have. We cannot say we have found that the answer to a problem we cannot intelligibly be presented with is ‘Yes’. I think exploring the conditions of perceptual knowledge would uncover an understanding of human perception on which the traditional problem of the external world could not even seem to present us with a possibility. We would see how that apparent problem represents a demand to understand knowledge in general from a position no one can possibly occupy. That would be a real advance in understanding, but not because it is a positive answer to the epistemological problem. It would be a way of overcoming and so abandoning a view of the limits of perceptual knowledge that has long seemed to present an obstacle to understanding human knowledge. We might then simply lose the traditional epistemological problem altogether. It would have gone away with the restrictive view of perception it appears to rely on. But if that problem in that form has gone away, could we say that we now do understand what we feel we want to understand about ourselves? Or would we still feel that we cannot have something it seems to make sense to aspire to: an understanding of human knowledge of the world in general that does not itself presuppose any of the knowledge we want to understand? Finding that we cannot achieve such an understanding of ourselves can leave us still dissatisfied. I do not want to deny that. But in the face of that

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

SCEPTICISM AND THE SENSES



perhaps inevitable dissatisfaction, I think it is better to concentrate on explaining more and more fully why we can never reach such a nonengaged standpoint than to hold out the hope of eventually explaining how all knowledge of the world is possible from a position recognizably outside all human knowledge.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

7 Seeing What is So I think any satisfactory understanding of human perceptual knowledge must make room for the fact that we know things about the world around us by perception alone.¹ Most philosophical theories of knowledge have denied that. They have said in one way or another that the most we can know by perception alone is something that in itself implies nothing about the way things are in the world around us. This has not been taken to show that therefore we can know nothing about the world. It has been concluded, rather, that perceptual knowledge of the world is a combination of what we know by sense-perception alone plus other things we know from other sources. I think this way of understanding our knowledge of the world, when pressed, cannot really explain to those of us who have that knowledge how the knowledge we have is possible. It leaves us vulnerable to doubts about the basis for those steps it says we take beyond the knowledge we get from perception alone. That is a long story that I will not go into here.² Here I would like to suggest by contraposition that since we can and do know many things about the world around us by perception alone, all such restrictive views of perceptual knowledge must be rejected. The question is whether we can really understand perceptual knowledge of the world in this less restrictive way that I think we need. I will speak only about visual perception: seeing. I think we very often see and thereby know that p, where what fills the place of ‘p’ is something that is so in the world around us and would be so whether anyone ¹ This is the main proposal of my “Scepticism and the Senses”, European Journal of Philosophy 2009, which is reprinted in this volume. ² I have tried to elaborate these doubts in my “Understanding Human Knowledge in General” and “Scepticism, ‘Externalism’, and the Goal of Epistemology”, both reprinted in my Understanding Human Knowledge, Oxford University Press 2000, and in “The Epistemological Promise of Externalism” in this volume.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

SEEING WHAT IS SO



perceives it to be so or not. And I think it is possible to get knowledge of that kind that is not based on anything else we know that we somehow combine with what we get from sense-perception alone to give us knowledge that p. We simply see, and in seeing know, that p. So, I think we need the idea not just that seeing is believing, but that seeing is, or can be, knowing. I see, and in seeing know, that there is a chair in this room. Knowledge is knowledge that something or other is so. It involves a propositional thought that is true. So, the seeing that p that amounts to knowing that p involves a propositional thought that is true. The very same thought is involved in both attitudes: what you see to be so is what you know to be so. So, the seeing I have in mind could be called “propositional” seeing, or seeing that takes a propositional object. Putting it that way can be misleading if it suggests that what you see is a proposition. No; what I see is that there is a chair in this room. To say that that is a propositional object is only to say that my seeing is described in a sentence in which the complement of the perceptual verb ‘see’ is a sentence with a truth-value, not a singular term referring to an object. Of course, I also see that chair. That is stated in a sentence with a perceptual verb followed by a singular term. This could be called “objectual” seeing—the seeing of an object.³ “Propositional” seeing that p typically, but not always, involves seeing objects in this “objectual” sense. Seeing that the cat is on the mat typically involves seeing the cat and seeing the mat. But when I see only that it is foggy everywhere, for instance, I see and thereby know that that is how things are, but I do not see any objects at all. And it is also possible to know something about an object I do not see by seeing a certain other object. I can see that my neighbour is at home by seeing her car parked in front of her house. But in the typical case now before us, I know that there is a chair in this room because I see that chair. I might have come to know that there is a chair in this room in some other way, not by seeing that chair. What would explain my knowing what I know in that case would be something other than my seeing that chair. Seeing that chair is in fact how I come to

³ This is what Fred Dretske called “non-epistemic” seeing in his Seeing and Knowing, Routledege 1969.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



SEEING WHAT IS SO

know that there is a chair in this room right now. My knowledge is of something “propositional”, but my seeing that chair is not “propositional” seeing. I think what I am calling “objectual” seeing does not require knowing or perhaps even believing anything about the object you see. It could be true of me that I see that chair that is in this room even though I do not know it is a chair or know that it is in this room. Under those conditions, I could still see that chair. A creature can see an object that is blocking its way even if the creature has no knowledge of anything in the world beyond a capacity for moving around in it without obstruction. In a sentence of the form ‘He sees x’, any expression that is true of the object that ‘x’ stands for can be put in its place without changing the truth-value of the sentence. Many things can be true of the objects I see without my knowing that they are true. I can see something that costs $50 without knowing that it costs $50, just as a dog next to me can see that same thing that costs $50. But I could not see that there is a chair in this room without thinking and knowing that there is a chair in this room. And, as with all knowledge, I could not know it if there were no chair in this room. So I could not see that there is a chair in this room if there were no chair here. That is how this kind of seeing amounts to knowledge. I think we must be capable of this kind of seeing and knowing if we know things about the world by visual perception alone. Seeing an object in the “objectual” sense requires certain capacities or expertise on the part of the perceiver. An ability to see things involves having one’s attention visually drawn to the object, or discriminating it from its surroundings in some way, or responding to it attentively, perhaps tracking it with your gaze if it is moving. What exactly is involved in seeing an object is not easy to say. But a great many different kinds of creatures in the world all have that capacity; they can see and respond to objects. When I see that chair, and attend to it or find it within my awareness, it is available to me for thinking and saying things about. It is therefore available to me to come to know things about. But I don’t think I need to know anything about the object I see for it to be available to me in that way. Of course, typically I do believe and know many things about the objects I see. But I don’t need to know or believe anything about an object for it to be there and to be the object of my “objectual” awareness, my seeing.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

SEEING WHAT IS SO



The possibility of this kind of access makes “objectual” seeing very important, perhaps essential, to explaining the possibility of thought. It has been argued, for instance, that having de re attitudes towards some objects is a necessary condition of having any propositional attitudes at all.⁴ If that ‘priority’ of de re over de dicto attitudes in general carries over in particular to de re perceptual attitudes towards objects in one’s environment, “objectual” seeing or perceiving would be fundamental for thought. It makes available as object of a possible thought something that does not have to be thought about in any way to be available as the object of that thought. This is also especially important for language-learning.⁵ It is difficult to understand how language could be learned at all if we were not capable of perceiving things in a way that does not require belief or knowledge about the things seen. Objects must be perceptually available both to the conceptually unequipped pupil and to the sophisticated expert speaker if they are to serve as things to which the pupil can learn to apply new predicates that the expert can see the pupil understands correctly. If those concepts were needed even for seeing the objects in the first place, there would be no way to get started. A closely related thought lies behind John Campbell’s support for what he calls a “relational” view of perception of an object. If experience is the source of thought, as he argues, and perceptual experience of objects is what explains our ability to think about objects at all, there must be within our experience some “cognitive relation” to objects that is “more primitive” than having thoughts about them. “Conscious attention” alone, “by bringing the object itself into the subjective life of the thinker,” Campbell says, “makes it possible to think about that object”.⁶ So a “relational”, non-conceptual view of the perception of objects is essential to the explanatory potential of perceptual experience. It is a delicate question exactly how far the explanatory potential of this “relational” view of perception extends. Certainly perception of objects is needed to explain our ability to think about the objects we think about

⁴ See, e.g., Tyler Burge, “Belief De Re” and “Postscript to ‘Belief De Re’ ” in his Foundations of Mind, Oxford University Press 2007. ⁵ As Burge has also stressed in the same essay. ⁶ J. Campbell, Reference and Consciousness, Oxford University Press 2002, p. 6. (Page numbers in parentheses in the text refer to this volume.)

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



SEEING WHAT IS SO

when we see them. An object must stand in that relation to a perceiver to be the object of such a thought. And since the objects we see are typically objects in the world around us, our seeing them is part of the explanation of how we can think about the world around us. But that does not necessarily explain how we can think the things we actually think about the objects we see. Campbell sometimes speaks of the “relational” view of perception as explaining “what provides us with our concepts of the objects around us” (123) or “our grasp of the concepts of objects” (122) or “how it is possible for us to understand propositions about our surroundings” (123), or even what explains “how we have the conception of the world that we do” (122). I am doubtful that as much as all that can be explained by nothing more than our standing in a certain relation to an object we see. That is not to deny the fundamental importance of “relational” or “objectual” seeing for all thinking about objects. But we must also acknowledge that thinking something about the objects you see requires considerably more than just seeing them. Seeing objects puts us in a position to have thoughts about them, but a thought, a propositional thought, about an object you see comes only with the mastery of certain concepts or predicates. That conceptual capacity is a capacity to make judgements involving those concepts: an ability to apply the concept or predicate to items you take it to be true of. That is more than simply attaching a label or tag to the thing, or being able to attach that label to other similar things. It is also more than simply discriminating one kind of thing from another, or being able to sort things into different groups. It is a matter of being able to put something forward as true, and knowing how to do it under the appropriate conditions, and so understanding what you thereby think or say. And that capacity is not present simply in seeing objects in the “objectual” or “relational” sense. Knowledge requires propositional judgement, and judgement requires predication: the application of a concept to an object. “Objectual” seeing can make an object available for predication, and so for propositional judgement, without any prior thought or judgement about the object on the part of the perceiver. The kind of perceptual knowledge I want to draw attention to requires a capacity to recognize, in the right circumstances, that an item now within your awareness falls, or does not fall, under a concept you are master of and understand. And for certain concepts and certain situations, application of the appropriate concepts

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

SEEING WHAT IS SO



must be direct in the sense of not being made on the basis of any features of the present object other than its instantiating the concept in question. When an object is present, the perceiver’s recognitional conceptual capacities enable him to see immediately—without further assistance or guidance—that the concept applies, or does not apply, to the object. That is possible, of course, only for those perceivers who have the appropriate conceptual and recognitional capacities. Someone who has such concepts and can recognize that an object present to him is a chair, say, or is red, is in a position to make a judgement to that effect when an object shows up in his awareness and he pays attention to it. What he predicates of the object is something he must be capable of thinking of as true of other objects. Predicational capacity has a certain kind of generality; the same concept can be applied to different objects, and different predicates can be applied to the same object. The capacity is general in the sense of leaving a gap or an open place within each of the potential thoughts or sentences the thinker remains always poised to accept about chairs or red things or whatever it might be, whenever an appropriate object shows up in his perception. He does not have to know or think anything about that object before it shows up in order to come to have such a thought about it when he sees it. And when an object does show up in his awareness, he does not always need to deliberate or weigh competing factors to determine whether a certain concept applies to it. Perception of an object by anyone so equipped simply “brings facts into plain view”,⁷ in John McDowell’s phrase. A competent see-er and judger can see and thereby know by perception alone that a present object has such-and-such a property. It is tempting to think that there must be some concepts that we can sometimes see and thereby know by perception alone to be instantiated. If we could never know by perception alone that a certain concept applies to something we are aware of, it looks as if we could know it only if we know that a certain other concept applies to the thing, or a certain other concept applies to a certain other thing. And if that is true of every concept, it would go on forever. Even traditional theories with a more restrictive conception of perceptual knowledge did not reject the very ⁷ J. McDowell, “Conceptual Capacities in Perception” in his Having the World in View (Harvard University Press 2009), p. 139. (Page numbers with ‘M’ in parentheses in the text refer to this essay.)

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



SEEING WHAT IS SO

idea of knowing certain things by perception alone. They took that idea for granted but denied only that it applies to knowledge of anything that is so and would be so whether it is perceived to be so or not. Those theories say the most you can know by perception alone is that an object looks as if it has a certain property—is red, say—or that it seems to, or that it looks or seems as if there is a red object there, or that your perception is as of a red object, or some such less committal thing. If even those ways of hedging are still sticking one’s epistemic neck out too far, it is accepted that there is at least something we are directly aware of and know by perception alone.⁸ Only the scope or legitimate range of this kind of perceptual knowledge is in dispute, not whether there is any such knowledge at all. To know anything beyond that restricted range, those traditional theories say, we would need reason to believe that what we directly perceive is in some way or other a reliable guide to the way things are. But given the restriction, that is not something we can ever know by perception alone. This is the two-step conception of perceptual knowledge of the world that I think we must overcome if we are to understand how we know anything at all about the world by perception. One way to work towards overcoming that restriction would be to ask whether someone could have a capacity to recognize directly and without mediation that it looks or seems as if an object present to him is red if he did not understand what it is for an object to be red. And could someone who knows or understands what that is, and so could have a thought that a present object is red, lack the capacity ever to recognize directly, in the appropriate circumstances, that a present object is red? Could he lack the capacity not just on one or another particular occasion, but quite generally? It would seem that a person’s continued failure to recognize instantiations of the appropriate concepts when they are perceptually present to him would count against his possessing or understanding the concepts in question. This is not a line of argument I will pursue further here. But it is the kind of (dare I say ‘transcendental’?) investigation I think is needed to get to the bottom of, and to put behind us once and for all, the appeal of the

⁸ If what we are said to be directly aware of in perception are only objects, and never that such-and-such is so, that does not so far contribute to an explanation of perceptual knowledge of anything, since “objectual” seeing is consistent with knowing nothing about the object seen.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

SEEING WHAT IS SO



traditional restriction of perceptual knowledge to something always less than the world around us. This is where real work is needed. But I think there is no general cause for pessimism about the rewards to be found in that direction. Suppose we found after extended reflection that it is not possible in general for someone to recognize that an object actually present to him in perception looks red, or looks to be a red object, without his also having at least a capacity to recognize, and thereby to know, that an object present to him is red. That would mean that anyone already convinced that knowledge by perception alone is restricted only to how things look or seem to be, and who for that reason finds himself faced with the problem of knowledge of the external world, would in fact be capable of knowing things about that world by perception alone. He would be wrong to think he was faced by a general threat to his perceptual knowledge of the world. Despite his being convinced to the contrary, he could come to know that an object in the world is a chair, or is red, simply by seeing that it is. I think that is the position we are actually in, and it can have a liberating effect to realize it. But I do not think the fact that we can know things in that way amounts to a straightforward answer to the traditional epistemological question that the doubter thought he faced. It would not be an explanation of a person’s knowing that an object is red of the kind the traditional question seems to demand. That question appears to require that knowledge of the world is satisfactorily explained only by identifying a reason or ground or justification for the belief or knowledge-claim in question. It demands something on the basis of which the believer can be understood to justifiably accept what he does. And simply seeing and knowing that a red object is present does not satisfy that demand; it is the very thing that is meant to be explained. It is perhaps not surprising that the demand implicit in the traditional question is not satisfied, if perception is essential to knowledge of the world, but perceptual knowledge alone is thought never to reach as far as the world itself. When I see and in seeing know that an object present to me is red, it is not that what I see to be so is my “basis” or “reason” or “justification” for believing what I do. What I see to be so is that the object is red. It could be said that in those circumstances I “have every reason” to believe it. I certainly believe it “reasonably” or “justifiably”; I am (to put it mildly) “warranted” in believing it. There can be no better or stronger position

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



SEEING WHAT IS SO

for believing or knowing something than seeing that it is so right before your eyes. But what I see is not my reason or justification or warrant for believing what I do. There is no such thing in this case. What I see and thereby believe and know is that the object is red. I come to know that that is so by seeing that it is so. It is because I see what I see that I know what I know. My seeing what I see explains why I believe or how I know what I do. It could be said to be the reason why I believe and know what I do. But my seeing that the object is red is not my reason or justification for believing that it is red. It is what brings it about that I am justified or warranted in believing what I do. I have an eminently reasonable belief in the circumstances. But there is nothing that serves as my “basis” or “warrant” or “reason” for believing and knowing what I know. When I say that I see and thereby know that there is a chair in this room, or that a certain object is red, without relying on anything else I know or believe about the object, and so not on the basis of anything I know or believe about it, I do not mean that I could know what I do even if I knew nothing else at all. I could not see and know that the object is red without having the concept ‘red’, and so without knowing what I am saying when I say that something is red. I need a rich conceptual repertoire to be able to see and thereby know that p, whatever the ‘p’ in question might be. And I think we could not have the kind of conceptual repertoire we all have without knowing, or being capable of knowing, a great many things about the world around us. Learning to understand and think certain thoughts and learning things about the world those thoughts are about go hand in hand. So I must know or at least believe many things about the world even to be capable of seeing and thereby knowing that a certain object is red, or that there is a chair in this room. I have to know what chairs are, for instance, and how they behave, at least in general. This is another reason why my seeing and thereby knowing such a thing would not be a satisfactory answer to the traditional problem of the external world. That problem is understood as completely general, concerned with any knowledge of the world around us at all. So if I am in a position to see and know, of an object I see, that it is red only if I already know many things about the external world, that would not be a satisfactory answer to that completely general question. It would not fulfill the demands built into the very form of that question.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

SEEING WHAT IS SO



It is remarkable how those epistemological demands can seem to remain in force even for those who would claim to have no interest in that traditional question, or not to take it seriously. It remains difficult to overcome the feeling that knowledge of the world can be satisfactorily explained only in some such indirect, two-step way. It must be admitted that we are not infallible in coming to see and know such things as that there is a chair in this room. It is perfectly possible to think and say that I see and thereby know that there is a chair in this room and be wrong. When I am wrong I of course do not know; and since I don’t know, I do not see that there is a chair in this room. Either there is no chair, or if there is I do not see that there is. There remains a strong tendency to think that what I see when I don’t know must be the same as what I see when I do know by seeing that there is a chair in this room. So it seems that whatever I see in the case in which I do know cannot be that there is a chair in this room. This leads quickly down a very familiar and very slippery slope, at the end of which we are thought to be able to know only the very least committal thing we can be said to see and thereby know to be so. But there is no such thing, if the mere possibility of error is enough to disqualify it as not being seen to be so. This is what leads to the thought that we make errors only in judging something to be so, not in perceiving alone, or in seeing what we see. Judgement, or propositional or predicational thinking, is then thought to be no part of “the perceptual experience itself ”. And by now I think we are left with no possibility of explaining perceptual knowledge of the world. The demand for a step-wise explanation of knowledge, combined with this ‘non-judgemental’ conception of perception, leaves us with no satisfactory explanation of our perceptual knowledge of the world. When propositional thinking or the exercise of conceptual capacities are said not to be part of, or not essential to, “the perceptual experience itself ”, what is this thing called a perceptual experience? I think my seeing that there is a chair in this room is a perceptual experience. And that experience would not be what it is if I were not capable of propositional thought or the exercise of conceptual capacities in that very perception. Even if I see only that there looks to be a chair in this room, the perceptual experience I have also involves propositional thought or the exercise of conceptual capacities on my part. Which perceptual experiences that amount to knowing something do not involve any such capacities?

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



SEEING WHAT IS SO

Again, an indirect or two-step explanation of knowledge can seem required in accounting for the possibility of error. There is the feeling that I cannot know that there is a chair in this room by seeing that there is because I could have a perceptual experience of the very same kind when I see only that there looks to be a chair in this room. And I can have a perceptual experience of that kind even when I don’t know that there is a chair here. I think it cannot be denied that the two perceptual experiences, one when I know and one when I don’t know, are experiences of the same kind. But every two things are of the same kind in some respects. It does not mean that there is not also a great difference between a perceptual experience in which one sees and thereby knows that there is a chair here and an experience in which one sees and knows only that there looks to be a chair here but there isn’t. To say that, at least considered only as sources of knowledge, the two perceptual experiences are of the same kind, also seems wrong. They are very different in that respect. But to say they are of the same kind, considered only as “purely perceptual” experiences, leads right back to the familiar slippery slide to scepticism. As experiences understood in that way, they provide no knowledge of anything at all. A “perceptual experience itself ”, thought of as something with no conceptual ingredient, can perhaps be understood as simply a matter of being conscious and having your eyes open and seeing things in the “objectual” sense that does not imply that you know or believe anything about them. The objects seen in that way might be not only ordinary physical objects, but also colours, shapes, movement, and so on, which can also be seen in the “objectual” sense. It can be shown in your responses that a colour is what your visual attention is drawn to, or that you discriminate a certain shape from its surroundings, and so on. The things you are aware of in that way stand to you in the relation of being seen, but you know or think nothing about them. The objects you see are part of the world around you, but in that “perceptual experience itself ” you know nothing about the world around you. So whatever you might eventually come to know about the world could not be understood as something you know on the basis of what you know in perceptual experience alone. On this conception, there is nothing you know in having only the “perceptual experience itself ”. To have perceptual knowledge of anything at all, the perceptual experience in which you have it must involve some exercise of conceptual capacities involving propositional or predicational thought.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

SEEING WHAT IS SO



This requirement can seem to put perceptual knowledge of an independent world under threat from a completely different direction. John McDowell is one who thinks “conceptual capacities are operative in our perceptual experience” (M 141), even if he no longer holds that propositional thought is involved. But he thinks, as he characteristically puts it, that “there is justice in the thought that the idea can seem to work only in the context of an idealism” (M 141). I am puzzled by this. It would be disappointing to learn that the kind of perceptual knowledge I think we must have to avoid the scepticism inherent in the traditional problem of the external world is available only at the price of idealism. But maybe all is well. McDowell says that “any idealism with a chance of being credible . . . if thought through, . . . stands revealed as fully cohering with the realism of common sense” (M 141). So in his own thinking he works to “dislodge” from his idealism “the appearance that it does not genuinely acknowledge how reality is independent of our thinking” (M 142). Nonetheless, even when that independence is granted, he thinks his view is idealism “in an obvious sense”, “the label is a good fit” (M 143). What McDowell is impressed by is the special role or position we ‘rational animals’ occupy in the world. We have a “potential for selfdetermination”, for doing and thinking what we in our rational capacity can ourselves determine to do or to think. This is something not just any creature “whose senses inform it of things” can do (M 144). So our “rationality”, our capacity for assessing the reasons for and against something “enters into the possibility of describing ourselves as accepting what our senses give us”, as McDowell puts it (M 144). I agree that our having rational capacities enters into explaining how perception can provide us with something to accept. We have much richer and more varied conceptual capacities than other creatures. We can think, and think things to be true. And we can take one thing we think to be reason to believe something else we can think, and believe it for that reason. That is how we, and not other creatures, can come to believe the sorts of things we do. But in appealing to our distinctive capacities to explain our achievements I do not see any need, or role, for idealism. In the world we live in, perceivers with a capacity to think and reason and accept certain things as true can find that one thing they have come to think about the world is reason to believe another. That one thing is reason to believe another is something we find to be so in the world; it is

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



SEEING WHAT IS SO

part of what we take the world we believe in to be like. And in a world that is that way we can find by perception reasons to believe things about what that world is like. Other creatures cannot do that; they do not have our kinds of capacities for thought and the recognition of reasons. We are naturally equipped to find out things about the world in ways that are impossible for other creatures. Perhaps this is just the “realism of common sense” that McDowell says his idealism “fully coheres” with. But I do not see what is left of idealism. McDowell sums up his view in the slogan ‘The world is everything that is the case’. He says this expresses, truistically, what he calls an “unimpeachable way to use the notion of the world” (M 144). It is simply the idea of everything that is so—everything that can be truly thought or said. But he thinks this is idealism because “on this conception”, as he puts it, “the world itself is indeed structured by the form of judgment” (M 143). This is the point I do not get. I think McDowell is right to stress the importance of the idea that the world is everything that is the case. And I think he is right that it is ‘truistic’; there is no good reason to oppose it. I think the implications of that truistic thought have not really been seriously explored or faced. There has been an apparently undying predilection in philosophy for objects—as if what objects there are is enough to tell us what the world is like. But I take the point of the slogan to be that if there is a world at all, it is only because something or other is so. Even if objects are essential to any world, that the objects have certain properties, that they stand in certain relations to one another, that they behave in certain ways, and so on, are all things that must be so if there is a world like that. And that world—understood as all those things’ being so—cannot be simply a further object. So I think we need the idea that the world is everything that is the case in order to think of a world at all. But even when that idea is put together with a commitment to conceptual capacities as essential to perception of and thought about the world, I do not see how it amounts to any form of, or any encouragement to, idealism. That the world is everything that is the case does not mean that there are not a great many things that are so that we cannot even think, let alone have reason to believe, right now. By enriching our resources for thinking and recognizing reasons we could discover what some of those things are, and so come to think and know much more about the world than we do now. But what we found out in those new ways would still be something

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

SEEING WHAT IS SO



we could think, then. Our being able to think it, and our having found things we recognize to be reasons to believe it, would explain our believing what we do. None of that could be explained without appeal to our conceptual capacities. I think even the scope of propositional perception could be expected to expand in that way, and for the same reason. What you can see depends on what you can think and what you know about the world. We adults can see many things to be so that small children cannot see to be so. If you see my neighbour’s car and thereby see that there is a car parked in my street, the chances are that you do not thereby see that my neighbour is at home, as I do. I am in a better position for seeing that than you are, even if we are standing next to each other. And a physicist who sees a marker moving on a computer screen can see much more to be so than I can. Given his knowledge and his capacities, his seeing the marker is enough for him simply to see all those things to be so. Once it is granted that the range of perceptual knowledge can expand in this way to include anything that any adequately equipped perceiver can come to think and recognize to be true of the objects perceived, there seems no hope of specifying in advance any determinate general limits to what can be seen and known about the world by perception. But the fact that a rich body of conceptual capacities is needed to make such perceptual knowledge possible does not seem to me to lend support to any form of idealism.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

8 Perceptual Knowledge and the Primacy of Judgement Perception is our primary way of finding out about the world around us. We live as we do only because we can perceive and thereby come to know what is so in the world we live in, and act accordingly. Any adequate account of human perception should therefore explain, at least in general, how our perceiving what we do can give us the kind of knowledge we all know we have of the world around us. But since at least the time of Descartes, if not earlier, knowledge of the world by means of the senses has seemed to present a formidable philosophical problem. I don’t think there is any distinctive feature or requirement of knowledge itself that can make knowledge of the world in general seem unattainable, or even difficult. What I think has seemed to lead to that conclusion are certain apparently very natural ways of thinking about sense-perception. If there is a difficulty about perceptual knowledge of the world—and in philosophy there certainly has seemed to be—I think that is where the difficulty lies. Perception has been thought of in philosophy as a matter of having perceptual experiences. I don’t think there is anything problematic in that in itself either. Perceiving is a matter of having perceptual experiences, just as walking is a matter of having walking experiences or climbing mountains is having mountain-climbing experiences. But in philosophy, perceptual experiences have been understood in a certain distinctive way. Locke, Berkeley, and Hume, for instance, all spoke of perception as a matter of having something “present to the mind”: what they called “ideas”, or for Hume “impressions”. To have one of those things “present to the mind” was to be sensorily aware of what the idea or impression is an idea or impression of. And we were said to have ideas or

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

PERCEPTUAL KNOWLEDGE AND THE PRIMACY OF JUDGEMENT



impressions of perceptible qualities or properties like redness or roundness or warmth or smoothness. Those philosophers also thought we have ideas or impressions of an object such as a horse or a cherry or a billiard ball, but on this view of what is perceived there is a certain difficulty in that. Berkeley, for instance, spoke of an idea of a cherry as a “congeries of sensible impressions” all present together. But surely it is possible to be aware of several different properties all together without having an idea of their being properties of one and the same object. I think those philosophers did not really account for the possibility of perceptions like that, for reasons I will touch on later. In any case, this view says that perception is awareness of something— awareness of an object in the sense in which even qualities or properties are ‘objects’ or items rather than states of affairs. To put it linguistically, specification of what this view says a perceiver perceives takes the grammatical form of a psychological verb of perception followed by a noun or name or singular term: ‘a sees x’, ‘a hears x’ ‘a feels or is aware of x’, where what goes in for the ‘x’ is a noun or a name of some object or quality. This means that to perceive or be aware of something that is in fact an apple, or is in fact a property of an apple, is not necessarily to be aware that what you perceive is an apple or is a quality of an apple, or even is a quality of anything. It is simply to be aware of something that is an ‘object’ in the grammatical sense in which the accusative of the perceptual verb is a name or a term rather than a sentence. Any coextensive term could be substituted for that term without changing the truth-value of the sentence specifying the object of perception. If we take seriously this idea of what we strictly speaking perceive, I think we cannot explain how perceiving what we do gives us any knowledge of what is so. If we are to know things by perception we must be able think of and understand what we perceive in some way. We must be able to make sense of our perceiving what we do. For that we need some terms in which to express and think about what we perceive. Many philosophers have thought we could get the general terms we need by “abstracting” properties or general kinds of things from what is present in the individual perceptions we receive. But even if that process of “abstraction” as those philosophers understood it were successful (a big ‘if ’), it would leave us at best only with a collection of names or labels or general terms of certain qualities or properties or kinds of

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



PERCEPTUAL KNOWLEDGE AND THE PRIMACY OF JUDGEMENT

things, not an account of how those terms can be combined to say or think anything that is true or false. A series of names of things or properties does not make a sentence, and so does not say anything, even if they are names of things we are aware of. A list or collection of objects is not a thought that something or other is so. What is needed in explaining knowledge is some account of how propositional thought is possible. Awareness of things, even repeated awareness of things of the same kind, does not on its own give you the resources for having beliefs about them that are either true or false. I don’t think Locke or Berkeley or Hume ever succeeded in explaining how belief or judgement is possible. Hume saw the importance of the problem; he even thought he was the first person to notice it. But he could not really account for the distinctive character of belief or judgement with nothing more to work with than what he thought of as ideas. And I think he saw that he had not really explained it. Many twentieth-century philosophers of perception shared with their eighteenth-century predecessors the assumption that perception is awareness of or acquaintance with objects. They described what we are given in perception as “sense-data”: purely “sensory objects” that we are directly aware of. There was a real question what those purely “sensory objects” were supposed to be. Are they simply the very qualities we are aware of in perception: redness, roundness, warmth, and so on? Or are they objects that themselves have such qualities, so what we are aware of are sense-data that are red, or round, or warm, and so on? If so, do those objects also have qualities that we are not aware of, or is their esse percipi? I think sense-datum philosophers did not really explain the nature of those objects, but they nonetheless simply helped themselves to propositional thoughts about them. Even then, I think, they did not succeed in explaining how acquaintance with, or even propositional knowledge about, such purely “sensory objects” could give anyone any reason to believe the kinds of things we all know about the public world around us. Philosophers have continued to look carefully into what they think we are most directly aware of in perception. But in recent years not all philosophers interested in perception have been especially interested in knowledge. The nature or character of what they call “the very perceptual experience itself ” is what those philosophers tend to concentrate on. But most philosophers who are interested in perceptual knowledge still

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

PERCEPTUAL KNOWLEDGE AND THE PRIMACY OF JUDGEMENT



appear to start with what we can be sure perception alone gives us, and then try to account for the possibility of knowledge of the world on that basis. That is to proceed, so to speak, from the bottom up. But how is it to be determined what exactly that “bottom” actually is: what senseperception all on its own gives us? And can we be sure that if we settled that question first, we would then be able to reassure ourselves about how, on that basis alone, we can know the sorts of things we know about the world? At this point in the historical and philosophical development of this subject I would propose that we take a different direction and proceed not from the bottom up, as it were, but from the top down. I suggest that we start with the conditions that must be fulfilled if we have any knowledge of the world and ask how what we perceive must be understood if it is to fulfill the conditions of our knowing what we know about the world by perception. To know something is to know that such-andsuch is so. That involves a propositional thought that has a truth-value. So we must start in particular with the conditions of propositional thought. What does it take—what kinds of capacities or competences are needed—for thinkers to form and grasp thoughts that they understand to express something true or false? We could not have knowledge of anything at all without that. With some understanding of what is involved in that fundamental capacity, we could then go on to ask how perceiving what we do must be understood if what we perceive is to be accessible to the propositional capacities we exercise in believing and perhaps knowing the things we do. This was, in effect, the overall strategy of Kant, as against Descartes, Locke, Berkeley, Hume, and many others. The idea is to start, not with what seems to come to us in experience and ask how we know things about the world on that basis, but to start with the conditions of our even thinking about a world that perception could perhaps reveal to us. This is what puts judgement at the centre of the picture. To think at all—and so to be capable of knowing anything—is to be capable of asserting or endorsing or putting forward something as true. So to understand how perception can provide us with knowledge we must understand the connection between what we perceive and the possibility of thought and belief about what we perceive. How can what we perceive bear on the truth or acceptability of something we believe to be so? This is to acknowledge certain requirements on what we can be understood to

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



PERCEPTUAL KNOWLEDGE AND THE PRIMACY OF JUDGEMENT

perceive—what might be called the “proper objects of perception”—for perceptual knowledge of the world to be possible. Not just anything we perceive would be enough to explain how our perceiving what we do enables us to know the kinds of things we do about the world. If a capacity for judgement is a condition of having any thoughts at all, thinkers must be capable of thoughts with a certain distinctive structure. A thought cannot be simply a list or collection of separate items or objects, even things we are aware of. It must have a predicational structure in which part of the thought expresses something that is thought to be true of some object or item picked out by another part of that same thought. Two different ingredients or aspects of the thought must be put together in a way that yields something that is either true or false. We apply a concept to something we think of and thereby think of that thing as falling under or being characterized by that concept. That is what it is to have concepts. As Kant put it: “concepts are predicates of possible judgments”. Only someone with a capacity for judgement can think of things as being one way rather than another, and so can be said to possess concepts. Being capable of judgement is a condition of believing that something or other is so. It is therefore also a condition of knowing that something is so, and so of knowing something about the world. So being capable of judgement is a condition of knowing something about the world by perception. To understand perceptual knowledge we must understand the conditions of judgement, or belief. Belief is a much richer and more complex psychological state than many philosophers appear to have supposed. I think it is the key to understanding the possibility of knowledge, and so understanding the possibility of perceptual knowledge. Here I agree with Fred Dretske’s observation that “believing something requires precisely the skills involved in knowing”; “anyone who believes something thereby exhibits the cognitive resources for knowing”.¹ A thinker cannot believe something he does not understand. So a thinker who believes or knows something understands the thoughts he believes to be true, and he understands them to have a distinctively predicational structure. To have mastered that kind of structure one ¹ Fred Dretske, “The Epistemology of Belief ” in his Perception, Knowledge and Belief, Cambridge University Press 2000, pp. 64–79.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

PERCEPTUAL KNOWLEDGE AND THE PRIMACY OF JUDGEMENT



must understand many different sentences or thoughts with predicational structure. One must have a general competence in the application of different predicates to the same object and the same predicate to different objects. It is not possible for a thinker to be capable of only one or two, or half a dozen, thoughts. One must be master of many different sentences or thoughts to have the competence involved in understanding any one predicational sentence or thought. But to understand one’s thoughts it is not enough simply to understand their structure. One must also understand the terms used in those sentences or thoughts: the predicates and names and other ways of referring to the objects thought about. One must understand what those particular thoughts actually say: the conditions under which they would be true, or false, and the conditions under which they would be appropriately asserted, or denied. How the thinker understands those conditions is what determines what the thinker believes. Thinking or believing something is an intentional state: what the person believes is only what would be so as the person himself thinks of it or acknowledges it. Someone who believes that the man he sees is wearing a red shirt does not necessarily believe that the tallest man in town is wearing a red shirt even if the man he sees is the tallest man in town. A competent thinker who understands what he believes will be able to judge, in circumstances he finds to be appropriate, that certain thoughts he understands are true, or false, and so to judge accordingly. What Kant saw as essential to thought is not just a capacity for judgement, but actually judging in the appropriate circumstances that such-and-such is so-and-so. That involves positive endorsement or acceptance or putting forward as true something one understands. So correct or appropriate judgement is essential for possession of the competence or capacity. A putative thinker’s consistently making inappropriate or inexplicable so-called “judgements” would cast doubt either on his competence in judgement in general or on his understanding of the particular thoughts he so inappropriately appears to accept. There is even more than this involved in judgement or belief as it bears on the explanation of knowledge, and so of perceptual knowledge. It is crucial to belief that beliefs are held for reasons. We can come to believe one thing because we believe something else. If someone accepts the belief that p, his believing that q can sometimes be explained by his believing that p. To see his holding the second belief as explained by his

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



PERCEPTUAL KNOWLEDGE AND THE PRIMACY OF JUDGEMENT

holding the first is to see the thinker as taking what he believes in the first belief as reason to accept the belief that q. His reasons for endorsing the first belief are extended to his acceptance of the second belief. I think that to understand ourselves as believers we must have a correct understanding of this capacity to recognize and be responsive to reasons. One way not to understand that responsiveness is brilliantly illustrated by Lewis Carroll’s story of Achilles and the Tortoise.² The Tortoise says he accepts two statements, let’s call them (A) and (B), which we can see quite obviously imply a third statement, call it (Z). But the Tortoise says he does not yet accept that third statement, and he challenges Achilles to “force” him to accept it. Achilles says, “But you must accept (Z) if you accept (A) and (B), since if (A) and (B) are true then (Z) is true”. “I am willing to accept that conditional statement ‘If (A) and (B) then (Z)’, and add it to those statements I already accept”, replies the Tortoise, “but now, again, please force me to accept (Z).” Achilles points out that if everything you have now accepted is true, then (Z) is true. And the Tortoise replies, predictably, “I am willing to accept that longer conditional statement, and add it to what I have already accepted. And I ask again, please force to me to accept (Z).” And so on. And on. What I think this brilliant story illustrates is that the Tortoise has no sense of, or shows no recognition of, one thing he believes being reason for him to believe something else. He appears to think that the only way believing one thing can bear on a person’s believing something else is by the believer’s accepting some third proposition that states a relation between the two. But if that were the only way a believer could come to believe something on the basis of something else, another proposition would always have to be added to what the person believed at every point. That is the regress that goes on forever. I think we can learn something about belief from Lewis Carroll’s story. The Tortoise says he accepts (A) and (B). But if he accepts them he must understand them, and so he must understand their structure. So he must see that if (A) and (B) are true then (Z) must be true. And how could he understand and accept that conditional proposition without seeing that with (A) and (B) alone he already has reason to accept (Z)? Can we really make sense of the Tortoise as a believer at all, despite his apparent

² Lewis Carroll, “What the Tortoise Said to Achilles”, Mind 4, 1895, 278–80.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

PERCEPTUAL KNOWLEDGE AND THE PRIMACY OF JUDGEMENT



willingness, as he puts it, to “accept” certain things? I think the story shows that believing something for a reason cannot be understood simply as believing something as well as accepting some other proposition that states a relation between the proposition said to be the reason and the proposition said to be believed on the basis of that reason. Someone who believes that p and comes to believe that q for that reason takes p as a reason to believe that q. If that is what explains her coming to believe that q, her ‘taking something as a reason’ must be something more than her simply adding a certain proposition to all the propositions that she already accepts. In saying it must be ‘something more’ I mean that taking something as a reason to believe something must be a matter of one’s being ready or inclined or disposed to accept the second proposition in the light of one’s acceptance of the first. That readiness is expressed or exhibited in one’s accepting the second belief on the basis of the first. So regarding or taking one thing as reason to believe something else is an “active” or “productive” attitude. One is responsive to what one takes to be a reason, and one accepts the second belief in accord with that responsiveness. Simply “accepting” a proposition in the way the Tortoise says he “accepts” something is not enough, even if what he “accepts” is a truth about a relation between what is said by one proposition he “accepts” and what is said by another. Someone who sees that a billiard ball is moving rapidly towards another, stationary, ball can come to believe that the second ball will move when hit. Her taking what she sees to be true of the first ball as reason to believe what she does about the second ball explains her coming to have that second belief. That happens because she takes something she already accepts as reason to believe that the second ball will move. That is what Lewis Carroll’s Tortoise apparently never does. He always demands that one more proposition must be added to those he has accepted so far. We could perhaps say that, as a believer, the Tortoise appears to be ‘reason deaf ’, or ‘reason insensitive’, and so difficult to understand as a believer at all. The person who sees that the first ball is moving and for that reason believes that the second ball will move could be said to believe that what she sees to be so is reason to believe that the second ball will move. But in ascribing to her that belief about one thing’s being a reason to believe something else we ascribe to her a certain attitude—a responsiveness or readiness to respond in a certain way to something she believes. She sees

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



PERCEPTUAL KNOWLEDGE AND THE PRIMACY OF JUDGEMENT

that the first ball is moving, and in coming to believe on that basis that the second ball will move she exhibits or expresses an ‘active’ attitude of responsiveness that is present in her very acceptance of the fact that the first ball is moving as it is. I think arriving or being prepared to arrive at some beliefs on the basis of other beliefs we regard as reasons for them is essential to being a believer at all. I take this to show that a person’s believing something cannot be fully captured simply by listing the propositions the person believes. To believe something is sometimes described (in philosophy) as standing in what is called “the believing relation” to a set of propositions. But that could be at best only a purely formal characterization. It could perhaps identify what the person believes, but it would not fully characterize the psychological state the believer of those propositions is in. A list of propositions alone would be insufficient even if it contains propositions about the relations among the various propositions the believer already believes. Even if some of those propositions are statements to the effect that one thing on the list is reason to believe another, the list would still not fully capture the attitude of a believer who believes something for what the believer takes to be reason to believe it. I think what I am calling the attitude of responsiveness to reasons that is involved in a believer’s believing that q for the reason that p is best regarded as an evaluative judgement, and that such judgements are not equivalent or reducible to any non-evaluative, and in that sense “purely factual”, propositions about what states of affairs actually hold in the independent world.³ Regarding one thing as reason to believe another is not simply acknowledgement of a relation holding between what is said to be so by the propositions in question. That was the trouble with Lewis Carroll’s Tortoise. Believing something for a reason involves the believer’s recognition of something as a reason to believe it, and perhaps some assessment of the strength of that reason. It is an evaluative or “active” attitude in the sense that it is because of its presence that the believer is “moved” or “led” to believe something that is in accord with his evaluation of the reason. He comes to be in an intentional state of believing a certain thing because of his evaluation or assessment of the reasons other things he believes give him to believe it. ³ This view is clearly explained and convincingly defended in T. Scanlon, Being Realistic About Reasons, Oxford University Press 2014.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

PERCEPTUAL KNOWLEDGE AND THE PRIMACY OF JUDGEMENT



We take most things we believe to be supported by other things we believe in this way. If we try to trace the chains of beliefs and reasons backwards to their epistemic origins, we find that the chains do not go back forever. Sometimes the chain (pursuing it backwards) stops at something we perceive to be so. Sometimes the chain stops with something we recognize to be so not by sense-perception but in some other way. To say that the backward chain of supporting beliefs and reasons stops at a certain point is to say that at that point we simply recognize something or other to be so, and not on the basis of anything else we recognize to be so. This is what happens in sense-perception. So we come, at last, to the question of perceptual knowledge. I think we can know things about the world around us by directly perceiving that such-and-such is so. In saying we perceive and know such things directly I mean we perceive and know them not on the basis of perceiving or knowing anything else to be so. This kind of perceptual knowledge, for all its directness or immediacy, is of course something we can achieve only if all the necessary conditions of our perceiving and knowing such things are fulfilled. The elaborate set of necessary conditions we must fulfill to know things in this way is a matter of considerable complexity. I have been describing here only its general outlines. So knowing things in this way is no simple matter. To know something about the independent world we must be capable of thoughts that are understood to be true or false whether anyone judges them to be so or not, or even entertains them. They are in that sense thoughts of something objective. Take the simple thought of there being a red apple on a brown table. That is something that could be true, or false, whether anyone thinks it is so or not. And what we think in entertaining that thought is something I think we can sometimes see to be so right before our eyes. We see a red apple and we see that it is on a brown table that we see right here and now. And we do not see that that is so on the basis of anything else we see to be so. Of course, there simply being a red apple on a brown table right before your eyes is not enough for you to see or to know that there is, even if your eyes are open and the light is good. Even if you actually see a red apple and you see a brown table, and the apple you see is on the table you see, that is still not enough. To see and to know that there is a red apple on a brown table requires more than seeing those objects. It requires

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



PERCEPTUAL KNOWLEDGE AND THE PRIMACY OF JUDGEMENT

competence in and competent exercise of the perceptual and conceptual capacities required for propositional thought about what you perceive. But those who possess that rich set of capacities, as we all do, and exercise them competently in the right circumstances, can sometimes find themselves perceptually aware of a fact of the objective world that they thereby know to be so. This is what I call “seeing what is so”, and so knowing what is so by perception alone. Among the perceptual conditions of this kind of perceptual knowledge, I believe, is a capacity we all have to perceive particular objects, to be aware of an object and single it out perceptually, to discriminate it from its background or surroundings, to have our attention drawn to it, perhaps to track it if it is moving. It is not altogether easy to say exactly what is involved in seeing an object in this way, but it is a capacity we share with other animals, many of whom are probably better at it than we are. It is a matter of seeing an object in what could be called the purely “objectual” sense I mentioned earlier. What is seen in this way is specified by reference to an object rather than a state of affairs, with a noun or name or term as the accusative of the perceptual verb, not a sentence or sentential clause. Seeing an object in this way is what Fred Dretske called “nonepistemic” or “simple” seeing.⁴ It is a de re relation we can stand in to objects that show up in our perceptual awareness. To say it is de re is to say that we do not have to think something about an object we see in order to stand in that relation to it, and if we do think of the object in some way, what we think of it does not have to be true of it for us to see the object. In the statement that a person sees x, any expression that in fact refers to the object that is seen can be put in for the ‘x’ without changing the truth-value of the statement about what the person sees. As a matter of fact, of course, most of us do believe something or other about most of the objects we see or pay attention to, but we do not have to believe such things about them in order to see the objects in this sense. I think our standing in this relation to objects in the world is important, perhaps essential, to the possibility of propositional attitudes about them, and so to the possibility of objective thought. It is a way in which our words or thoughts are connected with things in the world that we

⁴ Fred Dretske, Seeing and Knowing, Routledge 1969, ch. 2.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

PERCEPTUAL KNOWLEDGE AND THE PRIMACY OF JUDGEMENT



perceive and think about.⁵ The possibility of this kind of perception is obviously crucial for language-learning: the beginner does not need any words in order to single out an object and attend to it in the course of eventually learning some words to apply to it or to things of that kind. Even those of us who already have a great many words and are thereby equipped with the resources for making sense of our experience are in a similar position with respect to seeing objects in this “direct” or de re sense. We do not need those words or all those resources in order simply to see the objects we see. But within those linguistic resources we possess a highly developed repertoire of predicates that we understand and are ready to find instantiated when the right circumstances present themselves. Our rich predicational competence could be thought of as our carrying around with us an enormous body of open sentences that we understand. When an object comes into our experience, when we “simply” see it, we can often see that certain open sentences we understand are true of that object. In seeing what we see, we see that certain concepts we are masters of are true of the object we see. We see that the object we see is a red apple, for instance, and perhaps that it sits on a brown table. What we see to be so in such a case is something we can also think or believe to be so when our eyes are closed or when we leave the room. What we see and thereby know to be so is something that is so whether anyone sees or otherwise perceives it to be so or not. We see and thereby know that such-and-such is so in the world, and we know it not on the basis of seeing or knowing something else. When I say it is not on the basis of knowing anything else I do not of course mean that someone could see and know in that way that there is a red apple on a brown table even if he did not know anything else about the world at all. I have been trying to explain why no one could know any such thing without having the elaborate conceptual and perceptual capacities I have been sketching. And anyone who possesses those capacities will come to know a great deal about the world in the competent exercise of them. So knowing anything at all about the world

⁵ This is brought out clearly in John Campbell’s purely “relational” view of the perception of an object. He explains how our attending to things we perceive in that sense, “by bringing the object itself into the subjective life of the thinker, makes it possible to think about that object”. John Campbell, Reference and Consciousness, Oxford University Press 2006, p. 6.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



PERCEPTUAL KNOWLEDGE AND THE PRIMACY OF JUDGEMENT

involves knowing quite a lot about the world. But that does not mean that in seeing and knowing that there is a red apple on a brown table one knows it on the basis of something else one knows: in some indirect way, or by inference or transition from something else one perceives or knows to be so. When we trace a chain of supporting reasons for beliefs back to something we perceive to be so in this way I think the chain stops with something that we perceive and thereby know to be so. What we know can then serve as reason to believe other things we come to believe. But when we see and thereby know that there is a red apple on a brown table I do not think that what we then perceive to be so is our reason to believe that there is a red apple on a brown table. What we perceive to be so is not a reason on the basis of which we make a reliable inference to the conclusion that there is a red apple on a brown table. What we perceive in those circumstances is the very fact that we thereby come to know: the very state of affairs we entertain in the thought of there being a red apple on a brown table. In that situation, there is no need for something to serve as our reason for believing that there is a red apple on a brown table. There is not even any room for such reason. That there is a red apple on a brown table is something that we see and know to be so right before our eyes. That does not mean that we cannot say that we believe that there is a red apple on a brown table. I can say, of any of the things I know, that I believe them. But to say in a case like this only that you believe that there is a red apple on a brown table, or that you have reason to believe it, even good reason, is too weak. You know it; you see and know how things are in the world. No circumstances could be more favourable for discovering how things are than seeing or otherwise perceiving that such-and-such (which you fully understand) is so right before your eyes. That is the point at which tracing back the chain of reasons to believe something comes to an end: in perceiving, and in that way knowing by perception alone, that such-and-such is so. For these reasons I think proceeding in this “top-down” direction, from the conditions necessary for competent thinkers and perceivers to perceive that certain objective thoughts are true, offers us greater hopes of finally understanding how perceptual knowledge of the world is possible than we can expect from the “bottom up” projects we are so familiar with but which continue to leave us dissatisfied.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

PERCEPTUAL KNOWLEDGE AND THE PRIMACY OF JUDGEMENT



I do not much care for labels in philosophy. Those who like labels will probably call the possibility I have been trying to describe here “Realism”. It says we can perceive and know things about the world directly, not on the basis of anything else, and that what we can know in that way holds independently of its being perceived or believed by anyone. Maybe that does make it Realism, twice over. But because of the elaborate set of conditions I have tried to show we must fulfill even to be capable of that kind of knowledge, it certainly cannot be called “Naïve Realism” (whatever that might be). The view I have been describing is that with properly functioning perceptual mechanisms, human beings who fulfill all the conditions of competent thought are capable of knowing by perception alone how things are in the independent world around them. That is what Kant called “the objective reality of outer intuition”. He took himself to have proved that doctrine, and so to have established something he too called “Realism”. But for him it was only “Empirical Realism”. And Kant held that what he called “Empirical Realism” could be established through an investigation of the necessary conditions of human thought and experience only if something called “Transcendental Idealism”—not a form of “Realism”—is true. Whether that applies as well to the conclusions I have tried to draw from the conditions of thought and experience is a question I leave for another occasion. It is not simply a matter of labelling.⁶

⁶ I am grateful to Christoph Pfisterer for valuable criticisms of several earlier versions of this paper.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

9 Doing Something Intentionally and Knowing that You are Doing it We all know how to do many different sorts of things, and we do lots of them, more or less expertly, every day. I think our having the talents and abilities we exercise in those ways involves our knowing certain things. I, along with almost everyone else, know how to tie my own shoelaces. And I usually know that I am doing it when I do it. When I have done it, my laces are tied because I tied them, and I usually know they are tied because I know that I tied them. I am also pretty good at putting the milk into the refrigerator. And when I want the milk I usually know it is in the refrigerator because I know I put it there. I think the knowledge we have in exercising these talents is not only knowing how to do something and knowing that we are doing it, but also knowing that such-and-such is so in the world. We know it because we bring certain things about and we know that we do. There is a traditional way of thinking about human knowledge in philosophy that seems to stand in the way of understanding how agents know these things about what they are doing. It has made it difficult even to recognize, let alone account for, what appears to be a distinct and fundamental kind of human knowledge. Here I take up the question of how this kind of knowledge is to be explained. I will concentrate on an agent’s knowing that he or she is doing suchand-such. When I am tying my shoelaces or putting the milk in the refrigerator I know that I am doing that. My question is: “What do I then know, and what kind of knowledge do I then have?”. Philosophy asks: “Is an agent’s knowledge that he is doing such-and-such a priori or empirical?”. It is implied that it must be one or the other. I think it is neither.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

DOING SOMETHING INTENTIONALLY . . .



For one thing, that I am tying my laces or putting the milk in the refrigerator are contingent facts. Things could have been otherwise. So if I know these things a priori it cannot be that I know them to hold necessarily. If you are there watching me, and you know what I am doing, you presumably do not know it a priori. But what you know about what I am doing is the same thing I know about what I am doing. You know it by observation, as we say. Does that mean that I know it not by observation? This question is more complicated than it looks, as we will see. But in any case it does not seem right to say that I know a priori that I am doing it. That would mean that I know it completely independently of all experience. Do I then know “by experience” that I am tying my shoelaces when I am? Well, there was a time when I could not tie my shoelaces, and the more experience I got in doing it, the better I became. So my present skill in tying my shoelaces, without which I could not know I am exercising that skill, does come from experience. But this is not what “by experience” usually means in philosophy. The word ‘experience’ is not typically used there as a general or mass term of undivided reference for something each of us accumulates more and more of as we get older. ‘Experience’ is typically used in philosophy in the singular, to refer to a particular experience, or an experience of this or that kind. If I know “by experience”, understood in this way, that I am tying my shoelaces, the idea is that there is some particular experience that I am having, and that it is somehow by means of having that experience that I know that I am tying my shoelaces when I am doing it. This raises two questions. First, what is that distinctive experience I have when I am tying my shoelaces? And second, how does that experience, whatever it is, yield my knowledge that I am tying my laces? These are different questions, but obviously they must be answered together. The experience to be identified must be such as to yield my knowledge that I am doing such-and-such. My having that experience must be, or at least must figure in, an explanation of how I know that I am doing what I am doing. This “philosophical” notion of ‘an experience’ is not simply the notion of an experience, or an experience of a certain particular kind, that we employ in everyday life. I have had many particular experiences of tying my own shoelaces – one almost every day for many years. But to say that I had those experiences, or experiences like that, is just to say that I have tied my own shoelaces many times while being aware that I was doing it.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



DOING SOMETHING INTENTIONALLY . . .

To see that this is no help on the question of knowledge think of a kind of experience many people have never had: the experience of walking across the Golden Gate Bridge in San Francisco. To say that is an experience you have never had is just to say that you have never walked across the Golden Gate Bridge. To have that experience, or an experience of that very kind, you have to be there and do it; you have to be walking across the Golden Gate Bridge. There is no difference between doing it and having the experience. Here we have identified an experience, or kind of experience—walking across the Golden Gate Bridge—but surely it is not by having that experience that you know that you are walking across the Golden Gate Bridge when you are doing it. It could perhaps be said that it is because you are having that experience that you know that you are walking across that bridge. That would mean only that it is because you are walking across the bridge that you know you are; you couldn’t know you are doing it if you weren’t doing it. But that doesn’t explain how you know that you are doing it. It doesn’t show that you know it by having that experience. This is not to deny that you have many different experiences while you are doing something. Seeing, hearing, and feeling things is part of any conscious person’s intentionally doing anything. But that does not mean that the agent’s knowledge that he is intentionally doing such-andsuch is based on such experiences even if they are all part of what we might call “the full experience” of, e.g., walking across the Golden Gate Bridge. What is thought to be the philosophically or epistemologically more promising idea of an experience is something that can contribute to your knowing something. Having an experience in that sense can inform you of something, or make plain or indicate to you something that could support a further opinion or belief about something or other. What you know in that way, you know on the basis of that experience. For an experience to play that role the experience must be thought of as something you can, at least theoretically, regard as separate or separable from what you claim to know on the basis of it. It is as if, at least in thought, you can somehow peel off the experience you are having from the rest of the world and attend only to the distinctive features of that experience on its own. If you had an experience like that every time you intentionally do something, that might seem to explain how you know by experience that you are doing what you are doing. The experience would provide the basis or ground of your knowledge.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

DOING SOMETHING INTENTIONALLY . . .



We saw the more-or-less everyday idea of what we might call the “thick” experience of walking across the Golden Gate Bridge is too “close” to actually walking across the bridge, so to speak, to explain how you know you are doing it. Having that experience is the same thing as doing it. On the other hand, this more philosophical idea of a perceptual experience seems too “thin”, or too detachable from the action itself, to explain how you know that you are doing it, even if you always have such an experience every time you are intentionally doing something. The difficulty should be familiar from centuries of trying to account for perceptual knowledge of the world on the basis of such “immediate data” of sense experience alone. On this conception, if I have what is called an “immediate visual experience” of a red tomato on a white plate before me, for instance, I could have had that experience, or an experience exactly like it, even if there were no tomato and no plate there. The whole point of this idea of experience is to insist that having such an experience does not imply the presence or existence of anything that the experience is ordinarily said to be an experience of. But if that is so, it seems obvious that I do not then know on the basis of that “visual experience” alone that there is a tomato on a plate there. That is the familiar challenge to the possibility of perceptual knowledge of the world that this conception of perceptual experience appears to leave us with. The same difficulty would arise for knowledge of one’s actions. Even if we have an experiences of this “thin”, non-committal kind whenever we intentionally do something, the experience alone would not explain how we know that we are doing what we are doing. Such experiences would not suffice precisely because it is possible to have them without actually performing the action, and so without knowing that you are. It seems equally obvious that no additional material or body of information available to us from experience could be combined with what we get from “immediate” experience alone to help explain how we know what we are doing. On this conception of strictly “non-committal” experience, we could never find anything in experience that would take us beyond it. I think this familiar difficulty is enough to show that no appeal to “experiences” or “immediate experiences”, understood in this special way, can successfully explain how we know that we are doing suchand-such when we are intentionally doing it. This means we need a better way of understanding, in general, how human beings know what

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



DOING SOMETHING INTENTIONALLY . . .

they know about their own intentional actions, just as I think we need a better account of how perceptual knowledge of the independent world is possible. One line of thought I find promising with knowledge of the world is an understanding of perception according to which it is possible for human beings to know what is so by perception alone without need for any further help.¹ I think a person can sometimes see, for example, that there is a red tomato on a white plate right before him, and in that way he knows it is so. It is not that he knows it on the basis of some experience he is having. No experience serves as his ground or reason for claiming or judging as he does. I want to say that in the case imagined there is nothing that serves as his ground or reason. He simply sees and thereby knows that there is a tomato on a plate there. We could say, perhaps, that the person has the experience of seeing and knowing that there is a tomato and a plate there. But that experience is not the ground or basis of his knowledge that there is a tomato there any more than having what we call the experience of walking across the Golden Gate Bridge is the ground or basis of an agent’s knowledge that he is intentionally walking across that bridge. In seeing that there is a red tomato on a white plate a person knows that that is so; there is no step from a ground or basis to his knowing. This does not mean that there can be no explanation of his knowing what he knows. It is not a mystery or a miracle. Since there are many different ways of coming to know something—even something as banal as there being a red tomato on a white plate—how this person comes to know it in this case can be explained accordingly. We could say “He knows it because he sees that it is so” or “He knows it by seeing that it is so”, or “He knows it by seeing”, for instance, not by being told that it is so or by reading it in the newspaper. But those explanations that appeal to seeing do not cite some “source” or “ground” that serves as the “basis” of the person’s perceptual knowledge of what is so. Of course, seeing that such-and-such is so is a more complex and more impressive achievement than simply seeing a tomato and seeing a plate. To be capable of what we might call propositional perception—seeing ¹ I have tried to explain why such an understanding of perception is needed, and to outline one way in which it could be defended, in my “Scepticism and the Senses” and “Seeing What is So”, both in this volume.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

DOING SOMETHING INTENTIONALLY . . .



that so-and-so—a perceiver must be capable of propositional thought. No such capacity is required for seeing a tomato or seeing a plate; a cat or a mouse can see a tomato or a plate. But no one could see that there is a red tomato on a white plate if he did not understand that propositional thought. That involves having some conception of the conditions under which the thought would be true, and some capacity to recognize or acknowledge its truth when appropriate, or at least a capacity to recognize something that is reason to believe it. A thinker’s understanding of propositional thoughts involves a capacity for judgement or putting forward something as true. It is only because an accomplished perceiver is in full possession of a rich conceptual structure of possible judgements that he can sometimes see and thereby know without mediation that suchand-such is so. He is confronted with the fact that a certain thought he understands is true right before his eyes. But he could not know that fact in that way if he did not have that rich capacity. I think this is the key to understanding the possibility of knowledge of the world by perception in the face of the persistent challenge that what it calls “perceptual experience” alone is never enough. Once we acknowledge that the possibility of some such knowledge of the world must be within the repertoire of competent perceivers, we can see through the apparent threat and understand how human perceptual knowledge of the world is, in general, possible. My parallel suggestion for knowledge of action is that the possibility of agents’ having some immediate, nonderivative knowledge that they are intentionally doing such-and-such must also be understood as involved in the very capacities and competences human beings must possess even to qualify as agents of intentional action at all. Someone who knows that she is doing a certain sort of thing must understand or have some conception of the kind of action she knows she is performing. She must have a way of thinking of it, and she must be able to think of it as something she does. Of course, she need not explicitly formulate the intention to herself at the time, or ever, but she will be doing something intentionally only if there is some conception she understands or can acknowledge that expresses an intention she can be truly said to have. The words ascribing the intention present a way the agent herself can think of what she would be doing in carrying it out. Those words appear in an intensional context which does not permit unlimited substitution. To intend to curtsy to the Queen is not

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



DOING SOMETHING INTENTIONALLY . . .

to intend to curtsy to that short old woman standing stiffly over there with her back turned. What an agent does intentionally is something that fulfills or executes some aspect of an intention expressed in a true attribution to her of an intention she understands. Since an agent needs some way of ascribing an action to herself, she must be a master of first-person thought. The knowledge that she is doing such-and-such is in that sense a kind of self-knowledge. This requirement is often understood as involving one’s consciousness or awareness of oneself or of something “within oneself” that is an exercise of one’s agency. This in turn has seemed to promise a different way in which an experience of a certain kind could provide a ground or basis of one’s knowledge of what one is intentionally doing. But I think we can see that there is no prospect of satisfaction in that direction either. For one thing, the self-knowledge that is required for self-attribution need not involve awareness or experience of anything. And appealing to an experience or awareness of something, whatever it might be, could never alone account for the kind of knowledge involved in intentional action. This is clear enough if the agent’s experience of something “within himself” is understood as one of those “immediate” or “non-committal” experiences of the traditional philosophical kind. The presence of such an experience could not explain the agent’s knowing anything beyond that very experience itself. Even if it is an experience of what the agent calls his “willing” or “initiating” the occurrence of something, he would not know from that experience alone that any “willing” or “initiating” was going on. But even an agent’s experience of something not solely “within himself”, even an experience of something actually going on in the action itself, so to speak, would not help explain how the agent knows that he is intentionally doing such-and-such. Take the experience of walking across the Golden Gate Bridge. What could you be aware of while having that experience that supports or grounds a claim to know that that is what you are intentionally doing? What could you be aware of in what is going on that would distinguish your intentionally walking across the bridge from your intentionally walking only half-way across, or intentionally walking on the bridge for only five minutes, or intentionally walking from San Francisco to Sausalito by the shortest route, and so on? Anything you were aware of at any time would be present in each of those different intentional actions. I think this is fatal to the idea

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

DOING SOMETHING INTENTIONALLY . . .



that an agent knows that he is intentionally doing such-and-such by being aware of or experiencing something present in the action. This shows the fundamental importance of the difference between being aware of something and knowing something. It is crucial to the proper understanding of human knowledge. Knowing something is a propositional state; it involves judging that p, something that is true or false. But you can be aware of some thing or feature or quality without judging anything. Nor does being aware of something imply knowing that one is aware of it. Ascriptions of knowledge are expressed in intensional terms; what an agent knows when she knows that she is doing such-and-such is something expressed in terms she understands. What she knows she is doing intentionally therefore cannot be simply read off whatever she is in fact doing; not just anything whatever that is true of what she is doing says what she is intentionally doing. That is why an agent’s knowing that she is intentionally doing such-and-such cannot be derived from her being aware of something or other while doing it. I think this is the real obstacle to explaining agents’ knowledge as based on experience or awareness of something, whatever it might be. And I think this is also the key to understanding the distinctive character of one’s knowledge of what one is intentionally doing and how that knowledge is possible. To know that you are intentionally doing such-and-such is not necessarily to know that you are succeeding in your action in every respect, especially with complex actions that extend over time and space. You usually need to be aware that one step in the process has been completed successfully in order to go on correctly to the next. You have to attend to what you are achieving with the objects you are dealing with in the situation. In monitoring your actions, what you become aware of is that such-and-such is so in the world you are acting in. That can tell you what you are doing, but only in the sense that it also tells you what you are bringing about but not doing intentionally. You come to know in that way what the results of your intentional actions are, but that is not necessarily to know that you are intentionally doing what you find you are doing or have done. Knowledge of what is so in the world is helpful in telling an agent to what extent what he has brought about accords with what he is intending to bring about. But it can tell him such a thing only if he knows what he is intending to bring about. And that is not knowledge he gets by monitoring what he is doing. He monitors what

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



DOING SOMETHING INTENTIONALLY . . .

he is doing to be sure that what he is doing accords with the intentions he independently knows he has in acting as he does. If you find that what you are doing or have done does not accord with your intention, your performance is faulty in some way. But that does not mean that you do not know that you are intentionally doing such-and-such. This can happen even with the apparently simplest actions. I might find I am tying the laces of my two different shoes together, or I suddenly realize I am about to put the milk into the dishwasher. What I know that I am intentionally doing is putting the milk in the refrigerator. Why, then, have I just opened the dishwasher? Good question. My mistake or error, if that is what it is, is possible— I am failing to do what I intend to do—only because I have an intention to put the milk in the refrigerator, and I have the ability to carry out that intention. If I did not have that intention, whatever I was doing with the milk at the dishwasher door would not be a mistake or faulty performance. Having the intention involves my having a way of thinking of what I am doing, and so having a capacity to recognize that I am doing it, or not doing it. Luckily, I recognized it just in time. With actions involving bodily movements, and tools, and other objects, monitoring the degree of success of your actions usually takes place through sense perception. You see or hear or otherwise perceive what you are bringing about, even if you do not strictly speaking know by perception that you are intentionally doing such-and-such. But the distinctive features of an agent’s knowledge that he is intentionally doing such-and-such stand out more clearly with actions that do not involve bodily movement, where sense perception is not essential. If I undertake to multiply 79 by 45, “in my head”, as we say, that is something I could not intend to do, and so could not be said to be intentionally doing, unless I understood what multiplication is and was competent in operating with numbers. When I do it, multiplying those numbers is what I am intentionally doing, and I know that that is what I am doing. I have certain multiplication abilities, and I know what it would be to arrive at the product of those two numbers. Of course, I might make a mistake along the way, and get the wrong answer. But even then I know that what I am doing is directed towards carrying out that intention, whatever answer I happen to get, just as I know I am intentionally putting the milk in the refrigerator even as I open the door of the dishwasher. When I have finished my calculations, someone could

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

DOING SOMETHING INTENTIONALLY . . .



report, with amused incredulity, “He multiplied 79 by 45 in his head, and look what he got!”. Nonetheless, that is what I was doing, and I knew that I was doing it. This shows that what I actually do does not serve as a ground or basis of my knowledge of what I am intentionally doing. I think my knowledge that I am intentionally multiplying 79 by 45 has no ground or basis in that sense. It is something I know, but not in any such derivative or indirect way, just as my perceptual knowledge that there is a red tomato on a white plate before me is not based on or derived from what I see. In that case, what I see is what I know to be so. But still, in the case of seeing, I could have come to know what I know about the tomato and the plate in some different way. My knowing what I do can be explained as having come about in one way rather than other ways in which I could have come to know it. But with knowing that one is intentionally doing such-and-such, there seems to be no such range of other possible explanations and so no other ways of knowing.² It is because the agent has the intention he has, or because he is acting under that intention, that he knows that he is intentionally doing suchand-such, but that does not explain how he knows. It is possible for an agent to come to have a certain intention, or to intentionally do suchand-such, for a certain reason, even though he could have had that intention, or acted intentionally in that same way, for a different reason. In that case there would be different possible explanations of how the agent came to have the intention or to act intentionally as he did, but not different possible explanations of how he knows that he is intentionally doing what he is doing. It is also possible for an agent to become distracted, or to forget, or in some other way no longer realize that he is intentionally doing such-andsuch. That was true of me at the dishwasher. One can then suddenly remember, or be reminded by something, or even be told by someone, that one is intentionally doing such-and-such, and pick up where one had left off. But in being reminded by something or somebody, your seeing or hearing what you then do does not explain how you know that you are intentionally doing such-and-such. It serves to remind you or bring back to your attention something you already knew but had for the moment forgotten or overlooked. ² This is one of several points at which I am grateful to Johannes Roessler for helpful comments on an earlier version of this essay.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



DOING SOMETHING INTENTIONALLY . . .

If you had not already known that you were intentionally doing suchand-such you could not have come to know in any such ‘observational’ way that that is what you are intentionally doing. You can come to know by observing or monitoring your actions what you are in fact doing. You can also come to know in other ways what you are in fact doing: friends might tell you, or you might discover the disastrous effects of your efforts right before your eyes. You can find out in such ways what you are doing but not intentionally doing. But you cannot find out in one or another of those ways that you are intentionally doing such-and-such. Even if you are intentionally doing it, and you observe or are told that you are in fact doing it, and you believe what you observe or are told, you do not thereby know that you are intentionally doing it. You cannot know that you are intentionally doing such-and-such without intending to do it. To have such an intention is to accept or endorse as one’s own an intention to do such-and-such. And that involves understanding and acknowledging an expression of that intention as ascribed to oneself. But one can see that one is in fact doing such-and-such without having or acknowledging any such intention. The fact that I can know that I am intentionally multiplying two numbers, or that I am intentionally putting the milk in the refrigerator, even though I do not succeed, does not impugn my knowledge that I am intentionally doing those things. The possibility of failure in actual performance can perhaps make it look as if what is called my ‘knowledge’ of what I am intentionally doing is really only knowledge of my having certain thoughts, or of my intending certain things—knowledge only of something going on “in my mind”, so to speak, not anything that is actually so beyond that. But the admitted possibility of failure to perform correctly is no more a threat to the general possibility of knowing that you are intentionally doing such-and-such in the publicly observable world than the admitted possibility of error is a threat to the general possibility of knowing by perception what is so in the world right before your eyes. In neither case does the possibility of failure or error show that the most one could ever know is restricted to something or other in your mind. The capacities an agent must possess and exercise even to be capable of intentional action are enough to ensure the general possibility of an accord between intention and performance. Someone with virtually no manual dexterity could not even intend to tie his own shoelaces; nothing

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

DOING SOMETHING INTENTIONALLY . . .



he did could express that intention. But even an extensive capacity for self-movement and bodily control is not enough for intentional action. An agent must also be able to think, and to think of what he does in certain ways. Someone who knows no arithmetic could not attempt to multiply two numbers; he does not have what it takes even to have such an intention. An agent of intentional action must be capable of regarding certain considerations as reasons to act or to judge in one way rather than another, and he must be able to bring those assessments to bear on thoughts and intentions and beliefs he knows he has as well as on bodily movements he is capable of. If a rich, intricate structure of thought, understanding, and assessment is essential to being capable of knowing that one is intentionally doing such-and-such, as I think it is, this means that animals, even so-called higher animals, cannot be said to know such things about themselves. It is not that they act unknowingly, or unwittingly, or without attention or full awareness of what they are doing. Most animals are extremely competent agents in what matters to them. They know how to do the things they have to do. They do what they do very well and, when it counts, they are highly successful. But I don’t think the successful lion knows that it is killing and eating an antelope, or eating its prey, or eating something it has killed, or even knows that it is eating. The lion, I believe, has no such ways of thinking about what it is doing. But that doesn’t matter; it has no effect on the lion’s expert performance. The lion also lacks any way of thinking of and attributing something to itself. If it did not sound faintly ludicrous, it could be said that a lion lacks mastery of the first person. Of course, each lion in fact occupies a unique place and time in the world, and each lion does what it does at each moment only from the perspective or point of view it then occupies. But the lion, I believe, does not think of itself as occupying such a point of view. And again that doesn’t matter. Lions do not need all the resources we have. They live the kinds of lives they live without them. Our lives are different. Much is known, and there is much to learn, about how lions and other animals actually do what they do—what they are aware of, what they attend to at different stages of the action, what kinds of perceptions and feedback mechanisms are at work in leading them to respond this way or that to what is happening, and how they are able to bring about what they succeed in doing. Similar studies of human beings and their actions

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



DOING SOMETHING INTENTIONALLY . . .

would explain exactly how we learn to respond as we do, and what we attend to, in getting our shoelaces tied in the ways we do, or getting a small container into a refrigerator. But I don’t think that kind of information, for all its interest in explaining how human agents manage to do the things they do, explains the knowledge human agents have that they are intentionally doing such-and-such. Similar factors are present in the performances of human beings and of non-speaking animals, but those factors do not serve to explain how those animals know that they are intentionally doing such-and-such. They know no such thing; only we can know that kind of thing about ourselves when we do things intentionally. To the original question ‘When a human agent knows that she is intentionally doing such-and-such, how does she know it?’ I think we cannot say that she knows it by, or on the basis of, some experience or type of experience she has when she is intentionally doing it. I think the kind of knowledge an agent has of what she is doing is not to be explained piecemeal, so to speak, from occasion to occasion, and on the basis of something the agent finds to be present each time she does something intentionally. Nor is the possibility of perceptual knowledge of the world around us to be explained in that way, on the basis of something less than the fact itself that is found to be present each time. On the contrary, to understand the kind of knowledge in question I think we must proceed from the top down, so to speak, and begin with what an agent or a perceiver must know, and what thoughts, reasonings, and activities he must be capable of, even to qualify as having the intentions and beliefs that are necessary for the kind of knowledge in question. No agent or perceiver could know what he knows on each such occasion if he had not had a lot of experience and training—experience of doing things, learning how to do things, and learning how to think about things, including the things he is doing and thinking. All that experience and learning eventually gives the potential human agent a rich repertoire and competence in doing and understanding many different kinds of things he seeks to do. That general competence, along with the associated knowledge and sets of abilities, is in play on every occasion of human intentional action; it is part of the very makeup of any effective human agent, and it carries with it the possibility of agents’ knowing what they are intentionally doing.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

DOING SOMETHING INTENTIONALLY . . .



This is the source, I believe, of the fundamental point stressed both by G. E. M. Anscombe and by Donald Davidson, each in their own way. Anscombe observes that, at least in the “more immediate descriptions” of what a person is doing, “the failure [of agents] to execute [their] intentions is necessarily the rare exception” because “it is the agent’s knowledge of what he is doing that gives the descriptions under which what is going on is the execution of an intention”.³ For Davidson, the fact that what an agent does “is intentional action under some description . . . requires that what the agent does is known to him under some description”.⁴ For any agent who moves his body in acting intentionally, there is some description of his movement under which he knows he makes it, e.g., I move my body in the way required to tie my shoelaces.⁵

This expresses a certain kind of necessity. It finds the general possibility of agents’ knowing that they are intentionally doing such-and-such to lie not in the careful observation or distinctive experiences of self-conscious agents, but in the very conditions of there being any agents of intentional action at all. One could not think of, or think of oneself as, acting intentionally without the capacities for understanding, for the assessment of reasons, and for the self-ascription of those attitudes required for competent intentional agency. We understand the possibility of the special knowledge we have of our own intentional actions only to the extent to which we understand the distinctive character, and the rich conditions, of informed human agency.

³ G. E. M. Anscombe, Intention, Blackwell 1959, p. 87. ⁴ D. Davidson, Essays on Actions and Events, Oxford University Press 1982, p. 50. ⁵ Davidson, p. 51.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

10 Judgement, Self-Consciousness, Idealism I am increasingly struck in new ways by the indispensability of a capacity for judgement for the possibility of thought and experience. The idea itself is hardly new or surprising to any reader of Kant’s Critique of Pure Reason. Nor do I think it stands in need of further justification. Thought and experience require some way of thinking of, or some way of being aware of, something. That requires concepts, and the Kantian idea that concepts are predicates of possible judgements means that only someone capable of holding or accepting or putting something forward as true— or at least of entertaining or considering something as being true or false—is capable of thought and experience of the appropriate kind. This is the Kantian idea at the heart of Frege’s conception of thought and of the fundamental role of predication. Predicates are the predicates they are, and express the concepts they do, only because of the specific roles they play in determining the conditions under which full sentences in which they appear are true. Any sentence must contain some expression that does not serve solely to name or refer to something or other. The distinctive contribution of predicates is to combine in a special way with other expressions to yield something that does not just name something but says something—something that is either true or false. Only thinkers who have a conception of something’s being so can therefore have concepts and be capable of thought in the fullest sense. One of the ways—perhaps the primary way—that conception or that capacity can be exercised is in judgement. Many people are willing to ascribe thought and the possession and application of concepts to any creatures that show in their behaviour that

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

JUDGEMENT , SELF - CONSCIOUSNESS , IDEALISM



they can distinguish one kind of thing from another. Whether that involves anything we can think of as thought, and if so what kind of thought it is, seems to me largely an idle question until we describe as carefully as possible exactly what the creatures in question are able to do, and what role those capacities play in their lives. What verbs, and in particular what psychological verbs, if any, do we have to use to describe and explain the behaviour of such creatures? And what are the typical objects or complements of those essential psychological verbs? I think these are the fundamental questions. With intelligent, mature human beings who come to know things by observing and interacting with the world and assessing the reasons for and against certain views of it, there is no question that they think and reason. We must ascribe such abilities to human beings to account for what we can see and understand them to do. And the thought we must ascribe to those human beings, I believe, is ‘propositional’ thought. Fullblown thinkers must be thought of as having attitudes of considering, assessing, adopting, judging, or taking a stand on whether or not something or other is so. We make sense of people by using psychological verbs the complements of which are ‘propositional’ in the sense of whole sentences that express something that is, or could be, either true or false. Of course, some of the human attitudes we observe can be described with psychological verbs whose complements are singular terms or names or words for things or kinds of things, not full sentences that express something that is true or false. We say ‘John sees a bird in the garden’, which does not imply that John thinks of what he sees as a bird, or even that he believes anything to be so. A cat, even a frog perhaps, can see that same bird in the garden. And Mary might be said to be looking for her eyeglasses even when she is looking right at them and in fact sees them right there on the table without believing that they are there. So there is no question that some psychological attitudes are attitudes to objects or things as opposed to attitudes with ‘propositional’ objects or complements. But in stressing the indispensability of ‘propositional’ thought and experience for human life, I mean that appealing only to psychological attitudes with things or objects as their complements would never be enough to describe and explain all the things we know mature human beings do. Human thought involves thought of something that can be true or false. What then is essential to the possibility of ‘propositional’ thought and experience, and what does it take for someone to be able to engage in it?

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



JUDGEMENT , SELF - CONSCIOUSNESS , IDEALISM

These are huge questions. I draw attention to them here not because I think anyone would explicitly deny the distinctive character of predicates and predicational thought, or would deny the centrality of the idea of truth. But I think such capacities are indispensable for thought, and I think this is not always sufficiently appreciated, and sometimes is even lost sight of, in much very active recent work in philosophy. I also draw attention to the capacity for judgement or propositional thought because of its connection with the self-consciousness of the thinking or judging subject. This is something Rolf Peter Horstmann has recently written about in “Propositional Activity in Kant and Hegel”¹ and in “The Limited Significance of Self-Consciousness”.² I have found what he says there very illuminating, and I want to explore it further. An underlying goal for both us is how far one can go in making clear sense of these matters without falling into any form of idealism. I have found that some people are much less afraid of falling than others. Just as I start with the idea that someone can be conscious and aware of an object without being conscious or aware of something or other’s being so, Horstmann starts with the idea that someone can be aware or conscious of something without being aware or conscious of themselves as the subject who is aware of it. Consciousness of something does not in general imply self-consciousness. But for Horstmann self-consciousness must be present with ‘propositional’ states of consciousness. A “self-conscious I”, he says, is a “necessary element” of such states (444): “the self-conscious I understood as the self-conscious subject is a necessary condition for being in a conscious state whose content is a proposition” (447). This necessary condition carries with it further necessary conditions. “A conscious propositional state”, Horstmann says, “consists in a mental attitude of a subject towards a propositionally structured content”, as in the normal everyday judgement ‘This house has a blue door’ (444). “Thus to be in a conscious propositional state means for a subject to be conscious in one or the other of the appropriate modes (believing, wishing, etc.) that such and such (e.g., this house) is so and so (e.g., has a blue door)” (450). This means that the “subject must have access to or must be able to

¹ In Dina Emundts (ed.), Self, World, and Art: Metaphysical Topics in Kant and Hegel, De Gruyter 2013. ² Revue de Métaphysique et de Morale 2010 (page numbers in parentheses in the text refer to this essay).

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

JUDGEMENT , SELF - CONSCIOUSNESS , IDEALISM



employ conceptual resources” ordered in a way to form a proposition (450). Furthermore, “this subject has to be conscious of itself as the subject of that state because otherwise it could not be its conscious state” (451). Horstmann explains why this must be so: this is a further necessary condition. To have a propositional thought I must somehow recognize a difference between the thought and the subject ‘I’ who thinks the thought. “I am not my thought but the subject of my thought”, he says (451). I can distinguish between me as thinker of this thought and the thought that I think. If that thought is a propositional thought, I can see, usually, that what I think could be so (or not) whether I think it or not. I recognize that the truth or falsity of the thought is independent of my thinking it. And that obviously requires both that I have a thought of something’s being so and that I am able to think of myself. I think of myself as a thinking subject or “I”. This is what I think Horstmann means by “the self-conscious I”. He goes on to enquire into what he calls “the origin and the ontological status of the self-conscious I” (450). He thinks it “comes into being in virtue of there being a propositionally-structured item which can function as its object, and that . . . the propositional structured item can occur as an object only if there is a self-conscious I” that has that item as its object (453). Horstmann then argues that if you accept the idea that there being something propositionally-structured is the result of some kind of activity on the part of the subject, and you accept the idea that there can be such a propositional object of thought only if there is also a self-conscious I, then “the most natural move”, he says, is “to attribute the formation of the self-conscious I to the very same activity which is responsible for our responsiveness to propositional structure” (453). This “most natural move”—explaining both the propositional thought and the self-conscious I who has it as generated or constituted by the very same activity—is then explored in the rest of the paper, especially as the activity in question is explained in Kant and in an apparently more tolerant form in Hegel. Horstmann finds serious difficulties in both attempts, and gives good reasons to doubt that either strategy as it stands can provide as robust and rich a conception of the thinking and experiencing subject as we need to make sense of all the things we know human subjects can think and experience and do. What is more, those attempts to explain the possibility of an epistemically accessible world that the self-conscious I can think of as an object of knowledge apparently lead in

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



JUDGEMENT , SELF - CONSCIOUSNESS , IDEALISM

each case to a world that is in one way or another “mind-dependent”— constructed or constituted as it is only in accord with our own forms of representation. So the result, in one form or another, is idealism. I would like to see to what extent we can understand selfconsciousness and its importance for propositional thought while stopping short of that idealism. So I would like to try to resist the assumption Horstmann thinks is “the most natural move” for explaining selfconsciousness. For what by now are familiar reasons, it is not easy to understand the kind of “activity” that constitutes both the subject of a thought and the propositional object that is the object of that thought. It is apparently an activity of synthesizing something or other into a kind of unity, but what is that activity, and what does it operate on? It operates on a “manifold” of something or other, but presumably not a “manifold” of experiences. Experiences are presumably themselves constituted by just such an activity. When the activity operates successfully, something new is said to be “constituted” by it, not simply by a blind process, but apparently by some kind of agency. Somebody, or something, does it. But then, what could that agent be? I would like to put most of these difficult questions to one side and ask about only one apparent implication of this idea of “synthesizing” as essential for propositional thought and self-consciousness. It seems at the very least that if this “synthesizing” activity is an activity that goes on whenever anyone has a thought of something or other’s being so, then each of the essential products of that activity must be generated or constituted at that time too. I am not sure about this, but it seems to follow if this “synthesizing” is an activity or process. But maybe it cannot be understood to take place in time. Perhaps talk of “synthesis” is only another way of saying that a person’s thought or experience must be unified in a certain way in order to be the thought or experience it is. That says nothing about an activity; and it says nothing about generating or constituting a self-conscious I either. Horstmann calls what he has in mind an activity. That is the assumption from which the “natural move” starts. And some of the things he says about the activity suggest that he does think of this object- and subject-“constituting activity” as something that goes on, something that happens, or perhaps even an activity that someone or something engages in. He says that when he is walking along the street and sees the house with the blue door, “this judgment ‘This house has a blue

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

JUDGEMENT , SELF - CONSCIOUSNESS , IDEALISM



door’ . . . cannot occur without my being conscious of me as its subject” (447). That is presumably because on that occasion the very activity of “constituting” the propositional thought also generates consciousness of himself as the subject or thinker of that thought. The activity is both object- and subject-constituting. But this seems to imply something that is not true. It is not true that whenever a thinker has a thought or holds a propositional attitude or is in some propositional state that thinker is also aware or conscious of himself as the subject of that state. He is the subject of the propositional state, but whether he thinks or is aware of himself as the subject of the state he is in is a separate question. And in most cases, I would say, a propositional thinker does not think, or is not aware, of himself as the subject of his thought. A thinker is, so to speak, present whenever there is a thought, but no “self-conscious I” need be present, or present to the thinker, each time. Horstmann walks along and thinks ‘This house has a blue door’. At that moment he does not, at least not typically, also think ‘I think this house has a blue door’. He is not conscious of himself as the subject of the state of thinking that he is actually in. But it seems as if he would have to be aware of himself as subject in that way each time he has a propositional thought if the activity of combining the ingredients of a propositional thought into a unity also and necessarily generates the presence of a self-conscious I which is conscious of itself as the subject of that thought. In denying that there is always, or even very often, such self-awareness when someone has a propositional thought I do not mean to deny that a self-conscious I is a “necessary element” in someone’s being in propositional states. I think it is a necessary element in the sense that any subject or agent of propositional thought is and must be a self-conscious being. He must be capable of thinking of himself as the subject or agent of his thoughts; if he did not have that capacity he would not be capable of propositional thought, for reasons like those Horstmann has given. So for propositional thought I think we have to accept something like Kant’s weaker condition that it must at least be possible for the ‘I think’ to accompany (I would say ‘attach itself to’) each of my propositionallystructured thoughts. But no self-awareness and no self-ascription of thought need be made on a particular occasion in order for a judgement or a propositional thought to occur. A self-conscious or self-aware I or subject is not necessarily present.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



JUDGEMENT , SELF - CONSCIOUSNESS , IDEALISM

This suggests that what is required for the presence of a propositionally-structured thought, and so for the possibility of the required self-consciousness that goes along with it, is something other than an activity or process that somehow generates or constitutes those elements. What does it take, then, for someone to have a propositionallystructured thought? What does it take for what someone does on a particular occasion to be an expression of such a thought? I think it would not be enough for the person simply to utter or to respond in some distinctive way to what is in fact a full sentence. The person might utter or respond to the sentence as a single unstructured signal or symbol, without recognizing or thinking of it or understanding it as a sentence at all. You could train your dog or a small child to run to the door every time he hears what is in fact a certain sentence from the Critique of Pure Reason. He would not then have understood the internal structure of that sentence or understood it as expressing something that is either true or false. I think whether someone grasps the internal structure of a sentence and understands it and so has a thought that says something true or false depends on what other sentences or thoughts the person can understand, and on what kinds of sentences or thoughts they are. A competent thinker brings to each particular sentence he hears or speaks a general ability or capacity to understand and respond in appropriate ways to a great many different sentences of similar or related forms. He understands what the sentence ‘This house has blue door’ says because he knows how it differs from other sentences he understands such as ‘This house has a black door’, ‘This car has a blue door’, ‘This house has a blue window’, and countless other sentences about objects and their parts and their colours and shapes and so on. It is because of the internal complexity and the wide generality of a speaker’s capacity to understand sentences and thoughts of many different kinds that he can discern the relevant structure of a particular sentence, and so have a particular thought with that structure. This ability or capacity is something we all acquire in learning to speak. It can be thought of as a kind of practical knowledge, or knowhow; a certain expertise that all speakers and thinkers have. It is a general competence in speaking, thinking, responding, and acting whose intricate structure is at least as complicated as all the differences among the great variety of sentences we can understand and discriminate from one

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

JUDGEMENT , SELF - CONSCIOUSNESS , IDEALISM



another. This understanding does not always involve judgement or assertion of the thought. A thinker can wonder whether a certain house has a blue door, or know that if the house has a blue door then so-and-so, without judging or believing that any house actually has a blue door. But if the speaker endorses, or judges, or puts the sentence forward as true, he can be said to think or believe that the house has a blue door. Horstmann apparently came to believe that the house has a blue door because he saw that the house has a blue door. What he saw to be so is the very thing he thinks to be so in the thought ‘This house has a blue door’. But he is able to see that that is so, and to think or believe it, only because there are a great many other thoughts he could also have, and many other things he could also see to be so (if they were so). Those other thoughts he is capable of are what we might call “possible judgements” for him. It is because such possible judgements are within his repertoire that he can be said to possess the concepts he employs in the particular thought he actually has. “Concepts are predicates of possible judgements.” So what you think, or whether you think such-and-such on a particular occasion, depends on what else you can understand and think and judge. It is a matter of what capacities you possess, and what you express in what you say or do, not a question of what processes or activities, even object- or subject-constituting activities, go on at the time of thinking the thought. When someone sees and so has the thought ‘This house has a blue door’, I think the person does not also have to have the thought ‘I think that this house has a blue door’; no self-conscious I or self-awareness on the part of the thinker has to be present at all. But, as Horstmann suggests, what the person thinks when he thinks ‘This house has a blue door’ is something that is true or false independently of whether the thinker or anyone else thinks it. And any thinker of propositional thoughts understands that that is true of most of his thoughts. He knows it simply by understanding the thoughts he has. Someone who thinks that this house has a blue door understands what it is for a house to have a door and for the door to be blue. In understanding that sentence or thought he knows the conditions under which it would be true, so he knows some of the things it implies and some of the things it does not imply. And he can see that it does not imply that he or anyone thinks that the house has a blue door. As soon as a competent thinker sees the door of that house, he knows that the house has a blue door. And

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



JUDGEMENT , SELF - CONSCIOUSNESS , IDEALISM

he knows that is something that would be so whether he or anyone sees or thinks it is so. For a thinker to recognize or acknowledge that what he thinks could be true (or false) independently of whether he thinks it, the person must be capable of thinking of himself as a thinker of thoughts. Having a conception of the independence or objectivity of what he thinks to be so carries with it the need for a capacity to think of himself as subject or bearer of those thoughts. Thought of oneself, or self-attribution of the thought, is not necessary each time one has a thought of an objective state of affairs. But without a capacity to think of myself as a thinker, and to attribute thoughts to myself, I could have no conception of something’s being so (or not) independently of whether I think it is so. The objectivity I ascribe to the things I think to be so is intelligible to me only because I have a way of thinking of their independence from their being thought or believed by me. And that requires that I have a special way of thinking of me. This way of thinking of myself is special in that it is not simply thinking of myself as a subject or bearer or thinker of thoughts. I do have the concept of such a subject or thinker; it is a concept I can apply to other people and so think of them as having certain thoughts and beliefs. With other people’s beliefs I can easily acknowledge the independence of the truth of what they think from its being thought or believed by those thinkers. That is because I can sometimes know or see right before my eyes that what those other thinkers think is not so. I know they believe it, but it is not true. But I can find such a thing to be so—I can recognize their error—only because I myself then believe or know something to be so that is incompatible with what I take the others to believe. And that means that I must then have a thought about what is so that I take to be objectively true. That is, I understand what I believe to be true or false independently of whether I think it. And my having that thought is what requires this special way of thinking of or referring to myself, not simply as one among many other agents or thinkers of thoughts. It is not that I am or must be something other than that; only that I must have a special way of thinking of or referring to the particular thinker that I am. This means that I must also think of other agents or thinkers as themselves capable of thinking of themselves in this same special way. To think of themselves as having thoughts that are true or false independently of whether they think them, they too must have some special

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

JUDGEMENT , SELF - CONSCIOUSNESS , IDEALISM



way of thinking of themselves. They must be capable of thinking of or referring to themselves not simply as one among many other agents or thinkers. Each of us has this special way of thinking of ourselves in the use of the first person, typically with a first-person pronoun. I can think of other thinkers, and even of myself, in the third person. But in so far as I think of others as thinkers of propositional thoughts that are true or false independently of their being thought, I think of them as also capable of thinking of themselves, and attributing thoughts to themselves, in this first-person way. I can think of them as having the thoughts I attribute to them in the third person only because I also think of them as having a capacity to attribute such thoughts to themselves in the first person. And they must think the same of me in so far as they see me as a thinker of propositional thoughts. This special way of thinking of or referring to ourselves must be within the capabilities of any thinker of propositional thoughts. This seems to me the sense in which self-consciousness is required for any thinker of propositional thoughts. Is this the “self-conscious I” Horstmann thinks is a “necessary element” in any “conscious state whose content is a proposition” (447)? Well, I don’t think there needs to be any self-conscious thought of oneself as the thinker of the thought every time one has a propositional thought. But if “having self-consciousness” or being a “self-conscious being” means only having a general capacity for thinking of or referring to oneself and one’s thoughts in this special first-persona way, then I think it is essential for propositional thought. One often thinks or says of oneself in this way that one thinks or feels or wants or hopes that such-and-such. Only beings capable of that way of thinking of themselves could have propositional thoughts that they understand. And no one can have thoughts they do not understand. So, every thinker of propositional thoughts must be capable of self-consciousness in this sense. The very words “self-conscious” or “conscious of myself ” suggest that when I am in such a state there is something I am then conscious of or aware of. This leads to familiar questions about what that something could be. Is it a “self ”, or “my self ”; is it an “Ich”, or “my Ich”? And if so, what sort of thing is that? When I try to look more closely, whatever it is seems strangely elusive, as Horstmann points out. But how can it be so elusive if I am conscious or aware of it every time I say or think something about myself in the first person?

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



JUDGEMENT , SELF - CONSCIOUSNESS , IDEALISM

I think that, as an object of awareness, the self is more than elusive. On the question of what I am aware of when I think about myself, I think Hume was closest to the truth. “For my part, when I enter most intimately into what I call myself,” Hume said, “I always stumble on some particular perception or other . . . I never can catch myself at any time without a perception, and never can observe anything but the perception”.³ As a point about awareness, or what Hume calls “observing” something, this seems to me right. I think it is not just that I never manage to find the thing I am looking for, however carefully I look. I think nothing I notice or find or am aware of could be something I can reliably depend on in picking myself out. For any such property or thing that I could notice and think about, I could conceivably doubt or deny its alleged connection with me, without thereby failing to think about myself. I might have wildly false beliefs about myself, or doubt or deny that whatever I had found was really me. I could even wonder, as apparently some people sometimes do, whether I am actually the thinker of this very thought. Of course, in having the thought ‘I wonder whether I am the thinker of this thought’, I refer with my first use of ‘I’ to me, the subject, the one who wonders. So I am the thinker of that very thought, whatever I might believe or wonder about myself. But in having that thought I do not secure reference to myself by way of the definite description ‘the thinker of this thought’. I don’t have to apply any concept or description. That reference to myself is secured simply by its being made by me in the first person, no matter what I might believe or doubt or wonder is true of me. Competent use of that ‘I’ is all that is needed for self-reference, and so for self-consciousness, if self-consciousness is first-person attribution of psychological states or attitudes to myself. That is a capacity I think is required for the possibility of propositional thought. But it does not require awareness or experience of an object of some recognizable kind, especially not an elusive object. The self-consciousness I exhibit in attributing states and attitudes to myself in the first person is itself a form of propositional thought. It involves saying or thinking something about myself that is expressed in a full sentence. I say ‘I think that house has a blue door’, which is ³ Hume, A Treatise of Human Nature, ed. Selby-Bigge, Oxford University Press 1958, p. 252.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

JUDGEMENT , SELF - CONSCIOUSNESS , IDEALISM



something that is either true or false. I can make and understand sentences like that because I am a master of the first person. I use firstperson forms in sentences to say and think things about myself. My firstperson pronoun, like everyone else’s, refers to whoever utters or thinks it in a sentence. So there is no possibility of failure of reference, or failure to ‘get hold’ of myself when I use it. The sentences of that kind that I say or think are true, or false, depending on what is true of the person they refer to as subject. This suggests that not only is the possibility of self-consciousness essential to propositional thought, as Horstmann and I agree it is, but the possibility of propositional thought is also essential to selfconsciousness. Only in having first-person propositional thoughts about myself do I possess self-consciousness. If that is true, it further confirms the idea that the self-consciousness we human beings possess is not to be understood as an awareness or consciousness of some object or thing in the way that a creature could be aware of something whether it was capable of propositional thought or not. A being without a capacity for propositional thought might be conscious, and conscious of many different things—pains and tickles and other sensations, even aware of a hawk in the sky overhead or aware of a swift movement in the grass underfoot. But whatever such a being was aware of, and whatever its psychological life was like, it could not think of itself as having that psychological life, or as being in any of those psychological states. It could not think that it is experiencing or thinking certain things. That requires a capacity for propositional thought. So such a creature could have no self-consciousness of the kind we human beings possess. That both requires and is required by a capacity for propositional thought. The intricately articulated capacity I possess as a propositional thinker enables me to understand and, on the right occasion, to assert or endorse the thought “This house has a blue door”. Walking down the street with that capacity, and with the concepts I have acquired in exercising it, only the briefest sight of the house is enough for me to see that the house has a blue door. I acquired my capacities as a propositional thinker, and they have taken the specific forms they have taken, partly because of the way things are in the world in which I acquired it. I learned some things to be so in the world around me at the same time as I was learning how to say and think that things are thus and so in the world. I don’t think that what is so in the world is “constituted” as being so by my having the capacities

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



JUDGEMENT , SELF - CONSCIOUSNESS , IDEALISM

I have to say or think that certain things are so. Nor is it “constituted” by anyone else’s having such capacities, or by everyone’s having them. Given our natural talents, we all acquire the capacity for propositional thoughts in a world in which certain things are so. But in the very idea of propositional thought, and in our having a capacity to engage in it, and in the fact that we must be capable of self-consciousness if we can think of an objective world at all, I do not see any encouragement to idealism. The indispensability of propositional thought, and therefore our fulfilling all the necessary conditions of our being capable of it, gives us no reason to conclude that what we thereby think and know about is a “minddependent” world.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

11 Feelings and the Ascription of Feelings A small child runs and falls and hurts himself; he begins to cry. One of the many virtues of David Finkelstein’s Experience and the Inner¹ is that it encourages us to start here if we wish eventually to understand how the things we think and say about ourselves are related to the states we are in when we think or say them. To really start here let us imagine a child so young he cannot say anything about how things are with him. He cannot say ‘I am in pain’ or even ‘It hurts’. But it does hurt, and he feels it; he is aware of the pain. That is the state he is in. We know he is in that state; that is what he expresses in crying as he does. But he does not think of himself as in pain. He does not know he is in pain. He is aware of the pain, but he is not aware that he is in pain. He does not ascribe a feeling of pain to himself. A slightly older child who has developed some linguistic and recognitional capacities might be very good with her colours. Hold an object of a certain basic colour up in front of her, and she immediately and confidently tells you what colour it is. If she says of a presented object “That’s red”, and we are sure she understands what she is saying, we can say that the child believes or knows that the object is red. But the child herself at that stage might not be able to say or understand such things about herself. She does know that the object is red; that is what she has learned from her training with the colour words. She expresses that state of mind in saying ‘That’s red’, but she doesn’t know she is in that state of mind. We know something about the child’s state of mind that the child herself does not know. ¹ Harvard University Press 2003. (Page numbers in parentheses in the text refer to this volume.)

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



FEELINGS AND THE ASCRIPTION OF FEELINGS

This child will learn to ascribe such attitudes to herself and to others by coming to understand herself and others as saying things like ‘That is red’ in the presence of objects she can see. She will see that other speakers are right, or wrong, in what they say, so she will make sense of them as knowing or believing things about the colours of the objects. And she will be able to say the same kind of thing about herself. She will understand sentences in which either her own name or personal pronoun or the name or pronoun of another person can be put in the blank space in sentences like ‘So-and-so knows that that object is red’. She thereby learns to ascribe such knowledge or belief to others and to herself simultaneously (roughly speaking). It is part of one and the same predicational capacity. Once she has acquired that predicational capacity, she can ascribe to herself knowledge of the colour of an object as soon as she has determined what colour that object is. No further investigation is needed. No ‘justification’ or ‘ground’ or ‘basis’ for ascribing that state of mind to herself is called for beyond her having seen what colour the object in front of her is. A competent self-ascriber can therefore know what she knows or believes without having to gaze into a special ‘inner’ or ‘private’ repository where her beliefs and other attitudes are thought to reside. In this simple case, she has only to look at an object and see and thereby know what colour it is. With more complicated questions it is not that easy. But even when complex thought and elaborate investigation are called for, a person’s coming to know or believe what she does is something she is responsible for in the sense that it is the outcome of her own active reflection and deliberation. That is what puts her in a position to ascribe that knowledge or belief to herself. All this is true of knowledge and belief and other so-called propositional attitudes. But with feelings and sensations and other such reactions it seems different. One seems to be only on the receiving end of such states of mind. Certain feelings and sensations simply come to us; one simply finds oneself with them. They are not the products of our deliberation or agency; we are not responsible for their presence even as we can be said to be responsible for our knowing or believing what we do. With feelings and sensations, it is much more as if we are simply a receptacle or a site for something that comes or goes independently of our wishes and of whatever linguistic or conceptual competences we happen to have mastered.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

FEELINGS AND THE ASCRIPTION OF FEELINGS



The child I imagined at the beginning was in pain even though he lacked the ability to say he is in pain. When he learns to ascribe pains to himself, the state he ascribes to himself is the kind of state we knew he was in before he could say that he was. We can see and hear him crying, and he can come to see and hear other people crying or expressing their states of mind in other ways. That will be part of his learning to ascribe such states to others, and so part of his learning to ascribe them to himself. He must understand that what he thereby ascribes to himself is the same kind of state that he can also ascribe to others—and vice versa. Like the child who knows the colours of things, he learns selfascription and other-ascription of certain states of mind simultaneously. He comes to understand sentences in which either his own name or personal pronoun or someone else’s name or pronoun can appear linked together with one and the same psychological predicate to ascribe feelings and sensations to someone. The child who is good with her colours knows that an object is red as soon as she sees it. The presence of a red object in plain view is all it takes. The child who learns to ascribe feelings of pain to himself learns to ascribe to himself the kind of state we know he was in before he could say that he is in it. He was aware of the pain even then; he felt it. Once he has the capacity to ascribe such states to himself, his being in such a state is all it takes for him to ascribe that state of mind to himself. That is what he learns from his training with the predicate ‘is in pain’. If he learns his lessons well, and so is a competent self-ascriber of pains and other such states, I think he can be said to know that he is in pain when he says he is in pain. This is something some people appear to have denied: Wittgenstein, perhaps. I think there is no reason to deny it, when it is properly understood. The question is how this knowing is to be properly understood. I think it is a matter of what a competent self-ascriber can be said to be competent in doing, and of what he can be said to know in exercising that competence. Think of the first-personal pronoun ‘I’ or the first-person singular inflection of a verb. I believe an experienced, competent user of the firstperson pronoun knows who he is talking about when he uses it. He knows he is talking about himself, and he knows he is distinguishing that person from someone else when he says ‘I went to Philadelphia last year, but you didn’t’. But he does not pick out the person he thinks went to

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



FEELINGS AND THE ASCRIPTION OF FEELINGS

Philadelphia by recognizing some distinctive feature or criterion on the basis of which he identifies that person and then applies that predicate to him. If he knows he is talking about himself, as I think he does, that is not the way he knows it. The fact is that he cannot fail to be talking about himself when he says ‘I’. No speaker can fail in that way, and that is something I think any speaker competent with the first person could be expected to know, whatever he might happen to believe to be true about himself. So I think there is no good reason to deny that a speaker competent with ‘I’ refers to himself, and knows he is referring to himself, when he uses that pronoun. This too is something that has been denied, perhaps by Wittgenstein, but certainly by Elizabeth Anscombe.² When we come to the capacity for self-predication—saying of myself ‘I am in pain’—I think knowledge of the truth of what is said is also part of competent self-ascription. Wittgenstein says that when I say ‘I am in pain’ I do not “identify my sensation by means of criteria”.³ The child who knows the colour of an object she sees does not identify the colour of that object by means of criteria either. She just knowingly says ‘That’s red’. Wittgenstein says that a person’s use of the expression ‘I am in pain’ is where “the language-game” “begins”. The “language-game” he is talking about is the practice or “language-game with the words ‘I am in pain’” (§300). In §290 someone protests, “But doesn’t [the language-game] begin with the sensation—which I describe?” (§290). The young child who falls and hurts himself certainly has a sensation, but since he cannot speak that is not the beginning of any language-game in which that child has a speaking part. So we can say, “No, the language-game does not begin with a sensation, but with the use of some linguistic expressions”. What then can the person who says “I am in pain” be understood to be doing? Wittgenstein wants to resist the idea that he is identifying an object and saying something about it. He raises doubts about whether the speaker can be said to be describing his sensation. But the protester insists that, whether or not ‘describe’ is the right word for the relation between the words and the sensation, “there is a something there all the same, which accompanies my cry of pain” (§296), or something that is said to be there by what I say when I say ‘I am in pain’ or ‘I have a pain’. “You ² G. E. M. Anscombe, “The First Person”, in S. Guttenplan (ed.), Mind and Language, Oxford University Press 1975. ³ L. Wittgenstein, Philosophical Investigations, Blackwell 2002, §290. (Section numbers in parentheses in the text refer to this volume.)

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

FEELINGS AND THE ASCRIPTION OF FEELINGS



will surely admit”, the protester goes on, “that there is a difference between pain-behaviour with pain and pain-behaviour without pain” (§304). “Admit it?”, Wittgenstein predictably says, “What greater difference could there be?” (§304). Here we come to passages from Wittgenstein that Finkelstein discusses carefully and helpfully. I will take up some of these remarks with a slightly different purpose, and so in a different order. Finkelstein is primarily concerned with John McDowell’s interpretation of these passages, and I agree with him completely in rejecting McDowell’s earlier idea that a sensation must be understood as “something that is not present prior to or independently of its being brought under a concept” (131). I have already rejected that idea in my story of the child who has a sensation and cries out in pain even before he can speak or apply any concepts to it. What can we make of Wittgenstein’s reply to the protester’s complaint “And yet you again and again reach the conclusion that the sensation itself is a nothing”? Wittgenstein says: Not at all. It is not a something, but not a nothing either! The conclusion was only that a nothing would serve just as well as a something about which nothing can be said. We’ve only rejected the grammar which tends to force itself on us here. (§304)

Why does the protester think Wittgenstein reaches the absurd conclusion that “the sensation itself is a nothing”? That is one question. And what is “the grammar” that Wittgenstein thinks “tends to force itself on us here”? These two questions are answered, I believe, in §293. Imagine that everyone has a box with something in it that they all call a ‘beetle’, and no one can ever look into anyone else’s box. If the word ‘beetle’ has a general use in that community, then even if everyone says he knows what a beetle is only by looking at his own beetle, the word ‘beetle’ would not be the name of a kind of thing. As Wittgenstein says, it would be quite possible for everyone to have something different in his box, or even for the thing in a particular box to keep changing, or for a box to be empty. So, as Wittgenstein says, “the thing in the box doesn’t belong to the language-game at all”. If there is something in each person’s box, it does not belong to the language-game “even as a something” (§293). That is to say, if we construe the grammar of the expression of sensation on the model of ‘object and name’, the object drops out of consideration as irrelevant. (§293)

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



FEELINGS AND THE ASCRIPTION OF FEELINGS

The only object that need be involved when someone says ‘I am in pain’ or ‘I have a pain’ is the person who says it. He says something about himself, but not by naming an object he has or thinks he has. He predicates something of himself. The established general use of the word ‘beetle’ as described in §293 shows that when someone in that community says ‘I have a beetle’, or ‘I have brought my beetle with me to the party’, the person is not naming an object in his box. His box might be empty. He is predicating something of himself: that he has a beetle, or that he has brought his beetle to the party. He speaks of his beetle in the way we might speak of our ‘earthly possessions’ or our ‘assets’ or ‘the contents’ of a box. The words ‘possessions’, ‘assets’, ‘contents’, and so on are not names of objects, or particular kinds of objects, or particular collections of objects. They are ways of speaking of something that everyone has, whatever his possessions or the contents of his box happen to be. If we resist the grammatical model of ‘object and name’, and take this description of the use of the term ‘beetle’ as the appropriate grammatical parallel for remarks like ‘I am in pain’ or ‘I have a pain’, then those remarks too are best understood as my predicating something of myself, ascribing a certain state to myself, and not naming an object, especially not a so-called ‘inner’ object. What I am talking about when I say I am in pain is in that way “not a something”, if that means an identifiable object of a certain kind. When I say ‘I feel ill’, for instance, I ascribe a certain state to myself, but I do not name or refer to any object or entity (except myself). I am saying that I am in a certain state, when my being in that state is all it takes for me, as a competent ascriber of such states, to be right in saying I am in that state. What I say is correct, but I need not be acquainted with or reporting on the presence of some object. To say that what I am talking about when I ascribe a sensation to myself or say how I feel is “not a something” is of course not to say it is “a nothing”. That would suggest, as the protester complains, that there is no difference between having a certain sensation or feeling a certain way and not having it, or not feeling that way. Saying it is “not a something” does not deny that obvious difference. What is said of someone said to be in pain is the denial of what is said of someone said not to be in pain. No difference could be greater. And what is true of the people described in §293—that each of them has a beetle—is not something that is true of any of us. None of us has a beetle in the sense in which each of them has one. But what all those people have and each of us lacks is “not a

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

FEELINGS AND THE ASCRIPTION OF FEELINGS



something”—a thing of a particular kind. The difference between us is that a predicate that is true of every one of them is not true of any of us. Finkelstein’s book is distinctive in “paying the right sort of attention to what Wittgenstein has to say about expression and the expressive dimension of mental state self-ascriptions” (148). He thinks one important outcome of taking that dimension seriously is that “one does not retain a desire to accommodate the idea that whenever one describes one’s own state of mind, one is epistemically justified” (136). He does not think that when someone honestly avows a headache, for instance, “she is, typically, expressing an epistemically justified belief ”. “When someone complains of a splitting headache, she does not judge on this or that basis that she is in pain” (p. 135). In fact, he says in a footnote on the same page, “She does not judge at all”. “Rather, she expresses her pain” (135). I think someone who says ‘I have got a splitting headache’ is certainly expressing her pain. But she does so not by holding her head or moaning, but by uttering an articulate sentence whose structure and meaning she understands and which is either true or false. She asserts that sentence, or puts it forward as true. That is what she learned to do when she learned to ascribe mental states to herself rather than simply expressing them non-linguistically. Her saying that she has a splitting headache is not to be understood as a “mere effect” of her being in that state of mind or feeling, as a wince or a moan might be. The person says what she says knowingly and intentionally, and she knows what she is doing. That is a much more elaborate way of expressing something than simply wincing or moaning. But if the person asserts that she has a headache, can it be said that she does not judge at all? Finkelstein rightly draws attention to the importance of what Wittgenstein calls a “plan for the treatment of psychological concepts”. As Wittgenstein describes that ‘plan’ in what is published as Remarks on the Philosophy of Psychology volume 2, it says: Psychological verbs characterized by the fact that the third person of the present is to be verified by observation, the first person not. Sentences in the third person of the present: information. In the first person present, expression. ((Not quite right))⁴ ⁴ L. Wittgenstein, Remarks on the Philosophy of Psychology, University of Chicago Press 1988, vol. II, §63 (p. 12).

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



FEELINGS AND THE ASCRIPTION OF FEELINGS

Why is it not quite right? Wittgenstein does not say, at least not as the plan is published there. That same plan, and apparently that same passage, as published in Wittgenstein’s Zettel, has another sentence following the ‘((Not quite right))’. It says “The first person of the present akin to an expression”.⁵ This is an important addition. Use of the first person present tense of a psychological verb is ‘akin’ to an expression in being an expression and also more than an expression. That is what was not quite right. Saying ‘I am in pain’ is an expression of pain made by a competent speaker who knows how to say of herself that she is in pain. Such a speaker knows how to say the correct thing about herself on the appropriate occasion while understanding what she is saying. She does not use the words ‘I am in pain’ wrongly or incorrectly. She knows how to get it right. If she did not have that capacity, she would not have the competence to ascribe pains to herself as she does. Finkelstein imagines a man who has a sharp pain in his knee and who infers on the basis of other reliable information that he is therefore likely to get a certain disease. The man would not draw that conclusion about himself if he did not have the pain, but Finkelstein thinks the man can be justified in believing that he is likely to get the disease even though he is not justified in making the statement that it appears to rest on, viz., that he has a pain in his knee. That is because the man who says ‘I have a pain’ does so “without epistemic justification but not without right” (149). To say he does it “not without right” is to say he does not “base” his selfascription of pain on anything; there is nothing that is his “epistemic justification” or his “ground” or “basis” for saying what he says. In that sense Finkelstein thinks we can reject “the idea that self-ascriptions of pain are, as a rule, epistemically justified” (148). In rejecting that idea of epistemic justification while nonetheless holding that the man ascribes pain to himself “not without right”, Finkelstein apparently wants to allow that the person is “so to speak, entitled”, to ascribe pain to himself (150). That is not to say that there is something that “entitles” the person to say or think what he does. There is nothing that serves as the “basis” or “ground” of his entitlement. But

⁵ L. Wittgenstein, Zettel, Basil Blackwell 1967, §472 (p. 84).

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

FEELINGS AND THE ASCRIPTION OF FEELINGS



the man is said to be “entitled” to say what he says even though he is not “epistemically justified” in saying or thinking it. I end with some questions about this. I wonder about the sense in which a fully competent self-ascriber of mental states can be said to be “entitled” to ascribe those states to himself as he does. Is his being entitled something more than his being fully competent in ascribing the states correctly? Is it more than his simply knowing how to do it, and doing it right? Not doing it ‘with right’ or ‘justifiably’, in the sense of something like legitimacy or warrant, but simply doing it right or correctly. If he is fully competent in saying he is in pain when he is in pain, then when he says it (if he is honest), he is right. For the most part, what he says is true. Is there a further question whether he is entitled to say it? Is the child who cries entitled to cry as he does? If he is too young, and cannot say he is in pain, is he nonetheless entitled to say he is in pain, even if he can’t yet say it? Or is he entitled only when he at least has the competence necessary for saying he is in pain? In gaining that competence does he thereby gain some entitlement that he didn’t have before, or does he just gain the competence? I have been talking about beliefs and knowledge and sensations and feelings. There is also self-ascription of actions and intentions. Suppose someone is looking for a book he was reading yesterday. He knows that is what he is doing. He is doing it intentionally. He can say to himself or others that that is what he is doing. Is that something he is entitled to say or think about what he is doing? Does putting it that way help us understand how he knows what he is doing? It looks as if he knows that he is looking for the book because that is what he is intentionally doing. In the case of sensations, is it not enough to say that we learn to ascribe them to ourselves by coming to understand what is said in ascriptions of feelings and sensations to people, and then intentionally saying such things about ourselves only when they are true (if we are honest). Finkelstein imagines three different possible responses to the man who has a pain in his knee and is thereby justified in believing that he is likely to get a certain disease. The first is to say that it is “unnatural” or “misleading” or even “sheer nonsense” to say the man knows (or does not know) that he has a pain in his knee. The second is to say that the man’s ascription of pain to himself “is not epistemically justified; he doesn’t know that he’s in pain”. The third is to say the man “does know

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



FEELINGS AND THE ASCRIPTION OF FEELINGS

he’s in pain; we can say this about him” (151). Given these three alternatives, I would go along with this third response. But Finkelstein elaborates it by adding: “The mistake is to think that knowledge always requires epistemic justification”. I agree that there are certain ‘grounds’ or ‘bases’ or ‘justifications’ the man does not have. But I think that in saying that he knows he is in pain we need to understand the kind of knowledge in question better than we now do. Finkelstein says philosophers who differ in these ways about whether the man can be said to know he is in pain need not be disagreeing about anything of genuine philosophical import. I am not so sure about that. I think anyone who is inclined towards the third response, as I am, can make good on that choice, philosophically and not just as something it is all right to say, only if more can be said about the kind of knowledge we have of ourselves and our states and our actions and intentions, and about what it takes for us to have that kind of knowledge. I think what is involved is a kind of knowledge of what we think and feel and what we are doing that we must have if we are to have any of the other more mundane, straightforwardly propositional knowledge that many philosophers appear to think is the only knowledge we have. If that is right, then the third response to the man who recognizes that he is in pain is the one to explore and to try to understand. Those who are inclined to one of the first two responses would then be missing something. And that would be a matter of genuine philosophical import.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

12 Kant’s ‘Transcendental Deduction’ Kant’s unprecedented ‘transcendental’ investigation of the human mind and its powers arises from a distinctive question. It is not the question how the ‘materials’ we employ in thinking get into the mind in the first place. John Locke had tried to show in detail that sense-experience alone can explain how the mind ‘comes to be furnished’ with that ‘vast store’ of ideas ‘which the busy and boundless Fancy of Man’ is capable of.¹ A successful explanation along those lines would leave no reason to regard some ideas as ‘native’ to the human mind, or ‘stamped upon’ it, or given ‘innately’ to it from some other source. Whether anything like that is so or not is in a very broad sense a question of fact. Locke meant to answer that question of fact, and Kant concedes that ‘we are indebted to the celebrated Locke for opening out this new line of enquiry’²—‘a certain physiology of the human understanding’ (Aix). But the question Kant starts from is different, as he explains. Jurists, when speaking of rights and claims, distinguish in a legal action the question of right (quid juris) from the question of fact (quid facti), and they demand that both be proved. Proof of the former, which has to state the right or the legal claim, they entitle the deduction. (A84 = B116)

The ‘deduction’ Kant thinks is needed for understanding the human mind would establish and explain our ‘right’ or ‘entitlement’ to something we seem in fact to possess: experience, concepts, and principles ¹ Locke, An Essay Concerning Human Understanding (ed. P. Nidditch), Oxford University Press 1975, II, 1 (p. 104). ² Immanuel Kant’s Critique of Pure Reason (tr. N. K. Smith), Macmillan 1953, A86=B119 (page numbers with A or B editions in parentheses in the text refer to this volume).

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



KANT ’ S ‘ TRANSCENDENTAL DEDUCTION ’

that we employ in ‘the highly complicated web of human knowledge’ (A85=B117). For many concepts—so-called ‘empirical’ concepts—no such question seems to arise, since ‘experience is always available for the proof of their objective reality’ (A84=B116). That could be loosely called an ‘empirical deduction’ of those concepts. But it would show only ‘the manner in which a concept is acquired through experience . . . not its legitimacy, but only its de facto mode of origination’ (A85=B117). The question of ‘right’ or ‘legitimacy’ demands a different kind of ‘deduction’. There is a special problem about this because Kant thinks some of our concepts ‘are marked out for pure a priori employment, in complete independence of all experience’ (A85=B118). Since ‘empirical proofs do not suffice to justify this kind of employment, we are faced by the problem how these concepts can relate to objects which they yet do not obtain from any experience’ (A85=B117). Explaining how this can be so is the task of what Kant calls the ‘transcendental deduction’ of those concepts (A85=B117). We will return to what he means by calling it ‘transcendental’. It is to be a ‘deduction’ in the sense Kant explains: a justifying or legitimating explanation of how our use of those concepts can yield knowledge of what is so, despite their independence from all experience. Whether any of our concepts are in fact ‘marked out for pure a priori employment’, and so present us with this special problem of how they can relate to objects independently of experience, is also, in a broad sense, a question of fact. Is that actually true of some of our concepts, or not? Kant has no doubt that we do have many a priori concepts, and that we do know many things a priori, because he thinks we all know many things to be necessarily true. And for Kant necessity is a ‘sure criterion’ of the a priori: ‘if we have a proposition which in being thought is thought as necessary, it is an a priori judgment’ (B3). So the task of the legitimating ‘deduction’ of our a priori concepts is to explain how we know the kinds of things Kant thinks we all know a priori. Kant does not address this problem in its full generality. The question of how a priori concepts ‘can relate to objects which they yet do not obtain from any experience’ arises only for what Kant calls ‘synthetic’ a priori knowledge. No such challenge arises for so-called ‘analytic’ truths which can be known a priori simply by reflecting on a given concept and ‘extracting’ its ‘constituent concepts’ from it ‘in accordance with the principle of contradiction’ (B12). But ‘synthetic’ a priori judgements

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

KANT ’ S ‘ TRANSCENDENTAL DEDUCTION ’



are ‘ampliative’ and ‘add to the concept of the subject a predicate which has not been in any wise thought in it’ (A7=B11). The problem then is to explain how those judgements can ‘relate to objects’ they are about completely independently of all experience and not simply in accord with the principle of non-contradiction. The question of how synthetic a priori knowledge of that kind is possible is for Kant the same as the question of how metaphysics as reliable philosophical knowledge of the world is possible. A successful ‘transcendental deduction’ that answered that question would therefore put metaphysics ‘on the secure path of a science’. The quid juris of the claims of metaphysics would have been answered positively and their ‘legitimacy’ would have been explained. Kant is optimistic about the prospects of this kind of metaphysics because ‘its subject-matter is not the nature of things, which is inexhaustible, but the understanding which passes judgment upon the nature of things’ (A12–13=B26). That subject-matter is ‘nothing other than reason itself and its pure thinking; and to obtain complete knowledge of these, there is no need to go far afield, since I come upon them in my own self ’ (Axiv). Kant believes that fruitful metaphysical conclusions about the nature of reality are to be expected from an investigation of our a priori ways of thinking about it because ‘these a priori possessions of the understanding, since they have not to be sought for without, cannot remain hidden from us, and in all probability are sufficiently small in extent to allow of our apprehending them in their completeness, of judging as to their value or lack of value, and so of rightly appraising them’ (A13=B26–27). So Kant’s ‘deduction’ is to vindicate metaphysical knowledge of objective reality through the proper understanding of ‘these a priori possessions of the understanding’. I will concentrate only on the single central line this ‘deduction’ follows to reach its desired end. It starts with the conditions of thought. Kant first establishes, of each of an identified set of very general concepts, that its employment is essential to the possibility of any thought. For Kant, thought is possible only for those with a capacity for judgment— for thought of something or other’s being so or of a certain thing’s being so-and-so. By reflection on what he regards as the fundamental forms that any possible judgment must take he arrives at an exhaustive set of extremely general concepts that he thinks must be applied to some things for any judgments at all to be possible. If all human thinking must

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



KANT ’ S ‘ TRANSCENDENTAL DEDUCTION ’

conform to those forms of judgment, those concepts can be known to be the fundamental ‘categories’ that must be present in ‘the understanding’ of anyone who can think about anything at all. The most intricate sections of Kant’s overall ‘deduction’ concentrate on the complex relations among a rich set of further conditions that must be fulfilled for any thoughts within even these most general and abstract ‘categories’ of thought to be possible. There is thought only if a capacity for thought is actually put to work on something. Any thought of objects, or any application of concepts to an object to make a judgment, requires at least that objects be present to us in some way. Objects that are present to us in experience obviously must conform to the ‘formal conditions’ of ‘sensibility’—the conditions necessary for our being sensorily aware of something. Otherwise, whatever affected us would be as nothing to us as far as having thoughts about it is concerned. But although whatever we are aware of in experience must conform to the conditions of our being aware of it, that alone does not imply that the objects we are aware of in experience must conform to the conditions necessary for the possibility of thought of an object, for a judgment that such-and-such is so-and-so. It seems possible at first that ‘appearances’ might present themselves in experience so randomly and in such confusion that ‘the understanding’ could make nothing of them, and would find no objects to form thoughts about, and so could not think of anything at all (A90=B123). The heart of Kant’s ‘deduction’ is the attempt to explain how this apparent possibility presents no real difficulty. The possibility of human thought and experience is to be accounted for by showing how and why objects can be given to us in experience only because the forms in accordance with which thoughts must be unified into judgments about something are also the forms within which whatever affects our ‘sensibility’ must be unified to yield experiences of objects. That would bring the conditions of thought and the conditions of experience together as necessary for each other. If there could be no experiences that do not conform to the conditions necessary for the possibility of judgment, and all judgment involves application of the fundamental and therefore a priori concepts of the human understanding, that would explain how those a priori concepts can ‘relate to objects which they yet do not obtain from any experience’. The application of those concepts to some objects of experience is held to

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

KANT ’ S ‘ TRANSCENDENTAL DEDUCTION ’



be a necessary condition of there being any thought or experience of any objects at all, so the ‘objective reality’ of those concepts in general is secured simply by the fulfillment of that necessary condition, not through having to have experience of those objects themselves. For Kant, thinkers who apply such a priori concepts to something, and not solely ‘in accordance with the principle of contradiction’, would thereby know a priori that certain things are so in the world they experience. Showing how and why that must be so would explain how synthetic or ‘ampliative’ metaphysical knowledge of the world is possible completely a priori. The quid juris of the claims of any metaphysical investigation pursued according to that plan would be answered positively and the ‘legitimacy’ of its results accounted for. The success of this ambitious philosophical enterprise obviously depends on the success of the arguments for the necessity of the fundamental ‘categories’ and of the many arguments throughout the Analytic of Principles by which Kant reaches the conclusion that this or that more specific way of thinking is a necessary condition of the possibility of thought and experience in general. Interpretation and criticism of those difficult arguments have kept philosophers busy ever since the Critique of Pure Reason was first published, and continues to this day. Our concern here is not with the validity of those arguments but only with the special character of the Kantian ‘transcendental’ project itself and the conditions of its success. That remains a formidable problem even if it is conceded that Kant’s arguments for the necessity of the conditions he identifies are all correct. The arguments are meant to establish a number of interrelated conclusions that could be summed up loosely and in general terms as follows. They start from the idea that for experiences and thoughts to belong to a conscious subject that subject must be able to think of those experiences and thoughts as belonging to him. That requires a capacity on the thinker’s part to think of those experiences and thoughts as belonging to him but distinct from him, and also to think of whatever he takes those experiences to be experiences of. That requires that thinkers be capable of thoughts about something other than themselves and their experiences, and so to that extent that they are capable of thoughts about something objective. Some of the objects of any thinker’s thoughts or experiences must be thought of as existing or as being as they are independently of their being thought about or perceived by anyone.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



KANT ’ S ‘ TRANSCENDENTAL DEDUCTION ’

This requires in turn that some of those objects be thought of as enduring independently in space and time and that states of affairs involving those objects be thought of as occurring in an order not necessarily the same as the order of thinkers’ perceptions and thoughts about them. That independent order must be thought of as containing enduring objects capable of entering into events which are connected with other events in accordance with causal laws. I think this loose, informal list of a few of the very general ‘results’ Kant takes himself to have reached about the necessary conditions of thought and experience will suffice for our purposes. To understand the general overall structure of the ‘deduction’ we must understand not only those ‘results’ themselves but also the precise role they are meant to play in the ‘deduction’ even if they are correct. It is important to observe that the conclusions of Kant’s arguments identify certain propositions or ways of thinking that all thinkers must understand and accept as a condition of their having any thoughts or experiences of anything at all. The idea is that whoever can think at all must think that each of the very general propositions Kant identifies describes some aspect of the way things are in the world, at least in general. That is, for Kant, a condition of thought. Even if Kant’s arguments for these conclusions are sound and the ‘results’ arrived at in the Analytic of Principles are taken as established, we can ask about the distinctive epistemological and metaphysical status those conclusions would thereby have been shown to have, and how their having that status serves Kant’s announced purposes. The arguments identify a number of propositions that all thinkers must accept or judge to be so. That certainly gives those propositions a special, distinctive status in any thinker’s conception of the world. But Kant’s ‘deduction’ seeks to establish and explain something stronger than that. It claims not only that certain propositions must be accepted or judged to be so by anyone who can think anything, but that what is said to be so by propositions that have that distinctive status is something all thinkers know a priori to be true of the world they live in. That is a strong epistemological conclusion about the knowledge human beings must have of the world; they know a priori that those propositions that Kant shows have that distinctive status are true in the world. And that implies corresponding metaphysical conclusions about what the world those thinkers know about must be like. What

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

KANT ’ S ‘ TRANSCENDENTAL DEDUCTION ’



is remarkable is that Kant appears to reach those conclusions about the world and our knowledge of it from nothing more than the necessity with which all thinkers must think of and experience things in certain ways. This calls for closer attention because the fact that someone thinks or judges that such-and-such is so, even if his thinking it is required for his being able to think of anything at all, does not on its own seem to imply that what that thinker thereby thinks is true. Nor does it imply that what that thinker thereby thinks is something he knows to be true either. Everyone’s having to think it is one thing; its being true, or its being known to be true, is something else. Kant holds not only that the distinctive propositions he has identified are known by all thinkers to be true, but that they are known a priori to be true. That is what the ‘transcendental deduction’ is meant to establish and explain. But it is not established, and certainly not explained, solely by the fact (even if it is a fact) that thinkers must think all such propositions are true if they can think and experience anything at all. We saw that for Kant necessity is a ‘sure criterion’ of the a priori: ‘if we have a proposition which in being thought is thought as necessary, it is an a priori judgment’ (B3). But if Kant were relying on that principle to support the view that the propositions he claims all thinkers must accept are known a priori, it would mean that he thinks those propositions, in being thought, are thought as necessary. If that means that those propositions are thought to hold necessarily in the sense of there being no possibility of their being or having been false, it would mean that Kant thinks that the distinctive general propositions he identifies—e.g., ‘Objects and their properties endure in space and time independently of whether they are perceived or thought about’ or ‘Objects enter into events that are connected by causal laws with other objects and events’— hold necessarily and could not possibly have been false under any circumstances. Since Kant holds that those distinctive general propositions must be accepted by anyone who can think, perhaps he means only that those propositions are ‘necessary’ in the sense that any thinker must accept them or judge them to be true. They would be ‘necessary’ in the sense of being necessary-for-thought; they must be held to be true by anyone who can think. But understanding their ‘necessity’ in that way does not imply that in being thought, those propositions are themselves thought to hold

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



KANT ’ S ‘ TRANSCENDENTAL DEDUCTION ’

with necessity. So that way of understanding their ‘necessity’ would not in itself support the idea that those propositions are even known to be true, let alone that they are known a priori. If we do know many general propositions of that kind to be true, as I think we do, and if Kant is right about the distinctive status his arguments show those propositions to have in our conception of the world, it still does not follow that we know them a priori. Kant explicitly argues for something he thinks does hold with necessity, and so could not possibly be or have been false. He claims in each case that if there is any thought and experience at all, then thinkers accept those distinctive propositions that he argues are required for the possibility of thought and experience. If Kant is right about that, and if he is also right that necessity is a ‘sure criterion’ of the a priori, then those claims he makes about the necessary conditions of thought are something Kant knows a priori. All the rest of us would also be able to know a priori that those propositions Kant identifies must be accepted by anyone who thinks. But what is thereby known or knowable a priori is a conditional claim. It does not imply that any of those propositions that must be accepted by thinkers are themselves known a priori, even if they are known. Kant thinks those propositions that must be accepted by anyone who can think involve the application of concepts that could not be acquired simply by finding instances of them in experience; they involve what he calls a priori concepts. That is one mark of the distinctive status of those propositions. But not every proposition involving the application of an a priori concept enjoys the distinctive status Kant identifies. Nor is every proposition involving application of an a priori concept knowable or known a priori. For instance, to know that an object of a certain kind caused such-and-such changes in an object of another kind at a certain place and time one must rely on observation and experience. Such truths about the world are knowable only a posteriori, not a priori, although what is thereby known involves application of a priori concepts like ‘enduring object’ and ‘causation’. Kant’s arguments establish only that a capacity for the application of such concepts to objects of experience in general is required of every thinker. He takes that to show that they are a priori concepts. But the ‘transcendental deduction’ aspires to more than that. It seeks to establish that all thinkers know a priori that those a priori concepts truly apply to objects in the world

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

KANT ’ S ‘ TRANSCENDENTAL DEDUCTION ’



they think about and experience, and then to explain how that a priori knowledge is possible. Knowing such completely general truths about the world a priori is compatible with knowing a posteriori that those fundamental concepts apply in particular to this or that item or kind of item in the world as things are. Kant’s ‘deduction’ does not aim to show that all knowledge expressed in terms of those fundamental concepts is a priori. And the fact that relatively specific propositions about enduring objects or causal connections, for instance, can be known only a posteriori might suggest (as it has to many philosophers) that we could come to know that those concepts apply to things in the world in general simply by generalizing from what we can find in particular cases. After all, if we have found some objects standing in causal relations with others, we know that, in general, there are objects standing in causal relations in the world. But Kant’s arguments rule out the possibility of knowing such completely general truths only a posteriori. Let us suppose that it is a necessary condition of thinking of and experiencing anything at all that one accepts and makes sense of one’s experience in general in terms of enduring objects and causal connections, for instance. Then one could not come to know by experience that more specific versions of those concepts apply to this or that kind of object or cause without already possessing the more general concepts of enduring object and causal connection. For Kant that necessity shows that they are a priori concepts. ‘There are enduring objects and causal connections in the world’ is a more general statement than any statement about particular kinds of objects or causes, but one could not come to know that more general statement a posteriori by generalizing from the more particular instances one finds to be true in one’s experience. What this shows, strictly speaking, is that the completely general propositions Kant argues must be accepted by anyone who can think and experience anything could not be known a posteriori. That is an important step in the overall strategy. But it does not alone achieve the declared goal of Kant’s ‘transcendental deduction’: to establish that such general structural propositions are known a priori to be true of the world, and to explain how that synthetic a priori knowledge is possible. To succeed in that ambitious task it must be shown not only, as Kant has argued, that those distinctive general propositions must be understood and accepted by anyone who can think and experience

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



KANT ’ S ‘ TRANSCENDENTAL DEDUCTION ’

anything at all, and not only that for that reason they cannot be known a posteriori. It must be shown that those distinctive general propositions are known a priori by all thinkers and perceivers to be true of the world they live in. Holding that anyone who can think at all must employ certain distinctive (a priori) concepts, and that the conditions of thought expressed in those concepts apply as well to the conditions of all possible objects of experience, puts Kant in position to explain how a posteriori knowledge of the world is possible in general. If those very general a priori concepts are true of things in the world, and if they hold as well of all objects of experience, perceivers could often know a posteriori that more determinate specifications of those general concepts apply to particular objects they are aware of in their experience. And since those conditions hold as well for the possibility of experiences of objects, empirical knowledge of those features of independent ‘outer objects’ could be gained by ‘immediate perception’ of them. That would give us direct knowledge of ‘a reality which does not permit of being inferred, but is immediately perceived’ (A371). Anything less than that as an account of our relation to the things around us would leave us vulnerable, in general, to what Kant calls ‘problematic idealism’, which he finds, for instance, in Descartes. On that view, the existence of external objects can be at best only inferred from one’s ‘inner perceptions’, and so remains always subject to doubt or uncertainty. Kant recognizes that all such—as he calls them, ‘idealist’—conceptions of experience and of perceptual knowledge of an independent world must be seen through and rejected. In his preface to the second edition of the Critique he draws special attention to the question. However harmless idealism may be considered in respect of the essential aims of metaphysics (though, in fact, it is not thus harmless), it still remains a scandal to philosophy and to human reason in general that the existence of things outside us (from which we derive the whole material of knowledge, even for our inner sense) must be accepted merely on faith, and that if anyone thinks good to doubt their existence, we are unable to counter his doubts by any satisfactory proof. (Bxl)

For Kant the way to ‘counter the doubts’ of those who find the existence of ‘things outside us’ doubtful or epistemically problematic in general

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

KANT ’ S ‘ TRANSCENDENTAL DEDUCTION ’



would not be to try to ‘prove’, somehow, that the ‘outer objects’ we perceive really do exist after all. How could such a thing be proved? And would whatever could be proved really silence the doubts of those who find the existence of such objects problematic? The strategy of Kant’s ‘transcendental deduction’ is rather to: show that we have experience, and not merely imagination of outer things; and that this, it would seem, cannot be achieved save through proof that even our inner experience, which for Descartes is indubitable, is possible only on the assumption of outer experience. (B275)

Such a ‘proof ’ would completely turn the tables on ‘problematic idealism’. On that view, ‘the only immediate experience is inner experience’, so ‘outer things’ can at best be known only by inference from what is experienced. But the ‘proof ’ embodied in Kant’s ‘deduction’ is meant to show that ‘outer experience is really immediate, and that only by means of it is inner experience . . . possible’ (B276–7). ‘Outer objects’ of experience as Kant’s ‘deduction’ explains them are required for the possibility of any experience of anything at all. What we experience directly are objects, causal relations between objects, etc., in the world as it is independently of our perceiving them. We could not even have the kind of ‘inner experience’ ‘problematic idealism’ takes for granted without being capable of ‘immediate perception’ of ‘outer objects’ that are as they are independently of us in that way. For knowing a posteriori how things are in the world there would be no need, in general, to rely on inference from what we experience to something else. With respect to a posteriori knowledge of the world around us, then, our position in the world must be understood as what Kant calls ‘empirical realism’. It is ‘realism’ in the metaphysical sense that objects and events and their properties exist and are as they are independently of us and our perceptions of them. And it is ‘realism’ in the epistemological sense that our knowledge of those ‘outer’ or independent objects and their properties is possible by direct perception of them. Perceptual experience is not restricted to what ‘problematic idealism’ calls ‘mere appearances’ or ‘inner experiences’ of objects, all of which depend for their existence on being perceived. Kant’s ‘empirical realism’ says we can know by experience that we are in direct perceptual and cognitive touch with the independent objects

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



KANT ’ S ‘ TRANSCENDENTAL DEDUCTION ’

around us, with no need to infer their existence from something we are said to be even more immediately aware of. This, of course, is meant to hold only in general. Whether this or that particular experience is an experience of how things independently are is something that can be doubted on this or that occasion, and so enquired into and settled empirically. But the completely general ‘proof ’ of ‘empirical realism’ would suffice to counter all ‘problematic’ doubts about the possibility of a posteriori knowledge of an independent world in general. The possibility of a posteriori knowledge of the world of nature would be successfully accounted for in this way only if it is true, as Kant claims, that all thinkers who employ the required a priori concepts know a priori that those concepts are true of objects in the world they live in. Everyone’s knowing a priori that that is the way the world is, at least in general, is what would account for the possibility of their knowing by experience that relatively specific determinations of those same general a priori concepts apply to certain objects within their experience. But has that guarantee of the truth of ‘empirical realism’ been secured? Is it true that everyone who employs a priori concepts knows a priori that those concepts have ‘objective reality’; that they truly apply to objects in the world? Has that been established by Kant’s ‘deduction’ as developed so far? And if so, has it also been explained how that a priori knowledge is possible, as a successful ‘deduction’ requires? Here we come at last to the sense in which Kant’s ‘deduction’ is to be understood as ‘transcendental’. I entitle transcendental all knowledge which is occupied not so much with objects as with the mode of our knowledge of objects in so far as this mode of knowledge is to be possible a priori. (A11–12=B25)

To have ‘transcendental’ knowledge, or to engage in a ‘transcendental’ inquiry, is to know or inquire into how a priori knowledge of certain kinds is possible: the conditions that must be fulfilled for there to be knowledge of that kind. Any knowledge or understanding yielded by such an inquiry must itself therefore be a priori, concerned as it is with the conditions and the possibility of a priori knowledge. The point of such a ‘transcendental’ inquiry or understanding is not simply to gain knowledge of the world but to achieve what turns out to be a distinctively philosophical understanding of how human knowledge of the world is

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

KANT ’ S ‘ TRANSCENDENTAL DEDUCTION ’



possible in general. ‘The distinction between the transcendental and the empirical belongs therefore only to the critique of knowledge; it does not concern the relation of that knowledge to its objects’ (A57=B81). We have seen that for Kant ‘realism’ is the only satisfactory explanation or ‘critique’ of a posteriori knowledge of things in the world around us. We can determine by empirical investigation that most of the objects we perceive exist and are as they are independently of us and our perceptions of them, and we can gain immediate, underivative knowledge of them and their properties by perceiving them. By contrast, any form of ‘idealism’ concerning this knowledge of the world would leave us at best only in indirect perceptual and cognitive touch with objects in the independent world, and so always in ‘problematic’ or ‘skeptical’ doubt, with no reliable knowledge of the way things are. But if we ask now in this same ‘transcendental’ spirit how the a priori knowledge of the general structure of the world that makes this ‘empirical realism’ possible is itself to be explained, Kant finds an apparently parallel ‘realism’ at this ‘transcendental’ level completely unacceptable. To accept, as part of a ‘transcendental’ explanation of our a priori knowledge, that the objects we perceive and know about exist and are as they are independently of us and our capacities for experiences of them, would leave that allegedly a priori knowledge completely inexplicable to us. Kant’s original problem was how those concepts that are marked out for a priori employment ‘can relate to objects which they yet do not obtain from any experience’ (A95=B117). Here Kant finds that if the objects we could have knowledge of were all completely independent of us and our capacities for knowing them, we could have no a priori concepts that we had any reason to suppose applied to those objects, and so no a priori knowledge of those objects at all. This would leave us with no explanation of the a priori knowledge Kant thinks every thinker has of the ‘objective reality’ of those a priori concepts he must employ if he can think at all. And that would mean that the ‘empirical realism’ just now thought to have been vindicated by that a priori knowledge would be left unexplained. To accept ‘realism’ as a ‘transcendental’ explanation of the possibility of a priori knowledge would leave both a priori and a posteriori knowledge of the world unintelligible to us. For Kant the only acceptable ‘transcendental’ doctrine—’the only refuge left open’—is therefore idealism. ‘Transcendental idealism’ is the

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



KANT ’ S ‘ TRANSCENDENTAL DEDUCTION ’

only way of avoiding ‘problematic’ or ‘sceptical’ idealism about the world of nature, and so the only way of guaranteeing the truth of ‘empirical realism’. We can be sure in general of reliable a posteriori knowledge of a world of nature that is ‘empirically’ independent of us because—as ‘transcendental idealism’ implies—the objects we are aware of in that world are one and all ‘appearances’, speaking ‘transcendentally’. They are not objects that are as they are independently of their fulfilling the conditions of being thought of and experienced by us. They are not, in the ‘transcendental’ sense, independent of their fulfillment of the necessary conditions of their being thought of and experienced. Kant’s ‘transcendental idealism’ explains the truth of ‘empirical realism’ only because it implies that the necessary conditions of thought and experience are ‘supplied by us’ to the objects we think of and experience. We thinkers ‘contribute’ to what we think of and experience those very conditions without which we could not think of or experience anything at all. It is because we ‘supply’ or ‘contribute’ those conditions to whatever we think of or experience that we cannot find anywhere in our thought or experience anything that fails to conform to them. This same understanding of idealism as our ‘contributing’ to the objects we know about is also involved in accounting for our knowing a priori that the fundamental categories of thought apply to anything we can think of or experience. That a priori knowledge must be accounted for if the truth of ‘empirical realism’ is to be explained. So, the ‘transcendental idealism’ that explains both a priori and a posteriori knowledge really is a robust form of idealism. Whatever we think about or experience is not something that is as it is independently of its fulfilling the conditions of our thinking of or experiencing it. What we think about or experience is dependent on something about us; it is in that sense not fully independent of ‘us’ and our ‘contribution’ to it. For Kant this is ‘the only refuge left open’ for explaining ‘transcendentally’ our knowing what we know a priori and our knowing what we know a posteriori. I think Kant must understand his ‘transcendental idealism’ in some such way if he is to regard his ‘deduction’ as successful and as having explained the possibility of our a priori and our a posteriori knowledge ‘transcendentally’. He does not appear to envisage or even find intelligible any form of idealism in which we do not ‘contribute’ something to our knowing or experiencing the things we do. He rejects ‘transcendental realism’ from the outset as even a candidate for a possible explanation of

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

KANT ’ S ‘ TRANSCENDENTAL DEDUCTION ’



our a priori knowledge of objects that are understood as independent in every way of us and our ‘contribution’ to them. He appears to find the possibility of a priori knowledge of such objects or states of affairs as beyond explanation. For him a priori concepts can be understood to ‘relate to objects which they yet do not obtain from experience’ only because we ‘supply’ the conditions that make possible the objects which we take those concepts to apply to. In calling ‘transcendental idealism’ a robust form of idealism I mean it is directly opposed to ‘realism’. It implies that we are never in cognitive or experiential contact with anything that is as it is independently of its fulfilling the conditions that ‘we’ ‘supply’ for being in cognitive or experiential contact with it. The tables and trees and other apparently ‘outer’ objects we are aware of in experience are, ‘transcendentally speaking’, only ‘appearances’ which depend for their nature on the ‘contribution’ we perceivers make to our being aware of them. As a ‘critique’ of knowledge, then, or as a philosophical understanding of how our knowledge is possible, the objects we experience cannot be understood as being as they are independently of whether we experience them or not. That same understanding of ‘idealism’ also means that the objects we take our purely a priori concepts to apply to are not objects that are as they are independently of whether we think of them in those a priori ways either. Kant finds no explanation of a priori knowledge possible along any such strictly ‘realist’ lines. He thinks the complete independence of such objects would leave no way to account for what he regards as the ‘objective reality’ of our a priori concepts. So if Kant’s ‘deduction’ is to explain the possibility of a priori knowledge of the world—as he requires—it must therefore see the necessary conditions of the possibility of thought and experience as something ‘supplied’ or ‘contributed’ by us. The explanation must lie in those ‘a priori possessions of the understanding’ which ‘have not to be sought for without’, and ‘cannot remain hidden from us’ because ‘I come upon them in my own self ’. This account of the conditions of success of Kant’s ‘deduction’ leaves two lingering questions worth pondering. (1) Has Kant accounted ‘transcendentally’ for a priori knowledge of the ‘objective reality’ of our a priori concepts—our knowing a priori that such concepts apply truly to things in the objective world? Or has he explained only the impossibility of our ever

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



KANT ’ S ‘ TRANSCENDENTAL DEDUCTION ’

finding anything within our thought or experience to which those concepts do not apply? (2) If the necessary conditions of thought and experience of things in the world are ‘supplied’ or ‘contributed’ by ‘us’, how is that itself known? If we have found those conditions to be necessary for the possibility of all thought and experience, as Kant claims, can we hope to explain their holding necessarily by appeal to something which does not appear to be necessary at all? To insist that we know a priori that those conditions must be ‘supplied’ or ‘contributed’ by us—that is ‘the only refuge left open’—because otherwise we would have no ‘transcendental’ explanation of a priori knowledge of the world, seems to leave us at best with an uncomfortable disjunction: either we do ‘supply’ the conditions that make thought and experience of the world possible or we have no satisfactory explanation of a priori knowledge at all.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

13 Unmasking and Dispositionalism Reply to Mark Johnston

Mark Johnston¹ acknowledges the potential significance of investigating the nature and prospects of metaphysical ambition as I try to do with respect to the colours of things in The Quest for Reality. But he is not impressed with the actual achievement. He thinks I spend too much time rejecting views that nobody nowadays, or perhaps ever, would take seriously, and too little time—and in some cases no time at all—opposing views contrary to mine that almost everybody who is anybody nowadays accepts. He raises some objections that I certainly have not answered or even explicitly forestalled in the book. The most I can do here is to indicate the directions in which I think adequate replies might be found, if they can be found at all.

Unmasking My resistance to the metaphysical unmasking of colour perceptions and colour beliefs as not capturing what is so in fully objective reality rests on the claim that without holding some beliefs about the colours of objects one could not acknowledge the presence in the world of perceptions of colour and beliefs about the colours of things. This bears on the project because a would-be unmasker must acknowledge such things in order to have something to unmask. But if he acknowledges them and that ¹ Mark Johnston, “Subjectivism and Unmasking”, Philosophy and Phenomenological Research, Vol 69, No. 1 (July 2004) (page numbers alone in parentheses in the text refer to this essay).

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



UNMASKING AND DISPOSITIONALISM

condition holds, he could reach the apparently unmasking conclusion he asserts only by believing something incompatible with it. I have tried to make a case for that central claim, and so for the impossibility of consistent unmasking of the colours of things. Johnston not only thinks I have not succeeded; he sees an “error in this whole approach” (192). He thinks the claim is too demanding, and that a more “holistic” conception of the conditions of understanding a family of uses of interrelated terms would be more “flexible” or more “forgiving”. He thinks it would allow us to understand our colour vocabulary as we do without accepting as true any sentences predicating colour terms of physical objects. Taking all the “platitudes” we all accept about colours, he thinks “we can divide the package into beliefs about the colours, whether or not external things have them, and beliefs attributing colours to external things” (196), and that assenting to the first part alone would be sufficient for grasping the contents of the other part, even if we denied all of them. That is the main issue between us on the question of unmasking. I think it is not easy to settle it decisively. Johnston gives the impression of thinking it is fairly straightforward; with “holism” and the “Ramsey/ Lewis” view that “the colors are the best satisfiers of . . . [all] those very plausible beliefs that are ostensibly about the colors” (188), an answer is at hand. But a disdain for “holism” cannot be the source of my “error”. My central claim of the indispensability of beliefs about the colours of things for understanding colour terms accepts the idea that colour terms, when not used to predicate colour of an object, mean what they do because they are “holistically” tied to the meanings they have in such predicative applications. They would not mean what they do in their non-predicational uses if that were not so. And it also accepts a connection running in the opposite direction: colour terms would not mean what they do in predicational application to physical objects if they did not mean what they do when not predicated of objects. It is a truly “holistic” package in which our understanding of each part of the package is inseparable from that of the other. Perhaps my “error” lies not so much in insufficient regard for “holism” as in my not appreciating the “flexibility” implicit in the “Ramsey/Lewis view” of colours. On that view, colours turn out to be whatever are “the best satisfiers” of all the sentences we accept involving colour terms. And, as in any theoretical enterprise, there is room for manoeuvre. Additional

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

UNMASKING AND DISPOSITIONALISM



argument, even what Johnston describes as “arcane philosophical considerations” (189), could recommend assigning a best fit between the sentences and the “satisfiers” that accords truth to most sentences with colour terms while denying it to all those sentences in which they are predicated of objects. I take the “Ramsey/Lewis view” of colours to be an ontological thesis. It is meant to tell us what the colours are, or which things in the world are the colours. It does so by arriving at the most satisfactory account of what entities, or what sorts of entities, are needed as referents of colour terms to make true a majority or at least a sufficient number of all the sentences we accept containing such terms. Johnston describes it as an “account of the content of our color beliefs” (188), and it is an account of the “contents” of those beliefs only in the sense of specifying what entities are the referents of their terms. It is in that sense that it is meant to tell us what entities the colours that we think and speak about really are. They will be certain kinds of properties, presumably ultimately certain sets. The metaphysical question the theory answers is which properties, or which sets. My interest is not in the metaphysics of colour in that sense. My question about unmasking is how or even whether one can reach a metaphysical conclusion to the effect that what we all believe when we believe that physical objects are coloured is never true. That requires understanding what we all believe when we believe that objects are coloured and acknowledging that we all believe it. And, on the question of what it takes to do that, I find a theory of the “Ramsey/Lewis” type in itself little direct help, for two different reasons. I am interested in our use of colour words as predicates, especially as predicated of physical objects. And I don’t think predicates designate or refer to anything in such uses; they function only as predicates. Of course, objects can always be assigned as referents of predicates for various purposes, but an assignment of objects alone does not answer the question what a person is doing or thinking in using a colour term predicatively, or what it takes for him to be able to do that. It does not help to say that in predicating colour of an object a person stands in a certain relation to a specified object designated by the predicate. To go on to say that the person also thinks that that designated object “belongs to” or “is a property of ” a physical object he is also related to leaves us again with the question of how such a predicational thought is possible.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



UNMASKING AND DISPOSITIONALISM

To answer that it is in virtue of the person’s standing in relation to a further object—the relation of belonging to or the relation of being a property of—is the same unhelpful answer one level up. This points to a large and deep issue which might well be the source of some of the main differences between Johnston and me. But putting that issue aside for now, there is another way in which the mere availability of a theory of the “Ramsey/Lewis” type does not help settle the question between us. A would-be unmasker must acknowledge that we perceive colours and believe that objects are coloured. And that acknowledgement in turn has its conditions. One condition is that he must understand colour predicates as they are understood by the perceivers and believers whose beliefs he would unmask. Whether that condition in turn includes the requirement that he must accordingly hold some beliefs about the colours of objects himself is the question on which Johnston and I differ. A “Ramsey/Lewis” theory of colours does not give a satisfactory answer to that question simply by assigning objects to colour terms in all the sentences we accept so that sentences predicating colour terms of physical objects all turn out to be false. Perhaps some such assignment of objects could be made. Johnston says that if we separated off all those sentences in which colour terms are predicated of objects “there would still be a massive set of platitudes drawn from [all other accepted colour sentences] that would serve to attach a sense to our thought and talk about the colors” (188). But the question is not whether that would leave those terms with some sense or other. The difficult question is whether the sense those words would have been given by that assignment would serve to express what we now express in sentences predicating colour terms of physical objects. Would the thought and talk that remained be about colours as we now understand them, as Johnston says they would be? The question is not easy to settle decisively because it requires judgment as to whether the speakers we are considering understand certain expressions in a certain way or not. And that is always a complicated question. It is a question, ultimately, of speakers’ competence: exactly what abilities or capacities they exercise in using those words in all the ways they do. That is what imposes constraints on any particular application of a “Ramsey/Lewis” theory of colour terms to those speakers. The “best satisfiers” for the colour terms in all those speakers’ beliefs will be those the assignment of which to their utterances most adequately

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

UNMASKING AND DISPOSITIONALISM



fits their competence. That is a matter of what best accounts for speakers’ accepting or asserting all the things they do, their being prepared to accept or assert certain things when they perceive what they do, their taking the colour they perceive to be related in certain ways to the colour they attribute to an object in the belief that it is coloured, and so on and so on. To speak as we do in all those ways is a truly massive achievement. It cannot be described simply by listing the sentences speakers accept, even a great many sentences. A theory of colours of the “Ramsey/Lewis” type in itself provides no shortcut to getting that description of our competence right, or as right as can be expected. Why then is Johnston optimistic that there is a “Ramsey/Lewis” assignment of objects that is “best” in the sense of being most adequate to our complex competence with colour terms while leaving all sentences predicating colours of physical objects false? I think it is because he is convinced that “massive error about the colours of external things could be made an intelligible error by appreciating [an unmasker’s] arcane philosophical argument” (189). The success of the philosophical argument he has in mind would provide a way of explaining and so making intelligible the massive error of believing in the colours of objects. That would show that I am wrong to think that a would-be unmasker faces a fatal dilemma. So Johnston’s optimism rests on something in addition to a theory of colours of the “Ramsey/Lewis” type. Its support can come now only from a plausible account of how a would-be unmasker could actually proceed while avoiding the dilemma. It is of course too much to expect Johnston to work out a thorough and convincing account of a possible unmasking strategy in just a few pages. But I do not find that the outline he sketches here gives any reasons for expecting success in the project beyond those I have already considered at some length. He says that what he calls a “de facto” unmasker could allow that it is possible that objects are coloured while holding that to believe that they actually are coloured would “violate some significant principle of parsimony”; it would be an “idle hypothesis” (192). He does not tell us here what such a “significant principle of parsimony” would actually say, or how our belief that objects are coloured would violate it. Nor does he explain why anyone should accept such a “principle” if it has that implication. Its rationale appears to be connected with the requirements of explanation. He says a world of coloured objects would be

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



UNMASKING AND DISPOSITIONALISM

an “idle hypothesis” because “the generation of our experience would proceed as if it were not so” (192). What Johnston appears to have in mind here is just the kind of “unmasking explanation” of our perceptions and beliefs about the colours of things that I have tried to assess. It brings to centre stage the conditions of anyone’s recognizing what “our experience” of colours and of the colours of objects and our beliefs about them actually is. That is what a would-be unmasker must have some conception of if he is going to explain it all without himself believing that any objects are coloured. And his being a potential unmasker puts constraints on what he can find that “experience” to be. If it is taken to include the experience of seeing that grass is green or that tomatoes are red, for instance—experiences I think we all have—then “our experience” could not be accounted for on the assumption that no objects are coloured. So the prospects of a successful unmasking explanation of that “experience” obviously depend on what an unmasker must believe or be prepared to accept in order to recognize our experience for what it is. Johnston says repeatedly that his unmasker could get by with much less in the way of positive assumptions than my “demanding” constraint on attribution says are needed. But he does not suggest any specific ways in which he thinks an unmasker would actually succeed while bypassing that constraint. He focusses on efforts to draw from one’s own “ostensible perceptions” what is needed in order to understand the contents of the colour perceptions and beliefs an unmasker must be able to attribute to others. He says “a variety of different kinds of beliefs or judgements might set the stage” (192) for abstracting the appropriate content from the “ostensive definition”. But he does not specify further what those beliefs or judgments might be. He says that in “ostensible perceptions of yellow lemons”, for example, a person might be aware of a “coloured patch” or of “a qualitative property ostensibly presented as a property had by external surfaces” (190). But he does not explain what a person must think or perhaps believe, or what capacities he must have, even to find a “qualitative” property “presented” to him as “ostensibly” a property of an external surface of an object. What does it take for a perceiver to have such an experience? It presumably involves more than simply seeing a coloured patch; that could happen to a creature who has no conception of anything and so no capacity to have properties presented to it as “ostensibly” properties of physical objects.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

UNMASKING AND DISPOSITIONALISM



Johnston does say that what he calls a “de facto” unmasker could retain “a conception of what it would be for things to be coloured” (192) despite his unmasking attitude towards that idea. But he cannot give good reasons in support of that claim without saying something about what a person’s having “a conception” of such-and-such amounts to, and how it is connected with what the person is prepared to accept under certain conditions, in particular under conditions in which he takes himself to perceive an instance of the very thing he has a conception of. Until more is said on such questions there is simply no telling whether the would-be unmasker has a conception of what it would be for objects to be coloured or not. Johnston does also mention a possible “de jure” unmasker who holds that there simply is “no satisfiable conception of how it is that material objects . . . could be coloured” (192). For such a theorist, “the colours are necessarily uninstantiated” (193). But even such an unmasker (if that is what he would be) “can be said to understand what is required for a surface to be coloured” (193). It would be just like someone’s understanding what is required for something to be round and square: it is required that the thing be round and that it be square. But if that is enough for understanding, we could equally well understand how someone could have a conception of what is required for a stone to think; it is required that it be a stone and that it thinks. I don’t see any encouragement to the prospect of successful unmasking in the possibility of a would-be unmasker’s having a conception in that sense of what is required for an object to be coloured. Johnston illustrates at some length a general argument to show that “if we accept that demanding constraint [which I think holds for attribution of colour perceptions and beliefs] we get the wrong interpretive results” (193); that the constraint “sets the standards for the attribution of color belief far too high” (197). He thinks my requirement implies that those who deny that objects are coloured, or who even vacillate about it while apparently saying a great many things about colours, are not really talking about colours after all. It would even mean that no one has ever believed the unmasking thesis; they could not really have been talking about colours in apparently asserting it! Johnston devotes the largest single portion of his remarks to a description of the speech and beliefs of various unmaskers or doubters who speak like this. In reply I can be brief. The absurd consequences about

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



UNMASKING AND DISPOSITIONALISM

interpretation do not follow from the requirement on attribution. It is not true, as Johnston claims, that if we accept the requirement “we should have to interpret [unmaskers] as not really thinking or talking about the colours” (194). The requirement on attribution does not demand that we put some particular interpretation on any speaker’s remarks. It does not imply that no one can deny that physical objects are coloured. It implies only that no one can consistently deny it by the unmasking route; one could assert such a conclusion on that basis only by also believing something incompatible with it. That does not mean that if someone says “Physical objects are not coloured” he is not saying what he seems to be saying, or he does not believe it, or we must interpret him as not saying or not believing it. It means only that he is not fully consistent in everything he says or believes. No one can consistently say “I am not now speaking”. But that does not mean that someone who says it is not really saying what he seems to be saying or that we are required to interpret him as not really saying it. In The Quest for Reality, I have tried to bring out how and why I think pursuing questions about the conditions of perception and belief about the colours of things, and so about the conditions of attributing them to others, will be found to support the claim that beliefs about the colours of some things are required for such attributions. Johnston does not offer competing answers to those questions. His opposition to that central claim is not based on a specific account of how unmasking can succeed without fulfilling the requirement. He argues only on general grounds that the requirement is too demanding and simply cannot be right. But the objection that it implies “the wrong interpretive results” does not work. Nor does general adherence to a theory of colours of the “Ramsey/ Lewis” type in itself support rejection of the requirement. And the additional “arcane philosophical argument” he has in mind would show that there is no such requirement only by showing that the colours of objects can in fact be successfully unmasked without fulfilling the requirement. Here we reach a kind of impasse at this level of generality. The dilemma I identify, if it is real, stands as an obstacle to the success of any consistent unmasking strategy. It cannot be overcome or shown to be unreal simply by declaring that our beliefs in the colours of things can indeed be unmasked as not capturing the way things are in reality.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

UNMASKING AND DISPOSITIONALISM



Some specific indication is needed of how that is to be done. There is no alternative now to getting down to detail and trying to show how we can think about the colours of things without having any beliefs about the colours of any objects. That is presumably one of the rewards to be expected from any investigation of the metaphysical enterprise: a better understanding of the ways we actually think about the world.

Dispositionalism Johnston might be right that my efforts to expound what I see as three essential conditions of success for any subjectivist dispositional theory of an object’s colour are carried out too exclusively in the light of formulations of theories of that type that many people no longer believe. It was not always so. Nor was it borne out by countless discussions of dispositionalism with philosophers over the years in which the book was being written. In any case, I think it is essential to stress that any successful dispositional view must imply (1) that the colours of objects, so understood, depend on the kinds of perceptual responses the objects are disposed to produce in perceiving subjects, (2) that the relation between an object’s having a certain specified disposition and its being a certain colour holds with necessity, and (3) the perceptual responses which serve to define the particular disposition with which an object’s having a certain colour is to be identified are specifiable independently of presupposing that any objects are coloured in a non-dispositional or non-subjectivist way. Good reasons must then be found for identifying having a colour with having a disposition that fulfills all those conditions. In illustrating the requirement that a dispositional theory must fulfill all three conditions, I consider connections between having a certain colour and having a certain disposition that are expressed in the form of certain biconditionals. The counter-examples I present to the various dispositional proposals I consider are all directed only to the question of that necessary condition of success. They turn on a candidate disposition’s failure either to meet conditions (1) and (3) if it meets (2), or, if it meets (1) and (3), its failure to meet (2). I do not believe that any of those arguments depend on the assumption, which Johnston says would be wrong, that an object’s having a certain disposition is to be identified with the associated biconditional’s being necessarily true.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



UNMASKING AND DISPOSITIONALISM

But I am embarrassed to find what is there on the page is much more obscure on that particular point than I thought it was when I wrote it. Johnston’s rejection of that assumption is based on what he takes to be successful dispositional analyses which do not have that implication. But in successful dispositional analyses of a certain property it is presumably already accepted that a necessary connection holds between an object’s having that property (e.g., fragility) and its being disposed to behave in a certain way; for Johnston they are one and the same property. My question is whether we can find any disposition an object has to produce perceptions of certain kinds with which its having a certain colour is necessarily connected in a way that supports subjectivist dispositionalism. But I certainly could have made it a lot clearer that that is what is at stake. I have come to see how unfortunate it is that I did not explicitly consider a version of dispositionalism that identifies an object’s being a certain colour with its having a disposition to produce perceptions of a certain kind in perceivers constituted just as we actually are in a world as it actually is right now. Several readers, Johnston now among them, have urged that such a view avoids whatever difficulties I might otherwise have raised for the versions of subjectivist dispositionalism that I consider. But even this “rigidified” version of dispositionalism must meet the third condition of success. It must specify the kinds of perception that serve to identify the particular disposition in question without assuming or presupposing that any objects are coloured in anything other than a dispositional sense. If I am right about the requirement for the attribution of colour perceptions and beliefs that I think bedevils all attempts at consistent unmasking, then that same obstacle challenges all dispositionalist theories of an object’s colour as well. The third of the three conditions of dispositionalism’s success could not be fulfilled. Only someone who believes that objects are coloured, and not in a dispositionalist way, could find in the world perceptions of the kind that dispositionalism says serve to define the particular disposition with which an object’s being a certain colour is to be identified. Strictly speaking, this implies only that no one could consistently find a dispositionalist view of an object’s colour to be true of all coloured objects in the world. No one could consistently assert that all objects are coloured in only a dispositionalist way.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

UNMASKING AND DISPOSITIONALISM



Johnston describes the disposition he has in mind as a “disposition to look yellow to standard perceivers as they actually are” (200). This leads us back to the question of what it takes for someone to have perceptions in which something looks a certain way, and how someone can understand and recognize that people have perceptions like that. If the perceptions are understood as having determinate intentional contents, there is the question of what it takes for something to look a certain way to a perceiver. Must the person have a conception of what it would be for things to be that way? If so, and if that in turn involves being prepared in general to accept unchallenged perceptions with just that content as veridical, then anyone who can see what we can see and who understands what it is for an object to be coloured will believe that objects are coloured. If even a person’s unchallenged perceptions would not be enough to prompt acceptance of that belief, there would be a question whether the person has a conception of such a perceivable state of affairs at all. There are no quick answers to these questions about the conditions of perception and belief about the colours of things and about the conditions of their attribution. I have tried to bring out how the complex relations we understand to hold among the contents of perception, thought, and belief stand as a challenge to all forms of dispositionalism, including the “mildly sophisticated” (200) version Johnston describes here. To oppose that theory is not to deny that, as things are, many objects have dispositions of just the kind Johnston describes. Many objects have a disposition to produce in us as we are, under appropriate conditions, perceptions of something’s looking yellow. And an object’s having such a disposition obviously depends in part on what human perceivers are actually like; if we had been different in certain ways there would have been no objects with such dispositions. The same is true of a disposition to produce in us, as we actually are, feelings of anxiety. Objects would not have such a disposition if we were not as we are. There are also objects that have a disposition to produce in us as we actually are, under appropriate conditions, perceptions of something’s looking ovoid, say, or perceptions of something’s looking like an elephant. There is no question that there are objects with just such dispositions, and that their having those dispositions also depends in part on facts about us.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



UNMASKING AND DISPOSITIONALISM

But presumably no one would think of identifying an object’s having either of these last two dispositions with the object’s being ovoid, or its being an elephant. Being an object of that kind is one thing, having a disposition to produce perceptions like that is something else. It is not one and the same property. In those cases, we can say that the object has the disposition because it is ovoid, or is an elephant—and also because of the way we are. We are actually such that we get perceptions like that from an object that is ovoid, or from an elephant, under the specified conditions. An object’s having the specified disposition depends in part on the way we actually are, but an object’s being ovoid, or being an elephant, does not depend at all on the way we actually are. This shows that the fact that an object has a “rigidly” specified disposition to produce perceptions of certain kinds in us as we are is not sufficient to support identifying the object’s having that disposition with its having the property that occurs essentially in giving the content of those perceptions. Some additional reason must therefore be found for identifying an object’s being yellow with its having such a disposition. What supports that particular proposal rather than holding that objects that have that “rigidly” specified disposition have it in part because they are yellow? There is nothing to recommend a comparable view of the property of being actually-anxiety-producing. In that case, unlike being ovoid or being an elephant, there is nothing other than an object’s having such a disposition for its having that property to be. And it makes no sense to attribute to an object the anxiety we feel when that disposition is activated. But when an object’s “rigidly” specified disposition to produce perceptions of yellow is activated, we see something yellow, and it makes perfect sense to say that an object is yellow. What in that case supports the idea of identifying an object’s having that property with its having that disposition? To say that there is nothing in the world other than that disposition for an object’s being yellow to be would raise the question of how that in turn is to be supported. To argue on grounds of “parsimony” that there is no such thing in the world because a world of nondispositionally-coloured objects is an “idle hypothesis” that is not needed to explain “the generation of our experience” (192), would take us back to the prospects of successfully unmasking the colours of things.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

14 Are the Colours of Things Secondary Qualities? I resist the familiar idea that the colours we believe to belong to the objects we see are “secondary” qualities of those objects: that whether an object is coloured at all, or what colour it is, depends essentially on what kinds of perceptions it does or would produce in perceivers like us in the kinds of circumstances in which we actually receive the perceptions of colour that we do. My resistance to that doctrine is not widely shared among philosophers. John McDowell continues to endorse the view, and he has explained why he thinks I am wrong to resist it.¹ Here I would like to consider his reasons. McDowell does regard it as misguided to try to “unmask” the colours of things as part of a general “quest for reality”, but when freed from the project of “debunking their reality”, he thinks there is an obvious and “innocuous” understanding of “the idea that colors are secondary qualities and as such differ from primary qualities in that concepts of them are in a certain sense subjective” (218). He thinks I have no reason to resist that “metaphysically innocent” version of the doctrine. I think McDowell slightly misunderstands my resistance. I do not think the doctrine that the colours of things are secondary qualities implies that the colours of things are not real, or not part of the world, or not part of what is so. The doctrine does not imply that objects are not coloured, or not really coloured, or that they are “mere appearance” or illusory. I agree that one could believe that the colours of things are ¹ See John McDowell, “Colors as Secondary Qualities”, in J. Bridges, N. Kolodny, W. Wong (ed.), The Possibility of Philosophical Understanding: Essays on the Thought of Barry Stroud, Oxford University Press 2011 (page numbers alone in parentheses in the text refer to this essay). For an earlier response along the same lines, see his “Reality and Colours: Comment on Stroud”, Philosophy and Phenomenological Research 2004.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



ARE THE COLOURS OF THINGS SECONDARY QUALITIES ?

secondary qualities without being concerned to distinguish in general between what is real and what is only apparent or unreal. But I do not regard the question of the nature or status of the colours of things as free of all metaphysical significance. In my own discussion of the colours of things, I have used the rubric “the quest for reality”, but I have not been concerned with debunking the reality of the colours of things.² I have had in mind their status with respect to what I call “independent” reality: what is so “independently” of perceivers responding to things in the world in all the ways they do or would. I take the doctrine that the colours of things are secondary qualities to mean that an object’s having a certain colour implies only that perceivers of certain specified kinds do or would have perceptual responses of certain specified kinds to objects of that kind in specified circumstances. The view is that objects do not have the colours they have independently of what perceivers of certain kinds do or would perceive under certain circumstances. And I take that to imply, contrapositively, that if perceivers of those kinds would not in fact get perceptions of the specified kinds in the specified circumstances, objects would not be coloured, or would not have the colours we now believe them to have. On this view, the colours of objects would not be part of what is so independently of perceivers’ responding to those objects in the ways they do or would. The truth of what we say in our apparently categorical ways of speaking of the colours of things would have been “unmasked” as dependent on the responses of perceivers. But the colours of things would still be part of reality. The truth of what we say would depend on what is true of certain kinds of perceivers, but what we say would still be true or false of what is so. The secondary qualities view says that the colours of things are essentially dependent on what human perceivers do or would perceive. It is a philosophical thesis about the nature of the colours of things; about their metaphysical status. That is the view I resist. I think we cannot consistently accept it. McDowell recognizes that my focus on the conditions of successfully unmasking the metaphysical status of the colours of things is a way of trying to make sense of whatever notion of reality is involved in it. But the notion of reality I am concerned with is “independent reality”, not ² Barry Stroud, The Quest for Reality: Subjectivism and the Metaphysics of Colour, Oxford University Press 2000.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

ARE THE COLOURS OF THINGS SECONDARY QUALITIES ?



simply the general notion of reality. He thinks the moral I would draw if my efforts to forestall the operation were successful is that “we do not understand the supposed special notion of reality . . . We do not know what it means to ask the metaphysical question” (218). I think that comes close, but it puts it too much in terms of meaning and understanding. It would be closer to say that I think we cannot find the alleged metaphysical conclusion fully satisfactory; we cannot consistently accept it. What does McDowell see in our concepts of the colours of things that leads him to think they “are in a certain sense subjective”? What he sees, first, is the dispositional character of the colours of objects. He wants to preserve the basic idea in Locke that having a certain colour is a power or disposition objects have to “occasion sensory goings-on in subjects” (219). McDowell describes those “sensory goings-on” as an object’s looking a certain way to someone. In my own reflections on the secondary quality view I have described the relevant “sensory goingson” as perceptions of an object’s being a certain colour. There might at first sight seem to be no real differences between us here. McDowell says, “an object that looks yellow presents itself as being yellow” (225). And I think a perceiver who sees an object that presents itself as being yellow can see that object to be yellow (not necessarily truly, of course). The “lookings yellow”, McDowell says, are “perceptions of a property of being yellow had by the objects that are being perceived” (221). And he thinks “we can capture the distinctive look of objects of a certain kind only by using a term for that color as a property of objects” (229). This suggests that a perceiver to whom certain objects look yellow can be said to see the property of being yellow to be a property of the objects seen. So in general, the kinds of perceptions a coloured object is disposed to produce in the specified circumstances are perceptions of the colour perceivers in those circumstances would see the object to have. I think both McDowell and I could assent to these descriptions of the relevant “sensory goings-on”. But there is at best only an appearance of agreement between us. On McDowell’s conception, the “property of being yellow” that perceivers see objects to have is a dispositional property: being such as to look yellow to such-and-such perceivers. That is the secondary qualities view. I think the property of being yellow that competent perceivers see objects to have cannot be that dispositional property. I think we cannot accept that view consistently with understanding how we actually perceive and

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



ARE THE COLOURS OF THINGS SECONDARY QUALITIES ?

believe what we do about the colours of things. Some of the colour perceptions coloured objects are disposed to produce must be understood as non-dispositional colour properties of the objects seen. What is in dispute between McDowell and me is the kinds of perceptions coloured objects are disposed to produce in perceivers. Of course, not all yellow objects look yellow, or are seen to be yellow, to all perceivers in all circumstances. McDowell’s view is that “for an object to be, say, yellow is for it to be such as to look yellow . . . to normally sighted human beings, in the sorts of circumstances in which we can directly determine what color something is by looking at it” (219). I think yellow objects do have dispositions like that. As things actually are in the world, and as normal human perceivers actually are, yellow objects would look yellow to normal perceivers like that in the sorts of circumstances in which we can directly determine what colour something is by looking at it. But that way of describing the dispositions coloured objects have does not capture the dispositional conception of the colours of things that McDowell has in mind. Rectangular objects (as long as they are not too large or too small) are also such as to look rectangular to normal human perceivers in the circumstances in which we can directly determine what shape something is by looking at it. In ascribing such dispositions to rectangular objects we mention certain subjective “goings-on”—an object’s looking rectangular to someone. But that is no reason to think that an object’s being rectangular is its having a disposition to produce certain perceptions. Nor does it suggest that there is something “subjective” in the idea of an object’s being rectangular. The secondary qualities view draws a fundamental distinction between the colours of things and the shapes of things in this respect. For McDowell, they differ metaphysically because of the different kinds of perceptions objects of the two different kinds are disposed to produce. With the colours of things, McDowell says, “the point of the dispositional formula lies in the connection it claims there is between the idea of being this or that color and the idea of its looking the way things of this or that color look” (224). That “connection” expresses “what it is” for an object to be a certain colour: “the formula can be cast as identifying an object’s being yellow with its being such as to look yellow to us” (222). “What I am urging”, McDowell says, “is that something’s being yellow is its being such as to look yellow” to appropriate perceivers in the

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

ARE THE COLOURS OF THINGS SECONDARY QUALITIES ?



specified circumstances as things actually are (226). But with rectangular objects, he agrees that there is no such identity. Rectangular objects do have dispositions to produce such perceptions, but that is not what it is for an object to be rectangular. It is the identity of the colours of things with their dispositions to look certain ways to specified perceivers that is the heart of McDowell’s conception of the colours of things as secondary qualities. McDowell concedes that in describing the disanalogy he sees between the colours and the shapes of things, he is doing no more at this point than affirming that the colours of things are secondary qualities in a way the shapes of things are not (227). He is not offering here positive reasons to accept that view of the colours of things. But he says the disanalogy “seems utterly intuitive, if we can separate it . . . from the supposed implication . . . that secondary qualities would be excluded from the results of a well-founded quest for reality” (227). I have separated my question about the status of the colours of things from that “supposed implication”. So I want to ask what McDowell finds “utterly intuitive” in the disanalogy he sees between the colours of things and the shapes of things. The answer to this question lies in McDowell’s conception of the looks of the colours of things, or of something’s looking a certain colour to perceivers. He says, first, that “the very idea of something’s being yellow . . . is an idea whose content cannot be actually captured without invoking the distinctive look of yellow things” (227). Here I think he draws attention to something important and distinctive about concepts of the colours of things. Coloured things look a certain way when they are perceived, and one cannot acquire or understand the concept of something’s being yellow, for instance, without being familiar with the distinctive look of yellow things, or at least with the distinctive looks of some coloured things. This means, as McDowell puts it, that “what ascriptions of a color mean cannot be conveyed except ostensively, which, colors being what they are, requires that the learner sees an instance” (229). This is true of the colours of things in general, but not of the shapes of things. Although we must be familiar with the looks of the colours of things in order to have concepts of the different colours, it is too strong to say that for each colour term we understand, we must have seen an instance of that colour. Think of Hume’s shade of blue which we might get an

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



ARE THE COLOURS OF THINGS SECONDARY QUALITIES ?

adequate idea of without having seen it. But there does seem to be a very general requirement: colours being what they are, possession of the concepts perceivers apply to objects in the thought that they are this or that colour, requires that perceivers have actually seen the distinctive look of certain colours. No such requirement holds for rectangularity or other shape concepts. I think this is true, and there is a good explanation of it. The colours of things are available in perception to only one sense; they are only visual qualities of objects. Since our only sensory access to them is by seeing them, what is distinctive of each colour can be understood only through seeing: by ostension in visual perception of the colours of objects and the systematic relations among them. That is why it is impossible to give a satisfactory definition of ‘yellow’ or of any particular colour term without making use of some colour terms. Determinate shades of a colour were traditionally regarded as “simple” ideas for just that reason. Concepts of the colours of things are irreducible in general to anything expressed in purely non-colour terms. This obviously does not mean that we cannot know or cannot say what it is for something to be yellow, for instance. Within the competence we all have with colour terms we can say, and even indicate, what it is for a thing to be yellow. It is for the thing to have that distinctive visual property we know to be denoted by the term ‘yellow’ which we understand to be true of certain things we see. We might not even have to use the term ‘yellow’ if the relations among other colour terms we understand served to specify yellow uniquely. Of course, indicating in this way what it is to be yellow could not serve to explain the concept of something’s being yellow, or to impart that concept, to someone who does not already have concepts of the colours of things. But in indicating what it is for something to be yellow by using colour terms we already understand, what we thereby say is something we can know to be true of certain objects. Someone who already has concepts of the colours of things, but not yet the concept of yellow, could acquire that concept, and so learn what it is for something to be yellow, by seeing what a ripe lemon looks like in good light and being authoritatively told that for something to be yellow is for it to be that colour. That is how we in fact learn most of the terms for the colours of things: by ostension. What we learn in that way is something new to us: that that colour is yellow, or that yellow things look

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

ARE THE COLOURS OF THINGS SECONDARY QUALITIES ?



like that. We can learn such things in that way only because we see objects to be a certain colour, or to look a certain way. The “ostension” requirement on concepts of the colours of things means that any perceivers who have such concepts must have seen the colours of some things or experienced their distinctive “looks”; they must have experienced some such “subjective” sensory “goings-on”. But that is a requirement on possessing and understanding concepts of the colours of things. Once we have mastered those concepts and use them to describe the colours of certain objects, we do not then say anything about how we acquired those colour concepts in the first place. We had some “subjective” perceptual experiences, but that does not mean there is something “subjective” in the “contents” of the concepts we apply to objects in thinking they are this or that colour. McDowell holds that ostension of the colours of some things is required for acquisition of concepts of the colours of things, but he does not take that to support the secondary qualities view. He reminds us that his view is a view about “what kind of property an object’s having this or that color is” (p. 224). It is not simply a question of what it is for an object to be yellow, or to be green, or some other particular colour. The essentially “subjective” character of the colours of things might not be made explicit in what we know in knowing what it is for an object to be yellow, or to be green, in particular. The secondary qualities view that McDowell defends is a completely general account. It is the view that: what it is for an object to be F, where “F” is replaced by any expression for a color, is for it to be such as to look F, to the appropriate subjects and in the appropriate circumstances. (224)

This tells us what it is for an object to have a certain colour F by “identifying” the object’s being the colour F with the object’s having a disposition to look a certain way (to look F) to certain perceivers. This is the crucial element in McDowell’s conception. An object’s being the colour F is the same as (identical with) the object’s being “such as to look F” in such-andsuch specified circumstances. That is what makes it “utterly intuitive” to McDowell that the colours of things are secondary qualities: nothing more than dispositions to produce certain “subjective” “sensory effects”. How is that identity revealed, or how is it known to hold? McDowell explains that the identity is expressed in the “connection there is between

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



ARE THE COLOURS OF THINGS SECONDARY QUALITIES ?

the idea of being this or that colour and the idea of looking the way things of this or that colour look” (224). What is that “connection” of ideas? Are those ideas one and the same idea, or two? Counting ideas is a discouraging task, but surely the idea of being such as to look yellow to certain perceivers is not the same idea as the idea of being yellow. The one idea—being such as to look yellow—contains more ingredient ideas than the other idea—being yellow. And there cannot be said to be an identity of meaning between the word ‘yellow’ and the words ‘such as to look yellow . . . ’. The relevant “connection” must be that those different ideas or words denote one and the same property. That is the identity McDowell asserts and relies on. What reasons are there to believe that such an identity holds? It is not easy to find McDowell’s answer to this question. One thing he appeals to in this connection is the sheer contingency of human beings’ having the kind of visual equipment we have, and of the world’s being such that we can make such effective use of it. As he puts it, that we have chromatic visual perception at all, that perceptual life exhibits sufficient uniformities to enable us to see colours that are widely enough shared to allow for a notion of normal vision, and that there is widespread agreement among human subjects about the colours of objects: all this, he reminds us, is true of human visual perception as it is. And it is all a huge and intricately related combination of contingent facts. Things could have turned out quite differently from this for inhabitants of this planet, even drastically so. McDowell finds within this huge human contingency: not just that the possibility of agreement in judgment about the colours of objects is a condition of the possibility of concepts of those properties. That contingent uniformity in chromatic color sensation enters into the very constitution of the properties themselves. There is nothing to those properties, those ways things can be, except being such as to look the appropriate way to subjects like us. (230)

He says there is “nothing to” the colour properties of objects “except being such as to look” a certain way to us. That is because he thinks, as he says, that the facts of chromatic colour sensation of human beings enter into the very “constitution” of colour properties. That expresses the identity of the colours of things with their dispositions that McDowell accepts. Nothing comparable is said to hold for shapes.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

ARE THE COLOURS OF THINGS SECONDARY QUALITIES ?



We can think about the shapes of things only because we can agree in judgments about the shapes of things. But that possibility of agreement in judgments does not enter into the constitution of shape properties themselves. What constitutes a shape property is not a propensity to occasion certain states in perceiving subjects, but a way of occupying space. (230)

But McDowell concedes that in this passage as well he is merely stating rather than arguing for the disanalogy he sees between the colours and the shapes of things. I think he is right to acknowledge the impressive complexity of the factors that contribute to our having a conception of a world of coloured objects at all, and the overwhelming contingency of those factors’ all having come together in the ways they have. It is remarkable, even thrilling. But I don’t think it is something special about the colours of things. A complex combination of contingent facts of human accord in “subjective” responses to the world, with shared judgements about what is so, lies behind our acceptance of almost everything we accept—about the shapes of things, the physical and biological features of the world, even mathematical truths. The contingencies which must hold for all that to be possible are equally striking, and equally pervasive, across the board. But the contingency of all those facts does not reveal anything “subjective” in the contents of our thoughts in those other domains. What reasons are there to think it is any different with our beliefs about the colours of things? When McDowell explains what he sees as the special character of the colours of things he is not drawing attention to a difficulty in explaining to someone what it is for something to be a particular colour. He is identifying what he thinks is the only thing there is in the world that an object’s being coloured could possibly be: “nothing except being such as to look the appropriate way to subjects like us”. He means to capture what “constitutes” the colours of objects, their essential nature. I do not regard this doctrine as “metaphysically innocent”, in McDowell’s phrase. I resist it, not because I think I have discovered something else that “constitutes” an object’s being a certain colour, or some other property for it to be “identified” with. I resist because I think we have been given no reason to believe that an object’s being a certain colour is to be identified with its being disposed to produce perceptions of certain kinds in certain specified circumstances. What is more, I think there is no such identity between those properties. I think that dispositional

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



ARE THE COLOURS OF THINGS SECONDARY QUALITIES ?

property of coloured objects is not the property we see an object to have when we see it to be a certain colour. So I disagree with McDowell’s understanding of the “sensory goings-on” that coloured objects are disposed to produce in perceivers. I do not share his conception of perceptual experience (or perception of the colours of things). We understand and apply concepts of the colours of objects to things we perceive. McDowell agrees that the perceptions coloured objects are disposed to produce are perceptions of an object’s being a certain colour. “On a naive understanding, our visual experience sometimes directly presents us with colors as properties of the objects we see” (221). We see objects to be yellow, or to be green, and we ascribe those colours to objects by using terms for colours as properties of those objects (229). It is because we see the colours of objects that we come to believe what we do about the colours of objects. Any adequate account of our perceptions of and beliefs about the colours of things must acknowledge these facts, and account for them. I think the secondary qualities view does not adequately account for these facts. I think we cannot accept that view while understanding how we do in fact perceive and believe what we do about the colours of things. We see objects to be a certain colour. What does it take for a perceiver to have such perceptions, and so to believe that the object is that colour? Could a perceiver see an object to be yellow if that perceiver did not have a concept of something’s being yellow, and so did not even understand what it is for an object to be yellow? Creatures with no such concepts can certainly see objects that are yellow, and see objects that are green. Perhaps that is enough to make a difference for those perceivers; the two perceptions are different, after all. But whether creatures who lack the relevant colour concepts see those objects to be yellow, and see them to be green, or even see them to be different in colour, is a further question. I am inclined to say that is not something creatures without colour concepts can see; they are not sufficiently equipped for perceptions of that kind. They can see an object of a certain colour, but they do not see the object to be that colour. If that is so, creatures without concepts of the colours of things never get the kinds of perceptions McDowell thinks coloured objects are disposed to produce in perceivers like us in the specified circumstances. They do not see objects to be yellow, or to be green. I think such creatures could not see objects to be such as to look yellow, or to look green, in certain circumstances either.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

ARE THE COLOURS OF THINGS SECONDARY QUALITIES ?



Whatever might be true of creatures without the relevant concepts, we human beings do have concepts of the colours of objects, and we understand what it is for something to be coloured. The secondary qualities view certainly must acknowledge that we have such concepts, since it holds that we have beliefs about the colours of things, and we could not have such beliefs without some way of thinking of objects as coloured. Those beliefs are just what the secondary qualities view is meant to account for. The point of the dispositional theory is to identify the content and explain the status of what human beings believe in believing that objects are coloured. The secondary qualities view says the concepts we apply to objects in the belief that they are a certain colour are dispositional concepts. But our beliefs about the colours of things come to us, ultimately, only from perception. Colours are available in perception only to sight, so we believe what we believe about the colours of objects only because of the colours we see. The dispositional theory says yellow objects are disposed to cause qualified perceivers to see objects to be yellow. But that view accepts only a dispositional sense of an object’s being yellow: being such as to produce perceptions of an object’s being yellow in appropriate perceivers. Is that the property we see to be a property of an object when we see an object to be yellow? I think it is not. I think it cannot be. I think that is not the property we see yellow objects to have, but not because yellow objects we see do not have that dispositional property. Yellow objects do have a disposition to produce just such perceptions in those circumstances. I think it is even possible to see an object to have that dispositional property. Someone who saw an object to be yellow, and knew he was in the specified perceptual circumstances, could perhaps then see the object to be disposed to produce perceptions of an object’s being yellow. He could see the object to have that disposition because he sees that disposition to be actualized. He sees that because he has a perception of the kind he knows objects with that disposition are disposed to produce in those circumstances. But that perception is a perception of an object’s being yellow, not simply a perception of an object’s being disposed to produce perceptions of certain kinds. I think an exclusively dispositional conception of an object’s being a certain colour cannot account for the perceptions we have of the colours of things. For an object to have a certain disposition, there must be something the object is disposed to do or to exhibit in the relevant

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



ARE THE COLOURS OF THINGS SECONDARY QUALITIES ?

circumstances that is distinct from its having that very disposition. Sugar that is disposed to dissolve in water, for instance, will dissolve when placed in water. The sugar’s actually dissolving when its disposition is actualized is not itself a dispositional property of the sugar, even if it involves the activation of many dispositions of many objects. If, in the specification of an object’s disposition to produce perceptions of an object’s being F, the ‘F’ denoted only that dispositional property itself, that would leave the disposition still unspecified: “being disposed to produce perceptions of an object’s being disposed to produce perceptions of an object’s being disposed”, etc. There must be some determinate kind of perception objects are disposed to produce other than perceptions of the object’s being disposed to produce perceptions of its having that very disposition. Some independent specification of the kinds of perceptions involved is needed. McDowell’s view identifies yellow objects with objects that are “such as to look yellow to normally sighted human beings, in the sorts of circumstances in which we can directly determine what color something is by looking at it” (219). I think a competent perceiver to whom an object looks yellow in those circumstances sees that object to be yellow; he sees it to have the colour property he sees. It is not simply that his “perceptual experience” exhibits a certain distinctive feature or “look”. He has a perception of the object’s looking to be yellow. And in those circumstances—when normal perceivers can directly determine the colour of an object by looking at it—the colours objects look to be are the colours they are. And that concept of an object’s being a certain colour must be a non-dispositional concept. Given the requirement of ostension for the possession of colourconcepts, all of us who are competent with such concepts have seen objects’ being a certain colour. I think some of those perceptions must be perceptions of an object’s being coloured in a categorical, nondispositional sense. If that is right, even someone who accepts the secondary qualities view must have had perceptions of that kind, and so must be proficient with that same non-dispositional concept. I think even to understand, without accepting, the secondary qualities view we must be competent in applying non-dispositional colour concepts to objects we perceive. I think this shows that a non-dispositional concept of an object’s being a certain colour occupies the primary position in

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

ARE THE COLOURS OF THINGS SECONDARY QUALITIES ?



the competence we all exercise in seeing and believing what we do about the colours of things. This was ably argued long ago by G. E. M. Anscombe. As she put it, the words giving the object of a verb of sense “are necessarily most often intended as giving material [rather than merely intentional] objects of sense: for this is their primary application”.³ She explains that a description of something’s looking or appearing a certain way must “in the first instance” be a description of how the thing would be if the description of the appearance were true of the object, “really, and not merely in appearance: this will be its primary application” (A 172). Further, we ought to say, not: “Being red is looking red in normal light to the normal-sighted,” but rather “Looking red is looking as a thing that is red looks in normal light to the normal-sighted.” For if we ought to say rather the first, then how do we understand “looking red”? Not by understanding “red” and “looking”. It would have to be explained as a simple idea; and so would looking any other colour. (A 172) It was plausible to say: Only if it looks red to him will he learn what is meant; but wrong to infer: What he then grasps as the correlate of the word “red” is a red look. (A 173)

When we see an object to be F, we take the object to be that colour; we take the concept of being F to apply to the object we see. In believing that the object we see is F we endorse that thought that the object is F. In general, what we see to be so of an object we see to be F is what we believe to be so of an object we believe to be F. The colour term ‘F’ is applied with the same meaning in each case. There is in that sense a direct connection between the contents of perceptions of the colours of things and the contents of beliefs about them. We get our beliefs about the colours of most objects directly from what we see the colours of those objects to be. McDowell accepts a direct connection between the objects of colour perception and the objects of colour belief (222). But he allows only a dispositional concept of an object’s being a certain colour; he thinks that is what we believe in believing an object to be a certain colour. On that ³ G. E. M. Anscombe, “The Intentionality of Sensation: A Grammatical Feature”, in R. J. Butler (ed.), Analytical Philosophy: Second Series, Blackwell 1965, p. 172 (page numbers with ‘A’ in parentheses in the text refer to this essay).

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



ARE THE COLOURS OF THINGS SECONDARY QUALITIES ?

view, with a direct connection, what we see in seeing an object to be a certain colour is the object’s being disposed to produce perceptions of certain kinds. I think that is not what we see; seeing an object to be a certain colour is not the same as seeing the object to have a certain disposition. So I think the secondary qualities view cannot accept a direct connection between what we see to be so of coloured objects and what we believe of such objects when we believe them to be a certain colour. If that is so, even to understand what kinds of perceptions the dispositional theory says F-coloured objects are disposed to produce, we must have a non-dispositional concept of an object’s being the colour F. We also need a capacity for perceiving an object to have that same nondispositional property. Both seeing objects to be F, and believing objects to be disposed to produce perceptions of an object’s being F, require possession of a non-dispositional concept of an object’s being F. I think an exclusively dispositional conception of the colours of things cannot account for these facts of perception and belief. If that is right, an exclusively dispositional view of the colours of things cannot account for its own intelligibility. We cannot consistently accept what that theory says while understanding our believing what we do about the colours of things and the perceptions by which we get those beliefs. If there were nonetheless good reasons to accept the secondary qualities view, we would perhaps have to face that inconsistency. But I think we have found no reason to accept that theory, or the alleged identity on which it rests. The inconsistency I see is not an inconsistency or contradiction in the statement of the theory itself. It is rather that we cannot accept that theory of the colours of things consistently with our making the sense we actually make of ourselves as perceiving and believing what we do about the colours of things.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

15 Concepts of Colour and Limits of Understanding Wittgenstein raises many difficult, puzzling questions in these very late notebooks published as Remarks on Colour.¹ “Why is it that something can be transparent green but not transparent white?” (I, 19). “Why is there no brown or grey light?” (III, 215). “If I look at pure red through glass and it looks grey, has the glass actually given the colour a grey content . . . or does it only appear so?” (III, 207). “What does it mean to say ‘Brown is akin to yellow’?” (III, 47). I will not take up questions like this directly. Instead, I will discuss a more general concern Wittgenstein raises about our concepts of colours and the way our possession of those concepts does or does not impose certain limitations on our understanding of other apparent possibilities. This is something Wittgenstein struggled with not only about colours but about almost every other aspect of our thought. I start with something that seems to have nothing to do with colour. Wittgenstein says, “Someone who has perfect pitch can learn a languagegame that I cannot learn” (III, 292). The language-game in question is presumably saying correctly what note you hear without relying on any independent standard. Wittgenstein cannot learn that language-game, he implies, because he does not have perfect pitch. Because of that limitation, which he shares with probably most other human beings, he cannot learn or participate in that practice. But although he does not have what it takes to engage in the practice, he presumably knows what the practice is. He knows and understands and can describe what those in that practice can do. ¹ L. Wittgenstein, Remarks on Colour (ed. G. E. M. Anscombe), Basil Blackwell 1977. (Parentheses in the text with roman numerals followed by numbers refer to the relevant part and section of this book.)

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



CONCEPTS OF COLOUR AND LIMITS OF UNDERSTANDING

In some remarks collected in Zettel,² probably from an earlier time than this material on colours, Wittgenstein draws a connection between perfect pitch and concepts of colours. He invites us to imagine people who speak of colours intermediate between red and yellow by means of fractions in a kind of binary notation representing different proportions of the colours at each end of the range from red to yellow. He presents this as a fully developed part of the life of these people. He says: These people learn how to describe shades of colour this way in the kindergarten, how to use such descriptions in picking colours out, in mixing them, etc. They would be related to us roughly as people with absolute pitch are to those who lack it. They can do what we cannot. (Z 368)

We cannot do what those with absolute pitch can do, but we at least understand what they do. Do we understand what these people who speak of colours in this different way are doing or saying when they speak as they do? Having described these people in this way, Wittgenstein says: here one would like to say: “But then is it imaginable? Of course, the behaviour is! But is the inner process, the experience of colour?” And it is difficult to see what to say in answer to such a question. Could people without absolute pitch have guessed at the existence of people with absolute pitch? (Z 369)

What makes it difficult—if it is—to say whether or not this culture is imaginable does not seem to be whether we could have guessed that there are people who speak of colours in this different way. Even if we could not have guessed that there are some people with perfect pitch, we now know that some people have it. We understand very well what people with that gift can do; we even have a test for it. We have concepts of the musical scale in terms of which we understand and can recognize the achievement of those people. The question about people described as speaking of the colours between red and yellow in these different ways is presumably whether

² L. Wittgenstein, Zettel (ed. G. E. M. Anscombe and G. H. von Wright), Basil Blackwell 1967 (parentheses in the text with ‘Z’ and a number refer to the relevant section of this book).

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

CONCEPTS OF COLOUR AND LIMITS OF UNDERSTANDING



we are in a position to understand what they do. Wittgenstein says what he calls their “behaviour” is imaginable, and so intelligible. His question is whether what he calls “the inner process, the experience of colour” of these people is imaginable. But whatever makes that question difficult, if it is, is surely not that we cannot gain direct access to the distinctive character of these people’s “inner processes” or “experiences”. The difficulty seems to lie rather in saying what experiences these people have when they see and speak in their special way of a certain colour. Wittgenstein says of these people that they can do what we cannot. But what can they do that we cannot? And can we understand what they do even if we cannot do it? We recognize a range of colours intermediate between red and yellow—from a strongly reddish yellow to a strongly yellowish red. Those people Wittgenstein describes presumably do that too. Is it that they have many different single words for different points along that spectrum, while we have very few, or mostly only compound words? Is the “behaviour” we are to imagine them engaging in a matter of applying single words to a great many different colours in the range between red and yellow? If so, there is a question whether they really have different concepts of colour from ours. Of course, we did not learn as many different colour-words in the kindergarten and in subsequent training as they did. But their language-game, as described, seems to be a language-game we could learn. We seem to have the resources and the capacities already. At least it seems that we would not be excluded from their practice as most of us are hopelessly excluded from the languagegame with perfect pitch. We can presumably make all the discriminations those imagined speakers make. So we would not be excluded from their practice in the way perceivers with various kinds of colourblindness are excluded from full participation in the language-game of those with more or less normal vision. The colour-blind cannot make all the discriminations the rest of us can make. Whatever can or cannot be understood about concepts of colours intermediate between red and yellow, Wittgenstein thinks things are radically different with something else we might think we can conceive of. He says: it is obvious at a glance that we aren’t willing to acknowledge anything as a colour intermediate between red and green. (Z 359)

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



CONCEPTS OF COLOUR AND LIMITS OF UNDERSTANDING

It is not that red and green cannot be mixed; if you put them together you get something. The point is, rather, as Wittgenstein puts it: we are not able to recognize straight off a colour that has come about by mixing red and green as one that can be produced in that way. (Z 365)

Does this mean we actually have no concept of a colour reddish green? At the very least, Wittgenstein says, we cannot see something as being that colour. If that is so, is that to be understood as a limitation on our part? Is it like the limitation most of us are under of not having perfect pitch? If it is a limitation, it would be a limitation everyone is under, if what Wittgenstein says is right. But is it a limitation at all? If someone were to report that certain people are acquainted with reddish green, Wittgenstein imagines one might reply, “But there is no such thing!”. Wittgenstein is immediately struck by that reply: “What an extraordinary sentence”, he says, “There is no such thing! (How do you know?)” (Z 362). There would be nothing extraordinary in replying to a report that certain people are acquainted with round squares, for instance, by saying that there are no such things, and we know why. In finding the corresponding denial of such a colour as reddish green “extraordinary” I think Wittgenstein is struck by the oddness of denying that there is such a thing as a certain colour. What is it for no such colour to exist? Even if we can make some sense of a language-game with concepts of colours intermediate between red and yellow, Wittgenstein thinks it is completely different with reddish green: can I describe the practice of people who have a concept, e.g. ‘reddish green’, that we don’t possess?—In any case, I certainly can’t teach this practice to anyone. (III, 122) Can I then only say: “These people call this (brown, for example) reddish green”? Wouldn’t that just be another word for something I have a word for? If they really have a different concept than I do, this must be shown by the fact that I can’t quite figure out their use of words. (III, 123)

This remark is immediately followed by the kind self-reproach found elsewhere in Wittgenstein’s writings. But I have kept on saying that it’s conceivable for our concepts to be different than they are. Was that all nonsense? (III, 124)

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

CONCEPTS OF COLOUR AND LIMITS OF UNDERSTANDING



It looks as if it would be nonsense to speak of the conceivability of concepts different from ours if the only way it could be shown was by our being unable to make sense of the imagined speakers’ uses of their words. That would not be to have conceived of a possibility of others possessing and using concepts recognizably different from ours. No determinate possibility at all would have been specified. Nonetheless, does it not seem possible—conceivable—for our concepts to be, or at least to have been, different from what they are? Couldn’t we who operate in all the ways we do have done certain things differently, and in ways we can now conceive of and understand? This is a more complicated question than it might seem. One way Wittgenstein puts the difficulty is: “Can’t we imagine people having a different geometry of colour than we do?”—That, of course, means: Can’t we imagine people who have colour concepts which are other than ours; and that in turn means: Can’t we imagine that people do not have our colour concepts and that they have concepts which are related to ours in such a way that we would also want to call them “colour concepts”? (III, 154)

This last condition is of course essential. We have to be able to recognize the different concepts that we conceive of as colour concepts in order to acknowledge the imagined possibility of our colour concepts being different from what they are. This means, as Wittgenstein says: We will, therefore, have to ask ourselves: What would it be like if people knew colours which our people with normal vision do not know? In general this question will not admit of an unambiguous answer. For it is by no means clear that we must say of this sort of abnormal people that they know other colours. There is, after all, no commonly accepted criterion for what is a colour, unless it is one of our colours. (III, 42)

We must be able to find something in what the people we conceive of can do that we can make sense of. And our resources for making sense of it, our resources for making sense of anything, lie in what we now know how to do and in our current ways of understanding things. All reflection or speculation about other conceivable possibilities faces a general kind of difficulty that Wittgenstein at one point describes this way: We say: “Let’s imagine human beings who don’t know this language-game”. But this does not give us any clear idea

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



CONCEPTS OF COLOUR AND LIMITS OF UNDERSTANDING

of the life of these people, of where it deviates from ours. We don’t yet know what we have to imagine; for the life of these people is [in other ways] supposed to correspond to ours . . . and it first has to be determined what we would call a life that corresponds to ours under the new [imagined] circumstances. . . . Questions immediately arise . . . You have to make further decisions which you did not foresee in that first statement. (III, 296)

Further decisions or specifications are needed to fill out the description of the otherwise ‘normal’, ‘familiar’ human life that the different ways of behaving we try to conceive of are to be imagined as fitting into. When the differences are small, or local, it is less clear that genuinely different concepts are involved. The colour-blind among us, for instance, cannot learn all the uses of colour words that the rest of us are masters of. But they understand a great many colour sentences; they just cannot use those sentences correctly in as many different ways as normal perceivers can (III, 278). It seems easier to make sense of others who are capable of less than what the rest of us know we can do, or even others who lack altogether certain concepts we know we have. That is to conceive of practices less rich or less complex than ours, not necessarily to conceive of concepts different from ours. The difficulties of understanding the use of concepts different from ours arise more clearly in attempting to conceive of others who do not less but more than we do, so to speak, or something completely different. To conceive of others who have and use concepts that we do not even possess we have to describe more and more fully the practices in which those concepts get their sense. And as a growing number of further decisions have to be made to help specify more fully exactly what the possibility in question amounts to, we begin to lose our grip on what the people we are trying to conceive of are really doing, or how they understand it. Is what they can be understood to be doing really a conceptual alternative to what we understand ourselves to be capable of doing and saying in our familiar language-games? When we press on for further specification, as we must, I think we can come to see that there is always much more at stake in the question of the intelligibility of concepts different from ours than appeared at the outset. Wittgenstein hinted in passing at part of the explanation of this in speaking of those with an elaborate vocabulary for colours intermediate

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

CONCEPTS OF COLOUR AND LIMITS OF UNDERSTANDING



between red and yellow as having learned it in the kindergarten. He generalizes the point in a couple of remarks in Zettel. I want to say: an education quite different from ours might also be the foundation for quite different concepts. (Z 387) For here life would run on differently.—What interests us would not interest them. Here different concepts would no longer be unimaginable. In fact, this is the only way in which essentially different concepts are imaginable. (Z 388)

This says the only way we can really imagine concepts essentially different from our own is by imagining human beings with very different upbringing and education, whose life accordingly “runs on differently” from ours, and who show little or no interest in what interests us. If that is really what it takes, it is not clear in advance how far we can get in conceiving of the use of concepts different from ours. Can we really imagine or describe a form of human life in which we understand the concepts expressed in it to be radically different from ours? There seem to be no firm stopping-points in such imaginings; we carry on, having to add further qualifications or specifications. But still it seems—doesn’t it?—that it is a contingent fact that we have the concepts and the conceptual resources that we do. So it seems hard to deny that, at least in very general terms, things could have been different from what they are. The indeterminateness we encounter in trying to conceive of people with radically different concepts and ways of living comes in part from our pushing against the very conditions of our understanding anything. This is something Wittgenstein touches on in these notebooks but takes up more directly elsewhere and earlier. In Remarks on Colour he draws attention to the difference between learning to do something, or being master of a certain use or practice, and being able to describe or understand a description of that practice. He imagines certain “mental defectives who cannot be taught the concept ‘tomorrow’ or the concept ‘I’ ” (III, 118). But then he asks: Now to whom can I communicate what this mental defective cannot learn? Just to whoever has learned it himself? (III, 119)

It looks as if the description of what such a person cannot do will be intelligible only to someone who has the very competence and understanding that the defective person is said to lack.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



CONCEPTS OF COLOUR AND LIMITS OF UNDERSTANDING

How can I describe to someone how we use the word “tomorrow”? I can teach it to a child; but this does not mean I’m describing its use to him. (III, 122)

To understand the possibility of concepts other than ours we must be able to describe and understand the imagined practice in which those concepts are used. But “[u]nderstanding the description itself already presupposes that [the person who understands] has learned something”, as Wittgenstein puts it (III, 121). I say to B, who cannot play chess: “A can’t learn chess”. B can understand that.—But now I say to someone who is absolutely unable to learn any game, so-and-so can’t learn a game. What does he know of the nature of a game? (III, 280)

It seems that what is said about other people with different concepts will be intelligible only to someone who understands and is proficient in the use of the very concepts those people are said to make use of. So when we ask, as Wittgenstein does, whether we can imagine people “having a geometry of colours different from our normal one”, that means: can we describe it, can we immediately respond to the request to describe it, that is, do we know unambiguously what is being demanded of us? The difficulty is obviously this: isn’t it precisely the geometry of colours that shows us what we are talking about, i.e. that we are talking about colours? (III, 86)

So to understand a description of what those with allegedly different concepts are doing and saying we must already be masters of languagegames in which we can make sense of their doing what they do. We must be able to “domesticate” those people to some extent, so to speak—to find them intelligible in our own terms—on pain of not finding them intelligible at all. But if we succeed in that effort, the people as described will not represent anything we can understand as radically different or ‘other’. I think there is an issue of more general significance behind all this that is worth drawing attention to. I think we come face to face here with the implications of an insight Wittgenstein expressed very early—in the 1930s—that remained fundamental to all his later thinking about understanding, meaning, and use. He wrote:

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

CONCEPTS OF COLOUR AND LIMITS OF UNDERSTANDING



The limit of language is shown by its being impossible to describe the fact that corresponds to (is the translation of) a sentence, without simply repeating the sentence.³

I take this to be directly relevant to the question of understanding the possibility of concepts different from ours. This remark says that it is impossible to say what a sentence says or means without making use of the very concepts expressed in that sentence. And it is impossible to understand what a sentence says without being competent in the use of those concepts. We could put it another way, as Wittgenstein sometimes does, by saying that it is only “from inside” understanding a language in which what the sentence says can be expressed that one can say and understand what the sentence says. So it is only “from inside” competence in and understanding of the structure of some conceptual practice or other that one can say what a particular sentence means or describe what people who utter and respond to that sentence are doing and saying. If this is part of what lies behind the general difficulty of understanding concepts that are genuine alternatives to our own, and we succeed in that daunting task only to the extent to which we can conceive of human beings with upbringing, education, and basic interests and attitudes radically different from ours, it is perhaps not surprising that we cannot finally satisfy ourselves one way or the other about their conceivability. But if we can begin to understand why that is so, why do we continue to raise the apparently unpromising question of the possibility of concepts different from ours? What is at stake? Is it perhaps that we think we can really understand our thinking and doing things in all the ways we do only if we can come to see how those ways of thinking and behaving are ‘grounded’ in or legitimately related to the world they enable us to think about? Wittgenstein asks in Zettel: Do I want to say, then, that certain facts are favourable to the formation of certain concepts; or again unfavourable? And does experience teach us this? It is a fact of experience that human beings alter their concepts, exchange them for others when they learn new facts; when in this way what was formerly important to them becomes unimportant, and vice versa. (Z 352)

³ L. Wittgenstein, Culture and Value (ed. G. H. Von Wright; tr. Peter Winch), Basil Blackwell 1978, p. 10.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



CONCEPTS OF COLOUR AND LIMITS OF UNDERSTANDING

Or, as he puts it in Remarks on Colour: We could say people’s concepts show what matters to them and what doesn’t. But it is not as if that explained the particular concepts we have. It is only to rule out the view that we have the right concepts and other people the wrong ones. (III, 293)

This talk of “the right concepts” echoes the well-known remark at the end of what used to be called Part II of Philosophical Investigations, now called ‘Philosophy of Psychology—A Fragment’: If anyone believes that certain concepts are absolutely the correct ones, and that having different ones would mean not realizing something that we realize—then let him imagine certain very general facts of nature to be different from what we are used to, and the formation of concepts different from ours will become intelligible to him.⁴

This is Wittgenstein’s target: the wish we apparently have to somehow vindicate our concepts or to show that they are “correct”, the “right” or the “only” concepts to have in a world like this. The demand or hope is that our concepts can be matched up with, or be seen or shown to have their ground or basis in, identifiable features or aspects of the independent world that the concepts are used to think and speak about. If that could be done, others with radically different concepts would be recognizable as missing something—not just something that we realize, but missing something that is really so in the way things are. Some such aspiration to understand the ‘ground’ or ‘basis’ of our concepts is what I think lies behind Wittgenstein’s provocative, challenging remarks in Zettel. We have a colour system as we have a number system. Do the systems reside in our nature or in the nature of things? How are we to put it? Not in the nature of numbers or colours. (Z 357) Then is there something arbitrary about the system? Yes and No. It is akin both to what is arbitrary and to what is non-arbitrary. (Z 358)

⁴ L. Wittgenstein, Philosophical Investigations (tr. G. E. M. Anscombe, P. M. S. Hacker, J. Schulte), Blackwell 2009, p. 241e. I discuss some implications of this remark and explore some of Wittgenstein’s examples of different concepts or practices in my ‘Wittgenstein and Logical Necessity’, Philosophical Review 1965 (reprinted in my Meaning, Understanding, and Practice Oxford University Press 2000).

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

CONCEPTS OF COLOUR AND LIMITS OF UNDERSTANDING



What does this ‘answer’ really say? That our “systems” are ‘akin’ to what is arbitrary, and ‘akin’ to what is non-arbitrary? But kinship, after all, is a very loose relation; it can extend great distances, and can even hold between incompatible extremes. One can see why Wittgenstein says what he says here, but I think there is something unsatisfyingly asymmetrical in the way he leaves it. To the question whether the number system and the colour system “reside in our nature or in the nature of things” he answers: “Not in the nature of numbers or colours”. That seems to mean that we cannot say that our system of numbers or colours is as it is because of the very nature of those things that the systems so successfully capture and express—numbers and colours. So we cannot say that different systems or ways of thinking of numbers or colours would miss something essential to those things we think about in our systems. It is not that the ‘ground’ of those systems of ours lies or “resides” in the nature of the things we think about in those ways. That is all Wittgenstein says here, but it is an answer to only one half of what is a two-part question. If the systems do not “reside” in the “natures” of those things themselves, are we to take it that the systems of numbers and colours do “reside in our nature” instead? Wittgenstein says nothing about that apparent possibility one way or the other. He does not conclude from his negative answer to the first half of the question that the number and the colours systems do “reside in our nature”. To draw that conclusion would presumably mean that the “geometry of colours”, the structural relations that hold among the different colours we acknowledge and understand, could be different, or would have been different if we had been different in certain ways. Is that what our accepting the system’s “residing in our nature” would amount to? If so, and if the system of numbers also “resides” not in the nature of numbers but in our nature, does that mean that what we now take to be truths of number would not have been true, or would have been different from what they are, if our nature had been different in certain ways? To accept this alternative seems no better as an answer to the question than the appeal to the “nature” of numbers and colours that Wittgenstein rejects. Can we accept that if we had been different in certain ways there might have been different truths about numbers and colours, or perhaps no systems of numbers or colours at all? If the necessities that hold in

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



CONCEPTS OF COLOUR AND LIMITS OF UNDERSTANDING

those systems as we now have them are truly necessary, we must acknowledge that they would hold in all circumstances, and so would hold even if we and our nature had been very different from what they are. I think we cannot consistently acknowledge that what we regard as genuine necessities could find their source in us. Nor would their residing in the very “natures” of numbers and colours themselves help explain the necessity of the necessities we recognize either.⁵ The very question itself about the ‘grounds’ of our concepts and their associated necessities is what seems to lead to the troubles and dissatisfactions we encounter. In asking whether or to what extent we can conceive of the use of concepts of number or of colour different from ours we seem unable to find a satisfactory resolution either way, given what we already think and know about numbers and colours. It is the very thoughts we now have about numbers and colours that we wish to have some deeper or more illuminating understanding of. But what exactly is that question about the ‘grounds’ or ‘basis’ of the concepts we know we have? And what could ever put us in a position to answer such a question?

⁵ I have tried to show why neither kind of answer can give a satisfactory account of several kinds of necessity or of value in my Engagement and Metaphysical Dissatisfaction: Modality and Value, Oxford University Press 2011.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

16 Logical Aliens and the ‘Ground’ of Logical Necessity At the beginning of ‘The Search for Alien Thought’¹ James Conant tells us that Descartes thought the laws of logic were only contingently necessary, Aquinas thought they were necessarily necessary, Kant thought they were simply necessary, Frege “wanted to agree” with Kant, but this “raised the worry that there was no way in which to express his agreement that made sense”, and Wittgenstein agreed with that “worry” (115–116). This is where the paper began, and now, twenty years later, how are we to respond to this battle of the giants? What I find most problematic in this array of views is the idea of the laws of logic as “contingently necessary”—or indeed, the idea of anything’s being “contingently necessary”. It could be said in favour of that option that at least it leaves room for the kind of question many philosophers have asked or wanted to ask about necessary truth. And it is the question from which Conant begins. He asks first, “What is the status of the laws of logic?”. That could be asking simply whether they are necessary or contingent, or perhaps how they are known—their “epistemic” status. But he also asks, “Wherein does their necessity lie?”. That seems to be a different question. Something different also seems to be at stake in his asking, pointedly, “In what sense does the negation of a basic law of logic represent an impossibility?” (116). I say he asks “pointedly” because his question seems to ask for something more than the flat-footed answer “the negation of a basic law of logic represents an impossibility in the sense that it represents something that is necessarily false, or could not possibly have been true”. ¹ Philosophical Topics, vol 20, no. 1, 1991 (page numbers in parentheses in the text refer to this essay).

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



LOGICAL ALIENS AND THE ‘ GROUND ’ OF LOGICAL NECESSITY

What Conant’s question apparently seeks is some explanation of the necessity of necessary truths, or of the impossibility of their being false. What seems to be wanted is some account of “the ground” (127) or “source” of their necessity. This is what the idea of necessary truths as only “contingently necessary” seems to leave room for. It appears to countenance the possibility of explaining how or why something is necessary or holds necessarily even though what explains it is something that could have been otherwise, or could have failed to be so. But can we understand the necessity of truths we now regard as necessary as due to something contingent? To do so it looks as if we would have to regard those truths as necessary while at the same time thinking, no doubt from an appropriately detached viewpoint, that although they are necessary, they are only “contingently necessary”. That seems to imply that they would have failed to hold, or failed to hold necessarily, under certain conditions. This is what I find problematic. Thinking that some necessary truths are only contingently necessary would not mean that we have to agree with what Descartes thought accounted for the necessity of necessary truths. But we would have to accept something we see could have been otherwise that nonetheless explains “wherein lies the necessity” of the necessary truths we accept, or what the “ground” or “source” of that necessity is. That explanation, whatever it might be, would presumably be offered as an answer to a certain question about the source or basis or explanation of necessary truth. But is there such a question we can envisage having a satisfactory answer to? What would an illuminating explanation of the necessity of necessary truths look like? And how, exactly, would it explain the necessity? If we can make no sense of having such an explanation, can anything be intelligibly said about the “source” or “ground” of the necessity of the necessary truths we accept? I don’t mean to deny that we can come to understand, and in that sense, explain to ourselves and others why a certain truth holds necessarily. We can come to understand why a certain thing must be so, even if we had not seen or understood that necessity at first. We can understand that it must be so, given that certain other things we know are so. When the college reunion is finally over, for instance, the number of graduates who have shaken hands at the party an odd number of times must be an even number. Given how many people it takes to make a handshake, we can see why the number in question must be even, and we can explain it

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

LOGICAL ALIENS AND THE ‘ GROUND ’ OF LOGICAL NECESSITY



to anyone who considers taking a survey or starts counting each person’s handshakes. But that kind of explanation does not seem possible in trying to explain the necessity of all necessary truths. A completely general explanation would presumably require that there be no unexplained or independent appeal to something already accepted as holding necessarily as part of the explanation of all necessities. The complete generality of the philosophical task appears to rule that out. It rules it out necessarily, one is tempted to add. It is that completely general question about the explanation or source of all necessity that I cannot envisage having a satisfactory answer to. Coming now to Kant and Frege, I have some doubts about what Conant calls the “worry” he thinks Frege faces in expressing what he “wants to agree” with in Kant. I think I see what the difficulty is, and I think I see how Wittgenstein in the Tractatus hopes to avoid it. But I don’t think I see how or why Frege faces that difficulty or has to fall into it. That is a question about the interpretation of Frege. Leaving Frege aside for the moment, I don’t see why such a “worry” must arise for anyone who wants to agree with what Conant says Kant believes. What Kant is said to believe is that the laws of logic are necessary (or “simply necessary”). It is difficult to tell whether Frege agrees or “wanted to agree” with that because Frege never speaks about the modality of the laws of logic at all. They are, he says, “the most general laws of truth”. (“Laws of truth”, not “laws of nature” as Hilary Putnam puts it in a passage Conant quotes twice (127, 139)). So what is it that Frege “wanted to agree” with in Kant? Kant is said to believe that the laws of logic are necessary (or “simply necessary”). That could mean that the laws of logic hold necessarily, or it could mean only that the laws of logic are necessary in the sense of being necessary or indispensable for something—for thought, say, or rational thought (if that is different). To say only that the laws are necessary or required for thought is not, strictly speaking, to say anything one way or the other about the modal status of those laws themselves. Even to say that they are necessary or required for the possibility of thought seems to me to leave open the question whether they themselves hold necessarily. That is not because I think there is any doubt about whether the laws hold necessarily, but because I think the laws’ holding necessarily does not follow from their being necessary for thought. Kant in his pervasive

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



LOGICAL ALIENS AND THE ‘ GROUND ’ OF LOGICAL NECESSITY

use of ‘necessary’ and ‘necessity’ is not always careful about the difference between these two readings. And I think the difference is relevant to the difficulty or “worry” Conant thinks Frege is confronted with in his wish to agree with Kant. Whether there is a difficulty, and if so what it is, depends on which of the two readings of Kant Frege is said to “want to agree” with. Conant says “Frege inherits the Kantian idea that accord with the laws of logic is constitutive of the possibility of thought”, and “therefore inherits (a great deal of) Kant’s philosophical conception of the status of the laws of logic (as constitutive of the possibility of rational thought)” (134). This is the source of the difficulty he thinks Frege falls into because Conant thinks it leads Frege eventually to the conclusion that “there cannot be logical aliens. . . . beings who fit this description are an impossibility” (149) The difficulty is that: This makes it seem as if, following Frege, what we have done is grasped . . . what it would be for beings to be able think in this remarkable way—and subsequently gone on to reject this possibility. We think of ourselves as rejecting the possibility of something: illogical thought . . . we take the sentences “illogical thought is impossible” or “we cannot think illogically” to indeed present us with thoughts . . . The attempt to say that illogical thought is something that cannot be, to say that it involves a transgression of the limits of thought, requires that we be able to draw the limit. (149–150)

If what leads into this quandary is the view that logic is “constitutive of the possibility of thought”, or “the possibility of rational discourse” (148), what reasons are there for attributing that view to Frege? Conant appeals to Frege’s remark that the laws of logic are “the most general laws of thought . . . [which] prescribe universally the way in which one ought to think if one is to think at all” (134). But I don’t see that passage as making a claim about something’s being “constitutive of the possibility of thought”. In the course of explaining the sense in which the laws of logic can be described as laws of thought, Frege says they prescribe or stipulate “the way in which one ought to think”, but he adds: Any law asserting what is, can be conceived as prescribing that one ought to think in conformity with it, and is thus in that sense a law of thought. This holds for laws of geometry and physics no less than for laws of logic. The

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

LOGICAL ALIENS AND THE ‘ GROUND ’ OF LOGICAL NECESSITY



latter have a special title to the name “laws of thought” only if we mean to assert that they are the most general laws, which prescribe universally the way in which one ought to think if one is to think at all.²

So what is special about laws of logic among all other laws that assert what is so is only that they “prescribe universally”; they are laws of truth or laws of thought that prescribe how one ought to think about any subject-matter whatsoever. Only in that sense are they “laws of thought” in any distinctive way, different from other laws that assert what is so. I don’t think this supports the idea that the laws of logic are in any other sense “constitutive of the possibility of thought”. In further support of attributing this ‘Kantian’ view to Frege, Conant says, “The absolute generality of the laws of logic, for Frege, is tied to their ultimate ground in pure thought alone” (134). This raises the question what such a “ground” or “ultimate ground” of the laws of logic would be. Is it something that would explain or account for the distinctive character or special modality of the laws of logic? And what would it be for that “ultimate ground” to be revealed as lying in “pure thought alone”? Is Frege concerned with these difficult, obscure questions? Frege does speak of the “ground” or even the “ultimate ground” of certain truths. A truth counts as analytic, he says, if “the ultimate ground upon which rests the justification for holding it to be true” is only laws of logic (and perhaps definitions).³ Laws of logic alone are the “ultimate ground” of analytic truths, but that says nothing about the “ground” or justification of laws of logic themselves, or even our “ground” for holding them to be true. “The question why and with what right we acknowledge a law of logic to be true,” Frege says, “logic can answer only by reducing it to another law of logic” (FBL 15). That leaves the “ground” or justification of those laws themselves so far unexplained. It says nothing to suggest that there even could be such an explanation. Nor does it suggest that the “ultimate ground” of the laws of logic, if there is one, lies in “pure thought alone”. The idea that their “ultimate ground” lies in “pure thought alone” or that they are “constitutive of the possibility of thought” is what seems to ² G. Frege, The Basic Laws of Arithmetic (tr. M. Furth), University of California Press 1967, p. 12 (page numbers with ‘FBL’ in parentheses in the text refer to this volume). ³ G. Frege, Foundations of Arithmetic (tr. J. L. Austin), Blackwell 1953, pp. 3–4 (page numbers with ‘FFA’ in parentheses in the text refer to this volume).

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



LOGICAL ALIENS AND THE ‘ GROUND ’ OF LOGICAL NECESSITY

lead into the quandary about the laws of logic. It can perhaps make it seem simply impossible to think (or think rationally) without accepting or abiding by those laws, that genuine thought could not possibly violate them. This is the way I think Conant understands Frege’s opposition to the so-called psychologistic logicians. Those ‘logicians’ hold that some men or other beings could be capable of judgements contradicting our laws of logic. And I think Conant sees Frege as wanting to argue in opposition that such beings are simply impossible; there could be no such logical aliens. Conant calls this “the conclusion of [Frege’s] argument against psychologism” (149). And the “worry” or difficulty he sees is that in reaching that conclusion and “affirming the negation” (149) of the psychologistic thesis Frege would have had to entertain that rejected doctrine as at least a judgeable content of thought. He would have to have had a thought that the conclusion of his own argument says is not a possible thought. He would have had to transcend what his view holds to be the limits of thought in order to think and assert that those limits lie where they do. Frege certainly opposes any such psychologistic conception of logic. But there is a question whether his opposition takes the form of an attempted demonstration of the impossibility of anyone’s having thoughts that contradict our laws of logic. Frege makes it clear that the real focus of his dispute with psychologism is the proper understanding of the idea of truth (FBL 14). The psychologistic logician appears to hold that the sense of the word ‘true’ includes a reference to the subjects who judge; but for Frege “one could scarcely falsify the sense of the word ‘true’ more mischievously” (FBL 14). That “falsification” of the idea of truth is presumably extended by the psychologistic logician to all judgements about anything whatever, not just laws of logic. But then there would appear to be no judging at all of anything objective or independent of the judging subject. Frege goes into this matter at considerable length in his attack on the views of Herr B. Erdmann (FBL 12–25). I do not see Frege’s opposition to psychologism as an attempt to demonstrate the negation of the psychologistic logician’s doctrine. Far from claiming that thought would be simply impossible without acceptance of our laws of logic, Frege appears to remain non-committal on that question. He admits that: If we step away from logic, we may say: we are compelled to make judgments by our own nature and by external circumstances; and if we do so, we cannot reject this

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

LOGICAL ALIENS AND THE ‘ GROUND ’ OF LOGICAL NECESSITY



law—of Identity, for example; we must acknowledge it unless we wish to reduce our thought to confusion and finally renounce all judgment whatever. I shall neither dispute nor support this view; I shall merely remark that what we have here is not a logical consequence. (FBL 15)

He goes further. He points out that even if it is impossible for us to reject the law of identity, for instance: this impossibility of our rejecting the law in question hinders us not at all in supposing beings who do reject it; where it hinders us is in supposing that these beings are right in so doing, it hinders us in having doubts whether we or they are right. (FBL 15)

This appears to grant the possibility of beings who do not accept the law of identity, for instance. We can suppose that there could be such people, or at least we are not in a position to deny it. But the question is what we could make of such beings, whether we could understand what they think and do. In response to the alleged discovery of beings “whose laws of thought flatly contradicted ours and therefore led to contrary results even in practice”, Frege tells us he would say: “we have here a hitherto unknown type of madness” (FBL 14). I take Frege to be saying that he could make no sense of those beings or of what they are doing. We are familiar with certain kinds of madness, and we think we sometimes have some idea of what those afflicted with it are up to. But to speak of a new and hitherto unknown type of madness suggests that in this case we are simply baffled and can make no real or lasting sense of its victims. Conant sees Frege’s invocation of “madness” as his reaching for a positive description of those who violate the laws of logic—something about them that might explain the “source” of the distinctive character or authority of those laws, or perhaps even reveal the “ultimate ground” of those laws in “pure thought alone”. But Frege’s use of the term “madness” does not seem to offer any positive explanation along those lines. In the end it is nothing more than what Conant calls a “philosophically innocuous idea” of “merely lunatic thought” (149)—as if we knew what that is. In his Remarks on the Foundations of Mathematics Wittgenstein presents many examples of what appear to be deviant ways of inferring, calculating, weighing things, or even selling wood. He quotes Frege’s remark about the apparent madness we encounter in such examples and

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



LOGICAL ALIENS AND THE ‘ GROUND ’ OF LOGICAL NECESSITY

observes that “Frege never said what this ‘madness’ would really be like”.⁴ I think there is good reason for that. When we press the descriptions of those apparently deviant ways of proceeding far enough, and try to get a consistent picture of them, I think we cannot really make full sense of such people. Up to a point, we can give them the benefit of the doubt, as we say. Or better, we can give them the benefit of humanity. But beyond a certain point in trying to conceive of such beings they become more and more impenetrable to us. As Wittgenstein puts it in another connection, “We cannot find our feet with them” [Or perhaps “We cannot find ourselves in them”].⁵ If this is right, it could explain Frege’s finding nothing more than a “hitherto unknown type of madness” in beings said to think in accord with laws of logic that flatly contradict our own. Even if we find the possibility of people like that more and more mysterious as the description of their practices is expanded, it does not mean we can make no sense at all of “illogical thought”. That is something we can recognize from time to time, in particular isolated invalid inferences, or in someone’s accepting the falsity of a logical law. Even the idea that the laws of logic are “constitutive of the possibility of thought” must presumably allow for occasional deviance. So I don’t think Frege’s attitude to “illogical thought” forces him into the difficulty or “worry” Conant describes. That difficulty is said to arise because if “there isn’t any sense to be made of the idea of someone (even God!) entertaining the falsity of a basic logical law”, Frege’s “account of judgment fails to leave room for anything which could count as judging a basic logical law to be true” (140). But is that really a difficulty for Frege? Surely Frege could recognize cases of logically fallacious reasoning when he sees them. He could see that the thinker had violated a general law of truth, and so could not be right in what he thinks. Frege himself at one time accepted something as a law of logic that is not true. He committed himself to the truth of his Basic Law V in the course of deducing the truths of arithmetic from truths of logic alone. He certainly acknowledged, and so made sense of, his having judged or put that Law forward as true, even though it is false and he came to see that it is.

⁴ L. Wittgenstein, Remarks on the Foundations of Mathematics (tr. G. E. M. Anscombe), MIT Press 1978, I, 152. ⁵ L. Wittgenstein, Philosophical Investigations (tr. G. E. M. Anscombe, P. M. S. Hacker, J. Schulte), Wiley-Blackwell 2009, Philosophy of Psychology: A Fragment }325.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

LOGICAL ALIENS AND THE ‘ GROUND ’ OF LOGICAL NECESSITY



So I don’t see a difficulty for Frege in making sense of someone’s judging a law of logic to be true when it is true. Someone who deduces ‘Socrates is mortal’ from ‘All men are mortal’ and ‘Socrates is a man’ could be understood to be doing just that. What Frege cannot grant is that someone he recognizes to be engaging in “illogical thought”, or accepting the falsity of a law of logic, is right in thinking as he does. In that sense, Frege could be said to accept the idea that the laws of logic are “constitutive of the possibility of rational (or correct) thought”. But that is just to say that the laws of logic are laws of truth. And what the “illogical” thinker accepts is in violation of the truth. That is something we can say about anyone to whom we attribute false beliefs, whatever the subject-matter of those beliefs happens to be. In general, the more falsity we have to attribute to thinkers’ beliefs in making sense of them, the more ‘deviant’ or difficult to understand they become. But because the laws of logic are universal laws of truth for all thinking about anything, attributing falsity to the logical beliefs of others threatens more quickly to render them more completely mysterious and impenetrable than attributing falsity to their other less general beliefs. With laws of logic, the most general truths of all, there is less room for tolerable but still intelligible deviance. We must attribute truth as much as possible to those whose thinking we are trying to make sense of because we are, perforce, trying to understand those others as part of the very world we also inhabit and understand. We cannot do that without ourselves accepting a great many things as true; that is what it is for us to have a world within which we can understand anything, including the thoughts of others. And given the complete generality of the laws of logic, the range of intelligibly deviant attribution is smallest with them. Does this show or suggest that the “ultimate ground” of the laws of logic lies “in pure thought alone”? I can’t see it does. Does it even show or suggest that there is such a thing as the “ultimate ground” of the truths of logic? Not having an answer to that question does not seem to prevent us from agreeing that the laws of logic are necessary in the sense of holding necessarily, with no possibility of their being or having been false. That is one way of taking what Kant is said to have believed. I mentioned that Frege appears to take no stand on the question whether the laws of logic are necessary in that sense. But I assume (perhaps rashly) that if Frege accepted a law of logic to the effect that if something P is true then

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



LOGICAL ALIENS AND THE ‘ GROUND ’ OF LOGICAL NECESSITY

something Q is true, he would not be reluctant to say that if P is true then Q must be true. I think we have to grant that some truths are necessary in this sense. But that idea of necessity, for those who accept it, does not seem to me to offer much promise of a satisfactory explanation of that distinctive modal status. I return to the question of what such an explanation would be. I take it that modal terms like ‘necessary’ and ‘possible’ are irreducible to others. There is no prospect of simply defining the idea of necessity or reducing it to some equivalent but fully non-modal terms. That is not to be expected. If the laws of logic are thought of as holding necessarily, is there some prospect of understanding the nature or source of that necessity in the thought that the laws of logic are “constitutive of the possibility of thought”? To say they are “constitutive of the possibility of rational or correct thought” would be to say only that they are laws of truth; they prescribe that one ought to think in conformity with them. But all truths are in that sense “constitutive of the possibility of correct thought”. Only those thoughts that are in accord with truths are true or correct thoughts. Saying that the laws of logic are “constitutive of the possibility of thought” could mean that the laws of logic state what must be so if thought is to be possible—there could be no thought, or no possibility of thought, if those laws were not true. One obstacle to explaining necessity in this way is that this is itself a claim of necessity: that the truths stated by the laws of logic are themselves necessary conditions of the possibility of thought. Even if that is true, it does not seem to help explain or account for the necessity of all necessary truths. It makes use of the very idea of necessity that is meant to be explained. This raises again the question what sort of explanation we are looking for, or what a fully satisfactory explanation of necessity is expected to do. Even leaving aside for the moment the question of the scope of the desired explanation of necessity, there is a question whether the truth of the laws of logic being a necessary condition of thought would really help explain the necessity of those laws, if they do indeed hold necessarily. Some of the necessary conditions of thought, or of the possibility of thought, it seems, hold only contingently. That there are thinkers is, I suppose, a necessary condition of there being thought, or the possibility of thought. But that there are any thinkers at all is something that could

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

LOGICAL ALIENS AND THE ‘ GROUND ’ OF LOGICAL NECESSITY



have been otherwise. Certainly each thinker can recognize that ‘I think’ is a necessary condition of his or her thinking anything, but that is not to recognize that what the thinker thereby sees to be true of himself is something that holds necessarily. The same for ‘I exist’, which is a necessary condition of ‘I think’, and so of ‘There are thinkers’ for each thinker. But that does not hold necessarily either. It also seems necessary for the possibility of thought that thinkers, or potential thinkers, have certain capacities and abilities. But whether there are or even could be creatures with such capacities again seems to be something contingent. In any case, the general point is that whether a truth is a necessary condition of thought or of the possibility of thought is not the same as the question whether that truth itself holds necessarily. One truth can be a necessary condition of another truth while both hold only contingently. That there is any thought at all seems to me to be something contingent; things could have been otherwise. Since there is thought, it must be possible for there to be thought. But is even that possibility of thought also something contingent; something that could have failed to be so? These are heady matters. I think it is safe to say, more modestly, that something’s being a necessary condition of the possibility of thought does not in itself explain or account for its holding necessarily, if it does. This distinction is obscured in the richly-developed Kantian reasoning to the effect that, for instance, there being thinking subjects with certain capacities and enduring objects with causal powers in space and time is a necessary condition of the possibility of thought and experience of an independent world, and so a condition of the possibility of thought itself. If that means that that complex necessary condition is something that holds necessarily, it would imply that it is true with necessity, with no possibility of its being or having been otherwise, that the universe contains thinking subjects and enduring objects with causal powers in space and time. There are indeed such thinkers and such objects in the universe as things actually are. Could things possibly have been different in some or all of those respects? If you think so, you must agree that something Kant thought is a necessary condition of the possibility of thought and experience, even if it is such a necessary condition, is not itself something that holds necessarily. So its having the privileged status of being a necessary condition of the possibility of thought is not the same as, nor does it explain, its holding necessarily.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



LOGICAL ALIENS AND THE ‘ GROUND ’ OF LOGICAL NECESSITY

It was a conclusion of Kantian metaphysics that there being a world of thinking subjects and enduring objects with causal powers in space and time, and so on, is not only the way things actually are, but the way the world must be. That seemed to involve a step from its being necessary for the possibility of thought and experience that thinkers think of and experience the world in certain ways, to the apparently stronger modal conclusion that that is the way the world must be. And to reach that stronger conclusion from the necessity of thinkers’ thinking of and experiencing the world in certain ways, Kant relied on his transcendental idealism. That doctrine would presumably explain or make intelligible to us how what we must think in order to think or experience anything at all must therefore also be true and must hold necessarily. If we could understand and accept this transcendental idealism, perhaps we could then understand and explain to ourselves the “ground” or “source” of the necessity we ascribe to the necessary truths we accept. Perhaps we could even come to see the “ultimate ground” of that necessity to lie in “pure thought alone”. I always find myself at least one step short of the illumination promised by that Kantian idealist doctrine. But I think all of us can appreciate the great force of the Kantian considerations that seem to reveal certain ways we must think or believe the world to be in order to think of a world at all, or even in order to think. We can perhaps see that our acceptance of certain things is a necessary condition of the possibility of thought. But if the laws of logic that we regard as holding necessarily are shown to be conditions of the possibility of thought in this sense, it would mean only that we must think certain truths hold necessarily if we are to think at all. That seems plausible to me; it is hard to see how we could move securely from one thought or belief to another without at least implicitly thinking of the two as necessarily connected. But even if the acceptance of some truths as necessary plays an essential role in all thinking, that would not explain the necessity of those truths we regard as necessary. Nor would it explain the “ground” or “source” of that necessity. This is what lies behind my doubts about the project of accounting for, or in some sense explaining, the necessity of necessary truths. Either we take the notion of necessity for granted in attempting to explain it in general, and so are left dissatisfied, or we try to avoid using it in explaining the “source” of necessity, and so fail to explain it at all. Or worse, we try to explain the necessity of necessary truths by appeal only

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

LOGICAL ALIENS AND THE ‘ GROUND ’ OF LOGICAL NECESSITY



to something contingent. The “psychologistic logicians” Frege has in mind grant that we cannot help taking certain laws of logic to be true, but they hold that they are only something we take to be true. Of this doctrine Frege observes that those ‘logicians’ thereby “presume to acknowledge and doubt a law in the same breath”, and this, he says, “seems to me an attempt to jump out of one’s own skin against which I can do no more than urgently warn them” (FBL 15). That same metaphor seems applicable to those who would account for the necessity of the truths we regard as necessary by appeal to something that could have been otherwise. They would appear both to acknowledge and to deny a necessity in the same breath. They could not consistently accept both that the truths hold necessarily and that if certain things had been otherwise they would not have held necessarily and perhaps would not even have been true. It is no doubt disappointing to leave the matter here. We find necessity philosophically puzzling or suspect. We think there must be some way to understand it or explain it or “ground” it in something more intelligible. It seems deeply unsatisfying simply to acquiesce in it as a basic part of our repertoire or as just part of the way things are. Necessity appears problematic or suspicious to us in comparison with other apparently simpler, more basic ways of thinking. I understand the feeling, and the dissatisfaction. But I think that very puzzlement or suspicion is what we should now focus on and try to get to the bottom of. What is that feeling of philosophical mystery about necessity? And where does it come from? Even if we cannot satisfy the desire or demand for a more satisfying philosophical understanding of it, we should not be led to deny or distort the idea of necessity that we actually have or the ways we think in modal terms.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

17 Ways of Meaning and Knowing Moral Realities I would like to take up in rather general terms some of the important and distinctive features of Mark Platts’ explanation and defence of his views of morality. Not his own personal moral views, which he warns us we are better off not knowing anything about. I mean his views of the nature of morality and of the kind of knowledge and understanding human beings have of it. I find those views and his defence of them very congenial, and I am happy to have a chance to pay tribute to them and to their author here. I will eventually draw attention to one place in his reflections where Platts speaks of a certain resistance one might feel to the kind of view he has been defending. He considers a way of accommodating the source of that resistance, and so in that way overcoming it. I raise the question whether what he suggests is really a successful response to the resistance he has in mind. That turns out to be a more complicated matter than it looks, and I will try to explain why. It depends on what the resistance Mark identifies actually is, and where it comes from. That raises some large, difficult questions in turn. In his Ways of Meaning,¹ Platts’ aim was “to present and discuss . . . the most important recent contributions to the philosophy of language” (WM xi). The most important recent contributions to that subject through the 1970s were Donald Davidson’s elaborations of the idea of a theory of meaning for a particular language. What came to be called a Davidsonic boom had been resounding through Oxford and beyond in those years, and Ways of Meaning tells us what it was all about. And it ¹ Mark Platts, Ways of Meaning, Routledge 1979. (Page numbers with WM in parentheses in the text refer to this book.)

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

WAYS OF MEANING AND KNOWING MORAL REALITIES



tells it, as it aspired to, “in a straightforward and accessible manner”. That book was the best and most thorough book-length treatment of that whole subject up to that time. It explained and developed the basic ideas of the new theory of meaning and investigated its implications as an account of a speaker’s understanding of the words of his language: what a speaker who knows English, or Spanish, knows simply by being a competent speaker of that language. Davidson’s idea was that a speaker of a language can be said to know the conditions under which each of the indicative sentences he understands would be true. The task of what he called a theory of meaning for a particular language was then to give a systematic description of the structure of the language in a way that would yield, for each sentence, a statement of what the truth-conditions of that sentence are. If all those statements about sentences of the language are derivable from the structure of the language itself they would all express something that someone who understands that language can be said to know. A speaker of English, for instance, knows that the sentence, “Ripe tomatoes are red” is true if and only if ripe tomatoes are red. It seems very hard to deny that a speaker of English does know such a thing, and that his knowing it is involved in his understanding the English words in the sentence. But however undeniable that might be, it was regarded by many philosophers at the time as completely inadequate as an account of a speaker’s understanding of the meanings of his words. It was thought to offer no explanation of how a competent speaker knows what he knows, or how his mere knowledge of the truth of such biconditional statements ‘issues in’ or is connected with the actual applications and responses he makes in real life with the words he is said to understand. Behind this objection was the thought that a speaker’s use of her words requires the grasp of certain ‘procedures’ or ‘grounds’ on the basis of which she applies or withholds her words on particular occasions. If some such ‘recognitional procedures’ or capacities are required for a speaker to use and understand her words, it was thought that they must be an essential part of any account of the meanings of those words. And a theory of meaning in terms of truth-conditions alone says nothing about that aspect of a speaker’s competence. Platts did everything he could in that book—short of the “evangelism” he assures us he would never stoop to—to disarm this line of resistance to the truth-conditional theory. And what he and others did was as much

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



WAYS OF MEANING AND KNOWING MORAL REALITIES

as I think should have been needed. The resistance came largely from those of a broadly verificationist persuasion, or those with a basically “epistemic” conception of truth. Platts accordingly called the kind of theory he defended a “realist” account of a speaker’s understanding, and a “realist” conception of meaning and truth. One can perhaps see why. But just how robustly metaphysical that talk of ‘realism’ was meant to be is not easy to say. I think this is connected with the difficulty I will consider. Platts calls the theory he defends a form of “realism”. That is presumably meant to imply more than the evident fact that the conditions under which each of the sentences a speaker understands would be true can be stated in the very words used in the sentences the speaker is said to understand. That is an important feature of the theory, but there seems to be nothing distinctively “realist” about it. It means that anyone who understands the words used on the right-hand side of the relevant biconditional statement will know the conditions under which the sentence mentioned on the left-hand side is true. What enables a speaker to know such a thing is the speaker’s competence in the language used in the biconditional statement. That is the language the theory of meaning is expressed in. Someone who does not understand that language could not learn from such biconditionals the truth-conditions of their mentioned sentences. This shows that a theory of meaning of this kind is not meant to provide a way into a language for someone who does not already speak and understand a language; it does not explain new words or concepts to someone who does not already have them. A theory of this kind could state the truth-conditions of sentences of one language in the words of another language. But only someone who already understands that other language could learn from that theory of meaning what competent speakers of the first language know and understand. This means that speakers must rely on the understanding they already have of some language to know what a theory of meaning for that language or for any other language says. A theory of this kind is perhaps more appropriately called “realist” in contrast with the “anti-realism” or “non-realism” of verificationist or “justificationist” views. Unlike those views, a theory of meaning of the kind Platts defends takes no account of any alleged “recognitional procedures” speakers make use of in applying their words or in determining the actual truth-values of the sentences they understand. The theory is in that sense also what he calls “austere”; it offers no explanation of

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

WAYS OF MEANING AND KNOWING MORAL REALITIES



speakers’ application of their words and no ‘basis’ or ‘ground’ of their knowledge of the meanings of words. The main point in the label “realism” is that the theory: assumes that sentences can be true (or false) independently of our capacity, or incapacity, to recognise them as true or false . . . that we can know what it is for a sentence to be true or false . . . even if it is beyond our capacities to recognise whether those truth-conditions apply or not”. (WM 5–6)

I think there is no question that this is the way Donald Davidson understood the implications of the kind of theory he had in mind. As far as I know, Mark Platts was one of the first philosophers to publicly and explicitly take up the question of the extent to which a theory of meaning and understanding of this kind is applicable or can be extended to the domain of evaluative language. He had in mind specifically morality, and he certainly raised the question before Donald Davidson himself had begun to discuss the issue publicly. David Wiggins and John McDowell were already working in nearby territory. I believe they both had an influence on Platts’ thinking at that time, perhaps even in the direction I am going to question.² In the early pages of Ways of Meaning, he expresses some doubts about the plausibility of extending the application of what he calls his “realistic reading . . . of the truth-predicate” to “an ethical sentence” (WM 12). But in the last chapter of that book, “Moral Reality”, he begins to go into the question directly. It cannot be said that he carries the investigation of that question very far in that chapter. But he insists that to apply the truth-conditional theory of meaning and understanding to moral sentences we must honour what both “utilitarianism” and “intuitionism” would preserve. Both those theories are “realist” in seeing ethical judgements as either true or false, and intuitionism is also “austerely realistic” in rejecting ‘justificationist’ demands for ‘explanatory’ accounts of conditions or procedures of application of moral terms, or for reductions of their contents to non-moral bases. The point to be preserved in any case is that “A speaker can know, have a grasp of, the truth-conditions of a moral sentence even if those ² Both Wiggins and McDowell were developing versions of what Wiggins first called “a sensible subjectivism” and then, after further reflection, “a sensible subjectivism?”. See ‘A Sensible Subjectivism?’ in D. Wiggins, Needs, Values, Truth, Oxford University Press 1998.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



WAYS OF MEANING AND KNOWING MORAL REALITIES

truth-conditions are beyond his (present) recognitional capacities” (WM 245). When this kind of account is generalized, it would mean that: By the process of careful attention to the world, we can improve our moral beliefs about the world, make them more approximately true; by the same process, we can improve our practical understanding, our sensitivity to the presence of instances of the moral concepts that figure in these beliefs. . . . Our moral language, like all the realistic part of that language, transcends our present practical comprehensions in trying to grapple with an independent, indefinitely complex reality. (WM 247)

This is a good expression of what the successful application of the truthconditional theory of meaning and understanding to moral language would give us. I think it captures the ways we actually think of moral assessment and helps explain whatever sense we have of discernible moral progress. Platts pursues this apparently intelligible goal in his Moral Realities.³ In that book he explains how a proper defence of the view depends on the correct understanding of what he calls, broadly speaking, “desires”— those “active” or “practical” attitudes that are essential to all intentional action. This leads into an intricate discussion of such attitudes, always with an eye on this central question of motivation and reasons. For our purposes here we can focus on one particular set of attitudes he draws attention to: so-called “motivated” or “reason-following” desires, as opposed to other so-called “unmotivated” or “reason-producing” desires. These latter are desires or wants we just find ourselves with which seem in their very presence to give us some reason to do something that we did not have reason to do before. “Motivated” or “reason following” desires, on the other hand, are present when an agent is moved or can be said to “want” to act in a certain way, not because he just feels like it or because it is a way of satisfying some desire he independently happens to have, but because he recognizes the desirability or value in acting in that way—either in its outcome or in the performance itself. The agent sees and takes something he recognizes to be true of the action or its outcome as reason to do

³ Mark Platts, Moral Realities, Routledge 1991. (Page numbers with ‘MR’ in parentheses in the text refer to this book.)

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

WAYS OF MEANING AND KNOWING MORAL REALITIES



it, and is so motivated by that recognition. For “desires” or attitudes of this distinctive kind, as Platts puts it: “The agent does what he does because he thinks that something—say, honesty, or scientific truth—is a value, or is of value, or matters, or is desirable, or is worthy of desire” (MR 76). The point is that, in such cases the agent acts, or is moved to act, as he does because he thinks or believes something about the action or object in question; he sees some feature he takes the action or object to have as reason to act in that way, and he is moved to act in the way he sees there is reason to act. The agent’s thought or belief in such a case could be called an “evaluative” thought about something. But an agent’s having such an “evaluative” thought or belief does not imply that the value she sees in the action or object is, or is thought to be, dependent on her or anyone else’s having that “evaluative” thought or belief, or “valuing” the thing in the way she does. The reason the agent sees in favour of doing such-andsuch is a reason she takes to hold of the action she considers, and to hold of it whether or not she or anyone else actually thinks it holds. This is, after all, a familiar aspect of propositional thought in general, about virtually any subject-matter. In the thought that ripe tomatoes are red, for instance, what one thinks is something that holds, or fails to hold, and is thought to hold or fail to hold, quite independently of whether anyone, including the present thinker, actually thinks that ripe tomatoes are red. If the same is true of evaluative thought, then whatever value- or desirability-characteristic an agent takes an action to have is something thought to be true of or present in that action independently of whether anyone, including the agent, actually thinks that characteristic is present, or even takes that characteristic to be a desirability-characteristic at all. Platts sums up his view of so-called “motivated” or “reason-following” valuings or desirings in this way: in these cases valuing can count as a case of being presented with some value-property or ‘quality in objects’ which is there anyway and which is at least worthy of being valued independently of actual human desirings and valuings . . . an agent’s desiring the thing concerned can be at least a reasonable response to the value independently had by the thing . . . (MR 95)

Having ably explained and defended this view, Platts then considers a certain objection he thinks many would make against it: the charge of

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



WAYS OF MEANING AND KNOWING MORAL REALITIES

what he calls “transcendentalism”. Here we approach the source of my doubts. Platts puts the objection like this: In view of what has been said in defence of the idea that . . . such valuing can count as a cognitive response to some value-property or value-feature which is independently there in objects without the mind, the thought might be invited that this account holds the existence of such value-properties or value-features . . . to be completely independent of human valuings and desirings. And that might be judged rather hard to swallow. (MR 98)

This is the charge or reaction Platts calls “transcendentalism”. Whatever it is called, it is an expression of resistance to the idea that the properties we ascribe to objects or actions in our evaluative judgements about them are independent of, or transcend, actual human valuings and desirings. Platts suggests that might be thought hard to swallow because it appears to be incompatible with what many regard as an obvious truth: that “actual human valuings and desirings are the source of all value” (MR 95). I find this last idea (“the source of all value”) a very slippery thought. We can begin to get a firmer grip on it by asking whether the thought is obviously true, or even true, in a way that threatens the idea of something’s having a certain value independently of anyone’s valuing or desiring it. There seems to be one way in which that slippery thought represents no threat to that form of independence, as we have just seen. Ripe tomatoes are red (or not) independently of whether anyone thinks they are red, so someone who thinks that ripe tomatoes are red thinks something that is true (or false) independently of whether he or anyone else thinks it. Of course, if a certain person thinks that ripe tomatoes are red, then someone does actually think it, but what that person then thinks does not itself imply that he thinks it or that anyone else does. Similarly, it seems, someone who thinks that slavery is wrong or deplorable can think it is wrong or deplorable independently of whether he or anyone actually thinks it is wrong, even if everyone (except maybe the slaves) actually thinks it is a wonderful idea. Of course, someone who condemns slavery does in fact make that particular evaluation of slavery, but what he thereby thinks about slavery does not imply that he or anyone else makes that or any other judgement about it. Platts gives other examples of valuings that are understood to hold independently of whether anyone actually makes such valuings or even would make them.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

WAYS OF MEANING AND KNOWING MORAL REALITIES



How, then, does the complaint of “transcendentalism” represent a threat or challenge to the idea that the value-properties we ascribe to objects in evaluative judgements can hold of them independently of all actual human valuings or desirings? What exactly makes that thought of independence hard to swallow, if it is? Platts does not try to answer that question directly. His first step in trying to “free” his view from what he calls “its appearance of (a certain kind of) ‘transcendentalism’ ” is to consider the well-known view of Hume that “Vice and virtue . . . 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”.⁴ The introduction of this view at this point is somewhat puzzling. Of course, Platts does not recommend the full Humean doctrine of values as something like so-called “secondary” qualities. That is precisely the kind of view he opposes. For his particular purpose here he means to endorse only that aspect of the view that says that so-called “secondary qualities” are dispositional properties. They are in effect dispositions objects have to produce certain responses in certain kinds of conscious subjects under certain specified conditions. Let us not ask for the moment what those responses might be, or under what conditions they would be produced. The point so far is that the ascription of such qualities to something takes a conditional form: the object would produce such-and-such effects if such-and-such conditions were fulfilled. Qualities of this kind are of course just as much qualities of the objects they are truly ascribed to as are all the categorical or non-conditional qualities objects also possess. So, this might look like a way of saving the idea that objects can have evaluative properties independently of whether anyone actually responds to them in any of the relevant ways. For objects to have evaluative properties in this way it would be sufficient for the objects simply to be disposed to produce valuings of certain kinds in certain kinds of subjects, even if none of those subjects ever actually makes any of the relevant valuings. Objects’ having value-properties would be in that sense independent of any actual human valuings or desirings in just the way something can have the dispositional property

⁴ D. Hume, A Treatise of Human Nature (ed. L. A. Selby-Bigge), Oxford University Press 1958, p. 469.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



WAYS OF MEANING AND KNOWING MORAL REALITIES

of being soluble in water independently of whether it is ever actually put in water or dissolves. Is this a way of defending the view that valuing is a “cognitive response to some value-property . . . which is independently there” against the charge of “transcendentalism”? The answer obviously depends on what that charge really is, or what its source is. I think there is a way of understanding the charge of “transcendentalism” in which the answer is, “No”; this move does not overcome the objection. Someone who objects to Platts’ view on grounds of its alleged “transcendentalism” could grant that the objects that affect us in evaluative ways have dispositional properties. We could be disposed to respond to objects in certain evaluative ways independently of our actually making any such evaluative judgements or “valuings”, even if “actual human valuings and desirings are the source of all value”. Platts’ view is not just that we can be disposed to make evaluative judgements even if no one ever in fact actually makes those judgements. His view is that the evaluative judgements we do or would make can be “cognitive responses” to some value-properties we think an object possesses independently of anyone’s either making or being disposed to make evaluative judgements about it. I take this to be what those who make the charge of “transcendentalism” object to. The objection is not that objects could not have dispositions to produce certain kinds of evaluative responses without someone’s actually making any of the evaluative judgements in question. The objection is presumably directed against the very conception of evaluative judgements or “valuings” of the kind Platts says we are disposed to make. I think Platts is right to understand evaluative judgements or valuings in that way. What, then, is the objectionable “transcendentalism” or transcendence that the objection complains is present in that conception of evaluative judgement? This is not a simple question. When Hume said, “Vice and virtue . . . 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”, he was invoking the “modern” view that the sounds, colours, heat and cold that we believe in can be understood only in terms of certain effects things in the world have on conscious subjects. The idea was that without such effects on perceivers, there is nothing in the independent world for sounds, colours, and so on to be. In drawing the parallel with values, Hume meant that there is nothing in the world, nothing in

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

WAYS OF MEANING AND KNOWING MORAL REALITIES



“objects as they really stand in nature, without addition or diminution”,⁵ for evaluative terms like ‘vice’ and ‘virtue’, ‘good’ and ‘bad’, and so on, to stand for or mean. Hume thought all such terms can be understood only as involving reference to certain kinds of responses conscious subjects would have to objects that affect them. Any other way of trying to understand those terms would involve something evaluative that is completely independent of, or transcends, all such possible responses. And what could that be? The objection does not seek an answer. The point is that this can be only a rhetorical question. Suppose we ask why those “modern” philosophers thought there is nothing in the independent world for terms for colours, sounds, and so on to pick out. Why did they think there could be nothing in the world independent of all possible human responses for sounds and colours and so on to be? I think that is actually a very difficult question that has still not been adequately answered. I suspect certain metaphysical preconceptions were at work; fixed ideas about the only kinds of things there could possibly be in reality. We need not go into that now. But in the case of values, when we ask, ‘Why do Hume and so many other excellent philosophers think there is simply nothing in the world independent of all possible human responses for values or value-properties to be?’ there is thought to be a convincing answer to the question. And acceptance of a certain version of that answer is what I think is the source of the charge of “transcendentalism” against views of evaluative attitudes of the kind Platts defends. The answer to the question that supports subjectivism about values has two parts. The first, which is the more convincing part, is that evaluative judgements or “valuings” must be understood as “active” or “productive” attitudes that play an essential role in human action. In order to act in one way rather than another an agent must prefer or assess or “value” that way of acting more highly than others. If a potential agent were completely indifferent among all available possibilities, with no preference for or evaluation in favour of any option over any others, there would be no intentional action, and so in that sense no agency. So, it looks as if agents’ attitudes can be understood as evaluative only in relation to their potential role in action. There could be nothing in the world for evaluative or value-properties to be that was not connected in ⁵ D. Hume, Enquiries Concerning Human Understanding and Concerning the Principles of Morals (ed. L. A. Selby-Bigge), Oxford University Press 1966, p. 290.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



WAYS OF MEANING AND KNOWING MORAL REALITIES

some way or other with possible intentional actions. That is the first step, and it seems convincing. It is in the second part of the answer, which is meant to explain exactly how evaluative attitudes are to be understood as connected with action, that there is room for resistance or disagreement. Not all versions of this second part of the story are equally convincing. For Hume and countless other philosophers, agents who act intentionally must be understood to have certain kinds of feelings or passions or sentiments. That is because only something like feelings, passions, or sentiments were thought to be capable of bringing about actions. Reason, or the discovery of truth alone, without any “addition” of anything else, as Hume put it, is “utterly impotent in this particular”.⁶ Belief in the truth of something, all on its own, could never bring about action. Because feelings or sentiments must be present for action, all “active” evaluative attitudes or “valuings” must involve the presence of feelings or sentiments. Such feelings or sentiments will be the evaluative attitudes of a potential agent only in so far as they “give” to objects or possible actions that an agent considers the only value or value-properties they can have for her. So, in this view, no object or action would have more value for an agent than anything else unless there were feelings or sentiments inclining her in one direction rather than another. Without some such feelings, she would find no more reason or inclination to do one thing rather than another thing, even, as we say, doing nothing. I think this idea of desires as feelings or sentiments that move one to act is what gives force to what I called the “slippery thought” that “actual human valuings and desirings are the source of all value”. The idea is that it is only our actually having certain desires and sentiments that ever gives us any reason to do the things we do. Those desires and “active” feelings are in that sense the “source” of all motivation, and so the source of whatever value we can find objects or actions to have. Platts suggests that his view that valuing is a “cognitive response to some valueproperty . . . which is independently there in objects without the mind” is “straightforwardly incompatible with the suggestion . . . that actual human valuings and desirings are the source of all value” (MR 95). I think that is right, if those valuings and desirings are understood as

⁶ Hume, Treatise, p. 457.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

WAYS OF MEANING AND KNOWING MORAL REALITIES



requiring certain feelings or sentiments. The two different views of evaluation or human motivation are incompatible. Attachment to the idea that feelings or sentiments are required for all evaluative intentional action is the real source of the charge of “transcendentalism” against more purely “cognitive” views. It is widely believed to be obvious that there is simply no room in the world for independent values or value-properties, and no need for them, since evaluation must be connected with action, and desires are feelings or sentiments that move us to action. If that were the only way to understand desires and acting for reasons, that “slippery thought” about “the source of all values” would perhaps have to be accepted. But that is precisely what the defence of so-called “motivated” or “reason-following” desires denies. That is the heart of Platts’ view; it is where the real dispute with those who complain of his “transcendentalism” is to be settled. That is a huge question. We can perhaps take a few beginning steps towards defence of a more “cognitive” view if we ask, as a matter of observed fact, whether it ever happens that someone is led to act in a certain way by recognizing something as a reason to act in that way. For example, does it ever happen that someone is moved to act in a certain way by seeing an unattended child struggling in the sea, or by learning that a friend of his is in some distress? We all know people, I hope, who would immediately act in such circumstances. It happens. Such people go to help, it seems, because they see the impending danger for the child or the distress of the friend as reason to do what they do. They recognize something they take to be true of the other person as reason to act in a certain way, and they act accordingly. And, it seems, those people would acknowledge that the child’s being in danger is reason to act in that way whatever anyone might think or feel about it or whether others would be moved to act as they do. It would be to no avail at this point to insist that if those people act to help it is only because they want to act in those ways. Of course they want to; they want to because they see the danger or the distress as reason to help. They have a so-called “motivated” or “reason-following” want or desire to do what they do. This does not mean that everyone who sees that the child is in danger or that the friend is in distress is moved to act in the same way. Seeing that a child is in danger or that a friend is in distress is never alone sufficient to bring about an action to help.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



WAYS OF MEANING AND KNOWING MORAL REALITIES

Whether such an action is forthcoming depends in part on who the person is who sees the child in danger, or what kind of person learns of the distress of his friend. But for those who do regard what they find to be true of the other person as reason to act, and who act accordingly, nothing more is needed to explain why they do what they do. They see what they take to be a reason, and straightaway they act. Here, I think, we can expect the complaint of “transcendentalism” to reappear. It will be objected that there is nothing in the facts described so far that explains why or how the helpful agents are moved to act as they do. It will be said that ‘recognizing a reason’ is not enough, since there is no such fact in the world, it will be said, as one thing’s being a reason to act in a certain way independently of the desires and interests of the agents concerned. It will be felt that more is needed to explain how the agent comes to take something like danger or distress for others as a reason to act in a certain way. It will be felt not to be enough to say that the agent takes the danger to the child as a “reason” because he regards a child’s death as something “bad”, “horrible”, or “one of the worst things that can happen”. Those are, admittedly, evaluative terms, and there is felt to remain a question of how an agent’s understanding and applying those evaluative terms in the way he does actually issues in the actions he performs when he thinks of things in those ways. There is a feeling that it has not yet been explained what leads the agent to act as he does, or what he is relying on. I simply observe, by way of conclusion, that here we seem to be confronted with a charge of unexplained mystery, obfuscation, or “transcendentalism” of the very kind raised against purely truth-conditional theories of meaning for a language. The complaint was that such theories give no account of how a speaker comes to know the things he knows by being a speaker of that language. Theories of meaning in terms of truthconditions alone do not explain how the understanding of the language they ascribe to speakers issues in speakers’ actual use of their terms. Those theories were said to give no account of the procedures or conditions on the basis of which speakers apply an expression to an item or a situation, or what speakers rely on in applying one expression in one situation and a different expression in another. And to that extent they were thought to leave speakers’ and agents’ knowledge and understanding unexplained.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

WAYS OF MEANING AND KNOWING MORAL REALITIES



I think there is a short and correct, but unfortunately almost universally unsatisfying, reply to this demand for an explanation of speakers’ and agents’ competence. Speakers come to know what they know about the meanings of the expressions of their language by learning that language and becoming experts in saying and responding to things expressed in its terms. They learn how to converse intelligibly and to assess reasons for saying and doing the things they do and say. All this does not come to them overnight, and it does not come easily, although it does come, on the whole, painlessly and without much effort. That is because we acquire the capacities we do only over time and in a culture surrounded by others who already have those capacities. We learn from others, and learn to act accordingly. We who learn English come to know that the English sentence ‘Ripe tomatoes are red’ is true if and only if ripe tomatoes are red. Of course, we could not know that without coming to understand the words ‘ripe’, ‘tomatoes’, ‘red’, and so on. What we have learned about English is something we can express in the very words we have come to understand. For many of us, those are the only words we have. And in learning to act and to speak as we do, we learn to recognize and to follow good reasons for doing or saying such-and-such. We come to understand what many people at least regard as good reasons to do or say things, and we eventually become sufficiently competent with the idea to make independent critical assessments ourselves of whether what are thought to be good reasons really are good reasons after all. As speakers of English we know that the sentence ‘Someone’s being in danger of drowning is reason to help’ is true if and only if someone’s being in danger of drowning is reason to help. To have the linguistic knowledge expressed in that biconditional statement we must have shown ourselves competent with the words that appear in it. There is no independent access to the understanding we have of those evaluative words beyond our having come to understand them as they are used in our culture. In understanding them we use them appropriately in acting and responding in all the ways we do. But what we have learned in learning the meanings of those words is something we can express only in some of the very words we have come to understand. We understand ourselves and others as prepared to act in ways we can see there is reason to act, and as sometimes acting accordingly. There is no prospect of

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



WAYS OF MEANING AND KNOWING MORAL REALITIES

reducing all those evaluative ways of thinking of ourselves to something fully non-evaluative, so we have no way of saying how we understand those ways of speaking and acting except by using some of those very evaluative words we are masters of. There is no neutral, non-evaluative way of expressing what we know by being masters of evaluative thought and language. There is no way of expressing what we know in understanding the words of our language except by using some of those very words we have come to understand.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

18 Meaning and Understanding The Preface to Wittgenstein’s Tractatus says that the aim of the book is “to set a limit to thought, or rather—not to thought, but to the expression of thoughts”. So “it will therefore only be in language that the limit can be set, and what lies on the other side of the limit will simply be nonsense”.¹ 4.12 Propositions [his word is ‘Der Satz’] can represent the whole of reality, but they cannot represent what they must have in common with reality in order to be able to represent it—logical form. In order to be able to represent logical form, we should have to be able to station ourselves with propositions somewhere outside logic, that is to say outside the world. 4.121 Propositions cannot represent logical form: it is mirrored in them. Propositions show the logical form of reality. They display it. A sentence cannot represent or say what it has in common with reality in order to represent it because sentences are pictures, and: 2.172

A picture cannot . . . depict its pictorial form: it displays it.

2.173 A picture represents its subject from a position outside it. Wittgenstein describes his distinction between “what can be expressed by propositions—i.e., by language— . . . and what can not be expressed by propositions, but only shown” as “the cardinal problem of philosophy”.² Almost from the beginning of his return to philosophy in Cambridge in 1929, Wittgenstein was concerned not only with the relation of ¹ L. Wittgenstein, Tractatus Logico-Philosophicus (tr. D. F. Pears and B. F. McGuinness), Routledge & Kegan Paul 1961, p. 3. ² Quoted in M. Black, A Companion to Wittgenstein’s ‘Tractatus’, Cambridge University Press 1964, p. 188.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



MEANING AND UNDERSTANDING

sentences or propositions to reality but also with the idea of someone’s thinking or meaning or understanding something in a particular way. He also returned to the idea of a ‘limit’, this time to ‘the limit of language’. In 1931 he wrote in a notebook: The limit of language is shown by its being impossible to describe the fact that corresponds to (is the translation of ) a sentence, without simply repeating the sentence. (This has to do with the Kantian solution of the problem of philosophy).³

This claim about the limit of language does not say that you simply cannot state or describe the fact that corresponds to a given sentence. It does not say that the relation between a sentence and the reality it represents is something that cannot be expressed in language at all. It allows that you can say what that relation is, but only by repeating that sentence. I take ‘describing the fact that corresponds to (is the translation of ) a sentence’ to be a matter of saying what the sentence says or what it means. A sentence or proposition that represents reality to be a certain way does not itself say how it manages to represent reality in that way. But that does not mean that no propositions or sentences can represent or state the way in which any proposition or sentence represents reality. Such a thing could even be stated without simply repeating the sentence in question: a sentence in a different language could be used to state the fact that corresponds to the original sentence. Even in one and the same language we would not have to repeat the sentence if there was another sentence in that language that meant the same as the first. But perhaps no two sentences in the same language ever mean exactly the same. If we describe the relation between the sentence ‘This book is red’ and reality by saying that the expression ‘this book’ refers to a certain demonstrated book, and that the predicate ‘is red’ is true of a thing if and only if the thing is red, we would have explained that that sentence is true if and only if the book designated by ‘this book’ is red. If we knew which book that is, we would know and we would have said what is stated by the original sentence. We would have a metalinguistic statement that tells us what is so in reality if and only if the first sentence is true. The metalinguistic sentence that gives the meaning of a given ³ L. Wittgenstein, Culture and Value (tr. P. Winch), Blackwell 1980, p. 10.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

MEANING AND UNDERSTANDING



sentence does not mean the same as that original sentence. And it does more than simply repeat that sentence. But it does repeat the original sentence in stating the conditions under which that sentence is true. It thereby does appear to say something about how the original sentence represents reality. If the semantical sentence does say what the original sentence means, or how it represents reality, this is not to say that it does so by representing “what the [original] sentence has in common with reality in order to be able to represent it”. But it does seem to state a relation between the sentence ‘This book is red’ and reality. Of course the semantical sentence does not say anything about the relation in which it itself stands to reality. Another sentence would be needed to do that. But if some such sentence could be found, and we could continue to go higher and higher up a hierarchy of sentences or languages, we could describe the relation to reality of all sentences lower down in the hierarchy. This is what Russell appears to have had in mind in his introduction to the Tractatus. For each sentence we encounter we could find a way to describe its relation to reality. We could say what would be so in reality if that sentence were true, and we could say it from a position “inside” language. Whether or not the original sentence must be repeated exactly, I take the point to be that in saying what a given sentence means or what fact corresponds to it we must use the words of the sentence, or at least some other words that equally serve to give the meaning of the sentence. We use the words with the meanings they actually have in the sentence or language in question. It would not be enough simply to refer to or name or mention a certain sentence and say that it means the same as the original sentence. To say of one sentence that it means the same as another is not to say anything about what either sentence means or what fact corresponds to them. I can know that a sentence in one language means the same as a certain sentence in another language without understanding a word of either language. Knowledge of the synonymy of mentioned expressions (even if there is such a thing) is not sufficient for knowing the meaning of either of those expressions. Obviously, in saying what a given sentence means, or in describing the fact that corresponds to it, we produce some words, typically a sentence. We produce those words in an effort to say what the original sentence means. But we do not do so by saying that the sentence we

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



MEANING AND UNDERSTANDING

produce means the same as the sentence in question. To do that would be to mention two sentences, or to mention the same sentence twice, and say that the two mean the same. When we repeat the original sentence by way of saying what that sentence means we do not at that point mention the sentence we are repeating. We utter the words we use in saying what the original sentence means, but the words we produce are not said to have the same meaning as the original sentence. We give the meaning in saying what we say. Wittgenstein observes that to know what a sentence means or what fact corresponds to it is “to be able to answer the question ‘what does this sentence say?’.⁴ And being able to answer that question involves being able to say what the sentence says, not simply knowing that it means or says the same as some other sentence. Wittgenstein says that its being impossible to describe the fact that corresponds to a sentence without simply repeating that sentence “shows” what he calls the ‘limit of language’. If what he says is impossible really is impossible, does it “show” a limit of language? It does seem to show a certain limit on explaining the meanings of sentences and on understanding or coming to understand a particular sentence. If you really have to repeat the sentence in order to say what it means, you could not explain the meaning of a sentence in that way to someone who does not already understand it. The same is true of a sentence different from the original, either in the same language or in a different language. And the same for a semantical metalinguistic sentence that gives a sentence’s meaning. Whatever sentence is used in stating the meaning of a given sentence, what is said could not explain the meaning of that sentence to someone who did not understand the sentence used in the explanation. It is obvious that no one could understand anything he is told unless he understands the words in which he is told it. So no one could understand what he is told about what a certain sentence means or what fact corresponds to it unless he understands the words in which he is told it. But this can still be said to express a certain kind of ‘limit’: this means that any kind of explanation of a language presupposes a language already. And in a certain sense, the use of language is something that cannot be taught, i.e., I cannot use language to teach it in the way in which language ⁴ L. Wittgenstein, Philosophical Grammar (tr. A. Kenny), Blackwell 1974, p. 44. (Page numbers with ‘PG’ in parentheses refer to this volume.)

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

MEANING AND UNDERSTANDING



could be used to teach someone to play the piano.—And that of course is just another way of saying: I cannot use language to get outside language.⁵

What Wittgenstein says “cannot be taught” is “the use of language”. Particular languages can be taught, and a teacher can use one language to teach someone a second language. Of course, a pupil has to understand the language the teacher is using. So that would not be a case of teaching the pupil “the use of language”, but only this or that language. Even if the use of language cannot be taught, it is certainly something that can be learned, or acquired. Wittgenstein does not say, strictly and without qualification, that the use of language cannot be taught. The “certain sense” in which he thinks it cannot be taught is that “I cannot use language to teach it in the way in which language could be used to teach someone to play the piano”. What a piano teacher can use language to get the pupil to do are things the pupil could do without the use of language—e.g., to play a particular chord. The pupil needs to be master of the use of language to understand what the piano teacher has told her, but not in order to do what the teacher tells her how to do. But a language-learner could not be simply told in the same way how to do the things involved in being a master of language. Or rather, she could be “told”, but it would do no good. What was said would make sense and would be understood only by someone who has already to some extent mastered the use of language. Wittgenstein sums this up by saying “I cannot use language to get outside language”. The Tractatus said that one would have to be “outside logic, that is to say outside the world” in order to represent what a proposition must have in common with reality in order to represent it. A picture was said to represent its subject “from a position outside it”. In the early parts of Philosophical Grammar Wittgenstein takes up the question of how one can talk about ‘understanding’ or ‘not understanding’ a sentence (PG 39). Understanding is what he calls “a psychological reaction while hearing, reading, uttering, etc. a sentence”; “the phenomenon that occurs when I hear a sentence in a familiar language and not when I hear a sentence in a strange language” (PG 41). As soon as I hear or see a sentence in a language familiar to me I understand it. What ⁵ L. Wittgenstein, Philosophical Remarks (tr. R. Hargreaves and Roger White), Blackwell 1975, p. 54. (Page numbers with ‘PR’ in parentheses in the text refer to this volume.)

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



MEANING AND UNDERSTANDING

I understand are those sounds I hear or those marks on the page. I understand what they express. He thinks someone could even express something by arranging a clump of trees in a certain way—as kind of code, say. It would make sense in those circumstances to ask “Do you understand what this clump of trees says?” (PG 39). And someone could see the trees and understand what they say. It could be said that someone who sees those trees and understands what they say knows what proposition they express. The same could be said of someone who sees and understands a sentence. But we cannot say that the person who knows what proposition is expressed by those trees or that sentence also understands that proposition in the way he understands the trees or the sentence. It isn’t that he knows what proposition the trees or the sentence express and in addition also knows what proposition that proposition in turn expresses. Apart from the clump of trees, or the sentence there on the page, there is no further item to mean or express anything. Once the person understands what the clump of trees or the sentence says—once he knows what proposition it expresses—there is nothing else in the situation for the person to understand. Considered simply as a clump of trees, or as sounds in the air or marks on a page, there is no difference between trees or sounds or marks that express propositions and can be understood in a certain way and those that mean or express nothing. Clumps of trees or sounds or marks can be considered simply as “a phenomenon or fact” as Wittgenstein puts it (PG 143). We do not thereby recognize that the trees were arranged with a certain purpose or that the sounds or marks express anything at all. The meaningful or expressive aspect of a clump of trees or of sounds or marks is “absent from the phenomenon as such” (PG 143). The same can be said of what a person can be observed to do in expressing an expectation, or a wish, or in carrying out an order, when those utterances or movements are considered in that same way—simply as “a phenomenon or fact”. To recognize that people think or expect or wish certain things we have to be able to read off from the expression of their thought or expectation or wish that it is the thought or expectation or wish that such-and-such. If we regarded the people around us and the marks on paper and the clumps of trees that we see, and so on, solely as “phenomena” in the sense Wittgenstein has in mind, we would never recognize anything as expressing anything or anyone as thinking or meaning or

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

MEANING AND UNDERSTANDING



understanding anything at all. To see things only as “phenomena” in that sense would be to see them, in Wittgenstein’s phrase, only “from outside” (PG 143). The expression ‘from outside’ used in this way obviously does not simply mean from a position outside another person or outside a sentence or outside a clump of trees. The idea is not that you can understand what a clump of trees says only by standing inside it. To see a clump of trees, or marks on paper, or people’s making noises or behaving in certain ways ‘from outside’ in this sense is to see them without recognizing or thinking of them as expressing anything or having any meaning. If we consider them ‘from outside’ we have to understand thoughts as thoughts, intentions as intentions and so on, without getting any information about something’s meaning. (PG 144)

Anything looked at only “from outside” in that way will appear “lifeless and isolated” (PG 146); “it does not point outside itself to a reality beyond” (PG 148). Looking at things only in that way would mean that we could not recognize any thoughts or intentions or meanings at all. That would leave us in the plight Wittgenstein describes like this: Whatever phenomenon we saw, it couldn’t ever be intention; for that has to contain the very thing that is intended, and any phenomenon would be something complete in itself and unconcerned with anything outside itself, something merely dead if considered by itself. (PG 144)

Seeing things only in that way is not a plight we are actually in. As things are, we can see a sentence or a clump of trees and know what it says (if there is something it says). We can see and hear people who think and mean things and recognize what they think and mean. But when we do that: If a thought is observed there can be no further question of an understanding; for if the thought is seen it must be recognized as a thought with a certain content; it doesn’t need to be interpreted!—That really is how it is; when we are thinking, there isn’t any interpretation going on. (PG 144)

When we observe a person’s thinking “there can be no further question of an understanding” because there is nothing further to be understood. The person’s thought has been recognized. When someone knows what proposition is expressed by a certain clump of trees there is not

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



MEANING AND UNDERSTANDING

something else—what that proposition in turn expresses—that also needs to be understood. So it is not just that the thought that has been recognized “doesn’t need to be interpreted”; there is nothing relevant that could be interpreted. When we see a person’s action or a sentence or a clump of trees and know what it means we do not see it “from outside” all thought and meaning. It is only because we do not see it only in that way that we understand it. In the very understanding or thinking anything at all “there isn’t any interpretation going on”. If we were restricted in our observations or descriptions to what Wittgenstein calls “mere phenomena” which give no “information about something’s meaning” we could never recognize or specify what a certain sentence expresses or how a person understands a certain expression. We could never mean or understand something or explain the meaning of something if we had to do so “from outside” all recognition of what some things mean or what some people mean by them. This is what I take it Wittgenstein has in mind in saying that it is impossible to describe the fact that corresponds to a given sentence or to say what that sentence means without repeating the sentence. In giving the meaning of a sentence we must use a sentence and not simply refer to or mention a sentence or some other object. Being restricted to “mere phenomena” without any “information about something’s meaning” would mean that even if one could somehow identify or refer to a certain item or object associated with or expressed by the sentence, that would be only to know that the sentence stands in a certain relation to some other mentioned thing. It would not be to know what that sentence says or means. To know and to say what the sentence means is to know and be able to say something “from within” a capacity to recognize and understand and say something about the meanings of expressions. The implications of these middle period reflections on the impossibility of understanding language and meaning “from outside” are illustrated and developed at many points and in many ways throughout Philosophical Investigations.⁶ In sections §185 and following, for instance, we have a pupil who by all the usual criteria has mastered the instructions he has been given in the continuation of different arithmetical

⁶ L. Wittgenstein, Philosophical Investigations (tr. G. E. M. Anscombe), Blackwell 1953. (Section numbers in parentheses in the text refer to this volume.)

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

MEANING AND UNDERSTANDING



series. On being given the number 2 and the order ‘Add 2 each time’, he writes down 4, 6, 8, 10, and so on, and when encouraged to continue eventually goes on to . . . 996, 998, 1000, 1004, 1008. It is clear that this pupil has not followed the order he was given. Or he has not followed it correctly, which comes to the same thing. If we ask whether he has understood the order he was given we could say he understood it if we think he just got tired of applying the rule or simply refused to follow it after 1000. But if he responds as Wittgenstein imagines him—by saying “Yes, isn’t that right? I thought that was how I was meant to do it”, or “But I went on in the same way after 1000 as before”—we recognize that he did not understand the order he was given. He did not follow that order correctly. We know this about this pupil because we know that what he did after 1000 is not in accord with the order as it was meant. We understand that order and we know what is and is not in accord with it. If, as the pupil claims, he was following the same order after 1000 as he had been before, then the order he was following was not the order to add two each time. That would mean that even before reaching 1000 he was not following the order to add two each time even though every number he wrote up to that point was in fact two more than the previous number. The sequence he had written up to 1000 was in accord with or in agreement with the rule to add two each time, but that would not have been the rule he was following. Developments of arithmetical series provide especially striking examples of the fundamental point Wittgenstein wishes to draw attention to in discussing meaning and understanding. What a person continuing a series has written down at any given time never uniquely determines what comes next, and so in itself it does not determine what rule the person is following. The person’s understanding of his instructions is shown or expressed in his writing down what he does, but the outcomes of those actions—the numbers he writes down—do not fix or indicate uniquely his understanding of the instructions. That remains so no matter how many numbers the person writes. His having written those numbers is compatible with his having understood the instructions in any one of indefinitely many different ways. With respect to fixing what the person understands, the numbers written on the page up to any given point could still be viewed, in Wittgenstein’s phrase, “from outside” any implication of the pupil’s having understood the instructions in

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



MEANING AND UNDERSTANDING

some particular way. The series of written numbers alone would be to that extent a “lifeless and isolated” “phenomenon or fact” which does not point beyond itself to the pupil’s having understood the instructions in one way rather than in countless other ways. Of course, in considering this example we do not view everything the pupil does “from outside” all thought and meaning. We take the pupil to understand the order he is given in some way or other and to have many other intentional attitudes. But even if we grant that the pupil intentionally writes numerals, and that he means each numeral to stand for a number, and we know which numbers they stand for, we cannot determine what rule he is following as long as we are restricted to describing only the numbers he has written down. We would then be confined to describing what he has done only “from outside” his having understood the instructions in some particular way. The same holds for any attempted specification of the meaning of the sound or marks ‘Add 2 each time’ by which the order was given to the pupil. Considered as a “phenomenon or fact” complete in itself, that sign or form of words—or the utterance of it—could mean, or could have meant, any number of different things, or nothing at all. Viewed “from outside” all information about thought and meaning and understanding, the sign or utterance itself obviously imposes no constraints on what someone is supposed to do when he hears or sees it. We know that the pupil violated the constraints imposed by the order as it was meant in the example; he did not follow it correctly. So, we obviously do not view that sound or mark only “from outside” in a way that carries with it no constraints. We understand the English expression ‘Add 2 each time’; and the meaning of that expression, or how it was meant, is what determines what step is to be taken at each point. What determines how the order was meant, as Wittgenstein says, is “the kind of way we always use it, the way we are taught to use it” (§190). As a general observation about meaning, understanding and use, this must be right. There can be nothing more to an expression’s having a certain meaning than its being used in a certain determinate way. And knowing what an expression means can be nothing other than knowing how it is or has been used. But to appeal to the use of an expression in accounting for our understanding of its meaning, the idea of use must be understood in a certain way. It is possible to say something true of all those occasions on which a certain expression was used in the past—and

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

MEANING AND UNDERSTANDING



in that sense to describe the past use of an expression—without thereby specifying the meaning of that expression. The same holds even if a similar description is true of all future applications of that expression. Describing the use of an expression only in such ways would not serve to specify its meaning. Even if everyone in the past had in fact always responded to the sound or marks ‘Add 2 each time’ with a series of numbers each of which is two more than the previous number, it would not follow that those sounds or marks mean that two is to be added each time. Even if that same correlation between utterance and behaviour continued forever into the future it would not follow that that is what those sounds or marks mean. This general point about meaning and understanding is illustrated by the simple case of the deviant pupil. Every number the pupil writes down before 1000 is two more than the previous number, but that does not imply that he is following the rule of adding two each time. If we grant that he could have understood the words he heard differently from the way we do and have been following some other order correctly even before he reached 1000, we thereby acknowledge that the expression ‘Add 2 each time’ could be, or could have been, an order to do something other than add two each time, even though every response to it in the past has involved adding two. The meaning of that expression cannot necessarily be read off a description of all the applications that have been or even will be made of it. A description made only “from outside” recognition of the expression’s meaning what it does will not serve to specify its meaning. We can describe the use of an expression in a way that suffices to fix its meaning or to specify what we know when we know its meaning. We can say “ ‘Add 2 each time’ means add two each time”; we simply repeat the original sentence exactly as it is. But that is not strictly required. We could say “The words ‘Add 2 each time’ are used by us to mean that two is to be added each time”. In these remarks we employ the idea of meaning in describing how the expression is used and what it is used to mean. We also make use of (and do not mention) the very words whose meaning we are specifying. The same is true in giving semantical metalinguistic explanations of meaning that state the conditions under which a given sentence is true. In understanding what order is given to the pupil, and so what he must do in order to comply with it, we also recognize that the teacher meant the words he uttered in a certain way.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



MEANING AND UNDERSTANDING

This can raise the question “how are we to judge whether someone meant such-and-such?”. To this Wittgenstein replies: The fact that he has, for example, mastered a particular technique in arithmetic and algebra, and that he taught someone else the expansion of a series in the usual way, is such a criterion. (§692)

That someone has such mastery and has taught it to others is not something we could recognize “from outside” all recognition of intention or of what people mean by the words they utter. In recognizing or stating the meaning of an expression in any of these ways we make use of the words in question. We do not just describe some aspect of their use that is available “from outside” all meaning or understanding of them. The idea that it is impossible to describe the fact that corresponds to a given sentence without repeating that very sentence is an expression of the requirement that we must describe the use of an expression in some such engaged way in order to succeed in specifying or explaining its meaning. By ‘engaged’ I mean “from within” recognition of certain intentions or meanings or by actually using words with determinate meanings to say what certain words mean or how they are to be understood. It is not enough just to mention words and say something only about a relation between them and certain other mentioned words. Nor is it enough to mention words and say something that is so in the world when and only when the words are uttered or heard. A description of the use of an expression cannot succeed in specifying what it means or how someone understands it if it is given only “from outside” all recognition of the meaning of expressions or of the intentions with which they are used. This imposes a certain kind of limit on the possibility of someone’s coming to understand an expression in a certain way. And it can seem to present a challenge to the very idea of a person’s understanding or meaning something in any way at all. If everything a person does or could do in applying a given expression is in accord with his understanding that expression in any one of many different but incompatible ways, how is the person’s understanding the expression in the way he does related to the particular applications he makes of it? In the case of the deviant pupil the challenge is expressed like this:

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

MEANING AND UNDERSTANDING



“But how can a rule shew me what I have to do at this point? Whatever I do is, on some interpretation, in accord with the rule.” (§198)

To this last sentence Wittgenstein replies: That is not what we ought to say, but rather: every interpretation, together with what is being interpreted, hangs in the air; the former cannot give the latter any support. Interpretations by themselves do not determine meaning. (§198)

Here Wittgenstein takes an “interpretation” of an expression to be another expression that is to serve as an interpretation of the first. So “interpretations by themselves do not determine meaning” because producing one expression as an interpretation of another is not enough to specify what the first expression means or how it is understood. Without any understanding or specification of the meaning of the second expression, both expressions would still “hang in the air” without the meaning of either of them having been explained. Saying of two mentioned expressions that they mean the same or that one is an interpretation of the other is not sufficient to specify what either of those expressions means. To regard interpreting or understanding or explaining the meaning of an expression as a matter only of substituting one mentioned expression for another would lead to an infinite regress of offered explanations. If explaining or understanding the meaning of any expression always required another expression, that expression in turn would have to be understood, which would require another expression, and so on. There would be no saying what any expression means and no understanding of anything. This is the source of the apparent “paradox” referred to in Philosophical Investigations §201: no course of action could be determined by a rule, because every course of action can be made out to accord with the rule. The answer was: if everything can be made out to accord with the rule, then it can also be made out to conflict with it. And so there would be neither accord nor conflict here.

There would be no such thing as a person’s understanding an expression in one way rather than another. Nothing would have any determinate meaning. This, Wittgenstein says, “was our paradox” (§201), growing out of the “challenge” expressed in §198.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



MEANING AND UNDERSTANDING

The sources and implications of this apparent “paradox” have generated a huge body of commentary, most of it in response to Saul Kripke’s Wittgenstein on Rules and Private Language⁷ in 1982. That book presented in striking form what it argued is an unanswerable challenge: to say what it is about a person that constitutes his meaning or understanding an expression in a certain particular way. No present or past facts involving the person were said to constitute his meaning or understanding an expression in one way rather than another when he uses or responds to it on a particular occasion. The unavoidable conclusion is the “sceptical paradox” that no one means anything by any expression he uses or responds to. The problem seemed easily generalizable from meaning or understanding words or expressions to any determinate thought or intentional attitudes whatever. Pressed to the limit, there appeared to be no such fact as anyone’s meaning or understanding or thinking anything at all in one particular way rather than another. After mentioning the “paradox” in Philosophical Investigations §201, Wittgenstein does not try to meet the challenge by saying what constitutes a person’s meaning or understanding something in a particular way. He says rather that there is a “misunderstanding” behind the alleged paradox; it rests on a confusion. We know that some thought lying behind the challenge must be confused because we know that it is not true, as the “paradox” claims, that every course of action can be made out to accord with any given rule. As we have just seen, we know that what the deviant pupil did after 1000 is not in accord with the rule to add two each time. Nor can it be “made out” to accord with that rule. The rule to add two each time requires that 1002 be put next after 1000; that is the course of action determined by that rule at that point, and the deviant pupil did not do that. It is true that what the deviant pupil did can be made out to accord with some rule. We can perhaps make sense of him as having understood the words ‘Add 2 each time’ as giving the order that two is to be added each time up to 1000 and four is to be added after that. That is a possible way to understand what he did, but it is not a way of understanding his actions as being in accord with the rule to add two each time. We know that his actions are not in accord with that rule. Nor does thinking of him ⁷ Saul A. Kripke, Wittgenstein on Rules and Private Language, Harvard University Press 1982.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

MEANING AND UNDERSTANDING



as having understood the instructions in that more complicated way show that any action can be made out to accord with the rule he would then be following. The rule he would be following on that more complicated understanding is a rule that determines that 1004 is the next step after 1000. No other course of action could be “made out” to accord with that rule at that point either. It is a confusion to say that the deviant pupil who follows that more complicated rule understands or interprets the rule to add two each time differently from the way we do. He does not interpret that rule at all; there is no such thing for him to interpret. We can say that he interprets or understands the words he hears as expressing the rule to add two up to 1000 and four thereafter. But to understand or interpret those words ‘Add 2 each time’ in that way is to understand or interpret an expression; it is not to interpret a rule. If the deviant pupil can be said to understand the words ‘Add 2 each time’ to express a certain rule, he does not in the same way understand that rule to express something in turn. We, who understand or interpret the words ‘Add 2 each time’ as they are normally meant in English, recognize what rule those words express, but it would be confusion to say that in addition we also understand or recognize what rule that rule expresses. Once we know what rule those words express, just as a person knows what proposition a certain clump of trees expresses, “there can be no further question of an understanding”; there is nothing left to be understood. “That really is how it is; when we are thinking, there isn’t any interpretation going on” (PG 144). What leads to the apparent “paradox” is the idea that understanding and following a rule imposes an apparently unsatisfiable demand for more and more rules for “interpreting” whatever rules have been accepted so far. One source of that demand in turn is the thought that someone who understands and follows a rule correctly must be instructed or guided in some way in how to follow it.⁸ The question is ⁸ This thought finds expression in Kripke’s argument when he observes, e.g., “Ordinarily, I suppose that, in computing ‘68 + 57’ as I do . . . I follow directions I previously gave myself ” or “When I respond in one way rather than another . . . I feel confident that there is something in my mind . . . that instructs me what I ought to do in all future cases” (Wittgenstein on Rules and Private Language, pp. 10, 21–2). The same idea, along with the threat of regress that it gives rise to, is present in Crispin Wright’s summary of the rule-following challenge as he understands it: “what makes it possible for there to be such things as rules . . . at all? . . . how do I create something which carries determinate instructions for an open range of situations which I do not contemplate

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



MEANING AND UNDERSTANDING

how his understanding and accepting the rule “issues in” or “yields” the particular applications he makes of it. As the challenge of §198 has it: “But how can a rule shew me what I have to do at this point?”. As before, in Wittgenstein’s words, “it can be seen that there is a misunderstanding here” (§201). If something were always needed to instruct or guide or show the rule-follower how to proceed, or if there had to be something in his mind that informs him what to do at each point, the thing that contains the instructions would also have to be understood by the person in some way or other. So that thing itself would stand in need of something further, something that gives instruction or guidance about how that first item of instruction or guidance is to be understood. And so on. If it really were a condition of a person’s meaning or understanding something in a certain way that there be something that tells or instructs the person how to mean or understand it in that way, it would be impossible for anyone to mean or understand anything. Once we see through this “misunderstanding” or confused demand there is no such threat of regress or paradox. When we say, from inside language, what rule is being followed or what certain words mean or how they are to be understood, there is nothing else to be understood or interpreted other than the words we utter. By ‘inside language’ I mean that the words we use (and do not just mention) in saying what rule is followed or what certain words mean are used with determinate meanings and are to be taken or understood in a certain way. Of course, the words we utter on a particular occasion might not be understood, so they might have to be further explained. And we will use words in explaining them. But if what we say in giving those explanations is understood, it is not something that in turn stands in need of further understanding or interpretation. There is no such thing to interpret. Even if the deviant pupil does understand the expression ‘Add 2 each time’ to mean that two is to be added up to 1000 and four thereafter, when we ask him to say how he understands that expression he might well reply by uttering the words “The expression ‘Add 2 each time’ means add two each time”. He might simply repeat the expression

in making it? What gives it this content, when anything I say or do in explaining it will be open to an indefinite variety of conflicting interpretations? And how is the content to be got ‘into mind’ and so made available to inform the successive responses of practitioners?” (C. Wright, Rails to Infinity, Harvard University Press 2001, pp. 2–3).

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

MEANING AND UNDERSTANDING



whose meaning he is explaining. That is also what we would do in expressing our own understanding of what that expression means; we would utter those same words. The difference between what we would thereby be saying and what the imagined deviant pupil would be saying about the meaning of that expression obviously does not lie in the words we would both utter in explaining its meaning. We both utter the same words, but the difference lies in what those words mean as uttered by us and as uttered by the deviant pupil. That is a matter of how each of us uses those words, and what we use them to say. And what someone who hears us would understand each of us to say depends on how that person understands the words he hears, or what he thinks we mean by them. This illustrates in another way how saying or understanding or being told something about what an expression means is possible only “from inside” language—only from an engaged position of understanding what one hears or says as meant or understood in a certain way. One source of resistance to Wittgenstein’s idea that meaning and understanding cannot be accounted for “from outside” all language is that it can seem to make it impossible for anyone to acquire a language in the first place. No one starts out being able to speak. So, there appears to be a question of how anyone could ever get into language or come to understand any words at all if any statement or explanation of the meaning of expressions can succeed only if the person already understands the words in which that explanation is expressed. This could be said to represent a certain kind of ‘limit’ on the acquisition of language, but it is a ‘limit’ only on the possibility of stating or explaining the meaning of something to somebody. The most that follows is that we do not get into language in the first place by having the meanings of all the words we come to understand explained or otherwise described to us. Nor does the fact that we can explain or describe those meanings only ‘from inside’ language represent a limit of language in the sense that to try to say what a sentence means or what fact corresponds to it would be to try to say something that “lies on the other side of the limit [and] will simply be nonsense”.⁹ This kind of ‘limit’ on the possibility of explaining the meaning of something does not imply that there are no facts of meaning or of correspondence between sentences and reality, or that we

⁹ Tractatus Logico-Philosophicus, p. 3.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



MEANING AND UNDERSTANDING

cannot know such facts, or cannot express that knowledge in language. We do know what our expressions mean, or what fact corresponds to a given sentence, and it can be expressed in language. We express it in those sentences in which we say what something means or what fact corresponds to a given sentence. We all have that knowledge in being able to speak. But it is not knowledge of meaning that could be given to someone who cannot speak to instruct or guide him into meaning or understanding something in a particular way. No one could receive the proffered information without already possessing at least some measure of what the instructions were trying to impart. We all start out with no linguistic competence and are brought to full mastery by training, interaction, and socialization with those who speak and live as we come to do. But we do not do so by being given directions or following instructions. Meaning and understanding things in certain ways is a fundamental fact of human life. Wittgenstein observes that commanding, questioning, recounting, chatting, and so on, which are all linguistic activities, “are as much a part of our natural history as walking, eating, drinking, playing”, which do not depend on language (§25). Appeal to the very ‘naturalness’ of these facts, and to their being open to view, can give rise to deeper dissatisfaction with the idea that meaning and understanding cannot be accounted for “from outside” all language. There is an understandable philosophical aspiration to explain language and meaning and understanding in general—to account for their presence in a world that is not already understood to contain language. This encourages the expectation that meaning and understanding must be explainable or intelligible in terms that do not themselves presuppose the ascription of meaning or understanding to anything or anybody in the world. This has even been held to be a condition of success of any properly philosophical theory of meaning: the theory must explain speakers’ possession of concepts and their expressing those concepts in the words of their language without itself attributing a grasp of any of those concepts to the speakers whose meanings it would account for. This is Michael Dummett’s idea of a “full-blooded” theory of meaning for a language. He thinks such a theory is required for “the purpose for which, philosophically, we require a theory of meaning”.¹⁰ ¹⁰ “What is a Theory of Meaning? (I)”, in M. Dummett, The Seas of Language, Oxford University Press 1993, p. 4. Dummett does not further specify what that philosophical

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

MEANING AND UNDERSTANDING



The idea that meaning and understanding cannot be accounted for “from outside” all language is in conflict with this aspiration for a “fullblooded” explanation of language and meaning. It implies that we cannot explain how speakers understand all the words they understand by observing and describing in non-intensional terms the conditions under which they utter or respond to those words. The fact that words have the meanings they do for particular speakers cannot be constructed in that way out of materials available “from outside” all thought and understanding on the part of those speakers. This does not imply that the meanings of expressions or the understanding speakers have of them is somehow hidden or cannot be described in terms of their overt use. It means only that the use of expressions that gives them their determinate meanings is not something that can be described “from outside” language in that way either. One obvious and important route into language is demonstrative or ostensive teaching and learning—coming to understand the meanings of certain words in attending to the presence of the very things the words stand for. At some point before any ostensive teaching has begun every prospective speaker is “outside language”. But we do learn words by ostension. It is fundamental to our understanding and meaning anything. It is difficult to see how any words would be connected with any of the things we talk about at all if that were not so. But ostensive learning or definition is no exception to the idea that intention or meaning or understanding is not determined by anything available “from outside” language in Wittgenstein’s sense. The point is illustrated again and again in different ways in the early sections of Philosophical Investigations. A sound’s being uttered over and over again in the presence of objects which are all of a certain kind does not alone fix the meaning of that sound or anyone’s understanding of it. It does not matter how long the ritual goes on. Taken on its own, as a “phenomenon or fact”, any correlation holding between a sound and a certain kind of object remains as “outside” the meaning of that sound or how it is to be understood as any other correlation between objects of that kind and anything else.

purpose is, or why philosophy requires it. In “Frege and Wittgenstein” (in his Frege and Other Philosophers, Oxford University Press 1991, p. 247), he describes the task as “giving an account of the language as from outside”, but he does not draw any connection between his use of that phrase and Wittgenstein’s earlier use of it.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



MEANING AND UNDERSTANDING

Getting a pupil to repeat the sound or to point to an object of a certain kind whenever he hears the sound might be part of what the pupil learns. That is what Wittgenstein calls “ostensive teaching”; it is not sufficient to fix the meaning of the sound or how it is to be understood. The sound the pupil hears might be uttered in the presence of objects that are all white, and all rectangular, and all slabs, for instance. How the sound he hears is to be understood, given the correlation involved in that “teaching”, depends on something more. What the pupil is meant to do with the word, or what he comes to do with it, determines what kind of word he takes it to be. That further step involves what Wittgenstein calls “ostensive training”. With different training the same ostensive teaching . . . would have brought about a quite different understanding. (§6)

An ostensive definition or explanation succeeds in giving the pupil the meaning of the word only “when the overall role of a word [of that kind] in language is clear” (§30) or only “if the place is already prepared” (§31) for words of that kind within whatever linguistic or other capacities the pupil already possesses. A speaker needs at least a rudimentary repertoire in order to mean or understand an expression in a determinate way. In Wittgenstein’s memorable phrase, there must be a “post” at which he can “station” a new word in order for it to mean anything to him (§29). It might seem that it would be enough for the pupil simply to see the teacher’s finger and recognize what the teacher is pointing at. But he cannot do that if he sees the teacher’s finger and the object only “from outside” all thought and meaning—only as two sorts of things that are present together. And to recognize the intention with which the ostension is carried out he must recognize it as an intention or thought “with a certain content” (PG 144). Contents or thoughts of that kind would have to be intelligible to anyone who could recognize them and learn the meaning of a word in that way. The person must have some idea of what a colour is, for example, or what a shape is, or what a buildingstone of a certain kind is, if he is to understand the sound he hears as a word for a colour rather than a shape or a kind of building-stone. That can reveal itself only in the kind of use he is prepared to make of the words he has heard. Describing the workings of language and what competent speakers mean and understand only “from within” language in this way is fully in

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

MEANING AND UNDERSTANDING



accord with what have been called more “modest” accounts of meaning and understanding. They make use of the very concepts they attribute to speakers whose linguistic competence they describe. To acknowledge the inevitability of understanding meaning in only that way is to acknowledge the poverty of what would be available “from outside” language in Wittgenstein’s sense.¹¹ To what extent can the function of language be described? If someone is not master of a language, I may bring him to a mastery of it by training. Someone who is master of it, I may remind of the kind of training, or I may describe it; for a particular purpose; thus already using a technique of the language. The difficult thing here is not, to dig down to the ground; no, it is to recognize the ground that lies before us as the ground. For the ground keeps on giving us the illusory image of a greater depth, and when we seek to reach this, we keep on finding ourselves on the old level. Our disease is one of wanting to explain.¹²

These remarks express resistance to the aspiration to “full-blooded” theories on the grounds that meaning and language can be understood or explained only “from within” language. But this resistance on Wittgenstein’s part has been regarded as mere “mystification”¹³ or as defeatist or “quietist”¹⁴ reluctance to take up a recognizably serious intellectual challenge. It is true that Wittgenstein’s philosophical treatment of the human fact of meaning and understanding, and his response to attempts to gain a finally satisfying general understanding of it “from outside”, does not yield a “constitutive analysis” of a person’s meaning or understanding something in a particular way. It does not seek or provide explanatorily revealing necessary and sufficient conditions of meaning or understanding something that would ¹¹ For elaboration and defence of the idea that requiring anything beyond “modest” accounts of meaning and understanding would leave linguistic behaviour altogether unintelligible, and for the sources of that idea in the writings of Wittgenstein, see John McDowell, “In Defence of Modesty” and “Another Plea for Modesty” in his Meaning, Knowledge, and Reality, Harvard University Press 1998. ¹² L. Wittgenstein, Remarks on the Foundations of Mathematics (tr. G. E. M. Anscombe), MIT Press 1978, VI §31. ¹³ M. Dummett, “Reply to John McDowell”, in B. Taylor (ed.), Michael Dummett: Contributions to Philosophy, Nijhoff 1987, p. 268. ¹⁴ Eg C. Wright, Rails to Infinity, p. 169.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



MEANING AND UNDERSTANDING

satisfy the demands of a “full-blooded” theory. But what Wittgenstein says does help explain, or at least exhibit, why he thinks such an “analysis” or theory is not to be had, and what he thinks is the most that can be said. This can be felt as a distinct philosophical disappointment. That feeling perhaps helps explain the impression of “quietist” or even complacent evasion. Wittgenstein’s remark that it is impossible to describe the fact that corresponds to a sentence without simply repeating the sentence is immediately followed by the observation that “this has to do with the Kantian solution of the problem of philosophy”. Not the problems of philosophy (plural); “the problem of philosophy” (singular). I take this to be the problem of how there can be such a thing as philosophy—how there can be, or what can be the upshot of, philosophical reflection on such matters as thought, meaning, and understanding. A “solution” is “Kantian”, presumably, to the extent to which the facts of human meaning and understanding can be understood and accepted only from within whatever meaning and understanding we are already capable of. Wittgenstein sums up the point of his observations in these words: What is spoken can only be explained in language, and so in this sense language cannot be explained. (PG 40)

What he has in mind might then be transposed into something Kant could well have said about thought: “what is thought, and our thinking it, can only be explained in thought, and so in this sense thought itself cannot be explained”. This Kantian observation could also perhaps be found philosophically disappointing. But presumably not because it is an expression of complacent or defeatist reluctance to engage in constructive philosophical reflection.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

19 Davidson and Wittgenstein on Meaning and Understanding I think there is a deep and distinctive accord between Donald Davidson’s conception of meaning and understanding and the treatment of those same notions in the later philosophy of Wittgenstein. Both philosophers presume that there can be nothing more to an expression’s meaning what it does than its being used in certain ways by speakers, and nothing more to speakers’ understanding an expression than their responding or being ready to respond to it in certain distinctive ways in interacting with others and with the shared world around them. This can seem beyond question. It is only because sounds or marks are used in some way or other that they mean what they do; no sounds or marks meant anything before human beings began to speak. But that very general connection between use and meaning is not the distinctive accord I see between Davidson and Wittgenstein. Whether an appeal to the use of an expression can account for its meaning what it does depends on what the use of an expression is understood to be, or how that use is to be described. It is on this question that I see a promising accord between Davidson and Wittgenstein. Not every conception of meaning as use is equally promising. Quine, for instance, also agreed that there is no more to the meaning of an expression than the use speakers and hearers make of it. For him, use was to be described in terms of the conditions under which speakers would utter an expression or would assent to or dissent from its utterance by others.¹ And for Quine those conditions could be specified by ¹ Quine, Word and Object, M.I.T. Press 1960, ch. 2.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



DAVIDSON AND WITTGENSTEIN ON MEANING AND UNDERSTANDING

any description that is true if and only if speakers would in fact respond in those ways. This is vulnerable to the so-called “indeterminacy of translation”, according to which any expression uttered when and only when a rabbit is present means the same as the English words ‘rabbit’, ‘undetached rabbit part’, ‘rabbit-stage’, ‘rabbithood’, or any other expressions that differ in meaning from one another but are true of every circumstance in which a rabbit is present. This leaves what the expression in question means at best indeterminate. It does not account for differences of meaning we all recognize. This is not a consequence simply of the idea that meaning is determined by use; it is a result of a certain way of understanding what the “use” of an expression can legitimately be said to be. It is perhaps not surprising that merely extensionally equivalent descriptions of the conditions for the appropriate use of a term do not suffice to fix what the term means. The same difficulty arises for conceptions of use understood as speakers’ dispositions to utter or to respond to a term under certain conditions, if those conditions could be specified in any extensionally equivalent terms. There is no such threat in Davidson’s or Wittgenstein’s conceptions of speakers’ use or understanding of expressions. Both Davidson and Wittgenstein see language as a public social practice. Speakers’ meaning what they do by their words can therefore be understood and explained by starting from the undeniable fact that each of us is born into a linguistic culture in which we learn to speak by coming to participate in that culture and so acquire competences we did not have before. To describe in detail the linguistic competence we all acquire would be to describe the use and meaning of expressions in languages we understand. In the early sections of Philosophical Investigations Wittgenstein explores the role of what he calls “demonstrative” or “ostensive” learning in the acquisition of language. That involves “ostensive teaching” in which a teacher draws a pupil’s attention to objects of a certain kind while uttering an expression whose use involves objects of that kind in some so-far-unspecified way. This requires more than the pupil’s simply noticing some similarity or other among the “ostended” objects. For one thing, there are too many similarities available; every object is similar to every other in many respects. If the objects the teacher points to when uttering the sound are both red and round, for instance, the question is

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

DAVIDSON AND WITTGENSTEIN ON MEANING AND UNDERSTANDING



which feature is relevant to what the expression means. Is the sound meant as a colour-word or a shape-word? Explicit instructions would be to no avail with a pupil who does not yet understand any words. By pointing, uttering, and gesturing, the teacher can eventually get the pupil to catch on to the similarity he is drawing his attention to. No one could learn to speak if he could not catch on in this way and so come to share judgements of relevant similarity with competent speakers. That competence starts with what Wittgenstein calls an “associative connection” in the mind of the pupil between the sound uttered and objects of a particular kind. That “association” is a step in the right direction, but it alone does not give the pupil a determinate understanding of what the expression means. There must also be what Wittgenstein calls “instruction” or “training”. The pupil must learn how to use the expression—how to apply it to relevant items and to respond to it appropriately in communication with others. The same “associative connection” between expression and object, if combined with one kind of “training”, would result in a quite different understanding of the term from that produced by a different “training”. The use a speaker learns to make of an expression is what determines what he means and understands by it. To catch on to the “training” and to know how to use the new word correctly is to recognize what Wittgenstein calls the “post at which we station the word”, the “place in language, in grammar, we assign to the word”.² That involves mastery of the structure of the sentences in which words of that kind appear, how it differs from the structure of other sentences, and under what conditions what is said to be so by such sentences is to be accepted or rejected. Competent speakers of a language could observe a pupil who is learning that language and see that by this kind of “training” she has come to understand a certain expression as a colour-word rather than a shape-word. The pupil might still lack the word ‘colour’, so she could not yet say such a thing about herself, but she would show in her use of the new word that she understands it as a colour-word. If we find that what she means by the colour-word she has learned is red, we make use of our own understanding of the word, and of what it is for an object to be red, in saying what she means by the word. When she is fully competent she ² L. Wittgenstein, Philosophical Investigations (tr. G. E. M. Anscombe), Blackwell 1953, §29. (Section numbers in parentheses in the text refer to this volume.)

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



DAVIDSON AND WITTGENSTEIN ON MEANING AND UNDERSTANDING

could say the same thing about herself: “By the word ‘red’ I mean red”. She then uses her recently-acquired colour-word to report what she knows the word she mentions means. She says something she knows, and is now able to say, about the meaning of the word by having learned that language. All mature speakers of a language are in a position to say such things about the words of their language. By making use of words we understand, we can describe correctly the use and meaning of words that speakers we understand use and understand. It is from just such a position that Wittgenstein describes the use and meaning of the words he is concerned with in his own investigations of what words mean or what speakers understand by them. The expressions he is concerned with are words of the languages we speak, and their use and meaning are described by using some of those very words. There is no restriction to describing the conditions of appropriate application of those words only “externally”, so to speak, in terms that presuppose or ascribe to the speaker no understanding of the meanings of any words. So there is no suggestion of behaviourism. This is part of the distinctive accord I find between Davidson and Wittgenstein. Each of them, by his own philosophical lights, describes speakers’ meaning and understanding of expressions from within their engaged, “internal”, and so intensionalist, understanding of the use of terms that we all employ in making sense of one another. Davidson, like Wittgenstein, also stresses the importance of ostension, not just in learning language but also for what the words one learns mean. It is because the meanings of words learned ostensively are “tied to the sorts of objects and events ostended”,³ that “ostensive learning is crucial to the existence of objective thought and language” (DR 694). That kind of learning involves what Davidson calls “triangulation”; a process of interaction between perceivers and objects in the world they perceive. The process in itself does not require language, nor is language always the result. Triangulation occurs whenever one creature sees and focusses on a certain object while also noticing another creature attending both to that same object and to the first creature’s attending to it. Two lionesses engage in triangulation when zeroing in on a prey while keeping each other in sight, for instance, or two baboons circling the ³ D. Davidson, “Responses to Stroud, McDowell, Burge”, Philosophy and Phenomenological Research 2003, p. 692. (Page numbers with ‘DR’ in parentheses in the text refer to this essay.)

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

DAVIDSON AND WITTGENSTEIN ON MEANING AND UNDERSTANDING



same female very much aware of the tactics of the other (DR 693). The point is that this is an essentially social process. No one can do it alone; it takes at least two agents, and it takes interaction between them and objects in their shared world. Although triangulation can occur without language, Davidson holds that triangulation is “a necessary condition of thought and language” (DR 694). What he calls triangulation is clearly present in the social interactions between teacher and pupil that Wittgenstein has in mind in his “ostensive learning”. In that situation triangulation serves to pick out the relevant similarities among the ostended objects that the pupil must catch on to. The teacher rewards what he finds to be similarities among the pupil’s responses to objects that the teacher regards as similar to one another in the relevant respect. Thus the child, learning the word ‘table’, has already in effect noted that the teacher’s responses are similar (rewarding) when its own responses (mouthing ‘table’) are similar. The teacher on his part is training the child to make similar responses to what he (the teacher) perceives as similar stimuli. For this to work, it is clear that the innate similarity responses of child and teacher—what they naturally group together—must be much alike; otherwise the child will respond to what the teacher takes to be similar in ways the teacher does not find similar. A condition for being a speaker is that there must be others enough like oneself.⁴

In the normal situation of a child learning a language from her elders, the pupil eventually comes to share the language with her teachers. But Davidson stresses that the success of a triangular “ostensive learning” procedure does not require that the speakers share a public vocabulary or common patterns of speech. The social conditions of language and thought can be recognized and appreciated even in the absence of a shared language. What the conditions of successful “ostensive learning” do imply, Davidson thinks, is that “there would be no saying what a speaker was talking or thinking about, no basis for claiming he could locate objects in an objective space and time, without interaction with a second person.”⁵

⁴ D. Davidson, “The Second Person” in his Subjective, Intersubjective, Objective, Oxford University Press 2001, p. 120. ⁵ Ibid., p. 121.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



DAVIDSON AND WITTGENSTEIN ON MEANING AND UNDERSTANDING

Davidson, like Wittgenstein, also insists that the triangulation involved in “ostensive learning” of words is not sufficient for understanding what those words mean. A speaker’s having simply learned to utter a certain expression under certain conditions, or to assent to or dissent from an utterance under certain conditions, is not sufficient to determine what the speaker means by that expression, or that he means anything at all by it. Meaning and understanding make their appearance only because of what Davidson calls the “wide variety of social interactions” that “cluster around ostension” (DR 694). Among those social interactions are the kinds of activities Wittgenstein speaks of as “training” or “instruction”. Speakers come to mean and understand something determinate by the expressions they are learning only when they have more or less mastered the grammatical structure of the language in which those expressions are used in sentences. Speakers’ mastery of that rich linguistic and cognitive structure is something that arises only gradually. It is a fundamental idea of Davidson’s that the competence speakers acquire in meaning and understanding expressions of a language can be represented by a description of the compositional structure of the language they have mastered. That would be a description of something all competent speakers of the language know, or know how to do, by having learned the language. Davidson’s idea is to describe the linguistic structure they have mastered in what he calls a “truth-conditional theory of meaning” for that language. That is not a theory of something called meanings which are somehow attached to words. It is a description of the structural interrelations among all the sentences in the language with respect to their expressing something that is either true or false. The idea is that, starting from specifications of the conditions under which certain sentences of relatively simple structure are true, a recursive account of the interrelations among the structures of all other sentences of the language would yield, for each sentence, a statement of the conditions under which that sentence is true. So for each sentence of the language there would be a “T-sentence” that states the conditions under which the sentence in question is true. Since each of those T-sentences is a consequence of the compositional structure of the language, they each could be said to state something that competent speakers know about the sentences they understand. This would explain how speakers can understand sentences they have never heard before. Their competence in their language would be embodied in the source of

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

DAVIDSON AND WITTGENSTEIN ON MEANING AND UNDERSTANDING



their knowledge of the truth of all those T-sentences: their grasp of the compositional structure of the language. Development of a theory of meaning for a language along these lines might seem alien to the spirit of Wittgenstein’s later philosophy— although not, perhaps, equally alien to the conception of language in the Tractatus. Crispin Wright, for instance, finds “a very fundamental tension” between Wittgenstein’s later conception of meaning and understanding and “thinking of language as available even to description by the sort of theory envisaged” by Davidson.⁶ But I think that is to misunderstand the point or function of a theory of meaning for a language as Davidson understands it. I will come back to this question. The goal of Davidson’s theory of meaning for a language is a systematic description of a competent speaker’s knowledge of the structure of a language he understands. That knowledge is what Wittgenstein also has in mind in his metaphor of a speaker’s grasp of the “post” at which different kinds of expression are “stationed” in his language—the different “place, in grammar”, assigned to different kinds of expression (§§ 29, 257). That grasp is something speakers acquire by learning the language. In the same way, on Davidson’s view, it is by learning the language that competent speakers come to know the truth-conditions of the sentences they understand. In knowing those truth-conditions they could be said to know the truth of the T-sentences that express those conditions. But knowing those T-sentences to be true is not speakers’ way into the language in the first place. T-sentences cannot convey the meanings of sentences to those who do not already understand them. You cannot know that the sentence ‘Snow is white’ is true if and only if snow is white, for instance, if you do not already understand the words ‘snow’ and ‘white’, and know what it is for something to be white. After all, those T-sentences that speakers come to know are contingent statements about certain linguistic expressions. The linguistic expressions themselves could have had different truth-conditions from those they actually have, or none at all. So the T-sentences speakers know to be true could have been false. As things are, the T-sentences of an actual language are not false, but that is because of the way the expressions they

⁶ Crispin Wright, “Rule-Following, Objectivity and the Theory of Meaning”, in S. Holtzman & C. Leich (ed.), Wittgenstein: To Follow a Rule, Routledge and Kegan Paul 1981, p.114.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



DAVIDSON AND WITTGENSTEIN ON MEANING AND UNDERSTANDING

mention are in fact actually used. And that is what speakers who learn the language come to know. We imagined earlier a beginning speaker who comes to learn what the word ‘red’ means by successful ostensive learning of the kind envisaged by Wittgenstein. She can report that by the word ‘red’ she means red, or that in applying the word ‘red’ to an object she means that the object is red. She expresses what she understands that word to mean by using the very word she understands within the very language in which the word means what it does. The T-sentences generated by Davidson’s truth-conditional theory of meaning for a language do not themselves say what any expression means, but they also state the truth-conditions of the sentences of the language in a similarly engaged, “internal” way: they use the very words of the sentences they mention with the meanings they have in that language. Even a theory of meaning for a different language states the truth-conditions of sentences of that other language by using (not simply mentioning) expressions in its own language in stating the conditions under which the mentioned foreign sentences are true. There is no other way of doing it. The fundamental idea about the specification of meaning and understanding that Davidson and Wittgenstein share is that you cannot say what an expression means, or what someone understands by it, without using that very expression or others with the same meaning to say what the expression or the person means. Being capable only of such “engaged, internal” specifications of meaning and understanding can be found dissatisfying. Michael Dummett, for instance, long insisted that a theory that gives the truth-conditions of sentences by using words of the very sentences whose truth-conditions are being specified is too “modest” to “satisfy the purpose for which, philosophically, we require a theory of meaning”.⁷ Dummett’s idea is that since a speaker’s employment of his language “rests on” his knowledge of it (DSL xi), possession of that knowledge or understanding must somehow be what explains a speaker’s successfully saying what he says and understanding the things he hears other speakers say. An adequate

⁷ M. Dummett, “What is a Theory of Meaning? (I)”, in his The Seas of Language, Oxford University Press 1993, p. 4. (Page numbers with ‘DSL’ in parentheses refer to this volume.)

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

DAVIDSON AND WITTGENSTEIN ON MEANING AND UNDERSTANDING



theory of meaning, for Dummett, must therefore be “a theory of understanding”. (DSL 4) A “modest” theory of meaning for a language does concern itself in a certain way with speakers’ understanding; it gives what Dummett calls “a theoretical representation of a practical ability” that competent speakers have mastered (DSL 36). In that sense, a “modest” theory of meaning describes what speakers who understand the language know. But Dummett complains that it does not explain “what it is” for speakers to have that knowledge, or what their understanding their language in that way “consists in” (DSL 6). Dummett thinks a philosophical theory of meaning should explain how a speaker’s “general understanding of the language” (DSL 101) is put to work in the speaker’s linguistic performances, or, as he sometimes puts it, it should explain how the “piece of internal equipment”—the understanding of the language possessed by the speaker—is “delivered” to him and “applied when the occasion for its application arises” (DSL xi). How does a speaker’s understanding the meanings of his words operate “to guide, prompt, or control the speaker’s utterances” (DSL xi)? A “modest” Davidsonian theory that simply describes speakers’ competence does not explain that. For Dummett only a more “full-blooded” “theory of meaning and understanding” can account for what he thinks we want to understand. It is difficult to see exactly what a “full-blooded theory of meaning” as Dummett understands it is meant to do, and how it is expected to do it. It looks as if the theory, being “full-blooded” and not merely “modest”, would have to give an account of a speaker’s understanding of his language without presupposing any grasp on the speaker’s part of any of the concepts needed in understanding the language. The theory would have to explain “what it is” for a speaker to understand the expressions of his language in the ways he does without itself making use of those expressions in a way that credits the speaker with some prior understanding of them, or with possession of any of the concepts expressed by those words. It is not easy to understand how any theory could succeed in doing that. I will come back to this difficulty when I return to Wittgenstein. For now it is enough to observe that Davidson does not think a theory of meaning as he understands it performs or is meant to perform the function Dummett complains it cannot fulfill. Davidson does not regard a theory of meaning for a language as a theory or device or “piece of

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



DAVIDSON AND WITTGENSTEIN ON MEANING AND UNDERSTANDING

internal equipment” that speakers consult in deciding what to say, or that they apply directly to utterances they hear to determine what those who utter them mean. Even when two speakers understand and speak the same language, as we say, Davidson denies that in communicating with one another they “operate on the basis of shared conventions, rules, or regularities” determining the meanings of their words.⁸ An “ability to operate in accord with a precise and specifiable set of syntactic and semantic rules” shared by different speakers is “neither necessary nor sufficient for successful linguistic communication”,⁹ Davidson holds. Understanding what another person’s words mean is of course important in communication, but understanding what a person says requires more than knowing the meanings of the words she utters. Davidson illustrates the point with eccentric speakers like Sheridan’s Mrs Malaprop. Mrs Malaprop describes an insulting letter as “a nice derangement of epitaphs”, and a certain person as “as headstrong as an allegory on the banks of the Nile”. Direct application of a theory of meaning for English to Mrs Malaprop’s utterances would have her speaking of epitaphs and allegories, but we all know what she means and says. Even when words are used normally and in accordance with the theory of meaning for the language, Davidson holds that the theory of meaning alone does not suffice for interpreting what a speaker says. That is because saying something is doing something, and no theory of meaning for a language can tell you what a person is doing or will do on a particular occasion. That is something a hearer has to recognize, or sometimes figure out, even if he knows the speaker’s language. Understanding what others say or mean is a matter of what Davidson calls “interpretation”. It involves making the best sense of what a person is intentionally doing in uttering the sounds he makes in the circumstances in which he makes them. That involves ascribing intentional attitudes of thought, meaning, and intention to the speaker. And to ascribe intentional attitudes to others the interpreter must himself possess the concepts he needs for ascribing any attitudes to others, and he must ascribe

⁸ D. Davidson, “A Nice Derangement of Epitaphs”, in his Truth, Language, and History, Oxford University Press 2005, p. 107. ⁹ D. Davidson, “The Social Aspect of Language”, in his Truth, Language, and History, Oxford University Press 2005, p. 110. (Page numbers with ‘DSA’ in parentheses in the text refer to this essay.)

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

DAVIDSON AND WITTGENSTEIN ON MEANING AND UNDERSTANDING



to others the concepts that those others must possess and understand in order to hold the intentional attitudes the interpreter ascribes to them. Davidson starts from the idea that the point of language or speech is communication: “getting across to someone else what you have in mind by means of words that they interpret (understand) as you want them to” (DSA 120). This he sees as common to all verbal behaviour. Meaning, in the special sense in which we are interested when we talk of what an utterance literally means, gets its life from those situations in which someone intends (or assumes or expects) that his words will be understood in a certain way, and they are. In such cases we can say without hesitation: how he intended to be understood, and was understood, is what he, and his words, literally meant on that occasion. (DSA 120–121)

This is clearly not meant as a definition of meaning in terms of speaker’s intentions.¹⁰ There is no elimination of meaning or a reduction of it to something else. Having the intention that one’s audience understand one in a certain way involves having the idea of an expression’s being understood in one way rather than another, and so “assumes the notion of meaning” (DSA 120). That is not surprising if the goal of interpretation is to recognize the intentional attitudes out of which a speaker says and does what he does. Davidson argues that interpreters must find, in general, that those speakers they come to understand share with them many concepts, and so share with them many judgements about what is so in the world. That is a condition of successful interpretation. It is the source of what Davidson calls his “social externalism”: what we all think about and speak about and know are the very objects and people and states of affairs that surround us all. Wittgenstein too makes sense of what people mean and say and think only from “within” the understanding we all have of people as active, speaking agents with intentional attitudes. A speaker who has caught on to her ostensive teaching and training understands the expressions she has learned. In her future use of those expressions she goes on in the same way—in accordance with the relevant similarities—that she learned in her training. We recognize that she understands those expressions because we too understand them in the same way, and we recognize the

¹⁰ As envisaged by, e.g., H. P. Grice in his ‘Meaning’, The Philosophical Review 1957.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



DAVIDSON AND WITTGENSTEIN ON MEANING AND UNDERSTANDING

similarities in terms of which she proceeds as essential to those expressions’ meaning what they do. Wittgenstein addresses repeatedly and from different directions the question of understanding or knowing the meaning of expressions. He shows over and over why understanding cannot be simply the presence before the mind of an image or picture or set of instructions that the speaker attends to or somehow draws on in applying the expression or responding to its use. Not because there are no images or pictures “present to the mind”, but because any object—anything that could be attended to or consulted for guidance in using or responding to a certain expression—would itself have to be understood in a certain way by the speaker, or would have to mean something to her, in order to be any help to her in understanding the expression in question. To take seriously the idea that something present to the mind either is, or is what produces, a person’s understanding an expression in a particular way, is to be led into a regress. A person’s knowing the meaning of an expression cannot be understood, in general, as his possessing something—some object or instruction—that directs or guides or instructs him to respond to a situation or a remark in a certain way. The threat of regress generated by an appeal to objects in the mind to account for meaning and understanding is a constant theme throughout the first half of Philosophical Investigations. It seems to me conclusive as an objection to the appeal to objects in the mind in accounting for understanding. What is needed in the place of objects in the mind is the idea that in coming to know a language and understand its expressions we acquire a certain ability or capacity. Such a capacity is not merely the presence or availability of certain expressions or other items. It is a general competence in using and responding to expressions in appropriate ways. In trying to understand such a capacity, Wittgenstein says it: dawned on him only later”¹¹ that: in philosophy we often compare the use of words with games, calculi with fixed rules, but cannot say that someone who is using language must be playing such a game. (§81)

¹¹ Later than what? After the Tractatus? Or perhaps after discussions with Ramsey? (see §81)

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

DAVIDSON AND WITTGENSTEIN ON MEANING AND UNDERSTANDING



To conclude in the face of this that “our languages only approximate to such calculi” is to stand, as Wittgenstein says, “on the very brink of a misunderstanding” (§81). The heart of the first part of Philosophical Investigations is devoted to identifying and exposing the source of that misunderstanding. That source will have been revealed, Wittgenstein says: only when one has attained greater clarity about the concepts of understanding, meaning something, and thinking. For it will then also become clear what may mislead us (and did mislead me) into thinking that if anyone utters a sentence and means or understands it, he is thereby operating a calculus according to definite rules. (§81)

In his discussions of understanding and meaning, Wittgenstein was fond of examples involving a rule for the development of an arithmetical series. To have caught on to the rule of such a series is to know “how to go on in the same way”, just as in the ostensive learning of words speakers acquire a capacity to “go on in the same way” in applying a word correctly to similar objects or situations beyond those they have encountered earlier. To best expose the potential misunderstanding of how this works, Wittgenstein takes as an example a strict arithmetical rule whose correct application at each point is fully determinate. He illustrates how, even in that case, a speaker who understands the rule, and understands the words that express that rule, and exhibits his understanding of them in applying the rule to particular cases, is not to be understood as operating a calculus according to definite rules. Someone who is given the number 2 and the instruction to “Add two each time”, and understands what she has been told, knows how to go on. In grasping the rule, she could be said to know what number appears at each place in that series, even if she herself continues the series only up to 10. Her understanding far outstrips any actual performance she makes, however long it might be. But her application of the rule at each particular point could be said to be drawn from, or to issue from, her understanding of the general rule. It is because she understands the general rule in the way she does that she knows to put ‘10’ right after ‘8’. The correctness of that step, and of each of the other steps she makes, is felt to be somehow “contained” in her understanding of the general rule, and needs only to be “extracted” from it. It is when we ask exactly how someone who understands the general rule “extracts” from that understanding a particular application of the

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



DAVIDSON AND WITTGENSTEIN ON MEANING AND UNDERSTANDING

rule that we stand on the brink of the misunderstanding Wittgenstein warns against. A person who understands the rule knows that 10 is the number right after 8 in that series, and we feel there must be some way the person knows that. We are tempted to think there must be some operation by which she knowingly makes the transition from the general rule she understands to the particular application of it: some other rule or principle she has mastered that explains how her understanding the general rule leads to her knowing in particular that 10 comes right after 8. This is the source of the question raised by the puzzled speaker in Philosophical Investigations §198: “How can a rule teach me what I have to do at this point?”. If the person understands the general instruction ‘Add two each time’, and she has just written down ‘8’, she wants to know how the rule shows her or tells her what to do next. If, as Wittgenstein warns, we have been “misled” into thinking that “if anyone utters a sentence and means or understands it, he is thereby operating a calculus according to definite rules”, we will try to find the “calculus” the person is operating, the “definite rules” she grasps that take her from her understanding of the general rule and the number ‘8’ she has just written down to her now writing ‘10’. What rules or procedure or instructions direct her from the one to the other? This question could be what lies behind Michael Dummett’s demands on a “philosophical theory of meaning”. He thinks any such theory must be “full-blooded”, as opposed to Davidson’s merely “modest” description only of what speakers know. According to Dummett the theory should explain how a speaker’s knowledge of what his words mean is “delivered” to him for application to particular cases, or how his understanding operates “to guide, prompt, or control” his performances. That is what the puzzled speaker in Philosophical Investigations §198 appears to be asking for. But we have seen that a “full-blooded” theory must apparently explain that procedure or guidance or instruction without itself presupposing any prior understanding on the part of the speaker of any of the expressions involved in the “delivery” of that guidance or instruction. The puzzled speaker who asks how the rule can teach him what to do at this point elaborates his quandary: “After all, [he says] whatever I do can, on some interpretation, be made compatible with the rule . . . So, is whatever I do compatible with the rule?” (§198). Here we have reached the fatal misunderstanding.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

DAVIDSON AND WITTGENSTEIN ON MEANING AND UNDERSTANDING



This speaker appears to seek just the kind of instruction or guidance that Dummett’s requirement on a “full-blooded” theory of meaning for a language is meant to provide: an explanation of how a speaker’s general understanding of his language is put to work by him in the particular applications he makes of it. This is what Dummett found missing from Davidson’s merely “modest” theory of meaning for a language. Much earlier, in his 1959 article “Wittgenstein’s Philosophy of Mathematics”, he also had found no such explanation in Wittgenstein’s accounts of inference and proof. In that paper, Dummett took what he referred to as the “considerations about rules presented in the Investigations and elsewhere” to carry the implication that even if the general rules of inference involved in a proof: had been explicitly formulated at the start, and we had given our assent to them, our doing so would not in itself constitute recognition of each transition as a correct application of the rules . . . Hence at each step we are free to accept or reject the proof; there is nothing in our formulation of the axioms and of the rules of inference, and nothing in our minds when we accepted these before the proof was given, which of itself shows whether we shall accept the proof or not; and hence there is nothing which forces us to accept the proof. If we accept the proof . . . we are making a new decision, and not merely making explicit a decision we had already made explicitly.¹²

On these grounds Dummett attributed to Wittgenstein what he called a “full-blooded conventionalism” according to which “the logical necessity of any statement is always the direct expression of a linguistic convention. That a given statement is necessary consists always in our having expressly decided to treat that very statement as unassailable”. As Dummett observes, “this account is very difficult to accept”.¹³ The puzzled speaker in Philosophical Investigations §198 apparently feels a need for just what Dummett thinks a “full-blooded” theory of meaning should provide. He appears to be looking for some particular way of understanding the expression of a rule he already accepts—‘Add two each time’—that will teach him or show him or in some way determine

¹² M. Dummett, “Wittgenstein’s Philosophy of Mathematics”, Philosophical Review 1959, p. 330. ¹³ Ibid., p. 329.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



DAVIDSON AND WITTGENSTEIN ON MEANING AND UNDERSTANDING

what application he is to make of that rule “at this point”. Wittgenstein observes in Philosophical Investigations §201 that if an interpretation of the expression is needed, and if each “interpretation” offered is itself another expression and must be understood in turn, then still another “interpretation” will be needed to give the meaning of that expression, and so on down a never-ending regress. But people do understand expressions of general rules, and we all can see that they do and that they know how to apply rules to particular cases. So, this quandary must be a result of a misunderstanding of how speakers do in fact understand the expressions of their language and know how to apply them. And there must be misunderstanding of how speakers reach conclusions that they know follow of necessity from something they understand. Wittgenstein says the threat of this regress of “interpretations” shows “that there is a way of grasping a rule which is not an interpretation, but which, from case to case of application, is exhibited in what we call ‘following the rule’ and ‘going against it’ ” (§201). To say it is “exhibited” in what we call “following a rule” and “going against it” is to say that when we understand a certain rule we can usually recognize when our fellow speakers are following it. We can see that a person has the general competence involved in understanding the rule and exhibits it in applying the rule correctly. We can recognize the correctness of what the person does because we too understand the rule and the words that express it. We can see that someone understands the words ‘Add two each time’ and that he follows that rule correctly and so does what we call “following the rule ‘Add two each time’ ”, just as we can see that someone understands the word ‘red’ and applies it correctly to a red apple right before us. That is what we call “understanding and saying truly that a thing is red”. Saul Kripke’s “skeptical paradox”, which is developed from materials he finds in Wittgenstein, leads to the conclusion that there is no such thing as a person’s using an expression with one determinate meaning rather than any other meaning compatible with all the applications he has made of the expression so far.¹⁴ The challenge is to find some “fact about a speaker” that could “constitute” his meaning the expression ‘+’ in one way rather than another. No objects or instructions in the mind, no ¹⁴ Saul A. Kripke, Wittgenstein on Rules and Private Language, Harvard University Press 1982.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

DAVIDSON AND WITTGENSTEIN ON MEANING AND UNDERSTANDING



current or past mental states or experiences, appear to suffice, since their mere presence does not show or imply that the speaker actually means plus by that expression. And adding further such states or expressions would lead only to regress. The fact that the person we are considering is a fellow speaker whom we understand and know to be proficient in arithmetic is disqualified as a candidate “constitutive” fact of his meaning plus by ‘+’ on grounds of “circularity”. Under that restriction, the paradox succeeds, and so draws attention to what still needs to be understood if we are to make sense of someone’s meaning a certain thing by an expression. It succeeds in this form because the range of potentially “constitutive” facts is not allowed to extend to any of those facts we all know about what fellow speakers whom we understand say and mean. We do make sense of others, and of what they mean, but we can say or specify what they mean only by using words of our own to specify it. And we can do that only because we and our fellow speakers understand the terms and concepts we all use in making sense of one another as speaking agents with intentional attitudes. On this fundamental point about the understanding of use and meaning Davidson and Wittgenstein are fully in accord. What lies behind their agreement is the truth of an important observation Wittgenstein wrote in a notebook in 1931, not long after his return to philosophy: The limit of language is shown by its being impossible to describe the fact that corresponds to (is the translation of) a sentence, without simply repeating the sentence.¹⁵

When Wittgenstein speaks here of “describing the fact that corresponds to (is the translation of) a sentence” I take him to mean saying what the sentence means. He says you cannot say what a sentence means without simply repeating the sentence. Perhaps you do not always strictly have to use exactly the same words. Maybe you could say what the sentence says slightly differently without repeating it, or saying in another language what it means. In saying you have to repeat the sentence, Wittgenstein is stressing that you cannot just mention the sentence, or quote it.

¹⁵ Wittgenstein, Culture and Value, p. 10. I have discussed some implications of this remark in my “Meaning and Understanding” in O. Kuusela and M. McGinn (ed.), The Oxford Handbook of Wittgenstein, Oxford University Press 2011 in this volume.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



DAVIDSON AND WITTGENSTEIN ON MEANING AND UNDERSTANDING

I take the main point to be that you must use some words whose meaning you understand in order to say what a sentence means. You give the meaning of the sentence in uttering the words you utter, and anyone who understands you thereby knows what that sentence means. If you said only that a certain sentence you refer to means the same as a certain other sentence that is simply named or quoted or otherwise displayed, you would not thereby say or specify what the sentence means. You would say only that the sentence means the same as some other expression whose meaning you have not stated or specified. It is possible to know, and so to say, that two sentences mean the same without knowing what either sentence means. When you do say what a certain expression means, or what a particular person means by an expression he uses, you must use some words whose meaning you yourself understand which you take to express what the expression, or the other person, means. And you could not recognize what the expression or the person means if you were not master of the grammar and vocabulary needed to say what the expression or the person means. To that extent speakers must share with other speakers they understand a capacity to use the terms or concepts needed to express their shared understanding of those meanings and thoughts. This is the basic idea at the heart of Davidson’s conception of meaning and understanding and his account of interpretation: the intensional vocabulary we use in speaking of what expressions or speakers mean cannot be reduced without remainder to something understood purely extensionally. There is no describing in extensional terms alone our saying and meaning and thinking the things we do. If we tried to understand the things human beings say and do “from outside” all meaning and intention, or as Wittgenstein puts it in Philosophical Grammar, if we treated those utterances and actions only as “a phenomenon or fact”: Whatever phenomenon we saw, it couldn’t ever be intention; for that has to contain the very thing that is intended, and any phenomenon would be something complete in itself and unconcerned with anything outside itself, something merely dead if considered by itself.¹⁶

¹⁶ L. Wittgenstein, Philosophical Grammar (tr. Anthony Kenny), Blackwell 1974, p. 144. (Page numbers with ‘PG’ in parentheses refer to this volume.)

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

DAVIDSON AND WITTGENSTEIN ON MEANING AND UNDERSTANDING



Every phenomenon seems dead in comparison with the living thought.¹⁷

This same idea is to be found in Davidson’s stress on what “interpreters” must share with those they “interpret” and understand. It is the source of his resistance to the idea of a “conceptual scheme” that is totally different from anything we can make sense of, and so is impenetrable to us but possibly available and intelligible to others.¹⁸ In this, Davidson concurs with Wittgenstein that “shared human behaviour is the system of reference by means of which we interpret an unknown language” (§206). But the “behaviour” referred to here is not “behaviour” as “philosophical behaviourism” understands it: something describable without attributing any intentional attitudes to agents. For Davidson, then, as well as for Wittgenstein, “It is not only agreement in definitions, but also (odd as it may sound) agreement in judgements that is required for communication by means of language” (§242).

¹⁷ Ibid. ¹⁸ D. Davidson, “On the Very Idea of a Conceptual Scheme”, in his Inquiries Into Truth and Interpretation, Oxford University Press 1984.

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

Index action and intention 119–27 and wanting 222–9 knowing what you are doing 114–25 Albritton, Rogers 30 Anscombe, G. E. M. 19, 127, 144, 191 on colour concepts 191 Aquinas, Thomas 205 Austin, J. L. 32, 66, 67 Berkeley CA 25, 31–2 Berkeley, George 100, 101, 102, 103 Broughton, Janet 33 Burge, Tyler 68, 89 Campbell, John 33, 89, 111 Carnap, Rudolf 16 Carroll, Lewis 106–8 Cassam, Quassim 62–70 on explaining knowledge 65–70 Cavell, Stanley 31 Chisholm, Roderick 42 Cimabue 15 Clarke, Thompson 31, 33–4 Code, Alan 33 colours as secondary qualities 167–78, 179–93 “deviant” concepts of colour 193–204 in perception and in belief 177–8, 188–92 reality of 167–75, 179–81 what are colours? vs. are objects coloured? 168–71 Conant, James 205–17 on Frege on logical laws 208–13 Davidson, Donald 32, 48, 127, 218–19, 221, 253–65, 271–3 on meaning and understanding 264–5 on theory of meaning for a language 218–21, 260–2 on triangulation (“ostension”) 258–60

Desires as sources of action 222–9 and the objectivity of values 222–9 “motivated” vs. “unmotivated” 222–3 Descartes, René 39, 63, 68, 71–80, 103, 205, 206 Dreben, Burton 29 Dretske, Fred 44, 65, 87, 104, 110 Dreyfus, Hubert 32 Dummett, Michael 250–4, 262–3, 268–9 on “full-blooded” theory of meaning 250–4, 262–3 on understanding meaning 268–9 externalism consequences of 48–54, 77–9 in definition of knowledge 43–7, 56–8, 83 of meaning 47–52, 59–61, 80–4, 92–3, 266–71 Finkelstein, David 141–50 on ascription of feelings 141–3 on expression of feelings 141–5, 147–8 knowing what you feel 148–50 self-ascription of feelings 142–5 Frede, Michael 32 Frege, Gottlob 128, 205–17 Conant on Frege on logical laws 208–13 on logical laws 207–17 Gettier, Edward 42, 58 Ginsborg, Hannah 33 Giotto 15 Goldman, Alvin 44 Grice, Paul 32 Harvard University 28, 30–1 Hegel, G. W. F. 39 Horstmann, Rolf Peter 130–40 on self-consciousness and propositional thought 130–5 Hume, David 39, 100, 102, 103, 138, 182, 225–7

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi



INDEX

Johnston, Mark 167–78 on Ramsey-Lewis theory of the colours 168–70 on “unmasking” the colours of things 171–3 what are colours? vs. are objects coloured? 168–71 Kant, Immanuel 18, 28, 29, 39, 56, 58, 80, 103–5, 113, 128, 130, 151–66, 205, 207, 209, 213, 215–16, 254 necessity and the a priori 152 necessary for thought vs. necessarily true 155–62, 165–6 knowledge a posteriori 114–15 a priori 114–15, 152 definition of 42–5 externalist theories of 43–7, 56 of the “external world” 40–8, 55–61, 62–70, 71–9, 86–8, 93–6, 102–3, 109–12, 119 of what you are doing 114–25 Kripke, Saul 246–8, 270–1 Lewis, C. I. 16, 41, 42 Lewis, David 168–70 Locke, John 100, 102, 103, 151, 181 McDowell, John 91, 97–8, 145, 179–92, 221, 253 on colours as “secondary” qualities 179–92 on perception of the colours of objects 181–4, 185–6, 188–92 on “perceptual experience” 187–9 Maddy, Penelope 75 Malraux, André 14–15 meaning and social practice 256–9, 265 and use 242–6, 255–9 “full-blooded” theory of 250–4, 262–3 giving the meaning 234–6, 248–9, 271–2 learned by ostension 251–3, 256–67 theory of, for a language 218–22, 260–4 understanding of 236–46, 266–71 what is intended 264–5 Moore, G. E. 15, 16, 21, 41 Myro, George 32

Nagel, Thomas 21, 32 necessity explanation of necessity? 152, 206–8, 213–17 necessary for thought 155–62, 165–6, 207–8, 213–17 necessary truth 55, 152–5 Nietzsche, Friedrich 24, 36 Nozick, Robert 44 Owen, G. E. L. 30 Oxford 17, 30, 32, 33 Pears, David 30 perception problems of 40–3, 45–7, 71–9 proper “objects” of 65–70, 86–95, 100–13 knowledge by 93–6, 109–10, 119–21 restricted scope of 41, 45, 47, 50–2, 58, 73, 92, 100–2 Pfisterer, Christoph 113 philosophy difference from religion 11–13 difficulty of ‘what is it?’ 7–8, 18 history of 14–15, 20–1, 26–8 not doctrine 11–14 “professional” state of 11–12, 17–18, 25–6 “treating” philosophical questions 18–23 what we seek in 36–9 Plato 27, 39, 40 Platts, Mark 213–32 on “motivated”, “reason-following” desires 221–3, 227–9 on Davidson’s theory of meaning 218–21 on meaning of moral sentences 221–3 Price, H. H. 41 Pryor, James 74 Putnam, Hilary 207 Quine, W. V. 16, 27, 29–30, 42 on meaning as use 255–6 Ramsey, Frank 168–70 reasons for action 222–3, 227–31 for belief 105–9 Ricks, Christopher 22

OUP CORRECTED PROOF – FINAL, 27/4/2018, SPi

INDEX

Roessler, Johannes 123 Russell, Bertrand 16, 39, 77 Scanlon, Thomas 108 Scheffler, Samuel 33 Searle, John 31 seeing “perceptual experience” 96–7, 100–3 seeing and knowing 93–6, 109–10, 119–21 seeing an object 65–9, 86–90, 100–11 seeing that p 69–70, 79–85, 88–97, 109–12 self-consciousness as a capacity 134–40 as a condition of propositional thought 130–5 what is one conscious of ? 137–8 Sheridan, Thomas 264 Shwayder, David 32 Sluga, Hans 32 Thales 27 Toronto, University of 27, 28 UCLA philosophy 33 understanding and meaning 218–21, 234–45, 249–54, 264–8



colour words 169–71, 175–7, 181–7, 189–92, 193–204 human knowledge 40–3, 47–8, 55–61, 62–70, 71–9, 86–8, 93–6, 102–3, 109–12, 119 in philosophy 12–14, 16–19, 21–3, 27, 34–7 oneself as an agent 121–8, 130–40 others as thinkers, believers 52–4, 81–3, 91–3, 136–7 White, Morton 29 Wiggins, David 221 Williams, Bernard 24, 32, 73 Wittgenstein, Ludwig 19, 20, 30, 31, 34, 143–50, 193–204, 205, 211–12, 233–54, 255–60, 266–73 on “ostensive” learning 256–8 on understanding 235–45, 249–54, 265–8 on understanding only “from inside” language 236–40, 244–5, 251, 265–8 Wollheim, Richard 32 Woodrow Wilson Foundation 28 Woods, John 27 Wright, Crispin 247–8, 253, 261