Acquaintance: New Essays 019880346X, 9780198803461

Citation preview

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

Acquaintance

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

Acquaintance New Essays

 

Jonathan Knowles and Thomas Raleigh

1

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

3

Great Clarendon Street, Oxford, OX2 6DP, United Kingdom Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide. Oxford is a registered trade mark of Oxford University Press in the UK and in certain other countries © the several contributors 2019 The moral rights of the authors have been asserted First Edition published in 2019 Impression: 1 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, by licence or under terms agreed with the appropriate reprographics rights organization. Enquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above You must not circulate this work in any other form and you must impose this same condition on any acquirer Published in the United States of America by Oxford University Press 198 Madison Avenue, New York, NY 10016, United States of America British Library Cataloguing in Publication Data Data available Library of Congress Control Number: 2019953447 ISBN 978–0–19–880346–1 DOI: 10.1093/oso/9780198803461.003.0001 Printed and bound in Great Britain by Clays Ltd, Elcograf S.p.A. Links to third party websites are provided by Oxford in good faith and for information only. Oxford disclaims any responsibility for the materials contained in any third party website referenced in this work.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

Contents Acknowledgements List of Contributors Introduction: The Recent Renaissance of Acquaintance Thomas Raleigh

vii ix 1

Part I. Phenomenal Consciousness 1. Acquaintance Is Consciousness and Consciousness Is Acquaintance Joseph Levine

33

2. Natural Acquaintance Sam Coleman

49

3. What Acquaintance Teaches Alex Grzankowski and Michael Tye

75

4. Betwixt Feeling and Thinking: Two-Level Accounts of Experience M. G. F. Martin

95

Part II. Perceptual Experience 5. Acquaintance in an Experience of Perception-cum-Action David Woodruff Smith

129

6. Dreaming, Phenomenal Character, and Acquaintance Tom Stoneham

145

7. Relationalism, Berkeley’s Puzzle, and Phenomenological Externalism Jonathan Knowles

169

8. Acquaintance, Conceptual Capacities, and Attention Anders Nes

191

Part III. Reference 9. Acquaintance as Grounded in Joint Attention John Campbell 10. Principles of Acquaintance Jessica Pepp

215 227

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

vi



Part IV. Epistemology 11. Acquaintance: The Foundation of Knowledge and Thought Richard Fumerton

245

12. Objectual Knowledge Katalin Farkas

260

13. Visual Experience, Revelation, and the Three Rs Bill Brewer

277

Index

293

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

Acknowledgements This collection of essays partly grew out of a series of workshops and conferences that were organized as part of the project ‘Representationalism or AntiRepresentationalism? Perspectives on Intentionality from Philosophy and Cognitive Science’, and the editors would like to gratefully acknowledge the Norwegian Research Council’s funding. We are also very grateful to the ‘New Directions in Philosophy of Mind’ initiative from Cambridge University and the Templeton Foundation, which also contributed funding for a conference on Acquaintance. Many of the contributors to this volume took part in these events in Trondheim and Fefor in Norway and at Senate House in London, and we hope that some sense of the interactions and excitement of those occasions—as well as the progress that was made concerning the notion of acquaintance—will be conveyed to the reader. The editors would like to thank the production team and editors at Oxford University Press for all their excellent work. In particular we would like to thank Peter Momtchiloff for all his help and patience throughout the long gestation of this project.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

List of Contributors B B is Susan Stebbing Professor of Philosophy at King’s College London. J C is Willis S. and Marion Slusser Professor of Philosophy at the University of California, Berkeley. S C is Reader in Philosophy at the University of Hertfordshire. K F is Professor of Philosophy at the Central European University. R F is Professor of Philosophy at the University of Iowa. A G is Lecturer in Philosophy at Birkbeck College, University of London. J K is Professor of Philosophy at the Norwegian University of Science and Technology. J L is Professor of Philosophy at the University of Massachusetts, Amherst. M. G. F. M is Wilde Professor of Mental Philosophy at the University of Oxford and Mills Adjunct Professor of Philosophy at the University of California, Berkeley. A N is Associate Professor of Philosophy at the Norwegian University of Science and Technology. J P is Researcher in Philosophy at Uppsala University. T R is Assistant Professor of Philosophy at the United Arab Emirates University. D W S is Distinguished Professor of Philosophy at the University of California, Irvine. T S is Professor of Philosophy at the University of York. M T is the Dallas TACA Centennial Professor in Liberal Arts at the University of Texas at Austin.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

Introduction The Recent Renaissance of Acquaintance Thomas Raleigh

1. Introduction That there is a distinctively philosophical usage of the term ‘acquaintance’ is, of course, due primarily to the influence of Bertrand Russell and in particular to the distinction he famously drew between ‘knowledge by acquaintance’ and ‘knowledge by description’. These phrases soon became part of the philosophical lexicon. For example, the Proceedings of the Aristotelian Society twice featured symposia on the question ‘Is there knowledge by acquaintance?’, first in 1919¹ and then again in 1949.² But then for much of the latter half of the twentieth century, as Russell’s influence waned, the notion of acquaintance came to be viewed with grave suspicion by many Anglophone philosophers. This was due in no small part to two hugely influential criticisms of a broadly Russellian picture:³ Wittgenstein’s (1953) ‘private language argument’ and Sellars’s (1956) attack on the ‘Myth of the Given’. However, over the last decade or two, the notion of ‘acquaintance’ has swung markedly back into fashion in philosophy. The concept has, it seems, become respectable again. This volume gathers together thirteen new essays, illustrating the wide range of topics in philosophy of mind, epistemology, philosophy of language, and metaphysics, for which the concept of acquaintance is nowadays being utilized.⁴ As ever in philosophy, this term of philosophical art has been used with a variety of different meanings, for various different philosophical projects and purposes, by many different philosophers. But as a first-pass characterization of acquaintance that I hope most parties could live with: acquaintance is a conscious mental relation that a subject can, supposedly, bear to particular items or features that is, somehow, fundamentally different from thinking a true thought about the item/feature in

¹ Featuring G. E. Moore, C. D. Broad, G. Dawes Hicks, and Beatrice Edgell. ² Featuring the great legal philosopher H. L. A. Hart along with G. E. Hughes and J. N Findlay. ³ Though neither Wittgenstein nor Sellars explicitly single out Russell as their target, it is clear that they are both criticizing a broadly Russellian, sense-data picture of experience. ⁴ See Wishon and Linsky (2015) for an excellent recent collection of essays focusing on Russell’s theory of knowledge by acquaintance at the time of Problems of Philosophy. Thomas Raleigh, Introduction: The Recent Renaissance of Acquaintance In: Acquaintance: New Essays. Edited by: Jonathan Knowles and Thomas Raleigh, Oxford University Press (2019). © Thomas Raleigh. DOI: 10.1093/oso/9780198803461.003.0001

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi



 

question. Rather than deploying concepts to form a mental state that is (merely) about something, when we are acquainted with something we are, in some sense, supposed to consciously confront that very thing itself. I suspect that any attempt to further clarify the contrast with conceptual thought, or the precise nature of the ‘conscious confrontation’, or to state the admissible categories of items/features that a subject can stand in the acquaintance relation to, is bound to step into contested territory. However, I would venture to suggest that acquaintance could illuminatingly be thought of as a way that the mind can supposedly be ‘directed’ at an object, which fundamentally contrasts with the notion of ‘intentionality’ that comes down to us from Brentano. For this latter, ‘quasi-relational’ notion does not require that the intentional object of a mental state be something that actually exists—e.g. I can think about my fairy godmother or about the present King of France—whereas acquaintance is supposed to be a genuine relation, so which could only obtain between a subject and something actual.⁵ The plan for this introductory chapter then will be to briefly consider the Russellian and pre-Russellian history of the concept, to consider a few questions and issues that the notion of acquaintance raises, and finally to survey some of the main philosophical topics for which ‘acquaintance’ has recently been invoked.

2. Historical Background Generations of budding Anglophone philosophers will have gained their first exposure to the philosophical notion of acquaintance from Russell’s immensely popular book The Problems of Philosophy (1912). Russell had earlier presented a paper to the Aristotelian society in 1911, also entitled ‘Knowledge by Acquaintance and Knowledge by Description’, which contains much of the same material that would make up chapter 5 of what he famously called his ‘penny dreadful’ book. In fact, this distinction between two forms of knowledge already appears at the beginning and end of his famous 1905 paper ‘On Denoting’. And a mention of acquaintance occurs even earlier in his Principles of Mathematics (1903). It is plausible that Russell’s notion of acquaintance was influenced by his teacher James Ward’s notion of ‘presentation’, which was in turn arguably drawn from Kant’s notion of an ‘intuition’ (Anschauung).⁶ But although it was Russell’s influence which undoubtedly enshrined the distinction as part of the standard terminology in ‘analytic’ philosophy, he was not the first English-language philosopher to use the word ‘acquaintance’ in this way. Passmore (1957) informs us that John Grote, a moral philosopher and opponent of Utilitarianism, who held the Knightbridge chair in philosophy at Cambridge,

⁵ Even this much is controversial—D. W. Smith (1989, this volume) develops a notion of acquaintance in terms of intentionality, complete with modes of presentation. Mark Johnston (2004) holds that we can, in hallucinatory episodes, be acquainted with uninstantiated universals. ⁶ Ward, who was both a philosopher and a psychologist and advocated a form of panpsychism, held the Chair of Mental Philosophy and Logic at Cambridge when Russell and Moore were students. See Hellie (2009) for the suggestion that Ward was drawing on Kant.

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi





distinguishes between ‘knowledge of acquaintance’ and ‘knowledge-about’ in his 1865 work Exploratorio Philosophica.⁷ And William James, in his Principles of Psychology, from 1890, likewise uses the term ‘acquaintance’ when drawing a philosophical distinction between different kinds of knowledge: I know the color blue when I see it, and the flavor of a pear when I taste it; I know an inch when I move my finger through it; a second of time, when I feel it pass; an effort of attention when I make it; a difference between two things when I notice it; but about the inner nature of these facts or what makes them what they are, I can say nothing at all. I cannot impart acquaintance with them to any one who has not already made it himself. I cannot describe them, make a blind man guess what blue is like, define to a child a syllogism, or tell a philosopher in just what respect distance is just what it is, and differs from other forms of relation. At most, I can say to my friends, Go to certain places and act in certain ways, and these objects will probably come. (James, 1890, 221)

Of course, in many languages other than English, some such distinction is linguistically marked by the existence of two different words—e.g. ‘savoir’ and ‘connaitre’ in French, ‘Kennen’ and ‘Wissen’ in German, ‘scire’ and ‘noscere’ in Latin—a linguistic fact that both Grote and Russell adduced in support of drawing a similar distinction in English. In Germany, at almost the same time as Grote, Helmholtz wrote in his 1868 paper ‘The Recent Progress of the Theory of Vision’ about a distinction between ‘das Kennen’ and ‘das Wissen’—the former being ‘familiarity with phenomena’, the latter being ‘the knowledge of [phenomena] which can be communicated by speech’. Unlike Russell, who held knowledge by acquaintance to be more fundamental than knowledge by description, Helmholtz thought that das Wissen was the more important and fundamental form of knowledge. For though he allowed that das Kennen has ‘the highest degree of certainty and precision’, he thought, in line with the passage quoted above from James, that it could not be expressed in words, even to ourselves, and was thus unfit to be a basis for science. I cannot say whether William James would have read Grote’s work, but he certainly did read and admire Helmholtz. Whatever the precise origins of the philosophical practice of using the English word ‘acquaintance’ as we do, it seems immensely plausible that something like the same idea or concept of an acquaintance relation has cropped up throughout the history of philosophy under different labels. For example: Sosa (2003) suggests that the notion of ‘direct’/‘unmediated’ cognitive contact that we find in chapter 5 of Problems of Philosophy bears a close resemblance to the views of Leibniz, who Russell was reading avidly at the time: Our direct awareness of our own existence and of our thoughts provides us with the primary truths a posteriori, the primary truths of fact . . . there is no mediation between the understanding and its objects. (Leibniz, 1765, Book IV, ch. 9) We shall say that we have acquaintance with anything of which we are directly aware, without the intermediary of any process of inference or any knowledge of truths. (Russell, 1912, 46) ⁷ John Grote is also credited with coining the term ‘relativism’. He is not to be confused with his brother George Grote, also a philosopher, for whom the Grote Chair in Philosophy and Logic at University College London was named.

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi



 

In the same vein, Descartes’s foundationalist project would seem to be naturally interpreted as being committed to some sort of especially secure and basic relation that one bears to one’s own conscious states—see Richard Fumerton’s chapter in this volume. Brewer (2011) points to Berkeley as an example of a philosopher whose theory of perceptual experience depends on something like a basic notion of conscious acquaintance with mind-dependent objects—see also Stoneham (2002). McLear (2016) and Gomes (2017) have both recently argued that Kant’s theory of perceptual experience is best interpreted in terms of an acquaintance relation with features in the external environment. Smith (1989 and this volume) employs the concept of acquaintance in his interpretation of Husserl—see also Jansen (2014). Gideon Makin (2000) suggests that Russell’s acquaintance is really just the very same relation that Frege called ‘grasping a sense’, though they disagreed over the objects of the relation. No doubt, many other similar precedents of something like the concept of an acquaintance relation could be found throughout the history of philosophy—but that is, I hope, enough history for present purposes.

3. Five Russellian Theses about Acquaintance Russell’s distinction was explicitly epistemological, between two kinds of knowledge, but Russell’s treatment of the distinction was also intimately bound up with claims both about reference/language and also about the metaphysics of mind/consciousness. Russell, at least around the time of Problems of Philosophy, had a number of commitments that may strike acquaintance theorists nowadays as controversial or just plain wrong. (i) Russell, at the time of ‘Knowledge by Acquaintance and Knowledge by Description’ (KAKD) and Problems of Philosophy (PP), notoriously held that we are never acquainted with familiar physical items in our shared environment—though it is important to bear in mind that Russell insisted that sense-data are nevertheless mind-independent and ontologically distinct from the subject’s conscious awareness of them. Concerning the admissible objects of the acquaintance relation, Russell held that we can only be acquainted with our current sense-data and mental states, plus some past sense-data and past mental states in memory, some universals, and possibly also ‘the self ’. We have acquaintance in sensation with the data of the outer senses, and in introspection with the data of what may be called the inner sense—thoughts, feelings, desires, etc.; we have acquaintance in memory with things which have been data either of the outer senses or of the inner sense. Further, it is probable, though not certain, that we have acquaintance with Self, as that which is aware of things or has desires towards things. In addition to our acquaintance with particular existing things, we also have acquaintance with what we shall call universals, that is to say, general ideas, such as whiteness, diversity, brotherhood, and so on . . . It will be seen that among the objects with which we are acquainted are not included physical objects (as opposed to sense-data), nor other people’s minds. (Russell, 1912, 51)

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi





However, Russell’s views on the possible objects of acquaintance had not always been so restrictive. As Peter Hylton points out, in their correspondence from around 1900–2 both Russell and Moore thought that there was no restriction whatsoever on what we can be acquainted with (see Hylton, 1990, ch. 4). Similarly in the preface to his 1903 Principles of Mathematics, Russell holds that we can be acquainted with abstract logical/mathematical items: The discussion of indefinable—which forms the chief part of philosophical logic—is the endeavour to see clearly, and to make others see clearly, the entities concerned, in order that the mind may have that kind of acquaintance with them which it has with redness or the taste of a pineapple. (Russell, 1903, xv)

Likewise, at the start of chapter 3, Russell maintains that we can be acquainted with inferential relations: it is plain that where we validly infer one proposition from another, we do so in virtue of a relation which holds between the two propositions whether we perceive it or not: the mind, in fact, is as purely receptive in inference as common sense supposes it to be in perception of sensible objects. (Russell, 1903, 33)

(ii) Knowledge by acquaintance is the more fundamental form of knowledge—all knowledge by description ultimately depends on knowledge by acquaintance. At the start of chapter 5 of problems of philosophy we are told: ‘All our knowledge, both knowledge of things and knowledge of truths, rests upon acquaintance as its foundation’ (Russell, 1912, 48). It seems clear that one might be attracted to the notion of acquaintance and want to assign it some kind of important epistemic role without going so far as to endorse Russell’s claim here that all knowledge depends on acquaintance. (iii) We can only understand propositions wholly made up of constituents we are acquainted with. This principle seems to be first stated in print at the end of ‘On Denoting’: ‘in every proposition which we can apprehend (i.e. not only in those whose truth or falsehood we can judge of, but in all that we can think about), all the constituents are really entities with which we have immediate acquaintance’ (Russell, 1905, 492). Russell reasserted this principle in many later works, including both KAKD and PP: Every proposition which we can understand must be composed wholly of constituents with which we are acquainted. (PP, 58) We must attach some meaning to the words we use, if we are to speak significantly and not utter mere noise; and the meaning we attach to our words must be something with which we are acquainted. (PP, 58) The chief reason for supposing the principle true is that it seems scarcely possible to believe that we can make a judgement or entertain a supposition without knowing what it is that we are judging or supposing about. (KAKD, 206)

A reformulated version of this principle was later influentially championed by Gareth Evans, in his ‘Varieties of Reference’, under the label ‘Russell’s Principle’: ‘in order to be thinking about an object or to make a judgement about an object, one must know

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi



 

which object is in question—one must know which object it is that one is thinking about. (I call this principle Russell’s Principle)’ (Evans, 1982, 65). (iv) When we are acquainted with a present sense-datum, we are perfectly/ completely acquainted with it; there is nothing we are missing, no ‘back side’ to a sense-datum. The particular shade of colour that I am seeing may have many things said about it—I may say that it is brown, that it is rather dark, and so on. But such statements, though they make me know truths about the colour, do not make me know the colour itself any better than I did before: so far as concerns knowledge of the colour itself, as opposed to knowledge of truths about it, I know the colour perfectly and completely when I see it, and no further knowledge of it itself is even theoretically possible. Thus the sense-data which make up the appearance of my table are things with which I have acquaintance, things immediately known to me just as they are. (Russell, 1912, 47)

A corollary of this thesis is that, for Russell, there is no such thing as different ways or manners of being acquainted with a present sense-datum, there is just the one simple/ brute acquaintance relation that can be borne contemporaneously to an item, which reveals it perfectly and entirely as it is in itself. The specific phenomenology of an experience then is determined entirely by the object side of the relation—the subject side cannot also contribute to the phenomenology by being acquainted in this or that specific way or manner. However, as Martin (this volume) points out, Russell did allow for different modes or manners of acquaintance insofar as he allowed that we can also be acquainted with a past sense-datum—and here the object of acquaintance does not determine the phenomenology of the episode of remembering, rather it is the accompanying imagery which determines the phenomenology. Later, in his Analysis of Mind (1921), Russell came to reject this fourth thesis,⁸ for by that date he treats sense-data as identical with the ultimate physical constituents of reality described by physics and so allowed that sense-data can have further unperceived/unexperienced qualities and characteristics (and indeed can exist unperceived). (v) Despite using the phrases ‘Knowledge by Acquaintance’, ‘Knowledge by Description’, Russell thought that acquaintance is not just a means or source for getting knowledge of a distinctive kind. He held that being acquainted with something is already in itself a form of knowledge. To quote again from the passage at the start of chapter 5 of PP: ‘so far as concerns knowledge of the colour itself, as opposed to knowledge of truths about it, I know the colour perfectly and completely when I see it, and no further knowledge of it itself is even theoretically possible’ (Russell, 1912, 47). Likewise in the previous chapter, Russell wrote: ‘Acquaintance with objects essentially consists in a relation between the mind and something other than the mind; it is this that constitutes the mind’s power of knowing things’ (Russell, 1912, 42, emphasis added). See Sam Coleman’s chapter in this volume for contemporary endorsement of this idea that acquaintance is itself a kind of knowledge. However, I think that the way most theorists use the term ‘acquaintance’ these days is

⁸ This move towards neutral monism arguably began even earlier, with his 1914 article ‘On the Nature of Acquaintance’.

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi





to denote a relation of direct conscious awareness that allows for a certain kind of knowledge or epistemic relation to an object but which does not in itself count as a state of knowledge. I.e. it is possible to be acquainted with O and yet fail to gain knowledge of any kind of O—one might be distracted or inattentive, one might be deluded or crazy, one could be infected with some reasonable doubt about one’s ability to form correct judgements. *

*

*

More generally, I think it’s fair to say that most recent theorists who have appealed to acquaintance would reject at least one, and quite possibly all five of these Russellian theses. Certainly, none of these five theses are obviously entailed by the core characterization of acquaintance I gave above in Section 1—viz. a relation of conscious awareness that is fundamentally distinct from thinking a true thought or forming an accurate judgement, in which the mind has some kind of unmediated confrontation with some portion of reality.

4. Five Questions for Acquaintance Theorists In this section, I briefly consider five important questions facing acquaintance theorists. (i) Firstly a methodological question: should we approach the relation of acquaintance primarily as an epistemological topic—i.e. as a relation whose essential nature we define in terms of its being a special source of knowledge or justification? Or should we approach acquaintance primarily as a notion in the metaphysics of mind—i.e. as a relation whose essential nature we define in psychological or phenomenological terms. In other words: should we take the alleged existence of a special kind of epistemic relation as the starting point for our philosophical investigations of acquaintance and only then go on to consider what kind of mental metaphysics or mechanisms that would be required to allow for that epistemic relation? Or do we start from the alleged existence of a special kind of conscious state or relation, posited initially on introspective or phenomenological grounds, and only latterly come to consider what epistemic or referential consequences might flow from such a mental state/relation? Of course, there is no need to treat these two different kinds of approaches to acquaintance as mutually exclusive—after all, it is normal enough in philosophy that we approach an issue or concept from various directions at the same time. Nevertheless, I think it is worthwhile to register that there are these different starting points for thinking about acquaintance and that different acquaintance theorists may have taken one or other route. For example: in his influential work on acquaintance, John Campbell seems to be primarily motivated by epistemological and referential concerns and this has led him to champion a relational model of experience on which the mind is acquainted with external features. In contrast, M. G. F. Martin seems to be motivated primarily by phenomenological considerations to endorse a form of naïverealism and he has been notably cautious about the epistemological benefits, if any, that might result from adopting this acquaintance-based model. Or to take another example, a number of discussions of acquaintance in the literature on singular thought/reference take their starting point to be the claim

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi



 

that there is some sort of special relation one must have to an item in order to think a singular thought about it (or to understand a singular term which refers to it)—with some theorists arguing for such a requirement and others against it. The precise nature of the mental states or mechanisms required for this relation is then a secondary question. It is not uncommon to read philosophers arguing that there must be some such requirement, but who are happy to leave its exact nature open at least for the time being. For example, Kent Bach writes: ‘A de re representation of a material object must be a percept or derive from a percept, either one’s own or someone else’s’ (Bach, 2010, 55). But how much of an object has to be perceived? Is it enough to have only seen a photo of it? What about on TV? What if the object is covered under a thin blanket but you can make out its rough shape and contours? Etc. Bach confesses he doesn’t really know—his hunch is to be reasonably inclusive, but he draws a line at merely hearing a sound produced by the object. But he’s also prepared to allow that the extent of singular thought is much more limited than his intuitions suggest (see Bach, 2010, 57–8). Its clear that Bach’s real interests here do not lie in the precise nature of the perceptual link that allows for singular thought so much as just establishing that there is some such constraint. (ii) Secondly: is the acquaintance relation non-representational? Or is being acquainted with an object just a distinctive kind—e.g. a non-conceptual or somehow sensory kind—of representational state? In general, those who are writing about acquaintance in the context of the traditional problem of perception have assumed that it is a non-representational relation that stands in opposition to the representational family of theories.⁹ But when it comes to the literature on the hard problem of consciousness this opposition between acquaintance and representation becomes less clear. On the one hand Levine (2007) and Chalmers (2003, 2004, 2007) both hold that we must acknowledge some kind of acquaintance relation to our own phenomenal properties/features, but that this cannot be explained by appeal to a special kind of phenomenal concept—i.e. to a special kind of representational vehicle. On the other hand, Balog (2012) thinks that a special kind of concept—the phenomenal kind, which uses itself to refer to itself— explains what acquaintance is. Being acquainted here then is treated as a special sensory kind of conceptual representation. Paul Churchland (1989), who asserts that ‘What Mary is missing is some form of “knowledge by acquaintance” ’ (71), provides a neuroscientific story on which ‘acquaintance’ is cashed out as a ‘distributed representation that is not remotely propositional or discursive’. Likewise, Bigelow and Pargetter (1990, 2006), in providing a physicalist response to Jackson’s Mary, appeal to acquaintance, which they understand in terms of a special kind of belief representation. Earl Conee, in yet another paper arguing that acquaintance holds the

⁹ There are by now various hybrid theories in the literature that combine relational and representational elements—e.g. Kennedy (2013), Logue (2014), Nanay (2016), and Langsam (2011). But these are hybrid theories precisely insofar as they combine a non-representational relation of acquaintance with the external world together with the notion of representational content. Thus these hybrid theorists are still treating acquaintance as a non-representational relation, even if it is somehow combined into a larger state with a different representational component. See also Bengson et al. (2011) for a ‘dual component’ theory of experience that also combines relational and representational elements.

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi





key to answering Jackson’s Knowledge Argument, is explicitly agnostic as to whether acquaintance is representational or not: ‘Perhaps awareness is experiential pure and unmediated; perhaps awareness of an experienced quality is mediated by some particularly transparent sensory form of representation. What matters for the present account is that experiencing a quality is the most direct way to apprehend the quality. That much seems beyond reasonable doubt’ (Conee, 1994, 145). For many epistemologists the point of appealing to an acquaintance relation is to account for noninferential justification of beliefs, or for a special kind of foundational epistemic security—which in itself does not obviously require one to take a stand on whether the relation essentially involves representational content. A number of recent defenders of traditional foundationalist epistemologies—e.g. Fumerton (1995, this volume), Bonjour (2003), and Fales (1996) all maintain that an episode of acquaintance can involve propositional content in the following way.¹⁰ In order to generate justification for a belief, one needs to be acquainted not only with a fact but also with: (i) a corresponding proposition (about that fact) and (ii) the relation of correspondence that obtains between the fact and the proposition. However, it is important to distinguish the idea that the object of the acquaintance relation is something representational—such as a thought or propositional content from the idea that the acquaintance relation itself is a form of representation. (iii) Thirdly, what sort of things are we (supposedly) acquainted with? What are the possible objects of acquaintance? We have already briefly discussed Russell’s views on the possible objects of acquaintance in Section 3. It is clear enough that any appeal to (or argument against) some notion of acquaintance needs to consider some of the following sorts of questions: are we supposed to be acquainted with mind-independent features out in the environment? Or are we only ever acquainted with mental or inner features? Or are we perhaps sometimes acquainted with the outer, sometimes with the inner? Or are we sometimes acquainted with both kinds of features in a single experience? E.g. Wishon (2012) argues that in perception we are simultaneously acquainted with external objects and also with sensory properties of our own experience. Should we treat the objects of acquaintance for each of the sense modalities along the same lines, or might it be that, say, vision acquaints us with external features whilst, say, taste or smell acquaint us with something inner? Are we acquainted with our own consciousness? Or our own self ? E.g. Duncan (2015) argues that we are acquainted with ourselves. Can we be acquainted with acquaintance itself ? (Fales (1996) suggests that we can be acquainted with given-ness.) As well as the inner vs. outer axis, we can also ask about the metaphysical structure of whatever it is we are acquainted with: are we acquainted with something (whether inner or outer) like a fact or state of affairs? E.g. Fumerton (1995) and Bonjour (2003) both endorse the idea that we can be acquainted not only with a fact and with a thought but also with the ‘fit’ between the fact and the thought. Or are we acquainted ¹⁰ Notice, the idea is that acquaintance can involve propositional content, not that it always or essentially does so. E.g. Fumerton explicitly allows that one can be acquainted with a fact in a way that does not involve propositional content. Many thanks to an anonymous reviewer for Oxford University Press for helpful discussion on this point.

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

   just with the particulars themselves? Or can we also be acquainted with token property instances? Assuming we are somehow or other acquainted with properties, we are obviously bound to ask which properties we can be acquainted with in experience—just as representational theorists are obliged to answer a parallel question as to which properties can be represented in experience. For example: are we only visually acquainted with 2-d shape and colour properties, or can we also be acquainted with other properties—e.g. 3-d shape and depth, perhaps temporal properties, perhaps relations such as identity over time, causation, meaning properties, perhaps the emotions of other people, etc.? And if we think the objects of acquaintance are always ‘inner’, we then face the question of whether these inner items can possess the same properties of shape and colour that we standardly ascribe to environmental items, or whether they instead possess distinct phenomenal properties—*, colour*. (iv) Fourthly, another important issue to consider is whether it is possible to stand in multiple different acquaintance relations to the one same object of acquaintance— i.e. can there be different ways or modes or manners of being acquainted with the very same thing? We have already seen Russell in PP insist that when one is acquainted with a sensedatum, one is perfectly acquainted with it—there is, as it were, no back side or hidden aspect to a sense-datum. And it seems that both Moore and Russell around this time held that whenever there is a phenomenological difference between two experiences this is always to be accounted for by a difference in the object side of the relation—see for example Moore’s (1903) famous discussion of the apparent diaphaneity of experience. (However, it should be bourn in mind that Moore eventually concluded that one can in fact attend to the act of consciousness itself, as opposed to the object, though it requires a rather special act of attention. It should also be remembered that whilst Russell held that there was only one way to be acquainted with a present sense-datum, he did allow that one could also be acquainted with it in a different way, via memory, once the sense-datum was past—see M. G. F. Martin’s chapter in this volume. I think this understanding of the acquaintance relation lives on in the literature on consciousness. For example: Chalmers (1996, 2003, 2007), Nida-Rumelin (2007), and Goff (2011, 2015) all claim that when one is acquainted with a phenomenal property—such as pain or red-looking-ness—one is thereby provided with knowledge of the essence of that property (or at least part of the essence). (Compare Bill Brewer’s chapter in this volume, in which he defends the thesis of ‘revelation’; the idea that conscious acquaintance with a property can provide knowledge of the nature of that property.) However, in the literature on perception and naïve-realism we find John Campbell (2002b, 2007) arguing that acquaintance is a 3-place relation, whose relata include not just the subject and the object but also a ‘viewpoint’ (see also Kennedy, 2007). Similarly Bill Fish (2009) allows that the nature and functioning of the subject’s visual system can contribute to the phenomenology of conscious acquaintance. More generally, if an acquaintance theorist holds that we can be acquainted with environmental features, they are going to have to deal with perceptual relativity, the changing appearances of an unchanging environmental object. And so then it seems one has to

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi





allow for different ways or manners of being acquainted with the same one item (see Raleigh (forthcoming) for further discussion of this point). One issue that then arises if one does allow for different manners or modes of acquaintance is the following: how do these modes of acquaintance differ from the orthodox representational notion of a mode of presentation or Fregean sense? You might worry here that once different modes of acquaintance are allowed we begin to lose the contrast between an acquaintance theory and a representational account of experience that treats the alleged contents of experience along Fregean lines. For now on both stories the perceptual experience is directed at an external object or scene, but this external target is presented to the subject via a distinctive mode or manner of presentation. A first thought here might be that representational modes of presentation do not require any actual existing object that the subject is presented with—the object can be merely intentional. Whereas acquaintance is genuine relation that requires both relata to actually exist for the relation to be instantiated. However, if you think, as McDowell (1982) and Evans (1982) long ago urged, that there can be special object-dependent, non-descriptive singular modes of presentation, then this dimension of contrast may be lost. Another obvious way to draw a contrast between the two approaches is to insist that manners/modes of acquaintance are not assessable as true or false, whereas Fregean modes of presentation are. However, not all representational theorists think of the content of experience as having truth conditions—e.g. Tim Crane (2009) insists that experiential content can be accurate/inaccurate but not true/false, since it is not propositional. And likewise, not all naïve-realists/acquaintance theorists deny that experiences are assessable as true/false—it is common to read people with naïve-realist sympathies, e.g. Martin (2002, 2004), continuing to use the term ‘veridical’ as applied to experience itself. A different possible dimension of contrast here is that on the orthodox Fregean account, sense determines reference—that is for any mode of presentation there is a unique (though possibly non-existent) intentional object. So we never find the same mode of presentation presenting different referents on different occasions (though perhaps the same mode of presentation can sometimes succeed in referring and sometimes fail). Whereas an acquaintance theorist can, and I think should, allow that a manner/mode of acquaintance—comprising the various factors both external and internal that contribute to the phenomenology of the experience, e.g. perspective, lighting, condition of one’s eyes, etc.—can be held constant over different occasions and yet acquaint the subject with numerically different items. If I swap one visually indistinguishable item for another and keep the viewing conditions identical, the mode or manner of acquaintance remains the same though the object of acquaintance has changed. Of course this dimension of contrast would also be lost if the representational theorist is prepared to abandon the orthodox Fregean doctrine that sense determines reference. Mark Johnston (2011), whilst arguing in favour of a relational view of perception, briefly discusses the contrast with Fregean senses: In an ASE [attentive sensory episode] the target is presented in a certain manner, and the target may or may not conform to its manner of presentation. Hence these manners of presentation

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

   are not Fregean senses, which determine something as their referent when and only when the referent satisfies the sense. A target can be presented in an ASE in an illusory fashion and yet the ASE may thereby allow its subject to makes his or her first demonstrative reference to the target, and so have the target as a topic of thought and talk. (Johnston, 2011, 173) The best model of the relation between manner of presentation and target in an ASE is given by ‘the theory of appearing’ where the manner of presentation just is the target presenting-in-acertain-manner. (174)

In contrast, Levine (2001) assimilates acquaintance to a special kind of mode of presentation—i.e. a phenomenal concept: our conception, or mode of presentation of property like redness is substantive and determinate in a way that the modes of presentation of other sorts of properties are not. When I think of what it is to be reddish, the reddishness itself is somehow included in the thought; its present to me. This is what I mean by saying it has a ‘substantive’ mode of presentation. In fact, it seems the right way to look at it is that reddishness itself is serving as its own mode of presentation. By saying that the conception is ‘determinate’, I mean that reddishness presents itself as a specific quality, identifiable in its own right, not merely by its relation to other qualities. . . . the mode of presentation of cathood [in contrast] lacks substance and determinacy. (Levine, 2001, 8)

And in a footnote he adds: The contrast I’m after between modes of presentation of qualitative properties and other properties (or objects) is perhaps captured in Russell’s (1959) distinction between ‘knowledge by acquaintance’ and ‘knowledge by description’. We are acquainted with the contents of experience, but not with anything else. (179 endnote 8)

(v) Finally, can acquaintance be further analysed and/or naturalized? It is not uncommon to hear the relation of acquaintance being described as ‘brute’ or ‘simple’—suggesting that in some sense it resists further illuminating analysis. For example, here is Evan Fales: the quality of being given is itself given; and moreover, that is itself something simple, not analysable into constituents, in the way in which the content of what is given on a particular occasion might be. As sentient beings, we apprehend, directly and immediately, what it is to be directly and immediately confronted by a perceptual experience . . . If someone did not already have acquaintance with given-ness, it would be quite futile to instruct him either by employing ostension or by saying such things as that the given is itself given, just as it would be futile to direct the blind person to color sensations by saying that they are visually sensed. (Fales, 1996, 147–8)

Likewise Richard Fumerton writes that Acquaintance: ‘cannot be informatively subsumed under a genus, and . . . cannot be analyzed into any less problematic concepts’ (Fumerton, 1995, 76). Of course, acquaintance is not meant to be simple or brute in the sense that fundamental particles or forces might be said to be simple or brute. I take it that the idea here is that the state or relation of being consciously acquainted with something cannot be further analysed or explained as

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi





being constituted by other, simpler conscious or personal level mental acts or states. And so I assume that neither Fumerton nor Fales mean to be suggesting here that the supposed unanalysability or ineffability of acquaintance shows that the relation could not in fact supervene on purely physical facts. Likewise, naïve-realist acquaintance theorists will presumably not want to deny that an episode of S being acquainted with O in some way supervenes on or is constituted by some complex extended physical process of interaction, a causal chain running from O to S, reflected light, subpersonal unconscious events in the retina and brain, etc. This might naturally lead one to wonder as to whether some kind of naturalizing project could or should be pursued along roughly the same lines as the various attempts to naturalize representational content. I.e. just as many theorists have attempted to state the conditions under which some physical or neural mechanism represents something, in terms that do not take for granted any kind of representational notion—meaning, truth, reference, etc., so one might attempt to state the conditions under which a subject is acquainted with something in terms that do not take for granted anything like consciousness—e.g. attention, or wakefulness, or ‘presence to mind’, etc. In the context of naïve-realist or relational versions of acquaintance, one might look towards enactive/embodied/extended approaches in cognitive science as natural allies when trying to give a naturalistic account of conscious acquaintance with one’s surroundings. And indeed Alva Noe has described his own enactivist position as being ‘as naïve-realist as one could hope to be’ (Noe, 2008, 703). Notice also that even if we accept that the state or relation of being acquainted cannot be analysed into personal-level constituents, nor that it can currently be given a reductive, naturalized account, that does not mean that nothing interesting or illuminating can be said about it. Compare: Williamson (2000) famously argued that knowledge is not analysable and should be taken as an epistemological primitive—but that does not mean that nothing interesting can be said about it, nor that we cannot specify necessary or sufficient conditions for knowledge. Indeed, Williamson himself holds that S knows that p iff S’s evidence includes the proposition p. And so it might be likewise for acquaintance—even if it is not analysable or reducible into simpler components, we might still be able to say various interesting or illuminating things about it.

5. Some Recent Acquaintance-Based Theorizing In this final section, I briefly survey some of the main uses to which the notion of acquaintance has been put in the recent literature—including the essays in the present volume.

5.1. Acquaintance and phenomenal properties On Russell’s original picture, we are acquainted with sense-data, which were supposed on the one hand to be distinct from familiar objects such as tables and chairs, but on the other hand were also supposed to be mind-independent and ontologically distinct from the subject’s conscious awareness of them. If we abandon this Russellian notion of sense-data, we might then think of acquaintance as a

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

   conscious relation to familiar objects and features in the external environment, or we might think of acquaintance as a relation that one bears to states or features of one’s own mind. A number of philosophers writing on the so-called ‘Hard Problem of Consciousness’, and in particular in the recent literature on phenomenal concepts, have had recourse to this latter ‘internalist’ conception of acquaintance. It is noteworthy that both physicalists and non-physicalists have made appeals to acquaintance in this regard. In the literature initially generated by Frank Jackson’s (1982) celebrated ‘Mary’ example, there were some theorists—e.g. Bigelow and Pargetter (1990, 2006), Conee (1994)—who suggested that what Mary gains when she leaves her black and white environment is knowledge by acquaintance of a phenomenal quality that she previously knew about only by description. This acquaintance-based response to Jackson’s ‘Knowledge Argument’ can thus be grouped alongside Lewis’s (1988) and Nemirov’s (1980, 1990) claim that Mary gains a new ability—in both cases the form of the response is to allow that Mary does indeed learn something but to deny that the knowledge in question is propositional knowledge of a fact. And so the physicalist thesis that all facts are physical facts can supposedly be saved. More recently, Michael Tye (2009) also endorsed an acquaintance-based response to the Knowledge Argument—though in Tye and Grzankowski’s contribution to the present volume, they argue that acquaintance is only part of the solution. Tye and Grzankowski hold that when Mary first experiences red she does indeed gain non-propositional knowledge by acquaintance of a simple sensible quality, but they point out that this nonpropositional state cannot in itself suffice for Mary’s epistemic progress. According to Tye and Grzankowski, knowledge of what it is like to experience red (an instance of knowledge-wh) is propositional knowledge. Tye and Grzankowski go on to argue that although non-propositional acquaintance cannot itself constitute a propositional answer to the question ‘what is it like to experience red?’, nevertheless acquaintance is vital to understanding Mary’s epistemic gain. Acquaintance with the sensible quality in question is that upon which Mary bases her phenomenal knowledge. In the context of the Knowledge Argument, a propositional answer to ‘what is it like to experience red?’ must be based in one’s own acquaintance in order to qualify as an appropriate answer. This position stands in contrast to the phenomenal concept strategy, a strategy that might seek to situate acquaintance in relation to the possession conditions of certain concepts. Since Tye and Grzankowski do not believe that there are any phenomenal concepts, they argue that their ‘epistemic basing’ approach is the preferred way to get Mary’s new acquaintance with a colour into the story. One obvious worry about the strategy of appealing to acquaintance as a way of defending physicalism is that this special relation of acquaintance that we can allegedly bear to our own conscious states is, prima facie, just as mysterious and difficult to explain for a physicalist as the ‘qualia’ of phenomenal redness with which we started. E.g. Gertler (1999) argues that it is only property dualists, and not physicalists, who can give a story as to how we can have this special form of access—acquaintance—with qualia. Bigelow and Pargetter’s acquaintance-based response to Mary allows that she would come to have a new belief as a result of being acquainted with phenomenal redness. By individuating beliefs in this fine-grained way, Bigelow and Pargetter’s

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi





acquaintance-based theory can be seen as presaging the later literature focusing on the ‘phenomenal concept strategy’ response to Mary. Phenomenal concepts are meant to be a special way of thinking about a phenomenal property that one can typically only acquire by having a conscious experience featuring that very property—i.e. by being acquainted with the very property that the concept is about. A defender of physicalism can then appeal to this special phenomenal kind of concept in explanation of how Mary can gain new propositional knowledge of the same old physical facts that she already knew couched in terms of neuroscientific concepts. This strategy relies then on slicing the propositional contents of knowledge states more finely than facts. A physicalist must insist that phenomenal properties are physical properties and that facts about what it is like to experience these phenomenal properties are just physical facts. The idea then is that Mary, who by hypothesis already knows all the physical facts, can gain a new phenomenal concept when she actually experiences red for the first time—and so she can reconceptualize a physical fact about the brain using this new phenomenal concept and thus gain new propositional knowledge of the same old fact. Of course, for this appeal to phenomenal concepts to count as a physicalist strategy, one must be able to give a purely physical account of what it is to acquire and deploy a phenomenal concept. That is, one must be able to give an account of what it is to think of a putatively neural property in a special, distinctively sensory way that does not just take for granted the existence of a special phenomenal aspect to the property. Many such theories, which attempt to spell out what is distinctive and special about phenomenal concepts in physicalistically respectable terms, have been offered in the literature. E.g. Lycan (1996), Perry (2001), and Ismael (2007) appeal to the idea that phenomenal concepts are a kind of indexical or demonstrative concept; Loar (1997 [1992]) and Levin (2007) claim that phenomenal concepts are recognitional concepts; Hawthorne (2002) and Braddon-Mitchell (2003) suggest that phenomenal concepts are conditional concepts. One important suggestion has been that phenomenal concepts are quotational concepts, which use a conscious state or property as part of the phenomenal concept whose referent is that very state or property—just as the device of quotation can embed a word as part of a larger meta-linguistic term that refers to that very word, e.g. ‘chair’ refers to the English word chair and is also partly constituted by it. Likewise then, the idea is that a token phenomenal concept uses (and is partly constituted by) a token phenomenal property in order to refer to that very property. This kind of approach has been pursued by Papineau (1993, 2002, 2007), Balog (1999, 2012), and Melnyk (2002). Balog in particular has claimed that a quotational account of phenomenal concepts can give a physicalist explanation of our acquaintance with phenomenal properties—i.e. how phenomenal concepts provide/involve a grasp of the phenomenal property’s essential phenomenal nature. Sam Coleman’s chapter in this volume also deploys the notion of quotation in order to give an explanation of our acquaintance with phenomenal properties that is consistent with physicalism. (Though Coleman distinguishes his quotational higher-order theory of consciousness from quotational theories of phenomenal concepts.) In contrast, a number of anti-physicalist theorists have claimed that recognizing a relation of conscious acquaintance is mandatory for a good account of consciousness and of our especially intimate cognitive contact with its manifest phenomenal nature,

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

   but that recognizing such a relation is a problem for physicalism. Levine (2007) argued that the phenomenal concept strategy as pursued in strictly physicalist terms fails to account for the special direct acquaintance we have with the phenomenal property. (Levine does not self-identify as an anti-physicalist, but he does think that the need to account for conscious acquaintance is a serious objection against physicalism.) See White (2007), Chalmers (2007), and Nida-Rumelin (2007) for related criticisms of the phenomenal concept strategy. In his contribution to the present volume, Joseph Levine argues that whilst a non-naturalistic theory of conscious acquaintance is required to explain the especially direct and intimate cognitive relation we have to our own consciousness, such a non-naturalistic theory cannot explain some of the other epistemological and semantic roles of experience that a naturalistic theory of acquaintance is well placed to explain. When theorizing about the metaphysics of consciousness, one might naturally try to start from uncontroversial claims about the manifest phenomenal appearance of experience to the conscious subject’s reflective point of view. But what seems uncontroversial and manifest to some philosophers may not seem nearly so obvious to others! For example, H. H. Price (in an oft-quoted passage) wrote: When I see a tomato there is much that I can doubt. I can doubt whether it is a tomato that I am seeing, and not a cleverly painted piece of wax. I can doubt whether there is any material thing there at all. Perhaps what I took for a tomato was really a reflection; perhaps I am even the victim of some hallucination. One thing however that I cannot doubt: that there exists a red patch of a round and somewhat bulgy shape, standing out from a background of other colourpatches . . . that something is red and round then and there I cannot doubt. (Price, 1932, 3)

But what seemed indubitable to Price—the existence of some actual red, round particular—has been denied by many adverbial and representational theorists. Likewise, a number of representational theorists have claimed that the representational nature of perceptual experience is obvious or manifest—e.g. Byrne (2001), Siegel (2010), Horgan and Tienson (2002)—though this will of course be disputed by their non-representational opponents. In his contribution to this volume, M. G. F. Martin considers whether there can be a description of the manifest subjective ‘facts of appearance’, which remains entirely neutral between the main rival competing metaphysical accounts of what grounds or determines those subjective facts—and in particular which remains neutral between representational vs. non-representational theories. Martin starts his chapter by considering a disagreement between two different sense-data theorists— Russell on the one hand, and John Foster on the other—concerning how to explain the phenomenal difference between episodes of sensory perception and episodes of sensory imagination and recollection. Whereas Russell sought to explain both perceptual experiences and conscious imagery in terms of acquaintance with sense-data, Foster proposed a ‘two-level’ account on which a relation of acquaintance with sense-data—as occurs in perceptual experience—is the more basic explanatory level, and then conscious imagining and remembering are to be explained in terms of a representation of the more basic relational sensory type of episode. An important moral that Martin draws from this debate is that Foster was correct to think that conscious episodes of imagining and remembering are manifestly representational in

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi





a way that episodes of perceptual experience are not. And thus, Martin argues, we should reject the idea that there can be a description of the subjective facts which remains neutral between representational and non-representational accounts of the metaphysical grounds of those facts.

5.2. Acquaintance and perception The past couple of decades have witnessed a surge of interest in ‘naïve-realist’ or ‘relational’ theories of perceptual experience, according to which we can be consciously acquainted with items and features in the external environment. Theorists who explicitly use the term ‘acquaintance’ in this naïve-realist context include: Campbell (2009), Fish (2009), Hellie (2010), Brewer (2011), and Soteriou (2013). But even when the term ‘acquaintance’ is not used—as in, for example, Travis (2004) Putnam (1999), Logue (2012a), and Martin (2002, 2004)—the relation of direct conscious awareness that is theorized to hold between subject and external object is very plausibly thought of as a variety of acquaintance relation. This surge of interest in naïve-realism has surely been in large part due to the development of new ‘disjunctivist’ strategies for responding to the argument from hallucination—a strategy that is standardly credited first to Hinton (1967, 1973) and then developed by Snowden (1980, 1990) and McDowell (1982, 1994) and receives its state-of-the-art defence in the work of M. G. F. Martin (2004, 2006). In its simplest form, the argument from hallucination moves from two premises to the conclusion that naïve-realism is false: (1) Hallucinations are conscious episodes that are not essentially relational. (2) (Common kind assumption): Hallucinations and perceptual experiences are the same essential kind of conscious episodes. Therefore: (3) Perceptual experiences are conscious episodes that are not essentially relational. The first premise has generally gone uncontested (though see Raleigh, 2014; Ali 2016; see also Johnston (2004) for the claim that in a hallucination one is consciously related to an uninstantiated universal), as it plausibly seems to simply fall out from the definition of a hallucination. The disjunctivist strategy is to deny (2), the common kind assumption, and to insist instead that hallucinations and perceptions, even when ‘subjectively indiscriminable’ and/or involving the same neural processes, are fundamentally different kinds of experience. And so a class of subjectively indiscriminable experiences is said to form a (merely) disjunctive kind—e.g. an experience as of a yellow lemon is either a perception of a yellow lemon or a hallucinatory episode that the subject cannot introspectively distinguish from a perception of yellow lemon. But there is, according to disjunctivists, no substantial/fundamental common conscious nature, other than this subjective indiscriminability, that is shared by both disjuncts—though there may be merely neurological similarities. Against disjunctivism, Robinson (1985, 1994) emphasizes the possibility that the neural causes/processes involved in both a perception and a subjectively indiscriminable hallucination might be exactly the same. The problem for a disjunctivist then is that once it is admitted that this neural event/process suffices by itself to give rise to

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

   the hallucinatory experience and to explain its phenomenology, it seems that in the perceptual case also the presence of this same neural event/process should likewise suffice to give rise to an experience of the same type as in the hallucinatory case. Which would seem to ‘screen off ’ the alleged acquaintance relation to the external scene from doing explanatory work vis-à-vis the phenomenology of the perceptual experience. In response, Martin (2004, 2006) argues that the disjunctivist should hold that there is nothing more to the conscious phenomenal nature of a causally matching hallucination than its indiscriminability from some perceptual experience—i.e. it is this negative epistemic condition which explains the phenomenal nature of the hallucination rather than vice versa. Martin holds that this ‘negative epistemic’ characterization of the hallucinatory case avoids the screening-off worry as such a characterization is essentially parasitic upon the phenomenal nature of the perceptual case—and so the alleged naïve-realist acquaintance relation with external features can still have a constitutive explanatory role vis-à-vis the phenomenology of the perceptual experience.¹¹ Just as hallucinations pose a potential threat to the claim that in perceptual experience we can be acquainted with the external environment, so likewise dreams—to the extent that they too are supposed to be episodes with a similarly sensory phenomenal character—might also be thought to potentially undermine ‘naïve-realist’ positions. Tom Stoneham’s chapter in this volume considers how a naïve-realist, acquaintance theorist should try to deal with dreams. One possibility would be to adopt a form of disjunctivism—which whilst allowing that dreams have some kind of phenomenal character, would insist that it is of a fundamentally different kind to the phenomenal character of perceptual experience. Stoneham considers this disjunctivist response to be viable but ‘requiring considerable confidence with respect to metaphysics’ and so ‘a comparative weakness of the [naïverealist] account’. His chapter thus explores an alternative non-disjunctive strategy when it comes to dreams: deny that dreams have phenomenal character at all. (Compare: Fish (2009) and Logue (2012b) both make parallel denials that hallucinations really have genuine sensory phenomenal character.) Whilst this will presumably strike many readers as a counter-intuitive and radical proposal, Stoneham argues that we lack any theory-neutral evidence that dreams really do possess phenomenal character as opposed to a rival view that we confabulate reports and construct false memories of dream phenomenology as a result of cultural influence and social pressures to conform. And so, if one antecedently thought that acquaintance provides the best account of normal perceptual experience, it would be legitimate to prefer the non-phenomenal theory of dreams on the basis that it fits with this best account of perception. As well as the various (disjunctivist or non-disjunctivist) strategies for defending the view against arguments from illusion, hallucination, or dreams, there are various positive motivations for the ‘naïve-realist’ idea that we can be directly acquainted

¹¹ See Fish (2009) for another presentation of the ‘negative epistemic’ account of hallucinations. See Sturgeon (1998), Siegel (2004, 2008), Pautz (2011) for criticism of the negative epistemic account.

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi





with external features. In Sections 5.3 and 5.4 we will consider some of the epistemic and referential virtues that such a view has been alleged to possess. But another important kind of motivation is phenomenological—i.e. the idea that we need to appeal to acquaintance with external features in order to do justice to the manifest phenomenology of experience. For example, in a series of important and influential papers, M. G. F. Martin (1997, 1998, 2002, 2004, 2006) has maintained that a naïverealist theory, according to which we can be directly acquainted with familiar mindindependent objects and features in our surroundings, provides the best articulation of how perceptual experience seems to first-personal reflection. Martin (2002) has also argued that naïve-realism provides the best account of the phenomenology of sensory imagining, such as visualizing. Another important (and perhaps somewhat unjustly neglected) example of acquaintance-based theorizing about perception, in which the motivation for appealing to acquaintance is phenomenological, is D. W. Smith’s 1989 book The Circle of Acquaintance. In his contribution to the present volume, Smith endorses the idea that in perception we can be directly acquainted with familiar features in our external environment, but he emphasizes that in our everyday actions we can also gain a kind of acquaintance with the objects in our surroundings. Moreover, perception and action are usually ‘intertwined’ to such an extent that they form a unified sort of experience—‘perception-cum-action’. In unfolding what he takes to be the complex phenomenological and intentional structure of this kind of experience, Smith suggests that acquaintance with the familiar features in our surroundings should be understood in the ‘embodied’ terms first laid down by Husserl and by MerleauPonty, which have more recently inspired theorists such as Varela et al. (1991), Gallagher (2006, 2017), or Noe (2004, 2012). This tradition of embodied and enactivist approaches to experience is also an important inspiration for Jonathan Knowles’s contribution to this volume. Knowles takes as his point of departure the recent debate between John Campbell and Quassim Cassam (2014) over the relative explanatory merits of an acquaintancebased, ‘relational’ account of experience vs. a representational account. He argues that Campbell’s acquaintance-based view is correct to insist that the phenomenal character of sensory experience is (at least partially) constituted by the external objects and features that we perceive. However, Knowles argues that the relational view does not sufficiently acknowledge the subject’s contribution to phenomenal character of experience. He thus advances what he takes to be a superior alternative theory, ‘Phenomenological Externalism’, according to which the objects and features that we are directly presented with in perceptual experience are part of a ‘world-for-me’ or ‘world-for-us’. Knowles’s theory is thus appealing to something like the idea of an ‘Umwelt’—a term that was due originally to the German biologist Jakob von Uexküll, and which was an important influence on Husserl’s notion of a lifeworld (Lebenswelt). Knowles’s chapter provides an extended discussion of how such a ‘world of experience’ relates to the world as described by our best physical sciences. Anders Nes’s chapter in this volume also engages closely with the work of John Campbell as well as with the work of John McDowell. Whilst McDowell can be classified as a kind of disjunctivist about perceptual experience, insisting as he does

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

   that in ‘good cases’ of normal perceptual experience we are ‘open to the world’ (McDowell, 1994, 111), he does not treat this direct perceptual awareness of our surroundings as a relation of acquaintance. This is because McDowell, following in the footsteps of Sellars and of Kant, insists that perceptual experience itself must already involve the operation of conceptual capacities if it is to be a source of knowledge and justified belief, whereas acquaintance is supposed precisely to contrast with the sort of thoughts and attitudes that employ concepts. However, Nes suggests that there can be significant convergence between the acquaintance-based, relational position of Campbell and the conceptualist position of McDowell if we consider the crucial role, for Campbell, of conscious attention. Nes argues that attention should be considered a conceptual capacity, in the sense relevant for McDowell’s purposes, and thus claims that many of the reasons Campbell provides to favour the relational account can be recast in conceptualist-friendly terms.

5.3. Acquaintance and reference There is a widely (though not universally) accepted distinction between singular or de re thoughts and descriptive or de dicto thoughts. The core intuition behind this distinction is that one way of forming a thought about a particular object is by thinking of some descriptive condition which that object (uniquely) satisfies. E.g. I may form a thought whose content includes the descriptive condition—the current heaviest sumo wrestler—and by doing so my thought will be about the particular individual (assuming there is one) who in fact uniquely satisfies this condition; in this case, the Mongolian rikishi Ichinojo Takashi. But it seems that at least sometimes one can think about a particular object in a more direct way—that is, not via the satisfaction of some descriptive, conceptual condition but simply as that thing. Granting that such direct, singular reference is sometimes possible, it is often assumed that the item itself would then figure in the propositional content of the thought, in contrast to the general, quantificational structure of a merely descriptive content. There is then a question concerning the conditions under which one can make singular reference to something and here it is often claimed that some kind of acquaintance relation with the item in question is required (at least by somebody, somewhere, at some time); though there is a whole spectrum of views on what counts as acquaintance, and there are also those who deny that acquaintance is a necessary condition on singular reference at all (Jeshion, 2010; Hawthorne and Manley, 2012). For example, Imogen Dickie (2010) has defined acquaintance as follows: ‘A subject, S, is “acquainted” with an object, o, iff S is in a position to think about o in virtue of a perceptual link with o and without the use of any conceptual or descriptive intermediary’ (213). Dickie then goes on to argue that we can be acquainted with familiar physical objects in our environment, as opposed to the idea that we can only be acquainted with mental items such as ‘sense-data’. Francois Recanati (2010) takes a slightly different approach, arguing that acquaintance is a not a strictly necessary condition on singular thought, but rather that acquaintance with a specific object is a normative/functional standard or constraint on singular thought—i.e. singular thoughts ought to be tokened on the basis of acquaintance with the referent for this is their proper purpose or functional role.

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi





In her contribution to the present volume, Jessica Pepp argues that we should distinguish more carefully between the claim that acquaintance is a necessary condition for a thought to have singular propositional content and the claim that acquaintance is a necessary condition for a thought to be about its object in a direct, ‘non-satisfactional’ way. Pepp suggests that these two conditions can come apart and so even if it is false that acquaintance is required for singular propositional content, it might yet be true that acquaintance is required in order to think about an item in this special non-satisfactional manner. Another, related way in which acquaintance has been thought to be important to reference is to allow for ostensive definitions of concepts/terms and so avoid a regress (or loop) of merely verbal definitions. If we accept that at least some concepts are not to be defined in terms of other concepts, then it seems we will need some other kind of non-conceptual relation to the referent. It is very natural to think that acquaintance with the referent would be ideally suited to play this kind of role—see both Fumerton’s and Levine’s contributions to this volume for further discussion. This sort of concern about the ultimate basis for our grasp of reference was an important motivation for John Campbell’s Reference and Consciousness (2002), a book which played a seminal role in re-establishing the notion of acquaintance as part of the mainstream conversation in analytic philosophy. Campbell famously argued that a non-conceptual, non-representational conscious relation to our surroundings—i.e. acquaintance—is required to ground our knowledge of the reference of our concepts, especially our basic demonstrative concepts. In later work, Campbell (2002a, 2005, 2009, 2011; Campbell and Cassam, 2014) has emphasized that conscious acquaintance with the external world is required in order to answer what he calls ‘Berkeley’s Puzzle’: Berkeley is trying to respect a principle about the relation between experience and concepts that is both important and difficult to keep in place. This is what I will call the explanatory role of experience. The principle is that concepts of individual physical objects, and concepts of the observable characteristics of such objects, are made available by our experience of the world. It is experience of the world that explains our grasp of these concepts. The puzzle that Berkeley is addressing is that it is hard to see how our concepts of mind-independent objects could have been made available by experience of them. (Campbell, 2002a)

Campbell’s claim is that experience cannot play the explanatory role we require of it—viz., giving us a conception of a mind-independent world of objects—unless we treat experience as a non-representational relation of acquaintance. In other words acquaintance is supposedly required to explain how it is we can even think about the mind-independent external features of the world, let alone know anything about them. (See Cassam, 2011; Campbell and Cassam, 2014; Rey, 2005; and Jonathan Knowles’s chapter in this volume for critical responses to Campbell.) Another concern that was already present in Reference and Consciousness was to understand our grasp not only of our own perceptually based demonstratives, but also other people’s perceptual demonstratives. And so we must consider cases of joint attention by two different subjects to the same object. Campbell’s suggestion was that the nature of one’s experience in such cases of joint attention is fundamentally different from cases of solitary perceptual attention—indeed Campbell claimed we

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

   should take a ‘relational view’ of joint attention: ‘Just as the object you see can be a constituent of your experience, so too it can be a constituent of your experience that the other person is, with you, jointly attending to the object’ (Campbell, 2002b, 161). In his contribution to the present volume John Campbell revisits his argument for treating experiences of joint attention as ‘primitive’. Drawing on empirical work by Michael Tomasello, Campbell argues that the sort of acquaintance with external objects that is required for a grasp of reference, and hence for communication, is grounded in very basic joint-attentional activities that we engage in with our parents or caregivers in early infancy. Campbell thus hopes to effect a rapprochement between the Wittgensteinian idea that shared, public language is explanatorily prior to gaining full-blown propositional attitudes and the core idea of Reference and Consciousness that it is perceptual acquaintance with objects and features in our environment that explains our grasp of reference.

5.4. Acquaintance and epistemology Russell employed the notion of acquaintance as part of the project of foundationalist epistemology which he saw himself as pursuing. One of the most ancient epistemological dialectics is a concern with a potential regress of justifications—sometimes called ‘Agrippa’s Trilemma’. Our beliefs stand in inferential relations (or rather the contents of those beliefs do). When the content of one belief entails the content of another, any justification for the first belief is ‘transmitted’ to the second. (Likewise when the relation is not full entailment but just probability raising, justification can be partially transmitted or bestowed.) But while inferential relations can transmit justification it is not at all obvious, despite what coherentists may claim, how they could create justification in the first place. An inferential chain or structure could go around in a circle (or some more complicated web-like structure), it could perhaps extend infinitely, or it can have a terminus—a belief that is not justified via inferential relations. A foundationalist holds that there are such terminal nodes, which are nevertheless justified non-inferentially. Some foundationalists in the past held that these nodes were self-justifying. But the more normal foundational position, these days, is that these terminal nodes are non-inferentially justified by something that is not a belief—viz, an experience. If the experience is to provide a reason for belief that is not itself just a further belief (not just another premise in a potential inference), then the experience must somehow make something available to the subject in a way that is different from just having another belief. And then one way to cash this out is that the subject is consciously acquainted with something—where this episode of acquaintance can provide a reason for believing that p but not by playing the role of a premise in an inference whose conclusion is p. Rather, the experience is held to present or reveal the very things, the truth makers, that the belief that p is about (see Raleigh (2017) for further discussion of such non-inferential justification). Recent theorists who have explicitly revived this traditional form of acquaintancebased foundationalism include Moser (1989), Fumerton (1995, 2001), Bonjour (2003), Fales (1996), Hasan (2011, 2013), and Chalmers (2012). In a similar vein, acquaintance has been appealed to in epistemological discussions of self-knowledge, where it might be supposed to explain or account for the distinctively secure status of

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi





knowledge of our mental states gained via introspection. E.g. Brie Gertler (2011, 2012) has based her account of first-person privileged knowledge on the idea that we are acquainted with our own conscious mental states. Chalmers (2003) likewise argues for a form of incorrigibility based on acquaintance with phenomenal properties. This revival of traditional, acquaintance-based foundationalism has faced various criticisms. For example, Poston (2010) and Huemer (2007) argue that acquaintance cannot allow for fallibly justified foundational beliefs. (See Fumerton (2010) and Hasan (2013) for pro-acquaintance responses.) Whilst Sosa (2003), Poston (2007), and Markie (2009) suggest that ‘speckled hen cases’¹² pose a problem for an acquaintance-based theory of justification. (See Bonjour (2003) and Fumerton (2005) for responses.) In Richard Fumerton’s chapter for this volume, he continues to pursue his own internalist foundational project in epistemology, a project that he explicitly connects with Descartes’s employment of the ‘method of doubt’ to search for secure foundations for our beliefs. Fumerton holds not only that we can be acquainted with conscious mental states of affairs (facts) and with propositional contents (thoughts) but also with the correspondence that can obtain between the fact and the thought such that the former is the truth maker for the latter. Fumerton claims that when we are acquainted with all three of these factors we would have ‘strong or ideal’ justification for believing the proposition in question. Fumerton argues that such an account can deal with familiar ‘speckled hen’-type objections by denying that we have genuine direct acquaintance with one or other of these three factors (fact, thought, or the correspondence between them). He also argues that acquaintance can provide a unified account of both empirical and a priori justification. To be clear—one does not have to be an acquaintance theorist in order to accept that there is non-inferential justification or in order to be a foundationalist. For example, Jim Pryor (2000) has influentially argued in favour of the non-inferential justification of beliefs by experience, but he does not make any sort of appeal to acquaintance (though he does think that it is the distinctively ‘presentational’ phenomenology of perceptual experience which allows it to provide defeasible, non-inferential justification for belief). More generally, one could, prima facie, be a thoroughgoing representational theorist about experience, eschewing all talk of acquaintance, and still maintain that a contentful, representational experience can stand in a justificatory relation to a belief that is not an inferential relation. A different kind of epistemic project for acquaintance, distinct from the foundationalist project sketched above, is to explain how it is possible for us to gain knowledge of the intrinsic, categorical nature of external mind-independent features—as opposed to knowing merely that some or other intrinsic/categorical feature occupies a certain position or role within a relational, theoretical structure. This ‘revelatory’ function for acquaintance has in recent years been championed both by John Campbell (1993, 2005) and Bill Brewer (2011).¹³ They have argued that ¹² The example of a speckled hen is presented by Chisholm (1942), though Chisholm credits the idea to Gilbert Ryle. ¹³ See also Dasgupta (2015) for another important recent work that appeals to the revelatory function of acquaintance.

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

   acquaintance with the external world allows us to avoid the sort of epistemic ‘Humility’ about the intrinsic nature of the external world that Rae Langton (1998) and David Lewis (2009) suggested we are bound to be limited to—a position that is very similar to epistemic structural realism in the philosophy of science (see, for example, Maxwell (1968) and Worrall (1989)—though it is arguable that Russell was already endorsing something like structural realism in Problems of Philosophy). In his contribution to the present volume, Bill Brewer argues that only an acquaintancebased account of visual experience can explain such an experience’s ability to be a source of knowledge about the intrinsic/categorical nature of mind-independent properties—e.g. a source of knowledge about what it is for something to be round or square, red or blue, etc. Brewer argues that two main rival theories, resemblance-based or representational accounts, cannot account for how experience can be a source of such knowledge. And so to the extent that we wish to avoid embracing humility and accept that experience is indeed a source of such ‘revelatory’ knowledge about the external world’s intrinsic nature, we have reason to embrace an acquaintance-based theory of experience. A rather more specific epistemological topic in which acquaintance has recently figured concerns aesthetic knowledge and judgement. Richard Wollheim (1980) proposed an acquaintance principle according to which it is impossible to gain certain kinds of aesthetic knowledge about O unless one is acquainted with O itself: ‘judgments of aesthetic value . . . must be based on first-hand experience of their objects’ (233). A number of theorists have argued that Wollheim’s formulation is too strong for it would rule out making judgements about a work’s aesthetic value based on a reproduction or photograph (see e.g. Livingston, 2003; Hopkins, 2006). Nevertheless, many philosophers have maintained that Wollheim’s acquaintance principle points to something correct about aesthetic judgement and have tried to provide improved reformulations—see e.g. Budd (2003), Todd (2004), Konigsberg (2012), Robson (2013, 2017), and Sauchelli (2016) for further discussion. It is commonplace to draw a distinction between knowing-that and knowing-how— i.e. between factual and practical knowledge. Acquaintance raises the prospect that there may be a third important kind of knowledge, which we might call objectual knowledge. We will then naturally wonder whether one or other of these kinds of knowledge reduces to or depends on one of the others. In the case of knowing-that and knowing-how, this question has generated a very large literature (see Stanley and Williamson (2001) and Stanley (2010) for recent influential discussion; see Fantl (2009) for a useful survey article). And as mentioned already, Russell held that knowledge of an object by acquaintance was a fundamentally different kind of knowledge, to knowing a truth about an object—indeed Russell held that knowledgeby-acquaintance was the more fundamental kind. Katalin Farkas, in her contribution to this volume, considers whether objectual knowledge forms a genuine, irreducibly distinct kind of knowledge. Farkas argues that our everyday, natural language talk of ‘knowing things’ does not express such a distinctive kind of objectual knowledge with its own special, uniform nature. Indeed, Farkas suggests that at least some of our familiar talk of ‘knowing things’ refers to relations that should not be classified as knowledge at all. However, Farkas concludes that a distinctively philosophical notion of acquaintance with one’s own conscious experience might

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi





allow us to carve out a narrower but genuinely unified, sui generis category of objectual knowledge. *

*

*

Whilst the foregoing survey of recent acquaintance-based theorizing has by no means been exhaustive, I hope that it will, like the essays gathered here, provide a sense of the rich variety of topics for which the notion of acquaintance is currently being employed.

References Ali, R. (2016) ‘Does Hallucinating Involve Perceiving?’, Philosophical Studies 175(3): 601–27. Bach, K. (2010) ‘Getting a Thing into Thought’, in Robin Jeshion (ed.), New Essays on Singular Thought, Oxford: Oxford University Press. Balog, K. (1999) ‘Conceivability, Possibility, and the Mind-Body Problem’, Philosophical Review 108(4): 497–528. Balog, K. (2012) ‘Acquaintance and the Mind-Body Problem’, in Simone Gozzano and Christopher S. Hill (eds), New Perspectives on Type Identity: The Mental and the Physical, Cambridge: Cambridge University Press. Bengson, J., Grube, E., and Korman, D. (2011) ‘A New Framework for Conceptualism’, Noûs 45(1): 167–89. Bigelow, J. and Pargetter, R. (1990) ‘Acquaintance with Qualia’, Theoria 61(3): 129–47. Bigelow, J. and Pargetter, R. (2006) ‘Re-Acquaintance with Qualia’, Australasian Journal of Philosophy 84(3): 353–78. Bonjour, L. (2003) ‘A Version of Internalist Foundationalism’, in L. BonJour and E. Sosa (eds), Epistemic Justification: Internalism vs. Externalism, Foundations vs. Virtues, Malden, MA: Blackwell. Braddon-Mitchell, D. (2003) ‘Qualia and Analytical Conditionals’, Journal of Philosophy 100(3): 111–35. Brewer, B. (2011) Perception and Its Objects, Oxford: Oxford University Press. Budd, M. (2003) ‘The Acquaintance Principle’, British Journal of Aesthetics 43(4): 386–92. Byrne, A. (2001) ‘Intentionalism Defended’, Philosophical Review 110: 199–240. Campbell, J. (1993) ‘A Simple View of Colour’, in John J. Haldane and C. Wright (eds), Reality: Representation and Projection, Oxford: Oxford University Press. Campbell, J. (2002a) Reference and Consciousness, Oxford: Oxford University Press. Campbell, J. (2002b) ‘Berkeley’s Puzzle’, in Tamar Szabo Gendler and John O’Leary Hawthorne (eds), Conceivability and Possibility, Oxford: Oxford University Press. Campbell, J. (2005) ‘Précis of Reference and Consciousness, and Replies to Neil Manson and Georges Rey’, Philosophical Studies 126: 103–14, 145–53, 155–62. Campbell, J. (2007) ‘The Metaphysics of Perception’, Philosophical Issues 17(1): 1–15. Campbell, J. (2009) ‘Consciousness and Reference’, in Brian McLaughlin, Ansgar Beckermann, and Sven Walter (eds), Oxford Handbook of Philosophy of Mind, Oxford: Oxford University Press. Campbell, J. (2011) ‘Relational vs. Kantian Responses to Berkeley’s Puzzle’, in N. Eilan, H. Lerman and J. Roessler (eds), Understanding Perception, Causation and Objectivity, Oxford: Oxford University Press. Campbell, J. and Cassam, Q. (2014) Berkeley’s Puzzle, Oxford: Oxford University Press. Cassam, Q. (2011) ‘Tackling Berkeley’s Puzzle’, in N. Eilan, H. Lerman, and J. Roessler (eds), Understanding Perception, Causation and Objectivity, Oxford: Oxford University Press.

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

   Chalmers, D. (1996) The Conscious Mind, Oxford: Oxford University Press. Chalmers, D. (2003) ‘The Content and Epistemology of Phenomenal Belief ’, in Quentin Smith and Aleksandar Jokic (eds), Consciousness: New Philosophical Perspectives, Oxford: Oxford University Press. Chalmers, D. (2004) ‘Phenomenal Concepts and the Knowledge Argument’, in Peter Ludlow, Yujin Nagasawa, and Daniel Stoljar (eds), There’s Something about Mary: Essays on Phenomenal Consciousness and Frank Jackson’s Knowledge Argument, Cambridge, MA: MIT Press. Chalmers, D. (2007) Phenomenal Concepts and the Explanatory Gap’, in Torin Alter and Sven Walter (eds), Phenomenal Concepts and Phenomenal Knowledge: New Essays on Consciousness and Physicalism, Oxford: Oxford University Press. Chalmers, D. (2012) Constructing the World, Oxford: Oxford University Press. Chisholm, R. (1942) ‘The Problem of the Speckled Hen’, Mind 51(204): 368–73. Churchland, P. (1989) ‘Knowing Qualia: A Reply to Jackson’, in P. Churchland (ed.), A Neurocomputational Perspective, Cambridge, MA: MIT Press. Conee, E. (1994) ‘Phenomenal Knowledge’, Australasian Journal of Philosophy 72: 136–50. Crane, T. (2009) ‘Is Perception a Propositional Attitude?’, Philosophical Quarterly 59: 452–69. Dasgupta, S. (2015) ‘Inexpressible Ignorance’, Philosophical Review 124(4): 441–80. Dickie, I. (2010) ‘We Are Acquainted with Ordinary Things’, in Robin Jeshion (ed.), New Essays on Singular Thought, Oxford: Oxford University Press. Duncan, M. (2015) ‘We Are Acquainted with Ourselves’, Philosophical Studies 172(9): 2531–49. Evans, G. (1982) The Varieties of Reference, Oxford: Oxford University Press. Fales, E. (1996) A Defense of the Given, Lanham, MD: Rowman and Littlefield. Fantl, J. (2009) ‘Knowing-How and Knowing-That’, Philosophy Compass 3(3): 451–70. Fish, W. (2009) Perception, Hallucination, and Illusion, Oxford: Oxford University Press. Fumerton, R. (1995) Metaepistemology and Skepticism, Lanham, MD: Rowman and Littlefield. Fumerton, R. (2001) ‘Classical Foundationalism’, in M. De Paul (ed.), Resurrecting OldFashioned Foundationalism, Lanham, MD: Rowman and Littlefield. Fumerton, R. (2005) ‘Speckled Hens and Objects of Acquaintance’, Philosophical Perspectives 19: 121–38. Fumerton, R. (2010) ‘Poston on Similarity and Acquaintance’, Philosophical Studies 147: 379–86. Gallagher, S. (2006) How the Body Shapes the Mind, Oxford: Oxford University Press. Gallagher, S. (2017) Enactivist Interventions: Rethinking the Mind, Oxford: Oxford University Press. Gertler, B. (1999) ‘A Defense of the Knowledge Argument’, Philosophical Studies 93(3): 317–36. Gertler, B. (2011) Self-Knowledge, New York: Routledge. Gertler, B (2012) ‘Renewed Acquaintance’, in Declan Smithies and Daniel Stoljar (eds), Introspection and Consciousness. Oxford: Oxford University Press, 93–128. Goff, P. (2011) ‘A posteriori Physicalists Get Our Phenomenal Concepts Wrong’, Australasian Journal of Philosophy 89(2): 191–209. Goff, P. (2015) ‘Real Acquaintance and Physicalism’, in P. Coates and S. Coleman (eds), Phenomenal Qualities: Sense, Perception, and Consciousness, Oxford: Oxford University Press. Gomes, A. (2017) ‘Naïve Realism in Kantian Phrase’, Mind 126(502): 529–78. Grote, J. (1865) Exploratio Philosophica, Cambridge: Deighton, Bell & Co. Hart, H. L. A. et al. (1949) ‘Symposium: Is There Knowledge by Acquaintance?’, Proceedings of the Aristotelian Society, Supplementary Volume 23(1): 69–128.

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi





Hasan, A. (2011) ‘Classical Foundationalism and Bergmann’s Dilemma for Internalism’, Journal of Philosophical Research 36: 391–410. Hasan, A. (2013) ‘Phenomenal Conservatism, Classical Foundationalism, and Internalist Justification’, Philosophical Studies 162: 119–41. Hawthorne, J. (2002) ‘Advice for Physicalists’, Philosophical Studies 109(1): 17–52. Hawthorne, J. and Manley, D. (2012) The Reference Book, Oxford: Oxford University Press. Hellie, B. (2009) ‘Acquaintance’, in Tim Bayne, Axel Cleeremans, and Patrick Wilken (eds), Oxford Companion to Consciousness, Oxford: Oxford University Press. Hellie, B. (2010) ‘An Externalist’s Guide to Inner Experience’, in Bence Nanay (ed.), Perceiving the World, Oxford: Oxford University Press. Helmholtz, H. (1995 [1868]) ‘The Recent Progress of the Theory of Vision’, in David Cahan (ed.), Science and Culture: Popular and Philosophical Essays, Chicago, IL: University of Chicago Press. Hinton, J. M. (1967) ‘Visual Experiences’, Mind 76: 217–27. Hinton, J. M. (1973) Experiences: An Inquiry into Some Ambiguities, Oxford: Clarendon Press. Hopkins, R. (2006) ‘How to Form Aesthetic Belief: Interpreting the Acquaintance Principle’, Postgraduate Journal of Aesthetics 3(3): 85–99. Horgan, T. and Tienson, J. (2002) ‘The Intentionality of Phenomenology and the Phenomenology of Intentionality’, in David J. Chalmers (ed.), Philosophy of Mind: Classical and Contemporary Readings, Oxford: Oxford University Press. Huemer, M. (2007) ‘Compassionate Phenomenal Conservatism’, Philosophy and Phenomenological Research 74(1): 30–55. Hylton, P. (1990) Russell, Idealism, and the Emergence of Analytic Philosophy, Oxford: Oxford University Press. Ismael, J. (2007) The Situated Self, Oxford: Oxford University Press. Jackson, F. (1982) ‘Epiphenomenal Qualia’, Philosophical Quarterly 32(April): 127–36. James, W. (1890) Principles of Psychology, New York: Henry Holt & Co. Jansen, J. (2014) ‘Taking a Transcendental Stance: Anti-Representationalism and Direct Realism in Kant and Husserl’, in Faustino Fabbianelli and Sebastian Luft (eds), Husserl und die Klassische Deutsche Philosophie, New York: Springer. Jeshion, R. (2010) ‘Singular Thought: Acquaintance, Semantic Instrumentalism, and Cognitivism’, in R. Jeshion (ed.), New Essays on Singular Thought, Oxford: Oxford University Press. Johnston, M. (2004) ‘The Obscure Object of Hallucination’, Philosophical Studies 103: 113–83. Johnston, M. (2011) ‘On a Neglected Epistemic Virtue’, Philosophical Issues 21(1): 165–218. Kennedy, M. (2007) ‘Visual Awareness of Properties’, Philosophy and Phenomenological Research 75(2): 298–325. Kennedy, M. (2013) ‘Explanation in Good and Bad Experiential Cases’, in Fiona Macpherson and Dimitris Platchias (eds), Hallucination: Philosophy and Psychology, Cambridge, MA: MIT Press. Konigsberg, A. (2012) ‘The Acquaintance Principle, Aesthetic Autonomy, and Aesthetic Appreciation’, British Journal of Aesthetics 52(2): 153–68. Langsam, H. (2011) The Wonder of Consciousness: Understanding the Mind Through Philosophical Reflection, Cambridge, MA: MIT Press. Langton, R. (1998) Kantian Humility: Our Ignorance of Things in Themselves, Oxford: Oxford University Press. Leibniz, G. (1765) New Essays on Human Understanding, trans. A. C. Langley, Cambridge: Cambridge University Press.

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

   Levin, J (2007) ‘What Is a Phenomenal Concept?’ in T. Alter and S. Walter (eds), Phenomenal Concepts and Phenomenal Knowledge: New Essays on Consciousness and Physicalism, Oxford: Oxford University Press. Levine, J. (2001) Purple Haze: The Puzzle of Consciousness, Oxford: Oxford University Press. Levine, J. (2007) ‘Phenomenal Concepts and the Materialist Constraint’, in T. Alter and S. Walter (eds), Phenomenal Concepts and Phenomenal Knowledge: New Essays on Consciousness and Physicalism, Oxford: Oxford University Press. Lewis, D. (1988) ‘What Experience Teaches’, Proceedings of the Russellian Society 13: 29–57. Lewis, D. (2009) ‘Ramseyan Humility’, in D. Braddon-Mitchell and R. Nola (eds), Conceptual Analysis and Philosophical Naturalism, Cambridge, MA: MIT Press. Livingston, P. (2003) ‘On an Apparent Truism in Aesthetics’, Journal of Aesthetics 43(3): 260–78. Loar, B. (1997 [1992]) ‘Phenomenal States’, in N. Block, O. Flanagan, and G. Güzeldere (eds), The Nature of Consciousness: Philosophical Debates, Cambridge, MA: MIT Press. Logue, H. (2012a) ‘Why Naive Realism?’, Proceedings of the Aristotelian Society 112(2pt2): 211–37. Logue, H. (2012b) ‘What Should the Naïve Realist Say about Total Hallucinations?’ Philosophical Perspectives 26(1): 173–99. Logue, H. (2014) ‘Experiential Content and Naive Realism: A Reconciliation’, in Berit Brogaard (ed.), Does Perception Have Content?, Oxford: Oxford University Press. Lycan, W. (1996) Consciousness and Experience, Cambridge, MA: MIT Press. Makin, G. (2000) The Metaphysicians of Meaning: Russell and Frege on Sense and Denotation, London: Routledge. Markie, P. (2009) ‘Classical Foundationalism and Speckled Hens’, Philosophy and Phenomenological Research 79(1): 190–206. Martin, M. G. F. (1997) ‘The Reality of Appearances’, in M. Sainsbury (ed.), Thought and Ontology, Milan: Franco Angeli, 77–96. Martin, M. G. F. (1998) ‘Setting Things before the Mind’, in A. O’Hear (ed.), Contemporary Issues in the Philosophy of Mind, Cambridge: Cambridge University Press. Martin, M. G. F. (2002) ‘The Transparency of Experience’, Mind and Language 17: 376–425. Martin, M. G. F. (2004) ‘The Limits of Self-Awareness’, Philosophical Studies 120: 37–89. Martin, M. G. F. (2006) ‘On Being Alienated’, in Tamar S. Gendler and John Hawthorne (eds), Perceptual Experience, Oxford: Oxford University Press. Maxwell, G. (1968) ‘Scientific Methodology and the Causal Theory of Perception’, in I. Lakatos and A. Musgrave (eds), Problems in the Philosophy of Science, Amsterdam: North-Holland Publishing. McDowell, J. (1982) ‘Criteria, Defeasibility and Knowledge’, Proceedings of the British Academy 68: 455–79. McDowell, J. (1994) Mind and World, Cambridge, MA: Harvard University Press. McLear, C. (2016) ‘Kant on Perceptual Content’, Mind 125(497): 95–144. Melnyk, A. (2002) Papineau on the Intuition of Distinctness’, SWIF Philosophy of Mind 4(1). Moore, G. E. (1903) ‘The Refutation of Idealism’, Mind 12(48): 433–53. Moore, G. E. et al. (1919) ‘Symposium: Is There Knowledge by Acquaintance?’, Proceedings of the Aristotelian Society, Supplementary Volume 2(1): 159–220. Moser, P. (1989) Knowledge and Evidence, Cambridge: Cambridge University Press. Nanay, B. (2016) ‘Philosophy of Perception: A Road-Map with Many Bypass Roads’, in Bence Nanay (eds), Current Controversies in Philosophy of Perception, London: Routledge. Nemirov, L. (1980) ‘Review of Thomas Nagel, Mortal Questions’, Philosophical Review 89: 473–7. Nemirov, L. (1990) ‘Physicalism and the Cognitive Role of Acquaintance’, in W. Lycan (ed.), Mind and Cognition, Oxford: Blackwell.

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi





Nida-Rumelin, M. (2007) ‘Grasping Phenomenal Properties’, in T. Alter and S. Walter (eds), Phenomenal Concepts and Phenomenal Knowledge: New Essays on Consciousness and Physicalism, Oxford: Oxford University Press. Noe, A. (2004) Action in Perception, Cambridge, MA: MIT Press. Noe, A. (2008) ‘Reply to Campbell, Martin and Kelly’, Philosophy and Phenomenological Research, 76(3): 691–706. Noe, A. (2012) Varieties of Presence, Cambridge, MA: Harvard University Press. Papineau, D. (1993) Philosophical Naturalism, Oxford: Blackwell. Papineau, D. (2002) Thinking about Consciousness, Oxford: Oxford University Press. Papineau, D. (2007) ‘Phenomenal and Perceptual Concepts’, in T. Alter and S. Walter (eds), Phenomenal Concepts and Phenomenal Knowledge: New Essays on Consciousness and Physicalism, Oxford: Oxford University Press. Passmore, J. (1957) A Hundred Years of Philosophy, New York: Basic Books. Pautz, A. (2011) ‘Can Disjunctivists Explain our Access to the Sensible World?’, Philosophical Issues 21: 384–433. Perry, J. (2001) Knowledge, Possibility and Consciousness, Cambridge, MA: MIT Press. Poston, T. (2007) ‘Acquaintance and the Problem of the Speckled Hen’, Philosophical Studies 132: 331–46. Poston, T. (2010) ‘Similarity and Acquaintance’, Philosophical Studies 147: 369–78. Price, H. H. (1932) Perception, London: Methuen. Pryor, J. (2000) ‘The Skeptic and the Dogmatist’, Noûs 34(4): 517–49. Putnam, H. (1999) The Threefold Cord, New York: Columbia University Press. Raleigh, T. (2014) ‘A New Approach to “Perfect” Hallucinations’, Journal of Consciousness Studies 21(11–12): 81–110. Raleigh, T. (2017) ‘Against an Inferentialist Dogma’, Synthese 194(4): 1397–421. Raleigh, T. (forthcoming) ‘Visual Acquaintance, Action and the Explanatory Gap’, Synthese. Recanati, F. (2010) ‘Singular Thought: In Defense of Acquaintance’, in Robin Jeshion (ed.), New Essays on Singular Thought, Oxford: Oxford University Press. Rey, G. (2005) ‘Explanation, Not Experience: Commentary on John Campbell, Reference and Consciousness’, Philosophical Studies 126(1): 131–43. Robinson, H. (1985) ‘The General Form of the Argument for Berkeleian Idealism’, in J. Foster and H. Robinson (eds), Essays on Berkeley: A Tercentennial Celebration, Oxford: Clarendon Press. Robinson, H. (1994) Perception, London: Routledge. Robson, J. (2013) ‘Appreciating the Acquaintance Principle: A Reply to Konigsberg’, British Journal of Aesthetics 53(2): 237–45. Robson, J. (2017) ‘Is Perception the Canonical Route to Aesthetic Judgement?’, Australasian Journal of Philosophy 1–12. Russell, B. (1903) Principles of Mathematics, Cambridge: Cambridge University Press. Russell, B. (1905) ‘On Denoting’, Mind 14: 479–93. Russell, B. (1911) ‘Knowledge by Acquaintance and Knowledge by Description’, Proceedings of the Aristotelian Society 11: 108–28. Russell, B. (1912) Problems of Philosophy, London: Williams and Norgate. Russell, B. (1914) ‘On the Nature of Acquaintance’, Monist 24: 1–16, 161–87, 435–53. Russell, B. (1921) Analysis of Mind, London: George Allen and Unwin. Sauchelli, A. (2016) ‘The Acquaintance Principle, Aesthetic Judgments, and Conceptual Art’, Journal of Aesthetic Education 50(1): 1–15. Sellars, W. (1956) ‘Empiricism and the Philosophy of Mind’, in H. Feigl and M. Scriven (eds), Minnesota Studies in the Philosophy of Science, Vol. I, Minneapolis, MN: University of Minnesota Press.

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

   Siegel, S. (2004) ‘Indiscriminability and the Phenomenal’, Philosophical Studies 120: 90–112. Siegel, S. (2008) ‘The Epistemic Conception of Hallucination’, in Fiona Macpherson and Adrian Haddock (eds), Disjunctivism: Perception, Action, Knowledge, Oxford: Oxford University Press. Siegel, S. (2010) The Contents of Visual Experience, New York: Oxford University Press. Smith, D. W. (1989) The Circle of Acquaintance, Cambridge: Cambridge University Press. Snowdon, P. F. (1980) ‘Perception, Vision and Causation’, Proceedings of the Aristotelian Society, New Series 81: 175–92. Snowdon, P. F. (1990) ‘The Objects of Perceptual Experience’, Proceedings of the Aristotelian Society, Supplementary Volume 64: 121–50. Sosa, E. (2003) ‘Beyond Internal Foundations to External Virtues’, in L. BonJour and E. Sosa (eds), Epistemic Justification: Internalism vs. Externalism, Foundations vs. Virtues, Malden, MA: Blackwell. Soteriou, M. (2013) The Mind’s Construction: The Ontology of Mind and Mental Action, Oxford: Oxford University Press. Stanley, J. (2010) Know How, Oxford: Oxford University Press. Stanley, J. and Williamson, T. (2001) ‘Knowing How’, Journal of Philosophy 98(8): 411–44. Stoneham, T. (2002) Berkeley’s World: An Examination of the Three Dialogues, Oxford: Oxford University Press. Sturgeon, S. (1998) ‘Visual Experience’, Proceedings of the Aristotelian Society, New Series 98: 179–200. Todd, C. (2004) ‘Quasi-Realism, Acquaintance, and the Normative Claims of Aesthetic Judgement’, British Journal of Aesthetics 44(3): 277–96. Travis, C. (2004) ‘The Silence of the Senses’, Mind 113(449): 57–94. Tye, M. (2009) Consciousness Revisited: Materialism without Phenomenal Concepts, Cambridge, MA: MIT Press. Varela, F., Thompson, E., and Rosch, E. (1991) The Embodied Mind: Cognitive Science and Human Experience, Cambridge, MA: MIT Press. White, S. (2007) ‘Property Dualism, Phenomenal Concepts, and the Semantic Premise’, in T. Alter and S. Walter (eds), Phenomenal Concepts and Phenomenal Knowledge: New Essays on Consciousness and Physicalism, Oxford: Oxford University Press. Williamson, T. (2000) Knowledge and Its Limits, Oxford: Oxford University Press. Wishon, D. (2012) ‘Perceptual Acquaintance and Informational Content’, in Sofia Miguens and Gerhard Preyer (eds), Consciousness and Subjectivity, Berlin: Ontos Verlag. Wishon, D. and Linsky, L. (eds) (2015) Acquaintance, Knowledge, and Logic: New Essays on Bertrand Russell’s The Problems of Philosophy, Stanford, CA: Center for the Study of Language and Information Press. Wittgenstein, L. (1953) Philosophical Investigations, ed. G. E. M. Anscombe and R. Rhees, trans. G. E. M. Anscombe, Oxford: Blackwell. Wollheim, R. (1980) Art and Its Objects, 2nd edn, Cambridge: Cambridge University Press. Worrall, J. (1989) ‘Structural Realism: The Best of Both Worlds?’, in D. Papineau (ed.), The Philosophy of Science, Oxford: Oxford University Press.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

PART I

Phenomenal Consciousness

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

1 Acquaintance Is Consciousness and Consciousness Is Acquaintance Joseph Levine

1. Introduction Acquaintance is a notion that has obviously played an important role in the history of philosophy of language, epistemology, and philosophy of mind.¹ With respect to philosophy of mind in particular, many have found it intuitive that there is some inherent connection between conscious experience and acquaintance. So, for instance, perception is often thought to be the primary vehicle of acquaintance, and is also the primary example of (phenomenal) consciousness. Russell indeed believed that sense-data—the principal denizens of conscious experience—were the only concrete entities with which we could be acquainted. On the other hand, many philosophers have hoped to reconstruct the phenomenon of acquaintance in a naturalistic manner that gives no special place to conscious experience in their account. In this chapter I want, first, to survey the various roles that acquaintance might play in philosophy of language, epistemology, and philosophy of mind; I will then explore the prospects for a naturalistic account of acquaintance to fill these roles; I will argue that while some roles can be filled by a naturalistic theory, others cannot; and finally, I will briefly present a non-naturalistic theory, according to which consciousness just is the relation of acquaintance, and show both how it accomplishes what a naturalistic theory could not but also how it cannot accomplish everything a naturalistic theory can.

2. Acquaintance Roles The first role for acquaintance concerns philosophy of language, one made prominent by Russell (1905). This is the role of semantic primitive. According to Russell, the only terms that are genuine referring expressions, as opposed to (disguised) descriptions, are terms for entities with which we are acquainted, like sense-data.

¹ For the locus classicus see Russell (1912).

Joseph Levine, Acquaintance Is Consciousness and Consciousness Is Acquaintance In: Acquaintance: New Essays. Edited by: Jonathan Knowles and Thomas Raleigh, Oxford University Press (2019). © Joseph Levine. DOI: 10.1093/oso/9780198803461.003.0002

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

   Normal proper names, for instance, only designate their bearers because they satisfy associated definite descriptions. But whether or not one adheres to such a strict notion of reference as Russell’s, something has to play the role of semantic primitive. It’s clear that the compositional mechanisms of our language cannot serve to generate an infinite class of interpretations for the infinite class of sentences unless there is first a primitive assignment to the atomic expressions of the language. But if a term is genuinely atomic, so not definable in other terms, then there must be a means for connecting the term to its referent without mediation by other conceptual mechanisms—that is, it must acquire its interpretation ‘directly’, as it’s sometimes put. It would be natural to call the mechanism of direct reference ‘acquaintance’. Another job for acquaintance comes from epistemology. A natural thought regarding justification, and one that underlies Descartes’s quest for indubitable foundations, is that justification for one belief cannot consist in an endless string of justificatory relations among beliefs, but must bottom out with beliefs that stand in no need of justification themselves—the foundational beliefs. One way to demarcate the set of foundational beliefs is to say that they consist in beliefs that are based on a relation of acquaintance. If I believe that there is a reddish, roundish object in front of me, that belief requires no more justification than my direct acquaintance with that object. Obviously nowadays foundationalism is not the only game in town, but it’s still one of the main contenders, and the role an acquaintance relation can play to provide the necessary foundational elements is straightforward. Another epistemological job for which acquaintance seems well suited is to provide a mechanism for ‘objectual’, as opposed to propositional knowledge—or ‘thing knowledge’, rather than ‘fact knowledge’. This role has recently been emphasized by Michael Tye (2009), and it’s one he exploits in order to respond to Jackson’s (1982) famous ‘Knowledge Argument’ against materialism. (I will discuss Tye’s position more below.) Generally it is assumed that items of knowledge are aptly described by ‘that-clauses’, expressing propositional contents. One knows that snow is white, that the sun is 93 million miles from the earth, and that 2 + 2 = 4. But we also often speak of knowing objects, people, and places. I know New York, Hilary Kornblith, and my car. Indeed, it’s pretty standard to describe such objectual knowledge as being acquainted with the place, person, or object in question. So to the extent one holds that objectual knowledge really is a distinct kind of knowledge, not reducible to propositional knowledge, one has reason to regard the relation of acquaintance with favour. The issue of objectual knowledge is related to a long-standing issue in philosophy of language concerning the distinction between de re and de dicto belief ascriptions.² De dicto belief ascriptions, which block quantifying in, attribute relations between subjects and propositions that are general in the sense of not containing individual objects. So with regard to the statement ‘John believes that the last person to leave the building should lock the door’, we don’t consider that

² A classic discussion is in Quine (1956).

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

  



John stands in any special relation to the last person to leave, but rather stands in the belief relation to the general proposition expressed by the ‘that’-clause. On the other hand, when we attribute the de re belief to John that Sheila should lock the door, we are prepared to say John believes something ‘of Sheila’. David Kaplan (1968) proposed that only statements employing ‘vivid names’ within the scope of the belief operator ascribe de re beliefs. A vivid name is one that puts the subject ‘en rapport’ with the object/person in question. Similarly, Gareth Evans (1982) enunciates a principle he calls ‘Russell’s Principle’: ‘that in order to be thinking about an object or to make a judgment about an object, one must know which object is in question—one must know which object it is that one is thinking about’, which assumes that there is a principled notion of ‘knowing who or what’ someone/something is. Being ‘en rapport’ with someone or something and ‘knowing who or what’ someone or something is, are usually taken to require acquaintance with the object in question. Related to all of these roles, perhaps underlying them all, is the role of serving as the mechanism of direct cognitive, intentional contact between mind and world. Whenever one thinks of acquaintance, the idea of cognitive immediacy or directness comes to mind. What we can think about might seem to be limited in some way by, or at least bootstrapped from, what we can be acquainted with. Cognitive immediacy is what enables foundational knowledge and primitive semantic interpretation. Acquaintance just isn’t acquaintance without cognitive immediacy.

3. A Naturalistic Acquaintance What I mean by ‘naturalism’ for our purposes is the doctrine that all of the basic properties and relations in nature are non-mental and non-intentional.³ Any new physical properties and relations discovered or postulated by practitioners of fundamental physics count as natural, for my purposes, so long as they aren’t recognizably mental or intentional. So charm and spin are natural, but a fundamental seeking or thinking are not. Naturalism about the mind, then, is the doctrine that all mental properties and relations are realized by non-mental properties and relations. Partly because I think it’s true, but also because it’s easier for my exposition, I will assume that the most promising naturalist theory of mind is the ComputationalRepresentational Theory of Mind (CRTM).⁴ On this view mental states are relations between subjects and mental representations. The relations are computational in the sense that they are realized in non-mental mechanisms, but serve to underwrite interpretations to the representations under which the operations defined over these representations come out semantically coherent. Thus the functional mechanisms that realize inference look like inferences when the representations involved are assigned the appropriate interpretations. What determines the interpretations is of course a big question for such theories, but for present purposes I will assume some form of ‘information-based semantics’, the idea that a mental representation is about ³ For a discussion of naturalism in philosophy of mind see Levine (2016). ⁴ See Fodor (1975) for the classic presentation and defence of the doctrine, and also Rey (1996), from which I took the term ‘CRTM’.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

   what it’s about in virtue of standing in the appropriate causal-covariational relation to its referent.⁵ Let’s see how the various roles for acquaintance characterized above would fit into CRTM. While I don’t want to rule out non-discursive, iconic representational formats, especially for perceptual systems, let’s make the simplifying assumption that all mental representations are language-like, with atomic symbols and complex symbols constructed according to combinatorial principles that determine both wellformedness and compositional semantic values. So the obvious first role for something like acquaintance is to determine the semantic values of the atomic terms, both singular and general. In Levine (2010a) I explored this question through a discussion of the nature of demonstrative thought. There I distinguished between two kinds of ‘meta-semantic’ mechanisms: IMMs (intentionally mediated meta-semantic mechanisms) and DMMs (direct meta-semantic mechanisms). Meta-semantics, a term taken from Kaplan (1989), is distinguished from semantics in that the latter tells us what the semantic values of the expressions of a language are, while the former explains how it is that the expressions acquired these semantic values. With regard to the complex expressions of the ‘language of thought’, we know that their values are determined by the values of the atomic symbols together with the compositional semantic rules that align with the combinatorial syntactic principles of the language. The main metasemantic question then comes down to how the values of the atomic symbols are determined. It’s plausible that the ultimate DMM, upon which all other meta-semantic mechanisms are based, involves perceptual demonstratives. However precisely we formulate the causal-covariational conditions for a symbol meaning a certain object or property, it seems plausible that our perceptual contact with the object, or instances of the property or kind, play a foundational role. My thinking ‘that x’ when perceptually confronted with an x seems to serve as the most basic point of intentional contact between mind and world. Assuming that an appropriate causal condition that is characterizable in non-intentional terms can be formulated on which a primitive percept is of a particular object, and one’s demonstrative thought employs that percept in demonstrating the object, we have the DMM we’re looking for. What’s more, the kind of cognitive contact we enjoy with the objects we can perceptually demonstrate seems appropriately described as a form of acquaintance. So here we have a naturalistic account of acquaintance that seems to fill one of its most important roles. Let’s turn to epistemological roles. The two related ones we described above were: serving as the foundational elements in justification and constituting a ground of certainty, thus allowing the ultimate justifiers to themselves not require justification. Sense-data clearly played these roles, as did ‘clear and distinct’ ideas for Descartes. Can the items we identified above, the objects of perceptual demonstratives, also play the epistemological roles? To an extent, and given some qualifications, I’d say that they can.

⁵ See Dretske (1981) and Fodor (1990).

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

  



When it comes to certainty, there is no naturalistically realizable cognitive relation that can deliver. The objects of perception, available for demonstration, are not guaranteed to be as I perceive them to be, nor to be there at all. But most epistemologists have long given up the quest for absolutely certain foundations, allowing error to creep in right from the very start. What does matter, however, if we are to preserve any notion of an epistemological foundation, is that there be elements in the chain of justification that are not themselves justified by, or even cognitively affected by, other elements in the chain. That is, that there be something that plays the role of ‘the Given’. One way to achieve this foundational character for perceptual judgement is to endorse a version of the modularity thesis. According to Fodor (1983), perceptual and other ‘input’ systems (specifically, language), are ‘informationally encapsulated’. That is, the process of forming a percept is essentially ‘bottom up’, so all that the perceptual systems have to go on in formulating their hypotheses concerning the distal environment are the impinging energy on the system’s receptors and a highly restricted data base of information concerning general features of the external world. What one expects, desires, or believes, on this view, does not affect what one sees. Fodor himself emphasizes how the modularity thesis helps deliver a principled observation-theory distinction, and therefore also a notion of intersubjective objectivity. Two people can differ in myriad ways in their world views, coming from perhaps wildly different cultures, yet still be constrained to see the litmus paper as blue (assuming it is and their visual system is working properly and in decent light). The primitive cognitive relation between world and percept will not deliver certainty, but it still can deliver, if the perceptual psychology works out right, a foundation. However one comes down on Fodor’s modularity thesis, it does seem that a Given can only be reconstructed within the naturalistic framework of CRTM if there is a principled cognition-perception boundary. If one assimilates perception into cognition, or vice versa, one loses any basis for isolating some representations as foundational, having the Given as their contents. But of course those who seek to blur this line also disparage the notion of the Given and so don’t mourn its loss. Let’s turn to the question of objectual knowledge and the related issue of de re knowledge. As described in Section 2, de re knowledge (or belief) entails a principled notion of knowing who or what something is. I can believe that the best student in the class (whoever it turns out to be) will get an A without knowing who that student is, but I can’t believe of the best student in the class that she will get an A without knowing who she is; or, as we might say, without knowing her. This seems to entail that there is a principled distinction to be made between those people, or objects generally, I can be said to know and those I can’t. But can such a distinction be maintained? To my mind this distinction is a pragmatic one, and applies differently depending on context. If asked whether I know Donald Trump, I’d say ‘no’. If asked, do I know who Donald Trump is, I’d say ‘yes’. Usually questions about knowing someone or something, or knowing who/what someone/something is involve a contextually determined set of criteria for an affirmative answer. Does one know the person’s name? Does one know the generally known facts about the person? Is one personally ‘acquainted with’ the person?—a question that itself has contextually determined

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

   criteria. To illustrate that last point, note that even if one has a perceptual encounter with Donald Trump, which seems the best candidate for a principled non-pragmatic sense of being acquainted with him, for most conversational purposes it doesn’t count as being acquainted with him. Otherwise all those people who attended his rallies—or have seen him on TV—would count as being acquainted with him. The problem I see for coming up with a principled notion of knowing who or what someone/something is underlies the reason I don’t think CRTM can underwrite a notion of objectual knowledge. The problem with finding a principled distinction between knowing what something is and not knowing what it is derives from the fact that knowing anything with respect to an object (or person) involves bringing it under a description. If that’s the case, then the principled distinction we’re looking for would have to be characterized in terms of a distinction between two kinds of description: those that capture what an object is and those that don’t. The only principled way I know of for making that distinction appeals to essential properties. But, assuming an object’s essential properties are given by scientific descriptions (if they exist at all), most cases of knowing what something is won’t count. But any other basis looks hopelessly pragmatic and contextual, appealing to conversational norms and the like. Notice that even using essential properties as the distinctive factor doesn’t quite work. For which descriptions count as expressing an object’s essential properties? Take, for instance, the notion of knowing what water is. A number of philosophers have worried about the apparent consequence of externalist semantics for selfknowledge of content.⁶ Supposedly when I believe that water is what comes out of my tap I know what I believe—so I know the content of my belief. Some have worried that if it’s essential to water that it’s composed of H₂O molecules, then I wouldn’t know what it is I believe when believing that water comes out of the tap unless I also believed that water is composed of H₂O molecules. But if water is in fact identical to a substance that is composed of H₂O molecules, when I apply the description (or term) ‘water’ to the stuff coming out of my tap, aren’t I in fact attributing to it the property of being composed of H₂O molecules? What is it I’m lacking? It seems to me that here too, what counts as my knowing what water is depends on contextually determined pragmatic considerations. It’s a matter of which and how many descriptions I’m able to apply to the stuff in question, and how explanatorily connected they are. A premise in the argument above was that the only way we have of characterizing knowledge of a thing involves bringing it under a description. Once we have predication, we have a structure in thought that can be truth-evaluable, and thus a candidate for the object of belief, so long as we also have the interpretations of the predicates and singular terms. We then need to distinguish belief from other propositional attitudes, such as desire, and the only way I know of that CRTM can do that is through functional role. Once we have true (and false) belief characterized, we then need the conditions that must be added to true belief to make it knowledge, and we all know how that endeavour goes. But the main point here is

⁶ See Wright et al. (1998) for a number of papers on the topic.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

  



that only with predication can a notion of belief emerge at all, and thus bringing under a description is the basis of any cognitive relation CRTM can underwrite. As Kant might say, to know an object is to bring it under a concept. But once we do that, especially on CRTM, the object of our cognition is not the object itself, but rather the propositional content that is expressed by the composed representation consisting of the representation of the object and the concept under which we brought it. On CRTM, it seems, the only sort of intentional relation involving a singular term in thought is the representational relation itself; when I use a token of a proper name, or demonstrative, in thought, I am representing the relevant person or object. But this relation of representation, though the fundamental intentional relation, is not really a mental relation itself. Rather, it is the necessary condition for there being any mental relations. I don’t represent as a mental act—it isn’t something I do—but rather it is something that happens by virtue of my brain’s standing in the relevant causal/covariational relation to the object/property. Put another way, it isn’t really a relation between the subject and the object, but rather a relation between a representation, a symbol, and the object. But cognition is something the mind—the subject—does, something that goes beyond mere representation, though it presupposes it. On CRTM the only way for that to happen is for a predicate to be applied to the singular representation and for the subject to stand in the appropriate functional/computational relation to the composed singularpredicate representation. To illustrate my point about the non-mental character of primitive representation, consider this contrast between two ways of representing exemplified by the practice of using ‘bitter herbs’ during the Passover Seder to symbolize the suffering of the Israelite slaves in Egypt. When I use the bitter herbs in this way, I am mentally associating two objects and deciding to let one represent the other. But I couldn’t do that unless the thoughts I had about the bitter herbs and the suffering already met the conditions for representing them. When I thought ‘bitter herbs’ and ‘suffering’, that I did without any intention to represent them—I just tokened those representations. What did the representing worked behind my back, as it were; again, it’s something that happened to me, I didn’t do it. On the other hand, the deliberate employment of the bitter herbs to represent the suffering was indeed something I did; it was a mental act of mine. My point is that where the acquaintance relation for CRTM is doing its work is below the threshold of mental activity, behind the scenes, and so cannot support a principled notion of knowing an object, or knowing what one is referring to or thinking about. We only get genuine mental activity, on CRTM, once representations are combined into a truth-evaluable unit and a functional attitude toward that complex representation is taken. If this is right, then it undermines a recent argument of Michael Tye’s (mentioned above) that seeks to employ the notion of acquaintance to support a principled distinction between ‘thing-knowledge’ and propositional knowledge. On Tye’s view, we can use this distinction to reply to Jackson’s ‘Knowledge Argument’ against materialism. According to the argument, Mary, the super-scientist in the black-andwhite room, knows all there is to know about colour vision and colour experience that can be learned from science texts and experiments. When she leaves the room and sees red for the first time she seems to gain some knowledge; namely,

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

   what it’s like to see red. According to the argument, since she already knew all of the physical-functional facts before leaving the room, what she learns upon leaving must involve a non-physical-functional fact. This conclusion stands in conflict with materialism. Tye claims that what happens upon leaving the room is that Mary becomes acquainted with red for the first time and thus acquires objectual knowledge. Knowing what it’s like to see red just is being acquainted with red, which can be characterized as acquiring objectual knowledge of red. Since everything she knew in the room about colour consisted of propositional knowledge, it is true that she gains new knowledge upon seeing red for the first time. However, it isn’t knowledge of something new—as the argument purports to show—but rather a new kind of knowledge of the very same thing. Since we can capture what Mary learns in terms of her becoming acquainted with red and thereby gaining objectual knowledge of it, we are not forced to posit non-physical-functional properties to adequately characterize Mary’s new cognitive state. While I do think something happens to Mary when she sees red for the first time that is accurately described as her ‘becoming acquainted’ with red, and thereby gaining objectual knowledge (or non-propositional knowledge, at any rate), as I argued above, I don’t believe that CRTM can support this claim—at least not in a principled way. Mere intentional contact, of the sort underwritten by the causal impingement from the red object to the visual system, is not acquaintance, not a kind of cognition at all, but rather the precondition of any cognition. Genuine cognition only begins with a truth-evaluable complex. The issue of objectual knowledge and the issue of cognitive immediacy, or directness, come down to the same thing I think. I mentioned earlier that underlying all the jobs for acquaintance is its serving as an immediate relation between subject, or mind, and object, or world. We have seen that within the framework of CRTM there is only one point of immediacy, at the foundational level of intentional contact, whereby a symbol comes to have its semantic value by virtue of a DMM. But this fundamental intentional relation, I have argued, is not a genuinely cognitive relation; it holds between symbol and object, not mind and object, and it happens to the subject, and is not something done by the subject. To the extent we want to hold on to a notion of genuine cognitive immediacy through acquaintance, it seems we will have to abandon the naturalistic framework. This is of course a pretty strong conclusion. One way for a naturalist to respond is to just give up on the notion of cognitive immediacy as an essential feature of acquaintance. After all, just like ‘the Given’, many items that seem intuitively compelling are subject to ontological expulsion in the name of best overall theory of the world, and why not this robust sort of acquaintance as well? To this I don’t have a counter-argument other than to say that I think anyway there are lots of reasons to be suspicious of naturalism about conscious experience, so the case against naturalism doesn’t rest here alone. On the other hand, one might want to push back against the argument above in the following way. What we’re looking for is a relation between mental subject and some object that is genuinely cognitive and also immediate so that it adequately conforms to what we intuitively think of as acquaintance. So why, the CRTM naturalist objects,

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

  



can’t perceiving an object count? After all, we can suppose, as described earlier, that the relation between the perceived object and the percept, by virtue of which the latter is about the former, is non-intentionally mediated. On the other hand, the perceived object isn’t merely represented, it’s perceived—and isn’t this just the kind of cognitive relation between subject and object we’re looking for? In reply, I want to make two points. First, I agree that when we perceive an object we do stand in this cognitively immediate relation of acquaintance to it. This may not be the only case of acquaintance, but it certainly is the paradigmatic one. But the question isn’t what happens in perception, but whether this thing that so compellingly seems to happen can be adequately explained in terms of CRTM. Second, I contend that it can’t. While the primitive intentional relation realized by causal-covariational mechanisms can, perhaps, adequately explain how the percept is about the object, it doesn’t yet amount to cognizing the object. What we need here is some attitude toward the representation, and the representation has to ‘say something’—if only, there is something out there in such-and-such a location. Again, CRTM seems to give us a way of building an ‘awareness that’ which is fed from perception to cognition, but I don’t see it gets us just ‘awareness of ’ per se. The only direct relation is between the percept and the object, and that’s just the ground-level representation relation, not, as emphasized above, a genuinely mental relation at all.

4. Acquaintance as a Primitive Non-Natural Relation For our purposes, what I mean by a ‘natural’ relation is any relation for which there is a realization theory adverting to only non-mental, non-intentional properties and relations. CRTM counts as a naturalized theory because all mental relations— crucially, the foundational representation relation and the attitude relations—are cashed out in terms of causal or evolutionary history and nomic covariation, or, for the attitudes themselves, functional relations, which are all non-intentionally characterizable. I’ve argued that a direct, immediate, cognizing relation between subject and object is not something that can be reconstructed in these terms. What’s the relation between acquaintance and consciousness? I propose that they are just one and the same. Of course there has been a lot of literature recently about phenomenal vs. access consciousness, along with much attention devoted to the seemingly peculiar metaphysical features of phenomenal properties.⁷ For a long time in philosophy of mind the problem of ‘qualia’, the older name for phenomenal properties, was strictly distinguished from the problem of intentionality. Qualia— the reddishness of what’s it’s like to see red, or the qualitative character experienced when smelling a rose—were treated as what Mackie (1977/90), speaking of moral properties, called ‘queer’ properties. They were thought to be metaphysically nonnatural properties that were instantiated in experiential states and what made those states the kinds of phenomenally conscious states they were. If they had any representational features—were about anything—what they represented was thought ⁷ For the phenomenal-access consciousness distinction, see Block (1995).

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

   to be accidental, not essential to them. Materialists then tried to show how such apparently queer properties could be realized in natural, ultimately physical properties, while anti-materialists resisted this reductive enterprise. More recently philosophers have come to question this strict distinction between phenomenal, or qualitative character and intentionality. This has come from both directions: those who want to reduce phenomenal character to intentional content as a way of executing the materialist reduction of the phenomenal,⁸ and those who see intentionality as inherent in phenomenal consciousness, and thus undermining even the naturalization of intentionality. The recent popularity of the ‘phenomenal intentionality programme’ is evidence of this latter trend.⁹ I see my current position as within the phenomenal intentionality fold, though with a crucial difference, which I will get to below. Basically, as I see it now, phenomenal consciousness just is a kind of intentional relation—it is pure awareness, or acquaintance, with whatever one is conscious of. There are two slogans of Sartre’s (1956/92) that to me beautifully capture the essential features of conscious awareness: first, that all consciousness is consciousness of, and second, that consciousness is ‘nothingness’. The first slogan captures the essentially relational and intentional character of conscious awareness, and the second captures the fact that all there is to conscious experience that is amenable to substantive characterization derives from its object, what it is we are conscious of. In this sense, the recent emphasis on the so-called ‘transparency’ of conscious experience, deriving from G. E. Moore’s discussion, and revived by Harman (1990), is also consonant with Sartre’s view of the ‘for itself ’ as non-substantial, a kind of ‘nothingness’. I propose, then, to treat consciousness as just the same thing as conscious awareness, which is the same thing as this special relation of acquaintance we’ve been trying to pin down. Consciousness (phenomenal consciousness, the only kind I’m talking about here) is not, then, a matter of instantiating certain ‘queer’, ‘feely’ properties, but rather a matter of experiencing, or being acquainted with, a world that is presented to the subject.¹⁰ I will call the relation ‘Conscious Awareness’ (or the CA relation), and my claim is that this relation is what constitutes both phenomenal consciousness and acquaintance. As I conceive of it, CA is a primitively intentional relation holding between the subject of conscious experience and whatever objects the subject is consciously aware of. One interesting question is what the nature of the object of conscious awareness is, and just which ontological categories are represented there—concrete objects, properties, or perhaps also propositional contents. By saying the relation is ‘primitively intentional’ I mean that its intentionality is not constituted by a nonintentional relation (or combination of such relations), such as causal covariation. It holds of its relata as a basic relation. I also mean to convey that the intentionality is inherent in the relation itself, and not a matter (as with CRTM) of inheriting its

⁸ See Harman (1990), Dretske (1995), and Tye (1995). ⁹ See Kriegel (2013) for a number of papers generally supporting this line. ¹⁰ I have discussed this view in three previous papers, Levine (2006, 2008, and 2010b). In those papers I referred to the CA relation as the ‘AA’ relation, standing for Acquaintance Appearance.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

  



intentionality from the intentionality of a representation tokened by the subject. So, to take a paradigmatic instance of conscious awareness, when I visually experience my computer screen as I type, I, the subject, am consciously aware of the screen. This is a relation I bear to the screen. This relation—being aware of the screen—is thereby intentional, as it is directed upon, in some sense about, the screen. But, to repeat, this isn’t a matter of my bearing an attitude toward a token in my head that represents the screen. The intentionality is primitively in the relation of awareness itself, not in some representation that mediates my relation to the screen. In Section 2 I listed a number of functions that acquaintance has been asked to play in the literature, so let’s see how CA does. Let’s begin with the job that seemed troublesome for the CRTM model of acquaintance, capturing the phenomenon of objectual knowledge. Now ‘knowledge’ is a tricky word, with relations to truth, belief, and justification built in. So whatever is meant by ‘objectual knowledge’, it’s important not to engender these commitments in analysing it. Rather, as Chomsky did a long time ago for another purpose, let’s consider a more neutral term, like ‘cognize’. I think what people mean by saying that acquaintance affords objectual knowledge of what we are acquainted with can be captured by saying that acquaintance with an object is a form of cognizing it. Cognizing is intentional, but it goes beyond mere representation in being cognitive—involving the mind’s ‘taking in’ the object in question, ‘grasping’ it. Whether it meets all the conditions one requires for fullfledged knowledge is irrelevant here. Consider again Tye’s attempt to respond to the Knowledge Argument by appealing to objectual knowledge. He argued that though Mary knew all the facts about red and experiences of red before leaving the room, she did gain a new kind of knowledge when she saw red for the first time, namely objectual knowledge. So there was no new subject matter, but rather a new mode of cognizing. Since the Knowledge Argument seemed to rely on there being new things known, not a new way of knowing, this response plausibly undermined the argument. But if what I’ve presented above is right, Tye’s response doesn’t help the materialist. Perhaps what it is Mary is acquainted with is in fact identical to some physical (or physically realized) property, though we will have reason to reject that presently. Still, the relation of acquaintance, the kind that supports cognizing an object (or property, taking ‘object’ here quite loosely), is itself a basic intentional relation, so we still have ontological commitments that take us beyond those of materialism. In this section I’ve introduced the CA relation and characterized it as a basic, intentional, cognitive relation between conscious subject and the objects of which the subject is conscious. I claimed that one of the chief features of the traditional notion of acquaintance—that it affords a different kind of knowledge, objectual knowledge—can be sustained for the CA relation, but not for any version of acquaintance constructible within the naturalistic CRTM. But what about the other features of acquaintance we discussed above, the ones which CRTM seemed capable of capturing? Interestingly, I think CA actually cannot do that work, and so I argue for a division of labour when it comes to acquaintance. The reasons I say this also bear on the difference between my view of phenomenal intentionality and that of many others who talk about it.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

  

5. The Division of Intentional Labour In order to develop the points I want to make here, it’s helpful to begin with a question I haven’t yet addressed: what are the objects of conscious awareness, or acquaintance? Acquaintance is normally thought to relate us at least to concrete objects and their properties. So when Mary sees a red tomato for the first time after leaving the black and white room she becomes acquainted both with the tomato and with its redness. Tomatoes are concrete physical objects, and redness, let’s assume for now, is some complex property of the surface involving its spectral reflectance profile. Does the CA relation hold between a conscious subject and these physical objects and properties? I think there are two reasons to answer in the negative. For one thing, though I have displaced ‘queer’ qualia from being the defining feature of phenomenal consciousness, focusing instead on the inherently and primitively intentional relation of conscious awareness, there is still something ‘queer’ about qualitative properties like redness, the odour of a rose, and the taste of coffee, not to mention the feeling of pain. For all the reasons I and other anti-materialists have argued there is an explanatory gap between the phenomenal and the physical/functional, it doesn’t seem plausible to identify qualities like redness with, say, surface spectral reflectances, or the qualitative characters of phenomenal states with internal physical/functional states.¹¹ The secondary qualities generally appear singular in two ways. First, as purely qualitative and intrinsic properties, they don’t seem identifiable with a causal role property, and it seems that all the properties that are appealed to in any naturalistic account of the world are ultimately reducible to something like a causal role. As Chalmers (1996) has emphasized, physical properties ultimately are either structural or functional, and the secondary qualities seem to be neither of these. The second singular feature of the secondary qualities is their especially intimate connection to conscious awareness. In (Levine 2006) I described properties like redness, loudness, and painfulness as ‘ways of appearing’. Following to an extent Berkeley’s view that for the sensible qualities ‘esse et percipi’, it seems essential to, say, colour that it is a way of appearing to a conscious subject, just as pain seems to be. If this is right, then what seems to be the case is that conscious awareness relates us to properties that themselves are inherently tied to their being objects of conscious awareness. Only within conscious experience do ways of appearing seem to be instantiated. If this is right, then it isn’t the actual external physical properties of objects with which we are acquainted, but rather properties that themselves have no existence outside of conscious experience. What then about the concrete objects themselves? Perhaps the redness of the tomato is a property it has only in the context of being the object of conscious awareness (in particular, visual experience), but one might think it’s still the case that the tomato itself is a direct object of acquaintance. By experiencing its redness I am aware of the tomato. However, what do we do about hallucination?

¹¹ See Chalmers (1996) and Levine (2001) for extended arguments on this question.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

  



As I am eschewing a representational theory of conscious awareness in favour of a direct intentional relation between subject and object, it might seem I should adopt the view of many ‘direct perception’ theorists on the question of hallucination; namely, ‘disjunctivism’.¹² On that view, the phenomenal state one is in when veridically viewing a tomato is fundamentally distinct from the one we are in when hallucinating a tomato, and the difference is that the former is essentially a relation between the subject and the tomato. I personally do not want to adopt disjunctivism, for two reasons. First, I do think there’s something fundamentally in common between veridical and hallucinatory experiences—I think the structure of conscious awareness is the same in both cases. I won’t argue for this here, but it’s one of my principal reasons for rejecting disjunctivism. Second, even if disjunctivism were right, there is still the question of how to characterize the objects of acquaintance in hallucinatory experiences. Of course one thing to say is that when hallucinating one is acquainted with nothing. But this seems inadequate. Imagine occupying a visual state that is qualitatively, for you, just the same as you would have were you looking at a ripe tomato in good light. Perhaps you’re a brain in a vat. Clearly we want to say that there is something it’s like for you to have this experience, and also that experiencing redness partly characterizes what it is like. But if conscious awareness is relational, and the contents of experience are what it is you are conscious of, then we have to say something about the status of the apparent red tomato in your hallucinatory experience. Once we find ourselves required to find an object for hallucinatory experiences, it only makes sense to provide the same sort of object for veridical experiences and unify the two categories. What I propose is that conscious awareness involves a relation between subjects of conscious experience and a ‘virtual world’, one that is composed of the (virtual) objects and properties we seem to encounter in experience. While this might sound like a sense-data doctrine—and indeed does bear some similarity to it—it is different in the following way. I don’t take there to be individual qualities that obtain in a mental realm and that somehow combine in a way to form the normal objects of experience. Rather, I take conscious experience to be the result of a constructive process which begins with a structured and complex description of a world surrounding the subject and which then is presented in its structural complexity to the subject as what she is consciously aware of. What is the input to this constructive process? I propose that the input is precisely the representations of the world produced by our naturally realized cognitive and perceptual systems. An analogy I like is to consider conscious experience to be a movie (or ‘Cartesian Theater’ if you like) and the ongoing, constantly updated cognitive/perceptual representation of the world to be its script. The reason I think you have to tie the contents of the virtual world—with the conscious subject as its sole audience member—to the systems studied by cognitive science is that so much of the fine-grained structure of conscious experience is explained by the natural cognitive and perceptual systems. One of the principal arguments materialists have been

¹² For direct perception and disjunctivism see Brewer (2011) and Martin (2004).

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

   making for decades is that by appeal to the functional structure of cognition and perception—as revealed through cognitive science—one can explain quite a lot about the nature of our experience. This has led many philosophers and cognitive scientists to then infer that conscious experience just is the tokening of the relevant cognitive and perceptual representations; a conclusion that then also allows them to dispense with the non-materialist metaphysics one is otherwise stuck with. I find the evidence of the intimate connection between the representational structure produced by cognition and perception and the fine-tuned structure of conscious experience to be extremely persuasive. However, rather than reduce the latter to the former, for all the reasons cited above, I am inclined to see the latter as determined by the former in the way just described. That is, in conscious experience we somehow (and this is the non-naturalistic part) get a conversion from a systematic representation of the world in our brains to a (virtual) world of objects and qualities presented to us, as subjects, to be aware of. If we see conscious experience as this output of the cognitive machine in our brain—that device captured by CRTM—then two interesting and related consequences follow. First, unlike most of the philosophers who have signed on to the ‘phenomenal intentionality programme’, I do not see phenomenal intentionality as the source of all intentionality, the ground of so-called ‘original intentionality’. Most philosophers who have discussed intentionality have distinguished between ‘derived’ intentionality and ‘original’ intentionality. Standard examples of the former are the intentional properties of natural language systems and the representational properties of various symbols and pictures. Usually the semantic features of computer languages are included in the derived category. What makes these vehicles possess only derived intentionality is that their meaning what they mean depends on the intentional states of people’s mental states (and intentionally sustained conventions). The idea, then, is that the intentional buck stops with our own minds. Now some philosophers have, understandably, despaired of there being a naturalistic account of original intentionality, finding the nomic-covariation and related accounts associated with CRTM to be inadequate. So some have looked to the intentionality of conscious experience—phenomenal intentionality—as the only truly original form of intentionality, and to then accord intentional status to unconscious brain states by virtue of their causal or other relations to conscious experience. Thus, on this view, all intentionality outside phenomenal intentionality is derived.¹³ However, on the view I’ve taken here—that conscious experience is the result of the computational-representational activity of the mind-brain—one cannot consider the intentionality of this very constructive activity to be derived from its result. In some sense the intentionality of the script must precede that of the movie. I am not saying that phenomenal intentionality, the intentionality of conscious experience, is derivative from that of the computational machine in the way that the intentionality of natural language and artificial symbol systems is. For one thing, as I emphasized above, phenomenal intentionality is intrinsically embedded in the relation of conscious awareness, which itself I treat as a basic, non-naturalistic

¹³ Again, see a number of papers in Kriegel (2013).

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

  



relation. Secondly, when it comes to the pure qualities like colour, odour, pains, and tickles, they first make their appearance within the virtual world of conscious experience, so intentional directedness on them can’t be derivative from the representational states realized in the computational mechanisms of the brain. I did say their fine-tuned structure is governed by the computational mechanism, but this doesn’t extend to their intrinsic qualitative character. So on my view, there are two sources of original intentionality: one that is grounded in whatever naturalistic causal or nomic relations that attach the computational mechanism to the world, and the other that arises within conscious experience. Neither one is derivative, in the requisite sense, from the other, though one is a (causal) result of the other. So while I hold that there is a kind of sui generis phenomenal intentionality, I do not hold that it is the basis for all intentionality. But this entails another consequence, taking us back to the original list of jobs that acquaintance is supposed to perform. Epistemologically, acquaintance was supposed to provide the Given for justification, and metaphysically and semantically it was supposed to provide the locus of mindworld connection. We saw earlier how the kind of acquaintance that is constructible within CRTM could plausibly fill these roles, to the extent one held they needed filling. (That is, many reject both foundationalism in epistemology and the very existence of a Given.) But if I am right in my characterization of conscious awareness, then it can’t fill these roles for acquaintance. The fundamental mind-world connection is provided on my model by the computational machine, while the relation holding between conscious subject and virtual world is a reflection of the base level relation. Similarly, when it comes to justification, which must advert to truth conditions that involve a genuinely mind-independent world, there too all the foundational work—if there be any—is done by the interaction between stimuli from the world impinging on our sense organs. What we are conscious of is a function of what impinges on us together with what the computational machine already represents, and so is not a further input into that machine. When I enjoy the visual experience of my typing on the computer screen in front of me, all the evidence for my beliefs has already been processed and cognitive conclusions drawn before my conscious awareness sets in. What I see in the Cartesian Theatre is, on this view, but an epiphenomenal display of the results of the work going on behind and under it.

6. Conclusion In this chapter I have surveyed various jobs that the notion of acquaintance has been assigned by philosophers over the years. These include both epistemological and semantic, or intentional jobs. I’ve argued that a naturalistic account of acquaintance consistent with the programme of CRTM can capture many of these roles, but not the role of exemplifying objectual knowledge. A non-naturalistic account of acquaintance that just takes it to be the non-naturalistic relation of conscious awareness does capture this immediate cognizing of an object by a conscious subject, but it cannot play the epistemic and semantic roles for which CRTM is well suited. If I am right, we need two different theories of intentionality to cover two different phenomena: the intentionality of the states studied in cognitive science and the intentionality of conscious experience.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

  

References Block, N. (1995) ‘On a Confusion about a Function of Consciousness’, Behavioral and Brain Sciences 18(2). Brewer, B. (2011) Perception and Its Objects, Oxford: Oxford University Press. Chalmers, D. (1996) The Conscious Mind, Oxford: Oxford University Press. Dretske, F. (1981) Knowledge and the Flow of Information, Cambridge, MA: Bradford Books/ MIT Press. Dretske, F. (1995) Naturalizing the Mind, Cambridge, MA: Bradford Books/MIT Press. Evans, G. (1982) The Varieties of Reference, Oxford: Oxford University Press. Fodor, J. A. (1975) The Language of Thought, New York: Thomas Crowell, and Cambridge, MA: Harvard University Press. Fodor, J. A. (1983) The Modularity of Mind, Cambridge, MA: Bradford Books/MIT Press. Fodor, J. A. (1990) A Theory of Content and Other Essays, Cambridge, MA: Bradford Books/ MIT Press. Harman, G. (1990) ‘The Intrinsic Quality of Experience’, in J. Tomberlin (ed.), Philosophical Perspectives, 4, Action Theory and Philosophy of Mind, Atascadero, CA: Ridgeview Publishing. Jackson, F. (1982) ‘Epiphenomenal Qualia’, Philosophical Quarterly 32: 127–36. Kaplan, D. (1968) ‘Quantifying In’, Synthese 19. Kaplan, D. (1989) ‘Demonstratives’, in J. Almog, J. Perry, and H. Wettstein (eds), Themes from Kaplan, Oxford: Oxford University Press. Kriegel, U., ed. (2013) Phenomenal Intentionality, Oxford: Oxford University Press. Levine, J. (2001) Purple Haze: The Puzzle of Consciousness, New York: Oxford University Press. Levine, J. (2006) ‘Color and Color Experience: Colors as Ways of Appearing’, Dialectica 60(3). Levine, J. (2008) ‘Secondary Qualities: Where Consciousness and Intentionality Meet’, Monist 91(2). Levine, J. (2010a) ‘Demonstrative Thought’, Mind and Language 25(2). Levine, J. (2010b) ‘Phenomenal Experience: A Cartesian Theater Revival’, Philosophical Issues 20: Philosophy of Mind: 209–25. Levine, J. (2016) ‘Naturalism and Dualism’, in K. Clark (ed.), The Blackwell Companion to Naturalism, Chichester: Wiley-Blackwell. Mackie, J. L. (1977/90) Ethics: Inventing Right and Wrong, London: Penguin. Martin, M. G. F. (2004) ‘The Limits of Self-Awareness’, Philosophical Studies 120. Quine, W. V. (1956) ‘Quantifiers and Propositional Attitudes’, Journal of Philosophy 53(5). Rey, G. (1996) Contemporary Philosophy of Mind: A Contentiously Classical Approach, Oxford: Blackwell. Russell, B. (1905) ‘On Denoting’, Mind 14(56): 479–93. Russell, B. (1912) The Problems of Philosophy, New York: Henry Holt. Sartre, J. P. (1956/92) Being and Nothingness, New York: Washington Square Press. Tye, M. (1995) Ten Problems of Consciousness: A Representational Theory of the Phenomenal Mind, Cambridge, MA: Bradford Books/MIT Press. Tye, M. (2009) Consciousness Revisited: Materialism without Phenomenal Concepts, Oxford: Oxford University Press. Wright, C., Smith, B.Macdonald, and C., eds (1998) Knowing Our Own Minds, Oxford: Oxford University Press.

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

2 Natural Acquaintance Sam Coleman

1. Introduction Though many philosophers find it phenomenologically plausible that we enjoy acquaintance, many also doubt that acquaintance is compatible with a naturalistic approach to the mind. This leads physicalists to deny that we have acquaintance, and to dismiss its phenomenological manifestation as a cognitive illusion. Antiphysicalists, for their part, have employed the phenomenological plausibility of acquaintance in arguing against physicalism. By offering a natural model of acquaintance, I show that none of this controversy around acquaintance is warranted. In Section 2 I narrow down the kind of acquaintance that interests me, and in Section 3 note some of its key epistemic and metaphysical features. Section 4 explores what I claim is a promising metaphysical framework for naturalizing acquaintance— higher-order theories of consciousness. I argue that extant higher-order theories are unable to capture acquaintance’s key features, and diagnose this failure as largely due to their reliance on representation as a naturalistically acceptable means of supplying mental content. Finally, in Section 5 I set out my own, non-representational variant of a higher-order theory, and explain how it successfully approximates acquaintance. Since my model is capable of physical implementation, this constitutes a possibility proof that acquaintance is naturalizable. This result does not show that acquaintance in fact involves no non-physical goings-on, for that depends on how acquaintance is actually implemented. For all I say here acquaintance may be realized in non-physical materials in our world—e.g. if its actual relata are physically irreducible. At this stage, we just don’t know. But what is shown is that nothing in the debate around naturalism and physicalism hangs on whether one posits acquaintance as such, because the acquaintance relation could perfectly well be given a physical implementation. Whether acquaintance, in addition, is wholly physical waits on the truth of physicalism. But it is not, as usually thought, physicalism’s truth that depends on the status of acquaintance; acquaintance itself is quite neutral on that issue.

2. Acquaintance: Preliminaries Acquaintance as Bertrand Russell explains it concerns consciousness—what it is like to be you at a given time. There is something it is like to be you, and there is Sam Coleman, Natural Acquaintance In: Acquaintance: New Essays. Edited by: Jonathan Knowles and Thomas Raleigh, Oxford University Press (2019). © Sam Coleman. DOI: 10.1093/oso/9780198803461.003.0003

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

   something it is like to be me. Assuming panpsychism is false, there is nothing it is like to be a chair.¹ What it is like to be you is characterized, at least partially, by a set of sensory qualities, things like the colours you visually experience and the bodily sensations you feel. You are aware of a different set of these properties at this moment to those I experience, and that is part of what makes us two different subjects with two distinct conscious perspectives on the world. To a first approximation each of us has an awareness, and within the purview of each awareness different sets of sensory qualities enter at a time and over time. Consciousness can be conceptually analysed into these two components, as an awareness of qualitative content. Though there is obviously more to conscious mentality than sensory states, even when perceptual states are included along with the sensory, I will focus on standard sensory qualities to make my points. Insofar as there is something it is like to have a thought, say, that goes beyond sensory qualities, I take what I say to apply mutatis mutandis.² In the relevant Russellian sense, to say you are acquainted with x implies that you are conscious of x.³ You might infer, or be told, that you are angry at a colleague based on your irascible behaviour, but that form of awareness is quite different to being conscious of, acquainted with—actually feeling—the anger. For Russell acquaintance was the peculiarly direct form of awareness each of us bears to his or her own sense-data, posited inner objects possessing the aforementioned sensory qualities, and which make up the contents of consciousness.⁴ For present purposes I eschew sense-data, and will talk neutrally of (instances of) qualities we know through experience, what I called sensory qualities, and what philosophers call phenomenal qualities: e.g. redness, itchiness, cold.⁵ I will assume without argument that a phenomenal quality can exist, and exist intrinsically unchanged, whether experienced or unexperienced.⁶ In other words phenomenal qualities are not essentially phenomenal, though they are essentially qualitative. This is one reason why I favour the term ‘sensory qualities’. On the other side of the acquaintance relation are subjects. I will talk in those terms without trying to say just what a subject is. Minimally, a subject is a being with conscious states: states such that there is something it is like to have them. I agree with Russell that we cannot be acquainted with items outside the head. What I think we can be acquainted with are sensory states and their properties, in consciousness. I will not consider our alleged acquaintance with conscious states; that is, states that are already conscious, and where we perform, beyond whatever operation makes them conscious, a further act of acquaintance. Even if we can be ¹ This sense of something-, or what-, it-is-like-ness, commonly invoked in explication of consciousness, is often traced to Nagel (1974). ² For arguments that thoughts have a proprietary ‘cognitive’ phenomenology, see Pitt (2004), Strawson (2008), and Montague (2015). ³ Or have been—Russell believes we can be acquainted with the contents of memories. I will not explore this claim. ⁴ It is, says Russell, direct awareness of the sense-datum object (1910–11, 108). ⁵ This fits with Russell (e.g. 1912), who allows acquaintance with properties of sense-data, like a table’s brownness. ⁶ Russell purportedly offers some argument for this claim, to the effect that sensory qualities are logically independent of the subject (e.g. 1917, 112), but as independent arguments they are unconvincing. See Rosenthal (1991) for empirical arguments.

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

 



acquainted with such states, which I doubt, this presupposes a more basic kind of acquaintance, simply that involved in consciousness of a sensory state in the first place. On this understanding being acquainted with an already conscious state demands two acts of acquaintance.⁷ The acquaintance I am interested in is what goes with our being conscious of a sensory state at all.⁸, ⁹ I will not be discussing introspection either. People have wondered whether in introspection we are acquainted with experiences, which might provide a sure foundation for phenomenal judgements.¹⁰ I am not wholly certain what introspection is meant to be, though of course we can make judgements about consciousness. These are not immune to error, and if they are less error-prone than judgements about the non-mental that is probably only due to the proximity of what we judge about. So: I am interested in acquaintance as an especially direct relation we bear to a sensory state, or at least to its properties, most importantly its sensory qualities, simply in being conscious of it. That is my target. People rejected Russell’s acquaintance because its special features, and its association with the sense-data doctrine, ill fitted the prevailing physicalist-empiricist-reductionist tenor of the twentieth century. Interestingly, philosophers have been more willing to accept that we might be acquainted with items outside the head, in ‘direct perception’. I will not discuss that topic, though it is intriguing that some find it more plausible that we might bear such an intimate relation to the properties of buses and prickly cacti than to our own brain states.¹¹ Regardless of that, my aim will be to rehabilitate Russellian acquaintance, offering it a respectable home in the brain.

3. Acquaintance: Key Features Consider a visual experience of a homogenous purple field, like that produced by seeing a well-lit purple painted canvas from close enough that purple fills your vision. I assume that the homogenous purple quality-field you visually experience is really an inner property of you. As well as whatever else occurs—conceptualization, judgement, memory—I say that you are acquainted with the purple quality instance. It is ⁷ Russell calls this ‘acquaintance with the present experience’ and, more perspicuously, ‘experience of experiencing’ (1914, 443). I do not believe, as do Zahavi and Kriegel (2015), Strawson (2015), and Brentano (1995 [1874]) that every experience comes with experience of that very experience (for some criticism of this ‘self-intimation’ doctrine see Coleman (2016)). I also deny a doctrine popular amongst advocates of acquaintance, including Russell, that we are acquainted with acquaintance itself. My main reasons are phenomenological: I do not find acquaintance in awareness, and I have no idea what it would feel like if I did (in addition to finding the sensory qualities with which I am acquainted). It is rather the way I am aware of sensory qualities, how they seem to be present to me, that makes me a believer in acquaintance. It might sound strange to say that I infer I have acquaintance, but I am happy to say it. ⁸ With Howell (2008, 132), I think what is of interest about consciousness arises well before we turn to take any mental or introspective stance on conscious states. ⁹ It is still true in a sense that we are acquainted with conscious states, but only because acquaintance goes with their being conscious: these are not states that are already conscious. Similarly, for Russell, we are acquainted with what is presented to us, but only because presentation is the flip-side of acquaintance: it is not that something is first presented and then becomes the object of acquaintance. For me consciousness is the relevant flip-side. ¹⁰ See e.g. Gertler (2012). ¹¹ For doubts about the direct perception relation see Coates (2007).

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

   there, for you, in a peculiarly direct way. Metaphysically it is not in virtue of being aware of something else that you are aware of the purple—so you are aware of it in that sense directly.¹² This metaphysical intimacy grounds direct cognitive access to the purple quality: what you access is simply the purple as it is, unembellished and unreported. When we talk of ‘acquaintance’ as a mental item it has this dual metaphysical and epistemic status: acquaintance is a something, and it also gives something to us, connects the mind to something else—which sorts of connection, whether direct or indirect, we call knowledge. Following Russell, I do think that acquaintance with the purple counts as a basic sort of knowledge. One reason for saying this is that this episode of acquaintance cuts down epistemic possibilities for you, i.e. ways the world could have been for all you were aware. Prior to the experience—literally a priori—it could have been green or blue or black present to you—there is an epistemic scenario corresponding to each of these colours.¹³ But in fact it is purple, and in the core sense of being aware of it, you know it to be purple and not any other quality. Knowledge has the essential function of narrowing down possibilities for the subject, and sheer acquaintance awareness of purple, as opposed to any other colour, fulfils this role. Note, for comparison, that my being acquainted with purple narrows nothing down for you. Nor does something’s simply being purple, outside of anyone’s awareness. The first of these comparison cases involves knowledge, because my acquaintance narrows things down for me. The second case, as far as described, lacks knowledge altogether. Of course we can still have propositional knowledge about things of which we, and even all other people, are currently unaware. The present point is that with the coming and going of acquaintance comes and goes a certain other kind of knowledge.¹⁴ Acquaintance is somewhat more than the brute specific being of some thing, but somewhat less than the explicit framing of that thing in propositional terms by the subject. It is a cognitive impact on the subject of an in-between kind. As such, acquaintance, though not itself propositional knowledge, is clearly an enabler of, a way into, propositional knowledge: knowledge of truths. Direct awareness of an existent puts one in a position, at least, to know truths about that thing.¹⁵ This is not to say that you do not have any propositional knowledge along with acquaintance, and perhaps you always do.¹⁶ But whatever you get, the most basic thing you get is knowledge of the purple in the sense of being acquainted with it, grasping it, mentally meeting it, which is not knowing that such-and-such is true of it. Nor is acquaintance conceptual—one can surely be aware of a colour without having the concepts COLOUR, SHADE, PURPLE, etc. I need not classify it.¹⁷ It is not ¹² Nor, as Russell (1912, ch. 5) stresses, is acquaintance mediated by inference or knowledge of truths. ¹³ For the notion of epistemic scenarios, and their relation to possible worlds, see Chalmers (2011). ¹⁴ Contrast with Balog (2012), for instance, who seems to think the basic form of knowledge of our sensory states through experience is conceptual (albeit non-inferential and ‘direct’). ¹⁵ Tye (2009, 102). Cf. Wishon (2017) on Russell on this point. ¹⁶ Russell (1912, 46): ‘it would be rash to assume that human beings ever, in fact, have acquaintance with things without at the same time knowing some truth about them’. ¹⁷ The content of such a state is not one regarding which the subject need possess relevant concepts (Tye, 1995, 139). Feigl (1958) believes all grasp of colour is classificatory, hence conceptual, requiring

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

 



presented, let alone represented, under a guise. It is just there. Even a demonstrative conception requires me already to be acquainted with the colour, in order to have something to demonstrate.¹⁸ This is not to deny that we routinely classify, or that such classifications can affect the overall character of experience. Perhaps for humans conceptualization of experiences is unavoidable. But I suspect that experience is in its core non-conceptual, and I maintain that the acquaintance with sensory qualities it involves counts as a form of knowing.¹⁹ I will say more about the place of concepts shortly. It is tempting to express the fact that nothing mediates your awareness of the purple by saying that there is ‘nothing between’ your awareness and the purple. But it is not that there is a gap between them, either. Rather, at least this is somewhat how it seems, the purple is jammed right up ‘against’, even somehow ‘into’, your awareness.²⁰ We can say, in a sense to be elucidated, that the quality instance you are aware of and your awareness of it are not wholly ontologically distinct items.²¹ Prima facie they are distinct in some way: on the assumption that your present awareness can take in other qualities in place of this one, your present awareness and the quality presently experienced can come apart.²² The alternative is that each awareness of a distinct quality is itself numerically different; in that case the present awareness and its qualitative object are inseparable. But if present awareness cannot survive its present qualitative object, it would seem that there could not easily be an awareness of a change in experienced qualities. I will therefore suppose a good case for distinctness—the claim that this very awareness can survive this experience of this quality.²³ Given that supposition, we seek a relationship between quality and awareness on which they are not identical, nor yet wholly separate—whatever that quite means! Of a colour we are acquainted with, Russell says: ‘so far as concerns knowledge of the colour itself, as opposed to knowledge of truths about it, I know the colour perfectly and completely when I see it, and no further knowledge of it itself is even theoretically possible’ (Russell, 1912). If he learns truths about the colour these, Russell reaffirms, ‘do not make me know the colour itself any better than I did before’ (46–7). Russell has been read as asserting here the thesis of revelation: that

explicit grasp of the similarity and difference relations in which a given colour stands. The ontological point subserving this claim seems just: that the being of a colour in fact consists in, or does not outrun, its relations to other colours. But that this ontological complexity must be explicitly registered by the subject in simple consciousness of the colour, that she must possess correspondingly complex mental terms, seems a further claim that can be resisted, e.g. on phenomenological grounds. Similarly, Russell (1914, 145) rules that classificatory knowledge concerning colours is propositional, hence not given in acquaintance: we can know (acquaintance-wise) different colours without knowing (propositionally) that they are different. ¹⁸ Tye (2009, 136) makes this point. ¹⁹ For a view of perception as non-conceptual knowledge, see Hoffman (2014). ²⁰ There are those who think sensory qualities are mere modes of awareness. I set aside this view, which I think is false—or worse, unhelpful. That construal of awareness prevents acquaintance being a cognitive achievement, in the sense explained below. ²¹ Cf. Kriegel (2009, 109). ²² I also assume that a sensory quality can exist unconsciously, so can come apart from awareness in that way. ²³ Experiences seem plausibly not to be independent in this way, but tied to their qualities essentially.

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

   one knows a quality through and through in acquaintance, in a way that cannot be improved upon, such that it has no hidden aspects at all. I want to adopt this thesis in qualified form. It does seem possible in principle to be acquainted with a wholly simple quality instance, taking it in its entirety. But it is unlikely that such ‘full revelation’ is guaranteed by acquaintance; one can be acquainted with x without x being fully revealed, in other words. There are two reasons to qualify revelation: a quality may have breadth, and it may have depth, that elude full grasp. These features flow from the ways in which the qualities we typically experience are complex. Breadth first: the fact is, a quality can be experienced without all of it being experienced, because it has extension—multiple qualitative parts.²⁴ Naturally, that portion of the complex quality outside of acquaintance is not conscious, so the upshot is that a multi-part quality can be partly in consciousness. Consider a feeling familiar to academics: a pervasive background stress, perhaps focused on a paper due for a short deadline against a large marking load. You are aware of this feeling during the day, going about life. But you do not pull the whole thing into view; that would be disruptive. I pull such feelings fully into view, to consider and dispose of them, in bedtime meditation. This is something like the difference between glancing at a book’s spine and opening it to have a good browse. During the day you are often aware of the ‘edge’ of the feeling, enough to mark it as a background state of tension, not enough perhaps to identify its intentional object. When perusing it at leisure, all its extent and richness can come into view, as well as its target.²⁵ So I say that during the day you are acquainted with the stress quality, and it is the same quality you later examine, but most of its extent is not yet in acquaintance. It seems to be excluded, or occluded, by other objects of daytime awareness. Thus acquaintance clearly has a ‘bandwidth’, or scope. It has a limited field of view. Russell would be wrong to claim what we are acquainted with is always all there is to a quality breadth-wise, if that’s what he means by complete knowledge.²⁶ Someone might not want to count the evening quality as one with the day quality—after all they feel different, in an intuitive sense. The objector may want, in other words, to restrict ‘the quality’ (e.g. Russell’s ‘colour’) to what is in awareness. But there are good reasons to defend the identity claim, that we have one relevant quality complex in play throughout the day and night episodes. For one thing, the two qualities share effects: both cause you to be forgetful and snappy, and your heart to race—and these effects are explicable with reference to this one unified quality complex, whose parts need not all be conscious to be efficacious. Further, the day feeling does not go anywhere—it appears to be part of the bigger whole (like the

²⁴ Broad says the same thing about sense-data: ‘a sense-datum with which I am acquainted may perfectly well have parts with which I am not acquainted. If therefore I say that a given sense-datum has no parts except those which I have noticed and mentioned I may quite well be wrong’ (1919, 218). ²⁵ This phenomenal contrast brings to mind the difference between having a word on the tip of your tongue and finally fully retrieving it—perhaps the mental mechanics are similar. See James’s justly celebrated description of the phenomenology (1890, 251). ²⁶ But Russell may well have this notion of qualitative breadth in mind when he asserts (1910–11, 109) that we can be acquainted with a ‘complex’ without being acquainted with all its constituents.

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

 



book’s spine) and is still there when you inspect the rest of the feeling.²⁷ And, more than the book’s spine, the day quality is integrated with the further aspects you access in meditation: they mesh as a qualitative whole (hence this sort of volume can be judged by its cover). During the day you are not conscious of the whole extent of the stress quality, and that goes with not being acquainted with its whole extent. Other feelings are competing for the limited window of acquaintance. The second qualification to revelation concerns a quality’s depth. Homogenous visual purple has red and blue as constituents. Once someone tells you there is red and blue in purple you can tell by inspection that it is somewhat red and somewhat blue—you can see them ‘in there’. Red and blue are not present in purple in the way either is present when on its own, though, or when next to each other, as in Barcelona’s home kit. No matter how hard you focus on a purple patch you will not see literal redness or blueness, as when looking at a fire engine or the evening sky. And it is perfectly possible to experience purple without knowing that it is made of red and blue—it is even possible to take yourself to be experiencing a simple quality, before anyone gives you a clue of its composition.²⁸ So you can miss that purple is complex. Nonetheless red and blue are in there, in the purple you experience; they are present. If a playful neuroscientist subtracted the redness from the purple, the patch you experience would not be purple anymore, only blue. Are you acquainted with the red and blue instances that are in the purple? I claim so. They inform your experience, what it’s like for you. If they change, your experience changes. So you have mental contact with them, and we can say that you are acquainted with them, these constituents of purple, in being acquainted with the purple. Clearly, though, you are not acquainted with a sheer redness, what you would experience seeing a red fire engine. That is because the red present to you is in an ongoing qualitative reaction, or mixed state, with the blue—they are interpenetrated and mutually informing, qualitatively. Time for a distinction: I will say that a heterogeneous quality, one with various qualitative aspects before the mind that are in themselves each relatively homogenous, has, in this respect, horizontal (qualitative) parts. A homogenous quality that nonetheless has qualitative complexity, being the product of further qualities blended together, I will say has, in this respect, vertical parts.²⁹ A simple quality before the mind has therefore neither horizontal nor vertical parts. A complex quality with horizontal parts may also have vertical parts—for instance if one of its horizontal parts, uniform if taken by itself, has vertical parts. To illustrate, a complex quality with horizontal parts would be present in an experience of Barcelona’s red and blue striped home strip. The stress quality described above is also of this sort. And we will say that homogenous purple, though lacking horizontal parts, has red and blue as vertical constituents.

²⁷ Unlike the spine it does not become hidden as you (mentally) revolve the whole it belongs to. It feels more as if it ‘diffuses’ over, or pervades, that whole. ²⁸ Hume famously claimed that all experienced colours are simple qualities. ²⁹ In the case of vertical as much as in that of horizontal parts, I will also use the terms ‘components’ and ‘constituents’ more or less interchangeably with ‘parts’.

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

   Returning to the purple instance, there seem to be degrees of acquaintance, even for items fully within the field of acquaintance. You are acquainted with the purple in a way such that you are not acquainted with the red or blue vertically composing it in that way, though you are acquainted with them. You are acquainted with the purple directly, but you are acquainted with the red and blue only in virtue of their composing the purple. So you are not fully acquainted with them. For full acquaintance it seems you would have to experience each homogenous quality disentangled. So we may say that you are only ever fully acquainted with a quality that is not, in your experience, vertically composing some further quality.³⁰ Such a quality I will call the ‘top quality’, the one ‘closest’ to your awareness. Hence it is false that you know red perfectly and completely by being acquainted with it in purple: acquaintance does not guarantee perfect and complete grasp. Only full acquaintance gives you that. Someone—probably the same someone as previously—might ask why we do not reserve the term ‘acquaintance’ for our relationship to the top quality. But the mind does meet the red in meeting the purple, since the purple is nothing but the red and the blue mixed, so this does still seem to be acquaintance.³¹ Whatever sub-qualities vertically compose some homogenous quality, they all make a difference to its character. Otherwise they either are not genuine qualities or do not compose it. The more qualities are vertical constituents, the smaller the difference each may make. And these qualities may have their own qualitative vertical constituents; and so on. Still, in being acquainted with the top quality you are acquainted, though to a lesser degree, with all its vertical constituents. They are such that your experience would change were they removed.³² And your mind is meeting with them, in meeting the quality they compose and their contributions to it. It seems quite possible that what we think of as basic ingredients of colours—hue, saturation, brightness—have themselves far more qualitative complexity than we realize. We are acquainted with these ‘micro-qualities’, but have not explicitly recognized them yet.³³ Concepts would help us to get a fix on them. If you inform someone that purple is a red and blue compound, they should be capable of seeing the blue and the red ‘in there’. Nothing is added to the purple experienced in such cases. Having previously seen isolated red and blue, one might now be led to imaginatively project them into

³⁰ Relatedly, here is a way of distinguishing vertical from horizontal constituents: when you are aware of horizontal constituents, it need not be in virtue of their composing some homogenous quality. But you are always aware of vertical constituents in virtue of their composing some homogenous quality. (Note that you can be aware of horizontal constituents in virtue of their composing some homogenous quality, when these are vertical constituents of homogenous horizontal constituents of a complex heterogeneous quality.) ³¹ Recalling that I said acquaintance does not present its object under aspects or conceptualize it, our someone might now complain when I say one is not fully acquainted with the whole qualitative depth of a vertically complex quality. Is this not getting the whole quality under an aspect or mode of presentation? No—otherwise we must say that cameras that present things at different levels of magnification have or create ‘modes of presentation’. Or, if they do, then modes of presentation are after all flimsy enough for me to be able to allow that acquaintance does involve them, in a sense that doesn’t implicate concepts. ³² Whether you would judge that it changes is irrelevant, except as defeasible evidence. ³³ See Schroer (2010) for speculations about what I am calling ‘micro-qualities’ (or ‘sub-qualities’): he suggests ‘strength’, ‘warmth’, and ‘coolness’ may be among their number, for the colours at least. Hartshorne (1934) argues that such ingredients are multi-modal.

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

 



the purple—that is something to watch for. But restricting experience to the purple itself, nothing changes. Why would it? After all the red and the blue mixed just make it the purple it is, so why would acknowledging them change it? This recognition (we cognize them already) involves wielding the concepts of blue and red as ingredients. Imaginative projection requires you to have had full acquaintance with red and blue, and forming completely adequate concepts of them likely also requires this. If we had experienced red exclusively as mixed—in purple, orange, etc.—we would have only a partial sense of its identity as an independent quality. It does seem possible in experience to discern qualitative elements we have never witnessed in isolation, even to form concepts of them, albeit not fully adequate ones. For instance, we seem to know what visual warmth is—red has it, orange, and brown, but not grey, silver, blue or black—yet we never catch visual warmth on its own. The same is true of brightness.³⁴ Examples like these tend to confirm that there are micro-qualities within acquaintance,³⁵ and that even if they run to some depth they need not elude the classificatory reach of conceptual cognition.³⁶ Conceptualizing the vertical constituents of a complex quality does not alter the quality experienced. But concepts can affect our experience of horizontally complex qualities. Rosenthal observes that in ‘wine tasting and musical experience . . . conceptual sophistication seems actually to generate experiences with more finely differentiated sensory qualities’ (1991, 34). The key question is whether concepts cause these experiences to change, or whether their application constitutes the change. I suggest their contribution is causal. Imagine that, though a relative novice, you manage to taste a wine properly—slowly, with the correct nasal action, and so on. After your first crude taste, someone says: ‘Look out for the acidity’. Isn’t it right that this prompts you to scan for an element that was present, but which was not the object of distinct focus? According to Smith: ‘Experienced tasters will learn more from their sensations about the . . . wine . . . by paying attention to particular aspects of their sensory experience . . . [such a taster] guides his attention towards certain aspects of his experience, selecting some for peculiar scrutiny’ (2007, 49). This description strongly suggests that the relevant qualities are already there for attention to be paid to them. Are you acquainted with the acidity before you are encouraged to look for it? Well, does it shape your experience? Is it an element of the experience? The answer to that seems to be ‘yes’—otherwise it is puzzling how you could select it for scrutiny. I propose that the application of concepts (like ACIDITY) causes the guiding Smith describes rather than constituting any change in the flavour ³⁴ Though Plato has this as one among the colours in Timaeus. ³⁵ Tye (2009) says we are acquainted with something only if we are put in a position to ask ‘What is that?’, and this would apparently not be true of micro-qualities—we do not normally notice them, in the relevant sense. Still, conscious animals must be acquainted with qualities, and they are in no position to ask such questions (this is not to deny that they have some basic concepts, in this I agree with Tye). Or are we in a position to ask the question, in fact? What is there to ‘being in position’ here? We are in position to ask about a micro-quality, since we are acquainted with it: it is subjectively present for the question to be asked about it. We just do not ask because we do not focus on it. Alternatively, perhaps by restricting what he says to full acquaintance Tye’s thought can be swallowed. ³⁶ Could our macro-qualities of experience, even our entire personal fields of qualities, be vertically composing something far bigger? Why not? A universe fabric: we are each ‘zoomed in’ to a patch, in our particular awareness. Pantheism lies yonder! See Coleman (2019).

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

   experienced. This seems intelligible as the concept prompting your awareness to narrow its focus, or scan along the complex quality—the horizontally extensive wine flavour—zooming in to pick out qualitative contours only peripheral until then. Such magnification can perhaps be understood as making a given horizontal qualitative part fill up proportionally more of the limited field of acquaintance, in the process excluding (or making peripheral) other qualities: literally raising its volume.³⁷ That is why experience changes. After conceptual prompting the quality in question takes the centre stage of awareness, and becomes more experientially prominent.³⁸ This is often the case with a quality that is myriad, with many horizontal constituents: the sensory equivalent of a multi-coloured ball of knitting, where taking in the whole is at the expense of resolution in specific parts. Still, though your experience is altered by the application of concepts, the wine flavour, that extensive horizontal quality, plausibly remains as it was initially. Experience is a function of the manner in which we access a given quality—e.g. which parts we focus on. But the horizontally complex qualities focused on need not be altered by such processes. Once we distinguish in this way an alteration of experience from an alteration in sensory qualities, we can say that even if conceptual sophistication seems to generate more finely differentiated sensory qualities, all that really becomes more finely differentiated is our awareness of existing sensory qualities. This fits better with Smith’s phenomenological description.³⁹ Overall, I restrict Russell’s cases of complete and perfect knowledge—revelation— to qualities with which full acquaintance is possible. That means a top quality, one not vertically constituting any further quality in one’s experience, and where this quality is relatively homogenous: it does not have horizontal qualitative parts, or not many, within consciousness or without. If they are out of consciousness then our knowledge of the whole is incomplete, as with the stress feeling. If they are in consciousness, as with the wine flavour with its tannin, acidity, and so on, then it seems there can be a ‘crowding effect’, where each individual aspect enjoys insufficient volume at a time. Strictly all these horizontal parts are within acquaintance. But the epistemic aspect of acquaintance benefits from a more monolithic focus. Of the homogenous purple field it does seem right to say that, at the level we peruse it, our knowledge of the purple is perfect and complete. Learning about its vertical qualitative texture—red and blue—does not increase knowledge of the purple qua purple.⁴⁰ For it is what it is, that very mixed state of red and blue. Revelation is simply the name we use when contemplating the epistemic aspect of full acquaintance.

³⁷ Listening to classical music one might focus now on the violins, and hear the rest more peripherally; of course one heard the violins already, only not as distinctly. ³⁸ You also become aware now that it is present, via the concept. The taster will likely have had previous experiences of acidity to focus on, and thereby form the corresponding concept, so as now to be capable of scanning for the quality. Short of that, acquaintance with a similar enough quality will likely suffice. ³⁹ That the description is phenomenologically apt may also make one doubt that conceptual sophistication even seems to generate further qualities. ⁴⁰ If you think learning about its vertical qualitative constituents counts as completing our knowledge of purple, then full acquaintance is possible only for vertically simple qualities. I do not know if we experience any such unblended qualities, but it seems possible that we do. Yet learning about is learning truths, and the topic here is (complete) non-propositional knowledge.

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

 



Is acquaintance then infallible? I have suggested that it has a limited field of view or capacity, so that detail in a complicated sensation may not all be brought into clear focus. But it doesn’t of itself get things wrong. Nothing acquaintance says about the purple, the wine quality, or the stress feeling for that matter, conflicts with what you find later. You only access in greater resolution what was there already, or discover parts that were not in awareness. You can easily err by judging that this is all there is to a quality, for example when some of it outruns awareness horizontally: but this is not something acquaintance declares, so not its error. You can certainly misperceive the blue sky as green; but you cannot be acquainted with a blue quality that is really, in its true nature, green.⁴¹, ⁴² Lastly, acquaintance is a cognitive achievement: it involves making substantive contact with something. This feature can be seen to depend upon the aforementioned fact that awareness and its target are non-identical. In acquaintance awareness achieves a metaphysical and epistemic connection with another existent—acquaintance thus in some degree enlarges the subject’s world. Had it not enlarged the subject’s world in this way, acquaintance’s claim to be a kind of knowledge would be jeopardized.⁴³ These are the features I have picked out as characterizing acquaintance: it is an especially direct relation the subject bears to sensorily qualitative states, so that one’s awareness seems to be not wholly distinct from its object. With certain notable qualifications, it provides complete and perfect, infallible, non-conceptual, and nonpropositional knowledge of its objects. Nonetheless acquaintance amounts to a substantive cognitive achievement. Given this special list of properties, it is perhaps not surprising that theorists have been dubious about acquaintance’s naturalistic credentials.

4. Reductive Models of Acquaintance In this section and the next I consider some reductive physicalist accounts of consciousness, seeking one with formal features suited to modelling acquaintance. I examine higher-order theories of consciousness, a field inspired by David Rosenthal’s

⁴¹ Kriegel (2009, 108): ‘It cannot be that, upon looking at the sky, my visual experience has a bluish qualitative character, but I am aware of it having a[n entirely greenish] qualitative character.’ Later I reject his explanation of this fact. ⁴² What are the implications for that familiar physicalist move—of saying that qualities of experience have a recognizably physical, but ‘hidden’, back side, something non-qualitative like being a c-fibre firing? My account is strictly compatible with such back sides—acquaintance need not make us aware that there’s a physical aspect to a quality, even if on the theory we must be aware of this aspect if it exists. Evidently we are not fully acquainted with any such aspect, and it is not revealed alongside the top quality (cf. Goff 2015). I would place a requirement on any constituting vertical constituent, that its relation to the whole— its contribution to the complex quality—be intelligible, in the way the contributions of red and blue to purple, or brightness and warmth to orange, are. If (e.g.) neural properties cannot meet this demand then they are not in the constitution of sensory qualities. As noted, however, the present chapter concerns the possibility of acquaintance being physical, not its relata. For my view of qualities see Coleman (2015a). ⁴³ Similarly, for Russell ‘Acquaintance . . . essentially consists in a relation between the mind and something other than the mind; it is this that constitutes the mind’s power of knowing things’ (1912: 22, my emphasis).

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

   work.⁴⁴ If acquaintance can be modelled by such a theory, that would show it to be capable of physical implementation. What makes higher-order theories a good place to search out a model, notwithstanding the common worry that acquaintance must be non-natural, is that they (typically) build in a two-part structure, and make a conscious mental state the object of awareness. These features already suggest echoes of acquaintance. Although higher-order theories are widely acknowledged as the most developed reductive accounts of consciousness, many philosophers doubt their power genuinely to explain subjective experience, i.e. to solve the ‘hard problem’.⁴⁵ My task is not to solve the hard problem, however, but to see whether acquaintance’s key features can be physically modelled. So I won’t defend higher-order theories on this score.⁴⁶ To the extent one finds my model of acquaintance in Section 5 appealing, though, and given the close connection between acquaintance and consciousness, one may find one’s credence in higher-order-style treatments of consciousness correspondingly boosted.⁴⁷ Consider first Rosenthal’s own higher-order thought theory. Rosenthal believes sensory qualities can exist unexperienced. Perhaps when a blindsighter perceives a purple painting she instantiates a purple quality in her visual cortex, and its presence explains her amazing capacity for colour judgement. Still, she isn’t conscious of that purple; there’s nothing it is like for her. Rosenthal claims that what makes a sensory state conscious is having a suitable thought about it. This thought counts as higher order because it is a mental state that represents another mental state. So a subject experiences her sensory state of purple just in case she tokens a higher-order thought (HOT) that represents her to be in such a purple sensory state. Note that for Rosenthal consciousness standardly involves two separate mental states: the sensory state, and the HOT, and each can exist without the other. Without a HOT to represent it, a sensory state is unconscious, like a blindsight visual state. But what if a HOT represents there to be a purple sensory state when there is no such state?⁴⁸ Rosenthal says that the subject experiences indistinguishably from the case where the purple state exists and is HOT-represented. This reveals that it is HOTs that call the shots for the stream of consciousness: the subject is aware of all and only what her HOTs represent. It is the manner of HOT representation that truly shapes the subject’s experience. That is confirmed by ‘illusion’ cases, where the HOT merely ⁴⁴ See his 2005 volume. As has been noted, the idea of awareness through meta-cognition has antecedents in Aristotle’s De Anima and in Locke (see e.g. Caston, 2002). See also Feigl’s (1971, 305) suggestion that one brain area ‘scans’ another. ⁴⁵ For instance, some object that higher-order theories analyse consciousness as the conjunction of two non-conscious things, and such a conjunction cannot add up to consciousness. But this objection is unfair: any attempt to explain consciousness as such must be in terms of the non-conscious, or it is no explanation. Thus any explanation of consciousness will involve a conjunction of non-conscious elements. So all the objector can be saying is that consciousness is brute. Perhaps it is, but purported explanations of it cannot be faulted simply on the ground that they involve components that lack consciousness, i.e. that they attempt to explain it at all! ⁴⁶ For relevant defence of my own higher-order theory of consciousness, see Coleman (2016). ⁴⁷ WARNING: Credence levels can fall as well as rise. ⁴⁸ Neander (1998) initiated a tradition of objection to Rosenthal via such cases, the existence of which he had long acknowledged.

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

 



misrepresents an existent sensory state. Say that it represents the purple state as yellow: here the subject experiences yellowly, Rosenthal rules. Can this model capture acquaintance? The main problem is that on HOT theory we turn out to be somewhat cut off from sensory reality. One is never in fact genuinely aware of one’s sensory state, even where it exists, but, instead, only of what the relevant HOT represents to be the case. This is what follows from Rosenthal’s treatment of the two sorts of mismatch case above. One certainly can be in the happy situation where the representation is accurate: the sensory state is as the HOT portrays it to be. But that seems analogous to the situation of seeing a projector image of a purple painting, which happens to be projected onto just such a painting. Although one is well informed about what is there, still in an intuitive sense one only sees the projection, and not what it represents, which is actually obscured by the very representation of it.⁴⁹ Rosenthal’s move to put HOTs in sole charge of experiential contents has the result that consciousness on his theory is not after all a relation to a sensory state.⁵⁰ So HOT theory does not give us the direct cognitive access to sensory states that marks acquaintance.⁵¹ A second problem is that Rosenthal explicitly affirms that awareness is wholly conceptual: it is concepts in HOTs that dictate experiential contents, and these contents are descriptive in form. This is the most basic form of consciousness, he maintains.⁵² So there is no possibility on HOT theory of that kind of sheer, nonconceptual awareness of qualities characteristic of acquaintance. We framed acquaintance as something underlying the application of concepts to experience, but for Rosenthal there is nothing more basic than this conceptualization. Hence, overall, Rosenthal’s theory is unsuitable for capturing acquaintance. This is hardly surprising: one of Rosenthal’s main preoccupations is the possibility of divergence between mental appearance and reality, and he is opposed to anything in the spirit of acquaintance. If the main problem with Rosenthal’s theory is the gap between awareness and its would-be target—between HOT and sensory state—a natural move is to bind them more tightly together. The theory that binds them maximally tightly is simple self-representationalism, where sensory state and higher-order representation are identical. This entails one state with two representational contents. First, it represents a purple quality, perhaps ascribing it to an external region of space. Second, it ⁴⁹ That’s so even if the projector gets a feed from the painting, via a camera, which causes the projection; attach any further conditions you feel are necessary for representation, the problem concerns the nature of representation itself. Similar reasoning prompts Block’s (2011) complaint concerning the ‘dual layer’ of sensory contents on HOT theory. See also Rosenthal’s statements that sensory states are only present in consciousness in the way HOTs represent them, which is to say they aren’t directly present at all. For more on this line of argument see Coleman (2015b, 2018). ⁵⁰ Cf. Brown (2012). This is as much as to say that a sensory state is never conscious on HOT theory. ⁵¹ This worry applies equally to Lycan’s (1996) higher-order perception (HOP) theory: it is the higherorder state that governs how experience manifests on this theory, and higher-order-style illusions/ hallucinations remain possible, so we are every bit as cut off from mental reality as on Rosenthal’s theory. Equally unsurprisingly, Lycan would also have no time for acquaintance. In the end I am not certain whether my positive theory is best classed as HOT or HOP—see Section 5, and Coleman (2015b). ⁵² Hence if animals and babies are conscious they must have some relevant concepts. For defence of this claim see Gennaro (2011).

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

   represents itself: its content being something like ‘I am a purple-representing visual state’. On this analysis, when a sensory state self-represents it becomes (state) conscious, is felt. Here we can say that there is no ontological gap between awareness and what one is aware of, since the awareness-supplying component and its sensory target are one. So simple self-representationalism might appear a promising way to capture acquaintance. This theory faces two difficulties, however. Our project is to devise a model of acquaintance that can be physically implemented, thereby naturalizing it. And because simple self-representationalism employs representation as the awareness relation it seems naturalistically respectable. But not every variety of representation is guaranteed to fit with naturalism. Representation is analysed by reductionist philosophers as a tracking relation: of a mental state to external targets in perception, or of one mental state to another for those who think representation can implement consciousness. But tracking mechanisms have a core causal component—the tracked causes the tracker (or did in a learning phase, etc.). Clearly, a single mental state cannot cause itself. Hence it cannot track itself. Nor, therefore, can it self-represent, in a naturalistically acceptable sense. So simple self-representationalism is not our desired model of acquaintance.⁵³ The second problem points towards a better theory. We earlier suggested that awareness, while not wholly separate from its target (as on Rosenthal’s theory), is nonetheless distinct from this target in some way. It does not seem phenomenologically as if every sensory state is identical to its own awareness. And it would be odd to hold that a state of awareness could not survive a change in sensory qualities. But this seems to be a consequence of simple self-representationalism, since the qualitative properties pertaining to a single self-representing state surely enter into its identity conditions. Given that this state is also a token state of awareness, it follows that a state of awareness cannot survive a change in the qualities in awareness.⁵⁴ Here, then, is another reason to look beyond simple self-representationalism: it binds awareness and sensory states a little too tightly. Kriegel defends a sophisticated self-representationalism. Suppressing some detail,⁵⁵ there is again a single state involved in consciousness, but now it has two parts. One is (say) our purple sensory state. The other is a higher-order component, representing the sensory state. But since sensory state and higher-order representation are bound in a further, conjoint, state, in representing the purple state the higher-order state represents a part of a mental state of which it is also part. According to Kriegel, this bi-partite mental state thereby self-represents—one part represents the whole, by representing another part of that whole. And since selfrepresentation is Kriegel’s essential recipe for consciousness, we have a conscious ⁵³ Kriegel, formerly an advocate, makes this criticism of simple self-representationalism in his (2009), and I repeat it in my (2015b). Perhaps this theory remains open to a non-naturalist modeller of acquaintance, e.g. if they can make out a non-tracking-based form of self-representation—notwithstanding the following objection. ⁵⁴ The picture on simple self-representationalism seems to be of a rapid, but disjointed, cascade of selfrepresenting states of awareness, each with its own sensory quality, to model the evolving stream of consciousness. This picture, though pretty, is phenomenologically inapt. ⁵⁵ For full details of Kriegel’s theory, as well as criticism on other counts, see Coleman (2015b).

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

 



state. Because there are now parts to the vehicle for a conscious state, these can be causally related, so no state need represent itself directly. The higher-order state’s representation of the whole is indirect, for Kriegel, hence non-causal: it is akin to the way a painting represents an entire house by directly depicting just the front. Generally, Kriegel argues, one can non-causally, indirectly, represent a whole by directly, causally, representing a significant part of that whole. So Kriegel’s theory avoids the naturalism-related worry confronting simple self-representationalism. At first sight this single, but bi-partite, state structure offers an appealing way of capturing the idea that in acquaintance one’s awareness is not identical to, but neither wholly distinct from, the target of awareness. Here the awareness-supplying component and its target are not wholly separate, since they compose a single conjoint state. But neither are they identical: they are two states that can come apart. This appears to be a pleasing in-between option. So is this the way to model acquaintance? Unfortunately it is not. With dual components the possibility recurs, familiar from Rosenthal’s account, of the awareness component misrepresenting the sensory state. What if the higher-order component represents the purple state as yellow? Kriegel explicitly understands higher-order representation as constitutive of experienced qualities: what the higher-order state says, goes, subjectively. Hence the subject experiences yellowly in this case. Kriegel seeks to make a virtue of this feature, specifically as regards approximating acquaintance: It may be possible . . . to capture both (a) the impossibility of ‘getting wrong’ qualitative properties in the right kind of inner awareness, and (b) the lack of ‘whole distinction’ between that awareness and what one is thereby aware of, by construing inner awareness in terms of constituting representation. The idea is that [experienced] qualitative properties are constituted by the inner awareness representation of the conscious state. (2009, 109)

But now not only is there no whole distinction between awareness and the quality one is aware of, there is simply no distinction at all—the component supplying awareness and the component supplying qualitative content are one. So Kriegel’s account has not ultimately delivered on its promise to construe awareness and its target as neither identical nor wholly distinct, in line with acquaintance. Another upshot of Kriegel’s constitutive representation is that the sensory state is screened off from awareness, as on Rosenthal’s theory.⁵⁶ Equally seriously, Kriegel’s embrace of constitutive representation prevents acquaintance from being a cognitive achievement. Kriegel’s model makes consciousness infallible about experienced qualities ‘not because of any cognitive achievement involved in the relevant awareness, but simply because [qualities] are constituted by the contents of the awareness’ (2009, 110). It is not that one’s awareness is guaranteed to be of purple because a purple sensory state directly and intimately confronts one in consciousness. Rather, it is the awareness itself that guarantees, by manufacturing, an experience of purple. But acquaintance carries definite connotations of an intimate and infallible encounter with something else. As noted, acquaintance enlarges the subject’s world. The

⁵⁶ Kriegel candidly concedes that on his theory the sensory state’s qualitative content is ‘not phenomenologically manifest in any way’ (2009, 110).

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

   relevant intimacy is no good as supplied by the internal lightshow of a mental projector shining back on itself—that makes the subject’s awareness meet no reality but its own. If we want acquaintance, we will therefore want something other than Kriegel’s theory. Kriegel is eventually content with the fact that his model differs from genuine acquaintance, which he finds a ‘somewhat mysterious notion’. He is far more comfortable resting with constitutive self-representation. His worry must be that acquaintance cannot be naturalized. But, though we have not yet been able to model acquaintance in a way that can be physically implemented, which might indeed support the view that it cannot be naturalized, we should not give up. I will next show how natural acquaintance is possible, by moving away from employing representation as the main cog in a higher-order theory.

5. Natural Acquaintance Favoured accounts of content transmission in the brain, e.g. of how visual information gets from the retina to V1 and beyond, are causal-cum-representational. Earlier processing is understood to impact causally on later processes; single neurons downstream have receptive fields of multiple upstream neurons, and in their firings carry—are said to represent—the contents pertaining to these earlier cells. Once the later neuron is set up to fire, earlier neurons and their activity are in theory dispensable: their key role is to cause later neurons to discharge, all the way up to those supposed to be directly involved in the production of a visual image.⁵⁷ We thus tend to think that a token visual state caused in the normal way—via retinal activity—could also have been produced, had we only the technical knowhow, by a direct cortical stimulation: bypassing earlier neurons while replicating their exact effects on later ones. Everything that matters content-wise is judged to be present at the last stages of visual processing, since it is held there in representations. Likewise, it is the electromagnetic goings-on most proximal to your television screen that really determine the image displayed, even though the typical causal chain extends back to the studio and cameras.⁵⁸ But the inadequacy of the higher-order theories just considered has already served us a warning that representation may not be the best way to render the mechanics of experience, given that consciousness involves acquaintance. There is another way to conceive of the contribution of early neurons and their contents to a visual state. On this model the content-transmitting relation, what ⁵⁷ At least some earlier neurons will be involved in the feedback loops of re-entrant processing, and this may well be part of the implementation of conscious acquaintance. However, the point stands if we restrict it to the activity of an earlier neuron at a time t: the neuron’s earlier activity is considered dispensable once effects downstream are primed, even if some of those effects involve that cell, and its subsequent activity, in re-entrant processing. ⁵⁸ Thinking of the visual state as perceptual, we may want to individuate it by its ultimate cause, the object seen. But qua sensory state, the conception relevant for now, it—this token—is considered capable of being caused in different ways. Similarly, we will identify the TV image by who is in front of the camera when thinking of it as part of a broadcast show, but qua mere coloured image on the screen it is susceptible also of multiple causes, e.g. bumping the TV.

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

 



gets visual information from the brute sub-personal levels of earlier processing up to the personal level of visual experience, is not causation/representation, but partwhole constitution. Feinberg (2000) contrasts what he calls compositional, or ‘nested’ hierarchical systems with ‘non-nested’ hierarchies. A military chain of command is a non-nested hierarchy. Here the system’s constituents are physically independent of one another: the general is not composed of the lieutenants, and the lieutenants are not composed of the privates. But in a nested hierarchy, Feinberg says, ‘the elements comprising the lower levels of the hierarchy are physically combined or nested within higher levels to create increasingly complex wholes’ (2000, 79). Content carried through such a system can behave in the same way. This leads Feinberg to suggest that the structure of the visual system is a nested hierarchy, as follows: Consider again the face-responsive ‘grandmother cells’. The topical convergence of the visual pathways creates cells that are so specific that they will selectively fire to a face. The existence of such cells might lead one to think that a single ‘grandmother’ cell, at the top of the perceptual hierarchy, embodies the representation of an entire face in consciousness, but the analysis of the visual hierarchy demonstrates why this is not so. A ‘grandmother cell’ might respond quite selectively to a face, but the conscious representation of the face of one’s actual grandmother requires contributions from diverse and widely separated brain regions. The neurons that code the lines for the grandmother’s nose, the color of her eyes and where her face is located in space all must make a contribution to consciousness, and this information is not and could not be contained in a single cell. Rather, the entire nested system of the brain functions interdependently to create the visual experience of the face. Just like each organelle makes a contribution to the life of a cell, in the nested hierarchy of a mind all the lower order elements—every line, shape and patch of color that make up total awareness of the face—continue to make a contribution to consciousness.⁵⁹

I wish to emphasize two features of such a nested hierarchy, expanding upon what Feinberg says. The first point concerns causal relations, and the second point dependence relations, in the two kinds of hierarchy. First, in the non-nested chain of military command it is clear that the relationship between the elements is merely causal: orders are passed down and reports passed up the chain, and these provoke actions by the privates or the general. In particular, the relationship between the doings of the ‘lower-level’ elements—the privates—and the final ‘higher-level’⁶⁰ product—whatever the general decides the army should do—is also causal. The privates, their actions and states of mind, do not compose, but are mere prompts for, whatever resolution the general forms or decision he takes.⁶¹ In a nested ⁵⁹ Feinberg (2000). See also his (2001). ⁶⁰ Feinberg’s talk of lower- and higher-level elements is not entirely clear in the context, but I retain it for ease. The sense is clear enough as concerns a nested hierarchy: lower-level elements compose higherlevel ones. Obviously this is precisely not the case in a non-nested hierarchy, where the terms’ application is unobvious. If we think of a system’s product or output (a decision by the general, a visual image) as the higher level, then elements are lower level by virtue of being its antecedents, whether causal or compositional. This gives a univocal sense to cover both kinds of case. ⁶¹ Likewise with what the army—i.e. the privates—ends up doing, say on the battlefield, with respect to the general: generals (notoriously) are not literally part of, but only prompt, an army’s action or end state.

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

   hierarchy the relationship between the lower-level elements may also involve causation: it is undeniable that neurons cause effects in subsequent neurons as we travel along the visual system in its generation of a visual state. But, crucially, the relationship of these elements to the final product is not causal, but is rather compositional. The processing of lower-level neurons does not prompt, but actually helps to make up, the final product, the visual state, when construing the visual system as a nested hierarchy. The second, more important feature is a consequence of this point. The final product in the nested hierarchy is peculiarly dependent on the lower-order elements. In a non-nested system, as we noted, once the later elements are causally primed, earlier elements, or at least their activity, can drop out—they are redundant, and for present purposes need not even continue to exist. Once the general has ordered a certain manoeuvre it does not much matter with respect to its prosecution whether he now dies. What matters is the effect of his order in the privates’ minds. But in the nested system earlier elements and their activity take up an ongoing place in the final product, so remove them and you remove part of it. Correspondingly, in a non-nested system various kinds of antecedent could have resulted in (numerically) the same final product; they need only have had the same effects in priming the final and crucial parts of the system. Analogously, one of Punch’s token movements in a puppet show can be variously caused, by different puppeteers tugging his strings in that particular manner, or by someone manipulating the puppet’s limbs directly, or even by a gust of wind. In the nested system you cannot vary or remove the lower-order elements without altering the final product; they are intimately bound up with it. If the visual system is a nested hierarchy then a visual state is massively distributed in the brain. The content assigned to earlier neurons is not preserved by later causation-based representation of this content, but by the content actually helping to constitute the final state. Thus the ‘grandmother neuron’ does not collate the earlier information, it rather signals, perhaps coordinates, the completion of the visual state. This is to describe what may go into building a visual state, but we have not yet said what makes such a state the object of awareness—what makes us have a visual experience. My theory of consciousness extends Feinberg’s nested hierarchy model, combining it with the state structure characteristic of higher-order theories. On this view, a sensory state is nested within the further state that provides awareness of it. We may imagine two components, an awareness component and a sensory state, whose coming together, integration, supplies a full conscious state of awareness of sensory content. This dual-component model, where one component is primarily responsible for awareness and the other for content, i.e. what we experience, is reminiscent of HOT theory. But the consciousness-enabling relationship I propose is that the sensory state is embedded within the awareness component, which thereby functions as a mental ‘display frame’ of sorts. This embedding directs awareness onto

In an army the idea of a ‘final’ product is somewhat ambiguous, thus also talk of lower and higher levels, though this doesn’t substantially affect the present point as it applies to a given episode involving the army.

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

 



the sensory component, and the completed two-part state is the vehicle for a full state of consciousness. So, in contrast with the higher-order views considered above, what gets sensory content into consciousness on the present theory is not being represented by a higher-order component, but this content actually slotting into the higher-order component to compose a single complex state of awareness-of-content. It is the sensory state’s qualitative content itself, that token content, which figures in consciousness. From the perspective of the higher-order component,⁶² its job is akin to quoting the sensorily qualitative content embedded within it. Just as we slot an item to be linguistically quoted into the quotational structure ‘He/She said “-----” ’, on this ‘quotational higher-order thought’ (QHOT) theory we slot the sensory state into the state that provides awareness of it. That higher-order state must therefore possess the psychological-functional equivalent of quotation marks. Without yet knowing what this feature is, physiologically, we can identify it by its functional upshot, which is likely that the embedded state of which we are made conscious becomes cognitively accessible, in Block’s sense.⁶³ Kriegel suggests—and I like the suggestion—that phenomenal consciousness is the categorical basis of (dispositional) access consciousness.⁶⁴ Without the mental quotational frame a sensory state remains unconscious. And without a sensory state to enclose, the quotational frame lacks qualitative content. Since consciousness is always consciousness of content—consciousness is in that sense intentional, there cannot be consciousness of nothing—the frame by itself fails to supply a conscious state: that requires the combination of sensory state and quotational frame. Just so, an empty linguistic quotational structure fails to say anything. Consider as another analogy a picture frame. There is the Mona Lisa, hanging in the Louvre. Its frame’s function is to display the picture. Analogously, the mental-quotational state’s function is to make the subject aware of the content it ⁶² Being a mental state intentionally directed on a second, sensory state, albeit by means of a compositional rather than a representational relation, this component still merits the term ‘higher-order state’. The slogan ‘higher-order cognitive access’ covers both relations. ⁶³ See e.g. his (1995). Analogously a linguistic quotation makes a certain content available to the audience. ⁶⁴ Kriegel (2006). To be clear, the claim would be that the right kind of embedding, which constitutes consciousness, enables accessibility for the broader cognitive system, and that no other kind of relation could do this. Cognitive accessibility is thus the hallmark, not the ground, of consciousness. Someone might object that they can imagine the mental-quotational embedding in question without wider cognitive accessibility being implemented (perhaps there is, likewise, such a thing as a quoting event nobody hears). I don’t have to take a stand on whether this is conceivable—the important claim for QHOT theory is that mental-quotational embedding supplies consciousness. I take it that in the normal case at least mental quotation/consciousness suffices for system-wide cognitive accessibility, and if the relation is one of ground to disposition this will be metaphysically necessarily the case (even if the disposition does not manifest—i.e. the state is never accessed). But if mental quotation is possible without accessibility, the result would be cognitively isolated (as regards the wider system) phenomenal states of the sort Block (1995) intriguingly posits: e.g. phenomenally conscious pains suffered by a patient under general anaesthetic, that she can neither react to nor remember afterwards. The embedding relation, I have also said, constitutes the higher-order state’s cognitive access to the sensory state. Note that this is a narrower, not system-wide, form of cognitive access, and that it is not dispositional: the sensory state is accessed, not merely accessible (it was by hypothesis accessible to the QHOT just prior to embedding, but consciousness did not then obtain). So there is a sense in which consciousness is a form of cognitive access: but this is the core claim of all higher-order theories.

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

   embeds—a kind of subjective inner display, with the functional upshot associated with awareness. There is, notably, no possibility that the Mona Lisa’s current frame could really turn out to be displaying some other picture in the gallery. Rather it displays exclusively, is directed upon, the picture it physically encloses. This answers the question: ‘But what ties a mental quotational frame to the sensory state it happens to enfold?’ The embedding relation plausibly has what it takes to ground a primitive form of intentionality.⁶⁵ The quotational frame directs awareness onto the sensory quality it contains: the sensory quality is in this literal sense a content. To be clear, I am not talking metaphorically: I am talking about a physical embedding, on some level of brain organization, of sensory states within the apparatus of awareness.⁶⁶ The proposal is effectively that the sense of ‘content’ normally in play in these discussions, for instance in connection with representational mechanisms posited for thought and perception, is to be cashed out via a physical, spatial sense of ‘content’, instead of via the notion of representation.⁶⁷ Non-representational QHOT theory, though resembling representational higherorder theories of consciousness, is also reminiscent of quotational accounts of phenomenal concepts, those concepts we use to think about our experiences in virtue of what it is like to have them. On such accounts, too, we are said to be cognitively unusually close to experiential qualities, for instance in introspecting a pain, as the thought actually ‘quotes’ the sensation it is about.⁶⁸ I should briefly distinguish my model from the quotational model of phenomenal concepts, as this matters for how my theory captures acquaintance. First, what phenomenal concepts quote are already experiences. My theory aims at analysing what turns sensory states into experiences. Accordingly, what are embedded by quotational mental states on QHOT theory are not yet conscious states, they are mere sensory states—in-themselves-unconscious contents. Second, it is important to note that with quotational phenomenal concepts one typically quotes an experience token in order to think about an experience type—e.g. one quotes a token pain to bring it under the classification ‘pain’, to cognize it as the kind of qualitative state it is.⁶⁹ One uses the token to represent the type it belongs to, in other words, and one’s real object of thought is the type, which is not present to the mind in the way the token is. But in theories of linguistic quotation there is the interesting ⁶⁵ There is a tendency to think that causation-based tracking theories of representation must be superior when it comes to latching a mental state securely to its intentional object. But it’s hard to see why: being caused by (plus whichever fancy embellishments on top) and containing are equally extensional relations. ⁶⁶ Since on the nested hierarchy model sensory contents, even in a single modality, may be widely distributed in the brain, not to mention multi-modal experiences, our QHOTs will have to be on the large side—perhaps this is the place for Van Gulick’s brain-level ‘global higher-order states’ (Van Gulick, 2004), which are able to embed states from multiple areas across the brain. In this way the binding problem receives a solution (Coleman, 2016). By comparison, why would synchronized neural oscillations unify anything? They may be a sign, but cannot be the basis, of binding. Synchronizing things does not spatially unify them, or even seem to (otherwise keeping up with distant friends would be far easier). But binding phenomenology is in large part spatial. ⁶⁷ Compare with Balog (2012), whose talk of part/whole constitution involving mental contents is meant non-literally, or at least non-spatially. ⁶⁸ For the quotational model of phenomenal concepts see Papineau (2002), and Balog (2012). ⁶⁹ Balog (2012).

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

 



idea of quoting a token in order to talk about that very token—for instance Searle imagines quoting the token sound of a bird by leaving a gap in your sentence for its call.⁷⁰ Here the semantic duties of the quotational structure begin and end with the embedded token (sound). This is the relevant model for consciousness, for while we may think about experience types, we experience only sensory tokens.⁷¹ Hence quotational phenomenal concepts involve representation in a way QHOT theory shuns.⁷² The third difference is that quotational phenomenal concepts provide conceptually mediated access to their targets. But I hypothesize that, although I call them ‘thoughts’ in a loose sense,⁷³ QHOTs do not, or certainly need not, conceptualize their embedded sensory targets. They are rather of the form of ‘slots’ into which sensory states enter as they are. Even this brief sketch of QHOT theory should be enough for us to see how it might be used to model acquaintance:⁷⁴ • Directness: One is aware of what slots into the QHOT, and this embedding is the mechanism of awareness. One is not aware of a sensory state by being first aware of something else—in particular, one does not represent it. • Non-propositional/non-conceptual knowledge: The cognitive and epistemic connection mental quotation provides is not of the form of making a judgement about the target, and it is not something truth evaluable.⁷⁵ Moreover you do not, just in being aware of a sensory state by embedding it, bring it under any particular concept. Rather the sensory content is simply present to consciousness, as it is. Acquaintance as I have described it is a minimal condition of experience, whatever conceptualization standardly goes on top. This counts as a basic form of knowledge, since epistemic possibilities are cut down for the cognizing subject in experience: she is aware of a certain specific way the world is. • Awareness and its object are not wholly ontologically distinct: The complete consciousness-supporting state has as parts the awareness component and the ⁷⁰ Searle (1969, 76). ⁷¹ See also Zemach’s (1985) notion of a ‘mental display sentence’, which features a token mental state ‘presenting itself ’ within the relevant display structure. Zemach does not entertain a role for mental display sentences in implementing consciousness, but Kriegel (2009) seizes upon the idea for this purpose. Where Kriegel seeks to construct mental display sentences using his sophisticated self-representational structure, earlier criticized for its inability to capture acquaintance, I favour my QHOT model. See Coleman (2015b) for more on mental display sentences in relation to QHOT theory. ⁷² This is not true of the Gertler-Chalmers model (Gertler 2001; Chalmers 2003), which also uses part/ whole constitution as the intentionality-grounding mechanism of a phenomenal concept. ⁷³ In some moods I am prepared to give up this term: it seems clear that a QHOT cannot be true or false. But must all thinking be propositional? One recalls Descartes’s broader use of the notion. ⁷⁴ For more on QHOT theory see Coleman (2015b), for objections to the theory see Rosenthal (2018). ⁷⁵ But isn’t quotation propositional? So does not a completed QHOT do something like assert that a certain sensory quality is present, i.e. it does have a truth-evaluable content? But quotation is not always propositional. We can for example quote something somebody has said in order to mock them, if that something is so obviously objectionable that merely displaying it is funny. The audience may well already know that the person said it, so an assertion that they did is not part of the quotational act. The suggestion that such an act is covertly assertive, equivalent to ‘The phrase here embedded is objectionable to the point of amusement’ is as unlikely as it is unwieldy, and loses the humour. Of course what’s quoted may be propositional, for all that.

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

   sensory state, and the sensory state is embedded in the awareness component. Thus we can say that a state of awareness is not wholly distinct from that of which it provides awareness: a state of awareness-of-content has the token content built into it.⁷⁶ Still, it is not built in in that undesirable way achieved by Kriegel’s constitutive representation of qualities. On QHOT theory, the qualitative content experienced does not exist only insofar as it is experienced, and it is not constituted by our awareness of it. A sensory content and its awareness can perfectly well come apart. So, when combined, we may happily say that they are not identical, nor yet wholly ontologically distinct. They form a complex single state. We have found the relation we were seeking to model acquaintance. • Connectedly, acquaintance is a clear cognitive achievement on QHOT theory. One is not caught in the internal lightshow of a quality-producing awareness; rather awareness makes contact with something ontologically beyond itself.⁷⁷ Russell relies on a distinction between the act and the object of awareness to make good on his claim that acquaintance connects the subject to something beyond herself.⁷⁸ Though in consciousness the sensory object of awareness becomes part of a complex, and undoubtedly mental, state of awareness-of-qualitative-content, QHOT theory can capture Russell’s distinction. For, following Russell, we can closely associate the subject with the episode or state of awareness, i.e. the QHOT frame. For its part the sensory object can exist outside of awareness⁷⁹ (and may enter the vehicle of a conscious state without intrinsic change). Further, since awareness is plausibly of the essence of mentality, there is a good sense in which awareness meets something non-mental (or: not intrinsically mental) in meeting the awareness-independent sensory object.⁸⁰ Considering what is there intrinsically, a completed QHOT may thus be viewed as a mental state with a non-mental component. Such a posit is not unusual—it compares to a direct realist conception of a perceptual mental state as featuring an external-world constituent. The striking difference on the current theory is that certain objects of awareness even in the head are considered as non-mental.⁸¹ The upshot is that on QHOT theory acquaintance is a bridge between the subject and the world beyond her.⁸² • Infallibility: There is no question of a quality being misrepresented in acquaintance, since the QHOT that makes it conscious does not represent it, nor bring it ⁷⁶ Note that the awareness component, the mental-quotational frame, is not something of which we are aware in consciousness: it is part of the vehicle, or machinery, of consciousness, not part of the experience, the conscious content. See further Coleman (2015b, 2016). ⁷⁷ Recall that on Kriegel’s theory, as much as on Rosenthal’s, we are not even conscious of our sensory states, merely of facsimile content manufactured by awareness. ⁷⁸ As noted this shores up his claim that acquaintance is knowledge. ⁷⁹ Cf. Russell (1951 [1917], 113). ⁸⁰ Cf. Russell (1951 [1917], 111). That this object becomes mental in a relational sense in consciousness is not a problem, but arguably a virtue, of the theory. ⁸¹ Gertler (2007) and Coleman (2011) entertain the view that all that belongs to mentality proper is awareness, in connection and contrast with the ‘extended mind’ thesis. Gertler labels this view ‘the narrow mind’. ⁸² Cf. Russell (1912, ch. iv).

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

 



under a mode of presentation in the usual weighty sense.⁸³ What you get is the naked quality, however it is—mental quotation cannot get things wrong. This is because what carries the sensory state’s content into consciousness is just that state itself, with its content. It does this by composing one’s conscious state.⁸⁴ • Revelation: A purple sensory state, taken by itself as an intrinsically unconscious item, is plausibly a nested hierarchy of levels of content, along the lines Feinberg suggests for the grandmother visual state. In acquaintance we are aware of the sensory state in its whole depth. But we are not equally aware of each level of depth: as noted, acquaintance comes in degrees. This is explained as follows. I have modelled acquaintance metaphysically by composition: we are acquainted with our purple sensory state since it becomes part of the overall consciousnesssupporting state, by embedding in the awareness component. Our consciousness is partially composed of the purple sensory state. I earlier said that if red and blue compose the purple we must be acquainted with them too. We can now explain this fact. Composition is transitive, so if acquaintance is modelled by composition, it inherits this transitivity: whatever composes the thing we are acquainted with is also an object of acquaintance. How then to account for the fact that the purple is fully present to awareness, and epistemically revealed, in a way that the red and blue are not? The answer is that we are aware of the purple because it composes consciousness, but we are only aware of the red and blue because they compose the purple. The composition relation between the purple and awareness is in other words direct, whereas the composition relation between the red and the blue and awareness is indirect: they compose consciousness by composing something else first. Red is present only as a contribution to purple. The same follows if red in turn has components like visual warmth: we are acquainted with those only by virtue of being acquainted with the red, which they compose, and with the red because it composes the purple. We are two levels of composition removed from the visual warmth quality; accordingly, we are even further from full acquaintance with it than we are from full acquaintance with the red, which is phenomenologically an apt result. Thus I capture degrees of acquaintance by the directness of the composition relation with respect to awareness. This is why being acquainted with an item even wholly within the field of awareness does not guarantee revelation, full and perfect grasp. That is so only for the top quality.

⁸³ QHOT theory allows that awareness can modify the presentation of a sensory state to consciousness (Coleman, 2016). But such effects are limited to partial presentation of what is there anyway—e.g. where a complex quality with many horizontal components is only partially in awareness. A QHOT thus does not present a sensory quality under a guise, add to it, or represent it. So QHOTs do not have, or provide, modes of presentation in the usual substantial sense. Again (see fn. 31), if it be said that QHOTS do in this way supply modes of presentation of sensory qualities, it must also be said that cameras provide ‘modes of presentation’ of the scenes they cover. But this is not a sense of ‘mode of presentation’ whereby such a mode can mislead. For more detail on this aspect of QHOT theory see Coleman (2015b, 2016). Someone might say that in completed QHOTs an embedded sensory quality provides its own ‘mode of presentation’. But, so far as I can see, what it really means for a single property instance to provide its own mode of presentation is that there is no ‘mode’ of presentation. ⁸⁴ For more on QHOTs’ invulnerability to error see Coleman (2015b).

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

  

6. Conclusion QHOT theory closely approximates the formal features of acquaintance. And the model uses only a mundane mechanism to implement awareness, part/whole constitution, which need not trouble physicalists.⁸⁵ So the foregoing amounts to a possibility proof of a natural model of acquaintance. This far from guarantees that acquaintance is physically implemented, since for all we know its relata, awareness, and sensory qualities are in actuality non-physical.⁸⁶ But it does mean that, as regards the acquaintance relation itself, it does not merit its supernatural aura: physicalists need not reject it; nor can anti-physicalists wield it polemically. A broader point of the discussion is that we ought to look beyond causation-based representation as the content-carrying mechanism of choice across philosophy of mind and the mind sciences, else we unduly restrict our imaginative possibilities for meshing the manifest and the scientific images. If one thinks acquaintance cannot be physical because it cannot be analysed in representational terms, that only shows the need to investigate physicalism-friendly content-transmission relations other than representation.

References Balog, K. (2012) ‘Acquaintance and the Mind-Body Problem’, in C. Hill and S. Gozzano (eds), New Perspectives on Type Identity: The Mental and the Physical, Cambridge: Cambridge University Press. Block, N. (1995) ‘On a Confusion about a Function of Consciousness’, Behavioral and Brain Sciences 18: 227–87. ⁸⁵ Levine (2006, 163) objects to a constitution-based ‘acquaintance’ model of phenomenal concepts, a model designed to address the explanatory gap, on the ground that it is unclear how ‘the presence of the relevant state within the physical implementation of the representation become something of which we are aware . . . The transition from physical containment to awareness is still an inexplicable transition’. QHOTs are not concepts, nor representational, and the present model is designed as an analysis of what happens in consciousness, not in thought about consciousness. Nonetheless it might seem that Levine’s criticism retains bite: just how does QHOT theory explain awareness? Levine writes that talk of physical containment can give rise to an irresistible instinct to think that the immediacy of awareness has been captured (cf. Kriegel, 2009, 164), but believes we should resist, and demand a full account—plausibly he means a deduction of the features of awareness from the model, by analogy with how he views the H₂O-to-water explanatory relationship. For my part I am unsure why we should resist the irresistible, especially as we probably cannot expect full-dress deductions of important properties. The kind of a priori inclination to ascribe awareness that physical containment talk inspires might be all we can reasonably hope for (cf. Chalmers and Jackson’s (2001) model of reductive explanation). But I should repeat that my aim in this chapter is not to solve the hard problem, to explain awareness. The aim has been only to provide a formal model of the features of acquaintance, and to make the case for the possibility of a physical implementation. ⁸⁶ I have not shown acquaintance to be physical, and there is nothing to stop a dualist adopting the formal features of my account (Balog (2012) makes the same point about her quotational phenomenal concepts-based account of acquaintance). Giving a model with physicalism-friendly formal features is one thing, showing that the model in fact receives a physical implementation is another. One might, for example, hold that awareness cannot be satisfactorily physically explained; I have not addressed this central element of the hard problem here. On the other hand, sensory qualities may turn out to be physically irreducible. Either result means consciousness as a whole—awareness-of-qualities—cannot be naturalized. And if the relata of the acquaintance relation are non-physical, then the relation, though mundane, receives a non-physical implementation: just like meeting your great-grandfather in heaven.

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

 



Block, N. (2011) ‘The Higher-Order Approach to Consciousness is Defunct’, Analysis 71 (3): 419–31. Brentano F. (1995 [1874]) Psychology from an Empirical Standpoint, 2nd edn, trans. A. C. Rancurello, D. B. Terrell, and L. McAlister, London: Routledge. Broad, C. D. (1919) ‘Is There “Knowledge by Acquaintance”? (IV)’, Proceedings of the Aristotelian Society, Supplementary Volumes 2: 206–20. Brown, R. (2012) ‘Review of Rocco J. Gennaro, The Consciousness Paradox: Consciousness, Concepts, and Higher-Order Thoughts’, Notre Dame Philosophical Reviews, http://ndpr. nd.edu/news/30848-theconsciousness-paradox-consciousness-concepts-and-higher-orderthoughts/. Caston, V. (2002) ‘Aristotle on Consciousness’, Mind 111: 751–815. Chalmers, D. (2003) ‘The Content and Epistemology of Phenomenal Belief ’, in Q. Smith and A. Jokic (eds), Consciousness: New Philosophical Perspectives, New York: Oxford University Press. Chalmers, D. (2011) ‘The Nature of Epistemic Space’, in Andy Egan and Brian Weatherson (eds), Epistemic Modality, Oxford: Oxford University Press. Chalmers, D. J. and Jackson, F. (2001) ‘Conceptual Analysis and Reductive Explanation’, Philosophical Review 110: 315–61. Coates, P. (2007) The Metaphysics of Perception, London: Routledge. Coleman, S. (2011) ‘There Is No Argument that the Mind Extends’, Journal of Philosophy 108: 100–8. Coleman, S. (2015a) ‘Neuro-Cosmology’, in P. Coates and S. Coleman (eds), Phenomenal Qualities: Sense, Perception and Consciousness, Oxford: Oxford University Press. Coleman, S. (2015b) ‘Quotational Higher-Order Thought Theory’, Philosophical Studies, 172(10): 2705–33. Coleman, S. (2016) ‘Panpsychism and Neutral Monism: How to Make Up One’s Mind’, in G. Bruntrup and L. Jaskolla (eds), Panpsychism, New York: Oxford University Press. Coleman, S. (2018) ‘The Merits of Higher-Order Thought Theories’, Transformação. Coleman, S. (2019) ‘Personhood, Consciousness, and God’, International Journal for Philosophy of Religion 85(1): 77–98. Feigl, H. (1958) ‘The “Mental” and the “Physical” ’, Minnesota Studies in the Philosophy of Science 2: 370–497. Feigl, H. (1971) ‘Some Crucial Issues of Mind-Body Monism’, Synthese 22: 295–312. Feinberg, T. E. (2000) ‘The Nested Hierarchy of Consciousness: A Neurobiological Solution to the Problem of Mental Unity’, Neurocase 6: 75–81. Feinberg, T. E. (2001) Altered Egos: How the Brain Creates the Self, New York: Oxford University Press. Gennaro, R. J. (2011) The Consciousness Paradox, Cambridge, MA: MIT Press. Gertler, B. (2001) ‘Introspecting Phenomenal States’, Philosophy and Phenomenological Research 63 (2): 305–28. Gertler, B. (2007) ‘Overextending the Mind’, in B. Gertler and L. Shapiro (eds), Arguing about the Mind, New York: Routledge, 192–206. Gertler, B. (2012) ‘Renewed Acquaintance’, in D. Smithies and D. Stoljar (eds), Introspection and Consciousness, New York: Oxford University Press. Goff, P. (2015) ‘Real Acquaintance and Physicalism’, in P. Coates and S. Coleman (eds), Phenomenal Qualities: Sense, Perception and Cognition, Oxford: Oxford University Press. Hartshorne, C. (1934) The Philosophy and Psychology of Sensation, Chicago, IL: University of Chicago Press. Hoffman, F. (2014) ‘Non-Conceptual Knowledge’, Philosophical Issues 24(1): 184–208.

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

   Howell, R. (2008) ‘Subjective Physicalism’, in E. Wright (ed.), The Case for Qualia, Cambridge: MA: MIT Press. James, W. (1890) The Principles of Psychology, Vol. I, New York: Holt and Co. Kriegel, U. (2006) ‘Consciousness: Phenomenal Consciousness, Access Consciousness, and Scientific Practice’, in P. Thagard (ed.), Handbook of Philosophy of Psychology and Cognitive Science, Amsterdam: North Holland, 195–217. Kriegel, U. (2009) Subjective Consciousness: A Self-Representational Theory, New York: Oxford University Press. Levine, J. (2006) ‘Phenomenal Concepts and the Materialist Constraint’, in Torin Alter and Sven Walter (eds), Phenomenal Concepts and Phenomenal Knowledge: New Essays on Consciousness and Physicalism, Oxford: Oxford University Press. Lycan, W. (1996) Conscious Experience, Cambridge, MA: MIT Press. Montague, M. (2015) ‘The Life of the Mind’, in P. Coates and S. Coleman (eds), Phenomenal Qualities: Sense, Perception, and Consciousness, Oxford: Oxford University Press. Nagel, T. (1974) ‘What Is It Like to Be a Bat?’, Philosophical Review, 83: 435–50. Neander, K. (1998) ‘The Division of Phenomenal Labor: A Problem for Representational Theories of Consciousness’, Philosophical Perspectives 12: 411–34. Papineau, D. (2002) Thinking about Consciousness, Oxford: Oxford University Press. Pitt, D. (2004) ‘The Phenomenology of Cognition, or, What Is It Like to Think that P?’ Philosophy and Phenomenological Research 69: 1–36. Rosenthal, D. M. (1991) ‘The Independence of Consciousness and Sensory Quality’, Philosophical Issues 1: 15–36. Rosenthal, D. M. (2005) Consciousness and Mind, New York: Oxford University Press. Russell, B. (1910–11) ‘Knowledge by Acquaintance and Knowledge by Description’, Proceedings of the Aristotelian Society, New Series, 11: 108–28. Russell, B. (1912) The Problems of Philosophy, London: Williams and Norgate. Russell, B. (1914) Our Knowledge of the External World as a Field for Scientific Method in Philosophy, Chicago, IL: Open Court. Russell, B. (1951 [1917]) ‘The Relation of Sense Data to Physics’, in Mysticism and Logic, London: George Allen & Unwin, reprinted Totowa, NJ: Barnes and Noble Books, 1951 (pagination from the latter). Searle, J. (1969) Speech Acts, Cambridge: Cambridge University Press. Schroer, R. (2010) ‘Where’s the Beef ? Phenomenal Concepts as both Demonstrative and Substantial’, Australasian Journal of Philosophy 88(3): 505–22. Smith, B. C. (2007) ‘The Objectivity of Tastes and Tasting’, in B. C. Smith (ed.), Questions of Taste: The Philosophy of Wine, New York: Oxford University Press. Strawson, G. (2008) ‘Real Intentionality 3: Why Intentionality Entails Consciousness’, in Real Materialism and Other Essays, Oxford: Clarendon Press, 281–305. Strawson, G. (2015) ‘Self-Intimation’, Phenomenology and the Cognitive Sciences 14(1): 1–31. Tye, M. (1995) Ten Problems of Consciousness, Cambridge, MA: MIT Press. Tye, M. (2009) Consciousness: Revisited: Materialism without Phenomenal Concepts, Cambridge, MA: MIT Press. Van Gulick, R. (2004) ‘Higher-Order Global States (Hogs): An Alternative Higher-Order Model of Consciousness’, in R. J. Gennaro (ed.), Higher-Order Theories of Consciousness: An Anthology, Amsterdam: John Benjamins. Wishon, D. (2017) ‘Russellian Acquaintance and Frege’s Puzzle’, Mind 126(502): 321–70. Zahavi, D. and Kriegel, U. (2015) ‘For-Me-Ness: What It Is and What It Is Not’, in D. Dahlstrom, A. Elpidorou, and W. Hopp (eds), Philosophy of Mind and Phenomenology, London: Routledge. Zemach, E. M. (1985) ‘De se and Descartes: A New Semantics for Indexicals’, Noûs 19(2): 181–204.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

3 What Acquaintance Teaches Alex Grzankowski and Michael Tye

1. Introduction It’s hard to see how someone who knows so much could learn something more. And yet Mary, who knows every physical fact before she experiences a colour, has room to grow. Before experiencing a colour, she doesn’t know what it’s like to experience red, but once she encounters red, she does know what it’s like. A straightforward answer to how it is that Mary could learn something new has it that she learns a non-physical fact, but physicalists don’t think there are any such facts, and so there’s a challenge. The challenge is one we think physicalists can meet, but it’s worth appreciating that matters are more difficult than one might have supposed. First, we think that it is necessary and sufficient for knowing what it is like to experience red that one have a propositional thought which constitutes knowledge and which is an appropriate answer to the question ‘What is it like to experience red?’. If this is correct, non-propositional epistemic gains of the sort offered by the Ability Hypothesis¹ or the Acquaintance Approach² don’t look to be of the right form to explain Mary’s epistemic growth. But perhaps there is some physicalist-friendly propositional thought which is an answer that Mary couldn’t know before her first experience—perhaps a thought featuring a phenomenal concept for example—which she is in a position to think and so learn only after having an experience. But we think that there are no concepts off limits to someone like Mary. In fact, we think that someone who knew every propositional answer to the question ‘What is it like to experience red?’ might still not know what it is like to see red. But this looks to clash with the necessary and sufficient conditions for knowing what it is like just mentioned. So we have the makings of a puzzle: (1) Before experiencing red, Mary does not know what it is like to experience red. (2) After experiencing red, Mary does know what it is like to experience red. (3) One knows what it is like to experience red just in case one has a propositional thought which constitutes knowledge and that is an appropriate answer to the question ‘What is it like to experience red?’. ¹ See Lewis (1983, 1988) and Nemirow (1980, 1990, 2007). ² See Conee (1994) and Tye (2009). The latter is not a pure acquaintance view. Rather it endorses a ‘mixed’ approach to Mary’s epistemic growth, one component of which is acquaintance.

Alex Grzankowski and Michael Tye, What Acquaintance Teaches In: Acquaintance: New Essays. Edited by: Jonathan Knowles and Thomas Raleigh, Oxford University Press (2019). © Alex Grzankowski and Michael Tye. DOI: 10.1093/oso/9780198803461.003.0004

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

      (4) In principle, even before undergoing colour experiences, there are no propositional thoughts off limits to Mary. It’s hard to see how 1–4 could all be true. We think the way out of the puzzle is to appreciate that some answers, although plausibly individuated by their contents or by the concepts that make them up, are made appropriate not by concepts or content but by how they are formed. In order to know what it is like, an appropriate answer to a question such as ‘What is it like to experience red’ must be based on one’s (sustained) acquaintance with red. This provides the link between Mary’s new acquaintance and her propositional knowledge and does so without invoking dubious modes of presentation or concepts that we think don’t exist. On its own, acquaintance doesn’t teach us much, but when properly connected to propositional knowledge, concerning the sensible qualities with which we are acquainted, it allows us to know what they are like.

2. What Mary Learns A slight reimagining of the case is helpful. Mary has been outfitted from birth with contact lenses that allow her to see only in black and white. She is free to roam the world, interact with objects of all sorts, attend a university where she studies colours and colour vision, interact with normal perceivers, and so on. But the lenses only have a thirty-year lifespan at which time they will dissolve. On her thirtieth birthday, Mary awakens to see the bright and vivid colours that she painted her room long ago. ‘So this is an experience of red!’ she thinks to herself. There is something Mary now knows that she didn’t know when she went to bed. But what is it, exactly, that Mary has learned? When she is shown a red object, Mary comes to know what it is like to experience red; when she is shown a green object, she comes to know what it is like to experience green; and so on. But mere showing is not enough. Mary painted her room herself and she knows which colours go where. If the paint cans were mislabelled, she wouldn’t know what it is like to experience red by thinking the thought that this is what it is like to experience red while looking at, say, a green surface. A false thought will not do.³ Nor will no thought. Even if she was experiencing red, it wouldn’t be enough for her to know what it is like to experience red if she formed no thought at all. Mary comes to know what it is like to see red in part because she gains some propositional knowledge. But that knowledge has to be tied in the right way to the right kind of experience. The standard (and indeed well-supported) semantic treatment of knowledge-wh has it that sentences containing embedded questions require, for their truth, propositional knowledge which constitutes an appropriate answer to the embedded question.⁴ In order to know why white objects reflect light, one must have propositional ³ See Nida-Rümelin’s (1996, 1998) example of Marianna. See also Tye (2011) for further discussion. ⁴ The position has been discussed in detail. A helpful overview can be found in Parent (2014). See Bach (2005), Boër and Lycan (1986), Lewis (1982), Karttunen (1977), Groenendijk and Stokhof (1982), and Hintikka (1975) for discussion of the general type of approach we endorse. Higginbotham (1996) puts the point very nearly as we wish to: ‘knowledge-wh sentences may be assigned the following meta-linguistic

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

  



knowledge that constitutes an appropriate answer to the question, ‘Why do white objects reflect light?’. Knowing what it is like is no exception. It’s not enough for Mary to clap her eyes on something red. She must also come to have propositional knowledge which constitutes an appropriate answer to the question, ‘What is it like to see red?’. But by physicalists’ lights, what propositional knowledge could Mary possibly be missing?

3. Acquaintance Won’t Work, But It Might Help It’s tempting to react to the problem posed by Mary by looking for some nonpropositional knowledge that Mary might gain. We won’t defend further that knowing what it is like requires propositional knowledge—others have defended that position to our satisfaction.⁵ So, if there is non-propositional knowledge, gaining it won’t capture what Mary gains. But that’s not to say that non-propositional knowledge couldn’t help. In Section 4 we will argue that there is a kind of nonpropositional knowledge by acquaintance, that it cannot be reduced to propositional knowledge, and that Mary (before experiencing red) doesn’t have it. Since it’s nonpropositional, non-propositionally knowing red isn’t sufficient for knowing what it is like to see red, but if it is something that Mary lacks, it may be helpful in some other way, for one challenge facing physicalists is simply to find something the very knowledgeable Mary is missing.

4. Knowledge by Acquaintance Propositional knowledge is importantly similar to any other propositional attitude. When one believes or desires or hopes that p, one represents that p. The same is true when one knows that p (although when one knows it, it’s got to be the case that p). Propositional mental states represent things as being some way and hence are evaluable for accuracy, satisfaction, truth, and so on. When one fears that p, if p is true, things are as one fears them to be. When one desires that p, if p is true, things are as one desires them to be. Since knowing is factive, all instances of knowing are true or accurate but this does not prevent them from representing things as being some way. It’s just that things must be that way if the representational state one is in truth-conditions: there is a proposition p such that s knows that p, and p is a true and contextually appropriate answer to the indirect question of the wh-clause’ (381). We put the point in terms of thoughts since, as will become apparent below, we want to make room for a discussion both of the contents of thought (i.e. propositions when they are propositional thoughts) as well as their vehicles (i.e. concepts in the case of conceptual mental representation). See Braun (2006) for a dissenting voice on the connection between appropriateness and context. ⁵ See footnote 4 for general remarks about knowledge-wh. We take knowing what it is like to be a special case. For further defence see Brogaard (2011) and Stoljar (2016). Their own final views are influenced by their take on ‘what it is like’. See also Lormand (2004) and Hellie (2007) for additional discussion of ‘what it is like’. Importantly for us, all of the aforementioned views lead to a treatment of knowing what it is like in terms of knowing a propositional answer. But knowing a propositional answer can still fall short since we think it is possible to satisfy the condition of having propositional knowledge of an answer and yet to fail to know what it is like.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

      is indeed knowledge. When one knows that p, if p is true (and indeed it must be!), things are as one knows them to be. But alongside propositional mental states, there are objectual mental states which simply represent things.⁶ Such states aren’t evaluable for truth, accuracy, satisfaction, or so on—they don’t represent things as being some way. For example, suppose that John loves Bill. When is John’s love accurate? When is it true or satisfied? Under what conditions would things be as John loves them to be? Such questions seem misplaced. Non-propositional attitudes don’t have accuracy or satisfaction conditions. We should of course grant that there are propositional varieties of many of the non-propositional attitudes and they do have satisfaction or accuracy conditions. Thinking-that, loving-that, and so on can be true, satisfied, and so on. But the non-propositional instances are not like this. If these states did have propositional contents, we would expect them to be evaluable for accuracy or satisfaction and we would be able to say under what conditions things are as they are represented as being. Since they are not, we have good reason to believe that they don’t have propositional objects (mutatis mutandis for other candidate entities that would wrongly imbue the non-propositional attitudes with accuracy/ satisfaction conditions such as sentences or structures of concepts that form truthevaluable units). We think that objectual knowledge by acquaintance is another example of a non-propositional, intentional state. When one knows a thing, one needn’t represent it as being some way. One’s object knowledge isn’t true or accurate. Why think such knowledge is intentional? Because knowledge, much like thinking-of or loving, has aboutness or directedness. When one knows the colour red, the city of Austin, or Brad Pitt, one’s knowledge is directed at the colour, the city, or the person.⁷ What does it take to know a thing?⁸ It might depend on the kind of thing known. Can I know London simply by catching a glance of it while flying over? How about with a good look? It just doesn’t seem sufficient. And in fact, if I get off the plane and sit myself down in Trafalgar Square for a month, I still won’t know London. Similarly for people. If I see Brad Pitt across the room at a party, I wouldn’t be able to truthfully ⁶ See Grzankowski (2013, 2016, forthcoming) for further discussion of non-propositional, objectual mental states. For general discussion of non-propositional intentionality, see Grzanowski and Montague (2018). ⁷ One can think of or love things that don’t exist such as the Fountain of Youth or Pegasus. It’s less clear whether one can know things that don’t exist and it sounds like a stretch to say that one can be acquainted with things that don’t exist. But this should be no bar on such states being intentional, for they still have directness or aboutness. If it is correct that one cannot know or be acquainted with things that don’t exist, a comparative observation with propositional knowledge is worth making. Perhaps part of what it takes to know that p is to believe and so represent that p, but it can’t be knowledge if p isn’t true. Similarly, perhaps to know an object one must represent it (we think it is a representational state in any event), but if it is knowledge, the object must exist. In light of this, we might say that propositional knowledge is factive and knowledge by acquaintance is ‘existive’. Nothing we wish to argue in the present chapter turns on these choice points. ⁸ There is a further question—what’s knowledge? One might worry that objectual knowledge doesn’t meet the usual standards (Farkas, this volume), but we think that those standards are the standards of propositional knowledge. It seems hard to object to the claim that ‘S knows o’ is true just in case S knows o. Objectual knowledge attributions, then, are made true by objectual knowledge. It might be correct that the kind of knowledge that makes true a claim of the form ‘S knows that p’ is of a different sort, but that doesn’t speak against objectual knowledge being knowledge.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

  



claim that I now know Brad Pitt. If he and I sit silently in a small room getting a good look at each other for a few hours, I still won’t know Brad Britt. One tempting reaction to this failure of sufficiency is to hold that in order to know a person or a place, one must know some sufficient number of or some sufficient kind of facts about the person or place. Perhaps one must also meet the objects in experience (since knowing all the facts about Brad Pitt by reading about him in books isn’t enough to know him either). Such a view would have it that knowing a person or a place isn’t merely a matter of having propositional knowledge (we just saw reasons above concerning accuracy and satisfaction for resisting such an equation) but it does require it. Or perhaps knowing a person or a place is a matter of being acquainted with many parts or aspects of the object known. In order to know London—a very complex object—we might hold that one must be acquainted with many parts or aspects of it. One must be acquainted with a sufficient number of key places, one must be acquainted with the sounds and smells, one must be acquainted with the general lay of the land, and so on, the suggestion continues. With a person, perhaps one must be acquainted with some of the characteristics of their personality and some of the ways they tend to behave. Such a view would have it that acquaintance with complex things depends on acquaintance with simpler things but not on propositional knowledge. These gestures and observations can make a perfectly ordinary notion of ‘knowing’ look like a bit of a mess and things may, in many cases, be as messy as they appear. For our purposes, things needn’t be quite so complicated. When our attention is restricted to the most basic sensible qualities, one can, in what we think is a perfectly ordinary sense, know a thing simply by being conscious of it.⁹ Catch a quick whiff of skunk or take a glance at red. That’s all it takes to know the smell of skunk or the colour red. As you smell it or see it, you know it. To know a basic sensible quality is to be acquainted with it, and one is acquainted with such a quality in the first instance just in case one meets it in experience. (More on sustaining acquaintance below.) Of course some sensible qualities are complex. The taste of a good wine from Burgundy may be correctly described as complex because there are many aspects to be detected and appreciated. But there are sensible qualities which have no further sensible parts or aspects that one can meet in experience. The taste of salt, the look of a specific shade of red, and so on. When you experience such a sensible quality, you’ve experienced all of it (as it were) and it is natural to say that you, at that moment, know it. The knowledge may be fleeting, but as one’s eyes are trained on the colour, one knows the colour. And no more seems to be required. Even if I know no truths about such a quality, it is natural to say that I know it and as I learn an array of truths ⁹ Crane (2012) is sceptical of the ordinary notion of knowledge by acquaintance and offers reasons for thinking there is nothing clearly answering to the putative ordinary notion. But we think that what really follows from Crane’s observations is that knowing things is more fine grained than one might have supposed. What it takes to know a person is different from what it takes to know a place and those are different yet again from what it takes to know a basic sensible quality. Talbert (2015) holds that knowing a person requires that they also know you. Suppose that is correct. That can’t also be a requirement on knowing a city. We think that knowledge by acquaintance of basic sensible qualities requires no more than meeting the sensible quality in experience and that this is a perfectly ordinary notion. When you taste salt, you know the taste. When you smell skunk, you know the smell.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

      about the colour, I don’t come to know it (in the sense at hand) any better. I will know more about it, but won’t know it any better. This is in contrast to knowing, say, Brad Pitt. Just seeing him isn’t enough and perhaps by learning more about him, it will be correct to say that I come to know him better. But not so for basic sensible qualities. It is exceedingly plausible that it is sufficient for knowing the most basic colours, smells, tastes, and so on that one need only experience them. But of even more importance for the present chapter, it is also necessary for knowing them that one experiences them. Mary has never met red in experience and because of this she isn’t acquainted with it and so she doesn’t know it. A noteworthy aside. Suppose one agrees that when one knows a basic sensible quality (as one experiences it), one is not in a position to know it (in the sense at hand) any better. One should not then conclude that one thereby knows the thing’s nature or essence completely or even in part. The essence or nature of something is what makes it the thing that it is, and so to know the essence or nature of something one must know what makes it the kind of thing that it is. Such knowledge looks to be propositional (another instance of ‘knows-wh’). Three is what the square root of nine is, but knowing what the square root of nine is requires more than acquaintance with the number three. One might answer the question by saying, ‘Three’ or by saying, ‘Three is the square root of nine’, but one’s relationship to the thoughts that are given voice by those utterances must constitute propositional knowledge. Similarly for sensible qualities or felt qualities. When one substitutes between ‘the hurting sensation’ and ‘what pain is’, one also shifts between non-propositional and propositional knowledge attributions, and so (supposing for a moment that the hurting sensation is what pain is) it doesn’t follow that one knows what pain is from one’s knowing the hurting sensation.¹⁰ Where does all of this leave us? Besides propositional knowledge, there is nonpropositional knowledge by acquaintance. To know simple sensible qualities, it is necessary and sufficient that one meet them in experience. Mary has never met the colour red in experience, so she doesn’t know it. Although tempting to say that this gain in acquaintance knowledge is the epistemic gain physicalists have been looking for in Mary’s case, it isn’t. Mary could know green by acquaintance but think she is looking at red. When she thinks that this is an experience of red (pointing at green), she doesn’t know an answer to the question, ‘What is it like to experience red?’ since the putative answer isn’t true. Mary could also know red by acquaintance and form no thought at all and so again fail to know an answer to the question. The crux of the issue it that since knowledge by acquaintance isn’t propositional knowledge and since what Mary learns in the thought experiment is what it is like to experience red, her coming to be acquainted with red isn’t enough to explain her epistemic progress. But Mary does indeed lack knowledge by acquaintance with red—no one should doubt

¹⁰ There is a connection here to the revelation thesis put forward by Mark Johnston (1992) according to which ‘the intrinsic nature of canary yellow is fully revealed by a standard visual experience as of a canary yellow thing’ (223). There is some sense in which one cannot know canary yellow any better when one is getting a good look at it, but no propositional knowledge thereby follows. Johnston continues, ‘Hence, canary yellow is a simple non-relational property pervading surfaces, volumes and light sources’ (223). But knowing that canary yellow is simple and so on (if it is) isn’t acquaintance knowledge.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

  



that—and if we could only leverage that gap to show why it is that she must also lack some propositional knowledge before she experiences red, we may have all we need. Unfortunately, we think this connection is very hard to make. Worse, there don’t seem to be any thoughts Mary isn’t (in principle) in a position to have even when her contact lenses are in, so it’s hard to see how (by physicalist lights) any connection with acquaintance is supposed to help Mary make the right kind of epistemic gain.

5. Cut Them Coarsely or Cut Them Finely, They’re All Available to Mary One well-known approach to the Knowledge Argument is to distinguish between coarse-grained and fine-grained facts. Coarse-grained facts are worldly situations or perhaps true Russellian propositions. They are potential objects of knowledge individuated in terms of objects, properties, relations, and their arrangements. Finegrained facts are potential objects of knowledge individuated more finely, perhaps in terms of senses, ways of entertaining, modes of presentation, or concepts and their arrangements. Fine-grained facts look to provide a promising physicalist answer to the Knowledge Argument, since a physicalist might argue that Mary learns a new fine-grained fact by coming to think in a new way, but this new knowledge places no further requirements on the world. The move is familiar and its most well-known implementation can be found in the phenomenal concept strategy.¹¹ The crucial idea common to those who advance the phenomenal concept strategy is that some concepts require experience for their possession. If true, this is good news for physicalists aiming to explain Mary’s epistemic growth: Mary’s contacts dissolve and she sees colours; when she sees colours, she meets them in experience and becomes acquainted with them and so she is finally in a position to possess phenomenal colour concepts. And once she has those concepts at her disposal, she is in a position to form new thoughts, thoughts she couldn’t have had before experiencing the colours. If we individuate knowledge in a fine-grained way, in terms of finegrained facts, we can explain how it could be that Mary doesn’t know what it is like with her lenses in but does know what it is like once they are out. But this is no threat to physicalism—there are no non-physical things needed in the world to tell this story, only ways of thinking about the physical things. Phenomenal concepts are an attractive candidate for tying Mary’s acquaintance to her propositional knowledge. Mary learns what it is like because she comes to have a propositional thought which constitutes knowledge and which is an appropriate answer to the question ‘What is it like to see red?’. This thought, the appropriate thought, was unavailable to her inside the room because it is a thought which is individuated in a fine-grained way and which requires of its thinker the possession of

¹¹ See, for example, Balog (1999, 2009), Loar (1990), Lycan (1996), Papineau (1993), Perry (2001), Stoljar (2005), Sturgeon (1994), and Tye (1995, 2000a). See Alter and Walter (2007) for a recent collection of essays on phenomenal concepts.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

      the relevant concepts which in turn requires acquaintance with the colour red. The problem with this approach is that there aren’t any phenomenal concepts.¹² Concepts are mental representations of worldly entities—things, events, states, properties, etc.—and they are individuated in a fine-grained way.¹³ They are exercised whenever we undergo thoughts or other comparable mental states. One cannot notice something, recognize it, or make a judgement about it without conceptualizing it in some way, without bringing it under a concept. Concepts are, in short, representational constituents of thoughts. Thoughts are made up of concepts, and what thoughts as a whole represent is a function of their component concepts: what they represent and how they are combined. Some concepts may represent phenomenal things such as the phenomenal character of red or whatever it is that we think about when we introspect on experiences and form judgements about them. But these concepts, though about the phenomenal, are nothing special. Concepts are relatively easy to possess and easy to share. Moreover, which concepts we possess is very often a matter that depends on things external to us. Consider Burge’s well-known arthritis example (1979). The patient who goes to the doctor and complains of arthritis in his thigh could refuse to accept the doctor’s correction and insist that, whatever the doctor may think, he really does have arthritis in his thigh. Such a person would be highly atypical. The usual response would be to accept the doctor’s correction, thereby indicating that there is a shared concept in play about whose applications conditions the doctor knows more. One who rejects the doctor’s claim that arthritis is found only in the joints is operating with another concept—the concept tharthritis, as we may call it. And that concept is non-deferential. What Burge’s discussion brings out is that the concepts one deploys in thought are very often determined not by how things seem to the thinker, the kinds of descriptions one might associate with various thoughts, or other individualistic matters, but by who one engages with and with the environment in which one finds herself. And one very important upshot about this is that even without full understanding of the application conditions of the concept one is deploying in thought, one can nevertheless possess it and deploy it. And this holds true for concepts about the phenomenal as well. Consider a point made by Burge (1979): colour concepts can be over or underextended. For example, someone might have the usual beliefs as to which common objects are red, and in many cases this person might agree with others about which presented colour patches are red while also thinking that in one particular case the shade of that object over there is clearly red even though everyone else agrees that it is on the border between orange and red. Such a person would likely accept correction from others who confidently agree about the right way to classify the given shade. In this way, colour concepts are deferential. Typically, their users do not understand

¹² See Ball (2009) and Tye (2009) for more detailed defences of this claim. See Alter (2013), Ball (2013), and Veillet (2012) for additional discussion. ¹³ Although there is an array of well-known options for individuating concepts more finely than their referents, our preferred view is one according to which concepts are individuated by their origins. See Sainsbury and Tye (2012).

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

  



their conditions for application fully and are willing to accept corrections about how to apply them in some cases. But if the concepts we apply via introspection to our phenomenal states are deferential, they can be possessed even if they are only partially understood. If this is the case, it is not at all obvious that it is necessary to have undergone the relevant experiences in order to possess such concepts, any more than it is necessary to have undergone certain experiences in order to possess such concepts as arthritis. Return to Mary, wearing her contact lenses. Because concepts are easy to possess and very sharable, there don’t seem to be any concepts in principle off limits to Mary in her room. Mary spends time out in the world discussing at length colour and colour vision. She interacts with objects in the world which have colours and while chatting with her friends she says things such as, ‘This is dull brown paint and so we shouldn’t use it on the warning signs in the lab’, and ‘Red is described by all the participants as more similar to orange than it is to green’. Mary is able to discuss colours in detail and engage in disagreement over them, she gains concepts in the usual way by interacting with others and reading books, and she thinks thoughts about experiences, colours, and brains. Looking for thoughts Mary cannot have starts looking like a dead end.¹⁴ But things are now starting to seem a bit puzzling. As we said at the outset, all of the following appear true: (1) Before experiencing red, Mary does not know what it is like to experience red. (2) After experiencing red, Mary does know what it is like to experience red. (3) One knows what it is like to experience red just in case one has a propositional thought which constitutes knowledge and that is an appropriate answer to the question ‘What is it like to experience red?’. (4) In principle, even before undergoing colour experiences, there are no propositional thoughts off limits to Mary. How could it be that Mary lacks an appropriate answer if every thought is in principle available to Mary? Perhaps one way forward here is to tease apart aspects of the fine-grained and coarse-grained views by focusing on demonstrative concepts. Mary’s thoughts are composed of concepts which individuate more finely than referents and concepts are easy to posses, but her concepts have referents (at least in non-empty cases). Her concept red refers to the property of being red and her concept this refers to (roughly) Mary’s perceived and intended referent. Although Mary is capable of deploying any concept in thought, perhaps there are referents unavailable to her or conditions on the use of a concept that she cannot meet. We can imagine a scenario in which she possesses all of the concepts she might need but we can suppose that she cannot use,

¹⁴ One suggestion might be that although Mary possesses relevant phenomenal concepts, she hasn’t mastered them. On this view, when Mary leaves her room she acquires mastery of a fine-grained fact that she already knew in her room. But this won’t solve the Mary problem since on this approach Mary fails to increase her knowledge. Moreover, we are sceptical of partial concept possession and concept mastery. The relevant distinctions are better captured under conceptions of the things we bring under concepts. See Sainsbury and Tye (2012) for further discussion.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

      for example, a demonstrative concept this to think true thoughts which refer to experiences. Perhaps a more promising line then would be to look not to possession conditions the way the phenomenal concept strategist suggests but rather to what concepts refer to or perhaps to their conditions of use in terms of the availability of a referent. The suggestion is that although no thoughts construed of as structures of concepts are off limits to Mary, some thoughts of Mary’s can’t make contact with experiences in the right way to constitute true thoughts or appropriate thoughts which answer the embedded question.¹⁵ But no forthcoming way of spelling this out looks plausible. Option 1: Mary must deploy in thought a demonstrative concept which refers to an experience of red. This cannot solve the puzzle about Mary because Mary is in a position, even with the contact lenses in, to demonstrate experiences of red. She might point to a live video feed of someone in a paint shop or to a readout on a brain scanner of someone looking at a red rose, for example. ‘This is an experience of red’ she might truly say. But Mary isn’t yet in a position to know what it is like to experience red in such a case. Physicalists think that experiences are physical occurrences, so Mary, with her contacts in, should have no trouble pointing to one (be it in someone’s head or elsewhere). Option 2: Mary must not merely refer to an experience, she must refer to her own experience with the demonstrative concept. This is more promising, but the shareable nature of concepts and the fragility of acquaintance shows that this suggestion can’t quite work. Imagine that Mary’s contacts dissolve and she sees a red rose. She truly thinks to herself, ‘So this is an experience of red’. Mary now knows what it is like to experience red. But now imagine that Mary, while she is seeing red, is connected to a cerebroscope—a device which is recording her brain activity in great detail. After undergoing her experience and forming her true demonstrative thought on its basis, Mary is outfitted with new colour-blocking lenses. Moreover, through a bit of manipulation to her brain, she is made to completely forget her experience of the red rose. Mary is, in effect, right back where she started. She wonders again, ‘what is it like to experience red?’. Mary briefly knew what it is like, but she knows no longer. So now Mary doesn’t know what it is like to experience red. Not knowing what it is like, Mary then points to the cerebroscope recording on screen. She can truly say, ‘This is an experience of red’, and she can even say, ‘This is my experience of red’, but in such a case she might sensibly also claim, ‘I wish I knew what it was like to have

¹⁵ See Perry (2001) for such an approach. See also Crane (2003) who argues that in order to make sense of Jackson’s Mary (as well as various indexical cases) we need a category of ‘subjective facts’: those facts the learning of which requires that one has certain kinds of experience, or occupies a certain position in the world. According to Crane, the book of the world which aims to express all the facts cannot express the proposition that Mary expresses when, now experiencing red, she says ‘red looks like this!’. But we think the proposition so expressed is one Mary could know even in her room and so this approach cannot explain what Mary didn’t know.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

  



one of those’, or ‘I wish I could remember what it was like’. In this case, Mary’s demonstrative concept refers to an experience of hers and yet she does not know what it is like.¹⁶ Option 3: Mary must refer to an experience of hers which she is presently undergoing and presently introspecting/attending to. This looks to be the most natural recourse but it is a non-starter. One might know what it is like to taste pineapple without having had any recently. One can now know what it is like to smell the ocean even as one is too far away to catch so much as a whiff. It isn’t a requirement on knowing what it is like that one now undergo the experience. And this is true of mental images or imaginings as well. One might know what it is like to taste pineapple even though she is not now imagining the experience and referring to it in thought. While asleep, the normally visioned know what it is like to experience red. We still are in a tough spot. We seem unable to land on a condition both unavailable to Mary and yet required for knowing what it is like. So how could it be that Mary fails to know what it is like in her room? She’s got all the answers one could hope for.

6. A More General Worry about Demonstratives The appeal of demonstrative concepts is that they can only be used successfully when proper relations hold between thinkers and the referents of the concepts. Unlike the concept red which one can deploy in thought not in the presence of red things, the concept this as in ‘this red cup is too small’ can’t be successfully used if there is no red cup in the demonstrable vicinity. But a focus on demonstratives is only going to help if we really think that Mary must deploy one in order to come to know what it is like to experience red. Plenty of philosophers have been attracted to this idea and ‘This is what it is like to experience red’ (said or thought in the right situations) is a very attractive answer to the question ‘What is it like to experience red?’. But it’s not the only kind of answer that seems available. What if Mary, as she undergoes an experience of red, says to her lab-mates, ‘Red is a fantastic colour’, ‘The experience of red is even better than I had hoped’, or ‘Colour isn’t all it is cracked up to be, Dennett was right, ho hum to red’. We think it would require a bit of dogma for one to deny that Mary would hence know what it is like to experience red. She is undergoing the experience and she is forming a true thought on its basis. She knows an answer to the question and it’s very hard (in our estimation) to see why it shouldn’t be an appropriate one. So a view that requires that Mary have a demonstrative thought in order to know what it is like seems to miss the mark. The preferred view must be more flexible.

¹⁶ It is worth noting that this type of case also makes trouble for an approach in terms of concepts individuated by origin. See Sainsbury and Tye (2012). Suppose we hold that concepts are individuated by the originating use. We might hold that Mary must think a thought which has as a constituent a concept which was introduced by her on the basis of her own experience. But since concepts can outlast the experiences, forgetful Mary may still possess such a concept and yet not know what it is like.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

     

7. Acquaintance, Sustained Acquaintance, and Knowing What It Is Like We think that what Mary is missing when she hasn’t had an experience of red is an appropriate answer to the question, ‘What is it like to experience red?’ but that the failure of appropriateness is not something to be explained in terms of which concepts are being deployed in thought nor in terms of the contents of the thoughts. Rather, Mary must have a thought which answers the question and which is based in acquaintance. Return to forgetful Mary who has seen red but no longer remembers it or her experience at all. She’s now watching who she believes to be her friend on a video call experiencing red for the first time. Mary monitors the brain activity by cerebroscope. ‘This experience is amazing,’ the person on the screen tells her. ‘This experience is amazing,’ Mary repeats to herself while pointing at the screen.¹⁷ Unbeknownst to Mary, it is she who is on the screen and it is her brain that was scanned. But Mary was made to forget all about the experience and she thinks she is watching a live feed of someone else undergoing an experience. In both cases her demonstrative refers to the same experience. And as we saw, a demonstrative isn’t needed. Change the claim to, ‘Experiencing red is incredible’, and the relevant points are just the same. With her contacts in, Mary knows an answer, but she doesn’t know what it is like. When the contacts dissolved, when she was looking at red, she knew the same answer and she did know what it is like. But then she was made to forget and, once again, she did not know what it is like. The difference is that without the contacts, before she forgot her experience, Mary came to her answer on the basis of her own acquaintance with red. When the contacts were in, Mary knew an answer to the question, ‘What is it like to experience red?’ but she knew it on the basis of testimony. A case like this shows that the problem for Mary doesn’t arise from issues concerning the things she is thinking about or the concepts under which she brings them (for they are the same throughout the phases of the case), but rather from the way she forms her thoughts. It is intuitively attractive to say that knowing what it is like to experience red requires knowing an answer based on one’s acquaintance with red. But what might this ‘basis’ be? The relevant relation we have in mind is the epistemic relation of having as one’s reason.¹⁸ Mary’s reason for believing that this is an experience of red (or that experiencing red is incredible) is the fact that she is acquainted with red through her experience of it. It is our view that in the context of the Knowledge Argument, in order to know what it is like, one must think a thought which constitutes knowledge and which is based in this way on one’s acquaintance.¹⁹ ¹⁷ Physicalists can pick the preferred screen. If one thinks that Mary needs to be thinking about a sensible quality out in the world, let her point to the screen of her phone. If it is a brain state, let it be the cerebroscope screen. If neither of those cover one’s preference, put the right physical thing in view. ¹⁸ According to Harman (1973, 26), the basing relation is the epistemic relation which holds between a reason and a belief when the reason is the reason for which the belief is held. ¹⁹ Perhaps what-it-is-like questions always create such a context, but we don’t think one needs to or clearly should commit to this. In some contexts, it seems correct to say that Mary, even before experiencing red, knows what it is like to experience red. Suppose for example that Mary and her lab-mates are running an experiment on a very nervous subject. The subject has seen many colours but not red and his

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

  



Precisely how to understand epistemic basing is certainly controversial, but we think that its existence is very hard to deny.²⁰ Some beliefs are formed on the basis of reasons and some aren’t. When the formation unfolds in the right way and when the reasons are good ones, one’s belief is justified. The epistemic basing relation connects Mary’s acquaintance and her propositional thoughts had on its basis in an attractive way. It is pretty clear that the basing relation requires more than mere coincidence or mere causation. Suppose one has as an available reason for the belief that there are three cookies on the table the fact that there are three cookies on the table. But suppose further that one simply ignores this fact. If one forms the belief that there are three cookies on the table on a mere whim, the belief is not knowledge. Similar problems may haunt Mary. Suppose Mary is presently acquainted with red but is paying no attention to her experience whatsoever. On a whim, she thinks the thought that experiencing red is amazing. Although Mary is thinking a thought which answers the question, ‘What is it like to experience red?’, Mary does not know what it is like. Or suppose that Mary is presently acquainted with red and this instance of acquaintance causes a momentary abnormality in her brain that causes her to think (or perhaps constitutes her thought) that this experience is amazing. Such wayward causal chains seem insufficient for knowing what it is like and they also seem insufficient to connect reasons and beliefs. The basing relation connects acquaintance and belief in a way that avoids such problems. The idea that how one arrives at a thought might influence whether one knows something isn’t a new idea. Some true, justified beliefs are arrived at by luck and they do not constitute knowledge because of this. When Gettier’s Smith forms the belief that the man who will get the job has ten coins in his pocket, his belief is true and justified but because Smith lucks into a true belief, the belief isn’t an instance of knowledge. The present suggestion is that in order to know what it is like to, say, experience red or smell skunk, one needs to arrive at an answer in the right way. In Gettier’s case, one has a justified belief that isn’t knowledge. In the present case, Mary has knowledge all right—she knows that experiencing red is exciting, she knows that this is an experience of red, and so on. But those instances of knowledge fail to yield knowing what it is like because they aren’t appropriate pieces of knowledge in the context of the Knowledge Argument. One upshot of this view is that it is possible for Mary to know an answer (indeed many answers) to the question at hand and yet to fail to know what it is like—she might, for example, know that experiencing red is exciting without knowing what it is like to experience red. And one can move from failing to know what it is like to mischievous friends have told him that seeing red is a lot like experiencing wild, frightening hallucinations. Nervously, the subject cries out, ‘What’s it going to be like!?’. Marty, Mary’s lensed lab-mate, isn’t as studied up as Mary. ‘Mary knows what it is like,’ Marty tells the subject, ‘let me go get her.’ ‘Experiencing red is a lot like experiencing orange,’ Mary tells the subject. In such a context, it doesn’t sound far fetched to us to say that Mary, but not Marty (neither of whom have experienced red), knows what it is like to experience red (and now the subject knows too). But the Knowledge Argument sets up a context where more is needed and those are the contexts on which we will focus in the main text. ²⁰ For overview and discussion of the basing relation see Korcz (1997) and Sylvan (2016). See Audi (1993 [1986]), Moser (1989), and Swain (1979, 1985) for detailed accounts and see Bondy (2015) and Evans (2013) for more recent work.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

      knowing what it is like without forming a new or different type of thought. What one must do is base a thought (be it a new one or an old one) in the right way on one’s acquaintance. This may indeed require tokening a new thought, but no new thought type is required. Now, this might strike one as odd, so it is worth taking a moment to defend this upshot. Our claim is that it is possible for someone like Mary to possess all the concepts any of us do and, further, that it is possible for her to know an array of answers to the question ‘what is it like to experience red?’ and yet for her to not know what it is like to experience red. In other cases of knowing-wh, this seems wrongheaded, so isn’t our view ad hoc? Could one, for example, know that Mark is gardening and yet not know what Mark is doing? Or could one know that Jerry ate the cake and yet fail to know who ate the cake? It’s hard to find cases where the basis on which one knows those answers matters to whether Mark and Jerry know-wh. So, our view appears to have it that knowing what it is like is an outlier. For two reasons we think this is a defensible position and not ad hoc. First, given the cases discussed above, especially the case of Mary watching herself on the screen and taking the testimony of the person she sees, something like our position seems required. Just about everyone agrees that in order for Mary to know what it is like to experience red, she needs to undergo the right kind of experience. It’s a shared assumption that no amount of book learning, for example, will provide Mary with all she needs in order to know what it is like. But there is plenty of disagreement about how her missing experience of red stands in the way. Even if there is agreement that in order to know what it is like to experience red, one must have a propositional thought that constitutes knowledge and that is an appropriate answer to the question ‘What is it like to experience red?’, there will be plenty of room to disagree over which answers are appropriate and how they come to have that status. For example, a proponent of the phenomenal concept strategy, as we’ve seen, may ague that an appropriate answer is a thought, the having of which requires the deployment of phenomenal concepts. But we’ve seen reasons for thinking there aren’t any such concepts. Or one might maintain that an appropriate answer must demonstrate the right kind of thing, perhaps one’s own occurrent experience. But we also saw that this approach and variations upon it cannot work. The phenomenal concept strategy suffers because the thoughts apparently off limits to Mary aren’t off limits and although the demonstrative approach yields candidate answers Mary isn’t in a position to have, they are answers one who knows what it is like needn’t have available. It’s hard to see, then, how Mary’s lack of experience could stand in the way of her knowing what it is like to experience red. An unexplored avenue—and we think it is the right one—has it that Mary doesn’t know what it is like not because there are answers unavailable to her, but because she arrives at those answers in the wrong way before she has experienced red. Our proposal may look like a surprising divergence, but in light of what has come above, it seems to be exactly what’s called for. Nothing else appears to be left over. Before experiencing red, Mary forms all the thoughts about experiences of red any of the rest of us form but she does so through testimony and inference. Once Mary experiences red, she is acquainted with it and can provide an answer on that basis. So even if an outlier, we think our position is supported by its ability to navigate the puzzles presented by Mary.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

  



Second, it simply isn’t ad hoc to hold that knowing what it is like diverges from other cases of knowing-wh. As Stoljar (2015) argues, there are good reasons, for thinking that ‘what it is like’ does not pattern with other ‘wh’-phrases. Moreover, ‘what it is like’ seems to be closely related to ‘how’-questions in a way that other ‘wh’-phrases are not. Stoljar considers a number of translations of the title of Nagel’s ‘What is it like to be a bat?’ and notes the following: In at least four of these cases (German, Hungarian, Portuguese, and Russian) the interrogative word is something that is best rendered as ‘how’ in English, which strongly suggests that ‘what it is like’ questions are closely related to ‘how’ questions. Indeed, this connection is borne out in English too. ‘How does it feel to be one of the beautiful people?’ is a close variant on ‘What is it like to be one of the beautiful people?’ (1169–1170).

This fact helps our case in two ways. First, our view has it that knowing what it is like stands in contrast with other instances of knowing-wh and this contrast is in keeping with Stoljar’s observation and so not ad hoc. Second, the connection with ‘how’phrases and knowing-how is suggestive, since there are cases of knowing-how that seem sensitive to how one comes to her knowledge (just as we think knowing what it is like can be sensitive to how one comes to her knowledge). Consider recent attempts to understand knowing-how in terms of knowing-that rather than in terms of ability. A number of authors have argued persuasively that there are instances of knowing-how that fail to require ability or performance.²¹ For example, a famous pianist who has lost his arms in a terrible accident may still be said, in certain contexts, to know how to play the piano.²² In place of non-intellectual, ability-based views of knowing how, there is growing consensus that knowing how should be treated much like knowing-wh. Stanley and Williamson (2001) have argued, for instance, that S knows how to φ if, and only if, there is some contextually relevant way w for S to φ such that S stands in the knowledge-that relation to the proposition that w is away for S to φ (441).²³ But there are cases that look to be counter-examples that turn not on the general necessity of ability, but necessity in certain contexts. That is, there is good reason to think that, in at least some contexts, knowing how does require ability and merely knowing that w is a way for S to φ comes up short. Brogaard (2011) provides a Mary-esque example: As Tim is an excellent scholar, Tim was, prior to his skiing vacation, in the possession of a vast amount of knowledge-that concerning skiing. Tim knew that to slow your speed as a beginner you should use the snow plow position, that to snow plow you must stand with the tips of the

²¹ See especially Bengson and Moffett (2007) and Stanley and Williamson (2001). See the introductory chapter to Bengson and Moffett (2011a) for a general overview of many of the key issues. ²² See Glick (2012) for some complications concerning this claim. One salient feature of a case like this is that the pianist once had the ability. ²³ Interestingly, in (2007) Bengson and Moffett argue for a view that is a cousin of Stanley and Williamson’s which makes use of acquaintance with a way of performing actions. In Bengson and Moffett (2011b), acquaintance plays an even more central role. The present chapter cannot be the place to explore this thought, but perhaps the similarities between knowing how and knowing what it is like run even deeper than what’s gestured at in the main text.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

      skis closer together than the tails, that to turn right your head should move toward the tip of your right ski, and so on. But he still didn’t know how to ski. After ten days on the slope with his private skiing instructor Tim had acquired the ability to ski. Only then could Tim claim to know how to ski. (137)

To avoid these kinds of cases,²⁴ Stanley and Williamson argue that knowledge-how sometimes requires having the knowledge in question presented under a certain ‘practical mode of presentation’.²⁵ In Stanley and Williamson (2001), one is told relatively little about practical modes of presentation,²⁶ but it is clear that in the context in question concerning Tim, knowing how seems to require an answer appropriately connected to ability, action, or performance. These counter-examples to the simple intellectualist view are noteworthy. Notice that in the cases in which knowing-how seems to require being able, even if one knew, of every way, that it is a way to ride a bicycle or, of every way, that it is a way to do a backflip, there are contexts that call for more still. Knowing all of those propositions just isn’t enough. How one comes to know the proposition seems to be exactly what is called for when meeting standards of appropriateness. This parallels Mary needing more than correct answers in order to know what it is like. On our view, Mary’s knowing what it is like patterns with cases such as Tim’s knowing how to ski. In light of the connection between ‘what it is like’ and ‘how’phrases pointed out by Stoljar, rather than making ‘what it is like’ look like a worrisome outlier, our view connects it with familiar discussions of knowing-how. So our view does not predict an ad hoc abnormality. To the contrary, it predicts a pleasing similarity with knowing-how. Although the present paper isn’t the place to work out the details of knowing-how, it is worth mentioning, just briefly, that a view like ours could be offered for knowledge-how and it would allow one to avoid appeal to practical modes of presentation altogether.²⁷ One way we might connect ability and action to propositional knowledge is in terms of how one thinks of a way of doing something. But why must it be a mode of presentation that makes the difference? It seems to us even more natural to account for what we all agree on—that Tim’s knowing how to ski is connected to ability, action, or performance—by holding that Tim must know an answer to the embedded question and must know it on the basis of performance or action. Just as our view concerning Mary’s knowledge allows us to avoid phenomenal concepts, an analogous move allows one to connect action to propositional knowledge without appealing to practical modes of presenation. ²⁴ See Bengson and Moffett (2007), Cath (2011), and Glick (2012) for additional cases as well as others showing the complex connections between knowing how and ability. ²⁵ They provide their own worrying case that looks to require ability (428–9) as well and on its basis argue for practical modes of presentation. ²⁶ In (2011), Stanley goes into additional detail about how it is that he thinks of practical modes but also offers a modal treatment aimed at doing without them. ²⁷ See Pavese (2015) for a development of practical modes of presentation though see Bengson and Moffet (2007) (especially footnote 32) as well as Glick (2015) for concerns. Brogaard (2011) discusses the possibility of practical abilities serving as a justificatory ground of knowledge and is in the spirit of our preferred position.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

  



Returning to the Knowledge Argument, we aren’t quite out of the woods yet. What’s come above will cover cases in which Mary is experiencing red and on that basis thinking a thought such as that this experience is amazing, but as we saw above, one needn’t be undergoing an experience in order to know what it is like to experience red. But this creates no deep problem for the suggestion on offer since knowledge by acquaintance is something sustainable. Until recently, one of us (Alex) hadn’t met the taste of Marmite in experience. He was expecting a salty, beefy flavour and that’s not at all what he got. But it wasn’t so long ago and he still knows the taste. He can recall it, imagine it, and he’s confident he could identify it if he had some right now. So although he’s not now meeting the taste of Marmite in experience, he still knows it and his acquaintance with it is sustained in its absence. But just as it is possible to sustain acquaintance, it is possible to lose it. Think of a taste you haven’t tasted in ages such as a favourite childhood candy. It may be long enough ago that you don’t remember the taste. You might know that you really liked the candy and you might remember that it tasted good, but take a case where you don’t remember the taste. In such a case you no longer know their taste and no longer have acquaintance with the taste. What is such a person missing? One is no longer able to recall the taste, imagine the taste, and, if presented with it again, one might not identify it. Such a person is missing the abilities one must have to sustain acquaintance with a sensible quality. One who has those abilities remains acquainted and when such a person also knows an answer to the question ‘What is it like to experience X?’ on the basis of that acquaintance, then one knows what it is like even in the absence of a presentation of that sensible quality. The ties to the Ability Hypothesis should be striking here. The kinds of abilities offered by Lewis (1988) in support of the Ability Hypothesis had something going for them, but they were wrongly taken to be abilities that constituted the knowledge (knowledge-how) that Mary was missing in her room. That was a mistake since Mary can know what it is like by looking at red and thinking an appropriate thought on its basis while lacking all those abilities.²⁸ So the Ability Hypothesis fails to provide a necessary condition for knowing what it is like. But we do think that someone who has the abilities to recall, imagine, and identify has something important. These abilities sustain their acquaintance and allow one to retain the link between a sensible quality and a retained propositional answer. So what is it that acquaintance teaches? On its own, not very much. But in the hands of someone capable of reflecting on her experiences and the world around her, it puts her in a position to come to know what it is like.

References Alter, Torin (2013) ‘Social Externalism and the Knowledge Argument’, Mind 122 (486): 481–96. Alter, Torin and Walter, Sven (2007) Phenomenal Concepts and Phenomenal Knowledge: New Essays on Consciousness and Physicalism, Oxford: Oxford University Press.

²⁸ See Tye (2000b) for further discussion of the Ability Hypothesis.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

      Audi, Robert (1993 [1986]) ‘Belief, Reason And Inference’, in The Structure of Justification. Cambridge: Cambridge University Press. Bach, Kent (2005) ‘Questions and Answers’, comments on Jonathan Schaffer’s ‘Knowing the Answer’, Bellingham Summer Philosophy Conference, August. Ball, Derek (2009) ‘There Are No Phenomenal Concepts’, Mind 118 (472): 935–62. Ball, Derek (2013) ‘Consciousness and Conceptual Mastery’, Mind 122 (486): 497–508. Balog, Katlin (1999) ‘Conceivability, Possibility, and the Mind-Body Problem’, Philosophical Review 108: 497–528. Balog, Katlin (2009) ‘Phenomenal Concepts’, in Brian McLaughlin, Ansgar Beckermann, and Sven Walter (eds), Oxford Handbook in the Philosophy of Mind, Oxford: Oxford University Press, 292–312. Bengson, John and Moffett, Marc (2007) ‘Know-How and Concept Possession’, Philosophical Studies 136 (1): 31–57. Bengson, John and Moffett, Marc (eds) (2011a) Knowing How: Essays on Knowledge, Mind, and Action, Oxford: Oxford University Press. Bengson, John and Moffett, Marc (2011b) ‘Non-Propositional Intellectualism’, in John Bengson and Marc Moffett (eds), Knowing How: Essays on Knowledge, Mind, and Action, Oxford: Oxford University Press, 161–95. Boër, Steven and Lycan, William (1986) Knowing Who, Cambridge: Cambridge University Press. Bondy, Patrick (2015) ‘Counterfactuals and Epistemic Basing Relations’, Pacific Philosophical Quarterly 97 (4): 542–69. Braun, David (2006) ‘Now You Know Who Hong Oak Yun Is’, Philosophical Issues 16 (1): 24–42. Brogaard, Berit (2011) ‘Knowledge-How: A Unified Account’, in John Bengson and Marc Moffett (eds), Knowing How: Essays on Knowledge, Mind, and Action, Oxford: Oxford University Press, 136–60. Burge, Tyler (1979) ‘Individualism and the Mental’, Midwest Studies in Philosophy 4(1): 73–122. Cath, Yuri (2011) ‘Knowing How without Knowing That’, in John Bengson and Marc Moffett (eds), Knowing How: Essays on Knowledge, Mind, and Action, Oxford: Oxford University Press, 113–35. Conee, Earl (1994) ‘Phenomenal Knowledge’, Australasian Journal of Philosophy 72: 136–50. Crane, Tim (2003) ‘Subjective Facts’, in Gonzalo Rodriguez-Pereyra and Hallvard Lillehammer (eds), Real Metaphysics: Essays in Honour of D. H. Mellor, New York: Routledge, 68–83. Crane, Tim (2012) ‘Tye on Acquaintance and the Problem of Consciousness’, Philosophy and Phenomenological Research 84(1): 190–8. Evans, Ian (2013) ‘The Problem of the Basing Relation’, Synthese 190(14): 2943–57. Glick, Ephraim (2015) Practical Modes of Presentation. Noûs 49 (3):538–559. Groenendijk, Joroen and Stokhof, Martin (1982) ‘Semantic Analysis of Wh-Complements’, Linguistics and Philosophy 5: 117–233. Grzankowski, Alex (2013) ‘Non-Propositional Attitudes’, Philosophy Compass 8(12): 1123–37. Grzankowski, Alex (2016) ‘Attitudes towards Objects’, Noûs 50(2): 314–28. Grzankowski, Alex (forthcoming) ‘Non-Propositional Contents and How to Find Them’, Journal of Consciousness Studies. Grzankowski, Alex & Montague, Michelle (eds.) (2018) Non-Propositional Intentionality. Oxford University Press. Harman, Gilbert (1973) Thought, Princeton, NJ: Princeton University Press. Hellie, Benj (2007) ‘ “There’s Something It’s Like” and the Structure of Consciousness’, Philosophical Review 116(3): 441–63.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

  



Higginbotham, James (1996) ‘The Semantics of Questions’, in Shalom Lappin (ed.), The Handbook of Contemporary Semantic Theory, Oxford: Oxford University Press, 361–83. Hintikka, Jaakko (1975) The Intensions of Intentionality and Other New Models for Modalities, Dordrecht: Springer. Johnston, Mark (1992) ‘How to Speak of the Colors’, Philosophical Studies 68(3): 221–63. Karttunen, Lauri (1977) ‘Syntax and Semantics of Questions’, Linguistics and Philosophy, 1: 3–44. Korcz, Keith (1997) ‘Recent Work on the Basing Relation’, American Philosophical Quarterly, 34(2): 171–91. Lewis, David (1982) ‘Whether Report’, in Tom Pauli (ed.), Philosophical Essays Dedicated to Lennart Åqvist on his Fiftieth Birthday, Uppsala: Filosofiska Studier, 194–206. Lewis, David (1983) ‘Postscript to “Mad Pain and Martian Pain” ’, in Philosophical Papers, Vol. 1, Oxford: Oxford University Press, 130–2. Lewis, David (1988) ‘What Experience Teaches’, Proceedings of the Russellian Society, 13: 29–57. Loar, Brian (1990) ‘Phenomenal States’, in Ned Block, Owen Flanagan, and Güven Güzeldere (eds), The Nature of Consciousness, Cambridge, MA: MIT Press, 597–618. Lormand, Eric (2004) ‘The Explanatory Stopgap’, Philosophical Review 113: 303–57. Lycan, William (1996) Consciousness and Experience, Cambridge, MA: MIT Press. Moser, Paul (1989) Knowledge and Evidence, New York: Cambridge University Press. Nemirow, Lewis (1980) ‘Review of Thomas Nagel, Mortal Questions’, Philosophical Review 89: 473–7. Nemirow, Lewis (1990) ‘Physicalism and the Cognitive Role of Acquaintance’, in William Lycan (ed.), Mind and Cognition: A Reader, Oxford: Blackwell, 490–9. Nemirow, Lewis (2007) ‘So This Is What It’s Like: A Defense of the Ability Hypothesis’, in Torin Alter and Sven Walter (eds), Phenomenal Concepts and Phenomenal Knowledge: New Essays on Consciousness and Physicalism, Oxford: Oxford University Press, 32–51. Nida-Rümelin, Martine (1996) ‘What Mary Couldn’t Know: Belief about Phenomenal States’, in Thomas Metzinger (ed.), Phenomenal Consciousness, Schoenigh: Paderborn, 219–41. Nida-Rümelin, Martine (1998) ‘On Belief about Experiences: An Epistemological Distinction Applied to the Knowledge Argument’, Philosophy and Phenomenological Research 58(1): 51–73. Papineau, David (1993) Philosophical Naturalism, Oxford: Blackwell. Parent, Ted (2014) ‘Knowing-Wh and Embedded Questions’, Philosophy Compass 9(2): 81–95. Pavese, Carlotta (2015) Practical Senses. Philosophers’ Imprint 15 No. 29, pp. 1–25. Perry, John (2001) Possibility, Consciousness, and Conceivability, Cambridge, MA: MIT Press. Sainsbury, Richard Mark and Tye, Michael (2012) Seven Puzzles of Thought and How to Solve Them: An Originalist Theory of Concepts, Oxford: Oxford University Press. Stanley, Jason (2011) Know How. Oxford University Press. Stanley, Jason and Williamson, Timothy (2001) ‘Knowing How’, Journal of Philosophy 98(8): 411–44. Stoljar, Daniel (2005) ‘Physicalism and Phenomenal Concepts’, Mind and Language 20(2): 296–302. Stoljar, Daniel (2016) ‘The Semantics of “What It’s like” and the Nature of Consciousness’, Mind 125(500), 1161–98. Sturgeon, Scott (1994) ‘The Epistemic View of Subjectivity’, Journal of Philosophy 91: 221–35. Swain, Marshall (1979) ‘Justification and the Basis of Belief ’, in George Pappas (ed.), Justification and Knowledge, Boston, MA: D. Reidel, 25–50. Swain, Marshall (1985) ‘Justification, Reasons and Reliability’, Synthese, 64(1): 69–92.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

      Sylvan, Kurt (2016) ‘Epistemic Reasons II: Basing’, Philosophy Compass 11(7): 377–89. Talbert, Bonnie (2015) ‘Knowing Other People: A Second-Person Framework’, Ratio 28(2): 190–206. Tye, Michael (1995) Ten Problems of Consciousness, Cambridge, MA: MIT Press. Tye, Michael (2000a) Consciousness, Color, and Content, Cambridge, MA: MIT Press. Tye, Michael (2000b) ‘Knowing What It Is Like: The Ability Hypothesis and the Knowledge Argument’, in Gerhard Preyer (ed.), Reality and Humean Supervenience: Essays on the Philosophy of David Lewis, New York: Rowman and Littlefield. Tye, Michael (2009) Consciousness Revisited: Materialism without Phenomenal Concepts, Cambridge, MA: MIT Press. Tye, Michael (2011) ‘Knowing What It Is Like’, in John Bengson and Marc Moffett (eds), Knowing How: Essays on Knowledge, Mind, and Action, Oxford: Oxford University Press, 300–13. Veillet, Benedicte (2012) ‘In Defense of Phenomenal Concepts’, Philosophical Papers 41(1): 97–127.

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

4 Betwixt Feeling and Thinking Two-Level Accounts of Experience M. G. F. Martin

I believe it will not be very necessary to employ many words in explaining this distinction. Every one of himself will readily perceive the difference betwixt feeling and thinking. (Hume, Treatise 1.1.1)

Few readers can share David Hume’s confidence that the scant words he devotes to explaining the distinction between impressions and ideas really do succeed in the task. This distinction lies at the heart of his theory of the human mind, and commentators continue to puzzle over the question whether it is a distinction fundamentally in kind, or just one of degree; and much ink has been wasted over whether Hume’s criteria of force and vivacity succeed in sorting mental items into the one category or the other. My concern here, though, is not with what we should make of Hume’s own use of the distinction but rather to explore what others have done with it. In particular, I shall focus on the Oxford philosopher, John Foster, who was inspired to develop what I shall call a ‘two-level account’ of experience on the basis of Hume’s distinction. First in his The Case for Idealism (1982) and then, nearly twenty years later, in The Nature of Perception (2000), Foster argues for a view of experience which he claims to be ‘subjectively manifest’ to us all, and which explains Hume’s notorious distinction. In the earlier work, Foster’s concern is with what he labels ‘sensations’ and ‘images’ (Foster, 1982, 102–3); in the later work, Foster’s terms are ‘phenomenalexperiential states’, PE-states for short (Foster, 2000, 97) and ‘explicit imagistic conceiving’, EI-states for short (121). And in the latter, more extended discussion, it is clear that the second category includes at least some examples of memory. Foster’s categories of sensations and PE-states echo Hume’s category of impressions, or more exactly, impressions of sensation, or original impressions. Foster’s categories of images and EI-states echo Hume’s category of ideas, which includes both ideas of imagination and ideas of memory.

M. G. F. Martin, Betwixt Feeling and Thinking: Two-Level Accounts of Experience In: Acquaintance: New Essays. Edited by: Jonathan Knowles and Thomas Raleigh, Oxford University Press (2019). © M. G. F. Martin. DOI: 10.1093/oso/9780198803461.003.0005

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

 . . .  In both discussions, Foster insists that the pattern of similarity and difference between his two categories leads to contrasting accounts of these phenomena. His succinct statement of this in the earlier work goes as follows: in imaging, qualia are merely (albeit transparently) conceived, in sensing they are, in some way or other, realized. By occurring in the content of a sensation, a quale achieves, in some way, a concrete realization; by occurring in the content of a mental image, its realization is only (though with a peculiar vividness) represented. ((Foster, 1982), 103)

What this passage hints at is made more explicit in the later discussion’s gloss on EI-states, that the representation in question for images is a representation of the corresponding sensation: With whatever sense-realm it is associated, and in whatever form it occurs, imagistic conception is just a special—a specially focused—form of introspective conception. It is to conceive of a type of sensible item or situation in terms of what it is like, subjectively, to encounter it in the content of a certain type of experience. It is to conceive introspectively of this type of experience, but with the spotlight on the sensible universal which features in its content. (Foster, 2000, 129)

This reveals that Foster’s account is indeed a two-level one. On the base level, Foster supposes that sensation, or phenomenal experience, is at heart a relational phenomenon: the subject of experience stands in a relation of awareness to some instance of a sensible quality. There are experiential phenomena, according to Foster, which are not relational in this way: sensuous imagery and experiential memory are to be construed instead as representational. In imagining and remembering, one represents the corresponding kind of sensory act at the base level. Why only a two-level account of experience? I can not only see a red square and visualize such a red square, I can remember visualizing such a red square, or imagine remembering visualizing a red square. Once we pay attention to the possibility of such embedding of experiential episodes one within another, we must also recognize that there is a potentially indefinite hierarchy of such embedded experiences. But Foster’s account needn’t deny or neglect the extension of the hierarchy. Rather, on his view, there is a fundamental contrast between the first level and all of the subsequent levels in this hierarchy: sensations are non-representational while all of the examples of embedding will be examples of representational consciousness. There are two levels because experience is exemplified in two forms: there is the realization of sensible qualities, where a subject is related through awareness to some actual red square; then there is the representation of such qualities, where so representing things does not require them to be realized. This will be true of multiply embedded cases, such as remembering imagining seeing a red square as for the first step up, when one imagines seeing a red square. Consider the following analogy. Images of images of Fs are typically themselves also images of Fs. So, one can construct a parallel hierarchy of embedding images. One displays a red cube on a table top. One then takes a photograph of the scene and has a photographic image of a red cube. One displays a print of this image and takes a photograph of the displayed image. One now has a photograph of a photographic image of a red cube. This image too one could print, and photograph again. One now

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

   



has a photograph of a photograph of a photograph of a red cube. As long as each photographic image captures the photographic image of the print photographed, then each of these images will also be an image of a red cube. Looking at any of these images, one will confront the red cube visually, but not in a way which requires the red cube to be present in one’s visual environment. This contrasts with where we started, the original scene: one only perceives the red cube itself in its setting and not any image of the red cube. I doubt that we should seek to find in Hume’s own text this account of the distinction between impressions and ideas.¹ But note that Foster does not claim to offer it as an interpretation, but rather that to which Hume’s account answers in reality. Anyway, the theory is interesting in its own right. Foster’s account of experience has been little discussed, but it offers an intriguing mix of elements which warrant further investigation. There is the combination of a relational view of sense experience and a representational view of sensory imagination and experiential memory. Moreover, Foster endorses a form of the sense-datum theory of sensory awareness. That is, he not only holds that sensation involves a relation of awareness, but he takes the instances of sensible qualities which are the objects of such awareness to be non-environmental, and indeed to be mind-dependent. He proposes a relational conception of sense experience, but rejects such an account of sensory imagination. This stands in striking contrast to many discussions of mental imagery which assume that a sense-datum theorist of sense perception, a relational view of sense experience, should in turn adopt a relational view of sensory imagination and posit the existence of non-physical mental images.² But in its initial framing, the theory can seem peculiarly puzzling. Why should Foster assume that there is a contrast between experience as involving the realization of sensible qualities and experience as the conception or representation of this? Why should there be an opposition between the notions of realization and representation? And why should those exhaust our options when talking of the nature of experience? Furthermore, Foster does not just propose that sensory imagination and memory are forms of representation, but more specifically he proposes that we should think of them as representing the corresponding types of sensory experience. What motivates such a picture of experience as embedding other experiences? Since Foster is intent on promoting a sense-datum account of sense perception, we may profitably compare and contrast Foster’s position with earlier sense-datum theorists. Bertrand Russell presents a detailed acquaintance account not only of

¹ There is decisive evidence that this is not Hume’s view of the relation of impressions and ideas. Hume takes ideas to be individual existences: although they represent impressions and objects this is an ‘extraneous denomination’ of them, rather than an aspect of their nature (see 1.1.7 ¶6). In addition, Hume denies that in having the idea of x we thereby conceive of x perceived: see 1.4.2, where in explaining the non-contradictory nature of our conception of insensible existence, Hume repudiates the Berkeleian thought that to have the idea of an object is to conceive it as perceived (see, in particular, ¶¶39–40). But even if Foster’s picture is not an accurate rendering of Hume’s official account, still it might still count as the best understanding of the distinction which Hume wishes to draw our attention to. However, one can assess the claims of such a reading of Hume only after one had concluded that Foster’s account of the distinction is the correct way to view the mind in the first place. ² Cf. Tye (1991).

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

 . . .  sensation, but also imagination and memory in his Theory of Knowledge manuscript from 1913 (Russell, 1986b).³ This offers Russell’s most elaborated psychological theory during the period he affirmed the existence of acquaintance. Foster suggests that the combination of accounts, a relational one for sense experience, a representational one for imagination and memory, is forced on us by the need to do justice to the subjectively manifest contrast between these different kinds of experience. In The Nature of Perception, Foster presses that claim as an objection to an intentionalist or representational account of sense perception. And it puts him also, implicitly at least, at odds with Russell’s account from much earlier in the century: for Foster, recognizing the contrast between the two categories requires of us an interpretation of imagery as representational rather than relational. This disagreement between them is of interest beyond the somewhat scholastic question of what form a sense-datum theory of perception should take. For Foster’s account, together with his claim that it is grounded in what is subjectively manifest to us, challenge a methodology common to much of the recent discussion about sense perception. Many writers assume that whatever the subjective facts are, the facts about what it is like for one to see, feel, touch, or hear, that is, the phenomenological or phenomenal facts of sense perception and conscious life, such facts are immediately evident to us all on suitable reflection on the matter. This approach assumes that whatever rational disputes there can be about the nature of sense perception, such disagreement cannot be over the subjective facts: either we have all enjoyed much the same, or very similar, experiences and so disagreement is a symptom only of confusion; or our experiences turn out to be varied, some of us have encountered aspects of life denied to others; and hence apparent disagreement simply reveals talking past one another. According to this approach, we need to distinguish between the subjective facts, which are beyond dispute, and other facts which are the real target of philosophical debate: the facts which determine or constitute the subjective facts being so. Philosophers may disagree with each other whether one should endorse an intentionalist, a qualia, or a sense-datum view of the nature of sense experience, but in this they are not disagreeing about what experience is like, but only about what constitutes or explains experience being thus and so. Taking this methodological proscription seriously invites us to suppose that what we know from introspective reflection about our experience forms a kind of veil of ³ One can view the history of Oxford philosophy in the twentieth century as being in thrall to Cambridge philosophy at the start of the century. Post-World War II, the dominant influences were G. E. Moore on J. L. Austin, and Ludwig Wittgenstein on G. E. M. Anscombe, P. F. Strawson, Michael Dummett, and many others. J. O. Urmson, A. J. Ayer, and Foster all belong in a minority strand who champion Russell. Note that although Foster is a sense-datum theorist, like other post-World War II sense-datum theorists, he assumes that non-physical sense-data would have to be mind-dependent. Moore and Russell, on the other hand, insisted that sense-data, being the objects of sensory knowledge, had to be mind-independent. Introductions to these debates often suggest that this disagreement is a matter concerning optional extras to one’s theory—much as one might choose between go-faster stripes and a retractable roof when curing middle-aged despair with a shiny new car. What that misses is that the early twentieth-century debate took place against a backdrop in which the causal closure of the physical world was not assumed, while most of the philosophical debate from the 1950s on supposed that one had no other option but to grant this starting point. With that assumption settled, the issue of mind-dependence becomes settled too.

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

   



appearance: we must move beyond the subjective in order to determine its underlying nature. In turn, this restriction echoes an old complaint against the original sense-datum theorists, in particular Russell and his colleague G. E. Moore, when they sought to express the certainty of various claims about the nature of the objects of perception through appeals to a kind of first-personal reflection on experience. Critics complained that one can derive nothing of objective import from considering solely the subjective facts. The first traces of this kind of critique turns up as one strand in the debate about whether the posits of sense-datum theorists reflect a genuine disagreement about the nature of the world, or should rather be interpreted simply as an invitation to talk in a different and novel way about the world. But it is also at the root of the kind of rejection of the ontological commitments of sensedatum theories which gets associated with the label ‘adverbialism’.⁴ Although proscriptions to avoid deriving an ‘is’ from a ‘seems’ were prompted initially by a focus on the sins of sense-datum theorists, their scope is arguably of broader consequence. If it is clear that the sense-datum theory violates such a proscription by claiming the introspective character of our sense experience reveals the existence of actually existing objects of awareness, then isn’t there also a worry here for opposing representational views which deny that connection between subjective character and objects of awareness? A significant sub-set of the proponents of representational theories of experience take their views to be supported by the introspective evidence: the representational nature of our experience is revealed to us in reflection. But the claim that the occurrence of experience is independent of the actual existence of any object of awareness is no less a claim about the nature of the experiential situation than is the denial of any such independence. And in this context, we see the wider relevance of Foster’s contrasting attitudes towards sensation and imagination. If Foster’s position is tenable, then we should reconsider these methodological proscriptions. According to Foster, we can only adequately articulate the difference that is manifest between seeing and merely visualizing a red square once we recognize the very different natures of these two experiences. And if this is correct, then we can have no way of fully articulating the subjective facts of the situation while remaining neutral about the underlying nature of sense experience. I invite the reader to focus on Foster in order better to get clear on the question whether there is such a veil of appearance, an adequate description of what sense experience is like for us which remains entirely neutral about its underlying nature. I first want to highlight the basic elements of Russell’s position, and where the fundamental disagreement with Foster lies. In Section 2, I’ll consider Russell’s response to the kind of objection Foster poses for Russell’s view of imagery. I’ll suggest at this point that as unsatisfactory as Russell’s position may ultimately be, the mere insistence on a subjective difference between sense experience and imagery does not articulate where Russell goes wrong. In the third and fourth sections I’ll ⁴ One can locate the origins of this critique in Paul (1965), Barnes (1965), Ayer (1936), Ducasse (1942), and Chisholm (1950). There is an echo of this criticism in Wittgenstein’s lectures at the start of the 1930s, see (Moore 1954–5). A much later and very sophisticated articulation of the thought is to be found in Williams (1978, ch. 8).

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi



. . . 

articulate an alternative strategy which focuses on the even more controversial example of memory experience. This, I’ll suggest, explains what is so costly in Russell’s view and why we should group imagination and memory together in the way Foster proposes. In the concluding section, I’ll return to the question of what morals we should draw from this debate for the supposed veil of appearance.

1. Acquaintance enters Russell’s philosophy in 1903 (Russell, 1994 [1903]), but his first detailed discussion of its psychology does not arrive until ‘Knowledge by Acquaintance and Knowledge by Description’ (Russell, 1910–11).⁵ At that point, Russell is keen to stress the close connection between acquaintance and sense experience: acquaintance, we are told, is the converse relation of presentation. By the next year, in his avowedly popular treatment, The Problems of Philosophy (1912), Russell has revised his view and admitted the existence of memory acquaintance which contrasts with sensation or presentation. But the details of the view are worked out only later in the 1913 manuscript, The Theory of Knowledge. At this point, Russell’s account of acquaintance with particulars and the various determinate varieties of this becomes part of his account not just of experience per se but temporal experience. According to Russell we have acquaintance with universals, with logical forms, and with particulars. Particular-acquaintance divides into three varieties: sensation, memory, and imagination. In sensation, we have acquaintance with a sense-datum which exists contemporaneously with the subject of acquaintance, and which is recognized as being present. In memory, we have acquaintance with a sense-datum which is in the subject’s past and is recognized as past. In imagination, we have acquaintance with an image which is contemporaneous with the subject but without any recognition of its temporal relation to the subject. And for Russell, experience is no more than the occurrence (or co-occurrence) of such acts of acquaintance, with or without attention, and in all cases is a dual relation, presupposing the existence of both a subject and an object (see 1913, 35–7). Russell’s three-fold division of acquaintance contrasts with the two-level account we find in Foster. At first sight, Russell’s three kinds of particular-acquaintance are all on a level, in contrast to the privileged status of sense experience, and the dependent status of imagination and memory within Foster’s story. But the real contrasts between the approaches will be shown to be somewhat different, once we see the underlying principles governing Russell’s account. As I noted above, Russell’s discussion of acquaintance is part of his account of temporal experience and our recognition of a contrast between past and present. And that raises two immediate questions of interpretation. First, in simple theories of tense, we recognize not only past and present, but also future: so why doesn’t Russell’s account of acquaintance have any role for acquaintance with future sense-data? Second, Russell’s distinction among the varieties of acquaintance seems to accommodate two parameters of variation: the temporal location of the object of awareness relative to its subject on ⁵ For illuminating discussion of Russell’s conception of acquaintance see Proops (2014, 2015).

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

   



the one hand, and the subject’s recognition of that temporal location on the other. So, even given Russell’s restriction to past and present, this would predict four options and not just the three he offers: we should add acquaintance with a past image where the subject doesn’t recognize the temporal relation of the past image to him or her. Russell’s theory does not arbitrarily exclude either future objects or unrecognizedly past images, and articulating why explains the principles behind the three options he does endorse. In response to the first puzzle, it should be noted that Russell would not rule out the logical possibility of acquaintance with future sense-data. He insists that it is just a fact about the laws of nature that there is no backwards causation. Although it is not written into the definitions of mental acts and acquaintance, Russell supposes that acquaintance with any particular requires a suitable causal history by which some impression has been made on the mind. Future sense-data could so impress on the mind only if causal influence ran from future to past. Russell assumes, in keeping with common thought, that we cannot have knowledge of future particulars given how the world works (Russell 1986a [1917]). In response to the second puzzle, it is important to highlight the equivocal attitude Russell evinces towards acquaintance. Given the initial association of this notion with sensory awareness, it is tempting to interpret Russell as employing a substantive notion which helps explain how we single out some terms in thought rather than others. But, even in the 1910–11 paper, it is clear that Russell is prepared in some cases to affirm that we must have acquaintance with an entity simply on the basis that we do make judgements about it. For example, Russell claims that we are acquainted with the self because a descriptive theory of self-awareness would be viciously circular (and later, when he supposes a descriptive theory is cogent, he gives up the claim that there is acquaintance with the self). The same holds in relation to his views about memory acquaintance. When Russell first introduces memory acquaintance in The Problems of Philosophy, he makes clear that we should not confuse memory proper with any imagery which might occur at the same time as recall. The latter imagery must be in the present, Russell insists, while the object of memory acquaintance is in the subject’s past (1912, 114–15). So, what shows that we do have acquaintance with past objects is just that we have facility with the past tense and can make judgements about objects as past. Although Russell is not fully explicit in this commitment, it appears that he accepts a principle that the character of experience can be determined only by present objects of awareness. On Russell’s view, sensation and imagination both contribute to what a subject’s stream of consciousness is like, but the objects of memory contribute only through the judgements that we make. And this means that according to Russell, while there is a logical possibility of acquaintance with past images which are not recognized as such, there could be no trace of them within one’s psychology. In the images lying in the subject’s past, acquaintance with them could contribute no positive character to one’s stream of consciousness. In the lack of recognition of their temporal position, a subject would be unable to make any judgement concerning them. Such images, even if they existed, would be nothing to us. Russell’s selection of three varieties of particular-acquaintance is not arbitrary after all. Given the causal structure of the world, there is no room for acquaintance with future individuals. Given the assumption that only present objects, present sense-

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi



. . . 

data or images, can contribute to the character of our experience, then past sense-data can make a contribution to our cognitive lives only through being recognized as past. Out of six logically possible options, the only three viable ones are the three varieties of acquaintance Russell explicitly mentions. This makes clear that there is a privileged role for sensation among the three varieties of acquaintance, even if it is also true that the others are not dependent or derivative on it. Sensation exhibits two distinctive features: the objects of sensation help determine the qualitative character of one’s stream of consciousness; and in recognizing that such objects are present, a subject is placed by acquaintance in the situation of being able to have knowledge of truths about these objects, and to make true judgements of them. In this story, both memory acquaintance and imagination are defined by the ways that each differ from sensation. Memory is such that it makes no contribution to the character of our experience, although it does support knowledge of truths and the making of judgements. Imagination contributes nothing to our knowledge or our judgement, although it colours our inner lives. The most striking contrast, therefore, between Russell’s three-fold approach and Foster’s two-level account lies not in what either says about sensation, but in their attitudes towards the other two terms. Foster, echoing at least Hume’s terminology, sees a commonality among experiential memories and acts of pure sensory imagination; for him both of these occurrences are examples of imagistic conception, in keeping with Hume’s original taxonomy of ideas.⁶ For Russell, there is no such interesting commonality among memory and imagination: each is just a different way of failing to be sensation. As I’ll suggest in Section 4, this contrast is actually fundamental to the case Foster can make against Russell’s picture. Taking seriously the need to account for aspects of our temporal experience requires one to embrace the commonality of memory and imagination in the way Foster’s account suggests, and it gives us a way of explaining how it makes sense to think of the status of an act as representational can be a mode of how things appear to a subject.

2. In recent commentary, it is Russell’s extension of the notion of acquaintance to memory that is the source of most controversy concerning his psychological theories of experience. But the view of imagination he offers is equally remarkable. It is quite common in discussions of the nature of imagination to take the sense-datum theorist about perception as a model for the parallel view of imagery and mental images: if one accepts the existence of non-physical objects of sensory awareness, why not too for the case of sensory imagination? From an external critic’s point of view, the costs of a sense-datum account of sensuous imagination are no greater than the costs of a sense-datum theory of perception: positing non-physical items of awareness raises questions about the causal conditions both for sensory awareness ⁶ At a neurological level there is increasing evidence of overlap of function between imagery and episodic memory, for a recent review see Schacter et al. (2012). But in itself this doesn’t settle any of the questions raised in our dispute between Foster and Russell.

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

   



and for acts of imagination; not to mention the consistency of such theories with one’s other commitments about the fundamental nature of the world. But adopting such an external perspective misses something about sense-datum theories’ internal motivations, and how that implies a different attitude towards sense perception and imagination. Typically, sense-datum theorists do not claim that the possibility of thinking about things which are not the case, or hypothesizing the existence of merely imaginary creatures, requires us to posit any new kind of entity as the object of our psychological states. Rather, they claim that there is something distinctive and special about the case of sensory awareness, per se. HH Price, for example, admits that he cannot offer a proof that there must be something brown whenever he has visual experience, or that there must be two items of which he is aware, when he sees double, but nonetheless insists that this claim is self-evident and indubitable (Price, 1932, 63). This strategy seems distinctively directed at sensory awareness, and not at all at imagination. So, whatever exactly sense-datum theorists suppose is the introspectable, or phenomenological, ground for their position, they typically act as if this is evidence for sense experience alone, rather than for both it and imagination. From that perspective, Russell’s position concerning imagery is non-conservative among sense-datum theorists.⁷ And so it is not, perhaps, so surprising to see an implicit rejection of his view in Foster’s account. Foster’s intended target, of course, is not Russell’s account of imagery, but a representationalist account of sense experience. Foster claims that there is a manifest subjective difference between explicit imagistic conceiving and phenomenal experiences: Hume’s terminology may not succeed in making the nature of the distinction clear. But that there is a real subjective distinction, and indeed one which we might initially think of describing in terms of a difference in ‘force and liveliness’, cannot be denied. Now the problem for the imagist proposal is that it does not seem to have the resources to account for this distinction. For if the sensible awareness in both cases is that of EI-conceiving [explicit imaging conceiving], it is hard to see how there could be this striking difference in the way things introspectively appear. (Foster, 2000, 125)

Foster supposes that there is a clear subjective difference in the character of sense experience and imagination which shows that the former should be taken to involve the experiential presence of an object in a way that requires its actual existence, in contrast to the latter kind of experiential act where such a requirement is absent.⁸ ⁷ Though, it must be said that Price accepts the existence of images just as Russell did. ⁸ Of course, Foster’s insistence that there is a general subjective difference here will be questioned by some. Many philosophers (for example, Tye, 1991; Kind, 2001) have thought that the Perky experiments from the very beginning of the twentieth century support the contention that sense experience and sensory imagination can be phenomenologically identical (Perky 1910). To derive this conclusion, one both needs to hypothesize that the subjects’ inability to detect whether they were seeing rather than imagining the images projected on frosted glass can only be explained by the indistinguishability of seeing from visually imagining in this case, and that such indistinguishability in the test case is to be explained in terms of the sameness of phenomenal character. The original experiments do not really speak to either of these assumptions. And while the Perky results have been replicated in more recent studies (cf. Segal and Nathan, 1964), it is common to treat them as examples of competition for resources between visual

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi



. . . 

This is presented as an objection to the view that sense experience is a kind of imagistic conception, with that interpreted along the lines that Foster himself thinks of imagery. According to such a view, neither sense experience nor imagistic conception require the existence of an object of awareness. Foster complains that this treats the two kinds of state as too much alike, and so fails to explain adequately the systematic differences between them. And Foster takes these essential differences to be manifest to us all. If Foster’s objection is a good against his intended target, it should also be decisive against Russell. If the manifest subjective difference between sense experience and imagery reflects the first as requiring the actual existence of an object, in contrast to the second avoiding any such requirement, then it is as problematic to suppose both involve this requirement as to insist that both avoid it. Suppose Foster is right that there is a difference which applies across the board to sense experience and to imagination, and which is subjectively manifest, as he claims, in at least some cases. What would show that we had to give an account of that difference in the terms he provides? Russell is a good foil to explore this question further. For Russell does acknowledge that there is a systematic difference, and also that common deflationary explanations of the difference do not work. But he does not explain it using Hume’s vocabulary of force and vivacity, nor yet in Foster’s terms. Russell emphasizes that whatever we say about imagery we must avoid trying to characterize it in terms of non-reality or non-existence: ‘A colour visualized, or a sound heard in imagination must exist on the same level as regards “reality” as a colour seen or a sound heard in sensation: it must be equally one of the particulars which would have to be enumerated in an inventory of the universe’ (Russell, 1986b, 53). And he immediately goes on to insist that there is an important difference nonetheless between sensation and imagination which we mark with talk of ‘real’ and ‘unreal’. According to Russell, there are two key marks of the difference between images and sense experience. Russell’s principal contention is that we can mark the difference between the two in terms of the account of acquaintance we sketched above: imagination, like sensation, involves the apprehension of an image which is simultaneous with the subject’s acquaintance with it. However, in the case of sensation, the subject recognizes the simultaneity, the presentness of the sensedatum; in the case of imagery, the subject is left ignorant of the temporal position of the image (56). Russell does not take himself to be in a position to demonstrate as certain the correctness of this hypothesis, however. He rejects common suggestions that the difference in vividness or being subject to the will explain our difference of attitude, since these criteria will not be sufficiently general: it is at least possible, he claims that a sense-datum and image might be intrinsically alike. So he closes his discussion with a further possible explanation: that images lack the correlation with the physical world which we find with sense-data (62). And it is our sense of the lack of coherence here which leads us to classify them as unreal.

imagination and visual perception (see, for example, Cramer-Lemley and Reeves, 1992). For a recent and rather different hypothesis about how to interpret the Perky results, see Hopkins (2012).

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

   



Russell’s two hypotheses both treat the contrast between sensation and imagination as an epistemological one, rather than in terms which require a fundamental difference in the structure of the experiences in question. The unreality associated with imagery is articulated in terms of one or other form of ignorance that the subject has in respect of the image. One’s sense of the unreality of the image of a red square is not a symptom of its non-existence, or that one is merely conceiving such a type of thing, rather than standing in a relation of awareness to some actual red square; the sense of unreality is one’s ignorance of its temporal position, or one’s ignorance of how it fits into a systematic causal order. Moreover, if Russell can rely on his first hypothesis, then this ignorance will systematically mark the contrast in general between sensation and imagination. The first moral to draw from Russell’s account is that Foster’s explicit argument, that we recognize a subjectively manifest difference between the two, does not by itself furnish us with the materials to reject Russell’s hypothesis. Merely noting that there is a subjectively accessible difference does not settle what that difference is, nor yet how we should evaluate competing hypotheses about what the difference amounts to. Of course, one might respond that all that this shows is the weakness of Foster’s idiosyncratic strategy. One might search for some other feature of imagery which requires us to offer a representational account of it. For example, one might point out that imagery allows of a certain kind of indeterminacy which only representations permit.⁹ But as popular as such a line of argument is, it misses the real force of Russell’s position. For Russell can accept the subjective data, but question why they need to be understood in representational terms, rather than being explained solely in epistemological terms. For example, Russell might well point out that the limits to our sensory powers of discrimination are enough to show that for a sense-datum theorist intent on avoiding contradiction, some apparent samenesses among sense experiences reflect only our uncertainty of their distinctness rather than qualitative sameness. Likewise, what might otherwise be treated as qualitative indeterminacy, the number of speckles on a hen, or faces in a crowd, may be treated as simply our ignorance of the determinate fact of the matter in question. This is something Foster himself concedes in discussion of the case of sensory experience and sense-data (1985, ch. 2, sec. 10, particularly 169–74). Applied to the challenge about images, Russell can insist that what seems to be indeterminacy in the object of awareness in imagery should best be construed as the subject’s uncertainty about the determinate characteristics of that image. And the problem Russell poses is more general in its implications. Even if we can find some distinction which holds with sufficient generality among perceptual experiences and sensory images, we then need to articulate why that particular feature distinctively needs to be interpreted in representational terms. Even if Russell’s appeal to uncertainty does not in itself appear to be an adequate explanation, it highlights the issue of what it would be to detect a distinctively representational character to imagery in opposition to Russell’s metaphysics. To return to our

⁹ This is the strategy Daniel Dennett endorses (1969, 135–7).

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi



. . . 

initial concerns: does the mere fact that one is adopting a first-person perspective on experience, and seeking to articulate how things are for the subject of experience, bring the consequence that no question can be raised about the nature or existence of objects of awareness, or how we are, or are not, related to them? If it does, then it is difficult to see what resources we have to definitively show Russell to be wrong. But if we are to reject this, what is it in our discussion that will show that the issues remain one about the appearances themselves, and not just some non-experiential ground for those appearance facts? The adequacy or inadequacy of Russell’s discussions of the temporal dimension of experience will provide us with the required materials.

3. I noted above that Russell’s views on memory acquaintance seem to have generated the most controversy among commentators. But that, I suspect, is because the commentators follow Russell in treating as near equivalents ‘acquaintance’, ‘awareness’, and ‘experience’ and suppose that Russell intends us to take memory acquaintance to be a form of memory experience. As I remarked when first outlining Russell’s views about memory and sensation, Russell takes pains in both The Problems of Philosophy and The Theory of Knowledge to distinguish between memory acquaintance and any form of imagery that might occur when one recalls: There is some danger of confusion as to the nature of memory, owing to the fact that memory of an object is apt to be accompanied by an image of the object, and yet the image cannot be what constitutes memory. This is easily seen by merely noticing that the image is in the present, whereas what is remembered is known to be in the past . . . the essence of memory is not constituted by the image, but by having immediately before the mind an object which is recognized as past. (Russell, 1912, 114–15) In the first place, we must not confound true memory with present images of past things. I may call up now before my mind an image of a man I saw yesterday; the image is not in the past, and I certainly experience it now, but the image itself is not memory. The remembering refers to something known to be in the past, to what I saw yesterday, not to the image which I call up now . . . But in the immediate memory of something which has just happened, the thing itself seems to remain in experience, in spite of the fact that it is known no longer to be present. (Russell, 1986b, 9–10)

In part, Russell is concerned to alert his readers to the possibility of memory which does not involve acquaintance with the past at all (see 1986b, 57 and also ch. 7’s discussion of certainty and the varieties of memory judgement). But there is a more fundamental equivocation here about memory acquaintance as a distinctive form of experience. Russell grants that there need be no intrinsic difference between the objects of sensation, memory, and imagination. So, the only differences that are allowed are their location in time relative to the subject, and the subject’s recognition of this fact. The red patch which is now an object of attention in immediate memory is no longer part of one’s present experience, as Russell defines the notion. But the past sense-datum is both actual, as are all particulars with which one has

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

   



acquaintance, and also comes with a sense of reality, in contrast to mere images. After all, the main feature that Russell wishes to secure is that the subject be put in a position to make the judgement ‘there was such-and-such a patch of red’ parallel to the present-tense judgement elicited by sensation. Although Russell talks here of memory acquaintance, and in the manuscript draws a distinction between remote and immediate memory, his focus is really on a very narrow range of memory phenomena. The rare philosophical discussion of memory typically draws distinctions among factual memory, personal memory, and habit memory; this compares to a distinction within psychological research into memory among semantic, episodic, and practical memory (though it connects not at all with the ideas, for example, of iconic memory or working memory).¹⁰ So, Russell’s focus here is really on a sub-set of personal or episodic memory, where a subject recalls some specific past scene, and in recalling that enjoys some imagistic episode. To make matters more concrete, consider a particular example. A few days ago, I was reminiscing with a colleague about underground magazines and student activism in Britain during the early 1970s. At one moment, she was suddenly struck by the recall of an episode from when she was fourteen: an occasion, she insisted, that she had not thought about for more than thirty years, and which she had certainly not strived to recall. As a would-be militant teenager, she had written away to the School Action Union requesting various materials. Not knowing how to close the letter, she’d enquired of her grandmother a suitable closing line, and ‘thanking you in anticipation’ was offered as a suitably polite close to a letter starting out, ‘Comrades!’. The episode of recall clearly brought back a vivid sense of intense embarrassment. And this reflects, as is common with memory, that what is here recalled is not merely the original incident, but a later recall of that incident. In her memory on this occasion there was also a memory of some earlier recall of that incident: the innocent fourteen year old is recalled through the lens of the incredibly embarrassed slightly older teenage, or, again, awkward university student; and the current adult academic can take ironic distance on both responses. Now the present occasion of recall clearly had an imagistic aspect to it: on the one hand, there was the affective dimension of recall, where an aspect of the memory was an evaluative stance that differs from the recaller’s current take on the situation; on the other, my colleague stressed elements of visual and voice recollection, for example the exact tone of her grandmother’s voice in offering the advice. These imagistic elements in episodes of recall are just what Russell insists must belong only in the present, and need to be distinguished from memory proper, which connects to the past. At first sight, this insistence that all memory needs to be distinguished from the imagery that is involved in it appears unwarranted. Grant that in employing mnemonics one can associate arbitrary images with facts one wants easily to recall; and that, in response to reminiscence, one can engage in visualizing, the better to enjoy a story. In both cases, one would have the sense that on the one hand there is the imaging one engages in, and that on the other and in

¹⁰ See, for example, Malcolm (1963), Martin and Deutscher (1966), Tulving (1982), Wheeler, et al. (1997), Storm and Jobe (2012), and Schacter et al. (2012).

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi



. . . 

addition there is distinctively the recall which can notionally be distinguished from it. In contrast to such cases, there can also be occasions of recall where there is no such distinction to be made. For my colleague, the sense that this was a memory of that past occasion is conveyed in the visual image of her grandmother, and in the mixture of innocence and embarrassment that were her own past reactions. In as much as imagery is an aspect of the stream of consciousness, these elements are experiential, but in their significance, they impute the past event and not simply some current activity my colleague was engaged in. If our guide is solely what we can reflectively articulate of recall, then this kind of example gives us reason positively to affirm imagistic, or experiential, memory as a distinctive aspect of our conscious lives, and to be contrasted with mnemonics and day dreaming. We should revise this picture, and reject the idea of inherently imagistic memory experience, only if there are structural reasons to prefer a revisionary account which go beyond anything that can be introspectively accessed. The point can be underlined by exploiting a contrast to which Hume drew his readers’ attention.¹¹ Reconfigure the actual episode of accidental recall with a hypothetical situation in which my colleague’s sister teases her about her teenage activism. Imagine, for the moment, that the sister knows the details of the case, and can retell it in cringe-worthy accuracy, while my colleague has completely blanked on the occasion and can remember nothing of that period of her childhood. Her sister’s reminiscences might certainly provoke her to imagine in various ways the scene being retold and, in being told it and vividly imagining it, my colleague might vicariously feel all kinds of reaction to the scene within the story. But in this case, she would accord her imagery no authority over her sister’s words, nor yet the feelings she has or imagines as having any weight. Having blanked on her early life, she is taking in the story just as one might a biographical account of a stranger. Now imagine that the sister’s, perhaps overpointed, telling of the tale comes to stimulate the genuine memory, just as the accidental conversation did in actuality. In that case, my colleague would shift her stance, and start to give authority to how she visualized the kitchen in recalling, and the sense she now has of how she felt. Once she has the sense of recalling, she no longer needs to rely on her sister’s words. And there need be no internal mark of any source with authority distinct from the imagistic aspects of the remembering. Initially, we can draw a contrast between the memorial information that derives from the sister’s words and the imagery that colours in the contours; after the shift, there is just the imagistic recall, in all of its glorious embarrassment. If our concerns were with the mechanics of memory and recall, then there would be some purpose in distinguishing between those elements of a memory system which are responsible for providing imagistic details and those which draw on a data base of details about one’s past. But Russell’s original discussion was certainly not presented in those terms, and it is unclear why the terms of boxology or implementation are the only terms in which we can raise questions of memory. Our underlying

¹¹ See the paragraph inserted by the appendix to 1.3.5.3 about the reminiscence of two men. The example is picked up by A. J. Ayer (1956) and used also by Gareth Evans for a similar purpose (1982).

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

   



problematic was to make sense of the ways in which experience can subjectively manifest its representational, or non-representational, nature. Even if the underlying realization of episodes of recall invites us to distinguish between different elements, it is unclear why this by itself should deliver a verdict about the conscious act of recall. Moreover, Russell, and his later commentators, seem to be moved by some thoughts concerning just the character of experience, and not particularly anything about how the neurophysiology of memory might be organized. Our problem here is that while there are examples of experiential recall which fit Russell’s model, there are others which do not match his description so well. What is the further reason to discount this second set of cases? Why should we reinterpret experience which we would otherwise label as memory experience in terms of different elements, some distinctively imagery, others distinctively memorial? What is the problem in taking memory acquaintance to provide us with a distinctive kind of experience of its objects? Despite a long tradition of commentary criticizing Russell for believing in memory acquaintance, critics remain inarticulate about the problem with supposing memory experience to be a distinctive kind of acquaintance with the past.¹² So, we might do best to look outside of that debate, and look for other criticisms of empiricism. For example, one might be reminded of the puzzle in explaining our concept of the past that G. E. M. Anscombe raises in her paper ‘The Reality of the Past’ (Anscombe 1981). What would it take for someone to learn to apply their concept of the past in response to an experience which is characteristically experience of the past, which is characteristically memory experience? Go back to Russell’s example of the red patch. In the case of sensation, we have the presentation of the present sense-datum which prompts the intuitive knowledge that there is a red patch. In the case of immediate memory, no red sense-datum forms part of the subject’s total experience, but the subject does have memory acquaintance with a past, red patch. The subject is prompted to formulate a judgement concerning red, but at the same time they are to recognize that no red patch is within the present scene. Why is this not simply a situation in which they are prompted to offer contradictory judgements, ‘Some red patch’, ‘No red patch’? Anscombe notes that the ready response is to reply that the one judgement is ‘There was a red patch’ in contrast to the judgement, ‘There is no red patch’. The appearance of contradiction is removed by suitable employment of different tense inflections. But this does not satisfy her: she complains that this explanation simply presupposes the very notion of the past which the story about experience was intended to explain. Does Anscombe’s hypothetical opponent really need to offer some characterization of the situation in terms more primitive than one which already uses the notion of the past? The equivocal stance that Russell takes helps us articulate why Anscombe’s demand is reasonable. The account that Anscombe complains the lack

¹² Urmson found the idea so absurd he lambasted Pears for (correctly) attributing the view to Russell (Urmson, 1969); Pears himself attempts a number of different diagnoses of the problem (Pears, 1975), while acknowledging that according to Russell memory acquaintance cannot be the same kind of thing as sensational acquaintance.

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi



. . . 

of would be one which requires of our subject a sense of why they should be applying the concept of past in response to their experiences. If we characterize our experiences simply in terms of the objects and qualities of which we are aware, then we can make no sense of the contrast in qualitative terms alone, between an experienced object’s presently being red and some object’s pastly being red. The additional contrast made between present and past, between the tense inflections ‘is red’ and ‘was red’, are contrasts made only at the level of thought. But, what Anscombe demands of her empiricist target is some account of how marking that difference in thought might rationally be guided by a difference in experience. Anscombe herself takes this challenge to be unanswerable. According to her, there could be nothing about experience which could explain why someone would be compelled to make the appropriate pattern of judgements about present and past. And a key element of her position is the assumption that the only relevant features which might be experiential would be some quality assigned to the object, or some feeling that the subject has associated with the object. For it should be clear that no simple quality or feeling could explain the complex patterns of judgement one needs to make in revealing competence with the contrast between present and past. I suggest that we can use Anscombe’s anti-empiricism about memory and the past as an explanation of Russell’s equivocations over memory acquaintance. While Russell is inclined typically to identify particular-acquaintance with awareness and experience, on a number of occasions he allows that we have merely top-down reason to posit acquaintance: namely, our capacity to make judgements about the phenomenon in question. As we noted before, in his early discussion of self-knowledge in (1910–11), Russell embraces the claim that we are acquainted with the self only via a negative route, rejecting a descriptive theory as viciously circular; and he later rejects this conclusion in Theory of Knowledge, where he embraces a descriptive theory instead. In both places, Russell alludes to Hume’s discussion of the absence of an impression of self. So, Russell does not take self-acquaintance to involve any distinctive phenomenology of self-observation. Russell’s reflections on memory acquaintance reveal the same pattern. Russell is moved to insist on acquaintance with past sense-data because the relation of past-acquaintance gives us the sole basis for acquiring the concept of the past. But, he is also keen to avoid justifying his account of immediate memory by reference to any distinctive phenomenology at all. For Russell, there are some connotations of experience which are not guaranteed to be present in all acquaintance with particulars. Hume’s famous animadversions about the elusive self explain that attitude in the case of self-consciousness. Anscombe’s complaints about empiricism offer Russell a rationale for this stance in the case of memory. If all we can make use of to characterize the subject’s perspective are the objects and qualities of which the subject is aware, then given the logical possibility that the object of sensation and the object of memory might be identical, there is no further feature to which one could appeal to explain what is experientially distinctive of something past as opposed to present. To render his position consistent, Russell is forced back to the position that memory as such has no experiential character: the presence of memory is reflected just in the knowledge

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

   



or judgements we have in response to it. And, this explains why Russell is so insistent on distinguishing memory proper from any imagery which might be associated with it.¹³ But is Anscombe right that there is no possible response to her challenge? At this point we can see the distinctive resources Foster’s two-level account provides. Foster’s position offers a suitable response to Anscombe’s challenge. Foster insists that the distinction between the representational consciousness of imagery (imagistic conception or introspective conception as he labels it) and the presentational character of sense experience is subjectively manifest. At the very least, then, this involves rejecting the assumption mentioned above, that all that we can appeal to in accounting for what our experience is like are the objects and qualities of which we are aware. We need additionally to accommodate the way, or the mode, in which objects and qualities belong together. Foster insists that a quality can be before the mind in at least two different ways: as actualized in sensory presentation or as unactualized in mere imagistic conception. And he insists that these two ways of being present to the mind are manifestly different. It is the fact that there are two marks here, each accessible to the subject in reflection, which provides a resource to respond to Anscombe’s challenge. Foster is committed to the presence of two positive features, one each for sensation and imagery. In the case of sensation, the object is realized and that is manifest to a subject. The imaginary is not merely characterized by a lack of awareness whether the object is realized: in imagination, the object is merely conceived or represented, and the condition of being represented is equally as manifest to the subject of experience as is the condition of realization in the case of sensation. This contrasts with Russell’s explanation of the subjective difference between sense perception and imagination, where he seeks to make do with just one positive mark associated with sensation, and the possibility of its absence in the case of imagination. In sensation, the subject is aware of the manifest presentness of sense-data, while in imagination the subject is ignorant of the time position of the image. The manifest difference turns on one’s knowledge of a feature in the case of sensation, the present existence of the sense-datum, and one’s corresponding ignorance of the temporal location of the image. It should be clear that structurally such a one-way contrast is insufficient to answer Anscombe’s challenge: one’s mere ignorance of the temporal location of an image leaves open whether its being some way is consistent or inconsistent with what one’s current sense experience tells one about the present situation. For example, if one is currently aware of a red vertical on the left with a red horizontal at half way, but the right-hand side of this is obscured, then one knows that the scene one presently encounters is consistent with the presence of red H partly occluded, and is not

¹³ This is not to say that Russell rejects empiricism as such, however. Russell offers an account of concept acquisition which exploits acquaintance with instances of relational universals, being pastly acquainted being one such relation. What Russell rejects, and Anscombe insists on, is there being some qualitative, experiential, or phenomenological route into the empiricist account which does not presuppose the conceptual capacity to be explained.

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi



. . . 

consistent with the presence of a purely green H partly occluded. On the other hand, if one recalls that complex, the red vertical and the red horizontal, then one’s awareness of this past complex is quite consistent with both situations in the present, that one is currently confronting either a red or a green H. Russell’s treatment of the status of the image, as something of which one is uncertain of its temporal location, leaves one equally uncertain what possibilities it rules in or out for other objects in the present. One cannot positively know that what one is imaging is quite consistent with what one is perceiving given that they must concern different spatial or temporal locations: Russell insists that one is simply ignorant of the location of the image, and so ignorant of its relation to what holds in the present. Anscombe’s challenge demands more than this. It demands that one can grasp that the way some past object was is quite consistent with how present objects are, something that contrasts with present experience of present objects. Locating an object in one’s past frees that content from the need for consistency with any experience of the present that one may have. A suitable account of the difference between memory and sensation needs to give some explanation of this positive freedom. Foster’s account, which assigns positive features to both memory and sensation, provides us with a way of answering this challenge. According to Foster, what is subjectively manifest to one in sense perception or sensation is that the relevant qualities, e.g. shape and colour, are realized in the object of awareness. In relation to Anscombe’s challenge, one can hypothesize that this feature guides a subject to recognize that the object’s being the way it is must be compatible with the way any other thing is within the present. In contrast, for Foster, it is manifest when remembering that the object experienced is represented rather than realized in the experience. In what way could that status be experientially manifest? Since the object is, at the very least, recalled, then that object and its remembered qualities (that it is red, that it is turnstile-shaped) are objects of awareness, available to attention. And since in this case, the fact that one is remembering is evident to one, the status of these objects as being recalled is also, according to Foster, subjectively manifest. In being represented rather than presented, they are given to the subject as absent: to echo a familiar phrase, they are present in absence. And this mark, Foster can insist, guides the subject to recognize that the objects of recall have an independence of any objects in the present. In sum, the manner in which objects are presented to one as realized in sensation manifests one location within which there must be consistency, and the manner in which objects are recalled as merely represented manifests another location, within which there must also be consistency though there need be no consistency across these two locations. In this picture, the experiential difference between sensation and memory is not merely some ineffable quality, but a structural difference which echoes exactly the pattern of consistency and inconsistency that mastery of the concept of the past demands. Anscombe took her challenge to show the poverty of empiricism. Whatever the merits of empiricism, her argument is too hasty: it relies on an impoverished conception of experience. For all she says, our sense of the contrast between present and past can be reflected in structural differences among experiences of the present and experiences of the past.

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

   



4. Yet, even if Foster’s account provides sufficient materials to answer Anscombe’s challenge, we have done nothing to show that it is the only, or even the most salient way of responding to Anscombe. What shows, for example, that the only way to address the challenge is by treating memory as representational? After all, we can develop an aspect of Russell’s official position to make much the same structural point against Anscombe from a relational point of view. For Russell, acquaintance is a determinable, with sensation and immediate memory as two of its determinations. One could propose a theory on which the mode of acquaintance marks a further parameter in explaining the subject’s response to experience. Such an account introduces a further element in the story to explain the difference between sensing and remembering, one which goes beyond the objects and qualities with which one is acquainted, but which is also not some simple, ineffable, further quality. This Russellian theorist may point out that a subject can recognize that the way a sensedatum is presented in sensation is through present-acquaintance, and that the way a past sense-datum is remembered is through past-acquaintance. And the Russellian can exploit this different manner of acquaintance to offer the same story as above about how one’s grasp of the different logical locations is guided by structural features of one’s experience. This response holds on to Russell’s original idea that both sensation and memory should be thought of in terms of relations to the object rather than being treated as representational. What reason is there for preferring the Foster account to this?¹⁴ Furthermore, as I remarked above, the deep difference between Russell and Foster concerns the relation between imagination and memory, rather than the priority of sense perception. On Russell’s account, imagination and memory contrast with each other through each lacking one of the key connotations of sense perception. For Foster, echoing the original Humean taxonomy of ideas, imagination and memory are both modes of the same mental category, representational experience. So far, we have done nothing to show that one should also think of imagination as representational. Indeed, if the representational character of experiential memory is the mark of the pastness of the objects of awareness, one may wonder how imagery could also have a representational character without thereby conveying pastness. Addressing the first challenge will place us in a better position to address this second concern. First, note an aspect of experience which tends to be downplayed by the early sense-datum theorists, together with more contemporary theorists who characterize their view of sense perception as relational: it is common in philosophical discussions of consciousness and experience to talk of the ‘what it is like for

¹⁴ This view could not be Russell’s own. Although he allows for determinations of the acquaintance relation, in contrast to Moore, there is no evidence that Russell gives up the Moorean condition that acquaintance be ‘colourless’. That condition is needed to respect the idea that particular-acquaintance, as a mode of knowing, adds nothing to its object, and so shows that cognition need not alter what is cognized. For modern versions of such revised Russellianism, though, see Dorothea Debus (2008, 2016) and also John Campbell’s discussion of autobiographical memory (2002), where Campbell places much weight on his third parameter, the perspective or way in which an object is given in awareness.

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi



. . . 

someone’ to have a visual experience, or to imagine, or to remember.¹⁵ One connotation of this way of talking is that it underlines that we think of experience as something a subject undergoes, as something that happens to the subject, and that it is part of their life, something which belongs in their autobiography. In contrast, we do not suppose that an object which is seen or which is heard thereby undergoes the viewing or the hearing. In this way, there is an asymmetry between subject and object with respect to experience, even if we follow Russell in taking it to be a dual relation. Note that when we talk of spatial relations among objects, or familial relations among people, we do not take the relation to belong to one of the parties more than the other. When I am to the left of the Eiffel Tower, I do not undergo being in this spatial position any more than the Eiffel Tower undergoes being so related to me. Being a grandparent may have as a consequence many experiences I undergo, but I do not undergo being the grandparent of a child in a way that that child does not undergo being my grandson. Overemphasis on the claim that experience is relational, an element of Russell’s polemic against Meinong in The Theory of Knowledge manuscript (41–4), may blind one to this important dimension of our conception of experience. But any adequate theory of sense experience must treat this kind of experience, like others that we enjoy, as something that subjects undergo, as aspects of their subjectivity. In the case of sense experience, that aspect of experience which we associate with the subject’s undergoing also connects with other important connotations of experience and sense perception. As we commonly conceive of sense perception, our senses put us in a position to know various things about the world around us, and sense perception also puts us in a position to think about the particular objects and events in the world in our near vicinity within perceptual reach. Moreover, it is not just that we think that, in general, sense perception provides for these cognitive capacities, but also that on various occasions, an exercise of the relevant capacity is explained by reference to sense perception, and what we know of our experience. Standing in a lecture hall you raise a board marker. University logistics being what they are, there are thousands of near-identical board markers in seminar rooms and lecture theatres throughout the world. Yet of all the board markers that one could be thinking about, each one of your audience, with the exception of those asleep or daydreaming, latch on to the same one. A mysterious coincidence? We are not likely to think so: there is just one board marker which is salient to each member of the whole group: one board marker which each sees, and to which each has his or her attention drawn. And it is not just that each can see the board marker, but each is in a position to tell that he or she can see, and pay attention to, that marker. So, an explanation of why one latches on to one object and to no other is provided by our sensory perception; and in such sense perception being conscious, and open to self-aware reflection, that explanation is

¹⁵ This way of talking about the subjective aspects of experience was popularized by Nagel (1979), although it originates in Farrell (1950). But see Snowdon (2010) for scepticism about the utility of distinctively singling out the experiential in this way.

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

   



available immediately to any subject who formulates the thought and raises a question about it.¹⁶ As I remarked in Section 2, when Russell first formulates his psychological claims about acquaintance, his focus is on the case of sensation, in (1910–11), he goes as far as to claim that acquaintance is the converse of presentation. His readers, in reflecting on their own viewing of his text, will have thoughts about what visual perception provides them in respect of knowledge concerning what they can see. If we stick just with paradigm cases of visual perception, independently of Russell’s stipulations, we have some handle on what presentation of objects might be: it is that on which one can now reflect and with respect to which there is an answer to the question, What is it like to see a red H, or a board marker? Hence the claim that when one is in this kind of situation one is thereby put in a position to think about particulars and, moreover, that one is solely put in a position to think about particulars in this way, is liable to seem both to be a substantive claim, and to impose a genuine limitation on what can be thought about. On the other hand, as we noted above, Russell’s discussion of acquaintance is not restricted to cases in which he can expect his reader to have any independent handle on whether acquaintance is present or not. This is so for the early discussion of selfknowledge, which parallels what Russell has to say about memory, and memory experience. Non-acquaintance-based memory judgements may evoke a feeling of familiarity and a sense of things having been in the past, but such judgements presuppose our grasp of the concept of the past. This, according to Russell, can only be explained by immediate memory and the dual relation of past-acquaintance. But the existence of this relation is not vouchsafed by any distinctive qualitative experience of the past, but just by the availability of judgements which invoke our competence with the past tense. So, Russell doesn’t suppose that we get insight in all cases as to how acquaintance provides for reference. And in particular, he doesn’t suppose that there is any special kind of memory experience attention to which would explain our access to the past. But in our discussion of Anscombe, we have seen how we could make sense of the idea that there is full-blown memorial experience, and not just episodes of remembering to which imagery comes to be attached. That is, not just that there is some privileged relation to a past object, but that the subject undergoes an experience whose qualitative character is distinctive of being so privileged. Suppose we take the idea seriously, and following the example I gave earlier of my colleague, we insist that there are experiences which it is natural for us to treat as experiences of past events, experiential memories of one’s life. What should we say about what it is like to undergo such an experience? And how should we connect such undergoing with our understanding of the judgements it puts us in a position to make? The first thing to note is that the current memory experience involves a subjective perspective, indeed multiple subjective perspectives, which are not my colleague’s current take on the past: thinking first about the affective or evaluative dimension of

¹⁶ The question of how sense perception, consciousness, and attention fit with demonstrative reference is a core theme of Campbell (2002) and Dickie (2015).

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi



. . . 

the episode of recall, we have the innocent optimism of the misled fourteen year old, but that stance is itself embedded within a more intense embarrassment or shame at that naïveté; both of these reactions are part of the recall, but neither is how my colleague feels now about the original occasion. She has more distance, and hence more irony about the value of misdirected advice. Although she can access the optimism and the shame, she does so in a way which does not require of her that she now be ashamed or that she now be naïve. So, the past emotions, however vividly present, are not realized in the present act of recall, even though, since they are available for attention and comment, they are nonetheless in some way present to the mind. This gives us exactly the kind of experiential feature that Foster wants to highlight as a matter of representation: something which is present in absence. The rationale for drawing such a distinction between subjective quality as an object of experience and and subjective quality as realized is keenly evident in such examples of affect: remembered shame and relived shame are contrasting experiences. But once we have found one such example of this contrast, it is easy to highlight others, in addition. If the episodic recall includes a perspectival take, if it is a field memory in current jargon, such that my colleague remembers as from the chair at the kitchen table, then there is a visual perspective on her grandmother, and a sense of the direction from which the advice comes. The relevant perspectives here are not those of some stuffy academic common room in California in the twenty-first century, but of the original room back in Edinburgh some forty years ago. Again, the relevant experiential aspects, parts of what it is like for her to be remembering just so, are not aspects of her current undergoing, or only partly so; they are rather aspects of the past experiences, the past undergoings, which are now revived in the current act of recall. The past perspectives are present in the current recall in the sense that the recaller can attend to them, and can give expression to them, but they are not realized in the current experience, in as much as my colleague is not under the delusion that she has been transported back to the austerity of 1970s Edinburgh.¹⁷ So, all of these elements too are present in absence. Thinking of the experience of experiential recall in this way fits equally with the understanding that we have of how memory, just like sense perception, can put one in a position to think of some particular individual or event. Of all the occasions in which my colleague talked with her grandmother and gained advice, sound or questionable, there was just one such occasion that was brought to mind by her sudden recall. Remembering that event put her in a position to think about the occasion. Parallel to the example of seeing the board marker, we can imagine a case of recalling that board marker. Just as present visual experience of the board marker enables you to single out a particular board marker in contrast to all of

¹⁷ The issue about multiple perspectives here needs careful unpacking. The contrast drawn here is not addressed simply by hypothesising that sense experiences and imagining both have perspectival contents, or scenario contents, or contents as centred worlds. The issue raised is rather the connection between the information, or mode of information present in the experience, and the subject’s sense of location relative to those objects or events. In the case of sense experience, the perspectival character connects the objects to where one actually is, no such connection obtains for the memory.

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

   



its duplicates in the universe, later recall of the occasion enables you at that later time to think about the board marker of earlier encounter, singling it out again, in contrast to any of its duplicates, and in contrast to any other board marker which exists at the later time. Again, the understanding we have of these cognitive powers is underwritten by our reflective knowledge of the experiential episodes. What reflection reveals is that the recalled event is present to one through the earlier experience: the subjective features of that earlier experience are part of what is now recalled. This echoes those constraints on episodic or personal memory that philosophers have often taken to be distinctively characteristic of it. Namely, what Sydney Shoemaker dubs, ‘The Previous Awareness Condition’, that one can only recall episodes which one had oneself previously witnessed, been the agent of, or undergone (Shoemaker, 1984). As many writers have remarked, memory is a recursive or two-step cognitive resource: it provides us with a connection to the object of knowledge or awareness only through some prior connection with that object. The example of my colleague’s memory of embarrassment offers us an experiential echo of this: the current act of recall has within it the trace of the earlier experiential encounter. The pastness of the original encounter is reflected in the recall through the way in which it is merely represented and not realized in current recall, it is present in absence. Hence, in recognizing this structural element in our experience of recall, we underwrite the recursive conception of what memory does for us. In turn, this explains why the Foster approach is the only adequate way of addressing the Anscombe challenge, and why the modified Russellian approach is inadequate. If the previous awareness condition is to be reflected in our experience of recall, then prior experience must be an aspect of what it is like for us to recall. At the same time, for us to comprehend that this is prior experience, it cannot be sufficient for us to undergo a matching experience now: we need some phenomenological mark of the pastness of the experience, and not just of its objects. Parallel to what was said above, our recognition of the experience as merely represented and not realized can guide the subject into recognizing the different temporal location of that undergoing from their current temporal perspective on these events. Even if Russell treated memory as a form of introspection of past experience, this would not give him the resources to explain how my undergoing an experience is both an object of awareness in memory and yet is not itself undergone again. For Russell, something is part of experience through being an object of acquaintance. Either the object’s character determines the character of experiential undergoing (as in sensation and imagination), or it does not (as in memory). No further option is available. And this does not seem simply to be a limitation in Russell’s own toolbox: the relevant feature seems distinctively to be representational in character. No purely relational account of memory experience has the structural resources to satisfy this further condition. That provides an answer to our first challenge, of explaining why the Foster account uniquely satisfies the demands of Anscombe’s challenge. How does it help with the second problem? How does this discussion bear on the case of imagination? Can we see what it is about the two experiences of memory and imagination that should lead us to treat both as representational? And if imagining is thought of as

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi



. . . 

representational in the manner that Foster claims memory is, how can that be consistent with the suggestion made above that memory’s manifest representational character reveals its pastness? If we follow the same strategy as in the case of memory, and shift our attention from the object of experience to its undergoing, then we can see clearly the limitations of Russell’s epistemic account of unreality. When I visualize a red H, my act of visualizing has some dimensions in common with what I undergo in seeing a red H, and yet there are also crucial differences. To the extent that Russell insists that an image can be shaped and coloured just as a sense-datum can, the only difference he can admit between the acts of imagining and sensing will be differences in the act, the relation of acquaintance or awareness. And that is his official position: the sensing relation involves the recognition of being contemporaneous with the object, while the imaging relation lacks that element. This difference is just a difference in how one relates to the object, but is not a difference in what one undergoes in being related to the object. Why should one’s ignorance of the temporal position of the image constitute any difference in what it is like to be aware of its redness, or its H shape? But in typical cases of willed visualizing, one is quite aware that one is not seeing when one visualizes the H. The kind of experience one has is visual in some sense while clearly being other than a case of seeing. The subjective similarities between the two acts are not a matter of the two undergoings being qualitatively the same: that would require imaging to be the same kind of undergoing as seeing, and typically it is not that. So, the subjective aspects of seeing are alive in visualizing without being realized. As with the case of memory, the subjective aspects of vision are present in the act of imagining not through being realized but by being represented, by being the object of imagining itself. If this gives us a commonality between memory and imagining, it appears to create a problem for the hypothesis offered above, that the subject is guided to appreciate the contrast between the past and present through recognizing the representational nature of memory. For when we imagine something, such as visualizing a red H, we certainly do not have to take the H to be located in our past. Indeed, we do not need to take any attitude towards its temporal location. So, why should the representational character of memory connect with the sense of experienced pastness, but the representational character of imagining fail to do so? The answer to this comes from an element we have already seen in Russell’s competing theory of these experiences, albeit an element which he can’t make proper application of because of his other commitments. In denying that there is any acquaintance with future data, Russell acknowledges that our awareness of particular items is constrained by how the world can influence us. We think of our sense experience as presenting that world, in as much as some specific, or particular, event, or object, which actually exists comes to be the object of awareness. Note that the question which object counts as that which one sees, or hears, or feels, turns on the causal history and the spatial and temporal relations the subject bears to the world around them; it is not settled by how the subject conceives, or takes, or supposes the world to be. For any case of genuine perception, it makes sense to ask the question, Which F was the subject aware of? And this question is settled by the facts about how the world was then arranged, which particular object had come into the subject’s

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

   



orbit, which one of them had impinged in some way on the subject’s sensuous receptivity and prompted attention and thought. The specificity, or particularity, of one’s experience is a feature that experiential memory shares with sense perception. In recalling the conversation with her grandmother, my colleague was relating back to a specific event which happened at one time, and at no other, and in one location and in no other. Indeed, my colleague might have become confused about where or when the conversation took place, and so mistook one event for the other in her verbal report. In such a case, it makes sense to insist that one is recalling one incident and not another, where that is fixed not by the recaller’s expectations of what is recalled, or how they are inclined to describe it, but rather the actual causal history of what was originally perceived and what memory is being drawn upon.¹⁸ The question, Which F?, does not have the same application in the case of imagery. Invited to imagine an apple, or a pebble, a table, or a chair, the further question, ‘Which apple/pebble/table/chair?’ need have no answer. One can, of course, imagine particular objects which already exist and events which have already happened in the world. Imagining can be focused on objects encountered in the past, or on determinate courses of events; it can be directed even at specific future beings, and occasions which we can predict with certainty. But in such cases, what connects one’s act with the particular entity in question is the initial specification of the act of mind, how one is to imagine in order to comply with instruction or intention: here conceiving does the work that in perception and memory is done by causal connection with the world. There is no room in such conceiving for the further discovery that in fact, although one took oneself to be imagining one object, given how one was situated in relation to the various entities and happenings in the world, it turns out that one must have been imagining another such object.¹⁹ In the light of this contrast between imagery on the one hand, and sense perception and memory on the other, we can formulate the response to the Anscombe challenge more exactly. In recalling some actual incident, and so being presented with a particular occasion, or a particular object recalled, the subject is guided to appreciate that that occasion has a distinct temporal location from the present in their thereby recognizing that the occasion is present to the mind merely representationally, as opposed to being before the mind through an experience currently realized. No such recognition is part of enjoying mere imagery, despite its representational character, because there is no particular object experienced for which the question, How am I related to this?, arises. Given imagery’s inherent generality, there is no particular event or object presented in the imagery for which the question of ¹⁸ Of course, this is not to deny that memories can become conflated, and so one’s recall may no longer be strictly a faithful episodic memory of one occasion, but something more generic about the past. What is crucial here is just to recognize that there can be memories which are distinctly targeted on single occasions, and are so independent of any associated beliefs we may have about the temporal location or relation of that occasion. ¹⁹ However, this is not to deny the possibility of a kind of case made much of in Martin and Deutscher (1966), where one takes what is in fact memory of specific objects or events for mere imagining. That there should be a structural difference in the kinds of experiences we enjoy is consistent with denying infallible introspection of these differences.

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi



. . . 

temporal location arises. In this way, the particularity which sensation and memory share come to be the experiential mark of reality, which imagery lacks.²⁰ Again, unlike Russell’s own proposal, this is not an evidential contrast. We do not merely lack knowledge of the individuality of the object of imagination: in recognizing an occasion as one of imagination, we know that the question makes no sense. Memory and imagination differ with respect to experienced reality in being, or failing to be, inherently experience of a particular entity. But this is quite consistent with them also being similar in another dimension. And recognizing the commonality of memory and imagery as modes of representing acknowledges our common practice of treating false memory as a form of imagination. Consider Russell’s favourite example of George IV falsely recalling his exploits on the fields of Waterloo (Russell, 1912, 116). The king never having shown his mettle in battle, this was an entirely false memory. It is perfectly natural to describe the king in this case as merely imagining, and not recalling any such event. King George didn’t remember being at Waterloo, he simply imagined it. King George not having been on the battlefield, there is no past experience of his to be witness to his current apparent recall. The experience he now enjoys fails to have the specificity required of genuine episodic memory. Our experience may fail to have specificity either because one simply engages in imagining an F without thereby linking to some particular one, or because, although one takes oneself to be in contact with some particular F, there is in fact none such for one to be connected to. In either way of lacking such connection, we count the mental act as imagination. To have before the mind experientially some specific person or event, other than through current perception, is just to recall them. And this completes the spelling out of Foster’s two-level account of experience. The example of memorial experience explains why we must treat some experience as representational in nature, and more specifically, as the representation of other experience. Experience which manifests the aspect of presence in absence, which manifests a representational character, thereby provides an experiential ground for one’s grasp of the contrast between the past and present as different locations of objects and qualities. Moreover, memorial experience is experience of one’s past encounter with what one remembers: past experience, what one underwent, is present in current recall. Russell’s account of imagination treats the unreality of imagery as an epistemological matter: the subject is ignorant of his or her temporal relation to the image. While this may be adequate as an account of our relation to the objects of imagery, it proposes an implausible view of imagining as an experience one undergoes. There is a similarity between seeing and visualizing, but normally in visualizing it is not subjectively just as if one is seeing. In this, imagery has a commonality with memorial experience: the features of seeing are present in absence.

²⁰ Note that this explanation draws on the same assumption as Russell that we have no experience of future particulars (and perhaps, for the same reason, that we lack the appropriate causal connections to purely future objects). The account of the distinctively structural features of experience of past provides no element of directionality. This is not a worrying lack if there can be no corresponding experience in the opposite direction of time from the present.

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

   



5. In contrasting Foster’s sense-datum account with that of Russell we have gained a better understanding of the form of, and motivations for, the two-level account. Foster projects Hume’s contrast between impressions and ideas into a contrast between occasions of sensible presentation, and episodes which involve the conception or representation of the former. Foster claims that the contrast he draws is subjectively manifest. And the puzzles with which we started were, Why should Foster suppose that there is a choice between presentation and representation? and, Why should he suppose that the representational mode of experience should involve representing experiences at the bottom of the hierarchy? Our discussion of Anscombe’s challenge and of the role of episodic memory in our knowledge of the past addresses these two concerns. Foster claims that the distinction he proposes is subjectively manifest to us all. And as he first presents matters, this amounts to the claim that we can recognize that some sensible quality is realized in sensation but only represented in imagination or memory. However, our discussion of the implicit disagreements with Russell enables us to articulate the position more precisely. Our route in was to take seriously the idea that there could be genuinely memorial experience. Something Russell is committed dogmatically to denying. Our discussion, in effect, spells out the full consequences of acknowledging such temporal experience, and how the existence of such distinctive experience is inconsistent with a pure sense-datum theory applied to all varieties of experience. The key element which is manifestly not realized in recalling a past event is the subjective undergoing of the original experience. Nonetheless, that experience is part of what is recalled, to the extent that it is manifest to a subject why he or she can think about the occasion in question: the original encounter with the episode is part of what one experiences now, only it is not realized again in the present, but merely represented. We are forced to employ the contrast between something subjective being realized and its merely being represented, or present in absence, in order consistently to acknowledge that the past visual experience or emotional response is at once an element of current recall and at the same time not something that one undergoes again. That distance is equally an aspect of episodes of imagination: in imagining pain or anger one need not thereby feel pain or anger. And this supports Foster’s classification of both as forms of representational experience. On this picture, each is manifestly representational because one experiences various subjective elements without those elements being realized in one’s current experience itself. And this provides answers to the questions about Foster’s two-level account. There is a simple opposition between presentation and representation simply because each is associated with one answer to the binary question whether some aspect of the object of experience is realized in the experience itself. Experience is presentational where its object must be actual for experience to be so; it is representational where an experience being of such an object allows for the non-actuality of the object in question. In turn, our discussion of the case of memorial experience, and the parallel between it and sensuous imagining, explains why Foster should insist for these kinds of experience that the representation in question is of an experiential act. As argued

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi



. . . 

above, to do justice to the experiential character of memory, we need to recognize how subjective aspects of past experience are present in current recall without thereby being realized. So, some experiences involve the embedding of other experiences, and they do so representationally rather than presentationally. Moreover, we cannot do justice to the epistemological status of memory without acknowledging a role for past experience recognized as past being an aspect of current recall. Once we acknowledge that there is a distinctive kind of memorial experience that we enjoy, then in addition we must accept some form of two-level account of experience. And this returns us to the theme raised at the outset: to what extent can we, or must we, trade in questions about the nature, or constitution, of experience when we properly describe its subjective character? Right from when sense-data were first posited as objects of awareness, critics have raised worries about a methodology of deriving substantive conclusions about how things objectively are from considerations deriving just from a first-personal reflection on how things appear. If we take this methodological scruple to an extreme, one may wonder not only whether one can derive from appearances the requirement that objects of awareness exist, a view characteristic of sense-datum theories, but also the absence of any requirement that objects of awareness exist, a claim that is characteristic of intentional theories of perception. And this highlights a question whether there is anything introspectable about conscious experience which recommends that we think of it in representational, or non-representational terms. If we side with Foster in the above discussion of temporal experience, then we do have a reason to suppose that there are such marks of representational experience. According to the two-level account, experiential memory and, hence in turn imagination, must strike us as representational modes of consciousness. And moreover, this representational import is manifest to us in the contrast between being part of the present and, in the case of memory, being real but non-present. In turn, Foster supposes that this provides an argument to show that sense experience itself must be presentational and not representational (Foster, 2000, 127–30). That argument presupposes that the only way to conceive of sense experience as representational is to suppose that what Foster claims holds of imagery holds of sense experience too. That is, that if sense experience is representational, or imagistic, then it involves the conceiving of a type of sensory state. Given this, we get a vicious regress in treating sense experience itself as an instance of imagistic representation: such representation will be one of introspectively conceiving the introspective conceiving, and so on. I doubt, however, that any of Foster’s intended targets would be much moved by this complaint. It requires not only that sense experience be representational in some manner analogous to imagistic conception, but that it be identical in representational character. What should convince us that we should accept such a stringent condition? The reasons we have given above for supposing imagery to involve the embedding of first-order experience draw on characteristics which are distinctive of memory and imagination, but are absent in the case of sense perception. For it is precisely the non-actualization of the subjective features of imagined seeing or feeling which lead us to suppose that acts of sensuous imagining relate to sense experiences of those types representationally.

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

   



The intentionalist will hardly be inclined to accept that it is manifest to us that we are not undergoing a visual experience when we are having a visual experience! At best, Foster could be interpreted as articulating what it takes for a representational experience to be manifestly representational: that the distinctive way in which experience keys the subject to presence in absence is by having subjective aspects which are not realized in the experience but still available to awareness. This may be taken as a sufficient condition of experiential representational character. And, given absence of any viable alternatives, arguably a necessary condition on experience being manifestly representational. However, what Foster ignores is the possibility that his opponent may insist on the representational nature of experience while denying that that feature of the experience is in any way manifest to the subject of experience. So, Foster either needs to show that all representational experience is manifestly representational, or that his account of imagery offers the only possible model of representational consciousness. He ignores, and so can hardly discharge, this additional argumentative burden. Now that is not to say that one could not attempt to develop an argument in the context of weaker assumptions to show that a two-level account of experience requires that the bottom level be relational rather than representational. And elsewhere I have explored considerations in that direction (see Martin, 2001, 2002). But it would be a distraction in the current context to pursue that question further. Let us content ourselves here with exploring the weaker consequences of our discussion. Foster has provided us with a model of how some experience can not only be representational in nature but manifestly so. That conclusion alone is sufficient to challenge the methodological constraints we raised at the outset. On the one hand, we have seen that there are reasons to think of some experiential states as manifestly representational, as the two-level account suggests in the case of experiential memory and sensory imagination. So, the subjective facts of appearance cannot be so etiolated that they do not allow us to draw some distinction between experiences which purport to occur independently of the objects of awareness, and those which do not. On the other hand, the way in which we have spelled out the representational character of memory and then imagination does not offer a good model for how we might think of sense experience as also being representational. Memory and imagination are manifestly representational in the sense that the subject appreciates that the objects and qualities before the mind in undergoing such experience are, at the same time, not present. The subject is not currently having the experience which one would have were they present. If we propose that sense experience is representational, is our conception of representation along the same lines as that we find in the case of imagery? If it was supposed to involve the manifest non-presence of the objects, then it would be puzzling why we should respond to sense experience as we do. When I imagine a red H, I do not act on my immediate environment as if there is such a shape present. But when I see one, or seem to see one, then typically I will. Those who hold that sense experience is representational do not want to deny that from the subject’s perspective, the world is such as to contain the object of experience. So, presumably, the proposal for sense experience must concede that, although the nature of our sense experience allows for the non-existence of its objects, just as our imagery manifestly

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi



. . . 

does, that feature of experience is not itself salient just in our having the experience, in contrast to most imagery. This suggests that we should not think of intentional approaches to sense perception as providing an account principally from the first-personal perspective on our experience, despite their disagreements with sense-datum theories. Rather, intentionalism is offered as a gloss on what makes it the case that one’s experience is the way it is, articulating its qualitative character. On this interpretation, a defender of intentionalism can reasonably claim that intentionalism does not violate the veil of appearances, unlike the sense-datum theory. It does not read off the nature of sense perception from what is introspectively given, but is concerned solely with the facts which ground or constitute the subjective facts that can be so read off. Such an interpretation of intentionalism is not costless. If intentionalism takes such a form, then what distinguishes it from an account which is avowedly revisionary, an account intended as a kind of error theory? Consider a view which concedes that the manner in which sense experience strikes us to be leads us to suppose we stand in some relation to the objects of perception, but that that conception of the nature of our experience must be mistaken (perhaps given reflection on the causal argument). Errortheory intentionalism would then be offered as an account of how it comes to be that we misconceive the character of our perceptual experiences as relational. Most formulations and discussions of intentionalism fail to compare and contrast the theory on offer to some such error theory. And one might remark at this point that what it would take to show that the view is not an error theory is precisely some positive account of how sense experience should be conceived of as representational rather than presentational. Typically, though, no such positive account has been given of what makes us classify sensory experience as a mode of representational experience. Now, one might think that this gap in the formulation of intentional theories of sense perception simply reflects the methodological constraints with which we started: it recognizes the importance of a veil of appearances. If there can be nothing about the subjective facts which forces us to choose between a representational and a non-representational account of some mode of experiencing, then the debates about intentionalism or representational experience must all be carried out at the level of the grounds of subjective facts, and in no way disputes about those subjective facts themselves. However, our discussion of the two-level account of experience shows that this position is untenable. For to make sense of the contrast between memorial experience and sensory experience of the present, we need to recognize ways in which some of our experience is manifestly representational. We can make no principled distinction between the subjective aspects of experience on the one hand, and its nature or ground on the other. One might hope for a veil of appearances to separate neatly these matters. But Foster’s rejection of Russell’s simpler sense-datum theory illustrates for us why there could be no such veil.²¹ ²¹ This chapter is a sequel to Martin (2015). It was originally written for the Modern Workshop on Imagination at New York University, where Amy Kind gave apposite commentary for which I am immensely grateful. Versions of the material were also presented in Turin, Glasgow, Ligerz, Bonn, Paris,

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

   



References Anscombe, G. E. M. (1981) ‘The Reality of the Past’, in The Collected Papers of Gem Anscombe, Vol. 2: Mind and Metaphysics, Oxford: Basil Blackwell, 103–19. Ayer, A. J. (1936) Language, Truth and Logic, London: Victor Gollancz. Ayer, A. J. (1956) The Problem of Knowledge, London: Macmillan. Barnes, W. H. (1965) ‘The Myth of Sense-Data’, in Robert J Swartz (ed.), Perceiving, Sensing, and Knowing. Berkeley, CA: University of California Press, 138–67. Campbell, John (2002) Reference and Consciousness, Oxford: Clarendon Press. Chisholm, Roderick M. (1950) ‘The Theory of Appearing’, in Max Black (ed.), Philosophical Analysis, Upper Saddle River, NJ: Prentice Hall. Cramer-Lemley, Catherine, and Adam Reeves (1992) ‘How Visual Imagery Interferes with Vision’, Psychological Review 99(4): 633–49. Debus, Dorothea (2008) ‘Experiencing the Past: A Relational Account of Recollective Memory’, Dialectica 62(4): 405–32. Debus, Dorothea (2016) ‘Imagination and Memory’, in Amy Kind (ed.), The Routledge Handbook to Imagination, London: Routledge, 135–48. Dennett, D. C. (1969) Content and Consciousness, London: Routledge and Kegan Paul. Dickie, Imogen (2015) Fixing Reference, Oxford: Oxford University Press. Ducasse, C. J. (1942) ‘Moore’s “Refutation of Idealism” ’, in Paul Arthur Schilpp (ed.), The Philosophy of Ge Moore, La Salle, IL: Open Court. Evans, G. (1982) The Varieties of Reference, ed. J. McDowell, Oxford: Clarendon Press. Farrell, Brian (1950) ‘Experience’, Mind 59: 178–98. Foster, John (1982) The Case for Idealism, Boston, MA: Routledge and Kegan Paul. Foster, John (1985) A. J. Ayer, London: Routledge. Foster, John (2000) The Nature of Perception, Oxford: Clarendon Press. Hopkins, Robert (2012) ‘What Perky Did Not Show’, Analysis 72, 3: 439–43. Hume, David (1978) A Treatise of Human Nature, Oxford: Oxford University Press. Kind, Amy (2001) ‘Putting the Image Back in Imagination’, Philosophy and Phenomenological Research 62: 85–109. Malcolm, Norman (1963) ‘Three Lectures on Memory’, in Knowledge and Certainty, New York: Cornell University Press. Martin, C. B. and Deutscher, Max (1966) ‘Remembering’, Philosophical Review 75: 161–96. Martin, M. G. F. (2001) ‘Out of the Past: Episodic Memory as Retained Acquaintance’, in Christoph Hoerl and Teresa McCormack (eds), Time and Memory, Oxford: Clarendon Press. Martin, M. G. F. (2002) ‘The Transparency of Experience’, Mind and Language 17(4): 376–425. Martin, M. G. F. (2015) ‘Old Acquaintance’, Analytic Philosophy 56(1): 1–44. Moore, G. E. (1954–5) ‘Wittgenstein’s Lectures on Experience 1930–1933’, Mind 63, 64. Nagel, T. (1979) ‘What Is It Like to Be a Bat?’, in Mortal Questions, Cambridge: Cambridge University Press, 165–80. Paul, G. A. (1965) ‘Is There a Problem about Sense-Data?’, in Robert J. Swartz (ed.), Perceiving, Sensing and Knowing, Los Angeles: University of California Press, 271–87. Pears, David (1975) ‘Russell’s Theories of Memory’, in Questions in the Philosophy of Mind, London: Duckworth, 224–50.

Austin, Texas. I wish to thank the several audiences for questions and criticisms. I am also grateful to Ian Proops for discussion of Russell on acquaintance, Rory Madden for written comments and discussion of Russell on self-knowledge, and Jackson Kernion for a provocative question about Russell’s theory.

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi



. . . 

Perky, C. W. (1910) ‘An Experimental Study of Imagination’, American Journal of Psychology 21: 422–52. Price, H. H. (1932) Perception, London: Methuen. Proops, Ian (2014) ‘Russellian Acquaintance Revisited’, Journal of the History of Philosophy 52(4): 779–811. Proops, Ian (2015) ‘Certainty, Error and Acquaintance in Russell’s Problems of Philosophy’, in Donavon Wishon and Bernard Linsky (eds), Knowledge, Acquaintance, and Logic, Stanford, CA: CSLI. Russell, Bertrand (1910–11) ‘Knowledge by Acquaintance and Knowledge by Description’, Proceedings of the Aristotelian Society 11: 108–28. Russell, Bertrand (1912) The Problems of Philosophy, ed. G. N. Clark, Gilbert Murray, and G. De Beer, London: Williams and Norgate. Russell, Bertrand (1986a [1917]) ‘Idealism on the Defensive’, in John G. Slater (ed.), The Collected Papers of Bertrand Russell, Vol. 8, London: Routledge, 106–10. Russell, Bertrand (1986b) Theory of Knowledge the 1913 Manuscript, London: Routledge. Russell, Bertrand (1994 [1903]) ‘Points about Denoting’, in Alasdair Urquhart with Albert C Lewis (eds), The Collected Papers of Bertrand Russell, Vol. 4, London: Routledge, 306–13. Schacter, D. L., Addis, D. R., Hassabis, D., Martin, V.C., Spreng, R. N., and Szpunar, K. K. (2012) ‘The Future of Memory: Remembering, Imagining, and the Brain’, Neuron 76(4): 677–94. Segal, S. J. and Nathan, S. (1964) ‘The Perky Effect: Incorporation of an External Stimulus into Imagery Experience under Placebo and Control Conditions’, Perceptual and Motor Skills 18: 385–95. Shoemaker, Sydney (1984) ‘Persons and Their Pasts’, in Identity, Cause and Mind, Cambridge: Cambridge University Press, 19–48. Snowdon, Paul (2010) ‘On the What-It-Is-Like-Ness of Experience’, Southern Journal of Philosophy 48: 8–27. Storm, B. C. and Jobe, T. A. (2012) ‘Remembering the Past and Imagining the Future: Examining the Consequences of Mental Time Travel on Memory’, Memory 20(3): 224–35. Tulving, Endel (1982) Elements of Episodic Memory, Oxford: Clarendon Press. Tye, Michael (1991) The Imagery Debate, Cambridge, MA: MIT Press. Urmson, J. O. (1969) ‘Russell on Acquaintance with the Past’, Philosophical Review lxxxviii: 510–15. Wheeler, Mark A., Stuss, Donald T., and Tulving, Endel (1997) ‘Toward a Theory of Episodic Memory: The Frontal Lobes and Autonoetic Consciousness’, Psychological Bulletin 121(3): 331–54. Williams, Bernard (1978) Descartes: The Project of Pure Enquiry, Harmondsworth: Penguin.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

PART II

Perceptual Experience

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

5 Acquaintance in an Experience of Perception-cum-Action David Woodruff Smith

1. A Paradigm Experience of Acquaintance Acquaintance, we say, is direct awareness. What exactly is this form of awareness? The best way to understand the phenomenon is to analyse a concrete form of activity in which we experience acquaintance within a familiar environment. The paradigm I choose is drawn from everyday experience and assumes a good deal of contextual understanding. Here is a scenario characterized in the first person: while playing a game of tennis, preparing to hit a serve, I pick up a tennis ball, bounce it a couple of times, then gently toss it skyward and, with the form of the modern service motion, swing my racket upward, striking the ball a glancing blow, sending it arcing toward my opponent’s side of the court. Within this scenario we observe, among other things, a complex structure of experience featuring several different forms of acquaintance: • I see this yellow tennis ball on the ground before me. • I bend down and pick it up with my left hand. • I toss it upward with my left hand, toward a point above my head, as I turn my eyes upward toward its controlled movement. • I hit the ball, wielding my racket in my right hand, i.e. springing upward from my bent knees, I swing my racket up and over the ball, striking it a glancing blow imparting spin, thereby executing my serve. And so goes my activity of launching play on the given point, within the context of playing the social game of tennis with my playing partner, my opponent. My activity of retrieving the ball and hitting the serve is a dynamic pattern of movement in spacetime featuring a complex neurophysiological-biomechanical process of bodily movement. Moreover, an intrinsic part of this activity is my awareness of certain aspects of the activity, an awareness that renders those parts of the activity conscious. As indicated, I see this ball on the ground, I pick it up, I toss it, I hit it with my racket in a topspin service motion. In these differing phases of my activity, I experience a direct acquaintance with the ball and eo ipso with my action. All these phenomena transpire just as I experience them: phenomenally, within my David Woodruff Smith, Acquaintance in an Experience of Perception-cum-Action In: Acquaintance: New Essays. Edited by: Jonathan Knowles and Thomas Raleigh, Oxford University Press (2019). © David Woodruff Smith. DOI: 10.1093/oso/9780198803461.003.0006

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



  

stream of embodied conscious perception-cum-action. Each of these forms of experience involves a direct awareness: of the ball as I see it, of the ball as I pick it up, of the ball as I toss it, of the ball as I feel my racket striking it in a glancing blow. Moreover, in each of these phases of my experience of hitting a spin serve, I have an inner awareness of that phase of my experience, whence a direct awareness of my phenomenal, conscious, intentional experience in that phase of my unfolding action. Self-consciousness (in that sense) is already a form of acquaintance, even where present in such different types of experience as seeing a ball and throwing a ball and hitting a ball with a racket. Our task here is to parse and analyse the forms of awareness in this form of perception-cum-action: each forming an experience of acquaintance.

2. Toward a Phenomenology and Ontology of Acquaintance In The Circle of Acquaintance (1989) I sought to develop an analysis of specific forms of phenomenological structure in experiences of direct acquaintance: with perceptual awareness of objects I see before me, with immediate awareness of my own passing experiences, and with empathic awareness of other persons in my presence. The approach was to model these forms of consciousness—forms of intentionality—in light of the semantics of indexical expressions: as I see ‘this tennis ball (actually now here visually before me)’, as I experience an inner awareness of ‘this very experience’, I am now having (say, in seeing ‘this tennis ball’), as I see ‘this person’ addressing me (say, from across the net in a tennis match), another subject of conscious experience. This work was a study in phenomenology, drawing on Husserl’s theory of intentionality and developing a specific analysis with an eye to formal semantic structures. In recent years philosophers of mind have explored anew the characters of phenomenal consciousness, phenomenal intentionality, ‘cognitive’ and ‘agentive’ phenomenology, and embodied consciousness. These phenomena (pardon the virtual puns) were studied extensively by Husserl and then Merleau-Ponty and are now appearing—rediscovered—in new lights in the wake of contemporary philosophy of mind. Here I propose to explore structures of phenomenal intentionality in experiences of acquaintance featuring the ways in which perception and action are interdependent in the structure of perception-cum-action. Thus we feature the paradigm of a familiar experience of seeing and grasping and throwing and hitting a tennis ball. This paradigm serves to bring out complexities of structure, especially the interaction of the phenomenology and the ontology of such a case of perception-cum-action. By the phenomenology of the experience, I mean the phenomenological structure of the lived experience, appraised ‘as’ I experience it from a first-person perspective. By the ontology of the experience, I mean the structure of the experience as it transpires in the wider world beyond the bounds of my own lived phenomenal experience. Of course, the ontology of an activity of perception-cum-action includes

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

  --



the neurophysiology and biomechanics of the bodily movement—and ultimately the algorithmic patterns of neural activity well beyond the subject or agent’s awareness (as well as the ultimate quantum-mechanical-relativistic structure of the movement’s constituent bosons and dark energy and so on). But the range of ontology to be explicated here is much closer to home, closer to the lived experience, part of the form of acquaintance at hand. Specifically, we need to appraise the intentional relation between the lived phenomenological structure (in my stream of consciousness) and its satisfaction in the surrounding world (as in my bodily movement in playing tennis). In the semantic theory of intentionality, the conditions of satisfaction of my activity’s phenomenal intentional content are defined in the ‘logic’ of my consciousness—quite as the truth of a statement or (better) of a thought is captured in the truth conditions of the statement or thought (following the model of truth set forth in logic and formal semantics). Accordingly, I propose to weave together strands of theory focused on intentionality, phenomenality, and embodiment: each playing its role in the structure of acquaintance exemplified in our paradigm case of perception-cum-action. I should note a recent shift in terminology. Classically, phenomenology is defined as the study of varieties of conscious experience, notably including the theory of intentionality, whereas ontology is defined as the study of being or what exists, notably including the theory of mind and body. As philosophy of mind has rediscovered the problem of consciousness, the ‘hard problem’ in the theory of mind, philosophers have spoken of the ‘phenomenology’ of a conscious experience as its phenomenal character, or what it is like to experience that form of consciousness: part of the object of study in the discipline of phenomenology. If you will, phenomenology, the discipline, studies the ‘phenomenology’ of varieties of experience, as the discipline of ontology studies, inter alia, the ‘ontology’ of mental acts, their ontological status, etc. I have followed the newer usage just above in scare quotes, and I shall often speak below of the phenomenological structure or character of perception or action, as opposed to their ontological structures. Where I refer to the discipline of phenomenology, this should be clear from context. (In recent philosophy of mind, key issues relevant to the present study are pursued in many works including: Varela et al., 1993; Siewert, 1998; Smith, 2004; Smith and Thomasson, 2005; Gallagher, 2005; Janzen, 2008; Chalmers, 2010; Hopp, 2011; Bayne and Montague, 2011; Kriegel, 2013, 2015; Searle, 2015; Dahlstrom et al., 2016; Montague, 2016. The role of phenomenological content in perceptual acquaintance is articulated in somewhat different ways in Hopp (2011), in a neoHusserlian model, and Montague (2016), in a neo-Brentanian model. My own understanding of the relations between the phenomenology and the ontology of acquaintance will draw on the reconstruction of Husserl’s complex system in my book, Husserl (Smith, 2013). The present chapter will extend and amplify the basic approach to acquaintance I took in 1989, looking further to phenomenality and embodiment in the paradigm case set out above. I shall assume a model of intentionality that parallels logical models of truth, featuring truth conditions, along lines explored in Smith (2016a.)

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



  

3. Emerging Perspectives on Acquaintance The perennial paradigm of acquaintance is an experience of perception, as in seeing something directly before one. Indeed, the classical terminology was ‘cognitio intuitiva’ in Scotus and Ockham in the Middle Ages, which was rendered ‘Anschauung’ in Kant, Bolzano, and Husserl. Literally, both terms mean seeing into something—an image reflected in our contemporary English terms ‘intuition’ and ‘insight’. Bertrand Russell adapted the English word ‘acquaintance’ to cover the phenomenon at issue. Russell defined acquaintance within a sense-datum model of visual experience: We shall say that we have acquaintance with anything of which we are directly aware, without the intermediary of any process of inference or any knowledge of truths. Thus in the presence of my table I am acquainted with the sense-data that make up the appearance of my table—its colour, shape, hardness, smoothness, etc.; all these are things of which I am immediately conscious when I am seeing and touching my table. (Russell, 1977 [1912], 46)

Russell speaks aptly of acquaintance in a form of perceptual experience, say, wherein ‘I am immediately conscious [of its color, shape, etc.] when I am seeing and touching my table’. However, when I approach my study table, typically I am ‘immediately conscious’ not of its black-lacquer colour or its rectangular-top shape. Rather, I am immediately aware of my table itself, quite oblivious, for the moment, of its elegant black appearance and taking for granted its textured rectangular top even as I touch it. In fact, it takes a considerable shift of my consciousness to be able to see or to feel the pure sensible qualities of the table, its colours and shapes separated from the table I see and touch. A painter like Claude Monet sought to capture seeing just the blue of the water beside his lily pads, and when I view one of his famous paintings I too have to attend carefully to the hue and apparent texture in the painting’s image of the water. I must ‘bracket’ the richer content of my visual experience presenting the bluehued water buoying the lily pads. Russell approached the phenomenon of acquaintance within an epistemological project: My knowledge of the table as a physical object . . . is not direct knowledge . . . There is no state of mind in which we are directly aware of the table; all our knowledge of the table is really knowledge of truths, and the actual thing which is the table is not, strictly speaking, known to us at all. We know a description, and we know that there is just one object to which this description applies, though the object itself is not directly known to us. (Russell, 1977 [1912], 47–8)

However, around the corner from Russell’s epistemological concerns lie phenomenological aspects of intentional experience. Russell’s distinction between ‘knowledge by acquaintance’ and ‘knowledge by description’ rests fundamentally, we should see, on a distinction between two forms of intentional content that lead into types of ‘knowledge of truths’. In this light we may distinguish between intentionality by acquaintance (say, in my seeing ‘this black table before me’) and intentionality by description (say, in my thinking that ‘this is the table I ordered from St Louis’). The fundamental distinction here is

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

  --



not about types or grounds of knowledge of truths. Rather, it is about structures of phenomenal intentional consciousness that are at work in various forms of acquaintance. Accordingly, the paradigm we chose—a case of perception-cumaction—is designed to bring out distinctions in the content of experiences of acquaintance. Epistemic features of my activity of perception-cum-action are a further aspect of acquaintance, concerning how my acquaintance with the ball ties into my knowledge about it. (On relevant distinctions between knowledge— knowing that p—and intentionality, see Smith and McIntyre (1982), and subsequently in Smith (1989).) In an experience of seeing-and-touching a table, we should see, palpably, that Russell was quite wrong in saying, ‘There is no state of mind in which we are directly aware of the table’. Russell was of course in the grip of the then fashionable sensedatum theory. The most compelling counter-argument is simply to experience the state of consciousness in which (in the first person) I am directly aware of the object I am currently seeing-and-touching—the yellow tennis ball or the black table, as the case may be. My intentional consciousness of the tennis ball—in this complex form of phenomenal, embodied, intentional experience—is ‘direct’ in several different senses. First, the experience does not involve inference: I do not consciously see a colour patch and consciously infer the existence of a ball in my presence. Second, the experience does not involve a sense-datum that stands between me and the ball: my experience links me to the ball itself—there is no round patch of yellow ‘between’ me or my experience and the yellow ball itself. Here we need a proper theory of intentionality, distinguishing object and content of perceptual experience. Third, my consciousness of the ball, where veridical, involves a contextual relation to the ball, a relationship wherein I am ‘in the presence of ’ the object of perception (cf. Russell’s words above): I am in the actual presence of the tennis ball itself as I see it and pick it up and serve it. Acquaintance is thus a ‘direct’ phenomenal intentional experience of something in my presence—and this complex structure is brought home by reflection on a familiar form of experience such as that in my activity with the tennis ball. (Compare Fiocco 2017 on the relational form of acquaintance in ‘knowing things in themselves’. I emphasize that the acquaintance relation at issue is both intentional and phenomenal.) In everyday life we do not experience a sequence of sense-datum phenomena, like the still photos that are reeled together in a motion picture projection. Instead, we experience a familiar yet complex activity: from walking along a New York street, amid a bustle of people and rushing of cars; to hammering a nail into a wooden board, amid a carpentry project; to fetching a tennis ball and hitting a serve, amid a game of tennis. The experience of direct awareness, or acquaintance, is most aptly featured in activities like these. Accordingly, we should approach the phenomenon with tools of classical phenomenological analysis exercised by Husserl and Merleau-Ponty. Husserl’s detailed theory of intentionality in perception, set in relation to one’s ‘lived body’, was not known to Russell. And Merleau-Ponty’s nuanced account of perceptual experience, placed in a ‘phenomenal field’ centred on one’s body in everyday action, followed decades later. In these phenomenological analyses—of perceptual experience and the role of bodily awareness in everyday perception—we

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



  

find key contours of a theory of acquaintance as a familiar form of direct awareness of things around us: as when I see this table on which my hands are resting, or when I pick up this tennis ball and hit a serve.

4. Acquaintance as a Phenomenal Intentional Relation Acquaintance, we say, is direct awareness of something: thus, following Husserl, a form of intentionality, a ‘direct’ consciousness-of-something. On Husserl’s analysis, the structure of intentionality is articulated in a complex relationship as follows (cf. Smith 2013): background—subject—act—content——> object—context. The elements in the intentional relationship may be characterized as follows: • • • •

the act of consciousness is a concrete experience; the subject of consciousness is the person experiencing the act; the object of consciousness is that of which the subject is conscious in the act; the content of the act is the ideal meaning content or sense (Sinn) that prescribes ‘what’ the subject is conscious of in the experience; • the background is the ‘horizon’ of implicit meaning conditioning the experience; • the context is the appropriate environment (or ‘Umwelt’) in which the experience is directed toward the relevant object (if you will, the Umwelt as constrained and configured in accord with the background ‘horizon’ of meaning). Husserl—former mathematician and logician—focused on intentional content and its semantic force. The type of ideal meaning (Sinn) at work in intentionality Husserl called ‘noematic content’, or ‘noema’, borrowing the ancient Greek for what is given in mind. Husserl evolved a strong ontology of ideal contents, which aim acts of consciousness toward appropriate objects. Our analysis of acquaintance here need not assume the full ontology Husserl developed, but we shall observe the fundamentals of a Husserlian model of the structure of subject-act-content-object. Within this Husserlian model we may distinguish between an intentional experience and an intentional relation. An experience is intentional insofar as it carries a content and so is aimed as if toward an indicated object: regardless of whether there exists an appropriate object answering to the content. And then, by contrast, an experience is intentionally related to an object insofar as the content of the experience aims at and is accordingly satisfied—semantically—by the object: rendering the experience veridical, or successfully intentionally related to the object. For Husserl, an act of consciousness is assumed to be a phenomenal experience, ‘appearing’ in the subject’s stream of consciousness. And where the experience is veridical, we should say, the intentionality consists in a phenomenal intentional relation to the object, a successful consciousness of the object. Turning to acquaintance per se, we may thus distinguish an acquainting experience from an intentional relation of acquaintance. And accordingly we look to the specific forms of content at work in acquaintance. We define acquaintance itself, then, as a successful or veridical intentional relation. Accordingly, if the content of the

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

  --



acquainting experience is not satisfied by anything in the actual world, as in hallucination, then there is no successful relation of acquaintance between the experience and its ‘intended’ object. Thus, where the phenomenology of acquaintance lies with the acquainting experience and its content, the ontology of acquaintance lies with the successful relation among act, content, and object—a phenomenal intentional relation to the object in the subject’s surrounding world. The logical or semantic force of indexical content is the focus of analysis in The Circle of Acquaintance (Smith, 1989). For instance, where ‘I see this yellow Dunlop tennis ball (actually now before me)’, the indexical content plays a special role in my visual acquaintance with the particular ball actually present before me. Basically, the object satisfies the indexical content in my visual experience precisely because it is properly situated in the context of my visual encounter with the ball. This notion of ‘satisfaction’ derives from Tarski’s famous theory of truth, applied here not to the linguistic expression ‘this tennis ball’, but to the form of experience, or the form of intentional content, we may express with these words: a form of consciousness ‘directed’ toward the particular ball visually before the subject on the occasion of perception. (Cf. Smith (2016a) on the semantic conceptions of truth and intentionality. Searle too has considered perception and its satisfaction conditions akin to indexical expressions: see Searle, 1983, 2015.) Given this basic model of intentionality, then, acquaintance is a contextualized form of intentional relation. For, in paradigm cases, the intentional experience transpires within an environing context where the subject is in direct contact with the object, that is, in a ‘direct’ intentional relation to the object within the context of awareness. In vision the object is directly impacting one’s eyes, while in touch the object is directly contacting one’s hands, and the object is experienced as directly present in that way. Because these two forms of experience are integral parts of a familiar experience like seeing and grasping a ball, this form of experience is a particularly compelling paradigm for the theory of acquaintance. What this paradigm brings to the phenomenology of acquaintance is the role of the body in typical forms of acquaintance. In the first person, I am directly aware of the tennis ball I see and grasp and toss and strike. And what makes the awareness direct, in a paradigmatic way, is the ‘bodily presence’ of the tennis ball to me, and that means in relation to my own body as I experience seeing and grasping and tossing and striking the ball. This paradigm of acquaintance is thus a natural form of embodied consciousness of something directly present in my environs. In this familiar form of everyday experience, then, an object is phenomenally present to me as I see and touch and interact with it. We turn shortly to the special role that the body plays in such an experience of perception-cum-action, but we need to bear in mind that embodied consciousness in perception-cum-action is a form of conscious intentional activity—embodiment does not preclude consciousness. Indeed, acquaintance is, by definition, a form of awareness. And so, we emphasize, the intentional relation that constitutes acquaintance is a phenomenal intentional relation. The intentional relation is itself ‘appearing’ in consciousness. Thus, my seeing and grasping and tossing and striking the tennis ball form a complex phenomenal intentional relation to the ball. If I could be doing all these things

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



  

without conscious awareness, by definition my activity would not constitute a form of acquaintance. My activity then would have no phenomenal character of lived experience even if successful—it would be, in philosopher’s parlance, a totally zombie activity.

5. The Case of Perception-cum-Action: The Role of the Body in Acquaintance Assuming a broadly Husserlian model of acquaintance as a phenomenal intentional relation, we now look to special features of an experience of perception-cum-action— our designated paradigm of acquaintance. This form of experience is a complex dynamic structure within which we may distinguish several distinct ‘acts’ of consciousness, here refining our initial characterization of the experience: • Turning my eyes downward, I see this yellow tennis ball on the ground before me. • With my left hand, I pick up this yellow tennis ball I see on the ground before me. • With my left hand, I toss this ball upward, keeping my eyes on it as it rises. • With my racket in my right hand, I swing my racket upward into the ball, keeping my eye on it until my racket strikes it. Each of these four acts of consciousness yields an instance of acquaintance, assuming the content of the experience is appropriately satisfied by its ‘intended’ object. The first act is a relatively simply visual experience, presenting an object nearby. The next three acts are intentional actions, each more complex than the predecessor—picking up a tennis ball, tossing it to a point above my head, and striking it there with my racket. Each of these actions is aimed at a bodily movement: my grasping the ball, my tossing the ball, my striking it with my racket. The content of my visual perception aims at the ball on the ground before me, while the content of each of my subsequent actions aims at my bodily activity. Okay, that’s an obvious characterization of the flow of experience through these four acts of phenomenal intentional consciousness. But there is more! First, there is a pattern of interdependence among these four ‘acts’. They are all proper parts of the flow of phenomenal intentional activity in my playing tennis, and accordingly each act presupposes each of the other acts in that continuing activity. Second, each of these ‘acts’ involves my body. In order to see the ball properly, I turn my eyes upon it, volitionally moving my eyes in its direction, though my attention lies almost wholly with my impending service action. In order to grasp the ball in my left hand, I look toward it as I reach for it and I feel it in my hand as I grasp it. In order to toss the ball where I want to, I watch it rise out of my hand. And in order to hit the ball successfully with my racket, I watch it as I swing my racket upward and into the ball. Moreover, as I strike the ball, I feel the ball through my racket. In tennis parlance, I experience my racket as an extension of my hand and arm, and so the ball has a certain ‘feel’ as I strike it. In

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

  --



short, as my activity ensues, my visual experience and my tactile experience and my volitional-kinaesthetic experience are thoroughly intertwined. Obviously, these distinguishable phases of perception and action are interdependent parts within the overall structure of my perception-cum-action. In this characterization of my experience readers will hear echoes of Husserl and Merleau-Ponty on the experience of embodiment. In Ideas I (2014 [1913]) Husserl laid out his conception of phenomenology, his theory of intentionality, and his methodology of epoché, framing the discipline in a high-level perspective. However, in Ideas II (1989 [1912]), written conjointly with Ideas I but remaining unpublished, Husserl laid out concrete details including phenomenological analysis of our everyday experience of embodiment. Subsequently, in Phenomenology of Perception (2012 [1945]) Merleau-Ponty developed a rich and impressionistic analysis of the ways in which our everyday perceptual experience is centred and grounded in our experience of our bodies. Accordingly, our paradigm experience of acquaintance, as recounted above, features the essential role of the body in an experience of perception-cum-action. Husserl distinguished one’s ‘lived body’, or Leib, from one’s ‘physical body’, or Körper. These are not two types of body. Rather, my Leib is my body as I experience it, for example, as I walk around and pick up a tennis ball and hit a spin serve. By contrast, my Körper is my body itself, a physiological and neurological organism in nature. My ‘lived’ body is articulated in the phenomenal content of my experience of my body in perception and action, whereas my ‘physical’ body is an object in nature, an integrated system of muscles, bones, organs, and neural pathways largely unknown to me even in principle, that is, until scientific research informs me. Merleau-Ponty amplified on Husserl’s distinction, stressing the role of one’s sense of body in centring one’s ‘phenomenal field’, and noting the meaning or sense (sens) of things in relation to one’s body. Shifting focus, he talks of the physiology of the body, in contrast to its phenomenal role in relation to objects of perceptual consciousness. Assuming a basic Husserlian model of intentionality, then, the distinction between ‘lived body’ and ‘physical body’ is a distinction between content and object of consciousness: between my body as it appears, in experiences of perception and action, and my body in itself. In perception and in action, Husserl says, the object before me is presented as ‘bodily present’, or leibhaftig. Thus, when I see this tennis ball at my feet, or when I reach down and grasp it, I experience the ball as ‘bodily present’, that is, in a proximal relation to my own lived body as I see it, as I touch it, as I grasp it, as I hit it. Here is a core feature in the phenomenon of acquaintance: in perception, in action, in perception-cum-action, the ball itself is experienced as in relation to my body. Thus, the phenomenal intentional content of my experience presents the ball in proximal relation to my body, and correlatively my body in relation to the ball, as I see and grasp and hit the ball. (By etymology, ‘Leib’ derives from ‘Leben’, meaning life, and ‘Haft’ derives from ‘Heft’, meaning ‘handle’, i.e. an act of handling or seizing something. So Husserl’s neologism ‘leibhaftig’ would literally mean the state of an object’s being available in everyday life for grasping or handling in some way, that is, via one’s Leib.)

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



  

This image of ‘bodily presence’ aptly informs the phenomenology and ontology of acquaintance. Thus, the paradigm experience we chose places the subject in a ‘direct’ phenomenal intentional relation to the object of acquaintance. For the indexical content in the experience semantically prescribes just that contextual circumstance in which the object of acquaintance is ‘bodily present’ to the subject in the experience. (Husserl’s mature theory of intentionality is detailed in Husserl, Ideas I (2014 [1913]). Husserl’s distinction between the body as lived, Leib, and the physical body, Körper, is detailed in Husserl, Ideas II (1989 [1912]). Merleau-Ponty’s account of the body’s role in the phenomenal field is developed in Phenomenology of Perception (2012 [1945]). See Smith (2013) for my own reconstruction of Husserl’s model of intentionality and the distinction between Leib and Körper within that model. Cf. Dermot Moran (2016) for a detailed account of Husserl’s notion of the ‘lived body’, or Leib, and Merleau-Ponty’s amplification of the role of the body in everyday experience.)

6. Inner Awareness as a Form of Acquaintance Acquaintance, we say, is direct awareness of something. In perception I am acquainted with objects in my environment. In action I am acquainted with objects with which I am interacting in my environment. And in an experience of perceptioncum-action I am acquainted with a complex situation comprising my activity in my surrounding world, for example, the dynamic situation including this tennis ball and my seeing and grasping and hitting it in a certain way. As these experiences of acquaintance are forms of consciousness, they are phenomenal intentional experiences—and, where successful, they stand in phenomenal intentional relations, as to the tennis ball. Now, these forms of acquainting experience typically include a certain inner awareness of the experience as it transpires or is enacted: a form of ‘selfconsciousness’, or consciousness of the experience itself. Thus, in our paradigm case, I am aware of my experience of perception-cum-action regarding the tennis ball. Importantly, this awareness is a ‘pre-reflective’ form of consciousness—as JeanPaul Sartre emphasized. For this form of inner awareness is an integral part of the given experience, all of which occurs prior to any reflection on the experience, as in the practice of phenomenology or for that matter cognitive neuroscience. In his Psychology from an Empirical Standpoint (2015 [1874]), Franz Brentano proposed that every act of consciousness includes two constituent parts: a presentation of an object and an accompanying presentation of that primary presentation itself. In a simple act of perception, we might say, the act thus consists in a consciousness of the object and an integral consciousness of that consciousness— that is, a perceptual consciousness of the object and an ‘inner consciousness’ of that perceptual consciousness. Brentano’s model of inner consciousness has been pursued recently in light of contemporary philosophy of mind (see Montague, 2016, 2017; Textor, 2017). As rightly argued, in Brentano’s wake, inner consciousness—whatever its precise form—is part of the given act, whereas ‘introspection’ requires a separate further act of ‘observation’ or ‘reflection’ directed upon the first.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

  --



In The Circle of Acquaintance (Smith, 1989), with Brentano in mind, I sought to appraise the form of inner awareness as a form of acquaintance. Thus, through inner awareness of my seeing this yellow tennis ball, I am directly acquainted with my visual experience itself. Again: in a pre-reflective way indigenous to the visual experience, normally involving little or no attention focused on my visual experience itself. In my model, this form of awareness—‘awareness of awareness’ (per Montague, 2016, 2017)—is not itself an intentional relation, but rather a constituent part of the experience achieving perceptual acquaintance with the yellow tennis ball. The exact form of inner awareness we consider below, departing from Brentano’s own model. (Smith (forthcoming) explores the Brentanian roots of my preferred ‘modal’ approach to intentionality and inner awareness.) Treating inner awareness as a form of acquaintance plays into our current story in interesting ways. For, as I see and grasp and hit this tennis ball, I am directly aware of my so doing. My ongoing experience is immediately ‘present’ to me in this integral form of awareness, and that awareness renders the experience conscious, or selfconscious in this way. However, my experience is aimed not at itself, but at the tennis ball I am dealing with. Accordingly, as Sartre famously observed, this form of self-consciousness—consciousness’ consciousness of itself or ‘conscience de soi’—is ‘pre-reflective’. Only in a distinct act of phenomenological reflection do I turn from the tennis ball to my consciousness of the ball. (For details of the phenomenological structure of inner awareness, see Smith (1989, 2004, 2016b). For Sartre’s account of pre-reflective self-consciousness, see (1992 [1943]). Approaching problems of consciousness in recent philosophy of mind, David Chalmers has addressed ways in which we may have ‘acquaintance’ with our own conscious mental states (2010).) This basic awareness of my passing experience does not involve bodily activity per se: say, as in tilting my head and raising my eyes in a pose of quiet reflection. Consider by contrast the experience of sitting in a posture of Vipassana ‘insight’ meditation. In this experience of meditation, I am consciously sitting upright and breathing rhythmically, and this volitional bodily activity facilitates my observing the thoughts, images, and feelings that pass through my stream of consciousness. But this act of meditative observation is distinguishable from the passing act of consciousness that I am observing or ‘witnessing’. By contrast, my observed thought or image or feeling already carries its own intrinsic inner awareness, by virtue of which it is ‘present’ for me in meditation. We might say the observed experience is ‘bodily present’ in my embodied practice of meditating, by virtue of the work of sitting and breathing. Yet is there any way in which a conscious experience is ‘bodily present’ in inner awareness? If we count inner awareness as a form of acquaintance, self-acquaintance within a passing act of consciousness, then we should see that the context of acquaintance is the lived experience itself. We might say the experience is ‘bodily present’ to itself in self-consciousness, where the ‘lived body’—the ‘body’ that feels the experience—is the lived experience itself, and that lived experience is itself the ‘body’ that is ‘present’ to consciousness in inner awareness. (Bracketing, of course, normal physical embodiment of the experience in neurocortical activity.) Thus: in inner awareness I am aware of my experience as ‘bodily present’ within the context of the full structure of the experience itself.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



  

This image of ‘bodily presence’ aligns with the model of (self-)consciousness proposed in The Circle of Acquaintance (Smith, 1989, amplified in 2004 and 2016b). On that model we articulate the formal structure of a simple visual experience as follows: . The left part of this expression articulates the structure of my inner awareness of my visual experience in seeing the ball, while the right part articulates the structure of my visual experience of the ball. The full expression in brackets articulates the full phenomenological structure of the experience. Within that structure, the experience is experienced as ‘present’ to itself—self-conscious or self-intimating—by virtue of that integral part of the experience which consists in inner awareness: , where this structure modifies the further structure, . In this light Husserl’s idiom ‘leibhaftig’, literally ‘bodily’, may be suggestive of a wider sense of ‘presence’. For ‘Leib’ is etymologically linked to ‘leben’, ‘to live’. We might say that inner awareness gives an experience its ‘lively’, or animated, phenomenal character. So inner awareness is experienced not as embodied in the way that visual or tactile perception is experienced as embodied. Rather, in inner awareness an act of consciousness is ‘lived’, it is ‘lively’, ‘bodily’ in a primordial sense. As the ‘body politic’ is a complex of people interacting, so the ‘body experience’ is a complex of phenomenal items interacting, constituting consciousness per se. When Russell launched the philosophical conception of ‘acquaintance’, he restricted perceptual acquaintance to sensory impressions of ‘sense-data’. He likely had in mind Hume’s talk of ‘vividness’ in sensory ‘impressions’, from which the mind must infer more abstract ‘ideas’. In the present analysis, considering phenomenological analyses of perception-cum-action, we find acquaintance in a notably wider range of consciousness.

7. Phenomenality in Acquaintance For much of the twentieth century philosophy of mind pursued models of behaviourism, physicalism, and functionalism. The phenomenon of consciousness seemed to elude the going theories, however, precisely because the phenomenal character of conscious experience resisted reduction to behavioural, neural, and functional or computational properties of the mental states we experience in our own consciousness. In the present analysis of acquaintance, I have assumed there is a proper phenomenal character in experiences of perception, of action, and of perceptioncum-action. Without phenomenality, we stressed, these experiences would not qualify as acquaintance—direct awareness of something—even if successful in reaching intended objects. In recent debates about ‘cognitive phenomenology’, some philosophers of mind have argued that only purely sensory experience has a bona fide phenomenal

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

  --



character, or ‘phenomenology’. Others have argued that experiences of consciously thinking something have a distinctively ‘cognitive’ phenomenal character that outruns purely sensory experience. Others have argued further that experiences of volitional action have a ‘conative’ or ‘agentive’ phenomenal character: what it is like to try to execute an action as an agent—as in throwing a ball. The present analysis of acquaintance is designed, in part, to bring home an appreciation of the phenomenal intentional structure of familiar forms of acquaintance: in perception, in action, and paradigmatically in everyday perception-cum-action. Thus, we are experientially connected with our surrounding world, in acquaintance, precisely insofar as we experience such intentional activities phenomenally—as we are directly aware of things we deal with in perception, in action, in perception-cum-action, far beyond our experience of pure sense-data. (On these issues see Bayne and Montague, 2011; Kriegel, 2013, 2015; Smith, 2016b.) In the preceding analysis we focused on the role of the body in experiences of acquaintance—thus the paradigm of seeing, grasping, and hitting a tennis ball. However, the body is not only sitting at the terminus of a successful relation of acquaintance. Rather, the body is experienced phenomenally but implicitly in, say, our paradigm of acquaintance—I live ‘through’ my body as I see and grasp and hit the tennis ball. In terms Husserlian, my lived body is central to the ‘horizon’ of my consciousness as I interact with the ball in perception-cum-action; in MerleauPonty’s terms, my body is central to the ‘phenomenal field’ in which I see and grasp and hit the ball. My body is indeed the centre of my horizon of consciousness, the phenomenal field within which I experience the ball I see and grasp and strike. All that said, we need to carefully distinguish the phenomenology and the ontology of acquaintance. Indeed, the phenomenological and ontological aspects of acquaintance tend to blur together in a variety of recent programmes, to which we turn in closing.

8. Consciousness in Acquaintance: Embodied, Embedded, Enactive, Extended? In his 1949 classic, The Concept of Mind, Gilbert Ryle parodied Descartes’s metaphysics as the doctrine of ‘the ghost in the machine’ (cf. Ryle 2002, [1949]). Ever since, philosophers of mind have desperately tried to avoid all things sounding ‘Cartesian’. Actually, Ryle himself noted the conception of mind as ‘self-intimating’ in consciousness (158), and accordingly we ought to give Descartes credit toward launching the discipline of phenomenology as the science of consciousness (cf. Smith, 2004, including the essays, ‘The Cogito circa AD 2000’ and ‘Return to Consciousness’). However, the ontology of or in phenomenology is a further matter—beyond dualism, materialism, functionalism, etc. (cf. Smith, 2013). In more recent times, when consciousness could no longer be avoided, philosophers found new ways to weave mind into the world beyond the ‘inner’ Cartesian theatre. Along these lines, mind has been held to be ‘external’ in ‘content’: where the object of perception or thought—as in seeing ‘this ball’ or thinking that ‘this is a Dunlop tennis ball’—has been taken to be a proper part of the ‘content’ of perception or thought.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



  

The essence of intentional mental states has thus been held to lie partly beyond ‘what’s in the head’. Carrying matters of mind further, the very structure of mental activity has been held to be: ‘embodied’ in its very essence, ‘embedded’ in a natural environment, ‘enactive’ even in perception, and ‘extended’ into the information we draw from the surrounding world and store in paper and in silicon. This externalizing perspective has been dubbed the 4e approach to theory of mind. What might we make of these externalized properties of mind itself, from the perspective of the present model of acquaintance? (On the four features characterized in various 4e models, see inter alia: Varela et al. (1993) and Gallagher (2005) on embodied and thus embedded characters of experience; Noé (2004) on enactive characters of perception; and Rowlands (2010) on various extended characters of cognition.) Acquaintance is an instructive phenomenon for the theory of mind. For, on the present model, our direct awareness of things in our immediate environment— notably in perception-cum-action—involves elements of pure phenomenal consciousness, elements of the embodiment of these acts of pure consciousness, and elements of the concrete intentional relation of these acts to things in the context of those acts—all these items taking their place in the Umwelt of consciousness (in Husserl’s apt phrase). As recounted above, the ontology of acquaintance is formed by the structure of a phenomenal intentional relation in which consciousness is in direct contact with things in the surrounding world. In this way, the present model integrates the phenomenology and the ontology of acquaintance. But then the externalizing characters of mind take on a distinctive look, a look that is not radically anti-Cartesian or indeed anti-Husserlian or anti-phenomenological— arguably running against the grain of various externalizing approaches in recent philosophy of mind. In The Circle of Acquaintance I sought a third way between ‘internalist’ and ‘externalist’ models of intentionality, e.g. in perception. The satisfaction conditions for the indexical content in a simple visual experience specify that the act is intentionally related to an object in the context of the visual experience. The intentionality is internalist, then, in that its phenomenological content specifies what the object would be in the appropriate context. Yet the intentionality is externalist in that the proper object is determined by the context of experience, its external circumstance. I called this ‘semantic’ approach to intentionality a context-sensitive internalist approach in the case of perception. Some philosophers have argued for a ‘disjunctivist’ theory of perception, where veridical perception and hallucinatory perception are two ontologically distinct phenomena. This theory of perception in effect follows out the line leading from Brentano into Husserl, to the effect that veridical perception consists in a bona fide relation whereas hallucinatory perception has a character merely as if relational or merely relation-like. The semantic approach to intentionality obviates a strict disjunctivist model. For every intentional perceptual experience has a content which if satisfied yields a bona fide intentional relation to the object perceived, but which if not satisfied yields only an intentional character but no intentional relation to an existing object. If you will, the ontological difference between intentional character and intentional relation yields a disjunctive structure in perception without dividing

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

  --



perception per se into two unrelated phenomena. (Cf. Soteriou (2016) on the disjunctivist theory of perception.) Different philosophers have outlined different forms of ‘embodied’ or ‘embedded’ or ‘enactive’ or ‘extended’ structure for a mental activity, but one version is closer than most to the present approach. Mark Rowlands (2010) begins with a basic account of intentionality that distinguishes act, content, and object. Cognition, Rowlands argues, is achieved as the mental act is related to an object in such a way that the mental act is ‘extended’ into the surrounding world, as it reaches its object via information linked to the object itself. (Cf. Varela et al. (1993) on embodiment in a Merleau-Pontian approach, Gallagher (2005) on embodiment, Noë (2004) on the enactive character of perception, and Rowlands (2016) on 4e models differing from his own.) In the course of our study of acquaintance, with eyes toward the case of perception-cum-action, we see specific ways in which acquaintance essentially involves consciousness in embodiment, in embedment within the surrounding world, in enactive dependence between perception and action, and in intentional extension to the actual object of perception and action. However, in our model above, we insist on carefully distinguishing the phenomenology and the ontology of acquaintance. We cannot respect the interdependence of the phenomenological and the ontological in acquaintance, I believe, without a careful distinction among subject, act, content, and object of acquaintance. The semantic model of intentionality facilitates this distinction and reveals very different roles of the subject, the subject’s body, and the object of direct awareness—all within the lived context of experience. Such is the complex structure of acquaintance.

References Bayne, Tim and Montague, Michelle (eds) (2011) Cognitive Phenomenology, Oxford: Oxford University Press. Brentano, Franz (2015 [1874]) Psychology from an Empirical Standpoint, London: Routledge. Chalmers, David J. (2010) The Character of Consciousness, Oxford: Oxford University Press. Dahlstrom, Daniel O., Elpidorou, AndreasHopp, and Walter (eds) (2016) Philosophy of Mind and Phenomenology: Conceptual and Empirical Approaches, New York: Routledge. Fiocco, M. Oreste (2017) ‘Knowing Things in Themselves: Mind, Brentano and Acquaintance’, Grazer Philosophische Studien, 94(3), 332–58. Gallagher, Shaun (2005) How the Body Shapes the Mind, Oxford: Oxford University Press. Hopp, Walter (2011) Perception and Knowledge: A Phenomenological Account, Cambridge: Cambridge University Press. Husserl, Edmund (1989 [1912]) Ideas II, Ideas pertaining to a Pure Phenomenology and to a Phenomenological Philosophy, Second Book: Studies in the Phenomenology of Constitution, trans. Richard Rojcewicz and André Schuwer, Dordrecht: Kluwer Academic Publishers. Husserl, Edmund (2014 [1913]) Ideas I, Ideas for a Pure Phenomenology and Phenomenological Philosophy. First Book: General Introduction to Pure Phenomenology, trans. Daniel O. Dahlstrom, Indianapolis, IN: Hackett Publishing. Janzen, Greg (2008) The Reflexive Nature of Consciousness, Amsterdam: John Benjamins. Kriegel, Uriah (ed.) (2013) Phenomenal Intentionality, Oxford: Oxford University Press. Kriegel, Uriah (2015) The Varieties of Consciousness, Oxford: Oxford University Press.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



  

Merleau-Ponty, Maurice (2012 [1945]) Phenomenology of Perception, trans. Donald A. Landes. London: Routledge. Montague, Michelle (2016) The Given: Experience and Its Content, Oxford: Oxford University Press. Montague, Michelle (2017) ‘What Kind of Awareness Is Awareness of Awareness?’, Grazer Philosophische Studien, 94(3), 359–80. Moran, Dermot (2016) ‘Lived Body, Intercorporeality, Intersubjectivity: The Body as a Phenomenological Theme’, in Daniel O. Dahlstrom, Andreas Elipdorou, and Walter Hopp (eds), Philosophy of Mind and Phenomenology: Conceptual and Empirical Approaches, New York: Routledge, 57–78. Noë, Alva (2004) Action in Perception, Cambridge, MA: MIT Press. Rowlands, Mark (2010) The New Science of the Mind: From Extended Mind to Embodied Phenomenology, Cambridge, MA: MIT Press. Rowlands, Mark (2016) ‘Bringing Philosophy Back: 4e Cognition and the Argument from Phenomenology’, in Daniel O. Dahlstrom, Andreas Elipdorou, and Walter Hopp (eds), Philosophy of Mind and Phenomenology: Conceptual and Empirical Approaches, New York: Routldege, 310–25. Russell, Bertrand (1912) The Problems of Philosophy, Oxford: Oxford University Press. Ryle, Gilbert (2002 [1949]) The Concept of Mind, Chicago, IL: University of Chicago Press. Sartre, Jean-Paul (1992 [1943]) Being and Nothingness, trans. Hazel E. Barnes, New York: Washington Square Press. Searle, John R. (1983) Intentionality, Cambridge: Cambridge University Press. Searle, John R. (2015) Seeing Things as They Are: A Theory of Perception, Oxford: Oxford University Press. Siewert, Charles (1998) The Significance of Consciousness, Princeton, NJ: Princeton University Press. Smith, David Woodruff (1989) The Circle of Acquaintance: Perception, Consciousness, and Empathy, Dordrecht: Kluwer Academic Publishers. Smith, David Woodruff (2004) Mind World: Essays in Phenomenology and Ontology, Cambridge: Cambridge University Press. Smith, David Woodruff (2013) Husserl, 2nd edn, London: Routledge. Smith, David Woodruff (2016a) ‘Truth and Epoché: The Semantic Conception of Truth in Phenomenology’, in Jeffrey A. Bell, Andrew Cutrofello, and Paul Livingston (eds), Beyond the Analytic-Continental Divide: Pluralist Philosophy in the Twenty-First Century, London: Routledge, 111–28. Smith, David Woodruff (2016b) ‘Cognitive Phenomenology’, in Daniel O. Dahlstrom, Andreas Elipdorou, and Walter Hopp (eds), Philosophy of Mind and Phenomenology: Conceptual and Empirical Approaches, New York: Routldege, 15–35. Smith, David Woodruff (forthcoming) ‘Descriptive Psychology and Phenomenology: From Brentano to Husserl to the Logic of Consciousness’. Smith, David Woodruff and McIntyre, Ronald (1982) Husserl and Intentionality: A Study of Mind, Meaning, and Language, Dordrecht: D. Reidel Publishing. Smith, David Woodruff and Thomasson, Amie L. (eds) (2005) Phenomenology and Philosophy of Mind, Oxford: Oxford University Press. Soteriou, Matthew (2016) ‘The Disjunctive Theory of Perception’, Stanford Encyclopedia of Philosophy (Summer), ed. Edward N. Zalta ed.), . Textor, Mark (2017) Brentano’s Mind, Oxford: Oxford University Press. Varela, Francisco J., Thompson, EvanRosch, and Eleanor (1993) The Embodied Mind, Cambridge, MA: MIT Press.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

6 Dreaming, Phenomenal Character, and Acquaintance Tom Stoneham

1. Naïve-Realism, Acquaintance, and Phenomenal Character Naïve-realism about perception is the view that—when all is going well¹—perception directly acquaints us with objects in the external world. It is usually thought to deserve the title ‘naïve’ because it says that perception is—at least when all is going well—just how it seems to the non-philosophical subject. It is a form of realism because it claims that the objects we are acquainted with in perception are not mental, are not within the perceiver’s mind, but exist outwith the subject. That naïve-realism actually captures the ‘non-philosophical’ view of perception can be, and has been, challenged. For example, Dummett (1979, 24–9) has argued that an ordinary perceiver would notice various effects of perspective and observation conditions which would lead her to realize that we are not directly acquainted with the external world but that our perceptual experience, that is how we perceive the environment,² is the causal product of many factors. Thus the non-philosophical view is not that we are simply acquainted with reality in perception. Furthermore, most phenomenological arguments for naïve-realism work by inviting philosophically sophisticated subjects to engage in careful introspection (Martin, 2002). It does not seem that acquaintance accounts are usefully called ‘naïve’ and it would be better to reserve the title for the Berkeley-Hume view that in all cases things are as they seem and seem as they are, not in virtue of the controversial claim that this captures commonsense (on that question Berkeley and

¹ I shall use the expression ‘when all is going well’ as a neutral way of picking out cases of perception which are not illusory or delusory or otherwise inclined to mislead an unsuspecting subject. ² The noun phrase ‘perceptual experience’ is used by philosophers in many different ways. I am restricting it throughout this chapter to be a label for answers to the question ‘How does S perceive the world to be?’ While every perceptual experience in this sense is always some specific subject’s experience, it does not follow that it is an event in or state of that subject. To take a description of how someone perceives the world to also be a description of how things are with her is to make a philosophical move (Stoneham, 2008). Tom Stoneham, Dreaming, Phenomenal Character, and Acquaintance In: Acquaintance: New Essays. Edited by: Jonathan Knowles and Thomas Raleigh, Oxford University Press (2019). © Tom Stoneham. DOI: 10.1093/oso/9780198803461.003.0007

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

Hume disagreed), but in virtue of its rejection of the first step in construction of a theory of perception, namely the appearance–reality distinction. Such a genuinely ‘naïve’ view of perception would not be able to make sense of the concept of a veridical or correct perception because there would be nothing on the subject side of the occurrence of perceiving which could be said to line up correctly with the world. There would remain a question whether such an account of perception could be a form of realism, whether it could be consistent with regard to these things that seem as they are and are as they seem to be external to the mind, and this is also something that Berkeley and Hume disagreed upon. However, we will keep to contemporary usage and use ‘naïve-realism’ to denote the view that when all is going well, perception directly acquaints us with objects and properties in the external world. Instead of following Berkeley and Hume, contemporary naïve-realists accept an appearance–reality distinction, but not in all cases of perceptual experience. Thus in good cases—when all is going well—the appearance just is reality, or perhaps we should say, reality is what appears to the subject. That is acquaintance. But in bad cases—cases such as illusions or hallucinations where our perceptual experiences—how we perceive the world to be, do not match up with reality, contemporary naïve-realists offer a distinction between appearance and reality. Thus the class of perceptual experiences is divided in two for the purposes of explanation: in the good cases, how things seem to us is explained by how they are; but in the bad cases, how they seem needs a different explanation. This view is now standardly called ‘disjunctivism’ (Hinton, 1967; Martin, 2004; Fish, 2009). Many disjunctivists regard their biggest challenge to be the ‘screening-off argument’ (e.g. Martin, 2004, 2006; Farkas, 2006): (1) For every good case, there is a possible bad case with the same proximal cause and the same phenomenal character of experience. (2) In the bad case, that proximal cause is a sufficient explanation of the phenomenal character. (3) So in the good case where the same proximal cause is present, that is a sufficient explanation of the phenomenal character. (4) So in the good case, the object with which we are allegedly acquainted, being a distal cause, plays no role in explaining the phenomenal character. The disjunctivist typically blocks this argument by rejecting premise (1), and there are many ways to do that but most turn upon denying that the bad case has the same, rather than merely indistinguishable, phenomenal character. Rather than discuss these attempts, I want to note that the force of the argument in the first place relies upon the disjunctivist accepting that her view is an attempt to give an explanation of phenomenal character in the good case by reference to the external object with which we are acquainted. If the disjunctivist was not interested in explaining phenomenal character by acquaintance, the argument would not be an objection to her view. Thus the seriousness with which a disjunctivist takes the screening-off argument is an indicator of the value she places upon explaining phenomenal character by acquaintance.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

,  ,  



2. Merely Successful vs. Good Explanations Rather than respond to the screening-off argument directly,³ I want to raise a slightly different question, which is whether acquaintance is a good explanation of phenomenal character. I do not want to cast doubt on the thought that acquaintance succeeds in explaining phenomenal character, or is at least an essential element of a successful explanation of phenomenal character, in the good cases. Rather I want to ask about the value of this explanation. How can there be successful explanations which are not good explanations? One way might be an explanation which invokes a particular feature of the situation when in fact explanation by reference to a more general kind which that feature instantiated would also be possible. For example, an explanation of a large harvest which referred to the high spring rainfall might be successful, but it is less valuable than one which referred instead to the amount of water the plants received, since the method of delivery of that water is irrelevant and it would be possible to replicate the effect artificially.⁴ Here, a deeper explanation would unify the naturally and artificially irrigated crops. Another way would be an explanation which made little progress with the larger explanatory project. For example, if we are seeking to explain a famine and we discover the crop failed in one village because a corrupt supplier had sold them sterile seeds, then while that is a successful explanation of that crop failure, it does not help us explain the general famine, especially if we think the corruption was localized. We have explained one crop failure but made almost no progress with the larger project of explaining the famine. A third example might be the successful explanation of a local variation. For an example here, consider again the famine but suppose that the crop in one village had not failed, but for geographic reasons which make that village unlike any other. We have a good explanation of why this crop did not fail, but this is of little interest when we are trying to prevent future famines, because the causes of success cannot be replicated elsewhere. Returning to the explanation of the phenomenal character of perception by reference to acquaintance, there are three reasons one might doubt that this explanation is of much value, each corresponding to the three examples just given: (1) The similarity of phenomenal character in the good and bad cases suggests that a deeper explanation would unify the two. (2) If good cases are sufficiently uncommon, then the bulk of the explanatory work for a theory of consciousness is still to be done. (3) If good cases are relatively uncommon, then acquaintance is not a part of our subjective mental lives which we place a high value on as subjects. Considering (1), the disjunctivist might point out that it is entirely possible that no deeper explanation will be forthcoming, and this seems to underlie the assertion many make that disjunctivism is a claim about fundamental metaphysical kinds (Martin, 2002; Snowden, 2005). For if the good and the bad cases of perceiving are of

³ The conclusion of this chapter amounts to a denial of (1) for the case of dreaming, so the argument will receive a partial response. ⁴ That artificial irrigation is as good for plant growth as natural humectation is so obvious to us that the point is easily obscured. But it is not hard to imagine a society where that was not so obvious.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

different fundamental kinds, then there would be good reason to doubt a deeper explanation will be forthcoming. But this approach requires considerable selfconfidence with respect to metaphysics: a more humble approach might think that our natural kinds are tentative at best until we have done the explanatory work. Anyone who insisted that trees and bushes were of different fundamental kinds would now look ridiculous, and someone who insisted that we could classify whales as fish without knowing how they reproduced would be operating with a pre-modern biology. Thus the explanatory question and the question of fundamental kinds go together, and the disjunctivist’s deliberate closing off of the possibility of a deeper explanation looks like a comparative weakness of the account. Considering (2), the disjunctivist might argue that even if the good cases are uncommon, they are explanatorily and metaphysically prior: there could only be bad cases because there are good cases and the character of those bad cases depends upon the character of the correlative good cases. The classical empiricist account of imagination has this form, though it is not disjunctivist about the resulting phenomenal characters: for example, Hume tells us that we can only form the idea of a golden mountain by perceiving gold and perceiving mountains and (somehow) combining the memories or copies of these perceptions, so the perceptions are explanatorily and metaphysically prior to the imaginings. If the disjunctivist’s account of the bad cases has a structure somewhat like this, then they may argue that (2) is not such a weakness after all. However, that is to slightly misread the point. It remains the case that the bulk of the explanatory work is not done by acquaintance but by whatever mechanism is in place to create phenomenal character in the absence of acquaintance, even if that mechanism requires prior acquaintance. In Hume’s theory, for example, what does the bulk of the explanatory work is the mysterious process by which impressions are ‘copied’ into ideas which have similar phenomenal character, and then those ideas are stored—non-phenomenally—to be mysteriously reactivated in consciousness when we try to imagine the golden mountain. Finally, turning to (3), the disjunctivist might argue that uncommon experiences might still be of great subjective value. For example, if we are stuck in the Matrix but very occasionally ‘let out’ to be reacquainted with friends, lovers, or favourite foods, then we might place a very high value on those infrequent moments of acquaintance.⁵ But for that to apply to our actual situation, it would have to be the case that when our attention is drawn to how uncommon the good cases are, we align our values with the good cases rather than some mix of good and bad. It is far from obvious that this will happen, just as it is very hard to pin down exactly what is so bad about living in the Matrix (Pryor, 2005) or to persuade someone to change their mind about entering Nozick’s Experience Machine. Furthermore, there remains the fact that insofar as we place a high value on any so-called ‘bad experience’—such as a wonderful dream—then acquaintance cannot explain what it is about that experience which we find so valuable. (1)–(3) are not being offered as an objection to disjunctivism, but as a motivation for the thought that what the disjuntivist has to say about bad cases affects not the

⁵ I owe this example to Dorothea Debus, who has pushed me hard on all of (1)–(3).

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

,  ,  



success of her explanation of phenomenal character in the good cases but the value of that explanation. If we think that acquaintance is a good explanation of phenomenal character in the cases where we can use it, then we should take seriously the challenges posed by (1)–(3), for if we can deal with them, then we have made that good explanation even more valuable. It is possible to give a bite-sized example of how this might work. One ‘bad case’ which the naïve-realist has to deal with is negative after-images: these occur when you fixate on a bright, predominantly single-coloured object for approximately thirty seconds and then, upon looking at a pale surface immediately afterwards, experience a colour patch of the opposite colour, apparently floating above the surface of the perceived objects and moving with your direction of gaze. Can we explain this by acquaintance? Most contemporary naïve-realists will be inclined to say not, for when I turn my eyes from the green screen to the white paper and see a magenta patch, there is no magenta for me to be acquainted with.⁶ Consequently, the phenomenal character of this experience needs to be explained some other way, and typically it will be explained as created by the overstimulation of the photo-receptors. But if overstimulation of photo-receptors is sufficient to create phenomenal character, and if these disjunctivist naïve-realists are also tempted to say that understimulation of the same photo-receptors causes ‘brain grey’ experiences, then the explanation of the good case by reference to acquaintance looks like a very special case where the photoreceptors have just the right level of stimulation. Objection (2) is pressing. An alternative approach to after-images would be to propose that the overstimulation of the photo-receptors causes an insensitivity to light of certain wavelengths (Stoneham, 2012). If staring at the green screen causes an insensitivity to green light, then if I look at a source reflecting or emitting predominantly white light, it will look magenta. Imagine looking at a piece of white paper in sunlight: there are three things we can do to make you see magenta. First, we could filter the sunlight with a magenta filter, which will block the green part of the spectrum. Second, we could colour the paper magenta, which will prevent it reflecting light in the green part of the spectrum. Third, we could make the eye less sensitive to green light, perhaps by overstimulating the photo-receptors for green light. All that differs between these three cases is where in the causal story the green light is removed from the process. In all three cases, the magenta was there before the filtering and the change merely makes it visible by blocking the green parts of the spectrum. If this were the right form of explanation, the naïve-realist avoids disjunctivism in their explanation of negative after-images because the phenomenal character is explained by acquaintance with the magenta elements of the white light, an acquaintance made possible by the blocking of the green elements. In the example of negative after-images, the alternative approach extended the explanation of phenomenal character by acquaintance to what initially appears to be a ‘bad case’ of the sort that forces the naïve-realist into disjunctivism. Another ⁶ A good example of this is Philips (2013), who gives a very thorough and careful account of the phenomenology of after-image experiences and why these do not support a ‘sensation’ account, but then proposes that they are ‘illusory presentations [hallucinations?] of pure visibilia’ (427), ensuring they fall squarely into the list of bad cases.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

way to block the move from naïve-realism to disjunctivism would be to restrict the explananda for the theory of phenomenal character: if the alleged bad case where the phenomenal character cannot be explained by acquaintance does not in fact involve phenomenal character, then it provides no reason to be a disjunctivist. For an example, consider the problem of ‘cognitive phenomenology’, which is the claim that ‘some cognitive states put one in phenomenal states for which no wholly sensory states suffice’ (Chudnoff, 2015, 15). If we accept this and also that acquaintance is the preferred explanation for the phenomenal character of sensory states, then it seems we have a whole class of states with phenomenal character which cannot be explained by acquaintance, since they involve no acquaintance and do not depend upon the phenomenal character of acquaintance. Disjunctivism seems the only option, but the lack of unity makes objection (1) pressing. However, the existence of cognitive phenomenology is contested, so if a contemporary naïve-realist was in a position to plausibly deny that there is such a thing as cognitive phenomenology, that would, given objections (1) to (3), make their explanation of the phenomenal character of perception a little bit better. In what follows, we consider the issue of explaining the phenomenal character of dreaming and show how the contemporary naïve-realist might avoid disjunctivism by following Norman Malcolm and Daniel Dennett in denying the standard model of dreams and dreaming on which dreaming is a phenomenally conscious nocturnal experience remembered upon wakening.

3. The Standard Model of Dreaming With these thoughts in mind, I turn to the question of dreams. Apparently most people dream at least four to six times per night and 95–99 per cent of dreams are forgotten ‘for the very ordinary reason that we sleep right through them and aren’t paying attention to remembering anything’ (Schneider and Domhoff, 2017). Quantitatively, then, dreams are a significant part of our mental lives. Furthermore, the dreams we do recall have significant value for many people: most people have strong positive and negative emotional responses to many of their dreams; many people enjoy telling their dreams; poets and painters draw upon them for inspiration; and to not dream at all, or at least to not remember any dreams, is seen as missing out on an important part of human life—so much so that pet owners are very ready to project dreaming onto their animals. But it seems that acquaintance can play no role in the explanation of the phenomenal character of dreaming because when dreaming we are asleep and acquainted neither with the fantastical objects of our dreams nor even our mundane environment. Thus dreams present a significant problem for the disjunctivist. The standard model of dreaming identifies a four-stage process, with stages 1 and 2 happening during sleep but 3 and 4 during wakefulness: (1) (2) (3) (4)

Dream experiences. Encoding in memory. Recall from memory. Report.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

,  ,  

rēvevécu

sommeil

rēvesouvenir

encodage stockage rappel



rēverécit

traduction éveil

Figure 6.1 Standard model of dreaming Source: Guénolé and Nicolas, 2010, 194

For example, we find the model in Figure 6.1 used explicitly in neuroscientific research on dreaming.⁷ For clarity, let us call stage 1, the conscious experiences which happen during sleep, ‘dreaming’, and stage 4, the public report of a dream through some act of communication, whether directed at others or oneself, ‘dreams’. This distinction corresponds to David Foulkes’s distinction between Type-A and Type-B dreams: • Type-A dreams (= dreaming): ‘the involuntary conscious experience of mentation during sleep’; • Type-B dreams (= dreams): ‘a person’s everyday account of an experience, described with greater or lesser accuracy and with greater or lesser conviction, as having occurred during sleep’ (Foulkes, 1999, 35–6). The standard model is almost universally accepted by both dreamers and dream researchers for three reasons. First, and most influentially, stage 3 feels like memory of prior experiences and if it is such a memory, then stages 1 and 2 must exist. Second, between stages 2 and 3 there is often an experience of awaking with a strong emotion and stage 3 sometimes—but not always—provides an explanation of that emotion in terms of a stage 1 experience. Third, there seems no other way to explain the ubiquity of stage 4, namely dream reporting. However, the model is not subject to direct testing because there is no way to access dreaming apart from subjective reports based on apparent memory (cf. Section 5).⁸ The standard model has various problems. First, stage 1, the dreaming, is an entirely private experience. Not only are there good philosophical concerns about the intelligibility of such entirely private experiences,⁹ but this seems to require us to

⁷ Note that this model has a fifth stage, translation from recall to report, to explain some of the phenomena discussed below. ⁸ Foulkes adds to his definition of Type-A dreams, or dreaming, that ‘[o]ur best, and perhaps only reliable, knowledge of this mentation comes from a person’s report recorded immediately following abrupt arousal from the experience in question’ (1999, 35). ⁹ While it would be misleading to say there is consensus, objections to entirely private experiences have been raised by philosophers working in such different twentieth-century traditions as Wittgenstein, Dennett, Churchland, McDowell, and Shoemaker that it is reasonable to conclude that private experiences are highly controversial.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

explain the content of dreams in terms of some internal process, be it the Freudian unconscious or a sub-personal tidying up of memories and emotions. None of those explanations succeed in explaining most, let alone all, of what is actually reported by subjects (Botman and Crovitz, 1989; Harlow and Roll, 1992; Hartmann, 1968; Marquardt et al., 1996; Nielsen and Powell, 1992). Second, there are phenomena which do not fit into the standard model, such as precognitive dreams (e.g. Maury’s dream (Freud, 1976 [1900], 87)) and time compression in dreams which seem too long for the time spent asleep.¹⁰ Third, while stage 3 feels like memory to the subject, it has many features which are different from memory of waking experiences, such as fading very quickly, being resistant to recollective effort, and not being strengthened at second recall. There are many reported phenomena which may strike the subject as being the exercise of a cognitive ability which just cannot be (e.g. déjà vu, clairvoyance, telepathy) and in such cases good science overrides subjective conviction, so we do not need to take the fact it feels like memory as decisive. Fourth, large amounts of research have shown that episodic memory is notoriously unreliable and that false memories can be created at both the encoding and recall stage, with the prime determinants being the social and linguistic context in which the ‘recall’ is triggered (Loftus and Ketcham, 1992, 1996; Loftus and Palmer, 1996; Hyman et al., 1995; Kassin and Kiechel, 1996; Rubin, 1999). There is plenty of evidence that both labbased and diary-based dream research is heavily affected by these factors, which distort memory. And yet, despite all these problems, the standard model remains widely accepted simply because there is no plausible alternative. Another feature of the standard model which is rarely commented upon, because it is so deeply embedded in our thinking about dreams, is how it attributes the responsibility for dream content. For much of recorded human history, it seems that dreams were regarded as either symptoms of pathology or as externally caused (i.e. ‘visions’ with good or evil supernatural causes). Thus the natural response to disturbing dreams would be either ‘What is happening to me? Am I unwell?’¹¹ or ‘What will happen to me (or sometimes others)?’, and perhaps where the dream appeared to be a vision to do with other people, ‘Why have I been given this dream?’ But as the attraction of supernatural explanations waned, the mechanistic and physiological explanations seemed inadequate to capture the apparent meaning of dreams, and so the intellectual context was primed for Freud’s appeal to subconscious sources in the dreamer’s mind. The psychological character of this explanation, in terms of hidden desires, fears, and urges, enabled dreams to have meaning and significance. But they also added an additional feature to the standard model: the content of stage 1, the dreaming, is determined by something internal to the subject. And even those dream researchers who reject Freud’s theory maintain this additional element and seek explanations for dream content within the dreamer’s psychology (e.g. theories of the function of dreams in terms of processing memories or problem solving). This leads to extraordinary debates, such as whether O. J. Simpson’s dream ¹⁰ Freud mentions both these problems (1976 [1900], ch. 1) but sticks to the standard model. ¹¹ Hippocrates’ Regimen IV on dreams is a good example of treating dreams as diagnostic of physical health that has had a profound cultural influence on the West. It is worth noting that the cures he offers involve emetics and special diets.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

,  ,  



of killing his wife was admissible evidence of intention in his trial for murder.¹² More generally, it can lead to anxiety, especially in adolescents, as dreamers subject to unpleasant dreams become concerned that this may reflect a hidden or suppressed truth about their personality.

4. Dreams without Dreaming Most of the problems with the standard model derive from the fact that it takes the content of stages 3 and 4, what is recalled and what is reported, to correlate with the content of the nocturnal dreaming experiences, stage 1: dreams require there to have been correlated dreaming. To find a plausible alternative, we must drop this assumption. The alternative I will explore has been proposed in various forms (e.g. Malcolm, 1959; Dennett, 1976) but is usually traced back to Goblot (1896), who responded to Maury’s dream by proposing that dreams are made up very quickly while we are waking up. Using my terminology, we can say that this is a version of a general position that we have dreams without dreaming. While Goblot’s proposal may work for the specific case of a dream report which appears to be prompted by an event immediately prior to wakening, we need a more general account to explain dream recollection and reporting as a widespread phenomenon. Consider the fact that since at least the time of Galen (130–210 CE), who thought that eating cabbage gave us especially bad dreams,¹³ people have believed that there are correlations between what you dream and what you eat. How could a food cause dreaming with a certain type of content? What mechanism could relate eating cabbage or cheese to reporting bad dreams? Notice how different this is from the claim that traumatic experiences cause bad dreams—there the mechanism is not quite so magical because it appears to be mediated by an intelligible cognitive process (Domhoff, 1996; Hartmann, 1998; Kramer, 2000; Zadra, 1996): if dream content is partially caused by your psychological state prior to and during sleep, then that appears to be an example of a familiar causal pathway, namely one psychological state caused by another and the fact that the effect occurs during sleep has no particular significance. We also have familiar ingestion-to-psychological-effect causal pathways with psychotropic substances, but these do not apply here because neither cabbage nor cheese appear to be psychotropic.¹⁴ And to suggest that these foods are in fact psychotropic but that they only affect the sleeping mind—perhaps only the REM-sleeping mind—is no explanation until we have a mechanism by which the psychotropic effects could be sensitive to the sleeping state of the subject. So the longestablished correlation between what you eat and what you dream is, if true, puzzling on the standard model. ¹² http://articles.latimes.com/1995-02-02/news/mn-27322_1_dream-expert (4 April 2017). ¹³ Burton cites Galen on cabbage (1989, I.II.II.I) as well as reporting the effects of eating hare and pulses, and he also notes that sleeping in the daytime ‘upon a full stomach . . . increaseth fearful dreams’. ¹⁴ Care is needed here because some foods do cause involuntary and unpleasant images in some people even while awake, possibly because their high tyramine content causes a release of catecholamines. But unless we have reason to think this is a mechanism which is common in sleeping subjects and rare in waking subjects, it will not explain the ubiquity of the alleged correlation between foods and bad dreams. Thanks to Louise Moody for pointing out this phenomenon.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

The solution to the puzzle is surely that certain food types can cause digestive discomfort, which causes familiar, unpleasant bodily sensations.¹⁵ But how would an unpleasant bodily sensation cause dreaming experiences of bad things? Perhaps it could be incorporated into the dream in the way that perceptions of external events such as loud noises are sometimes incorporated into dreams. However, then we would expect the bad dreams caused by food to always involve the stomach—which they don’t. Thus on the standard model we seem to have a process like this:

bodily sensation

digestion

dreaming

encoded in memory

recalled

dream report

with no apparent correlation between the content of the dreaming and the bodily sensation, but then strong correlations between the content of the dreaming and what is recalled and reported. This is an inadequate explanation because while there is only a weak correlation between the bodily sensation and the dream content, there is some: unpleasant bodily sensations like indigestion produce unpleasant dreams. How are we to explain that? After all, diurnal indigestion has no such effects. By recognizing the role of physiological processes in dream formation, this variant on the standard model gets a partial explanation of the phenomenon, but there is more work to be done. Here is an alternative: (1) the encoding in memory of the bodily sensation of indigestion—a sensation which has a phenomenal character explained by treating interoception as a form of acquaintance with the body—is distorted by the fact that you are asleep at the time and not good at recognizing bodily sensations for what they are; (2) the recollection of that sensation is distorted by the fact that either it occurs during the process of awakening when cognitive functions are again disturbed, or it occurs in response to a deliberate prompt or involuntary cue to recall a dream, when social and cultural factors affect recall; (3) the dream report—whether public or part of an internal stream of conscious thought—nearly always occurs in a social and cultural context¹⁶ in which there are conventions about what sort of dreams one might have, what those might say about you, and to whom it is appropriate to report dreams of different types (e.g. we expect phenomena such as a character in a dream being X but looking like Y). So the bodily sensation of indigestion is the cause of the ¹⁵ Nielsen and Powell (2015, 12–13) find the most plausible hypotheses to be food distress (e.g. lactose intolerance for dairy-related dreaming), folklore, and misattribution. If we distinguish the explanation of a particular case of food-influenced dreaming, where many factors may be in play, from the explanation of the phenomenon itself, food distress has the advantage of being able to explain the existence of folklore and patterns of misattribution. ¹⁶ The reflexivity movement in psychotherapy is a good example of systematically working with those factors—called the ‘Social GGRRAAACCEEESSS’, standing for Gender, Geography, Race, Religion, Age, Ability, Appearance, Class, Culture, Ethnicity, Education, Employment, Sexuality, Sexual Orientation, Spirituality—in a therapeutic context (e.g. Burnham, 2012). But they are present in all conversations where we expect to be interpreted, assessed, or judged and reporting dreams is, in our culture, one such context.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

,  ,  



dream report, but the content of that report is explained by all the other factors that influence memory and reporting. The nocturnal discomfort, a form of acquaintance with your body, is causally responsible for a confused waking state which disposes subjects who have the concept of a dream to explain by confabulating a rich series of conscious perceptual experiences during their sleep. The content of the confabulation is only partially related to the actual nocturnal awareness (discomfort causes bad dreams) but filled out by cultural conventions, social expectations, and of course how those apply to the individual context of the dreamer, which might vary from day to day.¹⁷ Thus we have dreams without dreaming. This is just one particular kind of case, though one which is very common. When we take into account that nocturnal interoceptive and sensory experiences include not only indigestion, but also other bodily sensations such as warmth, coldness, stiffness, the touch of the sheets, etc. and restricted perceptions of the environment, such as light levels and noises, then we can find a huge range of initial causes for dreams, namely real perceptions of real things which are confused in the process of encoding and recall, subsequently feeding into a confabulatory process, which also draws upon a variety of psychological, social, and cultural influences, of engaging in the phenomenon of dream reporting.¹⁸ The bulk of the confabulatory work concerns the content of the dream, for the dream report is usually prompted by remembered nocturnal interoception and sensation, but the model can also allow cases where the social or environmental context triggers a confabulation without a confused memory of a nocturnal experience. An example of this is when one ‘remembers’ a dream one had some time previously and not before recalled, having an experience remarkably like déjà-vu and often described as ‘I am sure I had a dream about that’. If you wonder how the cultural influences work, consider the phenomenon of oedipal dreams in pubescent males. For physiological reasons, pubescent males are particularly prone to nocturnal emissions and concomitant reports of erotic dreams.¹⁹ How does this come about? While asleep the young man has some awareness of physiological changes such as the erection and ejaculation; when he wakes he is confused for he has been asleep, alone, and not masturbating; the concept of a dream—taught to him at a young age to help soothe the night terrors—provides a handy explanation; so he confabulates a sequence of experiences which would explain the ejaculation (were they real experiences, of course); this confabulation conforms to culturally expected norms, further embedding the problematic experience of nocturnal ejaculation into a familiar narrative. What norms might these be?

¹⁷ Here it is assumed that the mechanism of confabulation which is well documented in clinical cases might be available when there is no ‘underlying pathological condition’ (McVittie et al., 2014). Issues around extending the concept of confabulation in this way are discussed in Moody and Stoneham (forthcoming). ¹⁸ Rosen (2013) gives an excellent account of how the content of dream reports is constructed from these factors. However, she thinks that there is empirical evidence that something like dreaming occurs during sleep, merely denying that dream recollections are reliable guides to what that experience is like. I discuss the issue of empirical evidence for dreaming in Section 5, but note now that the view here proposed looks like a version of Rosen’s which restricts nocturnal consciousness to interoception and sensation, and for that reason declines to call it dreaming in Section 5. ¹⁹ See Lucretius, 1975, Book 4, 1025–36) for an early poetic description.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

Well, given the widespread Freudian idea that dreams are revelatory of suppressed desires, and the equally widespread belief that these include sexual desire for the mother (for a famous example, one only needs to listen to David Bowie’s ‘Sister Midnight’ in which he recounts a series of dreams of ‘mother in my bed’), we could predict some oedipal confabulations as a response to nocturnal emissions. On this model of dreams without dreaming, we do away with the postulation of large amounts of nocturnal experiences with vivid phenomenal character unexplainable by acquaintance with anything external to the subject. This is clearly of benefit to someone who wants to explain phenomenal character by reference to acquaintance, since acquaintance now explains much more of the data than it appeared to when we accepted the standard model of dreaming.²⁰ But one might fear that by denying dreaming, we are implausibly denying something that is too central to our conscious mental lives to be denied. However, it is a mistake to think that by denying that dreams correspond to nocturnal, conscious, non-perceptual experiences, that is, that dream reports taken as reports of such experiences are even roughly accurate, we thereby recommend putting an end to dream reporting.²¹ On the contrary, this view keeps firmly in place the cultural importance of dreams: whatever first caused people to report having had vivid nocturnal experiences of unreal scenarios, and we can only speculate, it quickly caught on. Sometimes it is useful, such as when ‘it is only a dream’ soothes the infant’s night terrors; sometimes enjoyable; sometimes fascinating in a way we want to share; but it is never a window onto a lively but hidden mental world. This does not mean that dreams are not revealing about you, but they are revealing about you in the way your jokes or choice of conversational topics are revealing about you: they say something about your personality, your cultural context, your social relations, your current concerns, and generally what is holding your attention or worrying you at the moment. It does not even hurt to continue falsely thinking of dreams as reports on largely accurate memories of nocturnal experiences, so long as you do not build a psychological theory or therapeutic practice out of the significance of these ‘hidden’ conscious events. In fact, all the main conclusions of large-scale content-analysis of dream reports would be equally predicted on the cultural-social model of dreams without dreaming (Domhoff, 2000): First, dreaming [i.e. whatever causes dream reports] is a cognitive process that draws on memory schemas, episodic memories, and general knowledge to produce reasonable simulations of the real world (Antrobus, 1991; Foulkes, 1985; Foulkes, 1999), with due allowance for the occasional highly unusual or extremely memorable dream (Bulkeley, 1999; Hunt, 1989; Knudson & Minier, 1999; Kuiken & Sikora, 1993). Second, dreams have psychological meaning in the sense of coherency, correlations with other psychological variables, and correspondences with waking thought (Domhoff, 1996; Foulkes, 1985; Hall, 1953b). Third, ²⁰ This is also the main reason to prefer dreams without dreaming to Rosen’s (2013) view that dreaming consists in sparse conscious experience which is confabulated into rich narratives on wakening. ²¹ The long-standing debate in popular culture about reason vs. emotion, personified in Star Trek’s Spock and Kirk, reminds us that classing a phenomenon as a cognitive imperfection does not rule out finding it non-instrumentally valuable.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

,  ,  



the unusual features of dreams, such as unlikely juxtapositions, metamorphoses, and impossible acts, may be the product of figurative thought. (Hall 1953a; Lakoff 1997)

5. Objections and Replies At this point I aim to have shown that there is a viable alternative to the standard model of dreaming, an alternative which should be attractive to anyone who wants to explain phenomenal character by reference to acquaintance with extra-mental items. But that falls short of showing the alternative model is correct. The evidence we would need to decide in favour of the standard model would be that there are conscious perceptual experiences during sleep which cannot be explained by acquaintance and which are remembered, sometimes inaccurately, in dream recall. And here we hit a problem: the normal evidence for conscious experiences is the subject’s report, whether verbal or via some behavioural response test, but in the case of dreaming there is no simultaneous subject report, only the later memory. Now in general we can take subjects’ memories of conscious experiences at face value, pace the normal issues about reliability of episodic memory, but in the case of dreaming we have a double problem: first, it is not possible to assess the general reliability of memories for dreaming by comparing reports against independent evidence of the dreaming because we have no other access to the dreaming;²² and second, we are seeking evidence to decide between two theories, one of which denies that dream reports are reports of memories (operations of the faculty of memory), however accurate or inaccurate. If what is at stake is whether someone is remembering or confabulating, we cannot decide the issue by appealing to what they ‘remember’. This should not be confused with a general form of memory scepticism: the problem is that dream recall is memory of something which was private at the time it occurred. It does not even extend to memories of what is in fact unverifiable, such as childhood events which occurred when one was alone, because we can assess the general degree of reliability and susceptibility to confabulation for memories of that type of event. With dreaming, however, we have no way of assessing, at individual or group level, how reliable we are and how much we confabulate, because we can never compare dream reports with dreaming. (1) Earlier I referred to the assertion that we forget 95–99 per cent of our dreaming. How is such a claim established if we only have the subjects’ memories to go upon for whether dreaming occurred? One method is to look at nocturnal neural activity prior to dream reports and to see how often similar activity occurs even though there is no subsequent report (e.g. Horikawa et al., 2013). There are several problems with this. One is a general methodological issue with neuroscience: there are no non-arbitrary grounds (yet) to determine when two patterns of neural activity are sufficiently similar. Everything is similar to everything ²² Claims about the reliability of dream reporting in the psychological literature often cite studies which compare reports given during the night with those given the next day (Baekland and Lasky, 1968; Meier et al., 1968; Strauch, 1969; Trinder and Kramer, 1971). Clearly this presupposes the accuracy of the nocturnal report, or at least takes that as a methodological starting point.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

else in some respect and anything as complex as a working brain will be similar to another working brain in multiple respects. Just mining the terabytes of data a neuroimaging generates for specific patterns will produce false positives (Kilner, 2013). Another is that the grounds for thinking this particular pattern of activity correlates with conscious experiences rests entirely on taking the dream report to be based on a largely accurate memory of such experiences. The problems which affect this particular use of brain-scanning data will affect any use of scanning techniques to determine that there are conscious experiences during sleep. Suppose the methodological issues were resolved and we had reliable data correlating patterns of neural activity with waking conscious experience. Could we then look at the sleeping brain to determine when it was having similar conscious experiences? This is the infamous and methodologically dubious ‘reverse inference’ (Poldrack, 2006). In general, it is hard not to be impressed by evidence that the brain is ‘doing the same thing’, in a context where we are trying to understand the psychological processes going on, as it is doing in a context where we have a reasonably good understanding of the psychology. Consequently we tend to infer from neural state A being associated with psychological function F in one context and the occurrence of A (or perhaps a similar A*) in another context to the occurrence of F (or perhaps F*) in that context.²³ However, there are two significant problems with inferring in this way from the context of conscious perceiving to the context of dreaming. The first is that because the dreaming subject is asleep, we will not find exactly neural state A (typically there will be no alpha rhythm across the occipital lobe—though see (3) below for detailed discussion of sleep states) and so we will always be comparing similar states, A and A*. That leaves the inference open to challenge that the states are not sufficiently similar and we should be particularly cautious of claims of similarity in the neuroscience literature based on scoring systems which are designed for different purposes. To classify A and A* as being of the same type for the purpose of inferring psychological function will always be tentative until we have a complete understanding of the grounding of psychological function in the brain, and we are a long way from that. Furthermore, philosophical assumptions about that grounding relation are often brought into the typology without being defended or even made explicit. The second is over the relation between the psychological functions. Someone who wishes to explain (some) cases of phenomenal character by reference to the subject being acquainted with the external world will regard the psychological function F in the conscious perceiving subject as something necessarily unattainable in the dreaming subject. Consequently, the inference must be to a psychological function F* which does not involve acquaintance, and it is thus an open question whether F* does involve conscious experiences. This is not just a point about local supervenience. Of course any phenomenal character which is partially constituted by acquaintance will not be locally supervenient, but the inference we are considering is not to exactly the same psychological state but to an allegedly similar one. Rather the point is to do with

²³ A recent example is the widely reported research by Siclari et al. (2017).

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

,  ,  



the oddity of postulating quasi-perceptual conscious experiences in the sleeping subject given the background view that it is the operation of the senses in acquainting the subject with their environment that produces phenomenal consciousness. The point was made eloquently nearly a thousand years ago by Al-Ghazali: If a man had no personal experience of dreaming and someone were to tell him: “There are some men who fall down unconscious as though they were dead, and their perception, hearing, and sight leave them, and they perceive what is ‘hidden,’ ” he would deny it and give apodeictic proof of its impossibility by saying: “The sensory powers are the causes of perception. Therefore one who does not perceive such things when his powers are present and functioning a fortiori will not perceive them when his powers are suspended.” (Ghazzālī, and McCarthy, 2010 [1105], 110)

(2) It might be suggested that we do not need to appeal to neuroscience to find evidence of dreaming, whether it is remembered or not. For example, most people are convinced that dogs dream because they have seen sleeping dogs twitching and whimpering as if they were ‘chasing rabbits’. Similarly, many people sleeptalk and some sleepwalk. It is worth carefully unpacking the inference here: (i) The evidence is some behaviour in sleeping subjects which is similar to some type of waking behaviour (though not necessarily one that occurs in the same subject—dogs who never chase anything while awake still twitch their legs in their sleep). (ii) The waking behaviour, when it occurs, is explained by reference to a conscious, perceptual experience: the dog is running that way because she can see the rabbit. (iii) The sleeping behaviour is similar enough to the waking behaviour for it to need a similar explanation. (iv) Therefore, we postulate a conscious perceptual experience in the sleeping subject to explain the behaviour, for example, the dog is dreaming of chasing a rabbit. It should be obvious that (iii) is problematic in many ways: it ignores the possibility of equifinality; it overestimates the similarity of the sleeping and waking behaviours;²⁴ it assumes sleeping and waking subjects are both intentional agents whose behaviour needs explaining by reference to conscious mental states. Let us concentrate on the third. Why think that a sleeping animal’s behaviours are susceptible to the same kinds of explanation as a waking one’s? It seems that we have tacitly assumed a kind of biological mechanism here, namely that the body is a machine subject to a single set of explanatory laws at all times. But when we consider animals

²⁴ The similarities are quite limited: the jerky twitching of legs is not at all like the graceful running of the greyhound, most sleeptalking consists of disconnected phrases unlike waking speech, and the sleepwalker is usually easy to spot from their gait and posture. More plausibly similar, but less familiar, are cases of REM Sleep Disorder, however in all but the rarest cases there is ‘complex motor behaviour’ rather than actual acting out of complex scenarios which requires something more clearly under intentional control (which might be what is going on in this famous video: https://youtu.be/Js50Orx94iM).

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

less mechanistically, the states of sleeping and waking seem to mark precisely the sort of difference we might expect to be reflected in explanations of their behaviour. Perhaps, what a waking animal does is to be explained one way (with reference to conscious experiences and intentions) but a sleeping animal does not do things in the same sense and is thus not to have its behaviours explained in the same way. In particular, when a sleeping dog moves as if running, that will have an explanation relevant to the fact that it is asleep and precisely not having perceptions of rabbits. To explain it by a dream is to try to force the sleeping state into the framework for explaining waking behaviours—it is tempting but not obligatory. The alternative model of dreaming fits well with a particular way of thinking about the relation between the sleeping and waking subject. It is easiest to explain this by a thought experiment. Suppose some people occasionally went into a state of heightened awareness, call it ‘super-waking’, in which they could control their digestion in much the same way that we can control our breathing by intentionally holding our breath or panting. Further suppose that changes very much like the ones the superawake bring about happen in the intestine of some people merely awake: we would not be in the slightest bit tempted to explain those changes by ‘hidden’ intentions but would instead treat them as purely physiological. But explaining the physical behaviours of sleeping subjects by postulating dreams is doing just that: it is taking something which is in the intentional control of a waking subject and postulating hidden experiences and intentions to explain it in the sleeping subject. Rather, when we find phenomena which occur in the sleeping subject and also under intentional control in the waking subject, we should see the waking occurrences as being susceptible of a different type of explanation, one in terms of perceptions and intentions, that is simply unavailable in explaining the sleeping occurrences. (3) The appeal to these familiar sleeping behaviours looks promising evidence for the standard model because it appears to provide contemporaneous evidence of the dreaming. However, that evidence was not decisive because we were choosing to interpret the behaviours as under conscious control based on their similarity to waking behaviours. An ingenious experimental method invented by Stephen LaBerge (1980) gets around this problem by focusing on lucid dreamers. Lucid dreams occur when the subject becomes aware that they are dreaming. For most people who experience this, the awareness is followed by immediate wakening, but some subjects report that they sometimes continue dreaming in this lucid state. Furthermore, some expert lucid dreamers report being able to induce lucidity and even choose the dream scenario. LaBerge’s idea was that subjects who were able to continue dreaming having attained lucidity might be able to signal that they were dreaming using a pre-arranged signal. Given that ‘stable lucid dreams are nearly exclusively found in REM’ sleep (LaBerge, 2000, 963) and most motor functions are paralysed in that sleep state, the signals were originally eye movements though later experiments looked at the onset of, for example, arm movements which had been inhibited. Using these signals, LaBerge was able to determine when these lucid dreams occurred, their duration—up to fifty minutes (LaBerge, 1990, 129)—and even ask the subjects to perform tasks to assess their level of cognition, such as recalling keywords and where they were sleeping (Levitan and LaBerge, 1993).

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

,  ,  



The phenomenon of lucid dreaming itself is not evidence for the standard model because lucidity in dreams is still only a phenomenon reported on the basis of recollection of a nocturnal conscious experience, and thus no different from any other dream content. What is crucial in the LaBerge paradigm is that we have prearranged signals the interpretation of which does not depend upon a subsequent dream report nor upon parallels with waking behaviours. While it remains possible that the development of a suitable measurement scale will reveal that phenomena associated with lucidity are more common than usually thought (Voss et al., 2013), for present purposes it would seem that even a few cases of signalling by expert lucid dreamers would provide the crucial evidence for the standard model and present a problem for acquaintance accounts of phenomenal character. Rather than try to pick apart particular experiments, I want to consider the experimental method more generally. The first thing to note is that the pre-arranged signal must have a determinate meaning. The most obvious one to start with is ‘I am now lucid dreaming’. Where the subject is able to not merely attain lucidity but also choose the dream scenario, a signal might mean ‘I am now lucid dreaming about X’. Finally, where a task to be performed during the dream had been agreed, a signal might mean ‘I have completed task Y’. However, simply agreeing a signal does not always succeed in giving that signal meaning: the agreement needs to take a certain form. Assuming that in this case, the subject does want to signal the event whenever it occurs, that is, she has no desire to conceal information from the experimenter, then the agreement must be of the form ‘signal s IFF e is occurring’. Now consider the simple case where we are attempting to create a signal for the occurrence of the lucid dream state. Can the subject sincerely agree to signal s if and only if she is lucid dreaming? To agree that she would have to believe that when lucid dreaming she will be able to signal s and when not lucid dreaming she will not ‘accidentally’ signal s. The second issue is addressed by choosing a signal which is very unlikely to happen during normal sleep, like a specific sequence of eye movements. The first is addressed by trials. But notice the structure the trial would have to take: the subject would sleep with the instruction of signalling when she is lucid dreaming, the signal would occur and the subject would be woken and asked if she was lucid dreaming. A strong correlation between the signal and the subsequent report would be evidence she has the ability to signal s if e is occurring. But now pressure is put onto our resolution of the second issue, namely whether she has signalled ‘accidentally’. Originally it looked sufficient for this to be the case that the signal be something which does not normally occur in sleeping subjects. However, for the trial of her ability to signal when lucid dreaming to be convincing, it must be the case that the signal is unlikely to occur accidentally in a sleeping subject who has been primed to make that particular signal during that particular night of sleep. Which is to say that we only have the subject’s report as evidence that the signal was caused by the lucid dreaming. The alternative explanation of the trial results is that she is primed to make the signal while asleep, on waking she has a memory of the act of making the signal, and—due to pressure to perform and perhaps a belief in the existence of lucid dreaming—confabulates a dream to explain the signalling. The problem we have with the LaBerge methodology is that the subject goes to sleep with the intention of

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

performing a certain action only when lucid dreaming, and what we know for sure is that she performs it when in REM sleep and reports a lucid dream, but that leaves open the question of whether she did what she intended to do, or whether there is some other explanation of what happened. The incredulous question ‘What else would explain why she did it?’ takes us back to the problem discussed in (2) above, namely that alleged evidence for the standard model often presupposes that the explanation of waking behaviours in terms of conscious experiences and intentions should also be used to explain sleeping behaviours. It is precisely this desire to make the sleeping subject psychologically just like the waking subject—except that a few physiological functions are ‘turned off ’—which makes the standard model so attractive and which the alternative model is challenging. That is not the end of the matter, however, because the self-identified expert lucid dreamers recruited for LaBerge’s experiments clearly have some unusual abilities on any theory of dreaming. In particular, they seem able to do things in their sleep which go beyond what the rest of us are able to do when asleep.²⁵ What is in question is whether we have evidence for LaBerge’s interpretation of these abilities in terms of lucid dreaming. When considering this, it is worth remarking that LaBerge often takes his research to show that common assertions to the effect that self-reflection, meta-cognition, reasoning, recall of waking knowledge, and volitional control are ‘greatly attenuated in dreams’ (Hobson et al., 2000, 799) are exaggerated. If his data do show that lucid dreamers exercise these competences, then that might just emphasize that these people have some very unusual abilities to perform complex behaviours while asleep. However, we should be cautious about drawing these conclusions without investigating another aspect of the methodology of lab-based dream research, namely the operational definitions of sleep stages. Consider the American Academy of Sleep Medicine’s Manual for the Scoring of Sleep and Associated Events (Iber et al., 2007), which is the standard for scoring polysomnographic data. This defines five stages of sleep (24): Stage W (wakefulness), Stages N1–3 (non-REM sleep), and Stage R (REM sleep). What is striking is the comparison of the definitions of Stages W and R. For example, Stage W’s definition includes exactly the same eye movements as Stage R with the following observation: ‘While rapid eye movements are characteristic of stage R sleep, they may also be seen in wakefulness with eyes open when subjects scan the environment’ (25). In fact, the crucial differences between stages W and R seem to be the presence of alpha rhythm and normal or high chin muscle tone in the former. Since rule A makes alpha rhythm over the occipital region sufficient for stage W and rule B allows scoring as W ‘without visually discernable alpha rhythm’ (25)

²⁵ We should not underestimate what normal humans are able to learn to do in their sleep. One striking example is that children learn not to roll out of bed while asleep, but there are also quite common abilities to learn more particular actions. Consider the way that when sleeping in an unfamiliar location or without a familiar sleep partner, some of our habitual sleep behaviours fail and we wake in surprise. Now note that this does not always happen, and happens less to some people. Which is to say some of us, some of the time, are able to control our sleeping behavioural habits to adapt to a new environment. And many people find it easy to sleep through a recurring noise, such as traffic or a clock chime once they know what it is. The abilities of expert lucid dreamers are more striking, but not obviously of a different kind to these abilities.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

,  ,  



when there are rapid eye movements with normal or high chin muscle tone, it seems neither is necessary.²⁶ In fact, the similarities at a neurophysiological level between stages W and R have led to REM sleep being commonly referred to as ‘paradoxical sleep’ (Jouvet, 1965). This raises complex questions about how expert lucid dreamers engaged in making pre-arranged signals to the experimenter should be scored. Let us assume that LaBerge is following the manual precisely and these subjects are correctly scored as being in stage R. That does not in fact show that they are asleep in the sense in which normal dreamers are asleep, precisely because the rules for scoring are not designed to take into account these unusual subjects in an unusual state. Suppose we extended the manual to include definitions of day-dreaming, in particular the state we sometimes call ‘tuning out’ or ‘being away with the fairies’, and hypnagogic/hypnopompic hallucinations, with the explicit intention of distinguishing these from the five stages already defined. How would these relate to the state of the expert lucid dreamer engaged in signalling? I have no idea what the polysomnographic data would show, but it is highly likely from what we know about these states that we would find several dimensions of similarity and difference. Consequently, any operational scoring of stages would depend to some extent on decisions made relative to our interests in these states. The primary interest here is likely to be clinical—diagnosis and treatment of sleep disorders—and from that perspective whether an epoch is preceded or succeeded by a sleep stage is highly relevant and likely to result in the special states of lucidity being classed as more similar to the hypnagogic/hypnopompic hallucinations than to day-dreaming. However, that method of scoring tells us nothing at all about the intrinsic character of the state and if applied too rigidly will have trouble identifying micro-sleeps and microawakenings. It seems then that LaBerge’s experimental method using expert lucid dreamers promises to provide the crucial evidence in favour of the standard model, but only succeeds in doing that against the context of certain assumptions about the differences between sleep and wakefulness. We could express the argument of this section as a dilemma: expert lucid dreamers engaged in signalling to an experimenter are either awake or asleep. If they are awake, then they are undergoing a rare form of daydreaming with REM sleep paralysis, but as such are not relevant to discussions of normal, sleeping dreams. If they are asleep, then while they have unusual abilities to learn sleep behaviours, we are not obliged to explain those behaviours using the model of conscious experiences and intentions which applies to waking subjects. (4) There remains a more purely philosophical concern over the dialectical position. Sensory imagination also appears to involve phenomenal character not explainable by acquaintance and the reasons for questioning the standard model of dreaming do not apply in such cases. Consequently, the philosopher interested in using acquaintance to explain phenomenal character will have to be a disjunctivist when it comes to sensory imagination. Furthermore, there are familiar arguments that dreaming should be classified as a form of sensory imagination (e.g.

²⁶ Rule B also refers to ‘reading eye movements’ which is a criterion that can only be applied in certain contexts and I will set aside for this discussion.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

Ichikawa, 2009), and this would thereby preserve the standard model without creating new challenges for the acquaintance theorist. This objection presupposes that we must accept the phenomenal character of sensory imagination as a given. It is, of course, open to the acquaintance theorist to deny this and to want to do so for precisely the reasons we discussed in Section 2. But that would be the subject of another paper. Equally, the acquaintance theorist might argue, along the lines of Section 2, that the benefits of not rejecting the standard model are outweighed by the benefits of rejecting it, namely denying that there is such a large stretch of our conscious mental lives, rich in phenomenal character, outside the explanatory scope of their theory of phenomenal character. Setting these responses on one side, let us consider how plausible it is that dreaming on the standard model is a form of sensory imagination. Ichikawa notes there are various options for the role of imagination here since dreams appear to involve both perceptual experiences and beliefs based on them: imagined perception with actual belief; actual (but not veridical!) perception with imagined belief; imagined perception with imagined belief. The main objection to the imagined perception is that dreams are mainly involuntary whereas sensory imagination is voluntary. The main objection to imagined belief is that dream beliefs seem to have the functional role of actual beliefs. What he overlooks is that what is driving the objection to dream beliefs being imaginings in fact provides a deeper objection than involuntariness to the claim we are currently considering, namely that dream perceptions are sensory imaginings. Consider Hume’s famous claim that ‘in the imagination the perception is faint and languid, and cannot without difficulty be preserv’d by the mind steddy and uniform for any considerable time. Here then is a sensible difference betwixt one species of ideas and another’ (1978 [1739], 1.1.3). This is standardly interpreted²⁷ as saying that imaginings are qualitatively distinct from perceivings (and rememberings) and it is objected that perceptions might lack vividness and imaginings possess it. But in fact our sense that the view is absurd does not derive from the existence of counterexamples but from the fact that perception is presentational whereas imagination is not. To use a common early modern phrase, in imagination we ‘conceive as possible’, emphasizing that the objects of imagination are presented as possibilities, whereas to perceive is to be related to actuality. Thus an account of dreaming as sensory imagination would have to find a supplementary explanation of why the subjects report their dreaming experiences as perceptions, as presenting actuality rather than possibility.²⁸ While this would be possible, it shows that the move does not in fact preserve the standard model intact and thus increases the dialectical force of the second response noted above.

²⁷ But arguably incorrectly (Everson, 1988). One person who certainly believed that an imagining made more vivid and coloured would turn into a perception was Arthur Collier (1713, 17–18). ²⁸ Many philosophers appeal at this point to the famous Perky experiments (1910). However, they fail to take into account Segal’s work (1971, 1972) which is only able to replicate Perky in ways which suggest that the phenomenon is not one of mistaking a perception for a sensory imagining at all, but interference of the perception in the process of imagining.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

,  ,  



6. Conclusion I argued that a naïve-realist who appealed to disjunctivism in order to relegate dreaming to the ‘bad cases’ thereby weakened the value of their explanation of phenomenal character in the good case of acquaintance with the external world. In the light of this, I considered what options there are for the naïve-realist when considering dreaming and noted that it is the standard model of dreams and dreaming which causes the problem. There are some well-known problems with the standard model, on the basis of which alternatives have been suggested. I tried to motivate an alternative which preserves and explains our culture and practice of dream reporting without regarding those reports to be expressions of roughly accurate memories of conscious nocturnal experiences. While it is possible to construct such an alternative, neither the problems with the standard model nor the existence of alternatives have resulted in more than a handful of mavericks rejecting the standard model. So I explored what evidence there might be to support the standard model and concluded that it was inconclusive and in many cases question begging. Where does this leave the dialectic? We have no direct evidence for or against the occurrence of dreaming, though we are all subjectively convinced it happens. But we have two empirically inconsistent models which are roughly equal in their explanatory adequacy (depending how you weight phenomena like Maury’s dream against the conviction that we are remembering our dreams). We could hold out for new direct evidence or new differences in explanatory adequacy, but absent such developments, it seems like a reasonable basis for choice is fit with our best theories elsewhere. And here the naïve-realist who wishes to explain the phenomenal character of perceiving by our being acquainted with external objects has a reason to reject the standard model of dreaming.²⁹

References Baekeland, F. and Lasky, R. (1968) ‘The Morning Recall of Rapid Eye Movement Period Reports Given Earlier in the Night’, Journal of Nervous and Mental Disease 147(6): 570–9. Botman, H. I. and Crovitz, H. F. (1989) ‘Facilitating the Reportage of Dreams with Semantic Cues’, Imagination, Cognition and Personality 9(2): 115–29. Burnham, J. (2012) Developments in Social GGRRAAACCEEESSS: Visible-Invisible and Voiced-Unvoiced. In I.-B. Krause (ed.), Culture and Reflexivity in Systemic Psychotherapy: Mutual Perspectives, London: Karnac, 139–60. Burton, R. (1989) The Anatomy of Melancholy, Vol. I, Oxford: Clarendon Press. ²⁹ An early version of this chapter was read to colleagues at the University of York and I am very grateful for their suggestions and scepticism. Louise Moody, David Austin, and Ben Springett have all provided fantastically helpful written comments and suggestions. Rob Davies combined excellent copy-editing with his usual insightful comments. Declan Hartness helped prepare the bibliography. A talk on the culturalsocial model of dreams has been presented to many and varied audiences in recent years and has always generated valuable discussions, especially in public lectures and schools. Finally, thanks to Dzmitry Karpuk of the Complex Trauma Therapists’ Network for helping me understand the relation between the culturalsocial model and his Systemic Embodied Experiential Re-processing (SEER) protocol for therapeutic interventions.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

Chudnoff, E. (2015) Cognitive Phenomenology, New York: Routledge. Collier, A. (1713) Clavis Universalis: Or, a New Inquiry after Truth: Being a Demonstration of the Non-Existence, or Impossibility, of an External World, New York: Garland. Dennett, D. (1976) ‘Are Dreams Experiences?’, Philosophical Review 85(2): 151. Domhoff, G. W. (1996) ‘The Repetition Dimension in Dreams and Waking Cognition’, in Finding Meaning in Dreams, New York: Springer, 191–212. Domhoff, G. W. (2000) ‘Moving Dream Theory Beyond Freud and Jung’. Presented to the symposium ‘Beyond Freud and Jung?’, Graduate Theological Union, Berkeley, CA, 23 September. Dummett, M. (1979) ‘Common Sense and Physics’, in Graham Macdonald (ed.), Perception and Identity: Essays Presented to A. J. Ayer, with His Replies, New York: Cornell University Press. Everson, S. (1988) ‘The Difference between Feeling and Thinking’, Mind, 97(387): 401–13. Farkas, K. (2006) ‘Indiscriminability and the Sameness of Appearance’, Proceedings of the Aristotelian Society 106(2): 39–59. Fish, W. (2009) Perception, Hallucination, and Illusion. Oxford. Oxford University Press. Foulkes, D. (1999) Children’s Dreaming and the Development of Consciousness. Cambridge, MA: Harvard University Press. Freud, S. (1976 [1900]) The Interpretation of Dreams, Harmondsworth: Penguin. Ghazzālī, and McCarthy, R. J. (2010 [1105]) Deliverance from Error: An Annotated Translation of Al-Munqidh Min Al Dal-al and Other Relevant Works of Al-Ghaz-al-i. Louisville, KY: Fons Vitae. Goblot, E. (1896) ‘Sur le Souvenir des Rêves’, Revue Philosophique de la France Et de l’Etranger 42: 288. Guénolé, F. and Nicolas, A. (2010) ‘Dreaming Is a Hypnic state of Consciousness: Getting Rid of the Goblot Hypothesis and its Modern Avatars’, Clinical Neuroscience 40: 193–9. Harlow, J. and Roll, S. (1992) ‘Frequency of Day Residue in Dreams of Young Adults’, Perceptual and Motor Skills 74(3): 832–4. Hartmann, E. (1968) ‘The Day Residue: Time Distribution of Waking Events’, Psychophysiology 5(2): 222. Hartmann, E. (1998) ‘Nightmare after Trauma as Paradigm for All Dreams: A New Approach to the Nature and Functions of Dreaming’, Psychiatry 61(3): 223–38. Hinton, J. M. (1967) ‘Visual Experiences’, Mind 76(2): 217–27. Hobson, J., Pace-Schott, E., and Stickgold, R. (2000) ‘Dreaming and the Brain: Toward a Cognitive Neuroscience of Conscious States’, Behavioral and Brain Sciences 23(6): 793–842. Horikawa, T., Tamaki, M., Miyawaki, Y., and Kamitani, Y. (2013) ‘Neural Decoding of Visual Imagery During Sleep’, Science 340(6132): 639–42. Hume, D. (1978 [1739]) A Treatise of Human Nature, London: John Noon. Hyman, I. E., Husband, T. H., and Billings, F. J. (1995) ‘False Memories of Childhood Experiences’, Applied Cognitive Psychology 9(3): 181–97. Iber, C., Ancoli-Israel, S., Chesson, A., and Quan, S. F. (2007) The AASM Manual for the Scoring of Sleep and Associated Events: Rules, Terminology and Technical Specifications, Vol. 1, Westchester, IL: American Academy of Sleep Medicine. Ichikawa, J. (2009) ‘Dreaming and Imagination’, Mind and Language 24(1): 103–21. Jouvet, M. (1965) ‘Paradoxical Sleep: A Study of Its Nature and Mechanisms’, Progress in Brain Research 18: 20–62. Kassin, S. M. and Kiechel, K. L. (1996) ‘The Social Psychology of False Confessions: Compliance, Internalization, and Confabulation’, Psychological Science 7(3): 125–8. Kilner, J. M. (2013) ‘Bias in a Common EEG and MEG Statistical Analysis and How to Avoid it’, Clinical Neurophysiology 124(10): 2062–3.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

,  ,  



Kramer, M. (2000) ‘Dreaming and Illness’, in Paul Raphael Duberstein and Joseph M. (eds), Psychodynamic Perspectives on Sickness and Health, Washington, DC: American Psychological Association, 73–120. LaBerge, S. P. (1980) Lucid Dreaming: An Exploratory Study of Consciousness During Sleep. Doctoral dissertation, Stanford University. LaBerge, S. (1990) ‘Lucid Dreaming: Psychophysiological Studies of Consciousness during REM Sleep’, in R. Bootsen, John F. Kihlstrom, and Daniel L. Schacter (eds), Sleep and Cognition, Washington, DC: American Psychological Association Press. LaBerge, S. (2000) ‘Lucid Dreaming: Evidence and Methodology’, Behavioral and Brain Sciences 23(6): 962–4. Levitan, L. and LaBerge, S. (1993) ‘Dream Times and Remembrances’, NightLight 5(4): 9–14. Loftus, E. and Ketcham, K. (1992) Witness for the Defence: The Accused, the Eyewitness, and the Expert Who Puts Memory on Trial, London: Macmillan. Loftus, E. and Ketcham, K. (1996) The Myth of Repressed Memory: False Memories and Allegations of Sexual Abuse, London: Macmillan. Loftus, E. and Palmer, J. C. (1996) ‘Eyewitness Testimony’, in Introducing Psychological Research, London: Macmillan, 305–9. Lucretius (1975) On the Nature of Things, trans. W. H. D Rouse, Cambridge, MA: Harvard University Press. Malcolm, N. (1959) Dreaming, London: Routledge. Marquardt, C. J., Bonato, R. A., and Hoffmann, R. F. (1996) ‘An Empirical Investigation into the Day-Residue and Dream-Lag Effects’, Dreaming 6(1): 57. Martin, M. G. F. (2002) ‘Beyond Dispute: Sense-Data, Intentionality, and the Mind–Body Problem’, in Tim Crane and Sarah A. Patterson (eds), The History of the Mind-Body Problem, New York: Routledge. Martin, M. G. F. (2004) ‘The Limits of Self-Awareness’, Philosophical Studies 120(1–3): 37–89. Martin, M. G. F. (2006) ‘On Being Alienated’, in Tamar S. Gendler and John Hawthorne (eds), Perceptual Experience, Oxford: Oxford University Press. McVittie, C., McKinlay, A., Della Sala, S., and MacPherson, S. E. (2014) ‘The Dog that Didn’t Growl: The Interactional Negotiations of Momentary Confabulations’, Memory 22(7): 824–38. Meier, C. A., Ruef, H., Ziegler, A., and Hall, C. S. (1968) ‘Forgetting of Dreams in the Laboratory’, Perceptual and Motor Skills 26(2): 551–7. Moody, L. J. and Stoneham, T. (forthcoming) ‘Conflabulating Your Dreams’, Topoi. Nielsen, T. A. and Powell, R. A. (1992) ‘The Day-Residue and Dream-Lag Effects: A Literature Review and Limited Replication of Two Temporal Effects in Dream Formation’, Dreaming 2(2): 67. Nielsen, T. A. and Powell, R. A. (2015) ‘Dreams of the Rarebit Fiend: Food and Diet as Instigators of Bizarre and Disturbing Dreams’, Frontiers in Psychology 6(47): 1–18. Perky, C. (1910) ‘An Experimental Study of Imagination’, American Journal of Psychology 21(3): 422–52. Phillips, I. (2013) ‘Afterimages and Sensation’, Philosophy and Phenomenological Research 87(2): 417–53. Poldrack, R. A. (2006) ‘Can Cognitive Processes Be Inferred from Neuroimaging Data?’, Trends in Cognitive Sciences 10(2): 59–63. Pryor, J. (2005) ‘What’s so Bad about Living in the Matrix?’, in Christopher Grau (ed.), Philosophers Explore the Matrix, Oxford: Oxford University Press, 40. Rosen, M. G. (2013) ‘What I Make Up When I Wake Up: Anti-Experience Views and Narrative Fabrication of Dreams’, Frontiers in Psychology 4: 514.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

Rubin, D. C. (1999) Remembering Our Past: Studies in Autobiographical Memory, Cambridge: Cambridge University Press. Schneider, A. and Domhoff, G. W. (2017) The Quantitative Study of Dreams, 24 January. Available at: http://www.dreamresearch.net/. Segal, S. J. (1971) ‘Processing of the Stimulus in Imagery and Perception’, Imagery: Current Cognitive Approaches: 69–100. Segal, S. J. (1972) ‘Assimilation of a Stimulus in the Construction of an Image: The Perky Effect Revisited’, in The Function and Nature of Imagery, New York: Academic Press, 203–30. Siclari, F., Baird, B., Perogamvros, L. et al. (2017) ‘The Neural Correlates of Dreaming’, Nature Neuroscience 20(6): 872–8. Snowdon, P. F. (2005) ‘The Formulation of Disjunctivism: A Response to Fish’, Proceedings of the Aristotelian Society 105(1): 129–41. Stoneham, T. (2008) ‘A Neglected Account of Perception’, Dialectica 62(3): 307–22. Stoneham, T. (2012) ‘-ve Afterimages Caused by Desensitized Photoreceptors for +ve Colour ∴ Naïve Realist Can Explain as Blocking of Ability to See +ve Colour’, ἀκαρές, 7 January. Available from: https://twitter.com/philosjournal. Strauch, I. (1969) Psychological Aspects of Dream Recall. In Sleep and Dreaming, Symposium at the Nineteenth International Congress of Psychology, London. Trinder, J. and Kramer, M (1971) ‘Dream Recall’, American Journal of Psychiatry 128(3): 296–301. Voss, U., Schermelleh-Engel, K., Windt, J., Frenzel, C., and Hobson, A. (2013) ‘Measuring Consciousness in Dreams: The Lucidity and Consciousness in Dreams Scale’, Consciousness and Cognition 22(1): 8–21. Zadra, A. (1996) ‘Recurrent Dreams and Their Relation to Life Events and Well-Being’, Trauma and Dreams: 231–47.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

7 Relationalism, Berkeley’s Puzzle, and Phenomenological Externalism Jonathan Knowles

1. Introduction According to John Campbell, Berkeley’s puzzle (henceforth, BP) is to understand how perceptual or sensory experience can form the justificatory basis of our knowledge and conception of mind-independent things (Campbell, 2002a). While Campbell does not sympathize with Berkeley’s idealism, he does think he bequeathed us an important insight, namely that ‘concepts of individual objects, and concepts of the observable characteristics of such objects, are made available by our experience of the world’. Thus: ‘The puzzle that Berkeley is addressing is that it is hard to see how our concepts of mind-independent objects could be made available by our experience of them’ (128). The resolution of this puzzle that Campbell proposes, consistent with the insight, involves what he calls the relational view of experience (also known, and referred to hereafter, as relationalism). According to relationalism, a sensory experience, such as a visual experience of a red apple in front of one, essentially involves the object the experience is about, and has its phenomenal character in large part constituted by this object and its qualities. Moreover, a sensory experience does not, as many have recently averred, involve representing the object and its qualities as being certain ways; rather, it is a more basic kind of epistemic relation to the world akin to the kind of acquaintance Russell took us to bear to non-physical objects like sense-data and universals. For Campbell relationalism provides a way of respecting Berkeley’s insight in a way representationalist accounts of perceptual experience cannot. Though Campbell’s work has attracted critical attention from a variety of quarters, he has had a particularly sustained debate with Quassim Cassam. Cassam argues, contra Campbell, that a representationalist view can be defended as a solution to BP, whereas relationalism fails to deliver on this front. He also has some doubts about the terms of the puzzle itself. Their exchanges can be found in Cassam (2011), Campbell (2011a), and most recently a jointly authored book which presents their to-date most considered views and points of disagreement (Campbell and Cassam, 2014). In this chapter, I will be critically considering Campbell’s relationalism, both as an answer to BP and as an adequate account of the nature of experience, aiming instead Jonathan Knowles, Relationalism, Berkeley’s Puzzle, and Phenomenological Externalism In: Acquaintance: New Essays. Edited by: Jonathan Knowles and Thomas Raleigh, Oxford University Press (2019). © Jonathan Knowles. DOI: 10.1093/oso/9780198803461.003.0008

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

to develop a somewhat similar but alternative account of this. I take as my point of departure Cassam’s critiques, and thus to an extent the chapter can be viewed as a critical discussion of their joint contributions to the relationalism-representationalism debate. However, my objections to relationalism go beyond what Cassam has to say; moreover, I have no interest in defending any of his representationalist alternatives, nor any kind of compromise position between relationalism and representationalism. With respect to the resolution of BP, I will argue that consideration of this does not provide any strong argument in favour of relationalism in view of the possibility of explaining our concept of mind-independence by reference to (something like) innate ideas. The underlying aim here is deflationary, suggesting, as Cassam himself does, that the alleged puzzle is less of a pressing concern than Campbell takes it to be. With respect to the issue of the nature of experience, I will argue that relationalism is challenged by a somewhat similar view I call ‘phenomenological externalism’ (henceforth PE). While acknowledging ordinary objects and their qualities as constitutive of phenomenal character, phenomenal externalism lays more emphasis on the experiencing subject in our understanding of what such objects and qualities amount to, viewing them as part of something like a ‘world-for-me’ or ‘world-for-us’ rather than the fully objective world that fundamental physics describes. PE shares many of the core motivations of relationalism, but it seeks to develop these in closer relation to ideas that can be identified with (and hence the name) the phenomenological tradition in philosophy: that originating in Kant and pursued most famously by Husserl, Heidegger, and Merleau-Ponty, whose ideas in turn inform contemporary work in so-called ‘embodied’ and ‘enactivist’ cognitive science by figures such as Franciso Varela, Hubert Dreyfus, and Alva Noë.¹ PE per se is not meant to coincide with any particular extant view; rather, it is aimed at being a general position in the debate on perceptual experience that many of these individual approaches could be seen as exemplifying.² I will not be offering anything like a knock-down argument for PE over relationalism, but rather suggesting ways it can seem to provide a more satisfactory externalist view of experience, and how some of its apparent drawbacks might not be so serious after all. Further, I will suggest it fits naturally with the kind of

¹ See Chemero and Kaufer (2014) for a recent introductory text on phenomenology structured along these lines. ² The term ‘phenomenological externalism’ is used by several other authors. One is Dan Zahavi, who employs it in a sense very similar to that I will be assuming here (cf. Zahavi, 2008). For him it is a variety of externalism which whilst stressing the constitutive dependence of experience on non-mental objects also sees the latter as constitutively dependent on subjectivity, in the manner first suggested by Kant. Zahavi also sees the notion as applicable to all the main thinkers of the phenomenological tradition, their disagreements notwithstanding. Max Velmans also uses the expression (see 2017, 147), again in a way that, at least as I understand it, is very similar to mine. A slightly different use of the term is Gregory McCullouch’s in his The Mind and Its World (1995), where it designates something closer to what I here mean by externalism about experience. At this juncture it is also appropriate to stress that my understanding of phenomenology is a (non-reductive) naturalistic one, seeing phenomenological study both as a theoretical one concerning a particular ‘phenomenon’ (viz. experience), and as at least potentially responsive to and in engagement with studies from neuroscience, cognitive science, and so on—and hence not as a (at least purely) transcendental, a priori study of the conditions of all meaningful thought and enquiry. This will become apparent in Section 4. For defence of this conception of phenomenology and its significance see e.g. Wheeler (2011), Knowles (2013), and Reynolds (2016).

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

,  ’  



deflationary perspective on BP that I argue for. Insofar as it is a view that divorces questions about experience from epistemology, as well as mitigating the sense in which objects contribute to phenomenal character simply by our being related to them, PE can also be seen as rejecting the centrality of the idea of acquaintance to the theory of experience. The chapter is structured as follows. In Section 2 I give a brief overview of relationalism and BP, based on Campbell’s contribution to Campbell and Cassam (2014). Section 3 then critically discusses relationalism as a response to BP, building out from Cassam’s objections. Section 4 also starts with Cassam’s objections to relationalism as an adequate account of experience, but goes considerably beyond these in its attempt to motivate PE. The conclusion reflects on how phenomenological externalism relates to the issue of mind-independence we find in BP.³

2. Relationalism For Campbell, confusion about the role of experience began with the scientific revolution. On Newton’s view of nature there is in the physical world nothing like the colours, smells, even the familiar solid objects of ordinary experience. The reaction was to ‘push sensory experience inside the head’. But ‘now we have the problem of explaining how this stuff inside the head can be playing any privileged or distinctive role in generating our knowledge of the world around us’ (2). Some people have concluded that it can’t, indeed that this whole way of thinking is wrong-headed. One of these was Berkeley. But while Campbell thinks it reasonable to hold that knowledge of the world derives from sensory experience, he takes Berkeley’s idealism to be an overreaction to the modern problematic. Berkeley’s significant legacy is rather this puzzle: we have at least a prima facie commitment to both (i) the idea that sensory experience is crucial to our conception and knowledge of mind-independent things and properties, and (ii) the idea that sensory experience only gives us a conception and knowledge of such experience itself; and these ideas seem to be in serious tension.⁴ Cassam labels (i) ‘experientialism’ and (ii) ‘sensationism’ (101), and I will follow him in this. Campbell thinks a solution to BP can be had by rejecting sensationism, but that this requires one embraces relationalism about experience. Relationalism however can also be independently motivated, and Campbell emphasizes two points in this regard. The first is that reality can be described at many different levels, which allows us to say that the mind-independent world is literally—at some level—populated by solid, moving objects that are, in themselves, coloured, smelly, noisy, and so on (3). The rise of the special sciences as well as the so-called ‘hard problem of consciousness’ (Chalmers, 1995)—finding a place for qualitative experience in a physical

³ In view of the centrality of Campbell and Cassam’s book to my presentation, otherwise unreferenced page numbers refer to it hereafter. ⁴ Saying Berkeley bequeathed us this puzzle might be misleading. Being more precise we could perhaps say that he had an insight (cf. the introduction) that on the assumption of a non-idealist picture leads to a puzzle. For of course Berkeley didn’t think his own view puzzling at all, but mere commonsense.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

world—in turn motivate this levels talk.⁵ The second is G. E. Moore’s notion of the ‘diaphanousness’ or transparency of experience (18 ff.): when we introspect on what our experience is like, we find essentially nothing (i.e. no qualia),⁶ and so must recoil to the worldly objects and properties of which we are aware. But since now, by the first point, there is nothing in the way of thinking of the world as having these properties, we can uphold a ‘naïve’ view of experience, as a mere relation of consciousness or awareness of this world, or better, of its multitudinous objects and qualities (22–3). Moreover, in the same blow, we can explain, or at least start to explain, how sensory experience grounds our conception of mind-independent objects, since such experience is nothing other than our being cognitively ‘in touch’ with them—‘acquainted’, to use Russell’s term.⁷ If one doesn’t opt for relationalism (and eschews Berkeley’s idealism), Campbell thinks one is forced into accepting ‘epiphenomenalism’, according to which sensory experience plays no role per se in our understanding of the concepts of things around us (27). He sees representationalist views of experience as committed to epiphenomenalism since they involve a notion of content that is understood in terms of causal covariation (plus a few twirls) between neural structures and environmental conditions, and as such is a proprietarily sub-personal phenomenon (43). So on this view a ‘super blindsighter’, who has learned that her responses to stimuli that are not registered consciously are nevertheless reliable (cf. Block, 1997), could gain a conception of mind-independent objects; whilst Mary (cf. Jackson, 1986) could have a conception of mind-independent colour just by studying the physics and physiology underlying colour and colour vision. But these are consequences Campbell takes to be deeply counter-intuitive, and hence he rejects representationalism. Cassam calls this the redundancy problem for representationalism (107)—i.e. the problem that representational views fail to find any explanatory role for conscious experience as such to play in giving us knowledge of the world—a usage I will again follow.⁸ Just what is a mind-independent object for Campbell, and how does relationalism help to explain our grasp of the concept of this? His examples are everyday objects like a table, tree, or knife, which Campbell theorizes as ‘causal unities’ involving ‘internal causal connectedness, which is independent of its relation to a mind’ (26). Another way of saying this is that mind-independent objects constitute a kind of mechanism underlying familiar, causally interconnected series of events—such as sharpening a knife at point A, taking it to point B, and it cutting tomatoes more ⁵ I return to the idea of levels of reality and the motivation it provides for relationalism in Section 4. ⁶ Campbell has a good deal to say by way of critique of qualia theories of consciousness, much of which I think could also be related to Cassam’s defence of a phenomenal intentionalist version of representationalism (172 ff.). I find what Campbell says on these matters largely convincing, and note Cassam also concedes quite a lot to Campbell on this score (see e.g. 200–1). Neither kind of view will be discussed further here. ⁷ Campbell does not in fact use this concept in the 2014 book, by contrast with earlier publications (see e.g. Campbell, 2002a, 2002b, 2011b) but I take it this indicates no substantive change in position. ⁸ Campbell has another main objection to representationalism, namely, that in invoking a notion of content it presupposes that we already grasp concepts of mind-independent things, as constituents of the contents of sensory experiences (45). Though this issue was central in earlier work by Campbell and Cassam (see Campbell, 2002a, 2011a), it would seem to have a subordinate status in their book and will not figure in my discussion here.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

,  ’  



efficiently at B (31) (here the mechanism is the knife and its movement from A to B). Relationalism, says Campbell, explains how perceptual experience might play a role in our having this conception of objects as mind-independent by allowing us to think of ourselves as in direct contact with them and their functioning as such mechanisms. Representationalism by contrast only secures contact with them via the idea of an essentially non-experiential, sub-personal system of representation; or perhaps ourselves utilizing theoretical, descriptive knowledge—neither of which secures the required first-personal contact with worldly objects, nor the intuitive sense in which experience is necessary to have knowledge of, say, colour. Thus, in briefest outline, is Campbell’s case for relationalism.⁹ However, in order to be acceptable relationalism needs to be refined beyond the intuitive idea that sensory experience is a relation between a subject and mind-independent reality. After all, we don’t experience everything in the world, and those things we do experience are not experienced in their entirety or always in the same way. To account for this, Campbell first says we must think of sensory experience in terms of ‘a three-place relation holding between: (i) the observer, (ii) the point of view from which the scene is observed, and (iii) the scene observed’ (27). Further, the experience ‘may itself be adverbially modified’ (28). What exactly Campbell puts in the latter idea is left somewhat unclear at the point it is introduced; it is exemplified there by the idea of ‘experiencing watchfully’. In the following chapter (chapter 3), Campbell fills out the idea of adverbial modification by drawing on empirical work by Huang and Pashler (2007) on the distinction between selection and access in visual attention.

3. Berkeley’s Puzzle In this section I will try to show that consideration of BP (restricted in the way just indicated) does not mandate relationalism. My reasons build on those Cassam adduces, but also go beyond these. Cassam thinks that relationalism doesn’t give a good or the only resolution to BP, and also that the very terms of BP are questionable in relying on both the thesis of experientialism, and on a certain, in his view faulty, conception of ‘mindindependence’. Starting with experientialism, Cassam argues that while it seems reasonable to think that experience plays an important role in grounding at least some of our concepts of things and properties in the world, Campbell grants it too large a role (124–6). Experience of colours seems necessary for our grasp of colour concepts; but what about concepts like ‘hexagon’, ‘gold’, or ‘apple’? For these and many others it seems a purely theoretical (i.e. non-experiential) understanding is ⁹ Campbell in fact gives a further argument for relationalism from the structure of the acquisition of propositional knowledge, which also forms part of BP as he defines it (cf. ch. 4). Following J. L. Austin (1962), he claims that under certain conditions sensory encounters are decisive in our making up our minds on certain matters, but that this explanatory role only makes sense on a relationalist view of experience since representationalism, again, can give no essential role to conscious nature of such encounters. Though again of inherent interest, I do not pursue the issue here, firstly, because the Austinian datum is not obviously compelling (cf. Cassam’s comments on pp. 132–3) and secondly because it does not seem to have any impact on how Campbell otherwise uses BP to argue for relationalism (as detailed above), which moreover seems to constitute its main source of support.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

possible: one could grasp them without having experienced the corresponding things. Cassam admits that for at least some of these concepts a full understanding might require experience, if not of the object types themselves, then something they can be understood as being composed of, e.g. simpler spatial qualities in the case of hexagons. But the dialectical situation leaves it at least far less clear than Campbell would want it to be that sensory experience is essential to concepts of mindindependent things and their properties. Indeed, Cassam also notes, the distinction between partial and full understanding might allow one to argue that one could have a partial understanding even of colours purely through theoretical knowledge (126). Cassam’s next group of arguments focus on Campbell’s specific conception of a mind-independent object and the role experience plays in our grasp of this (126 ff.). As we have seen, for Campbell our concept of a mind-independent object is the concept of an underlying causal unity or mechanism; further, to grasp such a concept, we have to experience some of its instances. Cassam objects that the reasoning here is unclear since we can have concepts of underlying mechanisms that we do not and even cannot experience (127–8). Campbell must of course be assuming that the experiential cases are basic, and somehow provide a model for the others; but, as with the general thesis of experientialism, it is not totally clear why this should be so. Cassam continues with a potentially deeper worry. As he sees things Campbell’s notion of a mind-independent object is one which even a Berkeleyan idealist could accept, for the property of being an underlying causal unity could it seems be attributed to objects viewed as constituted by sensations, i.e. things that there are necessarily perceived or mind-dependent (128). Building on work by Gareth Evans (1980), Cassam then develops a more robust notion of a mind-independent object, a discussion which issues in the idea of this both as something there to be perceived— thereby something that can exist unperceived—and material, i.e. a bearer of primary and not merely secondary qualities (150–2). Though Cassam acknowledges the role that experience might play in explaining our grasp of the idea of perception, which is central to this definition, he also raises questions about whether we could claim it impossible for us to attain such a conception independently of experience, at least in part (129, 153). Finally, Cassam argues that however precisely one views the concept of a mindindependent object, relationalism does not explain how this property could be as he puts it ‘registered’ in experience in such a way as to ground our conception of such objects (155–7). Indeed it is unclear how any property could be registered in experience for a relationalist. By contrast, at least in relation to Cassam’s preferred, Evansian conception of a mind-independent object, representationalism can, through its invocation of content, explain how this property is registered and thereby acquired, along with other properties we experience (160 f.). How devastating a critique of relationalism does all this amount to? Campbell himself offers responses in his epilogue to the book, in turn addressed by Cassam’s own epilogue. Though there is much of interest in these exchanges, what I want to do here is offer an outline of a more general response one might give to Cassam’s objections on relationalism’s behalf that will ultimately serve to bring out what I think is a more principled problem with it as a response to BP.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

,  ’  



Cassam’s objections can be summarized as follows: firstly, he questions the general acceptability of experientialism, both in relation to concepts of particular things in the world, and the concept of mind-independence itself; secondly, he points out the apparent compatibility of Campbell’s notion of mind-independent objects with Berkeleyan idealism; thirdly, he queries how relationalism can understand the experiential registration of properties in the world. With regard to the first point, relationalists could presumably demur at Cassam’s suggestion that it has the burden of proof to show that our concepts are dependent in a significant and deep sense on experience.¹⁰ With regard to the second, relationalists could insist that mindindependent things also are, and have to be seen as, potentially unperceived by us, and perhaps also as material; or alternatively, that the notion of a causal unity, understood aright, is not something an idealist can actually accommodate (this is also the substance of some of Campbell’s remarks in his epilogue). With regard to the third charge, relationalism could invoke, as Campbell has in the past, something like the idea of knowledge by acquaintance to explain how properties are registered in experience, extending the Russellian idea from objects like sense-data to everyday tables and chairs. One might not be wholly convinced by these replies, but it at least makes sense to ask: how strong a case for relationalism as a solution to BP are we left with, assuming they are cogent? Throughout the book Campbell presses the redundancy problem for representationalism. Now in fact, though Cassam does not fully accept this, he does acknowledge it has some force (200–2). I suspect this is because he is attracted to a kind of epistemological internalism: the view that the grounds we have for our knowledge or conception of something have to be consciously available to the subject. Certainly Campbell seems to accept this, and I can register personal sympathy with this part of his view. If we now, being generous to relationalism, and also accepting that epistemological internalism does create a serious redundancy problem for representationalism, does the former after all emerge, if not problem-free, nevertheless as our best response to BP as things stand?¹¹ I think not. For what is still on the table is the possibility of a more principled rejection of experientialism, motivated not piecemeal through the kinds of counterexamples Cassam gives, but through an emphasis on the idea of a conceptual understanding that is essentially independent of experience for its grounding, resting rather on what we might call (if only for want of a better term) innate ideas. Such ¹⁰ In the rest of this chapter, my focus will be exclusively on our concept of mind-independence, which is the more fundamental issue. I thus allow, at least for the sake of argument, that some of our concepts of particular objects and properties may rely on experience, at least in part; I take it that this is of no use to relationalism insofar as it is consistent with explicitly idealistic views (that is, experientialism should not be understood in such a way that it is something an idealist could accept). ¹¹ I should stress that as I understand the dialectic, representationalism is not ruled out simply by accepting that the redundancy problem is a problem—which might (along with other assumptions we have made, i.e. against the qualia and the phenomenal intentionality view) seem to render relationalism the winner by default—but rather by the fact that the redundancy problem creates problems for seeing how it can solve BP. (In fact given that I think there is a further view of experience consistent with the negative claims of relationalism but opposed to its positive ones—see Section 4—this assumption is not ultimately that important for me, but dropping it would require a tedious complication of the arguments in this section.)

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

‘ideas’ may of course require that we have experience for them to ‘come on line’, but that is quite a different matter from their being thus grounded. Moreover, though modern conceptions of the doctrine of innate ideas deriving from cognitive science see them as sub-personal structures of the brain, there seems no barrier to an epistemological internalist making use of the notion through the idea of somehow drawing on such structures in a conscious act or acts of (non-sensorily grounded) conceptual understanding.¹² Now in fact Georges Rey (2005), in his commentary on Campbell (2002a), takes up the idea of innate ideas or knowledge as the possible source of our conception of mind-independence. Campbell’s response to Rey is that stressing such knowledge doesn’t exclude the idea that experience plays a vital role in the acquisition of the concept of mind-independence; and moreover, that that is all relationalism ever wanted to say, i.e. that it never denied the contribution of innate knowledge (Campbell, 2005, 162). However, nowhere in this exchange is there any consideration of the idea that our conception of mind-independence might be grounded purely in non-experientially-based conceptual activity. This is odd: the view after all doesn’t threaten the idea that experience has an important role to play in acquiring concepts of particular mind-independent things, and knowledge about them, but only that experience can ground our conception of mind-independence per se. Elsewhere (2011a), Campbell criticizes what I think could be viewed as a somewhat similar account of the origin of mind-independence that Cassam puts forward, presenting it as inspired by Kant. According to this time slice of Cassam, our conception of mind-independence derives both from experience of objects and some non-experientially grounded conceptual component akin to a Kantian category. Campbell’s critique, very briefly, takes the form of a dilemma: either this view acknowledges the importance of our experiential contact with mind-independent objects, and hence is just a variant on relationalism; or it does not, and hence lapses into a kind of idealism, in which our conscious mental lives might in principle fail to be in epistemic touch with anything ‘outer’, and our notion of mind-independence will concern only a phenomenal world. I think as far as the first horn goes Campbell is right. However, it is the second which interests me, in that the interpretation might be seen as corresponding to my suggestion that our conception of mindindependence is grounded wholly in innate ideas, and only enabled by experience (which enabling job might then presumably be done without experience literally making contact with the external world).¹³ However, construed thus, Campbell’s objection seems wholly gratuitous. If there are mind-independent objects, and we have an innate concept of mind-independent objects, then we can reasonably expect this concept will latch onto them and not anything phenomenal, whatever our

¹² The model here is the Chomskyan idea that a suitably motivated individual can draw on her innate knowledge of syntax through intuitions and thereby consciously construct a theory of her language that (in the limit) coincides with that innate knowledge. ¹³ I should also say that though Cassam doesn’t talk in terms of innate ideas himself, there would seem no barrier to this interpretation of his Kantian view. (And for the record: though Kant himself of course espouses a form of idealism, I take it this and his reasons for it go beyond what is at issue in the current discussion.)

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

,  ’  



experience is like. Moreover, if the role acknowledged by Campbell to be played by innate ideas (or knowledge) in grounding our conception of mind-independence does not lead to an idealist construal conception of this on his relationalist account, then why should it do so if the role afforded experience is gradually reduced, ultimately to the point where it acts merely as a trigger? It is important here to remember Cassam’s objection that Campbell’s initial statement of relationalism is not clearly a non-idealist position, in that what we are aware of might be construable in mind-dependent terms (e.g. as bundles of sensations). Now I suggested above, on behalf of relationalism, that one can block that interpretation by insisting on a meatier notion of mind-independence. But where is that to come from, if not from non-experientially-based understanding? And given that, how can relationalism avoid idealism if another view which simply lays exclusive stress on such understanding apparently cannot?¹⁴ The debate thus comes down to something like a clash of intuitions over whether you think experience does or doesn’t play a vital role in grounding our conception of mind-independence. Relationalism might be seen as the default view if one starts out with the first. But given there is a coherent, non-idealist alternative, and given also that there are many challenges that relationalism still has to face down—for example, telling us precisely what acquaintance is and how it allows us to pick up information about the world around us—it strikes me that the prospects of cogently arguing this are not that great. I conclude that a consideration of BP does not at all clearly support relationalism.

4. Phenomenological Externalism In spite of failing to be mandated by a consideration of BP, relationalism qua theory about the nature of experience is, in my view, motivated by many sound insights (as are other versions of so-called ‘naïve-realism’ in the literature). Most fundamentally, the phenomenal character of sensory experience cannot credibly be understood in terms of qualia or representational content, but is, rather, at least to a large extent, constituted by the very things in the world we experience and their qualities. Though this kind of externalism—‘externalism about experience’, or simply ‘externalism’ as I shall call it in the following—is in the first instance a phenomenological claim, derived from the first-personal analysis of our own experience, I see no reason not to accept it as part of a theory of what sensory experience is, absent strong reason to do otherwise. On the other hand, I also, like Cassam and representationalists, think that there must be more emphasis on the subjective side of experience than relationalism allows. However, I do not think what is needed is a return to representationalism, even in part. We need instead something different from both representationalism and relationalism.

¹⁴ In putting the point this way I am also indicating that I am open to a certain deflation of the whole mind-independence issue; see the conclusion.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

To enter these issues, we can start by considering a generic objection standardly raised against relationalism (and naïve-realism more generally) that we can call ‘the mismatch problem’: quite simply, the way the world is and the way we experience it need not match up. Holding the world constant, experience can vary (cf. aspect perception, attentional shifts, effects of differing past experience, etc.); holding experience constant, the world can vary (cf. hallucinations and dreams); and, finally, the way the world is and the way we experience it may fail to coincide (illusions). If the phenomenal character of experience is given largely by the external world, how can all this be (cf. Cassam at 141 ff.)? Now relationalism, as we have seen, already has some answers to this question insofar as our experiences of objects, or scenes, are always from a particular point of view, as well as being subject to shifts in attention, degrees of watchfulness, and the like. Campbell also takes up hallucinations (90 ff.) and illusions (86 ff.), and discusses various responses relationalists have made or might make consistently with resisting representationalism. Against this background one can I think identify relationalism’s responses to the mismatch problem as falling into one of three broad categories. Firstly, it makes recourse to a broadly intentionalist position about experience which, in addition to the objects of perception, acknowledges different ways or modes in which these objects are perceived (cf. Crane, 2009). Most intentionalists also include representational content amongst the determiners of phenomenal character; that of course is precisely what relationalism denies, but I think it is useful nevertheless to see it as a form of intentionalism in the broad sense just characterized, insofar as things like point of view, attentional structuring, and so on contribute to the character of experience.¹⁵ Secondly, relationalism makes use of the broad ideas behind disjunctivism: that genuine perceiving and at least some other forms of experience that the individual may not be able to identify as differing from perceiving nevertheless must be seen as being fundamentally different kinds of state and as differing in experiential character. Thirdly and finally relationalism can make ontological gambits about the nature of the external world, as for example Bill Brewer’s (2011) account of illusion appears to in embracing the idea that things in the world can in and of themselves look certain ways (in certain contexts). There are of course many issues one might take up here in relation to the relationalism-representationalism debate, as Cassam does. As noted my sympathies are firmly with relationalism, and think a lot of what it has to say by way of resisting the pressure to go representationalist is either correct or at least promising (this applies in particular to its particular take on the intentional structure of experience, and its disjunctivism, at least when it comes to things like hallucinations and dreams). What I want to do now however is focus on a further aspect of the mismatch problem that is I think less often or at least less extensively discussed

¹⁵ Campbell also talks of ‘modes of presentation’ in a slightly different sense in chapter 3 in his discussion of selection vs. access, arguing that insofar as a ‘mode of presentation’ is understood as a property use to select some object in perception it should be seen as something external (67). Here I abstract away from this possibility, noting merely that insofar as he acknowledges modes of a more standard kind (point of view, watchfulness, etc.) he also acknowledges a significant role for things on the subject side as determinants of phenomenal character.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

,  ’  



than those mentioned above; and then on a response of Campbell’s to it that falls into the third kind of strategy identified above, but that I think fails. Cassam again gives expression to the problem (without going on systematically to pursue it) as follows: ‘Doesn’t physics show that the “qualitative character of the world is quite unlike anything that shows up in your experience?” (p. 3)’ (110, citing Campbell). Campbell’s response to this is to invoke the idea of different levels of reality, claiming that this can be used to vindicate the idea of experiences as a relation to mindindependent things and qualities in the world that constitute their phenomenal characters (see Section 2). As we have seen, Campbell sees the idea of levels of reality in turn as warranted by the rise of the special sciences, as well as the hard problem of consciousness. However, he surely moves too quickly here. Though the special sciences certainly show a need for various levels of explanation, the idea that these have ontological implications is another step. Moreover, even if they do have such implications, that this should vindicate the idea that things like coloured, solid, noisy objects are literally there in objective, physical reality in the same way electrons, quarks, and black holes are seems far less secure. Now Campbell has elsewhere (1993) defended a view known as ‘colour primitivism’ on which colour (as we experience it) is just there in the world along with electrons and the rest. However, this is highly controversial as a philosophical position, as well as conflicting with what I take to be a consensus amongst practising scientists—at least since the time of the Scientific Revolution— that colours and much else in the everyday world of experience simply cannot be aligned alongside the entities of fundamental physics in one common reality.¹⁶ Of course none of this shows colour primitivism—or other forms of primitivism about other qualities and objects of the world we experience—are false. Nevertheless, the way Campbell argues for relationalism in Campbell and Cassam (2014) is seriously compromised, for as we have seen, he argues for this position in part by appeal to the idea that such qualities do exist mind-independently, ‘out there’ for us to experience; whereas if this idea is controversial, it can hardly serve this function. Furthermore, I take it that a, if not the, primary motivation for primitivism (of various varieties) is precisely the way colours (etc.) seem to figure in our experience: as external qualities and objects, which qualities and objects literally constitute the phenomenal character of this experience. However, I think in fact that we do not need to endorse the idea of colours, etc. being there in the world independently of the way they figure in our experience to uphold this intuition.¹⁷ To see this requires a somewhat different starting point from that most contemporary philosophical discussions of these issues adopt. Consider then first ¹⁶ Byrne and Hilbert (2007) offer critical discussion of primitivism. I lay greater weight on the general scientific point insofar as people like Byrne and Hilbert’s positive views tend towards a kind of physicalism or anti-realism about colour, which I would also reject. As we shall see, my own view of the nature of experience, PE, itself embraces a kind of colour realism insofar as it upholds the idea the colours as we perceive them are out there in the world and not inside our minds or heads (and indeed much else that realists defend, such as what we find defended in Allen (2016)). What I reject is primitivism in Campbell’s sense. ¹⁷ The idea that the hard problem can be solved through relationalism, which was Campbell’s other motivation for it, also rests as far as I can see on the tenability of various kinds of primitivism—see e.g. Fish (2008) for a view along these lines—so will not be taken up separately here.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

(fundamental) physics. Many philosophers see physics as charting the fundamental structure of reality, and thereby placing heavy constraints on what else we might want to populate reality with. However, it is arguable that a more plausible view of what physics qua scientific enterprise actually amounts to is simply the charting of complex interrelations between its proprietary posits—charting ‘physical reality’ if you will, but where the idea of this being a let alone the ‘reality’ plays no substantive role (this idea is related to the deflationary ‘natural ontological attitude’ Arthur Fine finds in science; cf. Fine 1986).¹⁸ Given this conception, physics itself is simply silent on the existence of properties corresponding to, say, how things look, smell, or sound, for there is, as far as its reality is concerned, just what physics ‘finds there’; which again is not to say that such and such things don’t exist, rather, these issues are simply moot. Now as we know, philosophers have sought to take the world view of physics (or something like it) and ‘locate’ various non-physical qualities in it, either by reducing them to it or eliminating them from it—or, as Campbell does, adding them in somehow. However, at least to judge by the extent to which any particular such strategy has achieved consensus over the years, we might wonder whether there is much point to it. My first suggestion is then that we should start by letting physics be physics. The second thing to note now, pursuant on this, is that physics is—of course—not conversant about everything we talk or think about. In particular, it is not conversant about subjective experience,¹⁹ something that nevertheless surely deserves serious consideration. If one gives it this, and on its own terms—that is, looks at experience as a phenomenologist, concerned with the structures of experience per se²⁰—what one plausibly finds, at least if something along the lines of what relationalists and naïverealists have averred is on the right lines, is the presentation to the subject of a world of certain kinds of objects and qualities (in fact not just this, but certainly this in some way, and certainly not anything warranting talk of qualia or representational content). But now if this is what experience is (‘is like’) and this description is not answerable to any ‘physical reality’, then what need is there for a theory of experience to posit a separate ontology of colours, sounds, and everyday objects alongside electrons, quarks, and the rest? If the theory of experience finds these things there for us, in our experience, then that is, at least prima facie, good reason to have them there, insofar as understanding experience is concerned. We might indeed say that what a study of experience reveals is precisely a world of experience—a world-for-me or, perhaps, a world-for-us. This is not relationalism because we are not merely related to—acquainted with—something wholly independent of us, existing alongside electrons and the like. The kind of existence physical entities ¹⁸ Fine sees his ‘natural ontological attitude’ position as simultaneously deflating anti-realist or instrumentalist tendencies in philosophy of science. ¹⁹ I am abstracting away here from Bohr’s idea that quantum mechanics shows that reality must be understood as involving some kind of intervention from a subjective force; suffice it to say here that this so-called ‘Copenhagen interpretation’ is far from standard let alone uncontroversial as a take on the implications of modern fundamental physics. ²⁰ Note then that I am thinking here of ‘experience’ as denoting a particular kind of phenomenon, not (as in the Husserlian tradition) as simply coextensive with ‘phenomena’, in line with my taking phenomenology to be a kind of theoretical science, broadly naturalistic in character (see note 2).

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

,  ’  



have no more concerns the world-for-us than the latter concerns the world of physics.²¹ This sketch gives us a rough outline of what I mean by ‘phenomenological externalism’. To clarify it somewhat further, we can first consider in what way PE itself involves a relational conception of experience. Now some might see PE as aptly described by the term ‘relationalism’; Evan Thompson, for example, whose views I would see as cleaving closely to PE, uses this term to characterize his theory of colour (see Thompson, 1995). However, we need to tread carefully here. To start with, PE is in any case a different kind of relationalism from Campbell’s, in being a view on which phenomenal character consists in a relation—or at least results from such—not in what we are related to (acquainted with). That is, PE will be able to and want to say that experience and its phenomenal character can be seen as the joint upshot of what organisms bring to the physical world by way of their neural and somatic apparata, on the one hand, and the nature of the physical world, on the other, and is in part illuminable by these things.²² However nothing in this prohibits us from thinking of experience itself fundamentally in terms of objects and properties in a world around us, in accord with externalism. Relatedly, we should be clear at what level PE is in any case relationalist, in the sense outlined. Many so-called ‘relationalist’ views about things like colour typically operate with a picture on which they are dispositions to give rise to certain kinds of inner experiences, the intrinsic nature of which are what explain phenomenal character. However, this is clearly inconsistent with the externalist commitments of PE (as well as Campbell’s relationalism). Nor should PE be seen as committed to there being a self-subsistent mind or self that either projects its contents onto the physical world or constructs its own world out of materials given by this latter. The idea of a world-for-us—the world of experience—need not and probably should not be cashed out in these or any other terms, but rather should be regarded in the first instance as a primitive and fundamental feature of the phenomenological account of experience.²³ Though explicit expressions of PE are not that easy to come by, they do I think surface from time to time. Here for example is Alva Noë: The perceptual world is not a world of effects produced in us in our minds by the actual world. But the perceptual world is the world for us. We can say that the world for us is not the physical

²¹ The world of physics might thus be thought of as something like the world-in-itself, or the view from nowhere—though only something like, and certainly not in the sense Kant envisaged in speaking of Ding an sich (more on this below). Having said that, the idea of a world-for-us I am recommending here does plausibly have resonances with some aspects of Kant’s thought about the phenomenal world as something out there to be experienced but also dependent on us—as consisting of secondary qualities, in one particular sense of that notion (cf. McDowell, 1985 and, in particular, Allais, 2015). ²² Illuminable, but not reducible. More generally, between fundamental physics and the world-for-us there will be many levels of description that can cast light on the nature of our experience, without making in-roads into explaining conscious experience itself (insofar as PE respects, without trying to solve, the hard problem). ²³ There is no doubt more to say here—for example one might see subject and world as mutually presupposing structures or poles within the manifold of experience as a whole—but I hope the idea behind PE is at least sufficiently clear to make use of for present purposes. (For the record: ‘relationalism’ will refer to only Campbell’s view in what follows.)

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

world, in that it is not the world of items introduced and catalogued in physical theory. But it is the natural world (and perhaps also the cultural world) . . . One consequence of this is that different animals inhabit different perceptual worlds, even though they inhabit the same physical world. The sights, sounds, colours and so on that are available to humans may be unavailable to some creatures, and likewise, there is much that we cannot ourselves perceive. We lack the sensorimotor tuning and the understanding to encounter these possibilities. (Noë, 2004, 156)

The kinds of objects and properties that ‘show up’ for us and other experiencing organisms is dependent in part on what we variously bring to the physical world, as well as that world itself. However, this does not mean the world-for-us per se can be reconstructed from more primitive terms like ‘mind’.²⁴ No doubt many readers will find this talk of the ‘world-for-me’ or ‘world-for-us’ (indeed, the very glib movement back and forth between these expressions) as so obfuscatory or absurd as to render any apparent problem relationalism has making sense of primitivism about colours and the like anodyne by comparison. Below I will make some remarks that I hope will assuage this impression. However, I also hope the initial repugnance does not apply to all readers. In any case, I think it is clear that what I am calling ‘phenomenal externalism’ is an extant view in the literature that bears a strong affinity with relationalism and draws on much of what is attractive about it, whilst avoiding an unwelcome commitment of it. At the very least, then, I think it deserves an airing as a further possible position within the literature on perceptual experience. To further promote that end I will in what’s left of this section do the following. Firstly, I will suggest how PE plausibly provides a more coherent version of a non-representationalist intentionalism than relationalism. Thereafter I shall consider and try to answer two possible objections to PE. As suggested, particular views cleaving respectively to PE and to relationalist conceptions of experience (relationalism itself as well as other forms of naïverealism) will tend to have much in common at the level of phenomenological analysis and motivation. Both class of views start from a broad-based intentionalist view of experience which lays stress both on the things in the world we experience as well as how we experience them—albeit not how we represent them—as constitutive of phenomenal character. They will also tend, I think, to be sympathetic to certain

²⁴ A commitment to the idea of a ‘world-for-us’ is also plausibly attributable to Varela et al.’s (1991) autopoetic enactivism (see e.g. 172–3), as well as being a natural gloss on Gibson’s seminal ideas about ‘affordances’ as the immediate objects of perception (Gibson, 1979). Another prominent enactivist, Hubert Dreyfus, also seems to commit to this idea in his article with Stephen Dreyfus on existential phenomenology and the Matrix, where they argue that in a Matrix world people ‘are directly coping with perceived reality, and that reality isn’t inner’ (Dreyfus and Dreyfus, 2005, 75–6). (Having said that, in Dreyfus’s recent book with Charles Taylor (2015) a position is defended on which physics tells us more about the same reality as our ‘embodied copings’ cope with. I will return to this view and more generally the relationship between the world-for-us and physical reality below.) Finally I can mention the work of Max Velmans (2017), who less clearly qualifies as an enactivist but whose conception of consciousness as concerning a distinct phenomenal world, neither in the brain nor identifiable with the world of physics, would qualify him to the ranks of PE (indeed, as mentioned in note 2, he uses the phrase ‘phenomenological externalism’ himself, as well as referring approvingly to the phenomenological school as a precursor of his view).

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

,  ’  



disjunctivist treatments of at least things like hallucination or dreaming. Of course, they will not and do not need to coincide with respect to the analysis of every element of experience; for example, illusion might be more amenable to a more straightforward analysis assuming PE rather than relationalism, insofar as the former operates with the notion of a world-for-us, though I won’t be pursuing this line here, or seeing it as giving the former any kind of advantage over the latter. What does I think give PE an advantage is the following: experience has an essentially unitary character; it is, at least at a certain high level of abstraction, one homogenous kind of thing. The problem this creates for relationalism is that for it experience is fundamentally a product of two very different things: objects and properties in the objective world, on the one hand, and certain features of our subjective access to this world, on the other. But then it becomes very hard to understand exactly how these things could come together to create the unitary phenomenon that experience is. (Note that this is not question begging against relationalism: the charge is not that objective things and properties cannot be part of phenomenal character, but rather that it is obscure how these kinds of thing could combine with subjective properties to produce unified experience.) PE by contrast faces no such explanatory challenge: the things in the world-for-us that we see, hear, and touch are already part and parcel of the unified intentional structure that is experience. Maybe this is just a variant on the expression of puzzlement over the very idea of acquaintance as a kind of epistemic relation that I briefly discussed in Section 3. Just as it is unclear how I get to know about something by being acquainted with it, or even what this means, it might seem unclear how something could be part of my experience just by my being acquainted with, or again what this might mean. Relationalists might accuse the opposition of lack of imagination or charity here; and certainly I do not mean to suggest the point as a knock-down argument against it. Nevertheless it does strike me as an advantage for a view of experience that this kind of question is avoided, and that the components of experience, if you will, are already within the same ontological category. This is the case for PE: experience is most basically a world-for-us: a world shaped by our subjectivity, though a world of things and qualities nevertheless. I turn now to the two objections. The first can be seen as turning the justmentioned advantage I claimed for PE on its head and seeing its unified account of experience as in fact falsifying our phenomenology. Though there are I think different ways one might make this point, I will focus on a particular version of the objection that Campbell makes to Noë’s theory of experience, though one that I think he would see as hitting any form of PE insofar as it sees important links between action and experience (which it at least in most versions does). Now according to Noë (2004) certain aspects of at least visual phenomenology—such as the seeming ‘presence’ of whole, three-dimensional objects, certain parts of which are nevertheless out of view at any time—can be explained in terms of our implicit knowledge of sensorimotor contingencies such as ‘If I were to move my gaze to such and such a location on this object, I would experience so and so’. Understood in a suitably nonreductive way Campbell thinks this is insightful; indeed, that this kind of knowledge can be built into the ‘point of view’ parameter in an overall account of visual phenomenology. However, if one understands it as an account of what visual

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

experience constitutively is, Campbell thinks it amounts to a form of phenomenalism, in which experience of the world is reduced to a complex body of conditional knowledge about what sensations one would have if one did so and so. This implies that nowhere does experience of the particular or categorical get into the picture; as Campbell puts it: ‘The ordinary world, there independently of us, there for us to explore, has simply disappeared’ (Campbell, 2008, 667; see also Leddington, 2009). Now I am inclined to agree that this is a telling objection to Noë’s enactivist theory insofar as this is conceived—as Noë, at times at least, seems to conceive it—as a constitutive or reductionist one. However, though PE would, like Campbell, want to draw on Noë’s ideas, it is important to understand it is no more committed to these as a reductionist view of experience than relationalism is. What PE can and should rather insist on is that what we might call presentational and action-oriented elements are inextricably intertwined in our experience as a whole, experience which is not per se factorizable into these different components. Importantly, this does not mean that PE cannot acknowledge that sensory experiences present us with particular objects (or particular properties) along, as it were, one dimension or axis of such complex experiences. Moreover, just as the action-oriented dimension may be explained from a lower level—implicit knowledge of sensorimotor conditionals will presumably have some kind of neural underpinning—so in explaining the presentational axis we might advert to the fact that in perception we—that is, our physical bodies and sensory surfaces—are in causal interaction with distinct physical structures outside our bodies, with their own physical properties. The existence of these structures and properties might then be invoked as an illumination or ‘vindication’ of the presentational dimension of experience. Of course, this would be an explanation of an experiential, personal-level fact from the sub-personal level: the physical object in question is not in the world we experience, as it is for relationalism. But since according to PE experience is not in its essence identified with a relation to something pregiven, as it is for relationalism, there seems no reason to think it should not satisfy us. (There is no doubt more that could be said here, but my aim, again, is primarily to motivate and defend PE as a viable contender in the philosophy of perception debate, not vindicate it.) The second objection concerns what to many may seem the most alarming feature of PE, namely, its apparent commitment to a plurality of different worlds. To start with we have the split between the phenomenal world-for-us and the world-initself—the world of fundamental physics. Secondly, we seem to have to contend with a plurality of phenomenal worlds assuming different species and maybe even different individuals are sufficiently different in their sensory and somatic makeup. But these may seem simply bizarre and untenable commitments. Now I would not deny that PE faces a challenge, indeed several challenges here. However, I also think that any registration of utter consternation can be ameliorated. Let us start with the second broad challenge above. In relation to this, I would say firstly that in fact it hardly seems absurd at all to suppose that some different species literally inhabit different phenomenal worlds—say, human beings and frogs; or at least, to the extent there is, as they say, ‘something it’s like’ to be a frog at all, this will, surely, correspond to the experience of a quite different kind of world from ours (though still, given externalism, some kind of world nevertheless). Secondly, on

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

,  ’  



closer phenomenological analysis of our experience, we plausibly find, as Husserl stressed, something like a common human lifeworld that our individual subjectivities necessarily relate to: an intersubjective background of significance that we must see ourselves and our individual perspectives as part of to make sense of them as such (cf. Dreyfus and Taylor, 2015, ch. 6). Any world-for-me is thus always only comprehensible against the backdrop of a common world-for-us. So the idea that PE brings with it a brute plurality of subjective worlds simply because our experience is all somewhat different should not be countenanced. Moreover—and now returning to animals—it doesn’t, at least today, seem completely absurd to suggest some species themselves might have common lifeworlds, or indeed that certain animals might share their lifeworld with one other or even with us—their lack of linguistic abilities notwithstanding. In fact, these issues strike me as ultimately empirical questions, albeit ones that may be very difficult to adjudicate. I don’t suppose these considerations will render phenomenal worlds less unpalatable to many, but I do think that if one isn’t a priori disposed to reject the very idea, they do suggest ways of making progress in thinking about it. This leaves us with the question of the relationship between any such world and the world of physics. How should we conceive of this? By way of contrast to my own view, I will discuss briefly a stance on this question that might seem the most natural and that has recently been put forward by Hubert Dreyfus and Charles Taylor in their book Retrieving Realism. The main target of this is what the authors call ‘the mediational picture’ of the mind-world relation, which sees our contact with the world as mediated by a third item, such as an idea, sense-datum, thought, or utterance, raising the traditional spectre of mind-body dualism and external world scepticism. Drawing on work by Heidegger, Merleau-Ponty, as well as Samuel Todes’s (2001) seminal ideas, they argue for an alternative picture of the mindworld relation on which the body and skilled, absorbed coping has a foundational role to play in understanding how experience connects us with the world, a connection which is moreover direct and constitutive. At the same time, the world thus revealed has to be understood, at least in the first instance, in terms of our capacities for embodied coping. This is very much in the spirit of PE. However, they are also keen to point out that our coping always operates within boundary conditions: conditions that limit the possibilities for action and experience ‘from without’, so to speak, like the rigidness of a stick or the softness of a bed of leaves (Dreyfus and Taylor, 2015, 138). Dreyfus and Taylor then conjecture that natural science can be understood as an endeavour that investigates these boundary conditions on their own terms. On their line, talk of a ‘world-for-us’ (or an ant’s world, or whatever) thus does not enunciate a completely different reality from that which physics delineates, but a kind of view on reality whose nature in and of itself physics can tell us about. Now it also turns out that Dreyfus and Taylor do not hold that the ‘view from nowhere’ that physics provides is necessarily superior to any perspectival conception of reality (this is the ‘pluralistic’ part of their ‘pluralistic robust realism’; cf. chs 7 and 8). This idea I find plausible in itself. However, in their hands it is very hard to understand, since for them any conception aims to capture the one and only ‘reality’; but if physics captures this reality as it is in itself, how can alternative views of the same reality fail to be in some way a distortion or misrepresentation

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

of it? Furthermore, I think it is hard on their assumptions to make sense of the idea that any world of experience, albeit object-involving in its phenomenology, can in fact be anything other than a virtual world: a kind of (perhaps collective) representation or even hallucination generated by our brains. In view of this, it seems to me that the position Dreyfus and Taylor end up in is in fact very close to a version of the meditational view that they wanted to get away from. If, as we are assuming here, one wants to uphold a genuine externalism about experience, this strikes me as a very bad result.²⁵ Dreyfus and Taylor’s attempt to ‘retrieve realism’ (as they put it), at least along with upholding PE, thus plausibly fails. This suggests that a defender of PE has to face head on the question of how one should understand the idea that there is both a world (or, possibly, worlds) of experience, and a physical world, neither being reducible to the other. How should one proceed here? To start with, I think one can see the former (assuming for the moment its singularity) as the only world there is; that is, our ordinary experience corresponds, precisely, to what we ordinarily think of as a world, extended in Euclidian space and with directional time, containing solid objects interacting with each other causally in all the different ways we are accustomed to and exhibiting an independence from our individual subjective takes on them. Physics by contrast involves a highly rarefied conception of things, indeed, one so at odds with commonsense that it arguably doesn’t qualify as a description of a world as we would think of this at all. This might tempt some towards an anti-realist conception of physics, but insofar as one can understand its claims in truth-conditional terms, it seems unclear what grounds one could have for this (this point is connected to Fine’s defence of natural ontological attitude discussed briefly above). What we are not committed to saying is that there is both a world of experience and a world of physics.²⁶ In spite of this, I think PE must acknowledge phenomenology and physics as something like two ‘non-overlapping magisteria’, to borrow Stephen J. Gould’s metaphor (cf. Gould, 1997—something he applied to the divide between science and religion/ethics): two fundamental but also fundamentally different schemes or discourses of understanding—two ‘sciences’ in a suitably loose sense of the phrase. To this one might then object that they nevertheless must relate to each other somehow: after all, as well as being a highly mathematized and theoretical discourse, fundamental physics is also an empirical one; moreover, much of what I have said by way of motivation for and defence of PE has assumed they will relate via various intermediary levels of understanding. Doesn’t this force one into a more substantively realist position? It may suggest it but I think the answer is: not necessarily. Sciences that furnish in the first instance autonomous modes of understanding can be used to describe or redescribe existing phenomena—where the latter are themselves characterized in terms of some preexisting discourse—in the service of some or

²⁵ This is a somewhat truncated critique of Dreyfus and Taylor; for more details (as well as further discussion of the significance of ‘boundary conditions’) see Knowles (2019). ²⁶ I should also stress that under the general heading ‘physics’ one should allow for the possibility that there is an irreducible plurality of natural scientific theoretical disciplines, such as physics-of-the-very-big, physics-of-the-very-small, chemistry, and even certain theoretical kinds of biology and psychology.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

,  ’  



other explanatory goal. Thus, the world of experience is one that we can and do often seek to redescribe in purely physical terms, understanding this in terms of a view of ourselves as physical systems with a central nervous system stimulated by outer objects that initiate internal processing by neural mechanisms. In doing this we may also cast light on the ‘higher’-level phenomenon, here experience, insofar as we can ascertain significant correlations between this and features of neural processing— though, on the view defended here, without any goal of fully explaining the former in terms of the latter, or making metaphysical claims to the effect that consciousness is ‘in the head’ or anything of that kind. This non-reductionist stance is not a priori; reductions, or at least ‘unifications’, as one might rather call them, do occur between different disciplines in science, though they are not criterial for its success.²⁷ In the case of conscious experience it seems clear—at least, I contend that in the light of the hard problem it is clear—that we should not be expecting anything like a reduction or even unification of our understanding of this with what we know from fundamental physics. But reduction is not the only possible goal of relating different descriptions to one another. What I have said here only gestures at an account of how we should think about experience and physical science along the lines of PE.²⁸ What I think and hope it does show is that there is a complexity to the issues, forced upon us by deep problems with other available pictures, that makes simply pushing PE aside as a non-starter inappropriate.²⁹

5. Conclusion Let us recur finally to BP: is there any relationship between PE and my treatment of this? I think there is. Now I have spent relatively little time discussing the concept of a mind-independent object, but in fact it seems to me there is probably more than one idea to which this form of words might correspond. Crudely, one might say there are, on the one hand, ‘mind-independent objects’ like tables, chairs, and trees; and on the other, ‘mind-independent objects’ like electrons, quarks, and bosons. Insofar as tables, etc. are in a world-for-us they are also in some sense mind-dependent, or (as I would prefer to say) organism-dependent; however, if PE is right, they are no less external and intersubjective for that. Thus we can say that BP simply doesn’t arise for them; or perhaps that a (naturalized) form of Berkeleyan idealism about precisely these kinds of objects is correct—though in thus restricting the idealism, and in any case vindicating the objects’ status as external and intersubjective, there should be

²⁷ See Chomsky (2001, 106–7) on both these points. His first point, illustrated in relation to the famous ‘reduction’ of chemistry to quantum physics, is that in such cases adjustments are standardly needed at both levels of explanation. Rorty (1991) defends a strongly non-reductive view of science somewhat similar to that offered here, partly through the lens of Fine (op. cit.), whilst somewhat similar accounts can be found in Dupré (1993) and Ladyman et al. (2007). ²⁸ Again, for more on these matters see Knowles (2019). ²⁹ A further important aspect of the dialectical situation concerns whether we can so much as make sense of realism as ordinarily understood (see Putnam, 1983; Price, 2011; Knowles, 2014). If we can’t, I think a further intuitive resistance to what can seem like metaphysical extravagance will naturally fall away.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

nothing objectionable about that. Thus a supporter of PE will also say that things like electrons and bosons are not in the world for us and are not in any meaningful way external to us,³⁰ though they are quite real, making clear that PE is not a form of idealism in toto. But now does BP arise for these posits of physics? Arguably not, for the idea that everyday experience can give us any conception of what these things are is neither very plausible nor problematic, if we accept what physicists themselves say about them. And then we can say that whatever conception we do have of them, as the peculiarly organism-independent things they are, must in some sense come from non-experientially-based thought, i.e. precisely the innate ideas account I sketched.³¹

References Allais, Lucy (2015) Manifest Reality: Kant’s Idealism and His Realism, Oxford: Oxford University Press. Allen, Keith (2016) A Naïve Realist Theory of Colour, Oxford: Oxford University Press. Austin, J. L. (1962) Sense and Sensibilia, Oxford: Oxford University Press. Block, Ned (1997) ‘On a Confusion about a Function of Consciousness’, in Ned Block, Owen Flanagan, and Güven Guzeldere (eds), The Nature of Consciousness, Cambridge, MA: MIT Press. Brewer, Bill (2011) Perception and Its Objects, Oxford: Oxford University Press. Byrne, Alex and Hilbert, David (2007) ‘Colour Primitivism’, Erkenntnis 26: 73–105. Campbell, John (1993) ‘A Simple View of Colour’, in John Haldane and Crispin Wright (eds), Reality, Representation, and Projection, Oxford: Oxford University Press. Campbell, John (2002a) Reference and Consciousness, Oxford: Blackwell. Campbell, John (2002b) ‘Berkeley’s Puzzle’, in Tamar Gendler and John Hawthorne (eds), Conceivability and Possibility, Cambridge, MA: MIT Press. Campbell, John (2005) ‘Reply to Rey’, Philosophical Studies 126(1): 155–62. Campbell, John (2008) ‘Sensorimotor Knowledge and Naïve Realism’, Philosophy and Phenomenological Research 76(3): 666–73. Campbell, John (2011a) ‘Relational vs. Kantian Responses to Berkeley’s Puzzle’, in Johannes Roessler, Hemdat Lerman, and Naomi Eilan (eds), Perception, Causation, and Objectivity, Oxford: Oxford University Press. Campbell, John (2011b) ‘Consciousness and Reference’, in Brian McLaughlin, Ansgar Beckermann, and Sven Walter (eds), The Oxford Handbook of Philosophy of Mind, Oxford: Oxford University Press. Campbell, John and Cassam, Quassim (2014) Berkeley’s Puzzle: What Does Experience Teach Us? Oxford: Oxford University Press.

³⁰ I would uphold this even in the face of Ian Hacking’s famous defence of entity—in particular, electron—realism on the grounds that ‘if you can spray them, they are real’ (Hacking, 1983, 24). This might seem to suggest that electrons, at least, are in the world-for-us, but all it in fact shows is that we can relate this world to that of sub-atomic particles (see the discussion towards the end of Section 4). The thoroughly outré nature of electrons (and the rest) which grounds their constitutive separation from the things of this world remains. (I should perhaps add that I have never understood how Hacking’s argument was meant to function as an argument for realism about electrons in the face of anti-realist doubt.) ³¹ Thanks to Walter Hopp, Solveig Aasen, Sebastian Watzl, John Campbell, Ronny Myhre, Thomas Raleigh, Michael Amundsen, Truls Wyller, and Jussi Haukioja for discussion of and feedback on various aspects of this material over the last few years.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

,  ’  



Cassam, Quassim (2011) ‘Tackling Berkeley’s Puzzle’, in Johannes Roessler, Hemdat Lerman, and Naomi Eilan (eds), Perception, Causation, and Objectivity, Oxford: Oxford University Press. Chalmers, David (1995) The Conscious Mind, Oxford: Oxford University Press. Chemero, Anthony and Kaufer, Stephan (2015) Phenomenology: An Introduction, London: Routledge. Chomsky, Noam (2000) New Horizons in the Study of Language and Mind, Cambridge: Cambridge University Press. Crane, Tim (2009) ‘Intentionalism’, in Brian McLaughlin, Ansgar Beckermann, and Sven Walter (eds), The Oxford Handbook of Philosophy of Mind, Oxford: Oxford University Press, 474–93. Dreyfus, Hubert and Dreyfus, Stephen (2005) ‘Existential Phenomenology and the Brave New World of The Matrix’, in Christopher Grau (ed.), Philosophers Explore the Matrix, Oxford: Oxford University Press. Dreyfus, Hubert and Taylor, Charles (2015) Retrieving Realism, Cambridge, MA: Harvard University Press. Dupré, John (1993) The Disorder of Things: Metaphysical Foundations of the Disunity of Science, Cambridge, MA: Harvard University Press. Evans, Gareth (1980) The Varieties of Reference, Oxford: Oxford University Press. Fine, Arthur (1986) The Shaky Game: Einstein, Realism and the Quantum Theory, Oxford: Oxford University Press. Fish, William (2008) ‘Relationalism and the Problems of Consciousness’, Teorema 28: 167–80. Gibson, James (1979) The Ecological Approach to Visual Perception, Boston, MA: Houghton Mifflin. Gould, Stephen Jay (1997) ‘Non-Overlapping Magisteria’, Natural History 106(March): 16–22. Hacking, Ian (1983) Representing and Intervening, Cambridge: Cambridge University Press. Huang, Liqiang and Pashler, Harold (2007) ‘A Boolean Map Theory of Visual Attention’, Psychological Review 114: 599–631. Jackson, Frank (1986) ‘What Mary Didn’t Know’, Journal of Philosophy 83(5): 291–5. Knowles, Jonathan (2013) ‘Challenging Cartesian Materialism: Understanding Naturalism and the Mind-World Relation’, in Marcin Milkowski and Konrad Talmont-Kaminski (eds), Regarding Mind, Naturally, London: College Publications. Knowles, Jonathan (2014) ‘Naturalism without Metaphysics’, in Kenneth Westphal (ed.), Realism, Science and Pragmatism, London: Routledge. Knowles, Jonathan (2019) ‘Anti-representasjonalisme og realisme: Dreyfus & Taylor versus Rorty’. Norsk Filosofisk Tidsskrift 54(1–2): 55–69. Ladyman, James and Ross, Don, with Spurrett, David and Collier, John (2007) Everything Must Go: Metaphysics Naturalised, Oxford: Oxford University Press. Leddington, Jason (2009) ‘Perceptual Presence’, Pacific Philosophical Quarterly 90(4): 482–502. McCullouch, Gregory (1995) The Mind and Its World, London: Routledge. McDowell, John (1985) ‘Values and Secondary Qualities’, in Ted Honderich (ed.), Morality and Objectivity: A Tribute to J. L. Mackie, London: Routledge, 110–29. Noë, Alva (2004) Action in Perception, Cambridge, MA: MIT Press. Price, Huw (2011) Naturalism without Mirrors, Oxford: Oxford University Press. Putnam, Hilary (1983) Realism and Reason: Philosophical Papers, Vol. 3, Cambridge: Cambridge University Press. Rey, Georges (2005) ‘Explanation, Not Experience: Commentary on John Campbell, Reference and Consciousness’, Philosophical Studies 126(1): 131–43. Reynolds, Jack (2016) ‘Phenomenology and Naturalism: A Hybrid and Heretical Proposal’. International Journal of Philosophical Studies 24(3): 393–412.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

Rorty, Richard (1991) Objectivity, Relativism, and Truth: Philosophical Papers, Volume 1, Cambridge: Cambridge University Press. Thompson, Evan (1995) Colour Vision: A Study in Cognitive Science and Philosophy of Science, London: Routledge. Todes, Samuel (2001) Body and World, Cambridge, MA: MIT Press. Varela, Francisco, Thompson, Evan, and Rosch, Eeleanor (1991) The Embodied Mind: Cognitive Science and Human Experience, Cambridge, MA: MIT Press. Velmans, Max (2017) Towards a Deeper Understanding of Consciousness, London: Routledge. Wheeler, Michael (2011) ‘Science Friction: Phenomenology, Naturalism and Cognitive Science’, in Havi Carel and Darian. Meacham (eds), Phenomenology and Naturalism: Examining the Relationship between Human Experience and Nature, Cambridge: Cambridge University Press. Zahavi, Dan (2008) ‘Internalism, Externalism and Transcendental Idealism’, Synthese 160(3): 355–74.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

8 Acquaintance, Conceptual Capacities, and Attention Anders Nes

1. Introduction Russell’s theory of acquaintance offers a stark view of the relation of perception to thought. At its heart lie two conditions placed on perceptual awareness (the basic form of acquaintance with particulars, for Russell) in its relation to propositional knowledge. On the one hand, perceptual awareness is viewed as ‘essentially simpler than’ and ‘logically independent of ’ propositional knowledge (Russell, 1912, 72). On the other, it is held to be a source of propositional knowledge (48). A key question in the reception of the theory of acquaintance has been to what extent perception can jointly meet these conditions. In his attack on the ‘Myth of the Given’, Wilfrid Sellars voiced scepticism.¹ Perception can be a source of knowledge, he argued, only if perceptual awareness depends on the possession and operation of conceptual capacities. Conceptual capacities depend in turn on their use in the context of making knowledgeable claims about things. Therefore, Sellars concluded, perception turns out not to be independent of knowledge, in some suitable, ‘logical’, or constitutive sense of independence. In recent decades, this conceptualist line of critique has been advanced notably by John McDowell (1994, 2009). The influence of the conceptualist critique contributed to a neglect of the theory of acquaintance for much of the second half of the last century. In this century, however, resurgent relationist views of perception have rekindled interest in and even sympathy for the theory. My leading exemplar here will be the views of John Campbell (2002, 2009, 2012).² He departs from Russell in taking the objects of perceptual awareness to be not private sense-data but public objects (a view shared by most recent relationists). Campbell lauds, though, Russell’s ambition to articulate a view of perception as a relationship that is at once somehow more primitive than

¹ Sellars (1956), the locus classicus of the attack on the ‘Myth of the Given’, mentions neither Russell nor acquaintance, under that name. Many of its central arguments are however anticipated in Sellars (1949), which expressly targets Russell on acquaintance. ² Other relationists include Brewer (2011) and Travis (2013). Anders Nes, Acquaintance, Conceptual Capacities, and Attention In: Acquaintance: New Essays. Edited by: Jonathan Knowles and Thomas Raleigh, Oxford University Press (2019). © Anders Nes. DOI: 10.1093/oso/9780198803461.003.0009

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

thought about an object whilst, at the same time, playing a vital role in enabling such thought (Campbell, 2002, 6). My aim in this chapter is to spell out, and defend, a claim that offers the prospects for an attractive, unacknowledged element of common ground in the debate between conceptualist and relationist responses to the theory of acquaintance. The claim is that conceptual capacities, at least in a certain, fairly minimal sense to be made clear, must be operative in perceptual experience, if it is to rationalize judgement. That claim is on the face of it friendly to conceptualists. Setting aside the qualification about the ‘fairly minimal sense’, the claim has long been central to Sellars’s and McDowell’s conceptualism. More importantly, the relevant ‘fairly minimal sense’ is one that, I will argue, is implicit in McDowell’s recent, recast conceptualism, outlined in his paper ‘Avoiding the Myth of the Given’ (AMG) (2008) and later work. Section 2 addresses this recast conceptualism, pointing out some differences from McDowell’s earlier, propositionalist conceptualism in Mind and World (1994). These differences turn out to call for a reconsideration of the notion of a ‘conceptual capacity’. On McDowell’s current understanding of that notion, which corresponds to what I describe as ‘a fairly minimal sense’, conceptual capacities are distinguished by (1) belonging to the understanding and (2) bestowing content with categorial unity and form. Section 2 outlines readings of these conditions. The target conclusion here, that conceptual capacities, in the sense outlined, must be operative in experience, if it is to rationalize judgement, will be supported on the basis of two premises. Each premise will be defended, inter alia, on the basis of considerations stressed by Campbell. While he may or may not be entirely happy about endorsing both, each is at least not clearly inconsistent with his view. At least this weak sense, then, the defence of the target conclusion is friendly towards at least one notable relationist. The first premise, defended in Section 3, relates to a respect in which Campbell differs from Russell. While Russell denied attention is needed for acquaintance (Russell, 1914, 445–6), Campbell argues perceptual experience constitutes acquaintance with an object, enabling rational thought about it, only if the subject’s consciously attends to that object (Campbell, 2002, 6, 2012, 71). The premise defended in Section 3 relates closely to Campbell’s view here. It states that perceptual experience of an object rationalizes judgement about it only if it is an experience in which one attends to the object. Beside Campbell’s (2002, 8–9) nice case of ‘the sea of faces’, the defence draws on reflections on the phenomena of change- and inattentional blindness and visual extinction. I also address recent arguments, due to Susanna Siegel and Nico Silins (2014, 2019), that inattentive experience can justify belief. The second premise is defended in Section 4. It states that attention qualifies as a conceptual capacity, in the current, fairly minimal sense. That is to say that attention belongs to the understanding (a view I show to be held by Kant—a chief source of both Sellars’s and McDowell’s conceptualism), and bestows content with categorial unity and form. Again, the defence of these points relies inter alia on themes from Campbell. The conjunction of the premises defended in Sections 3 and 4 might be dubbed ‘attentional conceptualism’. They imply that conceptual capacities, in the current fairly minimal sense, must be operative in experience, if it is to rationalize judgement.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

,  ,  



Attentional conceptualism offers a novel route to this conclusion, with potential to appeal far outside the ranks of card-carrying conceptualists.

2. McDowell’s Recast Conceptualism If Russell’s conception of acquaintance is mythical, as Sellars and McDowell contend, what would a more promising view of perceptual experience look like? Sellars (1968) and, following him, McDowell (1994, 2009) take their cue here from Kant’s conception of intuition. At first blush, this might seem surprising, for Kant opposes intuitions to concepts. Sellars and McDowell argue, though, that Kant’s opposition is an artefact of his restriction of ‘concepts’ to general, mediate representations, while intuitions are singular and relate immediately to their objects (Kant, 1998: A320/ B376–7). They prefer a broader sense of ‘conceptual’, linking it to capacities belonging to Kant’s ‘higher’ cognitive faculty, the understanding. This allows Kant to emerge as a pioneering conceptualist. This is so because Kant, as Sellars and McDowell read him, takes intuitions of objects to be due not to sensibility (our passive, receptive faculty) alone but also constitutively to depend on the operation of the understanding.³ In one of McDowell’s more recent, summary statements of his view, the Sellarsian and Kantian backdrop is evident: In the experience of rational subjects, things are given to them to be known, in knowledge of a kind only rational subjects can have, knowledge that is a standing in the space of reasons. Is this givenness a case of what Sellars rejects as the Myth of the Given? No, but that is only because the experience of rational subjects, experience in which things are given for rational knowledge, itself draws on capacities that belong to the rational intellect, the understanding. (McDowell, 2009, vii)

This summary highlights two key features of McDowell’s conceptualism. First, the interest in experience is epistemic. The conceptualist requirements placed on experience are held to apply to it in its capacity of being a source of knowledge. No argument is given for imposing them on aspects of sensory awareness (if any) prior to, or irrelevant to, its status of providing knowledge. Second, ‘knowledge’ is taken in a demanding sense, on which it is to be understood in terms of knowledgeable, rational judgements: mental acts of making up one’s mind, open to reflection on their grounds. Putative experience that is realized in creatures incapable of such judgements, or that merely feeds into forms of knowledge not to be understood in terms of such judgements, is set aside. I will adopt these limitations tacitly for the rest of this section. When I speak of experience, I mean experience insofar as it is a source of knowledge, understood in this demanding sense. In Mind and World and various later essays (e.g. his 1998, 2006), McDowell set out a highly influential and much-discussed form of conceptualism. However, in AMG (2008), he abandons important components of the earlier view. Before turning to ³ Sellars and McDowell both think Kant needs to be corrected at some points, but disagree on just which, see McDowell (2009).

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

these changes, though, it is useful to distinguish three central claims that have remained constant throughout. Perhaps most centrally: Operation in Experience: In perceptual experience, conceptual capacities are already actualized; they are not merely actualized in judgements based on experience.⁴ In AMG, he glosses conceptual capacities variously as capacities ‘belonging to the understanding’, or ‘to reason’, and affirms Operation in Experience as follows: [W]e must suppose capacities that belong to [the understanding]—conceptual capacities—are in play in the way experience makes knowledge available to us . . . Avoiding the Myth requires capacities that belong to reason to be operative in experiencing itself, not just in judgement in which we respond to experience. (McDowell, 2008, 258)

A second, closely related claim, retained throughout, is: Same Kind of Content: judgements.

Perceptual experience has the same kind of content as

In AMG, McDowell describes perceptual experiences as intuitions and endorses Same Kind of Content in the following terms: ‘[T]he content unified in intuitions is of the same kind as the content unified in judgments: that is, conceptual content’ (2008, 265). Thirdly, McDowell adopts throughout a Fregean view of the content of judgements: Fregeanism: The contents of judgements are Fregean propositions: truthevaluable, abstract entities, with a logical structure, whose constituents are modes of presentation of objects, properties, or functions.⁵ Now, if we ask what kind of content judgements have, for McDowell, it is natural to turn to Fregeanism for an answer. It is natural to think, moreover, ‘the same kind’ in Same Kind of Content picks up on that answer, and so to infer that Fregeanism thereby generalizes pari passu to experience. Before AMG, McDowell endorsed this apparent implication: Propositionalism: Perceptual experiences have propositional content; specifically: contents taking the form of Fregean propositions. Propositionalism suggests, moreover, a natural interpretation of the conceptual capacities that Operation in Experience claims are actualized in experience. Gareth Evans (1982, 100–5) argued that the status of Fregean propositions as structured goes hand in hand with the fact that having such a proposition in mind is the joint exercise of several distinguishable cognitive capacities, which he dubbed ‘Ideas’. In entertaining the proposition That cup is blue, an Idea of mine of the relevant cup (in which the cup is thought of in a certain way) is jointly activated with an Idea of mine of the ⁴ McDowell speaks variously of conceptual capacities being ‘actualized’, ‘operative’, ‘drawn on’, or ‘in play’ in experience. I assume these come to the same thing. ⁵ For McDowell, Fregean propositions may well be object-dependent. Fregeanism does not imply a descriptivist view of singular thought or reference.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

,  ,  



property of being blue (in which blue is thought of in a certain way). The former Idea could also be activated jointly with an Idea of mine of red, say, thereby thinking That cup is red, and so on. In general, having in mind a Fregean proposition is, for Evans, the joint actualization, in a certain logical combination, of Ideas corresponding to each of the modes of presentation of an object, property, or function that are the constituents of the proposition. Now, if perceptual experiences have Fregean propositional content, and such content can be entertained in a subject’s mental life only through the joint activation of Ideas, this makes it extremely natural to identify the conceptual capacities that, according to Operation in Experience, is actualized in experience with Ideas. This identification of conceptual capacities in general, and of those operative in experience in particular, is explicitly endorsed in McDowell’s pre-AGM view (cf. McDowell 1998, 10–12). In AMG, however, McDowell rejects propositonalism.⁶ Arguing that any representational content would be propositional (whereby he seems to mean of the Fregean sort), he denies that experiences have representational content tout court (2008, 266–7). McDowell denies moreover that experiences have sub-propositional contents that can be identified as being fragments of, or potential constituents of, complete Fregean propositions (2008, 270). These revisions raise questions about how now to understand Operation in Experience. It seems hard to retain the identification of conceptual capacities with Evansian Ideas. If Ideas—capacities for thinking of things, in certain ways—are actualized in experience, how could that not bestow on experience a complete Fregean content, or at least a constituent thereof? If conceptual capacities are not Ideas, what are they? Questions also arise over Same Kind of Content. If ‘the kind of content’ such that experiences and judgements both have content of that kind is neither Fregean propositions, nor other kinds of propositions or representational contents (if any), nor constituents of Fregean propositions, then what could the relevant same kind be? I shall begin with the question raised over Operation in Experience. This will, in due course, lead us to that raised over Same Kind of Content. McDowell is committed, I take it, to a somewhat broadened conception of conceptual capacities that includes, but is not limited to, Evansian Ideas. Such a conception is not articulated very explicitly in AMG or other later work. However, two recurrent themes, or strands, in his glosses on the conceptual are, first, that it belongs to the understanding (2008, 257, 262, 265) and, second, that it is characterized by content having what he calls categorial unity and form (261, 263). McDowell does not explicitly argue that these conditions jointly suffice for a something to be a conceptual capacity, but I cannot see that he offers any further necessary conditions, not supposed to be entailed by those two. He thus seems to treat them as sufficient; I will here simply suppose they are.

⁶ His chief reason for rejecting propositionalism, it seems, is that it assimilates experiences too closely to beliefs, thereby threatening to move his view into the target range of his own arguments, in Mind and World, against Davidson’s coherentist claim that only a belief can be a reason for a belief, cf. McDowell (2008, 268–72).

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

The first strand, then, invokes a Kantian distinction between sensibility and understanding, and identifies conceptual capacities as ones belonging to the understanding. The understanding, for McDowell, is as a capacity paradigmatically manifested in our being responsive to reasons as such (McDowell, 2006, 128–34). In such responsiveness, we not merely think, or do, something for a reason, but in so doing are capable of reflecting upon whether the putative reason in question really is an adequate reason for thinking, or doing, that. Such a capacity for reflection on reasons, and for changing one’s view or course accordingly (if that is what reflection suggests), is the cash value McDowell finds behind Kant’s talk of the ‘spontaneity’ of the understanding. In belonging to the understanding, conceptual capacities are exercisable in mental acts responsive to reason in this way. The second strand, invoking categorial unity and form, requires a bit more settingup discussion. A useful starting point here, I think, is with the platitude that, whatever a conceptual capacity may be, its actualization should be a case of having something in mind—a case of something’s being presented to, or represented by, one—in one way or another. Let’s use ‘content’ in a broad sense so that a mental episode has content just in case it is an episode in which something is had in mind/ presented to/represented by whoever is the subject of the episode. ‘Content’ here is not definitionally restricted to propositional or representational content. To talk of the content of an experience is simply to talk of the experience in its aspect of being an experience of things. Even such anti-representationalists as Travis (2013), Brewer (2011), or Campbell would grant that experiences have content in this sense. (I shall assume that if one draws a distinction between the objects of one’s experience and how one experiences them, then a specification of content might be sensitive to facts of both kinds.) On this broad sense of ‘content’, our starting platitude can be rephrased as saying that the actualization of a conceptual capacity, in a mental episode, entails it has content. We may now ask: if the content of a mental episode is to be bestowed by the operation of conceptual capacities (as opposed to other candidate content-bestowing capacities, if any), what conditions does this content have to meet? At this point, Same Kind of Content can be invoked to deliver one such condition, viz. the content must be of ‘the same kind’ as that of judgement.⁷ Notice that, if this is a necessary condition for content-bestowing capacities to be conceptual, as I will assume it is, Operation in Experience entails Same Kind of Content. Yet what is the relevant ‘same kind’? As we saw, the revisions McDowell makes in AMG rules some familiar candidate answers out. He does not elaborate at any length about what the relevant ‘same kind’ positively might be. However, a central claim, as we noted, is that experiential content, like judgemental content, has categorial unity and form (2008, 261, 263). By categorial unity, he alludes to a unity that is of a kind with the unity of propositions, i.e. roughly, the unity consisting in the fact that ⁷ Strictly speaking, Same Kind of Content, along with Operation in Experience, only implies that the conceptual capacities operative in experience bestow ‘the same kind’ of content as judgement. I am here assuming (for simplicity, but inessentially to the overall argument) that Same Kind of Content is a special case of a generalization to the effect that any mental state in which conceptual contents are operative has the same kind of content as judgement.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

,  ,  



propositions, in spite of being structured entities, are not a mere list of elements but fit together logically so as to make an evaluable claim. Similarly, the notion of categorial form is supposed to be of a kind with, and presumably to be understood in terms of, the logical forms of propositions (cf. 2009, 127). How, though, should these notions of form and unity, explained in terms of propositions, apply to experiential content, if the latter is precisely not a proposition, or even a constituent thereof? I will here suggest the following reply, which McDowell does not explicitly endorse, but which seems to cohere with what he says. Although the most straightforward, and primary, way for a content to have categorial unity and form is for it to simply be a proposition, a content can also have these features, in a perhaps derivative, or secondary way, if it is not a proposition but essentially depends on, and is individuated in terms of, a certain relation in which it stands to a proposition or set of propositions. The suggestion would be that this latter, secondary option applies to experiential contents. These contents are to be individuated in terms of certain propositions, which the experiencer herself is in a position to grasp, and which the relevant experience makes it rational for her to endorse. Moreover, in their capacity of playing this individuating role, it matters that the relevant propositions, rationalized by the experience, are structured, having a logical form. That is why it is apt to think of the experiential content as having categorial form. The logical form of the rationalized propositions is non-redundant to bringing out some vital features of the rationalizing experiential contents. The suggested reading here is, of course, highly abstract and schematic. In Section 4, we will seek to flesh it out in terms of an example from Campbell, involving the experience of object identity across time, or across sensory modalities. As we shall see, in identifying just what this experience of object identity amounts to, Campbell himself invokes inferences, of such-and-such logical forms, to which the experience is held characteristically to entitle the thinker. For now, we may observe that the just-suggested way of individuating experiential contents, in terms of propositions rationalized, would be an alternative to some other options in the literature. For example, it contrasts with how a naïve-realist like Bill Brewer would individuate at least the fundamental experiential content of perception, viz. in terms of the physical objects experienced, and the point of view from which one experiences them, where point of view is understood in terms of such parameters as viewing angle, lighting conditions, etc.⁸ It also contrasts with an individuation of content in terms of the objects perceived and the properties perceptually attributed thereto, as might be proposed by a representationalist trading exclusively in Russellian contents.⁹ The outlined reading of categorial unity and form goes beyond what McDowell explicitly says, as we noted. It seems, though, to fit how he often sets out his view. ⁸ See Brewer (2011, 92–137). Brewer restricts ‘content’ to representational content, whereas, on our usage here, a specification of content is simply a specification of the experience in its aspect of being an experience of things (where this might include not just facts about the objects experienced, but also about how they are experienced). The qualification ‘at least the fundamental’ is included since Brewer holds there is also an aspect to perceptual consciousness, viz. what he terms ‘thick looks’, that need to be individuated in conceptual and perhaps propositional terms, paradigmatically exemplified by our perceptual recognition of things as belonging to various categories (Brewer, 2011, 121–2). ⁹ One such theorist might be Tye (2009).

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

When characterizing the content of experience, he does not go in for a specification in either the naïve-realist or Russellian representationalist style just indicated. Instead, he invokes precisely the knowledgeable judgements experience enables. Here is one illustrative example: ‘Though they are not discursive, intuitions have content of a sort that embodies an immediate potential for exploiting that same content in knowledgeable judgments. Intuitions immediately reveal things to be the way they would be judged to be in those judgments’ (McDowell, 2008, 267). To get a feel for the reading we have proposed, a rough analogy might be helpful, even one from a quite different domain. Let’s compare the contents of experiences and judgements, respectively, with the moral statuses of infants and adults, and consider the following view of the latter statuses: Adults, qua autonomous persons, have a special moral status, call it dignity, that no creature not (yet) a person could have. Still, infants do have a weighty moral status of their own, nonequivalent with dignity but not necessarily therefore morally second-class. The latter status, call it nobleness, essentially depends on, and, in part, is what it is in virtue of, the dignity that belongs to the person the infant normally would grow into. It would be a mistake—akin, perhaps, to the Myth of the Given—to try to secure an apt moral status for infants by looking merely at their already actualized, pre-personal properties (which is not to say, of course, that those properties are irrelevant or unnecessary to understanding how they can have nobleness). Although there is a specific kind of status of adults, dignity, that infants lack, they do have a vitally important kind of status, nobleness, in common with adults, since the dignity of adults entails (more or less trivially perhaps) their nobleness.

This moral view, if coherent, would offer an at least a partial analogy with the experience/judgement relationship, on the suggested construal, as follows. In each case, a certain normatively important status of a less intellectual being (the content of the experience/the moral status of the infant) is what it is in virtue of a similar, but more specific, status of a more intellectual being (the content of the judgement/the moral status of persons), to which the former being stands in some broadly functional/teleological relation of tending to be taken up/develop into. It is a good question just how deep the analogy might go (and how many grains of salt with which it should be taken). These must however be left for another occasion. For now, suffice it to gesture at the analogy as a way of conveying a rough, intuitive sense of the reading here proposed of how, for McDowell, experiential contents can be of a certain ‘same kind’ as propositions, having categorial unity and form, yet not be propositions.

3. Attention and the Rational Role of Experience The discussion of experience in Section 2 tacitly embodied a restriction to experience in its role of making judgements rational, a restriction reflecting the focus, in McDowell (and before him in Sellars), on this epistemic role of experience. In this section, I defend the claim that, in order for experience to play such an epistemic role, it must involve attention. More precisely, I will defend: Attentive Rationalizing: Perceptual experience of an object rationalizes a judgement about it only if it is an experience in which one attends to that object.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

,  ,  



By ‘an experience in which one attends to’ an object I mean one where it is an aspect of one’s overall perceptual consciousness that one attends to that object. Thus conscious attention is in question. I follow the widespread and plausible view that it is possible to attend to more than one object at a given time (it is commonly suggested one might at least attend to three or four, see, e.g. Scholl, 2009), and that, if attention is graded, one can attend to one object whilst attending even more to something else. In saying an experience ‘rationalizes a judgement’, e.g. the judgement that this thing before one is white, I mean that the experience occurrently exerts some rational pressure in favour of that judgement. This is not to say one would necessarily be rationally sub-par in not making the judgement (perhaps you could, quite rationally, refrain from even considering the question whether the relevant object is white). The point is, rather, that if you positively adopt some cognitive attitude towards the proposition in question that is incompatible with judging it true, such as consciously suspending judgement or rejecting it, you are under some rational pressure to discount your experience, by positing reasons to think your experience illusory or untrustworthy. A rational judgement, I will assume, can be regarded as one in which the subject is competently moved to make the judgement by such rational pressure. Now, on some views, perceptual experience is limited to what one attends to: there is no perceptual experience of an object apart from attending to it.¹⁰ If this is so, Attentive Rationalizing trivially follows. Thus, if one makes the contrary supposition that there is, or may well be, conscious perceptual experience of objects to which one does not attend, Attentive Rationalizing is if anything harder to defend. In this section I shall, therefore, make that supposition. Attentive Rationalizing may be motivated by reflection on imagined or empirically attested cases. A nice imagined case is Campbell’s story of ‘the sea of faces’ (2002, 8–9).¹¹ Suppose you are confronted with a small crowd, and an interlocutor asks for your view of the hair colour (say) of ‘that person’ before you. The indicated person is within your field of view, but your experience is as of a sea of faces; you do not attend to any one in particular, and you do not know which person is in question. Yet, when forced to guess, it turns out that you are right. Indeed, you are able reliably to give right answers for a range of the visible characteristics of the person in question. So you have a blindslight-like capacity for reliable answers about her. However, intuitively, when you venture, say, ‘That woman is red-haired’, you are not rational in making that judgement in the way you normally are in making perceptual judgements. Compare with a case where you consciously attend to the person in question, ‘highlighting’ her in your experience, now judging the same characteristics of her. Intuitively, there is difference in kind in the rationality of your judgements about her. Attentive Rationalizing fits the intuition here that the mere experience of an object is not enough for rational perceptual judgement about it.

¹⁰ For defence, see, e.g. Prinz (2011). ¹¹ In using Campbell’s ‘sea of faces’ case to support the importance of attention to the rational role of experience, I follow Smithies (2011).

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

Empirically attested cases in support of Attentive Rationalizing may be drawn from change- and inattentional blindness (for review of these phenomena, see, e.g. Rensink (2009)). Reflect on a case where you have been change-blind of some change, or inattentionally blind of some object, in your field of view, which then catches your attention and escapes change/inattentional blindness. Specifically, reflect back on your situation before the relevant change or object all of a sudden stood out for you. If you are like me, it will strike you that you would then have been quite irrational, making as it were epistemic leaps in the dark, if you had made judgements to the effect that those changes or objects were present before you. Moreover, this intuition, that you could not then reasonably have judged, going by vision, that those changes or objects were there, would seem closely related to the ease with which we accept the description of these cases as forms of blindness. It might be inveighed here that we accept that description because we think, perhaps rightly, that we completely lack conscious experience of those items. Yet suppose, as we do, that we do have experience outside of attention On that supposition, an attractive alternative explanation of our admitting these cases as forms of blindness is the noted intuition that we cannot make use of the experience to come to know of the relevant objects/changes.¹² Attentive Rationalizing permits us to allow perceptual experience of items towards which we are change-/inattentionally blind, without the uncomfortable implication that this experience places some rational pressure on us to accept the presence of those items before us, or else discount the experience as illusory, or hallucinatory. It is worth stressing that we are here reflecting on a rather special sequence of experience, viz. a case where you are at first change-/inattentionally blind of an object or change before then noticing it, realizing that it was there before your eyes all along. I do not mean to suggest that untutored reflection on any old perceptual experience would or should elicit acceptance of Attentive Rationalizing. On the contrary, it might well be that we are prone to think, upon untutored reflection, that experience of unattended objects ought to be available to us as input to knowledge and rational judgement pretty much as experience of attended objects are (with due allowance, to be sure, for the perhaps less determinate and detailed, or more vague, way in which unattended items are experienced). Some such conception might explain, in part, why change/inattentional blindness is liable to come as a surprise. The neurological disorder of extinction offers further relevant empirical cases (for reviews on extinction, see Driver and Vuilleumier (2001) and Riddoch et al. (2009)). Extinction patients have unilateral lesions to the parietal lobe. Lesions are typically on the left side, and for ease of presentation I shall pretend they always are. When a stimulus is presented on the contralesional, right side on its own, it is perceived in an approximately normal manner. Patients do not, then, have a major sensory deficit in their right hemifield (unlike, say, patients with hemianopia). Yet if another stimulus is simultaneously presented on the left side, the stimulus on the right will often not be ¹² For closely related arguments from change- and inattentional blindness to the conclusion that attention is necessary for perceptual information to be available to rational control of action, reasoning, or verbal report, see Smithies (2011). Smithies considers alternative accounts of inattentional and changeblindness, including amnesia and agnosia accounts, that I pass over here.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

,  ,  



noticed: the concurrent left-side stimulus ‘extinguishes’ the right-side one. The extinguished stimulus still receives considerable sensory-perceptual processing. For example, the meaning of an extinguished word may prime later decisions. Furthermore, if stimuli are concurrently presented on each side, and the right-side stimulus belongs to a category that tends strongly to attract attention, such as faces, the right-side stimulus is more likely to escape extinction than a neutral stimulus would have been. This exemplifies a range of evidence indicating that extinction is, at least in some of its central manifestations, an attentional deficit, in which left-side stimuli are excessively prone to attract or retain attention at the expense of rightside ones. Patients have often been assumed to lack consciousness of extinguished stimuli, yet, as Block (2001, 198–203) argues, if we make a conceptual distinction between phenomenal and access consciousness, it is not obviously ruled out that patients have phenomenal consciousness, and so perceptual experience, of extinguished items. Let’s suppose they do, and consider a patient undergoing a standard test, where stimuli sometimes are presented on the left side, sometimes the right, and sometimes on both sides, being asked to tell which condition obtains. If the patient, in a given case, suffers extinction and, accordingly, incorrectly judges that a stimulus appears on the left only, it seems we would not deem her irrational in doing so. This is so even if the subject lacks any belief to the effect that she may be subject to positive illusions or hallucinations in which things appear on her right although nothing is there. Yet, if her supposed perceptual experience of the extinguished stimulus on the right rationalized her in judging a stimulus to be present there, and she lacked any such beliefs about the risk of illusions or hallucinations, her judgement would seem rationally sub-par. Attentive Rationalizing prevents this untoward result. Recently, however, Siegel and Silins (2014, 2019) have argued that, assuming there is perceptual experience outside of attention, such inattentive experience can give reasons for beliefs, and both propositionally and doxastically justify beliefs, about the inattentively experienced objects. Their arguments might also be taken to support the conclusion that, contrary to Attentive Rationalizing, non-attentive experience can rationalize judgement (although Siegel and Silins do not explicitly make this further claim). I will argue, though, that their arguments do not disprove Attentive Rationalizing. Some of Siegel and Silins’s arguments for the optionality of attention for propositional justification rely on an analogy with beliefs. Consider beliefs held in long-term memory, that are recallable but not actually recalled at a given moment, i.e. that are not occurrent at that time. Siegel and Silins argue such beliefs may, at that time, constitute reasons for certain further beliefs, and propositionally justify them. For example, suppose I first make an appointment for a one-on-one meeting with Liza at noon, and some time later make a similar appointment with Frank. Knowledge of each appointment is stored in my memory (e.g. if asked ‘Do you have an appointment with Liza at noon?’, I will immediately affirm). Yet it is only when both show up outside my door at noon that I notice the conflict. Siegel and Silins (2014, 155) argue I could then rightly kick myself, admonishing myself that I had reason to believe I had conflicting appointments. Now, the kind of ‘mental grip’ I have here on the information stored in long-term memory, but not occurrently recalled, is

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

epistemically analogous, they suggest, to the kind of ‘mental grip’ we have on the contents of inattentive experience. If the former gives reasons and propositionally justifies belief, why not also the latter? I grant you are right to kick yourself in this example, and that there are various senses in which you should have realized the conflict, e.g. because you should have been more careful in keeping track of your appointments, or have failed in some epistemic virtue. I also grant that you possess evidence that you have conflicting appointments and, since evidence arguably constitutes reasons to believe, that you have reasons to believe that you do. If having evidential reasons to believe p suffices for having propositional justification for that belief, I even grant that you have such justification. I deny, though, that your stored knowledge or beliefs rationalize you in judging, at a given moment, that you have conflicting appointments at noon, independently of whether or not you are, at that moment, recalling any of that stored information, in the sense of occurrently bringing it to mind. To see this, suppose the thought strikes you, sometime before noon, that you have conflicting appointments at noon. When this thought strikes you, however, you are neither recalling your appointment with Liza, nor recalling that with Frank, nor is there any other specific commitment of yours that is occurring to you.¹³ In this situation, it seems you would not be rational in endorsing the thought that you have conflicting appointments. If you did endorse it, that would seem an eccentric leap, not rationally intelligible, at least not from your perspective at the time. This is not to say that you need occurrently to recall everything that might be rationally relevant to your judgement (for example, perhaps you do not need occurrently to think that Frank and Liza are distinct), only (to put it roughly) that some significant parts of what would be your reason should be occurrent to you. By analogy, Attentive Rationalizing does not require you to attend to everything that might be rationally relevant to your judgement, only to the object of your judgement. So if you would not be rational in making the relevant judgement in this scenario, then it seems the analogy Siegel and Silins rely on between our mental grip on information stored in long-term memory but not occurrently recalled and the contents of non-attentive experience, adds up to a case in favour of, rather than against, Attentive Rationalizing. Siegel and Silins (2014, 158–9) also argue that inattentive experience can doxastically justify belief, i.e. that beliefs can be based on such experience, and thereby qualify as well founded (assuming the experience propositionally justifies). In support of this claim they observe that behaviour might be responsive, in a putatively intelligent manner, to objects or facts that are inattentively sensed, such as when a driver makes apt adjustments to the wheel upon inattentively sensing a widening of a bend in the road ahead. They argue well-founded beliefs mediate between

¹³ This is not to say that you have forgotten about either appointment. For a wide range of cues, including being asked whether you have an appointment at noon with Frank/Liza, you would immediately recall the appointment on being presented with the cue. There can be such a wide range of cues even if the precise occurrent thought that strikes you here—that you have conflicting appointments at noon—is not itself such a cue.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

,  ,  



sensory-perceptual states and behaviour here. However, I doubt there are compelling grounds here for attributing beliefs, as mediating cognitive states distinct both from sensory-perceptual states on the one hand and low-level motor-control representations on the other. At least, there is good evidence that sensory input can smoothly fine-tune one’s ongoing action, through the dorsal visual stream, without passing through a layer of conscious belief (cf. Milner and Goodale, 2006). It might be replied that even action guided through the dorsal stream could be regarded as involving a layer of unconscious beliefs. However, even if this were admitted, and such unconscious beliefs were granted to be reliably based on inattentive visual input, it would not discredit Attentive Rationalization, which concerns the preconditions for rationalized judgements, where judgements are precisely conscious mental acts. Another argument Siegel and Silins (2014, 159) offer for thinking that beliefs might be based on inattentive experience is inspired by a case described by Martin. Archie is looking for a cufflink. Looking into a drawer, he fails to notice the cufflink lying there, and continues searching the room. Eventually he gives up and leaves for dinner. On the way to dinner, Archie agitatedly recalls his search. Having a relatively good visual memory, he recalls how things looked as he checked the drawer. Suddenly he realizes that the cufflink was lying there (Martin, 1992, 749–50). On Siegel and Silins’s construal, Archie is supposed not to attend to the object that is the cufflink when looking through the drawer.¹⁴ When he later judges that the cufflink was in the drawer, they argue, his judgement is based on his earlier inattentive experience of the same. The assumption here that we can later consciously recall seeing something we did not at the time attend to is highly controversial. In change- and inattentional blindness, unattended items are not typically available for later conscious recall. Similarly, although extinguished items in visual extinction may enter implicit memory, as indicated by priming effects, I am not aware of evidence that subjects later can consciously recall them (cf. Vuilleumier et al., 2002). Insofar as Archie did not attend to the cufflink, that might precisely be a reason to doubt that he will later be able to recall seeing it there.¹⁵ Even if that controversial assumption is granted, problems remain, stemming from the fact that Archie’s judgement, on his way to dinner, is immediately based on a form of visual episodic memory. How could that show that his judgement is (also) based on the earlier inattentive perceptual experience? Siegel and Silins do not go into this, and so I am led to speculate, but I can only see candidate routes here. First, perhaps Siegel and Silins intend the case so that, although Archie relies on an episodic memory, in that act of remembering his earlier perceptual experience, or a suitable part thereof, is relived or re-experienced, thus allowing his judgement to be immediately based on his (relived) experience after all. As against this proposal,

¹⁴ This supposition is not clearly built into the case as Martin conceives it. For Martin’s purposes, it is consistent to assume Archie attends to the cufflink, as long as he does not recognize it as such. ¹⁵ The likeliest cases of consciously recall of previously inattentively perceived items are perhaps ‘partial report’ tasks in Sperling-type paradigms (Siegel and Silins elsewhere rely on such examples). For recent arguments that attention is needed for recall even here (and references to the contrary view), see Mack et al. (2016).

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

however, Siegel and Silins do not argue that Archie fails to attend to the cufflink within the context of his episodic memory. The default supposition here would certainly be that he does attend it. If this is so, any relived visual experience of the cufflink would, by supposition, differ relevantly from the original one, viz. precisely in what attentional distribution that goes with or characterizes it. Second, Siegel and Silins might have some mediate form of basing and rationalization in mind. Perhaps they would argue as follows. ‘Suppose you start out believing P, infer therefrom that Q, and then later, believing Q, infer that R. Then, although your belief that R is immediately based on your belief that Q, we would, and should, also recognize it as being based on, and rationalized by, your belief that P, if only mediately. Analogously, Archie’s visual memory, on which his judgement is immediately based, derives from his previous visual experience. So there are grounds here too for saying his judgement is based on, and rationalized by, the latter, if only mediately.’ However, in the belief!belief!belief process, both transitions are of a kind. Both are rationally assessable, reason-responsive inferences. In the visual-experience!visualmemory!judgement process, they are not of a kind. The latter but not the former is a rationally assessable, reason-responsive act.¹⁶ It might be fortunate, useful, unpleasant, etc. to have, or lack, such-and-such visual episodic memories. We might be more or less rational in relying on them. Yet, on the face of it, we are not epistemically rational or irrational already by virtue of having, or lacking, such-and-such visual episodic memories. Siegel (2017) would perhaps disagree with this, as she thinks even perceptual experiences are rationally assessable. This is not the place to rebut her radical view on this matter. I remain unconvinced, though, that we should abandon the traditional, and (I think) compelling, view of perceptual experience, and visual episodic memory, as arational. If this is right, there is a pertinent structural difference between belief!belief!belief and visual-experience!visual-memory!judgement. Even if two linked, rationally assessible basing relations allow for a sense in which the third state is mediately based, in a rationally assessible sense, on the first, this gives no grounds for finding such structure in Archie’s case, where the first step to the visual memory is of a quite different, arational kind.

4. Attention as a Conceptual Capacity This section argues that attention qualifies as a conceptual capacity, in the sense relevant to McDowell’s recast conceptualism. That is to say that attention belongs to the understanding, and bestows content with categorial form and unity, on the reading of these conditions outlined in Section 2. Let’s consider, first, whether the capacity of attention belongs to the understanding. Since the notion of the understanding in play here derives from Kant, and since, ¹⁶ In saying the transition to the judgement is reason-responsive, I do not mean to commit myself to the controversial claim that the judgement is based on reasons provided by the relevant memory experience. I only mean that the rationality of the judgement is sensitive to reasons possessed by the thinker, e.g. having to do with the reliability of what she seems to be recalling.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

,  ,  



more generally, McDowell, following Sellars, has the ambition that his conceptualism emerges through a critical reading of Kant, it is pertinent very briefly to review here how Kant himself view attention and its relation to the understanding. Kant clearly regards attention as faculty of the understanding. At the outset of his treatment of understanding, in the first Critique, he offers ‘attention, its hindrance and consequences’ as a leading topic of ‘applied logic’, a discipline that is said to study the rules of the understanding, insofar as it is subject to contingent, empirically ascertained conditions (Kant, 1998, A54–5/B78–9). Attention is invoked again at a central juncture of the B deduction, where (to put it broadly) Kant argues that the understanding, in the guise of what he inter alia calls the productive imagination, affects sensibility, thereby enabling empirical knowledge of objects intelligibly ordered in space and time. Kant there claims that this sort of affection of sensibility by the understanding is something everyone can recognize in herself since ‘every act of attention can give us an example of this’ (B156–7, nt.).¹⁷ Central to Kant’s grounds for subsuming attention under the understanding is a conception of attention as an active power, subject to choice. Thus, in metaphysics lectures delivered as he was publishing the first edition of the first Critique, he describes attention as ‘consciousness with the power of choice’, a characterization that comes on the heels of an account of the distinction between sensibility and the understanding in terms of the former being passive, a case of receptivity, the latter ‘self-active’, having spontaneity (Kant, 1997, 29.877–8). This characterization links closely with McDowell’s favoured construal of the notion of spontaneity, on which it adverts to a sensitivity to, and responsibility in the light of, reasons, appreciated as such (cf. McDowell, 1998, 6, 1994, 4–5, 60). The power of choice for Kant plausibly involves such a sensitivity and responsibility to reasons. Now, attention clearly is exercisable in mental acts, responsive to reasons, grasped as such (which is not to deny that attention also can be passively drawn to stimuli). As Campbell observes: ‘We may have to appeal to the deepest aspects of an agent’s personal life in explaining why his conscious attention has just the focus that it does, and we have no way of recasting this causal-explanatory work in informationprocessing terms’ (Campbell 2002, 14). This is arguably no merely accidental feature of the capacity of attention had by reflective, yet limited thinkers such as us. If attention were forever at the mercy of what happened to be the most powerful outer stimulus or inner biological drive, it could not perform the job of putting our limited cognitive resources to work on topics or objects relevant to questions or ends we are setting for ourselves. That job—of securing that attention be paid to things that matter to us—seems vital to what we think of as attention. It shows up, e.g. in how we ask or even admonish people to pay it in such-and-such ways. If a mental capacity did not perform that job, we would, I think, reckon that capacity ipso facto differ in

¹⁷ Kant also repeatedly addresses attention in the Anthropology, where attention is given an interesting role in rendering otherwise ‘obscure’ representations clear (or ‘illuminated’) or even distinct, and as such ripe for knowledge (Kant, 2006, 7.135). He comes at least close to making attention necessary for knowledge, and affirms that the understanding must contain the faculty of attention (7.138). For a discussion of Kant on attention, congenial to this chapter, see Merritt and Valaris (2017).

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

kind and nature from what we consider the capacity of attention.¹⁸ This, to be sure, is just a stab in a direction of a defence of this claim. It suggests, though, that attention belongs to the understanding, in that it is actualizable in reason-responsive mental acts, in which attention is paid to things, because they matter to us. A clarification may be in order here. I have just been stressing the reasonresponsiveness of attending, while Section 3 defended Attentive Rationalizing, i.e. (to simplify slightly) that attention is needed for perceptual experience to rationalize judgement. Two questions must be kept firmly apart, though: (1) for what reasons does one attend to a perceived object, and (2) what is the rationalizing role of one’s attentive experience of that object? The answer to (1) may and perhaps even must include practical or telic considerations; the answer to (2) concerns the epistemically rationalizing role of experience. Though it is important to be clear about the distinction between these questions, they can of course be linked. One attends to an object, at least often and in part, out of interest in it, because one wants to know (more) of it. Specifically, the reasonresponsiveness of attention allows one’s manner of attending to be responsive to grounds for (to put it roughly) having another look or looking more carefully at things. Thus, we can see how the reason-responsiveness of attention is, at least, relevant to enabling (attentive) perceptual experience to meet the requirement of openness to self-critical scrutiny that McDowell places on would-be rationalizers (cf. e.g. 1994, 40, 2006, 128–34).¹⁹ The second strand we distinguished in McDowell’s gloss on conceptual capacities was that they bestow content with categorial unity and form. In our discussion of this claim, we set out from the platitude that actualizations of conceptual capacities are cases of having something in mind. They are cases of something’s being presented to, or represented by, whoever actualizes the capacities. Attention meets this platitude. In actualizing the capacity of attention, one attends to something, and thereby has it in mind, in that it is either presented to, or represented by, one. In our broad sense of ‘content’, the actualization of attention in a mental episode thus entails it has content. The question now is whether that content has categorial unity and form. On our reading of the latter phrase, that is to ask: is the relevant content either itself propositional or to be individuated in terms of certain propositions rationalized by the attentive experience, where the logical form of these propositions matter to their individuating role?

¹⁸ For a recent view of attention stressing its character of being, in rational subjects, under their rational control, and aiming to account for how that character could belong to the very nature of attention, see Watzl (2017). ¹⁹ It is a further step to argue that only attention can ensure that perceptual experience meets the requirements for rationalizing judgement and that only attention can do so at least in part because of the reason-responsiveness, etc. that make it a capacity of the understanding. To make this further step would, in effect, be to support and explain Attentive Rationalizing in terms of the features of attention that, as we here argue, make it a conceptual capacity, in the sense relevant to recast conceptualism. Taking that step would however require going into the connections between attention, reflection, and reasoning, in ways that easily would need a paper to itself. I must, then, leave these questions about the inner explanatory links between the view of attention as a conceptual capacity and Attentive Rationalizing for another occasion.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

,  ,  



Campbell has repeatedly stressed a feature of our attentive perceptual awareness of objects that, I shall argue, supports an affirmative reply. When we perceptually attend to an object, the identity of that object, across at least brief moments of time, or sensory modalities, or perhaps in other respects, is apparent to us (Campbell 2009, 654, 657–8). For example, looking at a car moving towards me, it is apparent to me that this thing now here is one and the same as an object that moments before was a little further away. Selecting a bottle from the fridge, it is manifest that this cold, round thing I am feeling is one and the same as this green, ampullaceous thing I am seeing. The apparent identity of the object here does not boil down to the mere identity of the object perceived at various times, or through different modalities. Consider a dense swarm of flies. At first I notice one fly, call him Billy, close to a curtain. Moments later, I see Billy again, near the middle of the swarm. Yet, in the meantime, I have, as we say, lost track of Billy: it is not evident to me that this fly here is the same as that I saw moments before close to the curtain (as it might well have been had Billy been flying unaccompanied before me). Similarly, if I am facing a tall, narrow wine rack, full of bottles, and I have to reach around to pull one out from behind, it may well not be obvious to me, in my overall perceptual experience, that this cold round thing I am grabbing is the same as this brown one I am looking at. As Campbell underscores (2009, 654–5), this feature of attentive perceptual experience of objects—that we experience objects as one and the same, across at least brief periods of time, or over modalities—seems not merely incidental; it is, arguably, essential for us to have an experience as of a moderately stable and coherent world of surrounding physical objects. What, then, does the apparent identity of an object, to which we are perceptually attending, come to? Campbell approaches this question in terms of a conception of perceptual experiences as individuated not merely by the objects they are of and their qualities, but also by the ‘standpoint’, as he calls it, from which one experiences them. The term ‘standpoint’ brings to mind the point of view in space from which one perceives things, and Campbell certainly intends the notion to capture something of the perspectival nature of perception. However, Campbell links the notion of a standpoint with the idea that our experience of an object, from a given standpoint, provides us with a corresponding understanding of a term referring to that object, where these understandings are individuated along Fregean lines: Suppose your understanding of t1 is provided by your experience of the thing from standpoint X, and your understanding of t2 is also provided by your experience of the thing from standpoint X. That is constitutive of your understanding the identity statement, ‘t1 is identical to t2’ as uninformative, as an instance of the logical law of identity. Understanding the terms in this way, you have the right to trade on identity in inferring from ‘t1 is F’ and ‘t2 is G’, to ‘something is both F and G’. In contrast, suppose your experiencing the object from standpoint X provides your understanding of t1, and your experiencing the object from some quite different standpoint Y provides your understanding of the coreferential term t2. This constitutes your understanding the identity statement, ‘t1 is identical to t2’ as informative; it is not merely an instance of the logical law of identity . . . The points in this paragraph are all laid down in advance of any substantive description of the notion of a ‘standpoint’. (Campbell, 2009, 658)

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

Campbell, then, invokes certain propositional contents—specifically, certain logical forms of certain propositions, and corresponding patterns of valid inference—to characterize the notion of a standpoint he sees as partially individuative of our perceptual experience of things.²⁰ The propositional contents invoked, with the relevant forms and corresponding inferential potentials, are propositional contents that the experience, from the relevant standpoint, entitles one to or makes rational. Let’s suppose this Fregean characterization could be construed as purporting to individuate standpoints, in the sense of articulating part of what makes them the features of our attentive experience that they are. Now standpoint is partly individuative of the content of experience (on our broad sense of ‘content’). It would follow that the content of experience is at least partly individuated by the cited Fregean characterization. Thus, the content of attentive experience of objects would turn out to have categorial unity and form, on our reading. It is not clear, though, whether Campbell intends the cited Fregean characterization to be individuative of standpoint in the sense of articulating what makes something the standpoint it is. This depends on how we are to interpret his remark that the Fregean condition is ‘laid down in advance of any substantive description’ of standpoint. On an alternative interpretation of this remark, and its wider context, he reasons as follows. Let a ‘Substantive Account’ be an account of what makes something the standpoint it is that appeals to various more or less familiar parameters of the perceptual situation, such as the perceiver’s vantage point upon the object, the lightning conditions, general visual acuity, and the like, but not to propositions or inferences made rational by the experience, from the relevant standpoint.²¹ The Fregean condition that is ‘laid down in advance’ does not tell us what standpoints are. Rather, it articulates a condition of success for a Substantive Account, viz: that it will turn out that sameness of standpoint upon an object makes available propositions about that object wherein the identity of that object is uninformative, entitling us to trade on the identity of the object, as per the Fregean condition.

If the content of experience is adequately characterized by the objects of experience and their qualities, as they are perceived from a standpoint specified by a Substantive Account, so understood, it would follow that we have not secured a categorial unity or form for that content. As Campbell recognizes, though, it is far from straightforward to articulate a Substantive Account. For one thing, sameness of spatial perspective on an object is neither necessary nor sufficient for sameness of standpoint. It is not necessary, as Campbell observes, since the sameness of an object might be perceptually manifest to me even as I move in relation to it, or it in relation to me, in which case I perceive it ²⁰ Campbell does not explicitly mention attention in this context, but, since he takes conscious attention to an object to be needed to understand a perceptual demonstrative referring to it, it seems that, when he says that ‘your understanding of t1 is provided by your experience of the thing from standpoint X’ he must implicitly be assuming your experience of the thing to be attentive. ²¹ This is how Brewer (2011, 95–9) specifies what corresponds on his view to Campbell’s notion of standpoint, insofar as it is construed as a third relatum on the relation of awareness, beside object and subject of awareness. Brewer (96, n. 2) explicitly does not follow Campbell in laying down a Fregean condition on the individuation of this third relatum.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

,  ,  



from one and the same standpoint. Sameness of perspective is not sufficient either. I may fail to keep track of an object, not because of any movement on its or my part, but because of various sorts of distractions, say from objects moving nearby, or even from engaging in a concurrent non-visual, attention-demanding task.²² What goes for spatial perspective goes for other parameters of perception, like lighting conditions and acuity. An object I am looking at might manifestly remain one and the same before me, even as the light goes darker or as my acuity drops. Thus sameness of standpoint, over time, must be compatible with shifts in these parameters. Conversely, sameness of these parameters, even jointly with spatial perspective and with each other, does not guarantee that one will keep track of an object, since, again, non-visual, attentiondemanding tasks can disrupt one’s ability visually to keep track of the object. Campbell acknowledges, in effect, that a general, explicit articulation of a Substantive Account is not readily forthcoming, but claims such an articulation is not needed. We have an implicit grip on what standpoints are, he argues. In a given case, we will often be able to point to specific features of perceptual parameters (vantage point, lighting, distance, etc.) that are constitutive of someone’s standpoint, in that case. Even if we cannot articulate relevant features of a given standpoint, we can often imagine them. However, Campbell might here seem to be confusing questions about our epistemic grip on standpoint, as ordinary reflective perceivers, with questions about the proper metaphysical account of what makes something the standpoint it is. The implicit abilities Campbell invokes do not allay the suspicion that a general individuating account of what standpoints are, couched in terms of the noted familiar sorts of perceptual parameters, will take the form of an unsystematic disjunction of different conjunctions of determinates for these parameters. By contrast, the Fregean condition Campbell ‘lays down in advance’ is nicely unified and systematic. This speaks in favour of the view that it is the latter, rather than the unsystematic disjunction, that tells us what makes something the standpoint it is. Is there a way of accepting this conclusion and still validating Campbell’s ambition that the account of standpoint will ‘show how we can characterize a way of experiencing an object without appealing to . . . the idea that “ways” are characterized by associated representations’ (Campbell, 2009, 659)? If ‘associated’ is taken broadly, the answer must be ‘No’. An individuating link with propositional contents and their forms, as articulated in the Fregean condition, would surely count as some form of association. However, if ‘associated’ is taken more narrowly, as implying that experiencing something from a certain standpoint (or in a certain ‘way’, as Campbell puts it here) is already a matter of entertaining propositional or other representational contents about it, then perhaps the answer can be ‘Yes’. At least, if McDowell’s recast conceptualism, on the reading we have proposed, has any promise, that view may be invoked here. Although experiential content (via, in this case, the factor of standpoint) is individuated in terms of propositions rationalized, that content need not itself be propositional or even representational, or so the recast conceptualism proposes. In other words: that recast conceptualism is an attractive bet for one who

²² For a study of how visual object tracking can be disrupted by distractors, or concurrent non-visual tasks, such as judging pitch of tones, see Tombu and Seiffert (2008).

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

accepts our conclusions about standpoint, yet wants to retain as much as possible of Campbell’s anti-representationalism.²³

5. Conclusion The main claims of Sections 3 and 4 respectively were (i) that experience rationalizes judgement only if the capacity of attention is operative in experience in a certain way, so that one consciously attends to the object one judges about, and (ii) that the capacity of attention is a conceptual capacity, in the sense of that term relevant to McDowell’s recast conceptualism. Their conjunction may be dubbed attentional conceptualism, in as much as they imply that experience rationalizes judgement only if conceptual capacities, in that sense, are operative therein.²⁴ Attentional conceptualism offers a novel route to this traditional conceptualist contention, resting on considerations that are at least non-trivially related to those advanced by McDowell or Sellars, and that should have appeal far outside the ranks of cardcarrying conceptualists.²⁵

²³ What light would the categorial unity and form of attentive experience throw on Attentive Rationalizing? Now, if (a) experiential content needs to have categorial unity and form to rationalize judgement, and (b) only attentive experience has such form, Attentive Rationalizing follows. We have not, though, defended (a) or (b). We have not even defended a certain weakening of (b) to the effect that attention is needed for experience to have categorial unity and form on the ground outlined in the text, viz. of presenting object identity (for discussion of related questions within psychology, see e.g. Scholl (2009)). If attention is not needed here, the argument for Attentive Rationalizing from (a) and (b) won’t get off the ground. It would not, though, disprove Attentive Rationalizing, since attention may, for other reasons, be required for perceptual experience to rationalize judgement, cf. nt. 19 above. ²⁴ Objection: ‘ “conceptual capacities” is plural, while “the capacity of attention” is singular. So (i) and (ii) does not imply that conceptual capacities must be operative.’ Reply: First, although ‘capacities’ perhaps conversationally implicates ‘two or more’, a strict implication to this effect has not been made explicit or motivated, as an official part, right from the outset, of the core conceptualist claim that conceptual capacities must be operative. It is true that, on McDowell’s pre-AMG, propositionalist conceptualism, at least two such capacities would have to be operative, since experience was supposed to have structured propositional content, the possession of which was supposed to manifest the joint actualization of two or more Evansian Ideas. That argument, though, falls with the fall of propositionalism. Second, the capacity of attention can be regarded as a super-capacity with sub-capacities, individuated in various more or less specific ways: e.g. capacities to attend to objects, to attend to several objects at once, to attend to this particular thing before me, to attend to properties, etc. This is, arguably, not altogether unlike how the capacity for conceptual thought, more generally, can be considered as a super-capacity with various subcapacities, individuated in various more or less specific ways. The case here presented for attentional conceptualism can, arguably, be fleshed out with more specific arguments for thinking (1) that several such sub-capacities need to be operative, for experience to rationalize judgement, and (2) that each of these subcapacities qualify as conceptual. This must however be left for another occasion. Third, even if the latter project were to fail, and ‘conceptual capacities’ should be construed as requiring ‘at least two’, it would be interesting to see how very close one can come to establishing a form of conceptualism through the route here sketched. ²⁵ Acknowledgements: I have been much helped by discussion with, or comments from, Bill Brewer, John Campbell, Jonas Jervell Indregard, Jonathan Knowles, Hemdat Lerman, Mike Martin, Jessica Pepp, Thomas Raleigh, Camilla Serck-Hanssen, Kristoffer Sundberg, Pär Sundström, Tom Stoneham, Sebastian Watzl, and Timothy Williamson. I thank audiences at the Norwegian University of Science and Technology, the Centre for the Study of Mind in Nature, and the Universities of York and Umeå. My research has been supported by RCN grants 217425 and 240645.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

,  ,  



References Block, Ned (2001) ‘Paradox and Cross Purposes in Recent Work on Consciousness’, Cognition 79: 197–219. Brewer, Bill (2011) Perception and Its Objects, Oxford: Oxford University Press. Campbell, John (2002) Reference and Consciousness, Oxford: Oxford University Press. Campbell, John (2009) ‘Consciousness and Reference’, in B. P. McLaughlin (ed.), The Oxford Handbook of the Philosophy of Mind, Oxford: Oxford University Press. Campbell, John (2012) Wittgenstein on the Role of Experience in Understanding Language Wittgenstein and the Philosophy of Mind, Oxford: Oxford University Press. Driver, Jon and Vuilleumier, Patrik (2001) ‘Perceptual Awareness and Its Loss in Unilateral Neglect and Extinction’, Cognition 79(1–2), 39–88. Evans, Gareth (1982) The Varieties of Reference, Oxford: Oxford University Press. Kant, Immanuel (1997) Lectures on Metaphysics, Cambridge: Cambridge University Press. Kant, Immanuel (1998) Critique of Pure Reason, Cambridge: Cambridge University Press. Kant, Immanuel (2006) Anthropology from a Pragmatic Point of View, Cambridge: Cambridge University Press. Mack, Arien, Erol, Muge, Clarke, Jason and Bert, John (2016) ‘No Iconic Memory without Attention’, Consciousness and Cognition 40, 1–8. Martin, Michael G. F. (1992) ‘Perception, Concepts, and Memory’, Philosophical Review 101(4): 745–63. McDowell, John (1994) Mind and World, Cambridge, MA: Harvard University Press. McDowell, John (1998) ‘Lecture I: Sellars on Perceptual Experience’, Journal of Philosophy 95(9): 431–450. Page references to reprint in McDowell 2009. McDowell, John (2006) ‘Conceptual Capacities in Perception’, in G. Abel (ed.), Kreativität, Hamburg: Felix Meiner Verlag, 1065–79. Page references to reprint in McDowell 2009. McDowell, John (2008) ‘Avoiding the Myth of the Given’, in J. Lindgaard (ed.), John McDowell: Experience, Norm and Nature, Oxford: Blackwell, 1–14. Page references to reprint in McDowell 2009. McDowell, John (2009) Having the World in View, Cambridge, MA: Harvard University Press. Merritt, Melissa and Valaris, Markos (2017) ‘Attention and Synthesis in Kant’s Conception of Experience’, Philosophical Quarterly 67(268): 571–92. Milner, David and Goodale, Melvyn A. (2006) The Visual Brain in Action, 2nd edn, Oxford: Oxford University Press. Prinz, Jesse (2011) ‘Is Attention Necessary and Sufficient for Consciousness?’, in C. Mole, D. Smithies, and W. Wu (eds), Attention: Philosophical and Psychological Essays, Oxford: Oxford University Press, 174–204. Riddoch, M. Jane, Rappaport, Sarah J., and Humphreys, Glyn W. (2009) ‘Extinction: A Window into Attentional Competition’, in S. Narayanan (ed.), Progress in Brain Research, Vol. 176, New York: Elsevier, 149–59. Rensink, Ronald (2009) ‘Attention: Change Blindness and Inattentional Blindness’, in W. Banks (ed.), Encyclopedia of Consciousness, New York: Elsevier, 47–59. Russell, Bertrand (1912) Problems of Philosophy, Oxford: Oxford University Press. Russell, Bertrand (1914) ‘On the Nature of Acquaintance. III. Analysis of Experience’, Monist 24(3), 435–53. Scholl, Brian J. (2009) ‘What Have We Learned about Attention from Multiple-Object Tracking (and Vice Versa)?’, in D. Dedrick and L. Trick (eds), Computation, Cognition, and Pylyshyn, Cambridge, MA: MIT Press, 49–77. Sellars, Wilfrid (1949) ‘Acquaintance and Description Again’, Journal of Philosophy 46(16), 496–504.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

Sellars, Wilfred (1956) ‘Empiricism and the Philosophy of Mind’, Minnesota Studies in Philosophy of Science 1, 253–329. Sellars, Wilfrid (1968) Science and Metaphysics: Variations on Kantian Themes, London: Routledge. Siegel, Susanna (2017) The Rationality of Perception, Oxford: Oxford University Press. Siegel, Susanna and Silins, Nicholas (2014) ‘Consciousness, Attention, and Justification’, in E. Zardini and D. Dodd (eds), Contemporary Perspectives on Scepticism and Perceptual Justification, Oxford: Oxford University Press. Siegel, Susanna and Silins, Nicholas (2019) ‘Attention and Perceptual Justification’, in A. Pautz and D. Stoljar (eds), Blockheads! Essays on Ned Block’s Philosophy of Mind and Consciousness, Cambridge, MA: MIT Press. Smithies, Declan (2011) ‘Attention Is Rational-Access Consciousness’, in C. Mole, D. Smithies, and W. Wu (eds), Attention: Philosophical and Psychological Essays, Oxford: Oxford University Press, 247–73. Tombu, Michael and Seiffert, Adriane E. (2008) ‘Attentional Costs in Multiple-Object Tracking’, Cognition 108(1): 1–25. Travis, Charles (2013) Perception: Essays after Frege, Oxford: Oxford University Press. Tye, Michael (2009) Consciousness Revisited: Materialism without Phenomenal Concepts, Cambridge, MA: MIT Press. Vuilleumier, Patrik, Schwartz, Sophie, Clarke, Karen, Husain, Masud, and Driver, Jon (2002) ‘Testing Memory for Unseen Visual Stimuli in Patients with Extinction and Spatial Neglect’, Journal of Cognitive Neuroscience 14(6): 875–86. Watzl, Sebastian (2017) Structuring Mind: The Nature of Attention and How It Shapes Consciousness, Oxford: Oxford University Press.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

PART III

Reference

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

9 Acquaintance as Grounded in Joint Attention John Campbell

1. The Role of the Shared Language in Cognition The later Wittgenstein (1953), and Benjamin Lee Whorf (1941), forcefully articulated pictures of the mental life on which the subject’s grasp of a shared language—a language like Chinese or German—is what grounds the subject’s ability to think about the world. In Wittgenstein’s work, the argument is that there being such a thing as the subject going right or wrong in the use of a word or concept cannot be grounded in the psychology of the subject; in particular, it can’t be grounded in the subject’s knowledge of some rule governing the use of the concept. You and the subject might agree on a form of words as expressing the rule governing the use of the concept. But if the subject goes off the tracks—as it seems to you—and you try to appeal to the rule to get the subject back in line, it turns out that the appeal to the rule is powerless in the case in which it’s needed. For the subject may agree with you that this form of words does indeed express the rule that they are following, but insist that their use of the word in this particular case is fully in accord with the rule. In fact, the subject may disagree with you about the application of the concept in arbitrarily many cases, but insist that they are correctly following the rule, as stated using that form of words. The trouble is that the only way you have of explaining what the verbal expression of the rule means is by giving the application of the word in particular cases. Suppose you say, ‘the concept applies to an object if and only if that object is F’. How are you to explain what you mean when you say that an object is F? In the end, the only explanation you have is by giving examples: you say, look, this is F and that is F and that one is not F, and so on. So the appeal to norms as grounded in an individual’s understanding of the rules governing a concept is just a mistake. The only way in which it can happen that there is such a thing as going right or wrong in your use of a concept is that you are using a part of a social practice in which there is widespread agreement on the application of the concept to particular cases. In Whorf, the argument against individualism about the mental life is empirical. The argument is that the shared language one speaks dictates which concepts one can grasp. Speakers of different languages may organize their conceptualizations

John Campbell, Acquaintance as Grounded in Joint Attention In: Acquaintance: New Essays. Edited by: Jonathan Knowles and Thomas Raleigh, Oxford University Press (2019). © John Campbell. DOI: 10.1093/oso/9780198803461.003.0010

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

of the world quite differently, because an individual’s grasp of concepts is constituted by their understanding of the shared language. These ideas about the priority of the shared language in constituting the mental life of the subject were generally abandoned from about the late 1950s on. There were three reasons for this. One was that both approaches seemed to lead to an implausible kind of linguistic idealism, a picture on which the world was merely the shadow of language; Wittgenstein had always attempted to resist this apparent implication of his picture, whereas Whorf embraced it with some verve. Another was the apparent implication of limits on what the subject could conceptualize, limits dictated by the limits of the shared language the subject uses. And it was not obvious that there really were any such limitations on the possibilities of thought. The third reason was the rise of cognitive science, which became increasingly self-conscious about its use of a quite different, individualistic way of grounding mentalistic content. Suppose you find a set of cells in some part of the visual cortex and want to understand their operation. You ask: (a) which external phenomena cause these cells to fire? And let’s suppose you give the answer, for instance, ‘the cells fire in response to the presence of characteristic F (for example, motion, or particular colours) at place p’. And you also ask (b) ‘Is it adaptive to have these cells firing in response to F-ness at p?’. If it makes sense to suppose that the organism is able to make adaptive use of cells firing in response to F-ness at p, then it’s compelling to say that the cell firings represent the presence of F-ness at p. Cognitive science is deeply committed to this commonsensical conception of brain representation, and the correlative individualism about mentalistic content is generally taken for granted now. In the study of language, for example, it is usually assumed that the working of a shared language is grounded in the brain representations of the individuals who speak the language, where these brain representations are understood in the causal-teleological way I just indicated. In the last couple of decades, however, there has been a wary return to ideas about the constitutive role of the shared language in cognition. One of the most theoretically general and intriguing approaches has been provided by Michael Tomasello’s work. Here’s Tomasello explaining how to approach the source of distinctively human aspects of cognition, which he argues are grounded in the distinctively human capacity for collaboration and cooperation: In the child-development literature, the earliest collaborative activities are often called ‘joint attentional activities’. At about nine months of age, infants begin to do things with adults like roll a ball back and forth or stack blocks together—activities that involve a very simple joint goal . . . By all indications—including several experiments that looked quite carefully for it— great apes do not engage in joint attention . . . In our several collaboration studies with great apes, they have never made any attempt at overt communication to establish joint goals and attention, whereas human children engage in all kinds of verbal and nonverbal communication for forming joint goals and attention and for coordinating their various roles in the activity . . . The species-unique structure of human collaborative activities is that of a joint goal with individual roles, coordinated by joint attention and individual perspectives. (Tomasello, 2009, 68–74)

Now on the face of it, much of Tomasello’s elaboration of this arresting idea aims ultimately to ground joint attention in the psychologies of the individuals: in their

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

     



beliefs about one another’s beliefs, and their individual desires. What I want to propose, though, is that this way of developing the key idea about collaboration and cooperation as the foundation of distinctively human modes of cognition loses something of the force of the insight. We ought to be taking joint attention—a social phenomenon, involving at least two subjects and a shared task—as more primitive than these individualistic beliefs about beliefs. If we take the capacity for joint attention as in this way relatively primitive, we can recognize it as grounding our capacity for linguistic communication, and our imaginative understanding of one another. The acquaintance with objects that is the basis for referential communication emerges first in the context of joint attention.

2. Acquaintance as Grounded in Joint Attention It is certainly a natural idea that communication is grounded in our capacity for joint attention. By ‘joint attention’ here I mean the kinds of phenomena that first begin to appear in humans when they are around one year old. Up to this point, the child will characteristically have been heavily engaged in one–one interactions with its caregivers. But the kind of thing that emerges at one year is when the child and the caregiver are focused on some third thing, and in effect are communicating about it. This can happen because the child first looks at the caregiver, catching their eye, then have drawn attention, turns to look at the target, perhaps pointing, taking the caregiver’s attention with them. Or it can happen that the target—for example, a loud explosion—is so salient that it draws the attention of both child and caregiver independently, who then look at one another. In either scenario, child and caregiver will typically oscillate attention back and forth between the object and one another. Tomasello writes, ‘However it is best characterized, the attentional loop [between two subjects] initially is made possible by having a joint goal . . . Later in life, infants can enter into joint attention without a joint goal’. Now these types of scenario obviously form the basis for the ordinary learning of language, from childhood through to old age. Tomasello writes: Human cooperative communication thus evolved within the bounds of collaborative activities because these activities provided the needed common ground for establishing joint topics . . . As Wittgenstein first noted, I may point to a piece of paper, its color, its shape, or any of its many different aspects, depending on the lebensform (form of life) in which the communicative act is embedded. Making contact with some lebensform—a collaborative activity would be a prototype, perhaps—grounds the act of pointing in a shared social practice, which gives meaning to the otherwise empty gesture. And without this grounding, conventional communication using ‘arbitrary’ linguistic symbols is simply noise. (Tomasello, 2009, 73–4)

The big difference between this kind of picture and Wittgenstein’s emerges when we reflect on the role that is being given to perception and perceptual attention by Tomasello. The whole thrust of Wittgenstein’s discussion is to downgrade the role that perception has traditionally been given, by empiricists from Locke and Berkeley through to Russell, in our understanding of language. As Wittgenstein put it, when I apply a word to an object, I do not do so on the basis of a perception that displays

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

the object meeting the condition for that word to apply to it: ‘When I obey a rule, I do not choose. I obey the rule blindly’ (1953, §219). Crispin Wright gives the following explanation of Wittgenstein’s picture here: Basic cases—where rule-following is ‘blind’—are cases where rule-following is uninformed by anterior reason-giving judgement—just like the attempts of a blind man to navigate in a strange environment. In such a case one follows a rule ‘without reasons’ in the precise sense that one’s judgements about the input condition for correct application of the rule are not informed by the exercise of concepts other than that which the rule concerns—that is, the concept whose expression the rule regulates and grasp of which consists in competence with that very expression. (Wright, 2007, 496)

Wittgenstein draws an analogy between the alleged role of ostension in explaining meaning and the role of ostension in explaining the rules of chess. Can we, for example, explain what it is for a piece in chess to be the king, merely by pointing to a particular piece and saying, ‘This one is the king’? Wittgenstein’s point is that you could have learnt the rules of the game without having encountered any of the physical pieces. Someone who has learnt the rules up to that last point, the shape of the king, will learn something from seeing the piece. But the shape of the king does not in any sense explain or justify the rules for the use of the king (1953, §31). The shape of the king can be varied arbitrarily while keeping constant the rules for the use of the king. Suppose you had the rules of chess explained to you before you had seen any of the pieces, as Wittgenstein supposes. You would experience no sense of illumination on finally seeing the shape of the king: ‘So that’s why it can move only one square at a time!’. In contrast, however, in our ordinary learning and communication in joint attention scenarios, it is because we are perceptually attending to the object that we know why we are using the words as we do. Recall Wittgenstein’s builders, who issue and respond appropriately to commands such as ‘Block!’, ‘Beam!’, and ‘Slab!’. Suppose we had among the builders one who does not have perceptual experience at all, but has a blindsight-style capacity to respond correctly to such commands, and can give such commands herself. As and when such a builder achieves conscious experience of her surroundings, and becomes able to consciously attend to various aspects of the blocks and slabs, she will achieve an understanding of why the various commands are used as they are: why blocks and slabs can be stacked, for example, but generally not together, and so on. This is quite unlike the case in which a practised chess player who had never seen a physical piece finally gets to see the shape of the king. This kind of approach allows us to see how Wittgenstein could be right about the fundamental nature of the shared language in cognition, while still holding on to the basic idea that a knowledge of reference and truth conditions is basic to a grasp of language and thought. In the Investigations, the picture is that all there is is the use of words; the idea that one has further to know what the terms stand for is taken to be part of an individualistic conception of understanding, on which an individual’s perceptions, outside of any social context, can ground their knowledge of reference. But a picture which highlights the basic place of joint attention in our understanding of concepts and language can acknowledge that the social context of a shared

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

     



language is required for thought, while still finding a basic place for knowledge of what words stand for. This, however, raises a further question. In Wittgenstein the pressure to recognize the role of the shared language comes from the abandonment of the idea that there’s a basic place for notions of reference and truth in our understanding of language. Once you’ve abandoned that idea, you face the problem of explaining how there can be such a thing as going right or wrong in your use of a word, since you’re not appealing to the idea of the subject’s grasp of reference and truth conditions as constituting what the subject has to keep faith with in the use of a term. It’s here, in trying to reconstruct the difference between getting it wrong and getting it right, that Wittgenstein finds the place of the shared language. But if we have a picture that takes the shared engagement with the world in joint attention as fundamental to thought and language, and knowledge of reference is provided by the individual’s attending to the object (in the context of joint attention), then why exactly does the shared language matter? If we can have an acquaintance with the object sufficient to provide knowledge of reference, why does it matter whether or not it happens in the context of joint attention? In fact, you might make the point even more strongly. Suppose you hold, with Tomasello, that joint attention has to be grounded in the individual psychological states of each of the attenders. So I have to know what you are attending to, you have to know what I am attending to, I have to know that you know what I am attending to, and so on. Then joint attention seems to rest on a rich capacity for thought, and perhaps language, that is not itself grounded in exercises of joint attention, but rather in the autonomous psychology of the individual. So can anything be left of the idea that the shared language is fundamental to thought and language?

3. The Mind’s Need to Represent One of the most intriguing, puzzling, and original ideas in Imogen Dickie’s Fixing Reference is that the mind has a ‘basic need to represent things outside itself ’ (p. 103). This is compared to other ‘basic needs’ that humans are said to have, such as: ‘basic emotional needs, for example, the need to avoid loneliness; the need to conform; the need to feel secure’. Before looking at any of the work that Dickie asks of this notion of a ‘need to represent’, I want to just dwell on the very idea for a moment. You might compare the idea of a ‘need to represent’ with the idea of a ‘need to read’. Some people do seem to be ‘print addicts’: on waking, they scan their watch, then the side of the toothpaste tube and the carton of breakfast cereal; on the way to work they inspect the advertisements they’ve already passed a hundred times; and on reaching the office, hooray, thousands of documents (O’Brien, 1968, 237–8). Now you might think of the mind’s ‘need to represent’ on the model of this kind of thing; it’s a need that perhaps isn’t often pointed out, but that some people do seem to have. There is surely a lot of individual variation, though. Some people read only when they need to use, for example, an instruction manual to help them put something together. And similarly, some people might represent only when some practical purpose, such as engaging in conversation, requires them to do it (as I understand him, Richard Heck,

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

in his comments on the book, suggests something like this). And some people manage without reading at all; couldn’t there be, similarly, some people who could manage contentedly without ever representing at all? After all, don’t some states of trance or meditation seem to aim at an absence of representation of anything external, and to regard such a state as a corrective to our ordinary reflexive representing? In fact the analogy with reading suggests that we might think of ordinary human representation as a reflex, leaving open the question whether it really is a need we have. Just as we find it hard (in the Stroop task, for example) to suppress the tendency to read any writing in our visual field, so we might find it hard to suppress our tendency to represent things outside the mind. But that doesn’t mean that it’s correctly described as a ‘need’ that we have, rather than a reflex that we’d be better off without. I think that this matters for Dickie’s account because the notion of a ‘need to represent’ is given a lot of work to do in grounding the normativity of reference and justification. And I want to propose a way of filling in the gap here. I want to suggest that we should think of the ‘need to represent’ as grounded in a distinctively human concern: the need to cooperate with others. As I understand it, the basic move in the book is to connect the fact that a singular belief is about a particular object, to the ways in which the belief is justified. The idea is that the facts about reference explain the facts about how the belief is justified, and the facts about how the belief is justified explain the belief ’s making the reference that it does. The natural thought here is that this is viciously circular. Dickie argues that the circularity is virtuous, but acknowledges that this needs some elaboration: ‘Vicious circles are all alike, but a virtuous circle is virtuous in its own way’ (2015, 113). She draws the analogy with a biconditional about what it is to have ‘legitimate credentials’. We can say: ‘X is a legitimately accredited expert if and only if X is approved by a legitimate credentialing institution’. Here the direction of explanation runs both ways, she says (112). But what stops this from being viciously circular is that we have a ‘story about the substructure that sustains the rich reading of the biconditional’ (112). This, I take it, would be a story that explains the basic point of having experts at all: the kind of work that they do in designing bridges or whatever it is, the things that an expert ought to be able to do for us. Consequent upon their role in designing bridges—what we need from the expert—we have the roles of accreditation and credentialing institutions. So far as I can see, the main thing that is missing from Fixing Reference is anything analogous to the explanation of what an expert ought to be doing for us in designing bridges, making them so they can take heavy loads, withstand wind, and so on. What does the activity of representation do for us? Why is it properly described as a need that we have, rather than a neurotic reflex that we would be better off suppressing? This is where I’m suggesting that we could appeal to the distinctively human phenomena of collaboration and cooperation. It’s because we have a human need to engage in cooperative tasks—that’s how humans work—that we need to be able to represent external objects. We wouldn’t be able to engage in joint tasks unless we could represent the objects of those tasks. When Dickie makes her key move to establish that the direction of explanation runs in both directions, from the ascription of reference to the structure of justification, and conversely, she seems to think that

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

     



we can do without any broader characterization of the context within which the need to represent can be seen to be a need. I quote fairly fully, because I really may be missing something here: The most basic information-marshalling routines associated with formation and maintenance of bodies of belief are guided by the mind’s need to represent things outside itself. The fact that they are guided by a need gives these moves one, thin, kind of normative status: the moves are weakly justified by the need that guides them. But, as a matter of empirical fact, these weakly justified moves tend to generate bodies of belief that stand in ‘homing in’ relations to specific objects . . . Given REFERENCE AND JUSTIFICATION [approximately: S’s beliefs are about o iff justification for holding them converges on o], the fact that (weakly) justified moves made in maintaining a body of beliefs entails that it is o that those beliefs are about . . . This fact now reinforces the weak claim about justification that we started with: the moves made in maintaining the body of beliefs are strongly justified because they are reliable generators of fulfilment of the need that guides them . . . And the fact that the moves are strongly justified in turn reinforces the claim that the beliefs are about o: the beliefs are about o because they are formed by a route which both confers strong justification and tends to get o’s properties right . . . So we have a biconditional whose two sides are explained in terms of one another: the beliefs are about o because they are formed by a route which both confers strong justification and tends to get o’s properties right . . . the route to formation of beliefs confers strong justification because it determines o as the object those beliefs are about. (Dickie, 2015, 112–13)

So far as I can see, what we have here is an argument that, given Dickie’s biconditional connecting reference and justification, a propositional state that meets the need to represent by having a suitable justificational structure will also refer to an object. And a propositional state that meets the need to represent by being about a particular object will also have a suitable justificational structure. But we don’t yet know whether one way of meeting the need is more basic than the other. That is, it may be that what we really need are states with a particular justificational structure, and if they refer, well that is guaranteed by Dickie’s biconditional, but has no significance in itself. Or it may be that what we really need are states that refer to a particular object, and the significance of the justificational structure is entirely consequent upon that. We have to know more about what the need is before we can address these questions. If we locate the need in the context of collaboration and cooperation, however, it seems natural to take it that the basic need is for states that refer to external objects, so that we can engage in joint behaviours with regard to those objects. The importance of the justificational structure will be entirely dependent on its role in enabling us to think jointly about, and carry out joint tasks in relation to, external objects. At the moment at any rate, I don’t really see how the justificational structure can be taken to have any significance of its own in meeting our basic needs. Of any particular justificational structure we can ask, Why care about that? And without

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

an appeal to facts about reference, we have no way of answering. In contrast, an appeal to the possibility of joint action does seem on its own to explain why we ought to engage in reference to external objects. It seems to me that something has gone wrong with the picture of the relation between a ‘need to represent’ and motivational phenomenology. Suppose I’m looking at and thinking about Suzy, love of my life, apple of my eye. I haven’t seen her for a while and it’s a pleasure to see her again and find out how she is and how she’s been. But at the level of individual phenomenology, why am I doing this? According to the account we’ve just been given, it’s because my mind has a need to represent and, casting about, has fallen on Suzy with a cry of relief, marshalling information into bundles in the awareness that this activity is guided by my mind’s need to represent. Now it seems to me that this is completely wrong about me and, for that matter, Suzy. The whole point of my life has been to find and think about someone like Suzy, and there isn’t any level, however far back in my instrumental motivations, at which my current thinking about Suzy is being motivated merely by a need to be thinking about something or other rather than not thinking about anything. It’s Suzy herself that motivates my thinking about her. In terms that Dickie herself takes over from Fred Dretske (1988) it seems to me that the need to represent ought to be thought of as a structuring cause rather than a triggering cause. This is particularly clear if we locate the need to represent in the context of the human need for cooperation and collaboration. Humans have evolved as a cooperative, collaborative species. This evolutionary context provides the space in which humans form distinctively human motivations, such as concern for other people or interest in the objects of their joint tasks. But for the individual human, the concerns for other people and the interest in the objects of joint tasks are motivationally fundamental. It’s a distortion to talk as though we form our beliefs in awareness of the evolutionary needs that are ultimately responsible for our having the basic motivations that we do. The situation seems to me rather different with REFERENCE AND JUSTIFICATION. The notions of ‘forming a body of beliefs about o’, and ‘using a proposition whose proprietary means of justification converges non-accidentally on the properties of o’, seem to me to be rather abstract, structural ideas about general human characteristics, that really can be grounded in the abstract, structural idea of the mind’s need to represent.

4. Joint Attention as Primitive I first sketched the argument for taking joint attention as primitive—rather than grounded in the individualistic psychological states of the individuals attending—in Campbell (2002). I’ll rehearse the basic argument here, then look at two recent analyses whose effect, it seems to me, is to reinforce the anti-individualistic point. (This is I think an unintended effect.) Suppose that you and I are engaged in a joint task that puts to work our capacity for joint attention. Suppose we both have to fire at the same target, whichever one we choose, out of a number of possible targets. Suppose that if we fire together, the target will be destroyed and we reap a significant but limited reward. On the other hand, if only one of us attacks, the result is disaster—we will both be detected and destroyed ourselves. Now the key datum is

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

     



that in this kind of situation, it seems entirely possible that you and I could rationally achieve a successful coordinated attack. If a potential target comes into the scene and you, catching my eye, say ‘That one!’, then it seems evident that we could then both blaze away and quite rationally achieve the destruction of the target. The puzzle is to understand how this can be so. To make the question vivid, suppose that you and I are in separate booths, and that though we have a way of communicating, messages get through only 50 per cent of the time. (The point of the booths and the imperfect communication apparatus is to allow us to be controlled and explicit in considering what knowledge is individualistically available to each of the people involved.) Groups of potential targets come into view and loom for a while before moving on. As the latest group shows up, I send a message to you, ‘Let’s attack the red one’. Let’s suppose that, though the apparatus works only 50 per cent of the time, you do get this message. Is that enough to make it rational for you to attack? Clearly not, because you know that I don’t know whether my message got through. So if I’m rational, I won’t attack. And you know all this, so you won’t attack either. What you have to do is send me a message letting me know you got mine. Then I’ll know that you know my choice of target. Is that enough to make it rational for you to fire? Well, no, because you don’t know whether your message got through. And I know that you don’t know whether your message got through, so I don’t expect you to fire, so I won’t fire. And so the long day wears on. We can put the point like this. Suppose we say that we have first-level mutual knowledge that the red one is the target if we both know that the red one is the target. We have level 2 mutual knowledge that the red one is the target if I know that you know that the red one is the target, and you know that I know that the red one is the target. We have level 3 mutual knowledge if I know that you know that I know that the red one is the target, and you know that I know that you know that the red one is the target. The moral of the example is that we seem to need an actual infinity of pieces of n-level mutual knowledge for it to be rational for us both to fire. As we’ve just seen, it wouldn’t be enough for us merely to have the potential to achieve n-level mutual knowledge, for any arbitrary n, because the messaging system is there for us to use. There is no finite n such that n uses of the messaging system would make it rational for us both to fire. This is, of course, deeply puzzling, because it seems so obvious that I think that the correct reaction here is to say that we shall take the three-place relation: ‘x and y are jointly attending to z’ as primitive, in that it is not reducible to any collection of individualistic psychological states on the part each of x and y. We should regard ‘x and y are jointly attending to z’ as a fundamentally social-psychological state, and as capable of grounding rational action on the part of x and on the part of y. In particular, when x and y have jointly identified z as the target, that can be enough to make it rational for both of them to fire at z, even when the payoff structure is as specified above, significant but limited gains if both fire, and utter disaster if only one fires. There have been two recent attempts to explain how it can be rational for x and y to fire in this situation. One is by Harvey Lederman (2018), who points out that our two subjects may each be individually rational even though neither supposes that the other is rational. After all, even at the entry levels of our puzzle, when I initially send to you, saying ‘Let’s fire at the red one’, what stops you from firing is that you know

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

that I don’t know whether my message has got through. And you assume that I’m rational, and consequently won’t fire, so you won’t fire yourself. But suppose you don’t assume that I’m rational. Then what’s to stop you firing, on hearing my message? And if I don’t assume that you’re rational, what’s to stop me blazing away too as I attempt the sending of my message? The trouble with this is that it’s not a small departure from rationality that’s envisaged here. You have to be assuming that I will risk utter disaster, in the hope of a significant but limited benefit, on the basis of a 50 per cent chance of my message getting through. And I have to be assuming that you are going to just blaze away at the target, without having any rational basis for it, thereby knowingly risking utter disaster. Going up the hierarchy a little will make the situation more complicated, but it won’t significantly alleviate the level of craziness that we have to be arbitrarily ascribing to one another for us both to fire. We are not talking here about some minor departures from ideal rationality. But when we go back to the original scenario, without booths or messaging systems, when I catch your eye and say, ‘Let’s go for the red one’, it seems quite evident that it can be possible for us both rationally to fire, without having to ascribe such craziness to one another. Lederman’s account can’t explain this. Another attempt to ground the rationality of coordinated attack in the individualist psychologies of the two protagonists is provided by Christopher Peacocke. Peacocke defines ‘Full joint attention by x and y to o’ as follows: (a) x and y are attending to o; (b) x and y are each aware that their attention in (a) has mutual open-ended perceptual availability; and (c) x and y are each aware that this whole complex state of awareness (a)–(c) exists. (Peacocke 2005)

This definition uses the technical term ‘mutual open-ended availability of a state of affairs s to two people’, which Peacocke defines as follows: ‘Each perceives that the other perceives that s obtains; and if either is occurrently aware that the other is aware that he is aware . . . that s obtains, then the state of affairs of his being so occurrently aware is available to the other’s occurrent awareness’ (Peacocke 2005). Now the idea is that this will get the force, in effect, of infinitary common knowledge, the key point being the reflexive condition (c) in the definition, which allows us to give a definition of ‘full joint attention’ in terms of finite individual psychological states possessed by the two individuals. The trouble here is evidently that in trading infinitary common knowledge for this kind of compact reflexive condition, we arrive at a condition that is conceptually far more complex than anything the ordinary speaker, or indeed the ordinary philosopher, might use in everyday reasoning with any facility. Nor is there the slightest evidence that this kind of formulation might give the content of some nonconceptual state implemented in the brain, and routinely operated on by some type of non-conceptual information processing. So this can’t explain how the ordinary speaker can be rational in engaging in coordinated attack. Perhaps the most generous interpretation is that Peacocke’s account is intended as a computational account in the sense of Marr (1982). It merely specifies the normatively correct outcomes of

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

     



some psychological processes in the two individuals engaged in full joint attention, rather than telling us what those processes actually are. But that simply gives up on the puzzle, which was to understand how it can in fact be rational for ordinary speakers to engage in coordinated attack. Peacocke’s approach simply cannot explain this. I mention the work by Lederman and Peacocke because these are technically sophisticated discussions of the topic, trying to explain in individualistic terms what we find in cases of ordinary joint attention. It seems to me that despite the sophistication of the work, the objections to each are obvious and decisive. The conclusion I draw is that there is no straightforward way to explain in individualistic terms how it can be rational to engage in coordinated attack. Rather than pursuing this strategy, we should consider taking ‘x and y are jointly attending to z’ as primitive, not to be explained in individualistic terms, and as grounding rational action on the part of each of the attendees. Elizabeth Spelke pointed out to me that joint attention behaviours in young children seem to be grounded in earlier behaviours, such as mutual gaze between child and caregiver. It is indeed a natural thought that the joint attention that emerges at about one year old is grounded in the interplay that earlier exists directly between child and caregiver. You might argue that this implies that there is some analysis to be given of triadic joint attention in terms of the earlier dyadic relation between child and caregiver. This is an intriguing idea, but for present purposes the point is that it suggests at most that we might consider analysing the triadic social relation of joint attention in terms of a dyadic social relation between child and caregiver. It does not show that we are going to be able to analyse either relation in terms of individualistic psychological states. This approach gives us a new way to think about what is going on in referential communication. Stalnaker (2002) remarked on the importance of ‘common ground’ in communication but tried to explain the notion of common ground in terms of possibly infinitary mutual knowledge. I am suggesting that we should hold on to the importance of the idea that there are some matters that are ‘out in the open’ in ordinary communication, but explain this idea of things being ‘out in the open’ in terms of irreducibly social relations. In particular, in simple referential communication about a jointly perceived object, what is ‘out in the open’ has to be merely which object we are talking about. And that can be explained in terms of speaker and hearer jointly attending to the object. It’s instructive here to contrast the approach I’m recommending with Grice’s (1957) model. On a simple Gricean account, discussion of an object that you and I are both perceiving has to be analysed in terms of our individualistic psychological states. You are intending that I form beliefs about a particular object, as a consequence of my recognition of your intention. I for my part am recognizing your intention and no doubt forming suitable beliefs about the object in response. This approach has no way of explaining how it could be normatively correct for you and I to launch a coordinated attack on an object we are talking about, in a case with the payoff structure explained above. That means that the Gricean approach simply cannot explain the sense in which, when you and I are talking about a jointly perceived object, it is ‘out in the open’ between us which thing we are talking about.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

The arresting thing about the transition from the dyadic relation between child and caregiver in infancy to the triadic relation of joint attention that emerges at around one year of age is that there seems to be some sense in which triadic joint attention is referential. When child and caregiver focus on one and the same thing and look back and forth, exchanging emotional and cognitive reactions to it, there is evidently a sense in which they are communicating about that thing. This kind of communication is not peculiar to young children: when you and I are proudly shown a new painting by the artist, the briefest of glances between us may communicate a lot. But in addition to being very basic in development, this kind of communication resists analysis in Gricean terms. In a way that is perfectly reasonable, for there is nothing here that we could call meaning. There is no vehicle for meaning, no string of words that we are using with one intention rather than another. It can be, of course, that I intentionally pull a face as we both look at the thing, leaving you to interpret its meaning. But in the most basic cases, that is not what is going on. Rather, we both jointly attend to the same object, and we openly share our reactions with one another, rather than taking our reactions under intentional control. So our use of language in communication can build on this prelinguistic referential platform. It being out in the open which thing we are communicating about is something that we have prior to any use of language. But this is a social structure. So we can sustain Wittgenstein’s idea that the shared language is prior to the individualistic psychological states of the speakers, while keeping the idea that knowledge of reference and truth conditions is fundamental to our understanding of thought and language.

References Campbell, J. (2002) Reference and Consciousness, Oxford: Oxford University Press. Dickie, I. (2015) Fixing Reference, Oxford: Oxford University Press. Dretske, F. (1988) Explaining Behavior: Reasons in a World of Causes, Cambridge, MA: MIT Press. Grice, H. P. (1957) ‘Meaning’, Philosophical Review 66: 377–88. Lederman, H. (2018) ‘Two Paradoxes of Common Knowledge: Coordinated Attack and Electronic Mail’, Nous. Marr, D. (1982) Vision, New York: W. H. Freeman & Co. O’Brien, F. (1968) The Best of Myles, New York: Walker. Peacocke, C. (2005) ‘Joint Attention: Its Nature, Reflexivity, and Relation to Common Knowledge’, in Naomi M. Eilan, Christoph Hoerl, Teresa McCormack, and Johannes Roessler (eds), Joint Attention: Communication and Other Minds, Oxford: Oxford University Press. Stalnaker, R. (2002) ‘Common Ground’, Linguistics and Philosophy 25: 701–21. Tomasello, M. (2009) Why We Cooperate, Cambridge, MA: MIT Press. Whorf, B. L. (1941) ‘The Relation of Habitual Thought and Behavior to Language’, in Leslie Spier (ed.), Language, Culture and Personality: Essays in Memory of Edward Sapir, Menasha, WI: Banta. Wittgenstein, L. (1953) Philosophical Investigations, ed. G. E. M. Anscombe and R. Rhees, trans. G. E. M. Anscombe, Oxford: Blackwell. Wright, C. (2007) ‘Rule-Following without Reasons: Wittgenstein’s Quietism and the Constitutive Question’, Ratio 20(4): 481–502.

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

10 Principles of Acquaintance Jessica Pepp

1. Suppose Jane is thinking about tigers—she is thinking that they are beautiful and scary. Greta is a tiger whom Jane has never encountered. Is Jane thinking about Greta? There is a sense in which she is. Her thought applies to Greta. But she is thinking about Greta only insofar as she is thinking about tigers and Greta is a tiger. Greta is not in her thought in any way over and above this. Now suppose Jane is thinking that the biggest of all tigers must be beautiful and scary, and suppose Greta is in fact the biggest of all tigers. Is Jane thinking about Greta? Again, there is a sense in which she is: her thought applies to Greta; Greta is the (unique) satisfier of the condition (being the biggest tiger) that Jane is thinking of. But again, Jane is thinking about Greta only insofar as she is thinking about being the biggest of all tigers and Greta is the biggest of all tigers. Greta is not in her thought in any way over and above this. In general, we can distinguish a loose sense in which thoughts are about things to which they apply from a more substantial notion of a thing’s independently being in thought. What can we say about this more substantial notion? What is it for a thing to be in thought (and not merely to be thought about, in the looser sense)? (In what follows I will sometimes use ‘being-in-thought’ to refer to the phenomenon of genuinely being in thought, as distinguished from being thought about, in the looser sense.) Bertrand Russell gave a partial answer to this question: for a thing to be in thought, the thinker must be acquainted with the thing. This is now known as Russell’s ‘Principle of Acquaintance’. I take Russell’s basic statement of this principle to be the following: Russell’s Basic Principle of Acquaintance All thinking has to start from acquaintance; but it succeeds in thinking about many things with which we have no acquaintance. (1905, 874) The principle is the part in italics. The continuation of the sentence serves to contrast the substantial notion of being-in-thought that is claimed to require acquaintance, from the looser notion of being thought about that includes satisfying a condition that is in thought. This looser, satisfactional way of being thought about requires a ‘start’ in things that are thought about in the more substantial way. Russell’s Basic Jessica Pepp, Principles of Acquaintance In: Acquaintance: New Essays. Edited by: Jonathan Knowles and Thomas Raleigh, Oxford University Press (2019). © Jessica Pepp. DOI: 10.1093/oso/9780198803461.003.0011

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi



 

Principle of Acquaintance says that for a thing to be in thought in this starting, nonsatisfactional way, the thinker must be acquainted with it. Russell also endorsed the following principle of acquaintance: Russell’s Propositional Principle of Acquaintance Every proposition which we can understand must be composed wholly of constituents with which we are acquainted. (1910/11, 117) For Russell, this principle is another way of stating his Basic Principle of Acquaintance, since he held that being in thought in the more substantial, non-satisfactional sense is necessary and sufficient for being a constituent of a proposition that is grasped in thought. But at least on the surface, these are different principles of acquaintance. Russell’s Propositional Principle of Acquaintance puts an acquaintance requirement on being a constituent of a proposition grasped in thought. Russell’s Basic Principle of Acquaintance, by contrast, puts an acquaintance requirement on being at the non-satisfactional ‘start’ of thinking about. These principles of acquaintance are supposed to be informative about how things get into thought. Their point is not simply that grasping propositions happens to require acquaintance with the constituents of those propositions, or that it happens to be the case that only things with which we are acquainted can be in thought nonsatisfactionally. Rather, the idea behind the principles is that acquaintance is required because a thing’s being thought about non-satisfactionally or being a constituent of a proposition grasped in thought derives from the thinker’s being acquainted with it. The strongest form of this idea is that a thing’s being-in-thought (understood in either way) just is the thinker’s acquaintance with the thing. This sounds wrong if being-in-thought is considered as an occurrent phenomenon while acquaintance is considered as a non-occurrent phenomenon. For, as Russell noted, one might at a given time be acquainted with many things that are not, in an occurrent way, in one’s thought at that time (1910/11, 109). However, acquaintance can also be considered as an occurrent phenomenon. Russell described occurrent acquaintance with a thing as being directly aware of or being presented with the thing. So, we can characterize the idea behind the principles as follows: a thing’s being-inthought derives from its being presented to the thinker—from an episode of occurrent acquaintance with the thing.

2. Of course, these principles are not informative until ‘acquaintance’ is defined. Philosophers from Russell onward have had different ideas about what counts as acquaintance. A common thread is that perception is critical, at least with respect to acquaintance with particulars. Russell thought that the particulars with which we are acquainted are our own sense-data because, as a sense-datum theorist, he thought that these are what we really perceive. Current writers tend to go in the other direction, extending acquaintance to cover not only perception of external objects, but also awareness of them deriving from perception, such as perception-based memory of them, uptake from others’ linguistic reference to them (where this linguistic reference is ultimately based on someone’s perception of the object), or

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

  



memory of them based on earlier linguistic uptake.¹ Robin Jeshion gives a useful formulation of a liberal version of this extension: Jeshion’s ‘Standard-Standard on Acquaintance’ One can be acquainted with an object O only by perception, memory, and communication chains. (2010, 109) Jeshion’s Standard-Standard suggests the following requirement for occurrent acquaintance. To be occurrently acquainted with an object one must either be: (1) perceiving the object, (2) being referred to the object² via a communication chain originating in someone’s perception of the object, or (3) remembering the object (with the memory deriving either from one’s past perceptions of the object or from one’s past uptake of the object via language). This requirement may raise concerns because it seems to exclude abstract objects from being objects of occurrent acquaintance, given that (as many believe) they cannot be perceived. If so, then the truth of either acquaintance principle would exclude abstract objects from being in thought. It is a substantive question whether abstract objects can be perceived, and it depends to some degree on how one understands perception. Charles Parsons (1980) and Penelope Maddy (1980), for instance, develop accounts of mathematical intuition that treats the intuiting of certain abstract mathematical objects as, in effect, a variety of perception. Treatment of this matter is beyond the scope of the present chapter. I will follow the fairly standard practice in current discussions of acquaintance requirements and restrict my discussion to concrete objects.³ In addition to adopting something like the Standard-Standard on acquaintance, current discussions focus on acquaintance requirements that are more in the vein of Russell’s Propositional Principle of Acquaintance than his Basic Principle of Acquaintance. Current writers typically are committed neither to the wholly propositional nature of thought nor to Russell’s particular metaphysics of propositions, according to which propositions are abstract objects having worldly objects and properties as literal constituents. A more general descendent of the Propositional Principle of Acquaintance, which also takes account of the refinements just discussed, can be stated as follows: Content Principle of Acquaintance (CPA) For a concrete object to figure in the content of one’s occurrent thought, one must be occurrently acquainted with the object—that is, one must be perceiving the ¹ Dickie (2016) calls this the ‘extended acquaintance tradition’. Hawthorne and Manley (2012) call it ‘causal acquaintance’. ² Here I use ‘being referred to the object’ in roughly the same way as Bach (2008): in understanding a use of a word to refer to an object, one is referred to that object. ³ If abstract objects cannot be perceived, then rejection of a Standard-Standard acquaintance requirement on concrete objects being-in-thought might be a positive development, as it would allow a unified treatment of being-in-thought across both abstract and concrete objects (Jeshion (2002, 57) suggests this). Nonetheless, if an acquaintance requirement is defensible with respect to concrete objects, this cannot be ignored simply in the interest of ensuring a unified account.

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi



 

object, being referred to it via a perception-based linguistic communication chain, or remembering it via one of those two kinds of earlier experience. An object’s figuring in the content of thought is often described as the thought being a ‘singular thought’ about that object. Singular thought and CPA have been the focus for recent investigations of the role for acquaintance in accounting for the substantial kind of being-in-thought. CPA has been the subject of much criticism. Some have even suggested that problems with CPA show that the notion of acquaintance ‘is a dispensable relic of a bygone era in the philosophy of language and mind’ (Hawthorne and Manley, 2012, 25). In Section 4, I will argue that however powerful the evidence against CPA may be, a claim like this is at best premature. The reason is that the Basic Principle of Acquaintance affords another potentially important role for the putative relic of acquaintance. It needs to be assessed separately from CPA. Nonetheless, the kinds of problems that arise for CPA are instructive, both in making the case that the Basic Principle of Acquaintance deserves separate attention, and in making a start on evaluating the Basic Principle of Acquaintance. Given this, I will briefly review the case against CPA in Section 3.

3. Criticisms of CPA can be divided into two categories: arguments by counter-example and arguments by failure of motivation. The counter-example-based arguments adduce cases in which, intuitively, an object figures in the content of thought (the thought is a ‘singular thought’ about the object), but in which the thinker has no acquaintance (in the extended Standard-Standard sense) with the object. The arguments by failure of motivation aim to show that traditional explanatory roles for acquaintance requirements on singular thought can be filled in other ways, or that acquaintance requirements do not in fact play those roles. Counter-examples to CPA need to evoke the intuition that an object is figuring in the content of occurrent thought, although the thinker is not perceiving the object, being referred to it in language, or remembering it via one of those two kinds of earlier experience. The first part of this intuition concerns a theoretical status: an object’s figuring in the content of thought, or the thought’s being singular with respect to the thing. Critics of CPA acknowledge that one’s intuitions about this matter will be shaped by one’s theory of singular thought.⁴ Their general strategy, however, is to use a rich diet of examples to issue the challenge: if you are going to draw a distinction between thought whose content is singular as opposed to general with respect to an object, do you not want at least some of these cases to fall on the singular side? If you do, and you endorse CPA, you have a problem, because in none of these cases is the thinker perceiving the object, being referred linguistically to the object via a perception-based communication chain, or remembering the object through either of these kinds of earlier experience. The kinds of examples used to issue this challenge include inference cases, future existence cases, linguistic cases, and map cases. First, there seem to be cases ⁴ See, for instance, Jeshion (2010, 112–13) and Hawthorne and Manley (2012, 3–4).

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

  



of thinking singularly about a thing whose existence one infers, though one has no perception-based link to it. For instance, a car mechanic might infer from the appearance of a car that it has an engine which has not been well treated, and think that that engine is destroyed. Second, there seem to be cases of thinking singularly about a thing that one has reason to believe will come to exist, despite there being no possibility of a perception-based link, as the thing does not yet exist. For instance, one might intend to have a child, and think that that child will have light hair. Third, there seem to be cases of thinking singularly about a thing using a name or indexical expression, despite having no perception-based link to it. For instance, one might use the word ‘I’ to think of oneself in a way that is (arguably) not based on any perception of one by oneself or by anyone else. Or one might use a name like ‘Mitochondrial Eve’, which was introduced without any perceptual link to the individual it names,⁵ to think of a particular individual. Fourth, there seem to be cases of thinking singularly about a location using a map (physical or mental), despite having no extended perceptual link to that location. For instance, a cartographer might map the coastline of an island and use it to think that that (spatially identified) inland location, which neither he nor anyone he is in contact with has ever perceived, is windy.⁶ Putative counter-examples such as these may cast doubt on CPA. But one can also ask what reason there is to endorse CPA in the first place. A cluster of traditional motivations relate to what I will call the ‘focus-shifting manoeuvre’. It seems that if Jane is able to think about there being a biggest of all tigers, then nothing stops her from shifting her cognitive focus from this condition to its satisfier. She might do this by introducing a name for whoever is the biggest tiger and going on to think with that name, or by thinking in a way that she might express as concerning ‘that animal—the one who is in fact the biggest tiger’.⁷ If Jane can shift her cognitive focus in this way, then it is not clear why the content of her thought should not also shift. Why should it not be that Jane goes from thinking that among the tigers the one who is biggest is beautiful and scary, to thinking that that tiger—the biggest one—is beautiful and scary? The focus-shifting manoeuvre has suggested to some that ‘semantic instrumentalism’⁸ is true. This is the view that if a thinker can entertain a content containing a condition she believes satisfied, and if she can introduce a name (or other term) whose reference she stipulates to be fixed by that condition, then she can use that name to entertain a content containing the satisfier of that condition. But most theorists reject semantic instrumentalism, presumably because it lets things into ‘singular thought’ too easily. This is one motivation for positing substantive requirements on figuring in the content of thought, such as CPA.⁹ ⁵ In human genetics, ‘Mitochondrial Eve’ is the name given to the most recent matrilineal common ancestor of all currently living humans. ⁶ These examples are drawn from or inspired by the discussions in Jeshion (2010) and Hawthorne and Manley (2012). As Hawthorne and Manley note, similar examples have been discussed by a range of other philosophers, notably Sosa (1970), Jeshion (2002), McGinn (1981), and Sutton (2001). ⁷ This would be a mental equivalent of introducing a Kaplanian ‘dthat’ term (Kaplan, 1989). ⁸ ‘Semantic instrumentalism’ is Jeshion’s name for the view, which is typically traced to Kaplan (1989), and is also attributed to Harman (1977) and Borg (2007). ⁹ A different supplementary requirement is Jeshion’s (2010) requirement that anything figuring in the content of thought be thought about using a mental file.

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi



 

Another problem that derives from the focus-shifting manoeuvre is that the manoeuvre seems to allow us to gain new knowledge too easily. If Jane knows that among all the tigers, one is biggest, why can she not shift her focus to that individual—Greta—and come to know something she did not know before: namely, that Greta is the biggest tiger? CPA seems to offer an explanation. Jane does not come to know anything new by this procedure because she does not come to entertain a new content that attributes being the biggest tiger to the individual who is in fact the biggest tiger, Greta. For her thought to have this content, the thought would have to be derived from acquaintance with Greta. But Jane’s shift in cognitive focus has done nothing to remedy her lack of acquaintance with Greta. Hawthorne and Manley challenge this explanation by imagining similar procedures which do remedy Jane’s lack of acquaintance with Greta. Suppose Jane has a magic arm that will reach out and allow her to touch whoever is the biggest tiger. Jane can then name this individual ‘Biggie’. This naming, and Jane’s subsequent thought using the name, do derive from Jane’s (tactile) perception-based acquaintance with Greta. Alternatively, suppose Jane orders a magical fairy to go and find the biggest tiger, become perceptually acquainted with that individual, and name her ‘Biggie’. When the fairy returns and tells Jane ‘Biggie is the biggest tiger’, Jane is referred via a perception-based linguistic communication chain to Greta. Despite all this, Hawthorne and Manley think, we should be no happier attributing new knowledge to Jane in these cases than in a case where she simply performs the focus-shifting manoeuvre. And if Jane’s gaining acquaintance via these procedures does not create new knowledge, then Jane’s not gaining acquaintance via the original procedure does not explain why the original procedure does not create new knowledge. A similar point could be made about using CPA to avoid semantic instrumentalism. If using the magic arm or the magic fairy would still be too easy a way to have singular thoughts about things, then the problem with semantic instrumentalism is not that the semantic version of the focus-shifting manoeuvre does not acquaint the thinker with the satisfier of the condition. For even if it did, our intuition would still be that singular thought is not so easily obtained.

4. With this brief review I have provided only a partial sketch of the case against CPA, and I have made no evaluation of it. This is because my interest is not in rejecting or defending CPA but in pointing out that however strong the case against CPA may be, it is not the whole story of whether acquaintance requirements have a place in a right account of being-in-thought. Although Russell treated his version of CPA as another way of expressing the Basic Principle of Acquaintance, they are different principles. If CPA is false, this does not entail that the Basic Principle of Acquaintance is false. Here is a version of Russell’s Basic Principle of Acquaintance, that ‘all thinking has to start from acquaintance’, revised and restricted in parallel with the statement of CPA in section B. Basic Principle of Acquaintance (BPA)

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

  



For a concrete object to be in one’s occurrent thought in a non-satisfactional way, one must be occurrently acquainted with the object—that is, one must be perceiving the object, being referred to it via a perception-based linguistic communication chain, or remembering it via one of those two kinds of earlier experience. For BPA to entail CPA, figuring in the content of thought would have to entail being thought about in a non-satisfactional way. For CPA to entail BPA, being thought about in a non-satisfactional way would have to entail figuring in the content of thought. It is not obvious that either of these is the case. Take the car mechanic case mentioned above: in the mechanic’s thought that that engine is destroyed, it is supposed to be the engine that figures in the content of the thought, even though, at least arguably, the engine is thought about only by satisfying a condition that is in thought, such as being the engine in that car. If this is right, then figuring in the content of thought does not entail being thought about in a non-satisfactional way, hence BPA does not entail CPA. It also seems at least possible that, although the car is (let us suppose) thought about in a non-satisfactional way, it does not figure in the content of thought. (The thought is about the engine, not the car.) Of course, this depends on the details of one’s theory of thought contents. But if it is coherent, then it is not obvious that being thought about in a non-satisfactional way entails figuring in the content of thought. Thus it is coherent to endorse CPA but reject BPA. The important point for my purposes, though, is that BPA does not entail CPA; hence if CPA is false, this does not entail that BPA is false. Furthermore, the question with which I began the chapter—what can we say about being-in-thought in the substantial sense that contrasts with satisfying a condition in thought?—is of interest even if one does not restate the question in terms of content. Maintaining non-satisfactional aboutness as the important mark of genuine beingin-thought, instead of focusing on singular content, has some advantages. One is that no problem of semantic instrumentalism arises. One can think about something via a condition that the thing satisfies, and one can introduce a name whose reference one stipulates to be fixed by that condition. One can go on to use that name in thinking about the thing. None of this changes the fact that one is thinking about the thing by means of its satisfying a condition that is in thought. Even if one goes on to think about the thing using the name without keeping the condition in occurrent thought, one’s previous occurrent thought about the condition is still part of the means by which one is now thinking about the thing. No semantic manoeuvre can kick away the essential role for the condition in making one’s thought be about the thing. The only way for this to change is for one to acquire a different, non-satisfactional way of thinking about the thing. Another advantage of focusing on non-satisfactional aboutness is that it may present a different way to deal with the problem of too-easy knowledge. Once we think of knowing something as knowing a content, we are faced with the problem that the focus-shifting manoeuvre seems to allow one to entertain a content that one is thereby (in virtue of having used that very procedure) in a position to know. If CPA and related substantive requirements on figuring in singular contents turn out to be unmotivated, the question remains of why knowledge cannot be gained by the

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi



 

focus-shifting procedure. Hawthorne and Manley’s suggestion seems to be that new knowledge can be gained in this way, it is just that the knowledge is not ‘interesting’ or ‘momentous’ (2012, 67). This is not altogether satisfying. Suppose we instead treat knowledge that (e.g.) Greta is the biggest tiger as knowledge of the fact of Greta’s being the biggest tiger?¹⁰ One way to conceive of knowing a fact is as having that fact in thought. Having a fact in thought is different from having (in the simple case) an object in thought and a property in thought, and doing a cognitive operation to ascribe the property to the object. Even if the ascription happens to be true, this is not enough for having the fact (that the object has the property) in thought. This is analogous to the intuition that having tigers in thought and bigness in thought, and doing a cognitive operation to form the condition of being the biggest tiger, is not sufficient for having Greta in thought. I would then suggest that in the same way that Jane thinks of Greta only by thinking with the condition the biggest of all tigers, she thinks of (the fact of) Greta’s being the biggest of all tigers only by thinking with the condition someone’s being the biggest of all tigers.¹¹ The former condition is satisfied by Greta; the latter is satisfied by the fact of Greta’s being the biggest of all tigers. Thus, neither Greta nor the fact of Greta’s being the biggest of all tigers are genuinely in Jane’s thought after she uses the focus-shifting procedure. Accordingly, no new knowledge of facts is gained. One might worry that this approach to the problem still leaves it open that in the case where Jane employs the magic fairy to gain acquaintance with the satisfier of the condition, she does gain new knowledge. For it seems that when the fairy says, ‘Biggie is the biggest tiger’, this enables Jane to think about the fact of Biggie’s (Greta’s) being the biggest tiger in a way that is arguably non-satisfactional.¹² This is a plausible upshot, but I do not think it is a problem for the approach. For the approach offers an explanation of why (despite potentially countervailing intuitions) the fairy procedure might create new knowledge of facts while the initial focus-shifting procedure does not. The explanation is that the fairy procedure gives one a non-satisfactional way of thinking of certain facts, which is, on this approach, what it is to know them. And while one can build the case (as Hawthorne and Manley do with the fairy case) so

¹⁰ Admittedly, treating knowledge that such-and-such as knowledge of worldly facts is a minority approach. (See Kratzer (2002) for a relatively recent defence, though she does not conceive of knowing a fact in the way I suggest below.) What follows in the text should only be taken as suggestive of a different sort of approach to the problem of too-easy knowledge that does not involve a focus on content. ¹¹ Note that one might also think of Greta’s being the biggest of all tigers satisfactionally in the following way. Suppose one thinks of Greta (non-satisfactionally, we may suppose) and, because of her large size, assumes that she is the biggest of all tigers. One thus forms a condition in thought, (the fact of) Greta’s being the biggest of all tigers, which is in fact satisfied by (the fact of) Greta’s being the biggest of all tigers. But this fact is not in thought in any way over and above satisfying this condition. ¹² Notice that the magic arm case does not obviously have the same result. It may be that Jane thinks about Biggie/Greta in a non-satisfactional way through touching Greta with the hand of her magic arm, but it is less clear that she thinks about the fact of Greta’s being the biggest tiger in a non-satisfactional way. More plausibly, she thinks about her touching the biggest tiger in a non-satisfactional way, and thinks about the fact that Greta is the biggest tiger only by means of thinking with the condition, this thing’s being the biggest tiger, which the latter fact satisfies. By contrast, she can now think of the fact of Greta’s being solid, or having other tactile features, in a non-satisfactional way.

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

  



that this knowledge is of little instrumental significance, the knowledge is significant in other ways. This is illustrated by an example that Jeshion (2010) discusses, of an adoptee who yearns to meet his biological mother.¹³ The adoptee knows that he has a unique biological mother, so he can use the focus-shifting procedure to think about the fact that she is his biological mother. But he thinks about this fact only by means of thinking of someone’s being my unique biological mother.¹⁴ The man goes on this way for many years, marvelling at the fact that he does have a mother, and pondering over what she might be like. Now let us add to Jeshion’s example that, at long last, the man saves up enough money to hire a private detective to go out and locate his biological mother. The detective returns and tells the man, ‘I found her. She is your mother, all right.’ It seems to me that even if nothing further follows from this (maybe the detective suddenly drops dead and none of his records can be located), something has changed. Although this change does not open up new possibilities for the man to act, it is nonetheless interesting and momentous. The fact of a particular person’s being his mother has come into the man’s thought in a new way. He is no longer thinking of this fact by thinking with the condition, someone’s being my unique biological mother. Rather, another person has encountered this fact and reported it to him. He thus gains knowledge of that fact that he did not have before. These reflections on the problem of too-easy knowledge are far from conclusive. They are only intended to give the flavour of an approach one might take if one maintains non-satisfactional aboutness, rather than singular content, as the primary mark of being-in-thought. When singular content is adopted as the primary mark of being-in-thought, problems like semantic instrumentalism and too-easy knowledge invite theorists to place substantive requirements on content inclusion, such as CPA. By contrast, when non-satisfactional aboutness is the focus, the problem of semantic instrumentalism does not arise, and other avenues are available for addressing the problem of too-easy knowledge. Since these problems connect to basic commitments about being-in-thought, this provides some inducement to focus on non-satisfactional aboutness as the mark of being-in-thought, and to assess BPA rather than—or at least as well as—CPA in the effort to give an account of being-inthought. In the final part of the chapter, I will take a first step in this direction by outlining the initial plausibility of BPA and the kinds of challenges against which it should be tested.

5. If concrete objects can be thought about in a non-satisfactional way, what would that way be? An answer that immediately presents itself is that we can perceive concrete objects, or be told about them by others who have perceived them, and we can remember them through these experiences. In other words, we think about them by ¹³ Jeshion adapts the example from one used in Velleman (2008). ¹⁴ Jeshion uses this case to evoke the intuition that the adoptee is able to entertain singular contents about his mother (presumably, despite the fact that his means of thinking of her is satisfactional). Jeshion does not discuss the case in terms of facts or ways of thinking about them.

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi



 

means of our acquaintance (in the Standard-Standard sense) with them. Acquaintance with concrete objects seems to be a paradigm means of thinking about them in a non-satisfactional way.¹⁵ This does not imply that all non-satisfactional thinking about concrete objects is via acquaintance. But it provides some initial motivation for BPA and sets the task of challenging BPA by identifying ways of thinking about concrete objects that are not via acquaintance. One place to look is among the putative counter-examples to CPA mentioned above. These are examples in which it is supposed to be intuitive that the content of thought is singular with respect to some object, although the thought is not derived from the thinker’s acquaintance with that object. If it is also intuitive in some of these cases that the relevant object is thought about in a non-satisfactional way, then these cases could serve as counter-examples to BPA. Inference cases, future existence cases, and maps cases do not fit this bill. As described, the cases do not elicit the intuition that the things are thought about nonsatisfactionally. Indeed, it seems plausible (for example) that the car mechanic thinks of a particular car engine by means of thinking with the condition the engine of this car. It seems plausible that an individual contemplating his future thinks of some particular future child of his by means of thinking of the condition the (first/only, etc.) child that I will have. It seems plausible that the cartographer thinks of a particular area on the island he charts by means of thinking with the condition the area that is such-and-such distance in such-and-such direction from this point on this coastline. Cases of thinking of things using linguistic devices seem more promising as potential counter-examples to BPA. A thinker might understand sentences like ‘I am hungry’ or ‘Mitochondrial Eve lived over 100,000 years ago’, and thereby think of herself or of a certain human being, but not in virtue of thinking of, or with, conditions that she or that human being satisfies. (She might not know the reference-fixing condition associated with the name ‘Mitochondrial Eve’, despite having picked up the name in biology class.) Rather, it might be suggested, the linguistic conventions governing the expressions determine referents for uses of those expressions. If she can refer to those things, and understand these claims about them, it is not clear why we should not also allow that she is thinking of those things. Some have questioned the view that being able to refer in language to a thing, and to understand talk about it, is sufficient for being able to think about it.¹⁶ Even if we instead accept this view (as I am inclined to), we may question whether the semantic operation of introducing a name like ‘Mitochondrial Eve’ via a reference-fixing description enables anyone to think about the satisfier of that description in a nonsatisfactional way. If such semantic manoeuvres do not have this power (as suggested in Section 4), then a speaker who picks up this name from another speaker acquires (at most) the ability to think about Mitochondrial Eve in a satisfactional way, albeit

¹⁵ Some hold that perceiving concrete objects only lets us think about them in a satisfactional way (e.g. Searle, 1983). One who had such a view might reject BPA, if (but only if) she also held that there are other, non-perception-based, non-satisfactional ways of thinking about concrete objects. I discuss some possibilities in this vein below. ¹⁶ Most notably, Gareth Evans (1982, chapter 11). See also McGinn (1981).

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

  



by means of the source speaker’s thinking with the condition the matrilineal most recent common ancestor of all currently living humans. In addition, there is a question about the order of determination. When thinking with words, does one think of things in virtue of thinking with words governed by linguistic conventions that (perhaps relative to context) determine those things as referents? Or does one refer using those words to those things in virtue of the fact that one’s use of the words results from thinking about those things? I think there is reason to favour the latter view.¹⁷ Further, it is not clear what it takes for a use of a word to be governed by a reference-determining convention.¹⁸ Such considerations suggest that it is difficult to make the case that linguistic conventions can make things be in thought. Nonetheless, the example of indexical expressions in particular points to a stronger challenge to BPA. Independent of conventional linguistic expressions, we seem to have a special, ‘first-personal’ way of thinking of persons, places, and times. We think of persons as I, places as here, and times as now. (By which I do not imply that we have to use those words, or any words.) At least to many philosophers, this firstpersonal way of thinking of things has seemed to be independent of any perception, memory, or linguistic reference to the things thought of. But it also seems clear that at least some things thought of in this way must be in thought in a non-satisfactional way. For if I think of (for instance) my present location in a satisfactional way, by means of thinking with the condition, being the place where I am now, then the question arises of how I am thinking of myself and of the present time, which figure in that condition. If it is by thinking with the condition, being the person who is here now, a circularity immediately emerges. Either first-person thought about myself, my time, and my location comes as a unit,¹⁹ or some of these are thought about in a nonsatisfactional way while others are thought about satisfactionally. Either way, there is some first-person thinking that is non-satisfactional. If this is right, and if some of the things thought of in this way are concrete objects, then BPA is false. For the sake of argument, I will accept that some of the things thought of in this way (for example, people) are concrete objects. Then the main point I want to make here is that BPA depends on the correctness of a broadly perceptual account of what makes first-person thought be about things. This is a substantial challenge to BPA, as such accounts are widely believed to run afoul of a central desideratum. The desideratum is that an account of what makes first-person thought be about things should entail that an individual’s first-person thought—be it I-thought, now-thought, or here-thought—cannot fail to be about that person, or the time or place of her thought. Yet, if things are brought into first-person thought via perception, it is not clear how this desideratum could be guaranteed. If one lacks perceptual contact with oneself or one’s spatiotemporal location, one’s first-person thought could not be about these things, violating the desideratum.²⁰ Similarly, if one’s perceptual contact

¹⁷ I make the case for this in Pepp (2019a). ¹⁸ The difficulty of making out what it is for a convention to be ‘in force’ for a use of a word has been emphasized by Pagin (1987) and Glüer and Pagin (1999). I develop this difficulty with respect to the reference of words, as used specifically in Pepp (2019a). ¹⁹ As suggested by Evans (1982, 153). ²⁰ Anscombe, 1975, 58.

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi



 

is with something other than oneself or one’s spatiotemporal location, one’s first-person thought could not be about these things, again violating the desideratum.²¹ Evaluation of this challenge to BPA requires (1) assessing whether it is really a desideratum for an account of first-person being-in-thought that it should entail this kind of stickiness to the thinker, and (2) assessing whether a perceptual account really violates this desideratum. These are big questions with a literature of their own, into which I will not enter here.²² For the purposes of the present chapter, I am content to tie the truth of BPA to these questions. They are key questions to answer in the effort to understand whether acquaintance, in the broad perceptual sense that I have been using that term, is essential to genuine being-in-thought for concrete objects. Another source of counter-examples might be what David Kaplan (2012) calls ‘evidence cases’. In these cases, one thinks about a thing by means of some evidence one has of it, but the condition one uses to think about the thing is not in fact satisfied by the thing. Additionally, it is supposed to be intuitive about these cases that one has no perception-based acquaintance with the thing. Kaplan describes a case in which his computer appears to have been stolen. He goes in search of what he would describe as ‘the person who stole my computer’, and finds the person he was looking for. However, it turns out that ‘unbeknown to either of us, he has an ownership right in the computer, so it wasn’t actually a theft or (on the basis of some even more fantastic scenario) it isn’t actually a computer’ (2012, 144). It seems that all along Kaplan was thinking of that person (plotting revenge upon him, wondering how he could be so selfish, and so on) though not in virtue of the person satisfying the condition that Kaplan used to think about him, and not by means of having perceived the person, or having been told about the person by anyone who has perceived him. Thus, this appears to be a case of non-satisfactional but nonacquaintance-based thinking about the person, and so a counter-example to BPA. However, both the ‘non-satisfactional’ and the ‘non-acquaintance-based’ parts of this judgement can be questioned. One might argue that in such cases there is always a ‘back-up’ condition in thought that is satisfied by the thing thought about, and that this is really the means by which the thinker thinks about the thing. For instance, it might be claimed that Kaplan is also thinking with the condition the person or persons that removed the thing that used to sit in that spot on my desk, and the person he is thinking about does satisfy that condition.²³ One might also question whether these are in fact cases in which the thinker does not perceive the object. In the case that Kaplan discusses, Kaplan’s thinking of the computer taker is not derived from his (or anyone else’s) seeing, hearing, or touching that person. But this is not decisive on the question of whether his thinking is derived from his perceiving that person. For he might be perceiving that person indirectly, much as he might perceive him by perceiving footage from a closed-circuit television

²¹ See Campbell (1994) and O’Brien (2007, chapter 3). ²² See, for instance, Campbell (1994), O’Brien (2007, 1995), Morgan (2015), and Peacocke (2008, 2014). ²³ For an extended discussion of these kinds of cases and this kind of response to them, see Dickie (2016, chapter 6). Dickie argues that in such cases the objects are thought about neither by condition satisfaction (even of back-up conditions) nor by acquaintance.

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

  



monitor that caught the taker in the act, or through a still image captured from such footage. The crime scene might similarly be a kind of record through which one can, indirectly, perceive that person. The question of whether there is genuine perception through photograph and video images remains a live one in philosophical aesthetics.²⁴ Even if this debate were settled in favour of there being such indirect perception, it is a further question whether there is also perception through other sorts of records or traces. These two lines of questioning need to be explored more fully to assess (1) whether thinking about things via evidence relies on those things satisfying conditions that are in thought, even if they fail to satisfy some other conditions that are in thought, and (2) whether evidence provides an indirect perceptual link to things. If either of these claims proves defensible, then evidence cases are not a clear counter-example to BPA. This is an additional important line of inquiry in assessing BPA.

6. Conclusion The purpose of advancing, challenging, and defending principles of acquaintance is to investigate the broader question of what it is for things genuinely to be in our thought. Much of the effort to do this since Russell has focused on CPA. This principle adopts being in, or figuring in, content as the mark of genuinely being in thought. There is no entailment from BPA, which focuses on the difference between satisfactional and non-satisfactional thinking about things, to CPA. Thus, even if criticisms of CPA are decisive, they do not straightforwardly apply to BPA. Further, there are problems associated with using content to divide genuine from derivative being-in-thought. Taken together, these considerations favour at least taking BPA seriously as a route to shedding light on genuine being-in-thought. I have suggested that two key lines of investigation in evaluating BPA are (1) the question of whether a perceptual account of first-person thinking is viable, and (2) the question of whether Kaplan’s and Dickie’s ‘evidence cases’ are really cases of both non-satisfactional and non-acquaintance-based thinking of things. These questions are of interest in their own right, and gain additional interest from their role in evaluating BPA. They should be approached with the distinction between CPA and BPA, and the choice of whether to use the notion of content in conceiving of genuine being-in-thought, firmly in mind.²⁵ ²⁴ The classic discussion is Walton (1984), where perception through photographs is defended as genuine perception. Criticisms and defences of Walton’s position include Currie (1991), Carroll (1996), Friday (1996), Walton (1997), Lopes (2003), Cohen and Meskin (2004), Nanay (2010), Yetter-Chappell (2018), and Pepp (2019b). ²⁵ I am grateful for valuable feedback on earlier versions of this chapter provided by audiences at: The Acquaintance Workshop at NTNU Trondheim in 2014, the Research Seminar in Theoretical Philosophy at the University of Gothenburg in 2016, and both the Umeå Higher Seminar in Philosophy and the Umeå/ CSMN joint workshop on Acquaintance held at Umeå University in 2017. I owe particular thanks to Pär Sundström, Torfinn Huvenes, Sebastian Watzl, Anders Nes, Bram Vassen and Kristoffer Sundberg for discussion. I also thank the editors of this volume for helpful discussions of the topic of acquaintance as well as feedback on both earlier and later versions of this work. Finally, thanks to David Kaplan and Joseph Almog for many conversations that have inspired my reflections on thinking about and referring to things.

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi



 

References Anscombe, Elizabeth (1975) ‘The First Person’, in Samuel D. Guttenplan (ed.), Mind and Language, Oxford: Oxford University Press, 45–65. Bach, Kent (2008) ‘On Referring and Not Referring’, in J. Gundel and N. Hedberg (eds), Reference: Interdisciplinary Perspectives, Oxford: Oxford University Press, 13–59. Borg, Emma (2007) Minimal Semantics, Oxford: Oxford University Press. Campbell, John (1994) Past, Space, and Self, Cambridge, MA: MIT Press. Carroll, Noel (1996) Theorizing the Moving Image, Cambridge: Cambridge University Press. Cohen, Jonathan and Meskin, Aaron (2004) ‘On the Epistemic Value of Photographs’, Journal of Aesthetics and Art Criticism 62(2): 197–210. Currie, Gregory (1991) ‘Photography, Painting, and Perception’. Journal of Aesthetics and Art Criticism 49(1): 23–9. Dickie, Imogen (2016) Fixing Reference, Oxford: Oxford University Press. Evans, Gareth (1982) The Varieties of Reference, ed. John McDowell, Oxford: Oxford University Press. Friday, Jonathan (1996) ‘Transparency and the Photographic Image’, British Journal of Aesthetics 36(1): 30–42. Glüer, Kathrin and Pagin, Peter (1999) ‘Rules of Meaning and Practical Reasoning’, Synthèse 117: 207–27. Harman, Gilbert (1977) ‘How to Use Propositions’, American Philosophical Quarterly 14: 173–6. Hawthorne, John and Manley, David (2012) The Reference Book, Oxford: Oxford University Press. Jeshion, Robin (2002) ‘Acquiantanceless De Re Belief ’, in Joseph Keim Campbell, Michael O’Rourke, and David Shier (eds), Meaning and Truth: Investigations in Philosophical Semantics, New York: Seven Bridges Press, 53–74. Jeshion, Robin (2010) ‘Singular Thought: Acquaintance, Semantic Instrumentalism, and Cognitivism’, in Robin Jeshion (ed.), New Essays on Singular Thought, Oxford: Oxford University Press, 105–40. Kaplan, David (1989) ‘Demonstratives’, in Joseph Almog, Howard Wettstein, and John Perry (eds), Themes from Kaplan, New York: Oxford University Press, 481–563. Kaplan, David (2012) ‘An Idea of Donnellan’, in Joseph Almog and Paolo Leonardi (eds), The Philosophy of Keith Donnellan, Oxford: Oxford University Press, 122–75. Kratzer, Angelika (2002) ‘Facts: Particulars or Information Units?’, Linguistics and Philosophy 25(5–6): 655–70. Lopes, Domenic MacIver (2003) ‘The Aesthetics of Photographic Transparency’, Mind 112(447): 434–48. Maddy, Penelope (1980) ‘Perception and Mathematical Intuition’, Philosophical Review 89(2): 163–96. McGinn, Colin (1981) ‘The Mechanism of Reference’, Synthese 49(2): 157–86. Morgan, Daniel (2015) ‘The Demonstrative Model of First-Person Thought’, Philosophical Studies 172(7): 1795–811. Nanay, Bence (2010) ‘Transparency and Sensorimotor Contingencies: Do We See through Photographs?’, Pacific Philosophical Quarterly 91(4): 463–80. O’Brien, Lucy (1995) ‘Evans on Self-Identification’, Noûs 29(2): 232–47. O’Brien, Lucy (2007) Self-Knowing Agents, Oxford: Oxford University Press. Pagin, Pagin (1987) Ideas for a Theory of Rules, PhD thesis, Department of Philosophy, Stockholm University. Parsons, Charles (1980) ‘Mathematical Intuition’, Proceedings of the Aristotelian Society 80(1): 145–68.

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

  



Peacocke, Christopher (2008) Truly Understood, Oxford: Oxford University Press. Peacocke, Christopher (2014) The Mirror of the World, Oxford: Oxford University Press. Pepp, Jessica (2019a) ‘What Determines the Reference of Names? What Determines the Objects of Thought?’ Erkenntnis 84: 741–759. Pepp, Jessica (2019b) ‘On Pictorially Mediated Mind-Object Relations’ Inquiry, DOI: 10.1080/ 0020174X.2018.1562372. Russell, Bertrand (1905) ‘On Denoting’, Mind, New Series 14: 479–93. Russell, Bertrand (1910–11) ‘Knowledge by Acquaintance and Knowledge by Description’, Proceedings of the Aristotelian Society, New Series 11(1): 108–28. Searle, John (1983) Intentionality: An Essay in the Philosophy of Mind, Cambridge: Cambridge University Press. Sosa, Ernest (1970) ‘Propositional Attitudes de Dicto and de Re’, Journal of Philosophy 67(21): 883–96. Sutton, Jonathan (2001) ‘The Contingent A Priori and Implicit Knowledge’, Philosophy and Phenomenological Research 63(2): 251–77. Velleman, David (2008) ‘The Gift of Life’, Philosophy and Public Affairs 36: 245–66. Walton, Kendall (1997) ‘On Pictures and Photographs: Objections Answered’, in Richard Allen and Murray Smith (eds), Film Theory and Philosophy, Oxford: Oxford University Press, 60–75. Walton, Kendall (1984) ‘Transparent Pictures: On the Nature of Photographic Realism’, Critical Inquiry 18(1): 67–72. Yetter-Chappell, Helen (2018) ‘Seeing Through Eyes, Mirrors, Shadows and Pictures’ Philosophical Studies 175(8): 2017–2042.

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

PART IV

Epistemology

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

11 Acquaintance The Foundation of Knowledge and Thought Richard Fumerton

1. Introduction Those of us who think that there is such a thing as direct acquaintance tend to believe that it is the foundation of all knowledge, and even the foundation of all thought. It is only in virtue of the fact that we can become directly acquainted with facts that we can know truths without needing to infer those truths from other different propositions that are known. It is only in virtue of the fact that we can become directly acquainted with features of the world that we can form thoughts that are about those features of the world. Both assertions, however, invoke a modal operator and there might be an important difference between the two claims depending on how we interpret that modal operator.

2. Epistemological Foundationalism Foundationalism is best known as an epistemological theory about the structure of knowledge and justified belief. The metaphor upon which the label is based is apt. The basic idea is that all knowledge and justified belief rest on a foundation of knowledge and justified belief that does not result from inference.¹ And the most familiar argument for the view is the famous regress argument. If the only way to know a given proposition is to infer that proposition from another proposition, problematic regress threatens. One can form a justified belief that P by inferring P from E only if one has justification for believing E. Much more controversially, one might also argue that to be justified in inferring P from E one must also have some sort of epistemic access to a relation of making probable holding between E and P (where entailment can be viewed as the upper limit of making probable). But if the only way to justify a belief is to infer what is believed from some other different ¹ In this chapter I’ll gloss over a number of important distinctions one might make among kinds of inference. In Fumerton (forthcoming) I distinguish (following Huemer, 2016) fully explicit inference, inference against a background, and non-inferential dependence. I also introduce the idea that there one might allow a sense of inferential dependence where what is relevant is only a disposition to believe (something that is different from an unconscious belief ). Richard Fumerton, Acquaintance: The Foundation of Knowledge and Thought In: Acquaintance: New Essays. Edited by: Jonathan Knowles and Thomas Raleigh, Oxford University Press (2019). © Richard Fumerton. DOI: 10.1093/oso/9780198803461.003.0012

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

proposition, then one could be justified in believing E only if one in turn justifiably inferred E from some other proposition F, and so on, ad infinitum. If we also need justification for believing that the relevant evidential connections hold, the regresses proliferate indefinitely. If all justification is inferential, then to be justified in believing anything one would need to complete not just one, but an infinite number of infinitely long chains of reasoning. And it doesn’t even make sense to talk about completing an infinitely long chain of reasoning. Just as God couldn’t finish counting the natural numbers, so also even God couldn’t complete an infinitely long chain of reasoning. So either there are beliefs that are non-inferentially justified or the most radical of all scepticisms is true—one has no justification for believing anything at all to any extent whatsoever. The argument, of course, is not uncontroversial. Both coherentists and infinitists offer alternatives. I’ve argued against these views elsewhere (1992, 2014), but here I’ll turn immediately to in-house disputes among foundationalists.

3. Acquaintance and the Analysis of Non-Inferential Justification It is one thing to be convinced that there must be non-inferentially justified beliefs if there are to be any justified beliefs at all. It is another to get clear about what makes a belief non-inferentially justified, and to get clear about which beliefs are noninferentially justified. It is an understatement to suggest that there is nothing even approaching a consensus on these critical meta-epistemological controversies. And given that there is no consensus on what makes a belief non-inferentially justified, it is surely not surprising that there is no consensus on which beliefs are noninferentially justified. For years I have defended an acquaintance theory of non-inferential justification, one that is inspired, at least, by Bertrand Russell’s work. One has strong or ideal noninferential justification for believing P when one is directly acquainted with the thought that P, the fact that P, and the correspondence or fit between the thought that P and the fact that P. I’m still inclined to think that one can get at least some degree of non-inferential justification for believing P when one is directly acquainted with the thought that P as one is also directly acquainted with a fact very similar to the fact that P and a relation very similar to fit or correspondence between the thought that P and that fact that is very similar to the fact that P. For example, as the experience of looking red slowly changes to one of looking orange, it is not implausible to suppose that one can move out of the looking red range and into the looking orange range while one still has some degree of introspectively based justification for believing that the experience is one of looking red. At the very least, it seems that one has more justification for thinking that it is an experience of looking red that one has for believing that it is an experience of looking navy blue.²

² For an argument against the view that an acquaintance theory is consistent with the existence of noninferentially justified but false beliefs, see Tucker (2016). For a response, see Fumerton (2016).

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

:      



But what is this acquaintance that is supposed to work its wonders in producing non-inferential justification? First, it is a relation—a real relation that can only hold given the existence of its relata. Acquaintance is not an intentional state like belief, fear, or desire. One can fear what doesn’t exist. One can want what one isn’t going to get. And one can believe what isn’t so.³ You cannot be directly acquainted with your pain unless both you and your pain exist. But what precisely is the relation under discussion? The disappointing answer has always seemed to me that it is indefinable. It is, as Moore (1903, 61) would say, one of those innumerable conceptual atoms out of which more complex concepts are built. It might be possible to ostend acquaintance though the attempt to do so is arguably question begging. Think about the familiar situation in which one is in pain, gets distracted by an interesting conversation or an exciting movie, and for a while doesn’t notice the pain. There are at least two possibilities. One is that the pain one felt temporarily ceased. The other is that the pain continued, but for a period of time one was no longer aware of it. I’m inclined to think that the latter is at least a plausible characterization of what happened, and if it is we can denote acquaintance as the relation we had to the pain before the distraction, a relation that ceased, and that began again when we were no longer distracted. And if we are happy with this characterization of the situation then we can recognize that there are all kinds of experiences to which one has potential introspective access, but where the access (the acquaintance) is intermittent. Sosa (2003) has argued that while one can make a distinction between experiences that one is aware of, and experience that one is not aware of, this amounts to nothing but the distinction between having a conscious experience when one knows that one is having the experience, and having a conscious experience when one does not know that one has the experience. I’m arguing that the phenomenology of the situation suggests that we make three distinctions. There is the having of an experience of kind X without our being directly aware of the experience, and without our knowing that we are having an experience of kind X. There is the having of an experience of kind X while one is directly aware of the experience but still doesn’t know that one is having an experience of kind X. And finally there is the having of an experience of kind X while one is directly aware of the experience, and while one also knows that one is having an experience of kind X. I suspect that almost any animal with sense organs (eyes, for example) has experiences—even fruit flies and shrimp. I also suspect (and hope, for their sakes) that at least some animals have experiences but are unaware of the experiences (just as I suspect that I am sometimes in pain without being aware of the pain). I am also almost certain that some animals have experiences, are aware of the experiences, but have no knowledge that they are having the experiences. Knowledge—at least propositional knowledge—requires thought (the application of concepts). Thought doesn’t require language, but it is highly likely that as a matter of contingent fact one wouldn’t have certain complex thoughts without language. And even thought ³ I’m inclined to think that one should infer from these truisms that intentional states are not relational. The search for ‘objects’ of baseless fears and frustrated desires will take one to a philosophical landscape that is more than a bit cluttered with exotic entities. The alternative is an adverbial account. I’ve defended such a view in Fumerton (2002).

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

that might initially seem simple is often really quite complex. Consider the thought that the surface of a physical object is red. One can’t identify the thought that the surface is red with a thought about the way in which the object currently appears. The visual appearance an object presents varies significantly depending on the external conditions of perception and the physical characteristics of the perceiver. That’s not to say that the concept of an object’s being red is independent of our understanding of appearance. It is only to assert that the connection between being red and looking red is exceedingly complex. While not a terribly popular view these days, I am still convinced that our understanding of physical redness has something to do with how the object would appear under certain conditions. But it is an understatement to suggest that the thought of the relation expressed by contingent subjunctive conditionals is complex. Philosophers are still arguing about how to understand the truth conditions of subjunctive conditionals. I doubt that most animals, or even young humans, are capable of forming such thought. Of course, the knowledge that we are primarily concerned with here is, arguably, knowledge that one is being appeared to in a certain way—not knowledge that an object has a certain colour. And one might suppose that this sort of knowledge requires very little by way of sophisticated thought. But I’m not sure that even this is so. While I think that there are appearances and that we are usually directly aware of them, it requires philosophical sophistication to form the thought of an appearance that is distinct from physical reality. With the right sort of philosophical training one can form such thoughts, but it does require philosophical sophistication. It requires, for example, the kind of dialectical sophistication that the moderns employed in arguing for a distinction between primary and secondary qualities. The above distinctions are often muddied by language. We talk about feeling pain and raise the question of whether there can be unfelt pains, a question that can easily be confused with the question of whether there can be pain of which one is unaware (of which one is not conscious). But as the so-called adverbial theory maintains, it is plausible to suppose that the grammatical structure of the sentence ‘John feels pain’ is misleading with respect to the thought that sentence expresses. Grammatically, the sentence has a subject, ‘John’, a transitive verb, ‘feels’, and an object expression, ‘pain’. And that might lead one to suppose that the sentence asserts a relation holding between John and pain. On such a view one can certainly wonder why John and his pain couldn’t continue to exist while the relation ceases. But the adverbial theorist argues that the noun ‘pain’ is really a kind of disguised adverb that modifies the kind of feeling we ascribe to John. To feel pain is to have a feeling of a certain sort. Feeling pain, feeling happy, feeling anxious, feeling sad, are all kinds of feelings. On the adverbial theory, it really doesn’t make sense to suppose that there could be unfelt pains. That would be to suppose that there can be feelings of the pain sort without there being any feelings. But if I am right and acquaintance is a genuine relation in which one stands to such states (among others) as feelings, one can still sensibly raise the question of whether there can be feelings of pain without the person who feels the pain being aware of the feeling in question. And one can certainly make sense of the different possibility that one can feel pain without knowing that one feels pain (perhaps because one lacks the relevant awareness, but also, perhaps, because one lacks the conceptual sophistication to form the thought that one is in pain).

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

:      



With the above distinctions in place, the acquaintance theorist has a number of plausible responses to the well-known problem of the appearance presented by a speckled hen. It seems entirely plausible to suppose that when one sees a manyspeckled hen, one has a visual experience with a complex but perfectly determinate character. If one is a sense-datum theorist one might argue that one’s visual experience is of a sense-datum having, say, forty-eight speckles. If one is an adverbial theorist, one will need to introduce the exceedingly ugly term ‘being appeared to forty-eight-speckledly’ to describe the determinate character of the experience. While the experience has that determinate character, few people with such an experience will know what the determinate character of the experience is. That seems almost undeniably right. But we have a number of explanations of why this is so. One is that one isn’t directly acquainted with the determinate character of the experience. If there is a distinction between determinate properties (like the specific nature of this pain) and the more abstract properties they determine (like the property of being in pain), there is no reason why one might not be aware of the abstract property without being aware of the determinate property. Furthermore, as Richard Feldman has argued (2006), one might not even have a ‘direct’ thought of the determinate experiential feature in question. I actually suspect that thoughts about large numbers are really complex thoughts that pick out non-linguistic properties through language and operations one performs at a linguistic level (employing numerals). And lastly, one might lack the propositional knowledge even if one has the relevant thought and is directly aware of the determinate character of the experience because one lacks direct awareness of the correspondence or fit between the thought and the relevant experiential character. This might be an appropriate point to acknowledge that the acquaintance conception of direct knowledge requires not just an intelligible concept of acquaintance, but also an intelligible concept of correspondence or fit between thought and world. This is not the place to defend a correspondence theory of truth. I have done so elsewhere (2002). But the long and the short of it is that correspondence is just as indefinable as is acquaintance. We can know precisely what it is because we are directly aware of correspondence, and with the right sort of philosophical training one can form the thought of correspondence. Fortunately for our ability to know, knowing that one is in pain doesn’t require having the concept of acquaintance, nor does it require having the concept of correspondence. It does require being acquainted with correspondence but that is a different matter.

4. Acquaintance, Non-Inferential Knowledge, and the Cartesian Concept of Foundational Knowledge It is always tempting for contemporary philosophers to interpret important historical figures in such a way that those figures either explicitly or implicitly support the view of the contemporary philosophers offering the interpretation. Descartes was uncontroversially one of the most prominent and influential epistemologists. While certainly not the first foundationalist, he provided one of the most compelling accounts of how to find secure foundations upon which to build the rest of what

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

we know. That account employed the famous ‘method of doubt’ and culminated in an identification of foundational knowledge with ideas that are ‘clear and distinct’. I’ve never understood what clarity and distinctness are supposed to be, and while it would be nice to suppose that it has something to do with a representation of reality coupled with direct awareness of the very reality that is the truth maker for the representation, I couldn’t offer such an interpretation with any sort of confidence. Nevertheless, there is, I think, a connection between how an acquaintance theorist would search for foundations and Descartes’s method of doubt as a criterion for discovering secure foundations. As I understand the method of doubt, the way to find the appropriate foundations upon which to build an edifice of knowledge is to sort through one’s beliefs, think about the justification one has for those beliefs, and reject for inclusion in the foundations of knowledge all those beliefs whose falsehood is consistent with the available justification.⁴ So, rightly or wrongly, Descartes suggested (at least initially) that the best justification we have for believing any contingent proposition describing the external world is compatible with such beliefs being false.⁵ The available justification is experiential and we can at least conceive of a world in which the experiences are caused by something other than the objects of ‘commonsense’. The traditional acquaintance theory often reaches precisely the same conclusion, and for similar reasons. Once the traditional acquaintance theorist becomes convinced that our best justification for believing what we do about the physical world is compatible with those beliefs being false, the acquaintance theorist concludes that one is not directly acquainted with the truth makers for those beliefs. You will recall that acquaintance is a real relation and that the obtaining of that relation requires the existence of its relata. Trivially, one cannot be acquainted with the fact that there is a table before one unless there is a table before one. If my justification for believing that there is a table before me now is compatible with its not being the case that the table exists, then I don’t have ideal foundational justification/knowledge for believing that the table is there. Now given that I have tentatively allowed for non-inferentially justified false belief that P when I am acquainted with a fact very similar to the fact that P, I couldn’t immediately reach the conclusion that we have no non-inferential justification for believing that the table exists. But if our reasons for concluding that we are not directly acquainted with the fact that the table exists have something to do with the intelligibility of evil demons, vivid hallucination, mad scientists stimulating brains in

⁴ This has to be stated a bit more carefully. It is not really the logical consistency that is crucial. After all, there is a sense in which we can conceive of our justification for believing a very complex mathematical proposition being consistent with that belief ’s being false. But if the mathematical proposition is true, it is, presumably, necessarily true and its falsehood isn’t consistent with anything. The proposition will fail the test for inclusion in the foundations because we can make sense of having our justification while the proposition is false. To say this more formally might require the introduction of a so-called relevance logic. ⁵ I say ‘initially’ because Descartes was no sceptic. He seemed to think that you can get yourself into an epistemic state where you can’t make sense of certain beliefs about the physical world being false. But you can only get yourself into such state by appreciating arguments for the existence of God and understanding what the existence of such a God implies vis à vis your being deceived with respect to beliefs about the physical world that you find ‘irresistible’.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

:      



a vat, and the like, it is hard to conjure up the right sort of fact to serve as a close relative to the fact of the table’s existence.⁶ Because our best justification for believing what we do about the physical world, the past, other minds, and the future is all consistent with the respective beliefs being false, the acquaintance foundationalist concludes that none of those beliefs are candidates for foundational knowledge. There is also a sense in which the acquaintance theory bears a striking resemblance not only to Descartes’s foundationalism but also to the now very popular theory of knowledge advanced by Tim Williamson. Williamson (2000) famously identifies evidence with knowledge, and argues that what is known always has a probability of 1. If we identify direct knowledge with the possession of strong or ideal justification as understood by the acquaintance theorist, the acquaintance theorist will also identify what is known with what has a probability of 1. As indicated above, one cannot be directly acquainted with the fact that P as one believes that P while one’s belief is false.⁷ Externalist versions of foundationalism are far more popular these days than are acquaintance conceptions of foundational knowledge. And some externalists will also secure a connection between strong justification and the impossibility of error. A causal theorist, for example, might identify my belief that P as foundationally justified when my belief that P is caused by the fact that P where there are no other belief states that are intermediate links in the causal chain leading from the fact to the belief. And again, if such causal connections count as a kind of justification, then trivially one couldn’t have that kind of justification for a belief while the belief is false. There is nothing to stop a reliabilist from insisting that one has foundational knowledge that P when one’s belief that P is formed by a belief-independent process (one that doesn’t take as ‘input’ belief states that need to be themselves justified) that is 100 per cent reliable. And if being formed by a reliable process counts as a kind of justification, then the belief formed by a process that is 100 per cent reliable couldn’t enjoy that justification while it is false. Most externalist views about foundational justification and knowledge make it much easier to secure foundational justification and knowledge. The foundations such views allow are much broader and, one might suppose, one will have a much easier time building upon them. I won’t rehearse here all the reasons that more traditional foundationalists have for rejecting externalism. There are all kinds of technical problems the views face. Causal theorists have a nightmare trying to handle deviant causal chains. Reliabilists need to deal with the generality problem. But the

⁶ It depends on one’s ontology. Harman once suggested to me in conversation (which he might not remember) that if one is a Meinongian one might be able to find a table that has being but not existence to serve as the object of direct awareness in the case of vivid hallucination of a table. I don’t really understand Meinongianism, but I suppose a subsistent table and an existing table might in some sense resemble one another? Johnston (2004) has suggested that in vivid hallucination of a physical object one is acquainted with the same ‘array’ of properties that are exemplified in veridical experience but are unexemplified in hallucinatory experience—or if they are exemplified they are just not exemplified by a physical particular. There is a sense in which that view might render the object of direct awareness in hallucination similar (at least in certain respects) to the object of direct awareness in veridical experience. ⁷ Of course Williamson wouldn’t embrace an acquaintance theory of non-inferential justification, and he has a much more expansive view of what has an epistemic probability of 1 for a given person.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

heart of the traditional foundationalist’s concern with externalism is the alleged inability of externalists to secure a connection between possessing non-inferential justification/knowledge and gaining philosophically satisfying assurance that what one believes is true. My belief that God exists might be caused by God’s existence; my belief that God exists might be produced by a process that is belief-independent and that is 100 per cent reliable (perhaps divine inspiration of some kind). But lacking any reason for thinking that the causal connection exists or that the belief is produced by the reliable process, how does that help me satisfy my philosophical curiosity about whether or not God exists? The acquaintance theorist is convinced that only when one gets directly before one’s mind the very truth maker for one’s belief does one get the kind of ideal, philosophically satisfying justification that one seeks.

5. Acquaintance and a priori Knowledge The examples I have been given of foundational knowledge whose sources are direct acquaintance have all been knowledge of contingent truths. If an acquaintance theory is on the right track, one of its great virtues might be that it can, in principle, provide a unified account of foundational justification and knowledge. It can find a common denominator to both a priori and empirical foundational knowledge. Both will have their source in direct acquaintance. The difference will be understood in terms of the kind of truth maker with which one is directly acquainted. So just as I am directly acquainted with my pain as I think that I am in pain and am acquainted with the correspondence between by thought and the pain. So I can find myself directly acquainted with a relation of darker than holding between the property of being red and the property of being pink, as I think that red is darker than pink and am directly acquainted with the relevant fact. Alternatively, as Hume would argue,⁸ the source of my knowledge that red is darker than pink might be my direct acquaintance with a relation between the idea of red and the idea of pink—more about this later.

6. Direct Acquaintance and Inferential Justification Everything I’ve said above concerns foundational justification and knowledge. But for the philosopher convinced that one needs some sort of access to evidential connections in order to form inferentially justified beliefs, acquaintance also plays a critical role. There are at least two possibilities. One is to allow for basic inferential justification for believing P on the basis of E only when one has a non-inferentially justified belief (whose source is acquaintance) that E makes probable P (where entailing is the upper limit of making probable). The other is to insist only that ⁸ Hume (1888, 458) argued that all truth depends either on matters of fact or on relations between ideas. Hume would tell you that your knowledge that red is different from blue is based on your awareness of the relation of difference holding between your idea of red and your idea of blue. I think he believed that you could ‘read off ’ that nature of facts from the nature of thoughts. And that is because he thought that there was a correspondence between thoughts and the facts that make them true. When such correspondence exists the facts and the thoughts share a kind of structure. The thoughts really are a kind of picture of the reality that makes them true.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

:      



one needs direct awareness of the relevant evidential connection—Fumerton (2016).⁹ Either view will introduce a relation of making probable with which one can be directly acquainted. The best hope is Keynes’s (1921) view of epistemic probability, but it is an understatement to suggest that the view is controversial.

7. Direct vs. Indirect Thought I suggested in my opening remarks that direct acquaintance is not only the source of all foundational knowledge but also the source of all foundational thought. This dramatic claim about a foundation for thought is plausible because of the following prima facie compelling idea. Much of what we think about we think about only indirectly—only through our thought of something else. I can think of your favourite colour without knowing what it is because I can think of you, I can think of being coloured, and I can think of preferences. It may be that I can think of you, in turn, only by thinking of that which is the bearer of certain properties. It may be that I can think of certain properties only by thinking of those properties as the bearers of still higher-level properties (e.g. the property whose exemplification causes certain things to happen). But this can’t go on forever. There must be some things I can think of directly. There must be thoughts that have the capacity to correspond to some aspect of reality—thoughts that have an intrinsic nature that allow them to correspond to one and only one aspect of reality. My thought of pain, it seems to me, is a good candidate for foundational thought. When I think of a certain sort of searing pain, I’m not thinking of the pain only as whatever it is that plays a certain functional role, or whatever it is that causes a certain sort of behaviour, or whatever it is that is typically caused by damage to a body. I probably do think that pain has a functional role to play, that pain does cause behaviour, and that it is usually caused by damage to bodily tissue, but my access through thought to the pain isn’t through any of these other thoughts I have. It is not like my thought of your favourite colour when I don’t know what that colour is. So there is a kind of regress argument for foundational thought that resembles in some respects the regress argument for foundational justification. There is such a thing as indirect thought. I can think of some things only through thinking of others. I can think of Jack the Ripper and I can form the thought that he was insane. When I think of Jack the Ripper, I am thinking of that person (whoever he or she was) who committed those atrocities in London in the late nineteenth century. But it can’t be the case that all thought is like that. If for all x, thinking of x always involves thinking of something else y, then every thought would be infinitely complex—complex in a way that precludes its very possibility. Just as the conclusion that there is a foundation for justification and knowledge leaves open the question of how to understand such foundations, and leaves open the question of which beliefs are foundationally justified/known, so also the conclusion that there is a foundation for thought leaves open the question of how to understand foundational thought, and leaves open the question of which ⁹ See also Sam Taylor (2015).

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

thoughts are foundational. And it is here that many philosophers think that the relation of acquaintance again plays a fundamental philosophical role. It seems plausible to some of us to suppose that the extension of foundational thought is restricted to items with which we are or have been directly acquainted. So to return to the example of pain, I am able to think of pain directly, but only because I have been directly acquainted with pain. I am able to think of phenomenal redness, but only because I have been directly acquainted with phenomenal redness. An extreme version of empiricism might restrict the class of entities about which one can think directly to those that have been introspectively accessible. But if, like Russell, one concludes that one can be directly acquainted in thought with such abstract entities as numbers, one might also suppose that one can have direct thought of them. If one can form a direct thought of some particulars and properties, and one can also form a direct thought of causal connection, then one can ‘build’ indirect thought out of direct thought. I can think not only of pain, but of whatever it is that causes pain. I can think of being appeared to in that phenomenally red way, and also think of whatever it is that plays a certain causal role in producing that phenomenal redness. But this already oversimplifies the way in which direct thought can be about something. The immediate problem is that there is no one event or state of the world that causes pain or an appearance of phenomenal redness. There are causal chains leading to these effects, and each link in the causal chain has as good a claim to being a cause of the relevant effect as do the others. So if we hope to pick out surface properties of an object when we think about the cause of a given visual experience, we will need to ‘triangulate’ on the relevant property.¹⁰ We’ll need to think of the common cause of both this and other experiences of red. That common denominator (whatever it is) might be the property we are trying to think about when we think about the redness of a physical object. I have so far talked about acquaintance with X as a condition for having a direct thought of X. I still haven’t said anything by way of characterizing what direct thought is. There are at least two possibilities. One is that acquaintance figures into the very analysis of direct thought. This was, I think, Russell’s view.¹¹ To think of X directly just is to stand in a relation of direct acquaintance to X. To be plausible, such a view would probably require, among other things, that there be universals, and that we can be directly acquainted in thought at a given time with such universals even without our being directly acquainted at that time with the exemplification of

¹⁰ See Evan Fales (1990). ¹¹ Russell didn’t explicitly state the view in terms of the idea of direct thought. He often raised it in the context of his related distinction between genuine names and disguised descriptions. He says in his reflections on his philosophical development (1959, 125): ‘I have often maintained a principle, which still seems to me completely valid, to the effect that, if we can understand what a sentence means, it must be composed entirely of words denoting things with which we are acquainted or definable in terms of such words’. The principle is stated in terms of the conditions required for understanding a sentence, but it is clear from the corpus of his work that he believes that thought doesn’t conceptually require language, and that the view would be extended to any sort of complex thought. To be possible that thought would need to be constructed out of simple thoughts each of which relates the thinker to an object with which the thinker is directly acquainted.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

:      



the universals. So I can still think directly of the searing pain I experienced five years ago, because I can still hold ‘directly’ before my mind the property of being in such pain. The only alternative for the acquaintance theory is to suppose that something like that property of being in pain is exemplified in the very imagination of the pain. If Hume were right and simple ideas are ‘pale copies’ of what they are ideas of, then in forming the idea of pain we experience something very much like the pain—just less vivid. But phenomenologically that just seems wrong to me. Fortunately for us, we can remember (and thus think about) severe pain without experiencing anything remotely like the severe pain. If we analyse direct thought of, say, a property, as direct acquaintance with the property, we might still suspect that it would be causally impossible to stand in such a relation to the property without having at least once been directly acquainted with the exemplification of the property. The intuition that Jackson’s (1986) colourdeprived Mary would never have been able to think directly of phenomenal redness is shaped largely by the idea that without having actually had the experience in question Mary would never have been able to think of that property in the direct way she could after having had the experience. But I take it that having had the experience once, Mary would continue to be able to think directly of phenomenal redness as long as she was able to remember what that experience was like. The intuition pushed by Jackson’s thought experiment is perfectly compatible with its being the case that all along, even before her colour experiences, Mary was able to think indirectly of phenomenal redness. She was able to think of that experience (whatever it is) that people have introspective access to as they look at apples, as their brain undergoes certain changes, that causes them to say such things as ‘That looks red’, and so on. But we don’t think she could think of phenomenal redness directly—think of it in the way that we who are not colour-deprived can. As David Lewis pointed out (2004, 82), however, the subjunctive conditional that we think is true doesn’t seem to be necessarily true in any strong sense of ‘necessarily’. We can conceive of a world in which colour-deprived Mary did form the thought of phenomenal red experience even without ever having had it. We can conceive of a world in which a pain-deprived Mary was able to think directly of pain even without ever having experienced it. We (at least we who reject externalist accounts of thought and its object) just don’t think that as a matter of contingent fact Mary would form such thoughts.¹² So when we think that there is a connection between having been directly acquainted with something and being able to think of it directly, we might only be thinking that the former is something like a causally necessary condition for the latter. Or perhaps, more weakly still, we might be thinking only that the former, given certain conditions concerning how our brains currently work and the current state of

¹² A more nuanced view might allow that it is causally possible to form direct thought of properties that one has never been acquainted with provided that one has been acquainted with properties very much like the unexemplified property. Think about Hume’s (1888, 6) famous discussion of the missing shade of blue. While Hume generally embraced the principle that one can form simple ideas only of what one has directly experienced, he allowed that one might be able to fill in a noticeable ‘gap’ in a continuum of shades of blue. He dismissed this exception to his general principle, however, as insignificant.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

neurological science, is causally necessary for the latter.¹³ If the relation between direct acquaintance and direct thought is contingent in this way, then we will need to search elsewhere for an analysis of what direct thought is. That search might take us to a view similar to one embraced by Gustav Bergmann (1964, 1967) and his student Laird Addis (1989), a view that in some respects inspired my own view (2002) about the nature of thought. On their view to have a thought of X is to exemplify a certain property, where the exemplification of that property has the capacity to correspond (in the case of true thought) to some aspect of the world. There are thoughts that don’t correspond to anything, but they still have the capacity to ‘reach’ out to some feature of the world. The property, exemplification of which just is thinking, is unanalysable. One can identify properties of the property, but one can’t define the relevant property. But we are all in a position to know what the property is by virtue of our acquaintance with its exemplification. The above view leaves open the conditions that are causally relevant to the formation of such thought. Neither Bergmann nor Addis embrace my conception of acquaintance and thus, trivially, would not share speculations about direct acquaintance being a causally necessary condition for the formation of direct thought. They both treat direct acquaintance as just another species of thought (every thought has both a content and is of a given species for them), where on my interpretation of their views, to have a thought is just to exemplify a nonrelational property, albeit one that has the capacity to correspond to some feature of the world.¹⁴ Which ‘acquaintance’ conception of direct thought is more plausible—one that analyses direct thought in terms of direct acquaintance, or one that embraces the much more cautious thesis that direct acquaintance might be a contingently necessary condition for direct thought? The candid answer is I’ve never been sure. It is one of those questions in philosophy the answer to which is closely tied to the answer one should give to a host of other difficult philosophical issues. This much seems clear to me. As I said earlier, one can now think directly of properties the exemplification of which one is not currently acquainted. Furthermore, if that is possible it seems equally plausible to suppose that one can think directly of a property that has never been exemplified—Hume’s missing shade of blue, for example (discussed in note 12). That, by itself, is no obstacle to the Russellian conception of direct thought provided that one can identify properties with universals, provided that universals can exist even if they have never been instantiated, and provided that we can be directly acquainted with such universals. Although there are serious dialectical pressures to embrace the theory that there are uninstantiated universals, I am still attracted to the view that at least some properties are best treated as tropes, where

¹³ Perhaps we will sometime be able to routinely manipulate the brain so as to produce the direct thought of phenomenal redness even in unsighted people. ¹⁴ This is, I believe, the view that Addis settled on. Bergmann always included in his analysis of thought a ‘meaning relation’. The thought that P means P whether there is a fact that P or not. But I put ‘meaning relation’ in scare quotes precisely because Bergmann denies that the intentional relation of meaning requires any existing relata. But he also doesn’t seem to want to embrace a Meinongian realm of subsistence, so I’m not sure with what we are left.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

:      



what makes a class of tropes all red tropes is some sort of resemblance between them. We might need to acknowledge certain relations as universals, but the question is complicated. In any event, until I become convinced that there are uninstantiated universals with which one can be acquainted in thought, I’m still attracted to the adverbial theory of thought I described above—a theory that doesn’t analyse direct thought in terms of direct acquaintance with its object, but allows that past direct acquaintance might be a causally necessary condition for the formation of the thought.

8. Externalist Accounts of Thought Acquaintance theories of direct thought are not the only, nor are they now the most popular, attempts to understand how thought reaches out to the world. These days content externalism is coming close to being the received view. Kripke (1980), Putnam (1975, 1978), Dretske (1995), and many others are convinced that to understand how at least some language and thought succeeds in being about some entity X we need to look to the causal history of both language use and neural activity. After all these years, it still seems to me that the basic idea behind the causal theory is little more than a sketch. But the sketch is just this. When I use a term like ‘Josephus’ in the sentence ‘Josephus was an historian’, I am referring to that person (whoever it was) that figured in the initial link of a long causal chain culminating in my use of the token ‘Josephus’ in that sentence. When I think of Josephus there is some neural activity that is the ultimate effect of that same sort of causal chain that begins with some tokening of Josephus. Kripke always denied that he was giving an overall theory of reference, though his explanation of his reticence was always a bit cryptic. One suggestion, however, is that he had no real account of the initial ‘baptism’ that secures the initial reference of a term or a thought token. And here, one might suggest that acquaintance still has a role to play. One could (though one needn’t) insist that it is only in virtue of being acquainted with some item can one fix a label to it or secure a thought about it—it is only in virtue of acquaintance that one can secure the first critical link in a causal chain that enables people thousands of years later to refer to the same item. But even if one conceded this much, the resulting theory would be a far cry from the kind of acquaintance theory that Russell advanced. That acquaintance theory requires acquaintance that is cotemporaneous with direct reference and direct thought. This is not the place to rehearse arguments for and against content externalism. Elsewhere (Fumerton, 1989) I have argued that one can ‘steal’ all that is plausible from a causal theory and incorporate it into a Russellian descriptive theory of ordinary proper names. The resulting theory will have the best of two philosophical worlds. It will solve the problems of informative identity and referentially opaque sentences while accommodating the externalist’s insight that traditional descriptivists don’t have the resources to secure for many people plausible definite descriptions that can capture the meaning of the ordinary proper names they use. Furthermore, though the issue is complex, I have argued (2003) that content externalism can’t handle the datum that we sometimes have unproblematic access to the content of our thoughts.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

9. Conclusion There must be foundations for knowledge and thought if knowledge and thought are possible. Direct acquaintance enters into the most plausible story of what makes foundational knowledge possible. It is also figures into the most plausible story of what makes foundational thought possible. But there are two quite different theories of direct thought that identify quite distinct roles for acquaintance to play. On the Russellian view, direct thought is identified with direct acquaintance with what the thought is about. On the Fregeian view, past direct acquaintance with X is a plausible candidate for a causally necessary condition for direct thought of X.

References Addis, L. (1989) Natural Signs: A Theory of Intentionality, Philadelphia, PA: Temple University Press. Bergmann, G. (1964) Logic and Reality, Madison, WI: University of Wisconsin Press. Bergmann, G. (1967) Realism: A Critique of Brentano and Meinong, Madison, WI: University of Wisconsin Press. Dretske, Fred (1995) Naturalizing the Mind, Cambridge, MA: MIT Press. Fales, Evan (1990) Causation and Universals, London: Routledge. Feldman, Richard (2006) ‘BonJour and Sosa on Internalism, Externalism, and Basic Beliefs’, Philosophical Studies 131: 713–28. Fumerton, Richard (1989) ‘Russelling Causal Theories of Reference’, in Rereading Russell, Minneapolis, MN: University of Minnesota Press, 108–18. Fumerton, Richard (1992) ‘A Critique of Coherentism’, in Louis Pojman (ed.), The Theory of Knowledge: Classical and Contemporary Readings, 1st edn, Belmont, CA: Wadsworth, 245–50. Fumerton, Richard (2002) Realism and the Correspondence Theory of Truth, Boston, MA: Rowman and Littlefield. Fumerton, Richard (2003) ‘Introspection and Internalism’, in Susana Nuccetelli (ed.), New Essays on Semantic Externalism, and Self-Knowledge, Cambridge, MA: MIT Press, 257–76. Fumerton, Richard (2014) ‘Infinitism’, in John Turri and Peter Klein (eds), Ad Infinitum: New Essays on Epistemological Infinitism, Oxford: Oxford University Press, 75–86. Fumerton, Richard (2016) ‘Prospects for Traditional Internalism’, in Michael Bergmann and Brett Coppenger (eds), Intellectual Assurance: Essays on Traditional Epistemic Internalism, Oxford: Oxford University Press, 239–57. Fumerton, Richard (Forthcoming) ‘Inferential Internalism and the Problem of Unconscious Inference’. Huemer, Michael (2016) ‘Inferential Appearance’, in Brett Coppenger and Michael Bergmann (eds), In Intellectual Assurance, Oxford: Oxford University Press. Hume, David (1888) A Treatise of Human Nature, ed. L. A. Selby-Bigge, Oxford: Clarendon Press. Jackson, Frank (1986) ‘What Mary Didn’t Know’, Journal of Philosophy, 83: 291–5. Johnston, Mark (2004) ‘The Obscure Object of Hallucination’, Philosophical Studies 120: 113–83. Keynes. John (1921) A Treatise on Probability, London: Macmillan. Kripke, Saul (1980) Naming and Necessity, Cambridge, MA: Harvard University Press. Lewis, David (2004) ‘What Experiences Teaches’, in Peter Ludlow, Yujin Nagasawa, and Daniel Stoljar (eds), There Is Something about Mary, Cambridge, MA: MIT Press, 77–103.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

:      



Moore, G. E. (1903) Principia Ethica, Cambridge: Cambridge University Press. Putnam, Hilary (1975) ‘The Meaning of “Meaning” ’, in Philosophical Papers, Vol. 2, Cambridge: Cambridge University Press. Putnam, Hilary (1978) Reason, Truth and History, Cambridge: Cambridge University Press. Russell, Bertrand (1959) My Philosophical Development, London: Allen and Unwin. Sosa, Ernest (2003) ‘Privileged Access’, in Quentin Smith (ed.), Consciousness: New Philosophical Essays, Oxford: Oxford University Press, 273–92. Taylor, Samuel (2015) ‘Is Justification Easy or Impossible? Getting Acquainted with a Middle Road’, Synthese 192: 2987–3009. Tucker, Chris (2016) ‘Acquaintance and Fallible Non-Inferential Justification’, in Michael Bergmann and Brett Coppenger (eds), Intellectual Assurance: Essays on Traditional Epistemic Internalism, Oxford: Oxford University Press, 43–60. Williamson, Timothy (2000) Knowledge and Its Limits, Oxford: Oxford University Press.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

12 Objectual Knowledge Katalin Farkas

1. Kinds of Knowledge The Western philosophical tradition has focused on knowledge of truths as the central and paradigmatic case of knowledge. Here is a typical first sentence of an introduction to a book on epistemology: ‘Epistemology is the theory of episteme, of knowledge. Ever since Plato it has been thought that one knows only if one’s belief hits the mark of truth and does so with adequate justification’. (Bonjour and Sosa, 2003, 1)

Occasionally, before launching into the discussion of knowledge of truths, treatises on epistemology mention other kinds of knowledge. Again, the following characterization is typical: It is common in epistemology to distinguish among three kinds of knowledge. There’s the kind of knowledge you have when it is truly said of you that you know how to do something—say, ride a bicycle. There’s the kind of knowledge you have when it is truly said of you that you know a person—say, your best friend. And there’s the kind of knowledge you have when it is truly said of you that you know that some fact is true—say, that the Red Sox won the 2004 World Series. (Fantl, 2016)

I am going to call the three kinds of knowledge mentioned here practical, objectual, and factual or propositional knowledge. Since it has been in the centre of attention, our understanding of factual knowledge is quite advanced. Factual knowledge, on virtually everyone’s opinion, is a specific kind of possession of a truth. A few further ideas are widely accepted as well (though each have a few dissenters). Possessing a truth means to have a mental attitude towards something that can be true or false, usually regarded to be a proposition. According to many, the mental attitude in question is a belief, but true belief is not sufficient for knowledge, there are further conditions: justification, or production by a reliable process, or by a cognitive ability or intellectual virtue, integration into an agent’s cognitive character, a modal condition of ‘safety’ or ‘sensitivity’, and so on. One function of these further conditions is to exclude a merely accidental or lucky possession of truth which is deemed to be incompatible with knowledge. On a somewhat different view, the relevant mental attitude towards a truth is not belief, but rather the mental attitude of knowing. Katalin Farkas, Objectual Knowledge In: Acquaintance: New Essays. Edited by: Jonathan Knowles and Thomas Raleigh, Oxford University Press (2019). © Katalin Farkas. DOI: 10.1093/oso/9780198803461.003.0013

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

 



Having this attitude excludes a merely lucky possession of truth by itself. So on a widely accepted conception, factual knowledge is at least a non-accidental possession of a truth (and possibly more). The nature of practical knowledge has been the subject of some debate. On the so-called intellectualist view, practical knowledge is a sub-category of factual knowledge (Stanley and Williamson, 2001). According to the opposing, anti-intellectualist view—influentially defended by Gilbert Ryle (1949), among others—practical knowledge is not a species of factual knowledge, but rather some sort of ability or disposition to successfully perform a certain action. However, a plausible antiintellectualist account of practical knowledge will make it clear why it is a species of knowledge. Presumably, defenders of the intellectualist view don’t want to deny that we can meaningfully talk about abilities to perform actions; what they deny is that these abilities constitute knowing how to do something. So the antiintellectualist should be able to explain what is common to factual and practical knowledge that makes them both knowledge. One particular version goes as follows (Farkas, 2018). Knowing how to G involves the ability to reliably succeed in a mentally guided execution of G-ing. Factual knowledge is one kind of cognitive achievement. It is an achievement partly because it involves success, where the particular success component of factual knowledge is having a true belief. The parallel success component of practical knowledge is a mentally guided successful execution of a certain type of action. In both cases, there is a mental attitude towards some content that is typical of the piece of knowledge in question. In neither case can the success be a matter of mere accident or luck. Some further conditions make sure that the success is non-lucky. So both factual and practical knowledge are at least non-lucky cognitive achievements (and possibly more).

2. Objectual Knowledge: A Third Kind? On the conceptions sketched in Section 1, both factual and practical knowledge form a proper kind. Knowing different types of facts—e.g. perceptual or mathematical, past or present, modal or actual—may have quite different conditions, but still, all instances of factual knowledge share a nature, in being some sort of nonaccidental possession of a truth. Similarly, instances of practical knowledge—both on the intellectualist and anti-intellectualist account—share a certain nature. And since both kinds are properly regarded as knowledge, it looks like we have a genus—knowledge—with two species (or with a sub-species): factual and practical. Let us turn now to objectual knowledge, which is the subject of this chapter. I think it is fair to say that our understanding of objectual knowledge, as a kind, lags behind our understanding of factual and practical knowledge. There have been attempts to analyse some sub-categories of objectual knowledge: knowing persons, or knowing experiences, for example. But it is hard to find detailed competing analyses of objectual knowledge as a general category. As we shall see, the few existing accounts tend to rely on the notion of acquaintance: it’s been suggested that on a suitably defined philosophical concept, the relation of acquaintance constitutes objectual knowledge.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

If we follow the logic of the quote above, then we should expect objectual knowledge to be a third species in the genus of knowledge. This would mean, first, that objectual knowledge is in some sense the same with respect to different types of objects, and second, that it shares something with the other two kinds of knowledge. Further, similarly to practical knowledge, objectual knowledge is potentially irreducible to factual knowledge, forming a genuine separate kind. In this chapter, I will try to raise doubts about this picture. First, I will argue that if objectual knowledge is expressed by our usual talk of ‘knowing things’, then it is not a proper kind with a uniform nature: the relation that constitutes thing-knowledge is quite different for different types of things. Knowing a person, knowing a place, knowing our rights, and knowing feelings are all quite different relations. Second, I will question that ‘knowledge of things’ is a recognizable member of the genus knowledge: if we agree that factual knowledge is the central and paradigmatic case of knowledge, then at least some of the relations denoted by ‘knowing things’ are not usefully classified as knowledge that forms the subject matter of epistemology. Knowing a person, in particular, consists at least partly of a relation that is not a cognitive achievement. I do have some sympathy for the third assumption: at least some types of knowledge of things, while they involve and enable factual knowledge, are not reducible to it. However, it’s precisely the irreducible aspect that arguably places these relations outside the genus that includes factual and practical knowledge. I will conclude this chapter by suggesting that the most promising direction to develop an account of sui generis objectual knowledge is to depart from the ordinary ascriptions of ‘knowing things’ and rely on a philosophical notion of acquaintance. The range of things that can be known objectually will be limited to the kind of things with which we can be acquainted—for example, to features of conscious experiences. But a lot of further work will be needed to show that this relation is usefully classified together with factual and practical knowledge as a third kind.

3. Knowledge by Acquaintance and Description Contemporary discussions of objectual knowledge often invoke Bertrand Russell’s distinction between ‘knowledge by acquaintance‘ and ‘knowledge by description‘ (Russell, 1910, 1912). Russell has probably done more than anyone to muddy waters in this area, so before we start our discussion, we will do well by trying to clear up some of the confusions connected to his work. In The Problems of Philosophy, towards the end of the chapter on idealism, Russell introduces a distinction between two senses of the word ‘know‘: one applies to knowledge of truths, and the other to knowledge of things. He notes that the distinction roughly corresponds to the distinction between the German wissen and kennen, or the French savoir and connaître. The next chapter is devoted exclusively to knowledge of things, which, in turn, falls into two categories: knowledge by acquaintance and knowledge by description (also the title of that chapter). So on Russell’s classification, knowledge by description is a sub-category of knowledge of things, rather than knowledge of truths. But it’s clear from the examples he gives that much (or all) of knowledge by description is factual knowledge, rather than sui generis objectual knowledge. A more promising direction for our current purposes is to ask

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

 



whether Russell’s ‘knowledge by acquaintance’ is sui generis objectual knowledge. Indeed it is quite common in contemporary discussions to equate knowledge by acquaintance with objectual knowledge (e.g. Zagzebski, 1999; Martens, 2010). Russell characterizes the difference between knowledge by acquaintance and knowledge by description as a difference in their ground: knowledge by acquaintance is ‘logically independent from any knowledge of truths’, while knowledge by description ‘always involves . . . some knowledge of truths as its source and ground’ (Rusell, 1912, 19). On Russell’s view, acquaintance is a direct awareness of an object, without the mediation of inference or knowledge of truths. His primary example of acquaintance is our awareness of sense-data. It can be debated whether there is such a cognitive relation, and whether we are acquainted only with sense-data or possibly with other kinds of objects. But suppose there is such a relation; what is ‘knowledge by acquaintance’ as opposed to mere ‘acquaintance‘? Or are these perhaps the same? In other words, is being acquainted with something a form of knowledge in itself ? Some people deny this. John M. DePoe claims that direct acquaintance never suffices for knowledge; knowledge by acquaintance, like all knowledge, is knowledge of truths, and its distinguishing feature is that it’s based on being acquainted with something (DePoe, 2013). Richard Fumerton and Ali Hasan think that Russell probably ‘equivocates between the relation of acquaintance and the special kind of knowledge of truth (foundational knowledge) whose sole source is acquaintance’ (Hasan and Fumerton, 2014). To avoid confusion, they propose to restrict knowledge by acquaintance to foundational knowledge of truths. Knowledge by acquaintance, that is, knowledge of truths based on acquaintance, is an interesting topic that raises a number of important questions, but I want to put it aside for the moment because it’s not our central concern. My interest here is whether there is a separate kind of knowledge which relates to things rather than to truths. It’s not clear to me whether Russell is after such a knowledge. Clarifying this issue would require a closer look at the relevant passages, something I have no space to do here. If acquaintance is the key to objectual knowledge, I propose to consult others on this issue who address this point much more clearly and explicitly than Russell does (see Section 9). Before we move on, let me note another unhelpful idea put forward by Russell. In the paper ‘Knowledge by Acquaintance and Knowledge by Description’ (Russell, 1910), Russell introduces the contrast between the two types of knowledge by talking of a case where ‘we know propositions about “the so-and-so” without knowing who or what the so-and-so is’ (108). In these cases, he says, the subject is merely described. In the subsequent discussion, knowledge by (mere) description is again contrasted with knowing who or what someone or something is (113). It looks like Russell equates knowledge by acquaintance with cases of ‘knowing who’ or ‘knowing what’. This isn’t a course worth following. As we shall see below, for example, ‘knowing who’ is most plausibly a form of factual knowledge, and it clearly comes apart from ordinary cases of knowing someone. I know who Maryam Mirzakhani was (a mathematician, so far the only female winner of the Fields Medal), but I didn’t know her. The usual sense of ‘knowing what’ arguably also comes apart from Russell’s special sense of knowing something by acquaintance. I could know what you are experiencing at the moment—the taste of Marmite—without being

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

acquainted with the taste of Marmite. So I propose to steer clear of Russell’s discussion of the matter.

4. The Language of Knowledge Ascriptions The following should be uncontroversial: when we gain knowledge about portions of the world, we enter into all sorts of interesting relationships to the things that are included in that portion. Facts have things as constituents, and for example when we have a perceptual knowledge of a fact, we commonly perceive the things that constitute that fact. But this observation does not require postulating a third kind of knowledge. It would require a significant further argument to show that knowledge of things is more than, or different from, knowing facts about them. So why think that there is objectual knowledge? An initial motivation may come from the linguistic constructions we use to attribute knowledge. Factual knowledge is often attributed by using that-clauses, as in (1)

Isaac knows that apples fall to the ground.

This may be contrasted with attributions where the verb ‘know’ is combined with a noun phrase (or ‘NP’ for short), as in the following examples: (2) (3) (4) (5)

René knows Marin. Federico knows Rome. Meno knows the way to Larissa. Endre knows some good cafés in Budapest.

In (1), it looks like the object of knowledge is the true proposition that apples fall to the ground. On some accounts, knowledge is not a relation to a proposition, but to a fact, where a fact is understood to be more than a true proposition. This detail will not matter for our discussion: the important thing is that on the central notion of knowledge, knowledge aims at a truth, or something that bears an essential relation to a truth, like a truth maker. (I will omit this qualification from now on, but it should be understood implicitly.) The question I ask in this chapter is whether there is a distinct type of knowledge whose object is not a truth, but rather a thing (or some things). In examples (2)–(5), the apparent object of knowledge is not a proposition or fact, but a person, a city, a way, and some cafés. When the distinction between knowing facts and knowing things is introduced, it is often noted that in a number of languages other than English, ‘know’ is translated by two different words. For example, German distinguishes between ‘wissen’ and ‘kennen‘, French between ‘savoir’ and ‘connaître‘, and Hungarian between ‘tudni’ and ‘ismerni‘. In each case, the second of the pairs is used to translate (2) and (3), and at least some occurrences of (4) and (5) above. This phenomenon seems to lend immediate, if somewhat superficial support to my previously mentioned claim that there is no common genus for factual and objectual knowledge. The support is superficial, because it is possible that English has a classificatory insight here that other languages lack. And as Edward Craig notes (1990, 140 ff.), English is not the only language using one word for these two concepts; Craig reports having heard the same about a handful of Asian and African

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

 



languages, and we may add Russian as another example (while interestingly, Russian has a separate verb for knowing how). In addition, Craig recalls that in several languages (including German and Hungarian), the word for ‘theory of knowledge’ is in fact related to the second of the words in the pairs mentioned above, the one that’s used to express a relation to things. The French word for ‘knowledge’ is ‘conaissance’. Hence Craig reasonably claims that the second, object-oriented sense of ‘know’ seems to be somehow related to the first, and it would be a desirable feature of a theory of knowledge if it could account for this connection. I agree with this, and I will hopefully offer such an explanation (not dissimilar to the explanation offered by Craig). But I will claim that the relation between the two kinds is less straightforward than belonging to the same genus.

5. The Non-Acquaintance Sense of ‘Know + Noun Phrase’ Constructions We often attribute knowledge in English by using a ‘know + noun phrase’ (know-NP) construction, as we saw in the examples (2)–(5) above. The common grammatical form covers some important differences. Rene’s knowing Marin, or Federico’s knowing Rome are most plausibly understood as some sort of familiarity or acquaintance. But not all know-NP constructions express the kind of acquaintance that figures in these two sentences. For example, consider the sentence ‘Meno knows the way to Larissa’. On one understanding, this does mean that Meno is acquainted with the road: presumably, he has travelled to Larissa, he is familiar with the road’s various features. However, the same sentence can be perfectly well applied even if Meno has never been on the road, but merely knows which is the way to Larissa, under some appropriate description. Imagine that Meno is about to set off to Larissa for the first time in his life, after carefully studying and memorizing the directions. If someone is wondering how Meno will get to Larissa, we can reassure her that he’ll be fine, he knows the way. The same ambiguity is present in many other noun phrase attributions. (5) could imply that Endre has been to the cafés in question, but it is also applicable if Endre has never been to Budapest, and just learned the names of the places from a guidebook.¹ On one plausible proposal, the non-acquaintance reading of many of these sentences can be reformulated: Meno knows the way to Larissa just in case he knows which way leads to Larissa. This directs our attention to another group of attributions: know-wh attributions. These combine the verb ‘know’ with a so-called ‘wh-clause’ containing an implicit question, as in the following examples: (6)

Cain knows where his brother is.

¹ In Hungarian, the non-acquaintance and the acquaintance senses of know-NP attributions are often translated by the two different verbs that correspond to ‘know’ (‘tud/ismer’ as mentioned above; ‘Menón tudja az utat’/’Menón ismeri az utat’). As I understand, German is somewhat similar, at least for some objects, so ‘Menon weiss den Weg’ and ‘Menon kenne den Weg’ are both well formed, and while the second implies acquaintance, the first doesn’t. In French, ‘savoir’ (the non-acquaintance senses of knowledge) does not naturally take noun phrases (except for poetic contexts); in French, the natural translation of the non-acquaintance sense of a know-NP sentence is a know-wh sentence.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

(7) Eve knows which fruit is forbidden. According to the standard analysis, someone knows-wh only if she knows a/the proposition that answers the question implicit in the wh-clause. There may be additional conditions for sufficiency, but even if there are, know-wh is basically a type of propositional knowledge. Thus (7) is true because Eve knows that the apple from the tree of knowledge is forbidden, and this answers the question ‘Which fruit is forbidden?’² The proposal I just considered is that the non-acquaintance sense of know-NP attributions is equivalent to a know-wh attribution.³ Know-wh, in turn, is the same as (a possibly special kind of) propositional knowledge. Suppose that Meno knows that the E75 leads from Athens to Larissa; since this answers the question ‘Which way leads to Larissa?’, Meno knows which way leads to Larissa, which is equivalent to his knowing the way to Larissa. So the non-acquaintance sense of know-NP is equivalent to some sort of propositional (factual) knowledge. I suggested the outlines of an analysis for some knowledge-NP attributions via an analysis of know-wh attributions. It also makes sense to proceed the other way around. Many know-wh attributions are easily converted into a know-NP format, for example (6) can be converted to: (6’) Cain knows the whereabouts of his brother. A plausible analysis for know-the sentences: S knows the F only if, for some x, she knows that x is the F. Other quantified phrases follow a similar pattern, and additional conditions may be required for sufficiency. Then we can proceed to give this analysis for the equivalent know-wh sentence. On either way of proceeding, we find that the core of both knowledge-wh and a lot of knowledge-NP is knowledge of propositions. If the states attributed by many know-wh and know-NP sentences are in fact states of propositional knowledge, one may wonder why we have these other forms of attribution at all. There are good reasons: both forms have functions that know-that attributions cannot serve. First, one can attribute knowledge by using know-wh and know-NP even if she herself is ignorant of the matter. Someone who doesn’t know the way to Larissa can still claim that Meno does. Know-wh and know-NP can also be used to attribute knowledge to a person over a changing subject matter, where factual attributions would require specifying many distinct propositions; for example, in saying that Kate always knows (what) the latest fashion (is). Both uses are made possible by the fact that neither know-wh nor know-NP constructions identify explicitly a propositional object in their content—even if they are in fact made true by knowing a proposition. To sum up this section: some cases of knowing the Fs or knowing some Fs are best understood as cases of know-wh, which, in turn, are best understood as some kind of

² For a more detailed discussion of know-wh and for some proposals to refine the standard analysis, see Farkas (2016, 2017). ³ See Brogaard (2009) for a discussion and defence of a similar view for ‘know-the’ and ‘know-a’ attributions.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

 



propositional knowledge. What suffices for ‘knowing who’ or ‘knowing which’ is a highly context-dependent matter. For example, there doesn’t seem to be a privileged list of propositions about a person whose knowledge would be either necessary or sufficient for knowing who that person is, in all contexts. But even if we cannot generally single out these propositions, in each context, knowing who someone is will simply consist in knowing some propositions. If all thing-related knowledge was like this, then there would be no sui generis objectual knowledge. So we need to search further.

6. Knowledge in the Biblical Sense Within the category of knowledge of things, certain kinds of entities have received particular attention: for example, knowledge of sensory or mental items, knowledge of the self, or knowledge of persons. However, in my readings in epistemology, I haven’t yet come across a proper discussion of one sense or type of knowledge: knowledge in the biblical sense. What explains this glaring omission, and what should we do about it? One option would be to try to work out an analysis of this type of knowledge. Following a current trend, we could design a few vignettes, which all had labels printed in small capitals, like Eden: Eden. Adam and Eve were the first man and woman. They were expelled from the garden of Eden. Adam had sex with Eve, and Eve had a child. We could remark that intuitively, in Eden, Adam knew Eve in the biblical sense. On the basis of this and other cases, we could suggest necessary and sufficient conditions for knowledge in the biblical sense (henceforth knowledgeB): Intercourse. S knewB Z iff S had sex with Z. We could then wonder if knowledgeB was symmetrical. We could remark on the episodic nature of knowledgeB, and contrast it with the static nature of other kinds of knowledge. We could refine Intercourse in response to certain objections raised by Reviewer 2, clarifying that not every form of sex is sufficient for knowledgeB. We could consider looking at the issue from a feminist perspective. Tempting as it is, this isn’t the response I suggest. While knowledgeB is clearly an important relation, and its study raises many interesting issues, I don’t think it’s a useful category for the theory of knowledge. Admittedly, it is not a complete coincidence that we use the term ‘knowledge’ in these contexts: one possible explanation is that knowledgeB of someone plausibly produces factual (and practical) knowledge of the more familiar kind. Nonetheless, it doesn’t seem that the relation of knowledgeB itself is a sui generis type of knowledge in the sense that matters for epistemology. I claim that objectual knowledge is similar to knowledge in the biblical sense at least in the following respect: there are various forms of objectual knowledge that pick out important relations, and standing in these relations to things can well lead to the production of knowledge. But none of these relations themselves, nor all of them together, form a sui generis type of knowledge in the sense that matters for epistemology. This sense will of course need to be explained.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

7. Knowing People One of the central cases of objectual knowledge is knowledge of persons; all discussions of this topic that I have seen include an example of knowing a person. (Witness the quote in the Section 1, and the quotes in Section 9). What are the conditions for knowing someone? As a first approximation, we know people whom we have met (possibly a few times). Meeting someone involves mutual recognition of subjects as particular individuals. A similar thought is expressed by the Oxford English Dictionary which gives the primary meaning of ‘Know [with object]’ as follows: ‘Have developed a relationship with (someone) through meeting and spending time with them; be familiar or friendly with’ (Stevenson, 2010, 975). Philosophers will want to unpack these conditions further. Matthew Benton offers a plausible and carefully argued analysis of knowing people. Benton argues that the following is a necessary condition for ‘interpersonal knowledge’ (knowledgei): Encounter: S knowsi R only if (i) S has had reciprocal causal contact with R, in which (ii) S treats R second personally, and (iii) R treats S second-personally. (Benton, 2017, 822)

Elaborating on this, Benton writes: A treatment by a subject S toward its recipient R is second-personal in virtue of S treating R as a subject (a ‘you‘), where S offers some of S’s own thoughts, words, or emotions to R, and S is, or for the most part intends to be, attentive to R’s thoughts, words, or emotions. (Benton, 2017, 822)

This analysis accords with the Oxford English Dictionary definition. The condition of reciprocal causal contact is more flexible than the condition of meeting and spending time with someone; it allows for less direct forms of contact. Benton’s analysis is also much more informative on the kind of relationship that develops between people who know each other, by appealing to the idea of second-person treatment. On this conception, knowing persons is a symmetrical relationship: we know people who know us. It may be objected that we regularly talk about knowing people even if the knowledge is one-sided. Someone could claim ‘I know that woman, she lives next door’ after spotting her on a photograph, even if they haven’t met. As a turn of phrase, a political commentator could say ‘I know the president, she would never consent to this’ even if they had never met. In these cases, the knowledge in question plausibly consists in knowing relevant information about a person. I don’t want to deny that on some occasions, this is how we talk. But I do think that the central and primary meaning of knowing someone is the one that implies personal interaction. In both of the cases just mentioned, it would make sense to ask: ‘I didn’t realise you know her; when did you meet?’, and expect them to back down by clarifying that they don’t know her personally, just know who she is, or what she is like. Knowing someone produces propositional knowledge of the known person, and without some such knowledge gained, we would be reluctant to say that the subject knows the other (Crane, 2012). But while knowing someone involves propositional knowledge, it has been argued that it is not reducible to propositional knowledge.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

 



This has been stated briefly for example by Ernest Sosa. The reason he gives for this irreducibility is that ‘(k)nowing someone or something, knowing some “object’ in the broadest sense of this term, seems at least sometimes to require having had some special causal interaction with that “object” ’ (Bonjour and Sosa, 2003, 100). Benton expresses a similar idea: ‘I . . . locate the irreducibility of interpersonal knowing in the two-way interactions characteristic of two subjects treating each other as subjects’ (Benton, 2017, 823). The idea needs elaboration. In a sense, some types of factual knowledge ‘require a special causal interaction’ with an object. For example, it is a necessary condition for having perceptual knowledge that a subject perceives, or has perceived an object; otherwise the knowledge wouldn’t be perceptual (that is, produced or justified by perception). Perception, in turn, involves a certain causal interaction between an object and a subject. Seeing something involves an object visually stimulating the perceptual system of a perceiving subject, and visual stimulation is a causal process. That such a causal process is required by, or involved in, perceiving an object, can (and should) be admitted even by those who think that perception cannot be reductively analysed in causal terms. One could argue that the knowledge relation itself doesn’t involve causation—it’s only its production or justification that does. But since a certain type of production or justification is essential for this type of knowledge, there will be still some sense in which knowledge ‘requires’ a causal interaction with its object. Another interesting comparison is with testimonial knowledge, which requires a certain kind of interaction with another subject. Perceptual knowledge and testimonial knowledge thus arguably ‘require’ a specific kind of causal interaction with an object or with another subject, yet they are cases of propositional knowledge. Hence someone who wanted to argue for a reduction could push back in response to Sosa’s and Benton’s observations, by suggesting that knowing a person is simply propositional knowledge that is produced through a certain type of causal interaction, or justified in virtue of standing in a causal relation to an object. In other words, it is structurally similar to knowledge by acquaintance, as discussed in Section 6 (on a certain notion of acquaintance, it would not be merely similar to, but also an actual case of, knowledge by acquaintance). I do not think this is quite right, but I won’t try to present a conclusive argument against such a reduction. Instead, I will split my argument into two strands. My overall claim is that there is no sui generis objectual knowledge which is both distinct from factual knowledge and also recognizably a member of the same genus ‘knowledge’. If you think that interpersonal knowledge is reducible to propositional knowledge, you’ll agree with the conclusion. If you think that interpersonal knowledge is not reducible to propositional knowledge, I will try to show that the aspect that makes it irreducible also places it outside the genus of knowledge. Either way, the overall conclusion is supported. So why would someone think—plausibly, in my view—that the requirement of a specific causal interaction makes interpersonal knowledge irreducible to factual knowledge (rather than just qualify its production)? The beginning of the answer is that the interactions are important because they produce and sustain not only propositional knowledge in each of the participants, but also an interpersonal relationship between them (as suggested by the Oxford Dictionary entry). Other

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

examples of interpersonal relationships that are similarly produced include friendship, being lovers, being accomplices, or allies in a shared project.⁴ Indeed interpersonal knowledge seems to be the precondition of all other similar interpersonal relationships. I’d like to suggest that we are reluctant to regard this basic interpersonal knowledge as reducible to propositional knowledge because it has a non-cognitive aspect. Let us approach this first by reflecting on friendship. Pretty much all accounts agree that friendship requires that friends have the attitude of caring towards each other. This is an attitude that lies beyond the realm of cognitive achievements: it’s a social or emotional attitude towards another. I propose that a less demanding but equally noncognitive attitude is constitutive of interpersonal knowing: to use Benton’s terms, it is the attitude of considering another subject as a subject. This is not simply representing the other as a subject of certain kinds of mental states: it is an attitude that takes in or recognizes the other, and is the basis of developing further interpersonal attitudes like care, trust, or love. Imagine a subject who has highly developed cognitive capacities, but is very deficient in social and emotional aspects. I don’t think this person can consider someone as a subject, and hence get to know another the way we do. If you disagree with this, if purely cognitive systems can know each other, then I return to the first strand of this argument: I challenge you to point at something that these systems have that is not propositional knowledge (produced in a certain way) but constitutes their knowing each other. I started the chapter by noting that the paradigm of knowledge in our philosophical tradition is factual knowledge. This determines the genus of knowledge as some kind of cognitive achievement, and if what’s been said above is right, given that interpersonal knowledge has an essential non-cognitive aspect, this places it outside the genus of knowledge so understood. Encounter will turn out to be similar to Intercourse, insofar as it specifies a relation that is not reducible to propositional knowledge, but lies outside the central concerns of epistemology. At the same time, it is a relationship that tends to produce propositional (and possibly practical) knowledge in the relata, and this explains its connection to knowledge in the paradigmatic sense.

8. Knowing Places, Facts, Rights, Experiences In addition to knowing persons, knowing places is also frequently mentioned as an example of knowing things. Knowing places seems to share with knowing persons the feature that a special kind of causal interaction is required with the object. In order to know Budapest, one had to be there, for an extended time or several times, and plausibly, one also has to be able to get around at least to some extent. In this case, it is less clear than in the case of interpersonal knowledge that anything beyond factual and practical knowledge is required. As proposed before, in this case it ⁴ Benton also notes that ‘Through personal encounters, people can progress from being mere acquaintances to being friends, to close friends, to intimates or lovers’ (2017, 823).

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

 



is an additional condition that the factual knowledge has to be produced in a certain way, namely, through repeated personal interactions. Objectual knowledge attributions can take all sorts of objects, and in the case of some kind of objects, the knowledge is straightforwardly propositional, and there is no restriction on the production of knowledge. The most obvious case is knowledge of facts, or knowledge of the truth (I have in mind attributions which would be translated by ‘connaître’ and ‘kennen’ and so on). Knowing a fact is factual knowledge. The same holds for some other kinds of objects; for example, knowing your rights, or knowing the rules, or knowing a theory, all plausibly consist in knowing some propositions, however that knowledge is produced. These are clearly quite different relations from the relation of interpersonal knowledge, and also different from knowing a place (which has to have a specific mode of production). Yet there is no obvious argument to exclude them from the instances of objectual knowledge. This suggests that the relation of ‘knowing an object’ is different when it comes to different types of objects. Next I will consider yet another type of object where the relation is again arguably different. A frequently mentioned example of knowing things is knowledge of experiences. Sosa mentions knowing ‘the experience of a cold shower after a hard run’ (Bonjour and Sosa, 2003, 100), and the two philosophers considered in Section 9 talk about knowing ‘the agony of defeat’ (Conee, 1994, 14) and ‘the joy of victory’ (Tye, 2009, 96). To know an experience requires that one has had the experience. Unlike in the case of knowing people and knowing places, there is no requirement of repeated interactions. It is commonly assumed that once a person feels the agony of defeat, she will be at least in a good position to know the agony of defeat. However, there is something special in this case: namely, it is often claimed that the relation we have to features of experience when we have the experience has some features that deserve a separate discussion. This point leads us to the consideration of the philosophical notion of ‘acquaintance’. Most contemporary discussions trace back the notion to Russell, and as the definition in Section 3 suggests, it is commonly assumed to be a direct (noninferential) form of awareness of an object. It is also often assumed that one can have this relation only to specific type of entities, primarily of a mental nature— mental facts or sense-data or features of conscious experiences. Russell thought that in addition to sense-data, one can also be acquainted with universals, and possibly with the self, but he categorically denied that we can be acquainted with physical objects or with persons as far as they appear to us physically. So knowing one’s best friend is certainly not a relation of acquaintance on Russell’s definition. In the contemporary discussion, it is again frequently assumed that acquaintance is a special relationship we have to our conscious experiences; witness the following definition: ‘Acquaintance is a unique epistemological relation that relates a person to her own phenomenally conscious states and processes directly, incorrigibly, and in a way that seems to reveal their essence’. (Balog, 2012)

Balog makes some assumptions which may not be accepted by everyone: that acquaintance is incorrigible and reveals essences. But she shares the common view that the class of objects with which we can be acquainted is limited.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

On most accounts, acquaintance with mental items has an important role in producing knowledge about those items. But as we remarked earlier when discussing Russell’s notion, this knowledge can be understood as propositional knowledge that is based on acquaintance, rather than objectual knowledge which is directed at an object rather than at a truth. Here we are particularly interested in the suggestion that acquaintance in itself constitutes (rather than merely produces or justifies) objectual knowledge. The suggestion is that if one bears the relation of acquaintance to an object, then this is the same as knowing that object in the specific object-directed sense. I am going to consider two proposals along these lines in Section 9.

9. Acquaintance as Knowledge of Things Earl Conee and Michael Tye both suggest that acquaintance itself constitutes objectual knowledge. The context of both Conee’s and Tye’s discussion is the evaluation of the so-called ‘Knowledge Argument’. This argument is found already in Russell, and was then formulated independently by Howard Robinson (1982) and Frank Jackson (1982). It concerns someone who has a certain type of conscious experience for the first time in their life. Tye and Conee both argue that the new knowledge gained on such an occasion—often characterized as knowing what it’s like to have an experience—is non-propositional, objectual knowledge of (features of) the experience. In support of this position, they both argue that there is a general category of sui generis objectual knowledge. Conee claims that ‘(a)cquaintance constitutes a third category of knowledge, irreducible to factual knowledge or knowing how’ (Conee, 1994, 136) and Tye holds that ‘just as there is a sense of “see” such that one sees an object if and only if it looks some way to one, so there is a sense of “know” such one knows a thing if and only if one is acquainted with that thing’ (Tye, 2009, 96). Note that this position requires that instances of knowing different types of objects are all constituted by the same relation of acquaintance, for otherwise we don’t have an account of a general sui generis objectual knowledge. Indeed, both Conee and Tye seem to accept this, as suggested by the following quotes: To come to know a property is to become acquainted with the property, just as to come to know a city is to become acquainted with the city, and to come to know a problem is to become acquainted with the problem. It is uncontroversial that some knowledge attributions ascribe a relation of acquaintance, as when we say things like ‘Sam knows Bill’, or ‘Bob knows the agony of defeat’. These considerations suggest a hypothesis about the examples in question to the effect that the difference between ignorance of what an experience is like and knowledge of what an experience is like is a matter of acquaintance. (Conee, 1994, 140) In ordinary English we talk of knowing things and knowing facts. I know Brian McLaughlin, for example. I know the city of Athens. I know the joy of victory and I also know the thrill of driving very fast. I know the feeling of anger. (Tye, 2009, 95–6) why should consciousness of something, direct or indirect, yield knowledge of that thing? . . . it is simply incoherent to suppose that one might be genuinely (noninferentially) conscious of an entity and yet not know it at all. In being conscious of a particular shade of red at a particular moment, say, I know that shade of red. How could I not? I know it just by being conscious of it. I may not know that shade of red a few moments later, after turning away; I may not know any

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

 



truths about that shade of red; but, as I view the shade, know it I do in some ordinary, basic sense of the term ‘know’. (Tye, 2009, 99)

The line of reasoning presented here seems to go as follows. First, let us take typical ordinary examples of knowing things, such as knowing people or places. (Both Conee and Tye have examples of both kinds; Conee talks of knowing Cambridge as an instance of being acquainted with Cambridge). Second, it is uncontroversial, the argument proceeds, that there is knowledge in this sense; after all, we truly say that Sam knows Bill, or Cambridge, that Tye knows Brian Mclaughlin and Athens. Third, consider the relation of acquaintance we have to features of our conscious experiences: this is the same kind of relation as the relation between a subject and a person or place known by her. Fourth, just as the latter is uncontroversially an instance of knowledge, so is the former—the argument concludes. There are two problems with this argument. One is with the third step: in fact, the relation of acquaintance we have to conscious features is very different from the relation of knowing people in the ordinary sense. It is a controversial matter whether or not we can be ‘acquainted’ with people in the philosophical sense—that is, whether or not we can be directly aware of a person in the sense we are directly aware of features of experiences. Russell certainly thought that we cannot be. On a different view, seeing something may count as an instance of acquaintance, and since we can see people, we can be acquainted with them. But even if this is right, acquaintance will still be different from interpersonal knowing. I see the person sitting at the next table in the café as I type these lines, but I don’t know her. Acquaintance is also going to be different from other relations of knowing things. It is different from the relation we have to our rights when we know them. Suppose seeing is a type of acquaintance—seeing a place is not the same as knowing a place. I have seen a number of cities from the train on a journey two weeks ago, but I don’t know these cities. So unless something else is said on the issue, we cannot assume that the same relation of acquaintance constitutes all the relations we bear to things when we are said to know them. There is another problem with the argument, and it’s in the second step. If what is said in Section 7 is right, then interpersonal knowledge does not belong to the genus ‘knowledge’ whose nature is determined by factual knowledge as its central case. So we cannot use interpersonal knowing as one of the uncontroversial examples of, and model for, objectual knowledge, if our purpose is to give an account of knowledge that is the subject matter of epistemology. Initially, knowing persons may seem like a very plausible example of ‘a third category of knowledge, irreducible to factual knowledge or knowing how’, as Conee puts it. But I argued above that this picture is not right. Just what makes interpersonal knowing ‘irreducible to factual knowledge or knowing how’, will also plausibly prevent it from being a ‘third category of knowledge’, if ‘knowledge’ is used univocally.

10. Acquaintance as Objectual Knowledge, Second Try In Section 9, I considered an attempt to use the philosophical notion of acquaintance, understood as direct awareness of an object, to give an account of objectual

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

knowledge. The idea was that acquaintance constitutes the ordinary relation of knowing things, and since these uncontroversially include cases of knowledge, acquaintance itself constitutes knowledge. I had two complaints about this idea: first, acquaintance in the philosophical sense is quite different for example from the relation of knowing persons; second, ‘connaître’ may not be ‘knowledge’ in the same sense as ‘savoir’ is knowledge. But now note that the two problems could actually cancel out each other and hence open a new avenue for a theory of objectual knowledge making use of the notion of acquaintance. We could admit that acquaintance is quite different from the relation of interpersonal knowledge; but since interpersonal knowledge is not ‘knowledge’ in the epistemic sense anyway, this isn’t a big loss; in fact, it is an advantage. Next we could suggest that direct awareness of an object is relevantly similar to the central cases of knowledge, in being a non-accidental cognitive achievement. Objectual knowledge will then belong to the genus of knowledge, we just have to keep in mind that ordinary attributions of know-NP, or uses of ‘connaître’ and ‘kennen’, most of the time do not ascribe ‘objectual knowledge’ in the philosophical sense. Depending on the notion of acquaintance that is proposed to constitute objectual knowledge, the range of things that can be known in this sense will be more or less restricted. If we can be acquainted only with features of conscious experiences, then only those can be objectually known. If perception also provides acquaintance with the perceived objects, then we can objectually know everything we can perceive, but presumably not rights or rules. And objectually knowing for example a person will not be the same as knowing her. This, I believe, is the most promising avenue for developing an account of a sui generis objectual knowledge. But the account needs very significant further support. First, we would need an argument to show that knowledge constituted by acquaintance is needed in addition to (factual) knowledge based on acquaintance (i.e. justified or produced by acquaintance). Many phenomena about acquaintance can be arguably explained simply by appealing to the idea of knowledge based on acquaintance. For example, it is often claimed that that we can’t know what it’s like to have a certain experience without having had the experience. One might think that this is explained by the fact that experiencing (which is the episode when one is acquainted with a phenomenal quality) is the same as what-it’s-like knowledge. But that’s not the only possible explanation. Suppose that ‘what-it’s-like’ knowledge is always based on acquaintance; and the only way to become acquainted with a phenomenal quality is to experience it. This theory would also explain the claim. Second, even if acquaintance can be understood as a non-accidental cognitive achievement, this may not be sufficient for knowledge, even if it is necessary. For an example that is quite different from the case of acquaintance: perhaps certain products in the sub-personal processing of our perceptual system can be regarded as non-accidental cognitive achievements, yet at least on certain notions of knowledge, they do not qualify as knowledge. Or recall Wilfrid Sellars’s influential attack on the Myth of the Given—an argument to the effect that anything that ‘enters the space of reasons’ has to have a propositional format. These discussions require significant further commitments about the nature of knowledge, and it’s possible that all things considered, there will be a proper case for

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

 



a separate category of objectual knowledge. In this chapter, I tried to show that there is no easy route to establish the existence of a third kind of knowledge. The linguistic format of knowledge ascriptions has little significance; ordinary ascriptions of knowing things indicate a number of different relations to things, including, in a large part of the cases, the relation of having propositional knowledge about the object in question. Furthermore, we cannot assume without further ado that the relations expressed by ‘knowing things’ are ‘knowledge’ in the sense that concerns epistemology.⁵

References Balog, Katalin (2012) ‘Acquaintance and the Mind-Body Problem’, in Simone Gozzano and Christopher S. Hill (eds), New Perspectives on Type Identity: The Mental and the Physical, Cambridge: Cambridge University Press, 16–42. Benton, Matthew (2017) ‘Epistemology Personalised’, Philosophical Quarterly 67 (269): 813–34. BonJour, Laurence, and Sosa, Ernest (2003) Epistemic Justification: Internalism vs. Externalism, Foundations vs. Virtues, Oxford: Wiley-Blackwell. Brogaard, Berit (2009) ‘What Mary Did Yesterday: Reflections on Knowledge-wh’, Philosophy and Phenomenological Research 78: 439–67. Conee, Earl (1994) ‘Phenomenal Knowledge’, Australasian Journal of Philosophy 72(2): 136–50. Craig, Edward (1990) Knowledge and the State of Nature: An Essay in Conceptual Synthesis, Oxford: Clarendon Press. Crane, Tim (2012) ‘Tye on Acquaintance and the Problem of Consciousness’, Philosophy and Phenomenological Research 84(1): 190–8. DePoe, John M. (2013) ‘Knowledge by Acquaintance and Knowledge by Description’, Internet Encyclopedia of Philosophy. Accessed 29 April 2018, http://www.iep.utm.edu/knowacq/. Fantl, Jeremy (2016) ‘Knowledge How’, in Edward N. Zalta (ed.), Stanford Encyclopedia of Philosophy (Spring), . Farkas, Katalin (2016) ‘Know-Wh Does Not Reduce to Know That’, American Philosophical Quarterly 53(2): 109–22. Farkas, Katalin (2017) ‘Practical Know-Wh’, Noûs 51: 855–70. Farkas, Katalin (2018) ‘Know-How and Non-Propositional Intentionality’, in Alex Grzankowski and Michelle Montague (eds), Non-Propositional Intentionality, Oxford: Oxford University Press, 95–113. Hasan, Ali and Fumerton, Richard (2014) ‘Knowledge by Acquaintance vs. Description’, in Edward N. Zalta (ed.), Stanford Encyclopedia of Philosophy (Spring), . Jackson, Frank (1982) ‘Epiphenomenal Qualia’, Philosophical Quarterly 32(April): 127–36. Martens, David B. (2010) ‘Knowledge by Acquaintance/by Description’, in Jonathan Dancy, Ernest Sosa and Matthias Steup (eds), A Companion to Epistemology, 2nd edn, Oxford: Blackwell, 479–82. ⁵ Research on this chapter was supported by project K-112542 of the Hungarian Scientific Research Fund, and I am grateful for the support received. Versions of the chapter were given in the New Directions in the Study of the Mind project funded by the Templeton Foundation; at MIT; and at the University of Birmingham. I benefited greatly from the discussion after these talks. I’d like to thank Lukas Lewerentz for advice.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

Robinson, Howard M. (1982) Matter and Sense: A Critique of Contemporary Materialism, Cambridge: Cambridge University Press. Russell, Bertrand (1910) ‘Knowledge by Acquaintance and Knowledge by Description’, Proceedings of the Aristotelian Society 11(5): 108–28. Russell, Bertrand (1912) The Problems of Philosophy, London: Williams and Norgate. Ryle, Gilbert (1949) The Concept of Mind, London: Hutchinson & Co. Stanley, Jason and Williamson, Timothy (2001) ‘Knowing How’, Journal of Philosophy 98: 411–44. Stevenson, Angus (ed.) (2010) Oxford Dictionary of English, 3rd edn, Oxford: Oxford University Press. Tye, Michael (2009) Consciousness Revisited: Materialism without Phenomenal Concepts, Cambridge, MA: MIT Press. Zagzebski, Linda (1999) ‘What Is Knowledge?’, in John Greco and Ernest Sosa (eds), The Blackwell Guide to Epistemology, Oxford: Blackwell, 92–116.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

13 Visual Experience, Revelation, and the Three Rs Bill Brewer

1. Introduction Normal vision involves consciousness; and we often characterize the specific modification of consciousness involved in a particular case of visual experience by appeal to the way things look to the subject of that experience. My concern here is with the way in which such characterizations work. In my own case now, it looks to me as though there is a sheet of paper off to my left with many differently coloured lines on it forming a flattened loop with outwards radials that converge in a cross at the centre—it’s a London tube map centred on Oxford Circus. How exactly does this description serve to convey the character of my current conscious condition; and what might this tell us about the nature of visual consciousness itself ?¹ I begin in Section 2 with some important ideas about how we do and must specify the conscious character of vision. Section 3 aims to elucidate a further condition on the nature of visual experience, which I follow others in calling Revelation (e.g. Johnston, 1992; Byrne and Hilbert, 2007), by contrast with the humility that Lewis argues governs our quite different mode of access to the imperceptible fundamental microscopic nature of the world around us (Lewis, 2009).² Sections 4 and 5 exploit Revelation in a series of objections to the first two of my three Rs: what I call respectively the Resemblance and Representational accounts of visual experience. I sketch the principles of my third R, the Relational account that I myself favour, in Section 6 and conclude in Section 7.

¹ The argument that I develop here owes a great deal to John Campbell. It is especially influenced by his recent contributions to Campbell and Cassam (2014), extending earlier work in Campbell (1993, 2002). ² Although I adopt the name ‘Revelation’, the condition that I propose here is far less demanding than others intend by it. See note 3. Bill Brewer, Visual Experience, Revelation, and the Three Rs In: Acquaintance: New Essays. Edited by: Jonathan Knowles and Thomas Raleigh, Oxford University Press (2019). © Bill Brewer. DOI: 10.1093/oso/9780198803461.003.0014

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

2. The Conscious Character of Vision In ‘The Refutation of Idealism’, Moore famously writes as follows. The moment we try to fix our attention upon consciousness and to see what, distinctly, it is, it seems to vanish: it seems as if we had before us a mere emptiness. When we try to introspect the sensation of blue, all we can see is the blue: the other element is as if it were diaphanous. (Moore, 1903, 25)

The passage has prompted a great deal of commentary, concerning what Moore may mean by it, what truths it may convey, and what the implications may be of such truths. I begin my own discussion here with an assumption that is, I think, one plausible interpretation of Moore’s idea, that one can only discern the conscious character of one’s visual experience by attending to the worldly phenomena that one is apparently visually aware of. Strawson makes a similar claim in discussing Ayer’s views on perception as follows. Ayer says that we take a step beyond our sensible experience in making our perceptual judgements. I say rather that we take a step back (in general) from our perceptual judgements in framing accounts of our sensible experience; for we have (in general) to include reference to the former in framing a veridical description of the latter. (Strawson, 1979, 45–6)

I began by claiming that we often characterize the specific modification of consciousness involved in a particular case of visual perception by appeal to the way things look to the subject of that perception. The suggestion here is that this appeal to the way things look is a matter of characterizing a way the world might be and derivatively specifying the conscious character of perception by remarking that that is the way the world looks to be around the perceiver in the particular case of perception in question. Moore and Strawson, as I read them, claim that this is not only natural and normal, but also essential. I agree with them about this and proceed from here on the assumption that they are right. At the heart of this assumption is the idea of a world-looks order of explanation. Concepts that have their primary application to mind-independent objects in the world around us are secondarily applied in the characterization of our conscious visual experience of it; and this is essential if we are to give a faithful and accurate account of the conscious character of such experience. The canonical characterization of visual experience is thus of the form: it looks to me as though there are objects of such and such kinds with such and such properties arranged around me thus and so. In order to have a specific case to consider in what follows I take this to be a simplified instance of the general form: it looks to me as though there is something round and red before me. My aim here is to explore how this crucial role for looks talk in the specification of the conscious character of visual experience may constrain the correct account of the nature of such experience itself. On the face of it there is precious little to go on in such an enterprise. The essential characterization of visual experience by the way things look to be in the mind-independent world appears to be consistent with what I will call Resemblance, Representational, and Relational accounts. According to the Resemblance account, the conscious character of visual experience consists in the subject’s acquaintance with mind-dependent phenomena; and it

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

 , ,    



is correct to specify particular cases of such acquaintance by the way things thereby look to be in the mind-independent world because such mind-dependent phenomena are caused by and at least in certain basic respects resemble worldly things. According to the Representational account, the conscious character of visual experience consists in the subject’s visually representing things as being thus and so in the mind-independent world around her, any particular case of which may thereby be specified as things looking to be precisely that way out there. According to the Relational account, the conscious character of the most basic visual experience consists in the subject’s acquaintance with mind-independent worldly things themselves; and it is correct to specify particular cases of such acquaintance by the way things thereby look to be in the mind-independent world because these are some of the ways the mind-independent objects of her acquaintance are in the cases in question. I hope to make progress beyond this comfortable consistency by introducing a further constraint. To a first approximation, this is the idea that at least in the most basic cases, a visual experience whose conscious character is specified by its looking to the subject as though there is an F before her is revelatory of what being F is. The argument that I offer tentatively here is that, in the presence of this additional constraint, the Relational account of visual experience is preferable to Resemblance and Representational accounts.

3. Revelation and Humility To a second approximation, Revelation is the thesis that visual experience whose conscious character is specified by its looking to the subject as though there is an F before her is a source of knowledge of what being F is, knowledge expressible in this context, and with appropriate attentional direction, by ‘being F is being like that’.³ So an experience in which it looks to me as though there is something round and red before me is a source of knowledge of what being round and being red are: being round is being like that (attending to perceived shape); and being red is being like that (attending to perceived colour). Johnston (1992) and Chalmers (2006) are sceptical of Revelation. They argue, very roughly, that vision science has established that, insofar as the properties that worldly objects look to have are those that such things normally have when they look to do so and that explain our experiences in which they look to do so, these are absolutely not the properties apparently revealed by those very visual experiences.⁴ I reject their argument. For scientific explanations of visual processes do not compete in a single explanatory space with commonsense explanations of the nature of visual experience by the properties of mind-independent objects that such experiences reveal. So the correctness of the ³ See Johnston (1992) for a sceptical discussion of Revelation in connection with colour perception. I adopt the term ‘Revelation’ from him and elucidate my own use of it below. Byrne and Hilbert (2007) share Johnston’s scepticism. Chalmers (2006) contains an extended and equally sceptical discussion of the same idea far more generally, in terms of what he regards as a myth of ‘Edenic’ perception. See Campbell (2005) for articulation and defence of what he takes to be the fundamental insight that he calls ‘Transparency’ and contrasts with Byrne and Hilbert’s reading of ‘Revelation’. My own understanding of the basic idea is certainly closer to Campbell’s. ⁴ Johnston’s concern is only with the colours; Chalmers’s scepticism is quite general.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

former have no eliminatory force against the latter. I make this objection in detail elsewhere (2011a, 2011b; see also Campbell, 1993, 2005; and Stroud, 2000).⁵ My project here is to explore the consequences of Revelation for our understanding of the nature of visual consciousness itself. The second approximation of Revelation could still benefit from further clarification. I propose elucidation via its contrary, Humility: the thesis that we are irremediably ignorant of the intrinsic nature of mind-independent things.⁶ According to Lewis’s Ramseyan argument for Humility (2009), the ‘final theory’ to which scientific research ideally tends ought to deliver a complete inventory of the fundamental intrinsic properties that play an active role in the actual workings of nature.⁷ Call the true and complete such final theory T. This contains a good deal of our old, O-language, which is available and interpreted independently of T, and which suffices to express all possible observations. T also contains its own theoretical T-terms. These are implicitly defined by their role in the overall theory and name the fundamental properties in question. Furthermore, Lewis assumes that none of these causally basic intrinsic properties are named in O-language, ‘except as occupants of roles; in which case T will name them over again, and will say that the property named by so-and-so T-term is the occupant of such-and-such role’ (Lewis, 2009, 206). Suppose that T(t₁ . . . tn) is the simplest form of T, where t₁, . . . tn are the T-terms, thereby implicitly defined in terms of the O-language that constitutes the remainder of this expression for T. The Ramsey sentence of T is 9x₁ . . . 9xnT(x₁ xn). This logically implies all and only the O-language sentences that are theorems of T. Call this Ramsey sentence R. Since O-language alone suffices to express all possible observations, every possible observable prediction of T is equally a prediction of R. Thus any evidence for T is equally evidence for R: evidence for T cannot go beyond evidence merely for R. Now, it is extremely likely that, if there are any, then there will be more than one fundamental property in at least the most basic ontological categories: monadic properties, dyadic relations, and so on. That is to say, if there are any monadic fundamental properties, then there are very likely to be more than one. Similarly for dyadic relations, and so on. Suppose that is the n-tuple that actually realizes T; and suppose that is any n-tuple which results from permuting some of the pairs in which ai and aj are of the same ontological category. In other words, supposing that ap and aq are both monadic fundamental properties in the n-tuple that actually realizes T, let be the n-tuple . Combinatorialism is the thesis that possibility is preserved ⁵ Campbell (2005) also responds to further arguments from Byrne and Hilbert (2007) against Revelation in the case of the colours. ⁶ See Langton (1998) and Lewis (2009) for canonical presentations on behalf of Kant and Ramsey, respectively. ⁷ There are substantive and controversial issues concerning the correct precise characterization of intrinsic properties. For a helpful overview see Sosa (2001). It is unnecessary for my purposes in what follows to engage with these debates in detail. The provisional characterization of intrinsic properties as those that an object has of itself, independently of any other thing, those it would retain, or retain the lack of, if it were the only thing that existed, should suffice.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

 , ,    



under permutation or replacement of co-categorial items. So, on the assumption of combinatorialism, is a possible realization of T. Quidditism is the thesis that possibilities which differ simply by the permutation or replacement of properties are genuinely distinct. So, on the additional assumption of quidditism, is a distinct possible realization of T from the actual realization . Furthermore, since any evidence for T is evidence for R, and R is true in both the actual case, in which realizes T, and in the distinct possible case in which realizes T, then no possible evidence can tell us that is the actual realization of T, as opposed to .⁸ Though our theory T has a unique actual realization . . . it has multiple possible realizations . . . no possible observation can tell us which one is actual, because whichever one is actual the Ramsey sentence will be true. There is indeed a true contingent proposition about which of the possible realizations is actual, but we can never gain evidence for this proposition, and so can never know it . . . Humility follows. (Lewis, 2009, 207)

The Humility Thesis that follows, according to Lewis, is the thesis that we are irremediably ignorant of the fundamental properties of the world: we cannot possibly know the intrinsic nature of physical reality itself. Provided only that a fundamental property is not a categorial singleton—that is to say that there are others of the same category⁹—then we can never have any evidence that it—as opposed to any of these others—is the actual realizer of the theoretical role definitive of its name. We know that there is a property, so named, that does just that; but we cannot possibly know which it is, what the intrinsic nature of the property so named actually is. Since all intrinsic properties supervene upon these fundamental properties, we are in this sense irremediably ignorant of the intrinsic nature of mind-independent reality itself. The proponent of Revelation is motivated by the intuition that perception of the macroscopic world around us is unlike theorizing about its imperceptible fundamental microscopic constitution in precisely this respect. The evidence provided by our conscious visual experience that there is something F before us is absolutely not neutral on what being F is in such a way as to warrant the introduction of any alternative possibility, F’, in which it equally constitutes visual evidence of exactly the same standing that there is something F’ there instead. The character of visual consciousness itself constitutes a source of precisely the knowledge that Humility denies, of what being F actually is. Furthermore, Revelation explains the Moore-Strawson insight with which I began, that a faithful and accurate account of the conscious character of our visual experience makes essential use of concepts that have their primary application to mindindependent objects in the world around us of which this is our experience. For Revelation requires that visual experience in which it looks to the subject as though there is an F before her, which therefore constitutes her evidence that there is ⁸ Combinatorialism and quiditism are both substantive assumptions that may be questioned. I abstain from such questions here. ⁹ Later in the paper Lewis introduces additional assumptions that enable him to extend the argument to all fundamental properties; but it is unnecessary for my purposes to get involved with the additional complications.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

something F out there, is absolutely not neutral on what being F is. Its adequate characterization as the conscious condition that it is therefore makes essential use of that very concept, F, which has its primary application to mind-independent things in the world around her. This is precisely the requirement that I derive from Moore and Strawson above.

4. Against Resemblance The Resemblance account may appear at first sight well suited to respect Revelation in this sense; and this explains its persistence in one form or another in philosophical theories of visual perception. I begin the current section with an elucidation of the core idea and go on to argue that it is ultimately untenable. The starting point is the suggestion that visual experience has a conscious character that indicates the presence of various worldly properties by causal covariation. A determinable range of worldly properties, the colours or shapes, say, reliably cause in us visual experiences with a determinable range of conscious characters. Specific determinates of the latter thereby indicate the presence of the corresponding determinates of the former, making our visual experiences into natural signs of those properties of mind-independent objects (Ayers, 1991, ch. 7). The conscious character of visual experience that indicates the presence of an F in this way may therefore be specified in terms of its looking to the subject as though there is an F before her. As it stands, this picture is a paradigm of Humility. Visual experiences indicate the worldly properties that normally cause experiences with like conscious character, whatever those worldly properties may be. As yet, the experiences themselves display precisely the neutrality that drives Lewis’s argument and do nothing to reveal what these normal causes actually are. The Resemblance account as I understand it has a crucial further feature. The conscious character of visual experience resembles the worldly phenomena that it indicates in at least certain basic respects. The core idea of the Resemblance account is that the addition of such resemblance to the indication relation makes visual experience whose conscious character is specified by its looking to the subject as though there is an F before her non-neutral on what being F is in such a way as to block the argument for Humility and constitute a source of knowledge of what being F is. So, as I say, the Resemblance account may appear well suited to respect Revelation. The most natural reading of this idea that I take to be definitive of the Resemblance account as I understand it here construes resemblance as the sharing of basic properties at least between worldly objects and certain mind-dependent sensations that indicate the presence of those properties in experiences in which the subject is acquainted with the sensations in question. Specific such sensations thereby constitute the conscious character of experiences that are specified in terms of the worldly properties that they indicate as explained above. My experience in which it looks to me as though there is something round and red before me therefore involves my acquaintance with a mind-dependent sensation that is likewise round and red. It is therefore absolutely not intrinsically neutral as between indicating something round and red, on the one hand, and indicating something square and green, say, on the other. It is ‘made for’ the former in such a way as to block the argument for Humility

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

 , ,    



and to secure the Revelation in visual consciousness of what being round and being red are. For an instance of roundness and redness is right there before me, constituting the very conscious character of my experience of its looking to me as though there is something round and red out there in the world before me. My experience is therefore a source of knowledge that being round is being like that (attending to its shape) and that being red is being like that (attending to its colour). My objection is that this crucial assertion of resemblance between mindindependent worldly objects and the mind-dependent sensations that we are acquainted with in visual experience does not really make sense. Roundness and redness are properties of mind-independent, space-occupying material objects. This is the whole point of the initial insight from Moore and Strawson that we can only specify the conscious character of visual experience by embedding concepts that have their primary application directly to mind-independent objects in the world around us within the scope of the way things look to the subject in those experiences. The Resemblance account is that an experience in which it looks to me as though there is something round and red before me consists in my acquaintance with a minddependent sensation that is itself round and red in precisely this sense that applies to mind-independent objects themselves. Now the roundness and redness of mind-independent objects consists in a specific way in which they fill the space that they occupy—roundly—and a specific manner of colouration of their surfaces—redly. So it is entirely opaque to me how we are supposed to understand the idea that mind-independent sensations might be round and red in precisely this sense, how such things could possibly be round and red. Which space is it supposed to be in which mind-dependent sensations are roundly and redly extended? There is only one genuine space, namely this one, in which things may be round and red in the relevant sense; and its genuinely extended occupants are mind-independent objects and not mind-dependent sensations.¹⁰ There may perhaps be determinable dimensions along which mind-dependent sensations vary in their intrinsic nature; and we may metaphorically call a structure organized around such dimensions a space. We may further identify a given such dimension with the mind-independent determinable shape insofar as various sensations’ position on it thereby indicate the shapes of worldly objects in vision. But this does not suffice for any true resemblance between such sensations and the objects whose shapes they indicate. The literal sharing of properties constitutive of any genuine resemblance relation is inconsistent with the postulation of mind-dependent and mind-independent relata. To paraphrase Berkeley, a material object can be like nothing but a material object (1975, §8). Retreating to a pure indication account may ¹⁰ It might be replied on behalf of the Resemblance account that this one physical space contains both mind-independent material objects and the mind-dependent sensations that are the direct objects of our visual experience and whose modes of spatial extension resemble some of those of material objects indirectly seen as proposed (see e.g. Jackson, 1977). I find the basic idea that there may be determinately extended yet mind-dependent occupants of physical space outside the subject impossible to comprehend. Furthermore, once we acknowledge that the direct objects of visual experience may be spatially extended entities in physical space, then I have argued elsewhere that phenomena of illusion and even hallucination do nothing to persuade us that these need be anything other than the mind-independent material objects themselves that we all know and love (Brewer, 2011a). This Relational account is taken up in Section 5.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

avoid inconsistency but immediately invites Lewis’s argument for Humility. Repeated insistence on resemblance in addition to indication expresses a genuine insight that conscious visual experience avoids Humility and satisfies Revelation; but this is simply unavailable on the present approach. I conclude that the Resemblance account as articulated here fails to secure Revelation and so fails by the lights of the current investigation. The Relational account that I endorse in Section 5 offers an alternative conception of something that might be described as resemblance, and may even be lying behind various historical appeals to resemblance in this area—notably Locke’s discussion of primary qualities (1975, II.viii)—but this is a quite different view from the Resemblance account of the present section. The next proposal that I consider in Section 4 is that the key to securing Revelation lies in representation rather than resemblance.

5. Against Representation The core idea of the Representational account is that the conscious character of a visual experience in which it looks to the subject as though there is an F before her is correctly so specified in virtue of its representational content: roughly, that there is a suitably placed F out there.¹¹ This clearly accommodates the Moore-Strawson insight with which I began, that the conscious character of visual experience has to be specified in terms of the way the mind-independent world looks to be around the subject. In the case of my toy example here, it looks to me as though there is something round and red before me. According to the Representational account, this is a matter of my having a visual experience whose conscious character is given by the representational content that there is something round and red there. The Representational account also seems well placed to secure Revelation. For its characterization of visual consciousness in terms of the specific way the world around the subject is represented—as containing something round and red at p, say—is surely non-neutral in precisely the way required to block the argument for Humility concerning what it is for the world to be the way that such experience provides evidence that it is—namely as containing something round and red at p. I argue in the current section that this appearance of non-neutrality is not sufficient to secure Revelation, though. So the representational account must also be rejected. According to the Representational account, my having a visual experience with the conscious character of its looking to me as though there is something round and red before me is a matter of my being in a state that represents the presence of something round and red out there. Being in this state therefore places a determinate condition on the world if its content is to be correct, that it contains something round and red there. The question of Revelation concerns what being in this state offers me, though, in particular, whether it constitutes a source of knowledge of what being round and ¹¹ Representative proponents of the Representational account include Armstrong (1968, ch. 10), Dretske (1981, ch. 6), Searle (1983, ch. 2), Burge (1986, 2010), Peacocke (1989, 1992, ch. 3), McDowell (1986, 1994), Harman (1990), Tye (1992, 1995, 2000), Byrne (2001), Chalmers (2004), Pautz (2010), and Siegel (2010). There is of course significant controversy amongst these philosophers over the nature and range of representational contents involved in conscious vision.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

 , ,    



being red are; and this depends on how this determinate content is conveyed to me by the experience in question. My argument against the Representational account takes the form of a dilemma. Define a state with conceptual content to be a state with representational content such that the subject must actively deploy the concepts used in the canonical statement of its content in order to be in the state in question.¹² According to the Representational account, visual experiences are representational states. Either these have conceptual content or they do not. Suppose, first, that the Representational account proposes that visual experiences are states with conceptual content. Thus, I must actively deploy the concepts ‘round’ and ‘red’ in order to have a visual experience in which it looks to me as though there is something round and red before me. This proposal is surely set fair to secure Revelation. For the content of my experience is conveyed to me by my deployment of the concepts ‘round’ and ‘red’ used in its specification, which in turn requires my knowledge of what being round and being red are.¹³ So it follows from the fact that I have a visual experience whose conscious character is specified by its looking to me as though there is something round and red before me that I know what being round and being red are. This is not sufficient for my experience to be a source of such knowledge as Revelation claims, though. We may acquire knowledge of what being round and being red are otherwise—by definition or description, perhaps, or by some other means. But Revelation claims that this knowledge may be provided by a visual experience whose conscious character is specified by its looking to the subject as though there is something round and red before her as its source. That is to say, a subject of such an experience who previously lacked this knowledge may thereby come to know what being round and being red are. This is the sense in which her experience is supposed to be non-neutral with respect to what the roundness and redness of mind-independent objects are. The subject of a representational state with conceptual content that there is something round and red before her necessarily deploys the concepts ‘round’ and ‘red’ in that very state. She therefore either previously knew what being round and being red are or is the subject of a distinct but concurrent state that is her source of such knowledge. It follows that her representational state in itself is not a source of the knowledge of what being round and being red are. According to the Representational account, though, Revelation is supposed to be secured by the representational content of conscious visual experience. The current conceptual version of the Representational account therefore fails. These difficulties with the conceptualist Representational account may be illustrated by considering perceptual demonstrative representation. Suppose that a ¹² Various debates have been characterized as that between conceptualists and non-conceptualists concerning perceptual content. I intend my characterization in the text to be stipulative of what I mean by this here. For a sample of these debates, see Evans (1982, esp. ch. 6), Cussins (1990), Peacocke (1992, ch. 3, 2001), Crane (1992), McDowell (1994), Brewer (1999, ch. 5, 2005), Heck (2000), Kelly (2001), and Byrne (2005). ¹³ I assume throughout that active deployment of the concept ‘F ’ depends upon the subject’s knowledge of what being F is in precisely the sense in which Revelation claims that this knowledge may have its source in visual experience.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

person with normal vision is facing a single prominent object with a specific shape and colour in decent lighting conditions. Consider, first, the case in which she has no prior knowledge of what being that shape/colour is. Even if she comes to her situation with suitable background knowledge of what shapes and colours are in general, attempting to deploy a determinate shape/colour concept as a mere formality, as it were, by simply rehearsing an (internal) statement that ‘there is something that shape/colour out there’ is insufficient for her genuinely to represent there being so in the sense in which this really does involve actively deploying the specific shape/ colour concept in question. In addition she must actually have conscious sight of the instance before her and attend appropriately to its shape/colour. In which case, the true source of her knowledge of what being that shape/colour is is a concurrent conscious-attentional state distinct from any conceptual representation that may draw upon it in deploying the demonstrative concept in question. On the other hand, she may be in a position immediately to issue an (internal) assertion constituting a genuinely conceptual representation that there is something that shape/ colour before her if she already has the relevant concept, and hence the knowledge of what being that shape/colour is, perhaps on the basis of previous visual encounters with various instances. In the nature of this second case, though, her representational state is not itself the source of her crucial knowledge of what being that shape/colour is, but rather presupposes it. Her perceptual demonstrative reference anaphorically exploits her previous knowledge of the shape/colour in question. The Representational account therefore fails to capture the conscious character of visual experience in a way that truly secures Revelation. Suppose, secondly, then, that the Representational account proposes that visual experiences are states whose content is not conceptual. That is to say, the proposal is that visual experiences are states with representational content such that a person may have such an experience without actively deploying the concepts used in the canonical statement of its content, which must therefore be conveyed to the subject in some other way. My objection to this proposal is that the only real alternative as an account of how non-conceptual content may be conveyed to the subject in visual experience takes us straight back to the problems of the Resemblance account considered and rejected in Section 3. Consider a subject who has a visual experience with non-conceptual representational content that is correct if and only if there is something round and red before him. He need not deploy the concepts ‘round’ and ‘red’ in his experience; and suppose without loss of generality that he previously had no knowledge of what being round and being red are. According to the non-conceptualist Representational account currently under consideration, his visual experiential state constitutes a source of just such Revelatory knowledge. It is difficult to see how this is supposed to be accomplished, though. For being in a state with non-conceptual representational content is in general manifestly insufficient for knowledge of what it is for something to fall under the concepts employed in specifying its content. Appeal to sub-personal representational states with quite complex contents is essential to the explanation of a great deal of what we say and do. For example, Peter McLeod and Zoltan Dienes (1996) have argued convincingly that a skilled fielder is able to catch a cricket ball lofted towards her by moving in such a way that d²(tanα)/dt² = 0, where α

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

 , ,    



is the angle of elevation of gaze from the fielder to the ball. Her success involves the sub-personal representation of the second derivative with respect to time of the tangent of the angle of her gaze; but there is absolutely no presumption that she knows what having a zero such second derivative is. Many successful fielders clearly have no such knowledge. Similarly, a subject may be in a sub-personal state with precisely the non-conceptual representational content of a visual experience in which it looks to him that there is something round and red before him and not thereby have any knowledge of what being round and red are. Being in a state with that content is therefore no source of such knowledge. Quite generally, whatever the nonconceptual content is supposed to be of a person’s conscious visual experience, he may be in a sub-personal state with the very same content and not know what falling under the concepts employed in specifying its content actually is. That is to say, the non-conceptual representational contents of visual experience are in themselves no source of such knowledge. So once again the Representational account misidentifies the source of Revelation in conscious visual experience in its appeal to representational content. The insistent Representationalist might attempt to block this objection by insisting that, unlike McLeod’s evidently sub-personal ball-catching contents, the representational contents of conscious visual experience may not equally be the contents of unconscious sub-personal states, in spite of the fact that neither contents make the relevant demands on the subject’s active deployment of the concepts used in specifying them. So it does not follow from the fact that ball-catching contents are no source of knowledge for the subject of what having a zero second derivative with respect to time of the tangent of an angle is that the non-conceptual contents of visual experience are likewise no source of knowledge of what being round and being red are. This reply is only as good as its explanation of the crucial difference between the two kinds of content, though. Quassim Cassam offers such an explanation on behalf of the non-conceptual Representational account in his recent discussion of ‘Berkeley’s Puzzle’ (Campbell and Cassam, 2014, ch. 8, and Cassam’s epilogue). His proposal, drawing on the ‘phenomenal intentionality research programme’ (Kriegel, 2013), is that visual experiences have their contents in virtue of their phenomenology. Such contents are therefore essentially conscious and may not equally be the contents of unconscious sub-personal states that evidently fail to provide their subject with Revelatory knowledge of what being round and being red are, for example. The structure of this response immediately gives cause for concern. What makes visual experiential content essentially conscious is supposed to be the fact that experiences have their content in virtue of their phenomenology. This leaves perfectly open, it seems to me, the possibility that sub-personal states may have precisely the same contents in virtue of something else: their non-conscious intrinsic nature, or various relational or functional properties instead. So the initial objection goes through. Simply being in a state with the representational content in question fails to secure Revelation. Furthermore, Cassam’s proposal is that it is the phenomenal nature of visual experience, in virtue of which it has its non-conceptual representational content, that is really responsible for Revelation; and this leaves entirely

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

unanswered our fundamental question of how the conscious character of visual experience in which it looks to the subject as though there is an F before her is a source of her knowledge of what being F is. The current suggestion is effectively that it is the phenomenology that does it. But that is the datum to be explained rather than any adequate explanation of it. Cassam may reply that it is at least the beginning of an explanation because the phenomenology of the visual experience in question is what explains the fact that it has a representational content that is correct if and only if there is an F out there. Phenomenology provides an alternative to the conceptualist’s appeal to the active deployment of the concept F in experience as the means by which the non-conceptual content of conscious visual experience may be conveyed to the subject. But now we are going round in a circle. Phenomenology is supposed to secure Revelation because it explains this content, although the content fails to secure Revelation since the subject may just as well be in a sub-personal state with that same content and yet have no knowledge of what being F is. So this beginning of an explanation gets us precisely nowhere. The Resemblance account considered in Section 3 at least attempts to give a substantive account of how the conscious character of visual experience itself constitutes a source of such knowledge, and perhaps Cassam’s proposal is implicitly designed to draw on some notion of resemblance between phenomenology and worldly F-ness here too. But the Resemblance account fails. So any such appeal at this crucial point in the Representational account is equally unsuccessful. This completes my dilemma for the Representational account. It fails to secure Revelation if the representational contents of conscious visual experience are supposed to be conceptual and it fails if they are supposed not to be. What is required is an alternative to the Resemblance account as a substantive explanation of how the conscious character of visual experience itself secures Revelation directly. In Section 5 I suggest that this is provided by the recognition that visual consciousness consists in the subject standing in a relation of acquaintance with mind-independent worldly objects.

6. The Relational Account The Relational account offers a direct explanation of Revelation on the basis of the conscious character of visual experience by taking the initial Moore/Strawson point absolutely literally. Their insight is that specification of the conscious character of visual experience makes essential use of concepts that have their primary application to mind-independent objects in the world around the perceiver along the following lines: it looks to the subject as though there are objects of such and such kinds with such and such properties arranged around her thus and so. In our simplified example, it looks to her as though there is something round and red before her. According to the Relational account, having an experience whose conscious character is specified in this way is normally a matter of the subject actually standing in a relation of acquaintance with something round and red out there. That round red object itself, and in particular its roundness and redness, constitute its conscious character that is therefore correctly specified in terms of its looking to her as though

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

 , ,    



there is something like that there. In this way the conscious character of visual experience is given by the nature of the mind-independent objects themselves with which the subject is acquainted in such experience. As Campbell puts it, ‘The qualitative character of the sensory experience is constituted by the qualitative characters of the objects . . . in the scene observed’ (Campbell and Cassam, 2014, 28). So the Relational account has a direct explanation of Revelation. The datum is that visual experience in which it looks to the subject as though there is something round and red before her is a source of knowledge of what being round and being red are. According to the Relational account, having such experience is being acquainted with something round and red out there, a condition whose conscious character is given by its roundness and redness that is therefore revelatory of what roundness and redness are: being round is being like thats (attending to its shape); and being red is being like thatc (attending to its colour). Recall that Lewis’s argument for Humility depends on a neutrality in our evidence for the presence of an F as to what being F actually is that permits equally admissible permutations of worldly properties for the evidential states in question. Revelation insists, on the other hand, that conscious vision of an F before us is not like this. For it constitutes a source of knowledge of what being F is. According to the Relational account, visual experience in which it looks to the subject as though there is an F before her consists in her conscious acquaintance with a particular instance of F out there in the world: a condition whose conscious character is given by the F-ness of that very object amongst other things. This is precisely what it is for her experience to be non-neutrally F-related in such a way as to block the argument for Humility and secure Revelation. The Relational account provides an intelligible explanation of what the visual Revelation that blocks Lewis’s argument for Humility actually is. I suggested in Section 3 that the Relational account may offer a picture of visual Revelation that explains the prevalence and appeal of talk of resemblance in this area even in spite of my rejection of the Resemblance account as presented there. That account depends on the incoherent idea of two quite distinct things sharing properties like roundness and redness: mind-independent objects and corresponding minddependent visual sensations. According to the Relational account, on the other hand, only mind-independent objects are ever round or red: these are modifications of the spatial extension of those things. But our visual experiences are cases of our acquaintance with such things, whose conscious character is therefore given precisely by their roundness and redness. In this sense, our experiences may be said to ‘resemble’ those things. We legitimately talk about them both in the same terms because the experiences are partially constituted by the mind-independent objects in such a way that the conscious character of the former are given by the relevant properties of the latter. Again, this in the correct sense in which visual experiences reveal the nature of mind-independent objects, thereby constituting a source of our knowledge of what being round and being red actually are.

7. Conclusion The Resemblance and Representational accounts of the nature of conscious visual experience fail to secure Revelation. The Relational account succeeds in securing

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

Revelation. So Revelation supports the Relational account over Resemblance and Representation: it distinctively favours one of the three Rs over the other two. It would be natural at this point to ask why we should accept Revelation. I simply took this as an unargued pre-theoretic commitment from the start. That seems absolutely right to me; and we do need to start somewhere. Furthermore, it is certainly not possible to give any adequate theoretical defence of Revelation here. I do believe that such a defence is possible, though, on the basis of the premise that it is conscious perception that provides our empirical thought with its subject matter and basic content. This is of course a traditional empiricist conviction, prominent in Locke (1975) and Hume (1978), for example, and equally to the fore in the logical atomist arguments of Russell (1985) and, in his own way, the early Wittgenstein (1974).¹⁴ I hope to return to this topic in its own right on another occasion.¹⁵

References Armstrong, David (1968) A Materialist Theory of the Mind, London: Routledge. Ayers, Michael (1991) Locke: Epistemology and Ontology, London: Routledge. Berkeley, George (1975) A Treatise Concerning the Principles of Human Knowledge, in Michael Ayers (ed.), George Berkeley: Philosophical Works, London: Everyman. Brewer, Bill (1999) Perception and Reason, Oxford: Oxford University Press. Brewer, Bill (2005) ‘Perceptual Experience Has Conceptual Content’, in Ernest Sosa and Matthias Steup (eds), Contemporary Debates in Epistemology, Oxford: Blackwell. Brewer, Bill (2011a) Perception and Its Objects, Oxford: Oxford University Press. Brewer, Bill (2011b) ‘Realism and Explanation in Perception’, in Hemdat Lerman and Johannes Roessler (eds), Perception, Causation and Objectivity, Oxford: Oxford University Press. Burge, Tyler (1986) ‘Cartesian Error and the Objectivity of Perception’, in Philip Pettit and John McDowell (eds), Subject, Thought, and Context, Oxford: Oxford University Press. Burge, Tyler (2010) Origins of Objectivity, Oxford: Oxford University Press. Byrne, Alex (2001) ‘Intentionalism Defended’, Philosophical Review 110: 199–240. Byrne, Alex (2005) ‘Perception and Conceptual Content’, in Ernest Sosa and Matthias Steup (eds), Contemporary Debates in Epistemology, Oxford: Blackwell. Byrne, Alex and Hilbert, David (2007) ‘Colour Primitivism’, Erkenntnis 26: 3–21. Campbell, John (1993) ‘A Simple View of Colour’, in John Haldane and Crispin Wright (eds), Reality: Representation and Projection, Oxford: Oxford University Press. Campbell, John (2002) ‘Berkeley’s Puzzle’, in Tamar Gendler and John Hawthorne (eds), Conceivability and Possibility, Cambridge, MA: MIT Press. Campbell, John (2005) ‘Transparency vs. Revelation in Colour Perception’, Philosophical Topics 33: 105–15. Campbell, John and Cassam, Quassim (2014) Berkeley’s Puzzle, Oxford: Oxford University Press. ¹⁴ For illuminating discussion of the logical atomist arguments of Russell and Wittgenstein see Pears (1967). ¹⁵ Historical debts for formative discussions of these topics go back a long way, in particular, to Michael Ayers, John Campbell, Quassim Cassam, Imogen Dickie, Naomi Eilan, Mike Martin, Paul Snowdon, and Matt Soteriou. I am also grateful for helpful comments and suggestions on previous versions of this material to the following: Tom Baldwin, Mike Beaney, Jorgen Dyrstad, Johan Gersel, Andrea Giananti, Anil Gomes, Hemdat Lerman, Heather Logue, Chris Mole, Martine Nida-Rümelin, and Gianfranco Soldati.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi

 , ,    



Chalmers, David (2004) ‘The Representational Character of Experience’, in Brian Leiter (ed.), The Future for Philosophy, Oxford: Oxford University Press. Chalmers, David (2006) ‘Perception and the Fall from Eden’, in Tamar Gendler and John Hawthorne (eds), Perceptual Experience, Oxford: Oxford University Press. Crane, Tim (1992) ‘The Non-Conceptual Content of Experience’, in Tim Crane (ed.), The Contents of Experience, Cambridge: Cambridge University Press. Cussins, Adrian (1990) ‘The Connectionist Construction of Concepts’, in Margaret Boden (ed.), The Philosophy of Artificial Intelligence, Oxford: Oxford University Press. Dretske, Fred (1981) Knowledge and the Flow of Information, Oxford: Blackwell. Evans, Gareth (1982) The Varieties of Reference, Oxford: Oxford University Press. Harman, Gilbert (1990) ‘The Intrinsic Quality of Experience’, in James Tomberlin (ed.), Philosophical Perspectives, Vol. 4, Atascadero: Ridgeview. Heck, Richard (2000) ‘Non-Conceptual Content and the “Space of Reasons” ’, Philosophical Review 109: 483–523. Hume, David (1978) A Treatise of Human Nature, ed. Lewis Selby-Bigge, rev. Peter Nidditch, Oxford: Oxford University Press. Jackson, Frank (1977) Perception: A Representative Theory, Cambridge: Cambridge University Press. Johnston, Mark (1992) ‘How to Speak of the Colours’, Philosophical Studies 68: 221–63. Kelly, Sean (2001) ‘Demonstrative Concepts and Experience’, Philosophical Review 110: 397–420. Kriegel, Uriah (2013) ‘The Phenomenal Intentionality Research Programme’, in Uriah Kriegel (ed.), Phenomenal Intentionality, Oxford: Oxford University Press. Langton, Rae (1998) Kantian Humility, Oxford: Oxford University Press. Lewis, David (2009) ‘Ramseyan Humility’, in David Braddon-Mitchell and Robert Nola (eds), Conceptual Analysis and Philosophical Naturalism, Cambridge, MA: MIT Press. Locke, John (1975) An Essay Concerning Human Understanding, ed. Peter Niddich, Oxford: Oxford University Press. McDowell, John (1986) ‘Singular Thought and the Extent of Inner Space’, in Philip Pettit and John McDowell (eds), Subject, Thought, and Context, Oxford: Oxford University Press. McDowell, John (1994) Mind and World, Cambridge, MA: Harvard University Press. McLeod, Peter and Dienes, Zoltan (1996) ‘Do Fielders Know Where to Go to Catch the Ball or Only How to Get There?’, Journal of Experimental Psychology: Human Perception and Performance 22: 531–43. Moore, George (1903) ‘The Refutation of Idealism’, Mind 48: 433–53. Pautz, Adam (2010) ‘Why Explain Visual Experience in Terms of Content?’, in Bence Nanay (ed.), Perceiving the World, New York: Oxford University Press. Peacocke, Christopher (1989) ‘Perceptual Content’, in Joseph Almog, John Perry, and Howard Wettstein (eds), Themes from Kaplan, New York: Oxford University Press. Peacocke, Christopher (1992) A Study of Concepts, Oxford: Oxford University Press. Peacocke, Christopher (2001) ‘Does Perception Have a Nonconceptual Content?’, Journal of Philosophy 98: 239–64. Pears, David (1967) Bertrand Russell and the British Tradition in Philosophy, London: Collins. Russell, Bertrand (1985) The Philosophy of Logical Atomism, ed. David Pears, Chicago, IL: Open Court. Searle, John (1983) Intentionality, Cambridge: Cambridge University Press. Siegel, Susanna. 2010. The Contents of Visual Experience, New York: Oxford University Press. Sosa, Ernest (ed.) (2001) ‘Special Symposium: Defining Intrinsic’, Philosophy and Phenomenological Research 63: 347–403.

OUP CORRECTED PROOF – FINAL, 29/10/2019, SPi



 

Strawson, Peter (1979) ‘Perception and Its Objects’, in Graham MacDonald (ed.), Perception and Identity: Essays Presented to A. J. Ayer, London: Macmillan. Stroud, Barry (2000) The Quest for Reality: Subjectivism and the Metaphysics of Colour, Oxford: Oxford University Press. Tye, Michael (1992) ‘Visual Qualia and Visual Content’, in Tim Crane (ed.), The Contents of Experience, Cambridge: Cambridge University Press. Tye, Michael (1995) Ten Problems of Consciousness, Cambridge, MA: MIT Press. Tye, Michael (2000) Consciousness, Colour, and Content, Cambridge, MA: MIT Press. Wittgenstein, Ludwig (1974) Tractatus Logico-Philosophicus, London: Routledge.

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

Index Aasen, S. 188n.31 Ability Hypothesis 91 Ability Hypothesis 75 Addis, L. 255–6 adverbialism 98–9 after-images 149–50 Agrippa’s Trilemma 22 Allais, L. 181n.21 Allen, K. 179n.16 Al-Ghazali 158–9 American Academy of Sleep Medicine 162–3 Amundsen, M. 188n.31 Anscombe, G.E.M. 98n.3, 109–20, 237n.20 anti-empiricism 110 Antrobus 156–7 Aristotle 60n.44 Aristotelian Society 1–2 Armstrong, D. 284n.11 Attention 191, 215 Attentive rationalizing 198 Joint Attention 217 Audi, R. 87n.20 Austin, D. 165n.29 Austin, J. L. 98n.3 Ayer, A. J. 98n.3, 108n.11 Ayers. M. 282 Bach, K. 7–8, 76n.4 Balog, K. 8–9, 52n.14, 68nn.67–69, 72n.86, 81n.11, 271 Baekeland 157n.22 Barcelona 55 Bayne, T. 131, 140–1 Bengson, J. 89n.21–27 Benton, M. 268–70 Bergmann, G. 255–6 Berkeley, G. 4, 44, 97n.1, 145–6, 217–18, 283–4 Berkeley’s Puzzle 21, 169 (see also Campbell, J.) 287 Bigelow, J. 8–9, 14–15 Block, N. 61n.49, 66–7, 200–1 Boër, S. 76n.4 Bohr, N. 180n.19 Bolzano 132 Bondy, P. 87n.20 Bonjour, L. 8–10, 22–3, 260, 268–9, 271 Borg, E. 231n.8 Botman 151–2

Bowie, D. 155–6 Braddon, M. 15 brain in a vat 45 Braun, D. 76n.4 Brentano, F. 51n.7, 138–9, 142–3 neo-Brentanian model 131 (see Montague) Brewer, B. 4, 17, 23–4, 42n.10, 178, 196, 197n.8, 208n.21, 277, 283n.10 Brogaard, B. 77n.5, 89–90, 266n.3 Brown, R. 61n.50 Budapest 265 Budd, M. 24 Bulkeley 156–7 Burge, T. 82–3, 284n.11 Burnham 154n.16 Byrne, A. 16, 179n.16, 277, 280n.5, 284n.11 Cambridge 98n.3, 273 Campbell, J. 7, 10–11, 17, 19, 21–4, 113n.14, 115n.16, 169, 215, 238nn.21–22, 277, 288–9 And joint attention 191–2, 215 Sea of faces example 199 Berkeley’s puzzle 21, 169, 287–9 Carpenter, 218 Carroll, N. 239n.24 Cassam, Q. 19, 21, 169, 277, 287–9 Caston, V. 60n.44 Chalmers, D. 8–10, 15–16, 22–3, 44, 52n.13, 69n.72, 72n.85, 131, 139, 171–2, 279–80, 284n.11 Chemero, A. 170n.1 Churchland, P. 8–9 Chomsky, N. 43, 176n.12, 187n.27 Chudnoff, E. 149–50 Circle of Acquaintance 19 Coates, P. 51n.11 cognitive immediacy 35 Coleman, S. 49, 51n.7, 59n.42, 61nn.49–51, 68n.66, 69nn.71,74, 70n.76 Cohen, J. 239n.24 Collier, A. 164n.27 Computational Representational Theory of Mind (CRTM) 35–41 Conee, E. 8–9, 14, 75n.2, 271–2 Conditional concepts 15 Consciousness 10, 13, 21–2, 33, 41–2 Phenomenal Consciousness 33–95

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi





Consciousness (cont.) Higher-Order Theories of Consciousness 49, 59 Hard Problem of Consciousness 13–16 Conscious Awareness (or the CA relation) 42–3 Craig, E. 264–5 Cramer-Lemley 103n.8 Crane, T. 11, 79n.9, 84n.15, 178, 268–9 Crovitz, H. F. 151–2 Curie, G. 239n.24 Dahlstrom, D. O. 131 Davidson 195n.6 Davies, R. 165n.29 Debus, D. 113n.14, 148n.5 déjà vu 151–2 demonstratives 85 Dennett, D. 105n.9, 150, 153 DePoe, J. M. 263 Descartes 4, 23, 69n.73, 141–2 Cartesian Concept of Foundational Knowledge 249 De Re vs De Dicto Reference 20, 34–5, 37 Deutscher, M. 107n.10 Dickie, I. 20, 115n.16, 219–22, 238n.23 Dienes, Z. 286–7 disjunctivism 17, 45, 142–3, 146 Domhoff 150, 153, 156–7 dreaming 145, 178 see also Freud, S. oedipal dreams 155–6 Dretske, F. 42n.8, 222, 257, 284n.11 Dreyfus, H. 170–1, 182n.24, 185–6 Dreyfus, S. 182n.24 Driver, J. 200–1 DMMs (direct meta-semantic mechanisms) 36 Dummett, M. 98n.3, 145–6 Duncan, M. 9 Dupré, J. 187n.27 Eden 267 Edinburgh 115 Eiffel Tower 114 epistemological internalism 175 Epistemological Foundationalism 245 Error-theory intentionalism Evans, G. 5–6, 11, 35, 87n.20, 108n.11, 236n.16, 237n.19 And Fregean propositions 194–5 explicit imaging conceiving (EI) 95, 103 Fales, E. 8–9, 12, 22–3, 254n.10 Fantl, J. 260 Farkas, K. 146, 260–1, 266n.2 Farrell, B. 114n.15 Feigl, H. 52n.17, 60n.44

Feinberg, T. E. 64–7 Feldman, R. 249 Fine, A. 179–80, 187n.27 Fish, B. 10–11, 17, 146, 179n.17 Fodor 35n.4, 37 Foster, J. 16–17 In relation to Hume, Russell and Anscombe 95 Foulkes, D. 151, 156–7 Fregean propositions 11–12, 194–5, 207–10 Freud, S. 151–3, 155–6 Friday, J. 239n.24 Fumerton, R. 9–10, 12–13, 22–3, 245, 246n.2, 263 Gallagher, S. 19, 131, 141–3 George IV 119 Gennaro, R. 61n.52 Gertler, B. 14, 22–3, 51n.10, 69n.72, 70n.81 Gibson, J. 182n.24 The Given (the Myth of the Given) (see also Sellars, S.) 1, 37, 40, 47, 191–3, 198, 274 Glick, E. 89n.22, 90n.27 Glüer, K. 237n.18 Goblot, E. 153 Gomes 4 Goff, P. 10, 59n.42 Goodale, M. A. 202–3 Gould, Stephen J. 186–7 Guénolé 151 Grice, H. P. 225–6 Groenendijk, J. 76n.4 Grote, John 2–3 Grzankowski, A. 75, 78n.6 Hacking, I. 188n.30 Hallucination 17–19, 44–5, 142–3, 163, 178, 182–3, 250–1 Hall, C. S. 156–7 Harlow, J. 151–2 Harman, G. 42, 42n.8, 86n.18, 231n.8, 251n.6, 284n.11 Hartmann, E. 151–3 Hartness, D. 165n.29 Hartshorne, C. 56n.33 Hasan, A. 22–3, 263 Haukioja, J. 188n.31 Hawthorne, J. 15, 20, 230, 230n.4, 231n.6, 232–4 Heck, R. 219–20 Heidegger, M. 170–1, 185 Hellie, B. 17, 77n.5 Helmholtz, H. 3 Higginbotham, J. 76n.4 Hilbert, D. 179n.16, 277, 280n.5 Hinton, J. M. 17, 146

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

 Hintikka, J. 76n.4 Hobson, J. 162 Hoffman, F. 53n.19 Hopkins 24, 103n.8 Hopp, W. 131, 188n.31 Horgan T. 16 Horikawa 157 Howell, R. 51n.8 Huang, L. 173 Huemer, M. 22–3, 245n.1 Hume, D. 55n.28, 95, 145–6, 148, 164, 252, 254–5, 290 In relation to Foster, Russell and Anscombe 95 Missing shade of blue 255n.12, 256–7 Humility 279 Hunt 156–7 Husserl, E. 19, 131–4, 136, 170–1 neo- Husserlian model 131 (see Hopp) Hyman, I. E. 151–2 Ichikawa, J. 163–4 imagination 106 IMMs (intentionally mediated meta-semantic mechanisms) 36 innate ideas 175–6 interdependence 136–7 Introspective conception 96 intuition (Anschauung) 2, 20, 132, 179, 193, 199–200, 230, 232, 236, 255 mathematical intuition 229 Ismael, J. 15 Jack the Ripper 253 Jackson, F. 283n.10 Mary and ‘The Knowledge Argument’ 8–9, 14, 34, 39–40, 72n.85, 255 James, W. 2–3, 54n.25 Jansen, J. 4 Janzen, G. 131 Jeshion, R. 20, 228–9, 230n.4, 231nn.6,9, 235 Jobe, T. A. 107n.10 Johnston, M. 11–12, 251n.6, 277, 279–80 Kant, I. 2, 4, 132, 170–1, 176–7, 181n.21, 192–3, 196, 204–5, 280n.6 Kaplan, David 35, 231nn.7,8, 238 Karpuk, D. 165n.29 Karttunen, L. 76n.4 Kassin, S. M. 151–2 Kaufer, S. 170n.1 Kennedy, J. 10–11 Ketcham, K. 151–2 Keynes 252–3 Kilner, J. M. 157–8 Kind, A. 103n.8 Kiechel, K. L. 151–2 Konigsberg, A. 24



Korcz, K. 87n.20 Knowledge Argument 81 factual knowledge 77, 261 knowledge in the biblical sense 267 know-np 266–7 knowledge-wh 76–7, 76nn.4–5, 266 non-propositional knowledge 14, 40, 58n.40, 77 objectual knowledge 260 knowledge by description vs. knowledge by acquaintance 1–3, 5–7, 12, 100, 132–3, 262 Knowles, J. 169, 170n.2, 186n.25, 187nn.28–29 Knudson 156–7 Kramer, M. 153, 157n.22 Kratzer, A. 234n.10 Kriegel, U. 42n.9, 51n.7, 53n.21, 59n.41, 62–4, 62n.53, 66–7, 69–70, 69n.71, 70n.77, 72n.85, 131, 140–1 Kripke, S. 257 Kuiken 156–7 LaBerge, S. 160–3 Ladyman, J. 187n.27 Lakoff 156–7 Langton, R. 23–4, 280n.6 Lasky, R. 157n.22 Lederman, H. 223–5 Leibal 218 Leibniz 3 Leddington, J. 183–4 Levitan, L. 160 Levin, J. 15 Levine, J. 8–9, 12, 15–16, 33, 35n.3, 36, 42n.10, 44, 72n.85 Lewis, D. 14, 23–4, 75n.1, 76n.4, 255, 277, 280n.6 Lewis’s Ramseyan Argument 280–1, 283–4 linguistic idealism 216 linguistic communication 216–17, 229, 232 Livingston, P. 24 Lopes, D. M. 239n.24 Loftus, E. 151–2 Loar, B. 15, 81n.11 Locke, J. 60n.44, 217–18, 284, 290 Logue, H. 17 London 78–80, 253, 277 Lormand, E. 77n.5 Lucretius 155n.19 Lycan, W. 15, 61n.51, 76n.4, 81n.11 Mackie, J. 41–2 Maddy, P. 229 Makin, G. 4 Malcolm, N. 107n.10, 150, 153 Manley, D. 20, 230, 230n.4, 231n.6, 232–4 Markie, P. 22–3 Marmite 91, 263–4

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi





Marr, D. 224–5 Marquardt, C. J. 151–2 Martens, David B. 262–3 Martin C. B. 107n.10 Martin, M. G. F. 6–7, 10–11, 17–19, 42n.10, 95, 122, 145–8, 203 Matrix 148 Maxwell, G. 23–4 McCarthy 159 McCullouch 170n.2 McDowell, J. 11, 17, 19–20, 181n.21, 284n.11 And conceptualism 191–3, 204 McGinn, C. 231n.6, 236n.16 McIntyre, R. 132–3 McLear, C. 4 McLeod, P. 286–7 McVittie, C. 155n.17 Meier, C. A. 157n.22 Meinong 114, 251n.6 Meskin, A. 239n.24 Merleau-Ponty 19, 133–8, 141, 143, 170–1, 185 Merritt, M. 205n.17 Milner, D. 202–3 Minier 156–7 Mirzakhani, M. 263–4 Mitochondrial Eve 230–1, 236 mnemonics 107–8 Moffett, M. 89nn.21–27 Moody, L. 153, 155n.17, 165n.29 Moore, G. E. 10, 42, 98–9, 98n.3, 113n.14, 171–2, 247 And visual experience 278, 281–4, 288–9 Mona Lisa 67–8 Montague, M. 50n.2, 131, 138–41 Monet, C. 132 Morgan, D. 238n.22 Moser, P. 87n.20 Moser, P. 22–3 Myhre, R. 188n.31 Nagel, T. 50n.1, 114n.15 naïve-realism 7, 10–11, 17–19, 145, 149–78 Nanay, B. 239n.24 Nathan, S. 103n.8 Naturalism 35 Neander, K. 60n.48 Nemirow, L. 14, 75n.1 Nes, A. 191 Newton 171 Nicolas, A. 151 Nida-Rümelin, M. 15–16, 76n.3 Nielsen, T. A. 151–2, 154n.15 Noë, A. 13, 19, 141–3, 170–1, 181–4 Noematic content 134, see Husserl, E. Nozick’s Experience Machine 148 O’Brien, F. 219–20, 238nn.21–22 Ockham 132

Oxford 98n.3 Oxford Circus 277 Pagin, P. 237n.18 pain 10, 44, 68, 80, 121, 247–9, 252–5 Palmer, J. C. 151–2 Papineau, D. 81n.11 Pargetter, R. 8–9, 14–15 Parsons, C. 229 particular-acquaintance 100–2, 110, 113n.14 Parent, T. 76n.4 Pashler, H. 173 Passmore, J. 2–3 Passover 39 Pautz, A. 284n.11 Peacocke, C. 224–5, 238n.22, 284n.11 Pears, D. 109n.12 Pepp, J. 227 perception-cum-action 19, 129 Perky, C. W. 103n.8, 164n.28 Perry, J. 15, 81n.11, 84n.15 phenomenal concepts 13–15, 68–9, 72n.86, 81–2, 88 phenomenological externalism 19, 169 phenomenal intentionality 46–7, 130 phenomenology 130 cognitive phenomenology 130 agentive phenomenology 130 Philips, I. 149n.6 Philosophy of Mind (2x) 1, 35n.3, 41–2, 72, 131, 138–40, 142 Pitt, B. 78–80 Pitt, D. 50n.2 Plato 57n.34, 260 Poldrack, R. A. 158 Poston, T. 22–3 Powell, R. A. 151–2, 154n.15 Price, H. H. 16, 103, 187n.29 Principles of Acquaintance 227 Basic Principle of Acquaintance (BPA) 232–9 Content Principle of Acquaintance (CPA) 228–39 Prinz, J. 199n.10 pre-reflective self-consciousness 139 propositional mental states 77–8 Pryor, J. 23, 148 Putnam, H. 17, 187n.29, 257 Qualia 41–2, 96 Quotational concepts 15 Quine 34n.2 Raleigh, T. 188n.31 Ramseyan argument 280–1, 280n.6 Recanati, F. 20 redundancy problem 172, 175 Red Sox, The 260 Rensink, R. 200

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi

 Reeves, A. 103n.8 relationalism 171–7 see also Campbell, J. revelation 53–5, 58, 71, 277 Rey, G. 176 Reynolds, J. 170n.2 Riddoch, M. J. 200–1 Robinson, Howard M. 17–18, 272 Robson, J. 24 Roll, S. 151–2 Rosen, M. G. 155n.18, 156n.20 Rosenthal, D. M. 57–9, 70n.77 Rubin, D. C. 151–2 Rowlands, M. 141–3 Russell, B. 1, 3, 33–4, 49–51, 50nn.3–6, 51nn.7–9, 52n.16, 70, 132, 140, 169–71, 217–18, 254n.11, 262–4, 271–3 Five Russellian Theses on Acquaintance 4 In relation to arguments of Hume and Foster 97–124 Knowledge by description 262 Principles of Acquaintance 227 Problems of Philosophy 2, 4, 23–4, 100–1, 106, 262–3 Principles of Mathematics 2, 5 Ryle, G. 141–2, 261 Sainsbury 83n.14 Sartre, J. P. 42, 138–9 Sauchelli, A. 24 satisfactional and non-satisfactional aboutness 233–5 scepticism 119n.18, 165n.29, 185, 191, 279n.3 memory scepticism 157 Schacter, D. L. 107n.10 Scholl, B. 199, 210n.23 Schneider, A. 150 Schroer, R. 56n.33 Scotus 132 Searle, J. 68–9, 131, 135, 236n.15, 284n.11 Seiffert, A. 209n.22 Segal, S. J. 103n.8, 164n.28 Sellars, S. 1, 191–2, 274, see also the Given inspiring McDowell 193, 204–5 semantic instrumentalism 231–3, 235 sense-data (datum) 4, 10, 13–14, 16–17, 33–4, 36, 50–1, 54n.24, 97–9, 102–11, 133, 185, 191–2, 228–9, 271 future sense-data 100–1 semantic theory of intentionality 130–1 shared language 215–17, 226 Shoemaker, S. 116 Siegel, S. 16, 201–4, 284n.11 Siewart, C. 131 Sikora 156–7 Silins, N. 192, 201–4 singular thought 7–8, 20, 230 Simpson, O. J. 152–3 Sinn 134



sleep 145 see also dreaming non-REM sleep 162–3 REM sleep 153, 160–3 Smith, B. C. 57–8 Smith, D. W. 19, 131, 145 Snowden 17, 114n.15, 147–8 Sosa, E. 3, 22–3, 231n.6, 247, 260, 268–71 Soteriou, M. 17, 142–3 Speckled Hen (problem of the) 22–3, 249 Spelke, E. 225 Springett, B. 165n.29 Stalnaker, R. 225 Stanley, J. 24–5, 89n.21–27, 261 Star Trek 156n.21 Stevenson, A. 268 Strauch, I. 157n.22 Strawson, P. F. 98n.3 And visual experience 278, 281–4, 288–9 Stroud 279–80 Stoljar, D. 77n.5 Stoneham 4, 155n.17 Storm, B. C. 107n.10 Strawson, G. 50n.2, 51n.7 Stokhof, M. 76n.4 Stoneham, T. 145, 149 structural realism 23–4 Sutton, J. 231n.6 sustained acquaintance 86 Swain, M. 87n.20 Sylvan, K. 87n.20 Talbert, B. 79n.9 Tarski 135 taste 3–5, 9, 44, 57–8, 79–80, 85, 91, 263–4 Taylor, C. 182n.24, 185–6 Taylor, S. 253n.9 tennis 129–39 Textor, M. 138 Thompson, E. 181 Thomasson, A. 131 Tienson, J. 16 Todd, C. 24 Todes, S. 185 Tomasello, M. 21–2, 216–17, 219 Tombu, M. 209n.22 Trafalgar Square 78–9 Travis, C. 17, 196 Trinder, J. 157n.22 Trump, D. 37–8 Tucker 246n.2 Tulving, E. 107n.10 Tye, M. 14, 34, 39–40, 42n.8, 43, 52n.15, 52nn.17–18, 57n.35, 75, 75n.2, 97n.2, 103n.8, 197n.9, 271–2, 284n.11 von Uexküll, J. 19 Urmson, J. O. 98n.3, 109n.12 Utilitarianism 2–3

OUP CORRECTED PROOF – FINAL, 31/10/2019, SPi





Valaris, M. 205n.17 Van Gulick, R. 68n.66 Varela, Francisco 19, 131, 141–3, 170–1, 182n.24 veil of appearance 98–9 Velleman, D. 235n.13 Velmans, M. 170n.2, 182n.24 Veridical 11, 45, 164 Vipassana insight meditation 139 vision/visual experience 277 Voss, U. 161 Vuilleumier, P. 200–1, 203

White, S. 15–16 Whorf, B. L. 215–16 Williamson, T. 13, 24–5, 89nn.21–23, 251, 261 Wishon, D. 9, 52n.15 Wittgenstein, L. 1, 21–2, 98n.3, 215–19, 290 Wollheim, R. 24 Worrall 23–4 Wright, C. 38n.6, 217–18 Wyller, T. 188n.31

Walton, K. 239n.24 Waterloo, Battle of 119 Watzle, S. 188n.31, 206n.18 Wheeler 107n.10, 170n.2

Zadra, A. 153 Zagzebski, L. 262–3 Zahavi, D. 51n.7, 170n.2 Zemach, E. M. 69n.71

York, University of 165n.29