Perspectives on Perception 9783110327557, 9783110327298

Citation preview

Papers offered to Alan Lacey

Mary Margaret McCabe, Mark Textor (eds.) Perspectives on Perception

Philosophische Forschung Philosophical Research Herausgegeben von / Edited by Johannes Brandl • Andreas Kemmerling Wolfgang Künne • Mark Textor Band 6 / Volume 6

Mary Margaret McCabe, Mark Textor (Eds.)

Perspectives on Perception

ontos verlag Frankfurt I Paris I Ebikon I Lancaster I New Brunswick

Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliographie; detailed bibliographic data is available in the Internet at http://dnb.ddb.de

North and South America by Transaction Books Rutgers University Piscataway, NJ 08854-8042 [email protected]

United Kingdom, Ireland, Iceland, Turkey, Malta, Portugal by Gazelle Books Services Limited White Cross Mills Hightown LANCASTER, LA1 4XS [email protected]

Livraison pour la France et la Belgique: Librairie Philosophique J.Vrin 6, place de la Sorbonne ; F-75005 PARIS Tel. +33 (0)1 43 54 03 47 ; Fax +33 (0)1 43 54 48 18 www.vrin.fr



2007 ontos verlag P.O. Box 15 41, D-63133 Heusenstamm www.ontosverlag.com ISBN 13: 978-3-938793-73-2 2007 No part of this book may be reproduced, stored in retrieval systems or transmitted in any form or by any means, electronic, mechanical, photocopying, microfilming, recording or otherwise without written permission from the Publisher, with the exception of any material supplied specifically for the purpose of being entered and executed on a computer system, for exclusive use of the purchaser of the work

Printed on acid-free paper ISO-Norm 970-6 FSC-certified (Forest Stewardship Council) This hardcover binding meets the International Library standard Printed in Germany by buch bücher dd ag

Table of Contents

List of Contributors Preface

1

1. A bit of autobiography Alan Lacey

3

2. A puzzle about how things look R. M. Sainsbury

7

3. The problem of consciousness and the innerness of the mind Jim Hopkins

19

4. Seeing an individual Keith Hossack

47

5. Seeing something and believing IN it Mark Textor

65

6. Phenomenal concepts are not demonstrative David Papineau

87

7. Kant on the a priori content of perceptual experience Anthony Savile

111

8. Self-awareness Richard Sorabji

131

9. Perceiving that we see and hear: Aristotle on Plato on judgement and reflection Mary Margaret McCabe

143

List of Contributors Jim Hopkins is Reader in Philosophy at King’s College London. Keith Hossack is Reader in Philosophy at Birkbeck College London. Alan Lacey is Senior Lecturer Emeritus in Philosophy at King’s College London. Mary Margaret McCabe is Professor of Ancient Philosophy at King’s College London. David Papineau is Professor of the Philosophy of Science at King’s College London. R.M.Sainsbury is Susan Stebbing Professor of Philosophy at King’s College London and Professor at the University of Texas, Austin. Anthony Savile is Professor of Philosophy Emeritus at King’s College London. Richard Sorabji is Professor of Ancient Philosophy Emeritus at King’s College London. Mark Textor is Reader in Philosophy at King’s College London.

Preface On May 17th 2006 Alan Lacey had his 80th birthday. Alan is a greatly loved and valued member of the Department of Philosophy at King’s College, London. Accordingly we decided that we would like to mark this birthday in some way. Because Alan enjoys discussion so much – and because we all enjoy discussion with him – we planned a conference for his birthday, to be followed by a collection of papers arising from the conference. The conference took place on 10th and 11th May 2006; and this is that volume, which we offer to Alan with our affectionate respect. At Alan’s retirement dinner fifteen years earlier, Richard Sorabji read a poem which he had written for Alan. It was read again, with a supplement, at the conference dinner on 11th May. Here it is. Ode to Alan Lacey Though PLATO’s works include EPINOMIS Most scholars say ‘let’s give the thing a miss’; But ALAN showed his versatility By tackling this, and if the Will is Free: His first two papers. Next the dreaded FORMS Bowed to his pen, and SIDGWICK’S Moral Norms. Fresh woods to find, he asked ‘Is Life a Dream?’ And wrote on ‘Man and Robots’ in between. One saying he thinks true, if any is: That Thoughts are something SUI GENERIS. Entailment next absorbed his fertile mind, And paradoxes of the ‘Preface’ kind. ‘Can Change be Real?’ the ELEATICS cried. ‘If Substance is,’ the STAGIRITE replied. ‘What do we know of Socrates’ thought?’ The answer ALAN LACEY gives is ‘nought’. His articles conclude with JOSEPH BUTLER And thoughts on mind and body – even subtler. But these are just the OPERA MINORA, The KLEINE SCHRIFTEN of a Great Explorer. Now HENRI BERGSON fills his waking hours, And JOHN PHILOPONUS attests his powers.

[Addendum for 11th May 2006 This once I wrote, but now to recompose it I have to answer why the book on Nozick? The comprehensive range is why he chose it. There is no knowledge but that Alan knows it.] Once earth such minds as ISIDORE could show, Or his successor DENIS DIDEROT. What scholars now can match the mighty span Of ALAN, the ENCYCLOPAEDIC MAN? There is a saying that the Wise record: The Pen is ever mightier than the Sword. There’s one thing we should add of mortal men: The PERSON is still greater than the PEN. Richard Sorabji

In both the conference and in collecting the papers we incurred several debts of gratitude. First, the School of Humanities at King’s supported the conference with a grant: we are most grateful. Second, we should like warmly to thank all those who presented papers at the conference, and all those who contributed to the excellent discussions: the contributors to this volume; two other speakers whose papers were promised elsewhere, Peter Goldie and Scott Sturgeon; the audiences throughout; and of course Alan himself, who offered comments on each paper after it was delivered.

Mary Margaret McCabe Mark Textor Department of Philosophy King’s College London

2

A bit of autobiography Alan Lacey

I was born, so my mother assured me, in Birmingham, four hours after my sole (and stillborn) sibling, at 5.30 pm on Monday 17 May 1926, to a provincial bank manager and a hospital sister. When I came to boy’s estate I went to a public school where I specialised in Classics at an absurdly early age and as a result studied no science of any kind at all (a fate, I should add in fairness to the school, which would not be inflicted on any pupil going there now). School was followed by the army, where my job, which involved a tiny bit of Japanese at Bletchley (though nothing to do with Enigma, I’m afraid), ended suddenly when the war did. So in August 1945 I spent two glorious months reading Burnet and Gomperz on the Presocratics in the mornings, cycling around Buckinghamshire in search of blackberries in the afternoons, and curled up with a custard pie and a book in the peaceful cacophony of the NAAFI in the evenings (you could go in the Quiet Room, but there would be certain to be a whispered conversation going on, which did not make for easy concentration; also they didn’t serve custard pies in the Quiet Room). It was then that I found such books as Mark Twain’s What Is Man? which I described in my notebook as the finest book on ethics I’d ever read, which was strictly true, as it was the first book on ethics I’d ever read. Arriving eventually at Cambridge in 1946 I read Part I of the Tripos in Classics and Part II in English, changing mainly to find out at least something of what happened after the death of Marcus Aurelius. My favourite paper there was ‘The English Moralists’, affectionately known as ‘The English Moralists: Plato to Hegel’ (the first question I answered in finals was simple, at least to state: ‘Compare the ethics of Aristotle and St Paul’). My last tutor in English was Theodore Redpath, a personal friend of Wittgenstein (as I only discovered much later), and he put me on to doing a Ph.D.in Moral Sciences on the ethics of Sidgwick, supervised by C.D.Broad. (How easy it was to change faculties or departments in those days!) I was just too late to hear the big three, Moore, Russell and

Wittgenstein, at the Moral Sciences Club, though almost the first thing I did way back in 1946-7 was attend Russell’s lectures later published as Human Knowledge, along with what felt like almost the whole of the rest of Cambridge in the packed three ground floor lecture-rooms of the old Mill Lane building – taking comfort ever since from the fact that even as brilliant and eminent a lecturer as Russell could find his audience diminish to a third of its original size by the end of the second term. All good things end eventually, of course, and my halcyon days in Cambridge did so in 1952, whereupon, after two years in the grey wastes of Manchester (a comment on the city, of course, not on the University) learning about metaphysics and Greek philosophy from Dorothy Emmet and George Kerferd, I ended up at Bedford College London in 1954. I transferred with the Philosophy Department to King’s College London in 1984 and retired as a Senior Lecturer officially in 1991 and properly (i.e. stopping the part-time teaching I’d been doing) in 1998. Philosophers in the relevant area, it seems to me, could be motivationally divided into two main types (overlapping of course), which I’ll call metaphysicians and epistemologists. Metaphysicians primarily want to find out the truth about things. What is the universe really like? How many fundamentally different kinds of things does it contain? What is it to be real, and what kinds of things are real? What are we? These and similar questions look outwards from ourselves, and those for whom they are the questions of primary interest might be called extrovert in a relevant sense. Epistemologists on the other hand are introvert. They look inside ourselves and ask how we are related to the world; in particular what can we really know about the world or anything in it, and how? ‘How do you know?’ is their favourite response to any assertion, because they fear that unless we can answer it we shall be cut off from the world in solipsistic isolation. My own motivation is primarily extravert. I want to rush ahead and find out what the world is like, before coming back to the hard slog of how I can justify saying that I have found out. But the introvert approach, though secondary, does get a good look in too of course, especially on issues such as antirealism and realism. Perception is a mode of knowledge and so its relevance for the introvert motivation is clear enough. But it also bears on what the world is like, in its relevance to realism versus idealism

4

(and realism versus antirealism), and in particular to the mind/body problem (‘What are we?’). Take colours, for instance. Red, blue and yellow are associated with various sets of reflectances, but how do the purely qualitative primary hues, opaque to all analysis, it seems, relate to the quantitative notion of reflectances? So it strikes me that perception, though I did not choose it, is an excellent choice for the present volume. Naturally I am both highly gratified and flattered to be the volume’s dedicatee, a role which I am afraid I have done nothing to deserve except by becoming an octo, which I have done by following Groucho Marx’s advice to just keep on living.

5

A puzzle about how things look R. M. Sainsbury

1. Imagine you are looking at a uniformly colored object whose parts are differently illuminated. Perhaps it is someone’s orange shirt: he is sitting by the window, so that the upper half is brilliantly lit by direct sunshine, whereas the lower half is not. It does not look as if he is wearing a bicolor shirt: on the contrary, it looks orange all over, the same orange top and bottom. On the other hand, the top and bottom look different: the top half looks significantly lighter than the bottom. On the face of it, we have an inconsistency in the appearances: the shirt looks uniformly orange, and yet the parts look different in color. It may be that appearances are occasionally, and under extraordinary circumstances, inconsistent: one might cite the waterfall illusion or the appearance of one of Escher’s paradoxical drawings. But the present example is of the most ordinary possible kind. We hardly ever, in normal circumstances, see things which are completely evenly illuminated. It seems mad to suppose that there is hardly ever, in normal circumstances, consistency in visual appearances. This paper shows how madness can be avoided. The relevant phenomenon is related to the fairly well understood phenomenon of ‘lightness constancy’. Normally, as in the case of the orange shirt, the mechanism works well, ensuring that appearances are not deceived by different levels of illumination. But it can be fooled, as this example by Edward Adelson shows:

©1995, Edward H. Adelson http://web.mit.edu/persci/people/adelson/checker-shadow_illusion.html The very mechanism which makes the orange shirt look uniform makes squares A and B look disparate, though they are the same shade of grey (check the image on the right). In this case, the visual system provides no correction, and we have an illusion: in no sense and in no way do these squares look the same grey in the figure on the left. By contrast, in the case of the orange shirt, vision appears to deliver two verdicts, both in some sense correct: that the shirt is uniform orange, and that the two regions differ in color. The experience of the orange shirt is one we intuitively wish to count as veridical. There is no illusion, not the slightest tendency to form a false belief, and no sense of a divergence between appearance and reality. Yet on natural accounts of veridicality, we can shift the inconsistency from within the appearance into reality itself. V1: If an object, x, looks F, this experience is veridical iff x is F. V2: If objects x and y look different in color, this experience is veridical iff x and y are different in color. These principles about veridicality, given our assumptions about looks stated by (1) and (2) below, transport the inconsistency from within appearance into the world of fact: 1. The shirt looks uniformly orange. 2. The upper and lower parts of the shirt look different in color. 8

3. The experiences which make (1) and (2) true are veridical. 4. The shirt is uniformly orange. (1, 3, V1) 5. The upper and lower parts of the shirt are different in color. (2, 3, V2) The inconsistency between (4) and (5), a consequence of seemingly platitudinous premises, is the puzzle about how things look which this paper addresses. It is not difficult to come up with a redescription of the situation which has no obvious gaps yet which contains no inconsistency. The shirt looks to be uniform in color, though the plainly visible different levels of illumination ensure that different shades are presented. In the end, such redescriptions hold the solution. But just as they stand, they do not tell us what is wrong with the original way of putting things: the shirt looks uniformly orange all over, yet the top part looks different in color from the bottom. In addition to providing a consistent redescription, a proper account should show how and why (if at all) the original description goes wrong. 2. Any attempt to avoid inconsistency will consider the possibility that we have ambiguity or some similar phenomenon. In particular, it might be suggested that a familiar distinction between so-called epistemic and phenomenal senses of ‘look’ should be brought to bear on our puzzle. ‘Look’ in its phenomenal sense is supposed to be used merely to report appearances, not things we infer from them, whereas ‘look’ in its epistemic sense reports the result of an appearance-based inference. A standard example of the latter is 6. The neighbors look to be away which, if true, does not involve seeing the neighbors, and so cannot be regarded as a report of their visual appearances. Rather, I see other things (the overstuffed mailbox and empty garage), and these present appearances on the basis of which I infer the absence of neighbors. It is fairly hard to find a clear candidate for an epistemic ‘look’ which is followed simply by an adjective (like ‘orange’), or adjective

9

phrase (like ‘uniformly orange’). In (6), ‘look’ is followed by a verb in the infinitive. This might prompt the thought that in the following pair 7. The shirt looks uniformly orange 8. The shirt looks to be uniformly orange (8) represents an epistemic rather than a phenomenal ‘look’. This would be offered as an interpretation of (1), (2) would be interpreted phenomenally, and our contradiction in appearance would disappear. The conflict would be of a common enough kind, between how things look and how we are inclined to judge them to be. This easy solution does not work. First, it is clear that ‘looks’ plus infinitive can be used to make phenomenal as opposed to epistemic reports. Seeing a Müller-Lyer puzzle, it seems quite proper to say that the lines look to be different in length, even if one has not the slightest disposition to believe that they are different in length. Second, it is clear that both (7) and (8) can properly be used to make purely phenomenal reports. The aim in both cases may properly be simply to describe the visual appearances, and not to say what one has inferred. Similar points hold for associated constructions, for example ‘looks like’ and ‘looks as if’. Even when an epistemic interpretation has been set aside, we can still generate an apparent inconsistency with phenomenal readings. The different appearance of, on the one hand, a two-colored shirt and, on the other, a uniformly colored shirt differentially illuminated is fairly hard to describe in words, and considerable painterly skill is required to represent it on canvas. (Leonardo da Vinci said: ‘light and shade should blend without lines or borders, in the manner of smoke’ (Notebooks §492). Adelson (1995) says that the relatively gentle luminosity gradient generated by shadows is an important cue.) The difference, however exactly it is cued, is quite apparent: it is a difference in appearances, and is clearly phenomenally available. We can normally easily tell just by looking which case we are dealing with. In my envisaged situation, the shirt looks uniformly orange, and looks to be uniformly orange, and looks as if it is uniformly orange, and looks very different from a two-colored shirt.1 We can1

The infinitive is needed whenever something non-adjectival follows ‘looks’: ‘looks to be a car’, ‘looks to be dancing’. Adjectives can feature as complements to forms of the verb ‘to be’ (‘is orange’), and can also be attached directly to nouns (‘orange shirt’); this may explain why they can either directly follow ‘look’ or else be part of the infinitive construction.

10

not use the distinction between epistemic and phenomenal senses of “looks” to undermine our determination to affirm (1), understood as a phenomenal report. We would have a flat contradiction in appearances if the following was true: 9. The shirt looks (to be) non-uniform orange. But this seems not to be so, especially if the parentheses are deleted. (9) looks made for reporting the case in which the shirt looks multi-colored, and that is not how the shirt looks. Bearing this in mind, we need to reexamine (2) itself: The upper and lower parts of the shirt look different in color. We need to understand how the parts of the shirt could look that way, without the shirt looking non-uniform. The explanation seems to be this: sometimes colors appear as colors of surfaces, and sometimes they do not. When they appear as colors of surfaces, the surfaces appear to be colored. But when colors appear as, for example, colored lights, the appearance is not as of a colored surface. I suggest that the explanation of the consistency in the appearances reported by (1) and (2) is that (2) need not be understood as reporting a difference of apparent surface color, as opposed to a difference in apparent color; whereas it is very difficult to understand (1) except as reporting on an apparent surface color. The upper and lower parts of the shirt look different in color, but do not look to be of different colors. They present different colors, but do not present surfaces as differing in color. The inconsistency in appearance is revealed as illusory. But this is not the end of the story, for the proposed interpretation of (2) does not show it to be false; on the contrary, the interpretation explains what makes (2) true. So if the veridicality principles (V1) and (V2) are correct, we still have an inconsistency, one that cannot be ignored, for it is an inconsistency in fact: the inconsistency of (4) and (5). 3. With the distinction between surface colors and others in hand, we are going to need more than one principle of veridicality: at a minimum, one

11

which says that surface looks are veridical just when a surface has the apparent color in question, and another which says that the appearance of a color is veridical just when that color is present, not necessarily as a surface color. A surface color is simply the color possessed by some surface. For color-sighted people, seeing a colored object in suitable circumstances gives rise to an appearance which is not only the appearance of a surface color, but an appearance which is distinctive of surface colors. Typically, the appearance to which a surface color gives rise is manifestly different from the appearance to which a non-surface color gives rise: when we see colored light, or see the sky, or see water underwater, or water from above at night when it is illuminated from below, or a work of James Turrell, it does not look as if we are seeing a colored surface. In seeing something whose surface is red, one typically sees the color red, and also sees the surface as red (it looks red). But one may see a color without seeing anything looking like a surface of that color: looking at the sky on a clear day, it looks blue, but it does not look as if the sky, or anything else, has a blue surface. Awareness of a color as a color of a surface guarantees awareness of that color, but the converse does not hold. When we think about the color-appearances of colored objects, we naturally think about surface color. This is what underlies (V1) (If an object, x, looks F, this experience is veridical iff x is F). Typically when we ascribe a color to an object with colored surfaces, we think of our ascription as true iff a surface has the color we ascribe. It was in the nature of our example that this did not lead to trouble with the claim that the shirt looked uniformly orange. We took the reported experience to be an experience in which surface color is manifest. Veridicality was accordingly assumed to consist in the manifest surface color being the color of the shirt’s surface. For some kinds of colors with which we are presented, including those presented as non-surface colors, this account of veridicality will be inappropriate. This includes the case in which the shirt presents different colors, though not as surface colors. These are colors which do not look to be possessed by any surface. That is how things are with the apparently lighter upper part of the shirt and the apparently darker lower part. The point can be made in more detail by moving away from English to an extension of English which has no word ‘looks’ but a number of word-forms on the pattern ‘X-looks’. Here are three of these unfamiliar words, with approximate English equivalents alongside. ‘F’ is to be replaced by a word for a color or shade of color.

12

x S-looks F x V-looks F x P-looks F

x looks as if it has an F surface x looks like a volume suffused with color F x presents a color that looks F, though x neither S-looks F nor V-looks F

The table is not supposed to be complete. For example, the ‘V-looks’ category may need subdivision, to deal with the different looks of dyed water and water illuminated by colored light; and further main categories may need to be added. The sky V-looks blue, but does not S-look blue (no surface is visible). A ripe tomato typically S-looks red, even when nothing V-looks red. If we merely shine a magenta light on a piece of white card, the card continues to S-look white, even though it also P-looks magenta. Unless we take a lot of care, the result of our illumination is just to make the card look to be a white card with magenta light shining on it. (With care, we can make the card S-look magenta.) This example would tend to generate an apparent inconsistency in appearance when described in ordinary English, for we might incline to say that the card with the magenta light shining on it looked white (all over) and also looked magenta (all over); these colors are exclusive. The veridicality conditions for S-looking and V-looking are clearly distinct, so something (e.g. the sky) can V-look F without S-looking F (or anything else). One might at first think that P-looking cannot fail to be veridical, but there is no guarantee of this. For all I know, if one suffers from jaundice, white things present a pale yellow appearance without Slooking yellow. (This may not be how things are in the disease, but it is not ruled out a priori.) If so, we would wish to say that the yellow P-look is non-veridical. The envisaged X-look language enables us to describe some familiar features of appearance that are hard to describe in English. When we put on sunglasses things in a sense look darker but (as we might put it) do not look as if they are darker, a case which generates the original puzzle. The contrast is not merely that between phenomenal and epistemic looks. As I put on the sunglasses I have no tendency to judge that any object becomes darker. In addition, there is a clear sense in which the objects do not look darker. In X-language the threatened inconsistency in appearance is avoided: objects S-look the same before and after putting on the glasses, but they do not P-look the same.

13

P-looking is so defined as to exclude S-looking and V-looking. But is there not a more general notion, call it P*-looking, which includes the other kinds of looking, and which is the basic form of a mind’s experiential confrontation with a color? We might think of it as a color being presented, but without specification of whether it is presented as a surface color, a volume color, or whatever. For every X, if something X-looks F, it P*looks F. Then we seem to be saddled with contradiction in appearance, for it seems we must say both that the shirt P*-looks uniform in color and also P*-looks non-uniform. I think it is wrong to say that such a general notion of looking has any part to play in describing our visual experiences, at least where colors are concerned. When the shirt looks uniform orange, it is not that a uniform orange color is presented. S-looking does not guarantee P*-looking, if ‘P*’ connects in any natural way with the ordinary notion of presentation. What may have been overlooked is that there is a purely phenomenal notion of S-looking which does not always amount to the presentation of a color. It is not that we infer that the shirt is uniform orange: that is how it looks. Yet there is no presentation of a uniform orange color which extends across the shirt. That is just how things are with how things look; and we should welcome it, for this is what ensures that there is not really a contradiction in appearance. The veridicality conditions for the different X-look words are obvious: x veridically S-looks F x veridically V-looks F x veridically P-looks F

x has an F surface x is a volume suffused with color F x presents color F

It is also obvious how to extend these to comparative cases.2 The orange shirt veridically S-looks a single shade of orange all over, and veridically P-looks different shades. So the shirt has a surface which is a single shade all over, but presents different shades. What else could one expect from different levels of illumination?

2

There are other normative dimensions of assessment of appearances, for example, something may look just as it should (to a creature with such-and-such a perceptual system in such-and-such circumstances).

14

4. One might attempt to resolve the puzzle in other ways. I will mention two such alternatives. (A) ‘Darker than’ is unspecific or ambiguous between darker in color than, call this darkerc, and less brightly illuminated than, call this darkeri. The lower part of the shirt looks darkeri than the upper part, but does not look darkerc. So it does not look darker in color than the upper part, and so does not look different in color from the upper part; only differently illuminated. One problem with the proposal arises from the close semantic connection between the comparative and non-comparative forms: both or neither should display ambiguity or underspecification. Hence the present proposal entails that ‘dark’ in, for example, ‘dark orange’, can properly be interpreted as meaning ‘less brightly illuminated orange’. This result is incorrect: dark orange is a color, not some kind of pairing of a color and an illumination level. The proposal is also lacking in generality. Lacking the resources of the X-looks language, we can create similar puzzles that do not depend on ‘darker’ or ‘dark’. The swimming pool (or the water in it) illuminated by a pink underwater light at night veridically looks pink, but it is not pink; hence we must find an alternative account of veridical looking in this case. Likewise the sky veridically looks blue, even though it has no blue surface (and so counts as ‘not blue’, as this phrase is most often used), and an apple may veridically look red even though it has no volume suffused with red. (B) Color words are semantically underspecific or ambiguous or polysemous: ‘x is orange’ may be true just if x has an orange surface, or just if x is a volume suffused with orange, or just if x is an orange light. We don’t need different ways of looking, only different ways in which colors can find a place in the world. For example, instead of ‘x S-looks F’ we should say ‘x looks FS’, where ‘F’ is to be replaced by a color adjective, and the subscript shows that it is to taken as ascribing a surface color. Instead of ‘x V-looks F’ we should say ‘x looks FV’; and so on. There is one way of looking, but many ways color can relate to objects.

15

The idea has some appeal. For one thing there are certainly many ways for colors to exist, not just as surface colors. And there is independent evidence for the relevant behavior of color adjectives.3 But the proposal seems not to do justice to all the cases, including the original one: ‘orange’ is used just once in (1) and (2), so supposing the word to be polysemous does not immediately explain the phenomenon. Another conspicuously hard case for this approach is the change in look produced by putting on sunglasses. We can describe the effect without using any specific color words (‘Everything looks darker, though nothing looks to have changed color’), so polysemous color words don’t seem to be able to begin to address the case. 5. When we speak the X-looks language, there is no inconsistency, within or outside of appearances, as we saw in §III above. Nothing follows about how things are when we speak English. A simple conjecture extends the happy result for the X-look language to English: English ‘looks’ is ambiguous between, or unspecific between, or polysemous between, the meanings of the various X-looks expressions. The upshot is (i) that we have inconsistency in appearance only when there is inconsistency in the content of ‘looks’ sentences which have been rendered unambiguous or specific in the same way and (ii) that there is no single veridicality condition for English ‘looks’ sentences. In the original account of how the shirt looked, the two occurrences of ‘looks’ (in (1) and (2)) need to have different disambiguations or specifications if they are both to be true; each corresponds to a different veridicality condition. As a result, there is no inconsistency in appearance, and, a fortiori, no inconsistency in fact. A further conclusion is that all arguments (and there are many) which trade on a single veridicality condition for ‘looks’ need to be reconsidered.4

3

There are well-known cases in which surface and interior colors are contrasted, as in Travis (1997) and Bezuidenhout (2002). 4

Many thanks to Mike Martin and Michael Tye for valuable comments.

16

References Adelson, Edward H. 1995: http://web.mit.edu/persci/people/adelson/checkershadow_illusion.html Bezuidenhout, Anne 2002. ‘Truth conditional pragmatics’. Philosophical Perspectives, 16, pp. 105–34. Travis, Charles 1997: ‘Pragmatics’. In: Hale, Bob and Wright, Crispin (eds.) A Companion to the Philosophy of Language. Oxford: Blackwell Publishers, pp. 87–107. da Vinci, Leonardo, Notebooks. Various editions and translations, for example: http://www.gutenberg.org/etext/5000

17

The problem of consciousness and the innerness of the mind Jim Hopkins 1. Consciousness and innerness The problem of consciousness is commonly taken to be that of accounting for aspects of experience which we describe as inner, and also as phenomenal, subjective and private. Thus for example David Chalmers introduces what he calls ‘the real mystery of the mind’ as follows: From an objective viewpoint, the brain is relatively comprehensible. When you look at this page, there is a whirr of processing: photons strike your retina, electrical signals are passed up your optic nerve and between different areas of your brain … . But there is also a subjective aspect. When you look at the page you are conscious of it, directly experiencing the images and words as part of your private, mental life .... The hard problem...is the question of how physical processes in the brain give rise to subjective experience. This puzzle involves the inner aspect of thought and perception ... . It is these phenomena which constitute the real mystery of the mind. (Chalmers 1995, p. 92)

In speaking of ‘the inner aspect of thought and perception’ Chalmers draws on a familiar feature of our everyday conception of the mind. We speak of the mind as a locus which things are in, and these things as therefore having an internal aspect; likewise we speak of ourselves as aware of things from the inside, or by introspection; of our inner lives; and so on. We seem to speak and think in this way from early childhood, and to distinguish things which we take to be ‘in the mind’ from the physical things we can see or touch. Thus as Henry Wellman writes With regard to ontology, children as young as three years firmly divide the mental and physical worlds … if told about someone who has a dog that has run away, versus someone who is thinking of a dog, three-year-olds know that neither ‘dog’ can be seen or petted, but that none the less one is mental (‘just in his mind’, ‘only imagination’) whereas the other is physically real, and merely unavailable. In these ways young children appropriately distinguish between real and mental entities. (Wellman 1993, p. 14)

Our notion of things as internal to the mind thus seems to go with a natural intuitive dualism. Philosophical writers have developed this idea in many ways, often speaking as if the mind were a kind of inner place or space, in which mental phenomena are somehow housed. Augustine, for example, stressed how as a baby he found ‘my wants were within me and those others [who might satisfy them] were outside; nor had they any faculty enabling them to enter my mind. So I would fling my arms and legs about and utter signs ... .’ (Augustine, p. 6) And we can readily skip the centuries to comparable recent discussions, such as the following by Thomas Metzinger. To be able to speak seriously about a science of consciousness, a number of fundamental questions would have to be answered. It is interesting to note that with the emergence of consciousness private worlds – spaces of inner experiences – are opened up. These spaces, however, are individual spaces: ego-centres of experience that suddenly appear in a centerless universe. Each such centre of consciousness constitutes its own perspective on the world. This perspective is what philosophers sometimes like to call the 'first-person perspective'. A phenomenal world of its own is tied to each of these perspectives. These individual worlds of experience also possess a historical dimension: almost always a psychological biography emerges together with them – what we call our 'inner life'. This too can be seen as the history of the genesis of a world, or a phenomenal cosmology: within each of us a cosmos of consciousness unfolds temporarily, a subjective universe develops. The first part of the problem is to understand how a variety of subjective universes can constantly form and disappear in our objective universe ... . (Metzinger 1995, p. 6)

Widespread and apparently familiar as they are, these ideas of innerness ought surely to be regarded as puzzling. What does it mean to say that something exists within the mind, if this is to be contrasted with existence which is physical? And how does this relate to the idea of things, including the mind, being inside the physical body? In what follows I propose an approach to these questions, and try to show how it casts light on the problem of consciousness. 2. The double internality of the mental It is noteworthy that on this way of thinking mental phenomena are ascribed a double innerness. They are said to go on within the mind, and this in turn to be located within the body. To bring this out more fully we can illustrate it with a kind of cartoon, as follows:

20

The physical realm My body

The mental realm: my mind or soul

Now when we put the matter this graphically, we may wonder how far, despite speaking in this way, we actually do entertain such an image of ourselves. This is a serious question, and we often reject what Wittgenstein called the ‘pictures’ underlying our uses of words when we make them more explicit.1 We can gain a more vivid sense of this by considering the phenomena associated with a phantom limb. The bearer of a phantom arm, like a person with a real arm, feels himself to have an arm which is a locus of sensation, willing, etc. In the case of the phantom arm, however, the felt locus is actually empty, and can be seen to be so. The feelings connected with the phantom thus seem distributed in the physical space where the arm would be, but also not to be real occupants of this space, in the sense of things which actually have extension within it. Rather, and in accord with the conception we are considering, the feelings seem located in a subjective inner space – that of the bearer’s mind − which is encompassed by, but somehow distinct from, the physical space which surrounds it. Vilayanur Ramachandran has devised a way of enabling patients with phantom arms to have an experience as of seeing the phantom as well as feeling it. This is done using a box with holes in the sides and an internal mirror. The patient puts his intact arm in one of the holes and puts the phantom (or imagines putting it) into the other. The mirror is so placed as to allow the patient to see his intact arm, but also to reflect the image of that 1

This is the ‘picture’ of the mind which Wittgenstein discusses often in Philosophical Investigations, and caricatures as that of the ‘beetle in the box’. This is discussed further in my 2000a.

21

arm, so that it also appears (and in mirror-image right-to-left reversal) as in the place where the phantom has been ‘put’, and feels to be. When the patient looks into the box, he sees himself as having two intact arms, in just the places where he feels these arms to be. In effect his normal image of his missing arm – as having feelings going on within it, which are also encompassed by his physical body – is temporarily (and illusorily) regained. The phantom arm which he previously felt as from inside but could not see now appears as visible, and hence from outside. If asked to move both arms or hands in symmetrical ways, the patient may also have the experience both of moving the phantom and seeing the arm itself move accordingly. This, as Ramachandran notes, may enable the patient to manipulate the phantom as if it were real. He describes a patient who suffered discomfort from the apparent position of a phantom arm, which he had not been able to alter. Philip rotated his body, shifting his shoulder to ‘insert’ his lifeless phantom in the box. Then he put his right hand on the other side of the mirror and attempted to make synchronous movements. As he gazed into the mirror he gasped and then cried out, ‘Oh my God! Oh my God doctor! This is unbelievable. It’s mindboggling!’ He was jumping up and down like a kid. ‘My left arm is plugged in again… I can move my arm again. I can feel my elbow moving, my wrist moving. It’s all moving again.’ (Ramachandran 1998, p. 8)

In considering first the phantom limb and then this illusion of seeing it, we as it were leave and return to normal experience by rotating through two right angles. Doing this may remind us how apt it is to regard mental phenomena as both distinct from our bodies and yet also housed inside them. For this seems particularly clearly to be the nature of normal experience as it is recovered by the bearer of the phantom limb in Ramachandran’s box. His experience of himself from inside includes that of his arm, as a place of feeling and potential exercise of will. But although such a place is clearly there in his mind, it is ‘just in his mind’. For as the loss of his physical arm makes clear, this inner locus is not really part of the physical space in which extended things are to be found. Once placed in Ramachandran’s box, however, this psychological but non-physical place is re-clothed in flesh, and so again seems fully real. The space of the arm, as felt from inside, now seems integrated with a bodily arm which is subject to it’s owners will. The situation seems restored to that which we describe in our everyday language of the inner and the outer, and which is schematically depicted in the drawing above. So something like this image seems really to represent an aspect of our experience of mental phenomena,

22

even if we rarely make it fully explicit, and are surprised enough by its implications to be tempted to withdraw it when we do so. 3. The inner aspects of the mental as phenomenal, subjective, and private This example also indicates the ways we contrast the phenomenal characteristics of mental events with those of external physical things. Something physical and external to the mind, such as a tree, or again the flesh of an arm, can be externally perceived. Such things can be seen, touched, or otherwise made the object of the senses, and hence located in the encompassing physical space in which we find ourselves. Accordingly they have physical properties such as size, shape, and material composition, and admit further description in terms of the physical sciences, such as cellular and chemical structure, mass, weight, energy, and so on. So also such things are objective. They exist independently of our perceiving or thinking about them, and they have a real nature distinct from what is apparent in our perception or understanding. Again, and connectedly, such things are public, in the sense that they are capable of being perceived and investigated by more than one person. These features, moreover, seem intrinsic and interconnected. The world which we think of as external to the mind and perceivable by the senses is the physical world, and it is in the nature of what is physical to be objective and public in these or comparable ways. This, as it seems, stands in sharp contrast to the way we think of experiences such as the feeling of pain, or the visual experience of seeing a tree. Since these things exist within the mind, they cannot be seen by the eyes, touched, smelled, etc.; rather they are presented in introspection. Their being in the mind, moreover, is not (or not just) a matter of their being located within the physical space of the body, or in the brain or nervous system. Rather their innerness seems to be of an entirely different kind. We may be inclined to locate visual sensations, say, as just behind the eyes, and we may think of them as disposed in a spatial way, for example as constituting a visual field. But we do not think that if we physically examine the space behind the eyes we will see, or otherwise detect, this spatial field. As with the feelings of the phantom arm, visual experience and the visual field seem not to be in the same space as the retina, the optic nerve, etc. Something similar applies in the case of pains, which have a precise bodily location. We feel the pain of an aching tooth as in the tooth, but we also hold that no external examination of the space occupied by the tooth –

23

no x-ray, scanning process, or the like – will reveal the pain itself which we feel there. If I touch an aching tooth with the tip of my tongue – even if I manage to put my tongue exactly at the place where the ache is felt to be – the tip of the tongue and the pain still seem to be in different places, which as it were nearly overlap. The felt quality of the pain, unlike the tip of the tongue, seems to be in a purely internal locus, comparable to that we noted in connection with the phantom limb, and accessible only to the person who actually has the toothache. Thus, as we say, these things are in the mind; but unlike the nervous system or the brain they apparently cannot be seen or otherwise detected in the body or in the head. Like the sensations and exercise of will relating to the phantom limb, they seem in, and part of, the self, but not part of the physical body, even though, like the sensations in the non-existent phantom limb, they seem distributed in space. The inner aspects of sensations, experiences, and other conscious mental items are often called phenomenal qualities, or qualia. To many it seems that these phenomenal qualities are non-physical, in the sense that it is unintelligible or inexplicable that they should be the qualities of a physical thing. Again, these inner qualities seem subjective, in the sense (and by contrast to the physical things considered above) that they seem to depend for their existence or nature upon the person who apprehends them, and to reveal their whole nature in this subjective apprehension. When I feel a pain, it seems this pain could not exist if I were not feeling it – if I were to cease to feel the pain, it seems, the pain itself would cease to be, or cease to be painful. Likewise I seem to be acquainted with the whole inner nature of the pain. Unlike a physical object which I might see with my eyes, a pain has no inside which is concealed by its surface, nor any backside which is invisible to me. And since there is no more to the pain I am feeling than I now feel, I take the pain actually to be as I feel it to be.2 Also, and connectedly, the inner quality of a particular pain seems private, in the sense that it can be felt or apprehended only by the person 2

This notion of subjectivity was well expressed by Thomas Reid: When I am pained, I cannot say that the pain I feel is one thing, and that my feeling it is another thing. They are one and the same thing, and cannot be disjoined, even in imagination. Pain, when it is not felt, has no existence. It can be neither greater nor less in degree nor duration, nor anything else in kind than it is felt to be. It cannot exist by itself, nor in any subject but a sentient being. No quality of an inanimate being can have the least resemblance to it. What we have said of pain may be applied to any other sensation. The feeling and the thing felt are one and the same. (Reid 1785, 1.12.)

24

who has the pain. When I am aware of the inner nature of a pain, it seems that no one else can be aware of the inner nature of that particular pain; for, as noted above, my awareness seems to be of an internal locus which is available only to me (and the same holds for the sensations pertaining to a phantom limb). These features of experience again seem both intrinsic and interconnected; for it is the inner (introspectible) aspects of experience which we take to be phenomenal, and these phenomenal aspects which are subjective and private. 4. The inner and physicalism These considerations provide a more fully developed account of the intuitive distinction, which we apparently make from early childhood, as between things which are inside as opposed to outside the mind. The difference is marked by what we can regard as a connected and perhaps overlapping series of oppositions, including those between (i) Inner or internal to the mind as opposed to outer or external to it. (ii) Introspectible (internally perceivable) as opposed to visible, tangible, etc. (externally perceivable). (iii) Phenomenal as opposed to physical. (iv) Subjective as opposed to objective. (v) Private as opposed to public. We can illustrate these oppositions by redrawing and extending our initial diagram as follows:

25

External or Outer Externally Perceivable (Visible, Tangible, etc.) Internal or Inner Internally Perceivable (Introspectible)

Phenomenal Subjective Private

Physical Objective Public

The internality of the mental As this illustrates, the oppositions we are considering all seem rooted in that between the internal and the external. We conceive the mind as in the body, and mental phenomena, again, as occurring within the mind. Introspection is the internal analogue of perception, and phenomenal properties are those manifest in the already inner field of the mind. The subjectivity and privacy of these properties, therefore, seems consequent on their being internally perceivable by just one person, and from just one point of view. The objectivity and publicity of physical things, by contrast, seems secured by their existing in an external public space, and so being perceivable by more than one person and from more than one point of view. These oppositions are partly specified via negation, and they seem contradictory. It seems that no single thing could be both inner and outer, introspectible and externally visible (or tangible), phenomenal and physical, objective and subjective, or private and public. Hence, as Nagel, Chalmers, Metzinger and very many others have stressed, these internal aspects of mental phenomena seem particularly difficult to understand in physical terms, and give rise to the problem of consciousness. 5. The inner and the problem of consciousness 5.1 Two aspects of the problem: reconciliation and derivation Taking the problem of consciousness in this perspective, we can regard it as having two connected aspects. The first we may describe as a problem of reconciliation, and the second as one of derivation. To approach the problem it is important to distinguish them clearly.

26

Reconciliation: To reconcile things, in the sense involved here, is to show how they are compatible. So as regards consciousness, the problem of reconciliation is that of showing how the qualities we assign to conscious events (particularly those of being inner, phenomenal, subjective, and private) are compatible with those we assign to neural events (particularly those of being outer, physical, objective, and public). Nagel, for example, expresses this problem by saying that ‘We do not at present possess the conceptual equipment to understand how subjective and physical features could both be essential aspects of a single entity or process.’ (Nagel 2000, p. 444) As Nagel notes, this problem is not solved by maintaining that conscious events are in fact events in the brain, for the difficulty is to understand how this can be so − how the properties which we ascribe to conscious events are compatible with those which we ascribe to physical ones.3 Hence it is also clearly not solved by holding that neural events, while physical, also have aspects or properties which are inner, phenomenal, subjective, or private, for this again is a version of what requires to be explained. Derivation: The problem of derivation is related to that of reconciliation, but goes beyond it. This problem is raised by considering the kinds of explanation we attain in the physical sciences. In general, physical science enables us to understand how the manifest features of things can be derived from their physical nature, and so to exhibit how they are consequent on it. So if physicalism is true, it seems that this should hold for the manifest inner features of experience as well. If we describe the problem of opposition as that of seeing how certain physical events could possibly have the features of conscious events, or vice-versa, then we can describe the problem of derivation as that of seeing how the physical events in question ought certainly to have the mental features of conscious events. This is the problem of seeing how the inner, phenomenal, subjective, and private aspects of these events actually flow from, or are consequences of, their physical nature. 5.2 Derivation and explanation. We can spell this out more fully as follows. The natural sciences enable us to explain the manifest features of the physical world. We see the apparent 3

This of course was stressed by Nagel in his celebrated Nagel (1979).

27

movements of the stars across the night sky, and we have come to be able to explain these movements. We observe that ice and iron are solid, that water and glass are transparent, and we can explain why these things have these features. In each such explanation there is some form of derivation of the manifest and observable properties we explain, from other and more basic properties, which we find or hypothesize in the course of scientific investigation. In some cases we can cast these derivations in the form of rigorous logical or mathematical arguments which proceed from a few simple laws; and in others we are inclined to hold that such derivations are available in principle, even though they are too complex, or require too much information or calculation, for us actually to specify them. This process of derivation and explanation is at the core of our scientific view of the world. For we accept our scientific hypotheses, and so acknowledge the deeper and more basic properties in terms of which these explanations are cast, partly because this enables us to derive and so explain the observable properties we are familiar with. (Indeed our awareness of what we observe is now saturated with a sense of the deeper explanations which pertain to it.) So if conscious events are neural events, then it is natural to suppose that we should also be able to explain their manifest inner features in some such way. In accepting physicalism, therefore, we apparently incur the obligation of showing how the inner properties of experience can be derived from the physical properties of the neural events we take experiences to be. This is the problem which Nagel expresses by saying that ‘If physicalism is to be defended, the phenomenal features themselves must be given a physical account.’ (Nagel 1974, p. 478) Such an account would be a derivation, showing how the phenomenal properties of experience were to be seen as consequences of their physical nature. 5.3 The relation between reconciliation and derivation The problems of reconciliation and derivation are clearly connected. If it seems impossible from the outset to derive the manifest inner properties of experience from physical properties of the brain, this may be because the properties to be derived seem incompatible or incongruent with those from which we aim to derive them. Conversely, if we were to solve this problem of derivation, then this would perforce also constitute a solution to the problem of reconciliation. For clearly to derive B from A is thereby to show that A and B are fully compatible. Still the problems are partly distinct. In particular, it may be possible to show how two sets of properties are compatible with one another by

28

some means other than deriving one from the other. (A proof that A and B are consistent is different from a proof that A entails B.) So it might be possible to solve the problem of reconciliation independently of the problem of derivation, and so to treat the two separately. This is how we will approach them in what follows. From the outset, however, we should note that both these problems seem difficult, and that we have a strong intuition that the problem of derivation cannot possibly be solved. For when we contemplate the physical and chemical properties of brains (or neurons, or whatever) on the one hand, and the manifest inner felt property of pain on the other, we feel that there is no prospect of deriving the latter from the former. How could physics, chemistry, physiology, or the like, provide materials for deducing how this pain feels (internally, subjectively, privately) to me? We have no idea what such an explanatory derivation might be like, and we are inclined to feel confident in advance that none could be given.4 6. Mental concepts and physical phenomena Following Davidson many physicalists regard the mental as a conceptual rather than an ontological category. The idea is that we use a distinct and irreducible family of concepts – mental concepts – in thinking of mental phenomena as such, that is, as mental. Thus we use the concept of pain to think of events as pains, the concept of belief to think of states as beliefs, the concept of desire to think of states as desires, the concept of a self to think of things as selves, etc. These concepts are interwoven in their use, for when we think of a pain as a pain, we also think of it as something going on in ourselves, etc. The events, states, and other items to which we actually apply these concepts, however, are physical ones, centred in the brain and nervous system. (And generally, since mental and physical phenomena are identified via their causal roles, the physical phenomena to which mental concepts apply encompass the causal roles we assign to the mental.) Hence on this account mental concepts actually serve us as ways of thinking of physical phenomena. They are the mechanisms which evolution has given us for thinking about the causes within us of what we say and do – and hence for thinking about what happens in our own brains

4

Thus as Nagel also says, ‘it seems clear in advance that no amount of physical information about the spatiotemporal order will entail anything of a subjective, phenomenological character.’ (Nagel 2000, 439)

29

– in advance of any explicit knowledge about the insides of our bodies in general, or our nervous systems in particular.5 Putting the matter this way makes it clear that we can regard mental concepts both as irreducible to physical concepts, and also as physical concepts (that is, concepts for physical phenomena) of a particular kind. If we take physical concepts to be those which we use to think of things as physical (where we oppose this to thinking of things as mental) then mental concepts may be irreducible to physical concepts, in the sense that there are neither definitions nor strict natural laws relating their applications; and the same may hold for the types, properties, or states which we use these concepts to demarcate. A main lesson of philosophy, indeed, is that very many concepts – including everyday concepts like ‘clock’ or ‘cup’ – are indefinable or irreducible in this way. Reducibility is a form of strict redundancy, and this seems relatively uncommon among human concepts: like words, particular concepts have particular uses, and the use of one seems rarely to be just the same as another. So mental concepts can also rightly be regarded as physical concepts (and hence as determining physical types, properties, or states), in the sense that the concepts themselves are physical (neural) forms of representation, and these serve to represent things which are also physical. In this sense mental concepts and mental phenomena are as physical as rocks and stones and trees. They are part of the physical guiding mechanisms of physical living bodies. This is their biological nature as we know them and there is no more reason to hold that they might have an alternative non-biological or non-physical nature than there is to hold that there might be non-biological or non-physical trees. This holds even if the state- or property-types we impose in commonsense psychological thinking are distinct from those of everyday physicality or the physical sciences and irreducible to them. For as non-reductive physicalists stress, questions of conceptual reduction are distinct from questions of ontological nature. The irreducibility of the concept ‘cup’ to those of chemistry or physics does not tend to show that cups are not physical things. 7. Different kinds of irreducibility Mental concepts are, however, subject to more than one kind of irreducibility. To take one main kind, applicable to representations generally, consider photographs and the concepts/types we use in 5

I have discussed the representation of the mind as inner in relation to evolution in my 2000b.

30

describing them. We ordinarily describe photographs intentionally, that is, in terms of what they are about or of. These are the objects and properties represented in the photographs, and (where real objects are concerned) which played the appropriate causal role in their production. We also, however, speak of such photographs as in the camera, i.e. as now realized in the photographic apparatus itself, in the form of exposed film, or the registers of digital memory. So in addition to describing a photograph intentionally, we can also describe it intrinsically and physically, in terms of the chemistry of the exposed plate or film, the physical structure of the digital memory, and so on. These two forms of description – in terms of the things and situations represented, and in terms of the physical vehicles via which this is done – apply to representations of many kinds. Simple and straightforward as they are, they are characteristically irreducible, in the sense that they are neither interdefinable nor related by natural laws. For we can see that a given range of objects and properties – Aunt Maude swimming at a particular time and place at Brighton Beach for example – might photograph in many different ways (might cause and be registered in very many different sorts of chemical or electronic structure), and again that physically isomorphic photographs might nonetheless be of or about different things (e.g. of a wax model of Aunt Maude at Brighton Beach, etc.) So if someone told us that he had discovered a natural law linking photographs of Aunt Maude at Brighton Beach with a particular chemical or electronic structure, we would know in advance both that the claim was wrong and how to set about producing an actual falsification of it. So we should be prepared to acknowledge that there are no strict nomic generalizations – e.g. no photographico-physical laws – linking intentional and physical types. This kind of irreducibility can be said to hold over representational concepts, types, or properties; but it clearly shows no important gap in our understanding of such representations, nor any serious limitation on our scientific aspirations with regard to them. We understand what the relevant representations are and how they work, and in context we can determine what we want to know about a given representation in a way which is limited mainly by the material complexity of the case. Likewise such irreducibility gives rise to no problem of representational causality. We understand photographs as mechanisms established by the relevant physical processes, and relate their representational and causal roles (e.g. in driving a printer) as fully as we need be. This is enough for us to regard our practice of describing photographs or other comparable representations

31

(e.g. those in computers) as admitting an entirely unproblematic integration with science. In these cases we may lack (what philosophers have classically regarded as) reducing descriptions, but we can nonetheless frame realizing descriptions in a way which renders the lack of reduction scientifically irrelevant. This is also a main aspect of the irreducibility which Davidson noted as characteristic of mental concepts such as those of concept, intention, belief and desire.6 For concepts, beliefs, and desires are to be understood as involving forms of physical (neural) representation, which we describe in terms of the objects and situations they represent, and which can also be described in terms of their realization in our brains and nervous systems. As with photographs, we can know in advance that no strict laws will relate our commonsense intentional descriptions of these representations with the more intrinsic descriptions which neuroscience may ultimately enable us to frame. And as also with photographs, this kind of irreducibility in itself presents no barrier to the integration of commonsense psychology with neuroscience. But the innerness of the mental relates to an irreducibility of a deeper kind, as we will consider shortly. 8. Conceiving physical innerness as mental We began with our puzzling commonsense notion of the double innerness of the mental. So let us now note that we can apply the standard physicalistic approach above to this as well. In thinking of the ways mental items are in us, we are applying the commonsense mental concept of innerness which we have used from childhood, and which, as Wellman’s investigations suggest, may develop naturally in our species. Accordingly we should suppose that this concept has the same general function as our other mental concepts, namely that it serves for thinking about the causal mechanisms within us which regulate our behavior. There is an obvious suggestion as to how it does so. The mental concept of innerness enables us to think about the innerness of this system of causes, which in fact is a kind of physical innerness. So in using our mental concept of innerness, we are thinking of physical innerness, but thinking of it as mental rather than as physical. Just as we can think of internal neural mechanisms which prompt and guide our behavior as 6

I have simplified here by omitting Davidson’s early stress on rationality as a source of irreducibility in favour of his later emphasis on intentionality. I have discussed this in relation to Davidson in my 1995.

32

desires, beliefs, or feelings, so we can think of the internality of these mechanisms by thinking of them as in the mind. We are inclined to think of mental events as going on (somehow non-extendedly or non-physically) in the mind, and of the mind or self as (somehow non-extendedly or non-physically) inside the body. In fact, as physicalists acknowledge, mental events go on (extendedly and physically) in the nervous system, and the nervous system is (extendedly and physically) inside the body. So the dual but non-physical innerness which we ascribe to mental events – in the mind, in the body – parallels a dual physical innerness – in the nervous system, in the body – which mental events actually enjoy. This is how it would be if our conception of events as (somehow non-extendedly) in the mind served as a way of thinking of their occurrence in the nervous system, and our conception of the mind as (somehow non-extendedly) in the body served as a way of thinking of the innerness of the nervous system itself. (And the nervous system really is – as Descartes said of his mind or self – ‘not in my body as a pilot in his ship' but rather ‘very closely joined and, as it were, intermingled with [this body]…’ (Med. VI, p. 56))7 This is an hypothesis about the work done by a particular family of concepts. As our concepts of the mental enable us to think about (physical and neural) inner causes of our behavior, but in commonsense mental terms, so our concepts of the innerness of the mental enable us to think about the (physical and neural) innerness of these causes, and in commonsense mental terms as well. 9. Ratifying and reconciling the oppositions Once we consider this hypothesis, we can see that it provides a good account of the oppositions which seem at the core of the problem of consciousness. On the one hand it ratifies these oppositions, by showing that they are genuine, and relate to real differences in nature. On the other it reconciles them, by showing that the physical differences to which they relate are ones which we are conceiving in mental terms, and which are ultimately consistent with one another, as can be seen when we reflect on their physical reality. This again fits with the overall idea that mental concepts are ways of thinking about inner causes of behaviour.

7

Compare Nagel’s view that ‘the core of the self – what is essential to my existence – is my functioning brain.’ Nagel (1986, p. 40)

33

9.1 Inner and outer as related to the nervous system We can see this clearly in the case of the most basic opposition, that between the inner and the outer. According to our hypothesis this opposition marks a real distinction, between what goes on within the nervous system and what goes on outside it. But if this is so then it is clear that thinking of events as in the mind is actually consistent with also thinking of them as in the extra-mental physical world. The first problem of opposition was to see how, if mental events were physical, the same events could be both inner and outer, or both internal and external to the mind. This understanding of the opposition resolves the problem. There is a real opposition between the internal and the external, for events cannot in the relevant sense be both inside and outside the nervous system. But if mental innerness is understood as a way of conceiving occurrence within the nervous system, then mental events can rightly be regarded as both internal and external to the mind. A pain is internal to the mind (internal to the nervous system) of the person who has it, and external to all other minds, and part of the external world in general. 9.2 Introspection as internal awareness of the inner Consider next the opposition between what is (inner and hence) introspectible, and what is (outer and hence) externally perceivable. In this we can contrast the roles of what, following Kant, we can regard as outer and inner sense. Antonio Damasio introduces a neurological distinction related to this by asking the reader to look away from the page to the room in front and then back again. As you looked up many neural stations of your visual system, from the retinas to the cerebral cortex, shifted rapidly from making neural mappings of the page, to mapping the room in front of you. But when you returned to the page, these components resumed mapping the page again … resulting in different visual images… . However, several regions in your ‘body sensing’ brain, which has the job of mapping varied aspects of your body, did not change at all in terms of the kind of object they represent … . The moral of this story is that some parts of the brain are free to roam all over the world and to map whatever sound, shape, smell or texture that the organism’s design enables them to map. But some other brain parts – those that represent the organism’s own structure and internal state – are not free to roam at all. It is reasonable to hypothesize that this is the source of the sense of continuous being that anchors the mental self. (Damasio, 2003a, p. 227)8

8

For further discussion by Damasio see his 2003b, pp. 105ff.

34

The parts of the brain which ‘are free to roam all over the world and map whatever shape, smell or texture’ are those related to the systems of outer sense, including hearing, sight, smell, and touch. Those which map ‘the organism’s own structure and internal state’ can rightly be characterized in terms of inner sense. A. D. Craig has recently described them as constituting an interoceptive system, which ‘generates a direct thalamocortical representation of the state of the body…crucial for temperature, pain, itch and other somatic feelings.’ This, as he says, ‘constitutes a basis for the subjective evaluation of one’s condition, that is, “how you feel”.’ (Craig 2002, pp. 665-6) Kant’s notion of inner sense, like our commonsense conception of introspection, evidently encompasses the operations of this system, as well as the less bodily experiences, also conceived as inner, which result from outer sense. Awareness of things, whether internal or external, involves applying concepts to them, that is, thinking of them in certain ways. Thus in external perception we are aware of trees as trees, of computers as computers, etc.; and in introspection we are aware of pains as pains, of visual experiences as visual experiences, etc. again. From a physiological point of view, however, the acts of thinking or conceiving in the two cases are strikingly different. In the case of introspection or inner sense both the mental activity of conceiving the objects of awareness and the activities and structures conceived are internal to the same nervous system. Thus when we think of pains as pains from inside, the activity of conceiving, which seems centred in the cerebral cortex, is directed towards events in the same cortex, i.e. those which originate in the stimulation of nocioceptors. Such thinking is thus not mediated by the working of the external sensory systems, or the concepts applied in their use for outer things. (Likewise in conceiving visual experience as visual experience, the cortical activity of conceiving is directed to events in the visual system of the same cortex.)9 In external perception, by contrast, the conceiving of the things we are aware of is normally directed to objects which are presented to the outer senses. Such thinking therefore is mediated by the concepts we employ in using these senses. The objects of awareness in this case are presented to us as parts of the spatially extended world external to the nervous system (and hence to the mind) at work in conceiving them. Thus the distinction between introspection and external perception, like that between the inner and the 9

Visual experience, which is not part of interoception as discussed by Craig, requires more detailed treatment than space permits here. This is partly indicated in the discussion of visual experience in my 2000a and b.

35

outer, reflects both the distinctive neural systems involved, and the boundaries of the nervous system itself. In this perspective what we call introspection is a basic form of selfawareness, in which the neural structures involved in thinking are applied relatively directly to activities and structures in the very same nervous system. (So in this case ‘directly’ means without mediation by the activities of perceiving and conceiving involved in the working of the outer senses, which are normally not focussed on the nervous system itself.) It is in this way that our most fundamental sense of self – that of Descartes’ cogito, with its apparent non-extendedness – is generated. We noted above that the same event (a pain, say) can be both internal and external to the mind (internal to one nervous system and external to another). Now we can add that the same event (the same pain) can be conceived both internally (from inside) and externally (from outside). Such an event is normally internally conceived by the person who has it, and externally conceived – and so thought of as located in space external to the self – by others. But a person might feel pain while observing the very activity that he was thus internally conceiving in a scanning device, and so at the same time perceiving and conceiving it from outside, and as part of the world external to the self. In approaching the same events via these distinct perceptual and conceptual routes he would of course conceive them very differently; and this is a point on which our discussion will turn.10 10

It may be worth making some of the distinctions latent in this more explicit. We can do this by sketching an interpretation which is illustrative if overliteral. Let us use proper names such as ‘Tom’ to designate persons, capitalized words such as ‘PAIN’ to indicate the application of the concepts accociated with them, ‘P’ to designate the neural activity involved in simply having pain, ‘IC’ to designate conception from the inside, and ‘OC’ to designate conception from outside. Thus if we take Tom’s feeling pain, and take it that he internally conceives this pain as pain, we can designate the situation as: Tom(IC(P)PAIN) This in turn corresponds to a particular complex of realizing neural activity, namely that of P together with that involved in the IC of P as PAIN. We might, however, wish to stresss that Tom himself is internally aware of himself as feeling pain, and this would give: Tom(IC(Tom, P) I FEEL PAIN)

36

9.3 Phenomenal as physical but conceived introspectively Now let us take the difference between the phenomenal qualities of experience and the physical qualities of things external to the mind. When we apply concepts to things we thereby represent them as having properties or qualities. The items to which we apply the same concept are perforce alike, and in thinking of them as alike we thereby think of them as having the same property. In perceiving (and so conceiving) something we regard as physical, our application of concepts is characteristically mediated by our outer senses, and in particular by sight and touch. These senses are adapted to mapping things in the space exterior to the nervous system, and in using them and the concepts informed by them we represent the affordances of objects in the environment for our navigation and manipulation. In this we see and think of things as having colour; three-dimensional shape (and hence surface texture and interior volume and countless further internal properties ready to hand); location in relation to the perceiver’s body; and so on. In perceiving and thinking of things from outside, that is, we think This would explicitly designate a more inclusive pattern of neural activity, involving the IC of the self as I as well, arguably tacit in the previous formulation. By contrast when Fred ascribes pain to Tom (and taking Tom to be aware of his own feeling pain) the situation would be

Fred(OC(Tom(IC(Tom P)I FEEL PAIN)) TOM FEELS PAIN) This would correspond to a very different pattern of neural activity, involving an OC of Tom’s P and, plausibly, of his IC of himself as well. In these terms if Tom were both feeling pain and observing a scan of the neural activity involved in this he would be engaged in two sets of conceptualizing activity. The first would be the original Tom(IC(Tom, P) I FEEL PAIN) The second might be represented as

Tom(OC(Tom, P) I FEEL PAIN) THAT IS THE NEURAL EVENT OF MY FEELING PAIN The differences between these two activities of conceiving, according to the argument in the text, owe their deepest differences to the fact that no externally applied concept can have the causal and representational role, and so determine the same perspective on the same event, as one which is applied internally can have.

37

of them, and in a particularly full way, as objects extended and arranged in the physical space about us. (As Kant says, the form of outer sense is space.) All this contrasts with the introspective representation of items internal to the body and nervous system which manifest phenomenal properties. These are hidden beneath the skin, and their role in our lives is not that of things external to the body which it is relevant for us to see or touch. So we (or our brains) do not irrelevantly represent these inner events as visible or tangible, nor do we show their affordances for manipulation or navigation as parts of an extended spatial array. Instead, as in the case of pain, we represent them as manifesting properties which have a unitary and direct influence on behaviour, and which are therefore radically distinct from those which are visible, tactile, or the like. So although neural events of pain do occur in physical space, our interoceptive representations do not show them as spatial (or fully spatial): as having insides or outsides or surfaces or volume or as in a space populated by other extended things. Rather we tend to represent only part of the neural activity connected with the location of some injury, and that location mainly simply as hurting. This may explain why philosophers such as Nagel and McGinn so emphasize the problem of relating consciousness to space. Nagel urges that mental concepts ‘don’t give us the comfortable initial handle on the occupants of the familiar spatio-temporal world that prescientific physical substance concepts do’,11 while McGinn stresses the necessity for ‘ “containing” the non-spatial (as we now conceive it) phenomenon of consciousness.’12 The idea that phenomena such as pain are not spatial, however, would involve a non-sequitur: a move from the fact that pains (or conscious experiences more generally) are not conceived introspectively as having the spatial features linked with touch and sight, to the idea that they lack such features, and are therefore not extended, and so nonphysical. Since conscious events are neural events, consciousness is certainly a spatial phenomenon; and this is consistent with the fact that in introspection we do not represent things as having a full complement of spatial properties. All this – the distinct neural and conceptual systems involved, their different objects, and the different ways they work – entails that the internal properties we ascribe to experiences such as pain in our own case must perforce be radically different from the externally perceptible and spatially integrated properties of things we perceive as part 11 12

Nagel (1998, p. 339). McGinn (1995, p. 159).

38

of the external world. This deep sense of difference is partly registered in our regarding this property as phenomenal as opposed to physical. But of course the neural activity we represent as phenomenal is just as physical as anything else. What we regard as phenomenal is just an aspect of the physical, as we conceive it introspectively. (In this perspective, therefore, we can also regard dualism as mistaking a partial conceptualization of something which is physical for a full conceptualization of something which is non-physical.) 9.4 Subjective as objective but introspectively conceived Again, consider the difference between the subjective and the objective. When we feel a pain, as we have noted, the pain itself does not seem distinct from our feeling of it, and its whole nature seems disclosed in this feeling. This certainly contrasts with perceiving something external, whose existence is independent of us and our perceiving, and whose nature is only very partially revealed to our senses. But again the difference is best understood physically, and in terms of that between introspection and external perception. In the external case we naturally make a sharp distinction between our mental activity in conceiving and the thing itself which we conceive. As noted we see this thing as external to us, and as having spatial wholeness, form, and relations, and so a host of other properties bearing on what we can do with the thing or in relation to it. In the case of pain, by contrast, the internal activity of conceiving and the internal event conceived form parts of a single extended neural event. We think of this, moreover, mainly in terms of its direct effect on us, as opposed to what this effect tells us we might do via manipulating something else. Hence we think of it as a spatially uncomplex but behaviourally directive impingement of painfulness. So the objective and richly propertied neural occurrences we conceive as pain appear in our conception as simple and subjective: their existence seems to us scarcely distinguishable from our feeling of them, and we represent them as having no other nature than we feel them to have. This explains why pains and other experiences should seem subjective (or why they should seem to have, in John Searle’s phrase, a ‘first-person ontology’). As before, this may explain our tendency to think of pains, etc., as subjective as opposed to objective, and hence as not physical. But again the conclusion does not follow. Rather the first-person subjectivity of introspected events is a consequence of their objective neural nature;

39

and our sense of fully grasping their internal aspect is consistent with their having further qualities, of which in introspection we are simply unaware. 9.5 Private as public but introspectively conceived Finally, take the distinction between the private and the public. When I feel pain, it seems impossible that anyone else should feel this pain, or should be aware of how this pain feels in the same way as I am. So the pain seems private, and this seems inconsistent with its being physical and public. This, however, is another reflection of the facts we have already considered. A person who feels pain applies the concept of pain internally – to activity in her own nervous system – and thereby represents the instantiation of what we regard as a phenomenal property. It is in the physical nature of the case that no one else can likewise introspectively apply that concept, and so as to yield that instance of that property. Such first-person conceiving can be done by the subject alone. But this again derives from the neural nature of events like pain, and so is not inconsistent with, but rather explicated by, their physicality. The private is public, but introspectively conceived.13 10. Phenomenology as the presentation of causal role within the self The roles of internal and external perception are linked with a further topic, namely the way phenomenology manifests the working of causality within the self. One of the most striking features of mental concepts is the way they simultaneously face in two directions: inwards towards the self, and outwards towards the other, and hence also towards the world apart from the self. This again reflects the difference between applying such a concept to activity in our own nervous system, and applying it to activity 13

It is perhaps worth noting that this does not make the inner aspect of pain – or the inner application of the concept of pain to one’s own neural activity – private in the sense opposed by Wittgenstein’s arguments about private language. These were directed against a language which was private in the sense that it could not be interpreted by another, or again against the idea of a word denoting something that only one person could apply a concept to. But on the present account events of pain or feeling pain are public physical events, to which more than one person can apply concepts. Insofar as the inner conceptual activity of conceiving and the inner activity of pain itself are distinguishable, then each admits of investigation, as does their combination. So the inner aspect of pain manifest solely in the first-person perspective also admits of public consideration, if not of external conceptualization which would reduplicate the internal. The relation of the present account to Wittgenstein’s is briefly discussed in my 2000a cited above.

40

in the nervous systems of others, via the way this activity is related to observable behavior and circumstances. We regard the mental events which go on in us as having both phenomenology and causal role. In thinking of a sensation as pain, we think of that sensation both as having a certain internal character, and as thereby affecting the person who has it in ways related to that internal character itself. We take the person who has the pain to conceive the internal character of pain and to react to it from inside, whereas we others conceive it (and otherwise react to it, if we do) from outside. In this we make use of a single concept – the concept pain – which can be applied both from within and without. The concept which we use in these ways, moreover, is in both uses both a phenomenological and a causal concept. For the phenomenology of mental events – the way they feel to us, the way they are presented in consciousness – has direct causal import. In feeling pain, as noted, we feel our selves impinged upon, and in a way particular to the pain we feel; and in this we feel ourselves as immediately engaged in a causal process involving the self. So in feeling pain on touching something we thereby feel ourselves as moved to withdraw from the contact, in regarding something as a cause of pain we thereby feel ourselves as moved to avoid it, etc. (The matter is likewise, except opposite, when we feel pleasure in contact with something – in this case, by contrast, we feel ourselves moved to continue the contact or to renew it.14) The same holds, in a variety of ways, for all the varieties of experience. To experience pain or any other sensation is to respond to the operation of an internal cause. Thus phenomenology should not be regarded as distinct from causality; rather it is intrinsically causal. Phenomenology is the presentation of causes as internal to, and hence as part of, the self – the inner face of causality as it manifests itself from within the self. In encountering such causes from within we thereby gain further purchase on them, and so on our selves. Our conceptual awareness of these inner causes gives us the opportunity to work with them – and so to bring other internal causes to bear on them – in feeling, thought, and will.

14

We can put this point by saying that phenomenology is what Millikan calls ‘pushmipullyou’ representation: representation whose biological function is to have some effect or other, in this case an effect on the self.

41

11. The problem of derivation and the world apart from the self This bears on the second aspect of the problem of consciousness, namely the problem of derivation. This is the problem described by Leibniz in the 17th Century as the impossibility of using features of a mechanism ‘to explain a perception’, and by a long series of philosophers and scientists since. The kind of explanation apparently required is a derivation of the phenomenal features of experience from physical or physiological features of the nervous system: for example, a derivation of the phenomenal property of pain from the chemical or physical features of neural firings. And as we have already stressed, we have a strong intuition that no such derivation is possible. As Du Bois-Reymond asked in the 19th Century: What conceivable connection is there between certain movements of certain atoms in my brain on one side, and on the other the original, indefinable, undeniable facts: ‘I feel pain, feel lust; I taste sweetness; smell the scent of roses, hear the sound of an organ, see redness’? (Du Bois-Reymond, p. 45)

We can now see that this lack of ‘conceivable connection’ is a natural consequence of the way we distinguish the inner and the outer, and hence draw the boundary between the self and the world apart from it. For as we noted at the outset, we apply our notion of mental innerness at first level to the mind or soul, and hence to the self itself. In thinking of ourselves as selves, we regard ourselves as subjects who experience the world and agents who seek to change it, and we regard these phenomena of experience, thought, and will as going on in our selves and bodies. We saw that the subjectivity and privacy we ascribe to inner events – the way their existence and essential nature seems barely distinguishable from our apprehension of them – reflects the way the act of conceiving and the event conceived are so integrated as to constitute an extended event in a single nervous system. This same causal-conceptual integration evidently also integrates the range of neural events of which we are aware in such a way as to present them as aspects or parts of ourselves. So in thinking of events as inner we thereby also construe them as our own, as we do a pain, a visual or auditory experience, or a feeling of fear or pleasure. In thinking of things as outer and physical, by contrast, we think of them as parts or aspects of the world which is distinct from the self, and in this we set them apart from the self as internally demarcated. As we saw in considering the phantom limb, the boundary between self and not-self seems to correspond roughly with the skin. But since we draw this boundary from inside, by our inner use of mental concepts, it can seem

42

intact even in the absence of the body; and the innards of the body themselves can seem external by comparison with it. Hence when we think of our experiences from inside they seem part of ourselves, but when we think of them from outside, as events in our physical nervous systems, they again seem distinct and apart from us. It is because our distinction between the inner and the outer is also our distinction between the self and the world apart from the self that we feel in advance that the demand for a physical (outer) explanation of the inner qualities of experience cannot be met. The demand actually contains a conceptual or a priori absurdity, which we experience as a dissonance in imagining it fulfilled. For it is in effect a demand that we should somehow replicate the direct and spatially reduced inner mode of cognition by which we conceive aspects of the working of our own nervous systems as parts of our selves, by the external and more fully spatialized mode by which we conceive things as parts of the world distinct from the self. Asked, in effect, to envisage some future thinking of the inner self via the conceptual scheme we use in thinking of the outer non-self, we naturally draw a blank. For since these two ways of thinking actually have contradictory properties – and in particular since one is causally direct and immediate and the other is not – this is a demand which could not possibly be met. To see this more clearly it may be useful to consider the question in another form. C. D. Broad observed that an angel who knew the whole of mechanical physics would not be able to deduce from it how ammonia smelled, and Frank Jackson has likewise stressed that an omniscient but sensorily deprived neuroscientist would not be able to deduce how pain feels.15 If Broad’s angel, or Jackson’s Mary, were in possession of a physical explanation of the phenomenal properties of experience, they could use this to derive the knowledge they lacked, and also, if they wished, to investigate further topics in philosophical phenomenology, such as what it would be like to be a bat. Our conviction that they could not do this is the same as our conviction that there is no such explanation as might make it possible. But what tells us there can be no such derivation is not our knowledge of science and its limits. Rather it is our apprehension that when we feel pain we are in a kind of basic and direct causal-conceptual contact with an aspect of ourselves – a kind of contact realized, as we see, only among parts of a single nervous system; and we rightly feel that this self-involving causal-conceptual directness would be lost in thinking of any different or more scientific kind. 15

See Broad (1947, p. 71) and Jackson (1986).

43

12. Conclusion: reconciliation without reduction or derivation Examining the innerness of the mental thus enables us reframe the problem of consciousness in two ways. At the outset we distinguished two problems. The first was that of reconciling the inner and apparently nonphysical features of mental events with their physical ones, as represented by Nagel’s claim that ‘We do not at present possess the conceptual equipment to understand how subjective and physical features could both be essential aspects of a single entity or process.’ The second was that of deriving the inner and apparently non-physical features from physical ones, as represented by Nagel’s claim, carried forward from Leibniz and many others, that ‘If physicalism is to be defended, the phenomenal features themselves must be given a physical account.’ If our argument so far has been correct, we should reject both these claims, and reconsider what is problematic about consciousness accordingly. As regards the first, we have seen that if we take our remarkable but commonsense notion of the double innerness of the mental as a way of representing the physical innerness which mental events actually enjoy, then the opposition between the subjective and physical features of conscious events can be both ratified and resolved. The apparently anti-physical subjectivity of these events derives from the direct and spatially reduced way in which we think them as mental, and we accordingly take their inner properties as paradigmatic of their mental nature and essential to it. But since the events and properties we conceive in this way are actually neural and physical, we can rightly take this as essential to their nature as well. As regards the second claim, we see that in conceiving events as inner we also conceive them as parts or aspects of our selves, as opposed to the world apart from the self. This internal relation to the self is constituted by the same causal and conceptual directness and integration – the same confinement to a single nervous system – as underpins their subjectivity. This mode of conception is unique and part of the particular causal flow which we conceive in using it. The notion that we should somehow replicate it via another – by physical explanation or any form of reduction to the outer – is not a scientific ideal to which we should aspire (or which we might suppose others to attain) but a conceptual incoherence which we should recognize as such. This is not to say that consciousness and its relation to the physical presents no problems. But it suggests the importance of framing these problems in a way that enables us to recognize more clearly what they are.

44

References Augustine: Confessions. New York: Sheed and Ward, 1942. Broad, Charles D. 1947: The Mind and its Place in Nature. London: Kegan Paul, Trunch, Trubner & Co. Chalmers, David 1995: ‘The Puzzle of Conscious Experience’. In: ‘The Hidden Mind’, Scientific American, 273, pp. 80-86. Craig, A. D.: ‘How do you feel? Interoception: the sense of the physiological condition of the body’. Nature Reviews Neuroscience, 3, pp. 655-666. Damasio, Anthony R. 2003a: ‘The person within’. Nature, 423, 15, p. 227. — 2003b: Looking for Spinoza: Joy, Sorrow, and the Feeling Brain. New York: Harcourt. Descartes, Meditations on First Philosophy: With Selections from the Objections and Replies, ed. by Cottingham, John. Cambridge: Cambridge University Press 1996. Du Bois-Reymond, Emile: ‘About the Limits of Natural Knowledge’ in his Vorträge über Philosophie und Gesellschaft, Hamburg: Meiner 1974. Quoted and translated by Bieri, Peter in ‘Why is consciousness puzzling’. In: Metzinger, Thomas (ed.), Conscious Experience, Schöningh/Imprint Academic, pp. 45-60. Hopkins, Jim 1995: ‘Irrationality, Interpretation and Division’. In: MacDonald, Cynthia and MacDonald, Graham (eds.): Philosophy of Psychology: Debates on Psychological Explanation. Oxford: Blackwell Publishers. pp. 461-484. — 2000a: ‘Psychoanalysis, Metaphor, and the Concept of Mind’. In: Levine, M. (ed.), The Analytic Freud. London: Routledge, pp. 11-35. — 2000b: ‘Evolution, Consciousness, and the Internality of Mind’. In: Carruthers, Peter and Chamberlin, Andrew (eds.): Evolution and the Human Mind. Cambridge: Cambridge University Press, pp. 276-299. Jackson, Frank 1986: ‘What Mary Didn't Know’. The Journal of Philosophy, 83, pp. 291-95 McGinn, Colin 1995: ‘Consciousness and Space’. In: Metzinger, Thomas (ed.): Conscious Experience. Schöningh/Imprint Academic, pp. 149-163. Metzinger, Thomas 1995: ‘The Problem of Consciousness’. In: Metzinger, Thomas (ed.), Conscious Experience. Schöningh/Imprint Academic, pp. 3-37. Nagel, Thomas 1974: ‘What is it like to be a bat?’ reprinted in his Mortal Questions, Cambridge: Cambridge University Press, 1979, pp. 165-180. — 1986: The View from Nowhere. New York and Oxford: Oxford University Press. — 1998: ‘Conceiving the Impossible and the Mind-Body Problem’. Philosophy, 73, pp. 337-352. — 2000: ‘The Psychophysical Nexus’. In Boghossian, Paul, and Peacocke, Christopher (eds.): New Essays on the A priori. Oxford: Oxford University Press, pp. 433-478. Ramachandran, Vilayanur S. and Blakeslee, Sandra 1998: Phantoms in the Brain: probing the Mysteries of the Human Mind. New York: William Morrow Publishers.

45

Reid, T. 1785: Essays on the Intellectual Powers of Man, ed. by Brookes, Derek R. University Park, Pennsylvania State University Press 2002. Wellman, Henry 1993: ‘Early Understanding of the Mind: the Normal Case’. In: Baron-Cohen, Simon, Tager-Flusberger, Helen, and Cohen, Donald J. (eds.) Understanding Other Minds, Perspectives from Autism. Oxford: Oxford University Press, pp. 10-39.

46

Seeing an individual∗ Keith Hossack Alan Lacey has been a help and an inspiration to me throughout my career as a philosopher at King’s College London. When I was first appointed at King’s he very kindly invited me to join the reading group he runs, and made his nervous new colleague very welcome. I have faithfully attended his reading group ever since, never without benefiting from his wisdom and insights. He is remarkable in that his philosophical interests span the whole of the subject, so that he is at much at home when expounding the intricacies of Aristotle’s Metaphysics as he is when vigorously contesting the metaphysics of Kripke. Alan has a particular expertise in the philosophy of perception, so it is with much trepidation, but with great affection, that I have written what follows for this volume for his eightieth birthday. 1. Seeing Socrates What is it to see Socrates? An initial suggestion is that one sees Socrates if Socrates is part of the scene one sees. But for various reasons this answer will not do. One does not see every part of the scene one sees. One might see a large crowd, and Socrates might be part of the crowd; but it need not follow that one sees Socrates. Or on a dark night one might see a shadowy scene, and mistake Socrates for a shadow; Socrates is part of the scene one sees, but one does not see Socrates. Other examples: Socrates is in camouflage, so does not visually ‘pop out’ of the scene; or Socrates stands before one in plain view, but one does not see him, for one does not ‘register’ his presence, because one is absorbed in other thoughts. According to the sense-datum theory, we see the world in virtue of our visual sense-data. If Ayer’s (1964) version of phenomenalism is true, then to see an object is simply to experience certain sense-data. We can refute this by noting that there is a possible world where one has the same sense-data but does not see Socrates. For there is a possible world where there is a visual duplicate of Socrates; at this world one sees the duplicate, ∗ I gratefully acknowledge the help given me by members of my research seminar on this topic in the summer term of 2007.

not Socrates; yet one has exactly the same sense-data. Thus seeing Socrates is not simply a matter of experiencing certain sense-data. According to Russell (1912), one sees the world by inferring its existence from one’s sense-data. Is seeing Socrates a matter of drawing the right inferences from one’s sense data? Here we need the distinction between general and singular thought, i.e., the distinction between one’s thinking that something is thus and so, and there being something which one thinks is thus and so. In the former case one thinks the general thought that something is thus and so; in the latter case there is something concerning which one thinks the singular thought that it is thus and so. One may infer from one’s sense-data the general proposition that some socratic external object exists; and the socratic object that exists may in fact be Socrates; but one may nevertheless not be in a position to infer from one’s sense-data the singular proposition that Socrates exists. One would only be in a position to make this inference if one’s premisses ruled out the possibility of someone other than Socrates being the unique socratic external object. But the mere occurrence of one’s sense-data does not rule out this possibility. For there is a possible world where a visual duplicate of Socrates, and not Socrates himself, is the uniquely socratic external object. Nothing in one’s sense-data or one’s inferences from them yet excludes the possibility that this other world is actual. Therefore we do not have an account of how one can infer the existence of Socrates himself from one’s sense-data: there is more to seeing Socrates than the inferring of a merely general thought. According to the Causal Theory of Grice (1988), to see Socrates is to have a visual experience that represents Socrates, and which is caused in the right way by Socrates himself. The Causal Theory is an improvement on Russell’s theory because it is externalist, so it allows us to distinguish between seeing Socrates, and seeing someone who looks just like Socrates. Of course the visual experience would be the same if one were seeing a duplicate of Socrates, for all that we can ask of the content of the experience is that it matches Socrates (and of course also his duplicates). But the external causal connection to Socrates makes the experience one of seeing Socrates himself. The causal connection ties the experience uniquely to Socrates, but not just any causal connection will serve the purposes of the Causal Theory. For example, if Socrates is standing on a button that causes one to have a visual hallucination as of Socrates, one does not see Socrates, even though he is the cause of one’s visual experience. Thus we must refine the definition of causal connection, and this is not easy to do. Lewis (1986)

48

suggested it is a question of there being a good counterfactual match between the scene before one and the visual experience one has: over a reasonable range, if the scene were different, there would be a matching difference in one’s experience. But one might see the scene before one even if there was poor counterfactual match. Suppose one is seeing Socrates, but that if Socrates had not been here, a device would have been activated that would have caused one to have a hallucination as of Socrates. There is poor counterfactual match between one’s experience and the scene before one, for if Socrates had been absent, one’s experience would have been no different. Nevertheless, one does see Socrates. Thus seeing Socrates does not appear to be amenable to a purely causal analysis, since we must say what it is that makes a causal connection the right sort for successful seeing. Since it is not known how to provide such an analysis, the Causal Theory fails to tell us what it is to see Socrates. According to Snowdon (1981), to directly perceive Socrates is to be in a position, in virtue of one’s perceptual experience, to judge truly ‘That is Socrates.’ But not every creature that can see Socrates is in a position to make a true judgement about Socrates. For a simple animal might be able to see, without being able to judge or to speak, so seeing does not entail being in a position to judge. Snowdon replies to this that direct perception is the relation in which a subject stands to an object which guarantees that, if the subject is capable of demonstrative thought at all, then the subject is capable of a true demonstrative thought about the object. The type of relation we seek is one which when allied to a general capacity for demonstrative thought makes possible true demonstrative judgements. Evidently seeing is that type of relation, but what type is that? The obvious suggestion is that the type of relation in question is acquaintance. According to Russell’s ‘Principle of Acquaintance’: Every proposition which we can understand must be composed wholly of constituents with which we are acquainted. (Russell 1912, p. 58)

This entails that one is in a position to think a singular thought about an individual only if one is acquainted with that individual. Since a demonstrative judgement about Socrates is a singular thought, one would be able to think it only if one were at the time visually acquainted with Socrates, i.e., if one were seeing Socrates. Snowdon rightly says that seeing puts human thinkers in a position to make demonstrative reference; so by Russell’s Principle, we might surmise that seeing Socrates is (visual) acquaintance with Socrates.

49

What is visual acquaintance? Certainly it is not the capacity to recognise or identify Socrates. One can see Socrates and not recognise him, for he might be unrecognisable from his appearance, for example if he is in disguise; or because one has forgotten what he looks like; or because there are a lot of duplicates of Socrates around, so that one cannot identify the real Socrates. However, none of these circumstances would prevent one seeing Socrates, if he stood before one in a good light; so visual acquaintance with Socrates has nothing to do with a capacity to (re-)identify Socrates. According to Russell, ‘we have acquaintance with anything of which we are directly aware.’ (1912, p. 46) Thus one might propose that seeing Socrates is being visually acquainted with Socrates, i.e., it is being directly visually aware of Socrates; and that is indeed my proposal. But what is it to be aware of something, visually or otherwise? I suggest we should identify visual awareness with visual knowledge. This would explain the connection, insisted on by the sense-data theory, between seeing and visual experience; for we can have visual knowledge only in virtue of visual experience. The suggestion would also do what the Causal Theory failed to do, and define the ‘right’ kind of causal connection between the thing seen and the visual experience; namely, as that kind of causal connection which gives rise to direct visual knowledge of the thing seen. 2. Fact-knowledge. But what do we mean by ‘knowledge’ here? There is a distinction, pointed out by Russell, between what he calls ‘thing-knowledge’ and what he calls ‘knowledge of truths.’ Russell took thing-knowledge to be a simple relation between a mind and another entity. But knowledge of truths he took to be a more complicated matter: it is knowledge ‘that something is the case,’ and it ‘applies to our beliefs and convictions, i.e., to what are called judgements.’ (Russell 1912, p. 44) It is clear that we cannot analyse seeing Socrates as a kind of ‘knowledge of truths’ about Socrates. For in the first place, there is no one specific truth A about Socrates such that one sees Socrates only if one has visual knowledge that A. For example, to see Socrates is not to judge knowledgeably on the basis of visual experience that Socrates is here, since one might see Socrates but not recognise him, so that one does not judge that Socrates is here. Nor can we analyse one’s seeing Socrates as there being some truth or other A about Socrates, such that one knowledgeably judges that A on the basis of visual experience. For one might see

50

Socrates, yet not be in a position to make any knowledgeable judgements about Socrates. Suppose that a doctor one trusts tells one that Socrates is a purely fictional character; and also tells one that one has a medical condition that makes one prone to visual hallucinations as of Socrates. If later one sees Socrates, one will not believe one’s eyes, so despite one’s visual experience one will withhold the judgement that Socrates is here. There will be no sentence A about Socrates such that one judges that A on the basis of the visual experience; yet one sees Socrates. Visual awareness of Socrates is therefore not to be analysed as any ‘knowledge-of-truths’ about Socrates. We must distinguish ‘knowledge of’ from ‘knowledge that’, ‘??thing-knowledge?,’ from ‘?knowledge-of-truths?.’? Russell took acquaintance to be thing-knowledge, not knowledge-of-truths, and this certainly seems to be right in the case of visual acquaintance. However, he operated with an undifferentiated notion of thing-knowledge, treating knowledge of an individual and knowledge of a fact as exactly the same sort of thing. According to Russell, thing-knowledge is a relation in which a mind can stand to any entity whatever, of which it is aware: thus one might have thing-knowledge of the individual Socrates, or of the event of the birth of Socrates, or of the fact of the existence of Socrates. But this undifferentiated notion seems to me to elide an important distinction. Certainly we may speak with equal propriety in English of knowledge of an individual and knowledge of a fact, but the word ‘knowledge’ has different senses in the two uses. To know an individual is to recognise or identify it, or at least to have the capacity to do so; or to know it in the sense of being familiar with it, or having had personal experience of it. Knowledge of fact has none of these connotations: to know a fact is not to identify it, and need not imply familiarity with it, or personal experience of it. Therefore I think we need to distinguish two different sorts of thing-knowledge or ‘knowledge of’. Knowledge of an individual corresponds to German kennen and French connaître, whereas knowledge of fact corresponds to German wissen and French savoir. When used of an individual, the English verb ‘to know’ does not attribute a single simple relation, for its exact meaning on an occasion of use is a complex and context-dependent matter. But it appears plausible that ‘to know’ does attribute a single relation when used of knowledge of fact. I shall call this relation ‘fact-knowledge’: it is the relation that exists between mind and fact, when the mind is aware of the fact. I have suggested that seeing Socrates is visual awareness of Socrates, but have rejected the view that it is any kind of ‘knowledge that’, i.e.,

51

knowledge arising from judgement. Nor is visual awareness of Socrates ‘knowledge-of’, in the sense of knowledge of an individual: one can see Socrates without recognising or identifying him, or even having the capacity to do so; and one can see Socrates without being familiar with him, or having had personal experience of him. So my suggestion is that seeing Socrates is the other kind of ‘knowledge of’, namely (visual) factknowledge: to see Socrates is to know presently and visually some fact of which Socrates himself is a constituent; at a minimum, the fact of his existence. This suggestion makes seeing Socrates independent of the mode of presentation under which one is visually aware of him, so it leaves room for simple creatures to see Socrates even if they are incapable of demonstrative thought, and for people to see him without recognising him. 3. Is visual fact-knowledge constituted by visually warranted belief? The suggestion is that seeing Socrates is current visual fact-knowledge of some fact of which Socrates himself is a constituent. But what is visual fact-knowledge? According to many philosophers, knowledge is a kind of belief, namely the kind that is true and in some sense warranted. Visual knowledge would then be the kind of true belief whose warrant is visual. I shall discuss three difficulties with the theory that visual knowledge is belief. The first difficulty is the problem of singular belief. The difficulty is that the theory seems to lead to a circularity in the definition of seeing an individual, as follows. For definiteness, assume that believing is ‘saying in one’s heart’. (Substitution of a different account of belief will not affect the argument that follows.) Suppose one sees Socrates, because one visually knows that Socrates is here. If visual knowledge consists in saying in one’s heart, truly and with visual warrant, ‘Socrates is here’, then we must ask what makes one’s mental word ‘Socrates’ refer to Socrates. For several reasons, we should reject the doctrine that a proper name abbreviates a description. If ‘Socrates’ merely abbreviates a description, then the knowledge one expresses in saying in one’s heart ‘Socrates is here’ is knowledge of a general fact, viz., the fact that [the x: socratic (x)](x is here). This general fact does not have Socrates as a constituent, so it does not suffice to explain why it is specifically Socrates that one sees; for one could be knowing the same general fact in a world that did not even contain Socrates. An alternative version of descriptivism says that one ‘fixes the reference’ of ‘Socrates’ by a description which Socrates uniquely fits. But one

52

can fix the reference of ‘Socrates’ by a description that Socrates does not fit: e.g., by the description ‘the man drinking a martini’ even if it is water Socrates is drinking, and not a martini. One can also fix the reference by a description that is improper for a different reason, viz. that too many people fit it, for example, if one describes Socrates as ‘the man wearing a hat’. Nevertheless, these descriptions can work to fix the referent, if they focus visual attention on the right person, who then becomes the referent of ‘Socrates’. But this will not help us to analyse seeing Socrates, since the capacity to attend visually to Socrates presupposes seeing Socrates, which is what we were trying to analyse. We may also object to descriptivism that even if a description is not improper, it may fail to secure singular reference. For example, the name ‘Holmes’ refers to no one, since it occurs in a work of fiction. The story told by Conan Doyle generates a definite description of Holmes; the description is actually improper, because no one uniquely lived at Baker Street, noticed the remarkable thing about the dog in the night, etc., etc. But even if someone, say Schmidt, had by coincidence done all these things, still ‘Holmes’ would not refer to Schmidt. (Lewis 1983). The natural explanation of this is that the reason ‘Holmes’ does not refer to Schmidt is because Conan Doyle has never seen Schmidt, so he is not in a position to refer to Schmidt as ‘Holmes’, or indeed as anything else. According to the practice theory of Evans (1982) and Sainsbury (2005), one may use ‘Socrates’ to refer to Socrates in virtue of one’s participation in a linguistic practice. The reference is to Socrates, because the chain of co-referential uses of the name in the practice can be traced back to Socrates himself, since it is he who was called ‘Socrates’ in the initial act of dubbing. But when we consider the necessary conditions for successful dubbing, we come back to the problem of seeing again. For it is necessary to pick out the person who is to be dubbed, and the natural thing is to suppose that one points them out visually. It seems doubtful that one can dub by a mere description - can Conan Doyle dub Schmidt? It may be suggested that referential intention is what matters. It is certainly true that in the case imagined Conan Doyle does not dub Schmidt, but it might be argued that this is only because his intention in using the word ‘Holmes’ was not to refer to anyone, but merely to write a work of fiction. Had he told the Holmes stories with a genuine intention to refer to whoever fitted the description in the story, then he would in fact have referred to Schmidt. The claim here is that if in using ‘Holmes’ one intends to refer to the uniquely sherlockian person, then if x is uniquely sherlockian, then one refers by ‘Holmes’ to x. This seems implausible: the

53

fact that one desires to refer to the sherlockian person does not seem to entail that the sherlockian person is such that one desires to refer to him. To have such a desire, or such an intention, one would require to be already in a position to think singular thoughts about the sherlockian person. The definite description puts one in a position to have the general desire or intention to refer to the sherlockian person; but an account is still needed of how this general desire is to be fulfilled. Desires are not in general selffulfilling, and merely forming an intention to do F does not entail that one succeeds in doing F; this applies to referring to someone, just as it does to other actions. The obvious suggestion is that fulfillment of referential intention requires an information link, visual or otherwise, between oneself and the referent; but that would make the proposed analysis of seeing circular. A second problem with the belief theory is the problem of ‘indirect’ visual justification. Suppose it is granted that one’s mental word ‘Socrates’ does refer to Socrates, and suppose one says in one’s heart, truly, ‘Socrates is here’. When is this visual knowledge, in the sense required to guarantee that one sees Socrates? Clearly if one comes to believe that Socrates is here in virtue only of the testimony of another person, one’s warrant or justification is not visual, and one does not count as seeing Socrates. But even if one’s justification is visual, one may not be seeing Socrates. It may be that one sees only Socrates’ distinctive hat; one visually recognises the hat, so one has a justified belief that Socrates is here, but one does not see Socrates, because he is hidden behind a wall, and only his hat is visible. It may be said that this is indirect visual justification, and indirect justification does not count. But then we must say how to distinguish the direct kind of visual justification from the indirect kind; it seems clear that the direct kind will be defined only as the kind one has, if one is seeing Socrates. A third difficulty is that a simple creature may be incapable of beliefs, yet be perfectly capable of seeing Socrates. According to Davidson (1984), dogs lack beliefs. Since evidently dogs can see very well, it follows that if Davidson is right, then there can be can be visual knowledge that does not involve belief at all. If Davidson is wrong about dogs, we can repeat the argument with respect to chickens, gnats, etc. We should therefore reject the suggestion that visual awareness is any kind of belief. 4. Is visual knowledge constituted by veridical warranted visual experience? A different theory is that visual knowledge need not involve belief, the

54

visual experience itself being sufficient. Just as some have tried to define knowledge as the right kind of true belief, so we might try to define visual knowledge as the right kind of veridical visual experience. For example, we might say that visual knowledge that Socrates is here is a veridical experience with the content that Socrates is here, if this experience has arisen in the right way. But now we must repeat all over again the same question that we asked about the mental word ‘Socrates’. What makes the content that Socrates is here have anything to do with Socrates himself? If we characterise the content of the experience phenomenologically, in terms of the attendant sense-data, it looks as if one could have an experience with the same content, even if Socrates had never existed. We could say the connection with Socrates is that one is having an experience with a certain general content, e.g. that one faces something socratic, and that the experience is caused in the appropriate way by Socrates himself. But this is to embark on the vain attempt to analyse the ‘appropriate’ way without circularity, which we have already failed to do. In fact the suggestion gets things exactly backwards. It says the experience is about Socrates because Socrates is causally upstream of the experience. We do better to reverse the direction of explanation: it is what lies ‘downstream’ of the experience that makes it concern Socrates. It is because the experience confers knowledge of Socrates that its content has to do with Socrates. We characterise the content of the experience by the fact it causes us to know, or that it would cause us to know, if circumstances were favourable. Thus the very notion of the content of a visual experience seems to presuppose the notion of visual knowledge, and therefore we should give up the attempt to analyse visual knowledge as a species of experience. 5. Disjunctivism According to McDowell (1982), it is a mistake to think of perception and hallucination as two species of a single genus. We should not think of perception as the veridical warranted variety, and hallucination as the falsidical unwarranted variety, of something x called ‘experience’. There is, he thinks, no x which is the highest common factor. Rather the word ‘experience’ names a disjunctive or miscellaneous kind, which includes two quite distinct sorts of things, perception and hallucination. An example of a kind with a highest common factor are the owls. The brown owl and the tawny owl are two species of a common genus: every owl, of whatever species, intrinsically naturally resembles every

55

other owl, which is why the owls are a single natural kind. An example of a kind that lack a common factor are the pests. The wasp and the mouse are both pests, but they do not resemble each other in their own natures. Rather, they are both called ‘pests’ only because humans react to them in a similar way: the pests are therefore not a single natural kind. Disjunctivism is the doctrine that perception and hallucination do not share a common nature, but are both called ‘experience’ only because they evoke the same sort of reaction in human beings: their human subjects can’t tell them apart phenomenologically. One advantage of disjunctivism is that it takes to heart the lesson of Gettier, and once and for all gives up the fruitless attempt to analyse knowledge as belief + truth + something. Perception is a species of knowledge, for it involves ‘direct uptake of the fact itself’. So in the spirit of Gettier, disjunctivism once and for all gives up any attempt to analyse perception as experience + truth + something. A second advantage is that disjunctivism avoids the ‘veil of ideas’. On the Common Factor theory, the perceiver’s visual awareness (factknowledge) of the world is not a real relation in its own right, but is constituted by two other relations. The first of these is the has relation: a perceiver S has an experience e if S is the subject of e. The second relation is causation: the experience e must be appropriately caused by the fact p. The Common Factor theory says that seeing is not itself a fundamental relation but is constituted by the has relation and causation, as in the diagrams below: e

e is suitably caused by

has

p

S

has S

'sees' Experience - good case

Experience - bad case

On this account, ‘seeing’ does not attribute a genuine relation between the subject S and the fact p: rather, S and p satisfy the predicate in virtue of real relations in which each separately stands to the experience e. Therefore if the fact p were missing, for example if S were suffering a hallucination, then S would not be different in any real respect, for S would still

56

‘have’ the experience e. It is true that e would then be different, since it would no longer be caused by p; but that is a difference in e, not a difference in S. Thus on the Common Factor theory, the only thing to which S is really related in perception is the experience e itself. S is directly aware of e by consciousness, but S does not thereby stand in any direct cognitive relation to the fact p, or to any of its constituents. The ‘veil of ideas’ in perception exactly matches the ‘veil of descriptions’ in a descriptivist theory of reference. According to descriptivism, S refers to an object O if S has in mind a description d, and O uniquely fits d. On this account, ‘refers’ does not attribute a genuine relation, but is simply a semantically complex predicate satisfied by S and O if S has d in mind, and O fits d. Thus reference is constituted by the has-in-mind relation and the uniquely-fits relation, as in the diagram below: d has in mind

d uniquely fits O

S

has in mind

S

'refers to' Thought - good case

Thought - bad case

In the bad case where one has ‘strayed into the field of fiction without knowing it’, no object O uniquely fits the description, so there is nothing one is thinking of. However, one is no different mentally in the two cases, since reference is not a real relation. Disjunctivism has the advantage of decisively rejecting the veil of ideas. According to McDowell, the difference between good case and bad case ought not to be so ‘blankly external’ to the subject, and intuitively this seems right.1 The solution is to make seeing a real relation: the truth-maker for the predicate ‘sees’ is not a pair of relations, but the holding of the single real mental relation of seeing. In the bad case where one does not see but instead has a hallucination h, the seeing relation fails to hold. Thus there is a real mental difference in the subject in the two cases, according as the subject does or does not stand in the sees relation, as in these dia1

McDowell’s claim that scepticism is inescapable if the difference is ‘blankly external’ seems to be overstating the case, however; here I am persuaded by Martin (2006).

57

grams: h

e = hallucination

has

e = perception (fact-involving)

p

S

S sees Experience - good case

Experience - bad case

A similar story can be told in the theory of thought. In the good case where one refers to an object, one has a genuine singular thought; in the bad case one is merely the subject of a ‘mock thought’ m, thus: m

mock thought

has object-involving thought

O

S

S

thinks about Thought - good case

Thought - bad case

It might be objected that if perception and hallucination are really so different, and if thought and fiction are so different, then one should be able to notice the difference, and to tell which state one is in. Thus is the argument given by Ayer as the ‘Argument from Illusion’. (Ayer 1964, Chapt. 1) But the disjunctivist reply is that this is attributing to us more powers than we have: there are plenty of things that are different yet indiscriminable to one particular faculty. For example, Tweedledum and Tweedledee are visually indiscriminable, gunfire and a backfire are aurally indiscriminable, and red spheres and yellow spheres are tactually indiscriminable. None of these are absolutely indiscriminable, but they are indiscriminable with respect to one particular way of discriminating. The disjunctivist says perception and hallucination are consciously (phenomenologically) indiscriminable, and that the same goes for reference and fiction.

58

6. Difficulties for Disjunctivism Although disjunctivism has real advantages, there are also several difficulties with it. First, there is the problem of qualia. According to those who believe in them, a quale is a unary universal, i.e., a one-place property of an experience. If seeing is a relational state of the subject, and hallucinating is an intrinsic state, it seems implausible that two such different sorts of state can both be capable of instantiating the same unary universal. Of course, it might be held to be a virtue of disjunctivism that it is inhospitable to qualia. However, the arguments in favour of the reality of qualia are rather strong. For certainly the fact perceived does not on its own determine the qualitative character of experience. For example, Shoemaker (1996) says he enjoys a glass of cabernet sauvignon, his sense of taste making manifest to him certain facts about the chemical composition of the wine. The same facts might be made manifest to him visually, if he developed skill in telling a wine’s chemical composition from its appearance; but by just looking he would not enjoy all the pleasures of the oinophilist. Thus the fact perceived is not on its own enough to determine qualitative character. It can also be argued (Shoemaker 1984) that even the pair of the fact perceived and the sensory modality are together not enough to determine qualitative character, since we can easily imagine variations of qualitative character that depend on the biological species of the perceiver, or variations within species in cases of spectrum inversion, etc. Thus a case does exist for recognising qualia as a third and independent respect in which experiences can resemble or differ, which creates a difficulty for disjunctivism. A second difficulty for disjunctivism is the problem of consciousness. If one is the subject of a visual experience, one is consciously aware of the experience. Thus if one is having a hallucination as of a dagger, one is conscious of the experience of having the hallucination, though one need not be aware that the experience is a hallucination. And if one is perceiving a dagger, one is conscious of the experience of seeing the dagger, though one need not be aware that the experience is a perception. (Perhaps a doctor has told one that one is prone to hallucinations of daggers). What is the mechanism of consciousness supposed to be, according to disjunctivism? Here we must distinguish consciousness from perception: one may perceive a dagger, and in that sense be aware of it, but clearly it is not correct to say one is conscious of it. Consciousness delivers demon-proof knowledge of the occurrence of the experience. According to disjunctiv-

59

ism, the experience is either a case of perception, or it is a case of hallucination. So one knows by consciousness that one is perceiving-orhallucinating, i.e., that one is in either a dagger-involving relational state, or a daggerless intrinsic state. In all relevantly similar cases one knows a disjunction in virtue of knowing one of its disjuncts, i.e., K(A ∨ B) → (KA ∨ KB). But this seems strange here, for how can mere consciousness latch on to an external dagger, in the case that one’s experience is a veridical perception? A third difficulty for disjunctivism is that an explanation seems required of the phenomenal indiscriminability of each matching perception and hallucination pair, whether they are as of a pink elephant, or as of a dagger, or as of a little green man. For every veridical perception there is another experience type that is phenomenologically indiscriminable from the perceptual case. The Common Factor theory can easily explain the existence of the phenomenally indiscriminable ‘bad case’, so it is incumbent on disjunctivism to explain it too. A fourth difficulty is the narrow functional indiscriminability of the good and bad cases. If I face a bear and you are following me on the trail, I curl up in a ball and play dead and you run for help. (Perry 1990, p. 68) If I hallucinate facing a bear and you hallucinate following me on the trail, I still curl up in a ball and play dead and you still run for help. The narrow psychological role of the bear-involving perception, which has a bear as a constituent, and the bearless hallucination, are exactly the same. This gives prima facie support to the theory that there is a common factor at work here. A final difficulty is the broad functional indiscriminability of the good and bad cases. The rational subject S is characterised by the following two features, according to Bayesians. i) S maximises expected value, i.e., S chooses that action which maximises subjective probability of success times subjective value of outcome ii) S updates S’s beliefs in the light of experience by conditionalising. This means that if S experiences E, then S’s new subjective probability for H should equal S’s old subjective conditional probability for H given E: Pnew(H) = Pold(H | E)

For example, if one has a visual experience as of facing a bear, one’s newfound confidence that one faces a bear should equal one’s old conditional confidence that one would be facing a bear, given that one were having a

60

visual experience as of facing a bear. Therefore a Bayesian interpreter, e.g. Davidson (1984), must attribute a common ‘broad’ propositional content to the state of seeing a bear, and to the state of hallucinating a bear. This makes it natural to posit a single mental act here, with a single content, whether the experience as of a bear is a perception or a hallucination. 7. The Faculty Theory Indiscriminability is inescapable absence of knowledge of difference. If one cannot discriminate x from y, one cannot know that x ≠ y. The disjunctivist explains the indiscriminability of perception and hallucination as non-knowledge: there is a difference of which one can be ignorant. The common-factor theorist explains indiscriminability as non-difference: there is nothing to know, for one cannot know a ≠ a, because a = a. Instead of trying to decide between the two theories, I think we do best to seek a third theory that combines the advantages of both. The advantage of the Common Factor theory is that it says that a visual experience is a single kind of mental act (in the sense of Geach (1971)). It says that exactly the same kind of mental act is present in both the veridical and the hallucinatory cases. The advantage of this is that the single kind of mental act can do all of the following: it can be the bearer of qualitative character; it can be the object of phenomenal consciousness; it can be the occupier of the narrow functional role; it can be the occupier of the broad functional role; and it can be the vehicle of intentional content. The advantage of disjunctivism, in contrast, is its recognition that in the good case of visual ‘uptake’ something is really present in the subject that is missing in the ‘bad’ case of hallucination or dream: disjunctivism says that only in the case of veridical perception is there uptake of the fact itself. We seek a theory that combines these advantages. What I shall call the Faculty Theory of vision says that in the good case there is a mental act of visual experience, which is an act of ‘uptake’ because it causes visual awareness, whereas in the bad case there is the very same mental act but no ‘uptake’, because in this case the mental act fails to cause visual awareness. The Faculty Theory is a variant of the Common Factor theory; it differs from it only in declining to identify visual awareness with the visual experience itself. According to the Faculty Theory, in good case and bad case alike there is a common experience, but visual awareness is always something over and above the experience, because it is in its own right a genuine relation between mind and fact, namely, the relation of knowledge. This relation is caused to obtain by the

61

experience, but is not constituted by it. When one takes up an object in one’s hand, this event causes a relation to begin to obtain: one now holds the object. Similarly, according to the Faculty Theory, uptake of a fact is a mental act that causes a relation to begin to obtain: one now knows the fact. In veridical perception there are two quite different ‘awarenesses’ that need to be distinguished. First, there is awareness of the fact one perceives: for example, vision may make one aware of the fact that one faces Socrates. Secondly, there is one’s awareness of the fact of the visual experience itself. This second awareness is not perceptual, for one does not see one’s visual experience; rather, one is aware of it by consciousness. According to the Faculty Theory, in both the good and the bad case there is the same visual experience, which in both cases is the object of conscious awareness, and which has the same qualitative character, narrow functional role, and broad intentional content. The difference between the cases lies not in these features of the experience, but in whether or not it causes awareness (fact-knowledge) of the relevant environmental fact p. e

e visual faculty causes

has

p

S knows

Perception - good case

has

S

visual faculty causes p

Hallucination - bad case

In the good case, the fact p causes the visual faculty to cause the experience e, which in turn causes the subject to know the fact p. In the bad case some fact p causes the visual faculty to cause the same kind of visual experience e, but e fails to cause knowledge of p. ‘Veridical hallucination’ is the case where e has the broad content that p, and p exists and perhaps even causes e, yet e fails to cause knowledge of p. The Faculty Theory can explain the difference between actually seeing Socrates, and being merely aware of his presence by seeing his hat. In the case of seeing Socrates, it is the presence of Socrates that causes the visual faculty to cause the experience e that causes one to know that Socrates is here. In the case of seeing only his hat, it is the presence of the hat, and not the presence of Socrates, that causes the experience that causes one

62

to know that Socrates is here. One does not see Socrates, because one’s awareness of Socrates is not visual awareness, since the visual experience is not caused by the object of awareness. References Ayer, Alfred 1964: Foundations of Empirical Knowledge. London: MacMillan & Co. Dancy, Jonathan (ed.) 1988: Perceptual Knowledge. Oxford: Oxford University Press. Davidson, Donald 1984: ‘Thought and talk.’ In his Inquiries into Truth and Interpretation, Oxford: Oxford University Press. Evans, Gareth 1982: The Varieties of Reference. Oxford: Oxford University Press. See especially Chapters 4 and 6. Geach, Peter 1971: Mental acts: their contents and their objects. London: Routledge and Kegan-Paul. Grice, Herbert P. 1988: ‘The causal theory of perception’. In: Dancy (1988), pp. 66-78. Lewis, David 1983: ‘Truth in fiction’ In his Philosophical Papers Vol I, Oxford: Oxford University Press, pp. 261-280.  1986: ‘Veridical hallucination and prosthetic vision.’ In his Philosophical Papers Vol II, Oxford: Oxford University Press, pp. 273-290. Martin, Michael G.F. 2006: ‘On being alienated’. In: Gendler, Tamar and Hawthorne, John (eds.), Perceptual experience, Oxford: Oxford University Press, pp. 354-410. McDowell, John 1982: ‘Criteria, Defeasibility and Knowledge’. In: Dancy (1988), pp. 209-219. Perry, John 1990: ‘Frege on demonstratives’. In: Yourgrau, Palle (ed.) Demonstratives. Oxford: Oxford University Press, pp. 50-70. Russell, Bertrand 1912: Problems of Philosophy. Oxford: Oxford University Press. Sainsbury, Mark R. 2005: Reference without referents. Oxford: Oxford University Press. Shoemaker, Sydney 1984: ‘Functionalism and qualia’. In his Identity, Cause and Mind. Cambridge: Cambridge University Press, pp. 184-205.  1996: ‘Qualities and qualia; what’s in the mind?’ In his The First-Person Perspective and Other Essays, Cambridge: Cambridge University Press, pp. 97-120. Snowdon, Paul 1981: ‘Perception, Vision and Causation’. In: Dancy (1988), pp. 192-208.

63

Seeing something and believing IN it# Mark Textor [W]e need to reject as a mere dogma the presumption that belief can only be propositional in nature, since perceptual consciousness […] is intrinsically doxastic, or engaged. A.D. Smith

1. Introduction: The propositional dogma A mental state is only a belief if it has truth as its distinctive standard of correctness. Truth is explained by invoking propositions, the referents of ‘that’ clauses, as the contents of belief. If you believe that p, your belief is true iff p. Putting the points about belief and truth together, one arrives at the conclusion that a mental state can only be a belief if it has propositional content. That is, beliefs represent the world in a way that can only be completely captured by the nominalisation of an indicative sentence. You believe that London is in England; that roses are red etc. I will call this the propositional view of belief. The propositional view of belief is naturally extended to belief’s episodic counterpart, judgement. Judgements are mental events that initiate or re-activate beliefs. For example, judging that p is starting to believe that p. Today, the propositional view is widely accepted and seldom challenged. But this was not always the case. Descartes and Hume, among others, held a nominal view, that believing consists in (being disposed) to think an idea in a particular way. Hume wrote:

#

I wrote and presented this paper to honour the occasion of Alan Lacey’s 80th birthday. I am very grateful to Alan for inviting me to his Seriously Senior Research Seminar and for showing us youngsters the right way. I am also grateful for Alan’s comments at the conference. I especially remember Peter Adamson’s, Mike Martin’s and Mark Sainsbury’s contributions to the discussion. Many thanks go to Keith Allen and Johannes Brandl for written comments. Special thanks go to MM McCabe for giving me twice very helpful written comments. Thanks go to Tom Crowther, Jennifer Hornsby, Keith Hossack, Guy Longworth, Christian Nimtz, Kevin Mulligan, David Papineau, Francisco Pereira, Richard Samuels, Gabriel Segal for helpful criticism and suggestions that led to significant changes.

When I think of God, when I think of him as existent, and when I believe him to be existent, my idea of him neither increases nor diminishes. (Hume 1739, p. 94)

Thinking of God and believing that God exists are both relations to the idea of God.1 Talk of ideas is out of fashion today. The point of the nominal view is better brought out by saying that the content of a belief can be completely specified by a singular or plural term. A more recent spokesperson for the nominal view of belief is Franz Brentano (1838-1917), the father of Austrian Philosophy. He thus explains the nominal view by using the concept of a name: This symbolism [(A+), (A-)] contains everything which constitutes a simple judgement: a name that stands for what is judged; and a sign, which indicates whether what is judged (‘das Beurteilte’) is to be accepted or rejected. (LRU, p. 98. My translation)

In Brentano’s symbolism, ‘God+’ identifies a simple judgement completely by giving a term that expresses a mode of presentation of an object and a force indicator (+, -). A judgement and an idea can share their complete content. Gendler Szabó has recently re-discovered the same idea: Believing in Fs in the ontologically relevant sense requires more than merely believing that Fs exist. Believing in Fs is not even a propositional attitude, it is rather an attitude one bears to the term expressed by ‘F’. (Gendler Szabó 2003, p. 584. My emphasis)

I will call an intentional attitude that does not consist in a relation to a proposition a “nominal attitude”; I will call the corresponding view of belief the nominal view of belief.2 Further prima facie examples of nominal attitudes are admiration (‘John admires Napoleon’) or fear (‘Alonso fears Schumacher’s comeback’). Is every judgement a nominal attitude? Surely not. For which object do I endorse when I judge that if the table were made of gold, I would be rich? Because of this and similar problems one should reject any too gen1

For further sympathetic discussion of Hume’s theory of belief see Owen (2003). For Descartes see Principia I, 34. For readings of Descartes as a ‘judgement nominalist’ see Perler (1996, p. 257) and Kenny (1972, p. 13ff). 2

66

eral claim. However, the weaker claim that there are nominal judgements is, as we will see, still philosophically interesting.3 A further clarifying comment is necessary at this point. Take the prima facie nominal act of desiring something. Searle argues that one can only desire something x if one believes that x exists and that x is thus-andso. The desire is about x because the presupposed beliefs are about x. Searle goes on to generalise this idea: all nominal attitudes ‘inherit’ their intentionality and direction of fit from presupposed propositional attitudes.4 A proponent of the nominal view worth his salt will reject Searle’s treatment of nominal attitudes. The title ‘nominal view of belief’ covers a number of different options whose common core is the assumption that there is a subspecies of belief whose instances consist in thinking of an object in a particular psychological mode. For instance, one may take a nominal belief to be an undecomposable object-involving relation or a decomposable object-neutral relation (a judgement x is about an object x iff x’s representational content is a singular mode of presentation of x). My discussion will be neutral with respect to these issues. We can shed further light on the nominal view by considering a representative objection to it. Stroud argues that a judgement nominalist like Hume makes a grammatical mistake: [Hume speaks of ‘those ideas, to which we assent’ or ‘an idea assented to’, or when he says ‘we conceive many things, which we do not believe’.] [But] [s]trictly speaking, we do not believe or assent to the very same thing of which we have an idea. We have an idea of, e.g. the book’s being on the table, but when that idea figures in a belief it is a belief that the book is on the table’. (Stroud 1977, p. 257-8)

Fair enough. I have an idea of London, but I cannot assent to the thing, namely London, of which I have an idea. However, English and other natural languages have in addition to ‘believe’ which takes a that-clause as complement the intensional transitive verb ‘believe in’.5 ‘Believe in’ takes as complements singular (‘the monster of Loch Ness’) and plural terms (‘dwarves’). Sometimes ‘believe in’ is used to express a positive evaluative attitude (‘I believe in love’), but often we use it to convey our ontological 3

The weaker claim is endorsed by Husserl (1901 II/1, V, § 34ff) and Husserl (1901, p. 463). 4 See Searle (1983, p. 34). 5 The same distinction can be found in German (‘glauben an’ versus ‘glauben, daß’).

67

commitments.6 Here is a dramatic example that illustrates the ontological use of ‘believe in’: Holding on to the world is mostly an act of faith: you see a little bit in front of you and you believe in the rest of it both in time and space. If you’re scheduled for a jump to Hubble on Tuesday, you believe in you, in Hubble, in the jump, and in Tuesday. Sometimes it was hard for me to believe all of it. (Russell Hoban, Fremder)

Stroud and others have overlooked the possibility (which is also an actuality) that I may have an idea of Hubble and indeed believe in Hubble. Gendler Szabó has argued that we have to take this distinction seriously: ‘believe in’ locutions cannot be analysed in terms of ‘believe that’ locutions. If we take ‘believe in’ statements to ascribe mental states, we have therefore a reason to assume that belief-in is not a special form of belief-that.7 Believing-in is a nominal attitude distinct from the propositional attitude of believing-that. Believing-in is an attitude whose content can be completely specified by a singular or plural term.8 Now even if one can quell Stroud’s basic qualm, the nominal view will still strike many philosophers as hard to stomach. I will therefore argue in this paper that there are good reasons to hold that believing-in is a nominal attitude. I will develop in section 2 Brentano’s Argument from Visual Experience by working through its premises. Section 3 will complete Brentano’s argument by arguing that belief-in something is not the same as belief that something exists. If that is so, we need a positive account of belief-in; section 4 will contain the building blocks for such an account. Finally, Sections 5.1 and 5.2 will locate the nominal view in the space of alternatives. 2. Brentano’s ‘obvious proof’: The argument from visual experience In his Psychology from an Empirical Standpoint Brentano rejects the thesis that every belief involves predication. I can believe something without mentally ascribing a property to some object. If one can believe something 6

See Mulligan (2003, p. 28f). See Gendler Szabó (2003) and (2005). 8 Surprisingly, English and German only have verbs to ascribe doxastic nominal states: ‘believe in’. There is no corresponding verb for the episodic counterpart of belief-in: we don’t speak of judgement in, etc. Brentano, therefore, used ‘Anerkennung’ and ‘Verwerfung’ as quasi-technical terms. 7

68

without endorsing the content of an atomic sentence, the natural alternative is that the content of the belief is nominal. One of Brentano’s arguments is based on assumptions about visual experience. In this section I will expand on the basic premise of this argument. Here it is: That predication is not the essence of every judgement emerges quite clearly from the fact that all perceptions are judgements, whether they are instances of knowledge [‘Erkenntnis’] or just mistaken affirmations [‘Führwahrnehmen’]. […] [I]t is hard to think of anything more obvious and unmistakable than the fact that a perception is not a [connection] of a concept of the subject and a concept of a predicate, nor does it refer to such a connection. Rather, the object of an inner perception is simply a mental phenomenon, and the object of an external perception is simply a physical phenomenon, a sound, odor, or the like. We have here, then, a very obvious proof of the truth of our assertion. (PES-E, p. 209-10; PES-G II, p. 50-51)9

I have called Brentano’s argument ‘the Argument from Visual Experience’ and not ‘the Argument from Perception’, since he uses ‘perception’ (‘Wahrnehmung’) to cover the veridical and non-veridical cases in which something appears to us. Brentano takes the premise that every visual experience is a judgement to be an uncontroversial starting point. But a perception or a visual experience cannot be a judgement. For a judgement is a mental event, while perceiving is something one can do over a period of time (‘I have perceived the bird in the garden for some time’). Therefore Brentano would do better to argue that every perception is a belief. But again this seems not right. A belief is a dispositional mental state (You believe that the earth goes round the sun even when you sleep dreamlessly). In contrast, perceptions and visual experiences are paradigmatically conscious. Hence, we should say instead that there is a necessary connection between a perceptual experience as of an object and a belief. Indeed Brentano’s student Stumpf explains Brentano’s premise in this way: One must remember that for him [Brentano] every perception, inner as well as external, is a judgement, that this already constitutes for him an elementary affirmation, and that he is of the opinion that any kind of mental act, from the very beginning, is bound up with an evident self-affirmation, that is a judgement in the broadest sense of the word. This is all, of course, far removed from the standpoint that there are only verbally formulated judgements and

9

See also Stumpf (1919, p. 36).

69

then only those composed of subject and predicate. (Stumpf 1919, p. 36. My emphasis)

So does Marty: It is often said that everything that appears to the child is regarded by him instinctively as true, on the basis of innate necessity. On closer scrutiny, it can be seen that this instinctive belief is simply inseparable from sensation. […] It is not an act that is superimposed, for one-sided detachability is part of the concept of superimposition. The situation is such that sensation is an act which contains two mutually inseparable parts, namely the intuition of the physical phenomena and the assertoric acceptance of the same. (Marty 1895, p. 234. My emphasis.)

According to this account, a sensation or perceptual experience consists of two mutually inseparable (non-spatial) parts, an idea of something x and a belief-in that accepts x. Of course not any idea will qualify as a perceptual experience, the idea needs to be of a special kind. But for our purposes it is not necessary to say which kind this is. There is a strong prima facie reason to assume that perceptual experience and belief are connected in the way just described. If it visually seems to me as if a bird is directly before me, I am compelled to accept (in a sense to be clarified further) the bird’s existence. Hence, Reid has common sense on his side when he writes: In perception we not only have a notion more or less distinct of the object perceived, but also an irresistible conviction and belief of its existence. (Reid 1785, II, 5, p. 96)

Perceptual experience and belief seem to share an important feature: in belief as well as in perceptual experience one ‘takes’ things to be the way they are represented. It is for a good intuitive reason, then, that philosophers have tried to identify perception with belief, while no-one has ever been tempted to identify perception with desire or intention. Heck has borrowed the term ‘assertoric force’ from the philosophy of language to describe this commonality: perceptual experience and belief have assertoric force.10

10

See Heck (2000, p. 508) and Martin (2002, p. 390ff.).

70

However, there is also a well-known argument that seems to establish that perceptual experience cannot have assertoric force.11 You tell me that I will experience a hallucination: when it will look to me as if there is a man-eating tiger in the room, there will be nothing there. I trust your word: you have always been a reliable source of information. However, when it looks to me as if there is a man-eating tiger, there is indeed one and it causes my visual experiences in the right way. But since I trust you, I cannot rationally acquire the belief that there is a tiger. For I justifiably, but falsely believe that there is no tiger. Does this prevent me from having a visual experience of a tiger in this situation? No, I have such an experience and, moreover, I actually see a man-eating tiger. Hence, there can be no necessary connection between visual experience as of something and belief. I have raised this objection only to set it aside. There are a variety of ways to respond to it. Brentano and Marty have argued for the extreme position that in the situation just described you believe that there is a tiger before you AND you believe that there isn’t without being irrational. I have explored this response in another paper.12 In this paper I can fall back on a weaker premise that if one has a perceptual experience as of an object, and one does not have a reason to distrust one’s perception, one comes to have a belief that purports to be about the object. I also want to set aside another well-known difficulty. Take a look at figure 1: 13

11

See Bermudez (1995, p. 33); Craig (1975, p. 16); Evans (1982, p. 123); Martin (2002, p. 38); Siegel (2003, sec. 2.2). 12 See Textor (2007). 13 See Dretske (1993, p. 273).

71

If the situation is normal, you see all the spots and patches that compose Alpha and Beta. But Alpha contains one spot more than Beta. Normally, the additional spot is overlooked. However, it seems to be implausible to say that the additional spot is not seen. If one adds the assumption that belief-in x requires noticing x, one has an example of the perception of an object without belief.14 I am not persuaded that one can see an object without noticing it. There is a relaxed sense of ‘to see’ for which this may hold. However, this is not the sense of ‘to see’ that we use to describe our perceptual contact with the world. I will come back to this point in section 5.2. However, even if the reader remained unmoved, I could backtrack for the purposes of this paper to a weaker version of the original premise, namely that everything of which one is conscious in a perceptual experience is something one believes in.15 Now back to the main line of thought. Brentano’s argument started from the strong assumption that every perception is a judgement. We have now modified that to the weaker assumption that it is a ‘package’ that contains a belief. Brentano wants to argue that these beliefs are beliefs-in something, not beliefs that something is the case. Can we get Brentano’s conclusion from the weaker premise? Yes, I think so. Let us first distinguish propositional and non-propositional perceptual reports. First, propositional ones: Adam saw that the tomato is red. Adam heard that the TV set exploded. 14

See A.D. Smith (2001, p. 325ff). Brentano has an original answer to this difficulty. He says: Perception is an acceptance. And if the accepted thing is a whole with parts, then the parts are all, in a certain manner, concomitantly accepted [‘mitanerkannt’]. The denial of any of them would contradict the whole. Yet the individual part is, for this reason, by no means accepted let alone judged specifically [‘nicht ausdrücklich’] (by itself) and in particular. (DP-E, p. 36; DP-G, p. 34) In a concomitant acceptance the object is indistinctly presented. (‘indistinktes Vorstellen’, DP-G, p. 22f; DP-E, p. 25f) In the above counter-example, the perceiver believes in the complex figure. Thereby he believes in every part of the figure, but only concomitantly. Brentano rejects the assumption that believing in something requires, in general, noticing it. One can indistinctly believe in something without actually noticing it. Brentano’s answer needs more discussion than I can give it here. For example, we must make sure that the indistinct presentation of a is not just another way of saying that a is part of THE complex that is presented. For more, see Textor (2006). 15

72

Then, non-propositional or nominal ones. These contain derived nominals or definite descriptions: 16 Adam saw the tomato’s being red. Adam saw the tomato’s redness. Adam heard the explosion of the TV set. Nominal perceptual reports are logically independent from propositional perceptual reports. Adam can see the tomato’s being red without seeing that the tomato is red. Adam may see the tomato’s being red, but not come to the view that the tomato is red. Or Adam can see the tomato’s being red, without seeing that the tomato is red. An extreme view is that only nominal perceptual reports ascribe perceptions; every propositional perceptual report ascribes a judgement warranted by a perception.17 For my purposes the weaker thesis that there is a kind of sensory awareness that we don’t report by that-clauses is sufficient.18 If the reports are a reliable guide to the perceptual facts, there are perceptions and perceptual experiences with nominal content. When you say that Adam saw the tomato’s being red, this is an adequate description of how he perceptually represents things. Consider now the following example. You see the colours of the rainbow, but you don’t take your perception at face value, because you are convinced that you are suffering from an illusion. Hence, you don’t acquire a belief about the colours. Then you learn that your senses are working properly. Now you take your perceptual experience at face value. It seems very plausible to think that taking your perceptual experience at face value only involves losing the inhibiting belief. In particular, it does not require an additional propositional judgement. Taking a perceptual experience at face value only involves a change in the mode in which one and the same content is entertained: the same mode of presentation is now entertained with assertoric force. There is no additional content you need to grasp and judge in order to take your perceptual experience at face value. This example makes it plausible that a perceptual experience and a belief can represent the same object in the same way (under the same mode 16

Higginbotham (1983) and Parson (1987/88) discuss in addition perception sentences containing small clauses (‘Adam saw Eva naked’). Small clauses are also nonpropositional and it is tempting, although controversial, to treat perceptual reports with nominalisations, definite descriptions and small clauses in the same way. 17 See, for example, Collins (1967, p. 443). 18 See Johnston (2004, p. 270).

73

of presentation). The only difference between them is the psychological mode. Brentano seems to come close to this when he says that an idea of x and a belief-in x can have the same ‘content’ (‘Inhalt’). (PES-E, p. 221; PES-G II, p. 63) But unfortunately Brentano uses ‘content’ in different ways. The content of an idea is often just the object presented, sometimes it comes close to what is today called ‘representational content’. Minimally, Brentano commits himself to the view that an idea and a judgement can represent the same object and that the judgement does not represent an additional object. An idea presents an object, a belief-in is nothing but the acceptance of this object. We must remember here that according to Brentano, the object of an idea is an immanent object that exists in the idea. (PES-E, p. 88; PES-G I, p. 125) It would be good if we could discuss Brentano’s theories without committing ourselves to his theory of intentional inexistence. One can transpose Brentano’s main point into the contemporary views of intentionality, if one takes an idea and a judgement to share the same representational content. I have glossed representational content frequently as ‘mode of presentation’ and I take it that this gloss is sufficient for the purposes of this paper. The Brentanian view (that may not be exactly Brentano’s view) is then that an idea and a judgement may both have the same complete mode of presentation. For example, if I see a colour, I have a specific kind of idea of it. If I have an idea of this kind, I will, ceteris paribus, believe-in it. Belief-in consists in entertaining the singular mode of presentation involved in seeing the colour with assertoric force. Let us take stock. Brentano’s argument turned out to be less obvious than he thought it was, so we elaborated it in the following way: there is kind of sensory awareness that has nominal content (reported in statements, for example, like “S saw the redness of the rose”). It is sufficient for taking such an episode of sensory awareness at face value that one entertains its content (mode of presentation) with assertoric force. If one takes an episode of sensory awareness of the relevant kind at face value, one believes in the object the perceptual experiences purport to present. This belief has the same representational content as the perceptual experience. Hence, there are nominal beliefs, i.e. beliefs which refer to one object and don’t predicate anything of it. It is illuminating to compare this line of thought to an argument given by McDowell:

74

That things are thus and so is the content of the experience, and it can also be the content of a judgement: it becomes the content of a judgement if the subject decides to take the experience at face value. (McDowell 1994, p. 26)

McDowell and my Brentanian agree that a perceptual experience can share their complete representational content. McDowell assumes that every judgement has propositional content and concludes that perceptual experiences also have propositional content. The Brentanian assumes that perceptual experiences have nominal content. If we trust the way we report our sensory awareness of things, there are perceptual experiences that have nominal content. Therefore the Brentanian concludes that there are judgements that have nominal content. Brentano’s argument is based on a more plausible view of perceptual experience. Therefore, it seems preferable to give up on the dogma that judgement (belief) is a exclusively a relation to a proposition. 3. Belief-in and belief-that Section 2 supplied reasons for the conclusion that there are nominal beliefs. Is belief-in just a variety of belief-that? No, it seems implausible that belief-in a consists in a belief that a is F, for some feature F.19 Can the child’s seeing his mother consist in his acquiring or re-activating the belief that there is human being before him, something that feeds him, etc. This is rather unlikely. However, there is a possibility that seems not too far fetched. Brentano himself makes this point: Because we not only say that we perceive a colour, a sound, a seeing, a hearing, but also that we perceive that a seeing, a hearing exists, someone might be led to believe that perception, too, consists in the affirmation of the connection of the attribute ‘existence’ with the phenomenon in question. The untenability of such an opinion will, however, emerge with supreme clarity from a discussion of the concept of existence. (PES-E, p. 210; PES-G II, p. 51-52)20

Brentano’s pupil Meinong, for example, endorses that a perceiver S has a perceptual experience as of a iff S non-inferentially believes that a exists.21 If one adds to this that the existence belief is caused by the exercise of one 19

See Dretske (1969, p. 8). The translation is mine. The original translation spoils Brentano’s main points. 21 See Meinong (1921, p. 82). 20

75

of S’s senses, one has a good candidate for the analysis of perceptual experience as of an object. If the proposal can be defended, there would be no room for the nominal view of belief. I will now argue that the proposal made is implausible. Brentano himself argues that we first acquire the concept of existence by inner perception, hence, inner perception cannot require the concept of existence. When we have an experience, we accept it and abstract the concept of existence from these acceptances. Brentano’s argument is too speculative to bear conviction. It needs to be supplemented by arguments about the acquisition of the concept of existence. However, we can help Brentano to make his point. Assume that there is a first-order concept of existence that applies to everything. Assume that you perceive the colour of the rose before you and that your perception enables you to entertain a demonstrative mode of presentation of it (‘that colour’). Assume further for reductio that you predicate the concept of existence in your perceptual judgement that that colour exists. What would be the point of making this judgement? In special circumstances the judgement that that exists has a point. You took the little green man in your field of vision to be a figment of your imagination. Then you recognise that this little green man (really) exists. But normally there is no background assumption that something is a figment of the perceiver’s imagination. Since the judgement that this colour exists has no point, yet there is a judgement (belief) connected to the perception of the colour, we should take the judgement involved to be non-propositional. Brentano has shown us how we can do this. A further reason to distinguish belief-in so-and-so and belief that soand-so exists comes from pronouncements of ontological modesty.22 It seems reasonable to assert: (A)

I believe that things I don’t believe in exist,

but not: (B)

I believe in things I don’t believe in.

Yet, if believing in something were the same thing as believing that something exists, ‘x believes in y’ and ‘x believes that y exists’ should be replaceable salva veritate in all non-quotational contexts. The example above 22

See Gendler Szabó (2003, sec. 3).

76

shows that this is not the case. The difference in truth-conditions between (A) and (B) cannot be blamed on differences of scope. Even if we explicitly give (A) and (B) a reading in which all terms are within the scope of ‘I believe that’, (A) can be true, while (B) cannot. Now arguments that distinguish between Xs and Ys because certain expressions are not replaceable salva veritate have to be treated with caution, since there might be alternative explanations for the nonreplaceability. However, Brentano’s theory of belief offers one explanation for the difference. I can believe that there are things such that I don’t believe in them because believing that there are things is not the same as believing in some things. In order to believe in something I need to entertain a singular or plural mode of presentation of it. Believing in is entertaining such a mode of presentation in a particular way. This is not required for the quantification ‘there are things I don’t believe in’. ‘Things’ doesn’t express a mode of presentation of some things. No wonder I do not believe in some things there are; I have no mode of presentation that I could entertain in the assertoric mode. Although belief-in so-and-so and belief that so-and-so exists are different mental states, there is an intimate connection between them. If the thinker has a conception of Fs the following seem to hold: Necessarily, if x believes in Fs, and x is rational and possesses the concept of existence, x believes that Fs exist (and vice versa).23 If I believe in the colours of the rainbow, and I am rational and possess the concept of existence, I will also believe that the colours of the rainbow exist (and the other way around). But still believing-in so-and-so and believing that so-and-so exists are different states. Now one will feel inclined to ask: What is the point of the distinction between belief-in and belief that something exists? It is open to less sophisticated believers to see Fs and to believe in Fs, while it is not open to them to believe that Fs exist. Having such a notion of belief is of importance for understanding the nature of perception. In the next section I will show how belief-in can be more basic than belief-that.

23

Gendler Szabó (2003, p. 606ff) challenges this connection.

77

4. More about believing-in The result of the last two sections is that there are reasons to acknowledge that belief-in is a sui generis doxastic attitude. If there is such an attitude, we should be able to distinguish it from other attitudes in a non-questionbegging way. Moreover, the last section has already established a constraint on an account of belief-in. It must be possible for someone to believe in something x, although it is not possible for her to believe that x is so-and-so. What can we say positively about belief-in? First, as already said in section 1, judgment and belief have a distinctive relation to truth. If an intentional attitude cannot be evaluated in terms of truth (falsity), it is no judgement. But if we acknowledge nominal judgements, i.e. acceptances of objects full-stop, how can they be true? Brentano’s answer starts from the Aristotelian explanation of truth: To say of what is that it is not, or of what is not that it is, is false, while to say of what is that it is, or of what is not that it is not, is true. (Metaphysics Γ 7: 1011b 26-7) The standard interpretation reads the ‘is’ in Aristotle’s dictum as the ‘is’ of predication. If you say of what is (not) thus-and-so, that it is (not) thusand-so, your saying is true; if you say of what is not thus-and-so, that it is thus-and-so, your saying is false. For instance, if you say of blood that it is red that it is red, your saying is true. In contrast Brentano reads the ‘is’ in Aristotle’s formula as the ‘is’ of existence (‘God is’): [A] judgement is true if it asserts of some object that is, that the object is, or if it denies of some object that is not, that the object is. (WE, § 51)

A judgement accepts one or more things and the acceptance is true iff the thing(s) accepted exist. Brentano’s way of expounding the conception of truth for nominal judgements is misleading. For it sounds as if a judgement ascribes to an object existence. A better way of putting Brentano’s idea is to consider the following schema: If Y believes in x, then Y believes truly in x iff x exists. If we substitute for ‘Y’ a designation of the believer and for x a designation of the object believed in, every instance of the schema is true. Our grasp of the concept of truth consists in our disposition to accept instances of the 78

schema without requiring a reason for our acceptance. Fair enough, the schema may need refinement. For example, Gendler Szabó develops a similar idea, but he requires that the conception the believer connects with his designation for x is true of x.24 But the schema can serve at least as our starting point to understand the distinctive standard of correctness for believe-in without assuming that belief-in is a species of belief-that. One may want to reserve the predicate ‘is true’ for representations with propositional content. I sympathise with this terminological policy. The important point is that there is a distinctive standard for nominal belief analogous to truth for propositional belief. Second, can we also say something more about the psychological mode in which someone thinks of an object when she believes in it? Simple creatures can perceive things. Hence, if perceiving something is a species of believing in (acquiring or re-activating a belief-in), such creatures must be able to believe in things. This puts constraints on a characterisation of the psychological mode distinctive of belief-in. For example, we cannot characterise the distinctive psychological mode in terms of propositional attitudes. However, there is a whole family of nominal attitudes, especially emotions and feelings which we can make use of. Employing the general functionalist strategy with respect to mental concepts, we can try to define believe-in by outlining its place in a web of other basic nominal attitudes and feelings. Here is a part of the web of nominal attitudes: Ceteris paribus, if S believes in x, and S fears x, S will try to avoid x. Ceteris paribus, if S believes in x, and S hates x, S will seek to harm x. Ceteris paribus, if S believes in x, and S loves x, S will care for x. Ceteris paribus, if S’s attention is drawn to x, S acquires a belief-in x. Ceteris paribus, if S acts upon x, S believes in x. If we know the place of belief-in in the web of nominal attitudes, we know what belief-in is. Many of the nominal attitudes are attitudes that one may legitimately ascribe to simple minded beings. Hence, the functional explanation meets the constraint motivated above.

24

Gendler Szabó (2003, pp. 603-605).

79

5. A look at the competition 5.1 Non-conceptual content Many contemporary authors draw from the observation that in object perception one does not bring particulars under concepts - in addition with further arguments - the conclusion that perceptual experience has nonconceptual content. The question ‘What is non-conceptual content?’ is of course tricky. A satisfactory answer requires one to argue for an explication of the concept of a concept. I can’t argue for such an explication in this paper (nor in any other paper). But an acceptable starting point for an answer to this question is Evans’ ‘Generality Constraint’: For every propositional attitude φ, one can only φ that a is F if one can also φ that a is G, for every categorically adequate property G.25 An attitude φ has only conceptual content if it satisfies the Generality Constraint. Roughly speaking, an attitude has conceptual content if it is an element in a web of logically related mental representations. Mental representations with non-conceptual content don’t satisfy the Generality Constraint, but they still represent the world in a certain way. Many philosophers take perceptual experience and perception to have non-conceptual content that is propositionally structured. Here is an example: [I]n characterizing the fine-grained content of experience, we need the notion of the experience representing things or events or places or times, given in a certain way, as having certain properties or as standing in certain relations, also as given in a certain way. (Peacocke 2001, p. 241)

A perceptual experience non-conceptually represents something as thusand-so. Non-conceptualists argue that the non-conceptual content of perceptual experience is adequately expressed (approximated) by demonstratives. The non-conceptualist takes the observations of the previous section to suggest that our fundamental perceptual contact with the world is perceiving that this is thus or that this is over there. Perceptual experience is non-conceptual, but propositional. Byrne brings out the non-conceptualist’s commitment to the propositional nature of perception: All parties agree, in effect, that perceiving is very much like a propositional at25

See Evans (1982, p. 104).

80

titude, such as believing or intending, the issue is whether the contents or propositions that perceiving is a relation to are conceptual. When it is put like that, nonconceptualism is decidedly puzzling. When one has a perceptual experience, one bears the perception relation to a certain proposition p. The nonconceptualist claims that it is impossible to bear the believe relation to p-but why ever not? Absent some argument, the natural position is that the contents of perceptions can be believed. (This unappetizing feature of nonconceptualism is somewhat obscured because participants in the debate typically reserve ‘p’ for conceptual contents.) (Byrne 2005, p. 245)

Non-conceptualists take perception and perceptual experience to be representations that such-and-such is the case. This assumption gets the phenomenology of many perceptual episodes wrong. When I see the colours of the rainbow, I don’t stand in a relation to a propositional content, neither conceptual nor non-conceptual. Perceptual nominalists like Brentano can allow for non-conceptual content, but they will deny that our fundamental perceptual contact with the world is propositional in form. For example, Brentano holds that if something appears to you, you have a presentation of it. Something can appear to you, although you have no concept of it, that is, no mental representation that satisfies the Generality Constraint. What a ‘perceptual nominalist’ denies is that perceptual experience and perception are propositional. 5.2 Dretske’s non-cognitive theory of object Perception Dretske agrees with Brentano that object perception is not a species of judging that p nor founded on a propositional attitude. However, he draws another conclusion: object perception should not be explicated in terms of belief, judgement, etc. at all. If this is correct, the argument does not give us a reason to assume a nominal assertoric attitude. Dretske argues for a non-epistemic understanding of object perception by means of an example. S is looking for his cuff-link. The cuff-link is exactly in the place where S’ gaze is directed at, it is open to view and the only object at this position, S’ eyes are open, the light is turned on etc. In these conditions, Dretske claims, we take S to see the cuff-link.26 We are entitled to reprimand him with the words “But you must have seen it!” Why? If the physical and psychological conditions of perception are satisfied, we have a defeasible reason to say that S saw a. Dretske will therefore explicate the notion of object perception in terms of visual differentiation. 26

See Dretske (1969, p. 18).

81

(Dretske 1969, p. 20) But does S really see a under these conditions? We take the satisfaction of the physical and psychological conditions for perception only to be evidence for object seeing if the perceiver is noticing something.27 In Dretske’s example, S looked at the cuff-link, but didn’t see it. The cuff-link was so placed in his visual field that he would have seen it if he would have paid attention. Even if one wants to say that a perceptual relation obtains in this case, we have at best a peripheral or degenerate case of object perception.28 Why? There is a connection between seeing an object, (i) making justified perceptual judgements about it and (ii) referring demonstratively to it. If I see an object a, and I have the linguistic competence required to make a demonstrative reference, then I can demonstratively refer to a on the basis of my perception. A similar principle connects seeing a and being in a position to make justified judgements about it. These connections are severed if Dretske’s examples are cases of seeing an object. One could see the bird, master the practice of demonstrative reference and want to refer to every bird one sees demonstratively, yet not refer to the bird. Seeing something starts to lose its place in our common-sense framework if we accept Dretske’s theory. For these reasons I agree with Pitson: The condition that what is seen should be discriminated or differentiated from its environment points to a cognitive element in seeing that it is the purpose of nonepistemic theories of perception to deny. However this element is to be spelled out, its omission leaves us not with a kind of seeing that is basic or simple but with something that is simply not seeing. (Pitson 1984, pp. 128-9)

Dretske is right about the following point, however: in order to see an object one need not be able to bring it under concepts and to make judgements about it. He infers from this that object perception does not involve any epistemic or doxastic element. This is too quick. Dretske offers his readers examples in which someone is looking in the direction of an object in favourable external and internal conditions without noticing it. What factor is it whose absence makes these examples degenerate cases of seeing something? Easy, you will say, the object is not noticed; you are not aware of it. But if one holds that judgement and belief are propositional attitudes it is difficult to add something like noticing an object to Dretske’s account. For the addition will not be in harmony with the plausible view 27 28

See Pitson (1984) and Martin (1992, p. 749). See Siegel (2006, pp. 430-31).

82

that one can see something without being able to have propositional attitudes concerning it. If I notice something, I take it to exist; i.e. I believe that it exists. Hence, either we give up that object perception requires noticing an object or we assume that there is a form of belief that is not a propositional attitude. Dretske accepts the first option and gives for this reason an impoverished theory of object perception. Brentano can accept the second option and say: S perceives a iff (i) S perceives an event or state x whose subject is a; (ii) perceiving x differentiates a for S from a’s surroundings; (iii) S acquires or re-activates a non-inferential (distinct or indistinct) belief-in a, because a is differentiated for S from a’s immediate surroundings (iv) the acquired or re-activated belief constitutes knowledge. (iii) is the crucial new ingredient. This move enables Brentano to take the criticism of doxastic theories of object perception on board (Yes, perceiving a is not making a judgement about it that it is so-and-so), while preserving the intuition that object perception requires noticing and accepting an object. Hence, I take direct arguments against doxastic theories of object perception to be indirect arguments for the existence of belief-in. Please note also that the proposed explication preserves the core of Reid’s theory of perception: perceiving something is believing in it on the basis of a distinguishing conception. 6. Conclusion I hope to have made a case for the nominal view of judgement and belief. Of course, more needs to be done to develop the nominal view into a fruitful theoretical option. One important future task will be to understand the rational and logical connections between belief-in and other attitudes. Brentano developed term logic. This logic looks like a natural starting point for this project. Another task is to understand the epistemic role of beliefs-in. It is widely agreed that perception and knowledge are connected. Perception is our main source of empirical knowledge. We must ask whether the nominal view of judgement helps us to understand the justificatory force of object perception better. Important steps to get clear about this have been

83

taken by Husserl and recently by Johnston.29 These steps should be followed. References: Armstrong, David M. 1968: A Materialist Theory of the Mind. London: Routledge. Ayers, Michael 2004: ‘Objections to Davidson and McDowell’. In: Schuhmacher, Ralf (ed.): Perception and Reality. Paderborn: Mentis, pp. 239-263. Bermudez, Jose L. 1995: ‘Nonconceptual Content: From Perceptual Experience to Subcomputational States’. Mind and Language, 10, pp. 333-369. Brandl, Johannes: ‘Brentano’s Theory of Judgement’. Stanford Encyclopedia of Philosophy (Online). Brentano, Franz:  (PES-G) Psychologie vom Empirischen Standpunkt (1874) (two volumes). Hamburg: Meiner 1955. Trans.: (PES-E): Psychology from An Empirical Standpoint. London: Routledge 1995 (trans. Linda MacAlister).  (DP-G) Deskriptive Psychologie. Hamburg: Meiner 1982. Trans.: (DP-E) Descriptive Psychology. London: Routledge 1995, (trans. Benito Mueller).  (VNV) ‘Von der Natur der Vorstellung’ (Ein Diktat aus dem Nachlass (1903)). Conceptus 21 (1987): 25-31.  (LRU) Die Lehre vom Richtigen Urteil. Bern: Franke 1956.  (WE) Wahrheit und Evidenz. Leipzig: Meiner 1930. Byrne, Alexander 2005: ‘Perception and Conceptual Content’. In: Steup, Matthias and Sosa, Ernest (eds.): Contemporary Debates in Epistemology. Oxford: Blackwell, pp. 231-250. Campbell, John 2002: Reference and Consciousness. Oxford: Oxford University Press. Cohen, Jonathan L. 1990: An Essay on Belief and Acceptance. Oxford: Oxford University Press. Collins, Arthur W. 1967: ‘The Epistemological Status of the Concept of Perception’. The Philosophical Review, 76, pp. 436-459. Craig, Edward 1976: ‘Sensory Experience and the Foundations of Knowledge’. Synthese, 33, pp. 1-24. Crimmins, Mark 1992: Talk about Beliefs. Cambridge, Mass.: M.I.T. Press. Dretske, Fred 1969: Seeing and Knowing. London: Routledge and Keegan.  1993: ‘Conscious Experience’. Mind, 102, pp. 1-21. Evans, Gareth 1982: Varieties of Reference. Oxford: Oxford University Press. Gendler Szabó, Zoltan 2003: ‘Believing in Things’. Philosophy and Phenomenological Research, 66, pp. 584-611.  2005: ‘Sententialism and Berkeley’s Master Argument’. The Philosophical Quarterly, 55, pp. 462-475.

29

See Husserl’s sixth Logical Investigation and Johnston 2005.

84

Heck, Richard 2000: ‘Nonconceptual Content and the “Space of Reasons”’. The Philosophical Review, 109, pp. 483-523. Higginbotham, James (1983): ‘The Logic of Perceptual Reports: An Extensional Alternative to Situation Semantics’. The Journal of Philosophy, 80, pp. 100-127. Husserl, Edmund 1990: Logische Untersuchungen II, Halle a.S.: Max Niemeyer (Reprint Tübingen: Niemeyer 1980). Hume, David 1739: A Treatise of Human Nature. Oxford University Press: Oxford 1976. Johnston, Mark 2005: ‘Better than Mere Knowledge? The Function of Sensory Awareness’. In: Szabó Gendler, Tamar and Hawthorne, John (eds.): Perceptual Experience. Oxford: Clarendon Press, pp. 260-291. Kenny, Anthony 1980: Aquinas. Oxford: Oxford University Press.  1972: ‘Descartes on the Will’. In: Butler, Ronald J. (ed.) Cartesian Studies, Oxford: Basil Blackwell, pp. 1-32. Kraus, Oskar 1976: ‘Towards a Phenomenognosy of Time Consciousness’. In: McAlister, Linda (ed.): The Philosophy of Brentano, London: Duckworth, pp. 224-239. Martin, Michael G. F. (1992): ‘Perception, Concepts, and Memory’. The Philosophical Review, 101, pp. 745-763.  2002: ‘The Transparency of Experience’. Mind and Language, 17, pp. 376-425. McDowell, John 1994: Mind and World. Cambridge, Mass.: Harvard University Press. Mulligan, Kevin 2003: ‘Seeing, Certainty and Apprehension’. In: Fossheim, Hallvard, Mandt Larsen, Tarjei and Sapeng, John R. (eds.) Non-Conceptual Aspects of Experience. Oslo: Unipub forlag, pp. 27-44. Mulligan, Kevin; Simons, Peter; Smith, Barry 1984: ‘Truth-makers’. Philosophy and Phenomenological Research, 64, pp. 287–321. Owen, David 2003: ‘Locke and Hume on Belief, Judgement and Assent’. Topoi, 22, pp. 15-28. Parsons, Charles 2004: ‘Brentano on Judgement and Truth’. In: Jacquette, Dale (ed.): The Cambridge Companion to Brentano, Cambridge: Cambridge University Press, pp. 168-196. Parsons, Terence 1987/8: ‘Underlying States in the Semantical Analysis of English’. Proceedings of the Aristotelian Society, 88, pp. 13-30.  2000: ‘Underlying States and Time Travel’. In: Higgenbotham, James; F., Pianesi, Fabio; Varzi, Achille (eds.): Speaking of Events. New York: Oxford University Press, pp. 81-93. Peacocke, Christopher 2001: ‘Does Perception have a Nonconceptual Content?’ The Journal of Philosophy, 98, pp. 239-246. Perler, Dominik 1996: Repräsentation bei Descartes. Frankfurt a.M.: Klostermann. Pitson, Anthony E. 1984: ‘Basic Seeing’. Philosophy and Phenomenological Research, 45, pp. 121-29. Reid, T. 1764: An Inquiry into the Human Mind, ed. by Broackes, Justin. Edinburgh: Edinburgh University Press 1997.  1785: Essays on the Intellectual Powers of Man, ed. by Brookes, Derek R. University Park, Pennsylvania State University Press 2002.

85

Searle, John 1983: Intentionality. Cambridge: Cambridge University Press. Siegel, Susanna 2003: ‘The Contents of Perception’. Stanford Encyclopedia of Philosophy (Online).  2006: ‘How Does Visual Phenomenology Constrain Object-Perception?’ Australasian Journal of Philosophy, 84, pp. 429-442. Simons, P. 2004: ‘Judging Correctly: Brentano and the Reform of Elementary Logic’. In: Jacquette, Dale (ed.): The Cambridge Companion to Brentano. Cambridge: Cambridge University Press, pp. 45-65. Smith, David A. 2001: ‘Perception and Belief’. Philosophy and Phenomenological Research, 62, pp. 283-309. Stroud, Barry 1977: Hume. London: Routledge. Stumpf, Carl 1919: ‘Reminisciences of Brentano’. Reprinted in: McAlister, Linda (ed.): The Philosophy of Brentano, London: Duckworth 1976, pp. 10-46. Textor, Mark 2006: ‘Brentano (and some Neo-Brentanians) on Inner Consciousness’. Dialectica, 60, pp. 411-432.  2007: ‘Brentano on the Doxastic Nature of Perceptual Experience’. History of Philosophy and Logical Analysis, 11, pp. 137-157.

86

Phenomenal concepts are not demonstrative1 David Papineau

1. Introduction In this paper I want to explore the nature of phenomenal concepts by comparing them with perceptual concepts. Phenomenal concepts have been drawn to the attention of philosophers by recent debates in the philosophy of mind. Most obviously, their existence is demonstrated by Frank Jackson’s thought–experiment about Mary, the expert on the science of colour vision who has never had any colour experiences herself. It is widely agreed that, when Mary does first see something red, she acquires a new concept of red experiences, distinct from any of her previous scientific concepts of such experiences. This new mode of reference is an example of a phenomenal concept. Recent interest in phenomenal concepts is independent of views about the ontological significance of Jackson’s Mary argument. Thus phenomenal concepts are acknowledged both (a) by ontological dualists who take the Mary argument to demonstrate the non–physicality of conscious phenomena and (b) by physicalist monists who insist that Mary’s new concept refers to nothing but a material state that she could always refer to using her old scientific concepts. How then do phenomenal concepts work? Here there is far less consensus. Among those who trade in phenomenal concepts, some take them to be sui generis (Tye, 2003, Chalmers, 2003), while others have variously likened them to recognitional concepts (Loar, 1990), to demonstratives (Horgan 1984, Papineau 1993, Perry 2001), or to quotational terms (Papineau 2002, Balog forthcoming). In my Thinking about Consciousness (2002), I developed a ‘quotational–indexical’ account of phenomenal concepts on roughly the following lines. To have a phenomenal concept of some experience, you must be able introspectively to focus on it when you have it, and to recreate it imaginatively at other times; given these abilities, you can then form terms 1

This paper is adapted from some sections of my 2007.

with the structure the experience: —, where the gap is filled either by a current experience, or by an imaginative recreation of an experience; these terms then comprise a distinctive way of referring to the experience at issue. In this paper, I want to return to the topic of phenomenal concepts. It now seems to me that the treatment in Thinking about Consciousness was inadequate in various respects. Here I want to try to improve on that account. In particular, I shall develop an extended comparison of phenomenal concepts with what I shall call ‘perceptual concepts’, hoping thereby to throw the nature of phenomenal concepts into clearer focus. The revised account will enable me to deal with a common worry about phenomenal concepts.2 Suppose Mary has come out of her room, seen a red rose, and as a result acquired a phenomenal concept of the experience of seeing something red (though she mightn’t yet know that this experience is conventionally so–called). On most accounts of phenomenal concepts, including the one developed in my book, any exercise of this phenomenal concept will demand the presence of the experience itself or an imaginatively recreated exemplar thereof. The trouble, however, is that it seems quite possible for Mary to think truly, using her new phenomenal concept, I am not now having that experience (nor recreating it in my imagination) — but this would be ruled out if any exercise of her phenomenal concept did indeed depend on the presence of the experience or its imaginative recreation. The revised account of phenomenal concepts to be developed here will not require this, and so will be able to explain Mary’s problematic thought. 2. Perceptual concepts 2.1 Perceptual concepts are not demonstrative Let me turn away from phenomenal concepts for a while, and instead consider perceptual concepts. Getting clear about perceptual concepts will stand us in good stead when we turn to the closely related category of phenomenal concepts.

2

This worry has long been pressed on me by my London colleagues Tim Crane and Scott Sturgeon, and is developed in Crane’s contribution to a forthcoming symposium on Thinking about Consciousness in Philosophy and Phenomenological Research. The same objection is attributed to Kirk Ludwig in a typescript by Ned Block entitled ‘Max Black’s Objection to Mind–Body Identity’.

88

We can start with this kind of case. You see a bird at the bottom of your garden. You look at it closely, and at the same time think I haven’t seen THAT in here before. Later on you can recall the bird in visual imagination, perhaps thinking I wonder if THAT was a migrant. In addition, on further perceptual encounters with birds, you sometimes take some bird to be the same bird again, and can again form further thoughts about it, such as THAT bird has a pleasant song. (Let me leave it open for the moment whether you are thinking of a particular bird or a type of bird; I shall return to this shortly.) In examples like this, I shall say that subjects are exercising perceptual concepts. Perceptual concepts allow subjects to think about perceptible entities. Such concepts are formed when subjects initially perceive the relevant entities, and they are re–activated by latter perceptual encounters. Subjects can also use these concepts to think imaginatively about those entities even when they are not present. Now, it is tempting to view concepts of this kind as ‘demonstrative’. For one thing, it is natural to express these concepts using demonstrative words, as the above examples show (‘. . . THAT . . .’). Moreover, uses of perceptual concepts involve a kind of perceptual attention or imaginative focus, and this can seem analogous to the overt pointing or other indicative acts that accompany the use of verbal demonstratives. However, I think it is quite wrong to classify perceptual concepts as demonstratives. If anything is definitive of demonstrative terms, it is surely that they display some species of characterlikeness. By this I mean that the referential value of the term is context–dependent — the selfsame term will refer to different items in different contexts. However, there seems nothing characterlike about the kind of perceptual concept illustrated in the above examples. Whenever it is exercised, your perceptual concept refers to the same bird. When you use the concept in question, you don’t refer to one bird on the first encounter, yet some possibly different bird when later encountering or visually imagining it. Your concept picks out the same bird whenever it is exercised. It is possible to be distracted from this basic point by failing to distinguish clearly between perceptual concepts and their linguistic expression. If I want to express some perceptual thought in language, then there may be no alternative to the use of demonstrative words. In order to convey my thought to you, I may well say ‘That bird has a pleasant song’, while indicating some nearby bird. And I agree that the words here used — ‘that bird’ — are demonstrative, in that they will refer to different birds in

89

different contexts of use. But this does not mean that my concept itself is demonstrative. As I have just urged, my concept itself will refer to the same bird whenever it is exercised. The reason we often resort to demonstrative words to convey thoughts involving non–demonstrative perceptual concepts is simply that there is often no publicly established linguistic term to express our concept. In such cases, we can nevertheless get our ideas across by demonstratively indicating some instance of what we are thinking about. Of course, this possibility assumes that some such instance is available to be demonstrated—if there isn’t, then we may simply find ourselves unable to express what we are thinking to an audience. By insisting that perceptual concepts are not demonstrative, even if the words used to express them are, I do not necessarily want to exclude characterlikeness from every aspect of the mental realm. Millikan (1990) has argued that mental indexicality plays no ineliminable role in the explanation of action, against Perry (1979) and much current orthodoxy, and I find her case on this particular point persuasive. Even so, I am open to the possibility that primitive mental demonstratives may play some role in pre–conceptual attention (what was THAT?) and also to the possibility that there may be characterlike mental terms constructed with the help of predicates (I’m frightened of THAT DOG – i.e. the dog in the corner of the room).3 In both these kinds of case I allow that the capitalised expressions may express genuinely characterlike mental terms — that is, repeatable mental terms that have different referents on different occasions of use. My claim in this section has only been that perceptual concepts in particular are not characterlike in this sense, but carry the same referent with them from one occasion of use to another. 2.2 Perceptual concepts as stored templates I take perceptual concepts to involve a phylogenetically old mode of thought that is common to both humans and animals. We can helpfully think of perceptual concepts as involving stored sensory templates. These templates will be set up on initial encounters with the relevant referents. They will then be reactivated on later perceptual encounters, via matches between incoming stimuli and stored template — perhaps the incoming stimuli can be thought of as ‘resonating’ with the stored pattern and thereby being amplified. Such stored templates can also be activated 3

I will take no stand on whether or not such ‘mixed demonstratives’ are equivalent to definite descriptions.

90

autonomously even in the absence of any such incoming stimuli — these will then constitute ‘imaginative’ exercises of perceptual concepts.4 The function of the templates is to accumulate information about the relevant referents, and thereby guide the subject’s future interactions with them. We can suppose that various items of information about the referent will become attached to the template as a result of the subject’s experience. When the perceptual concept is activated, these items of information will be activated too. They may include features of the referent displayed in previous encounters. Or they may simply comprise behavioural information, in the form of practical knowledge that certain responses are appropriate to the presence of the referent. When the referent is re–encountered, the subject will thus not only perceive it as presently located at a certain position in egocentric space, but will also take it to possess certain features that were manifested in previous encounters, but may not yet be manifest in the re–encounter. Imaginative exercises of perceptual concepts may further allow subjects to process information about the referent even when it is not present. Note how this function of carrying information from one use to another highlights the distinction between perceptual concepts and demonstratives. Demonstrative terms do not so carry a body of information with them, for the obvious reason that they refer to different entities on different occasions of use. Information about an entity referred to by a demonstrative on one occasion will not in general apply to whatever entity happens to be the referent the next time the demonstrative is used. By contrast, perceptual concepts are suited to serve as repositories of information precisely because they refer to the same thing whenever they are exercised. 2.3 Perceptual semantics I have said that perceptual concepts refer to perceptible entities. However, what exactly determines this relation between perceptual concepts, conceived as stored sensory templates, and their referents? In particular, what determines whether such a concept refers to a type or a token? I suggested earlier that you might look at a bird, form some stored sensory template, and then use it to think either about that particular bird or about its species. But what decides between these two referents? At first pass, it seems that just the same sensory template might be pressed into either service. Some philosophers think of perceptual concepts as ‘recognitional concepts’ (Loar 1990). This terminology suggests that perceptual concepts 4

Cf. Prinz (2002) especially chapters 6 and 7.

91

should be viewed as referring to whichever entities their possessors would recognize as satisfying them. A stored sensory template will refer to just that entity which will activate it when encountered. If none but some particular bird will activate some template, then that particular bird is the referent. If any member of a bird species will activate a template, then the species is the referent. I do not think that this is a good account of perceptual reference. For one thing, it’s not clear that recognitional abilities are fine–grained enough to make the referential distinctions we want. Could not two people have just the same sensory template, and so be disposed to recognize just the same instances, and yet one be thinking about a particular bird, and the other about the species? It is not obvious, to say the least, that my inability to discriminate perceptually between the bird in my garden and its conspecifics means that I must be thinking about the whole species rather than my particular bird; nor, conversely, is it obvious that I must be thinking of my bird rather than its species if I mistakenly take some idiosyncratic marking of my bird to be a characteristic of the species. In any case, the equation of referential value with recognitional range faces the familiar problem that it seems to exclude any possibility of misrecognition: if the referent of my perceptual concept is that entity which includes all the items I recognize as satisfying the concept, then there is no room left for me to misapply the concept perceptually. However, this isn’t what we want — far from guaranteeing infallibility, perceptual concept possession seems consistent with very limited recognitional abilities. I think we will do better to approach reference by focusing on the function of perceptual concepts rather than their actual use. As I explained in the last subsection, the point of perceptual concepts is to accumulate information about certain entities and make it available for future encounters. Given this, we can think of the referential value of a perceptual concept as that entity which it is its function to accumulate information about. Give or take a bit, this will depend on two factors: the origin of the perceptual concept, and the kind of information that gets attached to it. Let me take the second factor first. Note that the kind of information that it is appropriate to carry from one encounter to another will vary, depending on what sort of entity is at issue.5 For example, if I see that some bird has a missing claw, then I should expect this to hold on other encounters with that particular bird, but not across other encounters with members of that species. By contrast, the information that the bird eats seeds is ap5

Here I am very much indebted to Millikan (2000).

92

propriately carried over to other members of the species. The point is that different sorts of information are projectible across encounters with different types of entity. If you are thinking about some metal, you can project melting point from one sample to another, but not the shape of the samples. If you are thinking about some species of shellfish, you can project shape, but not size. If you are thinking about individual humans, you can project ability to speak French, but not shirt colour. And so on. Given this, we can think of the referents of perceptual concepts as determined inter alia by what sort of information the subject is disposed to attach to that concept. If the subject is disposed to attach particular–bird– appropriate information, then the concept refers to a particular bird, while if the subject is disposed to attach bird–species–appropriate information, then reference is to a species. In general, we can suppose that the concept refers to an instance of that kind to which the sort of information accumulated is appropriate. To make this suggestion more graphic, we might think of the templates corresponding to perceptual concepts as being manufactured with a range of ‘slots’ ready to be filled by certain items of information. Thus a particular–bird–concept will have slots for bodily injuries and other visible abnormalities; a particular–person–concept will have slots for languages spoken; a metal–concept will have a slot for melting point; and so on. Which slots are present will then determine which kind of entity is at issue. The actual referent will then generally be whichever instance of that kind was responsible for originating the perceptual concept. As a rule, we can suppose that the purpose of any perceptual concept is to accumulate information about that item (of the relevant kind) that was responsible for its formation. This explains why there is a gap between referential value and recognitional range. I may not be particularly good at recognizing some entity. But if that entity is the source of my concept, then the concept’s function is still to accumulate information about it. Of course, if some perceptual concept comes to be regularly and systematically triggered by some entity other than its original source, and as a result information derived from this new entity comes to eclipse information about the original source, then no doubt the concept should come to be counted as referring to the new entity rather than the original source. But this special case does not undermine the point that a perceptual concept will normally refer to its origin, rather than to whichever entities we happen to recognize as fitting it.

93

Now that I have explained how it is possible for perceptual concepts to refer differentially to both particular tokens and general types, some readers might be wondering how things will work with subjects who have perceptual concepts both for some token and its type — for example, suppose that I have a perceptual concept both for some particular parrot and for its species. To deal with cases like this, we need to think of perceptual concepts as forming structured hierarchies. When someone has perceptual concepts both for a token and its type, the former will add perceptual detail to the latter, so to speak. The same will also apply when subjects have concepts of some determinate (mallard, say) of some determinable type (duck). In line with this, when some more detailed perceptual concept is activated, then so will any more general perceptual concepts which covers it, but not vice versa. Since any items of information that attach to such more general concepts will also apply to the more specific instances, this will work as it should, giving us any generic information about the case at hand along with any case–specific information. Before proceeding, let me make it clear how I am thinking about the relationship between perceptual concepts and conscious perceptual experience. I want to equate conscious perceptual experiences with the activation of perceptual concepts, due either to exogenous stimulation or to endogenous imagination. This does not necessarily mean that any perceptual concepts will be conscious when activated. There may be states that fit the specifications of perceptual concepts given so far, but whose activations are too low–level to constitute conscious states — early stages of visual processing, say. My assumption will only be that there is some range of perceptual concepts whose activations constitute conscious perceptual experience.6 In line with this, I shall restrict the term ‘perceptual experience’ to these cases — that is, I shall use ‘experience’ is a way that implies consciousness. In addition, I shall also assume that the phenomenology of these states goes with the sensory templates involved, independently of what information the subject attaches to those templates or is or is disposed to attach to them. (So if you and I use the same sensory pattern to think

6

It should be noted that my assumption that phenomenology goes with categorization, rather than with basic physical object representation, is denied by Jackendorf (1987) and Prinz (2000). However, I find their arguments uncompelling.

94

about a particular bird and a bird species respectively, the what–it’s– likeness of the resulting experiences will nevertheless be the same.) 2.4 Perceptually derived concepts The discussion so far has assumed that thoughts involving perceptual concepts will require the subject actually to be perceiving or imagining something. In order for the perceptual concept to be deployed, the relevant stored template needs either to be activated by a match with incoming stimuli, or to be autonomously activated in imagination.8 However, now consider this kind of case. You have previously visually encountered some entity — a particular bird, let us suppose — and have formed a perceptual concept of that bird. As before, you exercise this perceptual concept when you perceive further birds as the same bird again, or when you imagine the bird. However, now suppose that you think about the bird when it is not present, and without imaginatively recreating your earlier perception. You simply think ‘That bird must nest near here’, say, without any accompanying perceptual or imaginative act. I take it that such thoughts are possible.9 Having earlier established perceptual contact with some entity, you can subsequently refer to it without the active help of either perception or imagination. I shall say that such references are made via perceptually derived concepts. Here is one way to think about this. Initially your information about some referent was attached to a sensory template. But now you have further created some non–perceptual ‘file’, in which your store of information about that entity is now also housed. This now enables you to think about the entity even when you are not perceiving or imagining it. When you 7

It is consistent with this that there should be phenomenological differences when different patterns are involved within or across subjects when they identify an individual bird (Jemima), or its type (mallard), or indeed both (Jemima and a mallard). 8 Couldn’t you have a standing thought ‘that bird is a female’, say, even when you aren’t actively rehearsing the thought — and won’t this imply a sense in which you can think about the bird even when not perceiving or imagining it? Maybe so. But I am interested here in the involvement of concepts in occurrent thoughts, not in standing ones. 9 Let me be more specific about the possibility at issue here. The idea is not that you might refer non–perceptually using familiar indexical constructions, as in ‘the bird I saw in my garden yesterday morning’. Rather, the non–perceptual reference is supposed to derive more immediately from the prior perceptual contact, in such a way as to remain possible even if you have forgotten where and when you previously perceived the bird and so are unable to identify it indexically.

95

later activate the file, you automatically refer to the same entity as was referred to when the file was originally created.10 Perhaps the ability to create such non–perceptual files is peculiar to linguistic creatures. This is not to say that any such file must correspond to a term in a public language: you can think non–perceptually about things for which you have no name — for example, you may have no name for the bird that that you think nests nearby. Still, in evolutionary terms it seems likely that the ability to think non–perceptually depended on the emergence of language. In this connection, note that an ability to think about things that you have not perceived, and so cannot perceptually recognize or imagine, must play an essential part in mastery of a public language. For public languages are above all mechanisms that allow those who have first–hand acquaintance with certain items of information to share that information with others — which means that those who receive such information will often need to create non–perceptual ‘files’ for entities they have never perceived themselves. By contrast, languageless creatures will have no channels through which to acquire information about items beyond their perceptual ambit, and so no need to represent those items non–perceptually. This provides good reason to suppose that the ability to create non–perceptual ‘files’ arrived only with the emergence of language. If this is right, then only language–using human beings will be able to transcend perceptual concepts proper by constructing what I am calling ‘perceptually derived concepts’. Of course, as noted at the beginning of this paragraph, humans will also sometimes use this ability to cre10

Those who think in terms of ‘recognitional concepts’ may be inclined to argue that perceptually derived concepts, along with the perceptual concepts they derive from, require an ability to recognize referents perceptually. Against this, I have already argued that even ordinary perceptual concepts do not require their possessors to be any good at recognizing referents. With perceptually derived concepts I would say that recognitional powers can atrophy still further, even to the extent where there is no disposition perceptually to identify anything as the referent of your concept. Suppose that you previously referred to something perceptually. But your stored sensory template has faded, and you can no longer perceptually reidentify new instances when you come across them, and perhaps you can’t even perceptually imagine them. It doesn’t seem to me that this need stop you being able to rehearse your belief ‘That bird was female’, say, where the underlined phrase still refers to the original referent. Anaphora is perhaps a useful model here. Consider someone who first thinks about some entity perceptually, and then keeps coming back to it in thought, even after the ability to reidentify the entity has faded. It seems natural to suppose that these thoughts will continue to refer to the same entity, even in the absence of continued recognitional abilities.

96

ate non–perceptual ‘files’ that correspond to no word in a public language. But, still, when they do so, they may well be drawing on an ability that evolved only along with linguistic capacities. Perhaps there is an issue about counting concepts here. I have distinguished between ‘perceptual concepts’ and ‘perceptually derived concepts’. Do I therefore want to say that a thinker who has constructed a ‘perceptually derived concept’ from a prior ‘perceptual concept’ now has two concepts that refer to the same thing? From some perspectives, this might seem like double counting. In particular, it is not clear that the standard Fregean criterion of cognitive significance will tell us that there are two concepts here. After all, if the creation of a ‘perceptually derived concept’ is simply a matter of housing your store of information in a non– perceptual file, as I put it above, and if any subsequently acquired information about the relevant referent automatically gets attached to both sensory template and non–perceptual tag, then it seems that the subject will always make exactly the same judgements whether using the ‘perceptual’ or ‘perceptually derived’ concept, and so fail the Frege test for possession of distinct concepts. And this would suggest that we simply have one concept here, not two, albeit a concept which can be exercised in two ways — perceptually and non–perceptually.11 There is no substantial issue here. To the extent that the flow of information between the two ways of thinking is smooth, the Frege test gives us reason to say that there is only one concept. On the other hand, to the extent that there are cognitive operations that distinguish a perceptually derived concept from its originating perceptual concept, there is a rationale for speaking of two concepts, and I shall do so when this is convenient. 3. Phenomenal concepts 3.1 The quotational–indexical model Let me now turn to phenomenal concepts. My earlier ‘quotational– indexical’ model, recall, viewed phenomenal concepts as having the structure the experience: —, where the gap was filled either by an actual perceptual experience or by an imaginative recreation thereof. It now seems to me that this ‘quotational–indexical’ model ran together a good idea with a bad one. The good idea was to relate phenomenal concepts to perceptual concepts. The bad idea was to think that phenomenal concepts, along with perceptual ones, are some kind of ‘demonstrative’. 11

I would like to thank Dorothy Edgington for drawing my attention to this issue.

97

Let me first explain the bad idea. Suppose that perceptual concepts were demonstrative, contrary to my earlier arguments. Then presumably they would be constructions that, on each occasion of use, referred to whichever item in the external environment was somehow salient to the subject. By analogy, if phenomenal concepts worked similarly, then they too would refer to salient items, but now in the ‘internal’ conscious environment. This thus led me to the idea that phenomenal concepts were somehow akin to the mixed demonstrative construction that experience. On this model, phenomenal concepts would employ the same general demonstrative construction (that) as is employed by ordinary mixed demonstratives, but the qualifier experience would function to direct reference inwards, so to speak, ensuring that some salient element in the conscious realm is picked out. The ‘quotational’ suggestion then depended on the fact that this demonstrated experience would itself be present in the realm of conscious thought, unlike the non–mental items referred to by most demonstratives. This made it seem natural to view phenomenal concepts as ‘quoting’ their referents, rather than simply referring to distal items. Linguistic quotation marks, after all, are a species of demonstrative construction: a use of quotation marks will refer to that word, whatever it is, that happens to be made salient by being placed within the quotation marks. Similarly, I thought, phenomenal concepts can usefully be thought of as referring to that experience, whatever it is, that is currently made salient in thought. However, this now seems to me all wrong. Not only is it motivated by a mistaken view of perceptual concepts, but it runs into awkward objections about the nature of the notion of experience used to form the putative construction that experience. There seem two possible models for the concept of experience employed here. It might be abstracted from more specific phenomenal concepts (seeing something red, smelling roses, and so on); alternatively, it could be some kind of theoretical concept, constituted by its role in some theory of experiences. However, neither option seems acceptable. The obvious objection to the abstraction strategy is that it presupposes such specific phenomenal concepts as seeing something red, smelling roses, and so on, when it is supposed to explain them. If we are to acquire a generic concept of experience via first thinking phenomenally about more specific experiences, and then abstracting a concept of what they have in common, then it must be possible to think phenomenally about the more specific experiences prior to developing the generic con-

98

cept. But if thinking phenomenally about the more specific experiences requires us already to have the generic concept, as on the demonstrative account of phenomenal concepts, then we are caught in a circle. What if our notion of experience is constituted by its role in some theory of experiences (our folk psychological theory perhaps)? Given such a theoretically defined generic concept of experience, there would be no barrier to then combining it with a general–purpose ‘that’ to form demonstrative concepts of specific experiences. Since the generic concept wouldn’t be derived by abstraction from prior phenomenal concepts of specific experiences, there would be no circle in using it to form such specific phenomenal concepts. This picture may be cogent in principle, but it seems to be belied by the nature of our actual phenomenal concepts. If a generic concept of ‘experience’ were drawn from something like folk psychological theory, then we could expect it to involve some commitment to the assumption that experiences are internal causes of behaviour. Folk psychology surely conceives of experiences inter alia as internal states with characteristic causes and behavioural effects. But then it would seem to follow that anything demonstrated as that experience, where experience is the folk psychological concept, must analytically have some behavioural effects. It needn’t be analytic which specific behavioural effects that experience has — you could know that all experiences have characteristic effects without knowing what specific effects that experience has — but still, it would be analytic that that experience had some behavioural effects. However, this doesn’t seem the right thing to say about phenomenal concepts. There is surely nothing immediately contradictory in the idea that an experience picked out by some phenomenal concept has no subsequent effects on behaviour or anything else. Epiphenomenalism about phenomenal states doesn’t seem to be a priori contradictory.12 Yet it would be, if our ways of referring to phenomenal states analytically implied that they had behavioural effects. 3.2 Phenomenal concepts as perceptual concepts I said above that my old model of phenomenal concepts ran together the good idea that phenomenal concepts are related to perceptual concepts with the bad idea that both kinds of concepts are ‘demonstratives’. Let me now try to develop the good idea unencumbered by the bad one. 12

This leaves it open, of course, that there may be other good arguments against epiphenomenalism, apart from a priori arguments. Cf. Papineau (2002), sect. 1.4.

99

My current view is that phenomenal concepts are simply special cases of perceptual concepts. Consider once more the example where I perceptually identify some bird and make some judgement about it (THAT is a migrant). I earlier explained how the perceptual concept employed here could either be a concept of an individual bird or the concept of a species. I want now to suggest that we think of phenomenal concepts as simply a further deployment of the same sensory templates, but now being used to think about perceptual experiences themselves, rather than about the objects of those experiences. I see a bird, or visually imagine a bird, but now I think, not about that bird or species, but about the experience, the conscious awareness of a bird.13 The obvious question is — what makes it the case that I am here thinking about an experience, rather than an individual bird or a species? However, we can give the same answer here as before. I earlier explained how the subject’s dispositions to carry information from one encounter to another can decide whether a given sensory template is referring to an individual rather than a species, or vice versa — if the subject projects species–appropriate information, reference is to a species, while if the subject projects individual–appropriate information, reference is to an individual. So let us apply the same idea once more — if the subject is disposed to project experience–appropriate information from one encounter to another, then the sensory template in question is being used to think about an experience. For example, suppose I am disposed to project, from one encounter to another, such facts as that what I am encountering ceases when I close my eyes, goes fuzzy when I am tired, will be more detailed if I go closer, and so on. If this is how I am using the template as a repository of information, then I will be referring to the visual experience of seeing the bird, rather than the bird itself. More generally, if they are used in this kind 13

Does this mean that perceptual experiences are the only items that can be thought about phenomenally? This seems doubtful. To consider just a couple of further cases, what about emotions, and pains? At first pass, it certainly seems that these states too can be picked out by phenomenal concepts — yet they are not obviously examples of perceptual experiences. There are two ways to go here. One would be to understand perception in a broad enough way to include such states. After all, emotions and pains are arguably representational states, and so could on these grounds be held to be a species of perception. Alternatively, we might distinguish these states from perceptions, but nevertheless allow that they are similar enough for us to think about them in ways that parallels phenomenal thinking about obviously perceptual states. I have no strong views on this choice, but in what follows I shall simplify the exposition by sticking to perceptions.

100

of way — to gather experience–appropriate information, so to speak — the same sensory templates that are normally used to think about perceptible things will refer to experiences themselves. Can phenomenal concepts pick out experiential particulars as well as types? In the perceptual case, as we have seen, there is room for such differential reference to both particular objects and to types, due to the possibility of differing dispositions to carry information from one encounter to another. In principle it may seem that the same sort of thing could work in the phenomenal case. The trouble, however, is that particular experiences, by contrast with ordinary spatio–temporal particulars, do not seem to persist over time in the way required for re–encounters to be possible. Can the same particular pain, or particular visual sensation, or particular feeling of lassitude, re–occur after ceasing to be phenomenally present? It is true that we often say things like ‘Oh dear, there’s that pain again — I thought I was rid of it’. But nothing demands that we read such remarks as about quantitative rather than qualitative identity: nothing forces us to understand them as saying that the same particular experience has re–emerged, as it were, rather than that the same experiential type has been re–instantiated (note in particular that experiences do not seem to allow anything analogous to the spatio–temporal tracking of ordinary physical objects). In line with this, note that information about experiences, as opposed to information about spatio–temporal particulars, does not seem to divide into items that are projectible across encounters with a particular and items that are projectible across encounters with a type. Given all this, I am inclined to say that phenomenal concepts cannot refer differentially both to particulars and to types. Rather they always refer to types—that is, to the kind of mental item that can clearly re–occur. As I am conceiving of perceptual and phenomenal concepts, the function of a concept is to carry information about its referent from one encounter to another — and it seems that only phenomenal types and not particulars can be re–countered. The corollary is that, when we do refer to particular experiences, we cannot be using our basic apparatus of phenomenal concepts, given that these are only capable of referring to phenomenal types. Rather, we must be invoking more sophisticated conceptual powers, such as the ability to refer by description (thus the particular pain I am having now, or the particular experience of crimson I enjoyed at last night’s sunset).

101

3.3 Phenomenal use and mention This model of phenomenal concepts as a species of perceptual concept retains one crucial feature from my earlier quotational–indexical model, namely, that phenomenal references to an experience will involve an instance of that experience, and in this sense will use that experience in order to mention it. To see why, think about what happens when a phenomenal concept is exercised. Some sensory template is activated, and is used to think about an experience. This sensory activation will either be due to externally generated sensory stimuli or to autonomous imaginative activity. That is, you will either be perceiving the environment, or employing perceptual imagination. For example, either you will be perceiving a bird, or you will be perceptually imagining one. Except, when phenomenal thought is involved, this template is also used to think about perceptual experience, rather than just about the objects of perceptions. You look at a bird, or visually imagine that bird, but now use the sensory state to think about the visual experience of seeing the bird, and not only about the bird itself. This means that that any exercise of a phenomenal concept to think about a perceptual experience will inevitably either involve that experience itself or an imaginary recreation of that experience. If we count imaginary recreations as ‘versions’ of the experience being imagined, then we can say that phenomenal thinking about a given experience will always use a version of that experience in order to mention that experience. Note how this model accounts for the oft–remarked ‘transparency of experience’ (Harman 1990). If we try to focus our minds on the nature of our conscious experiences, all that happens is that we focus harder on the objects of those experiences. I try to concentrate on my visual experience of the bird, but all that happens is that I look harder at the bird itself. Now, there is much debate about exactly what this implies for the nature of conscious experience (cf. Stoljar, forthcoming). But we can by–pass this debate here, and simply attend to the basic phenomenon, which I take to be the phenomenological equivalence of (a) thinking phenomenally about an experience and (b) thinking perceptually with that experience. What it’s like to focus phenomenally on your visual experience of the bird is no different from what it’s like to see the bird. On my model of phenomenal thinking, this is just what we should expect. I said at the end of the section 2.3 that the phenomenology of perceptual experiences is determined by which sensory template they involve,

102

and not by what information they carry with them. I have now argued that just the same sensory templates underly both perceptual experiences and phenomenal thoughts about those experiences. It follows that perceptual experiences and phenomenal thoughts about them will have just the same phenomenology. This explains why thinking phenomenally about your visual experience of a bird feels no different from thinking perceptually about the bird itself. 3.4 A surprising implication The story I have told so far has an implication that some might find surprising. On my account, the semantic powers of phenomenal concepts would seem to depend on their cognitive function, rather than their phenomenal nature. I have argued that phenomenal concepts refer to conscious experiences because it is their purpose to accumulate information about those experiences. As it happens, exercises of such concepts will in part be constituted by versions of the conscious experiences they refer to, and so will share the ‘what–it’s–likeness’ of those experiences. But this latter, phenomenal fact seems to play no essential role in the semantic workings of phenomenal concepts. To see this, suppose that we had evolved to attach information about conscious experiences to states other than sensory templates — to words in some language of thought, perhaps. Wouldn’t these states refer equally to experiences, and for just the same reason, even though their activation did not share the phenomenology of their referents? However, this might seem in tension with the idea that phenomenal concepts involve some distinctive mode of phenomenal self–reference to experiences. If the phenomenality of phenomenal concepts is incidental to their referential powers, then in what sense are they distinctively phenomenal? (Cf. Block, forthcoming.) Note that my earlier ‘quotational–indexical’ account of phenomenal concepts is not open to this kind of worry. On that account, phenomenal concepts used experiences as exemplars, rather than as ways of implementing a cognitive role. Given this, it is essential to the phenomenal concept of seeing something red, say, that it ‘quotes’ some version of that experience, just as it is essential to the quotational referring expression ‘ “zymurgy” ’ that it contain the last word in the English dictionary within its quotation marks. On the quotational–indexical account then, there is no question of some state referring to an experience in the same way as a phenomenal concept does, yet its exercise not involve the experience.

103

Note also that the worrisome implication is not peculiar to the particular theory of the semantics of phenomenal concepts I have defended in this paper. It will arise on any theory that makes the semantic powers of phenomenal concepts purely a matter of their conceptual role, or their informational links to the external world, or any other facet of their causal– historical workings. For any theory of this kind will make it incidental to the referential powers of phenomenal concepts that they have the same phenomenology as their referents. Any such theory leaves it open that some other states, with different or no phenomenology, could have the same causal–historical features, and so refer to experiences for the same reason that phenomenal concepts do. My response to this worry is that there is no real problem here. On my account, it is indeed true that phenomenal concepts refer because of their cognitive function, not because of their phenomenology, and therefore that other states with different or no phenomenology, but with the same cognitive function, would refer to the same experiences for the same reasons. I see nothing wrong with this. Of course, it is a further question whether we would wish to include any such non–phenomenological states within the category of ‘phenomenal’ concepts, given their lack of what– it’s–likeness (cf. Tye 2003). But this is no grounds for denying that they would refer to experiences for just the same reason as phenomenal concepts do. I shall come back to the issue of what counts as a ‘phenomenal’ concept in the next section. But first let me ask a somewhat different question. Given that other items could in principle play the cognitive role that determines reference to experiences, why do we use experiences themselves for this purpose? What is it about conscious experiences that makes them such a good vehicle for referring to themselves? One possible answer is that this use of experiences is somehow well– suited to answering certain questions. To adapt an example of Michael Tye’s (2003, p. 102), suppose that we are wondering whether the England one–day cricket strip is visually darker than the Indian one. By thinking phenomenally about these colours, we will generate versions of the relevant experiences, and so be in a position to compare them directly. This makes some sense, but I think a simpler answer may be possible. Consider the analogous question: why do we use perceptual experience to represent perceptible items such as people, physical objects, animals, plants, shapes, colours, and so on? After all, in this case too the referential powers of these states are presumably determined by some type of

104

cognitive role, which could in principle have been played by something other than perceptual experiences themselves. Here the obvious answer seems to be that the perceptions are especially good for thinking about perceptible entities simply because they are characteristically activated by those entities, and so are well–suited to feature in judgements that those entities are present. It would unnecessarily duplicate cognitive mechanisms to use the perceptual system to identify perceptible entities, yet something other than perceptual experiences as the vehicle for occurrent thoughts that imply that those entities are present. This thought applies all the more in the phenomenal case. Conscious experiences are excellent vehicles for thinking about those selfsame experiences, simply because they are automatically present whenever their referents are. The fact that we use experiences to think about themselves means that we don’t have to find other cognitive resources to frame occurrent thoughts about the presence of experiences. 3.5 Phenomenally derived concepts I shall conclude this paper with some comments on the terminology of ‘phemonenal concepts’, and in particular on the question of exactly what qualifies a concept as ‘phenomenal’. As well as being of some intrinsic interest, this issue is relevant to the worry about phenomenal concepts raised at the beginning of the paper — namely, that someone might use a phenomenal concept of some type of experience to truly think I am not now having that experience (nor recreating it in my imagination). Following the discussion in the last section, let us take it as a basic requirement for the ‘phenomenality’ of a concept that it refer directly to a type of conscious experience in virtue of some aspect of its causal– historical workings. In this respect, phenomenal concepts will be akin to other direct perceptual concepts, and will be distinguished only by the fact that they refer to conscious experiences. We might call this the basic ‘semantic’ notion of a phenomenal concept. I shall consider two further conditions that might intuitively be required for the ‘phenomenality’ of a concept of experience. The first is the ‘use–mention’ feature discussed in section 3.3 above: this requires that exercises of the relevant concept use (a version of) the experience that they mention. The other condition I shall call ‘experience–dependence’: this requires that the acquisition of the concept depends on the thinker having previously undergone the experience it refers to.

105

Experience–dependence in this sense is illustrated by the ‘Mary’ thought–experiment. The story of Mary is built on the assumption that the relevant concept of red experience will only become available to Mary once she has herself seen something red. I would say that the reason for this experience–dependence is that Mary’s new concept demands the formation of a sensory template, and that the acquisition of this template depends on her visual system previously having been activated by some red surface. Note that, if this is right, the experience–dependence of the relevant concepts is a contingent feature of human beings. We can imagine beings who are born with the sensory templates that we only acquire as a result of prior colour experiences (cf. Papineau 2002 sect. 2.8). Still, as it happens, we humans are not like this. We are born with few, if any, sensory templates, but must rather acquire them from previous experiences.14 So now we have three possible requirements for ‘phenomenality’ — the basic semantic requirement, the use–mention requirement, and the experience–dependence requirement. At first sight it might seem that there is no great need for terminological adjudication here. It is true that in principle the requirements are dissociable in various ways. In the last section I observed that there is no absolute reason why items that satisfy the basic semantic requirement should display the use–mention feature, and I have just observed that items that display the use–mention feature could in principle fail to be experience–dependent. Still, it looks as if the three requirements always line up together for normal human beings — the only items that actually play the basic semantic role are experience–dependent concepts whose applications use the items they mention — and that to this extent nothing much will go wrong if we use ‘phenomenal’ indiscriminately for all three. However, there is one kind of real–world case where the requirements don’t all line up together, and which does call for some terminological precision. I am thinking here of concepts that satisfy the semantic and experience–dependence requirements, but not the use–mention one. This 14

If humans were born with the sensory templates that are in fact only activated by red surfaces, then physicalists could not answer the knowledge argument by saying that Mary needs a red experience in order to acquire the relevant concept of red. But if humans were born with the relevant templates, then physicalists wouldn’t need to answer the knowledge argument in the first place, since Mary would already have the relevant concept of red before she left her room, and so would already be in a position to know that scientifically described red experience is the same as THAT experience, where the italicized phrase expresses an imaginative exercise of the concept at issue.

106

kind of dissociation is not just an abstract possibility. Even though paradigm examples of phenomenal concepts, like the one Mary acquires on leaving her room, do display the use–mention feature along with the other two, there are other kinds of concepts which are phenomenal in the semantic and experience–dependent senses but which don’t use experiences to mention themselves. To see why, recall the earlier discussion of perceptually derived concepts. These derived concepts involved the creation of some non–sensory file to house the information associated with some perceptual concept, and they made it possible to think about perceptible entities even when those entities were not being perceived or perceptually imagined. Analogously, we can posit a species of ‘phenomenally derived concept’. Suppose someone starts off, like Mary, by thinking phenomenally using a sensory template instilled by previous experiences. But then she creates a non–sensory file in which to house the information that has become attached to that template, and which will henceforth allow her to think about the experience without any sensory activation. I say she now has a phenomenally derived concept. Exercises of this concept won’t activate the experience it mentions, and so this concept will fail to satisfy the use–mention requirement. But note that this phenomenally derived concept will still satisfy the experience–dependence requirement, in that its creation will depend on a prior phenomenal concept which will in turn depend on previous experiences. The possibility of phenomenally derived concepts offers an answer to the objection first raised in the Introduction. This was that standard accounts of phenomenal concepts seem to imply that any exercise of a phenomenal concept demands the presence of the experience it refers to or an imaginatively recreated exemplar thereof. However, this seemed too demanding. Surely someone like Mary can use her new concept to think truly that I am not now having THAT experience (nor recreating it in my imagination). Yet this should be impossible, if any exercise of her phenomenal concept does indeed require the relevant experience or its imaginative recreation. We are now is a position to respond to this objection: Mary thinks the problematic thought with the help of a phenomenally derived concept.15 15

It might occur to some readers that another answer to the challenge would be to insist that Mary must really be using some old pre–exposure material concept of red experience when she thinks the supposedly problematic thought. However, this will not serve. To see clearly why, it will help to vary the Mary thought–experiment slightly.

107

She starts with a phenomenal concept based on some sensory template, and then creates a non–sensory file to carry the information associated with the template. This allows her to think about the relevant experience without activating the associated sensory template — that is, without either having or imaginatively recreating the experience in question. She thinks I am not now having or imagining THAT experience — and since she is using a phenomenally derived concept, what she thinks can well be true. Let us briefly go back to the terminological issue. By this stage some readers may be wanting to reconsider the propriety of saying that derived non–sensory files constitute phenomenal concepts. After all, if exercises of such ‘phenomenally derived’ concepts don’t activate the experience in question, then are they really any more ‘phenomenal’ that the general run of ordinary concepts? Well, I have no principled objection if someone wants to withhold the description ‘phenomenal’ on these grounds. I don’t think that there is any substantial issue here. We have clearly distinguished various different cases — beyond that it is just terminological which ones we call ‘phenomenal’. It is true that, if phenomenally derived concepts don’t count as ‘phenomenal’ concepts, then we can scarcely appeal to them to explain how someone can use a phenomenal concept to think truly that I am not now having or imagining THAT experience. But that doesn’t matter. If phenomenally derived concepts don’t count as ‘phenomenal’, then the objection won’t arise in the first place, for the objection was precisely that a thinker can exercise a phenomenal concept while not having any version of the experience referred to, not just that a thinker can exercise any old concept of experience in such circumstances. For this objection to make any sense, ‘phenomenal’ cannot be understood as requiring the use–mention feature, for that is precisely what is absent in the supposedly problematic example. Rather, ‘phenomenal’ must presumably be understood as standing for those concepts whose acquisition depends on undergoing the releSuppose that, on her exposure, Mary was shown a coloured piece of paper, rather than a rose, and that she wasn’t told what colour it was. The objection would seem still to stand. The conceptual powers she acquires from her exposure would still seem to enable her later truly to think, I am not now having or imagining THAT experience. But now she can’t be using any of her old pre–exposure concepts to refer to the experience. For, if she doesn’t know what colour the paper was, she won’t know which of her old concepts to use.

108

vant experience. And in this sense of ‘phenomenal’ — experience– dependence — phenomenally derived concepts do explain how someone can think phenomenally without having any version of the corresponding experience.

References Alter, Torin and Walter, Sven (eds) 2006: Phenomenal Concepts and Phenomenal Knowlege: New Essays on Consciousness and Physicalism. New York: Oxford University Press. Balog, Katalin, Forthcoming: ‘The “Quotational” Account of Phenomenal Concepts’. Block, Ned, Forthcoming: ‘Max Black’s Objection to Mind–Body Identity’. Chalmers, David 2003: ‘The Content and Epistemology of Phenomenal Belief’. In: Smith, Quentin and Jokic, Aleksandar (eds.) Consciousness: New Philosophical Perspectives. Oxford: Oxford University Press, pp. 220-272. Crane, Tim, Forthcoming: ‘Papineau on Phenomenal Concepts’. Philosophy and Phenomenological Research. Harman, Gilbert 1990: ‘The Intrinsic Quality of Experience’. In: Tomberlin, James (ed.) Philosophical Perspectives: Action Theory and Philosophy of Mind, 4, pp. 31–52. Horgan, Terence 1984: ‘Jackson on Physical Information and Qualia’. Philosophical Quarterly, 34, pp. 147–152. Jackendorf, Ray 1987: Consciousness and the Computational Mind. Cambridge, Mass: MIT Press. Jackson, Frank 1986: ‘What Mary Didn’t Know’. Journal of Philosophy, 83, pp. 291–295. Loar, Brian 1990: ‘Phenomenal States’. In: Tomberlin, James (ed.) Philosophical Perspectives: Action Theory and Philosophy of Mind, 4, pp. 81–108. Millikan, Ruth 1990: ‘The Myth of the Essential Indexical’. Nous, 24, pp. 723–34. — 2000: On Clear and Confused Ideas. Cambridge: Cambridge University Press. Papineau, D. 1993. Philosophical Naturalism. Oxford: Basil Blackwell. — 2002: Thinking about Consciousness. Oxford: Oxford University Press. — 2007: ‘Phenomenal and Perceptual Concepts’. In: Alter, Torin and Walter, Sven (eds.) 2006: Phenomenal Concepts and Phenomenal Knowlege: New Essays on Consciousness and Physicalism, pp. 111-145. Perry, John 1979: ‘The Problem of the Essential Indexical’. Nous, 13, pp. 3–12. — 2001 Knowledge, Possibility and Consciousness. Cambridge, Mass: MIT Press. Prinz, Jesse 2000: ‘A Neurofunctional Theory of Visual Consciousness’, Consciousness and Cognition, 9, pp. 243–59. — 2002: Furnishing the Mind. Cambridge, Mass: MIT Press. Smith, Quentin and Jokic, Aleksandar (eds.): Consciousness: New Philosophical Perspectives. Oxford: Oxford University Press 2003.

109

Stainton, Robert and Escurdia, Maite and Viger, Chris (eds.): New Essays in Philosophy of Language and Mind. Canadian Journal of Philosophy, Supplementary Volume. Stoljar, Daniel, Forthcoming: ‘The Argument from Diaphanousness’. In: Stainton, Robert, M. Escurdia, Maita and Viger, Chris (eds.), New Essays in Philosophy of Language and Mind. Tye, Michael 1995. Ten Problems of Consciousness. Cambridge, Mass: MIT Press. — 2003. ‘A Theory of Phenomenal Concepts’. In: O’Hear, Anthony (ed.): Minds and Persons. Cambridge: Cambridge University Press, pp. 91-107.

110

Kant on the a priori content of perceptual experience Anthony Savile

1. Kant’s route to transcendental idealism passes through reflections that purport to show how perceptual experience of the world is pervaded by those pure concepts of the understanding he calls ‘categories’. These concepts are at once both non-empirical or a priori and necessary, necessary in that, dependent as we are on understanding in our grasp of the world, we have no way of experiencing it but by drawing on them. And then given their a priori origin – which it is the task of the Metaphysical Deduction to establish – the inevitable contribution they make to the representations we construct in synthesising sensory inputs ensures that the world we thereby perceive is the world of appearance, and not the world as it is in itself. So much is common knowledge. What is less well known is just how Kant thinks these claims can be made out, less well known, I think, because the short passage that ostensibly does the work, and which English translations of the Critique call ‘the clue to the discovery of all the pure concepts of the understanding’, is so very condensed. Here I offer a tentative explication de texte of the ‘Clue’s’ core paragraphs and towards the end propose a loose running paraphrase of them. To complete matters, I suggest that the resulting understanding of this passage may help resolve a standing puzzle about Kant’s notion of objective validity. Before broaching the text itself though, a word is in order about the way in which Kant himself must have viewed his Leitfaden, his ‘clue’. A literal translation of that German term is ‘guiding thread’, and that suggests something quite other than does the English word ‘clue’. Aridane’s thread led Theseus out of the labyrinth and provided it was properly followed could not have taken her hero anywhere other than to its mouth. By contrast, clues may mislead, either by pointing one in the wrong direction or by inviting misinterpretation. So if we are to understand Kant’s own thought we should resist saying, maybe on his behalf and in response to dissatisfaction with what he has to offer in this dense passage, that the clue

is just that, merely a clue.1 It is not. It is a guiding thread if it is anything. It is with this thought in mind that we should confront the text. 2. Text (CPR A78-9/B104-5): paragraph 1 Pure synthesis, represented in its most general aspect, gives us the pure concept of the understanding. By this pure synthesis I understand that which rests on a basis of a priori synthetic unity. Thus our counting, as is easily seen in the case of larger numbers, is a synthesis according to concepts, because it is executed according to common ground of unity, as, for instance, the decade. In terms of this concept, the unity of the synthesis is rendered necessary.

Four sentences: first, the claim to be established, that by reflection on what Kant calls ‘pure synthesis’ we shall be led to the pure concept of the understanding. That last term appears notably in the singular, so we should not expect to be taken straightway to the ultimately targeted twelve categories. Next, a sketchy elucidation of the term ‘pure synthesis’. Then, an illustration of how arithmetical concepts serve to give unity to a manifold in such a pure synthesis. Finally, as if by way of an aside, the observation that the employment of such concepts renders the unity in question necessary. The assumed background to all this is that perceptual experience, which Kant sometimes simply calls ‘knowledge’, should be seen as output of the mind’s processing system given appropriate sensible input. Synthesis is just the operation of combining the various sensible elements which affect us sequentially and which in its knowledge-yielding function generates representations of the world that are apt to reveal it to us, assessable as true to it or not as the case may be. Against this background, each member of the initial foursome invites brief comment. The first points forward, telling us that if we reflect on the way the mind works to achieve fully-fledged perceptual representations we shall see that it must draw on something non-empirical in order to do so. This reflection alone, Kant promises, will enable us to identify the understanding’s a priori contribution to experience. Moreover, since we are concerned with the mind’s operation of synthesis taken ‘in its most general aspect’, what he must be aiming for is some quite general truth about experience. It must be that whenever we have experience that we rate as complete, adequately unified and apt for assessment for accuracy, it can be shown to owe its unity to the operation of some unspecified a priori concept or other. We will have brought together the various sensory strands 1

Cf. McDowell’s discussion of the ‘Clue’, 1998, p. 465.

112

provided in intuition according to some a priori base, and since this is meant to be a thought that applies to the mind’s achievement of replete experience quite generally, the idea seems to be that there must be one such concept that is omnipresent and constantly at work in the construction of perceptual representation no matter what shape that takes in the particular case. Concentrating on the most general aspect of perceptual experience is, I surmise, what generates the initially surprising singularity of ‘the pure concept of the understanding’.2 If in the first sentence ‘pure synthesis’ may sound obscure, as if it were somehow to be understood in contrast to synthesis of an empirical kind, some light is shed on it in the next. Pure synthesis is just such synthesis as rests on an a priori base or is conducted in accordance with an a priori concept or rule. That may, of course, make the first sentence look entirely trivial, for in the light of the explanation offered it could seem as if we shall discover that when we reflect on synthesis which is conducted in accordance with an a priori concept we shall come across a priori concepts in the synthetic construction that we arrive at. However, what Kant has in mind is to affirm that we could not have replete experience unless we achieved it by pure synthesis, on the basis of some a priori element, that is, and this, I take it, is what he takes himself to be illustrating in a preliminary way by the arithmetical example we are given in the next sentence. Pure synthesis then is any synthesis of experiential intake that is apt to reveal the world to us, and it is always pure just because it inevitably and invariably relies on some a priori concept at its heart, something that Kant now proposes to show. At the very least we can say that his illustration exemplifies how in a particular sort of case our experience acquires the kind of unity it has from being constructed with a concept answering the question: How many?, a 2

One might suppose that the singular as it occurs both here and in the last paragraph I discuss is just a generic singular as are ‘the man’ and ‘family’ in such sentences as ‘In today’s family it is no longer automatically the man who wears the trousers’. If that were all there was to it Kant’s route to the categories would be more assertoric than argumentative, and the guiding thread pretty much otiose. Here, I take it, the singular ‘the pure concept of the understanding’ is indeed generic in so far as it points to the ultimately discoverable twelve categories, but it must also be seen as particular and as picking out the ubiquitous presence in experience of that single concept that Kant thinks will guide us to them. Only if it is so shall we understand the illustrative use he goes on to make of the decade, and later on, see just why he says that the pure concept of the transcendental object is ‘in reality throughout all our knowledge ... always one and the same’ (A109).

113

decade, to take Kant’s own example. Let us suppose I am counting my flock of goats as it comes in at dusk from pasture: I see there to be thirty of them. To do this Kant thinks that as I count the animals I need to make the past sightings of earlier goats that pass the gate present, and that I can hardly do this other than by drawing on a counting system that records and retains the past as it moves forward from one element to the next.3 It could not be achieved, for example, by a simple mental tally, marking ‘x’ each time a goat passes by me, and achieved only with difficulty for a larger herd without some systematic way of continuing the series in an openended way. (I would find it difficult enough if I had learnt to go ‘... eight, nine, ten, nyolcs, hétven, patranc, zontig’ and onwards for a while until my patience and ingenuity ran out, and downright impossible should I find myself possessed of more goats than numerals supplied by my patiently acquired rote.) One can see then how the use of a concept like the decade provides a unity for the construction of experience in this sort of quantitative case. Without some such base we could not find anything in our experience of those goats that would systematically give us an answer to the quantitative question. And of course the same would be true if we were interested not in numbers but in matters of distance. Today’s horses run races over miles and furlongs to win their trophies, but it might one day become handy to race hardier beasts over even longer courses. Then we might see a point to introducing a novel base, say one of some twenty furlongs, and giving that measure of distance a name, a lacey, perhaps, and then race the sturdier animals over three or four laceys. Using this novel concept would allow us to articulate a whole, complete and adequately unified experience to the effect, say, that the last lacey was hard going, or that the distance from Bruges to Ghent is just sixteen of them. And so on. It can hardly be denied that Kant’s example is less than perspicuous. If it is to help us understand why pure synthesis must be at work in such cases, the decade or the lacey needs to be a non-empirical a priori concept. 3

‘When I seek to draw a line in thought, or to think of the time from one noon to another, or even to represent to myself some particular number, obviously the various manifold representations that are involved must be apprehended by me in thought one after another. But if I were always to drop out of thought the preceding representations (the first part of the line, the antecedent parts of the time period, or the units in the order represented), and did not reproduce them when advancing to those that follow, a complete representation would never be obtained: none of the above mentioned thoughts, not even the purest and most elementary representations of space and time, could arise’ (A 102).

114

Yet it is scarcely compelling to suppose that either of them is that – though it could be said, eirenically enough, that the intimate connection there is between counting, mensuration and spatio-temporality would certainly have encouraged Kant to view them as a priori. Perhaps, though, we should not ask too much of an illustration, and be ready to take away from it the insight that any experience we have needs to be constructed on a basis that provides the manifold input with a kind of unity, and can only do that by drawing on some conceptual base that extends beyond what is given in the sensory manifold itself as it is enjoyed from moment to moment. If it is to make our experience apt to answer questions of quantity or distance, the concept that does this work must amplify the intuitive input to the system, and could colourably be thought of by Kant as a priori on that ground alone. About this I shall not speculate further and merely observe that some such analogous thought is what appears to motivate the Transcendental Aesthetic’s matching claim about the a priori nature of the concepts space and time. The last sentence of the paragraph says that once we construct experience in terms of the decade (or in our new case, the lacey) the resulting unity of the manifold is made necessary. There is a stronger and a weaker way of taking this, neither of which does anything to impugn the contingent nature of what our perceptual life reveals to us. The stronger of these is to suppose that given a determinate sensory input to the perceptual system and the understanding’s disciplined deployment of the unifying decade or lacey concept, my experience of the goats cannot but be as of thirty of them and my experience of the distance covered in the race cannot but be as of three laceys. For a well-functioning perceptual system, the choice of the unifying concept determines how in the particular case the construction of experience has to go. Given this input and this choice of unifying base, the decade, the lacey, the constructive answer to the How-many? question must be thirty, and to the How-far? one, three. The weaker, to my mind more strained, reading of the claim is merely that some such intuitionamplifying a priori concept as the decade is necessary for my experience to enjoy full experiential unity, while remaining silent about the mandatory nature of the experienced outcome. I shall return to this issue below, remarking now only that the weaker reading of the sentence muffles its exciting ring and effectively has it adding nothing to what is already implicit in the ones that precede it, recording merely that the function of these a priori concepts is to provide the manifold with that unity which renders it apt to represent the world truly.

115

What then should we take away from this initial paragraph? First, that the idea of anything like replete perceptual experience depends on its being constructed in accordance with some pervasive, constantly recurring but as yet unidentified a priori concept. Secondly, in a way that is given no immediate elucidation, that the resulting experiences, viz. absolutely any replete experience, enjoy a unity that Kant is ready to qualify as necessary. So we can move on to the following paragraphs with the following questions in mind: just what is the a priori concept that is so important? How does it relate to anything we can think of as a unity in experience? What is its connection with the putative necessity of the experiences to which it gives rise as it is employed in particular cases? And lastly, looming behind all these, lies the so far unraised question of what the connection might be between this requisite a priori concept and the categories themselves. 3. Text: paragraph 2 By means of analysis different representations are brought under one concept a procedure treated in general logic. What transcendental logic, on the other hand, teaches is how we bring to concepts not representations, but the pure synthesis of representations. What must first be given — with a view to the a priori knowledge of all objects — is the manifold of pure intuition; the second factor involved is the synthesis of this manifold by means of the imagination. But even this does not yield knowledge. The concepts which give unity to this pure synthesis, and which consist solely in the representation of this necessary synthetic unity, furnish the third requisite for the knowledge of an object; and they rest on the understanding.

The first thing of significance we meet is the distinction between bringing representations under concepts and bringing the pure synthesis of representations under concepts. We are offered no illustrative example, but I believe it is easy enough to see what Kant has in mind. I bring different representations under concepts as I see what is given to me now as the front of a house, say, for in thinking of it as that I am drawing on previous exposure to the building’s rear or intuitions I have had when emerging from its front door and so on. This is another, this time undisputedly empirical, example of the way in which, through imagination, Kant says, what I make of present sensory stimulation incorporates within it experiences of the past just as when earlier on I counted the last goat of my herd as the thirtieth. Kant thinks of this operation as depending on imagination because that is the faculty we have of representing as present what is in fact absent, some-

116

thing that he takes to go on all the time, and which he describes as ‘a blind but indispensable function of the soul, one without which synthesis in general would not be possible’ (A78/B103). By contrast, bringing the pure synthesis of representations to concepts must be a sort of second-order synthesis and one which we need to conduct if perception of the world is to provide us with knowledge of objects at all. And I take it this second-order synthesis is precisely what the synthesis that was the topic of the previous paragraph achieves, one that is conducted in accordance with the so far unidentified singular a priori concept of the understanding. That missing concept, Kant now repeats, is what gives the initial synthesis of representations (i.e. the manifold of intuition) the unity that it needs to have if it is to be apt to provide us in perception with knowledge of one sort or another, experience that arises from our bringing given sensory representations to concepts with the help of imagination as it operates at the level of a ‘first-order synthesis’, that of bringing representations under concepts.4 So what then is the concept that so represents individual things (inter alia) as to make our perceptual experience of them something we could think of as a replete whole and enjoying the unity that is at the focus of Kant’s attention? I take it that this second paragraph suggests an answer in twice saying that we are concerned with knowledge (or perception) of objects (Gegenstände). The a priori concept that Kant is concerned with just is, I propose, the concept of an object. Only what gets in the way of our seeing this is the very natural thought that objects are precisely what results from bringing representations to concepts at the second stage of the synthetic process, at the stage of what I called just now first-order synthesis. And of course that particular way of understanding the term ‘object’ is firmly anchored in philosophy well before Kant, as for example at the start of Leviathan, where Hobbes had written: ‘[The thoughts of men] are every one of them presentation or appearance of some quality or other accident of a body without us which is called an Object’ (Leviath. i. i). But what other way of reading those two occurrences of the expression ‘knowledge of an object’ (or ‘knowledge of all objects’) is there? Answer: one that treats the term ‘object’ simply as the internal object of replete perceptual experience, something that we articulate as a state of af4

Fidelity to the text demands notice that ‘concepts’ appears here in the plural. That reinforces the thought that in the previous paragraph the singular is merely generic. For the sake of the argument I insist on the specific singular once again, leaving the reader to judge the licence taken.

117

fairs, something in which Hobbesian objects are indeed usually involved, but which cannot be identified with such things or with simple experiences of them. That after all is surely what Kant must have had in mind when he insists that ‘even this [first-order synthesis of the manifold by means of imagination, the bringing of representations under concepts] does not yet yield knowledge.’ To the question of how Kantian objects rather than Hobbesian ones give our perceptual experiences their unity the answer must be that merely to perceive a Hobbesian object leaves entirely open how it is perceived and does not provide us with a representation that offers itself as true to the world. Taken by itself it could not be integrated satisfactorily into intentional action nor mesh smoothly with our wants, our needs and more general concerns. Just to see that horse in the paddock will not lead me to the Tote. On the other hand, if I perceive individuals as being this way or that, that is, if I represent states of affairs in my perceptual experience, the connection with desire and action is as close as it need be and warrants us thinking of such experiences as fully unified wholes if we are to think of anything in experience in that Kantian way. Seeing the horse out there as shaky on his legs will keep my hand firmly and wisely in my pocket. The idea of a fully unified experience can now be accounted for by its having the kind of Kantian object it does and so being an experience as of this or that being the case. (There is an interesting pre-echo here of the opening of Wittgenstein’s Tractatus, asserting that the world is everything that is the case; that it is the totality of facts, not of things (TLP 1 and 1.1).5 However, what for Wittgenstein could only be shown and not said is for Kant something that can be indeed be said, though only as a transcendental truth of pure reason.) Why though should Kant think that this concept of an object is a priori? In his usage a concept is a priori if it does not have an empirical origin and indeed in the only reference to the passage we are concerned with as ‘the metaphysical deduction’, at B 159, he explicitly says that there ‘the a priori origin of the categories has been proved’. As far as the present passage goes the a priori origin of the concept object should be clear. Empirical concepts are answerable to what is sensibly given in intuition. The synthesis of the intuitively given under the concept object cannot itself be supplied in intuition because if it were and were still to be thought of as pro5

However, while Wittgenstein’s idea occurs in metaphysical key, Kant’s own proposal starts off epistemically and in the Metaphysical Deduction goes no further, transition to the metaphysical mode waiting upon the Transcendental Deduction.

118

viding us with knowledge, it would itself need uniting with other intuitions into some sort of experienced whole. The concept that is to effect that, the concept object, or as I would rather say the concept of something’s being the case, has to have some other origin if regress is to be avoided. Resting solely on the understanding and not at all on intuition, it must eo ipso be a priori. To make my proposed identification of the singular pure concept of the understanding less off-putting, consider the etymology of the German word for object, ‘Gegenstand’. Kant says we speak of a Gegenstand ‘corresponding to, and consequently distinct from, our knowledge’ (A104), and certainly we think of states of affairs that we experience in perception in that way. Now ‘Gegenstand’ is literally what stands over against something, das was etwas gegenüber steht, and in that respect the term fits states of affairs just as well as Hobbesian individuals that stand over against our experiences of them. And while the educated English ear no longer hears it quite so clearly, exactly the same is true of our word ‘object’, descending as it does from the fourth declension Latin noun ‘objectus’ or the participle ‘objectum’, as they are used substantively in Lucretius’ ‘dare objectum parmae’ (to interpose one’s shield) (De Rerum Natura 4. 847) or verbally in Caesar’s ‘silva pro nativo muro objecta’ (the forest opposing (us) like a naturally grown wall). (De Bello Gallico 6.10, 5) Next, what about the thought of the deployment of the concept object or something’s being the case ‘render[ing] the unity of the synthesis of the manifold necessary’? The stronger reading of ‘necessary’ I offered before meant that any goat-counter using the decade as his base and obeying the discipline that possession of the concept brings with it would in this particular case be rationally bound to experience the animals’ number as thirty. To use an expression of an eminent sometime colleague of Alan’s, it would, in the circumstances, be ‘rationally non-discretionary’ that he should synthesise his intuitions in that way, see that as their number. Is there anything comparable that one might say about the concept that provides the second-order unity for the pure synthesis of the manifold? Obviously one cannot simply say that as one experiences things in terms of something’s being the case it is necessary or rationally non-discretionary to experience the world as being the case in some way or other. That is true enough, only trivially so. And if one sought to avoid triviality by saying that it is necessary that one experiences the world in the way that is determined by the first-order concepts that one uses, such as the decade or the

119

house-façade, then the necessity that is in question flows from those concepts, not from that of the Kantian object.6 At this point the analogy between the pure concept of the understanding and the decade is coming under strain even if it has not entirely broken down. Yet perhaps not in such a way that there is nothing to be made of Kant’s thought. A little later in the Critique’s A edition and in the course of his own elucidation of the term object Kant writes: Now we find that our thought of the relation of all knowledge to its object carries with it an element of necessity; the object is viewed as that which prevents our modes of knowledge from being haphazard or arbitrary, and which determines them a priori in a certain fashion. For in so far as they are to relate to an object, they must necessarily agree with one another, that is must possess that degree of unity which constitutes the concept of an object (A 104-5).

There is a worse and a better way of hearing these sentences. The worse is strongly suggested by the last of them: that the necessity that comes with relating our thoughts to objects is that they do so in such a way as makes them thoughts of objects. Clearly, that won’t do. But there is an alternative: namely, that if our thought is to be other than haphazard, it has to be disciplined in one way or another by concepts that represent something as being the case. It is a priori (and so, for Kant, necessary) that our synthetic construction of experience out of intuitively given data can only enjoy unity provided that we make sense of the data in terms of something external to us being the case. That is essential to our having anything that could be thought of as fully-fledged experience at all. For enthusiasts, I would say that what we have rescued from the second paragraph of my text helps us to understand the puzzling ‘transcendental object = x’ that has its place in the A version of the Transcendental Deduction and which disappears from the rewritten version in B. In the A edition, and tied to the idea that what we experience are appearances and so nothing other than representations, Kant expresses disquiet about how our knowledge can be of things capable of existing outside our power of representation (A104). The answer offered there is that we structure appearances as of such things, synthesising them in that way. ‘These appearances are not things in themselves; they are only representations, which in turn have their object – an object that cannot be intuited by us, and which may there6

Use of the Kantian second-order concept of an object would of course be necessary in any deployment in experience of a first-order concept in a way that was rationally non-discretionary.

120

fore, be named the non-empirical, that is, transcendental object = x’ He goes on ‘The pure concept of this transcendental object, which in reality throughout all our knowledge is always one and the same, is what can alone confer upon all our empirical concepts in general relation to an object, that is, objective reality’. Taking ‘object’ here as equivalent to ‘something’s being the case’, that is indeed one and the same thing throughout all our experience; and as Kant says, that is not anything empirical that we might intuit. Moreover, it is a necessary feature of all knowledge-directed perceptual experience. Where he says our representations themselves have their objects and yet baulks at saying that that is something which is capable of existing outside our power of representation, he is naturally drawn to something which serves as a way of making sense of those representations themselves, to wit, as something or other distinct from our experience itself being the case, their non-empirical and obligatory transcendental object. In the B edition, when bodies are no longer representations or appearances so much as things that are represented and things that appear, the transcendental object is no longer needed for its original purpose. Nevertheless, it remains a transcendental truth that our experience must always be of a Gegenstand, of something’s being the case, and the ‘transcendental object = x’ of the A edition merges silently into that. All this, you might perhaps say, is well and good, but what Kant is really after here is a deduction of the twelve bloodless categories, not of the singular pure concept of the understanding that we have been concerned with so far. Unless a connection between them is tightly secured everything that has happened so far is just a waste of time. So it must be the role of the last of our three obscure paragraphs to tie this connection firmly down. 4. Text: paragraph 3 The same function which gives unity to the various representations in a judgment also gives unity to the mere synthesis of various representations in an intuition; and this unity, in its most general expression, we entitle the pure concept of the understanding. The same understanding, through the same operations by which in concepts, by means of analytical unity, it produced the logical form of a judgment, also introduces a transcendental content into its representations, by means of the synthetic unity of the manifold in intuition in general. On this account we are entitled to call these representations pure concepts of the understanding, and to regard them as applying a priori to objects — a conclusion which general logic is not in a position to establish.

121

A casual reading of these lines might suggest a hiatus in the flow. Suddenly the subject appears to have changed, attention turning now to judgment rather than the synthesis of intuition or experience. Except, of course, one is brought up sharply by the explicit claim that the very same thing that gives unity to a judgment is what gives unity to the synthesis of representations in experience, in an intuition. And again, whatever that is, indeed, whatever concept that is, is to be called the pure concept of the understanding, appearing here once more in the singular. Kant glosses ‘function’ at A68/B93 as ‘the unity of the act of bringing various representations under one common representation’, and he illustrates this rather opaque idea with a brief discussion of the judgment ‘all bodies are divisible’, where the concept divisible (answering a Like-what? question) seems to play somewhat the same role as the concept decade in the discussion of the unity of experiences that yield answers to Howmany? questions. If we ask quite generally what it is for a judgment, be it linguistically expressed or just entertained in thought, to have a unity, that must come to the same thing as the question in what circumstances a concatenation of conceptual contents (or in the Lockean/Humean tradition that was Kant’s, ideas) constitutes a determinate whole, a question to which neither Locke nor Hume had any illuminating answer. But Kant’s answer is illuminating, and it is plain what it is. Such judgments constitute determinate wholes as they represent objects of knowledge, namely, when they present something or other as being the case. And this they do when they combine their constituting elements in such a way as to make up judgments that can be assessed as true or false. So, Kant reasons, just as we combine representations to form determinate judgments that can be assessed as true or false, so in experience we represent something as being the case by the very same sorts of combination as they have in judgment. Experience has the unity of judgment just because in perceptually representing something to be the case we do nothing other than combine representations in ways that are liable to capture a truth. Putting it in a way that Kant never did, but which happily expresses his thought, the unity of perceptual experience is supplied by its being experience of the form that p, and for experience to have that form is nothing more than its combining its elements as judgments to the effect that p, and, ‘represented in its most general aspect’, articulating the thought that this or that is the case.

122

To put this by saying that Kant’s crucial insight is that experience is essentially propositional may slightly obscure his idea that perceptual experiences and judgments both have one and the same pure concept of the understanding as their base. Putting it in his language we can say that both experience and judgment rest upon the concept of an object. And once we stress that the term ‘object’ stands for the concept of something’s being the case, it is plain that it is in play whenever we judge that p, since to judge that p just is to put p forward as being the case. Given that those experiences that Kant sees as unitary syntheses of perceptual representations take the form of representing something as being the case, they will inevitably do that by being cast in form of judgments. It is at this point, and only at this point, that Kant can introduce the categories, the pure concepts of the understanding in the plural, into the equation. To the extent that the combination of representations in judgmental structure necessarily involves non-empirical concepts, so the experiences that acquire their unity from their propositional form must also incorporate those same a priori concepts. That is why by means of the synthetic unity of the manifold in intuition in general (i.e. via the concept object) the understanding has to introduce a transcendental content into any whole or complete experience that it comes to construct. There remains just a little more unpacking to be done, unpacking that answers three questions: (a) With what justification does Kant suppose that judgments involve a priori concepts other than the one already discussed? (b) If there are any such, what reason has he to think that they must all contribute to the nature of our experience of the world? And (c), just what is the relation between the singular pure concept of the understanding, object, and the twelve categorial concepts that are also pure concepts of the understanding? As for the first of these, Kant’s understanding of the matter is straightforward enough. Judgments bring representations under representations. Thus, in the (slightly adjusted) example of ‘all these bodies are divisible’ members of an intuitively given demonstrated ensemble are brought under the empirical predicate concept divisible, and so brought in a way that produces a judgment presenting the thought as true. Schematically, we have a pair of empirical elements related by a relation R so as to yield p, thus R = p. Were we to suppose R to be a further empirical element in the judgment, that in its turn it would need to be related to the others in a way to produce a complete judgment R´ = p´, as perhaps in a judgment like ‘All Pygmies are smaller than any

123

Texan’. In a nutshell, for a judgment to present something as being the case the empirical elements have to be bound together by non-empirical ones in some way, and as far as Kant is concerned it is evident from inspecting the range of judgments that we make what the non-empirical, a priori, formal conceptual elements are that effect this combination. So much for the first question. The second question is also relatively unproblematic. Kant’s interest in the completeness of the table of categories is less an interest in the logical bounds of formal concepts of the understanding than interest in those formal concepts we must allude to in giving an account of the range of experience that is actually ours. Nothing can justify a claim that there could not possibly be other categories than the ones he lists, and I dare say he would not demur. But he would say that, if we were to suppose that some members on his list were redundant, they could only be redundant in the explanation of the range of experience that is narrower than that which we actually enjoy (I leave aside any thought about redundancy that is effected via interdefinability). The last of the three questions is of some interest in its own right. The identification of the singular pure concept, that of an object, was elusive just because like the air we breathe it is around us all the time and never obtrudes. To speak of experiencing something as being the case is more of a philosopher’s invention than anything else. In everyday speech we think of ourselves experiencing particular things, noticing a friend’s being off sick or the walls needing a lick of paint and the like, and if it comes to generalisation I suppose we would talk of seeing this or that, which smacks more of a Hobbesian object than the Kantian one. Given this marked reticence from view of the concept, it could be appealing to surmise that for Kant the idea is simply constituted by the twelve categorial concepts. Each of them in their different ways would enable us to put forward something as being the case or to represent that p in such a way that, as he put it elsewhere, the various empirical representations in the judgment are signalled not merely as having ‘always been conjoined in my perception ...[but] that they are combined in the object, no matter what the state of the subject may be’ (B142). That would be one way of connecting the singular pure concept with the plurality of the categories so close to Kant’s heart. But it cannot be the right one, since the idea of something’s being the case is itself in no way diminished as we restrict the number of ways in which our object-oriented judgments come to be made. Equally, were we to find novel ways of com-

124

bining empirical representations so as to construct hitherto unsuspected ways of judging something to be the case, the content of that singular concept would in no way have been enriched. It cannot be that the categories constitute the concept of an object as it is Kantianly conceived. A better thing to say, surely, is that all our perceptual experience rests on the basis of that a priori synthetic unity that is supplied by that recurrent concept. However, any experience enjoying that unity has also to draw upon the categorial concepts, not always the same ones, but always some of them, to attain its unity. That is the fundamental reason why as the understanding constructs experience of objects it is obliged to introduce a number of further transcendental concepts into our experience. Unless it did so, it would not succeed in bringing the data of intuition under the concept of an object at all. It always does that same thing in a certain way, and according to the way in which it does it, so it is obliged to structure the experience that it synthesises with the formal concepts that identify the way in question as the one it is. Thus it is that reflection of pure synthesis taken in its most general aspect gives us the pure concepts of the understanding, at last making their desired appearance in the plural. 5. The time has come for me to make good my promise to offer a loose paraphrase of the paragraphs we have been considering. Here it is, set out for you as the right hand column in face of the original text: CPR A78-9/B104-5

Proposed loose paraphrase

Pure synthesis, represented in its most general aspect, gives us the pure concept of the understanding. By this pure synthesis I understand that which rests on a basis of a priori synthetic unity. Thus our counting, as is easily seen in the case of larger numbers, is a synthesis according to concepts, because it is executed according to common ground of unity, as, for instance, the decade. In terms of this concept, the unity of the synthesis is rendered necessary.

We shall find pure concepts of the understanding pervading perceptual experience as we reflect quite generally on what goes into its construction. That must be carried out under an a priori constraint that lends instances of experience their unity and completeness, one that is apt to yield knowledge of this or that. So, for example, as we count things, we have to use a conceptual base like the decade or the mile to give a determinate answer to questions of quantity or distance. Only by drawing on some such concept can our experience be properly reflective of the world and demand of us a determinate

125

way of combining the sensible data that we process in our pursuit of it.

By means of analysis different representations are brought under one concept — a procedure treated of in general logic. What transcendental logic, on the other hand, teaches, is how we bring to concepts not representations, but the pure synthesis of representations. What must first be given— with a view to the a priori knowledge of all objects — is the manifold of pure intuition; the second factor involved is the synthesis of this manifold by means of the imagination. But even this does not yield knowledge. The concepts which give unity to this pure synthesis, and which consist solely in the representation of this necessary synthetic unity, furnish the third requisite for the knowledge of an object; and they rest on the understanding.

The concept decade illustrates how particular representations are united under some specific empirical concept. What we are seeking, though, is something that characterises experience quite generally and a priori. That we have in the last of three elements essential to the mind’s processing work. First there is disjointed sensory input; then comes the bringing together of that input under empirical concepts; finally, there is the notion that works to give experience that unity which makes it apt to represent states of affairs and supply knowledge of something or other’s being thus and so. This last is only achieved with the deployment of the a priori concept of something’s being the case, scilicet the concept object of experience.

The same function which gives unity to the various representations in a judgment also gives unity to the mere synthesis of various representations in an intuition; and this unity, in its most general expression, we entitle the pure concept of the understanding. The same understanding, through the same operations by which in concepts, by means of analytical unity, it produced the logical form of a judgment, also introduces a transcendental content into its representations, by means of the synthetic unity of the manifold in intuition in general. On this account we are entitled to call these representations pure concepts of the understanding, and to regard them as applying a priori to objects — a conclusion which general logic is not in a position to establish.

Now, what gives judgments their unity is just their putting forward in thought or speech something as being the case and assessable as true or false. And as we have just seen that is the very thing that gives instances of perceptual experience their unity. Because in judgment we can only present something as being the case by drawing on formal and a priori concepts to bind the constitutive elements together as complete propositional wholes, so too in the construction of replete perceptual experience we must apply those same a priori concepts that enable us to present something as being the case by drawing on formal and a priori concepts to bind the constitutive elements together as complete propositional wholes, so too in the construction of replete perceptual experience we must apply those same a priori concepts that enable us to make judgements of the various sorts available to us. Hence

126

experience of the world will itself be informed by those concepts, which are therefore rightly called ‘pure concepts of the understanding’.

Such then is Kant’s route to transcendental idealism pursued along the path of understanding. There is, it seems to me, so much good sense in it that its virtues deserve to be clearly signalled. But since the ultimate destination, Transcendental Idealism, does not exert any great pull for us today it is also worth pointing out at what point we may comfortably alight from the Königsberg express. The virtues first: Surely Kant’s greatest merit is to have seen how important it is to recognise that replete perceptual experience is fundamentally propositional in nature. In that he took a giant step beyond his predecessors, and we are in his debt for that. Then he seems to have seen that disciplined conduct of the concepts we possess brings with it an a priori normative commitment to acceptance of judgments that contain them, given only canonical stimuli as input. Thirdly, it can hardly be denied that the judgments that we make embody formal and syncategorematic elements in terms of which we are bound to inform the perceptual experiences that come our way. These are important truths for the recognition and articulation of which we are in Kant’s debt. What is far less certain, though, is that the formal concepts he put his finger on are a priori in any rich sense that we would wish to endorse. The claim that they are is, for him, the ultimate lever he relies on to secure the conclusion that the world we experience must be the world as it appears rather than the world as it is in itself. So as long as we do not take that view of them, that is the point at which, acknowledging our debt, we may think it right to part company with him and disembark. 6. Finally, if the discussion has been along the right lines what we have should enable us to settle a long-standing puzzle about Kant’s notion of objective validity. It can be put like this. For him, the presence of the categories in our experience is what lends it objective validity. Objective validity is, for him, coterminous with universal and necessary validity, and that

127

in turn with truth.7 Yet on occasion judgments that we come to make, perceptual ones among them, fully dependent on the categories, may be downright erroneous. So, one wonders, is Kant just over-optimistic in the connection he makes between the categories and the truth and objective validity of the judgments we make in using them? The first thing to say is that if we have regard merely to the idea that it is the concept of an object that gives the unity to the synthesis of the manifold of intuition, then that does nothing to secure the correctness of the perceptual judgments that we make. It merely insists that for the question of correctness to arise we have to synthesise our experience in the shape of a judgment to the effect that something or other is the case. (The pure synthesis is not itself a judgment, of course, merely a representation that is apt for acceptance or rejection, apt for judgment as true or false). However, we saw earlier that the second-order concept of an object can only be deployed if with the help of what Kant calls imagination we organise our experience in conformity with some first-order concept, goat, house-façade, maybe even decade and so on, answering questions of the What-sort? Like-what? and How-many? kinds and so on. Now, for Kant the concepts we employ at this first-order point are – at least in the kind of perceptual case we have been concerned with – introduced in the light of what our receptivity makes available to us, and disciplined by that and the degree of order which we are able spontaneously to bring to it. This is the fruit of Kant’s idealism, and I think he would even see the idea ultimately extending to the theoretical concepts (like magnetic field or planetary orbit) that he wants to find room for in any full characterisation of the world of appearance. That bold thought aside, though, the upshot is that, while we can of course make false judgments in line with the categories as we synthesise perceptual input in line with the first-order concepts that provide answers to our quantitative and qualitative curiosity, there is a range of judgments involving perceptual notions of which it holds a priori that when exercised with proper discipline and care they will be true, and will, as Kant expresses himself, furnish knowledge. Perhaps it is for this reason that Kant said in respect of the unity of the manifold that is provided by the concept decade, that it is rendered necessary, and necessary in the stronger of the ways I outlined. Counting my goats with all due care, I am bound to find that thirty is their number, and no-one else, proceeding with like care in that same situation, could come to any other total. As another of Alan’s distinguished quondam col7

The equation of objective validity and truth is explicit at A 125.

128

leagues has put it, there just is nothing else to think than that they are so many. Then the objective validity of the judgment I make to the effect that so many goats have returned home will be true and amount to knowledge itself. What this shows is that the form that experience has to take as it is governed by the categories cannot be divorced from universal and necessary validity, that is, from truth. And this is not impugned in the slightest by the anti-Cartesian and entirely Kantian reflection that there can be no guarantee that when within the appropriate range8 such judgments of mine are canonically disciplined and knowledge-yielding I shall know that they are so; nor, of course, is it threatened by the fact that if at day’s end I am a weary goatherd, I may all too easily miscount my flock. References McDowell, John 1998: ‘Having the World in View: Sellars, Kant, and Intentionality’ (the Woodbridge Lectures for 1997). Journal of Philosophy, 95, pp. 431-91. Peacocke, Christopher 1999: Being Known. Oxford: Oxford University Press.

8

Kant himself has nothing to say about the range of judgments over which the claim is most plausibly made out, and taken with the sort of generality that he most probably had in mind, it is surely false. A suggestion that has much to recommend it is Christopher Peacocke’s concerning those concepts he calls ‘epistemically individuated’ (1999, esp. chapter 2).

129

Self-awareness Richard Sorabji It is a great pleasure to have an opportunity to honour Alan, from whom I have benefitted for over 35 years through his philosophical range, his unfailing contributions to innumerable seminars in Ancient Philosophy, his valued comments on translations of late Greek Philosophy, and his ever supportive and genial presence. Characteristically, he has by a quiet but decisive objection caused me to modify one of the points I am making in the present chapter. When we talk of self-awareness, this can be awareness, among other things, of our activities, our motives, our character, or our true self. But these are inter-connected, so the idea of self-awareness is not just a random collection of ideas. It has been argued at various times that self-awareness is impossible, or at any rate so difficult that it is better achieved through awareness of others, or on the other hand that it is quite the reverse: inevitable and infallible. Further, it has been asked what faculty gives us a unified awareness of all our psychological activities. I have discussed these topics in chapters 11 to 14 of Self, which has been published since the conference in honour of Alan Lacey that formed the origin of the present book. But I think it might be useful to offer a streamlined overview of most of the topics, though I shall omit the discussion of inevitability. 1. The impossibility of self-awareness One argument for the impossibility of self-awareness is rehearsed by Plato in his Charmides at 167c-d; 168d-e. It is that sight cannot be sight of itself and not of colour. If sight is seen, it will have itself to be coloured. This is generalised to knowledge and belief not being just of themselves at 168a, and to other relations at 168d. On one interpretation, the argument is that sight that was just a perception of sight and of nothing else would have no content, for it is colour that gives sight its content. A further difficulty for self-awareness is discussed by Plato’s pupil Aristotle at On the Soul 3.2, 425b12-28, and it adds on to Plato’s difficulty of contentless sight the danger of an infinite regress. How are we aware that we are seeing (or, equivalently, of our seeing)? It must be by sight, for if we had to be aware

of an act of awareness (e.g. that we are seeing) by an act of another type (other than sight), we should have embarked on an infinite regress of types. There is another kind of regress that needs to be considered, but it is a regress of acts, not of types of act, and is exploited by the song about a lovers’ quarrel, ‘I thought you thought I was thinking’. If psychological activities are not self-luminous, but are known by further acts of awareness, then there has to be, for finite minds, at any given time, at least a temporary halt to a regress of acts of awareness of awareness. I can be aware that I am aware that I am seeing, but at any given moment the first awareness in any chain will always be something of which I am not currently aware. Of course, there is in principle no barrier, only the barrier of lack of concentration, to my becoming aware of it, and so adding an extra awareness to the beginning of the chain. But this new awareness will be something of which I am not yet aware. Whether Aristotle considers a regress of acts we shall consider shortly, but at present he is concerned with a regress of types of act. Aristotle has another objection to allowing that I have to be aware that I am seeing through a type of faculty other than sight. If this supposed other faculty is not to be contentless, as Plato’s argument threatened, it will have to be aware not only that I am seeing, but also of the colour seen. But in that case, there will be two faculties, sight and the supposed other faculty, directed to one type of object: colour. And it goes against a principle accepted by both Plato and Aristotle that there should be two faculties for the same type of object. For both philosophers distinguish faculties by the different objects they have.1 So far, it seems that I am aware that I am seeing by sight. But Aristotle considers a further difficulty. As Plato himself threatened, our seeing, or at least that which reveals our seeing, in other words the eye that sees, would have to be coloured. For sight by definition is aware of the coloured. Aristotle gives two answers. One is that the eye that sees is coloured in a way, and I take this to be because while it is seeing, it takes on colour patches, though not as intrinsic colours of its own, only as borrowed from the scene it is seeing. But the more important solution offered by Aristotle is that perceiving by sight is different from seeing. We do not see that we are seeing, but merely perceive it by the use of sight. He compares how we do not see darkness, because sight is defined as perception of colour or brightness. But we nonetheless perceive darkness by trying to use our sight (and failing), so we perceive darkness by sight. In the case that concerns 1

Plato Republic 477 c-d; Aristotle On the Soul 2.4, 415a20-2.

132

us, we perceive that we are seeing by using the generic faculty of perception which includes all five senses and more besides. But it is important that the generic faculty of perception should be brought to bear on our seeing not on our hearing. So it is by sight that we perceive that we are seeing, even though we do not see that we are seeing. Aristotle here refuses to generalise Plato’s suggestion about lack of content. He has agreed that seeing would lack content if it did not include colour among its objects, but perceiving, or perceiving by sight, is not treated as depending on colour for its content. According to Aristotle’s greatest ancient exponent, Alexander, if the passage is his,2 Aristotle finishes his discussion at 425b26ff by removing a regress of acts of awareness. Aristotle here says that there is a sense in which an act of perceiving is identical with what it perceives. Aristotle does not explain why he makes this point here. But the point would acquire relevance if it is meant to stop a regress of acts of awareness. So far, Plato has raised a doubt about whether self-awareness is possible and Aristotle has intensified the doubt, but sought to answer it. But the debate continued, for, around 200 A.D., more than 400 years after Aristotle, the sceptic Sextus Empiricus raised related, though slightly different, doubts, and 50 years later again Plotinus sought to answer them. 2. Is self-awareness so difficult that it is more easily achieved through awareness of another? It has been observed in modern psychology that at around 9 months, normal human children, unlike any of the great apes and unlike certain types of damaged children, play with mother, father or carer the game of ‘are we looking at the same thing?’. The desired state has been called one of shared attention. It has been argued that children get the idea of themselves as having a gaze only hand in hand with the idea of someone else having a possibly divergent gaze. They see themselves as having psychological characteristics only in connexion with seeing others as having possibly divergent characteristics.3 If so, Descartes has things the wrong way round. It is not that I know my mind alright, but as regards others, I do not know if they are automatons. Instead, I only have the idea of my mind insofar as I have the idea of another’s. I know myself through another. 2 3

Alexander(?) Quaestiones 3.7, 92, 31. See e.g. Tomasello (1994) and (1999).

133

This view is put forward by Plato in the First Alcibiades,132c-133c, if it is by him, although he is talking about knowing one’s true self as being one’s reason. Just as one cannot see oneself just by looking, but does better to see oneself reflected in the eye of another, so one recognises one’s true self by seeing reason at work in another. Augustine was to disagree 750 years later in On the Trinity 10.3.5. Incorporeal things are not known by mirroring. Nothing is more present to the soul than itself. Aristotle accepts Plato’s idea, though in a different connexion. He addresses the question what the value of friendship is, given the rather alien Greek idea that the ideal person would be self-sufficient. His answer turns on the idea that awareness of one’s friend is connected with awareness of oneself. The friend, it is suggested, is another self, and easier to observe. There are at least four versions of his argument,4 but in none is he, like Plato, discussing the true self. In some, he is talking about awareness of one’s friend’s activities as good and as one’s own. At two points, I believe, he is referring to the phenomenon I have mentioned of shared attention to activities and is treating it as being particularly pleasurable, Eudemian Ethics 7.12, 1244b24-28; 1245b21-24. The difficulty of direct self-awareness is evidenced by the Aristotelian author of pseudo-Aristotle Magna Moralia 2.15, 1213a13-26, through the point that it is much easier to recognise the faults of others than our own. The same point was to be made 500 years later by the doctor Galen,5 who drew the conclusion that to correct our faults we should rely not on introspection, but on getting someone else to criticise us. The Stoics introduced awareness of self through another in a quite different context to which Brad Inwood has drawn attention.6 A surviving papyrus fragment from an author of the second century AD, the Stoic Hierocles, asks why the chick does not fear the leaping bull close to it, but scuttles away from the small, distant hawk, Elements of Ethics col.3, lines 39-45, A. A. Long.7 His reply is that the chick has awareness of the liabilities and abilities of its physical self, and, as Inwood points out, this awareness about its physical self has to take into account the likely actions on it 4

See Aristotle Eudemian Ethics 7.12, 1244b24-1245a12; 1245a30-b1; 1245b21-24; Nicomachean Ethics 9.9, 1169b33-1170a4; 1170a29-b14; and pseudo-Aristotle Magna Moralia 2.15, 1213a13-26. 5 Galen On the Errors and Passions of the Soul, in Galen Scripta Minora, ed. Marquardt et al., Teubner edition, vol. 1, ch.2, p.4, line 11- p.5, line 2; ch.3, 6,17 – 7,1. 6 Inwood (1984). 7 Similar is Seneca Letter 121.19.

134

of other beings. Although these other beings do not have to be seen as having minds, they are seen as being safe or unsafe. This is how others come into the idea of self. Hierocles has an additional view that the chick’s awareness is not to be classified as the work of reason – the Stoics hold that animals lack reason. But neither is it mere unconscious instinct – nature in Greek terms. It is something in between these two: self-perception. Alan Lacey has pointed out to me that reaction to the hawk shape does look suspiciously like an instinctual response, which would not have to be mediated by self-awareness, but it was self-perception for which Hierocles was arguing. A final example has a rather different character. Plotinus at Enneads 5.1 [10] (1-17), is struck by the irony that the rebellious adolescent wants to belong to himself and to have self-determination, but by distancing himself from his parents becomes unaware who he is. This is used as a simile for describing the process by which inferior entities are formed in the universe, souls by apostasis from Intellect. Plotinus here sees knowledge of one’s parentage, hence knowledge of others, as entering into one’s very conception of who one is. 3. The infallibility of self-awareness It is Augustine, possibly influenced by Plotinus, who makes the shift towards the point of view we find later in Descartes. The soul, according to his On the Trinity, is present to itself, so does not have to seek itself. The difficult question for Augustine is how it ever makes a mistake about itself, but his answer is that, although it has a continuous knowledge (nosse; notitia; cf. nota) of itself even in infancy, it can make mistakes when it starts thinking (cogitare) about itself, 14.5.7; 14.6.8-9. It makes mistakes when it forms extended images to imagine itself, because it is not something extended, 10.3.5; 10.3.7 – 10.6.8; 10.8.11; 10.10.16 (72-85) ed. Mountain, CCL 50. Augustine gives one of his many versions of the Cogito argument that Descartes was to use later, according to which doubt is impossible about one’s own psychological activities: it comes in On the Trinity 10.10.14: But who will doubt that he lives, remembers, understands, wills, thinks, knows and judges? For even if he doubts, he lives. If he doubts where his doubts come from, he remembers. If he doubts, he

135

understands that he doubts. If he doubts, he wants to be certain. If he doubts, he thinks. If he doubts, he knows that he does not know. If he doubts, he judges that he ought not rashly to give assent. So whoever acquires a doubt from any source ought not to doubt any of these things whose non-existence would mean that he could not entertain doubt about anything. In some ways Augustine is more explanatory than Descartes was later to be in the second Meditation. He here shows very clearly the principle on which his Cogito rests. He is not relying on the supposed infallibility of introspection. Instead, he is confining himself to conditions which have to be met if there is to be any doubting at all. If he doubts the holding of such conditions, the doubt itself guarantees that they are met and that the doubt is untenable. He carefully avoids asserting anything unwarranted by putting everything in terms of an ‘if’: ‘if there is doubt’. He also explains better than Descartes why he thinks wanting is undoubtable by arguing that doubting implies wanting. He cannot be accused of giving us a weak argument by failing to express it in the first person with an ‘I’ like Descartes, because if the argument gains anything from being cast in the first person, which is a controversial claim, Augustine does cast it in the first person in some of his other formulations: si fallor, sum – ‘if I am mistaken, I exist’. It has been suggested by Dominic O’Meara8 that Augustine may have been influenced by a passage in Plotinus. Certainly Plotinus is looking for irrefutability and he uses an analogous strategy. Intellect is irrefutable because it confines itself to thinking what is identical with itself, 5.5 [32] 2 (18-21): So that the real truth, conforming with itself and not with something else, says nothing other besides itself, and what it is it also says, and what it says it also is. Who then could refute it? From where would he bring the refutation? Augustine continues in On the Trinity 10.10.16 to argue, like Descartes, for the incorporeality of the mind (mens). It has just been shown that the mind is certain (certus) about itself and so knows (nosse) itself. And then a controversial premiss is used, that a thing cannot be said to be known (sciri) so long as its essence is not known. Aristotle had already given examples which show this principle to be wrong. One can know that the moon is 8

O’Meara (2000).

136

eclipsed without knowing that the essence of lunar eclipse involves the earth’s shadow. Nonetheless, Augustine continues by saying that the mind must therefore know (nosse) and be certain of its essence. But it is not at all certain whether it is air, fire, or anything bodily. So it is not anything bodily (presumably in its essence). Descartes in the seventeenth century argues for this conclusion somewhat differently, because he completes the argument of the second Meditation, which initially sounds similar, only in the sixth, where he appeals to the further notion of clear and distinct ideas. But what does seem closer to Augustine’s argument for incorporeality is the Flying Man argument produced by someone who could not have known of Augustine’s Latin, the brilliant Islamic philosopher Avicenna, or Ibn Sina, in the eleventh century. A major presentation of the argument is given in Al-Shifâ (the Healing): Soul, 1.1, (Rahman 15.18 – 16.17) in Arabic, translated into medieval Latin in De Anima 1.1, in Avicenna Latinus, ed. S. Van Riet. In a letter to a former pupil, Bahmanyâr, published by Jean (=Yahya) Michot in Muséon 2000, he explains his premises – one which we can see to be interestingly different from Augustine’s, though equally unintuitive. Whoever knows something knows his own essence. Avicenna imagines a man who comes into existence floating through the air, feeling no contact even of one limb with another. Such a man would know directly that his essence existed. But he would not know that body existed, since he would have no evidence for its existence. Avicenna concludes that his essence, which he takes to be soul, is independent of body. There are further similarities between Avicenna and Augustine. Both use the idea that the soul is present to itself. Both argue that what is imagined cannot be part of one’s essence. Avicenna draws a distinction like that of Augustine between two kinds of awareness: the intellect has knowledge (shu’ûr) of existence, but there is no intelligising (‘aqala) of intellect. If there is an explanation of the similarities in the argument for incorporeality, it must be a common Greek source of influence, and they did both know the work of Plotinus’ pupil Porphyry. Porphyry, a little before Augustine, writes in Sentences 40-41 of being present to a self that is present to itself (parôn paronti). Further, speaking not of soul as a whole, but just of intellect, he writes that since intellect can turn away from the body and still both remain intact and know itself, it does not owe its essence to the body. One element of Porphyry’s argument was taken by the ancient Greek commentators to be contained already in Plato’s Phaedo. When Plato there

137

makes Socrates advocate withdrawing one’s soul from the body, he reveals that body is not in the essence of soul. 4. What unifies our awareness of our psychological activities? I come now to my final topic of the unification of self-awareness. Plato asks how we distinguish sweet from white, since no one sense can distinguish them. He infers that the five senses must converge (sunteinein) on something unitary (mia), the soul, Theaetetus 184d, and he stresses the soul’s reasoning activities, 186a – 187a. Aristotle makes the same point very graphically. Without sweet and white being presented to something unitary, it would be as if I perceived sweet and you perceived white, On the Soul 3.2, 426b17-23. But for the task of distinguishing sweet from white, he holds, the unitary thing is not the reasoning faculty of soul, but something sensory, the common sense, or sense perception taken generically as opposed to one of the five senses. Aristotle takes the subject further in On Sense Perception 7, 447b24 – 448a1. In order to tell whether we are perceiving qualities from the same range or (like sweet and white) from different ranges, we have to know whether we are using one sense or two. In order to tell which end of a range we are perceiving, we have to know whether we are perceiving a positive quality, or (like wet or cold) a privation. In order to tell whether the sweet and white belong to one thing, we have to know whether we are perceiving sweet and white simultaneously. This last point of Aristotle’s, which was drawn to my attention by Stephan Eberle, seems to me important. It implies that we get the idea of qualities belonging to the same object in the external world only through the idea of the same person perceiving them at the same time. Thus there will be no idea of a physical world of bodies possessing qualities without an idea of a single perceiver possessing simultaneous perceptions. This is an argument that we need the idea of a person possessing psychological states, not just the idea of streams of consciousness possessed by no one. I shall return at the end to another argument for a unitary self which possesses psychological states. This one has analogues in Kant and in Hindu thought. In the third and fifth centuries A.D., the subject is taken further by Plotinus and even more emphatically by Proclus, Commentary on Plato’s Parmenides 957, 28 – 958, 11, Cousin, when they point out that the range of psychological activities of which we can be aware is far wider than the

138

perception of sweet and white. In cognition, there are the five senses, the common sense and reason. Among desires there are appetite, anger and deliberate choice, to which a unitary vital faculty can give the nod, having enabled us to say, ‘I have appetite’, ‘I am angry’, ‘I exercise deliberate choice’. But prior to both cognitions and desires there is another unitary faculty conscious of all these activities, which says, ‘I am perceiving’, ‘I am reasoning’, ‘I have appetite’, ‘I will’. The debate on whether it is Plato’s reason or Aristotle’s common sense that serves as the unifying faculty has continued into modern philosophy, with David Rosenthal backing higher-order thought (HOT) and William Lycan higher-order perception (HOP), that is respectively thought about psychological activities versus perception of psychological activities.9 But HOT and HOP are only two views among many. By postulating a higher order of awareness, that is, an awareness of awareness, both views reject the idea that self-awareness of psychological activity is direct, and does not need a further level of awareness. Moreover, there are several different ways in which ancient philosophers suggested that awareness of psychological activities might be direct. Augustine’s idea of the soul being present to itself is only one. Another view was Aristotle’s, mentioned above, that in some sense an act of perceiving or thinking is identical with its object, so that when the object is itself an act of awareness, we do not need to postulate that it constitutes a second, lower level of awareness. Then there is the Neoplatonist view that intellect is self-aware by turning in on itself (epistrephesthai), a thing that bodily entities cannot do. But even if we accept that awareness of our psychological activities involves a higher level of awareness, the ancient philosophers offered more suggestions about the higher level than just perception or thought. Plotinus tended to suggest that the mirror of imagination was used for awareness of non-intellectual psychological activities. But by far the most interesting suggestion, recorded by pseudo-Philoponus, Commentary on Aristotle On the Soul 3, 464, 24 – 465, 31, is that we need to postulate a faculty of attention (prosektikon). This idea is ascribed to the ‘Attic commentators’, and he may have in mind the Athenian Proclus. It does seem to me that this, though not the whole story, is an important step forward. Attention deserves to be singled out as a distinctive capacity. For one thing, our attention can be subject to a global alert or heightening, so that after a shock received through one sense modality or 9

Rosenthal (2002-3); Lycan (2004).

139

through one piece of thinking, we remain for several hours or days on alert for further information through all our cognitive modalities, as if attention was something that could operate as a single unit, and furthermore as if it had the wide range that Proclus was looking for, because it can operate on all our modalities. Pseudo-Philoponus offers at 466, 30-35, an elegant argument that awareness of perception is not itself a kind of perception. Suppose I am busy reasoning out a problem while I walk along the street and I pass my friend while thus absorbed. When I get home, and am no longer in a position to see my friend, I reflect for the first time that I did see him. Previously, I was aware of my friend and even perhaps of my friend as seen from a certain angle. But this awareness of a view of my friend is not yet an awareness of seeing him. In this case at least it is not perception that recognises that I saw my friend. Pseudo-Philoponus suggests that it is reason. This is not incompatible with the suggestion that it is attention, if attention is classified, as Proclus may have classified it, as a kind of reason. This is an objection to HOP, but it can also be argued that sometimes awareness of psychological activity is not due to HOT. In the Republic Plato describes musical theorists who struck pairs of notes in order to see how small a difference in pitch could be detected. Thinking is involved in such a case, but only after attention has been directed to hearing. One first attends to one’s hearing and only then thinks about what that implies for the size of interval. These examples illustrate what I believe to be a general truth, that there is no single faculty by which we are aware of our own psychological activities. We use different faculties in different cases, even though it is a step forward to include attention on the list of faculties. So the unity which Aristotle rightly says is needed, if it is not to be as if you perceived one thing and I another, does not come from a unitary faculty, as both ancient Neoplatonists and modern philosophers suppose. It is the burden of my book on self that many problems remain insoluble until we recognise that there is such a thing as Self. In this case, I believe, it is not a unitary faculty, but the unity due to the faculties belonging to the same self, that gives unity to our awareness of diverse psychological activities.

140

References Inwood, Brad 1984: ‘Hierocles: theory and argument in the second century AD’, Oxford Studies in Ancient Philosophy, 2, pp.151-83. Lycan, William G. 2004: ‘The superiority of HOP to HOT’. In: Gennaro, Rocco W. (ed.): Higher Order Theories of Consciousness: An Anthology. Amsterdam: John Benjamins, pp. 93-113 O’Meara, Dominic J. 2000: ‘Scepticism and ineffability in Plotinus’. Phronesis, 45, pp. 240-51. Rosenthal, David 2002-3: ‘Unity of consciousness and the self’. Proceedings of the Aristotelian Society 103, pp. 325-52. Sorabji, Richard 2006: Self: Ancient and Modern Insights about Individuality, Life, and Death. Chicago and Oxford: Chicago and Oxford University Presses. Tomasello, Michael 1994: ‘On the interpersonal origins of self-concept’. In: Neisser, Ulrich (ed.): The Perceived Self. Cambridge: Cambridge University Press. Tomasello, Michael 1999: The Cultural Origins of Human Cognition. Cambridge, Mass: Harvard University Press.

141

Perceiving that we see and hear: Aristotle on Plato on judgement and reflection Mary Margaret McCabe

1. Perceiving that we see and hear: the opening gambit I begin with a short, and famously vexed,1 passage from Aristotle’s de anima: Since we perceive that we see and hear, it is necessary either to perceive by sight that one sees2 or by some other (sense). But the same sense will be of the sight and of its underlying colour. Then either there will be two senses of the same object or the same sense will be of itself. Further, if indeed the perception of sight is other [sc. than sight] then either the perceptions will go to infinity3 or some sense will perceive itself. So we should assume the latter in the first place. But this involves a puzzle: for if perceiving by sight is seeing, and what is seen is colour or what has colour, then if some sight is going to see the seer,4 the first seer will have colour. 1

See, for example, Ross (1961) ad loc.; Hamlyn (1968) ad loc.; Hicks (1907) ad loc.; Kosman (1975); Osborne (1983); Caston (2002, 2004); Sisko (2004); Johanssen (2005). I translate the text of Ross’ OCT (1956) except at 425b19, where I follow his (1961). In what follows, I have tried to avoid elephantiasasis of the footnotes by only mentioning points from the commentators which directly affect my argument. 2 The Greek shifts from ‘we perceive that we see….’ at 425b12 to a third person formula (‘to perceive that one sees’ at 425b13) which allows the construal that it is sight that perceives that it sees. 3 Kosman (1975) argues that the infinite regress implies that second-order perception must be a necessary condition of first-order perception; so too does Caston (2002, 2004). Johanssen disagrees; but argues that his interpretation (that this is an account of an ‘inner sense’) goes through anyway (2005). More below, §3. 4 Ross (1956) prints horan here, twice, with less authority than the majority of mss, which have horôn; contrariwise he prints the latter in his (1961). Support for the latter is supplied by horôn at 425b22, provided we can understand the role of the intervening sentence. Ross makes the proviso that in 19 the reference is to the faculty of sight, seeing, and to the absurdity of the faculty’s being coloured; while in 22 Aristotle must be talking about the organ’s being coloured. See Caston (2002, 2004) and Johanssen (2005) for different views of this issue.

Well, then, it is clear5 that perceiving by sight is not one.6 After all, when we do not see, it is by sight that we judge both the darkness and the light, but not in the same way.7 And again, that which sees is coloured8 in a way; for each sense-organ is receptive of what is perceptible without the matter. This is why even when the sense-objects are no longer there, there are (still) perceptions and appearances in the sense-organs. (Aristotle, de anima 425b12-25)

Aristotle’s opening gambit, ‘Since we perceive that we see and hear, it is necessary that we perceive that we see either by sight or by some other (sense)’, gives the appearance of being a fresh start – although that appearance may mislead. The previous chapter was about whether we have just the five senses; but now, or so it seems, Aristotle has shifted his attention to a different issue. If we read Aristotle as I was brought up to do, we might be unsurprised by his lacunose style (surely, we reassure ourselves, these are just lecture notes?9). And – however new the point – does Aristotle take it as just obvious that we perceive that we see and hear? The obvious is where he tends to begin, for new points and even some old ones.10 Is it obvious that we perceive that we see and hear? What is it to perceive that we see and hear? I shall wonder about the continuity, or otherwise, between what Aristotle thinks is obvious, and what we do; and I shall wonder just what he is advancing here as an explanation of perceiving that we see and hear. En route I shall ask a bit more about Aristotle’s method, and his ‘style’. 5

Osborne (1983, p. 402n.7) takes this to introduce a new point; below, I disagree. This is baldly put; what it does not say is that perception is ‘said in many ways’, vel sim. (compare and contrast the discussion of actuality and potentiality in perception, 417a9ff., 426a23-5, or the use of perception as an example of homonymy at Topics 1.15; and see Burnyeat (2002)): this need not be the claim that the expression ‘perception’ is ambiguous, rather than that the phenomenon of perception is not simple or uniform (compare the different ways of denying unity at Metaphysics 1017a3 ff.). 7 This must mean: ‘not in the same way as we see colours’, and not that we see light in a different way from seeing dark; see Ross (1961) ad loc. 8 The perfect tense here may recall the discussions of change, process and completion in de anima 2, especially chapter 5. 9 See, variously, Hamlyn (1968, p. 121); Kahn (1966, p. 50); Barnes (1995, p. 11 ff.); Osborne (1983, p. 401); Burnyeat (2002). 10 See e.g. Top. I.1, E.N.VII.1. One possible explanation of the shift from the first person plural here to the impersonal formulations of second-order perception might be that ‘we perceive that we see’ is taken to be one of the phenomena (as also the first person plural krinomen at 425b21) from which we should begin our investigation. 6

144

2. The snares of consciousness So: ‘Since we perceive that we see and hear’… do we? And when do we? When we see and hear, do we always perceive that we do? Or just sometimes?11 It has been argued12 that this passage does indeed start from the obvious: the obvious of the subjective side of perception. When I see, there is something that it is like for me to see – and this ‘something that it is like for me’ is itself perception.13 This conscious side of perception14 is a regular concomitant of ordinary acts of perceiving. So if I smell the baking bread, there is something that it is like for me to smell it; if I see the vivid pink scarf, there is something that it is like for me to see it; and so forth. I am sitting at my desk in the late afternoon, and the smell of buttered toast wafts through the door. When the smell hits me, it seems somehow or other to me, there is something that it is like for me to perceive the delicious aroma. My consciousness of the smell is not an inference from the smell (the inference is my muttering, as I leap from my hard chair, ‘Oh, good, teatime!’), but rather just the subjective side of the perceptual event,

11

There is nothing about the grammar of the present tense of ‘we perceive’ that determines the matter (and see below, §9, on a parallel passage in E.N. 9.9): both in English and in Greek the present tense can be aspectual: continuous (‘I breathe’); iterative (‘I get up in the morning’); expressive of a capacity (‘I play tennis’, ‘I solve crosswords’); conative (‘I fight off the Persians’). Kosman (1975) argues that Aristotle here is offering an account of perceptual awareness; Caston (2002, 2004) that for Aristotle perception is always complex, and reflexive; Johanssen (2005) that although the evidence to show that we always perceive that we see and hear is disputable, Aristotle here supposes that the complexity of perception needs to be explained by an inner sense. Connected here is the debate about what Aristotle might mean by a ‘common sense’ (see e.g. Kahn (1966), Osborne (1983)) e.g. at de somno 455a12 ff.: ‘…There is also a common faculty associated with them all, whereby one perceives (aisthanetai) that one sees and hears (for it is not by sight that one perceives that one sees; and one judges (krinei) and is capable of judging that sweet is different from white not by taste, or by sight, nor by a combination of the two, but by some part which is common to all the sense organs; for there is one sense-faculty, and one paramount sense-organ, but the mode of its sensitivity varies with each class of sensible objects, e.g. sound and colour)…’ [trans Hett (1964) modified]. 12 Kosman (1975). 13 Nagel’s famous formula (1979). 14 I shall call this what it is like to see, hear etc. ‘phenomenal consciousness’, following e.g. Tye (1995). On the various senses of ‘consciousness’ see Lycan (1996) ch. 1.

145

my side of the smelling that goes on: coextensive with the drift of buttered toast, but importantly my end of the business.15 Phenomenal consciousness might just be this subjectivity that marks me out as a perceiver, rather than as merely an object in the way of the buttered-toast-smell as it wafts. When we perceive, being conscious may be an irreducible and primitive feature of perceiving. Or consciousness may be intrinsic to perceiving, but susceptible to analysis or explanation (maybe I smell the toast, and at the same time I have some consciousness of the smelling, some perspective on it). However we account for the feeling involved in perceiving, it is distinct from something more reflective: the turning of my attention to what I perceive. I am sitting at my desk, and the argument I am trying to construct is stuck; and I think about what time it is – and turn my attention to the smell of buttered toast that my ennui makes compelling to me. Or someone asks me: ‘is that buttered toast or burned sausages?’ and I think about it (and am disappointed that it is sausages, not toast). This sort of reflection – reflecting on my smelling of sausages and toast – could be construed as something perceptual (although it need not: do I perceive my smelling of the sausages, or do I just think about it?). It may, further, mark a difference between the noticed smell of sausages and some unnoticed one, and so might be described in terms of consciousness. Nonetheless, reflection of this perceptual sort can be quite complicated (deliberate sniffing and attending); and it is not intrinsic to the first-order perception, which may frequently occur without it. So one of the broad lines of demarcation between phenomenal consciousness and reflective consciousness may be this: phenomenal consciousness should be a regular feature of ordinary perception, while reflection may be irregular, piecemeal, not necessary at all to the perceptions I ordinarily have. If Aristotle is indeed talking about phenomenal consciousness here, however, it might be a huge and exciting moment in the history of philosophy: the moment when it was first understood that there is something peculiar, special, strange about perception, from the subject’s point of view.16 Aristotle’s predecessors – so this moment would be described – missed what perception feels like and concentrated instead on how it comes about (notably by the affection of the perceiver by the perceived), while Aris-

15

Caston (e.g. 2002, 2004) takes Aristotle to account for consciousness in terms of the complex mental content of perception. 16 Aristotle himself may be making the claim to innovation at 426a20 ff; and compare 427a 21 ff.

146

totle, the forerunner of modernity, saw perception right.17 Is that correct? And is Aristotle, by contrast with his predecessors, taking (what we might call) phenomenal consciousness to be one of his phenomena? Or is he talking about something more reflective? 18 In what follows I shall suggest some reasons why reflectiveness may indeed be his focus of attention.19 I shall also suggest some reasons why we might not think that so dull. If this is about phenomenal consciousness, Aristotle must, I think, at least commit himself to the following conditions:20 • regularity: being (phenomenally) conscious of an act of perception is a regular, even if not a constant, feature of individual events of seeing, hearing etc.21 This just is what it is for perception to be conscious. So, here, ‘Since we perceive that we see and hear’ would translate as ‘since we regularly perceive that we see and hear…’ or, in Aristotelian, ‘since we always or for the most part perceive that we see and hear…’ • ‘what it is like for me’, part 1: when I perceive, there is a subjective or affective aspect of that perception, which distinguishes between this being a case of perception and its being a case where the perceptual quality of one thing is somehow directly transferred to something else.22 • ‘what it is like for me’, part 2: this subjective aspect is special to me. So phenomenal consciousness is (somehow or other) indexed to the subject who has it. Thus, here, ‘Since we perceive that we see and hear’ would parse as ‘since I perceive that I see and hear, and you perceive that you see and hear…’23 etc. Are these three features of phenomenal consciousness in Aristotle’s sights here?24 In what follows I argue that the regularity condition may not be 17

This is not Aristotle’s own account of his difference from his predecessors: see 427a17 ff. 18 See e.g. Kahn (1966). 19 Osborne (1983) takes a slightly different view of reflectiveness. 20 On consciousness in general see e.g. Tye (1995), Lycan (1996), Papineau (2002). 21 There are obvious tricky cases: just how conscious am I of the smell of buttered toast while I am concentrating on writing this sentence? 22 Here see the debate about spiritualism versus literalism, especially between Burnyeat (e.g. 1992) and Sorabji (e.g. 1992); and compare e.g. Caston (2005). 23 And you don’t perceive (in the relevant way) that I see or hear, nor vice versa. 24 I do not deny that Aristotle may sometimes be interested in something like consciousness; see, e.g., the cases where he contrasts alterations which ‘escape notice’ with those that do not (e.g. at Phys. 244b10-245a11, Sens.437a26-9) discussed by both

147

met; and that this may cause us to have some suspicions about how to understand Aristotle’s interest in what it is like for me to perceive. 3. The argument to ‘self-perception’ The construal of the argument is contested: here is just one account of it.25 1) We perceive that we see and hear (425b12). 2) We perceive that we see and hear either by sight26 or by some other sense (425b13). What follows is usually interpreted as an argument to resolve what seems to be the choice here (perception of sight is done either by sight or by some other sense) in favour of sight.27 This seems to be broadly right;28 but it does so by means of an interim conclusion, that when we perceive that we see or hear, it is a case of self-perception: of the perception’s perceiving itself.29 3) Sub-argument 1 (425b13-15): In either case (sc. when I perceive that I see),30 there will be the same sense of the sight and of the (firstorder) colour.31 Caston (2002, e.g. p. 757) and Johanssen (2005, p. 265); or compare Burnyeat’s account of perception (1992). But in these passages he uses the verb lanthanein to make his point and not the ‘perceiving that I see and hear’ locution. 25 Four notable, and different, detailed accounts are Kosman (1975); Osborne (1983); Caston (2002) and Johanssen (2005). In what follows my interpretation comes closest, in different ways, to those of Johanssen and Osborne. 26 Johanssen (2005, especially p. 246 ff., p. 252 ff.) argues that this must refer to the faculty of sight; Caston (2002, 2004) that we are talking here about individual ‘activities’. The Greek does not determine the matter, either for perception or individual sense-modalities: both aesthêsis and e.g. opsis can refer to the faculty or the activity. 27 Osborne disagrees (1983). 28 When Aristotle says, at 425b20, that ‘perceiving by sight’ is not one, this seems to follow from the argument about self-perception; it suggests, therefore, that selfperception happens somehow or other by sight (or whatever the appropriate sensemodality). 29 Notice the emphatic position of autê hautês at 425b15, and of autê tis estai hautês at 425b16. The reflexive, ‘self’, refers back to the perception, not to its subject: selfperception is perception of perception. 30 Osborne (1983) suggests that this applies to both options; not so e.g. Kosman (1975) or Hamlyn (1968: ad loc). 31 Why should the second-order perception include the first-order object? Osborne rightly observes that this is an easy consequence of Aristotle’s account of how (at least first-order) seeing and its object are actualised at once, in the sense-organ itself. Per-

148

4) So either there are two perceptions of the same object; 5) Or [there is one, and] perception perceives itself. 6) Sub-argument 2 (425b15-17): If indeed the perception of the seeing is done by some other sense than sight, either there will be a regress of perceptions (which is impossible); 7) Or the perception will be of itself. 8) So, when we perceive that we see and hear, this is self-perception.32 Sub-argument 1 focuses on the claim that any higher-order perception must have as its object not only the first-order perception, but also the first-order objects. Sub-argument 2 focuses on the claim that any higher-order perception must itself be either perceptible or perceived. Together they seek to show, twice over, that perception perceives itself. So either all or some perception is self-perception.33 Aristotle takes that conclusion to imply34 that higher-order perception is done by the same sense-modality: 9) [So when we perceive that we see, we perceive that we see by sight: and the same, mutatis mutandis, for the other sense-modalities]. The two sub-arguments exploit two assumptions, which themselves come under scrutiny in the sequel: i) manifest at 3, about the content of the perceptions: if there is a higher-order perception of a first-order seeing, the higher-order perception will include the content of the first-order one. Think about how it would be if this were false. When I perceive that I perceive, my first-order perception is both subject (of its own object) and object (of the higher-order perception). In perceiving that I perceive, however, I perceive at least the subjective aspect of the first-order perception (otherwise, it would not be perceiving that I perceive). Suppose, nonetheless, that I do not perceive what I perceive. My first-order perception’s being the subject of perception is explained by its being somehow altered by its object, on Aristotle’s view. So if ceiving the first-order subject must be perceiving it as seeing; so it will be perceived as actualised along with its object; see also Johanssen (2005, p. 243). 32 This is a gloss of ‘So we should assume the latter in the first place’: the interim conclusion, on the interpretation I offer here, from the thought that we perceive that we see and hear. 33 That this is problematic drives the dialectic of what follows. For the faculties of perception cannot be constantly self-actualising (417a3-5: see Burnyeat (2002, p. 38)). 34 Aristotle does not make 9) explicit; but he begins the next sequence with the assumption that we should be talking more generally about sight, 425b18. This may be consistent with the de somno passage (n.11 above) if both passages aim to deny that a sense-modality of a single order may both perceive, and perceive that it perceives.

149

the higher-order perception were to perceive the lower, without perceiving what it perceives, it could not perceive it as perceiving, would not perceive that it perceives. So it is plausible to suppose that the higher-order perception in some way includes the content of the first-order one. If the higher-order could be perception by sight, perhaps Aristotle invites us to think about this as though the firstorder seeing is somehow transparent, so that the colour it sees will be perceptible at the second-order.35 If seeing sees its proper object, higher-order perception perceives right through to the first-order object.36 Or if the first-order perception is understood in terms of change, then the change caused by the first-order object is somehow transitive to the second. (This is an assumption, then, about the objects of perception). ii) clearest in the second strategy, about counting the perceptions: if there is a higher-order perception of a first-order one, will we have two perceptions here? And if two, then perhaps three (if I perceive that I perceive that I see), or more.37 But this pluralizing is a bad thing (it doesn’t explain, after all, what the perceiving that I see and hear is): so we should block it at the first stage, and insist on there being just one, so that we may say that perception perceives itself.38 Call this (whatever it is of) the reducing assumption: 39 even if per35

Kosman’s objection (‘why from the fact that we see…that we see, should it follow that we also see …that which sees, and thus that…that which sees is coloured?’) rests on the prior assumption that it is phenomenal consciousness that is at stake here; the same may be said, perhaps, of Hamlyn (1968, p. 121). 36 This claim is itself as theory-laden as the ‘phenomenal consciousness’ interpretation, of course: for Aristotle, ‘seeing through’ would be understood in terms of perception as alteration of some kind. 37 If the arguments are parallel, as I suggest they are composed, the point about an infinite regress matches the point about there being two perceptions here – so pluralising, rather than an actual infinite regress, may be Aristotle’s real target here. 38 But: just one what? Perception in general? Sense-faculty? Activity of perception? Of course (see Barker (1981)), the discussion about faculties is itself parasitic on discussion about individual events, at least later in the chapter (e.g. 425b26 ff., 426a27 ff.); equally, the faculties are supposed to explain what happens in the individual cases; hence, perhaps, the instrumental datives e.g. at 425b13, 18. 39 That the reducing assumption is one focus of Aristotle’s attention is indicated by the apparent tensions in what he says: if higher-order perception is self-perception, perception is somehow just one; but in that case, also, somehow perceiving by sight is not one. If Aristotle writes with care, this tension is part of his argumentative strategy, not just a gap in his argument.

150

ception is ordered, it is not thereby merely plural; instead, the ordering of perception amounts to self-perception. The reducing assumption appears twice: in the choice of selfperception over there being two perceptions of the same object (at step 4); and over there being a regress of perceptions (at step 6). The symmetry of the two arguments suggests that steps 4 and 6 somehow deploy the same assumption: but the nature of the regress in step 6 is vigorously disputed.40 Is Aristotle’s claim that any perception (of any order) is also actually perceived – so that of any perception there is a higher-order perception? This would invite a regress, unless it is blocked by some self-perception. Or is the claim that any perception is also perceivable – so that of any perception there may be a higher-order perception? This – it has been suggested41 – may not invite a regress at all, but would just peter out when the higherorder perceiving happens to be done (so it supports no inference to selfperception). If the latter (perception is perceivable) is too weak to force the regress, then we may prefer to find the former here (perception is perceived) – after all, Aristotle does indeed conclude that perceiving that we see and hear, since it cannot be regressive, implies that perception perceives itself. But if his point is that perception is perceived, then self-perception will accompany all (or most) perceptions. That favours an interpretation of the argument as a whole that has it talk about the regularities of phenomenal consciousness. It favours, also, the view that the topic here is some individual event or activity of perception, rather than the faculty of perception, or the individual faculties of sight and hearing.42 For if any given event of perception is perceived, then (generalising) all perceptions will be perceived; and the inference to self-perception will generate an account of phenomenal consciousness. If, on the other hand, the chapter is more concerned with faculties of perception, it may allow the mere possibility of perception’s being perceived; and then we may need to account otherwise for this argument’s purpose.43 40

Notably by Kosman (1975), Caston (2002, 2004) and Johanssen (2005). Kosman (1975) and in his response to Johanssen (2005). 42 Caston (2002, 2004). 43 See Johanssen (2005, p. 245 ff.) on how we should construe the claim that ‘if some sight is going to see the seer, the first seer will have colour’ at 425b19, reading horôn twice; and, differently, Caston (2004, p. 524 ff.), who compares the Charmides, on which more below. 41

151

The suggestion that Aristotle is here concerned with actual individual perceptions sits uneasily with the place of this argument in de anima 3 – immediately after a discussion about how many sense faculties we have, and before an extended, and parallel, account of the faculties of imagination and thought.44 Further, both modifications of the puzzle in the immediate sequel turn on cases where the first-order perception is somehow missing (judging that it is dark, and retaining sense-impressions after their object has gone). These cases are taken to show that ‘perception is not one’. But that cannot then mean that individual actual perceptions are complex in the required sense, since in the cases in question that complexity is expressly missing. Instead, it implies that the faculty of perception in question is multiform: that is, we may call various different things ‘perception’, including first-order and second-order perceptual events. The generalisation that is made here, therefore, is about perception and its nature, not about all actual perceptions. This, in turn, may incline us to read the dominant argument of the passage as attending to the nature of the faculties of perception, multiform though they may turn out to be. But then how does the regress work? There will be, of course, individual actualisations of the faculties of perception; and of each of these we can say, not that it is perceived, but that it is perceivable (at least this, after all, is warranted by the premise ‘since we perceive that we see and hear’ however it is parsed). Indeed, that is a natural point to make in the context of the view, explicitly recalled in this passage (425b22), that perception is somehow an affection by the perceived as such (supported by the thought that first-order perceptions are either transparent, or transitive, to higher-order ones: see 426a2). For perception comes about by the affection of the perceiver by the object; however queer that affection may be, it is still a change of some kind in the perceiver. As such, it is quite reasonable to suppose that, for example, what is ‘coloured in some way’ is perceivable in some way (just as what is coloured is perceivable: Aristotle certainly does not maintain that everything coloured is perceived). That is, the very affection that is perception is perceivable; and if so, there must be a faculty that perceives it. But that faculty, whatever its sense-modality, will be affected by its object when it perceives, and so be

44

The datives at 425b13 and 20 to describe perceiving ‘by sight’ are most naturally read as instrumental or explanatory: and hence as referring to the faculty, not the act. Compare the point made at 426a13 that seeing is the activity of sight (opsis).

152

perceptible in its turn.45 If this chain of perceptible things is not to stretch to infinity – if, that is to say, it is to be explicable by positing faculties of perception – then some faculty of perception must be able to perceive itself. So the regress may work, I argue, on either construal of ‘since we perceive that we see and hear’: whether always or sometimes. Either way, the regress needs to be blocked by a reflexive: some perception perceives itself. That may be just as well, perhaps, to sustain the thrust of the opening chapter of the book, that there are only five senses. But it is problematic for Aristotle, severely exercised as he is by reflexivity. If an action is reflexive, and strictly so, it seems to be cause and caused at once – and that makes a nonsense of causation. 46 If a faculty is reflexive, it both explains its actualisations, and itself is one of its actualisations (and so needs explaining). Yet Aristotle posits sense-faculties to explain our ability to perceive; if those faculties are somehow reflexive, are they both explanantia and explananda at once?47 If we are to avoid the risk of regress, some account must be given of self-perception which is not so strictly reflexive as to be impossible. And yet the argument as Aristotle presents it may seem to emphasise that strictness. For there is, at first at least, nothing about the self here; and nothing, therefore, about the subject of the perceiving as such nor about what makes (any of) these perceptions mine.48 Instead, the argument is resolutely impersonal, marshalled around the two relata in perception: the object of perception (strategy 1) and the perception itself (strategy 2). Is this selfperception what we might call consciousness?

45

The interpretation I suggest, then, is a hybrid: it worries both about a regress of faculties and about the nature of individual events of perception. Quite right too – for Aristotle talks of faculties in order to explain the events. 46 Compare de anima 417a2 ff., or the worries about self-motion at Physics VII.5. 47 Johanssen is surely right to compare the Third Man Arguments of the Parmenides (2005, p. 244 ff. and note 16), which support two different regresses – one where the higher-order perception is the same sense-modality, the other where it is different. 48 The argument, notably, does not rest on any claims about the first-person features of this self-perception: the two personal verbs (at 425b12 and 425b21) seem to introduce, first, the phenomena to be explained, where the explanation notably lacks any indexing to persons; and, second, some additional data about our responses to light and dark.

153

4. A puzzle Whatever it is, Aristotle now confronts it with a puzzle – but a puzzle too condensed for clarity. Aristotle provides responses to it which consider both reduction (at lines 20-22); and transparency (at 22-25): but the point of the responses remains disputed. We might be forgiven for complaining about gappy lecture notes, once again. The puzzle runs like this: If perceiving by sight is seeing, and what is seen is colour or what has colour, then if some sight is going to see the seer, the first seer will have colour. Is the point that it is manifestly silly to think the seer is coloured? The puzzle would then be a parasite of transparency. Is it a satisfactory riposte to say, merely, that ‘that which sees is coloured in a way’ – as Aristotle does four lines later? Or is the puzzle focussed, as the preceding argument suggests, on self-perception and the problem of reflexivity? It might, thus, turn on the tension between reducing the perceptions we have here and making some sense of the paradoxes of the reflexive (how could something act on itself, without qualification?). But how reasonable a response to that is what Aristotle says next, that ‘this makes it clear that perceiving by sight is not one’? And how, anyway, does the puzzle fit into the economy of the chapter as a whole, or into its place in the de anima? How do the puzzle and the responses to it modify what Aristotle has said about self-perception? One might think that both responses – that ‘perceiving by sight is not one’ and that ‘that which sees is coloured in a way’ – rather fudge the issue. If the puzzle is how self-perception can occur, Aristotle’s solutions might seem to turn somehow on denying that this is self-perception in the strictest sense, and allowing merely that there are some loose ways in which perception perceives itself, and which are not vulnerable to the puzzle. Thus, if perception by sight is ‘not one’, then perhaps there are two quite different sorts of perception involved in our perceiving (sort 1) that we see (sort 2).49 The result, then, would be a quasi-self-perception (sort 1 perceives sort 2) fit to meet the reducing assumption only by sleight of word. Or if it is somehow objectionable that the primary seer is coloured, 49

Perhaps sort 1 sees the perceiver as perceiving, but sort 2 sees its object by virtue of a sheer affection of the sense-organ. On this see e.g. Caston (1998, p. 280).

154

again in some strict way, the transparency assumption could still be met by conceding that the seer is coloured in some way that is other than strict. I shall suggest that the qualifications Aristotle enters in answer to the puzzle are less disappointing. 5. The history of the puzzle Aristotle’s puzzle, however, is not original to him.50 In Plato’s Charmides (167-9), Socrates offers a tricky argument that there can be no perception of perception, and thence no perception that is ‘of itself’ (167d; 168d-e).51 From this he infers the implausibility of Critias’ claim that the virtue of sôphrosunê is in fact self-knowledge: and of his further claim that selfknowledge is reflexive: knowledge of knowledge, knowledge of itself.52 But this is rather a queer thing for Socrates to do. After all, he himself has a high stake in some sort of knowledge of knowledge, or at least in higher-order cognitive attitudes about knowledge. How else could he claim that he is wiser than others in knowing that he does not know? How else could he explain the reflective stance provided by the Socratic elenchus, the investigation of what people claim to know (expressly brought to mind at 166d)? His argument’s trajectory needs to be redrawn: especially, it needs to take account of the connections between knowledge and virtue. The problem (one of the problems…) here is how both knowledge and perception are construed in terms of a direct causal relation between subject and object: a relation that is both asymmetrical and intransitive.53 If it is asymmetrical, then there can be no reflexive relation between subject 50

This is regularly observed. e.g. Sorabji (1979, p. 49 n.23), Caston (2002, 2004, p. 524), Sisko (2004), Johanssen (2005, p.248). My contention is that Aristotle is engaged, not just with the short argument of Charm.167b-169a, but with the projects of the dialogue as a whole. 51 The argument as a whole is about knowledge, to which perception comes in as an analogue. It began, at 165c, by discussing knowledge of oneself; then ostentatiously shifts to a discussion of knowledge of itself (166c3). See McCabe (2007). 52 The marked impersonal formulae in the Socratic argument are mirrored in Aristotle’s response. 53 Asymmetrical – if x perceives y, then x is the perceiver, y the perceived, and as such the relation is not reversible; intransitive – when x is the perceiver and y the perceived that exhausts the perception relation, and x cannot see through y to z. Socrates’ argument is designed to show that if perception is like this, it cannot illuminate knowledge. Aristotle, on the account I give below, goes further – if perception is like this, it cannot account properly for the richness and complexity of perception.

155

and object; and if it is intransitive (and still directly causal), there can be little to be made of the content of a higher-order attitude to a lower-order state of mind. So perhaps Plato’s cunning plan in the Charmides is to expose this impoverished account of the subject-object relation, for knowledge at least; and to suggest that, in order to have both higher-order and reflective features, it must be understood as a far richer cognitive state. Knowledge, then, needs to be something like understanding, or wisdom; it needs to be reflective and broad based. Its reflexivity, then, will be modified and explained as reflection, where knowledge is not of itself, stricto sensu, but rather, higher-order knowledge has lower-order knowledge within its content. This suggestion affects knowledge’s analogue, perception, too: either perception needs to be understood in a similarly rich way (as highly cognitive, reflective, unlike a raw54 relation between subject and object); or it ceases to be an analogue for knowledge. There is a repeated use of perception as a reflective attitude in the Charmides, which suggests that it is richer than raw. But these suggestions appear in the frame dialogue: in the discussions between Socrates, Charmides and Critias about how the discussion should proceed.55 By contrast, the direct argument, discussing both hearing and sight, comes twice to a halt. On the first occasion, Critias agrees with Socrates that there is no sight: A: … which is not a sight of the things other sights are sights of, but is sight of itself and the other sights and likewise of the non-sights, and which although it is a sight, it sees no colour, but sees itself and the other sights. (167c8-d2)

On the second occasion, they agree that: B: …if it is going to see itself, [sight] must have some colour; for something colourless, sight could never see. (168d9-11)

Critias and Socrates, then, find it hard to fit their search for reflexivity (manifest in A) to their conviction that there must be some kind of transparency condition (if sight sees sight seeing, it sees also the colour it sees). As a consequence, either perception is not a proper analogue for knowl54

‘Raw’ may be about the feel of it (I smell the buttered toast in some irreducibly subjective way) or about its causal structure (the buttered toast somehow impinges on my sensation directly). The subject-object relation in which Socrates is interested involves the second kind of rawness. 55 Notably, at 159a; 167c; 168e.

156

edge; or it must be re-described. But nothing here suggests that higherorder perception – if it can be made to make sense at all – is intrinsic to, or necessary for, lower-order perception. So if we perceive that we perceive, on this account, this is not phenomenal consciousness. On the contrary, if perception can be re-described to account for the perception of perception, it should provide an analogue for wisdom – for that hard-won virtue that is the goal of philosophy. To perceive that we perceive may, therefore, be a hard task, not the natural occurrence of consciousness. 6. Aristotle reads Plato Aristotle offers a direct allusion to the Charmides.56 Compare, to A: A* But the same sense will be of the sight and of its underlying colour (425b1314);

and to B, the last two clauses of the puzzle: B* .. then if some sight57 is going to see the seer, the first seer will have colour (425b17-20).58

Like Socrates and Critias, Aristotle denies that there can be any seeing that is not seeing a colour. Unlike Socrates, who purports to solve the dilemma at this stage of the dialogue by denying that there can be any per-

56

Notice that Aristotle and Plato both start with the same two example senses – sight and hearing. 57 This construal, if there is an allusion to the Charmides going on here, is explained by opsis tis at Charmides 167c8. 58 That A* is an allusion explains several things: the thoroughly condensed form of Aristotle’s remark; the difficulty in identifying its subject (does this refer to both of the options canvassed in the previous sentence, or just one?); the emphatic ‘the same’ at 425a13, where the Platonic background shows that denying transitivity is absurd. In B*, the quotation is very close: notably, the verb forms (‘will see’, opsetai); the word order of the first clauses; in the second clauses, the emphatic position of ‘colour’ (chrôma) at Charmides 168d10 and de an. 425b19; and the shift from Plato’s ‘it must have’ to Aristotle’s ‘it will have’. By ‘allusion’ I mean that Aristotle intends his reader (some of his readers) to notice the connection, and that the dialectical relation with Plato is important to the development of Aristotle’s point. Surely (some of) Aristotle’s readers were, like him, careful readers of Plato (see Halliwell, 2006)?

157

ception of perception, any self-perception,59 Aristotle characteristically goes with the phenomena: we do perceive that we see colours and hear sounds. He infers from this that there is self-perception; and from this that perceiving by sight is ‘not one’.60 Aristotle’s puzzle is, I suggest, a reading of the Charmides’ argument, a reading of it that takes its conclusions not to be compelling, but problematic (just, I would say, as Plato set it up). I shall suggest that we have good reasons to ask what else Aristotle shares with the Charmides: in particular, the question whether perception is, or is not, analogous to knowledge, in circumstances where what is required of knowledge is that it be reflective. For – to repeat – what is not at issue, in the argument which Aristotle inherits, is phenomenal consciousness.61 If that were what it is to perceive that we perceive, it would not even begin to illuminate what Socrates wants to say about knowledge, about reflection, and about philosophical inquiry; nor would it bear any comparison with the context in which Socrates raises these issues – the discussion of virtue. Someone might complain that to suppose that this makes a difference to how we read Aristotle is to suppose that Aristotle reads Plato carefully and, as it were, dialogue by dialogue. On the contrary – the objection would run – Aristotle takes Plato on as a series of doctrines (about forms, about the good… and about perception) with which he engages, and which he rejects.62 So Aristotle could not care less about the overall structure of the aporia in the Charmides, just so long as its elements can be exploited in puzzles of his own. His engagement with the (many and the) wise is satisfied just by having in mind some version of some view which Plato might have held; neither the historical nor the textual accuracy of the views in question matters a bit.63 So Aristotle’s allusion to the Charmides would not show that he shares that dialogue’s background assumptions nor its dialectical structure: the form of words would tell us nothing about the shape and structure of Aristotle’s own argument. 59

This solution, however, may be what Socrates needs to avoid, in the outcome of the dialogue as a whole; he must at least be able to give some account of higher-order knowledge. 60 So the Charmides is in the background of Strategy 1 of the argument: Aristotle assumes what Critias denies, that higher-order sight is transitive. 61 Contra Kosman (1975, p. 517). 62 This is sometimes manifestly the case: see e.g. Fine (1993); but my claim is that direct arguments against Plato do not exhaust Aristotle’s interactions with the dialogues. 63 See the account of the starting points for dialectic in Metaphysics B1 and Topics 1.1, 10, 11.

158

This objection seems to me to make a broad assumption both about how the dialogues are written and about how they were read in the Academy within Plato’s lifetime, which is unjustified. Plato himself does not give us grounds to make any such assumption, nor to suppose that some bits of the dialogues are dispensable, others not. If his silence on this matter was – as I see no reason to doubt – deliberate, then Aristotle may not have been brought up to think the dialogues merely vehicles for doctrine, whose elaborate dialectical and dialogical structure is of no philosophical interest or importance at all. Quite the contrary: if Aristotle was taught by Plato how to read a Platonic dialogue, he was surely better at reading it than we are. And if the detail of the dialogue matters to the argument, we had better not suppose in advance that Aristotle did not see that, nor exploit it to his own advantage. So we cannot assume that Aristotle read the Charmides but ignored its reflections on the reflective character of knowledge when he comes to his own discussion of self-perception. No more can we assume that allusions thus detailed to Platonic originals would be lost on his audience. But now this allows to Aristotle the choice that Socrates puts before us. If perception is analogous to knowledge,64 and if it is conceived as a strict (asymmetrical, intransitive) relation, then there can be no reflexive perception (no perception of itself) and no reflexive knowledge. But knowledge (or, better, understanding) needs to be, if not reflexive in this strict sense, then reflective. For understanding needs to have a higher-order reflective dimension, where the first-order knowledge is its intentional object (if I know that I know, then part of the content of my higher-order knowledge is the content of the lower-order one; another part of its content is that I know that lower-order content). So either perception is not analogous to knowledge; or perception is not the strict relation conceived by Socrates’ argument. The Charmides read as a whole, I claim, suggests the latter; although it is easily read as proffering the former. Moreover Aristotle, in his readings of the Charmides, takes as his starting point the latter – that perception is not a strict, exclusively first order, relation, but is capable of higher-order dimensions. This, I suggest, is what he means by starting: ‘Since we perceive that we see and hear…’ At the start of the chapter he is already dialectically engaged with the Charmides: and it is from this engagement that he takes himself to be entitled to conclude that ‘perceiving by sight is not one’. I shall suggest that his engagement with the Charmides runs deep – to the ethical base on which Socrates’ argument is set. 64

For a subtle account of how this works in the peculiar 3.7, see Osborne (1998).

159

7. Aristotle’s responses to the puzzle Aristotle resists the aporia to which Socrates and Critias seem to be reduced. He does so by supposing that self-perception is proof against the apparent puzzle, that ‘if some sight is going to see the seer, the first seer will have colour.’ His two responses to the puzzle, therefore, should amplify just what he takes self-perception to be. Well, then, it is clear that perceiving by sight is not one.65 After all, when we do not see, it is by sight that we judge [krinomen]66 both the darkness and the light, but not in the same way (425b20-2). And again, that which sees is coloured in a way; for each sense-organ is receptive of what is perceptible without the matter (425b22-24).

The two responses seem pretty haphazard, at first, and still more evidence for gaps in Aristotle’s notes: how exactly does the second follow on from the first? And how does the first respond to the problem of selfperception? First, the first: to avoid the aporia, we need to say that ‘perceiving by sight is not one’. This is confirmed by the point about what happens when we don’t see (because it is dark) but nevertheless judge that it is dark. This, like the perceiving of light, is done by sight, even although it is not seeing in the way that we see colour. How does Aristotle’s point work? Consider the contrast between cases when we are blindfolded, so that it seems dark; and cases when we are able to see, but see nothing, so that we judge that it is dark. In the second case, unlike the first, the judgement is made – or so Aristotle suggests – by virtue of the sense-modality of sight: we survey, as it were, the nothing that is before us, and judge ‘Cripes, it’s pitch-black’. The same happens for light, Aristotle suggests (light, after all, is not a special object of sight): we are seeing, and there are

65

See above, note 6: this is not about linguistic issues (ambiguity), but about the real structure of things. Notice the later claim, 426a15 ff., that although the actuality of the perceived and the perceiver is one, their being is not: by the next stage of the chapter Aristotle is talking about the real complexity of perceptual events (compare de sensu 449a5-20), even if that complexity gives rise to linguistic error (426a26). 66 This expression appears also in the parallel passage from the de somno: see above n. 11.

160

things that we see,67 so that we judge that it is light – we say ‘Aargh, that’s bright!’. In both cases (darkness and light) we are said to ‘perceive by sight’, even although in neither case do we do so exactly by perceiving sight’s special objects. The expression ‘perceiving by sight’, therefore, includes this kind of judgement in its scope. And because it does so, perceiving by sight is not narrowly restricted to the direct perception of sight’s special objects; instead, we might say, that the general category of things that come under the description ‘perception by sight’ is ‘not one’.68 Aristotle’s point, I have argued, should be, not that individual perceptions are complex, but that perception, the faculty, is multiform: but what exactly does that mean? And how are these observations about perceptual judgement à propos?69 Maybe Aristotle distinguishes between sight proper (first-order seeing of the special objects of sight: my seeing this fuchsia pink, for example) and sight in some derivative sense – sight improper, where sensible judgements take the place of direct vision. On such an account, he only partly assimilates perceptual judgement to perception, allowing that the former can be called perception, but only in some etiolated way. This might be the point of his saying that perception is ‘not one’: he means us to understand that the expression ‘perception’ can refer to both perception proper, and perception improper, but not in the same way. ‘Perception by sight’ on that account, is ambiguous, it is ‘not one’ (and self-perception is then only of the ‘quasi’ sort).70 A passage from the previous book of the de anima might give us pause, however:

67

This elliptical sentence should, I think, read something like this: ‘when we are not seeing, it is by sight that we judge the darkness; in just the same way, we judge light (when we are seeing) by sight, but in neither case do we perceive in the same way as we perceive colours.’ We see both darkness and light in the same judgemental way (but see Johanssen 2005, p. 250 n.27). 68 The de somno (n.11) limits the point: perception by sight is not an ungoverned plurality, but held together by the ‘common faculty’ of perception. See Kosman (1975, p. 518) on understanding perception in terms of the whole organism. 69 This question, in my view, tells against Caston’s view that the chapter shows that conscious perception is a complex of the perceived object and our reflexive consciousness of it: in the case of darkness the judgement occurs in the absence of an object, albeit the presence of our ability to see. Instead the comparison with light and darkness amplifies the different and independent ways in which we may be said to ‘perceive by sight’ and decouples perceiving that we perceive from first-order perceiving. 70 See above, §4.

161

Now sight is both of the visible and of the invisible (for darkness cannot be seen, but sight judges that too), and again of what is too bright (this too cannot be seen, but in a different way from darkness)… (de anima 422a20-23).

Here Aristotle imagines a case where we can see by virtue of the fact that our faculties are working perfectly well; but we see nothing – so we judge that it is dark. He follows that up with a different point about light from the one he makes in 3.2, suggesting that excessive brightness is destructive of the faculty’s ability to see. The passage turns, then, on the contrast between cases when sight works, and cases where it does not; in the former, we should say even-handedly that sight is ‘both of the visible and the invisible’: seeing the invisible – judging that it is dark – seems to count as sight in a parallel way to seeing the visible,71 and to be properly contrasted with cases where the sense faculty is not working at all. The first clause of this sentence, therefore, does not suggest that sight’s operation in some cases of what cannot be seen is somehow etiolated, or metaphorical, or that the cases where sight manifestly judges are not proper cases of sight at all. Nor, indeed, would such a solution do justice to Aristotle’s earlier argument in 3.2, which insists that there is self-perception. For if this were the explanation of that phenomenon, it would only be self-perception by virtue of an ambiguity. If ‘perception’ were ambiguous between the proper and the improper sorts, and both sorts were in play in perceiving that we perceive, we would need to avoid the sophistical difficulties, and say that this is a case where I perceive-improperly that I perceive-properly. That is not self-perception at all (‘quasi-’ won’t help). Instead, perhaps, we should read the suggestion that ‘perception by sight is not one’ rather differently: as a warning that perception works in several different ways – a warning that Socrates and Critias would do well to heed. But if this is to avoid the complaint that it will only explain quasiself-perception, Aristotle needs to show at the same time that perception is somehow unified. Suppose he means the comparison with the perception of dark and light to suggest that perception generally involves judgement (as it obviously does in the case of dark and light). Suppose, further, that he intends, not to draw a contrast between perception-proper and perception-improper, but to show us that what holds of the most complicated cases holds also of the cases we might think simple, or perception-proper. I see this fuchsia 71

This, surely, is the effect of ‘sight judges that too’, krinei de kai touto hê opsis, 422a21.

162

pink before me. This might be quite basic, a seeing of fuchsia pink. But how does it relate to cases when I see that this is fuchsia pink; or when I see that this fuchsia pink is distinct from that burnt umber; or, in darkness, where I see no fuchsia pink at all? Perhaps – by virtue of Aristotle’s argument that the last case is indeed a perception by sight – we might come to agree that all of these cases are perception by sight; and that although they differ in various ways (for example, in the case of darkness there is no first-order seeing going on; and in the case of perceiving that I see, there is also second-order perceiving going on) they all count as genuine examples of perceiving by sight.72 Then, they are all indisputably perceptions proper (‘perception by sight’ is not equivocated) even although they differ in important ways. Now, so far from this separating off the basic case (I see fuchsia pink) from the rest, including the higher-order ones, the basic case turns out to have after all the essential features of the higher-order ones: they are all to be understood in terms of perceptual judgement, even if they vary widely in the details of their content. Perception, when we perceive that we see, has a complex object, as Aristotle’s argument brings out (both the first-order object and the firstorder subject: I perceive that I see fuchsia pink). But then perhaps the basic case works in a parallel way. When I see fuchsia pink, I see that it is fuchsia pink: the basic case can look as much like a judgement as the case where I contrast something fuchsia with something burnt umber. Instead, that is, of the perceptions that are clearly judgements (comparisons, for example, or higher-order perceptions) being perceptions-improper, we might say that they are all both proper and complex – and eschew the thought that there is something proper only about cases of the apparently basic kind. In that case, on the account that Aristotle offers here, we should count as ‘perception by sight’ all sorts of things, including the perception that I see; and we should do so by virtue of the fact that perception has judgemental content from the lowest level up.73 This, however, makes perception multiform, since types of perception can differ both in their content and in their order. Aristotle’s amplification of ‘perception by sight is not one’ makes it clear just how. 72

They may also have a phenomenology, awareness (see Burnyeat 1992). This does not imply that perceptual awareness is what is described by ‘perceiving that we see and hear’. 73 If perception in humans is continuous with that of animals, are we then to say that animals judge? or are we to allow that perception changes its content as it goes up the scala naturae? On this see e.g. Burnyeat (1992), Sorabji (1993) part I.

163

First, he compares (by the train of thought that runs from 425b19-20 to 425b21-2) the business of higher-order perception with the judgement involved in cases like seeing darkness. Both count as perceiving by sight; but they differ in order. He also compares seeing darkness with seeing a colour. Both count as perceiving by sight, but they differ in content because their objects are formally distinct. And yet all three (perceiving that I see, seeing darkness, seeing a colour) count somehow as perceiving by sight. If they are to be comparable at all, then first-order sight of colour needs to be, I have suggested, cognitive in content in something of the same way as seeing darkness and as perceiving that I see. But if they are distinct, as Aristotle also urges, there is a formality to their distinctness – they may be different in order; or different in their relations to the special objects of the sense modality in question – and still count as perception by sight. I will see colour under different conditions than I perceive darkness by sight; and I may see colour without perceiving that I do so: ‘perception by sight is not one’. Second, if perception by sight is a case of being affected by the object of perception, but this allows for cognitive content too, then we shall need to modify the sense in which we may say that the eye is coloured.74 The basic example needs to be made more sophisticated, and to allow that its content is cognitive (hence the modifications suggested at 425b23-5) while still retaining the central thesis, that sight in the first place occurs when the eye is affected by colour. This may be the point of Aristotle’s modification: ‘that which sees is coloured in a way.’ If this account of perception is the outcome of Aristotle’s argument about perceiving that we see, it fits well into the chapter as a whole, which brings out two aspects of perception thus understood. On the one hand, it occurs when the perceiver and the perceived are actually unified (425b26426a26): the realist cast of Aristotle’s account of perception is not damaged by insisting on perception’s cognitive features. On the other hand, what it is to perceive at least includes cases that are readily admitted to be perceptual judgments (for example, seeing that it is dark, 425b21; perceiving that something is both sweet and white without assimilating the two features, 426b12-21). It is for this reason that a large part of the rest of the 74

Hence, ‘that which sees is coloured in a way’, (425b22). Does this imply a literalist account of perception? No – all we need, both for Aristotle’s puzzles to get under way and for his solution to be consistent with what he says elsewhere, is that perception be sufficiently the same, whatever its level (so not raw at the first level, cooked at all the others).

164

chapter is concerned with understanding just how a perceptual judgement can be both complex and unified75 – the dual condition on its being as much perception as the first-order, basic case, which is a single event at a single time.76 Thus the perception of a concord or harmony is a single ratio (426a27-b7); the judgement of two sensibles as different must be done by a single faculty (426b20-1); and judgement may sometimes use the same element twice (as the point may appear twice in our definition of a line: 427a9-14). If perception is cognitive, that is to say, it is unified in judgement, not merely by the physical unity of the object with the subject, but by the conceptual, cognitive, unity of the perception itself. Where, now, does that leave my original question about whether this chapter is about phenomenal consciousness? The burden of Aristotle’s argument here is not to show the way in which the subject is aware, when it perceives, that it perceives. Instead, he emphasises the features of perception that will distance it from phenomenal consciousness: its judgemental, cognitive content. He also implies – in insisting that perception by sight is not one – that we should not expect that the different conditions of perception he describes in this chapter – perceiving that we see; judging difference – all occur in every instance of perception, or even just normally. Instead, perceiving that we see, the case which drives the argument against the singularity of perceiving by sight, should not be understood as a regular concomitant of first-order perception: it is not phenomenal consciousness. What, then, is going on here? 8. Some more Plato The question of the relation between perception and judgement is not original to Aristotle: it comes, instead, from the other Platonic dialogue in the background of these chapters of the de anima, the Theaetetus.77 There 75

The unity and complexity of perception is reiterated in the later stages of the chapter: the discussion of the perception of harmony (426a27-426b7: see Barker (1981)); the account of judgement done by one sense (426b8-14) and of judgement across senses (426b14-29); the puzzle about how this can be explained in terms of opposite motions (426a29-427a6); and the resolution in terms of actuality and potentiality (426a6-16). 76 See 426b28 ff. on the impossibility of some perceptual event’s being two opposite movements. 77 Verbal echoes include the introduction of Plato’s (term of art?) koina (from 425a6: compare Theaet. 185b8); both philosophers’ interest in sight and hearing as their example senses (compare Theaet. 184b10 ff.); and the shift from sight to the question of

165

Socrates examines a theory of perception (a theory associated with Protagoras, to whom I shall return) which is a raw relation between object and subject; and suggests that it cannot explain all of our attitudes or mental states, such as our judgement of the common terms (such as ‘same’ and ‘different’ or ‘good’ and ‘bad’). So … In respect of being, and that they [sc. perceptible things] are, and their opposition to each other, and, again, the being of this opposition, the soul itself, rising up78 and comparing them, attempts to make a judgement (krinein) for us. (Theaetetus 186b)

Socrates suggests that if perception is construed as a strict relation between object and subject, judgement will be done by the soul. The judgement itself is comparative, a weighing up of the relations between perceptible things, and reasoning abut those relations themselves. When the soul makes a judgement, however complex – as a famous later passage in the Theaetetus shows – it is involved in a ‘silent dialogue’. It scrutinises opposing views put to it (by reason, or by perception79), listens to each side, and itself comes to a decision: this is a belief (Theaetetus 190a) or a judgement (Philebus 38c).80 This account of the soul’s activity has a striking feature: its detachment. As the dialogue goes on in the soul, the soul itself stands aloof from the opposed views; only after consideration (Theaetetus 190a2-3) does it come to its judgement. So while the perception delivers reports, the soul considers them as if from the outside;81 this difference in stance separates perceiving by taste, e.g. something salty, (Theaet. 185b10, de anima 426b5). There are structural similarities, too: de an. 2.12 and the discussion of phantasia in 3.3. tackles the theory of perception; 3.1 tackles the unity of the soul and the common terms; and the theme of falsehood, which becomes more and more important in the early chapters of de anima 3, has the same emerging significance in the Theaetetus. This background is frequently observed; it often explains the lacunosity of these difficult Aristotelian chapters. 78 Compare the importance of the soul’s view of things outside the cave at Republic 516b. 79 Compare Philebus 38c ff.; Republic 523-5; Sophist 263e. 80 The ‘silent dialogue’ is designed, surely, to show the structure of the central case, even if some judgements may be far more exiguous. 81 Notice the odd echoes in the two models of the mind offered later in the Theaetetus: the wax tablet (191c ff.), which has to be held up and manipulated by someone other than it (notice 191d4-7); and the aviary (197c ff.), outside which there is the birdcatcher, in whose head most of the business of mistaking must occur.

166

perception from judgement. According to the Theaetetus, then, if any judging goes on, it is not perception that does it: perception would be a simple phenomenon, and self-perception would be ruled out.82 The soul’s activity, contrariwise, is something that philosophy promotes: there is a direct connection, that is to say, between the reflective stance of the soul when it judges, and the progress of the philosopher towards god (Theaetetus 176b). Aristotle dissents, up to a point. For him, perception is more than a simple raw relation, since it includes judgement, in various different ways. But it is not a mere collection of different phenomena, either, since it includes self-perception in some proper sense. This further suggests that the business of perceptual reflection is a continuum of cognitive activities, based on first-order perception of the special objects of a sense, but capable of greater and lesser reflective capacity. So in some measure the argument revises the reducing assumption with which it began by enlarging the scope of ‘perception by sight’. But if – in making this claim – Aristotle picks up not only the terminology of the Theaetetus, but its train of thought as well, then the judgement in question, now claimed to be part of perception, should exhibit the properties which Socrates there allocated to the soul. In particular, the perception that is self-perception should have the quality of detachment which Socrates wants for the soul: it is by this very detachment that reflectiveness (soul’s judgement, for the Theaetetus, perception’s judgement for the de anima) is ensured. Where then does that leave the second response to the puzzle: that ‘that which sees is coloured in a way’? For this, in its turn, invites the regress to run: is Aristotle’s account of self-perception sufficient to block it? The arguments for self-perception imagine that the first-order object (the colour, for sight) is perceptible both by the first-order and the second-order perception; the higher-order perception by sight sees through to the special object itself. But if the higher-order perception is reflective, this transparency may be modified; perhaps instead, the higher-order perception is thought of as somehow looking at the first-order perception, because that is ‘coloured in a way’. This ‘looking at’ might fit the thought that what the higher-order perception does is judging: for judging, like looking at something, supposes that the judge, or the looker, is somehow detached from

82

Much of the language for the soul’s detachment is perceptual: looking at, surveying, inspecting, etc. This should make one hesitate before committing Plato to a raw theory of perception.

167

what it sees.83 Still, on this account, it is perception; and can itself be the object of reflection. Reflection, however, is broad enough to include itself. Judgements, after all, can be self-referential without doing violence to the thought that nothing can be cause and caused, nor explanation and explained, at once. If self-perception is a judgement, the regress can be blocked. But the two responses to the puzzle are now seen to be connected, in amplifying the perceiving aspect (rather than the object aspect) of higherorder perception: higher-order perception considers, looks at, inspects, the lower-order and its content. Higher-order perception is detached. But it is still perception: far from the raw conception of perception offered both by Critias in the Charmides and by Socrates in the Theaetetus, Aristotle supposes that these perceptual judgments are themselves still part of the business of perception. It is a rich account, one which, by building detachment in, makes perception thoroughly cognitive. 9. Detachment and teleology It does more. Compare another passage where Aristotle speaks of perceiving that we perceive:84 …and if the one who sees perceives85 that he sees, the one who hears perceives that he hears, the one who walks perceives that he walks, and similarly in the other cases there is something that perceives that we are in activity, so that if we perceive it perceives that we perceive, and if we think it perceives that we think; and if perceiving that we perceive or think is perceiving that we exist (for as we said, existing is perceiving or thinking); and if perceiving that one is alive is pleasant in itself…. (Ethica Nicomachea 1170a29-b1 trans. Rowe ) 86

This is part of Aristotle’s argument to show that the self-sufficient person (the person of supreme virtue) still needs friends: not, for sure, friends for instrumental reasons (since the self-sufficient person has all she 83

Thus 425b23 talks about the sense-organs, where what is looked at (somehow or other) resides. 84 This passage is regularly cited in the discussion of de an 3.2; but its context underestimated. Compare Eudemian Ethics, 1244b25-34, which I discuss in more detail in McCabe (2008). 85 Notice here that the present tense has the same variable aspect as noted above, n.11. 86 See Johanssen (2005, p. 264 n.58) on the text.

168

needs), but friends who contribute to her sense of living the best life, and who thus enhance the pleasure she has in the life she lives. The full life she lives is determined by perception and thought (they are the actualisations of her nature), and so it is good and pleasant in itself, an object of everyone’s desire. In that case, the self-sufficient person gets pleasure from perceiving that she perceives, or that she walks – or that she lives; and she enhances that perception by joining with her friend in its contemplation. This argument rests its account of the self-sufficient person’s need for friends on a thesis about what living the best life is: the full actualisation of our capacities. But this full actualisation – as Aristotle argues both here and in the parallel passage in the Eudemian Ethics – is something we enjoy contemplating when we have it; it is something, therefore, that we desire to contemplate – on our own, or with our friends. Perception, as it is here construed, both marks the full actualisation of our capacities (it is not merely passive, nor a raw event impinging on our sense-organs), and reflects on that full actualisation: that is what happens when we perceive that we perceive (we perceive that we perceive, that we exist, and that this life is a good one). But as such, it is an object of desire and aspiration, something that involves the self-sufficient person in a wide spectrum of ethical activity (including the cultivation of friends, and the avoidance of certain distractions). So perception is a manifestation of human excellence, and perceiving it a source of pleasure. And in that case, the perception of perception simply cannot be the consciousness that necessarily follows, or is intrinsic to, our first-order perception of special sensibles. Perceiving that we perceive, on the contrary, is a normative ideal, something we endeavour to do, something we practise, habituate ourselves to, and take steps to achieve. But if perception that we perceive is thus teleological, it cannot be mere phenomenal consciousness. ‘Since we perceive that we see and hear…’ in de anima 3.2 does not, I conclude, describe phenomenal consciousness, either. For here too the discussion is enmeshed in a larger context, in which Aristotle lays out an account of natural function that underpins his ethics. If there is indeed continuity between what he says in the ethical works and his account of the nature of the soul, we should take the de anima account of self-perception to have similar normative features. Self-perception is not a necessary, nor a regular concomitant of first-order perceptions (this is one consequence of the argument that perception by sight is not one). It is not, either, a mark of what things are like for me, of the subjective side of perception, nor of its specialness to the perceiver herself. For the argument turns, not on ques-

169

tions about how I feel about my perceptions, nor yet about how my perceptions are ineradicably mine not yours, but on the reflective relation described by ‘self-perception’. In thus elaborating my perceptual judgments, the argument fails, that is to say, the tests for phenomenal consciousness I set for it at the outset. Instead, for Aristotle higher-order perception is reflective and detached: looking at, and adjudicating, what the first-order presents to it, as if from the outside. 10. The first person? In this Aristotle differs sharply from a lurking shadow in this text, Protagoras, once again from the Theaetetus.87 Protagoras’ relativist account of perception, of truth and of knowledge, one might be forgiven for supposing, is of nothing but consciousness. Protagoras – as Plato represents him in the Theaetetus – starts with perception, and insists that my perceptions are incorrigible by you; then extends this claim about perception to all appearances; and extends that to a claim about truth and the way things are: ‘as things seem for me, so they are for me’. This is universalised (always: as things seem for me, so they are for me) but ineliminably relativised (always: as things seem for me so they are for me). In Protagorean relativism, that is, what things are like for me is a permanent (and exhaustive – but that is another issue) feature of my account of truth and the way the world is; and the same goes for you, too. Protagoras cannot accommodate reflective detachment; for Protagoras, all my phenomena are ineluctably mine. 88 Aristotle rejects both this account of what it is to perceive, and any connected account of what it is to perceive that we see and hear: his account of appearances is designed to show how appearances themselves may be subject to reflection and correction.89 But the argument of 3.2, nonetheless, allows him to include these reflective judgements in the purview of perception (and so discussed further under the heading of phantasia in 3.3) rather than attributing them to some separate rational activity 87

Notice his presence at de anima 427b3; and compare Metaphysics 1009a6 ff. The ‘for me’ business could express the privileged view, the view especially from here, from my end of things (and so be congenial to claims about the incorrigibility of our perceptions or appearances by others); or it could be designed to show that a given perception is mine (not yours) and so an inalienable part of my world view. 89 We say ‘it appears to be a man’ just when there is some doubt about the matter (428a14 ff., echoing the Philebus). And our judgments may be corrections of what appears to us: the sun appears to be a foot wide, but we believe it to be greater than the inhabited world (428b3-4). 88

170

standing outside the perceptual process. Instead of consciousness, he proffers an account of perceptual judgement as indeed a part of perception; not regular or necessary for individual events of perception, but essential nonetheless to our understanding of what it is to perceive, of the collection of cognitive phenomena described by expressions like ‘perceiving by sight’. On such an account perception is both one and not one: each in a way. It is not one, at least because there is a contrast to be drawn between first-order perceptions (where the object and the subject are actualised together) and second-order ones – which may not even occur when a first-order perception takes place (first-order perception is only perceptible). It is one, nonetheless, because even these second-order perceptual judgments are perceptions: they are genuinely reflexive, in that they are perception properly socalled of perception. The connection with the Nicomachean Ethics, on the other hand, allows us to see just where (and how radically) Aristotle offers this account of perception to replace Protagorean relativism. Perceiving that we perceive is not phenomenal consciousness, because higher-order perception is not necessary for (nor intrinsic to) perceiving in a first-order way. But the account of perceiving that we perceive that does issue from the chapter is not focussed on the person who perceives – quite the contrary, the argument of 425b13-17 is notably impersonal. I suggested that the first-person formulations with which the chapter began mark their status as the phenomena to be described – where what catches Aristotle’s interest is the ordering of perception, not its indexing to the person who perceives. However, the role of Protagoras in the background to this chapter, and the significance of perceiving that we perceive in the arguments of EN 9.9 may give a better account of why Aristotle begins with what happens to us. For when I consider the development and flourishing of my faculties, I do so in terms of my own life (the puzzle about friendship, as Aristotle sets it up, rests on why, if I am self-sufficient, I should have any room for others in my life). It is essential, then, to understanding higher-order perception that we think about it in the context of a life belonging to a determinate person. Perceiving that I perceive, therefore, is essentially indexed to me; but this indexing comes from the normativity of perception, not from its connections to consciousness. In this Aristotle is true to his Platonic antecedents: for both in the Charmides and even in the Theaetetus Plato is attending to the nature, not just of cognition, but of virtue and wisdom. In the Charmides selfknowledge (and its analogue, self-perception) is of interest because it may

171

explain how best to live, and it may explain what the virtue is that is called sôphrosunê, self-control or temperance. In the Theaetetus, the discussion of perception and judgement comes at the end of a long engagement with Protagoras and Heraclitus, whose views of truth and reality are contrasted with the reflective life of the philosopher (172a-177b). That reflection, in turn, is contrasted with the state of mind of the clever politician, who is unable to see the way things really are (e.g. 176e4-5, 177a1). In Plato, no less than in Aristotle, perception has a normative cast. 11. A philosophical antique? What exactly would be in this for Aristotle? What would be in this for us? How far does Aristotle’s claim, as I have construed it, simply fall to the complaint that these are just ancient lecture notes, of no significance to us now? Why, oh why, didn’t Aristotle just see that consciousness is really important, much more important than this assimilation of perceptual judgement to perception proper? If Aristotle fails to see that there is something that it is like for me to perceive, and so fails (on the account I have given) to explain the difference between my smelling the garlic and the butter smelling of garlic, so much the worse, either for Aristotle, or for my interpretation of him. In response, I offer only a gesture, in the direction of Aristotle’s broader agenda. The current debate about this passage takes Aristotle to be interested here in some kind of account of the nature of perceptual awareness, whether that be by virtue of the subjective feel of perception (Kosman); the intentional structure of consciousness (Caston) or the inner sense which explains it (Johanssen). But this debate takes its agenda from the discussion of the mechanics of perception in earlier chapters, not only in de anima 2.12, but also earlier (2.5). However from the beginning of de anima 3, the argument has been about the number and nature of the faculties of soul; and it is no surprise that the book continues with a discussion of the queerness of the faculty that delivers phantasia, as well as the complex nature of thought and knowledge (3.4, 3.5). But this talk about faculties is connected to the talk about functions in the ethical works (and indeed it is one of the most significant features of the de anima that it has this kind of continuity with both the ethical and the metaphysical works).90 Both the de anima and the Platonic texts to which it is heir have the same presumption 90

Compare the notoriously difficult 3.5 with the discussion of theôria in E.N. X.7, and the account of the unmoved mover in Metaphysics XII. 9.

172

that questions about epistemology cannot be divorced from questions of ethics: indeed, to ignore this feature of the Platonic background risks missing the point of the complexity of de anima 3.2 altogether. Although perception by sight is ‘not one’, I have argued that nonetheless in de anima 3.2. higher-order perception is rendered continuous with the natural activity of perceiving first-order sensibles, by still qualifying as perception. This continuity allows Aristotle a naturalist account of perceptual judgement, wherein the activity of judgement is itself part of man’s natural function. Judging, at least at the perceptual level, is something that we do by nature. It is not immune to error, for sure; but it is something that can be explained by reference to the essence of man, and thence as something we can be good at. His claim, then, is not so much that we perceive that we perceive every time we perceive (not, as I have argued, an account of regular phenomenal consciousness, so not ‘always or for the most part’ in this sense) but rather an account of how we may develop our natures to exercise cognitive function of this kind – given the right development, then, man ‘always or for the most part’ makes perceptual judgments well: and he can come to do so by reflection on perception. Well, that may still look like an antique, until we recall the wider role Aristotle has for judgement, and for good judgement, in his account of how best to live. Consider the following: But since rhetoric is for the sake of the giving of judgments (krisis) (the hearers judge between one counsellor and another, and a legal verdict is a giving of judgement) rhetoric should look not only towards the judgement, that it should be demonstrative and convincing, but also to make both himself and the judge disposed in a certain way. (Rhetoric 1377b20-7, trans. from Barnes)

Being a good judge is a matter of how you are disposed, and how your dispositions are developed. This is true in both ethical matters and elsewhere: compare the account of dialectical development: Therefore one should have surveyed all the difficulties beforehand, both for the reasons we have stated and because people who inquire without first stating the difficulties are like those who do not know where they have to go; besides a man does not otherwise know whether he has found what he is looking for or not; for the end is not clear to such a man, while to him who has first discussed the difficulties it is clear. Further he who has heard all the contending arguments, as if they were parties to a case, must be in a better position for judging. (Metaphysics 995a33-b3, trans. from Barnes)

173

Aristotle supposes, then, that judgement is common to perception and to thought: Knowing and thinking seem to be like perception in a certain way (for in both of these [sc.too] the soul makes a judgement and recognises something of the things that are). (de anima 427a19-21)

Second, he supposes that judgement in general is normative. To understand how best to live, we look to the good man; and the good man exercises good judgement (E. N. 1099a23; 1113a4 ff.; 1143a6 ff.). This is something the good man does both as a matter of understanding and rationality (cf. E.N. 1141b9 ff.) and as a matter of developing his own nature best (cf. Politics 1.2, justice is a virtue of judgement, 1253a38; Rhetoric I.1-2). But good judgement in the ethical sphere is not simply analogous to the higher-order judgments of perception: just as perceptual judgement is continuous with first-order perception, so too the judgement we need for the best life is continuous with the good judgments we may make in perception. We can see all sorts of things: valid arguments (Prior Analytics 24b24); the right thing to do (E.N. 1143b5); first principles (E.N. 1098b4); the quality of a life (E.N.1100a33). And we can learn to see better, to be good at seeing these things.91 Aristotle neither supposes that there are two quite different things going on here (judgments of value, perhaps, as opposed to judgments of fact) nor that the development of our cognitive skills and abilities are separate, the ethical on the one side, the matter-of-fact on the other. Instead he imagines that the life of man, and his nature, is seamless.92 Is he wrong? 12. Coda: Aristotle reading Plato Return, finally then, to the old thought that Aristotle’s works are often just jottings for lectures, or the notes taken by a sharp student; and that this explains his rebarbative style. I have suggested that in fact this short passage from the de anima, apparently an extreme example of the gappiness of which he is so often accused, may be viewed quite differently if we read it as itself a complex reading of two Platonic dialogues. 91

Pericles is thought to be practically wise because he can see (theôrein) what is good for himself and for others (E.N.1140b8 ff.); and practical wisdom is a virtue or excellence, at which we can improve. 92 The Protagoras figure in the background is dismissed because since relativism rules out higher-order judgement, it rules out the possibility of a proper life, too.

174

First of all, these chapters of the de anima follow an agenda set by the Theaetetus. The Theaetetus raises three fat questions about perception (and, I think, remains unclear about just how far any of these questions are to be solved). The first is whether perception is to be understood as a raw mechanism, wherein the agent perceives just when affected by the patient. The second is how perceptions are somehow brought together to form a judgement which would include the ‘common terms’ as well as the objects of a special sense. The third is how perception can be mistaken: how can perception deliver falsehood? All of these questions are central to Aristotle’s discussion in de anima 3; and they are put in terms that make it quite clear that Aristotle is actually reflecting on the dialogue itself, not on doctrines we might think to have been abstracted from it. But the Theaetetus does not include a discussion of perceiving that I perceive; Aristotle gets that, as I have argued, from the Charmides. He puts it here because he supposes – what apparently the Theaetetus does not allow – that perception is not as simple as the Theaetetus allows. But the Charmides, read as a whole, takes perception to be just as complex as Aristotle would like: in the Charmides for example, Socrates might be charged with looking inside Charmides’ soul to see his virtue. Perceiving, in the Charmides, is a complex matter, not restricted to the raw feels of sensation, and careless, equally, of the problems of phenomenal consciousness. Aristotle introduces the Charmides to the Theaetetus; and perhaps in so doing he too reads the Charmides whole, realising that Plato was before him in taking the business of perception to be about our broad cognitive processes. For Aristotle’s text, the interplay between these two Platonic texts and his own arguments is responsible for the lacunae; this is not a set of lecture notes, but a complex reading of Plato.93

93

Many people have discussed the topics of this paper with me, in various different contexts. I am grateful to Melita Brownrigg for a discussion about the phainomena; to Hamid Hejawi for a discussion about the professionalisation of perception; to Peter Adamson, Verity Harte, Ursula Coope, Frisbee Sheffield and the other members of the KCL Thursday seminar on this text; to those participating in discussion at the conference, especially to Richard Sorabji and David Wiggins; to Peter Adamson and to Peter Baumann for their written comments; and to Mark Textor for extensive discussion. My special thanks to Alan Lacey, for his comments on this paper as well as for his sapience always in discussion of ancient texts, and for his shining example of the most virtuous way to do philosophy – with honesty, gentleness and an undeviating desire for the truth. I am also very grateful to the Leverhulme Trust, for the Major Research Fellowship during the tenure of which I wrote this paper and co-edited this volume.

175

References: Barker, Andrew 1981: ‘Aristotle on perception and ratios’. Phronesis, 26, pp. 248-266. Barnes, Jonathan (ed.) 1995: The Cambridge Companion to Aristotle. Cambridge: Cambridge University Press. Barnes, Jonathan, Schofield, Malcolm, Sorabji, Richard. (eds.) 1979: Articles on Aristotle: Vol IV, Psychology and Aesthetics. London: Duckworth. Burnyeat, Miles F. 1990: The Theaetetus of Plato (translation by M.J.Levett). Indianapolis: Hackett. — 1992: ‘Is an Aristotelian Philosophy of Mind still credible: a draft’. In: Nussbaum, and Rorty, (eds.) (1992), pp. 15-26. — 2002: ‘De Anima II.5’. Phronesis 47, pp. 28-90. Caston, Victor. 1998: ‘Aristotle and the problem of intentionality’. Philosophy and Phenomenological Research, 58, pp. 249-298. — 2002: ‘Aristotle on consciousness’. Mind, 111, pp. 751-815. — 2004: ‘More on Aristotle on consciousness: reply to Sisko’. Mind, 113, pp. 523533. — 2005: ‘The spirit and the letter: Aristotle on perception’. In: Salles, Ricardo (ed.), Metaphysics, soul and ethics in ancient thought. Oxford: Oxford University Press, pp. 245-320. Fine, Gail 1993: On Ideas: Aristotle’s criticism of Plato’s theory of forms. Oxford: Oxford University Press. Halliwell, Stephen 2006: ‘An Aristotelian perspective on Plato’s dialogues’. In: Hermann, Fritz-Gregor (ed.), New Essays on Plato. Swansea: University of Wales Press, pp. 189-211 Hamlyn, David W. 1968: Aristotle’s de anima Books 2 and 3. Oxford: Clarendon Press. Hett, W.S. 1964: Aristotle on the soul, parva naturalia. Cambridge Mass.: Loeb. Hicks, Robert D. 1907: Aristotle, de anima. Cambridge: Cambridge University Press. Johanssen, Thomas 2005: ‘In defense of inner sense’. Proceedings of the Boston Area Colloquium in Ancient Philosophy, 21, pp. 235-276. Kahn, Charles H. 1966: ‘Sensation and consciousness in Aristotle’s psychology’. Archiv für Geschichte der Philosophie, 48, pp. 43-81. Kosman, Aryeh L. 1975: ‘Perceiving that we perceive: On the Soul III, 2’. Philosophical Review, 84, pp. 499-519 — 2005: Commentary on Johanssen (2005), Proceedings of the Boston Area Colloquium in Ancient Philosophy, 21, pp. 277-283. Lycan, William G. 1996: Consciousness and experience. Cambridge, Mass.: M.I.T.Press. McCabe, Mary 1994: ‘Arguments in Context: Aristotle's Rhetoric and the Platonic challenge’. In: Nehamas, Alexander and Furley, David (eds.): Aristotle's Rhetoric: Philosophical Essays. Princeton: Princeton University Press, pp. 129-165.

176

— 2007: ‘Looking inside Charmides’ cloak: seeing others and oneself in Plato’s Charmides’. In: Scott, Dominik (ed.): Maieusis. Oxford: Oxford University Press. — 2008: ‘With mirrors or without? Self-perception in Eudemian Ethics vii.12’. In: Heinaman, Robert (ed.): Aristotle’s Eudemian Ethics, Proceedings of the Keeling Colloquium 2006. London: Ashgate Nagel, Thomas 1979: ‘What is it like to be a bat?’ in his Mortal Questions. Cambridge: Cambridge University Press, pp. 165-180. Nussbaum, Martha C. and Rorty, Amelie (eds.) 1992: Essays on Aristotle’s de anima. Oxford: Oxford University Press. Osborne, Catherine 1983: ‘Aristotle, De anima 3.2: how do we perceive that we see and hear?’ Classical Quarterly, 33, pp. 401-11. — 1998: ‘Perceiving white and sweet (again): Aristotle De anima 3.7, 431a20431b1’, Classical Quarterly, 48, pp. 433-46. Papineau, David 2002: Thinking about consciousness. Oxford: Oxford University Press. Ross, William D. 1956: Aristotelis de anima. Oxford: Oxford Classical Texts. — 1961: Aristotle: de anima. Oxford: Clarendon Press. Sisko, John E. 2004: ‘Reflexive awareness does belong to the main function of perception: reply to Caston’. Mind, 113, pp. 513-522. Sorabji, Richard R.K. 1979: ‘Body and soul in Aristotle’. In Barnes, Schofield, Sorabji, (1979), pp. 42-64. — 1992: ‘Intentionality and Physiological Processes: Aristotle’s theory of senseperception’. In: Nussbaum and Rorty (1992), pp. 195-226. — 1993: Animal minds and human morals. London: Duckworth. Tye, Michael 1995: Ten Problems of consciousness. Cambridge, Mass.: M.I.T.Press.

177

Philosophische Forschung Philosophical Research ______________________________________________________________________

Georg Meggle (ed.)

Edited by Johannes Brandl • Andreas Kemmerling Wolfgang Künne • Mark Textor

Georg Meggle (Ed.)

Social Facts & Collective Intentionality

Ethics of Terrorism & Counter Terrorism

3-937202-08-0, Hardcover, 478 pp., EUR 138,00

3-937202-68-4, Hardcover, 410 pp., EUR 98,00

Social Facts & Collective Intentionality: the combination of these two terms refers to a new field of basic research. Working mainly in the mood and by means of Analytical Philosophy, at the very heart of this new approach are conceptual explications of all the various versions of Social Facts & Collective Intentionality and the ramifications thereof. This approach tackles the topics of traditional social philosophy using new conceptual methods, including techniques of formal logics, computer simulations and artificial intelligence. Yet research on Social Facts & Collective Intentionality also includes ontological, epistemological, normative and - last but not least - methodological questions. This volume represents the state of the art in this new field.

We are supposed to wage war against Terrorism – but exactly what we are fighting against in this war, there is nearly no consensus about. And, much worse, nearly nobody cares about this conceptual disaster – the main thing being, whether or not you are taking sides with the good guys. This volume is an analytical attempt to end this disaster. What is Terrorism? Are terrorist acts to be defined exclusively on the basis of the characteristics of the respective actions? Or should we restrict such actions to acts performed by non-state organisations? And, most important, is terrorism already by its very nature to be morally condemned?

Mark Siebel • Mark Textor (Hrsg.)

René van Woudenberg, Sabine Roeser, Ron Rood (Eds.)

Semantik und Ontologie

Beiträge zur philosophischen Forschung ISBN 3-937202-43-9, Hardcover, 445 pp., EUR 93,00

Der zweite Band der Reihe Philosophische Forschung spannt zwei Kerngebiete der Analytischen Philosophie zusammen: die Semantik und die Ontologie. Was sind die Grundbausteine unserer Ontologie? Wie beziehen wir uns sprachlich bzw. geistig auf sie? Diese und weitere Fragen werden von international renommierten Philosophen aus historischer und systematischer Perspektive diskutiert. Die Beiträge sind in Deutsch und English verfasst. Sie stammen von Christian Beyer, Johannes Brandl, Dagfinn Føllesdal, Dorothea Frede, Rolf George, Gerd Graßhoff, Peter Hacker, Andreas Kemmerling, Edgar Morscher, Kevin Mulligan, Rolf Puster, Richard Schantz, Benjamin Schnieder, Oliver Scholz, Severin Schröder, Peter Simons, Thomas Spitzley, Markus Stepanians, Ralf Stoecker und Daniel von Wachter.

ontos verlag Frankfurt • Paris • Ebikon • Lancaster • New Brunswick www.ontosverlag.com

Basic Belief and Basic Knowledge Papers in Epistemology ISBN 3-937202-70-6, Hardcover, 293 pp., EUR 89,00

Over the last two decades foundationalism has been severely criticized. In response to this various alternatives to it have been advanced, notably coherentism. At the same time new versions of foundationalism were crafted, that were claimed to be immune to the earlier criticisms. This volume contains 12 papers in which various aspects of this dialectic are covered. A number of papers continue the trend to defend foundationalism, and foundationalism’s commitment to basic beliefs and basic knowledge, against various attacks. Others aim to show that one important objection against coherentism, viz. that the notion of ‘coherence’ is too vague to be useful, can be countered.

ontos verlag

Ontos

Philosophy

Andrea Bottani and Richard Davies (Eds.)

Modes of Existence Papers in Ontology and Philosophical Logic The volume collects essays by an international team of philosophers aimed at elucidating three fundamental and interconnected themes in ontology. In the first instance, there is the issue of the kind of thing that, in the primary sense, is or exists: must the primitive terms be particular or universal? Any reply will itself raise the question of how to treat discourse that appears to refer to things that cannot be met with in time and space: what difference is there between saying that someone is not sad and saying that something does not exist? If we can speak meaningfully about fictions, what makes those statements true (or false) and how can the entities in question be identified? Assessment of the options that have been opened up in these fields since the work of Bertrand Russell and Alexius Meinong at the beginning of the twentieth century remains an important testing-ground for metaphysical principles and intuitions.

ontos verlag

Frankfurt • Paris • Lancaster • New Brunswick 2006. 240 pages. Format 14,8 x 21 cm Hardcover EUR 79,00 ISBN 10: 3-938793-12-0 ISBN 13: 978-3-938793-12-1

P.O. Box 1541 • D-63133 Heusenstamm bei Frankfurt www.ontosverlag.com • [email protected] Tel. ++49-6104-66 57 33 • Fax ++49-6104-66 57 34