Phenomenal Intentionality 9780199764297, 2012026340

Citation preview

Phenomenal Intentionality

PHILOSOPHY OF MIND

Series Editor: David J. Chalmers, Australian National University and New York University Self Expression Owen Flanagan Deconstructing the Mind Stephen Stich The Conscious Mind David J. Chalmers Minds and Bodies Colin McGinn What’s Within? Fiona Cowie The Human Animal Eric T. Olson Dreaming Souls Owen Flanagan Consciousness and Cognition Michael Thau Thinking Without Words José Luis Bermúdez Identifying the Mind U.T. Place (author), George Graham, Elizabeth R. Valentine (editors)

Phenomenal Concepts and Phenomenal Knowledge Torin Alter, Sven Walter (editors) Beyond Reduction Steven Horst What Are We? Eric T. Olson Supersizing the Mind Andy Clark Perception, Hallucination, and Illusion William Fish Cognitive Systems and the Extended Mind Robert D. Rupert The Character of Consciousness David J. Chalmers Perceiving the World Bence Nanay (editor) The Senses Fiona Macpherson (editor) The Contents of Visual Experience Susanna Siegel

Purple Haze Joseph Levine

Attention is Cognitive Unison Christopher Mole

Three Faces of Desire Timothy Schroeder

Consciousness and the Prospects of Physicalism Derk Pereboom

A Place for Consciousness Gregg Rosenberg Ignorance and Imagination Daniel Stoljar Simulating Minds Alvin I. Goldman Gut Reactions Jesse J. Prinz

Introspection and Consciousness Declan Smithies and Daniel Stoljar (editors) Decomposing the Will Andy Clark, Julian Kiverstein, and Tillmann Vierkant (editors) Phenomenal Intentionality Uriah Kriegel (editor)

Phenomenal Intentionality EDIT ED BY U R I A H K R I EGEL

1

3 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 New York Auckland Cape Town Dar es Salaam Hong Kong Karachi Kuala Lumpur Madrid Melbourne Mexico City Nairobi New Delhi Shanghai Taipei Toronto With offices in Argentina Austria Brazil Chile Czech Republic France Greece Guatemala Hungary Italy Japan Poland Portugal Singapore South Korea Switzerland Th ailand Turkey Ukraine Vietnam Oxford is a registered trademark of Oxford University Press in the UK and in certain other countries. Published in the United States of America by Oxford University Press 198 Madison Avenue, New York, NY 10016

© Oxford University Press 2013 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 license, or under terms agreed with the appropriate reproduction rights organization. Inquiries 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. Library of Congress Cataloging-in-Publication Data Phenomenal intentionality / edited by Uriah Kriegel. p. cm. — (Philosophy of mind) Includes index. ISBN 978–0–19–976429–7 (alk. paper) 1. Intentionality (Philosophy) 2. Phenomenology. I. Kriegel, Uriah. B105.I56P53 2013 128c.2—dc23 2012026340 ISBN 978–0–19–976429–7

1 3 5 7 9 8 6 4 2 Printed in the United States of America on acid-free paper

CONTENTS

About the Contributors

vii

1. The Phenomenal Intentionality Research Program

1

uriah kriegel

2. The Access Problem

27

m ichelle mon tagu e

3. Indexical Thought

49

d av i d p i t t

4. Phenomenal Presence

71

chr istopher fr ey

5. Consciousness and Synthesis

93

colin mcginn

6. Constructing a World for the Senses

99

k a t a l i n fa r k a s

7. Phenomenal Objectivity and Phenomenal Intentionality: In Defense of a Kantian Account 116 fa r i d m a s r o u r

8. Phenomenal Intentionality and the Role of Intentional Objects 137 fr eder ick k roon

v

Contents

vi

9. Unconscious Belief and Conscious Thought

156

tim cr ane

10. Intellectual Gestalts

174

elija h chudnoff

11. Does Phenomenology Ground Mental Content? a d a m pa u t z

12. Phenomenality and Self-Consciousness charles siewert

Index

261

235

194

A BOU T T H E CON T R I BU TOR S

Elijah Chudnoff is assistant professor of philosophy at the University of Miami. His main areas of research are epistemology and philosophy of mind. He has written a number of papers defending a traditional Platonist view of intuition as a form of intellectual perception. He is currently working on a book about cognitive phenomenology. Tim Crane is Knightbridge Professor of Philosophy at the University of Cambridge, having previously taught at University College London. He works mainly on the philosophy of mind and metaphysics, and has defended a nonphysicalist account of the mind and an intentionalist conception of mental phenomena. His books include The Mechanical Mind (1995), and Elements of Mind (2001). He has just completed a book on the representation of the non-existent (The Objects of Thought) and a collection of his papers, Aspects of Psychologism, is forthcoming from Harvard University Press. Katalin Farkas was born and educated in Budapest, and has been teaching at the Central European University since 2000. Her main area of research is the philosophy of mind. In her book, The Subject’s Point of View (Oxford University Press, 2008) she defends an uncompromising internalism about the mental, and an equally uncompromising conception of the phenomenal availability of mental features. Perhaps unsurprisingly, she has great admiration for Descartes and hopes to make a modest contribution to restoring his reputation after a century or so of bad press. Christopher Frey is an assistant professor of philosophy at the University of Chicago. His research is in the philosophy of perception and ancient Greek philosophy, especially Aristotle’s metaphysics and natural philosophy. Uriah Kriegel is a research director at the Jean Nicod Institute in Paris, working mainly in philosophy of mind and metaphysics. His books include Subjective

vii

viii

About the Contributors

Consciousness: A Self-Representational Theory (Oxford University Press, 2009) and The Sources of Intentionality (Oxford University Press, 2011). Frederick Kroon teaches philosophy at the University of Auckland. He has published on a range of topics, including the theory of reference, the nature of rationality, semantic paradox, nonexistence, and the pretense approach to the nature and semantics of fiction. His current research involves applying fictionalist ideas to problems and puzzles in the philosophy of language. He is a subject editor for twentieth-century philosophy for the Stanford Encyclopedia of Philosophy, and holds the PhD from Princeton University. Colin McGinn teaches philosophy at the University of Miami, specializing in philosophy of mind, metaphysics, and philosophy of language. He has taught at Rutgers University, Oxford University, and London University. He has published some twenty books, ranging from consciousness to evil, Shakespeare to sport, fi lm to logic, Wittgenstein to imagination. He has written extensively for the general reading public and has published a novel. He is an avid tennis player and surfer. Farid Masrour is a College Fellow in the Philosophy Department at Harvard University. He is interested in philosophy of perception and in related areas of philosophy of mind, epistemology, history of philosophy, and cognitive sciences Michelle Montague is lecturer in philosophy at the University of Bristol and an associate professor of philosophy at the University of Texas, Austin. Her main interests are in philosophy of mind, philosophy of language, and metaphysics. She is an editor of Cognitive Phenomenology (Oxford University Press, 2011) and is currently working on a book on mental content, with particular reference to the relationship between phenomenology and intentionality. Adam Pautz is an associate professor of philosophy at the University of Texas at Austin. He works on consciousness, the philosophy of perception, the sensible qualities, and “the naturalization program.” David Pitt is professor of philosophy at California State University, Los Angeles. He has published articles in philosophy of mind, philosophy of language, and metaphysics. He is currently working on a book manuscript on thought and reference. Charles Siewert is Robert Alan and Kathryn Dunlevie Hayes Professor of Humanities and professor of philosophy at Rice University. He is the author of The Significance of Consciousness (Princeton University Press, 1998) and has written extensively on the interpretation and explanation of phenomenal consciousness and its relation to introspective self-knowledge, perceptual experience, and conceptual thought.

1

The Phenomenal Intentionality Research Program U r i a h K r i ege l

Introduction/Abstract Since the late seventies, the main research program for understanding intentionality has been based on the attempt to naturalize intentionality by identifying a natural relation that holds between internal states of the brain and external states of the world when and only when the former represent the latter. Call this the Naturalist-Externalist Research Program, or NERP. Different versions of NERP differ on how they construe the relevant natural relation. Typically, it is construed as involving in its core a type of tracking relation, whereby internal states occur sensitively to the presence of specific external conditions.1 Some philosophers, however, have remained skeptical of this entire approach. In particular, some have argued that phenomenal consciousness has an essential role to play in the theory of intentionality, a role it is not accorded in NERP. Thus a number of authors have recently brought to the fore the notion of phenomenal intentionality, as well as a cluster of nearby notions. There is a vague sense that their work is interrelated, complementary, and mutually reinforcing, in a way that suggests a germinal research program—what I call the Phenomenal Intentionality Research Program, or PIRP. My principal goals in this chapter are the following. First, I will explicitly articulate, and elaborate upon, what I take to be the fundamental claims at the heart of this nascent research program. Secondly, I will review some of the arguments in the recent philosophical literature that support key theses of PIRP, in the process illustrating some apparent advantages of the framework in comparison to NERP. Thirdly, I will point out some important theoretical options within the generic PIRP framework, specifying issues about which proponents of the general framework might disagree. 1

2

phenomenal intentionality

1. An Overview of PIRP Phenomenal intentionality is the intentionality a mental state exhibits purely in virtue of its phenomenal character. As far as I know, the term makes its first appearance in two papers circulating in the second half of the nineties, Brian Loar’s (2003) “Phenomenal Intentionality as the Basis of Mental Content” and Terry Horgan and John Tienson’s (2002) “The Phenomenology of Intentionality and the Intentionality of Phenomenology.” Loar’s paper started circulating in the late nineties, however, and in any case the idea of phenomenal intentionality is present in Loar’s work much earlier. His 1987 paper “Subjective Intentionality” may be regarded as the first explicit published discussion of phenomenal intentionality in analytic Anglo-American philosophy. Arguably, however, the lively debates on intentionality among Brentano and his students (see Brentano 1874, Twardowski 1895, Husserl 1901, and Meinong 1904) in fact concerned phenomenal intentionality, which was probably the only intentionality they recognized.2

1.1. Toward a New Research Program The term “research program” can be used to intimate two different things. In one sense, the term is used to refer to a degraded kind of theory—a cluster of ideas whose purpose is to become a comprehensive theory of some phenomenon but which has not yet reached the maturity and cohesion required. In another sense, the term is used to denote a perfectly clear general framework for the study of some phenomenon, a framework within which several distinct theories could be pursued that share a fundamental commonality (a “paradigm” perhaps). My claim here is that work on phenomenal intentionality is on the cusp of qualifying as a research program in the second sense. Compare NERP, where a number of distinct theories can be discerned—informational semantics, functional role semantics, teleosemantics—that nonetheless share the general idea that some kind of naturalistically kosher relation between brain events and world events must underlie intentionality. Put impressionistically, the basic idea shared by all theories within NERP is that intentionality is injected into the world with the appearance in nature of a certain kind of tracking relation. It is when the relevant tracking relation occurs between distinct states (including brain states and environmental states) in the world that intentionality makes its first appearance on the scene. Once intentionality has thus been injected into the world, it can start being “passed around” so it is somewhat “freed from” the relevant tracking relation. Linguistic expressions, paintings, and traffic signs, for example, may represent even in the absence of tracking, because they somehow derive their intentionality from things that do track. But the source of all intentionality is the relevant kind of tracking relation. Something like this is the paradigm that guides work on intentionality within NERP.

The Phenomenal Intentionality Research Program

3

The basic, guiding idea of PIRP could be thought of on the same model (Kriegel 2011). The cornerstone would be the idea that intentionality is injected into the world with the appearance of a certain kind of phenomenal character. It is when the relevant phenomenal character shows up that intentionality makes its first appearance on the scene. Here too, once this phenomenal character appears, and brings in its train “original intentionality,” intentionality can be “passed around” to things lacking this (or any) phenomenal character. But the source of all intentionality is the relevant phenomenal character. Within NERP, different comprehensive theories of intentionality may differ along two central dimensions. First, they may differ on what they identify as the relevant tracking relation that brings intentionality into the world. Is it a causal relation, an informational relation, a counterfactual-dependence relation, a teleological relation? Much of the debate on intentionality during the eighties concerned this question. Secondly, they may differ on how they choose to account for derived intentionality, in language and other forms of tracking-free intentionality.3 The leading account here is probably Grice’s (1957, 1969) intention-based theory (see also Schiffer 1982), which is most naturally applied to language, but may also be extended to pictorial representation (Abell 2005, Blumson 2006) and perhaps other forms of derived intentionality. One may say, then, that comprehensive theories of intentionality within NERP are composed of two chapters, namely, (i) an account of the relevant tracking relation and (ii) an account of the derivation relation (in the relevant sense).4 More generally, the two chapters concern source intentionality (if you will) and non-source intentionality. This general structure can be reproduced for PIRP. This would be to think of a comprehensive theory of intentionality as involving the following two chapters: (i) an account of the kind of phenomenal character that constitutes source intentionality and (ii) an account of the derivation relation that underlies non-source intentionality. Accordingly, different theories of intentionality within PIRP would differ in the kind of phenomenal character they identify as the source of all intentionality and/or in how they choose to account for phenomenality-free intentionality (in language, pictures, unconscious mentation, etc.).5 In the remainder of this section, I offer a first pass at understanding PIRP— what it attempts to do (§1.2) and how it attempts to do it (§1.3). Later, I will review some of the main ideas surrounding the notion of phenomenal intentionality (§2), then outline the structure of a general theory of intentionality within PIRP (§3).

1.2. Location Projects and the Theory of Intentionality Jackson (1998) usefully describes the core of what he calls “serious metaphysics” as the project of addressing the “location problem”: the problem of finding a

4

phenomenal intentionality

place for some phenomenon (typically familiar from the manifest image) in a description of the world cast entirely in some privileged vocabulary. In fact, however, there appear to be two kinds of location problem one might distinguish. I will call these naturalistic location and foundational location (respectively).6 Roughly and generically, naturalistic location of a puzzling phenomenon (e.g., intentionality) is a matter of identifying it with some phenomenon also describable in such a way that the phenomenon, as thus alternatively described, counts as “non-mysterious” and “kosher” within a metaphysical perspective that gives pride of place to natural science. In the case of naturalizing intentionality, the project is to locate intentionality in a world fully described by the natural sciences. To do so is to “naturalistically locate” intentionality. Naturalistic location thus involves taking the vocabulary of the natural sciences to be the privileged vocabulary in serious metaphysics. Foundational location, on the other hand, involves taking some other vocabulary, considered in some sense foundational, to be privileged. In the case of intentionality, for example, foundational location is, roughly, a matter of (i) identifying the fundamental source(s) of intentionality in the world (the way it gets “injected” into the world), and (ii) identifying the principal way(s) derivative kinds of intentionality arise from it. Within NERP, the two kinds of location project—foundational and naturalizing—are effectively pursued in tandem. Proposed “sources” are naturalistically respectable phenomena (causal, covariational, informational, or teleological relations); and their spread to non-source intentionality is construed as naturalistically respectable as well. Part of the philosophical motivation for proposing such sources is the desire to foundationally locate intentionality in a manner that also constitutes naturalistic location of these phenomena. Within PIRP, things proceed differently. Pride of place is given to the idea that phenomenal consciousness is implicated in the source of intentionality— it somehow founds, or grounds, all intentionality.7 Naturalization, if it is to come at all, needs to be a naturalized version of PIRP (“NPIRP,” if you will).

1.3. Fundamental Tenets of PIRP As noted earlier, work on phenomenal intentionality already qualifies as a research program in the looser sense, since many interrelated theses recur in discussions of the notion. I now propose, in rough and generic terms, an explicit articulation of the common basic ideas that are implicit in this work on phenomenal intentionality. The six theses below will be put somewhat vaguely, in order to be suitably generic.8 It would be too much to expect that all philosophers who qualify as pursuing PIRP subscribe to all six theses. But my suggestion is more modest: that (a) most would subscribe to all six theses and (b) all

The Phenomenal Intentionality Research Program

5

would subscribe to most theses.9 I first present the six theses, then offer some initial clarification of each. Phenomenal Grounding. There is a kind of intentionality—phenomenal intentionality—that is grounded in phenomenal character. Inseparatism. The phenomenal and the intentional do not form two separate mental realms, but are instead inseparably intertwined. Distinctiveness. Phenomenal intentionality is special and distinctive, in that it has certain important properties that non-phenomenal forms of intentionality do not. Narrowness. Phenomenal intentionality is narrow, that is, it is not constitutively dependent upon anything outside the experiencing subject. Subjectivity. Phenomenal intentionality is inherently subjective: it is built into the phenomenal character of a phenomenally intentional state that it (re)presents what it does to someone. Basicness. Phenomenal intentionality is a basic kind of intentionality and functions as a source of all intentionality. Some initial clarifications are in order. Regarding Phenomenal Grounding, what the pertinent kind of grounding comes to, metaphysically, is a matter about which different PIRPers could go different ways. One clarification is essential though: somewhat atypically, as I use the term “grounding,” the grounding relation need not be anti-symmetric. On the contrary, it could well be that the phenomenal property and the intentional property it grounds are strictly identical. (More on that in 1.4.) As for Inseparatism, a traditional picture divided the mind into two separate realms: sensory states, which are essentially phenomenal but non-intentional, and cognitive states, which are essentially intentional but non-phenomenal (Horgan and Tienson 2002). This is what Inseparatism rejects, holding instead (i) that paradigmatic sensory states in fact exhibit intentionality, which is moreover grounded by their phenomenality, and (ii) that paradigmatic cognitive states in fact boast a phenomenality, which moreover grounds their intentionality. Next consider Distinctiveness. This is the idea that there is something special about phenomenal intentionality that sets it apart from other kinds of intentionality. Phenomenal intentionality has certain significant peculiarities. Part of the reason the notion of phenomenal intentionality becomes the focus of dedicated research is precisely this distinctiveness. Concerning Narrowness, now, an intentional property is narrow just in case it supervenes on the subject’s non-relational properties, that is, is “locally supervenient.” Thus to say that phenomenal-intentional properties are narrow is to say that they are locally supervenient.

6

phenomenal intentionality

As for Subjectivity, I note that recent philosophical work on phenomenal consciousness has increasingly emphasized the fact that phenomenal character has to do not just what it is like but with what it is like for the subject. This feature is inherited by phenomenal intentionality: a phenomenally intentional state presents what it does to the subject. (More on this in 2.3.) Finally, Basicness is essentially the thesis that all intentionality derives from phenomenal intentionality. It appears to follow from this that in the absence of phenomenal intentionality there would be no intentionality at all. Clearly, this is a linchpin thesis for the foundational location project.

1.4. Some Theoretical Options within PIRP There are a number of important issues about which proponents of the generic PIRP framework could take different positions. Here I will briefly describe some of these issues and some principal theoretical options concerning them.10 First is the relation between phenomenal properties and fundamental intentional properties. One view is that the “relation” is outright identity: the intentional property just is the phenomenal property. An alternative view is that phenomenal properties are distinct from the relevant intentional properties, that the intentional properties supervene with metaphysical necessity upon the phenomenal properties, and/or that the intentional properties are realized by the phenomenal properties. A third option is that there is an anti-symmetric relation of metaphysical dependence that the intentional properties bear to the phenomenal properties and that goes beyond mere supervenience—a sort of “in virtue of” relation. There may be other options as well. Second is the extent of phenomenal intentionality. One view is that phenomenal intentionality is confined to perceptual experience, or to this plus somatic and emotional experience. Another view is that it is much more pervasive, including for instance the phenomenology of agency, the phenomenology of thought (so-called cognitive phenomenology), and so on.11 Embracing the leaner view presumably would greatly complicate the task of making a case for Basicness; the wider the extent of phenomenal intentionality, the more tenable Basicness is. However, there is nothing incoherent about pursuing PIRP while rejecting an expansive account of phenomenology. One could perfectly well hold that all phenomenology is sensory and nonetheless subscribe to all six theses listed earlier. This is important, because some commentators discuss the notion of phenomenal intentionality as though it is definitionally tied to the possibility of cognitive phenomenology (e.g., Lycan 2008). Third is the question of whether (some or all) phenomenally intentional states exhibit any kind of non-foundational, merely derived intentionality along with their underived source intentionality. One view is that some or all do, one that none do.

The Phenomenal Intentionality Research Program

7

Fourth is the question of whether phenomenal intentionality is in fact naturalistically, or even physicalistically, locatable. One view is that it is, even though the relevant kind of naturalization evidently would need to look rather different from NERP. The opposite view is that phenomenal-intentional properties are primitive and irreducible (or at least not reducible to physical properties).12

2. Arguments for Central PIRP Theses In this section, I survey various arguments from the recent philosophical literature pertaining to the existence of phenomenal intentionality, its scope and extent in human mentality, its distinctive features, and its basic role as the source of all intentionality.

2.1. The Existence of Phenomenal Intentionality Why think that there is such a thing as phenomenally grounded intentionality? One consideration unlikely to win converts but central in motivating sympathizers is the idea that phenomenal intentionality is simply introspectively manifest: attending to one’s stream of consciousness in the right way brings out that some conscious episodes are intentional, and intentional because phenomenal. Certainly introspection suggests that some mental states have both intentionality and phenomenality. Whether introspection reveals that sometimes the former is grounded in the latter is a harder question. Admittedly, it is implausible that introspection presents any grounding relation as a grounding relation. But it is much more plausible that introspection presents what is in fact a grounding relation under a simpler guise, but in such a way that a sufficiently sophisticated theoretician could justifiably conceptualize what is presented as grounding.13 An argument with a more neutral starting point is due to Charles Siewert (1998). He notes that, purely in virtue of their phenomenal character (and without need of interpretation), conscious experiences are often assessable for accuracy. Suppose you undergo an experience with a squarish phenomenal character. If nothing around you is square, your experience is assessable as inaccurate. If the right object or surface is square, your experience may be assessable as accurate. Thus phenomenal character can bring in its train accuracy conditions. Since having accuracy conditions is an intentional property, it appears that at least some phenomenal character can guarantee intentional properties. Finally, the existence of phenomenal intentionality may be supported with thought experiments. For instance, we can conceive of a disembodied soul in

8

phenomenal intentionality

an otherwise empty world who is phenomenally indistinguishable from us (Kriegel Ms). Intuitively, portions of the soul’s inner life are also intentionally indistinguishable (in the sense of instantiating the same intentional properties) from corresponding portions of our own inner life—we and the soul will have some intentional states in common. The fact that some phenomenal duplication secures intentional duplication suggests (perhaps entails) that some phenomenal properties are sufficient for intentional properties. Another thought experiment is the more familiar brain-in-vat one (Horgan et al. 2004). Intuitively, an envatted brain that is physically identical throughout its existence to your brain, with identical sensory inputs all the while, would have a conscious mental life that intentionally matches yours exactly. And intuitively, the basis for this intentional match would be that this envatted brain would have a mental life some of which is phenomenally exactly like yours.

2.2. The Scope of Phenomenal Intentionality Once one knows that there is phenomenal intentionality in the world, one wants to know just how much of it there is. Much of the work carried out by PIRPers concerns this question. The question may be profitably divided into two sub-questions, concerning (respectively) sensory and non-sensory phenomenal intentionality. The challenge presented by each question has been quite different. In the sensory domain, it is widely acknowledged that there is sensory phenomenology; the challenge has been to show that there is a sensory intentionality it grounds. In the non-sensory domain, it is widely acknowledged that there is non-sensory intentionality; the challenge has been to show that there is a non-sensory phenomenology that grounds it. In this section, I review some work on non-sensory phenomenal intentionality.14 This work has tended to fall in turn into two categories, concerned with phenomenal intentionality within and without the sphere of perceptual experience. Within the perceptual sphere, there are elements in perceptual experience that are intentionally rich but are claimed to nonetheless involve phenomenal intentionality. Perhaps the most systematic contribution to the study of perceptual phenomenal intentionality is due to Susanna Siegel (2005, 2006a), who argues that high-level properties are represented in perception. These include causation, meaning, and kind properties. The idea is that we not only understand, but can also perceive, that one billiard ball causes the motion of another; that some words on a page mean that the basketball game has been canceled; that something colorful and shapely is a parrot (see also Siewert 1998). A particularly intriguing debate in this area concerns the point at which perceptual experiences start presenting us with an objective world, a world whose character is independent of the subject’s perceptual activity. Siegel

The Phenomenal Intentionality Research Program

9

(2006b) herself argues that this occurs when part of the content of a perceptual experience is that its object will not move if the subject changes her spatial perspective on it. Farid Masrour (2008) argues that this occurs rather when the experience acquires a certain phenomenal feature he calls “dynamic unity structure.”15 Other views are also possible. A related and quite central debate concerns the representation of perspectival properties. Suppose you look at a tilted coin. Does your perceptual experience present the coin as (having the non-perspectival property of being) circular or as (having the perspectival property of being) elliptical? Kelly (2004) argues that only the (non-perspectival) circularity is presented in experience, whereas Noë (2004) claims that both the (non-perspectival) circularity and (perspectival) ellipticality are presented. For a variety of reasons, friends of phenomenal intentionality have often been keen to argue for the existence of non-perceptual experience with a purely intellectual or cognitive phenomenology that constitutes its intentionality. Perhaps the strongest thesis in this vein is Pitt’s (2004) claim that thoughts have a phenomenology which is both proprietary and individuative, that is, a phenomenology that both is different from all other types of phenomenology and varies whenever the content varies.16 The literature contains two main types of argument for claims of this sort (though there are at least three more minor types of argument I will not survey here). One is an argument from phenomenal contrast: two conscious episodes are contrasted, where (i) there is clearly an overall difference in what it is like to undergo these episodes and (ii) the best account of that difference is in terms of a difference in a purely cognitive phenomenal character. A much discussed argument of this form is due to Strawson (1994), though it is present in essence already in Moore (1953). Strawson argues for the existence of “understanding experience” by contrasting the overall phenomenologies of a French speaker and a non–French speaker listening to the news in French. Strawson claims that there is a difference in what it is like for them to listen to the news, and that the difference is best accounted for in terms of an element of understanding experience present only in the French speaker’s phenomenology.17 The second kind of argument appeals to an asymmetric access one has to one’s conscious cognitive states and their contents (Goldman 1993, Pitt 2004). Schematically, the argument proceeds as follows: one has a special, immediate access to some of one’s cognitive states (and their contents); only to phenomenal states (and contents) can one have this kind of special access; therefore, (some of) one’s cognitive states (and their contents) are phenomenal. This is a very partial survey of research in this area. The telos of this research appears to be to establish that sensory states’ phenomenology is inherently intentional and that non-sensory states’ intentionality is phenomenally grounded. If something like this is established, then the separatist picture

10

phenomenal intentionality

of the mind collapses, and a more unified picture emerges that conforms to Inseparatism (see §1.3). One way to develop the inseparatist picture is to construe the concept of mind as a prototype concept, such that a state or event qualifies as mental to the extent that it is appropriately related to prototypical mental states, but where all (and perhaps only) prototypical mental states are phenomenally intentional states (Horgan and Kriegel 2008).

2.3. The Distinctiveness of Phenomenal Intentionality Another focal point of much work on phenomenal intentionality is the thought that there is something special about phenomenal intentionality—that phenomenal intentionality exhibits significant distinctive features absent in other types of intentionality. This is important, because if phenomenal intentionality is indeed distinctive, it may resist theoretical treatment perfectly suitable for other types of intentionality. For example, one might hold that while teleosemantics is the true theory of non-phenomenal intentionality, some other theory would be needed to accommodate the distinctive features of phenomenal intentionality. One recurrent distinctiveness claim is that only phenomenal intentionality has determinate content in and of itself (Searle 1991, 1992, Loar 1995, Horgan and Tienson 2002, Strawson 2008, Horgan and Graham forthcoming). We may state this thesis as follows: Determinate Content. Necessarily, for any intentional state M with content C, if C is non-derivatively determinate, then M is phenomenally intentional. On this view, non-phenomenally intentional states can have determinate content, but not in and of themselves. Instead, they must derive their determinate content from phenomenally intentional states they are appropriately related to. By “determinate content,” I simply mean content which is as fine-grained as one’s intentional contents appear pre-theoretically to be. For example, pre-theoretically it seems that one’s thoughts are fine-grained enough to be about rabbits rather than undetached rabbit parts, about Phosphorus rather than Hesperus, about triangles rather than closed trilateral figures, and so on. If a kind of intentional state is not this fine-grained, I say that its content is indeterminate. And if it is this fine-grained, but not intrinsically so, I say that its content is only derivatively determinate. Determinate Content claims that only phenomenally intentional states have non-derivatively determinate content.18 One key argument for this thesis, in very rough outline, is that nothing other than phenomenal character can secure content determinacy.

The Phenomenal Intentionality Research Program

11

In particular, tracking relations cannot account for this determinacy—not even when teleologically augmented. Whenever an internal state bears tracking relations to rabbits or Phosphorus, it also bears them to undetached rabbit parts or Hesperus. And as Fodor (1984, 1990) argued long ago, whenever tracking a property F is adaptive for an organism, it is also and equally adaptive for it to track any property coextensive with F. Evolutionary benefit cannot discriminate between coextensive properties (let alone necessarily coextensive ones). Of course, for all I just said, there may be some other non-phenomenal feature that can secure content determinacy, but proponents of Determinate Content have attempted to consider all the initially plausible candidates and argue against them (see especially Horgan and Graham forthcoming). Another claim with the same general structure is that only phenomenal intentionality is intrinsically subjective (McGinn 1988, Kriegel 2003b, Georgalis 2006). Put in McGinn’s (1988) terms, the basic idea is that conscious content is Janus-faced, in that in addition to its outward-looking face of presenting some object or state of affairs in the world, it also possesses an inward-looking face involving an elusive presence to the subject. This can perhaps be interpreted more rigorously as the thesis that while unconscious intentional states instantiate in and of themselves only the two-place relation x represents y, conscious ones instantiate the three-place relation x represents y to z. Let us say that an intentional state M is non-derivatively subjective just in case it instantiates the three-place representation relation intrinsically, that is, in and of itself.19 Then we may put the thesis as follows: Intrinsic Subjectivity. Necessarily, for any intentional state M, if M is non-derivatively subjective, then M is phenomenally intentional. The thesis allows unconscious intentional states to instantiate the three-place representation relation, but not intrinsically. Rings on a tree trunk can represent the tree’s age to a botanist, and an unconscious state in the dorsal stream of the visual system can represent a circle to a neuroscientist. But in these cases, the representation-to is not inherent in the representation-of. Rather, the representation of a worldly feature represents what it does to someone simply because someone harbors a separate representation of it; it is only thanks to this other (second-order) representation that the original (first-order) representation represents to someone. Phenomenally intentional states, by contrast, are non-derivatively subjective in that they represent what they do to someone without requiring that “someone” to harbor a numerically distinct representation of them. It is not easy to discern in the existing literature a clear argument for Intrinsic Subjectivity. This is probably because the pull of the thesis is mostly phenomenological. It has sometimes been claimed, on broadly phenomenological

12

phenomenal intentionality

grounds, that all conscious states necessarily involve a kind of for-me-ness, or subjective significance, whereby their subject is aware of them in an immediate and somewhat elusive manner (Levine 2001, Kriegel 2005, 2009, Horgan et al. 2006). It is natural to think that when an intentional state is conscious, its for-me-ness will manifest itself as the kind of non-derivative subjectivity under discussion. Here too it is clear that the intrinsic subjectivity of phenomenally intentional states creates a prima facie problem for NERP, as the latter is geared to account mostly for representation-of, not representation-to. Perhaps one exception is Millikan’s (1989) consumer semantics, which may be naturally thought of as capable of accounting for representation-to. In consumer semantics, the content of a representation R is determined not by the way the system that produces R tracks conditions in the world, but by the way the downstream systems that consume R track them. In a way, we may say that it is what R represents to the consumer systems that determines R’s content. Thus consumer semantics offers a NERP-y gloss on representation-to. However, it is unclear that this type of representation-to is the one referred to by proponents of Intrinsic Subjectivity. For starters, the latter is always and necessarily a personal-level phenomenon, whereas the former can be a sub-personal phenomenon. Likewise, since the latter is a phenomenal element and the former is a matter of accessibility or consumability, it would appear that the latter is an occurrent property whereas the former is a dispositional one. Fuller discussion of these issues cannot be attempted in this review, but clearly there are some principled obstacles in the way of a consumer-semantic account of representation-to.20 A third distinctiveness claim that shows up often in the relevant literature is that phenomenal intentionality always involves narrow content, in the sense of being shared by intrinsic duplicates, whereas non-phenomenal intentionality often involves wide content (Loar 2003, Horgan and Tienson 2002, Horgan et al. 2004, Georgalis 2006, Kriegel 2007, 2011).21 If a kind of content internalism were generally true of phenomenal intentionality, while content externalism were commonly true of non-phenomenal intentionality, that would constitute another distinction of phenomenal intentionality. Any distinctiveness thesis based on narrowness, however, would have to stress a quantificational difference only: phenomenal intentionality is always narrow, whereas non-phenomenal intentionality is only sometimes narrow. For it is clear that non-phenomenal intentionality can be and often is narrow as well. For some non-phenomenal representations represent non-Twin-Earthable properties. For example, an unconscious representation of a chair is arguably shared by intrinsic duplicates, since arguably there are no Twin-Earth scenarios in which something looks superficially like a chair but lacks some chairly underlying nature. Arguments for internalism about phenomenal intentionality vary depending on the strength and nature of the internalism. But the basic argument for

The Phenomenal Intentionality Research Program

13

the narrowness of phenomenal intentionality is fairly straightforward (Horgan et al. 2004): phenomenal intentionality supervenes on phenomenal character; phenomenal character is locally supervenient; therefore, phenomenal intentionality is locally supervenient. The first premise is more or less definitional, the second is supported by brain-in-vat thought experiments (involving phenomenal duplicates that lack the relevant relations to the environment), and the inference is buttressed by the transitivity of supervenience. Content determinacy, intrinsic subjectivity, and narrowness are recurring central claims among philosophers working on phenomenal intentionality. There are other, more idiosyncratic claims that can and have been made, but which will not be discussed here.

2.4. The Basicness of Phenomenal Intentionality Perhaps the most important kind of claim made on behalf of phenomenal intentionality is that it is in some way basic among forms of intentionality (e.g., as in §1.2). In this section, I review first the main theses alleging basicness, then the kinds of argument that have been offered in their favor. The strongest possible basicness thesis is that, in reality, there is no intentionality but phenomenal intentionality—all intentionality is phenomenal (Strawson 2008, Georgalis 2006).22 On this view, it is simply false that there is any non-phenomenal intentionality. There may well be a variety of information-bearing states, but for one reason or another those do not qualify as intentional states. The only states that qualify as intentional are phenomenal ones. A more lenient view allows for some non-phenomenal intentionality but claims that any such must derive from phenomenal intentionality (McGinn 1988, Kriegel 2003b, 2007, 2011). The distinction between derived and underived intentionality was brought into modern discussions of intentionality by Grice (1957), who suggested that the intentionality of language derived from the intentionality of thought: the linguistic symbol c^a^t represents cats only in virtue of bearing a certain relation to cat thoughts, whereas cat thoughts have cat-representing content in and of themselves. More generally, it has often been suggested that all non-mental intentionality derives from mental intentionality (see Cummins 1979, Searle 1983, Dretske 1988). But proponents of phenomenal intentionality, while embracing the distinction between derived and underived intentionality, have sometimes sought to draw it more restrictively, claiming that the intentionality of non-phenomenal mental states derives from that of phenomenal ones. Thus non-phenomenal mental states and non-mental items may be intentional, but they must derive their intentionality from phenomenally intentional states. A slightly weaker thesis that might easily be confused with this holds that phenomenal intentionality is the only underived intentionality, and all non-phenomenal intentionality is derived, but does not claim that

14

phenomenal intentionality

non-phenomenal intentional states (and items) derive their intentionality from phenomenal-intentional states (Bourget 2010). Instead, it allows non-phenomenal states to derive their intentionality from each other, in contrast to phenomenal-intentional states, each of which has its intentionality independently of other intentional states. A significantly weaker claim allows non-phenomenal states to boast underived intentionality but requires that such states be potentially phenomenal-intentional (Searle 1991, 1992). In other words, only states that could potentially have phenomenal intentionality have underived intentionality. Presumably, these states have their intentionality precisely because they could potentially become phenomenal-intentional. Thus, a tacit and unconscious belief that 13.46>8.27 is endowed with underived intentionality, but only because it could be conscious. Indeed, its intentional content is partly determined by the phenomenal character it would have if it were conscious. A thesis weaker yet would require non-phenomenal states to bear some relation to phenomenal states in order to qualify as non-derivatively intentional, but not necessarily the relation of “potentially becoming” (or “potentially being”). Instead, it would allow various inferential and/or causal relations to experiential states to suffice for derived intentionality (Horgan and Tienson 2002, Horgan and Graham forthcoming, Loar 2003). As Davies (1995) notes, some sub-personal states, such as Marr’s (1982) 2.5D sketches, are naturally construed as non-derivatively intentional but are not even potentially conscious by any intuitive construal of “potentially.” However, even such states are cognitively integrated into a system of inferentially interrelated intentional states some of which are phenomenally conscious. It may therefore be suggested that being thus integrated is a necessary condition on a mental state’s being non-derivatively intentional. On this view, a mental state qualifies as intentional only if it is related to phenomenally conscious states by the relation of being integrated into a single inferential web.23 In sum, there are four main grades of basicness claims discernible in the existing literature: (B1) All intentionality is phenomenal intentionality. (Strawson, Georgalis) (B2) All intentionality derives from phenomenal intentionality. (McGinn, Kriegel) (B3) All intentionality derives from potentially phenomenal intentionality. (Searle) (B4) All intentionality derives from intentionality appropriately related to phenomenal intentionality. (Horgan et al., Loar) Although the theses are different, the arguments adduced in their favor are often neutral between them and can be wielded in defense of several or any.

The Phenomenal Intentionality Research Program

15

I end this section with a review of the main arguments to be found in the literature. In the present context, I am not endorsing any of them; my purpose is merely expository. Most of these arguments follow a similar schema (Kriegel 2003b). First, an asymmetry is established between phenomenal and non-phenomenal intentionality, along the lines of one of the distinctiveness claims surveyed in §2.3. Then, it is argued that the distinctive feature of phenomenal intentionality is mandatory in the basic kind of intentionality (the underived kind, the occurrently underived kind, or the not-merely-relationally underived kind). It is then concluded that only phenomenal intentionality is basic. Thus a schema for an argument for the primacy of phenomenal intentionality emerges: 1) Only phenomenal intentionality has feature F; 2) Only intentionality with feature F is basic; therefore, 3) Only phenomenal intentionality is basic. An argument from Determinate Content fitting this schema would proceed as follows: 1) Only phenomenal intentionality has non-derivatively determinate content; 2) Only intentionality with non-derivatively determinate content is basic; therefore, 3) Only phenomenal intentionality is basic. This is probably the most common type of argument for the basicness of phenomenal intentionality one finds in the relevant literature (see Loar 1987, 1995, Searle 1992, Horgan and Tienson 2002, Strawson 2008, Georgalis 2006, and Horgan and Graham forthcoming). The central claim of Loar’s argument, for instance, is that non-phenomenal states, left to their own devices, so to speak, would be referentially inscrutable—there would be nothing to make them about rabbits, say, rather than undetached rabbit parts (see especially Loar 1995). Only phenomenally intentional states are referentially “scrutable” (if you will) in and of themselves: there is something about their phenomenal character that makes them about rabbits rather than undetached rabbit parts. Yet we must suppose that intentional content is always determinate, hence “scrutable.” So non-phenomenally intentional states must derive their determinate content (or derive the determinacy of their content) from phenomenally intentional states.24 Searle (1992) also appeals to content (in)determinacy in his argumentation, though of a different type. For Searle, the threat is that intentional content turn out to be indeterminate between different ways one and the same worldly target could be presented (e.g., Hesperus and Phosphorus). But the distinctive feature of phenomenal intentionality he appeals to is essentially the same “inherent determinacy” invoked by Loar.25 Another type of argument fitting the same schema starts from the premise of Intrinsic Subjectivity (McGinn 1988, Kriegel 2003b, Georgalis 2006, Frey this volume). As McGinn puts it, phenomenal intentionality is Janus-faced: it has an outward-looking face, which has to do with what it presents, but also an inward-looking face, to do with who it presents it to.26 The thought,

16

phenomenal intentionality

presumably, is that such Janus-faced content, and the intrinsic subjectivity it bestows, are required for basic intentionality. If so, the fact that only phenomenal intentionality features it entails that only phenomenal intentionality is basic. Fitting this into the above schema, we obtain the following: 1) Only phenomenal intentionality is intrinsically subjective; 2) Only intrinsically subjective intentionality is basic; therefore, 3) Only phenomenal intentionality is basic. Interestingly, it is harder to identify in the literature a similar argument from Narrowness. Perhaps this is because the starting point of such an argument would have to be the premise that mental states with basic intentionality must have narrow content, and in the present philosophical climate, this would seem grossly question-begging. This is an interesting predicament, given that it is precisely the internalist promise of phenomenal intentionality that attracts its proponents in the first place. One argument that can be interpreted as revolving around narrow content appeals to the phenomenon of intentional inexistence: the fact that every intentional state can occur in the absence of that which it is about. On one version of the argument (Kriegel 2007), it is claimed that only adverbially intentional states exhibit intentional inexistence in and of themselves, and then that only phenomenally intentional states are adverbial. An adverbial intentional state, being entirely non-relational, would clearly have narrow content, so this can be seen as an argument for a basicness thesis from a narrowness one. Bourget (2010) offers an importantly different kind of argument. He claims that holism about intentionality, according to which intentional contents are assigned in the first instance to networks of interrelated states rather than individual states, is true of non-phenomenally intentional states but not of phenomenally intentional ones. In consequence, non-phenomenal states partially derive their intentional properties from other non-phenomenal states. By contrast, since holism is false of phenomenally intentional states, the latter’s intentionality is underived: each phenomenally intentional state has its content irrespective of any relations to other states. Here the designated distinctive feature of phenomenal intentionality appears to be a sort of intentional atomism, and the resulting argument takes the following form: 1) Only phenomenal intentionality is atomistic; 2) Only atomistic intentionality is basic; therefore, 3) Only phenomenal intentionality is basic. As stressed earlier, I am not concerned here with the persuasiveness of any of these arguments, nor for that matter with the plausibility of their conclusions.27 The purpose of this section has been to review the kind of work that has already been done on the notion of phenomenal intentionality. This work has focused on four main questions: concerning the existence, scope, distinctiveness, and basicness of phenomenal intentionality.

The Phenomenal Intentionality Research Program

17

3. The Structure of a General Theory of Intentionality In this section I consider how a general theory of intentionality would look within PIRP. As we saw in §1, within NERP the general theory of intentionality can be seen as involving two chapters: the theory of source intentionality and the theory of non-source intentionality. The same general structure can be replicated within PIRP. Since within PIRP phenomenal intentionality is the source intentionality, within PIRP a general theory of intentionality would comprise two chapters, the theory of phenomenal intentionality and the theory of non-phenomenal intentionality.

3.1. The Theory of Phenomenal Intentionality Recall that according to PIRP, intentionality is injected into the world with the appearance of the right kind of phenomenal character. The fundamental task of the theory of phenomenal intentionality is to identify the kind of phenomenal character whose appearance injects intentionality into the world. One way to think of the challenge is as seeking the phenomenological signature of directedness (Kriegel 2011 Ch.3). In the nature of things, it would have to be a very subtle phenomenal feature, one that may well be introspectively unimposing but quite pervasive in our stream of consciousness.28 One suggestion might be culled from Strawson (2008). As noted in §2.3, like others Strawson maintains that, distinctively, phenomenal intentionality has determinate content in and of itself. But unlike others, he makes a proposal about what it is that endows phenomenal intentionality with this content determinacy: it is the phenomenology of taking. The notion of “taking” is present already in Chisholm (1957), for whom the act through which the mind takes something to be thus-and-so is the fundamental intentional act, with other intentional states being elaborations or modifications of a core act of taking. For Strawson, taking is a component or aspect of cognitive phenomenology. There is a subtle phenomenological feature, cognitive rather than sensory in nature, whereby a conscious experience takes something to be thus-and-so.29 This phenomenal taking, as we might call it, determines the exact intentional content of a phenomenally intentional state. Every phenomenally intentional state has a phenomenal character that involves phenomenal taking as a component, and it is in virtue of this phenomenal taking that the state is intentional. A different suggestion is due to Masrour (2008), who claims that it is only when a conscious experience exhibits a structural phenomenal feature he calls “objectual unity structure” (OUS) that it becomes intentional. Two phenomenal items are objectually unified in a perceptual experience, according to Masrour, just in case the experience presents them as belonging to the same

18

phenomenal intentionality

object. Consider R. C. James’s famous Gestalt-switch-provoking picture of the Dalmatian (reproduced, e.g., in Marr 1982). Once the switch occurs and the Dalmatian is “seen,” certain black spots in one’s visual experience become objectually unified while other spots are not so unified. The sensory aspect of the experience is the same before and after the switch, but after the switch the experience also exhibits a richer objectual-unity structure than before; it is only then that its phenomenal character intentionally represents a dog.30 I have considered two possible subtle phenomenal features that may serve as the “intentional spark”—the kind of phenomenal feature whose appearance injects intentionality into the world. Interestingly, it is often thought that all phenomenal character is intentional. If that is the case, and phenomenal taking or OUS is indeed the intentional spark, then it follows that phenomenal taking or OUS is a component in every phenomenal character, that is, that all phenomenal character involves taking or OUS. But the view that all phenomenal character is intentional may also suggest a third and competing account of the subtle phenomenal feature that brings intentionality into the world. This is the suggestion that the relevant feature is phenomenality itself. By “phenomenality itself” I simply mean the most general phenomenal genus, that of which all others are species (or perhaps the phenomenal determinable of which every other phenomenal property is a determinate). On this view, the intentional spark is generic phenomenality. Philosophers attracted to the idea that all phenomenality is intentional are often motivated by the claim that phenomenal character is diaphanous or transparent: when one introspects one’s current conscious experience, one is only aware of what the experience represents (Harman 1990). So a fourth suggestion could be that transparency is the intentionality-injecting phenomenal feature.31 This suggestion would be coextensive with the last one if indeed all phenomenality is intentional. But it may be that not all phenomenality is intentional. Thus, it is sometimes held that while the transparency of experience is plausible for perceptual phenomenology (and certainly cognitive phenomenology), it is quite less plausible for somatic and emotional phenomenology.32 If this is the case, then the transparency suggestion and the generic-phenomenality suggestion would not coextend after all. Of course, another option always available is to hold that the phenomenological signature of directedness is a sui generis phenomenal feature, inexplicable in terms of any other, simpler phenomenology. The view would be that the feature is exhibited by some mental states, such that when (and only when) it is exhibited, the relevant states are endowed with source intentionality.33 We may call this feature sui generis phenomenal directedness.34 I have considered five options for identifying the phenomenal features that inject intentionality into the world: phenomenal taking, unity structure, generic phenomenality, transparency, and sui generis phenomenal directedness.

The Phenomenal Intentionality Research Program

19

There are certainly other options worth exploring.35 The theory of phenomenal intentionality would attend to all those options—and to the reasons for preferring one over others. Also within the province of the theory of phenomenal intentionality would be work on the scope and distinctiveness of phenomenal intentionality, reviewed in §2.3 and §2.4. In addition, the theory of phenomenal intentionality would ultimately consider contingent properties of phenomenal intentionality that are not necessarily distinctive or special but do matter to its intentional function. For example, I suspect that in phenomenal intentionality the referential connection to the world works roughly as suggested in the descriptive theory of linguistic reference, rather than as suggested by direct-reference theories. If this is right, this would be of central importance to the theory of phenomenal intentionality, even if it does not constitute a distinctive feature of phenomenal intentionality.

3.2. The Theory of Non-Phenomenal Intentionality A comprehensive theory of intentionality would require not only an account of the phenomenal character that serves as the source of intentionality but also an account of how that phenomenal character outsources intentionality. That is, a full understanding of source intentionality does not constitute an understanding of intentionality as such. The other chapter of the (PIRP-ly) theory of intentionality concerns the nature of non-phenomenal intentionality and how it is grounded in phenomenal intentionality. Surprisingly, perhaps, several options for such an account can already be found in the existing literature (see Kriegel 2011 Ch.4 for review and discussion).36 One account, due to Searle (1992), may be called potentialism. Recall that according to Searle, non-phenomenally intentional states are intentional in virtue of being potentially phenomenal. Thus, an unconscious belief that p has p as its content in virtue of the fact that if it were conscious it would have the phenomenal-intentional content that p. At the same time, linguistic expressions, pictures, and so on, are surely not even potentially conscious. So their intentionality must be grounded in mental intentionality in some other way, perhaps through the Gricean mechanism of intention-based semantics (see Grice 1957). We may thus propose that Searle’s overall account of how non-phenomenal intentionality is grounded in phenomenal intentionality proceeds in two phases: the first extends intentionality from the narrow base of phenomenally intentional states to the realm of non-phenomenal mentality via counterfactuals about merely potential phenomenal character; the second extends it from the mental realm to the realm of non-mental intentionality via the right kind of speaker intentions.37 A second account may be called inferentialism (Loar 2003, Horgan and Graham forthcoming). On this view, non-phenomenally intentional states

20

phenomenal intentionality

inherit their intentionality through inferential connections they bear to phenomenally intentional states within the overall cognitive architecture of the mind. Thus, an unconscious belief that p is assigned just this content because it interacts with a conscious belief that p→q to bring about a conscious belief that q. Again, because non-mental items are not inferentially connected to mental ones, this would require supplementation in the form of an account of how non-mental intentionality is grounded in mental intentionality. A third account may be called interpretivism (Kriegel 2011 Ch.4). This view proposes a unified account of non-mental and mental non-phenomenal intentionality according to which all such intentionality is bestowed by phenomenally conscious interpretation. For example, an unconscious belief that p has p as its content in virtue of the fact that the best (conscious) interpretation of the system whose state it is would assign to the system a belief that p; and ditto for any linguistic expression or picture with the content that p. Of course, there is also a fourth option, which is to deny the existence of non-phenomenal intentionality altogether; we may call it eliminativism (Georgalis 2006, Strawson 2008).38 On this view, there is no problem of grounding non-phenomenal intentionality in phenomenal intentionality, because the former does not exist. Thus a full account of phenomenal intentionality, perhaps along the lines of one of the options mentioned in §3.1, would constitute a comprehensive theory of intentionality. As before, I am not concerned here with how plausible these accounts are; their portraits given here are too summary to allow serious discussion of their merits anyway. My discussion here serves only to point out some available options for accounting for non-phenomenal intentionality. In different combinations with the accounts of phenomenal intentionality covered in §3.1, they would constitute different comprehensive theories of intentionality. For example, one comprehensive theory might combine the phenomenal-taking view of phenomenal intentionality with potentialism about non-phenomenal intentionality; another the sui generis phenomenal directedness view with inferentialism about non-phenomenal intentionality; yet another the transparency view of phenomenal intentionality with interpretivism about non-phenomenal intentionality; and so on. Some combinations may be somehow more natural than others, but all would appear coherent and thus antecedently viable as comprehensive theories of intentionality in the PIRP genre.

Conclusion Work on intentionality within analytic philosophy of mind has been dominated for the past four decades by the idea that intentionality comes into the world when a certain type of tracking relation appears. In the last decade or so,

The Phenomenal Intentionality Research Program

21

however, another approach has been gaining momentum, suggesting that it is rather the appearance of a certain type of phenomenal character that injects intentionality into the world. From and around this newer approach a cluster of observations, contentions, and debates has emerged that has enlivened philosophical discussions of the mind, prompting philosophers to take a fresh look at the phenomenon of intentionality. The claim I have been making in this chapter is that the work pertaining to the cluster is ready to be consolidated into a well-defined, clearly organized, self-conscious research program.39

Notes 1. First accounts in this vein, such as Dennis Stampe’s (1977) and Fred Dretske’s (1981), were focused on broadly causal or informational relations. Under the influence of Ruth Millikan (1984, 1989, 1993) and David Papineau (1984, 1993), many philosophers have incorporated a teleological aspect into their account of the relevant relation (see Dretske 1988, McGinn 1989, Neander 1995). 2. I cannot make the case for this claim here, but my suspicion that it is only phenomenal intentionality which is concerned in those debates arises from the fact that unconscious mental states were not widely recognized before Freud’s work on repression and the postulation in cognitive science of myriad sub-personal unconscious states. Brentano (1874) himself argued explicitly for the co-extension of the mental, the conscious, and the intentional. So it is only natural to interpret him, and hence his students, as concerned with phenomenal intentionality. There is interesting historical research to be done here. 3. I am assuming here that linguistic intentionality is derived. Some proponents of NERP disagree (e.g., Millikan 1984), arguing [that the tracking property in virtue of which mental states represent non-derivatively is exhibited by linguistic expressions as well.] 4. There is a limit-case version of NERP, too. According to Dennett (1990), all intentionality is derived (none is underived), so in general the intentionality of derivatively intentional states must derive from each other’s, without being ultimately grounded in underived intentionality. 5. In principle, the second chapter could be exactly the same in a NERP-ly theory and a PIRP-ly theory. In particular, there is no reason to suppose that the derivation relation cannot be thought of in the same way by both. However, the nature of the sources of derivation is likely to constrain theorizing about the nature of the derivation, so the radical difference between the sources might very well lead to very different accounts of derived intentionality. 6. I am indebted to Terry Horgan for this distinction. 7. One limit-case epistemic possibility is that there is not really any spread at all, because the only intentionality there is is the intentionality of phenomenal consciousness itself (more on this in 2.4). 8. I will offer some remarks in the present section about different ways of potentially cashing out the theses as vaguely and generically stated, but more will be said on such matters in §2. 9. I am indebted to Terry Horgan for this formulation. 10. Some of these matters will be further pursued later in the chapter. 11. This issue can get complicated by matters terminological. Some, like Georgalis (2006), reserve “phenomenal” for the sensory kind of what-it’s-like-ness (more or less stipulatively), while yet recognizing the what-it’s-like-ness of thought. This looks to be probably just a verbal disagreement with those who claim there is a phenomenology of thought.

22

phenomenal intentionality

12. This is the metaphysical picture of phenomenal consciousness embraced by Chalmers (1996), as I understand him. Chalmers holds that phenomenal properties supervene on physicalistically naturalizable ones with nomological necessity, and that the pertinent supervenience laws are ontologically brute. He also regards his metaphysical picture as a form of naturalism about phenomenal consciousness, but not as a physicalistic version of naturalism. For clearer examples of naturalistic PIRP, see McGinn 1988 and Kriegel 2011 Ch.2. 13. Horgan and Tienson’s (2002) descriptions of certain conscious episodes can be seen as attempting to tap into an idea of this sort. 14. The present use of “sensory” and “non-sensory” pertains, in effect by stipulation, to relatively raw aspects of one’s sense-induced experience. It is a non-trivial matter to get a more formal specification of these terms going. 15. What that feature is is elucidated in terms of paradigmatic examples, which Masrour describes in detail but which I will not reproduce here. 16. Weaker theses might suggest that there is an individuative but non-proprietary cognitive phenomenology (say, because every belief is accompanied by imagery, and the imagery varies with the belief’s content); or that there is a proprietary but non-individuative phenomenology (say, because although beliefs in general are phenomenally different from hopes, suppositions, and other propositional attitudes, it is not the case that a belief that p is phenomenally different from a belief that q for any p and q). 17. For other arguments of this form, see Peacocke 1998, Horgan and Tienson 2002, Kriegel 2003a, and Pitt 2004. 18. Different authors have focused on different types of threat to content determinacy. Searle (1992) focuses on the threat to the intensionality of content; Loar (1995), Georgalis (2006), and Horgan and Graham (forthcoming) on the Quinean threat of referential inscrutability; Strawson (2008) on the threat once referred to as the “horizontal disjunction problem,” the thought that tracking theories of content cannot discriminate between different links in a causal chain leading up to an intentional state. 19. Here as elsewhere, there are difficult questions surrounding the explication of intrinsicality. The relevant notion for our present purposes is that of a state instantiating a certain property not in virtue of standing in a relation to some other state, but in and of itself. 20. For a fuller discussion of these issues, see Kriegel forthcoming. 21. An exception is McGinn (1988), who argues that conscious intentionality is wide, and its being both wide and in some way present to the subject is part of what makes it so mysterious. 22. As mentioned in a previous note, Georgalis would not state the thesis this way but would probably subscribe to it as interpreted by us. Perhaps it is also worth noting that Strawson does not use the term “phenomenal intentionality” but instead the term “experiential intentionality.” Th is is potentially better in avoiding confusion with the more sensory reading of the term “phenomenal.” 23. An even weaker thesis would require non-experiential intentional states to bear the requisite relations not necessarily to phenomenal-intentional states but also possibly to merely potentially phenomenal-intentional states. I am not familiar with any defense, nor defender, of this latter claim. 24. The argument thus proceeds as follows: 1) Only phenomenal intentionality has inherently scrutable content; 2) Only intentionality with inherently scrutable content is basic; therefore, 3) Only phenomenal intentionality is basic. Essentially the same argument for the special status of phenomenal intentionality is presented by Horgan and Graham (forthcoming), who develop the argument quite a bit further. 25. Th is kind of fined-grained content is what Searle (1992) refers to as “aspectual shape.” His argument could thus be construed as follows: 1) Only phenomenal intentionality is non-derivatively aspectually shaped; 2) Only non-derivatively aspectually shaped intentionality is basic; therefore, 3) Only phenomenal intentionality is basic. 26. Similarly, Georgalis (2006) insists that only conscious intentionality involves essentially not only the two-place relation x represents y but also the three-place relation x represents

The Phenomenal Intentionality Research Program

27.

28.

29.

30.

31.

32.

33.

34.

35. 36.

37.

23

y to z. However, Georgalis argues that there is an important difference between phenomenal and conscious intentionality and does not make the same claim about phenomenal intentionality. One objection, due to Shani (2008), is that although proponents of phenomenal intentionality (most notably Loar and Georgalis) are right that a full theory of intentionality must incorporate a certain fi rst-person notion, there are ways to do so without appealing to phenomenal intentionality. In particular, Shani argues that the central aspect of fi rst-person intentionality would be a notion of idiosyncratic perspective, but such a perspective can be accounted for in naturalistic, indeed broadly teleological, terms that appeal solely to unconscious and non-phenomenal phenomena. These two may not be unrelated: the most pervasive features of phenomenology are often hardest to spot, because hardest to isolate (there may be no vivid contrasts between presence and absence). Note that if taking is a cognitive phenomenal feature, and is necessary for phenomenal intentionality, then all phenomenal intentionality is cognitive, at least in the sense that phenomenal character must have at least one cognitive element in order to be intentional. Th is may raise the objection that certain animals and neonates may be robbed of intentionality. There are probably reasonable responses to this objection, but a full discussion would take us too far afield. (Moreover, some of the moves here would parallel relatively familiar moves from discussions in the theory of animal consciousness and the problem it presents for higher-order thought theories.) The pre-switch visual experience is already intentional: it represents an externally situated pattern of black and white splotches. Still, it may be that there must be some unity structure present in order for the visual phenomenology to become intentional. It may be, too, that virtually all visual phenomenology already has some unity structure and is already intentional. Th is requires that we think of transparency as a positive phenomenal feature, exhibited by some (or all) experiences, rather than as the absence of certain phenomenal features, as it sometimes is. Th is is not to say that the transparency thesis has not been defended for such experiences as well (Shoemaker 1994, Tye 2000). But it is roundly accepted that such defenses tend to be more theoretical and rely less on the immediate intuitive conviction that the thesis enjoys in the perceptual domain. It would be a substantive question at this point whether all phenomenal states have this phenomenal feature. (It is, of course, coherent to hold either that only some do or that all do.) Which way one goes on this question will then have implications for any possible coextension between the present suggestion and the generic-phenomenality suggestion. Methodologically, it is probably preferable to seek a more explanatory account of the kind of phenomenal character we are seeking and settle for an account of it in terms of sui generis phenomenal directedness only as a last resort. See Kriegel 2011 Ch.3 for discussion of two more suggestions, one of which is developed from Frey’s (this volume) discussion. Gricean intention-based semantics, although the leading approach within NERP, cannot be straightforwardly applied within PIRP because the relevant intentions Grice identifies are presumably unconscious. It is very possible, however, to divide the theory of derived intentionality within PIRP into two parts, the fi rst concerned with the derivation of non-mental intentionality from non-phenomenal mental intentionality via the Gricean mechanism and the second concerned with the derivation of non-phenomenal mental intentionality from phenomenal intentionality via some other mechanism. (For more discussion, see Kriegel 2011.) Note that these intentions, at least those proposed by Grice, are typically unconscious. So the derivation of linguistic intentionality from conscious intentionality would have to proceed in two steps: it would not be possible for the former to derive directly from the latter.

24

phenomenal intentionality

38. It is also possible to interpret Georgalis (2006) and Pitt (Ms) as holding this position. Georgalis maintains that there is no unconscious intentionality, and although he allows that some conscious intentionality is non-phenomenal, this is mainly because, as remarked earlier, his notion of phenomenality is narrower and applies only to sensory quality. Pitt, by contrast, allows that there is unconscious intentionality but proposes that such intentionality is constituted by an unconscious phenomenality. Thus although there is such a thing as unconscious intentionality, there is no non-phenomenal intentionality. 39. For comments on a previous draft, I would like to thank Ben Blumson, Curtis Brown, Tamar Szabó Gendler, Kristin Hurlburt, Adam Pautz, Galen Strawson, and especially Brie Gertler and Susanna Siegel.

References Abell, C . 2005. “Pictorial Implicature.” Journal of Aesthetics and Art Criticism 63: 55–66. Blumson, B. 2006. Resemblance and Representation. PhD Dissertation, Australian National University. Bourget, D. 2010. “Consciousness Is Underived Intentionality.” Noûs 44: 32–58. Brentano, F. 1874. Psychology from Empirical Standpoint. Edited by O. Kraus. English edition L. L. McAlister, ed. Translated by A. C. Rancurello, D. B. Terrell, and L. L. McAlister. London: Routledge and Kegan Paul, 1973. Chalmers, D. J. 1996. The Conscious Mind. Oxford: Oxford University Press. Chisholm, R . 1957. Perceiving: A Philosophical Study. Ithaca: Cornell University Press. Cummins, R . 1979. “Intention, Meaning, and Truth Conditions.” Philosophical Studies 35: 345–360. Davies, M. 1995. “Consciousness and the Varieties of Aboutness.” In C. Macdonald and G. Macdonald (eds.), Connectionism: Debates on Psychological Explanation (Vol. II). Oxford: Blackwell. Dennett, D.C . 1990. “The Myth of Original Intentionality.” In K. A. Mohyeldin Said, W. H. Newton-Smith, R. Viale, and K. V. Wilkes (eds.), Modeling the Mind. Oxford: Oxford University Press. Dretske, F. I. 1981. Knowledge and the Flow of Information. Oxford: Clarendon. Dretske, F. I. 1988. Explaining Behavior. Cambridge, MA : MIT Press. Fodor, J. A . 1984. “Semantics, Wisconsin Style.” Synthese 59: 231–250. Fodor, J. A . 1990. A Theory of Content and Other Essays. Cambridge MA : MIT Press. Georgalis, N. 2006. The Primacy of the Subjective. Cambridge MA : MIT Press. Goldman, A . 1993. “The Psychology of Folk Psychology.” Behavioral and Brain Sciences 16: 15–28. Grice, H. P. 1957. “Meaning.” Philosophical Review 66: 377–388. Grice, H. P. 1969. “Utterer’s Meaning and Intention.” Philosophical Review 68: 147–177. Harman, G. 1990. “The Intrinsic Quality of Experience.” Philosophical Perspectives 4: 31–52. Horgan, T. and G. Graham Forthcoming. “Phenomenal Intentionality and Content Determinacy.” In R. Schantz, Prospects for Meaning. Amsterdam: de Gruyter. Horgan, T. and U. Kriegel 2008. “Phenomenal Intentionality Meets the Extended Mind.” Monist 91: 353–380 Horgan, T. and J. Tienson 2002. “The Intentionality of Phenomenology and the Phenomenology of Intentionality.” In D. J. Chalmers (ed.), Philosophy of Mind: Classical and Contemporary Readings. Oxford: Oxford University Press. Horgan, T., J. Tienson, and G. Graham 2004. “Phenomenal Intentionality and the Brain in a Vat.” In R. Schantz (ed.), The Externalist Challenge: New Studies on Cognition and Intentionality. Amsterdam: de Gruyter. Horgan, T., J. Tienson, and G. Graham 2006. “Internal-World Skepticism and Mental Self-Presentation.” In U. Kriegel and K. W. Williford (eds.), Self-Representational Approaches to Consciousness. Cambridge MA : MIT Press.

The Phenomenal Intentionality Research Program

25

Husserl, E. 1901. Logical Investigations, Vol. 2. Trans. J.N. Findlay. London: Routledge, 2001. Jackson, F. 1998. From Metaphysics to Ethics: A Defence of Conceptual Analysis. Cambridge: Cambridge University Press. Kelly, S. D. 2004. “On Seeing Th ings in Merleau-Ponty.” In T. Carmon (ed.), Cambridge Companion to Merleau-Ponty. Cambridge: Cambridge University Press. Kriegel, U. 2003a. “Consciousness as Sensory Quality and as Implicit Self-Awareness.” Phenomenology and the Cognitive Sciences 2: 1–26. Kriegel, U. 2003b. “Is Intentionality Dependent upon Consciousness?” Philosophical Studies 116: 271–307. Kriegel, U. 2005. “Naturalizing Subjective Character.” Philosophy and Phenomenological Research 71: 23–57. Kriegel, U. 2007. “Intentional Inexistence and Phenomenal Intentionality.” Philosophical Perspectives 21: 307–340. Kriegel, U. 2009. Subjective Consciousness: A Self-Representational Theory. Oxford: Oxford University Press. Kriegel, U. 2011. The Sources of Intentionality. Oxford: Oxford University Press. Kriegel, U. Forthcoming. “Brentano’s Most Striking Thesis: No Representation without Self-Representation.” In D. Fisette and G. Fréchette (eds.), Themes from Brentano. Kriegel, U. Ms. “Intentionality: Realization versus Essence.” Levine, J. 2001. Purple Haze. Oxford: Oxford University Press. Loar, B. 1987. “Subjective Intentionality.” Philosophical Topics 15: 89–124. Loar, B. 1995. “Reference from the First-Person Perspective.” Philosophical Issues 6: 53–72. Loar, B. 2003. “Phenomenal Intentionality as the Basis for Mental Content.” In M. Hahn and B. Ramberg (eds.), Reflections and Replies: Essays on the Philosophy of Tyler Burge. Cambridge MA : MIT Press. Lycan, W. G. 2008. “Phenomenal Intentionalities.” American Philosophical Quarterly 45: 233–252. McGinn, C . 1988. “Consciousness and Content.” Proceedings of the British Academy 76: 219–239. Reprinted in N. J. Block , O. Flanagan, and G. Gü zeldere (eds.), The Nature of Consciousness: Philosophical Debates. Cambridge, MA : MIT Press, 1997. McGinn, V. 1989. Mental Content. Oxford: Blackwell. Marr, D. 1982. Vision. San Francisco: W. H. Freeman. Masrour, F. 2008. Structuralism: In Defense of a Kantian Account of Perceptual Experience. PhD Dissertation, University of Arizona. Meinong, A. 1904. “The Theory of Objects,” in Roderick M. Chisholm (ed.), Realism and the Background of Phenomenology, Glencoe, IL: Free Press, 1960. Millikan, R. G. 1984. Language, Thought, and Other Biological Categories. Cambridge, MA: MIT Press. Millikan, R. G. 1989. “Biosemantics.” Journal of Philosophy 86: 281–297. Millikan, R. G. 1993. White Queen Psychology and Other Essays for Alice. Cambridge, MA : MIT Press. Moore, G. E . 1953. “Propositions.” In his Some Main Problems of Philosophy. Oxford: Routledge. Neander, K . 1995. “Misrepresenting and Malfunctioning.” Philosophical Studies 79: 109–141. Noë, A . 2004. Action in Perception. Cambridge MA : MIT Press. Papineau, D. 1993. Philosophical Naturalism. Oxford: Blackwell. Peacocke, C . 1998. “Conscious Attitudes, Attention, and Self-Knowledge.” In C. Wright, B. C. Smith, and C. Macdonald (eds.), Knowing Our Own Minds. 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. Pitt, D. Ms. “Unconscious Intentionality.” http://www.calstatela.edu/faculty/dpitt/Unconscious %20Intentionality.pdf Schiffer, S. 1982. “Intention Based Semantics.” Notre Dame Journal of Formal Logic 23: 119–159. Searle, J. R . 1983. Intentionality: An Essay in the Philosophy of Mind. Oxford: Blackwell. Searle, J. R . 1991. “Consciousness, Unconsciousness, and Intentionality.” Philosophical Issues 1: 45–66.

26

phenomenal intentionality

Searle, J. R . 1992. The Rediscovery of Mind. Cambridge MA : MIT Press. Shani, I. 2008. “Against Consciousness Chauvinism.” The Monist 91(2): 294–323. Shoemaker, S. 1994. “Phenomenal Character.” Nous 28: 21–38 Siegel, S. 2005. “The Contents of Perception.” Stanford Encyclopedia of Philosophy. Siegel, S. 2006a. “Which Properties Are Represented in Perception?” In T. Gendler Szabo and J. Hawthorne (eds.), Perceptual Experience. Oxford: Oxford University Press. Siegel, S. 2006b. “Subject and Object in the Contents of Visual Experience.” Philosophical Review 115: 355–388. Siewert, C. P. 1998. The Significance of Consciousness. Princeton NJ: Princeton University Press. Stampe D. 1977. “Towards a Causal Theory of Linguistic Representation.” Midwest Studies in Philosophy 2: 42–63. Strawson, G. 1994. Mental Reality. Cambridge, MA : MIT Press. Strawson, G. 2008. “Real Intentionality 3: Why Intentionality Entails Consciousness.” In his Real Materialism and Other Essays. Oxford: Oxford University Press. Twardowski, K. 1894. On the Content and Object of Presentations. Trans. R. Grossmann. The Hague: Martinus Nijhoff, 1977. Tye, M. 2000. Consciousness, Color, and Content. Cambridge, MA : MIT Press.

2

The Access Problem M ic h e l l e Mon tagu e

1. Introduction We perceive and think about many things: we enter into many mental states which involve intentional relations with particular objects. But how do we do it? How do we achieve access to the things with which we stand in mental intentional relations? When we think about or perceive some particular thing, what makes it the case that we have that very thing in mind? What mechanism determines what a thought or perception is of? I call this the “access problem.” For the purposes of this chapter I will restrict my attention to perceptions and thoughts of or about material objects other than oneself, and I will focus sometimes on thought, sometimes on perception. I will use the terms “internal” and “external” to mark off two common and opposed approaches to the access problem. There is, unsurprisingly, a connection between my use of these terms and the currently popular contrast between “internalist” and “externalist” conceptions of the mental, but the words “internalist” and “externalist” have been used in too many different ways in philosophy of mind to be stable. It will often be possible to substitute “externalist” for “external” in what follows, but this will not work nearly so well in the case of “internalist” and “internal.” As will become clear, the questions raised by the access problem—the problem of giving a philosophical characterization of the mechanism that determines which particular object a subject is perceiving or thinking of on a particular occasion—cut across many of the standard debates concerning internalism and externalism about mental content. The fundamental idea behind the internal approach to the access problem is that thinking of an object essentially involves conceiving of it in some manner, or characterizing it in some fashion, and that reference to the particular manner involved is essential for determining which object is being thought

27

28

phenomenal intentionality

of. It is important to distinguish between these two claims, which we may rephrase as [1] A subject, S, cannot think of an object without thinking of it in some way and [2] It is not possible to settle the question of which object (if any) S is thinking about without reference to the particular way of thinking of an object that is involved in the thought. [1] is a fairly weak claim, acceptable to most theories of intentionality, including many external(ist) approaches. It does not require that reference to one’s way of thinking of an object play an essential role in an adequate account of what determines which object one is thinking of, that is, an adequate account of how one has access to that object. [2] does require this. [2] requires that a particular way of thinking must be in place in order to count a thought or perception as being about a particular object. It imposes what one might call a “particular-way-of-thinking” condition on access, a “particular-way” condition, for short. This is an internal condition on thinking of or perceiving a particular object in the current sense, because it requires something over and above the physical object and a specification of the external relations that the subject stands in with respect to that object. External (and indeed externalist) theories of access seek to explain the phenomenon of having an object in mind solely in terms of external relations (causal or historical) that hold between a thinker’s relevant mental state and the world. According to the external approach, which object is being thought of can be determined without any appeal to the specific way in which one may be thinking of it (see for example, Donnellan 1966, 1974; Fitch 1990; and Dretske 1995). Fitch, for example, claims that “intrinsic features of a given thought have little or no bearing on the issue of how a thought refers to a particular object” (1990: 676). Dretske concurs: What determines the reference of a de re mode of representation (the object it is a representation of) is not how it is represented, but a certain external causal or contextual relation I will designate C. There is nothing in the content of the representation, nothing the representation says, which makes it about this object rather than that object or no object at all. De re modes of representation have their reference determined contextually, by the relation I am here calling C.1 A standard externalist causal theory seeks to solve the access problem by asserting that my thought is about John simply because he stands in the

The Access Problem

29

relevant position among the causes leading to whatever internal state it is in me that is a (the) candidate for being my John-thought. Gareth Evans calls this the “Photograph Model,” according to which “a mental state can represent a particular object simply in virtue of that object’s playing a suitable role in its causal ancestry” (1982: 81).2 Such a theory may accept [1]—that thinking of an object requires thinking of it in some way or other. What is crucial is its rejection of [2], its claim that there is no internal “particular-way” condition on access—that one does not need to consider which particular way of thinking of an object is in play in order to determine either which object is being thought of, or whether any object is being thought of at all. In this chapter I argue that there is an ineliminable internal “particularway” condition on access in addition to whatever external conditions there may be.3 The question is then, what is the nature of this “particular-way” condition? I will begin with a consideration of “Russell’s Principle” as articulated by Gareth Evans. It is uncontroversial that Evans places an external condition on thoughts about particular objects: there is a kind of thought we sometimes have, typically expressed in the form ‘This F is G’, and we may aim to have a thought of this kind when, in virtue of the absence of any appropriate object, there is no such thought to be had. (1982: 46) And it is no less clear that his version of Russell’s Principle places an internal condition on such thoughts, in holding that one cannot be said to be thinking about a particular object unless one knows which object one is thinking about, in the sense that one can distinguish it from all other objects. This is clearly an internal condition on access, as defined earlier, because it requires that something more than causal relations between the subject and the object be in place, that is, that the subject knows which object he is thinking of. I am going to argue that Russell’s Principle does not in fact succeed as an internal constraint on access. I will not, however, reject it for the usual reasons, which have to do with misrepresentation of the locations of objects, the possibility of subjective qualitative duplication, and so on.4 I will focus first on the particular case of seeing something, and argue that Russell’s Principle does not provide a sufficiently robust internal condition on access in cases of seeing. Then I will raise the more general question of whether it can ever provide a sufficiently robust internal condition on genuinely having an object in mind. I will call the internal view I defend the “matching view.” It states that for a perceptual experience to be about an object, there must be a certain degree of match between the properties an object has and the properties the experience represents the object as having. One distinctive feature of the matching view, as opposed to Russell’s Principle, is the emphasis it places on the

30

phenomenal intentionality

phenomenological or qualitative features of experience. It states that certain phenomenological features of an experience need to be in place in order for it to qualify as thinking of or perceiving a particular object. I construe phenomenological features of experience as “narrow.” So I reject “phenomenal externalism” as defended by Dretske 1996 and Lycan 2001. According to a narrow construal of the phenomenological features of experience, “phenomenal duplicates” can have phenomenologically identical experience despite having radically different environments. My claim, then, is that one has to consider phenomenology narrowly construed, when determining the object of a thought or perception.5 You may wonder how this can be right. For if reference to phenomenology is necessary when establishing which particular object a thought is about, then we can never establish this in the case of non-conscious thoughts, or indeed any dispositional states like beliefs and desires, since they have no phenomenology. This seems an unacceptable consequence of my proposal: surely non-conscious mental states or occurrences like thoughts and beliefs can have determinate intentional objects? In fact, though, this is not an objection to the present proposal. On the contrary: it is a virtue of the present proposal that it reveals this as a problem, and sets us the task of explaining how non-conscious mental states can have determinate objects. The idea that non-conscious mental occurrences or dispositional mental states can be determinately about particular objects is standardly taken for granted, but it stands in need of defence. Among those who have clearly recognized this problem, and offered solutions, are Searle 1992: ch. 7; Strawson 1994: §7.8, 2008: §6; and Kriegel 2011. It doesn’t arise for someone like Dennett, because of his anti-realism about mental states. It is, however, a topic for another time. In this chapter I will restrict attention to conscious mental episodes.

2. Russell’s Principle According to Russell “it is scarcely conceivable that we can make a judgment or entertain a supposition without knowing what it is that we are judging or supposing about” (1912: 32). In Evans’s words, “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” (1982: 65). This plainly imposes an internal condition on our access to objects in the sense defined earlier. To know which object one is thinking of, one need not know exactly what sort of object it is: it is plausible that I can think about the electron microscope in my school’s science lab without knowing that it is an electron microscope.6 What is required to satisfy the principle, as both Russell and Evans acknowledge, is rather

The Access Problem

31

“discriminating knowledge” or a “discriminating conception” of the relevant object: a capacity to uniquely identify the object, to distinguish it from all other things. Evans proposes that there are three ways in which one may be able to do this, three distinct kinds of modes of identification of objects in virtue of which one may possess the capacity to uniquely identify an object. One may [1] have knowledge of (uniquely) distinguishing facts about it; one may [2] be in a position to identify the object demonstratively; or one may [3] have a recognitional capacity, a capacity to recognize the object so as to be able to pick it out from all other objects.7 Evans explicitly links the concept of a mode of identification to a subject’s awareness. It is because a mode of identification essentially involves S’s being in a certain state of awareness that S is in a position to think about a particular object. And given that Evans distinguishes between one’s awareness and the modes of identification that it involves on the one hand, and the relevant external object and its properties on the other hand, it seems plain, again, that he is thinking of the character of the state of awareness in an essentially internalist way, at least in part. If Evans is right, then, we have to study modes of identification, internal features of thought, in any full investigation of the mechanisms of object-access, the mechanisms that determine which object a thought is of. One cannot remain wholly outside of a thought to explain how it manages to be about a particular object. In what follows I will focus on Evans’s account of demonstrative identification and the way in which it is supposed to guarantee the unique identification of an object. Although I reject Evans’s account, I will use it to articulate my own view, while at the same time showing the implausibility of the external solution to the access problem.

3. Perceptual Identification To begin to see how the “matching view” answers the access problem, I will first consider an apparent disanaology between the phenomenon of seeing and other sensory cases, in particular, auditory and olfactory cases. In the cases of audition and olfaction, but not initially in the case of seeing, when there is too much mismatch between the properties represented by an experience of an object x and the properties object x has, we are reluctant to say that “perceptual contact” is made, where by “perceptual contact with x” I mean seeing, hearing, tasting, touching, or smelling x. Since our theories tend to take visual cases as central and generalize from there, the importance of mismatch for whether perceptual contact is made is overlooked. Once the relevance of mismatch has been illustrated, I will go on to argue that our intuitions about seeing can be brought into line with these other sensory cases. Suppose for simplicity that colors are objective properties of objects, and consider Maria, who is looking at a red object in a normally attentive way and

32

phenomenal intentionality

is in no other sort of contact with it, sensory or otherwise, so that the only information link she has to the object is visual. Suppose that the mechanisms of visual experience operate normally except for the fact that she has, for whatever reason, and quite exceptionally, an experience of something green, on looking at the object. Question 1: Does she see the object’s color? Presumably not—not at all. Question 2: Can she think about the object’s color specifically on the basis of the particular character of her visual experience? Presumably not— not at all. The failure of match rules this out.8 Suppose now that the object is round and that she sees it, for whatever reason, and quite exceptionally, as square. Question 3: Does she see the object’s shape? Presumably not—not at all. Question 4: Can she think about the object’s shape specifically on the basis of her visual experience, in the sense just explained? Presumably not—not at all. Again this is explained by failure of match. Now combine the cases of color and shape. Question 5: Does she see the object? Most people will say Yes, even though they may agree that she can see neither the object’s color nor its shape, and cannot even think about its visually perceptible properties specifically on the basis of the character of her visual experience. Question 6: Can she think about the object on the basis of her visual experience? Again most philosophers will presumably answer Yes. For if she can see it, she can surely think about it.9 Now consider sounds on the same basis, that is, supposing that objects’ sounds are objective properties of objects. Suppose Maria, listening on her headphones, has, quite exceptionally, an experience as of Jimi Hendrix’s “Are You Experienced?” when the sound waves from a performance of the “Moonlight” sonata impact on her ears (she is in no other sort of contact, sensory or otherwise). Question 7 (parallel to Question 1 in the visual case): Does she hear the music’s sound? Again the answer is No. Question 8 (parallel to Question 2): Can she think about the music’s sound specifically on the basis of the character of her auditory experience, in the sense of “specifically on the basis of” just explained? Again the natural answer is No. Now consider Question 9 (parallel to Question 5 in the visual case): Does she hear the piece of music? The natural answer appears to be No—she just isn’t hearing the music, although her auditory experience is being caused by sound waves stemming from the playing of the “Moonlight” sonata—whereas the natural answer to Question 5 in the visual case (Does she see the object?) seemed to be Yes. So the visual and auditory cases appear to diverge. But with Question 10 (parallel to Question 6): Can she think about the music on the basis of her auditory experience?, the two cases may appear to fall back into step. For many, I think, will want to answer Yes to this question. Many take the visual case as central and reach for a theory based on that case. Thus those who answer Yes to the visual Question 5 may decide to answer Yes to its auditory parallel Question 9, on the grounds of general

The Access Problem

33

theoretical smoothness, and in spite of the seeming naturalness of answering No (they may also produce other considerations). But Thomas Reid was right to point out in his Inquiry that philosophers are far too ready to restrict their attention to the visual case, and that the other senses deserve separate attention. Now consider smells on the same basis, that is, supposing that objects’ smells are objective properties of objects. Suppose Maria has, quite exceptionally, an experience as of smelling freshly cut apples when the molecules from a freshly baked loaf of bread reach her nose (she is in no other sort of contact with it with it, sensory or otherwise). Question 11 (parallel to Question 1 in the visual case): Does she smell the bread’s smell? Again the answer is No. Question 12 (parallel to Question 2): Can she think about the bread’s smell specifically on the basis of the character of her olfactory experience, in the sense of “specifically on the basis of” just explained? Again the natural answer is No. Now consider Question 13 (parallel to Question 5 in the visual case): Does she smell the bread? or Is it the bread that Maria smells? The natural answer appears to be No—she just isn’t smelling the bread, although her olfactory experience is being caused by molecules stemming from the bread—whereas the natural answer to Question 5 in the visual case (Does she see the object?) seemed to be Yes. So the visual and olfactory cases also appear to diverge. But with Question 14 (parallel to Question 6): Can she think about the bread on the basis of her olfactory experience?, again the cases may appear to fall back into step. For most, I think, will want to answer Yes to this question. What the auditory and olfactory cases suggest is that if one’s experience of something x is wildly inaccurate, then that inaccuracy can have the consequence that it is incorrect to say that one is in perceptual contact with that something. Why do our intuitions seem to diverge with respect to the seeing case? One source of the disanalogy may be that in the visual case, although Maria does not see the red round object’s shape and color, she does see where it is located. And since it may seem that location properties play a role in individuating physical objects, we have the intuition that Maria can see the object in virtue of correctly identifying its location. With this background, let’s reconsider the phenomenon of seeing with the following two cases. [1] One is in causal, and sensory, and indeed visual, contact with a rock, but one’s visual-experience-based conception of it is inaccurate because of fog conditions—it appears bigger and of a different color than it really is. One can nevertheless locate and track the rock in spite of one’s inaccurate conception of it.

34

phenomenal intentionality

Does one see the rock? Most would say yes. [2] One is in causal, sensory, and indeed visual contact with a garden shed, but when one looks at it one has—due to a disorder in one’s visual system, or a distortion of the atmosphere in which the light-waves that would have reached one from the shed do reach one, but profoundly rearranged—an experience as of a pink elephant. In this case, too, one can locate and track the shed, in spite one’s inaccurate conception of it.10 Does one see the shed? (I am assuming that there is no other sort of link, causal or historical, between one and the shed.) My intuition is that one does not, because one’s apprehension or representation of it is simply too inaccurate. Visual contact with an object, in the present sense, is not enough—I suggest—to guarantee that one sees it, that one is in perceptual contact with it—even if standing in this relation makes it possible for one to locate and track it. To see an object, one’s experience (representation) of the object must not be wildly inaccurate. It is arguable, further, that if one cannot see the object, one cannot really think about it either (assuming one has no other access to it). But what do I mean by “visual contact”? For present purposes I’ll begin by defining it as follows. Visual contact is causal and sensory contact of a sort that [a] involves the impact of light on a sensory organ and [b] gives rise to experiences of color and shape of a kind that can be sufficiently indicated by saying that they are of the same sort qualitatively speaking as experiences of the kind we call “visual.”11 If a peculiarity of one’s brain has the consequence that the only sensory experiences one has when light waves impact on one’s eyes are olfactory, then one is not having visual experience. If aliens or robots can master our shape and color words on the basis of discriminating differences between sounds that they experience as a result of the impact of light waves on their sensory organs, then they do not have visual experience, nor do they see. Nor do bats see, given how we ordinarily conceive of their experience, however well they can negotiate their environment. Seeing, then, is not only an essentially conscious experience,12 it also essentially involves a certain specific kind of qualitative character, and my question is this: what is the relationship between the qualitative character of a particular act of seeing and what is seen?13 What I am suggesting with case [2] is that being in visual contact with an object does not entail that one sees it, that is, that one is in perceptual contact with it. That is, one can be in visual contact with an object, the object can be causing one to have a certain set of visual experiences, but that kind of contact is not yet enough for seeing. Is one subject to an illusion, in case [2]? The word “illusion” is defined in different ways. According to one definition, an illusion is a perception of an object, so that perceptual contact is made, but the content of one’s perception

The Access Problem

35

is a markedly inaccurate depiction of the object perceived, in some respect or respects—an object is perceived (seen, heard, felt, etc.), but some of the object’s properties appear other than they really are. In denying that one sees the shed at all, in case [2], I am denying that the case can be characterized as a case of illusion. Perceptual contact has not been made, one is not actually seeing the shed because there is simply too much inaccuracy with respect to the shed’s properties. So far I have only made a prima facie case for the matching view. Hopefully the case is strong enough so that anyone who thinks hard about these questions should at some point feel the pull of the intuitions behind them.

4. Perceptual Contact: How Wrong Can One Be? How wrong can one be before perceptual contact fails? The question cannot receive a sharp answer, but it seems that Evans must disagree with my intuition about the shed case—irrespective of the degree of inaccuracy. For he takes it that one can satisfy Russell’s Principle with respect to a particular object, given one’s present sensory contact with it, simply in virtue of being able to locate and track that object on the basis of that contact: our conditions for demonstrative identification do not require that the subject’s information link be functioning well—so long as it provides an effective route to the object. He can misperceive its colour, or its shape, or get altogether quite a wrong view of the thing, while still having a perfectly clear Idea of which thing is in question.14 Clearly Evans takes it that one can think about the object, in this case. Presumably, then, he must also think that one can see the object, given that one is able to think about it solely and specifically on the basis of one’s present visual contact with it. Evans’s focus on locating and tracking derives in part from his attempt to give a detailed explanation of what having discriminating knowledge of a physical object via perception amounts to. He argues that discriminating knowledge of a physical object requires that one have what he calls a “fundamental idea” of that object. To have a fundamental idea of an object is, centrally, to grasp its “fundamental ground of difference,” where its fundamental ground of difference is what differentiates it from all other objects of its own kind and all other objects. The fundamental ground of difference for physical objects, according to Evans, is their location in space.15 So to have a fundamental idea of a physical object is, by definition, to have discriminating knowledge of that object. The tracking element in Evans’s account records his commitment to

36

phenomenal intentionality

the idea that demonstrative identification is often dependent on a continuing information-link with an object. Some have rejected Evans’s theory on the grounds that one can demonstratively identify an object even if one gets the location wrong. Campbell, for example, says that Evans’s view “has no plausibility at all. It implies that you cannot even think demonstratively about an object you can see perfectly well, because you are subject to some illusion about its location” (2002: 112). But although such illusion cases may show that knowledge of location is not a necessary condition for demonstrative identification, they do not threaten the position I ultimately defend in this chapter—the matching view, according to which one must represent a sufficient number of the object’s properties correctly in order for it to be true that one sees it. Suppose that one’s visual experience is of nothing but flashing stars and circles, when one is in visual contact with a dog. Those who hold that the ability to track and locate an object on the basis of visual contact is a sufficient condition of seeing the object will presumably allow that one can be said to see the dog, so long as one can locate and track it via the movement of the flashing stars and circles.16 It seems to me, however, that one does not see the dog. If one rejects the view that the tracking/locating ability is necessary on the basis of considerations advanced by Campbell, then one presumably endorses the matching view. For if one sees the object “perfectly well” despite misrepresenting its location, one must presumably be representing some of its visible properties correctly. So far, Evans seems to be committed to accepting that one sees the shed and the dog, in the situations described. In the Appendix to chapter 6 of The Varieties of Reference, however, he writes if one sees the object as a woman . . . and if what one sees is not a woman (or, at least, a person), then the mistake deprives the attempted thought of content. (There would be a question about how seriously mistaken one has to be for this to be the consequence. But it seems that we do not attach the same importance to misperceiving a man as a woman, as we should attach to misperceiving, say, a stone as a woman. In the former case, a thought involving a demonstrative identification expressible by “that woman” is merely incorrect; in the latter case there is some inclination to say that the attempted thought lacks a content.) (1982: 196–7, my emphasis) This case is similar to the case of the shed and dog cases, and suggests that Evans might after all be disinclined to say that one was able to think about the shed and the dog in those situations. But given that it is plausible to say that one must be able to think about something if one can see it (if one is capable

The Access Problem

37

of thinking at all), and given that the shed and dog cases are cases in which one has no other causal or historical or sensory link with the object, so that there is no question of one’s being able to think about it unless one can see it, it appears to follow that if one cannot think about it then one cannot see it either. Evans does not explicitly say that one cannot see the object (the stone), in the quoted passage, only that any attempted thought about it would lack a content. To this extent he leaves open the possibility that one might be said to see it although one was unable to have a contentful thought about it. But this seems most implausible, for the reason just given. The notion of “seeing as” makes us think specifically of the deployment of concepts in perceptual experience, and it may be that the only point that Evans is intending to make, in describing this particular failure to meet a necessary internal condition on access, is a point about concept deployment—a point about the potentially radical interference effects that can arise from inappropriate concept deployment. It is, however, arguable that the point has a wider scope, and that failures of this general kind can arise even in the absence of inappropriate concept deployment. Consider two takes on the stone/woman case (I am assuming that the stone doesn’t look at all like a woman). According to the first [a] one possesses the concept woman, and is mistakenly representing properties associated with being a woman, when in visual/causal contact with a stone, in such a way as to be having experience that one takes to be experience of seeing (and thinking about) a woman. It is the incorrect deployment of the concept woman, or, more broadly, of concepts associated with woman, that prevents one from being able to have a genuine thought about, and so, presumably, and a fortiori, genuinely see, the stone. According to the second [b] whether or not one possesses the actual natural-kind concept woman, one’s experience mistakenly represents the stone as having “womanly” properties (shape and color properties, say) in a way that one might characterize by saying that if the visual content of one’s experience were transferred to a sheet of paper, the resulting picture would be judged to be a picture of a woman. The first interpretation of Evans is no doubt the most likely, but the second is also worth recording, even while it confl icts with his “official” view. My intuition is that one does not see the stone in either case, because the ground-floor reason that one does not see the stone is not that one’s experience involves specifically conceptual error, but simply that it involves too much error. Evans

38

phenomenal intentionality

might perhaps agree, given that he says that “there is some degree of incorrectness in a subject’s conception of an object that makes it pointless to ascribe thoughts about it to him” (p. 134). This, however, requires taking the word “conception” in a wide sense, and has to be set against his claim that one “can misperceive [an object’s] colour, or its shape, or get altogether quite a wrong view of the thing, while still having a perfectly clear Idea of which thing is in question” (p. 179). So I will now return to his “official” theory, and to the suggestion that the case of the shed (or dog) puts it in doubt. Differently put, the proposal is as follows: [i] one is in visual contact (as defined earlier) with the object o; a fortiori [ii] one is in sensory contact with o; a fortiori [iii] one is in causal contact with o; and by hypothesis one is no other sort of contact with o. Although it is true that [iv] one can track and locate o, given [i], and therefore that [v] one has discriminating knowledge of o, so that [vi] one satisfies Russell’s Principle with respect to o, and also true that [vii] one has a fundamental idea of o in Evans’s sense, even so, it still is not the case—according to the proposal—that [viii] one can see o, or (therefore) that

The Access Problem

39

[ix] one can think about o. There is, then, in this case, and most generally, no way in which [x] one can have o in mind and the impossibility of [x] rules out the possibility that [xi] one stands in an intentional relation to o; so one is not in an intentional relation with o.17 The dialectic so far may be summed up as follows. If one accepts my intuition about the dog and elephant cases, one may say either [a] the subject cannot see the dog (or shed), nor can she think about it or [b] the subject cannot see the dog (or shed), but she can think about it. If the intuition is denied, one may say either [c] the subject can see the dog (or shed) and think about it or [d] the subject can see the dog (or shed) but she cannot think about it (!) In considering option [b] it is worth comparing the case of (super) blindsight where, by hypothesis, [i]—[v] at least are true.18 One is in causal contact with the object and one can on that basis track and locate it even though one cannot see it (one can imagine giving correct answers when one prompts oneself about where one thinks the object is, then after it is moved, one prompts oneself again). Some would say that although one cannot see it one can think about it. Others would say that one cannot really be said to think about it either. I am more inclined to say that one can think about object o in the (super) blindsight case than the shed or dog case, because I think that the (super) blindsight case differs from the shed or dog case in that there is a respect in which the experience as of a pink elephant, for example, is positively preventing one from thinking of the shed. Some may feel the force of the intuition that one does not see the shed, but think that larger theoretical considerations oblige one to concede that this is not the best thing to say. Many philosophers accept across-the-board externalism

40

phenomenal intentionality

and want to show that they are hard, no-nonsense cognitive scientists whose theoretical claims about perception should apply to robots as easily as human beings. And even those who are half-minded to agree with the present proposal are likely to object that the case of the shed is extreme and does not provide a way of dealing with cases that fall somewhere between it and the case of the rock.

5. The Limits of Error: How Wrong Can One Be, continued The key difference between [1] and [2] is plainly the difference in how badly one gets things wrong. How many of an object’s properties does one have to get right in order to see it? It is, again, very hard to give a precise general answer to this question, but perhaps one can borrow an idea from the so-called cluster version of the description theory of names, which states that if most or a “weighted most” of a set of definite (uniquely identifying) descriptions associated with a proper name “a” are satisfied by a unique object x, then x is the referent of “a.” As is well known, this provision allows the cluster theory to avoid certain objections made to the original “Russell-Frege” version of the theory (see Searle 1958/1983); for as long as a speaker possesses a certain sufficient number of descriptions, or a certain number of the right descriptions, she can successfully refer to the object that satisfies the descriptions in virtue of possessing the properties attributed by the descriptions.19 One plausible way of understanding the notion of a “weighted most” is that there is a hierarchy of such properties, some mattering more than others, and the general idea is that one might treat “see” or “think about” as relevantly similar to “refer.” Thus one cannot be thinking about the shed if one takes the thing that one takes oneself to be thinking about to be an abstract object like a number, for example. That is too far off beam. But nor is it enough that one takes the thing that one takes oneself to be thinking about to be a concrete object, a physical object—if my intuition about the shed case is right. So what is enough? How many of an object’s properties must I get right? No less importantly, how many must I not get wrong—and which ones, exactly? Well, that is for discussion. And even if one can give a plausible “weighted cluster” theory of the conditions on seeing o or thinking about o, one cannot expect it to be more precise than the cluster version of the description theory of names.20 The question remains: how many of an object’s properties does one have to get right in order to see it? And this raises a new question: what is it to “get an object’s properties right,” in the perceptual cases in question? Plainly, my getting it right is not just a matter of my seeing what any ordinarily sighted person would see in my position. If I see a round coin from a certain position as a thin ellipse, and in so doing see what any ordinarily sighted person would see from that position, I do not get the coin’s properties right, in the relevant sense, if I judge it to be thinly

The Access Problem

41

elliptical. So too if I see a house from a distance as a tiny blob, and judge it to be tiny-blob-shaped-and-sized. Rightness is defined by reference to the object’s actual properties, and the matter of perceptual judgment is key.21 If I am seeing an object from a distance—a house that looks like a blob—there is a clear sense in which I am not misperceiving the house’s properties if I know what is going on. Houses look like blobs from a sufficient distance, and in this case it is part of the content of the visual experience that one is seeing an object from far away. Representing the house as far-off and thus blob-like will not involve an error in the content. This suggests that when considering cases of seeing at a distance, “distance factoring” should be considered to be part of the content of the experience, when evaluating the accuracy of what is seen. More generally, some conception of the form “Th is is what xs look like when seen from where I am relative to them” should be considered to be part of the content of visual perceptual experiences.22 Many might still think my claim quite implausible. The most likely response, perhaps, will be that one does see the shed, although it looks to one like a pink elephant; or that one sees the shed as a pink elephant. And it is true that if one sees an object, it must look a certain particular way. The question, though, is whether an object can look any particular way to one and still be said to be seen. I have suggested that the answer is no: that if the properties one represents and the properties an object has are too radically mismatched, one does not see the object. The present response simply denies this, insisting that one does see the shed despite its looking like a pink elephant. It is arguable that the “seeing/seeing as” locution has limited application: one can be said to see an object as something so long as the object is seen in the first place. The present claim, however, is that one cannot be said to see the object in the first place, in shed cases, let alone be said to be seeing it as something. One’s elephantine visual experience actually blocks one from seeing the shed. Suppose (as is wholly feasible) that I’ve been hypnotized in such a way that I see my cat as a wicker basket. I’m astonished to see the wicker basket suddenly move. I do not think I can see the cat in this case. I’m blocked from seeing the cat by my experience of the wicker basket. One can imagine saying in this case, “I know my cat is just there but I can’t see it!” What if my causal, visual contact with the shed begins to change, and I begin to have experiences that are less elephantine, and are in particular less unlike shed experiences. Then it may seem natural for me to say “When I first saw it I thought it was an elephant.” This may be the natural thing to say but this is because we naturally generalize backwards, and call it seeing, even if the best thing to say is that we didn’t really see it at all.

6. Object-Positing In this section I consider a series of objections that focus on the thoughts the subject in the shed case may be said to have, based on causal and visual contact

42

phenomenal intentionality

with the shed. In the first objection, my opponent argues that although it may be granted that one cannot see the shed, surely one can think of the shed. In the second version of the objection, my opponent presses further by asserting that if one can think of the shed, then surely this is evidence that one can see the shed! I argue against these claims by introducing the notion of “object-positing.” Once we become aware of the role of “object-positing” in our perceptual experiences and thoughts in general, it is, arguably, implausible to say that the subject can even think of the shed, however natural this may seem at first. Objection: “Whatever is the case with regard to seeing, you admit that in the pink elephant case, one has sensory and visual contact with the shed. It is this contact that allows one to locate and track the shed, despite the fact that one’s experience is as of a pink elephant. So why can’t one be said to be able to think of the shed simply on the basis of whatever it is that is allowing one to track and locate it?” In response, imagine one thinks that that elephant is pink and as it turns out the shed is in fact pink. Has one thought something true about the shed? (The idea is that the demonstrative [that elephant] somehow succeeds in referring to the shed.) Even when the property that the subject purports to attribute is in fact truly attributed to the object in question, it is not clear that “we would wish to say that the subject thought something true in this kind of case, because we would be extremely embarrassed if we had to provide an account of what it was that he thought” (Evans 1982: 134). I take the implication to be that if one were to have a thought at all, in this case, it would have to count as a true thought because the object does in fact have the property. So the only way it can fail to be a true thought is by failing to be a thought at all. But if one thinks that that elephant is pink (on the basis of one’s experience), then even if it turns out that the shed is pink, it seems incredibly counterintuitive to accredit the thinker with a true thought about the shed. Objection: “Surely I can simply point in the direction of the shed and say truly ‘I see something there, some thing, just over there”? And if the shed is actually pink, surely I have a true thought when I think that thing is pink, and a false thought if it’s green?’ It is the generality of the description—“that one sees some thing over there”—that lends it plausibility. But in doing so, I suggest, it simply rides over the actual phenomenon of seeing as it exists in this particular case, and accordingly misrepresents the actual phenomenal content of the thinker’s mental state. One never simply sees some thing over there, one always sees it as something, even if it is only as a blurry-something-or-other, and the present suggestion is that the generality of the description effectively robs it of any potential to describe this particular act of supposed seeing. And what goes for seeing seems to go for thinking: the thinker is not actually thinking that that thing is pink, but rather that the pink elephant is pink. One may respond that the subject is not just having a single thought involving the content elephant and an ascription of pinkness, but is grasping several

The Access Problem

43

thoughts, some of which are true and some of which are false. According to this proposal, the subject, for example, may grasp the true thought that [a] that thing is pink and the false thought that [b] that thing is a pink elephant This seems to avoid Evans’s doubt by giving an account of what the subject is thinking and the sense in which part of what the subject is thinking may be true. And by giving an account of the sense in which the subject is thinking something true, that is, that thing is pink, we may have reason to believe that the subject does in fact see the shed. The contents of the thoughts represented in [a] and [b] involve a “bare demonstrative,” that is, “that thing,” and we must keep in mind that they are ultimately based on, included in, or derived from the subject’s experience as of a pink elephant.23 So if this account of the subject’s thoughts is correct, there is a legitimate ascription of a thought of a very general nature on the basis of a perceptual experience with very particular content, that is, pinkish, elephant-shaped content. And of course if [a] is to be evidence that one sees the shed, the thought represented in [a] must be true. But is it? Answering this question requires analyzing the relationship between the bare demonstrative thought [that thing is pink] and the subject’s visual experience as of a pink elephant. It does indeed seem plausible that for creatures like us, many perceptual experiences (and thoughts) entail a bare demonstrative thought of the form [that thing is F] If I see a brown cow, for example, it is plausible that this perception entails the bare demonstrative thought [that thing is brown]. Moreover, given that I see the brown cow, the demonstrative thought is true. And this kind of entailment can exist even if the demonstrative thought is not true. If I am having a perceptual experience that is indistinguishable from a veridical perception of a brown cow but is in fact a hallucination of a brown cow, my hallucination also entails a demonstrative thought of the form [that thing is brown], although the demonstrative thought is not true because [that thing] fails to refer to, fails to be about, an external object. Such experiences, and the bare demonstrative thoughts they entail, manifest a fundamental thought-category of our thinking—the category object. Our thought and experience is intrinsically structured in such a way that it is fundamentally object-positing.24 The default setting is object-positing or taking-as-object.

44

phenomenal intentionality

Given that this is so, given that our thought is fundamentally “object-positing” in this way, what is true in the pink elephant/shed case is that the subject is having a bare demonstrative thought of the form [that thing is pink] simply in virtue of having a pink-elephant visual experience. But it is a further question whether this thought, [that thing is pink], is true, even if we continue to assume that the shed is pink. Plainly the thought will be true only if [that thing] succeeds in referring to, or being about, the shed. So the question, again, is whether the bare demonstrative [that thing] refers to the shed. And here, it seems, we are back at the clash of intuition. If the bare demonstrative refers to the shed (or is about the shed), it must do so in virtue of the pink-elephant experience. My intuition is that one does not in fact “get onto” the shed in virtue of the pink-elephant experience (despite locating and tracking abilities); my objector disagrees. Nonetheless, we are not back to Square One. The original objection appears plausible because it seems intuitive to say that the subject is having a thought of the form [that thing is pink], and it is then easy to pass uncritically from the claim that the subject is having a thought of this form to the claim that the subject is having a true thought of this form. My present aim, though, is to question this seemingly easy transition. It is indeed plausible that the subject is having a thought of the form [that thing is pink], but this is simply because so much of our thought and experience is intrinsically “object-positing” in its basic form, quite independently of whether it is veridical or hallucinatory. The subject is having a thought of the form [that thing is pink], which is syntactically identical to the thought the objector proposes the subject is having, and the two thoughts may even have the “same meaning.”25 However, I do not think the thought in question is true, because the bare demonstrative [that thing] fails to refer to, or be about, the shed.26

7. Not a Hallucination The shed/elephant case should—finally—be carefully distinguished from the case of hallucination. Although it is difficult to provide a precise account of hallucinations,27 most will agree that a hallucination, by definition, is the apparent perception of an object in the complete absence of any such object. Consider an adaptation of the original shed case which does count as a hallucination. The way the light reflects from the shed into the subject’s eyes causes the subject to hallucinate a pink elephant. Suppose also that the hallucinated elephant is seemingly located where the shed is, given the way in which the pink elephant hallucination is dependent on the way the light is reflecting off the shed. It seems to be true in this case that the subject does not see the shed. But then how does this case differ from the original shed/elephant case? The relevant difference, I propose, is that subject is in visual contact with the shed in the original case, but not in the hallucination version, although there is

The Access Problem

45

in both cases a causal link between the shed and the subject’s visual experience. Visual contact requires that one’s experience of an object change with eye and body movement (Noë 2003). In the original shed/elephant case, but not in the hallucination version, I assume that there is some systematic co-variation between the movements of the subject’s eyes and body and the shed, in allowing that the subject has locating and tracking abilities. If, for example, the subject were to move to the left, the “pink elephant” would shift to the right of her visual field; if the subject were to move some distance away, the “pink elephant” would appear farther away and take up less of her visual field. Certainly some counterfactual dependencies that ordinarily hold would fail to hold, and it may be difficult to determine the truth-value of others (if, for example, one were to reach out to touch the “pink elephant,” one’s tactile experience would not be elephant-like, ceteris paribus). But there are many ordinary cases of seeing in which counterfactual dependencies that typically hold fail to do so; one who visually misjudges the height of a picture on the wall may fail to take hold of it. In the original shed/elephant case, certain crucial counterfactual dependencies related to eye and body movements do hold, and so, I propose, there is visual contact with the shed. In the hallucination case, however, these counterfactual dependencies do not hold, and the subject does not have the ability to locate and track the shed. By hypothesis, the pink elephant hallucination depends on the way light is reflecting from the shed to the subject’s retinas, which explains why the pink elephant appears to be where the shed is. However, if the subject were to move to the left, forward or backward, for example, the reflection of light would change and the hallucination would end. The point of arguing that the original shed/elephant case is not a hallucination is to show that it cannot be dismissed as irrelevant to the overall debate on these grounds. If it were a hallucination, it could not be used to show that Russell’s Principle fails to provide a sufficient condition on seeing, with its requirement of locating and tracking. That is, if the shed/elephant case were a hallucination, one would not be in causal and visual contact with the shed and therefore would not be capable of the right kind of locating and tracking the shed.

8. Summary The question I have been concerned with in this chapter is how we achieve access to the objects we think of and perceive, and whether there is an internal condition on our access to objects. (I have restricted my attention to thinking of and perceiving material objects.) I rejected an external approach, according to which we achieve access to objects solely in virtue of external relations that hold between a subject’s mental state and objects, in favor of an internal approach. I began with Evans’s rejection of the purely external approach and

46

phenomenal intentionality

his internal constraint as formulated in his endorsement of Russell’s Principle, with its “knowing which one” requirement. I focused on demonstrative identification. According to Evans, one satisfies Russell’s Principle in demonstrative identification by having a “Fundamental Idea” of the object, grasping that object’s fundamental ground of difference, which is what distinguishes that object from all other things. For a material object, it is in virtue of locating and tracking it that one has the capacity to distinguish it from all other things. Although I believe Evans was right to take an internal approach to solving the access problem, I have argued that one may satisfy Russell’s Principle with respect to an object and still not see it. This suggests that Russell’s Principle does not provide a sufficiently robust internal condition for cases of seeing, and as a result, does not provide a sufficiently robust internal condition for having an object in mind. The internal constraint I have articulated is the “matching view”: we achieve access to a material object via perception by correctly representing enough of that object’s properties. It is difficult if not impossible to give a principle for determining when there is enough matching for perceptual contact. We may have to settle for proceeding on a case by case basis: sometimes getting a kind property incorrect may be enough to prevent perceptual contact—this is one proposal about the stone/woman case. Sometimes a radically distorted representation of shape may prevent genuine perceptual contact, as in the pink elephant/shed case. One virtue of the matching view is that it allows for some flexibility in deciding, in any particular case, whether enough of the right properties are represented in order to establish perceptual contact.

Acknowledgments I would like to thank George Bealer, Kati Farkas, Uriah Kriegel, Alva Noe, Susanna Siegel, and Galen Strawson for many helpful suggestions on earlier drafts of this chapter. I would also like to thank the AHRC for an early career fellowship in Spring 2011, which allowed me time to complete this chapter.

Notes 1. 1995: 24–5. Dretske is defi ning “de re mode of representation” as an experience whose object is determined solely by causal or historical relations. 2. A related problem for externalist theories, which I don’t have the space to discuss here, is how we are able to isolate the relevant object as the object of thought given the numerous causes that lie on the causal chain between the object and the thought. Strawson 2004 argues that only an internalist theory can hope to solve this problem 3. See Siegel 2006 for an example of a view that holds that visual phenomenology constrains object-seeing. 4. For example, Campell 2002. I’ll show that some of these worries about Russell’s Principle are consistent with the spirit of the internal constraint I am arguing for here.

The Access Problem

47

5. See also Horgan and Tienson 2002 who claim that there is an internal, phenomenological constraint on thought about and perception of material objects. 6. In section 4 I discuss how this claim may have to be amended for the phenomenon of seeing. 7. See also P. F. Strawson 1971 for this tripartite distinction. 8. If we allow that she is able to see and think about the object at all, on the basis of this visual experience, then we can allow that she can have a thought that is about the object’s color—about the object’s color-whatever-it-is—if only because she believes that all visible objects have some color or other. But she will not in this case be having a thought about the object’s color specifically on the basis of the character of her visual experience, but rather on the basis of her belief that all visible objects have some color or other. 9. My present concern is only with central cases of attentive perceptual experience, not with cases of peripheral vision. 10. The locution “as of” is designed to include the possibility that one can have an experience as of a pink elephant without possessing the concept pink elephant. 11. I am assuming that we are alike in respect of the fundamental type of qualitative character of our visual experience, though not in a way that rules out “color-spectrum inversion.” 12. I will briefly discuss blindsight cases in Section 4. 13. A functional defi nition of possessing vision might defi ne it simply as being able to do whatever vision as we ordinarily understand it makes possible. A slightly more restricted definition would insist that the reception of information must in addition be a matter of the impact of light waves on a sensory organ. On this see Strawson 1989: 208–12. I am not going to consider other “deviant” cases of the following familiar sort: light waves impact on X’s eyes causing olfactory experiences that then cause visual experiences wholly in virtue of being olfactory experiences; and so on. 14. 1982: 179. To have an “idea” of an object in Evans’s sense is to have discriminating knowledge of that object. I explain this in more detail later in the chapter. 15. Evans seems to qualify this by considering the possibility that a statue and a piece of clay, possibly two distinct objects, may occupy the same position in space, so that what differentiates a statue from everything else is its place in space and that it is a statue (see p. 107). 16. On this view there seems to be no good reason to deny that one is seeing the dog if the only sensory experiences one has when light waves impact on one’s eyes are olfactory—so long as those olfactory experiences allowed the subject to locate and track the dog. 17. In discussing steps [i]–[xi] I am only concerned with the possibility of standing in intentional relations to objects based on sensory experiences. Th at is, I am not concerned with the possibility of thinking of an object o and thus standing in an intentional relation to an object o on the basis of a purely descriptive thought. 18. Block 1995/2002 introduced the case of superblindsight as a souped-up version of the ordinary blindsight case. In the ordinary case of blindsight, the subject can only provide information about the object in his visual field when prompted by someone else, whereas the super blindsighter can prompt himself to report on what is in his visual field. As such, the superblindsighter, unlike the ordinary blindsighter, has rational control over the information about the object in his visual field. 19. Although Searle’s view proposes both a sufficient and a necessary condition on reference, my proposal only requires that the “weighted” cluster of descriptions be a necessary condition on seeing an object. 20. It may be argued that any descriptivist approach immediately faces the problem of accounting for the particularity involved in seeing. Given that one sees particular objects (for example, tables, chairs)—how can descriptions deliver this particularity? (For a discussion of this issue see Strawson’s 1959 massive duplication argument and Brewer’s 1999 attempt to address it. See also Martin 2003, Campbell 2002, Bach 2010.) I cannot hope to solve the problem of particularity here, or even adequately articulate what it is. I try to do so in “The phenomenology of particularity” (2011). 21. I understand “actual properties” in an everyday sense, so that I am not wrong in judging that an object is solid, conceiving solidity as I do in the everyday way, even though physics reveals the respect in which the object is 99.99999 percent “empty space.” Nor am I

48

22. 23.

24. 25. 26.

27.

phenomenal intentionality wrong in judging that it is round or smooth although inspection of its edge through an electron microscope reveals something that looks like a map of the Himalayas. It is easily adapted to fit the other sensory modalities. I am being purposely vague about the exact relationship between the perceptual experience as of a pink elephant and the bare demonstrative thought that thing is pink. I will speak loosely of “entailment.” See, for example, Spelke 1990 and Carey and Xu 2001. In David Kaplan’s terminology (1996/1989), these thoughts have the same character but different content. It is clear why Evans wishes to say that there is no thought if I am right about this kind of case. According to Evans, if a demonstrative phrase fails to refer, one would be having no thought at all! See Smith 2002 for an excellent discussion of hallucinations and illusions.

References Bach, K . (2010). “Getting a Th ing into Thought.” In New Essays on Singular Thought, ed. R. Jeshion. New York: Oxford University Press. Block , N. (2002). “Concepts of Consciousness.” In Philosophy of Mind: Classical and Contemporary Readings, ed. D. Chalmers. Oxford: Oxford University Press. Brewer, B. (1999). Perception and Reason. Oxford: Oxford University Press. Campbell, J. (2002). Reference and Consciousness. Oxford: Clarendon Press. Donnellan, K . (1974). “Speaking of Nothing.” Philosophical Review 83: 3–31. Donnellan, K . (1966). “Reference and Definite Descriptions.” Philosophical Review 75: 356–379. Dretske, F. (1995). Naturalizing the Mind. Cambridge: MIT Press. Dretske, F. (1996). “Phenomenal Externalism, or If Meanings Ain’t in the Head, Where Are Qualia.” Philosophical Issues 7: 143–158. Evans, G. (1982). The Varieties of Reference. Oxford: Clarendon Press. Fitch, G.W. (1990). “Th inking of Something.” Nous 24: 675–696. Kaplan, D. (1996). “Dthat.” In The Philosophy of Language, 3rd ed., ed. A. P. Martinich. Oxford: Oxford University Press. Kaplan, D. (1989). “Demonstratives.” In Themes from Kaplan, ed. J. Almog , J. Perry, and H. Wettstein. Oxford: Oxford University Press. Kriegel, U. (2011). “Cognitive Phenomenology as the Basis of Unconscious Content.” In Cognitive Phenomenology, ed. T. Bayne and M. Montague. Oxford: Oxford University Press. Lycan, B. (2001). “The Case for Phenomenal Externalism.” In Philosophical Perspectives 15: Metaphysics, ed. J. E. Tomberlin. Atascadero: Ridgeview Publishing , 17–36. Martin, M.G.F. (2003). “Particular Thoughts and Singular Thought.” In Logic, Thought and Language, ed. A. O’Hear. Cambridge: Cambridge University Press. Montague, M. (2011). “The Phenomenology of Particularity.” In Cognitive Phenomenology, eds. T. Bayne and M. Montague. Oxford: Oxford University Press. Noë A . (2003). “Causation and Perception: The Puzzle Unraveled.” Analysis 63.2: 93–100. Searle, J. (1983). Intentionality: An Essay in the Philosophy of Mind. Cambridge: Cambridge University Press. Searle, J. (1958). “Proper Names.” In The Philosophy of Language, 3rd ed., ed. A. P. Martinich. Oxford: Oxford University Press. Smith, A. D. (2002). The Problem of Perception. Cambridge: Harvard University Press. Strawson, G. (1989). “Red and ‘Red.’” Synthese 78, 193–232. Strawson, G. (2008). “Real Intentionality 3.” In Real Materialism and Other Essays, ed. G. Strawson. Oxford: Oxford University Press. Strawson, P. F. (1959). Individuals. London: Methuen. Strawson P. F. (1971). “Identifying Reference and Truth-Values.” In his Logico-Linguistic Papers. London: Methuen.

3

Indexical Thought Dav i d Pi t t

Call a thought whose expression involves the utterance of an indexical an indexical thought. Thus, my thoughts that I’m annoyed, that now is not the right time, that this is not acceptable, are all indexical thoughts. Such thoughts present a prima facie problem for the thesis that thought contents are phenomenally individuated—that is, that each distinct thought type has a proprietarily cognitive phenomenology such that its having that phenomenology makes it the thought that it is1—given the assumption that phenomenology is intrinsically determined (that is, that intrinsic duplicates are phenomenal duplicates). Let me introduce a notation—“thought quotes”—that will help make the presentation of the problem easier. Thought quotes—^’s (little thinking caps)—work like the quotes of direct quotation. If I want to report what you said the way you said it, I quote your very words: “She said ‘I’m annoyed,’” Similarly, if I want to report what you thought, the way you thought it, I thought-quote the very words you would use to express it: “He thought ^This is not acceptable^.”2 The problem is that there appear to be intuitively good reasons to think that the contents of indexical thoughts, like the contents of indexical utterances, are referent-dependent—that is, that they are determined not (or not only) by intrinsic properties of thoughts, but with respect to the referents of their constituent indexical concepts. Here are four such apparent reasons. 1. If two indexical thoughts t and tcare identical except that one is about—in the sense of having a constituent indexical concept that refers to—an object o, where the other is about (in the same sense) a distinct object oc, then what one thinks in entertaining t is not the same as what one thinks in entertaining tc. My thought ^I’m annoyed^ and your thought ^I’m annoyed^ are different thoughts, because mine is about me and yours is about you. And because our ^I^ concepts have different referents, our thoughts can, moreover, have different truth values. My thought ^Now is not the right time^ thought at noon and my thought ^Now is not the right time^ thought at 1 pm are different thoughts, because the first is 49

50

phenomenal intentionality

about (has a constituent concept referring to) noon and the second is about (ditto) 1 pm.3 Thus, indexical thought contents are, it would seem, individuated partly in terms of what their constituent indexical concepts refer to.4 But this is inconsistent with the thesis that thought contents are individuated wholly by their intrinsic phenomenology, on which our thoughts ^I’m annoyed^ and my thoughts ^Now is not the right time^ would be the same thoughts, if, as I would argue,5 intrinsic phenomenal properties are not affected by extrinsic features of referents. If we understand what is thought (like what is said )6 as the content of the thought (utterance), then we can say that the contents of indexical thoughts are individuated in terms of the referents of their constituent indexical concepts. One way of making this explicit is to say that the content of an indexical thought is referent-involving—that is, that it is a singular proposition, consisting at a minimum of an n-tuple of objects and an n-place property—and that indexical concepts are directly referential, in the sense that what they contribute to the propositions expressed by the thoughts they are constituents of is just their contextually determined referents.7 2. Counterfactual evaluation of what one has actually indexically thought also seems to require that indexical thought content be referentially individuated. Suppose I think ^I’m annoyed^. We can ask whether or not what I’ve thought in the actual world, the content of my actual thought, is true or not at some other world. But the counterfactual truth-value of what I actually thought is determined by whether or not I am annoyed in a counterfactual circumstance; and the only way to secure that I am the referent of ^I^ in another possible world is to, so to speak, bring me along for the otherwordly ride. The descriptive content of ^I^ can’t be relied on to pick out me when applied to another world. Indeed, it won’t pick out anyone without specification of a context; and the right context to specify is one in which I am the agent. Moreover, allowing modal operators to operate on the descriptive contents of indexical concepts results in absurdities such as that the thought ^Possibly I don’t exist^ is necessarily false (since it’s necessarily false that the agent of a context does not exist in that context).8 So, again, the descriptive/phenomenal “character” of ^I^ can’t be part of what I’ve thought when I think ^I’m annoyed^—it’s not a constituent of the content of my thought. 3. It has seemed to some philosophers (such as Evans 1982 and McDowell 1984) that unless one knows what the referents of the indexical concepts in an indexical thought are (by, for example, being acquainted with them), one doesn’t understand what has been thought—just as one wouldn’t understand what has been said by an utterance of an indexical sentence if one didn’t know what the referents of its constituent indexicals were. If I don’t know what you’re thinking about (what your indexical concept refers to) when you

Indexical Thought

51

think ^Th at’s odd^, then I won’t know what you’ve thought, just as I wouldn’t understand your utterance of “Th at’s odd” without knowing what the referent of your token of “that” is. And this would apply in my own case as well. I won’t understand what I’ve thought if I don’t have a referent in mind for ^that^. But if we assume that understanding a thought is knowing what its content is, then it would appear that the contents of indexical thoughts are referent-dependent, and, thus, cannot be purely phenomenal (that is, intrinsically determined). 4. The way we attribute propositional attitudes entails that the content of an indexical thought is individuated in terms of the referents of its constituent indexical concepts, and not by its intrinsic phenomenal features. The way for a third party to report the thought I express in uttering “I’m in agony” is not to say “He’s thinking [that] I’m in agony,” but, rather, “He’s thinking [that] he [himself] is in agony.” To get the content of my thought right, any third-person ascription must replace “I” with a term that, in the mouth of the ascriber, refers to what “I” referred to in my mouth (viz., me). The descriptive content of the part of the third-person ascriber’s utterance that specifies the content of my thought must be different from that of mine in order for the ascription to be correct. To capture the content of my thought your term must agree in reference, not conceptual content, with mine. (Cf. Frege 1918.) Hence, indexical thought contents are individuated referentially, and so not by any of their intrinsic properties.9 My concern in this chapter is to blunt these intuitions and to defend a conception of thought content on which it is entirely phenomenal and internalist. On this view, when you and I think ^I’m annoyed^, we’re thinking the same thought. The fact that my thought is about (refers to) me and yours is about (refers to) you, and that mine might thus be true while yours is false, doesn’t give them different contents and doesn’t make them different thoughts. Nor does modal evaluation of indexical thoughts require that their contents be individuated by the contextually determined referents of their constituent indexical concepts. One cannot determine the truth value of an indexical thought in the actual world without a specification of its context (a thinker, a time, a place, an addressee, etc.); and the same is true for evaluation of that thought in any other possible world. The fact that the relevant context is otherwordly doesn’t entail that indexical contents are referent-dependent any more than contextual sensitivity in the actual world does. And there is no semantic necessity that the object I pick out in another world be the very object referred to in the actual world. Moreover, ignorance of the referent of an indexical concept does not prevent understanding of its containing indexical thought. Knowledge of referents enriches one’s overall cognitive take on things, but only through the introduction of further thoughts. Finally, the constraints our propositional-attitude

52

phenomenal intentionality

ascribing practices are subject to are pragmatic rather than semantic. Our interest in what others are thinking is influenced by factors that do not affect the individuation of their contents.10

1. What Is Thought The intuition behind the first reason for thinking that indexical thoughts are referentially individuated is that what you think—the content of your thought—when you think, for example, ^I’m bored^ is different from what I think when I think ^I’m bored^, because we’re thinking about (referring to) different people, and what we think can differ in truth-value. Thus, the contents of our thoughts must be referent-dependent, and our thoughts cannot be individuated by their intrinsic features, phenomenal or otherwise. A standard move for anyone who accepts these intuitions but nonetheless thinks there’s something content-like that our ^I’m bored^ thoughts have in common is to make a distinction between so-called narrow content and broad (or wide) content. That is, it might seem that the best way for a phenomenal intentionality enthusiast to proceed here is to make a Kaplanesque character-content distinction for indexical thoughts, and identify cognitive phenomenology with (the cognitive equivalent of) character. One could then say that our ^I’m bored^ thoughts share a cognitive character, though they differ in content, and that they are the same thoughts in one sense, but distinct thoughts in another sense. Thoughts can be individuated in any number of ways, depending upon one’s purposes; and I suppose this way serves some purposes perfectly well. But I don’t think it’s the right way to individuate thoughts.11 On my view, it doesn’t capture what they are.12 Thoughts aren’t theoretical posits whose natures are determined by their containing theories, but objects of intimate acquaintance in experience. They’re states of minds, and, as such (on my view), are intrinsically constituted, and accessible and knowable from the first-person point of view. Th at a (mind- and language-independent) proposition could be referent-dependent or referent-involving is, perhaps, unproblematic. So if thoughts are representations of propositions, referential individuation of indexical thought contents need present no special issues. But if thought types (and their contents) are, as I maintain, cognitive phenomenal types (and their tokens are tokens of those types), then they can’t be individuated non-phenomenally. They can’t have non-phenomenal constituents or depend for their identity on non-phenomenal entities, and neither can their tokens. It couldn’t be literally true that one has a concrete (or abstract) object, such as an individual or a time (or a number or a universal), in mind, or that mere reference to an extrinsic, non-phenomenal entity could affect the intrinsic nature

Indexical Thought

53

of a mental state. If thought contents are phenomenal types, thoughts cannot be referent-involving; and if phenomenal properties are intrinsically determined, they can’t be referent-dependent.13 So I have reasons for questioning the motivations for making a narrow/ broad or character/content distinction in the first place—that is, for introducing a two-factor theory of content—of something like meaning (as opposed to reference).14 It seems to me that a theory that postulates only one kind of thought content is all the theory we need. To put it in Kaplanian terms, I want to argue that character is always content. In part, resistance to this sort of view is rooted in the intuition that, since content determines reference, if two concepts have different referents, they must have different contents. So, your token of ^I^ and my token of ^I^, having different referents, must have different contents. It can’t be that our thoughts have exactly the same contents but different referents, truth conditions and possible truth values. But I don’t think this is correct. The argument is that, for example, our ^I^ thoughts have different contents because their constituent ^I^ concepts have different referents. But it’s not true in general that thoughts with constituent concepts having different referents or extensions must have different contents, and be different thoughts. Consider the following examples. In 1972 I think ^All dogs are quadrupeds^. In 2002 I think ^All dogs are quadrupeds^.15 The extension of ^dogs^ in 1972 is not the same as the extension of ^dogs^ in 2002. But have I thought different thoughts? Has my concept ^dog^ changed its content? Again, I have a strong intuition that I haven’t, and it hasn’t. Do my thoughts have different truth conditions? Well, you might think they don’t, since they are, in both cases, just that the set of dogs is a subset of the set of quadrupedal things, and in both cases the conditions are fulfilled (so that difference of extension is not sufficient for difference of truth conditions). Though, on the other hand, if you think that the truth conditions of the 1972 thought ^All dogs are quadrupeds^ involve the set of the dogs that there are in 1972, and those of the 2002 thought the set of the dogs that there are in 2002, then the thoughts do have different truth conditions—though they are still, I maintain, tokens of the same thought type. In either case, however, the truth values of the thoughts could be different: dogs might have mutated into tripeds in the interim.16 And this ought to be enough, on the view I’m challenging, to conclude that they have different contents. But they don’t. General concepts don’t change their contents with changes in their extensions. Here’s another example. In 1972 I think ^The president of the United States is a criminal^. In 2002 I think ^The president of the United States is a criminal^. Have I thought the same thing or not? It seems to me that I’ve thought precisely the same thing on both occasions, in spite of the fact that in 1972 I’m

54

phenomenal intentionality

referring to a different criminal than the one I’m referring to in 2002, and, hence, that it’s possible for what I thought in 1972 to differ in truth value from what I thought in 2002.17 Difference in reference of definite-descriptive concepts, and the consequent difference of truth conditions and possible truth values of the thought tokens, doesn’t result in thought tokens with different contents either.18 Now, it might be maintained that the contents of the thoughts just mentioned are insufficiently specified, because they contain a hidden indexical concept—for example, ^current^—and, hence, are in fact indexical thoughts. (In which case it would be question-begging to deny that their contents are different.) But the concept ^president^ is not the same as the concept ^current president^, and, hence, it’s possible to think that the president of the United States is a criminal without thinking that the current president of the United States is a criminal. These are different thoughts; and the former doesn’t have a hidden indexical, and doesn’t change its content when thought at different times, about different presidents. (This sort of point would apply to any appeal to hidden indexical concepts.) You might also think that the contents of these thoughts are insufficiently specified because the complete content of any thought includes an indication of the time at which it is thought, so that, for example, my 1972 and 2002 dog thoughts in fact have different (non-indexical) contents, and that is what allows them to have different truth values. From a first-person perspective, however, the view that thought contents always include time indications, and hence are true or false once and for all (a doctrine sometimes called “eternalism,” and contrasted with “temporalism”) is quite implausible. I might think ^All dogs are quadrupeds^ and have no idea what the current date is. And even if I know the date, I need not think of it whenever I think ^All dogs are quadrupeds^. The thoughts ^All dogs are quadrupeds^ and ^All dogs are quadrupeds in 1972^ are different thoughts. Moreover, parity of reasoning would seem to require “ubiquitism”—the view that thought contents include place indications, and hence are true or false at every location—as well as “necessitarianism”—the view that thought contents include world indications, and hence are true or false at every possible world. But surely my thought ^The mayor is in his office^ (n.b.: not ^The local mayor is in his office^) could be true if I thought it in Los Angeles, though false if I had thought it (at the same time) in New York. And I could think it, yet have no idea where I am. And even if I do know where I am, I needn’t think that I’m there when I think ^the mayor is in his office^. Similarly, necessitarianism has the consequence that all of my thoughts are either necessarily true or necessarily false (it’s true (false) at every possible world that my thought is true (false) in the world in which I’m thinking it); and a parallel point holds about

Indexical Thought

55

the possibility of thinking a thought without thinking of the possible world I’m in (or even having the concept of a possible world). Analogous considerations hold for difference in reference or extension across worlds. We don’t suppose that general concepts change their meanings across worlds in which they have different extensions. My otherworldly twin may be thinking of other dogs, but he thinks exactly what I think when he thinks ^All dogs are quadrupeds^. Nor do we suppose that the content of definite-descriptive concepts changes with changes in their referents—for example, ^the president of the United States in 2011^ doesn’t mean something different at worlds in which Barack Obama isn’t president in 2011; and ^the president of the United States in 2011 is African American^ doesn’t mean something different at worlds at which it is false. Responses similar to those given earlier are available to the claim that such thoughts contain the implicit indexical concept ^actual^. Though I’ve given reasons for thinking that the (controversial) thesis that reference, extension and truth are always relative to something (be it a world or a context or contextual feature within a world) is correct, in fact my argument doesn’t depend on it. The essential point is that such relativity doesn’t entail relativity of content: one may consistently endorse the claim that the extensions of general and definite-descriptive concepts typically vary with respect to places, times and worlds while denying that their contents do. Hence, some rationale is required for thinking that indexical concepts are exceptional in this regard. In the absence of such a rationale, they may be treated as differing only with respect to the scope of their relativization. Whereas general and (indexical-free) definite-descriptive concepts have extensions relative to worlds, and times and places within worlds, indexical concepts have extensions relative to times, places, speakers, addressees, and so on within a world.19 The fact that two indexical thoughts have constituent indexical concepts with different referents, and hence different truth conditions or truth values, does not entail that they have different contents.20

2. Modality The standard modal motivation for a referent-dependent semantics of indexical expressions is that the actual referent of any indexical must be a constituent of what’s evaluated at other worlds in order to get the truth conditions of modal indexical sentences to come out right.21 Analogous considerations would seem to apply to indexical thoughts. This intuition is widely accepted; but I think some reflection shows that it’s not inevitable. There are intuitively satisfying ways to understand the counterfactual evaluation of indexical thoughts (and sentences) without individuating them referentially.

56

phenomenal intentionality

Suppose you and I both think ^I’m itchy^. On the view I’m defending, your token of the concept ^I^ has exactly the same content as mine, or anyone else’s. ^I^ is (like definite descriptions) referentially singular, but conceptually general: its tokens refer to individuals, but its content is such that different tokens can refer to numerically distinct but relevantly qualitatively identical individuals.22 What we have thought, the shared content of our token thoughts, is, on my view, the cognitive phenomenal type ^I’m itchy^. And our thoughts—our individual, unshareable, unrepeatable, dated cognitive episodes—are tokens of that type. (They have their contents in being tokens of their contents.) The truth-values of our thought tokens are determined by the states of the referents of their token ^I^ constituents; and the referents of our token ^I^’s are determined by their contents—relative, of course, to some parameter (a world, time, person, place, etc.). Insofar, there’s no difference from truth-value determination for non-indexical thoughts: contents determine referents and extensions relative to some parameter, and the relations of referents and extensions determine truth-values. As noted earlier, where indexical contents differ is only in the scope of their relativization—as it were, the extent of the context to which their reference is relativized. Concepts have referents or extensions relative to contexts, and contexts come in various sizes, from individual acts of attention to possible worlds to the space of all possible worlds (and perhaps beyond). General concepts have extensions at worlds and, typically, times within those worlds. Defi nite-descriptive and indexical concepts can have referents at proper parts of worlds—places (^the mayor^, ^here^), times (^the president^, ^now^), individuals (^I^, ^ you^) and acts of attention (^this^, ^that^). And just as you and I can think the same non-indexical thought—that is, token the same thought type—with different truth-values, we can think the same indexical thoughts with different truth-values: my token of ^I’m itchy^ can be true while yours is false—since the referents of their constituent ^I^ concepts are different, and the states of those referents may be relevantly different. The same parity holds for the evaluation of thoughts at other possible worlds. In order to evaluate a thought at a world, one must determine extensions for its constituent referential concepts. For some concepts, specification of a world may be enough. In the case of indexical concepts, specification of a world is normally not sufficient to determine referents, and so not sufficient to determine truth-value. One must also specify a relevant context within that world—a person, place, time, and so on. But it doesn’t follow that indexical referents must be constituents of the contents of indexical thoughts evaluated at a world—whether it be this world or some other. And it doesn’t follow from its being the case that an individual must be specified in order to evaluate an indexical thought at a counterfactual circumstance that the referent of its actual token must be specified.

Indexical Thought

57

I can sensibly ask whether what I thought in a particular context in the actual world is true in some context, in the actual world or in some other world. If I think ^I’m itchy^, I can ask if what I thought is true of you: not (n.b.), is my token true of you (we may assume it couldn’t be); but, supposing you’re also thinking ^I’m itchy^, is your token true of you. This is exactly parallel to the situation in which I, in Los Angeles, and you, in New York, think ^The mayor is in his office^. The thought type we’ve tokened contains the place- (and time-) sensitive concept ^the mayor^. Hence, since my token occurred in Los Angeles (in 2011), the default assumption is that it refers to Antonio Villaraigosa,23 while your token of the very same concept (at the same time) refers to Michael Bloomberg. And I can ask whether what I thought—the content I tokened—is true when tokened by you. And what’s true across contexts within a world is also true across worlds. I can sensibly ask with respect to some other world whether what I actually thought is true of me or you or someone else, or at some time or place, in that world. By which I would mean: if the thought type I tokened in some particular actual context were tokened in some context in some other world, would it be true in that context (of the chosen individual, time, place, etc.) in that world? If I ask whether what I thought when I thought ^I’m itchy^ is true at possible world w, the answer depends upon which individual in w is assigned as the referent of ^I^. Though I’m semantically constrained to refer to myself when I token the concept ^I^, I’m not semantically constrained evaluate the content of my ^I^ thought at another possible world with respect to me.24 The token ^I^ in my token of ^I’m itchy^ must refer to me. But since its content is general (in the sense specified earlier), it can have tokens that don’t refer to me. Which context I select at a given world will be determined by whom I’m interested in. If I want to know whether or not what I thought of me in the actual world is true of me at another possible world, then I must specify a context in that world of which I am the agent. I must find me there (or take me there) and examine my qualities there. It might be that when I or anyone else asks whether what I thought is true at world w, we’re most often concerned with whether or not what I actually thought of myself (necessarily) is true of me at w. But this is not semantically required, any more than it’s required of my thought ^The mayor is in his office^. There may be a pragmatic presupposition—even a very strong one—that questions about the (actual and) counterfactual truth value of what I thought will be anchored to the referent of my actual tokening of ^the mayor^, but there’s no semantic necessity that it be. Counterfactual reference to actual token referent is not necessitated by the content of the concept. The fact that ^I^ tokens must refer to their actual tokeners is a red herring.25 The evaluation of general thoughts at other worlds works the same way. Consider again ^All dogs are quadrupeds^. In order to determine its truth value at a world, we must determine the extensions of ^dogs^ and ^quadrupeds^ in

58

phenomenal intentionality

that world. (We may need to consult specific contexts as well, since the extensions of these terms vary over time.) It’s true at the actual world iff the actual extension of ^dogs^ is a subset of the actual extension of ^quadrupeds^. To determine the truth value of this thought at some other world, we must assign extensions to ^dogs^ and ^quadrupeds^ in that world. It’s perhaps the default assumption that we’re interested in whatever extensions those concepts happen to have in counterfactual—because we’re typically interested in the modal properties of dogs qua species. But we could also inquire about the modal properties of some dogs in particular. We might be interested in knowing whether ^Dogs are quadrupeds^ is true at some other world where ^dogs^ has the same extension it has here—that is, whether the dogs we’ve picked out in the actual world have four legs in some other world. (It might be that we suspect there’s something special about these dogs: they’re necessarily immune to leg loss, though no other dogs are.) The default assumption for indexical concepts is just the reverse: counterfactual questions concern their actual token referents. But in both cases it’s possible—it’s coherent (even if strange or typically pointless)— to cancel the default assumption. It’s not ruled out semantically. The pragmatic nature of the presupposition that, when one asks of an indexical thought (or utterance) whether it might have been true or false, one is asking about the actual referents of tokens of its constituent concepts, is perhaps plainer in the case of indexical concepts whose token reference is not semantically constrained in the manner of ^I^. Suppose there are two tall males before us, and I say to you “He’s tall.” If I don’t indicate to you whom I’m referring to (say, by directing my gaze, or pointing), you won’t know of whom I’ve said it. And if I’d said it twice (referring the second time to the second individual, without indicating him to you), you’d know that I’d said the same thing twice—though not whether about the same individual, or one then the other. As we saw earlier, the fact that these sentence tokens can have different truth values, and that their constituent ‘he’s’ can have different referents, does not, per se, give them different contents. (Indeed, I could utter the same sentence falsely, of one and the same individual, at different times. Should we conclude that what I’ve said has changed?) What you said is independent of whom you said it of. If I then say “Possibly, he’s not tall” (or, more colloquially, “He might not have been tall”), there’s nothing in the content of my utterance to determine which of the two guys I’m talking about. I might have been referring to either (or even to neither, some additional individual having come to my attention). And if I don’t direct your attention to one or the other, you won’t know to whom I was referring, or whether or not what I said is true. I could have said the very same thing about the other. Whom I’m referring to, in any case, including a counterfactual one, is determined by whom I’m interested in. But this is a pragmatic matter; it’s not determined by the content of my utterances. Your not knowing whom I’ve referred to by my utterance of “he” does not

Indexical Thought

59

entail that you don’t know what I said. You just don’t know whom I’ve said it of, or whether or not it’s true (supposing you have the relevant information). It might be odd to go from thinking ^He’s tall^ of one person to thinking ^Possibly, he’s not tall^ of some other person—especially if the latter were thought of someone not present. But it’s not semantically impossible.26 Indeed, it seems to me that it’s not conceptually impossible even to think ^He’s tall^ or ^Possibly he’s not tall^ about no one in particular—that is, just think it; just entertain the thought (the way you’re doing right now). In fact, such a non-specific thought might even be construed as having a determinate truth value. What one might be thinking in such a case—rather abstractly (and probably idly)—is that there is a possible world at which a male (at some time) isn’t tall. This would be to address the purely general content of the thought as such. There may not be much point to wondering this; and it’s certainly an unusual thing to think. But the fact that it’s possible shows that the contents of indexical concepts are as general as their non-indexical cousins. Of course, Kaplan objects to this sort of construal of indexical sentences with modal operators. He calls intensional operators on characters “monsters,” and accuses them of wreaking semantic and metaphysical havoc. If, for example, we suppose that “possibly” operates on the character of an indexical sentence instead of its referent-involving content, we end up with such absurdities as that “Possibly I don’t exist” (and, of course, ^Possibly I don’t exist^) is false, since, according to Kaplan, any context “appropriate” (to use his term) for the evaluation of an “I” sentence will contain an agent, which is the referent of the occurrence of “I” in that context, and, hence, “I exist” will be true in every such context. Thus, “I exist” will be true in every possible (appropriate) context, and “Possibly I don’t exist” will be false. But, obviously, it’s possible for such contingent beings as us not to exist (Kaplan 1989: 498). Such problems can be avoided by making indexical contents their contextually determined referents, and the content of my utterance of “I don’t exist” the singular proposition consisting of me, the property of existence (supposing arguendo that existence is a property) and negation. We can then approach a world (a circumstance of evaluation) with this proposition, and ask whether or not it is true at that world. If I’m not to be found in that world (at a specified time), then the proposition, hence what I said, is true at that world (at that time). But monsters are only a problem on the assumption that indexical contents are referent-involving. This is clear from the passage in “Demonstratives” (510) in which Kaplan responds to the question Are there such operators as ‘In some contexts it is true that,’ which when prefi xed to a sentence yields a truth if and only if in some context the contained sentence (not the content expressed by it) expresses a content that is true in the circumstances of that context?

phenomenal intentionality

60

as follows: Let us try it: (9) In some contexts it is true that I am not tired now. for (9) to be true in the present context it suffices that some agent of some context not be tired at the time of that context. (9), so interpreted, has nothing to do with me or the present moment. But this violates Principle 2! But Principle 2 is just the thesis that indexicals are directly referential. If we don’t accept it, then the fact that (9) evaluated with respect to other contexts might be true of someone else is not problematic. I might be most interested in what’s the case with me in some other context in evaluating (9), in which case I can pick contexts in which I am the agent. But—or so I’ve argued—this choice is pragmatically driven and not dictated by the semantics of indexicals. Likewise, if we don’t suppose that the content of a tokening of ^I^ by me is me, but, rather (something like) ^self^—a reflexive self-concept—then ^Possibly, I don’t exist^ can come out true even if ^possibly^ is monstrous. Of course this concept can’t be tokened in any context lacking a thinker (no concept can). But tokening of concepts is like utterance of sentences, and we can follow Kaplan in maintaining that it’s not utterances of sentences containing indexicals that get evaluated at counterfactual contexts but occurrences of them. If we took evaluation to concern utterances, then “Possibly I am not speaking to you now” would be in the same sinking boat as “Possibly I don’t exist.” So the question becomes, are there true occurrences of ^I don’t exist^ (understood as having (something like) the content ^there is no self^)? Clearly there are, namely, occurrences evaluated at any context in a world without selves. If there’s no self for ^I^ to refer to, then ^I don’t exist^ is true. No token of ^I exist^ can be false—just as no token of “I am speaking to you now” or ^I am thinking now^ can be false. But it doesn’t follow that its character can’t be false. It’s false at all contexts in worlds with no agents—just as “I’m speaking” is false at all contexts in worlds with no speakers. Again, what one is most likely interested in with respect to one’s thought ^Possibly I don’t exist^ is not worlds in which there is no one at all, but worlds in which one is oneself not to be found. But this can be accommodated. One can ask, “Is what I thought of me true of me?”—that is, is there a possible world in which the individual about whom I actually thought the thought, namely, me, does not exist? And the answer is (alas) yes. The fact that it might be extremely odd for someone to ask whom you are thinking about when you wonder whether ^Possibly I don’t exist^ is true doesn’t entail that it is semantically incoherent. If you’re interested in what’s true, at some other world, of the individual you actually referred to, you don’t have to put him in the content of the utterance in

Indexical Thought

61

order to take him there. You can take him along as the referent of your actual indexical token. I conclude that modal considerations don’t yield decisive reasons for thinking that the contents of indexical concepts are referent-dependent. One need suppose neither that indexical concepts are directly referential, in the sense that what they contribute to what is thought by their tokening (the content of their tokens) is just their contextually determined referents, nor that they’re rigidly referential, that is, that an indexical concept token must refer to the same thing in every possible world.27 They’re as general in content and as variable in their reference as their non-rigid relatives.28 This very token indexical concept that I entertain (except for ^I^), with its actual content, can have different referents in different possible worlds/contexts. So modal considerations don’t show that indexical contents are referent-involving, and don’t militate against the thesis that (the contents of) indexical thoughts are intrinsically, phenomenally individuated. The lesson here is that indexical contents are thin. But it’s their very thinness that makes them so useful—so portable (that is, applicable to things about which one has minimal information). We should resist the temptation to fatten them up with referents in response to extra-semantic considerations having to do with the way we typically use them.

3. Understanding The third motivation for supposing that the contents of indexical thoughts are referent-dependent is the claim that unless one knows the reference of an indexical concept one doesn’t understand the thought it’s a constituent of. If you or I think (or try to think), for example, ^That’s a nice one^ or ^She’s a little runaway^ without having any particular individual in mind as referents for ^that^ and ^she^, then neither of us will understand what either of us has thought.29 Moreover, if we tried to express our thoughts by uttering the relevant sentences without identifying referents for their constituent indexical terms, we would fail to understand what we had said, as would anyone else who could not identify such referents. My response here is a quick one. Understanding comes in degrees, and whereas there might not be much information associated with an indexical thought of whose indexical referents one is unapprised, or conveyed by its linguistic expression, it’s far from clear that nothing has been thought or understood or communicated. Say you hear an utterance of the sentence “She’s here!” and you have no idea who said it, where it was said, or whom it was said of. Do you nonetheless not know that someone has said that some female is located at some location? You’ll have a much richer idea of the nature of the state of

62

phenomenal intentionality

affairs that this utterance refers to if you know who she is and where here is. But it’s far from clear either that you don’t understand the utterance, or that if you repeated it yourself (perhaps replacing “here” with “there”) you wouldn’t understand what you said, or, indeed, that you haven’t succeeded in saying anything. Likewise in thought. I can very well think ^She’s a little runaway^ having no one in particular in mind as the referent of ^she^ and have thought something determinate—if only that some female is a small fugitive. And I can have absolutely no idea where I am (I was transported in a black hood) or what my surroundings are like (it’s pitch dark), and still think something determinate in thinking ^I really don’t like it here^. I know that ^here^ refers to the place where I am, even though that’s the only information I have. If someone turns the lights on, my understanding of ^here^ would increase, but only in the sense that further thoughts would be directed toward the place where I am, and would be known to be about the same place. My thought ^I really don’t like it here^, in the dark, with no knowledge of the referent of ^here^ beyond its being the place where I am, though surely paling in comparison with the same thought when there’s a referent in view, is nonetheless a real thought. I do succeed in thinking something in such a case, and I do understand what I’m thinking, even if what I’m thinking is relatively jejune and uninformative. What I’m thinking in the perceptually impoverished environment doesn’t change when the lights go on. Like I said, conceptual indexical contents are very thin. We don’t suppose that acquaintance with the extensions of definite-descriptive or general concepts, or knowledge of truth-values of the thoughts containing them, is required for determinate thought or understanding. It’s only the relative informational paucity of indexical concepts that tempts us to regard them as different.

4. Attitude Reports It appears that getting the contents of indexical thoughts right in third-person ascriptions depends upon identifying the referents of their constitutive indexical concepts, not their descriptive content. In order to report what you think when you think ^I’m hungry^ I must capture the referent of your tokening of ^I^, not its descriptive content (its character). Thus, cognitive phenomenology—a mental analogue of character—is not part of what is thought, the content of the thought. The contents of (token) indexical concepts are referent-involving, and so can’t be phenomenally individuated. Moreover, the way I’ve set up the issues begs the question against these facts in its use of thought-quotes, which refer to the vehicle of the thought rather than its content. These concerns can be quickly dismissed. It’s not a cheat for me to use thought quotes, since, unlike in the case of a sentence token and its content,

Indexical Thought

63

on the conception of intentional content I’m defending there’s no difference between the tokening of a thought and the tokening of its content. This is an important disanalogy with speech. The content of a sentence-token is not an intrinsic feature of it, whereas (or so I maintain) the content of a thought is. On my view, thoughts aren’t representations of contents; they’re tokens of contents. Further, I think the argument that the need to identify the referent of a reported belief entails that the belief’s content is referentially individuated is a non-sequitur. We could have a primary, even overriding interest in the referents of an individual’s concepts—which things in the world he is thinking about—without its being the case that content itself is referent-dependent. Interest in a thinker’s referents is perhaps to be expected, given the relation of propositional attitudes to behavior, and our interest—perhaps typically—in what someone is going to do, has done, or is doing, that is, which objects in our shared environment (including, especially, ourselves) the individual’s actions will affect. If our main reason for wanting to know what someone is thinking is that we wish to know which objects he might be acting upon, then we may be less concerned with how he is thinking of them, that is, with what his thought actually is. But our practical interests in the referents of an individual’s thoughts should not be the basis for an account of the nature of the thoughts themselves. The pragmatics of propositional-attitude ascriptions should not be allowed to dictate the metaphysics of propositional attitudes themselves. There are ascriber-independent facts about what individuates a thinker’s thoughts per se—independent as much from the interests of others as from the referents of their constituent concepts. (These facts are what are referred to using thought quotes.) We should not conflate interest in what someone is thinking about (in the sense of what the referents of his concepts are) with interest in what he is thinking. Indeed, even if our concern is primarily external, if our expectations, predictions or explanations are thwarted by an individual’s behavior, we will recur to an interest in what he is thinking.

5. Two Objections30 1. If we don’t appeal to the referents of indexical concepts in the individuation of indexical thoughts, then there will be cases in which we can’t make sense of thinkers agreeing in what they think. Suppose I think ^I’m sad and lonely^. If you want to agree with me, you’ll have to think the same thing I’m thinking. But in order to do this you’ll have to think (of me) ^He’s sad and lonely^. But the thought ^I’m sad and lonely^ and the thought ^He’s sad and lonely^ can only be the same thought if they’re individuated referentially—that is, as having the singular proposition featuring yours truly and the properties sadness and loneliness as constituents as their common content. Individuated

64

phenomenal intentionality

internally, our thoughts are different; but if our thoughts are different, then we don’t agree. My reply to this objection is that it rests on a false assumption. Thinking the same thought is not the only way in which two thinkers can agree. They can, for example, think thoughts with different referring concepts that have the same truth conditions. There’s a perfectly good sense in which you and I could agree on the mental status of George W. Bush, I by thinking ^The president of the United States is delusional^ and you by thinking ^The commander-in-chief of the armed forces is delusional^. (Whether or not we know that we agree depends on whether or not we know that the referents of our referring concepts are the same.) In the most basic case, agreement in thought involves merely attributing in thought the same properties to the same individuals. But this doesn’t require attributing in thought the same properties to the same individuals in the same way, and so doesn’t require that the thoughts we think be the same. So we don’t need recourse to extensional ways of individuating thought contents in order to accommodate agreement in thought. (This is obviously related to the points made earlier about the pragmatics of attitude ascription.) 2. Referential individuation of indexical thought contents would seem to be required in order to accommodate the possibility that a token of the thought ^This is this^ could be false, and a token of ^This is not this^ could be true. If the contents of all ^this^ tokens were the same, then it would seem that any thought of the form ^This is this^ would be necessarily true, and any thought of the form ^This is not this^ would be necessarily false. But there appears to be ample evidence that this is not the case. If I think of my left thumb that it’s my right thumb by thinking ^This [attending to my left thumb] is this [attending to my right thumb]^, what I’ve thought is false; and if I think of my left thumb that it’s not my right thumb by thinking ^This [attending to my left thumb] is not this [attending to my right thumb]^, what I’ve thought is true. So it can’t be that the contents of the two tokens of the indexical concept ^this^ are the same. If, however, we suppose that the contents of indexical tokens are individuated referentially, then the fact that the referents in the fi rst case are different and in the second case the same, smoothly accounts for the differences in truth value. However, the bump, having been thus depressed, reappears in another part of the carpet. For, if we suppose that the contents of indexical concepts are referentially individuated, then tokens that have the same referents have the same contents, and tokens that have different referents have different contents. But then, if I were to think ^This [attending to my left thumb] is not this [attending, again (inadvertently), to my left thumb]^, I would be thinking something contradictory, and we’d have to explain how I might nonetheless be

Indexical Thought

65

rational. Likewise if I were to think ^This [attending to my right thumb] is this [attending (inadvertently) to my left thumb]^. 31 I think the thing to do here is to return to the thesis that indexical contents are non-referentially individuated, and to rethink the relation between content and reference. If we suppose that ^this^ has a constant content, and that, therefore, ^this is this^ is analytic, are we constrained to think that it, or any of its tokens, is necessarily true? I think not. Analyticity need not be construed as necessary truth in virtue of meaning. In fact, there are thoughts which, I would argue, are analytic and not true—for example, ^The present king of France is male^, and ^Round squares are round^. Analyticity may be construed as an entirely internal, structural relation among the component contents of thoughts, which in itself does not determine reference or truth value. The content of ^this^ is not such as to determine the same referent on every occasion of its use. Certainly the contextual features relevant to the determination of the referent of an indexical concept can change mid-thought—as in, for example, the (true) thought ^Now is not now^; indeed, given that thinking takes time, the features relevant to the interpretation of tokens of ^now^ are constantly changing. Even if one observes Kaplan’s distinction between an utterance and an occurrence, one would not be constrained to assign the same referent to, for example, both occurrences of ^this^ in an occurrence of ^This is this^. Similar considerations apply to non-indexical singular concepts as well. Consider the thought ^The king is dead; long live the king^. Though the content of ^the king^ is the same in both of its occurrences, the sentence is not contradictory, since the referents of the occurrences are different. The content of ^the king^ does not determine that the referents of all of its occurrences must be the same. (Indeed, I would argue, the only expressions whose contents determine the same referents for every occurrence are de facto rigid designators.) It’s simply left open whether or not a thought like ^The king is the king^ or ^This is this^ is true, in spite of the fact that both tokens of its constituent concepts have the same contents. Likewise, it is left open whether a thought like ^The king is not the king^ or ^This is not this^ is false. In both cases sameness of conceptual content doesn’t entail sameness of conceptual referent; hence, analyticity doesn’t entail truth, and contradictoriness doesn’t entail falsity. And the same holds true for general thoughts such as ^Dogs are dogs^. If a thought is true iff the extension of the first occurrence of ^dogs^ is identical to the extension of the second, then, given that this might not be the case (some dog is born or dies in the middle of my thought), the thought might not be true. What of the thought ^I am here now^? Kaplan argued that the corresponding sentence is a truth of the logic of indexicals. Though the singular proposition it expresses is not a necessary truth, it’s necessarily the case that any occurrence of “I am here now” is true. It has a necessary character, but not a necessary content. If, however, one identifies character and content, as the account I’m defending does,

66

phenomenal intentionality

then would it not have to be that the content of my thought ^I am here now^ is necessarily true? But the intuition that it might not have been the case that I was in that place at that time, which is clearly correct, seems to rule this out. According to Kaplan, the character of “I” is a function from a context to the agent of the context, the character of “here” is a function from a context to the place of the context, and the character of “now” is a function from a context to the time of a context. Contexts are represented by indices, which are n-tuples of (perhaps among other things), a world, an agent, a place and a time. He therefore builds it into his system that any occurrence of “I am here now” will, necessarily, be true—since the agent, place and time of a given context are, necessarily, the agent place and time of that context. This construction is based on the intuition that “I am here now” is “deeply, and in some sense, . . . universally, true. . . . One need only understand the meaning of [it] to see that it cannot be uttered falsely.” (Kaplan 1989: 509). On the contrary, I would argue, intuitively, “I am here now” can be uttered falsely—if, for example, one moves while one is speaking. Intuitively, the characters of “I,” “here” and “now” are such as to determine as referents, respectively, the utterer of “I,” the place at which “here” is uttered, and the time at which “now” is uttered. But since there is no guarantee that a person who utters “I am here now” is at the time he utters “now” in the place he was in when he uttered “here,” the logic of indexicals does not guarantee that any utterance of “I am here now” is true. Of course Kaplan makes a distinction between utterance and occurrence, and one may say that it is not utterances of “I am here now” which cannot be false, but occurrences. But this just formalizes the intuition that one cannot utter “I am here now” falsely, and builds it into the system. While it does seem to be true is that the utterance as a whole will take place wherever it does, whenever it does, I don’t think this can be attributed to the characters of the indexicals. It’s a metaphysical fact that one is wherever one is when one is there; and this fact may be expressed by saying “I am here now”— but only if one doesn’t move! (If one takes the place denoted by “here” to be large enough, perhaps it would be (physically, not logically) impossible for one not to be at the place of utterance at the time of utterance.)

Conclusion Due to space limitations, I’ve said very little about what I take the contents of indexical concepts to be. It’s possible that these contents are all primitive; but it’s also possible that some are analyzable in terms of others (for example, ^here^ as ^this place^, ^now^ as ^the present time^). I’m inclined to think that this is the case, and I’m tempted to construe the content of ^I^ as (something

Indexical Thought

67

like) self, and the contents of ^this^ and ^that^ as the thing I’m attending to. But development of these ideas will have to await another occasion. Here I’ll have to content myself with having cleared the way (if indeed I have) for an internalist theory of indexical content, and, hence, for a phenomenally based theory of indexical content. Such a theory requires a novel way of thinking about indexical thought and its relation to reference, modality and propositional attitude ascription. This is not the currently most popular way of thinking about these issues. But I think it’s workable, and given that there are independent reasons for accepting an internalist, phenomenally-based theory of intentional content, well worth developing.

Acknowledgments For helpful discussion and criticism of earlier versions, my thanks to Mark Balaguer, Dave Chalmers, Elijah Chudnoff, Terry Horgan, Amy Kind, Uriah Kriegel, Peter Lewis, Jesse Prinz, Charles Siewert, Aaron Zimmerman, participants in the 2008 workshop on Consciousness and Thought at the Dubrovnik Inter-University Center, and the 2008 meeting of the Phenomenal Intentionality Research Group at the University of Arizona, as well as audiences at presentations at the CUNY Graduate Center, the University of Nevada, Las Vegas and the University of Miami. Thanks are also due an anonymous referee, for numerous helpful comments and suggestions.

Notes 1. I’ve articulated and defended this view in Pitt 2004, 2009, and 2011. 2. I introduce thought quotes rather than simply using the double quotes of direct quotation to avoid the temptation to construe thinking as inner speaking (a source of much mischief in philosophy of mind). Thought-quoted subsentential expressions refer to conceptual constituents of thoughts. 3. Compare Kaplan on the contents of indexical utterances: What is said in using a given indexical in different contexts may be different. Thus if I say, today I was insulted yesterday and you utter the same words tomorrow, what is said is different. If what we say differs in truth-value, that is enough to show that we say different things. (Kaplan 1989: 500; emphasis added.) 4. As a corollary, thoughts with referentially empty concepts would have no content at all. (See, for example, Evans 1982 and McDowell 1984; see also Crane 2011 for a rejoinder.) 5. Against phenomenal externalists like Dretske, Lycan, and Tye. (See my 2011 and “The Paraphenomenal Hypothesis” (ms).) 6. See, for example, Kaplan 1989, 503. 7. Another is to hold, with Evans (1982) and McDowell (1984), that contents are object-individuated senses, which are referent-dependent without being referent-involving.

68

phenomenal intentionality

8. The example is, of course, adapted from Kaplan 1989. 9. Furthermore, one might argue, my introduction of “thought quotes” obscures this fact, by retreating to the equivalent of direct quotation. One should no more think that ^I’m in agony^ gives the content of my thought than that “I’m in agony” gives the content of my utterance. 10. Needless to say, my arguments have direct application to the semantics of indexical expressions as well as to the defense of any internalist view of mental content. 11. See Pitt 2009 for a defense of the view that cognitive phenomenal types are identical to (as opposed to determinants of ) thought contents. I assume throughout that thought types are individuated by their contents. 12. You can individuate people by their distance from the northeast corner of 36th and Madison, and this might be useful for some purposes; but it doesn’t tell you anything about what people are. 13. Hence, thoughts about phenomenal states are not individuated by their (phenomenally individuated) referents, either. 14. See note 20 for my response to Putnam-style Twin-Earth thought experiments. I present a sustained brief against Burge’s anti-individualist arguments in my (unpublished) paper “The Burgean Intuitions.” 15. I’m not thinking about the species dog, but, rather, its members—the dogs that there are, the individuals we refer to when we say, for example, that too many dogs are victims of neglect and abuse. And I don’t think it’s plausible to suppose that the extension of ^dogs^ is all the dogs there are, were and will be as an explanation for our thinking the same thing. For, surely, the extensions of such concepts can change over time, as different individuals become their members. There are now dogs that did not exist 100 years ago. (And how would one then interpret a thought such as ^Dogs have recently begun talking^?) 16. I think the notion of truth conditions is in fact far too vague to be of any use in determining meaning or content. That a particular object has some property (or is a member of some set) places no constraint whatever on how the object or the property is thought of or described. That is, it doesn’t in itself determine any of the infinitely many ways of thinking of or describing the object or property that a thinker or describer might in fact deploy. 17. Th is intuition would not be defeated by a Russellian construal of defi nite descriptions, since it would still be the case that these thoughts would be about, in the sense of being made true by, different individuals (the values of the existentially quantified variables). 18. Compare John MacFarlane’s excellent paper “Non-Indexical Contextualism” (MacFarlane 2009). MacFarlane doesn’t accept the extension of his view to indexicals (personal communication). 19. True demonstrative concepts such as ^this^ and ^that^ have (or so I claim) extensions relative to an even more deeply world-embedded index, namely, a perceptual-attentive state of an individual. There isn’t space to develop this claim here. 20. There’s also a moral here for enthusiasts of Putnam-style thought experiments. Neither the principle that content determines reference, nor the principle that content (either linguistic or mental) supervenes on facts internal to the individual thinker, is threatened by the difference in extension of my and my twin’s terms or concepts. We can think and mean precisely the same thing in spite of the fact that we’re thinking or talking about (referring to) different substances. 21. From here on I’ll be focusing on the referent-involving version of referent-dependence (since, for one thing, it seems the more commonly accepted view). Adjustment of the points I make to address Evans-McDowell-style referent-dependence should be straightforward. 22. Concepts like ^dog^ are both referentially and conceptually general. 23. It’s at best a default assumption. Imagine an aide to the mayor of Chicago visiting Los Angeles and wondering if the mayor is in his office. Which mayor he refers to is not determined by where he is. Contexts of reference can be chosen by thinkers. 24. And even if I were so constrained, it wouldn’t follow that the content of my ^I^ thought is referent-involving. ^I^ could be construed as having, for each individual, a de facto rigid descriptive content—though at the high cost of each of us having our own unshareable

Indexical Thought

25.

26.

27. 28.

29. 30. 31.

69

self-concept. Indeed, the same would seem to be the case if we individuate our “I” concepts referentially, given that an individual’s ^I^ tokens are, necessarily, self-referential. But it’s a mistake to hold that no two individuals can have the same self-concept. It may be true that each of us is presented to him or herself in consciousness in a way in which we are not, and could not be, presented to anyone else, and in which no one else is, or could be, presented to us. But this is not due to the nature of our self- concepts: you or anyone else can think about yourself in exactly the same way as I can think about myself. It simply doesn’t follow from the fact that I can’t think about you by tokening ^I^ that you can’t think about you by tokening the very same concept. Indeed, the way in which one is presented to oneself in consciousness—namely, self-consciousness—is a type as well: though it’s a relation one can stand in only to oneself, others can stand to themselves in precisely the same relation. (It’s like identity. There isn’t a distinct identity relation for every object.) As we learned from Kaplan, it’s essential to distinguish the logical status of indexical tokens from the logical status of their contents. I can’t falsely think ^I am me^; but it’s not a necessary truth that ^I^ refers to me. Similar points can be made about names. Suppose I think, truly, ^Aristotle was a shipping magnate^, and then immediately wonder ^Is it possible that Aristotle was a shipping magnate?^ I might be wondering about Aristotle the shipping magnate (in which case the answer is, trivially, yes); but I might also be wondering about Aristotle the philosopher, in which case the answer isn’t trivial. What I’m wondering, in either case, is whether it’s possible that someone who bears the name “Aristotle” was a shipping magnate. Th at is, I’m wondering the same thing in both cases—albeit about different individuals. (Analogous considerations apply to names in speech.) With the exception of ^I^, of course—though as argued above this token referential rigidity doesn’t prevent ^I^ from having a constant content. Again, parallel considerations apply to names (both in thinking and speaking), beginning with the observation that the referent of a name (as it might be, which causal chain is relevant) is relative to context. Name-types are as referentially ambiguous (or indeterminate) as indexical types, as are their tokens. In the absence of a specified bearer of “John Smith,” ^Possibly John Smith is a philosopher^ is true iff there’s a possible world in which a bearer of “John Smith” is a philosopher. (This is analogous to ^Possibly he’s not tall^ being true at a world iff a male in that world isn’t tall.) It’s true of some particular John Smith we might have in mind iff there’s a possible world in which he is a philosopher. But there are lots of John Smiths (all, pace Kripke, with the same name). I can think the same thing of each and every one of them by thinking ^Possibly John Smith is a philosopher^, and I can sensibly wonder whether what I’ve thought of one of them in this world is true of another one of them in this or some other world. If I describe a possible world using the name “John Smith” as it is used in English, I might be referring to any of the many John Smiths. (I have a good deal more to say about names, and names in thought; but it will have to wait.) Analogously, the same result ensues if I think (or try to think) ^Glenn is an imbecile^ but have no individual in mind as the referent of ^Glenn^. I owe these objections to David Chalmers, in conversation. The problem is not dissolved by adopting a character-content-like distinction for concepts; for then thoughts with contradictory characters could have true content, and thoughts with analytic characters could have false contents (for example, the thoughts ^You are not you^ and ^You are you^) thought of two different individuals).

References Crane, T. 2011. “The Singularity of Singular Thought.” Proceedings of the Aristotelian Society Supplementary Volume 85: 21–43.

70

phenomenal intentionality

Evans, G. 1982. The Varieties of Reference. Oxford: Clarendon Press. Frege, G. 1918. “Thought,” reprinted in The Frege Reader, ed. M. Beany. Malden, MA : Blackwell, 1997: 325–345. Kaplan, D. 1989. “Demonstratives.” In Themes from Kaplan, ed. J. Almog, J. Perry, and H. Wettstein. Oxford: Oxford University Press, 481–614. MacFarlane, J. 2009. “Non-Indexical Contextualism.” Synthese 166: 231–250. McDowell, J. 1984. “De Re Senses.” Philosophical Quarterly 34: 283–294. Pitt, D. 2004. “The Phenomenology of Cognition, Or, What Is It Like to Th ink Th at P?” Philosophy and Phenomenological Research 69: 1–36. Pitt, D. 2009. “Intentional Psychologism.” Philosophical Studies 146: 117–138. Pitt, D. 2011. “Introspection, Phenomenality and the Availability of Intentional Content.” In Cognitive Phenomenology: New Essays, ed. M. Montague and T. Bayne. Oxford: Oxford University Press. Pitt, D. Unpublished manuscript. “The Burgean Intuitions,” www.calstatela.edu/faculty/ dpitt/ Pitt, D. Unpublished manuscript. “The Paraphenomenal Hypothesis,” www.calstatela.edu/ faculty/dpitt/.

4

Phenomenal Presence C h r istoph e r F r e y

1. Introduction One frequently encounters the thesis that experience is transparent or diaphanous in contemporary philosophical discussions of perceptual experience. Michael Tye provides a recent and representative statement of this thesis: When you introspect your visual experience, the only particulars of which you are aware are the external ones making up the scene before your eyes. . . . Your awareness is of the external surfaces and how they appear. The qualities you experience are the ones the surfaces apparently have. Your experience is thus transparent to you. When you try to focus upon it, you ‘see’ right through it, as it were, to the things apparently outside and their apparent qualities. (Tye [2002] 139) According to experiential transparency’s advocates, all that perceptual experience even seems to present us with are objective entities and their sensible characteristics. No amount of phenomenological reflection will enable us to appreciate the intrinsic features of perceptual experience as such. The question of whether all experiences are transparent occasions spirited disagreement. These disagreements, however, are premature. I will argue that the most common interpretation of experiential transparency’s significance is laden with substantive and ultimately extraneous metaphysical commitments. I divest this infl ated interpretation of its unwarranted encumbrances and consolidate the precipitate into a position I call Core Transparency (§2). Core Transparency is a thesis about experience’s presentational character. We open our eyes and a world is before us. Someone strikes a tuning fork, and a sound is simply present. In all sensory modalities, the objects of perceptual experience are there, present to us, in a way that the objects of most beliefs and judgments are not. According to Core Transparency, it is in the disclosure of 71

72

phenomenal intentionality

that which is central to experience’s presentational character, an intrinsic and irreducibly phenomenal aspect of experience I call phenomenal presence, that transparency’s significance principally consists (§3). Though valuable in itself, the consequences of this analysis extend well beyond the clarity it provides to discussions of transparency. For phenomenal presence is uniquely positioned to illuminate the relationship between perceptual experience’s most important features: its intentionality and its phenomenality. One popular account of how these features are related, namely, representationalism, asserts, roughly, that an experience’s representational features completely determine its phenomenal features. The thesis that all experiences are transparent is often taken to support representationalism; representationalists maintain that their account best (or uniquely) explains such transparency.1 I will argue that experiential transparency is far from being a solid foundation upon which representationalists can rest their arguments. The phenomenon is, when interpreted properly, not only among the view’s greatest obstacles, it supports a converse orientation. The position I defend comprises two main claims. 1. The representational features of experience, understood in isolation from experiential phenomenality, neither constitute nor explain phenomenal presence. Consequently, the representational features of experience neither determine completely nor explain exhaustively experiential phenomenality (§4). 2. Phenomenal presence is not representational, but is nevertheless the minimal realization of experiential intentionality (§5).

2. The Purported Significance of Experiential Transparency Though transparency’s significance is subject to a multiplicity of (often ambiguous and heterogeneous) interpretations, one can discern an emerging consensus over its general form. (T) The sensuous elements that one phenomenally appreciates in an experience are (with varying emphases) always appreciated as (i) public, (ii) objective, (iii) mind-independent, and/or (iv) external (that is, distally located).2 According to (T), when one undergoes, say, a visual experience as of a tree, the sensuous green one appreciates is appreciated as a quality of the tree’s

Phenomenal Presence

73

leaves. One appreciates the sensuous green as being, being instantiated in, or being about an objective entity in one’s external environment. Moreover, no amount of phenomenological reflection will alter the apparent externality of an experience’s sensuous elements or reveal novel elements with a different apparent ontological status. So if transparency holds generally, one will never appreciate an experience’s sensuous elements as intrinsic features of either the experience itself or oneself qua experiential subject.3 This interpretation is, at least in comparison to many characterizations that were advanced during the initial stages of transparency’s recent era of popularity, narrowly phenomenological. It concerns only apparent ontological classifications and does not involve a commitment to Revelation, that is, the thesis that the intrinsic nature of a sensuous element is fully revealed by the phenomenal appreciation of that element in a standard experience (cf. Johnston [1992]). According to these earlier interpretations, undergoing or reflecting upon an experience discloses the apparent ontological status of the sensuous elements one appreciates therein and Revelation guarantees that these elements are as one appreciates them to be.4 To abandon Revelation, as (T) does, is to recognize that phenomenological reflection, by itself, is metaphysically neutral. This neutrality is twofold. First, phenomenological reflection, by itself, does not disclose the ontological status of an experience’s sensuous elements. (T)’s advocates readily admit that phenomenological reflection does not immediately yield conclusions about the metaphysical status of what one appreciates in experience. For example, Gilbert Harman concedes that “one might be aware of intrinsic features of experience without being aware of them as intrinsic features of experience” (Harman [1990] 42). Similarly, both Michael Tye and Alex Byrne hold that transparency, on its own, does not speak against sense-datum accounts of experience.5 So even if we always appreciate an experience’s sensuous elements as being or qualifying physical objects in a publicly accessible environment, it may turn out that they are, as a matter of fact, instantiated by amalgams of sense-data or are intrinsic features of one’s experiential states.6 Second, phenomenological reflection, by itself, does not disclose the ontological status of experiential episodes themselves. Perhaps one has an experience with a particular phenomenal character in virtue of being in a representational state with a special kind of content or a special functionally specified role. Perhaps one has this experience in virtue of standing in some primitive relation of acquaintance or direct awareness to an appropriate class of entities. Perhaps one has this experience by virtue of the divine dispensation of an omnipotent god. Whatever the source, it is invisible. The means by which a scene becomes experientially present to one is not itself phenomenally appreciable. Despite the relative restraint of (T), many who accept it nevertheless aspire to the yield of its metaphysically indulgent counterpart. If (T) is correct, then

74

phenomenal intentionality

undergoing an experience always involves the appreciation of something as having certain features, namely, those that constitute membership in one or more of the classes enumerated in (T). This is best explained, the argument goes, if experiential phenomenality is representational in nature. That is, for an experience to possess a phenomenal character is for it to make a claim that something is a certain way and such claims are evaluable for veridicality. These contentful experiences, the argument continues, will be radically (and problematically) misleading unless things are, at least in standard cases, as we appreciate them to be.7 It should concern those who advance such arguments that parallel arguments exist for incompatible views. For example, many argue that transparency is best explained by an explicitly non-representational form of naïve realism.8 But of even greater concern is this: the narrowly phenomenological interpretation of transparency against which all parties have agreed to measure the explanatory virtues of their preferred accounts is, I contend, simply false. Let us begin with an examination of two rare but illustrative sorts of experience that serve as counterexamples to (T). Case 1: Ganzfeld Experiences. A ganzfeld is a visual field that is completely permeated with a constant and homogeneous sensuous color, say, a determinate shade of blue.9 The advocate of transparency is correct to this extent: the sensuous blue that is present in a ganzfeld experience is not appreciated as being an intrinsic property of the experience itself or of oneself qua experiential subject. Does this mean that one appreciates the sensuous blue as qualifying (or as being) an objective, mind-independent entity? Not at all. When one undergoes such an experience, one does not appreciate anything as possessed of the phenomenal characteristics that constitute the alleged world-disclosing phenomenality of our ordinary perceptual experiences. There is no figure-ground contrast, no manifestation of diachronic perceptual constancies, and no opportunity for perspectival variation. Relations of relative, spatial location are either severely limited or altogether absent, so one does not even appreciate the sensuous expanse as being a properly extended region of space.10 Finally, one need not (and often does not) appreciate the ganzfeld as distally located; at most, one can say that one appreciates the sensuous blue as being before one, where “before” does not connote “in front of” or any other notion that involves apparent spatial egocentricity.11 Case 2: Spatially Punctiliar Experiences. Damage to one’s occipital cortex can result in the diminution of one’s visual field. That which is lost is not replaced with darkness, the so-called brain grey that permeates our visual field when external optical stimulation is cut off. The visual system registers nothing in these lost regions; the field itself simply shrinks.12 It is possible in principle for one’s brain to atrophy in such a way that one undergoes visual experiences

Phenomenal Presence

75

that are spatially punctiliar.13 These are not experiences as of a point of light in a sea of darkness, that is, an isolated phosphene. One who possesses such a visual system would undergo visual experiences in which a single phenomenal point, and nothing more, is appreciable.14 A sensuous element is present in a spatially punctiliar experience. But one does not appreciate the sensuous point as possessed of features that constitute one or more of the classes enumerated in (T). Moreover, this case undermines not only the necessity of apparent three-dimensional spatiality in experience (as ganzfelds do), but that of apparent two-dimensional spatiality as well (as ganzfelds may not). That is, the possibility of spatially punctiliar experiences shows that a widely held constraint, namely, “If x is visually aware of y (if x sees y), then y must look extended to x,” is false.15 What conclusion should we draw from the possibility of these sorts of experience? One response is to maintain the interpretation of transparency expressed in (T), but abandon its universality. These and other atypical experiences show, according to this approach, that transparency is a normal but not ubiquitous phenomenon. But I think this approach is mistaken. The two sorts of experience we have considered are indeed counterexamples to (T). But (T)’s inadequacies extend well beyond these atypical cases. I noted earlier that an experience’s phenomenal character is compatible with a wide range of possibilities about both the experience’s nature and the nature of the sensuous elements one appreciates therein. But the retreat to apparent ontological status does not place its advocate on safe ground. For a typical experience’s phenomenal character is not only compatible with, say, a sense-datum account of perceptual experience, it is exactly what we ought to expect from a sense-datum account. Competing philosophical accounts of perception—for example, representationalism, naïve realism, projectivism, and sense-datum theories—issue identical phenomenological “predictions.” So even in unexceptional cases, experiential phenomenality, by itself, is silent with respect to the classifications in (T).16

3. Core Transparency and Phenomenal Presence Should we conclude then that no experiences are transparent? To do so would be to deny that we can educe any important insight from the sort of phenomenological observations to which the champions of experiential transparency appeal. This, I believe, is to go too far. Though transparency’s significance is far more modest than is commonly proclaimed, its advocates are pointing to a genuine phenomenon. The proper response, I contend, is to uphold the thesis that all experiences are transparent, but to employ an interpretation of transparency that is free of unwarranted commitments.

76

phenomenal intentionality

The seed of an appropriately parsimonious interpretation is already contained in our examinations of ganzfeld and spatially punctiliar experiences. It seems that no matter how simple or peculiar an experience may be, we always appreciate its sensuous elements as being present or before us in a way that the objects of most beliefs and judgments are not. That is, when we phenomenally appreciate a sensuous element in an experience, we appreciate it as being both something other than ourselves and as standing in opposition to ourselves. This view, which I call Core Transparency (CT), can be formulated in two interdependent ways. (CT1) The sensuous elements that one phenomenally appreciates in an experience are always appreciated as other. (CT2) The sensuous elements that one phenomenally appreciates in an experience are never appreciated as being, being instantiated in, or being about the self qua experiential subject (or a state/mode thereof). I will call the experiential nexus of self and other to which this pair of formulations refer phenomenal presence. Phenomenal presence is central to the presentational character of experience and it is in the disclosure of phenomenal presence that transparency’s significance principally consists.17 If Core Transparency is to avoid (T)’s fate, we must be careful to distinguish the predicates “__ appreciates x as other” and “__ appreciates x as (being . . . the) self” from similar expressions that we have already jettisoned, for example, “__ appreciates x as objective” and “__ appreciates x as distally located.” For (T)’s failure stems not from the particular classes or properties it employs, but from the very appeal to classification. If phenomenal presence involved the attribution of properties to, or the classification of, an experience’s sensuous elements, then it would determine a veridicality-evaluable claim. But as we have already seen, the metaphysical neutrality of phenomenological reflection guarantees that transparency, by itself, makes no such claims. What then in the significance of (CT)? We can, as a first pass, describe the situation as follows. To undergo a sensuous experience is (in part) to appreciate an element as other or as before one; it is (in part) to appreciate a manifest opposition between the self—that before which the other is present—and the other—that which is present before the self. But phenomenal presence does not consist in the instantiation of some relation, say, experience e presents y to z by members of two distinct kinds, namely, other and self. The distinction between self and other is rather an oblique communication of sensuous experience’s basic and intrinsic phenomenal structure. Sensuous experiences are phenomenally articulate unities and to appreciate something as other is to appreciate its invariant position within this bipartite, phenomenal articulation. I will elucidate this view by examining the two formulations of (CT) in turn.

Phenomenal Presence

77

(CT1)—According to (CT1), experience is necessarily presentational; there can be no appreciation of a sensuous element in experience that is not also an appreciation of it as other. This is true even of pain experiences and cases of perceptual imagination. Insofar as one appreciates a sensuous element in such experiences, one appreciates it as other. There may be, in addition to this manifest otherness, a genuine phenomenal basis for associating some of these elements more intimately with the self than others: an appearance of subjectivity or interiority may be due to, respectively, a distinctive sense of ownership in bodily sensation or an appreciation of the subordinacy of imaginative phenomenality to the spontaneity of our mental agencies. But the sensuousness of these experiences guarantees that they involve an apparent confrontation with something other than oneself. This is not because sensuous experience always involves the appreciation of something as having features that would ground a classification of kind, namely, the kind other, but because phenomenal presence does not involve veridicality-evaluable classifications at all. It is a category mistake to assess phenomenal presence in terms of success and failure. To appreciate a sensuous element as other is not to appreciate some mark possessed by that element; it does not consist in the apparent exemplification of some property or the apparent satisfaction of some criterion. It is rather to appreciate the sensuous element’s position within the experience’s basic phenomenal articulation. The occupation of such a phenomenal position by a sensuous element in an experience is no more subject to conditions of satisfaction or veridicality than is the occupation of a grammatical position by a proper name in a sentence.18 (CT2)—According to (CT2), the phenomenal appreciation of a sensuous element in experience can never be an appreciation of our selves, or the states/ modes thereof, as such. Hume makes this point vividly when he reports on his attempts to do so: When I enter most intimately into what I call myself, I always stumble on some particular perception or other, of heat or cold, light or shade, love or hatred, pain or pleasure. I never catch myself at any time without a perception, and never can observe any thing but the perception. (Hume [1739/1978] I.vi.6) As a matter of ontology, we are personal selves—living, embodied human beings with complex psychological histories located within an objective, ordered world of public, physical objects. Perhaps it is possible for one to appreciate a sensuous element in experience that is, as a matter of fact, a feature of oneself qua personal. That is, there is nothing incoherent about the sensuous manifestation (and the correlative appreciation as other) of an intrinsic feature of a personal self. But as a matter of what is revealed in experience’s

78

phenomenal intentionality

presentational character, experiential subjects are phenomenal selves. Thus conceived, the other is simply that which is present before the self; the nature of the phenomenal self is exhausted by the position it occupies within experience’s phenomenal articulation.19 So the phenomenal self is not sensuously manifest in experience. For to be sensuous is, in part, to be appreciated as other, and to be appreciated as other is to occupy a particular position within experience’s phenomenal articulation. The phenomenal self cannot occupy this position, it cannot be sensuous and present before one, without ceasing to be what it is. Any attempt to appreciate the phenomenal self or to appreciate the features, including structural features, of experience itself in the same manner as we appreciate an experience’s sensuous elements is guaranteed to fail.20 Phenomenal presence is not sensuous. But it is phenomenally appreciable. It is a mistake to suppose that an appreciation of phenomenal presence requires one to turn away from that which is present in experience and toward some isolated interiority. Sensuous experience is universally and intrinsically directed toward the other. But it is only through the presence of the other that the phenomenal self exists at all. The intrinsic orientation of experiential phenomenality toward the other phenomenally embodies its converse. To focus, as we must, on the other in experience is, ipso actu, to appreciate its position with respect to our phenomenal selves. This is what the phenomenological observations that have so impressed the advocates of transparency reveal—not the apparent natures of that which is present in experience, but the aspect of experiential phenomenality that constitutes this appreciable presence itself.

4. The Explanatory Inadequacy of Representationalism This essay’s ultimate goal is to illuminate the relationship between the intentionality and the phenomenality of perceptual experience. Phenomenal presence bears a twofold significance for this project. Its first consequence is negative. I will argue in this section that phenomenal presence neither depends essentially upon nor is explanatorily grounded in the non-phenomenal, representational features of experience. Its second consequence is positive. Phenomenal presence is not representational, but is nevertheless the minimal realization of experiential intentionality (§5). One account of how the intentionality and phenomenality of experience are related, namely representationalism, is overwhelmingly popular. It comprises (i) a representational account of experiential intentionality and (ii) a thesis of ontological supervenience.

Phenomenal Presence

79

Representationalism: (i) To undergo an experience is, inter alia, to be intentionally directed upon an entity in virtue of being in a state with one or more representational contents (that meet various conditions). (ii) An experience’s phenomenal features supervene upon either (a) features of the experience’s representational content(s), (b) features of the experiential state in virtue of which it has its representational content(s), and/ or (c) represented features of the entities the experience represents. As Alex Byrne correctly notes, representationalism “does not take a stand on whether phenomenal character can be explained in terms of, or reduced to, [representational features]—at least it doesn’t if these claims don’t follow from the mere fact of supervenience” (Byrne [2001] 204; cf. Horgan [1993]). But if the representationalist is to contribute to the project of understanding experiential phenomenality, she must explain why each basic and distinctive type of phenomenal feature supervenes upon a specific class of the experience’s representational features. Representationalism has the resources to provide adequate explanations for many of the relevant supervenience relations. In particular, it can explain how many of an experience’s sensuous elements supervene upon specific features of the experience’s representational content(s). That is, representationalism can in principle explain: (i)

The scope of phenomenal appreciation—Why do I appreciate a sensuous patch of blue (rather than a sensuous patch of red) in the experience I am currently undergoing and why do I appreciate it as being circular and located three feet to my left (rather than as being rectangular and located three feet to my right)? (ii) The phenomenal differences between experiences—In virtue of what does the phenomenality of my experience of a red patch differ from the phenomenality of my experience of a blue patch? In virtue of what does the phenomenality of my visual experience of a circular object differ from the phenomenality of my tactile experience of the same circular object? (iii) Many distinctive features of experiential phenomenality—What explains the complexity, richness, determinacy, and particularity of, say, my typical visual experiences?21 But even if we grant that experience’s representational features not only determine but explain these aspects of experiential phenomenality, we have not thereby conceded that the representationalist has successfully executed her explanatory project. For the representationalist must not only explain what is phenomenally present in experience; she must explain the phenomenality of this presence itself. There are numerous representational states, such

80

phenomenal intentionality

as beliefs, judgments, desires, and other propositional attitudes, that lack a presentational character altogether. What representational facts explain the manifestation of phenomenal presence in experience and the absence of phenomenal presence in other contentful states?22 There are several approaches available to the representationalist. I will focus on the shortcomings of three: (i) appeals to content, (ii) appeals to functional role, and (iii) appeals to primitive forms of representation. I will then offer reasons for thinking that these attempts do not fail because of their respective idiosyncrasies. Representationalism’s explanatory limitations are, I contend, endemic to the theory. Approach 1: Appeal to Content. The representationalist can advert to the distinctive kinds of representational content experiences possess, for example, non-conceptual, analog, pictorial, and so on, to explain experience’s presentational character. But this view is a non-starter. First, the candidate contents can be attributed to representational states that lack a presentational character. For example, it is not unreasonable to think that there are sub-personal states generated in vision that non-conceptually represent changes in light intensity (Tye [2000] 62; cf. Stalnaker [1998]). So the possession of such contents is not sufficient to demarcate phenomenally presentational states. Second, and more important, being present is not a property of experiences’ sensuous elements. According to representationalism, the phenomenal appreciation of x in an experience is constituted (in part) by the inclusion of x, or an entity that determines x, in the content of the experience’s constituent representation. To say that an experience, in addition to representing an object, must represent that object as being present, is to introduce an unnecessary and problematic redundancy. For these reasons, most representationalists do not attempt to explain an experience’s presentational character in terms of its representational content but rather offer explanations that appeal to the distinctive characteristics of experiential representation itself. That is, they maintain that, [representationalism] is not the view that the content of an intentional state determines its nature qua mental state without remainder . . . it is the doctrine that the content of an experience plus the fact that the experience represents the content as obtaining in the way distinctive of perceptual representation are what determines the experience’s nature without remainder. (Jackson [2007] 58; cf. Chalmers [2004] and Martin [2002] 378) These attempts typically appeal to the distinctive functional role of experiential representation (approach 2) or introduce primitive forms of experiential representation (approach 3).

Phenomenal Presence

81

Approach 2: Appeal to Functional Role. Representationalists often claim that it is experience’s functional role that makes experiential representation distinctive. For example, Tye claims that experiences and feelings, qua bearers of phenomenal character, play a certain distinctive functional role. They . . . stand ready and available to make direct impact on beliefs and/or desires. (Tye [2000]62) A subject phenomenally appreciates something in virtue of being in an experiential state that is, in the normal run of things, a maximally proximal causal trigger for the formation of beliefs, desires, and purposive actions. Representationalists can invoke functional roles of this sort, that is, roles that involve personal-level cognitive/practical agencies, to explain experience’s presentational character. For the subject it is as if the objects are right there before him. [The representationalist] seeks to explain this aspect of experience by reference to the kind of state of mind experiencing is. According to him, it is just that state of mind which is liable to fi x the subject’s beliefs about how his environment must be, and hence is a state of being presented to as if things are so. (Martin [2002] 399) It is true that our perceptual experiences are apt or poised to produce (authoritative) perceptual beliefs and influence action. But the representationalist’s approach reverses the proper order of explanation. Though I do not wish to endorse its representationalist accoutrements, the spirit of the following quote goes some way toward motivating this stance. In my view, it’s not the irresistibility of our perceptual beliefs, nor the nature of our concepts, which explains why our experiences give us the immediate justification they do. Rather, it’s the peculiar “phenomenal force” or way our experiences have of presenting propositions to us. Our experience represent propositions in such a way that it “feels as if” we could tell that those propositions are true—and that we’re perceiving them to be true—just by virtue of having them so represented. (Pryor [2000] 547 fn. 37) It is phenomenal presence itself that is both the source of our inclination to form beliefs on the basis of experience and the (partial) source of the warrant associated with these beliefs. The doxastic role of experience cannot be explained without adverting to experience’s presentational character. But representationalism, even versions according to which the phenomenal features

82

phenomenal intentionality

of experience are identical to (a subset of) its representational features, must be able to explain experiential phenomenality in a manner that avoids phenomenal notions and characterizations altogether. Approach 3: Appeal to Primitive Representational States. The representationalist can explain why experiences manifest a presentational character while other contentful states do not if experience involves a primitive and sui generis form of representation. On this approach, the distinctive manner in which an experience represents its content as obtaining is not determined by its functional role. In fact, it is not determined by anything at all. Though not himself a representationalist, Mark Johnston expresses this view when he says that, “visual experience is a sui generis propositional attitude—visually entertaining a content concerning the scene before one’s eyes” (Johnston [1992] 172–3; cf. Chalmers [2004]). According to this third approach, to be in a state in which something is phenomenally present is to represent experientially that something is the case. Nothing more can be said. If correct, the representationalist would be able to explain an experience’s presentational character in terms of its representational features. But this approach deprives the representationalist’s explanatory project of its value. The invocation of a primitive and sui generis kind of representational state that is essentially presentational in one’s explanation of experience’s presentational character is no better than the invocation of a virtus dormitiva; it is to abandon the view that there is an independent and relatively basic level of explanation for facts about experiential phenomenality. Though these three approaches do not exhaust representationalism’s explanatory resources, the remaining options will fare no better. For if our interpretation of transparency is correct, then it undermines any representationalist explanation of experience’s presentational character. I have argued that the phenomenological observations that ground the thesis of experiential transparency do not reveal the apparent natures of that which is present in experience or of experience itself, but disclose what is central to experience’s presentational character. This non-sensuous aspect of experiential phenomenality, namely phenomenal presence, is the manifestation of sensuous experience’s basic and intrinsic phenomenal structure. It may still be the case that experiences are essentially representational. But if the nature of experience is entirely determined by its representational features, then this class will necessarily include representations that are intrinsically and irreducibly phenomenal. That is, if representationalism is true, then it must take the form described in the third approach canvassed earlier. But if representationalism requires the introduction of primitive, presentational forms of representation, then it will be unable to provide an adequate and exhaustive explanation of experiential phenomenality. The representationalist can retreat to the less ambitious claim of ontological supervenience. For phenomenal presence poses no threat to the thesis that

Phenomenal Presence

83

experiential phenomenality supervenes upon experience’s representational features. But while relations of supervenience are often uninformative, this particular instance is exceptionally so. Phenomenal presence is not just a universal feature of sensuous experience, it is a necessary feature; one cannot appreciate a sensuous element in an experience that lacks a presentational character. So facts involving phenomenal presence are trivially entailed by facts involving experience’s representational features. Without an adequate explanation of this supervenience relation, representationalism can say next to nothing about a ubiquitous feature of experiential phenomenality. Perhaps the only option left is the bold recognition of representationalism’s limited explanatory power: Why then do experiences, including hallucinatory experiences, have a presentational phenomenology while thoughts do not? . . . My answer to this question is that there is no answer. (Pautz [2007] 519)

5. Phenomenal Presence and Experiential Intentionality Given our rejection of representationalism, the claim that phenomenal presence embodies an autonomous and basic form of experiential intentionality will strike many as confused. For “intentional” and “representational” are often used synonymously. I first provide a characterization of intentionality which clearly distinguishes it from representation (§5.1). I then argue that phenomenal presence satisfies this characterization (§5.2).

5.1. Intentionality Four features are commonly associated with (non-derivative) intentionality. (Int) Intentionality is that aspect of an occurrent, categorical state or event that consists in (i) its being of, about, or directed upon an entity [directedness] (ii) as an exemplar or instance of some general property, relation, kind, or category [ generality] (iii) from a particular perspective or under a particular aspect [aspectual shape] (iv) to or for its subject [personal subjectivity] Directedness. It is in the nature of intentionality to be directed beyond itself, beyond the individual (or a state thereof) that possess it. Intentional states that are, as a matter of fact, directed upon themselves are possible.

84

phenomenal intentionality

It is even possible, as Brentano held, that all intentional states are, in addition to being directed beyond themselves, self-directed. But when such cases occur, the identity that obtains between that which is intentionally directed and that upon which it is intentionally directed is entirely accidental. Such states are directed upon themselves qua other. Generality. Our intentional states are not directed upon entities simpliciter. They are always directed upon entities as exemplars or instances of some property, relation, kind, or category that is capable, in principle, of applying to various particulars. A capacity that yields states that are intentionally directed upon entities without being directed upon them as being some general way is traditionally called intellectual intuition (cf. Kant [1817/1996] 28:1051). Such a capacity cannot be found this side of Heaven. Aspectual shape. Our intentional states are always directed upon entities from a particular perspective or under a particular aspect.23 This is easily seen in perceptual experience. Our perceptual capacities are divided into distinct sensory modalities—a single property, say, sphericity, can be experienced either visually or tactilely. Moreover, they are always exercised from a particular point of view and provide, at best, a partial and incomplete perspective on that which we perceive.24 Personal subjectivity. Though it is convenient to attribute intentionality to an organism’s states and events, a convenience that I have already taken advantage of, the feature’s proper bearer is individual organisms. It is individual organisms inhabiting and coping with their environment that perceive, believe, judge, desire, reason, and know. To attribute an intentional state to an individual is to communicate indirectly that the individual is intentionally directed upon an entity. So if a system involved in an organism’s perceptual experience is entirely modular in its output, that is, if the states it yields are neither attributable to the whole organism nor first-personally available to guide the activity or other responses of the whole organism, then these states do not possess intentionality.25 Some of these conditions, especially personal subjectivity, are controversial. But if we momentarily prescind from this condition, it is clear that (Int) serves as the core of the comparatively substantive accounts of intentionality one typically encounters. I will briefly canvass the two most prevalent ways in which this minimal characterization of intentionality is developed and extended with respect to experience. First, as we have already seen, one can provide a broadly semantical/representational account of intentionality that takes the basic other-directedness of an intentional state to consist in the possession of one or more representational

Phenomenal Presence

85

contents that determine conditions of correctness/satisfaction.26 According to this conception, a state is intentional only if there are ways the world can be that render it veridical. This captures the generality of intentional directedness because such contents attribute some general property, relation, kind or category to one or more entities. Moreover, this conception can accommodate the aspectual shape of intentionality since a state’s representational contents can have principles of type-individuation that are more fine-grained than the experience’s satisfaction condition.27 Alternatively, one can provide a relational account of intentionality that takes the basic other-directedness of an intentional state to consist in the establishment of a simple, non-representational relation between the state (or the individual whose state it is) and one or more entities that partly constitute this intentional relation.28 On this conception, perceptual success does not presuppose experiential intentionality but consists in its establishment. That is, particular experiential episodes establish connections with entities and, in so doing, make them available to the perceiver as objects for attention, recognition, thought (especially singular, demonstratively expressible judgments), and purposive action. The relata upon which our experiences are intentionally directed occur as exemplars or instances of various general features and the intentional relation obtains from a particular point of view thereby satisfying the conditions of generality and aspectual shape. But standing in such an intentional relation to an entity is no more evaluable for correctness or veridicality than is standing in the relation kicking to a soccer ball.29 So directedness upon an entity can involve the possession of representational contents and directedness upon an entity can involve the establishment of certain non-representational relations. But (Int) captures the common commitment of both views; the representational and relational accounts simply provide different analyses of this conceptually prior characterization of intentional directedness. To analyze experiential intentionality in terms of veridicality-evaluable, representational contents is merely to embrace one among the several alternatives left open by our characterization. 30

5.2. Experiential Intentionality The neutrality of (Int) with respect to the various, relatively-substantive accounts of intentionality is mirrored by the neutrality of (CT) with respect to the various, relatively substantive accounts of the nature of experience’s presentational character. This opens up the possibility that phenomenological reflection reveals an essentially phenomenal form of intentional directedness. But it is more than a possibility. Phenomenal presence satisfies the conditions for intentionality encapsulated in (Int): (i) The phenomenal appreciation of sensuous elements in experience is universally and intrinsically other-directed. (ii) The

86

phenomenal intentionality

sensuous elements one appreciates exemplify one or more appreciable general characteristics. (iii) These sensuous elements are always appreciated under an aspect; phenomenal presence affords a primitive “point of view” on that which is present. (iv) The sensuous elements we appreciate in experience are present to the self; we appreciate these sensuous elements as being both other than ourselves and as standing in opposition to ourselves. Consequently, phenomenal presence realizes a basic and non-derivative form of intentional directedness. Several philosophers argue that experiences possess some form of phenomenal or experiential intentionality. But most of these accounts take this intentionality to consist in the possession of one or more phenomenal contents, that is, veridicality evaluable contents that are determined by a state’s phenomenal features. A subset of these philosophers make the stronger claim that these phenomenal contents are fully constituted by the state’s phenomenal features and cannot be reduced to its non-phenomenal features.31 Given the result of the previous section, namely, that the representational features of experience neither constitute nor explain phenomenal presence, it follows that accounts according to which the intentionality of experience is essentially representational are too demanding. This is the case regardless of whether these representational contents are phenomenal or not. Experiences may still possess phenomenal contents, but there is a form of phenomenal intentionality that is distinct from and prior to phenomenal representation.32 Despite these differences, the present proposal shares at least this much in common with alternative views of phenomenal intentionality: experiential phenomenality contains a basic form of intentional directedness and any attempt to provide an account of experiential intentionality that ignores phenomenality (and vice versa) cannot succeed.

Acknowledgments Earlier versions of this essay were presented at Barnard College and the University of Pittsburgh. I am grateful to the audiences for their responses. Discussions with Endre Begby, Jim Conant, Chris Hill, Uriah Kriegel, John McDowell, and Karl Schafer have led to significant improvements. Special thanks in this regard are due to Anil Gupta.

Notes 1. Prominent defenses of representationalism along these lines include Harman [1990], Tye [1995], and Byrne [2001]. The motivation for representationalism extends well beyond its alleged explanatory advantages with respect to transparency. For example,

Phenomenal Presence

2.

3.

4.

5. 6.

7. 8. 9.

87

many representationalists, especially those impressed by cognitive science’s explanatory successes, argue that their account facilitates the establishment of a materialist account of the mind. Four important terminological clarifications: (i) The sensuous elements in an experience are the most salient (and according to some the only) class of phenomenally appreciable elements in experience. This class is typically introduced by an enumeration of paradigmatic instances, say, the way the redness of a Red Delicious apple looks when one sees it, the way middle C sounds when one hears it being played on a Bösendorfer piano, and the way a pain feels when one experiences a pin pierce one’s fi nger. Our discussion will show that the total phenomenal character of an experience can contain more than its sensuous phenomenal character; we may phenomenally appreciate more than sensuous qualities in experience. (ii) I use the expression “phenomenal appreciation” in a way that is neutral with respect to different accounts of how one comes to be aware of an experience’s sensuous elements. On my own view, phenomenal appreciation places no demand on its subject beyond that of undergoing a relevant experience. To undergo an experience in which a sensuous element is present is to appreciate phenomenally that sensuous element. Phenomenal appreciation should therefore be distinguished from both introspection and conscious attention. But these further commitments are not required for the task at hand. (iii) I use both “element” and “entity” in a metaphysically neutral way to refer to any disjunction of object, state, event, property, and so on (context will usually make it clear which categories are relevant), with no restrictions on the natures of the referents—Meinongian objects, sense-data, abstracta, physical bodies—all count as entities/ elements. For ease of expression, however, I will often simply speak of sensuous qualities. (iv) I take no stand on how the adjectives “public,” “objective,” “mind-independent,” and “external” are related. (T) is stated in such a way as to cover various interpretations of these expressions. (i) Though phenomenological reflection can affect the sensuous elements that one phenomenally appreciates in an experience—it can, and commonly does, increase their intensity, vividness, and resolution—such reflection, according to (T), does not effect a change in the apparent nature of the experience’s appreciable elements and is not a window onto new ontological domains. (ii) Some attribute an epistemological significance to transparency, for example, the claim that one’s access to the intrinsic features of one’s experiences is indirect. Th is essay, however, is concerned primarily with transparency’s metaphysical import. (iii) The rider “qua experiential subject” allows a visual experience of, say, the color of one’s leg to count as transparent. “Appeals to transparency appear to involve the following thought: just by having a perceptual experience, the perceiver is placed in a position whereby he or she is able to classify the ontological category of what is manifest in experience. The nature of experience is supposed to be the kind of thing that can be discerned through introspection” (Coates [2007] 157). See Tye [2000] and Byrne [2001]; cf. Schroer [2007], Jackson [2007] 55, and Hill [2009] ch. 3. Phenomenological reflection is not only neutral with respect to classifi cation but is existentially silent as well. A hallucinatory experience can be subjectively indistinguishable from a veridical experience, but this phenomenological fact, on its own, requires neither existential profl igacy—through, say, the positing of Meinongian objects or the countenancing of uninstantiated properties/universals—nor existential parsimony—through, say, the acceptance of an adverbial or disjunctive account of experience. Arguments along these lines are widespread, for example, at Tye [2000] 46, 111ff. and Jackson [2007]. See Snowdon [1990] 136, Martin [2002], and Alston [2005]; cf. Schroer [2007] 405–7 and Smith [2008] 198–9. Th is discussion involves a slight but unproblematic idealization. In practice, there are no absolute ganzfeld experiences. Just as the darkness we appreciate when we close our eyes includes what Helmholtz called “optical dust,” that is, sundry points of light and dim patches, the experience of a ganzfeld will be subtly heterogeneous.

88

phenomenal intentionality

10. See Pylyshyn [2007] 66, Hochberg et al. [1951], and Avant [1965]. 11. Some individuals report that the ganzfeld vaguely resembles a surface; they estimate that it is located less than six inches in front of them (Gibson and Waddell [1952]). But just as many find such spatial descriptions inadequate or inappropriate (Cohen [1957]). The fact that apparent distance is not a necessary aspect of a ganzfeld experience means that its phenomenality is committed to even less than that of other atypical experiences, for example after-images. For we appreciate the sensuous color in an after image as qualifying a particular distal region in our visual field: according to Emmert’s law, an after-image always occupies a single visual solid angle, but the apparent linear size of the region fi lling that angle is directly proportional to the apparent distance of the surface onto which it is projected. 12. See Wright [1981] 589, Sorensen [2004] 462–3, and Magee and Milligan [1995] 11. 13. I suspect such experiences are nomologically possible, but metaphysical possibility is sufficient. 14. (i) A phenomenal point is not extended in the sense that one does not appreciate it as being spatially divisible. (ii) Spatially punctiliar experiences can occur in other sensory modalities. For example, if there were a creature with a single spine for a limb, and if the spine’s only sensory function is to extend outward in a single direction so as to register collisions with distinct bodies, then such collisions would yield non-haptic, tactile experiences that are spatially punctiliar (cf. Smith [2002] 154). 15. Pautz [2007] 517. In particular, the possibility of spatially punctiliar experiences undermines any interpretation of transparency that requires only two-dimensional spatiality, for example, Schroer [2007]. 16. Additionally (but relatedly), phenomenological reflection, by itself, places no conceptual or semantic constraints on discourse or judgment about experiential phenomenality. Most of the time, a normal perceiver undergoing an unexceptional experience will describe the sensuous qualities she appreciates as being instantiated in an objective, worldly entity. Furthermore, if she is justified in thinking that nothing is awry, she will be warranted when she judges that things are as she appreciates them to be. But we cannot accuse someone of inconsistency or misunderstanding simply because she judges, after reflecting upon the phenomenality of her experiences, that the sensuous qualities she appreciates are instantiated in one or more private, mind-dependent sense-data. If, for example, a perceiver were antecedently committed to a sense-datum account, it would not only be coherent for her to judge that a sensuous color she appreciates is an intrinsic property of an amalgam of sense-data, it would be reasonable for her to do so. It would, of course, be unreasonable for someone not antecedently committed to a sense-datum account to make such judgments. But this suggests only that the phenomenality of experience, by itself, does not epistemically privilege one set of judgments over its alternatives (cf. Gupta [2006]). 17. Experience’s presentational character has long been considered a basic datum that any adequate philosophical account of perception must accommodate. From sense-datum theorists—“That this whole field of colour is presented to my consciousness . . . cannot possibly be doubted. . . . This peculiar and ultimate manner of being present to consciousness is called being given, and that which is thus present is called a datum” (Price [1932] 3)—to direct realists—“[Perceptual] experience has a kind of directness, immediacy and involuntariness which is not shared by a belief which I might have about the object in its absence. It seems therefore unnatural to describe visual experiences as representations . . . because of the special features of perceptual experiences I propose to call them ‘presentations’” (Searle [1983] 46). See also, for example, Broad [1952], Strawson [1979] 97, Sturgeon [2000] 9, Smith [2002] 69, Martin [2002], Loar [2003] 82, Alston [2005] 255, Johnston [2007] 233, Burge [2007] 403–14, Pautz [2007], and McDowell [2008] 8. 18. It may help those familiar with Frege’s philosophy of logic to note that this account of experiences as phenomenally articulate unities with a structure describable in terms of the distinction between self and other resembles, in several important respects, Frege’s account of thoughts as logically articulate unities with a structure describable in terms

Phenomenal Presence

19.

20.

21.

22.

23. 24.

25.

89

of the categorial distinction between object and concept (Frege [1892/1997]; cf. Geach [1976], Diamond [1984], and fn. 20). The distinction between object and concept does not effect an ontological division of entities into kinds; the distinction captures thought’s logical articulation—an articulation which cannot be expressed in thought but is reflected in the syntactic segmentation of a properly constructed language’s well-formed formulas. Similarly, the distinction between self and other does not divide the world into two kinds; the distinction captures experience’s phenomenal articulation—an articulation which cannot be expressed by an experiential content but is phenomenally manifest in sensuous experience. “Self” and “other” signify, as it were, phenomenal categories. The phenomenal self needn’t be an attenuated Cartesian Ego or a merely formal transcendental subject. It is, as far as ontology is concerned, an aspect of the temporally extended life of a personal self. The parallels between the current proposal and Frege’s account of thought’s logical structure (cf. fn. 18) are especially strong on this point. According to Frege, we cannot truthfully (or sensibly) say of a concept that it is a concept; any attempt to make a singular claim about a concept will invariably result in a claim about an object. For “the three words ‘the concept horse’ do designate an object, but on that very account they do not designate a concept” (Frege [1892/1997] 184). According to the view I am elucidating, we cannot appreciate the phenomenal self (or the states/modes thereof) as such; any attempt to appreciate the self as such will invariably result in the appreciation of a sensuous element as other. For we can appreciate sensuous elements in experience, but on that very account they are not features of the phenomenal self. An example: the colors one appreciates in a typical visual experience are phenomenally rich. Th at is, the shades are absolutely determinate and assessments of their relative similarity yield a dense ordering. The representationalist can explain this aspect of visual phenomenality by attributing a non-discursive or analog content to the experience upon which sensuous colors supervene. The representational structure of such contents is isomorphic to the phenomenal structure of the similarity orderings and can thereby explain the latter’s manifestation. Th is challenge is similar to that posed by the alleged possibility of absent qualia. The representationalist must be able to explain, insists the proponent of absent qualia, why it isn’t possible for there to be functional and representational duplicates of sentient creatures that lack phenomenal consciousness. The present challenge does not focus on qualia themselves but on the phenomenally appreciable presence of such qualia. I believe that this reorientation avoids many of the problems associated with the hypothesis of absent qualia and other challenges that fall under the heading of “The Explanatory Gap” while simultaneously capturing what is central to the dissatisfaction of those who issue such challenges. Searle introduces the expression “aspectual shape” to refer to this feature of intentionality at Searle [1992] 155ff. Th is latter aspect of experiential intentionality’s perspectival nature involves more than there being features of perceived entities that are not themselves perceived, for example a visual experience of an opaque object doesn’t reveal its every side but only its facing surface. It also involves there being a perspective on the objects, properties, and relations we do perceive. The contextual parameters that contribute to the perspectival, aspectual shape of perceptual experience are legion. For example, a single perceived shape can appear differently as we move in relation to it and the appearance of a single perceived color will vary if subjected to differential illumination or if surrounded by objects with contrasting colors. The condition of personal subjectivity is intimately related to the restriction of non-derivative intentionality to occurrent, categorical states. Any intentionality we attribute to a dispositional state will be, at best, proleptically parasitic on the intentional directedness of its (perhaps merely possible) categorical manifestations (cf. Strawson [2005]). Millikan, among others, argues that an intentional state “must be one that functions as a sign or representation for the system itself ” (Millikan [1989] 284). But Millikan’s

90

26.

27.

28. 29.

30.

31.

32.

phenomenal intentionality account of “being for” encompasses dispositional, sub-personal phenomena and should thereby be distinguished from our notion of personal subjectivity. This account is often associated with conceptions of intentionality that focus on state ascriptions rather than states themselves. On these conceptions a state is intentional if and only if its ascription is susceptible to failures of (i) existential generalization and (ii) truth-preserving substitutions of extensionally equivalent expressions. Though the association of these sorts of expression with a state is often a good indication that the state possesses intentionality, such ascriptions are neither necessary nor sufficient for intentionality. This is easily done if one attributes Fregean contents —structured complexes of modes of presentation—to experiences. But one can also capture experience’s aspectual shape by invoking Russellian contents—structured complexes of objects, properties, and relations— that comprise appearance properties, that is, finely-individuated properties that reflect a subject’s perspective, rather than (or in addition to) properties simpliciter. See Alston [1999], Brewer [2006], Campbell [2002], Johnston [2006], Martin [2002], and Travis [2004]. (i) On many representational accounts of intentionality, if one is in a representational state, then one stands in a certain relation to a proposition (or a suitable non-propositional structure). Even if these propositions are Russellian, the relation to a proposition or to the entities within it must be distinguished from the relation to entities invoked by relational accounts of intentionality. Also, one can say that the satisfaction of a contentful state’s correctness condition places one in a relation to the entities the state represents. But the relatedness to entities that veridical representation affords, unlike its counterpart in the relational account, is not constitutive of the state’s being intentionally directed (even if the content comprises object-dependent elements); cf. Crane [2006]. (ii) Some prefer to maintain the synonymy between “intentionality” and “representationality” because representational states, unlike intentional relations, can be directed upon entities that do not exist (one’s capacity to enter into intentional relations is fallible, but one cannot stand in an intentional relation to a non-existent object). Though the permissibility of directedness upon non-existent objects is a common feature of representational states, I do not think that it is definitive of intentionality. Th is pair does not exhaust the possible extensions of (Int). For example, on one interpretation Aristotle takes experiential intentionality to consist in one becoming, in one manner of being, the entity one experiences: “Th at which can perceive is, as we have said, potentially such as the object of perception already is actually. It is not like the object, then, when it is being affected by it, but once it has been affected it becomes like it and is such as it is” (De Anima II.5 418a3–7). The motivation behind countenancing phenomenal contents is expressed well in Siewert [1998]. Examples of the stronger claim include Horgan and Tienson [2002], Loar [2003], and Kriegel [2007]. Consequently, the present account can avoid many of the controversies that surround the notion of phenomenal content, for example, whether such contents are narrow or wide. Additionally, this form of intentional directedness only occurs in sensuous experiences. So the present account can remain neutral about its relationship to the intentionality of non-sensuous states.

References Alston, W. P. 1999. “Back to the Theory of Appearing.” In Tomberlin, J. (ed.), Epistemology, vol. 13 of Philosophical Perspectives. Oxford: Blackwell. Alston, W. P. 2005. “Perception and Representation.” Philosophy and Phenomenological Research, 70(2):253–289. Avant, L. L . 1965. “Vision in the Ganzfeld.” Psychological Bulletin, 64(4):246–258. Brewer, B. 2006. “Perception and Content.” European Journal of Philosophy, 14(2):165–181.

Phenomenal Presence

91

Broad, C. D. 1952. “Some Elementary Reflexions on Sense-Perception.” Philosophy, 27:3–17. Burge, T. 2007. “Reflections on Two Kinds of Consciousness.” In Foundations of Mind: Philosophical Essays, Vol. 2, pages 392–419. Oxford: Clarendon Press. Byrne, A . 2001. “Intentionalism Defended.” Philosophical Review, 110:199–240. Campbell, J. 2002. “Berkeley’s Puzzle.” In Gendler, T. S. and Hawthorne, J. (eds.), Conceivability and Possibility. Oxford: Oxford University Press. Chalmers, D. J. 2004. “The Representational Character of Experience.” In Leiter, B. (ed.), The Future for Philosophy. Oxford: Oxford University Press. Coates, P. 2007. The Metaphysics of Perception: Wilfrid Sellars, Perceptual Consciousness and Critical Realism. New York: Routledge. Cohen, W. 1957. “Spatial and Textural Characteristics of the Ganzfeld.” American Journal of Psychology, 70:403–410. Crane, T. 2006. “Is There a Perceptual Relation?” In Gendler, T. S. and Hawthorne, J. (eds.), Perceptual Experience. Oxford: Oxford University Press. Diamond, C . 1984. “What Does a Concept-Script Do?” Philosophical Quarterly, 34: 343–368. Frege, G. 1892/1997. “On Concept and Object.” In Beaney, M. (ed.), The Frege Reader, pages 181–193. Oxford: Blackwell. Geach, P. T. 1976. “Saying and Showing in Frege and Wittgenstein.” In Hintikka, J. (ed.), Essays on Wittgenstein in Honour of G. H. von Wright, pages 54–70. Amsterdam: Acta Philosophica Fennica 28. Gibson, J. J. and Waddell, D. 1952. “Homogenous Retinal Stimulation and Visual Perception.” American Journal of Psychology, 62:63–70. Gupta, A . 2006. Empiricism and Experience. Oxford: Oxford University Press. Harman, G. 1990. “The Intrinsic Quality of Experience.” In Tomberlin, J. (ed.), Action Theory and the Philosophy of Mind, vol. 4 of Philosophical Perspectives, pages 31–52. Atascadero: Ridgeview. Hill, C. S. 2009. Consciousness. Cambridge: Cambridge University Press. Hochberg , J. E., Triebel, W., and Seaman, G. 1951. “Color Adaptation under Conditions of Homogeneous Visual Simulation (Ganzfeld).” Journal of Experimental Psychology, 41:153–159. Horgan, T. and Tienson, J. 2002. “The Intentionality of Phenomenology and the Phenomenology of Intentionality.” In Chalmers, D. J. (ed.), Philosophy of Mind: Classical and Contemporary Readings. Oxford: Oxford University Press. Horgan, T. 1993. “From Supervenience to Superdupervenience: Meeting the Demands of a Material World.” Mind, 102:555–586. Hume, D. 1739/1978. A Treatise of Human Nature, 2nd edition, ed. P. H. Nidditch. Oxford: Oxford University Press. Jackson, F. 2007. “The Knowledge Argument, Diaphanousness, Representationalism.” In Alter, T. and Walter, S. (eds.), Phenomenal Concepts and Phenomenal Knowledge, pages 52–64. Oxford: Oxford University Press. Johnston, M. 1992. “How to Speak of the Colors [and Postscript: Visual Experience].” In Byrne, A . and Hilbert, D. R . (eds.), Readings on Color Volume 1: The Philosophy of Color. Cambridge, MA : MIT Press. Johnston, M. 2006. “Better than Mere Knowledge? The Function of Sensory Awareness.” In Gendler, T. S. and Hawthorne, J. (eds.), Perceptual Experience. Oxford: Oxford University Press. Johnston, M. 2007. “Objective Mind and the Objectivity of Our Minds.” Philosophy and Phenomenological Research, 75(2):233–268. Kant, I. 1817/1996. “Lectures on the Philosophical Doctrine of Religion.” In Wood, A. W. and di Giovani, G. (eds.), Religion and Rational Theology. Cambridge: Cambridge University Press. Kriegel, U. 2007. “Intentional Inexistence and Phenomenal Intentionality.” Philosophical Perspectives, 21:307–340. Loar, B. 2003. “Phenomenal Intentionality as the Basis of Mental Content.” In Hahn, M. and Ramberg , B. (eds.), Reflections and Replies, pages 229–257. Cambridge, MA : MIT Press.

92

phenomenal intentionality

Magee, B. and Milligan, M. 1995. On Blindness. Oxford: Oxford University Press. Martin, M. G. F. 2002. “The Transparency of Experience.” Mind and Language, 17:376–425. McDowell, J. 2008. “Avoiding the Myth of the Given.” In Lindgaard, J. (ed.), John McDowell: Experience, Norm, and Nature, pages 1–14. Oxford: Blackwell. Millikan, R . 1989. “Biosemantics.” Journal of Philosophy, 86(6):281–297. Pautz, A . 2007. “Intentionalism and Perceptual Presence.” Philosophical Perspectives: Philosophy of Mind, 21(1):495–541. Price, H. H. 1932. Perception. London: George Allen and Unwin. Pryor, J. 2000. “The Skeptic and the Dogmatist.” Noûs, 34(4):517–549. Pylyshyn, Z. W. 2007. Things and Places: How the Mind Connects with the World. Cambridge, MA : MIT Press. Schroer, R . 2007. “The Reticence of Visual Phenomenal Character: A Spatial Interpretation of Transparency.” Australasian Journal of Philosophy, 85(3):393–414. Searle, J. 1983. Intentionality. Oxford: Oxford University Press. Searle, J. 1992. The Rediscovery of Mind. Cambridge, MA : MIT Press. Siewert, C . 1998. The Significance of Consciousness. Princeton, NJ: Princeton University Press. Smith, A. D. 2002. The Problem of Perception. Cambridge, MA : Harvard University Press. Smith, A. D. 2008. “Translucent Experiences.” Philosophical Studies, 140:197–212. Snowdon, P. 1990. “The Objects of Perceptual Experience.” Proceedings of the Aristotelian Society, supp. vol. 64:121–50. Sorensen, R . 2004. “We See in the Dark.” Noûs, 38(3):456–480. Stalnaker, R. C . 1998. “What Might Non-Conceptual Content Be?” In Villanueva, E . (ed.), Concepts, vol. 9 of Philosophical Issues, pages 339–52. Atascadero: Ridgeview. Strawson, G. 2005. “Intentionality and Experience: Terminological Preliminaries.” In Smith D. W. and Thomasson, A. L . (eds.), Phenomenology and Philosophy of Mind. Oxford: Oxford University Press. Strawson, P. F. 1979. “Percepton and Its Objects.” In Macdonald, G. (ed.), Perception and Identity: Essays Presented to A. J. Ayer with His Replies. London: Macmillan. Sturgeon, S. 2000. Matters of Mind: Consciousness, Reason, and Nature. London: Routledge. Travis, C . 2004. “The Silence of the Senses.” Mind, 113:57–94. Tye, M. 1995. Ten Problems of Consciousness. Cambridge, MA : MIT Press. Tye, M. 2000. Consciousness, Color, and Content. Cambridge, MA : MIT Press. Tye, M. 2002. “Representationalism and the Transparency of Experience.” Noûs, 36(1):137–151. Wright, E . 1981. “Yet More on Non-Epistemic Seeing.” Mind, 90(360):586–591.

5

Consciousness and Synthesis Col i n McGi n n

How does conscious intentionality differ from the unconscious kind? How does the intentionality of our conscious thoughts, in particular, differ from such unconscious representations as there might be? Even if there is no real intentionality of the latter kind, but only derivative or imposed intentionality, or quasi-intentionality, we can still ask what is characteristic of conscious intentionality. I shall here ask what unconscious intentionality would be like if there were any—how it would differ from conscious intentionality. No doubt many replies are possible, but I shall focus on just one, which has been neglected. Suppose I think, consciously and reflectively, that John is a bachelor, perhaps saying those words to myself; then my thought can be analyzed as having the content that John is an unmarried man. The intentional objects of my thought consist of John and the attribute of being a bachelor, but that attribute resolves into two further attributes (which may themselves resolve into yet further attributes upon deeper analysis). The intentionality of my thought is thus capable of analysis. If I had simply thought that John is a man who is unmarried, then my thought would not have needed such an analysis: it would already be analyzed (compare thinking that the king of France is bald versus thinking that there exists a unique king of France and he is bald). We can apparently think a proposition in an unanalyzed form or in an analyzed form—as we can say a proposition. If we thought every proposition in an analyzed form, then there would be no need for (or possibility of) conceptual analysis. So some of our intentionality is analyzable and some is not. The kind that is analyzable involves an operation we can call synthesis: a number of conceptual elements are brought together in thought so as to produce a unified concept that combines them. The elements have been synthesized into a whole, where before they were unconnected. There can be (interesting) analysis only where there has been synthesis.1 We can distinguish three degrees of conceptual connectedness: possession, conjunction, and synthesis. Thus a series of concepts C1 . . . Cn can simply be 93

94

phenomenal intentionality

jointly possessed by a subject, just when she “has” each of them. But it is also possible for the subject to conjoin the possessed concepts in thought, simply by use of the mental equivalent of “and.” However, there is a still stronger relation, which consists in the fact that the concepts may be synthesized into a unitary concept—where this goes beyond mere conjunction. Conceptual conjunction is not sufficient for real conceptual synthesis. This means that complex analyzable concepts are not psychologically equivalent to mental conjunctions of their components. When I think that John is a bachelor I don’t simply think the conjunction of unmarried and man—I think the concept that synthesizes these two concepts. Similarly for more interesting cases of conceptual analysis— such as Russell’s theory of descriptions or Bernard Suits’ definition of games. The mind has performed the operation of synthesis to produce a complex concept, where employing the concept is not simply reducible to thinking merely the conjunction of the conditions that define it. It is because of this mental operation that the analysis of thought is possible: the seeming simplicity of the concept is shown by analysis to conceal a hidden conceptual complexity. The surface unity is founded on an underlying multiplicity. This apparent unity is the upshot of the operation I am calling “synthesis.” No doubt synthesis is somewhat puzzling and mysterious. How can the generation of complex concepts be anything other than composition by conjunction? How can synthesis produce conceptual unity from disparate elements? I shall come back to this question, but for now the point is that synthesis appears to be a real psychological phenomenon: when we have a conscious thought of the analyzable kind, synthesis has occurred. I shall accordingly speak of “synthetic intentionality,” meaning that the intentionality in question is the product of synthesis—as with my thought about John being a bachelor. I shall contrast this with what I shall call “associative intentionality,” meaning the kind that is involved in consciously thinking that John is a man who is unmarried. This latter kind of intentionality can be understood as employing the mental operation of conjunction, without any accompanying synthesis into a new unity. We might think of the process of complex concept formation as a two-stage affair: first, the possessed concepts C1 . . . Cn are combined according to conjunction; second, the conjunction is subject to synthesis, whereupon we have a new concept C* that unifies or amalgamates the conjoined elements into a conceptual whole. When the second stage is complete the concept C* no longer looks like a mere conjunction, or feels like one introspectively—nor is it expressed by a syntactic conjunction in ordinary language. It has its own primitive predicate and its own corpuscular phenomenology—yet it is subject to an analysis in which the conjunction figures. We might say that it masquerades as primitive and could even turn out to be primitive—but it is in reality complex, structured, and analyzable. I hope I have said enough to establish the reality of synthesis as it applies to the concepts we employ in conscious thought. My thesis, then, is that conscious

Consciousness and Synthesis

95

intentionality involves synthetic intentionality, while unconscious intentionality (if there is any) involves merely associative intentionality. More exactly, conscious intentionality has the power of synthesis, while unconscious intentionality does not have this power. We need to put it that way because not all of conscious intentionality involves synthesis—notably, thinking that involves primitive concepts. Only complex analyzable concepts involve synthesis, not simple concepts. Then the idea is that consciousness has the power to synthesize simple intentionality into complex intentionality—but nothing else does. In effect, unconscious intentionality is just a conjunction machine: it can only bring concepts together by conjunctive association—it cannot really form synthetic wholes. In a certain sense, abbreviation is possible for conscious intentionality, but not for unconscious intentionality. Abbreviation, then, turns out to be a lot more interesting (and mysterious) than we thought: it results from the power to transcend mere conjunction in forming complex concepts. To put it differently, consciousness is what permits the possibility of conceptual analysis—while unconscious representations have no (interesting) analysis. The unconscious mental representation that John is a bachelor (that proposition) is just the representation that John is a man and unmarried—there is no defining to be done here. More interestingly, an unconscious representation that the king of France is bald (that proposition) just consists in the representation that someone is king of France and is uniquely so and is bald. The unconscious representation is already fully analyzed, so permits no further analysis. Consciousness, by contrast, is what conceals underlying structure—by creating synthesis—and hence makes informative analysis possible. Let me note, as an aside, that according to this conception analytic philosophy is possible only for conscious beings with the power of conceptual synthesis. Because of synthesis our knowledge of the analysis of our concepts is not an introspective given—we can’t just read the analysis off by introspecting our thoughts. Mental abbreviation (condensation) is necessary to the possibility of philosophy (conceived as the analysis of concepts). A being without synthetic consciousness would have no use for analytic philosophy, because the analyses of its concepts would already lie open to view. Or better: if all complex concepts appeared manifestly in thought as conjunctions of their primitive components, then the job of analysis would not need to be performed. If a being possessed all our complex concepts, but lacked the power of conscious synthesis, then those concepts could only exist in its mind as associative conjunctions, in which case the analysis has already been performed. Such a being could still pursue science in much the same way we do, but philosophy would be a dead end street—because already complete! This is why some analytic philosophers have supposed that our task is to discover the language of thought that lies unconsciously beneath our conscious thinking—for in that language all is revealed. Ironically, consciousness makes philosophy more difficult. The reason is that conscious

96

phenomenal intentionality

intentionality typically involves synthesis, and synthesis is what blocks the immediate recognition of conceptual structure. An unconscious thinker is a better philosopher than a conscious one! That is, Mentalese (if there is such a thing) is more logically transparent than the natural languages we consciously use—or the thoughts that flit through our teeming and economy-minded consciousness. In other terminology, our implicit knowledge is philosophically superior to our explicit knowledge—more analytically articulated. What is my ground for making this distinction between conscious and unconscious intentionality? I think it is apparent upon inspection that conscious intentionality is synthetic, but inspection is not so possible for unconscious intentionality—so why do I doubt it there? My reason for insisting on merely associative intentionality for unconscious representations is simply that I cannot see what could generate synthesis in the absence of consciousness: how do we get the surface-deep distinction for unconscious representations? We can’t appeal to introspection, because the representations are unconscious. How could complexity of representations in the unconscious mind ascend above the level of conjunction? All complexity there has to be, in a word, serial—a matter of concatenation—without the benefit of genuine synthesis. Computer code, say, resolves into a series of 1’s and 0’s: these basic representations never disappear into a higher unity as conscious concepts do. Unconscious intentionality can only mechanically combine; it cannot creatively synthesize. The conscious mind has the power to convert a conjunction into a synthesis (a mysterious power: see later in the chapter), but the unconscious mind cannot perform this operation—it must sluggishly conjoin and concatenate. The synthetic power gives rise to the appearance of simplicity in our complex concepts, but unconscious representations cannot have this appearance-reality distinction—there is no representational appearing down there. In the unconscious a conjunction is just a conjunction; in other words, the correct analysis is explicit. What is this power of synthesis, of generating or recognizing conceptual unities within pluralities? It is as if the conscious mind sees in a set of concepts a higher-level unity or coherence and consequently produces a synthesis of the elements. For example, the concepts truth, belief, and justification are apprehended as forming the complex concept knowledge (assuming this is a good analysis of the concept)—and not merely as capable of being conjoined. We can conjunctively combine any old concepts and not thereby produce anything with conceptual unity—as with “big and warm and triangular.” What we want to say here is that the three words “true” and “believed” and “justified” organically form a special unity—which we describe as “knowledge.” We thus see these three concepts as parts of a whole. What kind of capacity is this? Here we may be reminded of the Gestalt psychologists: they studied the way the human conscious mind imposes or discovers unities in arrays of stimuli, concentrating on perceptual unities. They tried to ascertain the particular laws of association

Consciousness and Synthesis

97

that produce apprehended unities, such as propinquity and continuity. They viewed consciousness as a unity-detecting device—something that takes mere plurality and converts it into unity. Using their terminology, we can describe the synthesis of concepts, as in the knowledge case, as one kind of Gestalt detection (or imposition). I would say that the perception of Gestalts is only possible for the conscious mind—because this is a matter of how the world appears in consciousness. The unconscious perceptual system could only compute geometric relations and possibly deliver verdicts as to function; it could not experience perceptual unity in the midst of plurality. Genuine Gestalt perception is essentially conscious perception. This is because the power of true synthesis is unique to consciousness—finding unity in plurality. Thus conscious perceptual intentionality is different from unconscious perceptual intentionality—the former being synthetic, the latter associative. The phenomenological Gestalt is the province of conscious perception; unconscious perception can only generate a simulacrum of it—a calculated unity, not an apparent unity. If this is right, then conceptual synthesis can be viewed as a special case of Gestalt synthesis—the ability to detect and formulate unities in pluralities. And the general thesis is that consciousness is the sine qua non of synthetic unification. It is not an accident that the Gestalt psychologists concentrated their efforts on conscious perception, because unconscious perception (if there is such a thing) has no Gestalt dimension—just as unconscious thoughts (if there are any) lack the kind of conceptual synthesis characteristic of conscious thoughts. I don’t think we have any very good theory of how mental synthesis works. It has a kind of magical feel to it—getting something from nothing, spontaneous generation. It evidently goes beyond mere conjunction, but we have no other model for how the simple gives rise to the complex in the conceptual domain. Consciousness just seems to whisk the new unity into existence—a kind of emergence takes place. (It is not unlike the way consciousness itself is mysteriously whisked into existence by the brain, apparently emerging from something insufficient to produce it.) But this very mysterious quality supports the thesis I am defending—that synthetic intentionality belongs exclusively to consciousness. Unconscious intentionality is not so mysterious, being both unconscious and lacking in magical-seeming synthesis. Machine intentionality, if such there could be, lacks the faculty of mysterious synthesis. The kind of puzzling synthesis I am discussing is akin to the kind identified by Frege, who likewise puzzled over the unity of the proposition: why is a proposition (or sentence) not just a list of elements—whence the unity?2 It is certainly made of parts, which may be detached from it—so how come the parts manage to cohere into new type of whole? Frege invoked the rather mysterious notion of “saturation” to handle the problem. I am talking about a psychological analogue: how the constituents of a complex concept cohere together into the higher unity represented by the concept itself—how do the parts arrange themselves into the whole (if not by

98

phenomenal intentionality

simple conjunction)? We want to speak here, not simply of composition, but of gestation—the parts giving birth to the whole by the mysterious mechanism of synthesis. But my aim in this chapter is not (fortunately) to explain synthesis; it is merely to use it to distinguish two kinds of intentionality. I can put the proposal very simply by saying that conscious intentionality is the kind that admits of analysis, whereas unconscious intentionality does not. Contrary to Freud, the unconscious has no analysis; but the conscious is deeply analyzable. That is, in the unconscious everything is analytically explicit, but consciousness contains implicit content—hidden by the synthesis it has imposed on collections of primitive concepts. Thus Brentano may be wrong to suggest that intentionality is the mark of the conscious mind, but we can easily amend his thesis to say that synthetic intentionality is the mark of the conscious mind. Only the conscious mind directs itself to things outside itself by synthesizing its more primitive states of intentionality. The essence of consciousness is not directedness, pure and simple, but condensed directedness.

Notes 1. The larger background to this chapter is my Truth by Analysis: Games, Names, and Philosophy (New York: Oxford University Press, 2011), in which I defend conceptual analysis; part of this defense is the introduction of the notion of conceptual synthesis (see chapter 4). 2. In fact, the problem goes back to Plato and was influentially discussed by Russell and Bradley, as well as Frege. I am not aware, though, that a comparable problem with sub-propositional elements has been recognized—how simple concepts are brought together to produce synthetic unities. If propositions threaten to collapse into lists, why don’t concepts? If consciousness is required for conceptual synthesis, one might wonder whether it is also required for propositional synthesis (that is, the having of unitary thoughts)—which would put in jeopardy the very idea of unconscious intentionality. Th roughout this chapter I have treated the claim of unconscious intentionality with some tolerance, but in fact I am quite skeptical of it. It may well be that all intentionality requires the faculty of synthesis, in which case intentionality requires consciousness, if the thesis of this chapter is right.

6

Constructing a World for the Senses K ata l i n Fa r k a s

1. Perceptual Intentionality We receive information about the world through our sensory organs. The surface of the organ receives some stimulus, and the complex mechanisms of our perceptual system produce perceptual experiences. It is a characteristic feature of these experiences that they present a world that seems to be outside our body, and to exist independently of our experiences. At least in one sense of the term “intentional”, this is to say that perceptual experiences are intentional. The term has been used in different ways, and sometimes people say that a mental state or event is intentional simply if it has an object, even if this object is experience-dependent, like a sensation. I am interested in a different phenomenon: I will confine my interest in this chapter to intentionality as the apparent directedness of a mental state at something beyond itself, moreover, at something which could exist independently of being experienced. I say “apparent” directedness, because hallucinations, that is, experiences that lack an object, can also be intentional in this sense: in hallucinations too, it can seem that an experience-independent world is presented to us. I take the question of whether an experience is intentional a phenomenal question, in the sense that it deals with what appears to be presented by an experience. Some philosophers are interested in a notion of intentionality or representation that separates the question of the apparent presence of the external world from the issue of representation. For example, some philosophers say that an experience of a brain in a vat may seem to represent the external world, but it does not, because it lacks an adequate causal-nomological connection to the world. This experience appears to have an object, but if the object of an experience is what is represented by the experience, then this appearance is misleading: the experience has no object. If the terms “representational” and “intentional” are used synonymously, then the experiences of brains in vats lack intentionality. I do not wish to take issue with such views here, nor do I claim that my sense 99

100

phenomenal intentionality

of “intentionality” is superior. I simply want to make clear what the focus of this chapter is: it is perceptual intentionality, understood in a specific way. To repeat then: a sensory experience exhibits perceptual intentionality if (i) it’s in a sensory mode and (ii) it seems to present an experience-independent world, including experience-independent objects and their experience-independent qualities. Understanding intentionality in this way fits well with the phenomenal intentionality program. According to defenders of the phenomenal intentionality view, there is a kind of intentionality, pervasive in conscious mental episodes, which is constituted by the phenomenal features of conscious experiences (Horgan and Tienson 2002). This chapter offers support for a specific version of this view. The basic idea is as follows. The simple phenomenal features of sensory experiences in themselves may amount to no more than modifications of the subject’s consciousness: they may present nothing beyond the experience that they are part of. A feeling may just be a feeling and not present or represent anything. However, when these sensory features are received by the subject in a highly organised and predictable structure, one that responds to actions and further inquiry in a systematic way, the experience may become suggestive of the presence of something beyond this experience, namely, an experience-independent object. Perceptual intentionality is thus constituted by the structure of sensory phenomenal features: by the way these features hang together and respond to movement and inquiry. I shall illustrate this thesis mainly on the case of pain. The views on the nature of pain vary: some suggest it is a pure sensation, which has no intentional object; others hold that pain experiences represent parts of our body; yet another view likens pain to exteroceptive perceptions like touch. According to the position defended here, there may be some truth in each of these views. The pure qualitative features of a pain-experience are not, in themselves, objectpresenting. However, if these features are organised in a certain order, they may come to present a part of the body, or even objects external to the body. Comparing these different scenarios hopefully sheds light on the nature of perceptual intentionality.

2. Cutaneous Senses Our largest sensory organ is the skin, and it contains various kinds of receptors. Mechanoreceptors are sensitive to pressure and vibration; thermoreceptors are sensitive to changes in temperature, and nociceptors react to various kinds of noxious stimuli. Each kind of receptor, when activated, sends out signals which travel, through dedicated pathways, to the brain, and when all goes well, result in an experience with a characteristic phenomenal character.1

Constructing a World for the Senses

101

Suppose that a group of Extra-Terrestrial scientists came to study the perceptual system of human beings. The sensory organs of these people are very different from those of ours, but they, just like we, have conscious experiences in response to events in the world, and they can form a perfectly good idea of how we learn about our environment through experiences with a certain specific kind of phenomenal character. They are now conducting a study of the human somatosensory system, and they have discovered the three kinds of receptors and how the receptors react to different kinds of stimuli. From the first-person reports of human subjects, they also learned that these stimuli give rise to different kinds of experiences. At this point, their hypothesis is that human beings perceive the world through their skin at least in three different ways. They think that what is presented in these experiences are properties and things that activate the receptors. They are not sure whether the three ways involve perceiving different sets of properties, or perhaps some of the same properties are perceived in different ways. Still, they are pretty confident that they have identified three exteroceptive perceptual systems: that is, three systems that seem to present the external world through characteristic conscious experiences. The Extra-Terrestrial scientists are subsequently somewhat surprised to discover that some philosophers have a different opinion. For example, Paul Grice, writing on the question of how to individuate different sensory modalities, condemns the practice of counting three cutaneous senses: “consider the assaults made by physiologists and psychologists on the so-called ‘sense of touch’. They wish, I think on neurological grounds, to distinguish three senses: a pressure-sense, a warm-and-cold sense, and a pain sense” (Grice 1962, p. 36). ˫ Grice points out that a characteristic kind of experience produced by a material thing is not always a perception of the thing. He thinks that feeling pain is a case of this kind: it’s true that objects produce pain experiences in us, but pain is not a perception of the pain-causing object. The qualities that we naturally use to describe these experiences – for example, ‘painful’ – do not identify a “relatively abiding characteristic which material things in general either possess or do not possess” (Grice 1962, p. 36) Locke also questioned that the three “senses of touch” are properly regarded as exteroceptive perceptual senses. Locke famously distinguished among primary, secondary, and – as they are occasionally called – tertiary qualities. Tertiary qualities are “bare powers”: the power of things to produce certain kinds of sensations in us. For example, fire has a power to produce a painful sensation. But when we reflect upon the experience, we are not inclined to think that the quality that characterises the experience is also to be found in the cause of the experience: we don’t think that pain is somehow in the fire. This is different from the way we are inclined to think of secondary qualities like warmth. The natural idea is that the warmth we feel in an experience is the

102

phenomenal intentionality

same quality as the fire has: fire is warm. Now Locke thinks that this is a mistake, because we should think of secondary qualities exactly the same way as we think of tertiary qualities: warmth is no more in the fire than pain is (Locke 1690, book II, chapter viii). Let us assume, as our first hypothesis, that if an experience is a perceptual experience, then the object one perceives (in case of actual perception, that is, excluding cases of hallucination) is an external object that causes the experience. That is, let us limit our inquiry to exteroceptive perception. Then we can summarise Grice’s and Locke’s view as follows. Grice thinks that touch, when it tells us about the shape, surface or warm-and-cold qualities of external things, is a form of perception. However, pain experiences are not forms of perception. Locke things that pain experiences don’t seem to be perceptions (only sensations), whereas warmth experiences do; yet this is misleading, because neither pain nor warmth experiences are perceptions. Only tactile experiences of shapes and surface qualities are perceptions. The Extra-Terrestrial scientists are very interested in these claims and would like to know what explains these differences. As far as they can see, the production of experiences of shapes, of warm-and-cold, and of pain, all have the same structure: external object – stimuli of appropriate receptors – neural mechanism – experience with a specific phenomenal character. They wonder why some of these count as perceptions, and why some don’t. The curiosity of the Extra-Terrestrial scientists is further intrigued when they discover that there are philosophers (different from the ones previously consulted) who apparently toyed with the idea of pain experiences constituting a case of external perception. In §312 of the Philosophical Investigations, Wittgenstein asks us to imagine the possibility that “the surfaces of the things around us (stones, plants, etc.) have patches and regions which produce pain in our skin when we touch them. . . . In this case we should speak of pain-patches on the leaf of a particular plant just as present we speak of red patches” (Wittgenstein 1953). We may imagine that we can trace the contours of a pain patch, just like tracing the contours of a smoother patch in a rough surface. The experience would not be like feeling the sting of nettles, because when one is stung by a nettle, there is no clear awareness of which area of the leaf caused the pain. In contrast, feeling the pain patch would involve being conscious of one’s finger leaving and entering the pain patch. The other example is based on an idea by A.J. Ayer.2 Imagine that there is a big building with a maze of corridors, and one of the corridors has the unpleasant property of giving you a strong headache as soon as you enter it. The headache stops once you leave the corridor. This is a very reliable feature, and after a few bad experiences, when considering how to get to another part of the building, you think it’s better if we avoid the ‘headachy’ corridor. One could say: this corridor is dark, this other is usually smelly, the third is noisy, and that one is very headachy.

Constructing a World for the Senses

103

If these examples are plausible, then we can see how to make sense of the idea that pain experiences could be exteroceptive perceptual experiences. If Grice and Locke are right, we don’t normally treat pain experiences as forms of external perception; but if the examples presented by Wittgenstein and Ayer are coherent, then at least in some sense, this isn’t a necessary feature of these experiences. One of the main aims of this chapter is to provide an explanation of this phenomenon: namely, the phenomenon that we do not regard certain sensory experiences as exteroceptive perceptual experiences, but we can imagine circumstances in which very similar experiences – exhibiting sensory features in the same modality – would in fact be regarded as exteroceptive perceptual experiences.

3. Felt Pain and Perceived Pain The first thing to notice is that the qualities that characterise our pain experiences and the qualities that characterise the pain patches on the leaf must be different qualities. On the most dominant notion of pain, pains are essentially felt. If one has an injury, and it hurts, the pain lasts as long as there is a feeling of pain. If one takes an anaesthetic, and it works, the pain ceases to be felt and hence it ceases altogether: a painkiller does indeed kill pain. Of course, the injury may remain, even if it doesn’t hurt any longer. I say that this is the dominant ordinary notion of pain, though there is perhaps some room for debate whether the way we talk of pains always follows this logic. Maybe sometimes it doesn’t. In any case, there is clearly a coherent notion of pain as essentially a feeling, and this notion is an important one. An “unfelt pain” (though I’d prefer to say “injury”) may cause certain kinds of concern: injuries are harmful, and we usually aim to heal them, or get rid of them in some way. Just making the feeling of pain disappear may not be enough. However, it is still true that felt pain is of special interest to us, and most of the time when are concerned about pain we are concerned about the painful feeling. According to some theories, a pain experience is the representation of tissue damage in the affected area of the body. Tissue damage can exist even when no one is representing it: damage to a bodily part is an experience-independent state of affairs. So the state of affairs I represent when my hand hurts is something that could obtain even if I were not conscious. This view cannot be the full account of what it is to be in pain, because it offers no explanation for the circumstance that pains are essentially felt. The phenomenal character of the experience cannot be entirely given by the representation of an experience-transcendent property. A much more plausible explanation is that pain is a kind of experience that is characterised by its own specific experiential

104

phenomenal intentionality

quality. I shall call this quality “Pain1’”; it is a quality that can be exemplified only if a conscious creature has the appropriate kind of experience.3 The second notion, Pain2, is employed when we talk of a pain-patch. Pain2 is an experience-independent quality of an experience-independent object. This quality can be exemplified even if no conscious creature has any actual relevant experience.4 Pain-patches and headachy corridors have their character even if no-one is around to experience this character. Pain1 and Pain2 are clearly different properties: Pain1 can be exemplified only if a certain kind of experience is occurring, whereas Pain2 can be exemplified in the absence of any relevant experience. Some of our usual notions of qualities of certain kinds of experiences and qualities of objects exhibit the same double nature. Consider, for example, taste. If someone has a bad head-cold, we sometimes say that she cannot taste anything. But people occasionally also say when they have a cold that everything is completely tasteless. There are two different notions here: Taste1 is a feature that can be exemplified only when a conscious creature has the appropriate experience, whereas Taste2 is a quality that can be exemplified even in the absence of any such experience. Clearly, there is an intimate connection between the qualities of Pain1 and Pain2: when I perceive something as having the quality Pain2, my experience has the experiential quality Pain1. Pain2 is a quality that figures in imaginary scenarios like the ones conceived by Wittgenstein and Ayer, but we don’t use it to characterise external objects in everyday life. As various philosophers put it, we don’t externalise pains. But why not? I contend that every theory of sensory experience should be able to answer this question. Perceiving pain patches and headachy corridors are experiences that exhibit perceptual intentionality in the sense I defined it above: they are sensory experiences that seem to present features of the world – the patch, or the corridor having the quality Pain2 – that could exist independently of being experienced. If we understand how an experience can acquire the feature of presenting external objects as having Pain2, we will be nearer to understanding the nature of perceptual intentionality.

4. The Contemporaneous Presence of the Object I argued in the previous two sections that there could be experiences which are perceptions of external objects, and which also have the same experiential qualities – of the type I called “Pain1” – as our ordinary pain experiences. These experiences are hypothetical, because our actual pain experiences are not like that. I suggested that we start to look for an explanation of this fact. In this section and the next, I will compare ordinary exteroceptive perceptual experiences and pain experiences, and I shall point out the features that seem to be present in external perceptions but missing from pain.

Constructing a World for the Senses

105

It is a characteristic feature of perceptual experiences that they seem to indicate the presence of objects and properties that are contemporaneous with the experience. Ordinarily, if we perceive an object and its having a property, the impression is that the object exists, and exemplifies the property at that moment, when the experience is taking place. When we cease to have the specific type of perceptual experience, this can be a result of either of two things: (i) the object ceases to exist or to have the property (for example, when we see the colour of traffic lights changing) or (ii) the circumstances of perception change (one turns one’s head and doesn’t see the traffic light any more). But in any case, as long as the perceptual experience persists, the object or property seems to be present. By contrast, pain is not usually contemporaneous with its external causes. This is probably one of the main reasons that we don’t regard pain as a case of external perception. You see the brightness of the fire, and feel its warmth; if the fire is extinguished, these experiences stop. But if you are careless enough to burn your hand in the fire, the pain remains, even if the fire ceases to exist. You cut your finger and you throw away the knife in frustration: the knife is not there (and perhaps its edge is blunted by the fall) but the pain lingers. In fact, the pain sensation can become even more intense after the harmful object is removed. Imagine that visual, auditory, or tactile perceptual experiences worked in the same way. You look at a picture, you turn away, and the picture remains with you etched on your visual field for days. You put down a warm cup of coffee, and take a box of cold milk from the fridge, but the feeling of warmth stays in your hand – perhaps becoming even more intense. And so on. It is true that a faint “echo” of a perceptual experience can linger for a bit even in these modalities: if you look at a bright source of light and turn away, bright patches seem to stay with you. But these don’t last long, and they obviously lack the intensity and detail of the original perceptual experience. In fact, afterimages are quite different from the visual experience that causes them: staring at a red patch causes a greenish afterimage. So it’s quite obvious that it’s not the original experience that lingers. As for auditory experiences, there may be a bit of ringing in your ears after hearing a loud noise, but the length, quality, and intensity of the lingering experience cannot be compared to the way pains linger. Let us distinguish typical cases of external and internal pain. By “external” pain, I mean pain caused by an injury like a cut, burn, or bruise on the surface of one’s body; by “internal” pain, I mean the kinds of pains that are often called “aches”: headache, stomach ache, pains in muscles or joints like backache, rheumatic pain, and so on. External pains are usually immediately caused by an easily identifiable external object, and if pain was a form of external perception, these would be the obvious candidates to be the perceived objects. We know that many internal pains are caused by the impact of certain external objects

106

phenomenal intentionality

or events: for example, a headache can be caused by too much drink the previous evening. In this case, the pain experience is not contemporaneous with the presence of the external cause because it is seriously delayed. In the case of both external and internal pain, we seem to have the following situation: an object external to the body causes a characteristic change in the body, and the experience seems to be contemporaneous with the presence of some sort of bodily condition rather than with the presence of the external object. So it is completely understandable that when people look for the intentional object of a pain experience, one of the most plausible candidates is the affected bodily part, rather than the offending external thing. I shall come back to this view in section 8. However, I contend that we can imagine situations where pain experiences indicate the presence of external objects or properties just like ordinary perceptual experiences do. Wittgenstein’s pain patches and Ayer’s headachy corridor seem to be illustrations of precisely this phenomenon. As I said when introducing the examples, the most plausible conception of the pain patch seems to involve the experience of our fingers tracing the pain patch: leaving and entering it. “Headachy” seems to be similar to “noisy,” “smelly,” and “dark” if the headache starts when we enter the corridor, and ceases when we leave.

5. Abiding Characteristics In the previous section, we saw that one reason why we do not regard pains as perceptions of external objects is that pain experiences are usually not contemporaneous with the presence of their external cause. Grice indicates a further reason: he thinks that the main reason we don’t externalise pains is that we do not regard “painful” as a relatively abiding characteristic of objects. We classify things according to their smells, he says, but not according to their painfulness. He thinks the practice has to do with the following circumstances: (a) Pains are not greatly variegated, except in intensity and location. Smells are. (b) There is no standard procedure for getting a pain: one can be cut, bumped, burned, scraped, and so on. There is a standard procedure for smelling, namely, inhaling. (c) Almost any type of object can infl ict pain upon us, often in more than one way. (Grice 1962, p. 36) As a consequence of these facts, Grice says, “pains are on the whole very poor guides to the character of things that hurt us” (ibid.). Let us assess these claims. The first point doesn’t seem to me decisive: pains do vary to some extent, and qualities of pain are helpful diagnostic tools. For

Constructing a World for the Senses

107

example, pains can be shooting, sharp, dull, burning, tingling, cramping or throbbing (see Victor et al. 2008). As for the second point, notice that Grice’s list of procedures concerns only external pain. In a certain sense, contrary to what Grice says, these various ways of encountering pain are intelligibly similar. Compare the procedure of getting pain to the procedure of getting tactile sensations: in the latter case, the procedure is basically for the object to have contact with the skin, and we regard this as fairly uniform, even though contact with the skin can take a number of forms: pressing (or being pressed), stroking, gripping, pushing (or being pushed), lifting. Being cut, burned, bumped, or scraped are all forms of violent and tissue-damaging contacts with the skin, and in this sense, they can also be regarded as similar. Grice’s point seems more plausible if, unlike Grice, we also consider internal pains, since the causes of internal pains are not similarly uniform: in many cases, they are not direct results of a certain type causal procedure. Grice’s point thus does not seem to be very effective against the point that external pains could be perceptual, but would be more effective against the claim that internal pains are perceptual. Grice’s third point is partly connected to the second. Consider the example of the pain patch on the leaf: of course, the “non-painy” parts of the leaf can also cause pain, for example, if someone sticks the leaf in your eye. Non-headachy corridors can cause a headache if they collapse on your head. Imagining a uniform method of sensing pain makes the cases of externalised pain qualities more plausible: we would be interested in the question of which surfaces are painful when touched, or which corridors radiate headache when we enter them. All in all, Grice may be right about some of the factors that contribute to qualifying a type of experience as perceptual: perceptual experiences will be those that arise through a meaningfully uniform procedure, and those we regard as guides to the nature of things. However, he seems to underestimate the suitability of specifically external pain experiences to qualify as perceptual on these grounds.

6. Structure, Order, Predictability, Exploration I’ve been arguing so far for two claims: first, that we do not normally externalise pains – that is, we don’t regard pain experiences as perceptions of external objects that cause the experience in question; nonetheless, second, that we can meaningfully imagine circumstances where we could or would externalise pains. A non-externalised pain experience is characterised by a specific phenomenal or experiential quality, the quality I called “Pain1”. The externalised pain experience has the same phenomenal quality, but in addition, it

108

phenomenal intentionality

also presents an external object as having the experience-independent quality Pain2. For the time being, I consider only the options that pain experiences are either internal sensations or perceptions of an external object. I shall consider the further possibility that pain experiences are perceptions of the injured bodily part in section 8. For the moment, I’d like to argue further that what is responsible for the appearance of the presented Pain2 quality is the structure of the experience: that is, the relation of Pain1 to further phenomenal and presentational features of the same, and other, experiences. Consider a typical visual experience like surveying the scene in front of you. This kind of experience is overwhelmingly presentational: objects and their properties appear as experience-independent. The scene provides a lot of details that you can explore in typical ways: if you turn your head, or approach an object, the quality of the experience changes in very familiar and highly predictable ways, suggesting the presence of an experience-independent state of affairs.5 The rich structure of the visual experience is in harmony, again, in very familiar and highly predictable ways, with the details of your perceptual experiences in other modalities. The feeling as you hold the book, the small noise the paper makes as you turn a page, all these features form a structure that, in itself, and also together with past experiences, clusters the experienced features around the well-known objects presented in the manifest image of our world. Contrast this with the experience of an afterimage. In an afterimage, the structure characteristic of the presentation of experience-independent objects is missing. There is no fine detail that returns even after an interruption: say, if you close and open your eye again. The features don’t seem to be available for further investigation since the experience is ephemeral, and the experience does not respond to movement in a way that’s typical in perceiving external objects. On the contrary, it responds in a way that we have learned to associate with “mere” experiences like phosphenes and afterimages. In the case of at least some afterimages, there is no overwhelming impression that some experience-independent object is presented: the nature of the experience can be naturally understood as a feeling, as nothing more than a modification of one’s consciousness. Now consider again the earlier example of the fire: the flames are bright, it feels warm, there’s a pleasant smell of burning wood, a sound of crackling. As you move towards, or away from, the fire, the intensity of all these experiences increases and decreases in a way we are accustomed to in the perception of objects. Our success in exploration also incorporates the point mentioned by Grice: that there is a uniform way to produce the experience, and we know how to move our body to enhance or diminish the intensity of a certain type of perception. All this is meant to be a comment on the phenomenal level: a brain in a vat, though obviously not actually moving closer to a fire, would seem

Constructing a World for the Senses

109

to move and have its experiences change accordingly. Our investigation is at the phenomenal level: we are interested in how objects appear in experience. Hence we shall disagree with Locke on this point: warmth experiences do seem to present the fire, and hence they will be considered as exhibiting perceptual intentionality. I suggest that the experiential qualities of ordinary pain experiences do not fit into the structure of experiential features clustering around objects, and that is why pain is not normally externalised. As I said earlier, if you burn your hand, the pain experience after the hand is withdrawn from the fire seems to be focused on you, rather than on the fire: the fire is extinguished, the brightness, the warmth, the sound, the smell are gone, but the pain stays. The experience doesn’t seem to point to anything external, because it is not coordinated with other experiential features whose source would appear to be the same object. Moreover, qualities of external pains have a subjective similarity to internal pains, and internal pains fit into the structure of external perceptual experience even less. I pointed out a difference between ordinary perceptual experiences on the one hand, and experiences that don’t seem to present experience-independent objects – afterimages, phoshpenes – on the other. The difference was that the simpler phenomenal features of perceptual experiences are organised into a systematic, cross-modally coherent and predictable order. This order is what I call the “structure” of the experience, and I shall suggest that this is responsible for the phenomenal appearance of an experience-independent object. We had two facts to explain: first, that we normally don’t externalise pains; second, that it seems conceivable that we could. If the structure of the experience accounts at least partly for the apparent presentation of the external world, then we can also explain why, under certain circumstances, pains may become a form of external perception: this would happen when the experiential pain feature did fit into the overall structure imposed upon us by the world. This suggestion is the beginning of a theory that explains how the phenomenal features of an experience can partly constitute perceptual intentionality. Representationalist and disjunctivist theories of perception usually take for granted that it is part of the phenomenology of a perceptual experience that it seems to present the world. A disjunctivist theory takes the presentational aspect to be a crucial feature of perceptual experiences: a veridical perception presents things, and a corresponding hallucination seems to do so. In a representationalist theory, a perceptual experience represents the world, and presumably this explains why a world seems to be presented. Perhaps we can’t even say that there is an explanatory relation here: an experience being representational simply means that the world appears in a certain way. Some versions of the representationalist view offer a reduction of representational properties in terms of causal-nomological or teleosemantic relations. In this

110

phenomenal intentionality

sense, they regard the (re)presentational aspect of the experience as reducible to further facts. However, this isn’t a specific explanation of the phenomenology of presentation. The view suggested here entails that the phenomenological presentational feature of perceptual experiences is not basic, but rather it is formed by the relations among phenomenal features that may not be presentational in themselves. I propose that experiences have phenomenal qualities of type “1’ – similar to Pain1 – and when these features are organised into a certain order and structure, the impression of an external world emerges. The manifest image of the world – the world as it is presented by the senses – is constructed out of basic experiential features.6

7. Spices It is instructive to compare ordinary pain experiences caused by external injuries to the experience of tasting hot chilli peppers. The hotness of chilli lingers in the mouth, sometimes for quite a while, yet we attribute hotness to the chilli. Prima facie, it looks like hotness is one of the qualities that we taste in food; indeed in some classifications, the basic tastes are supposed to be sweet, sour, salty, bitter, and hot.7 This may seem like a problem for my claim that one of the main reasons we don’t externalise pains is that pains linger. But hotness lingers too, yet we do perceive chillies as hot. To understand this issue more clearly, it is worth saying a few words about the physiology of tasting spicy food. The compound that causes the pungent taste of hot chilli is called “capsaicin”. Capsaicin activates the so-called transient receptor potential channel. Transient receptor potential (or “TRP”) ion channels are distributed all over the body, and they form an interface between the environment and the sensory system, by acting as transducers of external (chemical, thermal or mechanical) stimuli in sensory neurons. The Vanilloid 1 member of the TRP channel family is called TRPV1, and it is activated by (among other things) noxius heat above 43°C. This degree of heat activates the channel, which transduces the stimulus into inward currents in nociceptive neurons. The nociceptive neurons send further signals to the brain, and this results in thermal and pain sensations. The interesting fact is that the very same channel, whose activation can be the first step of producing pain sensation in response to noxious heat, is also activated by capsaicin. Capsaicin binds to the TRPV1 receptor, and sets into motion the same process as noxious heat does. This can explain why the taste of chilli is described as “hot” in English, and also why people feel hot and start to sweat when they eat spicy food.8 The TRPV1 channel is active as long as capsaicin binds to the receptor. In a normal dose, capsaicin does not cause tissue damage; the spicy-hot sensation

Constructing a World for the Senses

111

(caused by the activation of the TRPV1 channel) lasts as long as capsaicin is to be found on the sensory surface. Of course, capsaicin may cause actual tissue damage at very high doses, but the same is true of pretty much every kind of stimulus. So in fact, in ordinary cases, spiciness does not linger beyond the presence of the spicy substance: the experience lasts exactly as long as some of the spicy substance is in the mouth (or on other part of the skin). This fact is reflected in the popular advice on how to alleviate the burning caused by hot food: namely, drink milk rather than water, or eat yoghurt. The difference lies in the fact that milk has lipids that dissolve capsaicin, but water doesn’t. It is tempting to suggest that the hot quality in food could be regarded precisely as an externalised pain quality – that is, as a variety of Pain2. It is an abiding characteristic of objects, it is sensed through a uniform procedure, it activates nociceptors and gives rise to pain sensations, and the presence of the hot substance is contemporaneous with the experiences of hotness. However, I find it difficult to judge whether this is the right analysis of the situation, because it is unclear to what extent we are aware of the capsaicin’s presence in the mouth while having the sensation. The presence of capsaicin is not obvious, because it is not detected by other senses: we can’t see it, and we can’t feel it in the way we feel when a quantity of food is in our mouth. We may ask then why hotness is still attributed to chillis. I believe that part of the explanation is that the experience of hotness is assimilated to, or classified together with taste experiences because we experience hotness mostly when tasting food. Our talk of tastes themselves oscillates between regarding them as experience-dependent and as experience-independent qualities, as shown by the earlier example of having a cold. This is an example that illustrates the very complex nature of factors that determine whether a sensory experience presents an object as having an experience-independent quality. Not only the phenomenal quality in itself, but also its relation to the behaviour of other, similar qualities, will play a role in determining whether the experience is presentational or not.

8. The Intentionality of Bodily Sensations I argued in section 4 that we don’t regard the pain-causing external thing as the object of the pain-experience, partly because the presence of the thing is often not contemporaneous with the experience. I also remarked that on this ground, it is much more plausible to claim that the injured bodily part is the object of the pain-experience, and indeed, this suggestion has been made by various philosophers (Crane 2003, Tye 2002). However, I also added that we cannot regard the experience as merely representing the bodily part as having an experience-independent property. The bodily part having an

112

phenomenal intentionality

experience-independent property cannot exhaust the representational content of the experience, because this would leave unexplained why, on the dominant notion of pain, pains are essentially felt. The following little story may help to illustrate the point. The daughter of a friend of mine once said to him: “daddy, my foot has gone to sleep, come and feel it”. She seemed to think that if her father touched her foot, he would also be able to feel the sensation of numbness she was feeling. We sometimes experience bodily states which are perceptible to others. I can hear your stomach rumbling, I can feel your heart beating fast, I can see the injury on your hand. Pins and needles seem different. One explanation of why my friend could not oblige his daughter to feel the numbness in her foot is that the numbness is essentially felt by her, by the subject of the experience. This is only one explanation, but it is strengthened by our earlier considerations about pain. If one takes a pain-killer, the injury may well remain, but the pain is gone as soon as the painful feeling is gone. Many bodily sensations are of this “mixed” kind: they are directed at experience-independent objects (one’s foot, for example, which would clearly exist even if it didn’t hurt or went numb), but there is more to the experience than presenting the bodily part as having an experience-independent property. One is aware of the bodily part in question in a subjective way that depends essentially on one’s having an experience. Hence these episodes are not full-blown perceptual experiences: their object is not wholly independent of the experience. Pains are not externalised: they don’t present experience-independent features of things outside the body, but they don’t present experience-independent features of bodily parts either.

9. Sensory and Non-Sensory Intentionality In the last decades, a number of philosophers have suggested that intentionality – or in any case, one specific kind of intentionality – is constituted by the phenomenal character of a mental episode. (Horgan and Tienson 2002, Loar 2003) Many of those who are convinced that this is right base their conviction on fairly simple introspection: when we reflect on some of our experiences, on what it’s like to have them, their presentational aspect is obvious to notice. And the differences between what it’s like to have one experience as opposed to another often lies in what they seem to present to us. As Kriegel (this volume) notes, this is hardly an argument that would convert sceptics to believers in phenomenal intentionality, but it is certainly an important factor for those of us who have already joined the faith. Once someone agrees that some of our experiences present the world in virtue of their phenomenal character, one possibility is to say that intentionality

Constructing a World for the Senses

113

or directedness is a basic, irreducible feature of some type of phenomenal characters. Another possibility is to pursue the theory that is outlined here and claim that the external directedness of sensory experiences is not a basic fact, but it is rather constructed by a complex structure of phenomenal qualities which are not presentational in themselves. The main advantage of the latter view is that it can explain that experiences with a certain type of phenomenal feature can be either externally directional or non-directional, depending on some further factors. For example, in our case, experiences with the feature Pain1 are not normally directed at external objects, but we can imagine circumstances where they would be. I believe the same applies to the red in an afterimage and the red in the experience of seeing a tomato. The same sensational feature can be a mere sensation of one occasion, and a presentation of an external object on the other. Sensory substitution systems provide another example of how the structure of sensory features creates an object for the senses. Tactile-visual sensory substitution systems (originally designed by Paul Bach-Y-Rita) convert visual stimuli to tactile stimuli. The subject has a camera mounted on his forehead, which records a black-and white image of an object in front of him. The image is converted into tactile stimulus (for example pressure or vibration) on a plate which is in contact with the subject’s back or tongue. After some practice, subjects learn to locate the position of things recorded by the camera (Bach-Y-Rita et al. 1969). There must be some difference between random twinges on the back and the stimulus provided by the plate converting the camera’s image. The suggestion is that the orderly structure of the latter is responsible for the presentation of a new object. This is a story of the intentionality of sensory experiences. The intentionality of thought seems to have an entirely different nature: it is not constructed of more basic, non-intentional phenomenal features. Some defenders of phenomenal intentionality endorse the claim that cognition itself has a phenomenology. If this is right, then one characteristic of the phenomenal features that lends thoughts their intentionality is that it is basic and not constructed. This is a fundamental difference between sensory and non-sensory intentionality.

Acknowledgments I am greatly indebted to Howard Robinson for first suggesting to me that the intentionality of perceptual experiences has to do with their structure. Papers bearing more or less close relations to this paper were given at the universities of Stockholm and Turin, at the Philosophers’ Rally in Lodz and the SEFA congress in Tenerife. I have greatly benefited from questions and comments on these occasions. I’m very grateful to Tim Crane for discussions, to Zsuzsa Helyes for advice,

114

phenomenal intentionality

and to Brie Gertler for helpful comments on the penultimate draft. The research leading to these results has received funding from the European Commission’s Seventh Framework Programme FP7/2007–2013 under grant agreement no. FP7–238128, and from the project BETEGH09 supported by MAG Zrt.

Notes 1. Th is might be an oversimplification in the case of pain, because it may not be the case that the the stimulus of some receptors invariably gives rise to the feeling of pain, independently of contextual or cognitive factors (see Rollman 1991). But taking these factors into account, we can still classify receptors according to the typical stimuli they react to. 2. Th anks to Howard Robinson for calling my attention to this example. 3. Michael Tye suggests that pains represent bodily injuries, and he thinks that we need not appeal to the subjective character of experience. He also acknowledges that pain features are exemplified only when a subject undergoes a certain kind of experience. But once this acknowledgement is in place, I cannot see what motivation is left to deny that the relevant feature is an experiential feature, rather than a feature of an experience-independent object. See Tye 2002. 4 . It may be suggested that talking of this quality makes sense only relative to the possibility of some conscious creature having a specific kind of experience. In a world where there are no conscious creatures at all, Pain2 would never be exemplified, so in some sense, Pain2 is a mind-dependent quality (in the way secondary qualities or response-dependent qualities are supposed to be mind-dependent). Perhaps this is correct, and that is why I don’t call this quality mind -independent, but rather experience -independent, where I mean to refer this to episodes of experience, rather than the possibility of experiences in general. Even if the existence of certain kind of conscious creatures is required, no actual conscious episode is needed for the exemplification of Pain2. 5. One aspect of this condition is explored by David Smith (Smith 2002) who argues that the necessary condition for perceptual experience is an awareness of the spatial nature of experience, meaning that the experience would change in characteristic ways depending on our spatial movement with respect to perceived objects. 6. Hence a more precise, but unfortunately more cumbersome, title for this paper would have been ‘Constructing the manifest image of the world for the senses”. I do not propose to defend idealism: I do not think that the world itself consists of experiential features (though my view about perceptual experiences is compatible with an idealist theory). But I do think that the image of the world is constructed from basic experiential features. 7. Strictly speaking, we may not want to say that hotness (in the sense of spiciness) is a taste: all parts of the skin, and especially mucous membranes, are sensitive to hot chilli, as everyone who accidentally touched her eyes after slicing chilli will know. In contrast, if say, a few drops of Coke get in our eye, we won’t feel it’s sweet with the eye: to detect sweetness, we need the special apparatus of taste buds. However, whether hotness is a taste or not, it’s clear that it is attributed to chilli peppers. 8. Piperine, the compound responsible for the pungency in black peppers also activates the TRPV1 channel. So does zingerone, an ingredient of ginger. Other TRP channels are activated by other thermal stimuli and other food ingredients. TRPM8 is activated by cold temperature and by menthol. TRPA1 is activated by mustard oil and cinnamon; the range of noxious stimuli that activate TRPA1 is debated. On TRP receptors and pain see also Wang and Woolf 2005.

Constructing a World for the Senses

115

References Bach-y Rita, P., C. C. Collins, F. Saunders, B. White, and L. Scadden. 1969. “Vision Substitution by Tactile Image Projection.” Nature 221: 963–964. Crane, T. 2003. “The Intentional Structure of Consciousness.” In Alexander Jokic and Quentin Smith (eds.), Consciousness: New Philosophical Perspectives. Oxford: Oxford University Press. Grice, H. P. 1962. “Some Remarks about the Senses.” Reprinted in Alva Noë and Evan Thompson (eds.), Vision and Mind: Selected Readings in the Philosophy of Perception. Cambridge, MA: MIT Press, 2002, 36–54. Horgan, Terence, and John Tienson. 2002. “The Intentionality of Phenomenology and the Phenomenology of Intentionality.” In David Chalmers (ed.), Philosophy of Mind: Classical and Contemporary Readings. Oxford: Oxford University Press, 520–531. Loar, Brian. 2003. “Phenomenal Intentionality as the Basis of Mental Content.” In Martin Hahn and Bjørn Ramberg (eds.), Reflections and Replies: Essays on the Philosophy of Tyler Burge. Cambridge, MA: MIT Press, 229–258. Locke, John. 1690. An Essay Concerning Human Understanding. New York: Penguin Classics, 2004. Rollman, G. B. 1991. “Pain Responsiveness.” In M. A. Heller and W. Schiff (eds.), The Psychology of Touch. Mahwah, NJ: Erlbaum, 91–114. Smith, David.2002. The Problem of Perception. Cambridge, MA: Harvard University Press. Tye, M. 2002. “Representationalism and the Transparency of Experience.” Noûs 36, 137–151. Victor, T. W., M. P. Jensen, A. R. Gammaitoni, E. M. Gould, R. E. White, and B. S. Galer. 2008. “The Dimensions of Pain Quality: Factor Analysis of the Pain Quality Assessment Scale.” Clinical Journal of Pain 24(6):550–555. Wang, Haibin, and Clifford J. Woolf. 2005. “Pain TRPs Minireview.” Neuron 46, 9–12. Wittgenstein, Ludwig. 1953. Philosophical Investigations. Translated by G. E. M. Anscombe. New Saddle River, NJ: Prentice Hall, 1999.

7

Phenomenal Objectivity and Phenomenal Intentionality In Defense of a Kantian Account Fa r i d M a srou r

Perceptual experience has the phenomenal character of encountering a mindindependent objective world. What we encounter in perceptual experience is not presented to us as a state of our own mind. Rather, we seem to encounter facts, objects, and properties that are independent from our mind. In short, perceptual experience has phenomenal objectivity. Phenomenal objectivity distinguishes perceptual experience from those types of experience, for example mood experiences, that have the phenomenal character of presenting to one only the states of one’s own mind. An account of phenomenal objectivity would be useful to believers in phenomenal intentionality—a form of intentionality that is constituted by phenomenality. This chapter proposes and defends a Kantian account of phenomenal objectivity.

Introduction Some experiences, such as moods, emotions, nausea, and dizziness present our own states to us. We can call these experiences self-presenting experiences.1 In contrast to self-presenting experiences, perceptual experiences typically present external things to us. We can say that perceptual experience belongs to the class of other-presenting experiences.2 The distinction between self-presenting and other-presenting experiences is, in my view, a phenomenological distinction. Self-presenting and other-presenting experiences differ not only with respect to what they typically present, but also with respect to their phenomenal character.3 There is a general phenomenological feature in virtue of whose presence/absence experiences are 116

Phenomenal Objectivity and Phenomenal Intentionality

117

other-presenting/self-presenting. I call this feature phenomenal objectivity. The claim that perceptual experiences typically have phenomenal objectivity is a substantive claim and I shall say more in its defense later in the chapter. At this stage, however, I shall take it for granted.4 This chapter outlines and defends a phenomenological account of phenomenal objectivity. By this I mean an account that offers a reduction of phenomenal objectivity to allegedly more fundamental phenomenal features.5 I believe that this account would be beneficial for the phenomenal intentionalist. Let me explain why. Phenomenal intentionalists typically hold that there is a kind of intentionality that is constituted by phenomenality.6 This idea implies that the conditions that a state needs to satisfy in order to count as a state with phenomenal intentionality are phenomenal conditions. One task for the phenomenal intentionalist is thus to provide an account of the phenomenal conditions that constitute phenomenal intentionality. We can call this the constitution problem. One way to solve the constitution problem is to equate the conditions for phenomenal intentionality with the conditions for phenomenal objectivity. It is easy to see why one might be tempted to do so. Note that experiences with phenomenal objectivity have other-presenting phenomenal character. One might regard this feature as constituting the core of phenomenal intentionality. It is in virtue of having this phenomenal directedness toward the other that experience has phenomenal intentionality. So it is in virtue of having phenomenal objectivity that experience has phenomenal intentionality. Thus a theory that explains the phenomenal conditions in virtue of which experience has phenomenal objectivity would thereby explain the phenomenal conditions in virtue of which experience acquires phenomenal intentionality. In this construal, explaining phenomenal objectivity would be at the core of any phenomenal intentionality research program by providing a solution to the constitution problem. The phenomenal intentionalist might, however, refrain from equating the phenomenal-intentional with the phenomenal-objective. The phenomenal intentionalist might hold that some states that possess phenomenal intentionality lack phenomenal objectivity. For example, on some views, self-presenting states, such as mood experiences, display phenomenal intentionality. One who assumes this will not look at a theory of phenomenal objectivity as a theory of phenomenal intentionality. Nevertheless, she can regard such a theory as a theory of other-presenting phenomenal intentionality. This theory would not explain the phenomenal conditions that are constitutive of phenomenal intentionality. Rather, it would explain the phenomenal conditions in virtue of which phenomenal intentionality is differentiated into other-representing and self-representing types. We can call the problem of providing an account of the phenomenal conditions in virtue of which phenomenal intentionality gets

118

phenomenal intentionality

divided into its main types the division problem. Assuming that self-presenting and other-presenting forms of phenomenal intentionality are its main types, we can conclude that a theory of phenomenal objectivity would solve the division problem. An account of phenomenal objectivity can thus be beneficial to the phenomenal intentionalists by solving either the constitution problem or the division problem. I hold that a theory of phenomenal objectivity solves the constitution problem. But for the purposes of this chapter I shall not defend this idea against the alternative reading, which claims that a theory of phenomenal objectivity solves only the division problem. The reader can decide how she or he wants to look at this issue. On either choice, an account of phenomenal objectivity is important for the phenomenal intentionalist. The view that I will be offering in this chapter is primarily inspired by my reading of Kant’s Critique of Pure Reason. Thus, I regard it as a Kantian view. The next section explains the view and clarifies why I call it Kantian. The second section connects the view to contemporary psychology of perception and provides a preliminary motivation for it. The last section provides an argument for the view. In the remainder part of this section, I say more about what I mean by a phenomenological account. The account offered in this chapter aims at answering the following question: what are the phenomenal facts in virtue of which perceptual experience has phenomenal objectivity? This question is a phenomenological question. In saying this, I want to insist that an answer to this question should appeal to phenomenological facts. For example, an explanation in terms of the sub-personal mechanisms that distinguish experiences with phenomenal objectivity from those that lack it would not do. Such an answer, insofar as it leaves out matters at the personal level, is not a phenomenological answer.7 Similarly, one could not answer this question by appealing to the fact that the brain processes that underlie perceptual experience track external facts while the brain processes that underlie other experiences, such as emotions, only track the conditions of the brain and the body. For appealing to tracking relations is appealing to facts that fall outside what is given at the personal level. In general, a phenomenological question seeks an answer at the personal or phenomenological level. Thus, it can only be satisfactorily answered by appealing to how things are given in, or to, phenomenal consciousness.8 Obviously, such remarks do not offer a positive characterization of a phenomenological answer. They only explain what would not count as one. So let me say a few words by way of a positive characterization. As I’m using the term, phenomenological answers are given in terms of instantiations of phenomenal properties. Such answers sometimes appeal to the instantiation of primitive monadic phenomenal properties. For example, if you ask how visual experience presents redness, a short answer could be that it does so in virtue

Phenomenal Objectivity and Phenomenal Intentionality

119

of instantiating a monadic phenomenal property that some call phenomenal red. If one is a primitivist about phenomenal red then one’s phenomenological explanation ends here, because in the primitivist account the most fundamental phenomenal fact that explains the phenomenal character associated with experiences of red is the instantiation of phenomenal redness. However, phenomenological explanations can sometimes be reductive. For example, one might reduce phenomenal redness to more primitive phenomenal properties such as phenomenal hue, saturation, and brightness. Such a reductivist would hold that phenomenal hue, saturation, and brightness are more fundamental than phenomenal redness. Thus the instantiation of phenomenal redness is metaphysically explained in terms of instantiations of these more fundamental phenomenal properties. This would be a case where a phenomenological explanation is informative for being reductive. Other examples of this strategy are Hume’s reduction of the impression of necessary connection to the “determination of the mind” to move from one idea (or impression) to another and Berkeley’s view, which according to some commentators, reduced experiences of spatial relations to expectations involving tactile or proprioceptive experiences.9 The account that I shall be offering here is reductive in this sense. I shall argue that phenomenal objectivity is constituted by the instantiation of more fundamental properties. Explaining what these properties are requires some stage setting. This is what I shall do in the next section.

1. The Kantian Thesis There is strong textual evidence that a decade before the publication of the first Critique Kant had become interested in what we nowadays call the problem of intentionality. In a famous letter to Herz, he writes about this problem and its significance: I noticed that I still lacked something essential, something that in my long metaphysical studies I, as well as others, had failed to pay attention to and that, in fact, constitutes the key to the whole secret of hitherto still obscure metaphysics. I asked myself, namely: What is the ground of the relation of that in us which we call ‘representation’ to the object? (Correspondence, Letter to Herz, 1772, emphasis mine) Kant’s formulation of the problem of metaphysics changes in the first Critique. There, he describes the problem as that of explaining the possibility of synthetic a priori cognition. Arguably this change in formulation does not mark a radical change in Kant’s conception of the problem. Rather, it indicates his

120

phenomenal intentionality

realization that a solution to the problem will also show how we can have a type of cognition that is both a priori (independent from experience) and synthetic (not determined only by conceptual relations). There are still passages in the Critique, however, where Kant explicitly talks about the problem of the relationship between representations and objects. For example, in the Second Analogy, after remarking that pure representations are “inner determinations of our mind in this or that temporal relation,” he asks: Now how do we come to post an object for these representations, or ascribe to their subjective reality, as modifications, some sort of objective reality? (CPR, A197/B243) The answer comes in the same paragraph: If we investigate what new characteristic is given to our representations by the relation to an object, . . . we find that it does nothing beyond making the combination of representations necessary in a certain way, and subjecting them to a rule; and conversely that objective significance is conferred on our representations only insofar as a certain order in their temporal relation is necessary. (CPR, A197/B243)10 Kant’s question is how representations that are mere modifications of the mind acquire “objective significance.” His answer seems to be that representations acquire their objective significance in virtue of necessary temporal connections between them. In the Schematism section that precedes the Analogies, Kant has already argued that this necessary temporal connection depends on synthetic activities that somehow involve the schemas of the pure concepts of understanding or categories. My contention is that both Kant’s question and his answer can be read in a phenomenological-cum-psychological way. Accordingly, to ask how representations acquire “objective significance” is to ask how representations acquire phenomenal objectivity, and to say that “objective significance is conferred to our representations only insofar as a certain order in their temporal relation is necessary” is to say that representations acquire their objective significance in virtue of having the phenomenology of being combined in time in accordance with schematic rules. Schemas, in the interpretation that I wish to propose, embody the rules that guide the mental activity of synthesis. Th is activity unifies the members of the manifold of representations in time, and the phenomenology of combining mental representations in time in accordance with schematic rules is what confers objective significance to our representations which without this will be mere modifications of the subjective states of the mind and thus at best only self-presenting.

Phenomenal Objectivity and Phenomenal Intentionality

121

This interpretation can of course be challenged. First, Kant’s notions of objective significance and the necessary combinations of representations can both be interpreted non-phenomenologically. Second, the relationship between synthetic activities and the combination of representations can be understood in different ways. Should we understand Kant’s idea of combination as a by-product of an act of synthesis, or should we understand combination as identical with an ongoing activity? In the first interpretation, acts of synthesis produce a complex relational structure in our representational manifold, where the relationship between the act and the structure is analogous to the relationship between the activity of building and the structure that results from it. In the second interpretation, synthetic activity constitutes the relational structure in the manifold. On this view representations are related to each other in virtue of the fact that an ongoing synthetic activity in time somehow incorporates them. My proposed interpretation attributes to Kant the second type of view. There are also important interpretational puzzles about Kant’s account of schemas. In some passages, Kant writes as though schemas are rules that ground the activities of synthesis.11 It is tempting to rely on these passages and equate schemas with a type of representation that guides the activity that constitutes or produces combinations among other representations in time. However, there are passages in which Kant seems to equate schemas with the structure that is produced or constituted by synthetic activities.12 These passages suggest that “schema” is just Kant’s fancy term for talking about combination in time. My interpretation attributes to Kant the view that schemas are representations that guide activities of synthesis. A related puzzling issue is the relationship between schemas and concepts, in particular the relationship between transcendental schemas and the categories. Before the Schematism section Kant often speaks as though the categories furnish the rules that guide the activities of synthesis. However, in the Schematism section this role seems to be assigned to transcendental schemas.13 There are at least three ways to resolve this tension. One option is to equate the transcendental schemas with the categories. The other is to equate transcendental schemas with the structure that results from or is constituted by the activity of synthesis and assign pure concepts the role of the rules that guide these activities. A third option is to equate schemas with the guiding rules and interpret Kant’s earlier talk about the relation between the categories and the rules as presupposing schematic mediation. Here, my interpretational choice has been the third one. Another important feature of Kant’s view is his idea that schemas are neither sensible nor intellectual. In this view, a schema is a specific type of representation that is intermediary between conceptual and non-conceptual representations and belongs to the faculty of imagination.14 Kant’s account

122

phenomenal intentionality

of imagination has been the subject of intense exegetical controversy partly because of its apparent incompatibility with his division of mental faculties into sensibility, understanding, and reason. I want to claim here, again without argument, that Kant’s notion of a schematic representation can be equated with perceptually encapsulated innate knowledge. My interpretation attributes to Kant a set of phenomenological and psychological theses based on the mentioned interpretational choices. All of these interpretational choices require careful exegetical defense. However, this chapter is not the place to do so. Thus, rather than claiming that the view that I am describing is Kant’s, I shall call it a Kantian view. My main aim is to defend the phenomenological component of the Kantian view. As a first approximation, the thesis is that phenomenal objectivity is constituted by the phenomenology of a necessary combination of representations in time. I have explained what I mean by phenomenal objectivity before. Let me use an example to explain how I understand the idea of the phenomenology of a necessary combination of representations. Imagine walking toward a tree as you are looking at it. When you get closer, the visual angle through which you see the tree grows in size in an inverse relation to your distance from the tree. If things go well and your perception is veridical your representation of the visual angle through which you see the tree and your representation of your distance from the tree co-vary with each other in a law-like manner. How should we construe your visual phenomenology in this case? First, there is something that it is like to visually experience an object through a particular viewing angle. This aspect of your visual phenomenology changes as you get closer to or farther away from the tree. Second, there is something that it is like to visually experience an object to be at a particular distance from you. This aspect of your visual phenomenology also changes as you get closer to or farther away from the tree. These two aspects do not change in isolation. They co-vary in a law-like manner. Now, one might hold that there is a third phenomenological element here. There is something that it is like to experience the visual angle and the relative distance as co-varying in the particular law-like manner that they do and this additional phenomenological element is over and above the law-like covariance of the other two experiences and is constant throughout their change. My proposal is that we should understand Kant’s talk about the necessary combination of representations in time in this way. Let us call the phenomenology associated with experiencing two properties as co-varying in accordance with the rules embodied in schemas Schematic Dynamical Unity. The phenomenological component of the Kantian view claims that there is a constitutive relation between schematic dynamical unity and phenomenal objectivity. Throughout the chapter, I shall refer to this claim as the Kantian thesis and present it in the following way:

Phenomenal Objectivity and Phenomenal Intentionality

123

Kantian thesis: Experience has phenomenal objectivity in virtue of having Schematic Dynamical Unity.15 Why should we accept the Kantian thesis? Understanding Kant’s own argument for the thesis requires engaging with the complex dialectic structure of the Critique. Rather than trying to do so, I shall offer my own reasons for embracing the Kantian thesis. One of these reasons has to do with the psychological aspect of the Kantian thesis. In the next section I shall show that the empirical science of perception supports the psychological analogue of the Kantian thesis. The second reason is phenomenological. The last section offers a phenomenological argument in support of the Kantian thesis.

2. The Kantian Thesis and Empirical Psychology In this section I want to argue that experimental psychology seems to support what we might call the psychological counterpart of the Kantian thesis and thereby indirectly supports the Kantian thesis. By the psychological counterpart of the Kantian thesis I mean the thesis that the psychological mechanisms that underlie schematic dynamical unity are the same as (or an important component of) the psychological mechanisms that underlie phenomenal objectivity. If the psychological mechanisms are related in this intimate way then it is reasonable to assume that the phenomenal properties that are grounded in them are also related in an intimate way. What are the psychological mechanisms that underlie phenomenal objectivity? The answer to this question is not easy, because psychologists hardly ever use concepts such as phenomenal objectivity. In order to answer this question we need to connect our phenomenal talk to the psychological talk. In order to do so, I want to first draw your attention to an important feature of experiences with phenomenal objectivity. Our everyday perceptual experiences incorporate a duality in their presentational content. They inform us about the properties of objects around us, but that is not all they do. Our everyday perceptual experiences also inform us about the point of view that we occupy in relation to objects. When you look at a tilted coin, you typically experience it as a circular object. But you also experience something about the spatial relation that you have to the coin. For example, you experience the coin as tilted in relation to what psychologists often call the fronto-parallel panel, an imaginary plane that passes through your two eyes. Here, your point of view is determined mainly, but not exhaustively, by the location that you and the coin occupy in physical space and the physical shape of your perceptual apparatus. A similar point applies to color experience and experiences in other perceptual modalities. When you see a uniformly red table, you typically experience it as a red object. But your visual experience also

124

phenomenal intentionality

informs you of your relation to the table with respect to what we might call the illumination space. If the table is lit by non-uniform light, your experience of the table informs you of that. You occupy a point of view on the table in the illumination space. The general point here is twofold: firstly, experiences with phenomenal objectivity always have the duality that is illustrated in the examples given earlier and secondly, self-presenting experiences lack it.16 Nausea, pains, itches, and sadness are all self-presenting experiences. They inform you of your own properties and qualities. However, they do not inform you of anything like a point of view that you have in relation to these qualities. I want to insist that all and only experiences that have this duality are experiences with phenomenal objectivity. Thus, if we want to study the psychological mechanisms that underlie phenomenal objectivity we should study the psychological mechanisms that underlie this duality. The good news for us is that psychologists study the phenomenon of duality under the title of perceptual constancies. We encounter perceptual constancies when despite changes in our point of view the properties that we experience objects as having remain constant. Psychologists hold that visual representations of many properties such as shape and color display perceptual constancies. It is also widely held that perceptual constancies are not confined to vision. For example, representations of tactile texture, solidity, and auditory representations of volume display perceptual constancies. When we study the mechanisms that underlie perceptual constancies we can see that this characterization is exactly what we expect from a characterization of the mechanism underlying schematic dynamical unity. Let me illustrate this with an example. When you walk toward a tree as you are looking at it, the image that the tree projects on your retina grows in size. But you do not experience the tree as growing in size; you experience its size as constant. Psychologists often call this phenomenon size constancy and the aspect of size that remains constant intrinsic size. I shall refer to it as objective size. There is some controversy about the mechanisms underlying the computation of objective size, but most psychologists hold that our visual system computes objective size based on the relationship between relative distance and visual angle.17 According to this theory, the computation of objective size is based on the following mathematical formula: bc = 2d.tan(θ/2), where bc is the objective size of the object, d its relative distance from the subject, and θ the viewing angle (see Figure 7.1).18 This computational mechanism has the exact structure that we expect from a mechanism that would underlie schematic dynamical unity. I claimed earlier that schematic dynamical unity consists in experiencing properties as co-varying in accordance with a particular rule. Experiences with schematic

Phenomenal Objectivity and Phenomenal Intentionality

b

125

c’ θ b’

c

d

d’

Figure 7.1 Size Constancy.

dynamical unity thus have a specific structure that involves the experience of certain properties and the experience of their law-like correlation. Since it is reasonable to assume that the psychological phenomena that underlie experience mirror the structure of experience, we should expect the psychological mechanisms underlying schematic dynamical unity to contain a mechanism for determining whether the value of some of representations co-vary with each other in accordance with a specific rule. And this is exactly what we see in the standard account of size constancy. In this proposal, the brain determines whether the product of the value of relative distance and the viewing angle stays constant in order to represent the intrinsic size of the tree as constant. If this reasoning is correct, then it is reasonable to assume that at least in the case of size constancy the psychological mechanism that underlies phenomenal objectivity has the same structural signature that we expect from the psychological mechanism that would underlie schematic dynamical unity. I want to claim without further argument that this observation generalizes to other cases of constancy. If this claim is correct then we will have a strong empirical argument in support of the Kantian thesis. At the moment I do not know whether this claim is correct. But I have said enough to motivate the weaker claim that the empirical science of perception seems to support the Kantian thesis. Armed with this, I want to move to the next section where I give a phenomenological argument for the Kantian thesis.

3. The Abductive Argument According to the Kantian thesis, perceptual experience has phenomenal objectivity in virtue of instantiating schematic dynamical unity. In this section I want to provide an abductive phenomenological argument in support of this thesis. I argue that the Kantian thesis provides the best explanation for the absence of phenomenal objectivity in a paradigm case where it is absent. To do so, I want to first present the reader with the particular case in question.

126

phenomenal intentionality

Figure 7.2 Afterimage.

Focus on the black circle in the left rectangle in Figure 7.2 for a while and then shift your focus to the circle on the right. You will gradually experience an afterimage. The afterimage is not experienced as a pattern on the page in the same way that the left picture is. We do not experience the afterimage as an objective pattern on the page. The experience of the afterimage does not manifest phenomenal objectivity. What could account for the absence of the phenomenal objectivity? I believe that the Kantian thesis provides the best explanation. This idea can be articulated in the following argument (from here on I shall refer to schematic dynamical unity as SDU): Abductive argument: (1) Afterimages lack phenomenal objectivity. (2) The best explanation for (1) is that SDU is necessary for phenomenal objectivity. (3) SDU is sufficient for phenomenal objectivity. (4) The best explanation for (2) and (3) is that SDU constitutes phenomenal objectivity. Conclusion: SDU constitutes phenomenal objectivity. What follows defends each premise of the argument. According to the first premise, afterimages lack phenomenal objectivity. This idea is intuitive. Nevertheless, one can argue for it by holding that the primary reason for our beliefs that afterimages are not objective is that the perceptual presentation of afterimages lacks phenomenal objectivity. This argument can be resisted in two ways. First, one might hold that afterimages have phenomenal objectivity, but because of some post-perceptual inference we do not believe that they are objective. I shall call this proposal the post-perceptual defeat proposal. A second way to resist the idea that afterimages lack phenomenal objectivity is to hold that afterimages do not have a verdict on matters of objectivity/non-objectivity. This might be attractive to someone who denies that perceptual experience and sufficiently similar experiences, such as afterimages, have any objective/non-objective content. Experience is silent about matters of objectivity/non-objectivity. Our beliefs to the effect that

Phenomenal Objectivity and Phenomenal Intentionality

127

Figure 7.3 Mü ller-Lyer illusion.

afterimages are not objective should be explained in terms of post-perceptual inference. I shall call this the perceptual silence proposal. In what follows, I want to show that these proposals are problematic. Let us start with the post-perceptual defeat proposal. In this proposal, the verdict of experience is that afterimages are objective, but this verdict is defeated post-perceptually. However, this idea has counterintuitive implications. In particular, it implies that in the case of afterimages there is a discrepancy between the deliverances of experience and our beliefs regarding the status of afterimages. But this seems implausible. Let me elaborate. The typical example of discrepancies between perceptual content and beliefs is the Mü ller-Lyer illusion (Figure 7.3). The two arrows in Figure 7.3 seem to have unequal lengths, but in fact their lengths are equal. Your perceptual experience provides prima facie reason for you to believe that the lines are unequal. Thus, it is likely that you will form the belief that the lines are unequal if you do not know that this is an illusion. But if you know about the illusion your background beliefs might defeat the verdict of the experience and you might form the belief that the two lines are equal. As a result, there will be a discrepancy between what your experience presents and what you believe. And this discrepancy is apparent to you. To generalize, when the verdict of experience is defeated post-perceptually we normally encounter an apparent discrepancy between our beliefs and the verdict of experience. But this discrepancy is absent in the case of afterimages. When we form the belief that afterimages are not objective it does not feel as though there is a discrepancy between our belief and the verdict of experience. So the post-perceptual defeat proposal should be rejected. Let us turn to the second proposal. In this proposal, experience is silent about matters of objectivity/non-objectivity. It presents afterimages in the same way that it presents normal colors. The only difference is that we believe that the colors of afterimages are not objective because of post-perceptual reasoning. Now, “reasoning” could be understood in two different ways, namely, conscious or subconscious. The first option does not seem to be compatible with what we are conscious of. When we experience afterimages, it does not seem that we engage in any reasoning to the effect that they are not objective. So the only reasoning that might be going on here is subconscious reasoning.19

128

phenomenal intentionality

Suppose the reasoning is subconscious. When we experience an afterimage some sub-conscious reasoning leads us to believe that the afterimage is not objective. But this again has a counterintuitive implication. The reasoning that results in the belief is subconscious, and the experience does not present the afterimage as objective. One implication of this is that we should have no idea why we have such a belief. From our point of view, all that happens is that the belief that the afterimage is not objective pops up in our head. But we can point neither to our experience nor to any conscious reasoning as the ground for this belief. The above result, however, is very counterintuitive. An example helps us see why. Compare a person with normal vision with a person with blindsight. In the standard construal, blindsight patients do not have visual experiences, but if prompted in the right way they can correctly guess some of the features of objects at a rate higher than chance. If you ask a normal person why she thinks the painting in front of her is tilted, she can point to it and say: “because it looks that way.” Her experience presents her with something that she takes as evidence for her belief. But the same response is not available to a blindsight patient. From her point of view, the idea that the thing in front of her is tilted just pops up in her mind. Her belief is epistemically blind.20 This helps us see why the above proposal is problematic. For it implies that our situation with respect to afterimages is analogous to the case of blindsight. But it is very implausible that this is our situation. For example, learning that her belief concerning the orientation of the object is correct would surprise a patient with blindsight. For, from her point of view, her guess is groundless. In contrast, learning that afterimages are not objective does not surprise us. After all, they always looked that way. I thus hold that both the post-perceptual defeat and the perceptual silence proposals are problematic. So we have good support for the first premise of our argument: afterimages lack phenomenal objectivity. According to the second premise of the argument, the best explanation for the absence of phenomenal objectivity in the case of afterimages is that SDU is necessary for phenomenal objectivity. I shall defend this claim by showing that (a) afterimages lack SDU and (b) other explanations for why afterimages lack phenomenal objectivity are problematic. We can support the claim that afterimages lack SDU by the following observations. First, afterimages do not satisfy some implicit anticipatory conditionals that are satisfied in the case of normal objects. For example, if you blink as you are looking at the left rectangle in Figure 7.2 it will appear as soon as you open your eyes. But the afterimage does not appear the moment you open your eyes; it is absent for a short period after blinking and gradually appears again. If you tilt your head as you’re looking at the left rectangle it will not rotate. But if you tilt your head as you are experiencing the afterimage it will disappear for a short period and then appear rotated. If you move the page closer to yourself as you are looking at the rectangle its perceived size will not change. But if you

Phenomenal Objectivity and Phenomenal Intentionality

129

move the page closer, the afterimage appears smaller after a moment of disappearance. Second, we are not surprised that these anticipatory conditionals are not satisfied. Our little phenomenological experimentation with the afterimage does not tell us that the afterimage is behaving in a way that breaches our expectations. We simply do not have the implicit expectations that we have in the case of normal color patches. These two observations could be put in the service of an argument that shows that afterimages lack SDU. Since afterimages do not satisfy normal anticipatory conditionals, and they do so without surprising us, we should conclude that there are no such implicit anticipations in the first place. The experience of the afterimage does not have the normal anticipatory contents that the experience of real colors has. This in turn implies that afterimages lack schematic dynamical unity, because it is plausible to assume that there is an intimate relationship between SDU and the specific sorts of anticipatory conditionals about how experiences co-vary with each other.21 I have argued that afterimages lack SDU. This supports the claim that one explanation for the lack of phenomenal objectivity in the case of afterimages is that SDU is necessary for phenomenal objectivity. In order to show that this is the best explanation, we need to rule out the claim that the absence of other phenomenological features that are necessary for phenomenal objectivity explains the absence of phenomenal objectivity. I shall do so by considering a few initially plausible candidates. One option is that the afterimage is partially transparent. So the afterimage in Figure 7.2 is more transparent than the rectangle on the right side. But it seems implausible that non-transparency is necessary for phenomenal objectivity, for many patterned surfaces that we see in our everyday encounter with the world are partially transparent. But this does not make us experience the patterns as non-objective. We see through colored windows, glasses and so on. These are partially transparent surfaces. But the partial transparency of the patterns on these surfaces does not make them lose their apparent objectivity. Another candidate is that afterimages do not have the same degree of intensity as normal experiences. Sounds can be loud, and smells can be strong. Perhaps a similar feature could be ascribed to experiences of colors in terms of brightness or saturation. It might be plausible that afterimages are less intense in this sense; they are not as bright or saturated as normal color experiences. But again, this cannot be the reason for the absence of phenomenal objectivity. We experience many colored surfaces in our everyday encounters with objects that are not bright or saturated. But this does not make us experience those colors as non-objective. I conclude that explanations in terms of transparency or intensity cannot account for the absence of phenomenal objectivity in the case of afterimages.

130

phenomenal intentionality

These proposals have implausible implications. Let me thus turn to a proposal that does have implausible implications. I shall argue that, nevertheless, the Kantian view is preferable to this proposal. It is not uncommon to associate phenomenal objectivity with the representation of objective space.22 One might hold that to represent a property as objective is to represent it as instantiated out there, somewhere deep in space. Can we invoke this idea to account for the lack of phenomenal objectivity in afterimages? This might strike some as tempting, but, in my view, it is ultimately unsatisfactory. Before explaining the reason let us first see how this proposal might be applied to the case of afterimages. According to this proposal, afterimages are not strictly speaking experienced as having an objective location deep in space. Consider Figure 7.2 again. It is initially tempting to think that the afterimage is experienced as located on the sheet of paper. So it is initially tempting to hold that afterimages are experienced as located in space, namely, where the paper is. However, further reflection shows that this is wrong. Afterimages are in fact experienced as being overlaid on our visual field. The surface of the paper provides the background that is required for experiencing afterimages. So afterimages are not experienced as being “out there.” The proponent of the spatial location proposal appeals to this feature to explain why afterimages are not experienced as objective. The spatial location proposal assigns a central role to the experience of objective location. Now, what do we mean by experiences of objective location? If you move your head or eyes in different directions, the retinal location of the projected image of objects changes, but you do not experience the objects as moving. This is a case of location constancy. There is something like experiencing objects as having location constancy, which we can call the experience of objective location. According to the spatial location proposal, the phenomenology of objective location is necessary for phenomenal objectivity. This is why its absence in the case of afterimages explains the lack of phenomenal objectivity. We can distinguish between two versions of this proposal. In an unrestricted version, for any property, p, we experience p as objective partly in virtue of experiencing it as having the phenomenology of objective location. A more restricted version of the proposal limits it to a particular set of properties. Accordingly, there is a class of properties such that we experience them as objective partly in virtue of experiencing them as having objective locations. The unrestricted version of the spatial proposal seems problematic. Experiences of many properties display phenomenal objectivity. We can experience shapes, sizes, locations, colors, tactile textures, tactile solidity, auditory volume, and so on as objective. But it is not clear how experiencing something as having objective location would be necessary for experiencing some of its

Phenomenal Objectivity and Phenomenal Intentionality

131

other properties as objective. Consider color constancy, for example. Due to changes in illumination throughout a day, the reflected wavelength of objects changes. But we do not experience their objective colors as changing. On the face of it, this seems to be completely independent from experiencing these objects as located deep in space. The same could be said about experiences of objective tactile texture and similar encounters. The proponent of the spatial location proposal should thus adopt the restricted version of the view. The seemingly plausible option would be to restrict the view to spatial features such as shape and size. Maybe the phenomenology of objective location partly constitutes the phenomenology of objective size and shape. But it seems to me that the Kantian thesis is a better explanation than the restricted version of the objective location view. This has two reasons. First, the Kantian thesis is non-restricted. The idea of schematic dynamical unity applies to all cases of constancy. So unlike the restricted spatial view that needs to provide a different explanation for objective representations of non-spatial properties, the Kantian thesis is a unified theory. Second, the Kantian thesis can explain the phenomenology of objective location. But on the spatial location view, the phenomenology of objective location should be regarded as primitive. If there is an explanation for it, it is sub-personal. I thus conclude that the Kantian thesis provides a better explanation for the absence of phenomenology of objectivity than the spatial location view. This ends my reasoning against the alternative explanations of the lack of phenomenal objectivity in the case of afterimages. I have argued that explanations in terms of transparency, lack of intensity, and the absence of experienced objective location are either implausible or less desirable than the Kantian explanation. So, in the absence of other candidates, it is reasonable to suppose that schematic dynamical unity is necessary for phenomenal objectivity. Let us summarize what we have so far accomplished. My aim in this section has been to defend the following argument: Abductive argument (1) Afterimages lack phenomenal objectivity. (2) The best explanation for (1) is that SDU is necessary for phenomenal objectivity. (3) SDU is sufficient for phenomenal objectivity. (4) The best explanation for (2) and (3) is that SDU constitutes phenomenal objectivity. Conclusion: SDU constitutes phenomenal objectivity. So far, I have defended the fi rst two premises of the argument. In what follows I shall defend the last two premises.

132

phenomenal intentionality

According to the third premise, schematic dynamical unity is sufficient for phenomenal objectivity. My main support for this is that reflection on our experience supports it. Of course, this is an empirical claim. But let me provide some preliminary support for it by way of blocking an objection. I have characterized schematic dynamical unity as the experience of the covariance of properties in accordance with certain rules or laws. But one might hold that this characterization is too easy to satisfy, because rules are easy to come by. One can even gerrymander rules that apply to the way afterimages change. So it follows that schematic dynamical unity is present even in the case of afterimages and cannot be sufficient for phenomenal objectivity because afterimages lack phenomenal objectivity. To see the flaw in this reasoning it helps to distinguish between the Kantian thesis and a simple regularity account of phenomenal objectivity. Such an account identifies phenomenal objectivity with regularity in the course of experiences. On this account it suffices that representations of some properties co-vary in accordance with some rule. But it is not required that these rules be associated with a specific form of phenomenology nor that they depend on a specific psychological mechanism. In the Kantian view, in contrast, the experience of schematic dynamical unity requires the activation of schematic representations which are associated with a specific form of phenomenology. A schematic representation is in effect a detector whose job is to test whether certain dynamic relations between the values of certain representations obtain. An abstract arbitrary rule that does not correspond to such a detector does not count as a schema. Thus in the Kantian view, schemas are not arbitrary rules; they should correspond to psychologically real detectors whose activation is phenomenologically manifest. Our imaginary interlocutor might object that acquiring psychologically real detectors is easy. One can simply internalize rules based on past experience. For example, after a few encounters with afterimages one might form expectations that are satisfied by them and thereby experience them as conforming to certain rules. In response we should note that it is one thing to internalize a set of expectations and it is another thing to have perceptually encapsulated detectors.23 You can expect or believe many things. But, no matter what you believe or expect at the belief level, your perceptual system expects an increase in the apparent size of an object as you get closer to it. In the Kantian view, schemas are perceptually encapsulated sub-systems. More important, the phenomenology associated with the activation of schemas is, according to the Kantian view, different from the phenomenology of associations between representations due to past experience.24 The first three premise of the argument imply that there is a perfect correlation between phenomenal objectivity and schematic dynamical unity.

Phenomenal Objectivity and Phenomenal Intentionality

133

According to the last premise, the best explanation for this is that schematic dynamical unity constitutes phenomenal objectivity. The main threat to this premise is that the correlation between phenomenal objectivity and schematic dynamical unity is contingent. I shall defend the last premise of the argument by blocking this objection. There are two ways that one might defend the contingency claim. First, one might argue that for all we have said it could still be the case that the neural mechanisms that underlie schematic dynamical unity and those that underlie phenomenal objectivity are separate and the correlation between them is only due to the way our nervous system is wired. In principle, these mechanism can come apart and the association between phenomenal objectivity and schematic dynamical unity is contingent. I think we should reject this for the reasons that I offered in the previous section of this chapter. As far as I know there are no empirical studies on the neural mechanisms underlying what I have been calling schematic dynamical unities. But, as I argued in the previous section, there are good reasons to assume that the psychological mechanisms underlying schematic dynamical unity and those underlying phenomenal objectivity are the same. This suggests the sameness of the neural mechanisms. A second way to defend the contingency claim is to argue that the concepts of phenomenal objectivity and schematic dynamical unity are a priori independent. Thus, we can coherently conceive of phenomenal objectivity without schematic dynamical unity. But this argument is also problematic. As Chalmers (2003) argues, phenomenal properties enter phenomenal concepts as constitutive components. Accordingly, whether the phenomenal concepts of phenomenal objectivity and schematic dynamical unity are a priori independent depends on whether the phenomenal properties that they pick out coincide in us, and our observations support the claim that they do. Thus the two concepts are not a priori independent when we consider them as phenomenal concepts. This concludes my defense of the premise of the abductive argument for the Kantian thesis. Afterimages lack phenomenal objectivity and the best explanation for this is that schematic dynamical unity constitutes phenomenal objectivity.

Conclusion In this chapter I have proposed and defended a Kantian thesis about phenomenal objectivity. According to this thesis, perceptual experience has phenomenal objectivity in virtue of schematic dynamical unity. I have offered both empirical and phenomenological reasons in support of this thesis. I believe that this thesis and the broader view that it belongs to have significant implications for our understanding of phenomenal intentionality. The significance

134

phenomenal intentionality

of the account, however, is not restricted to its contribution to the phenomenal intentionality research program. Understanding how perceptual experience has objective content is important in its own sake. The account also offers a phenomenological interpretation of some of Kant’s central ideas about representation of objectivity, an interpretation which can be regarded as a competitor to common interpretations such as Strawson’s.25 In Strawson’s interpretation, the representation of objectivity requires highly sophisticated conceptual capacities such as the concept of the self and the conceptual capacity to distinguish between appearance and reality. Th is account has been criticized, in my view justly, as being hyper-intellectualized.26 My proposal provides the initial steps toward giving a non-intellectualist interpretation of Kant’s view.

Acknowledgments This paper has benefited from discussions with Ned Block, David Chalmers, Terry Horgan, Uriah Kriegel, Goeff Lee,Ted Sider, Susanna Siegel, Jonathan Simon, Houston Smit, Joseph Tolliver, the anonymous referee for Oxford University Press, and the audience attending talks on this paper at Cornell, NYU, Oxford, and the University of Cincinnati.

Notes 1. My usage of “self-presenting” diverges from Chisholm’s famous usage according to which a self-presenting mental state is one such that a subject who has it knows or is in the position to know that she has it. 2. The term “present” can be understood in different ways. Under one construal, it is an intentional verb that functions in the same way that ordinary intentional verbs like “believe” function. Sam can believe that Santa is fat, although there is no Santa. Analogously, on my usage sentences of the sort “experience presents x as p” imply neither the existence of x nor the instantiation of p by x. Also, on my usage “experience presents x as p” does not imply the inclination to form the belief that x is p on the part of the subject. My experience can be presenting a tree even if I know that I’m hallucinating and have no inclination to form the corresponding belief. Th roughout the chapter, I am using “presents” in this non-factive/non-epistemic sense. 3. Th is distinction is not meant to be exclusive. It is in principle possible for an experience to be both self-presenting and other-presenting. 4. Further characterization of this intuitive idea depends on one’s broader theoretical commitments about the nature of perceptual experience. For example, a representationalist who thinks that perceptual experiences are constituted by relations to contents might analyze phenomenal character by distinguishing between contents with objective purport and contents without objective purport, where a content with objective purport is one whose truth conditions go beyond the subject’s current mental state. For now, I shall use the notion of phenomenal objectivity in its under-defi ned form and avoid further theoretical commitments.

Phenomenal Objectivity and Phenomenal Intentionality

135

5. I shall say more about this shortly. 6. For explicit defenses of this idea see Horgan and Tienson (2002) and Loar (2002). See also Kriegel (this volume). 7. Another unsatisfactory answer is a na ïve realist one. For example, one might hold that perceptual experience has the phenomenology of objectivity because it is a relation to real objects and properties. Th is answer is unsatisfactory because phenomenological explanations are acceptable within the bounds of phenomena distinguishability, but the na ïve realist answer does not satisfy this condition. 8. Thus the typical externalist strategies for determining the contents of mental states by appealing to tracking relations cannot help us with this question. 9. See Schwartz (1994) for a discussion of Berkeley’s view. 10. All the quotations are based on the Wood and Guyer translation. 11. “Now, this representation of a general procedure of the imagination for providing a concept with its image is what I call the schema for this concept” (CPR A140/B180). 12. “The schema is in itself always only a product of imagination” (CPR B179/A140). 13. Ibid., A140/B180. 14. Th is is based on: “Obviously there must be some third thing, which is homogeneous on the one hand with the category, and on the other hand with the appearance, and which thus makes the application of the former to the latter possible. Th is mediating representation must be pure, that is, void of all empirical content, and yet at the same time, while it must in one respect be intellectual, it must in another be sensible. Such a representation is the transcendental schema” (A138/B177). 15. My account also gets its inspiration from Noë (2004) and Siegel (2006). 16. Burge (2009) also stresses the relationship between representation of objectivity and perceptual constancies. 17. The main alternative accounts are texture occlusion offered by Gibson (1950), relative size offered by Rock and Ebenholtz (1959), and the horizon ratio offered by Sedgwick (1986). 18. The viewing angle can be computed from the size of the retinal projection, b’c’, and the distance of the retina from the lens by the formula: θ = 2arctan(b’c’/2d’). 19. One might think that the reasoning is conscious but it is so fast that we ignore it. I don’t think that this option is plausible either. For a discussion of issues regarding fast reasoning processes and their manifestation in consciousness, see Masrour 2011. 20. Although I think this shows that the person with blindsight is not justified in her beliefs, this assumption is not required for the reasoning given. A belief is epistemically blind if the subject of the belief cannot provide an item that she regards as the basis or evidence for her belief. Th is presupposes, neither that the epistemically blind belief is not justified nor that the evidence that a subject provides in non-blind cases is really the evidence for the belief. 21. Th is intimate connection can be understood in different ways. For example, one might hold that schematic dynamical unity is partly constituted by implicit consciousness of these anticipatory conditionals. The other option is to hold that there is a causal relation between the mechanisms underlying schematic dynamical unity and the mechanisms that would ground these anticipatory conditionals. For our purposes at this stage, it does not matter which option we choose. 22. See Smith (2002), chapters 2 and 5. 23. By a perceptually encapsulated mechanism, process or representation I mean one whose causal interaction with other representations is mediated by the input or the output of the perceptual system. Perceptually encapsulated representations are immune from the influence of beliefs in the short term although they can be affected by beliefs in cases where perceptual learning happens. 24. Th is, in my view, is important for understanding Kant’s answer to Hume’s problem. 25. See Strawson (1966). 26. Burge (2009) offers the latest version of this criticism.

136

phenomenal intentionality

References Burge, T. 2009. “Perceptual Objectivity.” Philosophical Review 118(3): 285–324. Chalmers, D. 2003. “The Content and Epistemology of Phenomenal Belief.” In Consciousness: New Philosophical Perspectives, ed. Q. Smith and A. Jokic. Oxford: Oxford University Press, 2003. Horgan, T., and J. Tienson. 2002. “The Intentionality of Phenomenology and the Phenomenology of Intentionality.” In Philosophy of Mind: Classical and Contemporary Readings, ed. D. J. Chalmers. Oxford: Oxford University Press. Loar, B. 2002. “Phenomenal Intentionality as the Basis of Mental Content.” In Philosophy of Mind: Classical and Contemporary Readings, ed. D. J. Chalmers. Oxford: Oxford University Press. Masrour, F. 2011. “Is Perceptual Phenomenology Th in?” Philosophy and Phenomenological Research 83(2): 366–397 Noë, A . 2004. Action in Perception. Cambridge, MA : MIT Press. Schwartz, R . 1994. Vision: Variations on Some Berkeleian Themes. Cambridge, MA : Blackwell. Siegel, S. 2006. “Subject and Object in the Contents of Visual Experience.” Philosophical Review 115(3): 355–388. Smith, A. D. 2002. The Problem of Perception. Cambridge, MA : Harvard University Press. Strawson, P. F. 1966. The Bounds of Sense: An Essay on Kant’s Critique of Pure Reason. London: Methuen.

8

Phenomenal Intentionality and the Role of Intentional Objects F r e de r ic k K roon

1. Introduction The last decade or so has seen the rise of a new internalism about the mind. This new internalism offers a Cartesian perspective on experiential and cognitive states—at least conscious ones—in contrast to the externalism that has been orthodoxy for the last quarter of the 20th century. But instead of repackaging old empiricist notions of what is distinctive about conscious experiential and even cognitive states, this new internalism appeals to the phenomenology of such states and to that extent is far more reminiscent of Husserl than of the empiricists. Indeed, the new internalism advocates a notion of intentionality that is based on phenomenology alone, and this, at first blush, sounds positively Husserlian. But there is a quite fundamental difference. We get a hint of this difference when Loar, an early and prominent supporter of this phenomenology-based Cartesianism, offers this characterisation of internal intentionality: The internal intentionality of perceptions and thoughts consists in their apparent directedness, in their purporting subjectively to refer in various complex ways. (Loar 2003, p. 231; all emphases in this and other cited quotes are original) Husserl, one assumes, would have queried the qualifier “apparent,” and insisted that apparent directedness is always genuine directedness, although not always genuine directedness to existent objects. But Loar’s choice of words is deliberate, for he wants to allow for a conception of internal intentionality—phenomenal intentionality—that involves an intrinsic, nonrelational kind of directedness in a person’s experiences and thoughts, a kind of directedness that is independent 137

138

phenomenal intentionality

of the actual referents of these experiences and thoughts. Advocates of this kind of intentionality—call them “phenomenal intentionalists”—often present their case by appealing to the example of phenomenal states that can’t be distinguished in terms of phenomenal properties. Horgan and Tienson, for example, appeal to the states of “phenomenal duplicates,” creatures whose phenomenal life is just like ours, but which may exist in very different environments: they may exist in a Twin-Earth environment, for example, or they may exist as Brains in a Vat (Horgan and Tienson 2002; Horgan et al. 2004). Such creatures lead identical, or at least indistinguishable, mental lives, and corresponding mental states instantiate the same or exactly matching phenomenal properties of an intentional kind.1 Furthermore, these properties yield an appropriate notion of phenomenal-intentional content, a content that is narrow to the extent that it is fully determined by what goes on inside the head. By now it should be clear how unHusserlian this new form of phenomenology-invoking internalism is. First of all, because different actual referents are involved, Husserl (at least the young Husserl, before he developed the method of epoché) would have thought that the experiences of Earthlings and their Twin-Earth counterparts differ in their intentional objects.2 More fundamentally, for Husserl phenomenology teaches us that where there is phenomenal directedness it is always directedness to intentional objects, some existent, some nonexistent—phenomenal directedness is always object-involving. Call this view about intentional directedness intentional realism.3 By contrast, phenomenal intentionalists deny that such intentionality is essentially relational and object-involving. In particular, they deny that there can be relational directedness to merely intentional, nonexistent objects, since they deny that there are such objects. While most allow an object-involving externalist form of intentionality (the kind of intentionality possessed by experiences and thoughts that are genuinely of something ; that is, of something existent),4 they think that the most basic form of intentionality is an internalist form characterised by a nonrelational kind of directedness. So phenomenal intentionalists advocate an outlook on (basic) intentionality and content that, like Husserl’s phenomenology, places a strong emphasis on the internal, richly intentional life of the mind but understands this intentionality in a fundamentally different way from Husserl.5 Focusing on the case of sensory experience, the present chapter attempts to adjudicate this disagreement between the phenomenal intentionalist and intentional realist ways of understanding intentionality. It begins this task by giving a new reason for thinking that phenomenal intentionalists have underestimated the need for intentional objects in an account of the intentional content of sensory experience. To this end, it rehearses a thought experiment given by Loar in support of phenomenal intentionalism (section 2), rejecting an argument by Bill Lycan for an intentional realist interpretation of this thought experiment

The Role of Intentional Objects

139

before showing how Lycan’s argument can be strengthened to make the role of intentional objects seem almost inevitable (section 3). Section 4 argues that we should nonetheless resist this intentional realist conclusion. In its place, the chapter offers a compromise fictionalist approach that fits the spirit of the phenomenal intentionalist framework while retaining the benefits of the intentional realist’s way of invoking intentional objects.

2. Loar and Lycan on Lemons Why appeal to intentional objects, including mere intentional objects, in an account of the intentional content of sensory experiences? One reason comes from a surprising source. Bill Lycan is well known as a hard-headed representationalist about sensory experience. For Lycan, properties like the colour occupying such-and-such a region of your visual field now, the pitch or the volume of a heard sound, or the smell of an odor are properties represented in a subject’s sensory experience as belonging to objects, and these objects are in turn represented in the experience as having those properties. Not surprisingly, phenomenal intentionalists have objected to such a representationalist view of content. Their main argument involves the phenomenon of nonveridical experiences, since these are experiences with an apparently definite content—they seem to represent objects as having properties in a way that is indistinguishable from the way veridical experiences represent objects as having properties—even though they do not represent anything. Suppose, to take a well-known example of Loar’s, that you are visually presented with what you take to be a series of indistinguishable lemons, and that you have a singular demonstrative thought about each one (Loar 2003). Now suppose that it is revealed to you that some of the “lemons” were hallucinatory, produced by electrodes in the brain. Such a scenario is bread-and-butter stuff for the phenomenal intentionalist, of course. Loar comments that, regardless, “[your] successive visual demonstrative thoughts all visually presented their objects in the same way,” and adds: This presents itself as sameness in an intentional feature. . . . This is a nonrelational phenomenal feature, by which I mean something rather strong: we are aware of internally determined phenomenal features of visual experience, of their manifold felt aspects, and among those features—though not separable in imagination—is the directedness . . . mentioned [above]. (p. 239) For Loar, then, such states, do have content—in fact, they have identical or at least matching intentional content—despite the fact that some of them

140

phenomenal intentionality

represent no objects (at best they represent properties) and so don’t have their intentionality grounded in (intentional) objects. Lycan demurs at just this point. (He grants the first point, that your visual demonstrative thoughts all visually presented their objects in the same way.) According to Lycan, the representationalist reply is all too obvious. In each case, a particular lemon is represented. Some of the lemons are real, and some are unreal, but that does not change the fact that in each case you visually demonstrate an identical lemon. The experiences are introspectively indistinguishable because the lemons themselves are indistinguishable. That the experiences all represent a lemon just like that is indeed the sharing of an intentional feature, but the feature is intentional because it is unabashedly representational. (Lycan 2008, p. 243) So for Lycan, the yellowness represented in each illusory-lemon experience is the yellowness of a nonexistent lemon that is also represented in experience. Similarly, the properties represented in the sensory experiences of Brain-in-a-Vat subjects are properties of nonexistent physical objects. Lycan uses Harman’s (1990) transparency argument to support this view. Harman claims that when you turn your attention to your perceptual experiences, all you can discern phenomenologically are properties of the (apparent) object of the experiences: the properties we are aware of are attributed to the (apparent) object perceived, not to something more phenomenal: there is no phenomenal paint of which we are introspectively aware. In short, the intentionality of sensory experience is object-involving (and to that extent wide rather than narrow) because it is representational. While that usually means that we represent existent physical objects as having properties, in some cases it means that we represent nonexistent physical objects as having properties. Lycan does not deny that sensory or even nonsensory cognitive mental states have a higher-order kind of phenomenal character—a certain what-it-is-like-ness—that is not representational. But he thinks that phenomenal character so understood is akin to Kaplan’s linguistic notion of character; both, he thinks, are quite irrelevant to a notion of genuine content.6

3. Phenomenal Intentionality and “Consciousness of Identity” Confronted by this argument from representationalism, phenomenal intentionalists are likely to lodge the most vigorous part of their protest at Lycan’s view—a view he shares with Husserl—that intentionality is genuinely relational. Early in the debate, Loar had this to say about the lemon case: The veridical perception [i.e., where the subject really sees a lemon] has an intentional object in a transparently relational sense: it refers. It

The Role of Intentional Objects

141

could only be a fanciful Meinongianism that construes having a mere intentional object relationally. (Loar 2003, p. 241) So for Loar the problems facing a “fanciful Meinongianism” constitute the crucial reason to reject a relational conception of intentionality that can be used to explain what veridical and hallucinatory perceptions have in common.7 With the nonrelational feature of directedness at the forefront of the new understanding of intentionality, intentionality becomes “a matter of how one’s perceptions and thoughts represent things if they succeed, rather than of what is thereby represented” (p. 240), making this basic form of intentionality fully phenomenal rather than representational. We have already seen that Lycan resists this reorientation of the notion of mental content, claiming Loar and other phenomenal intentionalists confuse content with phenomenal character. But how can Lycan respond to the charge that underlying his thoroughly relational conception of intentionality there is only a “fanciful Meinongianism”? On this point he is inclined to take the dialectical high road. (“There are things that do not exist. . . . Get used to that plain fact” [Lycan 2008, p. 249, fn. 23].) How to understand this fact he takes to be less clear. As he points out, Meinongianism is not the only account of nonexistent objects in town: whatever the best account is, that is the one the representationalist should use. (Although he doesn’t rule out other views, Lycan himself is an “Ersatzer,” construing nonexistent objects as abstract entities.)8 But this response leaves Lycan-style representationalism in an uncertain position, for it is scarcely a Moorean fact that our talking as if there are things of a certain kind shows that we must be committed to there actually being things of that kind. Sometimes such talk is merely figurative, as when we say of someone that she is in the grip of the green monster or affected by lingering doubts (cf. Yablo 1998). Maybe talk of unreal lemons is like this too. Phenomenal intentionalists are adamant that there are no unreal, nonexistent objects, certainly not as objects of nonveridical experiences, and this view will strike many as eminently sensible. They will reject any tendency to take talk about unreal lemons as literally true. Rhetoric aside, how can Lycan respond? To begin with, by challenging, on directly phenomenological grounds, the idea of (mere) directedness as a purely nonrelational feature of experience. Mere directedness, in fact, seems too thin a base on which to rest a phenomenologically informed account of phenomenal content. Such directedness is supposedly the feature of somehow reaching out ( purporting to refer, as opposed to actually referring), but phenomenologically our experience and associated demonstrative concepts strike us not merely as directed, but directed in what seem to be irreducibly object-involving ways. We find a clue to the problem facing Loar when we return to his lemon-experiment. Loar did phenomenology a disservice when he invoked

142

phenomenal intentionality

an experiment involving a series of distinct lemon-experiences to introduce his new nonrelational form of directedness. Experience does not come as an experiment of this type. When we attend to the world, our experience shows both more differentiation and more unity. We discern, as a matter of phenomenology, distinctness: we discern the presence of distinct physical objects (not surprisingly, given spatial differentiation). We also discern, as a matter of phenomenology, identity: we see a person, then see the same person again, recognising her as the same. Such recognition can be very vivid, as can its converse (if we look at a certain spot expecting to see someone we know, we get the shock of the unexpected if it is a different person). There is no reason to count this sense of identity and distinctness as anything other than part of the very intentional structure of experience. (Those who prefer to count it as adding a cognitive element to the directedness of experience can count it as part of cognitive phenomenology.) The idea should be comfortingly familiar to those who know the phenomenological tradition. It is found in Husserl, but also in the early Heidegger (the Heidegger most indebted to Husserl). And Husserl is explicit that this “consciousness of identity” pertains as much to our experience of merely apparent objects as existing ones: I see a thing, e.g., this box. . . . I always see one and the same box, however it may be turned and tilted. . . . [T]he object’s real being or non-being is irrelevant to the true essence of the perceptual experience, and to its essence as a perceiving of an object as thus and thus appearing, and as thus and thus thought of. In the flux of experienced content, we imagine ourselves to be in perceptual touch with one and the same object: this itself belongs to the sphere of what we experience. For we experience a “consciousness of identity,” i.e., a claim to apprehend identity. On what does this consciousness depend? Must we not reply that different sensational contents are given, but that we apperceive or “take” them “in the same sense”, and that to take them in this sense is an experienced character through which the “being of the object for me” is first constituted ? (Husserl 2001, Investigation V, §14) According to Husserl, such a “consciousness of identity” is found not only in the course of a series of consecutive perceptual experiences such as the tilting-of-a-box series, but cuts across types of experience as well as temporal gaps between experiences. We might first catch a glimpse of something, then much later fleetingly recollect the thing, then later again contemplate the thing in imagination, and finally encounter the thing directly presented in perception. Husserl says that in the latter situation a series of partial intentions have finally reached intuitive fulfi lment.9 Heidegger agrees: “the intended identifies

The Role of Intentional Objects

143

itself in the intuited; selfsameness is experienced [erfahren]” (Heidegger 1992, p. 49).10 Here my focus will be on cases of “consciousness of identity” that involve a self-conscious awareness of identity—specifically, cases where we self-consciously recognise something as the same again. (I’ll use the abbreviation CI for this type of consciousness of identity.) Of course, even when thus restricted, consciousness of identity remains an utterly pervasive phenomenon. It is surprising, then, that phenomenal intentionalists don’t discuss it, for it does present them with a prima facie difficulty, at least in the case of hallucinations and, more radically, Brain-in-a-Vat type situations. Phenomenal intentionalists will of course resist its description in terms of objects—in terms of the idea that the subject recognises a certain intentional (albeit nonexistent) object. But how else should we describe CI? Note that we don’t avoid a commitment to such objects simply by recasting this idea as “experience presents a merely apparent object again,” for that sounds like a dodge: what is a merely apparent object, if not an intentional object that doesn’t exist? Now, some phenomenal intentionalists do write as if they approve of the idea of (merely) apparent objects. According to Horgan, Tienson, and Graham, for example, phenomenologically, perceptual experience is richly presentational: it presents, to the experiencing subject, a richly textured apparent world of apparent concrete objects apparently instantiating numerous properties and relations. (Horgan et al., 2004, p. 303) Context makes it clear, however, that the sole purpose of “apparent(ly)” is to mark out a feature of the “numerous properties and relations” that are genuinely given in experience: [In the BIV (Brain-in-the-Vat) case] your BIV’s duplicate’s . . . acquaintance [with shape-properties and relative-position properties] occurs via radically nonveridical experiences of merely apparent instantiations of these properties and relations by merely apparent objects. (Ibid, p. 305) For Horgan et al., then, the phenomenology of experiences should be understood as narrow: there is mental reference to properties and relations11 but no other reference (say, to a category of apparent objects) that is guaranteed of success on the basis of phenomenology alone. But it is hard to see how a narrow recasting of an appeal to apparent objects—a recasting that only involves notions that are wholly constituted phenomenologically and, in particular, no intentional objects—can deal with

144

phenomenal intentionality

CI. To fi x our understanding of this kind of phenomenal narrow content, suppose that we stay with the account that Horgan et al. have on offer. Applied to thought-constituents that purport to refer to objects or kinds, this account invokes “phenomenally constituted grounding presuppositions”: qualitatively specifiable presuppositions “concerning the existence of, the persistence of, and various features of, the sorts of entities presented in experience” (Horgan and Tienson 2002, p. 528).12 Consider the following toy example, involving the singular thought-constituent that picture, deployed in an experiencebased thought that we might express linguistically by “That picture is hanging crooked.” According to the relevant grounding presuppositions that must be satisfied for the singular thought-constituent to refer, roughly, there must be an object at a certain location relative to oneself (a location that one could designate linguistically by a specific use of the place-indexical “there”), this object must be a picture, and there must not be any other picture at that location that is an equally eligible potential referent of “that picture.” (Horgan et al. 2004, p. 307) Now consider an experience, or a thought based on an experience, that is characterised by CI (for example, one characterised linguistically as “There is that picture again, still hanging crooked”). If we try to formulate the appropriate grounding presuppositions, we strike a difficulty. The grounding presuppositions for the singular thought-constituent cannot simply have the structure: (R0) There is an object x in front of me that possesses property P, etc., and x existed earlier, for this is a trivial addition—I don’t need to be told that I take myself to be experiencing a persisting thing; what is crucial is that I take it to be a persisting thing that I recognise. Note also that phenomenal intentionalists will deny that any such recognitional element can be captured in object-invoking terms: (R1) There is an object x in front of me that possesses property P, etc., and x = a [where “a” designates some object encountered in an earlier experience]. For this structure would make the content of the subject’s state wide rather than narrow, leaving such presuppositions unable to characterise thoughts and experiences in which, say, there is only the appearance of something’s having been encountered earlier.

The Role of Intentional Objects

145

What is needed, it seems, is a formulation according to which there are objects that present themselves as objects that were encountered in previous experience(s). This suggests something like the following account: (R 2) There is an object x in front of me that now instantiates properties P, etc., and at some time in the past I experienced x as instantiating relevantly similar properties [where the standard of relevant similarity picks out properties that are relevant to recognizing the object as the same object]. Is (R 2), or something like it, what is needed to capture CI in the phenomenal-intentionalist terms preferred by Horgan et al.? (R 2) faces two problems that may suggest more is needed. First of all, it uses the notion of experiencing something, and phenomenal intentionalists may deny that this notion is genuinely “phenomenally constituted” in their sense since it appeals not only to the concept of experience itself (not just to concepts of properties encountered in experience) but to a relationally directed notion of experience. In my view, however, phenomenal intentionalists should just bite the bullet here: such a reflexive appeal to the notion of experience—furthermore, a notion of relationally directed experience—seems all but inescapable. (I return to this point later.) A rather more pressing worry is that while (R 2) may provide a promising account of CI if we focus on a single experience, it does poorly on CI taken across experiences. Remember that the quantifiers in (R 2) are local quantifiers: they don’t take in the content of other experiences. Consider the current experience E of my Brain-in-a-Vat surrogate, an experience in which he “sees O” (there really is no such thing, of course), which he recognises as something he has seen earlier. Now let EP be a previous experience in which he encountered this apparent object, an encounter that played a causal role in his recognition of the object in E. Clearly, there is no common element in the characterizations of the narrow contents of E and EP to capture the apparent cross-experiential role of this apparent object O. Putting the point in terms of a formal-linguistic formulation of grounding presuppositions: Any variable (“y,” say) used to identify O in the relevant grounding presuppositions appropriate to his earlier experience EP (“∃y Φ(y)”) falls outside the scope of the quantifiers used to identify O in the relevant grounding presuppositions appropriate to his later experience E (“∃xΨ(x)”). Now this may seem as it should be, for in the Brain-in-a-Vat case there is no common object O. But from the point of view of the experiencing subject this gets the phenomenology badly wrong. Recognition is often the occasion for the experiencing subject to reconstruct in memory (a sequence of) earlier encounters with the recognised object, allowing the recognised object to play a fuller

146

phenomenal intentionality

cognitive role in the subject’s life. A suitably enriched version of experience E might thus take in a vivid reconstruction in memory of what the experiencing subject saw in the course of previous experiences like EP (he may report: “I am now remembering how x was earlier; it looked so-and-so and did thus-and-so”). On the account of Horgan et al., however, none of this matters. Even in the case of someone with total recall, the impression that the memories share anything but bare qualitative similarity with the experiences they are memories of is just an illusion; absent external objects, there is no sense in which the subject can remember things experienced earlier. To that extent, Horgan et al. get the cross-experiential phenomenology badly wrong. Can Loar do better? Unlike Horgan et al., Loar does not think that a phenomenal state has a narrow content that fi xes what the world must be like for a singular thought-constituent to refer. He merely invokes the nonrelational directedness of the state and its constituents. But that leaves his account poorly placed to deal with CI. What is needed is a notion of co -directedness that is phenomenologically no less respectable than simple directedness. As applied to a single extended experiential state it is perhaps easy enough to see how to understand such a notion of co-directedness in an appropriately nonrelational way: just take a sequence of momentary visual demonstratives to be what carries the directedness. But this doesn’t help at all in making nonrelational sense of co-directedness when the relevant sequence of visual demonstratives contains some that are no longer accessible to the subject (perhaps not even in memory) because they occurred as constituents of past experiential states. In that case, the co-directedness is scarcely an “internally determined phenomenal feature” of the subject’s current visual experience. It is difficult to see what such co-directedness could be other than directedness at the same object on the part of constituents of different experiences, an idea not available to Loar since he explicitly rejects a relational notion of directedness.13 To sum up, there is an aspect of phenomenology (CI) that familiar accounts of phenomenal intentionality have trouble with. Adapting Horgan et al.’s earlier quoted comments, we might say that, phenomenologically, a subject’s perceptual experience taken as a whole (extending over her whole life-time) is richly presentational: it presents the experiencing subject with a richly textured apparent world of apparent concrete objects apparently instantiating numerous properties and relations, with the subject often recognizing the same apparent objects—recurring apparent objects, for short—in different experiences. Even in the Brain-in-a-Vat case, for example, the subject’s apparent world is such that ∃x(x is an object first encountered in experience E1 and then recognised in a later experience E2). We are thus left with a formulation committed to a wide rather than narrow characterisation of the phenomenologically available content of a subject’s experiential states.

The Role of Intentional Objects

147

4 As-if Directedness and the Role of Reflexive Narrow Content Lycan doesn’t mention CI, but it is evident, I think, how he would (should ) deal with it, given the intentional realism to which his version of representationalism is already committed. He should follow Husserl’s lead. In cases where there are no external objects, an intentional realist like Husserl is able to invoke a level of intentional object to ground consciousness of identity: one can be conscious of the identity of an object through different experiences, even when the experiences are nonveridical. Combining such a line with his representationalism would allow Lycan to capture CI, and would thus give him what he needs to respond to phenomenal intentionalists: a beefed-up phenomenological argument for merely apparent, nonexistent objects, one that is independent of any prior commitment to representationalism. Does this finally vindicate Lycan’s response to phenomenal intentionalists like Loar? Recall Loar’s dismissive remarks about the ontological cost of Lycan’s intentional realism. To vindicate Lycan’s response, the argument presented here would need to harbor a convincing rejoinder to such remarks. And on this score I remain unconvinced. For the question surely remains: what sense can we make of the appeal to nonexistent objects as genuine objects of experiences and thoughts? Our reservations about any such view might remain undiminished even in the face of the phenomenological argument made earlier. (We might wonder how in particular an Ersatzer like Lycan can make good sense of such objects. They are supposed to play a role in nonveridical experience as merely apparent objects, but how can abstract objects have an appearing role?) We seem to be faced with a conflict between two seemingly compelling views: (i) Aim to capture the apparent object-directedness of even nonveridical experiential states in terms of a relationship to objects; anything else will get the phenomenology wrong; and (ii) Repudiate any commitment to nonexistent, merely apparent objects in the attempt to account for intentionality; there are no such objects, and in any event their nature would preclude them from playing a role in an account of intentional directedness and content. Is there a way to resolve this confl ict? My sympathies lie with (ii), and so I need a way of countering the phenomenological attractiveness of (i). One thought is the following. When I said at the end of the previous section that perceptual experience presents a richly textured apparent world of recurring apparent objects, the focus remained on the subject’s experience: the characterisation didn’t locate the apparentness of apparent objects outside of the subject. So a tempting first move is to say that for the experiencing subject it is as

148

phenomenal intentionality

if in her experience taken as a whole she encounters a richly textured world of concrete recurring objects, without, perhaps, there being such objects. But while such a proposal may seem tempting, Loar shows why we should be suspicious. As it stands, the proposal is “phenomenologically blank” (p. 242); it does not say what it is about the course of my and my Brain-in-a-Vat duplicate’s experiences that makes it seem as if both of us encounter concrete recurring objects in these experiences, and, furthermore, concrete recurring objects that are experientially indistinguishable from the point of view of our two perspectives.14 I suggest the following response. On what we might call narrow as-if-ism, experiences have a type of internal, narrow content that makes it natural to re-describe them as representational states that represent objects and their instantiation of properties and relations (including the CI property of being recognised as having been encountered in previous experiences), even in cases where the experiencing subjects are hallucinating and there are no objects. No nonexistent, merely apparent concrete objects are represented in experience on this view, but doing as if there are concrete objects being represented is close to irresistible, given the kind of internal content our thoughts and experiences have. But how, then, should we think of narrow content? We have already seen considerable disagreement among phenomenal intentionalists on this point. Horgan, Tienson and Graham take the relevant contents to involve existentially quantified grounding presuppositions. Loar, by contrast, takes the feature of narrow content contributed by singular thought-constituents to be a kind of primitive that-ness: The content of those thoughts [the ones based on seeing a succession of lemons in the lemon experiment] is not itself existentially quantified: “I am seeing some lemon or other and it is yellow.” The thoughts in question are demonstrative and they are not self-consciously reflexive. (Loar 2003, p. 239) I claimed earlier that neither party is hereby able to capture CI. Can we improve on their accounts so that we can capture this feature, and in particular improve on them in a way that also allows us to capture its being as if the states and thoughts are about experientially recurring concrete objects? Only, I think, by rejecting theoretical positions that they seem to insist on. In my view, the internal contents of thoughts and underlying experiences that seem to make reference to concrete objects via singular thought-constituents are best understood as involving something like Horgan, Tienson and Graham’s existentially quantified grounding presuppositions—but with a crucial difference. I claim such contents are reflexive in a way explicitly denied by Loar and less explicitly by Horgan, Tienson and Graham.

The Role of Intentional Objects

149

Consider again the case of my looking at what seems to be a crooked picture. As we saw earlier, Horgan et al. think that corresponding to the singular thought-constituent expressible linguistically by “that picture” there are phenomenally constituted grounding presuppositions according to which there is a uniquely salient picture at a certain location relative to myself. In my view, that doesn’t go far enough. What is missing is a (further) reflexive feature: I presuppose that there must be an object of this sort that I am visually acquainted with. Now, Loar explicitly denies that there is any such feature to the content of my thoughts (“the thoughts in question . . . are not self-consciously reflexive,” p. 239), and the fact that Horgan et al. don’t mention such a feature suggests they think the same. I am not convinced. What argues for including such a reflexive feature in a suitably extended notion of what is phenomenologically available is that a subject implicitly recognises such a feature as a necessary part of what is required for a singular thought-constituent to refer to its intended external referent. Put differently, but in a way that comports with the willingness of many phenomenal intentionalists to allow that “felt” dispositions feature in phenomenology, we can say that among the potentialities that a subject “feels” in the phenomenology of such a thought is that she would not count something as the external referent of her thought-constituent if her experience didn’t acquaint her with this object.15 She knows a priori that were she caused to have a hallucinatory experience as of a picture placed directly in front of her, then even if there is in fact a picture in front of her, although one hidden from view, she would not have a visual experience of that picture; there would be no externalistic intentionality. And she knows this as something that is available to her on the basis of a suitably generous account of the phenomenology of her experience. I propose, then, that we let the narrow internal content of such apparently object-directed thoughts and experiences involve grounding presuppositions that make reflexive reference to relationships of acquaintance afforded by the experience itself (reflexive narrow content, for short).16 My focus in the remainder of the chapter will be on showing that the idea of reflexive narrow content yields the benefits promised earlier: (a) it provides an internal, narrow perspective on CI, able to capture not only single-experience CI but also cross-experiential CI; and (b) it allows us to make sense (in narrow as-if terms) of the way we so readily appeal to objects, even merely apparent ones, when describing the content of these states.17 The two tasks are related. I will make a start on (a). If the feature of narrow content contributed by singular thought-constituents reflexively refers to relationships of acquaintance afforded by the underlying experiences, then surely nothing prevents narrow content from referring in even richer ways to relationships of acquaintance. In particular, narrow content may make reflexive

150

phenomenal intentionality

reference to relationships of acquaintance afforded by earlier experiences, including earlier experiences where the relationships of acquaintance yielded information able to facilitate subsequent recognition of that same object— there may be phenomenally constituted grounding presuppositions according to which there is a unique object the experiencing subject takes to instantiate such-and-such properties on the basis of her being presently acquainted with it, and which she recognises on that basis as one she was previously acquainted with. (The claim is not that the subject explicitly thinks this, but that it articulates something that is clearly present and implicit in the distinctive phenomenology of her experience. Phenomenal intentionalists should agree that much else that is included in narrow content is merely implicit, and so they have no reason to be especially concerned about this aspect of the proposal.)18 If I am right, such grounding presuppositions capture CI (at least, single-experience CI) in terms it is entirely reasonable to describe as “narrow.” But what about cross-experiential CI, the phenomenon that takes our experience as a whole as involving encounters with recurring apparent objects? For that we need to make a start on task (b), that of accounting for the attractions of a representationalist view of experiential states that invokes concrete objects, real or merely apparent, as what is represented in such states. Now, the acquaintance-base of reflexive narrow content should help with this task. Consider a subject in an experiential state entertaining a thought whose singular thought-constituent is governed by appropriate acquaintance-based presuppositions. If this thought-constituent has an external referent, the conditions are right for the thought to be about that referent: there is then the right sort of acquaintance-based causal connection between experiencing subject and external referent for the experiential state and thought to represent that external referent as having certain properties. But what if there is no external referent? From the subject’s point of view, there is no difference. Given the phenomenology of her experiential state, even externalists like Lycan can agree that to her it will be as if she is acquainted with the external referent of her thought-constituent, and it will be as if her state represents this object as having properties.19 And because of the acquaintance-based terms in which a subject implicitly thinks of the objects encountered in previous experiences, to her it will be as if her current experiences represent objects that were also represented in earlier experiences. In sum, to the subject it is as if the course of her experience as a whole consists of experiences that involve encounters with recurring concrete objects. If she doesn’t know any better—doesn’t know, say, that she is hallucinating or is a Brain in a Vat—she will believe that the world is really like this. To commentators who are in the know about the absence of an external referent for a thought constituent—they might be theorists speculating about our Brain-in-a-Vat counterparts, say—the world is not like this at all. But such

The Role of Intentional Objects

151

commentators will have their own reasons for doing as if this is how the world is; they have good reason to adopt a narrow as-if stance. From their point of view, it is appropriate and natural to invoke objects when characterising the content of experiences and thoughts. In order to understand and explain the subject’s thoughts and actions from the inside, as it were, they will want to simulate being in a situation in which the subject’s narrow perspective is right rather than wrong. That perspective says that there is an object that she is visually acquainted with and which she judges on that basis to have such-and-such properties. But to simulate being in a situation in which the subject is acquainted with an external object in this way means simulating being in a situation in which there is the right sort of connection between the subject and this object for her thought to be a thought about the object, and for the experiential state underlying the thought to represent the object as having certain properties. On the basis of their simulation, these commentators will be tempted to describe the experience as representing a particular external object as having certain properties, even though they will also acknowledge—now speaking from outside the scope of their simulation—that the subject is wrong about this, that the represented object is, so to speak, hallucinatory.20 In taking such nonveridical experiences to represent external objects, these commentators will also re-describe experiences characterised by CI in objectinvolving terms: because of the acquaintance-based terms in which a subject implicitly thinks of the objects encountered in previous experiences, the commentators will re-describe the subject’s current experiences as representing objects that she recognises as objects represented in earlier experiences (even though they will again acknowledge that she is strictly wrong about this). In short, there is a natural interpretive standpoint—given by narrow as-if-ism— on which the subject’s experience as whole consists of experiences that involve encounters with recurring concrete objects. That is the sense in which we can objectively account for cross-experiential CI even in Brain-in-a-Vat type cases where there literally are no objects for subjects to re-experience.

5. Conclusion This chapter has focused on the conflict between a phenomenal conception of the intentionality of sensory experience and an older intentional realist conception that views intentional directedness as relational and invokes nonexistent intentional objects of experience to deal with nonveridical experience. I began by recasting Lycan’s representionalist challenge to phenomenal intentionalists like Loar in terms of a new phenomenological argument for nonexistent objects of experience, one that appeals to the phenomenon of consciousness of identity. Because of the way in which this argument focuses on the apparent

152

phenomenal intentionality

persistence, in consciousness, of intentional objects of sensory experience, I consider this about as good an argument for intentional realism about such objects as it is possible to get. I have now shown why I think that this new challenge fails. By letting narrow as-if-ism explain why it is both tempting and natural to re-describe the reflexive narrow content of a subject’s experiential states in terms of a wide representational content according to which these states represent recurring concrete objects, we not only explain the pull of this argument. 21 We also explain how phenomenal intentionalists can (and should) resist its conclusion.

Acknowledgments Earlier versions of this chapter were presented at a workshop titled “The Intentionality of Phenomenology and the Phenomenology of Intentionality” (University of Modena and Reggio Emilia, April 2008), and at the 2008 AAP (New Zealand) conference (Massey University, December 2008). I am grateful to participants, including Carola Barbero, John Bigelow, Bill Fish, Galen Strawson, and Alberto Voltolini, for their useful comments. Special thanks to Uriah Kriegel, Bill Robinson, and Matheson Russell.

Notes 1. Loar (along with most other phenomenal intentionalists) thinks identical properties; Horgan and Tienson think exactly matching ones. 2. As Husserl writes in an appendix to Investigation V (Husserl 2001), “the intentional object of a presentation is the same as its actual object, and on occasion as its external object, and . . . it is absurd to distinguish between them. The transcendent object would not be the object of this presentation, if it was not its intentional object” (Appendix to §11 and §20). 3. On this use of the phrase, intentional realism is a kind of entity realism about the realm of the intentional, not merely a view about the reality of the phenomenon of intentionality. My comments about Husserl presuppose an intentional realist interpretation, and such an interpretation is to a degree controversial. (Consider, in particular, the so-called West Coast quasi-Fregean interpretation of Husserl provided by commentators like Føllesdal and Dreyfus.) In my view, an intentional realist interpretation becomes almost inescapable when one considers what Husserl has to say in such places as Investigation V (“On intentional experiences and their contents”). Here we fi nd Husserl insisting (against the “image theory” that there is always a representative image in representation but sometimes not an object, and against the theory that intentional objects are somehow immanent) that “if I represent God to myself, or an angel, or an intelligible thing-in-itself, or a physical thing, or a round square etc., I mean the transcendent object named in each case, in other words my intentional object: it makes no difference whether this object exists or is imaginary or is absurd” (Husserl 2001, Investigation V, Appendix to §11 and §20). Cf. also Husserl (1973) on the parallels and differences between the world of actuality and of the imagination (§§39–41, §§74–75). 4. Some deny even externalist intentionality; see, for example, Farkas (2008).

The Role of Intentional Objects

153

5. Supporters of what Horgan and Kriegel (2008) call the “phenomenal intentionality outlook” also tend to ascribe a phenomenological dimension to more cognitive experiential states (see, for example, Strawson 1994 and Siewert 2006). For a critique of this thesis of cognitive phenomenology, see Lycan (2008), 234ff. The present chapter is focused on what Lycan sees as the “more specific and ambitious” (234) claims of recent intentionality-focused works such as Loar (2003) and Horgan et al. (2004). 6. See Lycan (2008), pp. 243–4, 247. Cf. also Lycan (1996), pp. 134–5, which allows that part of phenomenal character might be constituted by functional rather than representational properties. Lycan’s view of representational content is what Chalmers calls an object-involving Russellian representationalism (Chalmers 2004, p. 167f). 7. Loar also insists (2003, p. 241) that if we were to go Meinongian “having an intentional object” would stand for two very different relations in the veridical and hallucinatory perceptions cases. He is surely right. (For a start, there is no causal connection between a subject and the nonexistent Meinongian lemon he “sees.”) 8. The best-known such view is that of Zalta (1988). By contrast, Harman (1990) approvingly cites Parsons’ neo-Meinongianism (Parsons 1980). 9. Husserl’s notion of “intuitive fulfi lment” is crucial to his analysis of intuition, evidence and truth. The basic idea is that relatively indirect ways of thinking about an object (by representing it in imagination or recollection, say) may be followed by direct fi rst-hand experience of the object that in a sense captures what was really intended all along in the indirect ways of thinking. Husserl thinks of it as a marriage of meaning and the meant. (It is analogous to Russell’s idea of progressing from knowing something by description to knowing it by acquaintance.) 10. Husserl notes the further phenomenological distinction between experiencing an object as the same through a sequence of perceptual and other acts and coming to the sudden belief that this object is identical to that object (as in the identification of Hesperus and Phosphorus, say). About the former kind of experience he says: “Identity . . . is not fi rst dragged in through comparative, cogitatively mediated reflection: it is there from the start as experience, as unexpressed, unconceptualised experience” (Husserl 2001, Investigation VI, section 1, Ch. 1, §8), adding that “we are entitled to say that an identifying coincidence has been experienced, even if there is no conscious intention directed towards identity, and no relational identification [as in an identity claim]” (§8, Addendum). 11. Horgan et al. think this reference is “wholly constituted phenomenologically” (p. 304), something sharply disputed in Loar (2003). 12. Th is is how they introduce the notion in Horgan and Tienson (2002): Phenomenal intentional content presents to consciousness an apparent world that goes far beyond what one is conscious of at any one time; presuming so is itself an aspect of the overall phenomenal character of experience. Phenomenal intentionality thereby determines a complex set of presuppositions concerning the existence of, the persistence of, and various features of, the sorts of entities presented in experience: presuppositions about individuals (including flora, fauna, and other creatures like yourself), kinds, properties, relations, processes, and events of that world. . . . [W]e call these presuppositions grounding presuppositions. . . . In making a grounding presupposition, one takes it that there really exists an entity of a certain sort; and normally, one also presupposes that the (putative) entity in question has certain specific attributes. If there is an actual entity satisfying that presupposition (or satisfying it near enough), then one’s thoughts that are intentionally directed toward such a putative entity will refer to the actual entity in question; and so the properties of the satisfier will determine whether the beliefs about it are true or false, whether hopes and desires about it are satisfied, and so forth. (p. 528) 13. It might be thought that Loar can count co-directedness as simply another primitive phenomenal notion, not explicable in terms of other such notions. But this would make the notion of a state’s phenomenal-intentional content both more complex and more mysterious than it appeared at fi rst. How would such a view deal with the way the recognition

154

14.

15.

16.

17.

18.

19.

20.

21.

phenomenal intentionality of an object may be accompanied by feverish attempts to work out where it is recognised from? Or the contrast between that level of recognition and complete familiarity with the object? A view of this kind would seem far more fanciful than Lycan’s “fanciful Meinongianism.” Beyer (2007) suggests that, for Husserl, “the intentional act [of thinking about Pegasus, say] is merely ‘as if of’ an object” and lacks an actual object. But a portrayal of Husserl as an as-if-ist is hard to square with the evidence. Passages like the following abound in his work: “In real phenomenological treatment, objectivity . . . transcends the act. It makes no difference what sort of being we give our object, or with what sense or justification we do so, whether this being is real or ideal, genuine, possible or impossible, the act remains ‘directed upon’ its object ” (Husserl 2001, Investigation V, §20). As far as I can see, whenever Husserl talks about as-if, he is talking about the way in which merely intentional objects can present themselves as if they are real (thus, “[the concept of experience] also includes experience in the mode of as-if, i.e., the givenness of the individual in phantasy, which in an appropriate, always possible, free alteration of attitude turns into positional experience of a possible individual” [Husserl 1973, §6]). Th is second formulation renders the appeal to relationships of acquaintance more like the kind of appeal that Horgan et al. (2004) make to phenomenally constituted dispositions (cf. their comments about elms and the role played by what they call “counterfactual phenomenal facts,” such as facts about what you would fi nd out if you were to ask experts about elms; pp. 309–10). Searle (1983 is the locus classicus for such a reflexive conception of intentional content. (A related view in the philosophy of language is the doctrine of causal descriptivism, which lets causal descriptions feature in the meaning or reference-fi xing conditions of reference; see, for example, Kroon 1987, Lewis 1994.) I am here leaving open an important question about the status of the notion of acquaintance. I am inclined to think of it as a primitive intentional notion, but if this is wrong, then the view described in this chapter owes some kind of theoretical account of the notion. Th is doesn’t, as far as I can determine, affect the substance of the view or the argument that the chapter presents for the view. It also yields the following benefit: the part of narrow content that involves the attribution of properties (“There is something in front of me that is a picture . . . ”) will be recognised as having a fallible status, being an attribution that rests on a fallible route to knowledge (for example, visual acquaintance). So unlike Horgan et al., I do not think that we should expect the reference of a singular thought-constituent to satisfy the grounding presuppositions without qualification. See, for example, Horgan et al. 2004, pp. 313–14. In fact, the reflexive role of acquaintance tends to be relatively close to a subject’s consciousness, unlike some other phenomenal features that phenomenal intentionalists have appealed to (including some of the features cited by Horgan et al.). In the crooked picture case, for example, a subject does not need to be very articulate to hazard something like “The picture I am looking at is the only picture I see in front of me, and I recognise it as one I have seen before.” I am sympathetic to Horgan, Tienson, and Graham’s view that something is the referent of a singular thought-constituent if, and only if, it satisfies the (according to me, acquaintance-based) grounding presuppositions. But the point I am making goes through even on an externalistic theory of reference determination of the kind preferred by Loar. All that is needed is the right-to-left conditional—on its own, the truth of this conditional is consistent with there being reference even if the narrow presupposition-conditions are not fulfilled. The temptation to speak of seeing a “hallucinatory” object is the temptation of describing the experiencing subject’s error from inside the scope of the simulation. In reality, there is no object—it is merely hallucinated that there is an object. See Kriegel (2007) for a very different answer to the question of how (to use his example) we can represent Bigfoot even though Bigfoot does not exist. Kriegel answers this question in terms of an adverbial account of (conscious) representation. My answer, by contrast, has attempted to explain and even preserve—not explain away—the intuition that we represent Bigfoot.

The Role of Intentional Objects

155

References Beyer, C . (2007). “Edmund Husserl,” Stanford Encyclopedia of Philosophy (Winter 2007 edition), Edward N. Zalta (ed.), http://plato.stanford.edu/archives/win2007/entries/husserl/. Chalmers, D. J. (2004). “The Representational Character of Experience.” In B. Leiter, ed., The Future for Philosophy. Oxford: Oxford University Press. Farkas, K . (2008). “Phenomenal Intentionality without Compromise.” The Monist 91: 273–93. Harman, G. (1990). “The Intrinsic Quality of Experience.” In J. E. Tomberlin, ed., Philosophical Perspectives, Vol. 4: Action Theory and Philosophy of Mind. Atascadero: Ridgeview. Heidegger, M. (1992). History of the Concept of Time. Translated by Theodore Kisiel. Bloomington: Indiana University Press (Midland Book Edition). Horgan, T., and J. Tienson. (2002). “The Intentionality of Phenomenology and the Phenomenology of Intentionality.” In D. Chalmers, ed., Philosophy of Mind: Classical and Contemporary Readings. Oxford: Oxford University Press. Horgan, T., J. Tienson, and G. Graham.(2004). “Phenomenal Intentionality and the Brain in a Vat.” In R. Schantz, ed., The Externalist Challenge: New Studies on Cognition and Intentionality. Amsterdam: de Gruyter. Horgan, T., and U. Kriegel. (2008). “Phenomenal Intentionality Meets the Extended Mind.” The Monist 91: 347–73. Husserl, E . (1973). Experience and Judgment. Revised and edited by L. Landgrebe, and translated by J. S. Churchill and K. Ameriks. Evanston: Northwestern University Press. Husserl, E . (2001). Logical Investigations, Volumes 1 and 2. Translated by J. N. Findlay with a new Preface by Michael Dummett, edited with a new introduction by Dermot Moran. London: Routledge. Kroon, F. (1987). “Causal Descriptivism.” Australasian Journal of Philosophy 65: 1–17. Lewis, D. (1994). “David Lewis: Reduction of Mind.” In Samuel Guttenplan, ed., A Companion to the Philosophy of Mind, 412–31. Oxford: Blackwell. Kriegel, U. (2007). “Intentional Inexistence and Phenomenal Intentionality.” Philosophical Perspectives 21 (1): 307–40. Loar, B. (2003). “Phenomenal Intentionality as the Basis of Mental Content.” In M. Hahn and B. Ramberg , eds., Reflections and Replies: Essays on the Philosophy of Tyler Burge. Cambridge, MA : Bradford Books/MIT Press. Lycan, W. G. (1996). Consciousness and Experience. Cambridge, MA : Bradford Books/MIT Press. Lycan, W. G. (2008). “Phenomenal Intentionalities.” American Philosophical Quarterly 45: 233–52. Parsons, T. (1980). Nonexistent Objects. New Haven: Yale University Press. Searle, J. (1983). Intentionality, an Essay in the Philosophy of Mind. Cambridge: Cambridge University Press. Siewert, C . (2002/2006). “Consciousness and Intentionality.” Stanford Encyclopedia of Philosophy, http://plato.stanford.edu/entries/consciousness-intentionality/. Strawson, G. (1994). Mental Reality. Cambridge, MA : Bradford Books/MIT Press. Yablo, S. (1998). “Does Ontology Rest on a Mistake?” Proceedings of the Aristotelian Society, Suppl. Vol. 72: 229–261. Zalta, E . (1988). Intensional Logic and the Metaphysics of Intentionality. Cambridge, MA : MIT Press.

9

Unconscious Belief and Conscious Thought Tim Cr ane

1. Introduction We call our thoughts conscious, and we also say the same of our bodily sensations, perceptions, and other sensory experiences. But thoughts and sensory experiences are very different phenomena, both from the point of view of their subject and in their functional or cognitive role. Does this mean, then, that there are very different kinds or varieties of consciousness? Philosophers do often talk about different kinds of consciousness: Christopher Hill, for example, claims that “it is customary to distinguish five forms of consciousness” (Hill 2009: 1). These are: agent consciousness, propositional consciousness, introspective consciousness, relational consciousness and phenomenal consciousness—to which Hill adds experiential consciousness, making six in total. It is an important question for the theory of consciousness whether these labels pick out genuinely different psychological phenomena, or whether some (or all) of them are just different ways of talking about the same thing. I cannot address this large question here, but I will instead focus on a more specific question: is the kind of consciousness involved in sensation an entirely different thing from the kind of consciousness involved in thought? Or is there a single kind of consciousness which both these kinds of mental phenomena exhibit, regardless of how else they might differ? In this chapter I will attempt to answer this question. Some dominant theories of consciousness give it quite unsatisfactory answers. I will consider three recent theories: that consciousness consists in the instantiation of a simple quality; that consciousness consists in higher-order representation; and that there are two independent kinds of consciousness with wholly different natures (in this connection, I will consider Block’s well-known distinction between 156

Unconscious Belief and Conscious Thought

157

“phenomenal” and “access” consciousness). None of these gives entirely the right answer, although the last comes closest. I will argue for three claims: (1) that Block’s notion of access consciousness must be understood in terms of phenomenal consciousness, and in this sense phenomenal consciousness is the more fundamental notion; (2) beliefs are never phenomenally conscious, though episodes of thinking are; (3) the sense in which thoughts and experiences are conscious is that they are both a certain kind of episode, that I call an episode in the stream of consciousness. Th is last claim provides the answer to the question posed earlier: there is a single kind of consciousness which thoughts and experiences exhibit. I do not give a definition of what phenomenal consciousness is, nor a description of it in non-mental terms. Indeed, like many philosophers working in this area, I take the fact of consciousness for granted, and I do not believe that it can be reduced or “defined.” Rather, I will attempt to outline some of the relationships between conscious thoughts, other conscious episodes, and unconscious mental states like beliefs, with the aim of getting a clear overview of aspects of the phenomenology of consciousness. How do my claims connect to the currently popular idea of “phenomenal intentionality”? The term “phenomenal” has been understood in a number of ways in contemporary philosophy, but if we restrict it only by its etymology, then “phenomenal” means pertaining to appearances. So phenomenal intentionality is intentionality that relates to how things appear—but “appear” must mean “consciously appear” if the term’s current usage is to make any sense. Conscious thought is the most crucial phenomenon for theories of phenomenal intentionality. All theories of consciousness recognize the existence of phenomenal consciousness in the sense I have just described (even if they do not use this term). Sensations and sensory experiences are the paradigm examples. Many theories hold in addition that such experiences have intentional content—that they represent states of the world or states of the subject. Intentionalist theories of consciousness are now widely accepted, even if they are not entirely uncontroversial. But just as intentionalist theories of consciousness are not news, the phenomenal intentionality they talk about—the intentionality of bodily sensation and perception—is not news either. What is controversial is whether there is a distinctive kind of conscious or phenomenal intentionality associated with conscious thought (see, for example, Pitt 2004; Bayne and Montague 2011). This is not something which is implied by the very idea of an intentionalist theory of consciousness. It is possible to hold that all consciousness is intentional, or representational, without holding that conscious thinking has any distinctive phenomenology of its own. Michael Tye, for example, is a well-known defender of intentionalism about consciousness; but recently he has explicitly denied that thoughts have “a further phenomenology over and above the imagistic and sensory phenomenology that may

158

phenomenal intentionality

accompany conscious thoughts” (2009: 201). For Tye, there is only imagistic or sensory phenomenal consciousness; and since all consciousness is intentional, there is only imagistic or sensory phenomenal intentionality. However, I am persuaded by the arguments of Kriegel (2007), Pitt (2004), Strawson (1995) and others that there is a distinctive kind of phenomenal intentionality associated with conscious thought: thought has a so-called proprietary phenomenology. But I will not attempt to argue for this here. Rather, I am attempting something more modest: to sketch an account of the structure of conscious thinking in order to make intelligible how thought can be conscious in the same sense as sensations and sensory experiences. In the next section, I will argue that two well-known theories of consciousness fail to give an adequate answer to the question of why thoughts and sensations are both conscious, and that Block’s claim that “conscious” is ambiguous also fails to answer this question in a satisfactory way. I will then argue in section 3 that my approach to Block’s distinction also helps to illuminate what is sometimes called (I think misleadingly) conscious belief. In section 4 I show the bearing of these points about belief on the phenomenon of conscious thought, and hence on the thesis of phenomenal intentionality.

2. Two Inadequate Answers to Our Question Some popular theories of consciousness, it seems to me, are unable to give a single account of the consciousness involved in thought and in sensation. One is the theory that consciousness involves the instantiation of intrinsic, non-intentional properties, known as “qualia” (Block 1996). Another is the theory that consciousness is higher-order thought or representation: the “HOT” theory of consciousness (Rosenthal 1986). While the qualia theory may have some plausibility for conscious sensations, it is hard to apply it to conscious thought. And while the HOT theory might make sense in the case of conscious thought, it has little plausibility when applied to sensation. The qualia theory identifies consciousness with the instantiation of qualia: properties which are variously described as intrinsic, non-intentional, ineffable and private. The details of particular theories of qualia do not matter here, since my objection is quite general. The paradigms of qualia are bodily sensation properties, like the property of being in pain, or perhaps some specific properties of pain states or events. Sensory qualities like this have a particular “feel” which is essential to them. To each distinctive type of quale corresponds a distinctive sensory “feel.” Conscious thought cannot plausibly be described in this way. If you are currently thinking about the causes of the current financial crisis, there is no particular sensory feel that this episode has. In fact, it sounds strange to talk of

Unconscious Belief and Conscious Thought

159

this episode as having any “feel” at all. If it feels like anything to be thinking about the financial crisis, this kind of “feeling” has so little in common with sensory feelings that it is unilluminating to describe them both as having qualia. If we are inclined to say that thinking “feels like” something, then all we can really mean by this is that it is conscious. But this is what we are trying to understand, so it is hardly an explanation to say that there are “feelings” in this sense associated with thinking, or that thinking involves qualia. The HOT theory might appear to fare better in accounting for conscious thought. On this theory, a mental state’s being conscious is a matter of its being the object of a higher-order thought or other representation. Theories differ as to whether the higher-order state actually has to be instantiated or whether one needs only to have the capacity or disposition to have these higher-order representations (for the latter view, see Carruthers 2003). But again, the details do not really matter here, since my question is the general one of whether such a view can possibly apply both to conscious thought and to conscious sensation. In the case of thought, it might seem a plausible approach: what it is to have a conscious thought is to be conscious of your thought, and what is being conscious of a thought other than representing it? (Lycan 2001 uses this point as part of his “simple argument” for a higher-order theory of consciousness.) The point is not indisputable, but it has something to be said for it. (In §4 I will dispute it.) Yet in the case of sensation, being conscious of a sensation cannot be understood simply in terms of representing it. We can envisage creatures who are incapable of thought about their mental states but nonetheless can feel sensation—surely many animals are actually like this. And even in our own case, the consciousness of a sensation is plausibly something which is internal to the sensation itself, rather than a product of thought or representation of it. “Conscious of” does not always mean “representing.” Being conscious of a pain in one’s toe—in the sense of being aware of it when you have it—is a different thing from being conscious of the pain when you think about it and attend to it. Merely having the pain, in normal cases, is sufficient for awareness of it. One way to respond to the inadequacies of these two theories is to say that each is right in its own way—the qualia theory is on the right lines when it comes to conscious sensation, while the higher-order thought theory is on the right lines when it comes to conscious thought. This is either because “consciousness” is ambiguous, or because there are two (or more) kinds of consciousness; or these might be two ways of saying the same thing. I will briefly examine the best-known version of this proposal: Ned Block’s distinction between phenomenal and access consciousness. In his original (1995) presentation of the distinction, Block said that “‘consciousness’ is an ambiguous word, though the ambiguity I have in mind is not one I’ve found in any dictionary” (Block 1995: 391). Phenomenal consciousness

160

phenomenal intentionality

he defines in a usual way as “experience: what makes a state phenomenally conscious is that there is something ‘it is like’ to be in that state” (Block 1995: 380). Access consciousness is defined for representations, and in terms of these representations’ role in reasoning and control of action: “a representation is access-conscious if it is poised for free use in reasoning and for direct “rational” control of action and speech” (1995: 382). In later writings on this subject (for example, Block 2002), Block modified his presentation somewhat, to answer criticism suggested by his choice of terminology. To indicate that he is using the words in a more technical sense, he called phenomenal consciousness “P-consciousness” and access consciousness “A-consciousness.” For Block, sensations are the paradigms of P-conscious states, while propositional attitudes (for example, belief) are the paradigms of A-conscious states (2002: 281). Could this be the answer to my question, then: that sensations are phenomenally (P) conscious, and thoughts are merely access (A) conscious? On the face of it, the answer seems to be no, for two reasons. First, thoughts can be phenomenally conscious too, as we noted previously, in a perfectly ordinary sense of “phenomenal.” When consciously thinking about something, things can appear to you a certain way—you might be imagining something, reflecting on it intellectually, or preoccupied by memories of it. But all these cases involve the appearance of things in the world, and (as we noted earlier) this is the original meaning of “phenomenal.” It is, of course, common these days to use the word “phenomenal” to apply only to sensory consciousness. But even if one were to restrict the word in this way, visual imagination still counts as thinking, even though it involves sensory “phenomenal” qualities. So if visually imagining can be a kind of thinking, then some thinking can be phenomenally conscious, in the broader or narrower sense of that word. The second reason that Block’s distinction cannot be straightforwardly used for answering our question is that sensations themselves can be available to reasoning systems; they are accessible for verbal report; they are available to memory, and so on. So sensations can be A-conscious too. It might be said that sensations are not representations, so by definition are not A-conscious states. But it is not obvious that sensations are not representations: representationalists or intentionalists explicitly claim that they are (see Tye 1995; Crane 2001). Such representationalism may be false, but it is not obviously false. Block’s distinction does not provide a clear and simple answer to our question. The distinction (between being conscious as such and “accessing” one’s conscious states) is certainly something real, but I am skeptical of the way he tries to distinguish one from the other as independent phenomena, or distinct existences, by showing how they can be dissociated. The reason for my skepticism is that although there clearly is such a thing as “accessing” one’s states of mind, this cannot be fundamentally understood independently of the idea of phenomenal consciousness. I will now argue for this claim.

Unconscious Belief and Conscious Thought

161

When Block originally introduced the notion of access consciousness, he introduced it in term of the notion of being “poised” for control of action and speech. “Poised” in this sense is a dispositional notion—I can be poised to dive into the swimming pool even if I do not actually do it. In this respect it is like the ordinary notion of “access”: I can have access to the money in my bank account even if I am not actually “accessing” it. Yet as Block later acknowledged, a dispositional notion is not what is needed here since consciousness is an occurrence (2002: 279). So he revised his definition of A-consciousness by saying that a representation is A-conscious if it is broadcast for free use in reasoning and so on (2002: 278). Broadcasting seems to be an event-like thing, or a process, and hence something occurrent. (I will return to occurrences in §§3–4.) But when Block spells out what “broadcast” literally means, he appeals once again to dispositional ideas. At one point he glosses “broadcast” in terms of the senses “making available” representations to other “consuming systems,” such as systems of classification, memory and so on (2002: 278). But what if these representations are made available to other systems and these systems never take advantage of their availability? Are they still conscious in any ordinary sense? Block also says that an element in A-consciousness is “reportability” (2002: 278). But, again, what if these representations are reportable and no one actually reports on them? Are they still conscious? I don’t think it is particularly helpful here to identify A-consciousness with information-processing, as Block does. It’s not just that there is nothing in the idea of processing information which requires that it be conscious (although this is true). It’s rather that all the paradigm examples of cognitive information-processing which we have are examples of things that take place unconsciously. Block may say that this is to use the word “conscious” in the phenomenal sense. But the difficulty is that everyone will agree that much of our psychological information-processing is unconscious. What is it, then, that distinguishes the A-conscious information-processing from the unconscious? Block is sensitive to the criticism that a creature who only had A-consciousness in his sense would (in some intuitive sense) be an unconscious creature, but nonetheless insists that it is worth calling A-consciousness a kind of consciousness. As he points out, the access-related notion is the notion used in some psychoanalytic discourse—when one has an unconscious desire in something like the Freudian sense, it is a desire that one cannot access. And the notion of consciousness as accessibility or reportability is also used in much psychology and cognitive neuroscience. These are his reasons for insisting that if a creature were purely A-conscious, it would still be conscious in an ordinary sense. But he concedes that “A-consciousness can be a kind of consciousness even if it is parasitic on a core notion of P-consciousness” (2002: 282). I think Block is quite right to make this concession. But how should we spell out the idea that A-consciousness is parasitic on P-consciousness? I think the

162

phenomenal intentionality

connection is simpler than Block says. The idea that there is an interesting category of states which we can access is (in general) quite true. And we should understand access in terms of availability to other processes and cognitive systems in the way Block says. Assuming that accessing something is bringing something about, then we can ask: what kind of thing is brought about? The obvious answer is: a phenomenally conscious occurrence. To access or “broadcast” a mental state is to bring about a change in phenomenal consciousness. So understood, accessing our mental states is not a special kind of consciousness, but rather a matter of bringing those mental states to phenomenal consciousness. This understanding of “access” (or “broadcasting”) presupposes phenomenal consciousness. However, it does not follow from the fact that some essentially non-conscious state S brings about a change in phenomenal consciousness that S itself is made conscious. Many states or events in the brain bring about changes in phenomenal consciousness without being conscious in any way whatsoever. The important thing is not just that the state brings about the change in phenomenal consciousness but that the state itself “figures” (in some sense) in consciousness. And since mental states are individuated by their contents, what is important about the state itself “figuring” here is its content. Consciously accessing your belief that Mongolia had the second largest empire in the history of the world implies that this content is brought to phenomenal consciousness, since that is what distinguishes that belief from all others. So to bring a state of mind to phenomenal consciousness is (at the very least) to bring its content to consciousness: that is, to bring about an event in phenomenal consciousness in which that content figures. The next question, then is this: what is it to bring the content of some mental state to phenomenal consciousness. I will discuss this question in §4. But before doing this, I will explain how my way of understanding Block’s distinction makes sense of what is sometimes called “conscious belief.”

3. “Occurrent” and “Conscious” Belief It is frequently said that there is a distinction between beliefs which are conscious and those which are unconscious. Some go further: Christopher Peacocke has attempted to define the concept of belief partly in terms of the idea of conscious belief: A relational conception R is that of belief only if the following condition is met: (F) The thinker finds the first-person content that he stands in R to the content p primitively compelling whenever he has the conscious

Unconscious Belief and Conscious Thought

163

belief that p, and he finds it compelling because he has that conscious belief. (Peacocke 1993: 163) For Peacocke’s definition to be intelligible, there must be an intelligible distinction between conscious and unconscious belief. But what is this distinction? Some theories attempt to explain the distinction in terms of lower-order versus higher-order attitudes: D. H. Mellor, for example, has argued that a conscious belief in a proposition is simply the belief that one believes it (1977: 90). Mellor’s proposal resembles the higher-order thought theory defended by Rosenthal (1986). But unlike Rosenthal, he does not think it is the correct account of all forms of consciousness. Others explain conscious belief in terms of the idea of an “occurrent” belief, which is distinguished from “dispositional” belief. Here I will focus on this latter thesis: that conscious belief is occurrent belief. When we identify what is wrong with this thesis, it will show us what is wrong with the idea of conscious belief as higher-order belief. The reasons for taking beliefs to be dispositions—to combine with other mental states to cause actions or other mental occurrences—are well known, and I will not dwell on them here.1 Belief understood as a disposition is sometimes contrasted with occurrent belief, usually in order to identify and describe the phenomenon of “conscious belief.” Uriah Kriegel, for example, argues that we should “construe a conscious belief as one that is at once an occurrent first-order belief and an occurrent second-order belief” (Kriegel 2004: 108; see also BonJour 1999: 131). Kriegel implies that the distinction between firstand second-order beliefs is not sufficient to pick out the conscious beliefs: the beliefs must be occurrent too. An occurrent belief is presumably supposed to be a kind of occurrence, that is, an event or episode. A disposition is a different kind of thing: a state or property, something that persists beyond its manifestations (which are themselves occurrences). So an occurrent belief cannot be the very same thing as a dispositional belief, any more than a breaking can be the very same thing as an instance of fragility ( pace Gennaro 1996: 41). It cannot be right to say that the very same belief state is both conscious and unconscious, if “conscious belief” is understood as occurrent belief. What must be meant (at least) by saying that a conscious belief is the same as an unconscious belief is that it has the same content: these are states in which you believe the same thing (that is, the same proposition). But the state also needs to involve accepting or endorsing that content, if it is to be anything like a belief. As Laurence BonJour puts it, “what one is primarily conscious of in having such a belief is precisely its propositional content (together with one’s acceptance of that content)” (BonJour 1999: 131). There can be no objection in principle to the idea that mental phenomena of different types may have the same content. You can hope that the weather will improve, and you can believe that the weather can improve. But hope and belief

164

phenomenal intentionality

are unquestionably different mental phenomena. More controversial, but still widely endorsed, is John McDowell’s well-known claim that the content of a perceptual experience can also be the content of a judgement (1994: 9).2 So there cannot be a general objection to the idea of conscious and unconscious beliefs having the same content. However, in cases like these, the states are different because what (we can call) their “attitudinal” components are different: belief, hope, perception or judgement. But occurrent and dispositional belief are both supposed to be kinds of belief. So they must both have some or all of the distinguishing marks of belief. But it hard to see how whatever occurrence it is that is being labeled by the words “occurrent belief” can have these distinguishing marks. The essence of belief, I take it, is that it is taking something to be the case. We could also say that a belief is taking something to be true, so long as this does not imply that in order to have beliefs, a believer must have the concept of truth. This is a more controversial thesis, and should not be built into the minimal idea of belief which all should accept. Likewise, a believer (in this minimal sense) need not have the concept of something’s being the case. In this sense, a relatively simple creature can have beliefs, so long as it takes the world to be a certain way, or takes certain things to be the case. “Taking something to be the case” could be understood in terms of perceptual consciousness, where what is taken to the case is present to the mind at the moment it is perceived. But this is not how it is with belief: belief is not just a matter of taking something to be the case for the duration of (e.g.) a perceptual experience. Rather, it is essential to beliefs that they persist through changes of in current consciousness. Beliefs are stored in memory and can be called upon when future action is needed. It is crucial that they do this if they are to guide the actions of organisms in the way they do. An organism which could only take things to be the case as they perceptually seem at the present moment would have to learn anew each fact about its environment. It’s hard to know whether there are any such creatures, but it is clear enough that we and other believers are not like this. When a belief is formed, by perception or by other means, it can be stored in the organism’s mind and then can be put at the service of the organism’s projects in various ways—either by guiding action, by being used in reasoning or by being used in planning. For beliefs to play this functional role, they must continue to exist—that is, persist—beyond the moment of their acquisition. This feature of beliefs is reflected in our commonsense psychological understanding: we do not count someone as stopping believing something when they go to sleep, lose consciousnesss, or when the subject-matter of their belief is not before their conscious mind. A defender of “conscious beliefs” must say that this is one of the reasons that we should recognize the phenomenon of unconscious belief: in descriptions of this phenomenon, we are typically invited to

Unconscious Belief and Conscious Thought

165

think of all the things which we believe which have rarely, if ever, come before our conscious mind. This is quite right—but it is rarely recognized that since the persisting character of belief is essential to it, it makes conscious belief hard to understand. If a conscious belief were an occurrence, it would exist for as long as that occurrence were to exist. For a belief to cease to be conscious, then, on this understanding, would be for the occurrence to cease, or to go out of existence. But if the occurrence thus went out of existence—for example, when the subject paid attention to something else, or became unconscious—then it cannot play the essential role of belief as I have characterized it. So whatever such an occurrence is, it cannot be belief: “occurrent belief” is not belief at all. It might be objected that this is a merely verbal point: that all I am doing here is deciding to use the word “belief” for the persisting, unconscious functional state, and forbidding its use for occurrent, conscious episodes of judgement. But since I do not deny that that there are such conscious episodes, then what can be wrong with calling them “beliefs,” so long as we also distinguish episodes from persisting states? It is, of course, notoriously difficult to say when a thesis or question is a “merely verbal” one. But if I am right that there is a fundamental metaphysical distinction between episodes/occurrences and persisting states, then it is misleading at best to use the term “belief” for entities in both categories. For one kind of entity goes out of existence when we go to sleep and another does not. How much more different could these entities be? Calling these two states “belief1” and “belief2” would only obscure the fundamental difference between them. Some support for this view of belief—that conscious belief is not a kind of belief—comes from reflection on the phenomenon of knowledge. Although on most views, there is a close connection between belief and knowledge, it is not common to hear talk of “conscious knowledge,” as there is this tendency to talk of conscious belief. And no one talks of “occurrent” knowledge.3 This is revealing: for if belief can be occurrent, why can knowledge not be so too? And yet knowledge often plays the same kind of functional role as is often attributed to belief, in organizing an subject’s behaviour and reasoning (see Williamson 2000: 60–64). And we do not consider that people necessarily lose their knowledge when they go to sleep. Knowledge can be lost, of course: there is forgetting, and there are maybe other ways of losing knowledge too (see Dretske and Yourgrau 1983). But people do not lose knowledge in any sense simply because they think about something else, or lose consciousness. So for the same kinds of reason that belief cannot be an occurrence, knowledge cannot be an occurrence either. What is undeniable is that there is such a thing as bringing what you know to mind, or to consciousness. Bringing X to mind is without question an

166

phenomenal intentionality

occurrence; but this fact on its own does not make “bringing X to mind” a special kind of X. One could stipulate here, and say that all that is meant by calling a belief “occurrent” is that it is one which you are bringing to mind, or bringing to consciousness, or consciously attending to. This is how I understand a remark like this of Graham Oddie’s: “a belief is occurrent if it is the belief that you are consciously attending to” (Oddie 2005: 240). Whatever the merits of such a stipulation, it clearly isn’t supposed to be part of an explanation of consciousness, since it assumes the idea of “consciously attending” to something. This is not an objection to the stipulation as such; it is just to point out what the stipulation is, and is not, supposed to be doing. Nonetheless, whatever stipulations we make, we should accept that there is such a thing as attending to what we know and believe. “Bringing what we believe to consciousness” is a good description of this, however this notion should be further understood.4 And it is plausible that this is the same phenomenon as the phenomenon Block calls “A-consciousness.” I argued earlier that we should understand Block’s notion of A-consciousness in terms of bringing about a phenomenally conscious episode. That is, a state is A-conscious when it is accessible to consciousness, where this is understood in terms of being available to figure in an episode or occurrence in phenomenal consciousness. This is also how I think we should understand the talk of conscious belief. As I argued in the previous section, we should not say that A-consciousness is a distinct sort of consciousness simply because it is a disposition to bring about events in phenomenal consciousness. And the same applies to the idea of “conscious belief.” Just as, on Block’s view, a representation is “A-conscious” when it can be brought to phenomenal consciousness, so a belief is “conscious” or “occurrent” when it can be brought to phenomenal consciousness. “Occurrent” in “occurrent belief” is a sort of transferred epithet: what is occurrent is not the belief itself but the episode of being conscious of that belief. It is important to emphasise that bringing a belief to consciousness is not just a matter of being aware of its content—since, as we saw earlier, the same content can be the content of many different kinds of mental state. When a belief is brought to phenomenal consciousness, the “attitudinal component” (what in other work (2001) I have called, following Searle (1992), the “intentional mode”) must also be an object of phenomenal consciousness So my understanding of “bringing something to consciousness” is the same as my understanding of A-consciousness: what is conscious, first and foremost, is an event, not a persisting state. If this is right, then this tells against those accounts of “conscious belief” (like Mellor’s and Rosenthal’s) which attempt to explain the consciousness of a belief in terms of the existence of a certain kind of state (that is, a higher-order belief). For all belief states must have the essential characteristic of beliefs: namely, they must persist through changes in consciousness, or even through the absence of consciousness. But if this is

Unconscious Belief and Conscious Thought

167

an essential characteristic of all belief states, it will be possessed by higherorder belief states as well. So higher-order belief states cannot be conscious episodes.

4. Thought and the Stream of Consciousness In §2 I argued that what Block calls A-consciousness should not really be thought of as a different kind of consciousness, distinct from phenomenal consciousness. Rather, to call a state “A-conscious” is to say that it can give rise to phenomenally conscious episodes. I then applied the same idea to the distinction between conscious and unconscious belief. What are called “conscious beliefs” are not really beliefs at all, but rather phenomenally conscious episodes: episodes of being aware of what you believe and the fact that you believe it. The belief state itself is not conscious; rather, you can be conscious of what you believe when such a phenomenally conscious episode occurs. So when we call persisting mental states “conscious” what we really mean is that they can give rise to a certain kind of phenomenally conscious events. It follows that when Block talks of A-conscious states, he should be interpreted as talking about states which can give rise to—or manifest themselves in— certain phenomenally conscious events, or what can be called more generally “occurrences.” What the philosophy of consciousness needs, then, is the distinction between occurrences and persisting states.5 Events and processes are both occurrences: they take time, they can unfold over time and have temporal parts. My watching a bird catching a worm in the garden is an event, as is the event I am watching. Both events take time, they unfold or develop over time, they typically consist of earlier parts or periods. Maybe there are also instantaneous events; in which case, not all events have significant temporal parts. But nonetheless, an instantaneous event is still an occurrence: something that happens. States do not unfold or evolve over time: they consist in the instantiation of properties and/or relations in objects. If we think of an object as changing its state—for example, its temperature—this is not because the state itself evolves over time, but because one state is followed by a different one. This is because states do not have temporal parts. My having the height I do is a state of me; it makes no sense to talk about the earlier part or the later part of my having the height I do. D. H. Mellor once distinguished between objects and events by saying that unlike events, objects are “wholly present” at each moment of their existence (1981: 104). This distinction can be used also to distinguish between events and states. An event with temporal parts cannot be wholly present throughout each moment of its existence: at any moment at which the event is “going on”

168

phenomenal intentionality

or happening, some temporal parts of the event exist or are present, and others do not and are not. Th is is not true with states. Consider a state which consists in an instantiation of a property in a particular over a period of time (note that I am not saying that all states have this structure; this is just an example). The property and the particular are wholly present at each moment of the state’s existence or instantiation. If the particular ceases to have this property, then it undergoes a change of state. The change itself is an occurrence, but the resulting state is not. It is the distinction between occurrences and states which gives the nonmetaphorical backbone to William James’s famous image of a stream of consciousness: Consciousness, then, does not appear to itself chopped up in bits. Such words as “chain” or “train” do not describe it fitly as it presents itself in the first instant. It is nothing jointed: it flows. A “river” or a “stream” are the metaphors by which it is most naturally described. In talking of it hereafter, let is call it the stream of thought, of consciousness, or of subjective life. (James 1890: 233) The image of a stream flowing suggests motion, and one thing following another. But “following” in the case of consciousness must mean following in time. Events follow one other in time; objects do not (see Davidson 1967: 691). A cat can literally follow a mouse in space (for example, around a room), but the event of the cat’s eating the mouse “follows” the event of the cat’s catching the mouse only in its temporal ordering. Of course, states too can be ordered in time, since they are defined in terms of objects, properties and times. States can be ordered as their constituent times are ordered: as earlier than, later than or simultaneous with others. For some philosophers, this shows that all we need is the notion of a state, and no independent notion of an event. (Alternatively, they might, like Lewis (1986) and Kim (1976) use the word “event” for what I am calling states.) I cannot here show that such a reductionist view of events—a reduction of events to what I would call “states”—is incorrect. All I will assume here is that the distinction is phenomenologically plausible for the reason James says: a sequence of states is something more like a train or a chain than something flowing, and conscious processes do not seem like chains or trains. Of course, this phenomenological observation is not decisive: it could be said that what appears to be continuous and flowing at the phenomenological level might be constituted by states which are discrete at the level of the underlying reality (just as a fi lm in the cinema might represent continuous motion, despite the fact that it is made up of many discrete images shown one after another).

Unconscious Belief and Conscious Thought

169

Here though, my concern is with consciousness, with the phenomenal, with the apparent metaphysics of how things seem; and not with the fundamental metaphysics. The distinction between the state of something—(for example, the shape of the tree outside my window) and an event (for example, the woodpecker pecking against its bark)—is a distinction which it is easy to appreciate from within consciousness. Whether the fundamental reality of the universe is event- or state-like is a further question. In §§2–3 I argued that we should think of A-conscious states (for example, so-called conscious beliefs) as states which could be “brought” to phenomenal consciousness. But what does it mean to “bring” a state to phenomenal consciousness? One answer is clear: it is not to create another, conscious, version of the state itself. What is brought to mind when reflecting consciously on one’s beliefs is not itself another belief. My claim is that what is brought to mind is, rather, a conscious thought. A conscious thought is an episode of thinking something. “Thought” here refers to the episode or “act” of thinking, not to what is thought (the latter is a proposition, or a Gedanke in Frege’s terminology). Thoughts in the sense in which I use the word form a heterogeneous category: I include episodes of speculation, wondering, imagining, planning, as well as the central phenomenon of judgement. There are three broad kinds of conscious occurrence which can be called judgement. The first kind occurs when subjects are making up their minds about something. When they have not formed a belief about some subject-matter and need to do so, they may weigh up the evidence and come to one conclusion or other; or they may “suspend judgement.” Here judgement is the formation (or the attempted formation) of belief. Judgement stands to belief in this case as decision stands to intention. Just as a conscious judgement is (in this kind of case) an event which is the formation of a belief, so a conscious decision is an event which is the formation of an intention. Intentions, like beliefs, are persisting states, and like beliefs they can manifest themselves in consciousness. (Another similarity is that one can take oneself to have formed a belief, and yet future circumstances show that one did not really have this belief; the same can be true of intention.) Episodes of the second kind of judgement are those which occur when our beliefs are brought to mind without effort, or any act of will. We need our beliefs to guide our action in the world, and often this needs to be conscious. Or we reveal what we believe in consciousness when someone asks us a question, and the answer occurs to us immediately, without effort or without having to make up our minds. The third kind of case occurs when one deliberates explicitly in consciousness about what one already believes. This can occur when one is trying to work out what one believes, or remember some fact, or draw out some consequence

170

phenomenal intentionality

of what one believes. As with the second kind of case, this isn’t a case of forming a belief, but rather a case of revealing to oneself what one believes anyway. But this case involves effort, and what is sometimes called introspection. In contemporary terminology, occurrences of all three kinds are manifestations of our capacity for “self-knowledge” (see Cassam 1994). The term “self-knowledge” is misleading in a number of different ways: we are not talking here about knowledge of the self, but of the special kind of epistemic access that we have to our own states of mind, and events in our mind. Hence, the awareness we have of our own sensations and experiences is also a form of “self-knowledge” in this sense. Also, it is questionable whether such awareness is itself knowledge, or whether it is simply something that makes knowledge possible. (Compare: is perception itself knowledge, or is it something that makes knowledge possible?) Nonetheless, these qualifications made, we can take “self-knowledge” in the contemporary sense to be a term for the distinctive capacity to access one’s own mind, as opposed to the access one has to the minds of others. In the first kind of case, of the capacity is exercised in coming to believe something (making up one’s mind). In the second kind of case, it is exercised in our spontaneous access to what we believe anyway. And in the third kind of case, the capacity is exercised in our ability to introspect and examine our own beliefs. But it is important to emphasise that not all thoughts (in my sense) are judgements. Consciously thinking something can be pursued without the desire to commit to the truth of any particular proposition thought. One could simply be considering something, without needing or wanting to make up one’s mind. Such speculation takes place against a background of belief, of course; but not all of these beliefs need to be brought to mind in order for one to speculate. Conscious thoughts, then, are episodes. My claim is that without employing the idea of a phenomenally conscious episode, it is not possible to make sense of what underlies the talk of “conscious belief” or “access (A) consciousness.” It is true that information could be used by a cognitive system, and hence “accessed” even at the level of the information-processing in the brain. But there is no reason to call this kind of “accessing” a kind of consciousness. The only point of saying that these states are A-conscious, as far as I can see, is to mark the fact that they can manifest themselves in consciousness. And the same applies to the talk of “conscious belief.” We are now in a position to answer the question with which I began this chapter: are thoughts and sensations/sensory experiences conscious in the same way? Is there a single sense of “conscious” in which all these otherwise very different mental phenomena are conscious? My answer is that they are conscious in the same way, because thoughts and sensory experiences are episodes or events in the “stream” of a subject’s consciousness. Of course, this is not a definition of the sense in which they are conscious, because it appeals to

Unconscious Belief and Conscious Thought

171

the idea of a stream of consciousness. But the image of the stream is supposed to tell us something significant about the structure of conscious phenomena. It is essential that thoughts or episodes of thinking, like sensations and sensory experiences, are occurrences: they are events or processes that have a particular temporal extent and duration. This is compatible with sensations and thoughts differing in all sorts of other ways (for example, in the extent to which it is intelligible to think of them as subject to constraints of reason or rationality). But what I have been arguing for here is the fundamental similarity between them, rather than any differences. Finally, we have seen that the other thing sensations and conscious thoughts have in common is that they can both be known by the exercise of our capacity for self-knowledge—or, as I would prefer to say, they are both available to the subject through privileged access. Why this is, and what the exact connection is between non-conscious states of mind like beliefs and our capacity to know these states with privileged access, are questions which must be left to another occasion.

5. Conclusion: Phenomenal Intentionality and Conscious Thought What are the implications of the answer to our question for the idea of phenomenal intentionality? The first implication is that if we want to look for phenomenal intentionality outside the less controversial examples (for example, perceptual experience), we should not be asking ourselves whether beliefs exhibit a phenomenal character of their own. If phenomenal character is the character of phenomenal episodes, and these are episodes in the stream of consciousness, then since beliefs are not episodes in the stream of consciousness, beliefs can never have phenomenal character. But thoughts—in the sense characterized in §4—do have phenomenal character, since they are episodes in the stream of consciousness. The next question is whether thoughts have phenomenal character of their own: that is, phenomenal character which is not the character of some other kind of episode. It is true that some thoughts are accompanied by imagery or words (talking to yourself, words running through your mind, images coming to you). But part of the phenomenal character of the experience is something that cannot be conveyed by these sensory or linguistic representations: it is the attitudinal component of the thought. Maybe the “vehicle” of the content of a conscious thought is words running through your mind; but the attitude (judging, hoping, desiring, etc.) cannot be. There need be no words which correspond to the attitude when you judge that p rather than manifest hope that p: you do not have to verbally articulate your attitude in order for the thought

172

phenomenal intentionality

to involve that attitude. To develop this argument properly would be the next stage in the defence of the thesis that conscious thoughts have a phenomenal intentionality to call their own. But the first stage is to get a correct view of the metaphysics, or apparent metaphysics, of phenomenal intentionality. This is what I have attempted to do here.

Acknowledgments The ideas in this chapter were first aired at an ESF-sponsored conference in Sardinia, as part of the EUROCORES “Consciousness in Its Natural and Cultural Context” (CNCC) initiative. I’m very grateful to Pierre Jacob, Barry C Smith, and Till Vierkant for helpful discussion on that occasion, and to subsequent audiences at King’s College, London (Institute of Psychiatry, Philosophy Group); Oxford Brookes University and the University of Durham; the RIP conference in Edinburgh in honour of Timothy Sprigge; and the Conference in Liverpool on the work of Barry Dainton. Special thanks to Katalin Farkas and to an anonymous reader for Oxford University Press for their very useful suggestions; and thanks to Mike Martin and Matt Soteriou for discussion of these matters.

Notes 1. A distinction is sometimes made between dispositional belief and a disposition to believe (see, for example, Audi 1994). Th is is a useful distinction but it is not directly relevant to what I want to argue here. 2. Th is is a claim McDowell went on to reject: see his 2009. 3. We should ignore here Russell’s (1910–11) notion of “knowledge by acquaintance,” which is supposed to be an occurrent mental act. It is arguable how much it has to do with the ordinary concept of knowledge: see Crane 2011. 4. David-Hillel Ruben remarks, of the term “occurrent belief” that “I know of no plausible characterisation of what is intended by this term, except the one that ties it to present awareness or consciousness” (Ruben 2003: 114); and he further writes that “many, like Alvin Goldman, thought of ‘occurrent’ as I do, as elliptical for ‘occurrent to present consciousness” (Ruben 2003: 116). 5. I am indebted here to discussions with Mike Martin and Matt Soteriou: see Soteriou 2007.

References Audi, Robert. (1994). “Dispositional Belief and Dispositions to Believe.” Nous 28: 419–434. Bayne, Tim, and Michelle Montague, eds. ( 2011). Cognitive Phenomenology. Oxford: Oxford University Press. Block , Ned. (1996). “Mental Paint and Mental Latex.” Philosophical Issues 7: 19–49.

Unconscious Belief and Conscious Thought

173

Block, Ned. (2002). “Some Concepts of Consciousness.” In David Chalmers, ed., Philosophy of Mind: Classical and Contemporary Readings. Oxford: Oxford University Press. BonJour, Laurence. (1999). “The Dialectic of Foundationalism and Coherentism.” In John Greco and Ernest Sosa, eds., Blackwell Guide to Epistemology. Oxford: Blackwell. Carruthers, Peter. (2003). Phenomenal Consciousness. Cambridge: Cambridge University Press. Cassam, Quassim, ed. (1994). Self-Knowledge. Oxford: Oxford University Press. Crane, Tim. (2001). Elements of Mind. Oxford: Oxford University Press. Crane, Tim. (2011). “Tye on Acquaintance and the Problem of Consciousness.” Philosophy and Phenomenological Research 84: 190–198. Davidson, Donald. (1967). “Causal Relations.” Journal of Philosophy 64: 691–703. Dretske, Fred, and Palle Yourgrau. (1983). “Lost Knowledge.” Journal of Philosophy 80: 356–367. Gennaro, Rocco J. (1996). Consciousness and Self-Consciousness: A Defense of the Higher-Order Thought Theory of Consciousness. Amsterdam: John Benjamins. Hill, Christopher. (2009). Consciousness. Cambridge: Cambridge University Press. Kim, Jaegwon. (1976). “Events as Property Exemplifications.” In Myles Brand and Douglas Walton, eds., Action Theory. Dordrecht: Reidel, 159–178. Kriegel, Uriah. (2004). “Moore’s Paradox and the Structure of Conscious Belief.” Erkenntnis 61: 99–121. Kriegel, Uriah (2007). “Intentional Inexistence and Phenomenal Intentionality.” Philosophical Perspectives 21: 307–340. Lewis, David. (1986).“Events.” In Philosophical Papers, Volume II. Oxford: Oxford University Press, 241–269. Lycan, W. G. (2001). “A Simple Argument for a Higher-Order Representation Theory of Consciousness.” Analysis 61: 3–4. McDowell, John. (1994). Mind and World. Cambridge, MA : Harvard University Press. McDowell, John. (2009). “Avoiding the Myth of the Given.” The Engaged Intellect Cambridge, MA : Harvard University Press, 256–274. Mellor, D. H. (1981). Real Time. Cambridge: Cambridge University Press. Oddie, Graham. (2005). Value, Reality and Desire. Oxford: Oxford University Press. Peacocke, Christopher. (1993). A Study of Concepts. Cambridge, MA : MIT Press. Pitt, David. (2004). “The Phenomenology of Cognition or What Is It Like to Th ink Th at P?” Philosophy and Phenomenological Research 69: 1–36. Rosenthal, David. (1986). “Two Concepts of Consciousness.” Philosophical Studies 49: 329–359. Ruben, David-Hillel. (2003). Action and Its Explanation. Oxford: Oxford University Press. Russell, Bertrand. (1910–11). “Knowledge by Acquaintance and Knowledge by Description.” Proceedings of the Aristotelian Society XI: 108–128. Searle, John. (1992). The Rediscovery of the Mind. Cambridge, MA : MIT Press. Soteriou, Matthew. (2007). “Content and the Stream of Consciousness.” Philosophical Perspectives 21: 543–568. Strawson, Galen. (1995). Mental Reality. Cambridge, MA : MIT Press. Tye, Michael (1995) Ten Problems of Consciousness. Cambridge, MA: MIT Press WIlliamson, Timothy. (2000). Knowledge and Its Limits. Oxford: Oxford University Press.

10

Intellectual Gestalts E l i ja h C h u dnof f

Introduction By a phenomenal character I mean a property that types experiences by what they are like for their subjects. Two experiences have the same phenomenal character just in case what it is like for a subject to have one is the same as what it is like for a subject to have the other. There are three different theses about phenomenal characters that will figure in this chapter. Here are first pass articulations of them: Phenomenal Intentionality: (PI) Some phenomenal characters can only be instantiated by experiences with certain intentional contents. Cognitive Phenomenology: (CP) Some phenomenal characters can only be instantiated by experiences that are not purely sensory. Phenomenal Holism: (PH) Some phenomenal characters can only be instantiated by experiences that are parts of certain wholes. Though all three theses will figure in this chapter, they will play different roles. PI will be an assumption. I will take it for granted that one way to discuss the phenomenal character of an experience is to discuss its intentional content— taking it as understood that this is intentional content fi xed by phenomenal character. PH will be my focus. I will defend phenomenal holism by considering some examples that motivate it and responding to various worries one might have about it. CP will be a beneficiary. After discussing PH, I will show how to use it to defend CP. The plan of the chapter is this. In Section 1, I will consider some of the historical background to PH and in particular connect it to one strand— the phenomenological strand, not the psychological strand—in the gestalt theoretical tradition. In Section 2, I will argue that certain kinds of sensory

174

Intellectual Gestalts

175

experiences—sensory gestalts—are phenomenally holistic, and so motivate PH, and I will defend PH against some objections to it. In Section 3, I will argue that there are intellectual experiences—intellectual gestalts—that are also phenomenally holistic. And in Section 4, I will explain why I think my defense of PH, and in particular my extension of it to intellectual experiences, provides resources for defending CP.

1. Gestalt Psychology and Gestalt Phenomenology The main motivation for endorsing phenomenal holism derives from reflection on gestalt experiences—both sensory and intellectual. This raises the question: what is the relationship between phenomenal holism and gestalt theory? The question, however, is ambiguous. There are at least two sorts of gestalt theory. One is psychological, the other is phenomenological. Gestalt psychology and phenomenal holism are logically independent. Gestalt phenomenology presupposes phenomenal holism. Gestalt psychology is a theory about psychological explanation. Here is Wertheimer’s famous summary: The fundamental “formula” of Gestalt theory might be expressed this way: There are wholes, the behavior of which is not determined by that of their individual elements, but where the part-processes are themselves determined by the intrinsic nature of the whole.1 Wertheimer makes two claims. The fi rst is negative: there are mental states or events (wholes) certain properties of which are not explained by their composition out of certain other mental states or events (parts) that have certain properties. The second is positive: there are mental states or events (parts) certain properties of which are explained by the role they play in composing other mental states or events (wholes) that have certain properties. These claims could use further elaboration: the negative claim is trivial if there are mental states or events that lack parts; the positive claim is trivial if among the “certain properties” of parts is the property of being a part of a whole. The Gestalt psychologists did not bother about formulating principles immune to such worries. Their main agenda was to develop psychological explanations. And Wertheimer’s aim in the quoted passage was to highlight a certain feature of the kinds of explanations they pursued: the explanations are what we might call downward psychological explanations— they explain the properties of parts by the properties of the wholes those parts compose.

phenomenal intentionality

176

To fi x ideas, here is an example (Figure 10.1).2 Look at figures A and B:

(a)

(b)

Figure 10.1

Here are two facts about the visual experiences subjects typically have in looking at these figures: (1) In A, the diamond looks like an area where two hexagons overlap—not like a figure inscribed in an 11-sided polygon. (2) In B, the diamond looks like a figure inscribed in a single hexagon—not like an area where two hexagons overlap. On reflection and further examination, one might very well see the diamond in A as a figure inscribed in an 11-sided polygon, and the diamond in B as an area where two hexagons overlap. But this is not what typically happens. Why? The explanation that Wertheimer suggests invokes the Law of Prägnanz—that “psychological organization will always be as ‘good’ as the prevailing conditions allow,” where “the term ‘good’ is undefined,” but “embraces such properties as regularity, symmetry, simplicity.”3 Here is a sketch of how one might appeal to this law in explaining the typical responses to figures A and B: two overlapping hexagons are simpler than one 11-sided polygon, and so by the Law of Prägnanz that is how our visual experience organizes the stimuli in A; one hexagon is simpler than two, and so by the Law of Prägnanz that is how our visual experience organizes the stimuli in B. Set aside the question of whether this is a good explanation and whether the Law of Prägnanz is a genuine psychological law. The point of the example is to illustrate the sort of downward psychological explanation that the Gestalt psychologists favored: a property of the whole—greater simplicity in organization—explains a property of the part—the way the diamond looks. Gestalt phenomenology is not a theory about psychological explanation. It is a theory about psychological individuation, and in particular about the individuation of experiences. Psychologists such as Wertheimer, Koff ka, and Köhler endorsed gestalt phenomenology in addition to gestalt psychology. But the philosopher Aron Gurwitsch did the most to promote gestalt phenomenology. Here is a passage in which he discusses its principal tenet: [a] It is the functional significance of any part of a Gestalt-contexture that makes this part that which it is. The part is what it is only as a

Intellectual Gestalts

177

constituent of the Gestalt-contexture and as integrated into its unity. Any part of a Gestalt may then be said to be determined as to its existence by its functional significance in the sense that the part only exists in, and is defined by, its functional significance. [b] Properties and characters which qualify any constituent of a Gestalt as that which it is in a concrete case, belong to it on account of its functional significance, and of its integration into the Gestalt-contexture. Such determinations belong to the part in question only insofar, and as long, as it is thus integrated.4 I’ve split the passage into two parts, [a] and [b]. In part [a] Gurwitsch says that there are some part experiences that are metaphysically dependent on—that can only exist in—whole experiences; gestalts are structured whole experiences that have such metaphysically dependent parts. In part [b] Gurwitch says that there are some “characters,” that is, phenomenal characters, that an experience can have only insofar as it is part of a certain whole. Parts [a] and [b] fit together if we make the assumption that experiences have their phenomenal characters essentially. This is a plausible assumption, and Gurwitsch does seem to embrace it. On the other hand, he—and other proponents of gestalt phenomenology—do sometimes speak as if there are part experiences that feel one way in one whole experience and would have felt another way in another whole experience. Return, for example, to the visual experiences we typically have in looking at figures A and B—call them experience A and experience B. Suppose experience A actually occurs and experience B might have occurred, and focus on the diamond-presenting part of experience A. Consider three different claims about this part: (1) It could have been part of B (2) If it were part of B, it would have had a different phenomenal character (3) It has its phenomenal character essentially These three claims are mutually inconsistent. I will assume that (3) is non-negotiable; (2) is plausible: were the diamond-presenting part a part of experience B it would have represented the diamond as an inscribed figure, not a region of overlap. So the claim that should be given up is (1). It is not the case that there is a partial experience that while actually a part of experience A could have been a part of experience B. The most we should say is, had experience B occurred, it would have had a diamond-presenting part, and this part would have been phenomenally similar—phenomenally similar in phenomenally non-holistic ways—to the diamond-presenting part of experience A. Both gestalt psychologists and gestalt phenomenologists think that there are special kinds of experiences—gestalt experiences. According to gestalt

178

phenomenal intentionality

psychologists, gestalt experiences are special because they are structured wholes, the facts about which explain the facts about their parts. According to gestalt phenomenologists, gestalt experiences are special because they have parts that are phenomenally and so metaphysically dependent on them. The two views are logically independent: a gestalt psychologist might think that gestalts bear explanatory, not individuative, relations to their parts; and a gestalt phenomenologist might think that gestalts bear individuative, but not explanatory, relations to their parts. For the same reason, gestalt psychology is logically independent of phenomenal holism. Finally, it should be clear from the foregoing that gestalt phenomenology presupposes phenomenal holism: gestalt phenomenology entails phenomenal holism, and, further, it is phenomenal holism and essentialism about phenomenal character that provide the best account of why gestalt phenomenology should be accepted.

2. Sensory Gestalts Gestalt experiences are structured whole experiences whose parts seem to have their phenomenal characters—and perhaps other of their psychological properties—because of the role they play in composing the whole. One might agree that there are gestalt experiences but reject phenomenal holism, gestalt psychology, and gestalt phenomenology: one might not take what seems to be the case about gestalt experiences at face value, or one might develop an alternative theoretical account of it. In this section I will examine sensory gestalt experiences, argue that they motivate phenomenal holism about the phenomenal characters of their parts, and defend phenomenal holism against a range of objections. Experiences A and B discussed in the previous section are examples of sensory gestalts. The diamond-presenting part of experience A represents the diamond as a region of overlap. The diamond-presenting part of experience B represents the diamond as an inscribed figure. These representational differences are—I will assume, hereby invoking PI—phenomenal differences.5 Further, the diamond-presenting parts of the visual experiences seem to have their phenomenal characters because of the role they play in composing the respective whole visual experiences. So much is fairly neutral ground. Phenomenal holists think that the diamond-presenting parts of the two visual experiences have phenomenal characters that partial visual experiences can have only insofar as they play respectively similar roles in composing respectively similar whole visual experiences. What reason is there for making this additional claim? The main reason is that it is impossible to imagine a visual experience having the phenomenal character of the diamond-presenting part of experience A, or the diamond-presenting part of experience B, without also

Intellectual Gestalts

179

being part of a whole visual experience that is largely similar to experience A, or a whole visual experience that is largely similar to experience B. I think we should take the impossibility of imagining such partial visual experiences as evidence that there can be no such partial visual experiences. And this is precisely the phenomenal holist’s point. Consider some other examples. Compare the following figures (see Figure 10.2).

(c)

(d)

(e)

Figure 10.2

Compare the way the pie in figure C looks with the way the upper left pie in figure D looks. In figure C the pie looks like a pie with a wedge cut out. In figure D the pie looks like a pie that is partially occluded by a triangle. Now compare the triangle in figure D with the triangle in figure E. The triangle in figure D looks like it hovers above three pies. The triangle in figure E looks like it is cut out of a black patch. The differences I have pointed out are phenomenal differences. Further, they seem to derive from the role the respective partial experiences play in composing the whole experiences to which they belong. Let us just focus on figure D. According to phenomenal holists, the upper-left-pie-presenting part of our visual experience of figure D has a phenomenal character that only partial visual experiences that play similar roles in similar whole visual experiences can have—and the same goes for the triangle-presenting part of our visual experience. The main reason for endorsing the phenomenal holist view, again, is the impossibility of imagining visual experiences having the same phenomenal characters by themselves or as parts of very different whole visual experiences. This is the intuitive case for phenomenal holism. Now I will try to strengthen the case for phenomenal holism by responding to a few objections to it.6 Barry Dainton distinguishes between the following two theses: Strong Impingement: (SI) Phenomenal wholes have certain parts that possess intrinsic phenomenal features that reflect the character of that whole, and parts with the same character could not possibly occur except in a whole of the same or similar type. Weak Impingement: (WI) The character of the constituent parts of a phenomenal whole are partly dependent on their being such, but items with just the same intrinsic phenomenal characters as these parts could exist in wholes of a different type, or as perceived wholes in their own right.7

180

phenomenal intentionality

SI is more or less equivalent to PH; WI is incompatible with PH. Dainton argues that many putative illustrations of SI, and so PH, are at most illustrations of WI. One example he considers is the following (see Figure 10.3):

(f)

(g)

Figure 10.3

The triangle in F might appear to point in any of three directions. The left-most triangle in G also might appear to point in any of three directions, but there is a tendency for it to appear to point along the axis of symmetry of the whole of figure G. Consider two experiences: an F-experience representing the triangle in F as pointing rightward, and a G-experience representing the left-most triangle in G as pointing along the axis of symmetry of the whole of figure G—that is, rightward. It is natural to say of this G-experience that the axis of symmetry determined by the whole of figure G contributes to making the left-most triangle in G appear to point rightward. Dainton argues, however, that this is an example of at most weak impingement: the F-experience represents the triangle in F just as the G-experience represents the left-most triangle in G, and it does so even though the whole F-experience is quite different from the whole G-experience. Dainton’s observation fails to show that our G-experience of the left-most triangle in figure G is an example of weak rather than strong impingement. Focus on the triangle-presenting part of our F-experience and the left-most triangle-presenting part of our G-experience. Suppose—conceding Dainton’s observation—that these partial experiences have the same propositional content, roughly, that there is an equilateral triangle pointing rightward. Our partial F-experience and our partial G-experience might, nonetheless, possess different phenomenal characters. How? One way is for there to be differing non-intentional aspects of their phenomenal characters. But set this possibility aside. Another way is for the modes of presentation under which each partial experience presents the proposition that there is an equilateral triangle pointing rightward to be different. Take the property of pointing rightward. Both our partial F-experience and our partial G-experience represent the property of pointing rightward. But they might do so under different modes of presentation. One way this might happen is for the two experiences to use different frames of reference. The partial F-experience might represent the property of pointing rightward as the property of pointing at 3 o’clock on the circle, and the partial G-experience might represent the property of pointing rightward as the property of pointing along the axis of symmetry of the whole of

Intellectual Gestalts

181

figure G.8 Because of these possibilities, it follows that Dainton’s premise—that our partial F-experience and partial G-experience has the same propositional content—does not adequately support thinking that our partial F-experience and partial G-experience have the same phenomenal character, and so does not adequately support his conclusion—that our G-experience is an example of weak impingement, rather than strong impingement. Similar considerations apply to attempts to extend Dainton’s argument to cover other cases, such as our experiences of figures A through E. Bayne, for example, argues that our experience of figure E is at most an example of weak impingement.9 He writes: Cut around “the white triangle” that you see on this page and superimpose it on almost any solid background: aren’t you now having a type-identical white triangle experience in a markedly different phenomenal context? I’ve done what Bayne suggests in presenting figure E. And he is correct: we have a type-identical white triangle experience in a markedly different phenomenal context. The relevant type under which the two experiences are identical is this: they both represent white triangles. Even so, the white triangle experiences might be, and plausibly are, phenomenally different. First, Bayne—like Dainton—has not ruled out the possibility that even though both white triangle experiences represent some of the same properties, such as whiteness and triangularity, they do so under different modes of presentation. Second, Bayne’s argument suffers from an additional flaw: the white triangle experiences do not have the same propositional content, since the white triangle in our experience of figure D hovers and the white triangle in our experience of figure E does not. There is a second sort of challenge that Dainton presses against certain putative examples of strong impingement, and so phenomenal holism.10 Consider the following instructions: Focus on the diamond-presenting part of your experience of figure A. Take its phenomenal character. Now try to imagine that phenomenal character instantiated by experiences that occur in isolation or as parts of significantly different whole experiences. I claim that when you try to do this, you fail. And I claim, further, that this gives you a reason to think the phenomenal character of the diamond-presenting part of your experience of figure A is holistic. Applied to this bit of argumentation, Dainton’s second sort of challenge can be framed like this: The phenomenal character that you wind up isolating is the phenomenal character of your whole experience. What you find impossible

182

phenomenal intentionality

to imagine is the phenomenal character of your whole experience of figure A instantiated by significantly different whole experiences that also happen to have a diamond-presenting part. But if this is what you cannot imagine, your failure to do so does nothing to suggest phenomenal holism: rather it suggests that the phenomenal characters of two whole experiences can only be the same if they share phenomenally identical or similar parts. Where, however, do we fail to follow the instructions, according to Dainton? We might fail to focus on the diamond-presenting part of our experience of figure A, and focus on our whole experience of figure A instead. But it is not clear why we should fail in this way. Dainton has provided no reasons to be skeptical about our ability to focus on partial experiences rather than whole experiences. And doing so in this case seems fairly simple: we can pick out the relevant part of our experience by the description “the part of our experience that presents a diamond shape.” Suppose, then, that we succeed here, that is, we get the part, rather than the whole in mind as our object of attention. Then where we might fail is in passing from this to the part’s phenomenal character rather than the whole’s phenomenal character. That is, while having the partial experience in mind and trying to isolate its phenomenal character, we might fail to do so and, rather, isolate the phenomenal character of the whole experience instead. Again, it is not clear why we should fail. We can just add to our description and form the description “the phenomenal character of the part of our experience that presents a diamond shape.” What these considerations show is that it is a simple matter to isolate in thought both partial experiences and their phenomenal characters. What might not be such a simple matter, however, is to isolate partial experiences and their phenomenal characters in imagination. But what could isolating partial experiences and their phenomenal characters in imagination be other than imagining partial experiences and their phenomenal characters in isolation? If this is what we cannot do, according to Dainton, then he is simply granting the major premise in the argument for phenomenal holism.

3. Intellectual Gestalts In this section I want to present some examples that suggest that there are intellectual experiences whose parts instantiate holistic phenomenal characters. I should first say something about intellectual experiences. I do not know how to define what an intellectual as opposed to a sensory experience is. Here are some examples: intuiting that circles are symmetrical about their diameters, grasping a proof of the Pythagorean theorem, deciding

Intellectual Gestalts

183

to bike rather than walk to work, understanding what some passage is about. Let me illustrate the last with an example that I am quite fond of. Consider the following passage: A newspaper is better than a magazine. A seashore is a better place than the street. At first it is better to run than to walk. You may have to try several times. It takes some skill but it is easy to learn. Even young children can enjoy it. Once successful, complications are minimal. Birds seldom get too close. Rain, however, soaks in very fast. Too many people doing the same thing can also cause problems. One needs lots of room. If there are no complications it can be very peaceful. A rock will serve as an anchor. If things break loose from it, however, you will not get a second chance.11 The typical response is to find this passage incomprehensible. Now let me give you the key word: kite. Re-read the passage. Now when you read it, you should have a felt sense of understanding what it is about. This new experience which was missing at first and which is now present is the experience of understanding what some passage is about. The claim that there are intellectual experiences should be distinguished from CP, the cognitive phenomenological thesis that some phenomenal characters can only be instantiated by experiences that are not purely sensory. I take the claim that there are intellectual experiences to be uncontroversial. There just is this experience of understanding what the kite passage is about, and it is—in a perfectly ordinary, even if so far undefined—sense non-sensory. What is controversial is that this experience of understanding possesses a phenomenal character that no purely sensory experience can possess. One might agree that the experience of understanding what the passage is about is intellectual, but think that it is possible to have a sensory experience with the same phenomenal character. One might argue that all you have to do is have a sensory experience as of hearing the words of the passage in your mind’s ear and also seeing various kites in your mind’s eye. The idea is that this experience is a sensory experience with the same phenomenal character as the intellectual experience of understanding what the passage is about. Whatever differences between the two experiences there are must be non-phenomenal: perhaps when you read the passage with understanding you form certain beliefs, which you do not form when you only have the sensory experience.12 Th is is not a view that I endorse, but it is a view that someone who rejected CP might endorse. What I want to focus on now are experiences of grasping a proof. Consider the following proof that the sum of the first n positive integers is half of n × (n + 1).

phenomenal intentionality

184

Proof: The first n positive integers can be represented by a triangular array of dots, as in the first diagram (see Figure 10.4): 1

2

3

4

n

n

n+1

Figure 10.4

Two of these triangular arrays can be fit together to form a rectangular array containing n × (n + 1) dots, as in the second diagram. Each triangular array is half of the rectangular array. So, the sum of the first n positive integers is half of n × (n + 1). When you consider this proof you have an intellectual experience with a certain phenomenal character. Part of this intellectual experience is a certain visual experience—a visual experience that is similar to the visual experience you might have of a mere array of dots presented in Figure 10.5).

Figure 10.5

Though these two visual experiences—the visual experience of the mere array and the visual experience of the array as part of the proof—are similar, there are, I find, phenomenal differences between them. That is, the visual experience that I have as part of my intellectual experience of grasping the proof seems different from the visual experience that I have independent of my intellectual experience of grasping the proof. I am not sure how best to articulate the differences, but here are some things I would say about how I represent the array in the proof that I wouldn’t say about how I represent the mere array: it seems meaningful, open-ended, representative of something else, a portion of something larger. The visual experience of the array that occurs as part of the intellectual experience of grasping the proof has a distinctive phenomenal character. So much should be evident to naïve introspection. I want to claim, further, that

Intellectual Gestalts

185

the phenomenal character that this visual experience instantiates is holistic. That is, it is a phenomenal character that a visual experience can instantiate only as part of a largely similar intellectual experience. The main reason I have for thinking this is that I cannot imagine having a visual experience that represents an array of dots in just the way that the visual experience I have as part of grasping the proof does, but that occurs in isolation—as does the visual experience of the mere array with which we started—or that occurs as part of a significantly different experience, whether sensory or intellectual. There might be a range of alternative intellectual experiences—experiences of grasping different proofs for different theorems, for example—within which a similar visual experience of a similar array of dots might occur. It is not necessary to circumscribe the class of alternative intellectual experiences. What I want to emphasize, however, is the plausibility of the claim that there is no alternative purely sensory experience within which a similar visual experience of a similar array of dots might occur. The phenomenal characteristics I tried to pick out by talking about how the array seemed meaningful, open-ended, representative of something else, and a portion of something larger plausibly depend essentially on the surrounding intellectual context within which the visual experience of the array occurs. I do not see how a mere sensory experience can endow the visual experience of the array with these phenomenal characteristics. The example considered is one in which the phenomenal character of a sensory experience is holistically dependent on the intellectual experience of which it is a part. There are also cases in which the phenomenal character of an intellectual experience is holistically dependent on the larger intellectual experience of which it is part. Take, for example, the proposition that the first n positive integers can be represented by a triangular array of dots. You might have a thought with this propositional content in isolation. Or you might have it as part of grasping the proof that the sum of the first n positive integers is half of n × (n + 1). I conjecture that the two occurrences will differ in phenomenal character. This is only a conjecture because, unfortunately, I haven’t ever had the thought that the first n positive integers can be represented by a triangular array of dots just pop into mind in isolation. But I have had other thoughts just pop into mind in isolation. I know what that is like. And I can imagine the thought that the first n positive integers can be represented by a triangular array of dots popping into mind in a similar fashion. When I do so, it seems to me that if the thought were to pop into mind in this way, the thought episode would be experienced differently from the thought episode that occurs when I have the thought that the first n positive integers can be represented by a triangular array of dots as part of grasping the proof that the first n positive integers is half of n × (n + 1). Here is a worry. I claim that the thought that the first n positive integers can be represented by a triangular array of dots is experienced in a certain way that

186

phenomenal intentionality

depends on its occurrence as part of grasping the proof that the first n positive integers is half of n × (n + 1). Note, however, that the thought occurs early on in the experience of grasping the proof. So, according to my claim, an earlier experience—the thought—depends on later experiences—the parts of the whole experience of grasping the proof that occur after the thought. But how can an earlier experience depend on later experiences? A similar worry will arise about any temporally extended gestalt experience, whether it is sensory or intellectual. One of the early examples of a sensory gestalt was the experience of hearing a melody: it was often claimed that an auditory impression of an individual note feels differently depending on whether it occurs in isolation or as part of an experience of hearing a melody.13 Call the proposition that the first n positive integers can be represented by a triangular array of dots P. And let us contrast three cases. Case 1: you work through the entire proof and in doing so think that P. Case 2: you work through part of the proof—you think that P, but then black out before you can finish the proof. Case 3: the thought that P just pops into your mind. Here is a puzzle. (1) The Case 1 thought that P is phenomenally identical to the Case 2 thought that P (2) The Case 2 thought that P is phenomenally identical to the Case 3 thought that P (3) The Case 1 thought that P is phenomenally distinct from the Case 3 thought that P I am committed to (3) for the reasons presented earlier. (1) is plausible because Case 1 and Case 2 can be imagined so that you are in the same brain states in each case up to the time you black out, and it is plausible that if you are in the same brains states in two cases then you are in the same phenomenal states in those cases.14 (2) is plausible because the thought that P comes first when you are working through the proof, and if you make it no further, then it is plausible that you are having a thought which is like the thought you would have were you to just have the thought that P, say because it pops into your mind. Something has to give. Suppose we keep (3). Then we have two options. The radical option is to give up (1). This is radical in part because it commits one to the phenomenal externalist view that phenomenal character fails to supervene on brain states. But it is even more radical than that. It is more radical because it commits one to the view that phenomenal character at a time t fails to supervene on all non-phenomenal states—whether they are states of the brain or states of the environment—up to and including time t. The more conservative option, then, is to give up (2). This is what I recommend. In order to render the denial of (2) plausible we must find some feature of Case 2 that differentiates it from Case 3. And there is an obvious one: Case 2 is a case in which you attempt

Intellectual Gestalts

187

to work through the proof; Case 3 is not. Why should that matter? Plausibly, it matters because when you have the thought that P in Case 2 you do so in the context of various other mental states, such as intentions and anticipations, connecting it to the future unfolding of your effort to grasp the proof. The Case 2 thought that P, then, does have its phenomenal character partly because of the role it plays in a larger intellectual experience of grasping the proof, but this larger intellectual experience bears on the thought that P only insofar as it is represented in mental states that are contemporaneous with the thought that P. This seems to me the most promising way to respond to the puzzle about temporally extended gestalt experiences.15

4. Cognitive Phenomenology Proponents of CP, the cognitive phenomenological thesis that some phenomenal characters can only be instantiated by experiences that are not purely sensory, have generally given two sorts of argument for their view. First, there are epistemological arguments. They tend to look like this: You can tell by introspection, or self-awareness, or inner-sense that you are hearing, or at least seeming to hear, a fire engine’s siren rather than a dog’s barking. The way this works is that hearing a fire engine’s siren has a distinctive phenomenal character, different from the phenomenal character of hearing a dog’s barking. Similarly, you can tell by introspection, or self-awareness, or inner-sense that you are thinking that idle hands are the Devil’s workshop, rather than that misfortunes come on wings and depart on foot, or rather than having the auditory experience of hearing someone utter, “Idle hands are the Devil’s workshop.” Again, the way this works is that thinking that idle hands are the Devil’s workshop has a distinctive phenomenal character, different from the phenomenal character of thinking that misfortunes come on wings and depart on foot, and different from the phenomenal character of hearing someone utter, “idle hands are the Devil’s workshop.” The basic idea is that in order for us to have the kind of self-knowledge we have about our conscious thoughts, there must be phenomenal characters that can only be instantiated by those thoughts.16 I find this strategy for motivating CP attractive, but I will not discuss it further here. I want, rather, to focus on the second sort of argument—arguments by example.17 They tend to look like this: Example (1) Consider the felt difference between (a) hearing “Dogs dogs dog dog dogs” as a mere list of words and (b) hearing it as an

188

phenomenal intentionality

English sentence. Example (2) Consider the felt difference between (a) hearing “He is heading toward the bank” as entailing that he is heading toward a financial institution and (b) hearing it as entailing that he is heading toward a part of a river. All of the phenomenal characters that can be instantiated by purely sensory experiences are the same in the (a) cases and the (b) cases. But there are phenomenal differences between the (a) cases and the (b) cases. So these differences must be differences owed to the absence in the (a) cases and the presence in the (b) cases of phenomenal characters that cannot be instantiated by purely sensory experiences. Philosophers hostile to CP have argued against the claim that all of the phenomenal characters that can be instantiated by purely sensory experiences are the same in the (a) cases and the (b) cases. One might argue, for example, that the phenomenal differences between hearing “Dogs dogs dog dog dogs” as a mere list and as an English sentence are due to differences in perceptual grouping. One might argue, further, that the phenomenal differences between hearing “He is heading toward the bank” as about a financial institution or as about a part of a river are due to differences in associated mental imagery.18 The aim of this section is to explore some ways in which our previous reflections on phenomenal holism might bolster the case for cognitive phenomenology. The gestalt experiences that motivate phenomenal holism expand the range of examples that can be used to motivate cognitive phenomenology. These new examples are, I will argue, immune to objections that opponents of cognitive phenomenology have made against the standard examples. I will call a phenomenal character that can be instantiated by a purely sensory experience a sensory phenomenal character, and I will call a phenomenal character that cannot be instantiated by a purely sensory experience a cognitive phenomenal character. Arguments by example purport to motivate CP by exhibiting pairs of experiences that are the same in their sensory phenomenal character but differ in some phenomenal character, and so differ in their cognitive phenomenal character. The standard pairs of experiences used to motivate CP are experiences of some change in linguistic understanding—from none to some, or from one kind to another. Arguments by example that appeal to felt differences in linguistic understanding invite a range of objections, deriving from the fact that changes in linguistic understanding are often associated with differences in sensory phenomenal character, such as differences in perceptual grouping of words or other elements of the representation understood, or differences in accompanying mental imagery. It has proved difficult to find felt differences in linguistic understanding that are immune to such objections. The phenomenal character of linguistic understanding, however, is not the only cognitive phenomenal character there is. I think that proponents of CP

Intellectual Gestalts

189

should look for other examples to support their view. What I will suggest now is that the pair of visual experiences of arrays of dots discussed earlier is a good example in that it is immune to the objections that beleaguer experiences of linguistic understanding. Recall: there is a visual experience of the array of dots that you have in isolation—call it (a); and there is a visual experience of the array of dots that you have as part of grasping the proof that the sum of the first n positive integers is half of n × (n + 1)—call it (b). I claim that the (a) experience of the array and the (b) experience of the array are the same in their sensory phenomenal character. There is a felt difference between them, however. As I have reported, the (b) experience of the array represents it as meaningful, open-ended, representative of something else, a portion of something larger. These are merely suggestive phrases. The important point is that the (b) experience of the array of dots takes on a different phenomenal character because of the role it plays in a larger intellectual achievement of grasping a proof. Further, as argued earlier, it is plausible that an experience of an array of dots must play such a role in order to take on that distinctive phenomenal character exhibited in the (b) experience. So what we have here is a pair of experiences apt to occur in an argument by example for cognitive phenomenology.19 The objections to arguments by example that appeal to felt differences in linguistic understanding do not apply here. For one, there are no differences in the way the (a) experience and the (b) experience group the dots. And two, these experiences themselves are imagistic, so, while there is mental imagery in the (b) case, it is supposed to be there, and poses no problem to the argument. One might object that in the (b) case there is additional mental imagery: perhaps we tend to imagine various expansions of the array of dots that represent the first 5, 6, or whatever integers. This objection can be met by starting out with arrays of dots large enough that any attempt to imagine expansions of them would be too much of a burden on our visual buffer. In general, dependent parts of intellectual gestalts—whether they are themselves sensory or intellectual experiences—possess cognitive phenomenal characters. The reason is that they possess phenomenal characters that can only be instantiated by a part of an intellectual experience. One of the vexed issues about cognitive phenomenology that phenomenal holism might bear on is the relationship between conscious thought and inner speech. Suppose you consciously think that the first n positive integers can be represented by a triangular array of dots. This is an intellectual experience and it has a phenomenal character. Consider two theses about its phenomenal character. One thesis is that it is a cognitive phenomenal character—it cannot be instantiated by a purely sensory experience. Another thesis is that it is a sensory phenomenal character—it can be instantiated by a purely sensory experience, plausibly the purely sensory experience of imagining oneself saying

190

phenomenal intentionality

“The first n positive integers can be represented by a triangular array of dots.” These views generalize: Cognitivism: all conscious thoughts instantiate cognitive phenomenal characters. Sensationalism: all conscious thoughts instantiate only sensory phenomenal characters. Note that both cognitivists and sensationalists agree that there are conscious thoughts. What they disagree about is the nature of the phenomenal characters of conscious thoughts. The sensationalist thinks that they can be duplicated by purely sensory experiences, such as episodes of inner speech, and the cognitivist denies this. How might the cognitivist argue against the sensationalist? One strong thesis cognitivists might argue for is that any conscious thought could occur in the absence of any sensory phenomenal character. The thesis is difficult to establish. The typical stream of consciousness is fi lled with instances of sensory phenomenal characters, and I personally find it difficult to imagine a stream of consciousness absent all sensory phenomenal character. Let us suppose that every conscious thought is accompanied by some episode of inner speech. Cognitivists might argue that episodes of inner speech are distinct from conscious thoughts, and if you properly focus on a conscious thought, not an accompanying episode of inner speech, you will see that it has its own phenomenal character, which cannot be duplicated in a sensory experience. Again, it is not clear how cognitivists might establish such a thesis. Suppose a sensationalist responds: whenever I have a conscious thought and introspect all I find is an episode of inner speech expressing the thought. I do not see why we shouldn’t take this claim at face value. Cognitivists might at this point have recourse to variants on the standard arguments by example for cognitive phenomenology. Contrast the experience of a monolingual French speaker saying to himself “The first n positive integers can be represented by a triangular array of dots” with the experience of an English speaker saying the same to himself. The experiences are phenomenally different. But they share—the cognitivist might claim—all of their sensory phenomenal characters. So they differ in a cognitive phenomenal character, and this is the cognitive phenomenal character distinctive of conscious thought and which cannot be duplicated by a sensory experience. This argument is open to the standard worries about the standard arguments by example that appeal to changes in linguistic understanding. While the cognitivist thinks that all conscious thoughts instantiate cognitive phenomenal characters, in order to refute the sensationalist it suffices to show that some conscious thoughts instantiate cognitive phenomenal characters. Consider, then, the thesis that some conscious thoughts instantiate

Intellectual Gestalts

191

phenomenal characters that cannot be instantiated by purely sensory experiences. If this thesis is true, then sensationalism is false. Further, there are example conscious thoughts that suggest it is true, namely, conscious thoughts that occur as parts of intellectual gestalts. Take, for example, the conscious thought that the fi rst n positive integers can be represented by a triangular array of dots, which occurs as part of grasping the proof that the sum of the fi rst n positive integers is half of n × (n + 1). The phenomenal character of this conscious thought is holistically dependent on the whole experience of grasping the proof. Th at entails that its phenomenal character cannot be instantiated by a purely sensory experience. And that refutes sensationalism. The upshot of this discussion is that while it might be difficult to establish cognitivism, it is not difficult to refute sensationalism. Thoughts that occur as parts of intellectual gestalts are obvious counterexamples to sensationalism. To sum up, then, the defense of phenomenal holism and the extension of it to intellectual experiences pursued in the previous sections provides resources for defending two theses about cognitive phenomenology: first, there are phenomenal characters that cannot be instantiated by purely sensory experiences, and second, at least some of our conscious thoughts have such phenomenal characters. The argumentative resources that reflection on phenomenal holism brings into focus seem to me to have various dialectical advantages over some of the more familiar arguments for CP.

Acknowledgments I presented this chapter at the Phenomenal Intentionality Workshop hosted at the University of Arizona (2008), and the Annual Meeting of the Florida Philosophical Association (2008). I thank the paticipants in both events for helpful feedback. Thanks to Uriah Kriegel and Terry Horgan for extending an invitation to present at the Phenomenal Intentional Workshop. And thanks again to Uriah Kriegel for extensive comments on earlier drafts.

Notes 1. 2. 3. 4. 5.

Wertheimer (1924). The example is based on one from Wertheimer (1923). The quoted formulation is from Koff ka (1935), page 110. Gurwitsch (1964), page 121. PI must be true if these representational differences are phenomenal differences, though PI might be true even if these representational differences are not phenomenal differences.

192

phenomenal intentionality

6. The objections I will consider come from Barry Dainton (Dainton 2000) and Timothy Bayne (Bayne 2001). Neither Dainton nor Bayne rejects phenomenal holism. Both, however, argue that holistic phenomenal characters are very rare, and that most putative examples of holistic phenomenal characters turn out, on further consideration, not to be holistic. The balance of this section is dedicated to examining some of the considerations Dainton and Bayne discuss, and arguing that in general they fail to show that putative examples of holistic phenomenal character are not really holistic. 7. Dainton (2000); both (SI) and (WI) are presented on page 199. 8. Clearly 3 o’clock is no more distinguished than 7 o’clock or 11 o’clock. Plausibly, that is part of the reason that the F-experience representing the triangle as pointing 3 o’clock is no more likely to occur than alternative F-experiences representing the triangle as pointing 7 o’clock or 11 o’clock. The axis of symmetry of G is, however, a distinguished feature of figure G. Plausibly, that is part of the reason the G-experience representing the left-most triangle as pointing along the axis of symmetry of G is more likely to occur than alternative G-experiences representing the left-most triangle as pointing in its other possible directions. For an illuminating discussion of these matters from a psychological perspective see Palmer (1990). 9. Bayne (2001). 10. In (Dainton 2001), page 202. 11. Th is is taken from Burton (2009), page 5. 12. The beliefs themselves must—in this view—be non-phenomenal states. 13. “The flesh and blood of a tone depends from the start upon its role in the melody: a b as leading to tone c is something radically different from the b as tonic” (Wertheimer 1924, p. 5). 14. Phenomenal externalists deny this. As I will point out below, however, even phenomenal externalists might be reluctant to give up (1). 15. I have merely sketched a response to the puzzle. There are many difficult questions about time-consciousness connected with the present issue. Husserl’s discussion of retentions, impressions, and protentions in the stream of consciousness seems to me to be the most promising starting point for further exploration. See Husserl (2008). 16. See Pitt (2004) for a version of this argument. 17. See Horgan and Tienson (2002) and Pitt (2004) for various arguments by example. 18. Carruthers and Veillet (2011) press these objections. 19. It is unproblematic that the (b) experience is a sensory experience, specifically a visual experience. There are two distinctions that should be distinguished: there is the distinction between sensory and intellectual experiences, and there is the distinction between sensory and cognitive phenomenal characters. Though the (b) experience is a sensory experience, it has a cognitive phenomenal character. The reason it can is that it is not purely sensory because it is metaphysically dependent on the larger intellectual experience of which it is a part.

References Bayne, Timothy J. 2001. “Co-consciousness: Review of Barry Dainton’s Stream of Consciousness.” Journal of Consciousness Studies 8: 79–92. Burton, Robert. 2009. On Being Certain: Believing You Are Right Even When You’re Not. Reprint. New York: St. Martin’s Griffi n. Carruthers, P. and B. Veillet. 2011. “ The Case against Cognitive Phenomenology.” In T. Bayne and M. Montague 2011 (eds.), Cognitive Phenomenology. Oxford: Oxford University Press. Dainton, Barry. 2000. Stream of Consciousness. London: Routledge. Ellis, Willis. 1950. A Sourcebook of Gestalt Psychology. New York: Humanities Press. Gurwitsch, Aron. 1964. Field of Consciousness. Pittsburgh: Duquesne University Press.

Intellectual Gestalts

193

Horgan, Terry, and John Tienson. 2002. “The Phenomenology of Intentionality and the Intentionality of Phenomenology.” In Philosophy of Mind: Classical and Contemporary Readings. Oxford: Oxford University Press. Husserl, Edmund. 2008. On the Phenomenology of the Consciousness of Internal Time (1893–1917). 1st ed. New York: Springer. Koff ka, Kurt. 1935. Principles of Gestalt Psychology. First Printing. New York: Harcourt, Brace. Palmer, Stephen. 1990. “Modern Theories of Gestalt Perception.” Mind and Language 5, no. 4: 289–323. Pitt, David. 2004. “ The Phenomenology of Cognition, or, What Is It Like to Th ink Th at P?” Philosophy and Phenomenological Research 69, no. 1: 1–36. Werheimer, Max. (1938) 1950. “Laws of Organization in Perceptual Forms.” In A Sourcebook of Gestalt Psychology, ed. Willis D. Ellis. New York: Humanities Press. Wertheimer, Max. (1924) 1950. “Gestalt Theory.” In A Sourcebook of Gestalt Psychology, ed. Willis D. Ellis. New York: Humanities Press.

11

Does Phenomenology Ground Mental Content? A da m Pau t z Fortunately for the determinate character of intentional content, content determinacy is fi xed phenomenally. —Graham, Horgan, and Tienson (2007)

Many go in for the reductive externalist program concerning the mind and its intentionality. For instance, many have said that at least some beliefs have their contents because their neural realizers bear an appropriate physical relation to certain external conditions. We might call the relevant relation the tracking relation, leaving open whether it is to be explained in terms of causal covariation under optimal conditions (Stalnaker, Tye), asymmetric dependence (Fodor), indicator function (Dretske), or normal conditions for the proper function of output systems (Millikan).1 Many extend the same model to the phenomenal side of the mind. They accept intentionalism about experience: the phenomenology of an experience is determined by its intentional content. And they think the content of experience, like the content of belief, is fixed by tracking relations to the environment. The result is phenomenal externalism: the phenomenology of experience is not fi xed by what happens in the brain, but by what environmental states the brain tracks.2 The reductive externalist program faces many long-standing problems of detail. Among them are the disjunction problem, the distance (depth) problem, and problems about content determinacy due to Quine and Kripkenstein.3 Indeed Lycan (2009: note 1) has recently spoken of the “dismal failure” of all existing proposals within the reductive externalist program, suggesting that this provides the best argument for a non-reductive approach to the mind. I agree. The reductive externalist program is in a state of stagnation. Recently an alternative approach has come to the fore, the phenomenal intentionality program. This program gives a sense of revolution, of upsetting the

194

Does Phenomenology Ground Mental Content?

195

applecart. A number of ideas are loosely associated with it. First, prioritism: phenomenology is “explanatorily prior to” intentionality. Indeed, Horgan and Tienson (2002: 520) define “phenomenal intentionality” as intentionality that is possessed “in virtue of” phenomenology, where this stands for a relation of explanatory priority. So, for instance, sensory experience is richly intentional, and its intentionality is grounded in its phenomenology. Second, phenomenal internalism: roughly, phenomenology, and hence phenomenal intentionality, is internally determined. Th is is supposed to be intuitively obvious. In consequence, the reductive externalist program is a non-starter. For instance, an accidental, life-long “brain in a vat” might have a rich phenomenal life and stand in various intentional relations to various (false) intentional contents, even though its states do not have the function of tracking the environment. Third, phenomenal liberalism: in addition to sensory phenomenology, there is such a thing as cognitive phenomenology. And, just as sensory phenomenology grounds the content of sensory experience, cognitive phenomenology helps ground the content of cognitive states. This is supposed to finally solve the long-standing problems raised by Quine, Kripkenstein, and others concerning content determinacy. If one wanted a single slogan for the phenomenal intentionality program, it might be phenomenology first or maybe consciousness first.4 The core thesis is that phenomenology plays a foundational role in grounding all intentionality. As for the nature of phenomenology itself, proponents of the program have had very little to say. For instance, they have not said whether it is possible to provide a reductive internalist theory of phenomenology and intentionality, which can be put in the place of standard reductive externalist (“tracking”) theories. They tend to ignore the whole hard problem of naturalizing the mind. While the externalist program is reductive, the phenomenal intentionality program might go best with a non-reductive approach. In general, my own views fall within the vague boundaries of the phenomenal intentional program. While I think that the specific theses associated with the phenomenal intentionality program are very underdeveloped and poorly motivated, I also think that there are defensible theses in the vicinity.5 In this chapter, I will be focused on one thesis in particular, the thesis that “cognitive phenomenology” might help ground mental content. First (§1–2) I will argue that this thesis is very underdeveloped and poorly motivated. Then (§3) I will develop several arguments against it. (Here I will address the largely ignored issue of whether phenomenal intentionality might be naturalized.) Finally (§4), I will defend a claim in the same vicinity. On my view, it is sensory phenomenology, not “cognitive phenomenology,” that is the source of all determinate intentionality. To explain how, I will draw on some of David Lewis’s ideas concerning intentionality.

196

phenomenal intentionality

1. Preliminaries: Cognitive Phenomenology To begin with, I clarify the thesis that I will criticize, namely, the thesis that cognitive phenomenology plays an important role in grounding mental content. Consider an example due to Horgan and Tienson (2002). You hear “Visiting relatives can be boring” first as a remark about the people who are visiting and then as a remark about visiting certain people oneself. Horgan and Tienson say that the actual sound or auditory imagery may be the same, but the total experiences are phenomenally different. They conclude that the two occurrent beliefs differ in non-sensory phenomenology. Let cognitive phenomenology be the phenomenology (if such there be) that attaches to beliefs and other intentional states that is distinct from associated sensory phenomenology, where sensory phenomenology is understood broadly to include perceptual, bodily, imagistic, and emotional phenomenology. Horgan and Tienson say (2002: 522) that cognitive phenomenology is quite rich: “Change either the attitude-type (believing, desiring, wondering, hoping, etc.) or the particular intentional content, and the phenomenal character thereby changes too.” So, for instance, they believe in a special non-sensory, conative phenomenology that helps individuate our desires. (I will use “cognitive phenomenology” broadly, so that it also covers conative phenomenology.) Call the minimal thesis that there is such a thing as cognitive phenomenology the CP-existence thesis. This thesis has been widely discussed (see Montague 2010 for a helpful overview). Many put forward a second, much stronger thesis about cognitive phenomenology that has not been widely discussed. It will be my primary focus. Here are some representative passages: How can [cognitive] experience ever deliver determinateness? It just can. Cognitive experience in causal context can do just this. Such is its power. (When it comes to [thinking of] the number 2, it doesn’t even require causal context.) . . . If God could look into my mind and apprehend the cognitive [phenomenology] of my experience he would certainly know what I was thinking about, given that he also knew— and how could he not—about my causal circumstances. It is the same power that makes it the case that I can think determinately about the number 2 although there is no relevant causal context. Pff f! This is the correct account of how it is that content can be determinate in spite of all the problems raised for this idea by Kripke in his book Wittgenstein on Rules and Private Language. (Strawson 2010: 351, 354) The part of what is thought that is fully determined by [cognitive] phenomenal character [is] a kind of thought content. (Siewert 2011)

Does Phenomenology Ground Mental Content?

[E]ach specific occurrent intentional state with phenomenal intentional content is constitutively determined by its own distinct phenomenal character—viz., the what-it’s-like of undergoing that particular attitude-type vis-à-vis that particular phenomenal intentional content . . . Suppose that you are now undergoing a psychological state with the distinctive [non-sensory] phenomenal what-it’s-like of believing that a picture is hanging crooked on a wall directly behind you. Then you thereby believe that there is a picture hanging crooked on a wall directly behind you; undergoing this phenomenology constitutively determines that you are instantiating that belief-state. Any experiencing creature [e. g. a brain in a vat] undergoing this phenomenology would thereby instantiate the belief-state, even if its overall phenomenology is otherwise quite different from your own. (Horgan and Tienson 2002: 526) Physically and apart from phenomenology, there is no “one, determinate, right answer” to the question of what is the content of an intentional state. For . . . the content of each mental state is not determinately fi xed once the physical facts (including perhaps physical facts about the internal-environmental linkages) are fi xed. Fortunately, however, for the identity or determinate character of intentional content, content identity or determinacy is fi xed phenomenally. For example, the what-it’s-like of thinking “Lo, a rabbit” is different from the what-it’s-like of thinking “Lo, a collection of undetached rabbit parts” . . . [This] commitment to phenomenal individuation of intentional content, combined with rejection of physical individuation, [might be] tantamount to dualism. (Graham, Horgan and Tienson 2007: 476, 481) You know what you are thinking and what you mean by your utterance, and there is a determinate fact of the matter about what you are thinking and what you mean by your utterance, because there is something it is like to think a determinate thought and to make an utterance that expresses that thought. Developing in detail our proposed account of content determinacy is . . . an agenda item for the future. (Horgan and Graham 2010) [My view] maintains that the intentional content of a thought is determined by its intrinsic phenomenal properties, not its relational properties. My teachers will be very disappointed in me. (Pitt 2009: note 5, my italics)

197

198

phenomenal intentionality

Let a cognitive phenomenal property be a property of individuals of the form having a state with such-and-such cognitive phenomenology. All of the quoted philosophers apparently endorse the following CP-determination thesis: for at least some cognitive phenomenal properties P, there is a unique content c such that it is metaphysically necessary that, if an individual simply has P, then he has an occurrent belief (or desire) with content c.6 Now this idea is not new. Kripke (1982: 42) took it seriously. But he compared cognitive experiences to mere “raw feels” or bits of “mental paint,” like undirected depression. Against this, CP-determination says that they have built-in intentionality, just as perceptual experiences arguably have built-in intentionality. Let me make some clarifications. First, CP-determination is only meant to apply to our occurrent beliefs (also sometimes called thoughts or judgments) and our occurrent desires. It cannot be applied to our unconscious, standing beliefs and desires, which certainly lack phenomenology. To explain such beliefs and desires, the proponent of CP-determination needs a different account (for some options see Graham, Horgan, and Tienson 2007). This will not concern us here. Second, on the assumption that it is “intrinsic,” cognitive phenomenology of course cannot determine the “wide contents” of our beliefs and desires which can differ between internal duplicates: natural kind contents (for example, about water or rabbits), singular contents (for example, contents involving a particular river), and so on. Consequently, proponents of CP-determination typically only claim that cognitive phenomenology determines the narrow (and perhaps “centered,” de se) contents of beliefs and desires that do not differ between such duplicates. In this category they include mathematical contents (Strawson mentions contents about the number 2), certain general descriptive de se contents (the watery stuff of my acquaintance is wet), artifactual contents (Horgan and Tienson mention a content about a picture), and so on. To explain wide content, Horgan and Tienson (2002) adopt David Lewis’s (1994) view that wide content is derivative from narrow content and relations to the environment. The approach taken by proponents of the CP-determination thesis is interesting, for two reasons. First, it is unorthodox. The puzzle of intentionality can be put like this: how is it that one manages to stand in intentional relations to some contents rather than others? (Here and I assume that to have a particular belief or desire is to stand in the belief relation or the desire relation to a particular proposition or “content.”) According to orthodoxy: • The contents of the (occurrent and non-occurrent) beliefs and desires of a subject are always determined by features of the subject that go beyond his intrinsic properties at the time he has those beliefs and desires: factors such as environment-involving behavioral dispositions, tracking relations to the

Does Phenomenology Ground Mental Content?

199

environment, causal or inferential relations among internal states, what sentences he accepts and their contents as determined by their overall pattern of use in the language, and so on. Granted, many philosophers, for instance David Lewis and Frank Jackson, do recognize a notion of non-Twin-Earthable “narrow content.” But Lewis and Jackson repeatedly emphasize that, on their weak notion of “narrow content,” having a mental state with a particular narrow content is not entirely intrinsic, because narrow contents are determined by the typical world-involving functional roles of your internal states in your population, which are not intrinsic.7 By contrast, proponents of CP-determination typically accept the radical phenomenal internalism associated with the phenomenal intentionality program: all phenomenal properties (and hence cognitive phenomenal properties) are intrinsic, and hence non-functional, properties of individuals (see the quote from Pitt, and also Horgan, Tienson, and Graham 2004). (Th is is not obligatory: as we shall see in §3.1, one could accept the CP-determination thesis but reject the intrinsicness thesis.) When this is combined with CP-determination, we get: • Some of a subject’s occurrent beliefs and desires have contents that are fully determined by certain of his intrinsic properties at the time he has those occurrent beliefs and desires (namely, cognitive phenomenal properties), where the relevant intrinsic properties are distinct from all of his sensory and functional properties (past, present and future). There is a second reason that the CP-determination thesis is interesting: many declare that it finally solves content determinacy worries due to Quine and Kripkenstein. Here is a rough formulation of the Kripkenstein problem as it arises for physicalists. Let quus1, quus2, quus3 denote different functions defined over numbers that are just like the plus function but that differ from the plus function only when it comes to numbers that are too large for us to compute. Then, at least if we set aside the widely rejected view that necessarily equivalent propositions are identical, there are infinitely many distinct contents up in Plato’s heaven: that two plus two equals four, that two quus1 two equals four, that two quus2 two equals four, that two quus3 two equals four, and so on. If we say that the non-intentionally characterized functional and behavioral facts determine (in the sense of metaphysically necessitate) that one believes one of these contents (in particular, that two plus two equals four) to the exclusion of all the others, then we want some kind of explanation for this. How do these facts select or point to that particular content to be what you believe? Is there a physicalist-functionalist (perhaps a posteriori) account of the belief relation

200

phenomenal intentionality

that explains how the physical facts determine that this is the content you believe, as opposed to the other candidates? In my view, the main problem here is the problem of finitude. We have (non-intentionally characterizable) dispositions involving a certain finite set of numerals in language (and perhaps in the language of thought) but we do not have dispositions involving numerals that are too long for us to compute. Quine’s well-known problem about rabbits and undetached rabbit-parts is similar to the Kripkenstein problem. In the passages cited earlier, Strawson as well as Graham, Horgan, and Tienson declare that in such cases cognitive phenomenology saves the day: somehow it manages to determine what content you believe. The CP-determination thesis is underdeveloped. One does not get much more than the bare assertion that cognitive phenomenology determines content. This leaves many obvious questions unanswered. What exactly is the relationship between cognitive phenomenal properties and intentional properties? What is the relationship between cognitive phenomenal properties and physical-functional properties? Is phenomenal intentionality a natural phenomenon? I will be looking at these questions. I will argue that, even if the CP-existence thesis is true, the stronger and more interesting CP-determination thesis is not. Along the way I will present some novel reasons to doubt the CP-existence thesis as well.

2. The CP-Determination Thesis as Unmotivated Let us first look at arguments for the CP-determination thesis. Contrary to some proponents, it is not introspectively obvious, since many reject it. I will consider two arguments for CP-determination, and suggest that they fall short.

2.1. First Argument: The Determinacy Argument In several places, Graham, Horgan, and Tienson have briefly argued for CP-determination on the grounds that it provides the only adequate solution to the determinacy problems due Quine and Kripkenstein, as follows: 1 Against Quine and Kripkenstein, there generally are determinate facts about what we believe and mean. 2 But Quine and Kripkenstein are right that such facts could not be “determinately fi xed by” the physical facts. 3 If premise 2 is true, then there could be determinate facts about what we believe and mean only if they are fi xed by non-sensory, cognitive phenomenology: this is the only alternative to physical determination.

Does Phenomenology Ground Mental Content?

201

4 So, there could be determinate facts about what we believe and mean only if they are fi xed by cognitive phenomenology. (2, 3) 5 So, what we believe and mean is fi xed by cognitive phenomenology. (1, 4)8 There are several problems with this argument. (i) Against premise 2, there are numerous important physicalist answers to Quine and Kripkenstein’s arguments in the literature on “naturalizing intentionality.” But Graham, Horgan, and Tienson do not provide an in-principle reason to think that all they fail, nor do they eliminate them one by one. (ii) In support of premise 3, Graham, Horgan, and Tienson (2009: 531) say that if we “[agree] with Quinean misgivings about the physical determination of content, then all that’s left as a mode of individuation is phenomenological individuation.” From the examples they discuss (they speak of the what-it’s-like of thinking “Lo, a rabbit”) it is clear they mean cognitive phenomenology and not mere sensory phenomenology; and Horgan and Graham (2010) explicitly say that cognitive phenomenology over and above sensory phenomenology is needed to secure our actual level of content determinacy. Otherwise they do not have an argument from determinacy for cognitive phenomenology. But they largely ignore certain views on which sensory phenomenology is an important source of content determinacy. On some views, sensory phenomenology is richly intentional (for example, essentially involves the representation of observational properties), but its intentionality is not reducible to purely physical facts (Pautz 2010a; Chalmers 2006). Further, our actual and potential sensory experiences, together with the functional facts, are enough to secure our actual level of content determinacy. The proponent of this view shares Quine’s misgivings about the physical determination of content in a sense: he agrees that facts about determinate content are not fi xed merely by the third-person physical facts—for instance, non-intentionally characterizable behavioral and functional facts. But he denies that the only alternative to austere physicalism involves the appeal to non-sensory, cognitive phenomenology. Later (in §4) I will recommend this view as an alternative to CP-determination. How could Graham, Horgan, and Tienson show that this view is wrong, and that they are right in maintaining that cognitive phenomenology over and above sensory phenomenology is needed to secure the level of content determinacy that our intentional states actually possess? They would need to show that a community which is exactly like us in all physical respects and all actual and dispositional sensory respects (same sensory experiences with their built-in intentionality, same language, etc.), but which lacks the occasional extra bits of “cognitive phenomenology” that we allegedly enjoy (an absent cognitive qualia case), would differ profoundly from us in having intentional states whose contents are much less

202

phenomenal intentionality

determinate than the contents of our actual intentional states. They would have to show that sensory-functional supervenience fails: the total intentional facts about a population fail to supervene on the total sensory and functional facts about that population. (Uriah Kriegel asked me whether such an absent cognitive qualia case is nomically possible. But for the point I am making the case only needs to be metaphysically possible, and in §3.2 I will argue that it is.) But Graham, Horgan, and Tienson have certainly not shown this or even addressed the supervenience issue. How could they show this? Later on I will argue that the sensory-functional supervenience thesis is actually correct (§§3.2–3.3). (iii) In maintaining that the only alternative to physical determination is determination by cognitive phenomenology (premise 3), Horgan and Tienson ignore yet another option, one that is historically well known. As Boghossian (1989: 542) notes, Brentano’s thesis of the irreducibility of intentional idioms answers Kripkenstein and secures content determinacy but does not require cognitive phenomenology. So considerations concerning content determinacy alone simply do not justify acceptance of CP-determination at all. Horgan and Tienson cannot object to primitivism. As we will see (§3.1), their own brief remarks suggest a primitivist (or, as they themselves say in the passage cited earlier, “dualist”) view of content determinacy. (iv) Finally, against premise 1, one might say that, while perhaps some of our beliefs clearly determinately possess certain contents (for example, immediate perceptual beliefs) thanks to having an especially close connection to sensory experience, other beliefs (beliefs about electrons, arithmetical beliefs) must be admitted to be radically indeterminate in content. In response, Horgan and Graham (2010) would claim that in general content determinacy “is just obvious,” even though Quine and many others have rejected it. But how can introspection deliver certainty concerning this disputed, highly theoretical issue? Further, according to Horgan and Graham, in the absent cognitive qualia case mentioned earlier, when their own counterparts (sensory-functional duplicates) say (just as Horgan and Graham do) that “in general there is content determinacy,” their counterparts are giving expression to a false introspective belief, for their counterparts in this case lack the cognitive qualia which, according to them, are necessary for securing general content determinacy. So by their own lights introspection is highly fallible when it comes to the highly theoretical issue of content determinacy.

2.2. Second Argument: The Access Argument Very roughly, the first premise of this argument—defended for instance by Pitt (2011)—is that one has some kind of “special access” to one’s occurrently believing or desiring that p. The second premise is that this requires that occurrently

Does Phenomenology Ground Mental Content?

203

believing or desiring that p have an “individuative” cognitive phenomenology, in a sense that entails CP-determination. But there are two problems with the second premise of this interesting argument. First, many accounts of special access do not require CP-determination. Indeed, even Pitt’s own theory—which a kind of direct acquaintance theory—does not require that our occurrent thoughts be phenomenal states. There is nothing in the notion of acquaintance that demands that we are only acquainted with phenomenal states. Russell (1912: 105) explicitly allowed for acquaintance with non-phenomenal states, for instance two plus two equaling four. So what exactly is the argument against a view on which our occurrent thoughts are non-phenomenal states but we have immediate, acquaintance-based knowledge of them?9 Many other accounts of special access do not require CP-determination (for example, Byrne 2011, Carruthers 2011). Pitt criticizes some of these models. But (although I cannot discuss this here) I think his criticisms are not decisive. Further, he does not eliminate all possible models. A second problem is this. Consider once again the absent cognitive qualia case. It is exactly like the actual case in all respects (sensory, perceptual, functional, physical) except that our counterparts lack cognitive qualia. The second premise of the access argument has a very radical consequence: that in this case our counterparts could not enjoy the same kind of special, non-inferential access to occurrent beliefs and desires that we enjoy. But there are reasons to reject this consequence. To begin with, they clearly have occurrent beliefs and desires with some contents, despite lacking cognitive phenomenology. (Whether their contents are just as “determinate” as the contents of our beliefs and desires does not matter to the point I am trying to make here, namely, that special access to content doesn’t in general require CP-determination.) They are not cognitive zombies. So, for instance, suppose my counterpart, like me, currently has an experience of a round tomato; the sentence “that’s a round tomato” runs through his inner speech and he also assents to this sentence in public speech; and he manifests understanding this sentence just as I do. Suppose he also has strong hunger pangs, imagines food, and has experiences of seeking food; the sentence “food would be good” runs through his inner speech and he assents to this sentence in public speech; and he manifests understanding this sentence just as I do. No one competent with the concepts of belief and desire would deny that, like me, he believes a round tomato is present and that he desires food; and there is no theory-neutral reason to deny these things. Moreover, like me, without having to go through any conscious process of inference, my counterpart assents to the sentences “I believe that’s a round tomato” and “I desire food” and exhibits full understanding of these sentences and mastery of the concepts of belief and desire. So no one would deny that he also has introspective beliefs about his belief and desire. Now these introspective beliefs have all the earmarks of knowledge; they are, for instance, non-accidentally true. What

204

phenomenal intentionality

reason is there to deny that they constitute knowledge? Perhaps it will be said that from my counterpart’s point of view these introspective beliefs pop up “out of the blue” (as in familiar cases of unwitting reliability discussed by epistemologists). As against this, they are quite in line with his experiential life, for instance, his inner speech and experienced behavior. The conclusion I draw is that my counterpart also has immediate access to his beliefs and desires, despite his lack of “cognitive phenomenology.” Special access doesn’t require cognitive phenomenology.

3. Against CP-Determination No convincing argument for the CP-determination thesis has yet been provided. I will now develop several arguments against it.

3.1. First Argument: The Danglers Argument My first argument is that there are reasons to interpret proponents of CP-determination as “anti-reductionists” or “primitivists” about cognitive phenomenology and hence cognitive intentionality. They accept what is sometimes called Brentano’s thesis (Quine 1960: 221). But, if they reject reductionism and accept primitivism, they need brute principles or “danglers” (Smart 1959) connecting cognitive phenomenology with mainstream physical properties. Let me first argue in some detail for the “primitivist” interpretation. According to proponents of CP-determination, what is the nature of cognitive phenomenology, such that it delivers determinate content? They never say. They might take a reductive physicalist view of cognitive phenomenology and phenomenal intentionality in general. Here I understand reductive views broadly to include views that identify mental properties with physical properties and physically realized functional properties. There are two types of reductive physicalist theories of sensory phenomenology: biological theories and functional theories. Proponents of CP-determination might apply the same reductive theories to cognitive phenomenology, thereby avoiding the need to postulate brute principles or “danglers” to explain its presence in the world. But there is reason to think that they would not. To see this, consider first a biological theory of cognitive phenomenology, akin to the familiar biological (type-type identity) theory of sensory phenomenology. To illustrate, let P be the non-sensory phenomenal property you allegedly possess when you occurrently believe that two plus two equals four on a certain occasion. On the biological theory of cognitive phenomenology, P is necessarily identical with an intrinsic neuro-computational property M, which has its connections with the external world and behavioral outputs only contingently.

Does Phenomenology Ground Mental Content?

205

This might be the only naturalistic theory compatible with the radical phenomenal internalism favored by friends of the phenomenal intentionality program, according to which phenomenal properties are intrinsic properties of individuals. Let biological CP-determination be the conjunction of CP-determination and the biological theory of cognitive phenomenology. This conjunction entails that, for some narrow (non-Twin-Earthable) beliefs and desires with content c, there is a single neural state that metaphysically necessitates having a belief or desire content c. For instance, on CP-determination, having cognitive phenomenal property P determines as a matter of metaphysical necessity believing two plus two equals four. On the biological theory of cognitive phenomenology, P is necessarily identical with neuro-computational property M. Hence, according to biological CP-determination, merely having the mere neuro-computational property M determines as a matter of metaphysical necessity believing two plus two equals four. Another neural property metaphysically necessitates believing there is a picture on the wall behind oneself. And so on. Of course, the biological theory might also be applied to sensory intentionality. Maybe, for instance, there is a neural property that necessitates being visually acquainted with the general property roundness (Horgan, Tienson, and Graham 2004). Might proponents of CP-determination accept biological CP-determination? Is the biological theory a plausible model for “naturalizing” phenomenal intentionality? I think not. First, belief and other standard intentional states must be taken to be relations to intentional contents or properties or other abstract objects. Those attracted to non-relational accounts of intentionality (for example, Kriegel 2011: chap. 3) have not answered the very strong arguments for the standard relational view (for example, Schiffer 2006). So, for instance, the proponent of biological CP-determination must recognize the existence of a two-place belief relation Bn that we bear to “narrow contents.” Now the proponent of biological CP-determination already reduces monadic cognitive phenomenal properties to monadic neuro-computational properties. So he doesn’t need brute laws or “danglers” to explain the correlations between these properties. Could he also provide a reductive account of the dyadic relation Bn? Could he identify it with some dyadic physical or functional relation between individuals and narrow contents (abstract objects), such that what contents you bear this relation to (two plus two is four, there is a picture on the wall, etc.) are somehow fi xed merely by your intrinsic neural properties? The same problem arises concerning other intentional relations. Horgan, Tienson, and Graham (2004: 304–305) claim that there are many existing things (for instance, abstract objects like numbers and general properties like being a picture or being round ) such that your thinking about them is fi xed simply

206

phenomenal intentionality

by your narrow, intrinsic neutral properties. Might they provide a reductive account of this narrow thinking-about relation, Tn? This is a huge problem, but elsewhere I have argued that the answer is “No.”10 To appreciate the argument, consider that biological CP-determination has the consequence that an accidental, isolated brain-in-a-vat might bear the relation Tn to the number four, the property of being a picture, the property of being round, and so on. But by stipulation it does not bear the usual tracking, behavioral, or other physical relations to these entities, which are appealed to in all existing naturalistic theories. How then could Tn be a physical relation? If we confine ourselves to intrinsic neural features (firing rates, causal relations among uninterpreted mentalese sentences) and do not appeal to relations to the environment and behavior to pin down content, then a reduction is out of the question. There is no general algorithm, applying to all actual or possible individuals, that goes merely from these intrinsic features to what items (properties like shapes, mathematical objects) an individual is thinkingN of and what narrow contents he believesN. So the defender of biological CP-determination must take Tn and Bn to be primitive relations, thus frustrating his reductive ambitions. A corollary of this is that he would have to accept “brute necessities.” He wouldn’t after all avoid the kind of “modal danglers” that type-type identity is designed to avoid (Smart 1959). For instance, he would have to say that it is a brute necessity that neural duplicates must bear the primitive intentional relations Bn and Tn to the same contents (two plus two is four, there is a picture on the wall ) and other abstract objects (numbers, general properties).11 And it is just a brute (and fortunate) fact that they bear these relations to plus-contents rather than “bent,” quus-like contents, so that there is no interesting solution to Kripkenstein puzzle. It is not merely that he would have to say that these necessities are brute in the epistemic sense that they are not a priori knowable (most physicalists accept such epistemic bruteness). He would have to say that they are brute in a more objectionable metaphysical sense (roughly along the lines of Dorr 2007): they cannot be derived from any more basic modal truths. There is another problem. Biological CP-determination is open to what we might call the “separation argument.” (Pautz (2010c: sect. 4) develops the same line of argument against biological theories of sensory intentionality.) According to biological CP-determination, having the mere neuro-computational property M metaphysically necessitates believing that two plus two equals four. Against this, there are possible “separation cases” in which M (perhaps just a bit of brain-writing in the “cognitive phenomenology” center of the brain) is totally functionally isolated and plays no interesting functional role with respect to experiencing collections of objects and the experience of counting objects, imagery, and so on (cf. the “baby case” in §3.2). There are also worlds in which it plays such a functional role, but one appropriate to a quite different arithmetical belief (for example, five plus five equals ten). In these cases

Does Phenomenology Ground Mental Content?

207

M is “separated from” its actual sensory-functional role. It would be absurd to suggest that in these possible cases M realizes believing that two plus two equals four. So, contrary to biological CP-determination, possessing the mere neuro-computational property M does not metaphysically necessitate believing that two plus two equals four. So those who favor CP-determination should not accept biological CP-determination for the simple reason that it is a bad theory. Against biological CP-determination, most think that it is the connections our internal neural states bear to the world and behavior that link them to some contents and not others. Accordingly, the friend of CP-determination might instead consider providing a reductive functionalist theory of the cognitive phenomenal properties which in their view determine intentional content. To illustrate, consider a non-arithmetical example. In the passage quoted in §1, Horgan and Tienson say that there is a unique cognitive phenomenal property that metaphysically necessitates believing that there is a picture hanging on the wall behind one. Call it P for “picture.” By contrast to the proponent of a biological theory of phenomenal properties, the reductive functionalist I have in mind will identify P with some kind of functional property defined in terms of external inputs and behavioral outputs. Here I understand functional properties broadly to include historically determined “tracking” relations to the environment. Let functional CP-determination be the conjunction of the CP-determination thesis and some functionalist theory of cognitive phenomenal properties. Thus, suppose that L is the functional property that David Lewis (1994) would say constitutes occurrently believing that there is a picture hanging on the wall behind one. The proponent of functional CP-determination might say that the cognitive phenomenal property P just is the Lewisian functional property L. Or again, suppose that F is the functional property that Jerry Fodor (1994) would say constitutes having the same occurrent belief. It might involve the tokening of a mentalese sentence, functional role, and relations of “asymmetric-dependence” to the external world. According to another version of functional CP-determination, the cognitive phenomenal property P just is the environment-involving functional property F. So, cognitive phenomenology and hence cognitive intentionality is not in the head, just as some intentionalists about experience insist that sensory phenomenology is not in the head (as noted in the introduction). According to any version of functional CP-determination, cognitive phenomenal properties constitute intentional properties, because they are identical with functional properties and those functional properties constitute intentional properties. Now friends of CP-determination could accept functional CP-determination, thereby providing a “naturalistic” theory of how cognitive phenomenal properties ground intentionality. True, rejecting biological CP-determination and

208

phenomenal intentionality

accepting functional CP-determination would require rejecting the thesis that cognitive phenomenal properties are intrinsic; but I never built that into the formulation of CP-determination. As I have pointed out elsewhere (Pautz 2008: 272–273), what is arguably the core thesis of the phenomenal intentionality program—namely, that phenomenology is in some sense the source of all intentionality—is quite compatible with reductive externalist (for example, tracking) theories of phenomenology. Recently, Kriegel (2011: 69, 249) has made the same point and has developed it into a whole theory. He says he is strongly attracted to a version of functional CP-determination. In particular, he favors a higher-order tracking theory of phenomenal intentionality (109). However, while Kriegel favors functional CP-determination, I think other friends of phenomenal intentionality will reject it, given their other commitments. (i) As noted in §1, Horgan and coauthors as well as Pitt hold that cognitive phenomenal properties are intrinsic and “non-relational.” This rules out functional CP-determination, since the relevant functional properties are non-intrinsic.12 (ii) As noted (§2.1), Horgan and coauthors reject all orthodox physicalist solutions to the determinacy problems due to Quine and Kripkenstein. They appear to hold that the functional facts are insufficient to constitute (or even “fi x”) the intentional facts. But, according to functional CP-determination, the functional facts do constitute the intentional facts. For, according to functional CP-determination, the cognitive-phenomenal facts constitute the intentional facts; further, they are identical with certain functional facts; so (by Leibniz’s law) the functional facts constitute the intentional facts. (iii) Presumably, fans of the phenomenal intentionality program would say that their claim that phenomenology grounds intentionality plays an essential role in avoiding the well-known problems of detail plaguing the reductive externalist program and the naturalization program more generally. But those who favor functional CP-determination (Kriegel 2011: 109) cannot say this. As just noted, like standard naturalists, they are committed to saying that certain naturalistic facts (about functional role, inferential role, tracking, and so on) constitute the intentional facts. So they face the standard problems of detail with that claim. For instance, they are committed to the existence of a general, non-circular, counterexample-free naturalistic account of intentionality of the form state S has content C iff S satisfies naturalistic condition F, despite the “dismal history” of failed attempts to provide such an account. They face the disjunction problem, the depth (distance) problem, Kripke’s problem of finitude, and Quine’s inscrutability problem. How do they have any advantage concerning these problems over standard naturalists such as Fodor and Lewis who do not recognize cognitive phenomenology? True, they believe that some of the relevant content-grounding naturalistic properties are identical with non-sensory

Does Phenomenology Ground Mental Content?

209

phenomenal properties, while standard naturalists deny that they have any special “feel.” Perhaps it will be said that this gives them an advantage (thanks here to Uriah Kriegel). But how could this identity claim possibly help them solve these particular problems? What exactly are their solutions? Their acceptance of this identity claim does not mean that they believe that there are more facts out there in the world than those recognized by standard naturalists like Fodor and Lewis. Both sides agree that the only facts out there are the naturalistic facts: facts about functional role, inferential role, tracking, biological function, and so on. So they have exactly the same basic materials to work with. Therefore, if there are solutions to the aforementioned metaphysical problems about content determination compatible with functional CP-determination, those very same solutions are also compatible (mutatis mutandis) with standard naturalism. Since I can think of no plausible reductive theory of cognitive phenomenology consistent with the things proponents of CP-determination have said, I provisionally conclude that they must accept primitivist CP-determination. Indeed, the passages from Horgan and Tienson and Strawson quoted in §1 already suggest some kind of “primitivist” view of cognitive phenomenology and intentionality. Primitivist CP-determination comes in two possible versions. According to a quite extreme version, there is no real definition of a cognitive phenomenal property. In other words, there is no (a priori or a posteriori) answer, in neural or functional or other terms, to the question: what is it to have a given cognitive phenomenal property? In this sense, cognitive phenomenal properties are utterly simple, just as G. E. Moore said goodness is utterly simple. Yet they somehow “ground” determinate belief and desire properties, in accordance with CP-determination. For instance, it is just a brute fact that, if you have a certain utterly simple, unstructured cognitive phenomenal property, you stand in the belief relation to a plus-content, as opposed to a bent, quus-like content. According to another version of primitivist CP-determination, there is a partial answer to the question: what is it to have a particular cognitive phenomenal property? The idea is that having a particular cognitive phenomenal property simply consists in (and so trivially determines) occurrently standing in a particular intentional relation to a particular narrow content. So cognitive phenomenal properties have a relational structure. Of course this identity view would nicely explain the tight relationship that is supposed to obtain between cognitive phenomenology and intentionality. It would explain why it is necessary that if you “change either the attitude-type (believing, desiring, wondering, hoping, etc.) or the particular [narrow] intentional content” then “the phenomenal character thereby changes too” (Horgan and Tienson 2002: 522).13 What makes the view I have in mind primitivist is that it adds that there

210

phenomenal intentionality

is in turn no reductive account of our phenomenally constituted intentional relations toward narrow contents. For instance, there is no answer to the question: what is it for you to occurrently believe that two plus two equals four, as opposed to two quus two equals four? This kind of “constitutive” question cannot be answered by appealing to functional role or anything else. This view of cognitive phenomenal properties is analogous to non-reductive intentionalism about sensory phenomenal properties (Chalmers 2006; Pautz 2006, 2010a). Now I can finally state my “danglers” argument, which targets primitivist CP-determination. Let P be the phenomenal property which allegedly determines (perhaps just is) your occurrently believing that two plus two equals four, as opposed to two quus two equals four. According to primitivist CP-determination, P is not a physical property, at least in a narrow sense where physical properties just include mainstream physical properties like neural properties, functional properties, tracking properties, and so on. Yet even on this view your having this primitive cognitive phenomenal property is not a fluke; it is in some sense determined by one of your mainstream physical properties. Thus Strawson (2010: note 54) speculates that cognitive phenomenal properties are dependent on neural properties instantiated somewhere near the sensory regions of the brain. Alternatively, perhaps your having P is connected to (but distinct from) your having a more complex functional property involving your inferential dispositions. In any case, there is a true conditional of the form: necessarily, if any individual has physical property F, then that individual has distinct primitive cognitive phenomenal property P, which in turn determines (and perhaps just is) his occurrently believing the content two plus two equals four as opposed to the content two quus two equals four. These conditionals link the physical properties of individuals with their determinately believing (or desiring or wondering) some narrow contents rather than others. Now, according to any version of primitivist CP-determination, these conditionals simply admit of no explanation. On one version, they are merely nomically necessary. They are fundamental laws of nature, dangling from the body of fundamental physical laws of nature. For instance, maybe there is a basic law linking some neural state M with a plus-phenomenology and a certain plus-content (for example, two plus two equals four), rather than a quus-phenomenology and a quus-content (two quus two equals four). This amounts to dualism: there are possible zombie worlds, in which everyone is physically the same but no one has the relevant special phenomenal and intentional properties. On another possible version of primitivist CP-determination, the conditionals linking physical conditions with distinct intentional conditions are brute metaphysical necessities. What I mean by calling them “brute” is not that they are not a priori. I mean that they cannot be derived from more basic modal truths (roughly along the lines of Dorr 2007). In that sense, they “dangle from” the rest of the body of

Does Phenomenology Ground Mental Content?

211

modal truths. In this version, primitivist CP-determination amounts to what I have called primitivist physicalism (Pautz 2010a; see also Horgan 2010). Note that this view is exactly like dualism concerning what there is. It only differs from dualism only modally, since like physicalism it entails that zombie worlds are metaphysically impossible. Now we all must admit some brute nomic and metaphysical necessities: the fundamental laws of nature, the truths of logic, maybe certain non-logical metaphysical principles. However, we have general Occamist reasons to keep them to a minimum. The trouble with primitivism about cognitive intentionality is that it solves the determinacy problems due to Quine and Kripkenstein at the cost of obnoxiously multiplying the stock of basic modal truths. Th is is not a knockdown argument; but it gives us a strong reason to disbelieve the view.14 My point here should be put as a dilemma. Either proponents of CP-determination say that cognitive phenomenal properties are identical with mainstream physical or functional properties, or else they accept primitivism. If they accept the first horn (Kriegel 2011: chap. 2 expresses sympathy), then they might avoid “danglers,” because then they would have a respectable reductive view. But then they are in effect accepting one or another of the familiar attempts to “naturalize” intentionality (Fodor, Dretske, Lewis, etc.). They are only adding that the physical-functional properties that determine content have a non-sensory “feel.” This does not help them solve the standard problems of detail due to Kripkenstein and Quine and others.15 If, on the other hand, they take the second “primitivist” horn, then their solution to these problems is not novel: it is just the primitivist solution, which has been on the table from the start (Quine 1960: 221). They are just adding that the primitive intentional properties have a non-sensory feel. Moreover, they face the “danglers” argument.

3.2. The Separation Argument My remaining arguments against CP-determination—the “separation argument” and two more arguments to follow—have a common form. To begin with, let me introduce the notion of sensory-functional conditions. These are conditions concerning your actual or possible sensory experiences. They include your actual perceptual and bodily experiences, sensory images, your experiences of inner speech, and transitions among sentences in inner speech. They include your experiences of behaving in the world. They also include causal and dispositional conditions about your sensory experiences. Thus, supposing you believe that there is food in the fridge, they include your disposition to form an image of food in the fridge, your disposition to look in the cupboard should you not see food in the fridge, and so on.

212

phenomenal intentionality

Now built into CP-determination is a certain irreducibility thesis. This is the thesis that cognitive phenomenal properties, and hence some belief and desire properties, are distinct from all sensory properties, forming an autonomous level of mental reality. Indeed, following standard practice, I have stipulated that “cognitive phenomenal properties,” if such there be, are distinct from sensory properties. Now irreducibility typically goes with modal independence. So the CP-determination thesis entails that cognitive phenomenal properties, and hence some belief and desire properties, are totally modally independent of all sensory properties (Chalmers 2008). So they ought to be totally modally independent of sensory-functional conditions, for those are just conditions involving patterns in the instantiation of sensory properties. All cognitive phenomenal properties, and hence some belief and desire properties, ought to be combinable with any set of sensory-functional conditions. Consider an analogy: since auditory qualia are distinct from visual qualia, they can be combined with any series of visual qualia. What my next three arguments have in common is that they are supposed to show that it is not true that cognitive phenomenal properties, and hence some belief and desire properties, are completely modally independent of sensory-functional conditions. I will ask whether we can make sense of three types of scenarios in which cognitive phenomenal properties float free from sensory-functional conditions. I will argue that we cannot make sense of these scenarios. My first argument, the separation argument, generalizes my “separation argument” against biological CP-determination (§3.1) to any version of CP-determination. Consider an example. According to the CP-determination thesis defended by Horgan and Tienson, there is a cognitive phenomenal property P that metaphysically necessitates occurrently believing the narrow, de se content there is a picture on the wall behind one. Now, in the actual case, P is associated with certain sensory-functional conditions: for instance, having the sentence “there is a picture on the wall behind me” run through one’s interior monologue, and imagining (or being disposed to imagine) a picture on the wall behind one. The first step of my separation argument asserts that, if there is such a property as P distinct from all sensory properties, then there should be possible “separation cases” in which it is associated with a completely different set of sensory-functional conditions than those with which it is actually associated. In one such case, while Charlie has P, the sentence “there is a clock on the wall” runs through his interior monologue (where Charlie manifests understanding this sentence just as we do), he is disposed to form a sensory image of a clock on a wall, and in general has sensory experiences of behaving exactly as if there is a clock on the wall behind him (for example, if he were to experience someone asking for the time, he would experience himself turning around and looking at the wall).

Does Phenomenology Ground Mental Content?

213

The second step of my separation argument asserts that, in this case, even though by stipulation Charlie has P, he does not occurrently believe that there is a picture on the wall behind him, contrary to Horgan and Tienson’s CP-determination thesis. The clock-appropriate sensory-functional conditions present in the case are incompatible with his having this occurrent belief.16 In general, for any belief and desire, some sensory-functional conditions are a priori incompatible with having that belief or desire.17 In this sense, there are sensory-functional constraints on belief and desire.18 Consider a second separation case. Let P be the cognitive phenomenal property which, according to proponents of CP-determination like Strawson, metaphysically necessitates occurrently believing the narrow content two plus two equals four, thus solving the Kripkenstein problem. Again, my separation argument against CP-determination has two steps. First, if there is such a non-sensory property as P, then a baby (say) presumably might have P for a few seconds while otherwise remaining the same. This might involve temporarily modifying its neural state in the “cognitive phenomenology center” of its brain (Strawson 2010: note 54) while leaving everything else the same; but I see no reason that it should not be metaphysically possible. Second, even though the baby has cognitive phenomenal property P, it does not occurrently think that two plus two equals four, contrary to CP-determination. That would require that it have certain arithmetical concepts (for example, plus), which in turn would require that it have certain arithmetical abilities (for example, the ability to count). But by stipulation it lacks these abilities. In general, the sensory-functional conditions present (or rather absent) in the case are incompatible with the baby’s having the belief that two plus two equals four.19 Here is a fi nal separation case. First consider an actual person, Elmer. As we saw in §1, according to Horgan and Tienson, when an ordinary person like Elmer says “There’s a rabbit ” and then “There’s an undetached rabbit-part,” he enjoys different cognitive phenomenal properties, and this somehow solves Quine’s (1960) inscrutability problem. Call them R and U, respectively. Now consider an altered cognitive qualia case where these cognitive phenomenal properties are everywhere swapped with one another, but the sensory-functional facts remain exactly the same. I cannot imagine this and so am skeptical of the very existence of these properties; but if Horgan and Tienson are right that there are such cognitive phenomenal properties distinct from sensory properties, the case should be possible. Further, given their commitment to CP-determination, they must say that, in this counterfactual case, when Elmer says “there’s a rabbit,” he is really thinking there is an undetached rabbit-part, because he has cognitive phenomenal property U instead of R. There are radical intentional differences between the actual case and this case, even though they are completely identical in all sensory and functional respects. Th is is absurd: by stipulation, in this case as in the actual case, Elmer

214

phenomenal intentionality

visually focuses on the whole rabbit, he insists “I am thinking about the rabbit, not just a rabbit-part,” and the use-properties of his sentences are exactly as they are in the actual case.20 Now Siewert (1998: 285) at least addresses a case similar to one of my cases, the baby case. He simply asserts that the case is impossible, saying that he accepts “holism” about cognitive phenomenal properties and rejects “atomism.” He does not address cases like my Charlie case or my Elmer case in which cognitive phenomenal properties are associated with different sensory-functional conditions than those with which they are in fact associated. But he might just declare that they too are impossible. In general, proponents of CP-determination (Graham, Horgan, and Tienson, Pitt, Strawson) might say that, since there are necessary sensory-functional constraints on belief and desire, and since cognitive phenomenal properties determine belief and desire, there are necessary sensory-functional constraints on cognitive phenomenal properties. But this is not enough. Remember: I gave an argument for claiming that, if there are cognitive phenomenal properties, then separation cases are possible. Graham, Horgan, and Tienson as well as Pitt themselves say that cognitive phenomenal properties are distinct from all sensory properties. Given this, they should be totally modally independent of sensory properties. Hence they should be modally independent of sensory-functional conditions, since these simply concern the pattern of instantiation of sensory properties. If proponents of CP-determination say that my separation cases are not possible, they need to answer this argument. When something is impossible, we generally expect an explanation for why it is impossible. Here is one response. According to David Lewis’s theory of belief, occurrently believing that two plus two equals four is a complex functional property L involving relations to experiences, other beliefs, and behavior. This view explains why belief properties are holistic and not atomistic, so that a baby (for instance) cannot have arithmetical beliefs. It explains why there are necessary sensory-functional constraints on belief and desire. To explain why separation cases are not possible, the proponent of CP-determination might apply the same model to cognitive phenomenal properties. Thus he might simply identify the cognitive phenomenal property P (which allegedly determines believing two plus two equals four) with the Lewisian functional property L. In general, cognitive phenomenal properties reduce to functional properties, he might say. This would explain why cognitive phenomenal properties are holistic and not atomistic (why there are necessary sensory-functional constraints on having such properties), something Siewert asserts without explaining. It would explain why my separation cases are not possible for cognitive phenomenal properties, thus answering my separation argument against CP-determination. But now CP-determination has become functional CP-determination. It is now just functionalism about intentionality. As we saw in §3.1, this is not

Does Phenomenology Ground Mental Content?

215

compatible with the usual commitments of friends of the phenomenal intentionality program. Indeed, they seem to hold that cognitive phenomenal properties are primitive properties that cannot be reduced to mainstream physical properties of any kind. Let me suggest a different reply to my separation argument on behalf of proponents of CP-determination. I will introduce the reply by analogy. Swoyer (1982) accepts an interesting view about fundamental physical properties, like masses and charges. According to this view, they do not reduce to clusters of dispositions or functional properties. But they essentially possess certain (second-order) nomic properties and essentially stand in certain nomic relations; these are given by the fundamental physical laws. So they are necessarily connected with certain dispositional and functional properties, and “separation cases” in which they are recombined in ways that violate their nomic profi les are metaphysically impossible. There are unHumean metaphysically necessary connections between distinct existences. Likewise, proponents of CP-determination might say that cognitive phenomenal properties are distinct from sensory phenomenal properties. But they are necessarily connected with certain sensory-functional conditions. They essentially bear certain systematic relations to various other states and conditions: other cognitive phenomenal properties, behavior or attempts at behavior, mental dispositions and abilities (for example, inferential dispositions and dispositions to form certain images), sensory experiences, and so on. Hence, any given cognitive phenomenal property (realizing a particular belief or desire property) necessarily brings with it a system of states and abilities. This is not because cognitive phenomenal properties reduce to abstract functional properties or clusters of dispositions, as in functional CP-determination. Rather, it is just a brute modal fact, brute in the sense that it cannot be derived from real definitions and logic (Dorr 2007). Hence, under this view, the sensory-functional constraints on belief and desire are just brute modal facts. Call this version of CP-determination systemic CP-determination. Under systemic CP-determination, my separation cases involving cognitive phenomenal properties (Charlie, the baby, Elmer) are impossible. So my separation argument against CP-determination fails at the first step. But, since we have no reason to accept CP-determination in general (§2), we have no reason to accept systemic CP-determination. On the other hand, we have strong reasons to reject it. (i) Previously, we saw that proponents of CP-determination apparently must accept primitivist CP-determination, which requires brute modal principles connecting cognitive phenomenal properties with underlying physical states. Now we have seen that, to accommodate the sensory-functional constraints on belief and desire and avoid my separation argument, they must accept additional brute “systemic” principles connecting cognitive phenomenal properties with other mental conditions

216

phenomenal intentionality

and background abilities. But we ought to keep brute modal principles to a minimum.21 (ii) Presumably, a fundamental physical property might have a determinate functional profi le, given by the fundamental laws of nature. But what are the essential functional profi les of cognitive phenomenal properties? For instance, does the cognitive phenomenal property P discussed previously necessitate the disposition to form a visual image of a picture (rather than a clock) on the wall behind one (at least in those with the capacity for visual imagination)? Presumably, there will be a lot of vagueness here. So systemic CP-determinate requires objectionable “vagueness in the world” concerning the fundamental modal facts.22

3.3. Absent Cognitive Qualia Cases Before (§2) I used absent qualia cases to undermine the determinacy and access arguments for CP-determination. Now I will use such cases in arguments against cognitive phenomenology. My first argument casts doubt on the basic CP-existence thesis and has two steps. The first step asserts that, if there are cognitive phenomenal properties, then cases of the following kind are possible and indeed we should be able to positively imagine them: (i) you have exactly the same sensory properties and functional properties that you have in the actual case, and yet (ii) your phenomenal life is profoundly different from your actual phenomenal life in that you lack the phenomenal properties (namely, cognitive phenomenal properties) you actually enjoy. Indeed, presumably, if the CP-existence thesis is true, and cognitive phenomenal properties form a distinct experiential modality (Strawson 2010: note 54), then there might be among actual humans individuals who lack cognitive qualia, but otherwise have phenomenal lives very similar to normal individuals. Analogy: our auditory phenomenal properties are distinct from our visual phenomenal properties. So there are possible (indeed actual) individuals who see but do not hear. The second step of the argument is that we cannot positively imagine such cases. Just try. Suppose in the actual case you hear a friend say “Let’s go to the bar later” and you quickly form an image of the local bar and follow up with a question as to time. Now try to imagine a case that is completely identical to the actual case in all sensory and functional respects, and yet profoundly phenomenally different in many ways in that you lack the cognitive phenomenal properties that you allegedly actually possess. I honestly cannot do it. In short, the CP-existence thesis makes testable empirical predictions about what we can imagine. But the predictions are false. Perhaps the proponent of the CP-existence thesis will reply: just imagine hearing “Let’s go to the bar” while not being a speaker of English. But that case would differ in sensory and functional respects from the actual case. If I did not

Does Phenomenology Ground Mental Content?

217

understand English my auditory system would not parse the auditory stream “Let’s go to the bar” in the same way it actually does. So I would have a different auditory experience. I would also not form a quick image of the local bar. Not understanding what was said, I might also feel stressed. And I would not be disposed to follow up with certain questions. What the proponent of the CP-existence thesis is committed to is the possibility of a case in which all the past and present sensory and functional facts are held constant (the very same auditory experiences, the same imagery, the very same feelings, and so on), and yet my phenomenal life is profoundly different from my actual phenomenal life in myriad ways. It is this I cannot imagine. Galen Strawson suggested another reply to me (in discussion): that in the hypothetical case I would just be very bored. But, since in this case I would be a sensory-functional duplicate of my actual self, I do not think I would be bored. More importantly, if we actually have cognitive phenomenal properties, then the absent cognitive qualia case would differ from the actual case in many ways: each and every cognitive phenomenal property I allegedly actually possess would in this case be “subtracted out.” To simply say that I would be bored simply does not enable me to imagine the myriad ways in which this case would differ from the actual case. Maybe the defender of cognitive phenomenology could insist that this case is possible, even if I cannot imagine it. But why should this be? It is not only possible that my experiential life should differ in myriad ways because I lack all of my actual auditory phenomenal properties while the rest of my phenomenal life is held constant; I can easily imagine this case. So why cannot I easily imagine occupying a phenomenal world different in myriad ways because all of my actual cognitive phenomenal properties are removed, while all of the sensory and functional aspects of my life stay exactly the same?23 Now I turn to my second argument involving absent cognitive qualia cases. While my first argument was directed against proponents of the CP-existence thesis, my second argument is directed against what we might call the necessity thesis: non-sensory cognitive and conative qualia are necessary in order to have the level of content determinacy we actually enjoy. We have seen that Horgan and Tienson rely on this thesis in their determinacy argument for CP-determination (it corresponds to premise 3 of that argument). Strawson (in the quote in §1) also says that cognitive qualia are necessary to solving the Kripkenstein problem. Let’s focus on an example. Suppose that in the actual case while hunting Elmer has a visual experience of a rabbit and has the sensory phenomenology of attending to the whole rabbit, as opposed to a rabbit part. He says, “Lo, a rabbit.” He also has hunger pangs and imagines having rabbit for dinner. He is disposed to behave exactly as if he believes that a rabbit is present and desires to have it for dinner.

218

phenomenal intentionality

Next consider an absent cognitive qualia case in which the entire human population is identical to the actual population in all sensory and functional respects, but no one possesses any non-sensory cognitive and conative phenomenal properties.24 So in this case Elmer is just as I have described him to be in the actual case, only he lacks any cognitive and conative qualia. Now, arguably, in the absent cognitive qualia case as well as in the actual case, Elmer has some kind of rabbit beliefs and desires (see §2.2 for arguments that individuals in absent cognitive qualia cases are not “cognitive zombies”). The proponent of the necessity thesis need not say otherwise. What he must say is only that in the absent cognitive qualia case Elmer’s rabbit beliefs and desires are much more indeterminate in content than they are in the actual case. In particular, given what they say in the passage quoted in §1, Graham, Horgan, and Tienson are committed to saying that in the actual case Elmer determinately believes that a rabbit is present and desires to have a rabbit for dinner, whereas in the absent cognitive qualia case it is indeterminate whether Elmer’s belief and desire are about rabbits as opposed to undetached rabbit parts. So there is a difference in intentional facts across the cases. This case shows that the necessity thesis violates sensory-functional supervenience: the thesis that the total intentional facts about a population supervene on the total sensory and (wide and narrow) functional facts about that population. My argument against the necessity thesis is now this. In the absent cognitive qualia case as in the actual case, Elmer enjoys the sensory phenomenology of focusing on the whole rabbit (not just a part), he is disposed to respond to Quinean indeterminacy arguments by insisting that he has beliefs and desires about “rabbits” rather than “undetached rabbit parts,” the use-properties of his sentences are exactly the same, and so on. All the first-person and third-person evidence suggests that across these cases his belief and desire enjoy the same level of determinacy, contrary to the necessity thesis. Likewise, in an absent cognitive qualia case identical to the actual case in all functional and sensory respects (including experiences of counting objects and so on), all the first-person and third-person evidence suggest that our arithmetical beliefs have the same level of content determinacy that they have in the actual case. Therefore, contrary to Strawson, cognitive qualia are not necessary to answering the Kripkenstein problem. This supports sensory-functional supervenience: the total intentional facts about a population supervene on the total sensory and (wide and narrow) functional facts about that population, without any need for cognitive qualia.

3.4. Disembodied Cognitive Qualia Cases Finally, I have two related arguments against cognitive phenomenology involving what I shall call “disembodied cognitive qualia cases.”

Does Phenomenology Ground Mental Content?

219

My first argument casts doubt on the CP-existence thesis and has two steps. If there are cognitive qualia, then in the actual world they are embodied in the sense that they are accompanied by sensory properties, including experiences of having a body and acting on the world. The first step of my argument asserts that, if there are cognitive phenomenal properties distinct from all sensory properties, then disembodied cognitive qualia cases ought to be possible and indeed we should be able to positively imagine them. In such a case, we allegedly have a rich phenomenal life that overlaps with our actual phenomenal life because we have all the same cognitive phenomenal properties we have in the actual world; but our cognitive phenomenal properties are “disembodied” in the sense that they are not ever accompanied by any sensory properties. In other words, we have no visual experiences (including no experiences of having a body or engaging in apparent behaviors), auditory experiences, no mental imagery, no “inner speech,” and so on. Consider an analogy: auditory qualia are distinct from other kinds of qualia. So we can imagine cases in which such qualia occur in the absence of all other sorts of qualia.25 The second step of my disembodied cognitive qualia argument against the CP-existence thesis is that we cannot positively imagine such a case. At least I cannot. Just try. If the CP-existence thesis is true, then in such a case we have a rich phenomenal life that overlaps with our actual phenomenal life, only it is totally non-sensory. But what would it be like? Can you imagine this overlapping phenomenology? If you try to imagine what it would be like, you might imagine seeing all black, having an experience of inner speech (“nothing much is happening”), and so on. But then you will not be imagining a case in which you have cognitive phenomenal properties but no sensory properties. So the CP-existence thesis makes a false empirical predication about what we can imagine. In reply, the proponent of the CP-existence thesis might say that such cases are possible, but that for some reason we cannot positively imagine them. But why not? When it comes to other sorts of qualia, we can easily imagine cases in which they occur in the absence of distinct kinds of qualia: for instance, we can imagine having only auditory qualia. Now I turn to my second argument involving disembodied cognitive qualia cases. While my first argument was directed against the CP-existence thesis, my second argument is directed against the CP-determination thesis. Let Nemo be someone who enjoys a rich series of cognitive phenomenal properties of the kind we actually undergo but who has no sensory phenomenal properties at all. (For the sake of argument, let us just grant that the case is possible, contrary to my first argument.) Now, according to CP-determination, Nemo’s having such a rich series of cognitive phenomenal properties alone determines (perhaps just consists in) his having a series of our actual sophisticated narrow beliefs and desires: for instance, the belief that two plus two equals four (Strawson), the belief that a picture is hanging on the wall behind him (Horgan and Tienson), and so on. Hence, according to CP-determination, Nemo has a

220

phenomenal intentionality

rich series of occurrent narrow beliefs and desires, even though he has never had any sensory experiences whatever. But this can be ruled out a priori. Even if Siewert (1998: 277) is right that there are occasional cases of simple, purely non-sensory conscious thoughts in us (I would deny even that), the extreme case of Nemo is impossible. Like the baby discussed earlier, Nemo does not satisfy the a priori sensory-functional constraints on belief and desire. He has never had any visual or other experiences of walls or pictures or spatial properties (like being behind ), or sets of two or four objects. He does not use a language like English, or even have experiences of using a language like English. He does not have any dispositions to engage in physical behavior. In fact, he does not have any experiences of behaving in the world (for example, counting or adding). He does not even have any experiences remotely like these experiences, even if he has “cognitive experiences.” Intuitively, all this means that we cannot credit him with the belief that two plus two equals four, the belief that a picture is hanging on the wall that is behind him, and so on. Hence, even if the CP-existence thesis is true, consideration of disembodied cognitive qualia cases show that the CP-determination thesis is false. The proponent of CP-determination can avoid this argument only by accepting what I called systemic CP-determination. The idea would be that, even though cognitive phenomenal properties are distinct from all sensory phenomenal properties, they are somehow necessarily connected with sensory phenomenal properties, so that entirely “disembodied cognitive qualia” cases are impossible. But, as I said, there is no reason to accept this view and it requires brute modal connections. It is a bit like saying that, even though auditory qualia are distinct from visual qualia, they are necessarily connected with visual qualia. In sum, the CP-determination thesis holds that cognitive phenomenal properties, and hence some belief and desire properties, are entirely distinct from sensory properties. They form a distinct level of mental reality. But then they ought to be totally modally independent from sensory-functional conditions. But consideration of separation cases, altered cognitive qualia, absent cognitive qualia, and disembodied cognitive qualia suggest that this is not so. Indeed, these cases support sensory-functional supervenience: the total intentional facts about a population supervene on the total sensory and (wide and narrow) functional facts about that population, without any need for cognitive phenomenology. I will now sketch a view of this kind.

4. An Alternative: Phenomenal Functionalism The CP-determination thesis does not represent the only way of developing the plausible, broadly empiricist thought that phenomenology grounds mental content.

Does Phenomenology Ground Mental Content?

221

I will now sketch a modified version of David Lewis’s influential functionalist theory of intentionality, which I call phenomenal functionalism. It entails that sensory phenomenology grounds all determinate intentionality. I will not attempt to argue for Lewis’s holistic, functionalist approach; Lewis and others have done so and the virtues of a holistic approach to belief and desire should be clear from the preceding discussion. Instead, after briefly stating Lewis’s own functionalism, I will raise an overlooked problem for it involving the notion of sensory evidence. Then I will explain phenomenal functionalism and how it solves the problem. Finally, I will explain how it accommodates the main ideas of the phenomenal intentionality program while avoiding the problems I have raised for CP-determination. Roughly, according to Lewis’s own a priori functionalism (1974: 120), an individual has an (occurrent or standing) intentional state (for example, belief or desire) with content p iff it is part of the best interpretation (or all of the interpretations tied for best) that the individual has that intentional state. Lewis holds that the best interpretation is an objective notion: it is the assignment of intentional states that best satisfies a handful of general principles, given the functional facts about the individual and others of his kind. The principles are a priori truths about intentional states drawn from folk psychology.26 One principle is the behavior-rationalization principle: all else being equal, an individual tends to have beliefs and desires that make his behavior largely rational. But this cannot alone determine a reasonable best interpretation. Suppose Karl is in front of a round thing and reaches for it. One interpretation is that he wants a round thing, believes that a round thing is before him, and so believes that by reaching he will get it. Another, deviant interpretation is that he has a basic desire for a saucer of mud, believes that one is before him, and so believes that by reaching he will get it. Lewis therefore invokes a second principle, the humanity (or “charity”) principle. It has two parts. First, some basic beliefs and desires are objectively reasonable simpliciter, and people tend to have these. Second, some beliefs (and perhaps even some desires) are objectively reasonable, given one’s history of sensory experiences and evidence; and people tend to have these. This does not mean that a subject’s beliefs or experiences must be largely accurate: the beliefs of a brain in a vat are reasonable but false.27 Now return to Karl. Given that Karl is having an experience as of a round thing, the belief that there is a round thing before him is more reasonable than the belief that a saucer of mud is before him. And a basic desire for a saucer of mud would be unreasonable. So the humanity constraint will favor the first interpretation mentioned over the second, deviant interpretation. Indeed, Lewis relies on the humanity principle to rule out all deviant interpretations of belief and desire compatible with the behavior-rationalization principle. For instance, given the humanity principle, Lewis could easily rule

222

phenomenal intentionality

out Williams’s (2007) recent deviant global arithmetical reinterpretation of all thought and language (even perceptual reports like “that is square”), by noting that arithmetical beliefs are typically evidentially unrelated our perceptual experiences and other available evidence. 28 Using the humanity principle, Lewis can also rule out Kripke’s more local deviant interpretation of our arithmetical thought, involving the “quus” function. For Lewis explicitly says that his preeminent humanity principle concerning reasonable belief entails as a special case his much-discussed “naturalness principle,” which favors natural contents over unnatural ones.29 For instance, given our history of evidence, it is more reasonable to believe that all emeralds are blue than it is to believe that all emeralds are grue. In the same way, the humanity principle favors the assignment of more natural plus-contents over unnatural quus-contents. Thus the Kripkenstein puzzle is solved without “cognitive phenomenology.” 30 Now for the serious overlooked problem for Lewis’s view, which I will call the problem of evidence. Given Karl’s rich history of evidence, the humanity principle helps to rule out deviant interpretations. But what determines his evidence? At one point Lewis (1974: 112) speaks of “Karl’s life history of evidence according to [the physical facts about him],” suggesting that his evidence is simply part of the basic physical facts about Karl described in non-mental, non-intentional terms. But this is not true. Although there are different conceptions of evidence, Karl’s evidence clearly necessarily depends somehow on his history of experiences and their phenomenal characters. Further, there is strong reason to accept intentionalism about experience, according to which their phenomenal characters are determined by their very detailed intentional contents (Chalmers 2004; Pautz 2010c; Tye 2000). Thus the problem of evidence becomes an extremely difficult special case of the problem of intentionality. We might call it the problem of sensory intentionality. This is one element of the hard problem of consciousness. Now, since Lewis is physicalist and a global functionalist who applies his functionalism to all intentional states, in order to solve this problem he needs additional constitutive principles that go from the purely physical facts about Karl to the rich contents of his experiences. His behavior-rationalization principle and his humanity principle evidently only apply to beliefs and desires. Lewis never explicitly addresses the problem of evidence, that is, the problem of sensory intentionality. However, he does say something that suggests what his solution might be: “A state typically caused by round things before the eyes is a good candidate for interpretation as the visual experience of confronting something round” (1983: 374). Lewis also holds that an experience as of a red thing is a state that is caused by the color in fact possessed by pillar-boxes and other standard red things. This suggests a simple causal principle: very roughly, if, in the relevant population, state S would be caused by something’s being F,

Does Phenomenology Ground Mental Content?

223

in so-and-so range of actual and counterfactual cases, then S is an experience with the content that something is F is present. The idea is that suitable causal connections to the world pin down the rich contents of Karl’s experiences and hence (given intentionalism) what his experiences are like and what evidence is available to him. In turn, by the humanity constraint, Karl’s experiential evidence plays a crucial role in pinning down the contents of his beliefs and desires and ultimately language, by ruling out “deviant interpretations.” Thus, causal connections to the world very indirectly help fi x the contents of his beliefs and desires.31 It is worth mentioning that Lewis’s appeal to the actual and potential causes of a state is quite compatible with his well-known belief in narrow content and his intuitive motivation for it ( pace Stalnaker 2004: 211). Lewis explicitly says that, according to his conception of narrow content, it is not wholly intrinsic. For Lewis, narrow content is just content that is intuitively shared by individuals (for example, you and your Twin Earth duplicate) whose states are typically caused by the same external states and typically cause the same behavior across a range of actual and possible cases (1994: 425). Even if causal relations to the world determine content, it can be narrow in his very weak sense (Braddon-Mitchell and Jackson 2007: 240). However my view is that the causal solution to the problem of sensory intentionality fails, for two reasons. First, the simple causal principle is too simple. The problems are the usual ones for reductive psychosemantic theories: the disjunction problem, the depth (distance) problem, and other problems of detail. Further, as stated, the simple causal principle entails that all states that track the environment are experiences. This is wrong: Karl has perceptual beliefs and various sub-personal states that track the environment, but they are not experiences. These problems are especially serious for the a priori functionalist. You might think that, to solve these problems, he could just propose a much more sophisticated causal principle, appealing to things like asymmetric dependence, biologically normal conditions, and cognitive accessibility. The trouble is that such a principle would be far too recherché to be something that the folk implicitly know a priori. In fact, no such principle could be a priori for the simple reason that we evidently cannot a priori deduce what experiences Karl has merely from the causal and functional facts about him. My second problem for causal theories of sensory intentionality goes deeper. According to any causal theory, the content of experience, and hence (given intentionalism) the phenomenology of experience, can only be narrow in Lewis’s very weak sense: it is shared by duplicates whose states would be typically caused, in the relevant populations, by the same external conditions. But I agree with friends of the phenomenal intentionality program that we should accept some stronger version of phenomenal internalism. The only difference is that, while their arguments are based on dubious intuitions about brains-in-vats (Horgan,

224

phenomenal intentionality

Tienson, and Graham 2004), mine are firmly rooted in psychophysics and neuroscience. Since the phenomenal character of experience, and hence the content of experience, is internally determined in a strong sense, all causal theories of sensory content like Lewis’s are non-starters, even setting aside problems of detail. Indeed, elsewhere I have argued that none of our standard models for reducing intentionality apply to sensory intentionality (see notes 5 and 11). But while Lewis’s ambitious reductive functionalism may fail because it does not apply to the hard case of sensory intentionality, a less ambitious non-reductive version I call phenomenal functionalism may have some promise. Phenomenal functionalism gives up on the task of providing a functionalist or other reductive account of sensory experience and its intentionality. Here it takes an anti-reductive view, according to which experiences are irreducible intentional states. The anti-reductionist approach could take any of a variety of forms (Alston 1999; Johnston 2007; Levine forthcoming). My own view, non-reductive intentionalism, goes as follows (Pautz 2006, 2010a; Chalmers 2006). All experience (with the possible exception of undirected moods) is essentially intentional. For instance, Karl’s having an experience with the distinctive phenomenology of seeing a tomato is just a matter of standing in a special “experiential” intentional relation to a detailed intentional content involving simple perceptible properties like colors and shapes. I call this relation the “conscious-of relation.” (I do not mean we are literally conscious of contents, which are abstract objects.) What makes the view non-reductive is that it also holds that this relation is irreducible. Although what contents we are conscious of in some sense supervene on physical conditions (on my internalist view, internal physical conditions), the conscious-of relation is not identifiable with a relation characterizable in non-phenomenal, physical terms (for example, a causal or tracking relation). Phenomenal functionalism combines anti-reductionism about Karl’s sensory experiences with a functionalist theory of all of Karl’s intentional states other than his sensory experiences, for instance his occurrent and standing beliefs and desires. In particular, Karl has the belief or desire that p iff it is part of “best interpretation” that the agent believes or desires that p. Here the best interpretation is the one that best satisfies the a priori principles about beliefs and desires, given two sorts of facts about Karl: (i) his history of experiences, which according to non-reductive intentionalism are themselves fundamental intentional states; and (ii) the wide and narrow functional facts about Karl and others of his kind. The wide functional facts help determine the “wide” contents of his beliefs and desires. Thus, while Lewis’s functionalism is global, phenomenal functionalism is local, applying only to non-sensory intentional states. Since it takes sensory experiences to be irreducible intentional states, it agrees with Quine (1960: 221) that determinate intentionality is not reducible all the way down to third-person, physical facts.

Does Phenomenology Ground Mental Content?

225

On phenomenal functionalism, when it comes to the problem of intentionality, we start with a giant leg up, because Karl’s experiences, with their determinate world-directed intentionality, are among the fundamental facts about Karl. Karl’s experiences of the world, which determine his history of evidence, are anchor points that help to determine, via the humanity constraint, the contents of his downstream beliefs and desires and ultimately language, by ruling out deviant interpretations. Thus the present local functionalist theory solves the problem of evidence that plagues Lewis’s global functionalist theory. 32 Phenomenal functionalism should be attractive to friends of the phenomenal intentionality program for several reasons. (i) Phenomenal functionalism avoids the three main problems I raised for CP-determination. First, it does not require extra laws or “danglers” connecting the physical facts about an agent like Karl with his “cognitive qualia” and thereby his beliefs and desires. The picture is that the physical facts about Karl a posteriori determine his irreducible sensory-intentional states. Then, his irreducible sensory-intentional states, together with the functional facts about him, are enough to a priori determine his beliefs and desires, given general constitutive principles about belief and desire that we must all accept. Second, phenomenal functionalism accommodates and explains the necessary sensory-functional constraints on belief and desire: they follow from the real definition of the belief relation and desire relation in terms of the “best interpretation,” given the functional and sensory facts about an agent. Third, it accommodates the plausible thesis of sensory-functional supervenience: the total intentional facts about a population supervene on the total sensory and functional facts about that population. (ii) Phenomenal functionalism is also compatible with the claim of proponents of the phenomenal intentionality program that much intentionality supervenes on phenomenology alone. To illustrate, now suppose that Karl is a brain in a vat. He has an (illusory) experience as of a round thing. He does not exhibit physical bodily movements. But he attempts to behave as if a round thing is present, and he has experiences as of so behaving. Hence, by the humanity constraint and the behavior-rationalization principle and the other constitutive principles, he counts as having the (false) occurrent belief that a round thing is present. Indeed, maybe “round” in Karl’s language gets its content directly from the content of his (hallucinatory) experiences, by way of his accepting something like “round is that shape,” while demonstrating an uninstantiated shape property presented by his experience (see Speaks 2011 for an interesting discussion). And maybe the sentence “2+2=4” in Karl’s language gets a certain arithmetical content, thanks to its use and “inferential role” as well as considerations of naturalness. And maybe Karl counts as occur-

226

phenomenal intentionality

rently believing this content, on a certain occasion, by virtue of accepting and understanding this sentence.33 On my view, contrary to proponents of CP-determination (§1), there is no non-sensory phenomenal property that is alone necessarily sufficient for believing that two plus two equals four, or that there is a picture on the wall behind one. Rather, the contents of beliefs and desires are always determined by (i) a cluster of actual and dispositional facts about his past and present sensory experiences (including what sentences he accepts and their patterns of use), and (ii) his links to the environment (if such there be). (iii) Finally, phenomenal functionalism entails the following version of the core thesis of the phenomenal intentionality program that phenomenology is the source of all intentionality: necessarily, if a creature has never had the capacity for experiences (and does not belong to a community with such a capacity), it does not determinately stand in the belief or desire relation to particular contents (as I take it that we at least sometimes do). Phenomenology is at least the source of all determinate intentionality. To see this, imagine a community of zombies who have no experiences but who otherwise resemble us as much as possible. According to phenomenal functionalism, since the zombies have no experiential evidence, the all-important humanity principle cannot kick in to make it the case that deviant interpretations are mistaken. So, if we allow that zombies have beliefs at all, there is always massive indeterminacy concerning what their contents are. Many assert that phenomenology is the source of determinate intentionality; phenomenal functionalism explains why this is so. 34 I find the source thesis to be independently plausible. Here is an argument. An ideal interpreter given only the fundamental, third-person physical facts about a zombie that entirely lacks sensory experiences would not be able to deduce that they have beliefs and desires that determinately possess some contents and not others. Since the zombie lacks a rich phenomenal life, the ideal interpreter would have no way of ruling out “deviant interpretations.” I think that this epistemic gap provides some evidence that any such insentient system in fact has no determinate beliefs and desires.35 By contrast, if an ideal interpreter had access to the much richer set of basic facts concerning a sentient creature like Karl (including his first-person sensory experiences together with their rich, determinate intentional contents), he clearly could a priori deduce that he has at least some beliefs and desires with determinate contents. Phenomenal functionalism, then, has many virtues. However, while it might avoid “the problem of evidence” for Lewis’s theory, it does not dodge another overlooked problem for that theory (Pautz 2010a: 55). Like Lewis’s theory, it appeals to the nebulous notion of a “best interpretation.” Just what makes an interpretation “best”? Until this question is answered, phenomenal functionalism is incomplete.

Does Phenomenology Ground Mental Content?

227

5. Conclusion The CP-determination thesis is mistaken. Instead, reflection on various thoughtexperiments suggests that all intentionality is entirely grounded in functional facts as well as sensory and perceptual facts, which on my view are richly intentional and irreducible. The question is how all intentionality is grounded in such facts. I have just sketched an answer to this question, namely phenomenal functionalism, but it is programmatic. David Chalmers, Eric Schwitzgebel, and Jeff Speaks have proposed views in the vicinity.36 In any case, I think one thing we have learned is that it is not enough to simply maintain that “phenomenology grounds the contents of occurrent beliefs and desires,” as many proponents of CP-determination do. Those of us sympathetic to the phenomenal intentionality program need a systematic theory that tells us how phenomenology grounds content.

Acknowledgments I am very grateful to Phillip Goff, Uriah Kriegel, Angela Mendelovici, Galen Strawson, and an anonymous referee for many extremely good and helpful comments.

Notes 1. In Pautz (2006) and (2010b) I characterize the reductive externalist program in terms of a generic “tracking relation” which can be spelled out in various ways. Kriegel (in his 2011 and in this volume) characterizes the program in similar terms. 2. For phenomenal externalism, see Dretske (1995), Hill (2009), Lycan (2001), and Tye (2000). 3. See Quine (1960) and Kripke (1982). “Kripkenstein” refers to an imaginary proponent of the views that Kripke attributes (some think wrongly) to Wittgenstein. 4. My “phenomenology fi rst” slogan is inspired by Williamson’s (2000) “knowledge fi rst” slogan. Williamson holds that knowledge plays a foundational role in the epistemic domain and that it is (at least conceptually) irreducible. Likewise, proponents of the phenomenal intentionality program hold that phenomenology plays a foundational role in the mental domain and appear at least open to the view that it is (ontologically) irreducible. 5. In Pautz (2008), I reject prioritism but defend “integrativism” (in the sensory domain). Chalmers (2008), Mendelovici (2010), and Kriegel (2011: 63) defend the same view. Likewise, in number of papers (2006, 2010a, forthcoming) I reject the standard intuition-based, armchair arguments for phenomenal internalism (for example, Horgan, Tienson, and Graham 2004). But I also defend empirical arguments for a version of phenomenal internalism and against the reductive externalist program. For discussion of the empirical arguments for phenomenal internalism, see Chalmers (2005), Cohen (2009: 81–88), Hill (2012), and Tye and Cutter (2011). 6. I will ignore degrees of belief and desire. (Would advocates of CP-determination say that in some cases these too are constituted by non-sensory feel, rather than functional role?) 7. For this point, see Lewis (1994: 425) and Braddon-Mitchell and Jackson (2007: 240).

228

phenomenal intentionality

8. See Graham, Horgan, and Tienson (2007, 2009) and Horgan and Graham (2010). 9. The acquaintance model is in any case too simple (Pautz 2011). For instance, it has difficulty explaining why we have immediate introspective justification for believing some complex propositions about our experiences but not others (the “scope problem”); it also cannot explain the graded character of immediate introspective justification. 10. For the argument, see Pautz 2010a: sect. 7. I argue there against internalist reductions of the “consciousness relation” (Horgan, Tienson, and Graham (2004) call it the “acquaintance relation”), but the argument generalizes to internalist reductions intentional relations more generally. Goff (2012) also expresses skepticism about reductive internalist theories of phenomenal intentionality; however, unlike me, he does not develop an argument based on the unavailability of internalist reductions of dyadic intentional relations, which I think is crucial. 11. Even if he is a “non-relationist” about intentional states and indeed rejects the existence of intentional contents and other abstract objects (Kriegel 2011: chap. 3), the proponent of biological CP-determination must recognize the existence of some mental relations, namely mental relations to concrete items such as x has a thought that is accurate with respect to existing concrete scenario y. If he accepts that such relations supervene on the totality of physical facts about the world, and if he wants to avoid a primitivist view, he would at least need an account of these relations in purely physical, non-mental terms. 12. For the point that the only kinds of functional properties that might determine intentionality (and hence phenomenal intentionality, if such there be) are non-intrinsic and involve relations to the environment or behavior, see especially Braddon-Mitchell and Jackson (2007: 240). Kriegel (2011: 246) claims that his favored functionalist theory of phenomenal intentionality is compatible with the intrinsicness claim. But since Kriegel (2011: 73) explains phenomenal intentionality in terms of our states having the Dretskean function to indicate various externally instantiated response-dependent properties, and since Dretskean functions always require history (a non-intrinsic affair), this is unclear. See Pautz (2010b: 351ff ) for more discussion. 13. For discussion of the identity view, see Pautz (2008: 269), Pitt (2009), Mendelovici (2010), Kriegel (2011: 63). Th is view entails that slight differences in cognitive phenomenology between individuals (as it might be, between two people who say “Aristotle was smart”) would entail that they entertain slightly different “narrow” contents, even if they might entertain the same “wide content” (see Montague 2010 for this point). 14. According to the kind of view I will sketch in §4, all facts about beliefs and desires reduce to clusters of functional facts and sensory facts, but sensory facts are not in turn reducible to physical-functional facts. So my own view requires physical-sensory danglers: brute bridge principles connecting the physical with the sensory. My hope is that the correlations between the physical and the sensory are systematizable, so that only a handful of physical-sensory danglers will be required. My “simplicity” objection to CP-determination is that it is even more complex: it requires, not only physical-sensory danglers, but a whole slew of physical-cognitive danglers. Th at is a point against it. As we shall see, I have additional objections to CP-determination, which are independent of my present appeal to simplicity. 15. Indeed, elsewhere (Pautz 2010b: 351ff ) I argue that Kriegel’s (2011: 176, note 30) particular naturalistic theory of phenomenal intentionality (a higher-order tracking theory, where the tracked properties are response-dependent properties) faces serious problems of detail in addition to the usual problems, such as Kripke’s plus-quus problem and Quine’s inscrutability problem. 16. The proponent of CP-determination might grant that a wide range of bizarre separation cases involving a given cognitive phenomenal property are possible (not just the Charlie case, but the baby case discussed later), but insist that across these cases it determines the same narrow content. Th is is what Siewert (2011) calls “the part of what is thought that is fully determined by [cognitive] phenomenal character.” The proponent of CP-determination might suggest that my separation argument only shows that this is not a content of the sort proposed by Horgan and Tienson and Strawson that can be

Does Phenomenology Ground Mental Content?

17.

18.

19.

20.

21.

22.

23.

229

captured in ordinary language and that involves ordinary concepts like picture or two or plus. (Th anks to David Chalmers and Angela Mendelovici here.) But if he cannot even gesture at some specification of them, we have no reason to believe there are such contents. My claim that there are sensory-functional constraints on belief and desire does not amount to any crude behaviorism or functionalism; for instance, it is obviously quite compatible with the claim that Strawson’s (2010) weather watchers could have beliefs and desires. While I think that there are holistic a priori sensory-functional constraints on cognitive intentionality, I do not think that there are holistic a priori functional constraints of any kind on sensory intentionality (or at least those aspects of sensory intentionality fi xed by sensory phenomenology). See Pautz (2010c: 271). In my view, this is an important difference between cognitive intentionality and sensory intentionality. Horgan and Tienson in one place (2002: 526) appear to accept qualified CP-determination: necessarily, if an individual has a certain cognitive phenomenal property and “has a sufficiently rich network of actual and possible phenomenal/intentional states,” then the individual has a thought with a certain narrow content (for example, there is a picture on the wall, or two plus two equals four). Unfortunately, they do not explain what this means or why they say it. In any case, my Charlie case and my baby case are counterexamples even to qualified CP-determination. For Charlie and the baby do have a rich network of phenomenal/intentional states in addition to the relevant cognitive phenomenal properties (they both have many experiences and thoughts). Yet they do not have the occurrent thoughts which are allegedly necessitated by those cognitive phenomenal properties in the context of a rich network of phenomenal/intentional states. Indeed, if CP-determination is true, then there is a possible world where Elmer’s sensory-functional duplicate believes there is a spaceship when he sees a rabbit and says “there is a rabbit,” provided then he has the cognitive phenomenal property S that determines that belief. But this intentional difference would be unknowable as all the sensory-functional facts would suggest in this world he believes that a rabbit is present just as he does in the actual world. In reply, the proponent of CP-determination might say that the sensory-functional constraints on cognitive phenomenal properties are not brute. He might say that they follow from the simple fact that having cognitive phenomenal properties amounts to standing in certain intentional relations (occurrently believing or occurrently desiring) to certain contents, together with the fact that there are sensory-functional constraints on standing in such relations to contents. But this just passes the buck. The question then becomes “Why are there sensory-functional constraints on standing in such relations to contents—in other words, why are such intentional relations necessarily constrained by the sensory-functional facts?” Can this be explained by more basic modal truths? (To say that it is “part of their essence” would not be an explanation.) Siewert might offer a different reply to my separation argument (1998: 286). In one formulation of his claim about cognitive phenomenology determining intentionality, it just amounts to the claim that someone who “has, has had, and is disposed to have, experience with phenomenal character indistinguishable from my own” would have many thoughts. But this is compatible with the denial of CP-determination, and indeed with the non-existence of cognitive phenomenology. For Siewert’s total phenomenal twin has the same sensory experiences of conducting inferences in inner speech, the same sensory experiences as of his environment, and the same sensory experiences of acting on the world. Maybe it is such clusters of past and future sensory facts, and not his alleged “cognitive phenomenology,” which determine the contents of his beliefs and desires (see §4). Terry Horgan tells me that in forthcoming work he (independently) uses absent cognitive qualia cases, but unlike me he uses them to argue for the CP-existence thesis. It is worth mentioning that the allegedly “diaphanous” character of experience cannot explain why we cannot imagine having profoundly different phenomenal lives due merely to the absence of “cognitive experience.” After all, our visual experiences are arguably diaphanous in the sense that we cannot have object-awareness of them (Tye 2000: 51–52), but

230

24.

25.

26.

27.

28. 29.

30.

phenomenal intentionality this does not prevent us from being able to imagine cases in which they are absent but our other types of experiences are the same. For, as Tye points out, even if we are not aware of experiences in the sense that we do not have object-awareness of them (we cannot attend to our experiences in addition to attending to the objects and properties represented by experiences), we still are certainly quite aware of them in the sense that we effortlessly have fact-awareness to the effect that they are present. Everyone should believe in the possibility of absent cognitive qualia cases when they are understood in this way. If you think we actually fail to possess non-sensory cognitive and conative “qualia” over and above sensory qualia, as I am inclined to do, then you will think that the absent cognitive qualia case is simply the actual case. On the other hand, if you believe we actually possess such “qualia,” then you will also say that the absent cognitive qualia case is possible, but that it would differ phenomenally from the actual case in that we would lack qualia we actually enjoy. My argument will be neutral on this issue. By a “disembodied cognitive qualia case,” I do not mean a case in which an individual actually has no body or brain. Maybe an individual can have cognitive qualia only if he undergoes the right physical processing in the alleged cognitive qualia center of his brain. Even on this view, if there are cognitive qualia distinct from all sensory qualia, disembodied cognitive qualia cases should be possible in my sense. For there might be a very inactive brain which only ever undergoes activity in the cognitive qualia center but no other neural area. According to proponents of CP-determination, such an individual would have cognitive experiences but no other experiences, including visual experiences of the world or his body. (Compare: someone might only have auditory qualia throughout his lifetime.) In my view, two functionalist theories of mental intentionality can be extracted from Lewis. First, there is the best interpretation account described in the text. It uses only two or three general principles. And it does not necessarily provide functional definitions of our names for the belief and desire relation (or of the infi nitely many names of particular beliefs) of the form “the belief/desire relation is identical with the unique relation R that plays functional role F,” anymore than Lewis’s analogous best system theory of laws of nature provides functional defi nitions of names for laws. Second, there is the more familiar functional definition account (Lewis 1972: note 13), which does provides such functional defi nitions, based on a folk theory containing a huge number of specific platitudes (for example, “if someone tastes something bitter, they have the desire to spit it out”). Here I focus on the best interpretation account, because the functional defi nition account faces serious problems of detail pointed out by Loar (1981: 60). Loar himself (1981: 62) proposes to solve them by explaining the belief relation and the desire relation in terms of functions-in-extension, but I think his proposal fails. (Briefly, there will be infi nitely many “bent” functions that satisfy the functional definitions; and the proposal also faces the usual modal problems with trying to explain relations in terms of sets of their actual instances.) See Lewis (1983) and (1994). Lewis clearly would reject a weak subjective Bayesian interpretation of the humanity constraint. Indeed, he thinks objective Bayesianism is built into folk psychology. In a very helpful discussion, Weatherson (ms) independently suggests that Lewis could use his humanity principle to answer Williams (2007). See Lewis (1983: 375) and (1994: 428). Since Lewis derives the naturalness principle concerning mental content from his more general humanity constraint (which has nothing to do with “simplicity”), I do not think Lewis himself could accept Williams’s (2007: sect. 2) suggestion of deriving the naturalness principle concerning mental content from a general principle about simplicity as a theoretical virtue used to decide between theories that fit the data equally well. Here I am mainly focused on mental content but Lewis (1992) also appeals to naturalness in his theory of linguistic content in order to rule out “bent grammars.” Horgan and Graham (2010) raise an intriguing general objection to Lewis’s theory of intentionality: what makes it the case that naturalness enters into the determination of intentionality,

Does Phenomenology Ground Mental Content?

31.

32.

33.

34.

231

either in the mental case or the linguistic case? For instance, by virtue of what does naturalness enter into the determination of linguistic reference (or the semantic value relation)? But Lewis has an answer: naturalness helps determine reference because the reference relation just is a relation involving naturalness (the “most natural” grammar) among other things: call it the naturalness-plus relation. Maybe now Horgan and Graham will ask for an explanation of the identity between the reference relation and the naturalness-plus relation. But I do not think Lewis needs to supply one, because in general identities cannot be explained. Of course, it is sensible to ask for an explanation of the corresponding metalinguistic fact: the fact that “reference” refers to the naturalness-plus relation. And here Lewis can provide an explanation that invokes his own theory of reference: “reference” bears the naturalness-plus relation to the naturalness-plus relation. Th is kind of self-subsumption is not circular in any metaphysical sense (a point repeatedly made in the literature spawned by Putnam’s model-theoretic argument). An alternative solution to the problem of evidence would appeal to Karl’s behavioral dispositions. Maybe his behavioral dispositions, together with the behavior-rationalization principle, determine the contents of Karl’s immediate perceptual beliefs. Then, given the humanity constraint, the contents of his experiences (and hence, given intentionalism, what experiences he has) are those that make those perceptual beliefs rational. But this suggestion faces the problem of deviant interpretations (Pautz 2010a: 56–57). Moreover, it is at odds with Lewis’s solution to that problem. According to Lewis’s humanity principle, Karl’s history of experiences and hence evidence play a crucial role in constitutively determining the contents of his beliefs. Given this, what experiences and evidence he has cannot in turn be pinned down by his beliefs. Th at would be circular. To avoid circularity, his experiences and their contents must be determined in a belief-independent way, for instance, by causal connections to the world. On Kriegel’s intriguing and innovative version of interpretivism about unconscious intentional states, you have a certain unconscious intentional state iff (very roughly) you are disposed to get a fully informed ideal interpreter in ideal conditions to interpret you as having that state (for full details, see Kriegel 2011: chap. 4). There are important differences between phenomenal functionalism and Kriegel’s theory. First, phenomenal functionalism is more general, since it is meant to apply to all conscious intentional states as well, with the exception of sensory experiences. (Kriegel tells me that if he followed me in rejecting the idea that such conscious intentional states are constituted by cognitive phenomenology, then he would take a similar view.) Second, Kriegel’s theory is observer-dependent and so faces the usual prima facie counterexamples. Consider a community of killer believers, who have beliefs but are disposed to emit rays (or whatever) that destroy (ideal or other) interpreters when they are about to become informed about them. (Th is is analogous to Saul Kripke’s unpublished case of killer yellow.) Phenomenal functionalism avoids such problems because it does not appeal to an ideal interpreter. There are objective, observer-independent facts about your (conscious and unconscious) beliefs and desires and the semantic values of your expressions, because there are objective, observer-independent facts about what the best interpretation is (or, in case of ties, the set of interpretations tied for best). On Lewis’s own view (1974: 117), mental content is always prior to linguistic content. By contrast, as my examples here suggest, I think that in some cases (for example, perceptual content) mental content is prior to linguistic content, while in other cases (for example, arithmetical content) linguistic content is prior to mental content. However, as Lewis notes, his basic theory (and hence my modified version of his theory) is in fact compatible with a mixed view of this kind (he calls it “method 3”). Do our zombie twins count as having any beliefs or desires at all, despite not having the capacity for experience? Many would say they do. By contrast, I fi nd it intuitive that they do not. (Smithies (ms) develops an interesting argument against Zombie belief; but it relies on the CP-determination thesis that I have argued against.) Th is may be a semantic issue, since we agree about the fundamental facts of the case. However, even if we allow that the zombies have beliefs and desires (or beliefs* and desires*), the crucial

232

phenomenal intentionality

point is that phenomenal functionalism entails that they cannot have the same beliefs and desires that we have (since they lack acquaintance with conscious states and sensible qualities), and that it is at best radically indeterminate what their true contents are. 35. In response to this “epistemic gap” argument, those who think that zombies might have beliefs and desires that determinately possess certain contents might reply that it has such determinate beliefs and desires even if I am right that an ideal interpreter could not deduce them a priori from the fundamental physical facts about its situation. In other words, they might advocate a posteriori physicalism about the alleged determinate beliefs and desires of zombies. Soames (1998) takes a view of this kind about our own beliefs and desires. But to make this credible one would have to at least sketch physicalist theories (perhaps a posteriori physicalist theories) of the belief relation and the desire relation which explain how the purely non-phenomenal, non-intentional physical facts about the zombies determine that it comes to determinately bear these relations to some contents rather than others. The history of failed attempts to sketch such theories provides strong grounds for doubting that one can be provided. 36. Schwitzgebel (2002) defends “phenomenal dispositionalism.” Speaks (2011) proposes a Horwichian use theory, supplemented with facts about the contents of perceptual experiences. Chalmers (2008) fl irts with a “combined view” on which belief content is grounded in sensory/perceptual intentionality and inferential role. Unlike these authors, I provide a general analysis (in terms of the notion of a “best interpretation”) of what it is to stand in the belief relation or the desire relation to an arbitrary proposition. Th is provides a rule for going from the sensory-functional facts about any individual to what the individual believes and desires. I think that such a general algorithm is needed to explain how it is that an ideal interpreter, merely on the basis of his grasp of the concepts of belief and desire (and other relevant concepts), would in principle have the ability to work out any individual’s beliefs and desires, given the sensory and functional facts about that individual. Such abilities are not miracles (Braddon-Mitchell and Jackson 2007: 165–166).

References Alston, W. 1999. Back to the Theory of Appearing. Philosophical Perspectives 13: 181–203. Boghossian, P. 1989. The Rule Following Considerations. Mind 98: 507–549. Braddon-Mitchell, D. and F. Jackson. 2007. Philosophy of Mind and Cognition (2nd ed.). Oxford: Blackwell. Byrne, A . 2011. Knowing Th at I Am Th inking. In Self-Knowledge, ed. A. Hatzimoysis. Oxford: Oxford University Press. Carruthers, P. 2011. The Opacity of Mind. Oxford: Oxford University Press. Chalmers, D. 2004. The Representational Character of Experience. In The Future for Philosophy, ed. B. Leiter. Oxford: Oxford University Press. Chalmers, D. 2005. Representationalism Showdown. http://fragments.consc.net/djc/2005/09/ representationa.html. Chalmers, D. 2006. Perception and the Fall from Eden. In Perceptual Experience, ed. T. Gendler and J. Hawthorne. Oxford: Oxford University Press. Chalmers, D. 2008. The Critique of Pure Thought. http://consc.net/powerpoint.html. Cohen, J. 2009. The Red and the Real. Oxford: Oxford University Press. Dorr, C . 2007. There Are No Abstract Objects. In Contemporary Debates in Metaphysics, ed. T. Sider, J. Hawthorne, and D. Zimmerman. Malden, MA : Wiley-Blackwell. Dretske, F. 1995. Naturalizing the Mind. Cambridge: MIT Press. Fodor, J. 1994. The Elm and the Expert. Cambridge: MIT Press. Goff, P. 2012. Does Mary Know I Mean Plus rather than Quus? Philosophical Studies 160: 223–235.

Does Phenomenology Ground Mental Content?

233

Graham, G., T. Horgan, and J. Tienson. 2007. Consciousness and Intentionality. In The Blackwell Companion to Consciousness, ed. M. Velmans and S. Schneider. Oxford: Basil Blackwell. Graham, G., T. Horgan, and J. Tienson. 2009. Phenomenology, Intentionality, and the Unity of Mind. In Oxford Handbook of Philosophy of Mind, ed. A. Beckermann and B. McLaughlin. Oxford: Oxford University Pres. Hill, C . 2009. Consciousness. Cambridge: Cambridge University Press. Hill, C . 2012. Locating Qualia: Do They Reside in the Brain or in the Body and the World? In New Perspectives on Type Identity: The Mental and the Physical, ed. S. Gozzano and C. Hill. Cambridge: Cambridge University Press. Horgan, T. 2010. Materialism, Minimal Emergentism, and the Hard Problem of Consciousness. In The Waning of Materialism, ed. G. Bealer and R. Koons. Oxford: Oxford University Press. Horgan, T. and G. Graham. 2010. Phenomenal Intentionality and Content Determinacy. In Prospects for Meaning, ed. R. Shantz. Amsterdam: de Gruyter. Horgan, T. and J. Tienson. 2002. The Intentionality of Phenomenology and the Phenomenology of Intentionality. In Philosophy of Mind: Classical and Contemporary Readings, ed. D. Chalmers. Oxford: Oxford University Press. Horgan, T., J. Tienson, and G. Graham. 2004. Phenomenal Intentionality and the Brain in a Vat. In The Externalist Challenge: New Studies on Cognition and Intentionality, ed. R. Shantz. Amsterdam: de Gruyter. Johnston, M. 2007. Objective Mind and the Objectivity of Our Minds. Philosophy and Phenomenological Research 75: 233–68. Kripke, S. 1982. Wittgenstein on Rules and Private Language. Oxford: Blackwell. Kriegel, U. 2011. The Sources of Intentionality. Oxford: Oxford University Press. Levine, J. Forthcoming. On Phenomenal Access. In Themes from Block, ed. A. Pautz and D. Stoljar. Cambridge: MIT Press. Lewis, D. 1972. Psychophysical and Theoretical Defi nitions. Australasian Journal of Philosophy 50: 249–258. Lewis, D. 1974. Radical Interpretation. Synthese 21: 331–344. References are to the version reprinted with a postscript in D. Lewis Philosophical Papers: Vol. I (Oxford: Oxford University Press, 1983). Lewis, D. 1983. New Work for a Theory of Universals. Australasian Journal of Philosophy 63: 343–377. Lewis, D. 1992. Meaning without Use. Australasian Journal of Philosophy 70: 106–110. Lewis, D. 1994. Reduction of Mind. In A Companion to the Philosophy of Mind, ed. S. Guttenplan. Oxford: Blackwell. Loar, B. 1981. Mind and Meaning. Cambridge: Cambridge University Press. Lycan, W. 2001. The Case for Phenomenal Externalism. In Philosophical Perspectives, Vol. 15: Metaphysics, ed. J. E. Tomberlin. Atascadero, CA : Ridgeview. Lycan, W. 2009. Giving Dualism Its Due. Australasian Journal of Philosophy 87: 551–563. Mendelovici, A . 2010. Mental Representation and Closely Conflated Topics. PhD Dissertation, Princeton University. Montague, M. 2010. Recent Work on Intentionality. Analysis 70: 765–782. Pautz, A . 2006. Sensory Awareness Is not a Wide Physical Relation. Noûs 40: 205–240. Pautz, A . 2008. The Interdependence of Phenomenology and Intentionality. The Monist 91: 250–272. Pautz, A . 2010a. A Simple View of Consciousness. In The Waning of Materialism, ed. G. Bealer and R. Koons. Oxford: Oxford University Press. Pautz, A . 2010b. Do Theories of Consciousness Rest on a Mistake? Philosophical Issues 20: 333–367. Pautz, A . 2010c. Why Explain Visual Experience in Terms of Content? In Perceiving the World, ed. B. Nanay. Oxford: Oxford University Press. Pautz, A . 2011. Can Disjunctivists Explain Our Access to the Sensible World? Philosophical Issues 21(1):384–433.

234

phenomenal intentionality

Pautz, A . forthcoming. Phenomenal Intentionality: A New Argument against Materialism? In New Waves in Philosophy of Mind, ed. J. KalleStrup and M. Sprevak. Palgrave. Pitt, D. 2009. Intentional Psychologism. Philosophical Studies 146: 117–138. Pitt, D. 2011. Introspection, Phenomenality and the Availability of Intentional Content. In Cognitive Phenomenology, ed. T. Bayne and M. Montague. Oxford: Oxford University Press. Quine, W. 1960. Word and Object. Cambridge: MIT Press. Russell, B. 1912. The Problems of Philosophy. London: Williams & Norgate. Schiffer S. 2006. Propositional Content. In The Oxford Handbook of Philosophy of Language, ed. E. Lepore and B. Smith. Oxford: Oxford University Press. Schwitzgebel, E . 2002. A Phenomenal, Dispositional Account of Belief. Noûs 36: 249–275. Siewert, C . 1998. The Significance of Consciousness. Princeton: Princeton University Press. Siewert, C . 2011. Phenomenal Thought. In Cognitive Phenomenology, ed. T. Bayne and M. Montague. Oxford: Oxford University Press. Smart, J. C . 1959. Sensations and Brain Processes. Philosophical Review 68: 141–156. Smithies, D. MS. The Mental Lives of Zombies. Soames, S. 1998. Skepticism about Meaning, Indeterminacy, Normativity, and the Rule-Following Paradox. Canadian Journal of Philosophy, Supp 23: 211–250. Speaks, J. 2011. Can We Explain Linguistic Representation in Terms of Perceptual Representation? Unpublished manuscript. Stalnaker, R . 2004. Lewis on Intentionality. Australasian Journal of Philosophy 82: 199–212. Strawson, G. 2010. Mental Reality, 2nd ed. Cambridge: MIT Press. Swoyer, C . 1982. The Nature of Natural Laws. Australasian Journal of Philosophy 60: 203–223. Tye, M. 2000. Consciousness, Color and Content. Cambridge: MIT Press. Tye, M. and B. Cutter. 2011. Tracking Representationalism and the Painfulness of Pain. Philosophical Issues 21:90–109. Weatherson, B. Ms. The Role of Naturalness in David Lewis’s Theory of Meaning. http://brian. weatherson.org/papers.shtml. Williams, R . 2008. Eligibility and Inscrutability. Philosophical Review 116: 361–399. Williamson, T. 2000. Knowledge and Its Limits. Oxford: Oxford University Press.

12

Phenomenality and Self-Consciousness C h a r l e s Si e w e rt

1. Introduction Does consciousness essentially include self-consciousness? Some who say “yes” will eagerly press further: consciousness not only includes but is exhausted by a kind of self-consciousness—more specifically, by a way in which a mind refers to or represents itself. Thus it is to be explained in terms of a self-directed, inward-pointing intentionality. On the other hand, you may object that consciousness—at least in the phenomenal sense—either does not necessarily involve self-consciousness at all, or else not in a way that licenses us to absorb consciousness into self-representation; if anything, the order of explanation runs the other way: varieties of self-consciousness are to be understood as forms of phenomenal consciousness. These matters need more spelling out. But it’s fairly clear that how we deal with them will significantly shape our approach to the subject of this volume. Understanding how phenomenal consciousness relates to self-consciousness is central to understanding the relationship of consciousness and intentionality generally—which will do much to determine views about how either is to be explained. And it impinges on other broad issues: how to conceive of the basis of self-knowledge, and in what sense (if any) one is conscious of or experiences one’s “self”—which in turn bears on concerns about the nature and reality of the self. My opening question, however simple its wording, cannot be so simply approached; the variety and obscurity of the notions it invokes breed many complications. These shouldn’t put us off though—especially if (as I think) they provide much of what makes the topic fascinating. Here I will concentrate on some matters I think we should deal with first, given the current state of discussion. To get started, I need to make explicit a few basic points about how the word “conscious” will be used. 235

236

phenomenal intentionality

We may speak of our conscious states, or of our being conscious of something. More specifically, in the former case, some states of mind1 are said to be conscious in the aforementioned phenomenal sense: they are “phenomenal”; they have “phenomenality.” (I will explain how I interpret this notion momentarily.) And in the latter case, when we say you are conscious of something, we may speak of your being conscious of what is not a state of your own mind (for example, a noise or a glance), or of what is a mental state of yours (you may be conscious of thinking about something, or of seeing something, or of feeling somehow). Now, to say you are conscious of thinking (or seeing or feeling) might be taken to imply that you are somehow “representing your own state of mind” in the sense in which, for example, you are said to “mentally represent” whatever you are thinking about. But must one be—in some self-representational sense— conscious of one’s every phenomenal state of mind? This question provides my essay’s primary theme. However, we should recognize straight off that this isn’t the only way to ask whether consciousness is bound to self-consciousness. For it’s not clear that all self-consciousness involves the representation of one’s own mind. I may be conscious of my hands, and of how they are situated, and this too may be a kind of self-consciousness—a “consciousness of self.” And maybe some marginal or implicit consciousness of my own body or of my location is essential to my perceptual consciousness of the environment. But it’s hardly obvious that any of this involves a representation of my own mental state. In my final section I will say a little about perceptual self-consciousness. In most of what follows, however, I will be asking specifically about how phenomenality relates to having a mind that represents itself. According to “higher-order” theories of consciousness, the relationship is very close indeed: phenomenality is nothing more than a species of the mind’s self-representation, and is to be explained as such. Prominent defenders of such views include Carruthers (2000, 2004, 2007), Lycan (2004), and Rosenthal (2002). As I use the terminology, it is enough for a particular state of mind of kind R to be a “higher-order representation,” if this instance of R represents one to have some representation of a distinct kind R*. On this understanding, represented and representing states could be identical; conscious states of mind could be “higher-order” just by representing themselves. (This puts my use of “higher-order” at odds with that of some, such as Kriegel (2009), but I don’t think this affects the substance of the discussion.) At least some higher-order accounts of consciousness seem to lead to a vicious infinite regress of self-consciousness. To give the gist of the concern: such theories appear to say all consciousness must be self-conscious. But if that’s so, then doesn’t that include any such self-consciousness—mustn’t one be conscious of it as well? And are we then to say one is also conscious of the consciousness of self-consciousness, and so on, without end? This looks like

Phenomenality and Self-Consciousness

237

an unhappy result. And so we may be inclined to conclude consciousness is not always self-conscious after all—leaving higher-order theories in the lurch. But their defenders are typically unimpressed with such regress-based concerns, and see them as easily and decisively dispatched. When Lycan, for example, argues for his “higher-order perception” theory, he very briefly replies to the regress worry as one of a pair of “bad and ignorant” objections before going on to those he regards as more serious (Lycan 2004: 95). But I share with Kriegel the belief that regress problems are serious, and not so casually batted away. Unlike Kriegel, however, I think (and will argue here) that a proper response to them significantly undercuts support for the general idea—integral to many contemporary theories of consciousness—that phenomenality essentially involves the mind’s self-representation. I take my criticism then to bolster the view that no intentionality distinctive of self-consciousness explains phenomenal consciousness, and that we should instead understand certain forms of self-consciousness as forms of phenomenality.

2. Phenomenal Consciousness What is meant by “phenomenal consciousness”? The notion is sometimes introduced along these lines: states that are phenomenally conscious are states there is something it’s like for one to be in—in a sense at work when you say you know what it’s like for you to feel dizzy, or see colors, or hear a tune “in your head” Also, phenomenality (or a certain form of it) is said to be what’s conspicuously and revealingly missing in the condition known as blindsight. I myself endorse both of these points of entry. But we must proceed with caution, for their proper interpretation is not entirely clear, and we may construe them in ways that keep us from getting the initial neutrality we need in order to critically examine what’s at issue here: the idea that a mind’s self-representation is essential to (or even constitutive of) its phenomenality. In other words, we may from the start conceive of the topic in ways that make the truth of this deservedly controversial thesis seem a foregone conclusion. However, we can and should clarify what may legitimately be meant by “phenomenally conscious” without building in such bias. In this section I will first explain how to do this in connection with “something it’s like” talk, and then turn to the notion of blindsight. Ned Block puts the relevant idea like this: “what makes a state phenomenally conscious is that there is something it’s like to be in it” (Block 2002: 206). I would add: “something it’s like for the one who is in it.” But by itself this sort of formulation seems to me too obscure and liable to divergent interpretations to give us an adequate starting place. Considering a difficulty to which Block’s principle gives rise will, I think, help us to explain the notion of phenomenality

238

phenomenal intentionality

a bit more clearly, in a relatively unbiased way, via talk of “what something is like for someone.” Recall what Nagel (1974) says in his influential invocation of “something it’s like” talk to train our attention on the relevant understanding of consciousness. He suggests that coming to know what it’s like to be a bat would require “adopting a bat’s point of view”—if only in imagination. I propose to endorse this conceptual link between knowing what it’s like and adopting a point of view, while taking the focus away from wondering what it’s like to be an alien sort of creature. To see what I mean—suppose, for example, that I tell you that I ate durian fruit for dessert. You may intelligibly respond, “I’ve never eaten durian—I wonder what it’s like.” The curiosity you would express is not out of place. There was indeed something it was like for me to eat a durian—and seemingly in a sense relevant here. For there seems to be “something that it’s like” to be in a type of state, provided that it is one about which it is suitable to express a subjective curiosity, or claim a subjective knowledge—one that can be satisfied or possessed only by occupying (or being able imaginatively to occupy) the position of a subject of that state. Since there is something that answers to such curiosity about durian eating, there is something that it’s like to eat a durian. However—and this is the difficulty—we may reasonably think that eating durian is not itself a phenomenally conscious state in the sense in which tasting and smelling durian are. And yet, according to Block’s dictum, the fact that there’s something it’s like to eat durian should make it a phenomenally conscious state. Similar remarks would apply if we consider weighing three hundred pounds or having a conjoined twin. These can be fit targets of subjective curiosity—we can sensibly wonder what it’s like to have these features. But then are their instances themselves phenomenally conscious states? It would not be good to start out assuming so. For one thing it might seem counterintuitive to count eating a durian (and weighing three hundred pounds, etc.) among conscious states. To this one might respond that the assumption is problematic only if it tells us that all such states are conscious. Maybe it’s just fine to say, for instance, that in some cases eating is a conscious activity (and weighing 300 lbs. a conscious state)—those cases where there is something it’s like for the eater (for the heavy person). And Block’s principle should be taken to concern only what makes something a token or instance of phenomenal consciousness. It does not imply that every type of state that it says has conscious instances (since there’s something it’s like to be in some instance of that type), has only conscious instances. It does not assume that instances of these kinds are invariably conscious states, and thus suitable targets of subjective curiosity, regardless of what other features accompany them. Maybe, for instance, if ants eat durian their eating is not conscious, for there is just nothing there that would answer to a subjective curiosity about what that eating is like for them. And (panpsychism

Phenomenality and Self-Consciousness

239

aside) it seems out of place to wonder what it’s like to weigh three hundred pounds . . . for a bag of rocks. Perhaps there’s nothing wrong with counting eating (and so on) as (sometimes, contingently) conscious states in some sense. But it’s not initially clear whether this is the same or different sense as that in which tasting, smelling, and feeling are conscious—and if they are different senses, what the relationship is. So I don’t think accepting this usage would give us a good start on explaining what it means to be in a conscious state. And it leaves the question: since types such as eating x or weighing n pounds don’t make their instances conscious on the occasions when they are, what distinguishes the types that do? What sorts of features do make it the case that there is something it’s like for one to be in some state? It seems we still need a better grasp of these—what I’ll call “phenomenal features”—if we’re to understand what makes a state a phenomenally conscious one. What we have to work with so far is the idea that, where some features are concerned, there is something it’s like for one to have them only non-essentially, due to the accidental presence of other, detachable features. In the durian eating case, those other features—distinct from eating durian—are plausibly some already mentioned: its tasting and smelling somehow to the durian eaters, and its feeling somehow in their mouths. Durian eating could occur without these, and if it did there would be nothing it was like for the eater. Moreover, even if it could not occur unexperienced, we may still want to say there is only derivatively something it’s like to eat durian. Suppose, for example, you ate fake durian that tasted, smelled, and felt just the way real durian does in the eating of it. Then you could know what it’s like to eat durian by knowing what it was like to have the taste and other experience you could get from eating the fake stuff, and knowing (because you’ve been reliably informed) that real durian also would taste (etc.) the same way to you. And only by having (or by being able to imagine having) the sorts of experiences that could be had even with fake durian could one know what it’s like to eat real durian. So there is some feature or features which could be had without eating durian—for example, the aforementioned taste, smell, and touch appearances, such that you could know what it’s like for someone to eat durian, just in case you knew what it was like to have these other features (and some further conditions held which didn’t consist in having “what it’s like” knowledge). If that is the case, I want to say that there is only derivatively something it’s like for one to eat durian. And this would still be so, even if there were essentially something it was like. But now this leads us to ask: what about these features: features of something’s tasting, smelling, and feeling somehow to you? Is there also only non-essentially or derivatively something it’s like for one to have them? We can interpret talk of such features so that the answer is no. We might reasonably think: something’s tasting somehow to you does not need to have added to

240

phenomenal intentionality

it some further feature which—were it taken away—would leave such tasting no longer suitable for one to claim or desire knowledge of what it’s like for the taster. Nor is there some feature you could have in the absence of such tasting, such that you know what it’s like for durian to taste to you as it does, just in case you know what it is like to have that feature to which tasting was unnecessary. Thus we can interpret “taste” (and other appearance-talk—“looks,” “feels,” etc.) in such a way that there is essentially and non-derivatively something it’s like for one to be “appeared to” in a certain manner. But the crucial point (and the key to the initial conception of consciousness we require) is this: to have a “phenomenal feature” is to have a feature there is essentially and non-derivatively something it’s like for one to have. Any instance of such a feature is a phenomenally conscious state. Can we say anything further about the meaning of this “what it’s like” locution I’ve been relying on? We can. Briefly put: there is something it’s like for one to have a given feature just when that feature is suited for one to claim or desire a knowledge of what feature it is, which requires one either have it oneself, or be able to imagine having it (hence this is a “subjective” knowledge), but which does not require one be able to give a theoretically satisfying account of what constitutes the possession of this feature (hence this is a “non-theoretical” knowledge). Thus on the conception of phenomenality I advocate, phenomenally conscious states are instances of phenomenal features, that is to say, features essentially and non-derivatively suited for one to claim or want a subjective, non-theoretical knowledge of what features they are. And I would add: whenever such features differ phenomenally (that is, whenever their instances differ in phenomenal character, whenever what it’s like for one to have them differs), then what the difference is also is thus suitable for subjective non-theoretical knowledge. Elsewhere (Siewert 2011) I discuss in more detail this conception of phenomenality. Here I hope to have said enough to offer an interpretation of “what it’s like” as a way into phenomenal consciousness that does not immediately impose the notion that phenomenally conscious states must always somehow involve the mind’s self-representation. It is worth making this explicit because there are ways of construing “what it’s like” talk that do commit us in this regard. For example, you may think, with Peter Carruthers (2000, 2004), that this locution is to be seen in the light of a distinction between “what the world is like” for you (determined by how it appears to you) and “what your experience is like for you” (which constitutes its phenomenal character)—where you then go on to take the latter to be determined by how your experience is “presented” to you, in some sense like that in which the world appears to you. In this way higher-order representation may simply be built into one’s construal of “what it’s like” talk. But there is an alternative. My understanding of “what it’s like” allows us to say there is essentially and non-derivatively something it’s like for you

Phenomenality and Self-Consciousness

241

whenever a color looks to you as it does. And meeting that condition would be enough to make its looking to you just as it does a phenomenal feature. On my view, “what the experience is like for you” (its phenomenal character) is none other than the color’s looking to you just the way that it does. For “what the experience is like” is simply the manner of appearing that constitutes the experience. And to say this is an appearance “ for you” is to say that something appears in this very way to you. In other words, what the experience is like for you is what manner of appearance this is to you. This does not tell us that the phenomenal character is determined by something higher-order, distinct from and parallel to first-order appearances. We may want to add that what it’s like for you to have an experience consists in “how you experience it.” But again, it’s not obvious that it needs to be interpreted in higher-order terms, any more than does speaking of how a feeling of pain feels—where the feeling does not thereby appear as an object distinct from a manner of appearance. The type of feeling felt and the manner of feeling it simply coincide. Just so: how you experience the appearance of a color, and precisely how the color appears to you, may simply be identical. I’m not claiming that my interpretation of “what it’s like” talk shows that the mind’s self-representation is not built into phenomenality. For all I’ve said, that’s still possible: nothing thus far rules out the idea that buried within all phenomenal vision is a representation of a visual representation. The point is just to offer a legitimate interpretation of “what it’s like” talk that does not assume that this is the case. You may want to say that durian’s tasting as it does to someone is essentially and non-derivatively an apt target of subjective curiosity only if it entails that the taster internally represents the experience by being aware of it. All I’m saying is that we can grasp an interpretation of “tastes” that makes what it picks out phenomenal by the “something it’s like” standard, while leaving that issue undecided. Now let’s turn to what I mentioned as the second way to initially clarify the very idea of phenomenality—by reference to blindsight. This is a condition in which subjects suffering damage to the visual cortex deny seeing types of visual stimuli in circumstances where—pre-trauma—they would have affirmed it, even though (when “forced” to select from a list of set options) they show the ability to successfully identify which type of stimulus has been exposed. Here is one way to interpret what is going on. To say that such subjects have “blindsight” is to say that in one sense they do see the relevant stimulus, and in another they do not. Thus, in a sense they see something, and in a sense they are blind to it. How could this be so? Well, first, consider a specifically visual sense of “look” in which no object in a pitch dark room looks any way at all to a person. Second, interpret “see” in such a way that a person cannot be rightly said to see something that looks to her no way at all. Then, regarding blindsight, we say: in this sense, the blindsighter correctly denies seeing the stimulus, even though

242

phenomenal intentionality

she correctly discriminates it (in verbal judgments, in movement) because retinal stimulation from it triggers activity in what’s left of her visual system. So in a sense she’s blind to the stimulus—it doesn’t look anyhow to her—and in another she sees it. For the kind of discrimination she does have could also be regarded as a kind of “seeing.” Further refinements on this conception are possible. But for present purposes what I’ve said is enough to convey the basic strategy of using blindsight to clarify the meaning of “phenomenality.” Phenomenality is that feature exemplified in cases of something’s looking somehow to you, as it would not be in blindsight as just conceived. Necessarily, any instance of its looking somehow to you is a phenomenally conscious visual state. The interpretation of “look” that figures in these reflections on blindsight is none other than that on which we earlier found something’s looking somehow to you to be a feature essentially and non-derivatively suited for one to claim or desire a subjective, non-theoretical knowledge about what that feature is. There—in connection with “what it’s like” talk—I contrasted my interpretation with one that would commit us to linking phenomenality essentially to the mind’s self-representation. Similarly here—in connection with blindsight—I would also distinguish my interpretation from one that at least strongly suggests this linkage. You might construe blindsight in my fashion, as a deficit in consciousness where, in one sense you do see a stimulus, and in another you don’t. But you might also construe it differently. You might say that in blindsight you see the stimulus in the very same sense as you do in the normal case, though you are just not conscious or aware of your seeing it (and so you are “blind” only to your own seeing). If we accept only the second construal, and we assume, as many do, that to be conscious of your seeing must be to represent it somehow, it will seem inevitable that the phenomenality missing in blindsight necessarily has somehow to do with the mind’s self-representation. Thus, for example, Smith (2005: 94) introduces his notion of the inner awareness (the reflexive self-representation he holds is built into every conscious experience) by contrasting this “with blindsight, where I saw but had no awareness of seeing.” And Carruthers (2004: 127) characterizes “blindsight percepts” as “ones to which the subjects of the states are blind, and of which they cannot be aware.” Thus the difference between the phenomenal vision blindsighters lack and the kind we have is immediately construed in terms of the lack or presence of an awareness of—hence a representation of—one’s own seeing. But we have seen that we do not have to interpret blindsight only in this way. We can leave open the question of whether something’s looking somehow to you (missing in blindsight) necessarily brings with it a representation of your own visual state. And we can interpret “looking somehow to someone” so that it is sufficient for phenomenality. So again, we can identify the notion we want without simply taking it for granted that phenomenality is a matter of having a self-representing mind.

Phenomenality and Self-Consciousness

243

2. The Regress Challenge Now we have an understanding of the notion of phenomenal consciousness that does not depend on precipitously endorsing the claim I want to examine. That claim is: The SRM (Self-Representing Mind) Thesis. Necessarily, a state c of a subject S is phenomenally conscious, only if S has the appropriate sort of mental representation of c. Among accounts committed to SRM I would include the self-styled “higherorder” theories of consciousness cited earlier—such as we find in Rosenthal, Lycan, and Carruthers—as well as Kriegel’s (2009) “self-representational” theory. These views differ among themselves in many significant ways. But what I now want to emphasize is something they share, the appeal to a principle routinely invoked in introducing and supporting their approach (and which crops up repeatedly in discussions of consciousness generally). Th is is put forward as something commonsensical or intuitive, or at any rate not merely stipulative—a basic pretheoretical commitment our theories should respect. The CO (Conscious-Of) Principle. One is in a conscious state c, only if one is conscious (or aware) of c. One can see this principle explicitly at work in the writings of higher-order theorists. After classifying various forms of higher-order theories (including his own), Carruthers (2007: 278) says they all “have as one of their main motivations the intuition that conscious mental states are representational states of which the subject is aware.” Rosenthal (2002: 407) speaks approvingly of what he describes as Descartes’s conformity to the “commonsense observation” that “the states we call conscious” are “states we are immediately conscious of”—which observation he says “provides a useful start toward a theory of consciousness.” In this context, I will (like Rosenthal) treat “conscious” and “aware” as interchangeable, and the two versions of the principle they would provide as mere terminological variants. (Later I will comment on the idea that we should endorse only the “aware” version of the principle.) We can readily see how CO is thought to support SRM. Suppose your visual perception of a given stimulus is conscious. Then (according to CO) you are conscious of seeing the stimulus. And (so the story goes) to be conscious of seeing is to form some sort of representation of your visual state. For the “of” in “conscious of seeing” is the “of” that indicates what follows is an object of representation. Thus, if we treat CO as a necessary truth regarding phenomenal consciousness,

244

phenomenal intentionality

we recognize that phenomenally conscious states must be represented by the very mind in which they occur; that is, we recognize the truth of SRM. Now here is another principle that we might find commonsensical or intuitive. The IC (Is-Conscious) Principle. Every state of being conscious of something is a conscious state. In support of this, consider the sorts of cases that are used to argue for unconscious perception. Blindsight subjects, for example, are said to have unconscious visual perception of the stimulus they correctly report. Now apparently it’s easy to assume that subjects who have only unconscious visual perception of a stimulus are not visually conscious of it. For example, Merikle, Smilek, and Eastwood (2001) canvass evidence that subjects are sometimes “[visually] unaware of critical stimuli” that they do, nonetheless, perceive. And these psychologists call this “perception without awareness,” perception where there is, “in other words, unconsciously perceived information” (117). So for them, an unconsciously perceived x is simply an x perceived without an awareness of x. But then it seems, a consciously perceived x will just be an x perceived with an awareness of x. No difference is recognized between the conscious perception of x, and being perceptually aware/conscious of x. This should not be so surprising. It would be hard, I think, to find any instances of putative unconscious thought about or perception of something that people generally would find it “intuitively” or “commonsensically” correct to describe as cases in which subjects are conscious of something by unconsciously thinking of it or unconsciously perceiving it. To be conscious of x by perceiving it is to consciously perceive it. But then (assuming that to consciously perceive something is to have a conscious perception of it) the IC principle also expresses some ordinary assumption about consciousness: being conscious of something is a conscious state.2 However, once we’ve got both CO and IC on board, there is at least apparently a regress problem. For we will say (according to CO): if you have a conscious state c, then you are conscious of c. But (according to IC), your state of being conscious of c will also be conscious. And thus (applying CO again) one will be conscious of being conscious of c—and so on, ad infinitum. This is a problem, at least if we construe being conscious of c as forming a mental representation of it, and we assume that we do not in fact instantaneously generate some infi nite hierarchy of conscious representations of representations for every measly conscious state we are in, every moment of our waking lives. Now I only say that there is apparently a regress problem, because there are various ways to block the regress. The question is: which of various ways of doing this is most defensible? Contemporary higher-order theorists often respond to this concern by simply rejecting the IC principle. There is no regress, they say, because the

Phenomenality and Self-Consciousness

245

higher-order mental states that make their representational targets conscious are themselves typically not conscious states. They may allow that, most of the time I’m awake, I have conscious visual perception of what is before me without forming conscious thoughts about my visual states, or consciously internally sensing them. Nevertheless, they will say, I am during those times conscious of my seeing by thinking unconscious thoughts (or unconsciously sensing) my own visual states. This, however, leaves me wondering: why is CO such a rock solid first principle, while IC is so utterly dismissible? We are told it is counterintuitive or uncommonsensical to think one can be in a conscious state of which one is in no way conscious. But it seems to me at least as strange to suggest that one can be conscious of something by having a totally unconscious perception or thought of it. On the face of it, IC has at least as much initial claim on our allegiance as CO. Why not instead sacrifice CO to IC? Or why not seek and defend interpretations that would allow us to retain a right to claim both consistently? Since “Is-Conscious” is as least as intuitive as “Conscious-Of,” if we are to give weight to intuitions in these matters, we should give at least as much to the former as to the latter. So if there are accounts that avoid the regress while preserving both principles on some reasonable interpretation, these are to be preferred. Now some higher-order theorists may respond that they simply do not and never have found IC at all intuitive, and owe it no respect. Or they may say that even if it is intuitive for them that you can’t be visually conscious of something by having an unconscious visual perception of it, they have all along sharply distinguished consciousness from awareness, and find nothing odd in the idea that one might be visually aware of something by unconsciously seeing it. Thus it would seem intuitively right to them say that, while of course blindsighters are not visually conscious of the stimulus, they are visually aware of it. And it is specifically just the principle that one is (perhaps blindly) aware of one’s every conscious state that they find intuitive or commonsensical. But I don’t think one should be complacent about such responses. For those (like myself) whose intuitions do not sufficiently match up with these, the theorists’ invocation of CO to motivate their accounts will be unsatisfying, until the apparent conflict with IC has been dealt with in a non-dismissive manner. And if intuitions diverge on crucial points, this casts doubt on the claim that either the “awareness only” interpretation of CO or the rejection of IC should really count as commonsensical, or as some datum to be honored. Moreover, to borrow an important point from Kriegel (2009)—if IC is false, or CO often true only in virtue of some unconscious awareness of our mental states, then it becomes mysterious how CO could have ever seemed to us like a plausible principle. That is, it wouldn’t seem intuitively or commonsensically plausible to us that we are always conscious (or aware) of our own conscious vision, if we usually were introspectively oblivious to this very fact about ourselves. But

246

phenomenal intentionality

we would be introspectively oblivious to being conscious/aware of seeing if this were unconscious. Given this, we should explore the possibility that we can reasonably preserve both CO and IC in a way that avoids the regress, without the “aware, not conscious” move, while still maintaining CO supports SRM—the claim that a self-representing mind is essential to phenomenality. Here is a way to try to do this. Start again with the thought that, necessarily, every phenomenally conscious state is a state one is conscious of (in a sense that entails one mentally represents it somehow). But now—don’t let go of IC: one’s being conscious of this state is also itself a conscious state. However, now let’s argue: contrary to appearances, this doesn’t entail that one is conscious of being conscious of it—and so no regress gets started. For what is required for your representation of your mental state to be conscious is not that you are also conscious of such self-representation. No, all that is required is that there is some type M of which the state of representing your own mind is also an instance, and that state of self-representation can be correctly reported by saying: you are conscious of M. So, consider for example: you hear a tone. Th at hearing is conscious. By CO this implies that you are conscious of hearing a tone. And (by IC) your being conscious of hearing a tone is conscious too. But that does not require that you are conscious of being conscious of hearing a tone. No—it only requires that there is some type M such that your being conscious of hearing a tone is also an instance of M, and you are conscious of M. And lo and behold, there is such an M, namely: hearing a tone. In other words: when your hearing a tone is conscious, your being conscious of hearing a tone is also conscious, not because you are also conscious of being conscious of hearing a tone, but because your being conscious of hearing a tone is somehow one with the very state of which you are thus conscious—your hearing a tone. This general way of cutting off the regress—by saying that the conscious state in some way refers to or represents itself—is not new. It is, I believe, more or less the position advocated in Brentano (1972), who was influenced partly by his interpretation of Aristotle. More recently Williford (2006) and Kriegel (2009)—acknowledging the historical affi nities—have articulated and defended in detail views in the neighborhood. However, to assess this general strategy, I think we need to ask: just what species of conscious self-representation is allegedly built into every conscious state? There seem to be three basic options. (a) This self-representation is some form of conscious thought, expressible in an assertion, capable of playing a direct role in voluntary inference. (My hearing a tone is not just a hearing but—at least in part—a thought about itself.)

Phenomenality and Self-Consciousness

247

(b) The self-representation is, by contrast with this, somehow broadly speaking perceptual or sensory in nature—and constitutes a kind of conscious “inner” perception or sense. (My hearing the tone is also a sensing or perception (but of course not a hearing) of itself.) (c) It is neither of these, but a sui generis form of conscious representation, specific to consciousness of one’s own mind. (My hearing the tone is also a representation of itself that is neither a thinking nor a sensing.) But each of these options is problematic. A problem with (a) is that I find that consciously thinking about my conscious experience is only a “sometime thing”— hardly ubiquitous. So I would say, on the basis of first-person reflection, that while I can and do begin to consciously think about how something looks to me (and in so doing am thinking about its looking to me as it does), when I do this I am only reflecting on an experience of a type that had already been happening. It’s not as if things in my surroundings just start looking somehow to me the moment I consciously reflect on their looking somehow to me (as if previously I had been operating with blindsight). I can recall that they already were visually apparent to me. But I don’t recall consciously thinking about them at the just-previous time—the reflection seems to be something new. Such reflection needn’t be very elaborate or articulated: I may just suddenly be “struck by” or “just notice” how something looks to me (the way moving shadows on the window jamb look to me just now, in the sunset)—and I may articulate this (if at all) only by saying something like, “I just noticed how that looks to me.” And it seems plausible to suppose that such noticing is not a wholly distinct state, detachable from the “looking” noticed. For how could I even think the specific thought I would express with the phrase, “how that looks to me,” unless it just then did look to me the way I was at that moment thinking of—even if only in illusory or hallucinatory fashion? So we have here a relatively inarticulate sort of conscious thought, intimately joined with the experience it is about in some fashion. But this thought arises as something new, something which hadn’t been happening just before—though visual consciousness was already there. So anyone who would defend option (a) needs to say that even prior to this conscious noticing, there was already some other sort of conscious thought also intimately joined with the experience I was only a moment later consciously struck by. But just how am I to distinguish this from the conscious noticing, the “being struck by”? Both thoughts would be conscious, both relatively inarticulate, both dependent on the experience they’re about. Are we to say that we can recall in some sense having been already “marginally” or “peripherally” consciously thinking some thought we are “focally” thinking when “noticing” happens?3 I am not sure how to make sense of the idea that I fi rst “marginally” or “peripherally” consciously think that p, and then “focally” do so, unless this means that the thought first occurs to me inarticulately (and thus as yet only

248

phenomenal intentionality

in some “dim” way), and then I verbally articulate it (achieving a fuller grasp of what I am thinking). But clearly this isn’t the contrast we’re looking for, since I may be (newly, consciously) struck by how something looks to me without yet articulating it (if we admit the occurrence of wordless thoughts at all). It’s true I would say I am at least sometimes (non-sensorily) marginally conscious of something or someone. For example, while I am thinking (“focally”) about what to say in an email I am writing, I am “marginally conscious of” the need to finish soon to get ready to go to the airport. But when this thought does “come to the fore,” breaking my concentration on the message, am I to say I recall that even before this I had been continually having the conscious thought that I needed to finish to get to the airport on time? Perhaps all there was to my having been marginally conscious of the need to finish (and all there was to having the need to go to the airport “dimly before my mind”) is that the very character of my experience of thinking while focused on composing the message was also such as to make me more readily inclined to consciously think that I needed to finish soon than to think many other thoughts of which I was capable, so that thought keeps coming back to me even after I turn my attention elsewhere. (A vivid case: when in love you keep intermittently slipping back to thoughts of the one you love—but that you are so inclined can be continually part of what it’s like for you to be thinking as you are—for your thoughts to be always being “pulled towards” your beloved, for this person to be always “hovering” on the “horizon” of your thoughts.) Thus, what it’s like for me to think doesn’t just constrain what thought I am currently thinking, but also what else I am especially liable to think. This is partly analogous to how what my peripheral vision is like for me readies and inclines me to look at things I am not yet looking at (even “marginally”—as if there were such a thing as peripherally looking at something). Even if this account of what it is to be marginally conscious of something/someone in thought is not endorsed, we can recognize the reality of “marginal thought” without saying we are always at least marginally consciously thinking about our own visual experience as long as it’s conscious. And I have no inclination to say that whenever I am struck by how something looks to me, I can recall that I had already been consciously—but only marginally—thinking something about how it looked to me. What about option (b) then—every conscious state includes a sort of representational “sensing” of itself? The problem is this. What’s wanted here is some sort of conscious sensing of one’s own sensing of color and shape (say) where this (higher-order) sensing of sensing is discernible as a feature distinct from the (first-order) sensing of color and shape. To find any second-order sensory representation, we need to find, at the second-order level, either something corresponding to the experience of shape and color constancy we find at the first, which allows us to distinguish an object sensed from its manner of being sensed (“objectual sensing”), or else something like a sensing distinct from

Phenomenality and Self-Consciousness

249

the condition sensed—a feeling of pain is distinct from the tissue damage it “registers.” But sensory object constancy is phenomenologically indiscernible, once the objects in question are not, for example, the shaped-and-colored things, but their looking to us as they do. We may speak of our experience of these things looking to us as they do—but here changes in the manner of experience are not distinguishable from changes in the manner of the visual appearance. And there is to be found no separate feeling that registers the feeling of pain, as that feeling may register a cut or a tear in one’s flesh. The points I’m making here need further explanation and defense—which I offer elsewhere (Siewert 2012). But hopefully what I’ve just said is enough to convey my basic objection to option (b): the sort of representational inner “sensing of sensing” hypothesized, being conscious, should be phenomenologically discernible—but once we are clear about what this would require (namely, higher-order sensory constancy or a distinguishable sensory registration of sensing), we can see that it just isn’t. That leaves option (c): endorsing the special sui generis self-representation— one that’s neither thinking nor sensing. A problem here is that it’s hard even to understand just what this allegedly ubiquitous form of representation positively distinctively is—I mean, what it is, other than being not this, and not that. But further: we seem to have no warrant from first-person reflection for thinking that this sort of conscious self-representation occurs, except indirectly— for example, through consideration of the regress problem. Reflection allegedly tells us that we are in some sense conscious of our every conscious state. From this we then argue: we can maintain this, while solving the regress problem, only by positing a unique form of representation to constitute this constant consciousness of our own conscious states. This suggests a potential vulnerability for SRM theorists who would appeal to the “Conscious-Of” Principle even if they take route (c). For suppose we can find reasonable alternative ways to interpret CO that don’t support SRM, while still allowing us to affirm “Is-Conscious” as well—and which (unlike option (c)) don’t require that we posit any conscious experiences of a type otherwise reflectively indiscernible to us. Then, other things being equal, those interpretations are to be preferred as both more parsimonious and less mysterious. In the next section, I will propose such alternatives.

3. Solving the Regress Problem. We may allow that there is some initial appeal to the idea that if you have a conscious state, you are somehow conscious of it. (Or, in the negative formulation that higher-order theorists like to invoke: if you are “in no way” conscious of it, it is not a conscious mental state.) But there are several ways to account

250

phenomenal intentionality

for this appeal, while dealing with the regress problem, none of which provide support to SRM. First, there is what I’ll call “the ambiguity proposal”: CO and IC concern distinct but closely related senses of “conscious.” It’s true that there’s some sense in which, if you have a conscious mental state, then you are conscious of that mental state. We may even recognize that there’s a sense for which this is necessarily true. We may even say that being conscious of one’s mental state involves having some sort of thought about it. However, when we explicitly interpret the principle this way and reflect on the regress problem, what we discover is that CO, unlike IC, is really not a principle about state consciousness in the basic phenomenal sense, but some distinct—let’s call it “reflective”—sense derived from the phenomenal one. CO says in effect: if one is in a reflectively conscious state c, then one must be consciously thinking of c. And IC says if one is conscious of something, then one’s being conscious of it is itself a phenomenally conscious state. That is not to deny there is an intimate relationship between being a reflectively conscious state and being a phenomenally conscious state. For we may say that one has a reflectively conscious mental state only if one is conscious of it in a phenomenal sense—by having phenomenally conscious thought about it. But since there is no necessity that one be similarly reflectively conscious of one’s every phenomenally conscious thought, no regress ensues. Now one may object: “Part of what makes CO seem attractive is the idea that our phenomenally conscious states are quite generally ones we are, in some sense, conscious of. And you have not done justice to that.” There is a way of accommodating this reaction as well, which I’ll call “relaxed availability.” When you say, yes, if you have a conscious experience, then you are conscious of having it—it is hardly clear that you have to claim that no conscious state could possibly occur, anywhere, anytime, except in some being that is then and there thinking about or otherwise representing that state to him, her, or itself. No—what makes this CO principle “sound right” to us on hearing it (if it does) may be something considerably less stringent. Th at is, we may have in mind nothing more than the idea that phenomenally conscious states are usually available to reflect upon consciously, when they occur in beings, like ourselves, who are capable of reflective thought. Th at is, if you have a general ability to form “higher-order thoughts,” then, usually you will be able and disposed to think phenomenal thoughts of some kind about your own actual phenomenal thoughts and experiences, if only retrospectively, should the issue of what you’re thinking or experiencing arise. Now it’s true that even this relaxed CO principle commits us to saying that usually one is capable of forming conscious thoughts about one’s own actual conscious states. But that is not unrealistic and does not lead to an infinite regress. For the principle is compatible with recognizing that actual human beings cannot perform many iterations of conscious reflection. Maybe typically, after

Phenomenality and Self-Consciousness

251

as few as, say, four “conscious of’s,” we become bored, forgetful, distracted, or unable to distinguish levels. This doesn’t conflict with the claim of relaxed availability, because usually we only actually get up to the second level (we consciously think about how we feel) or less often, the third (we consciously think about our consciously thinking about how we feel). So it’s true that we’re usually capable of at least one more (perhaps retrospective) reflection, about whatever level we’re actually at. And that’s all the principle says. Maybe someone will insist there is still something unaccounted for that makes the “Conscious-Of” Principle seem attractive. They might say: “Some animals—dogs, for example—can’t think about their own minds. Still, they feel some way, and things look some way to them. And these are phenomenally conscious sense experiences they have: there is something it’s like for them to see shape, movement, colors, and to feel pain. Furthermore, in these cases it sounds right to say that in some sense they are conscious of their feelings or conscious of seeing things. So the problem is this: the CO principle applies to dogs too. It’s also true that they are conscious of their every conscious state. But your account of the principle offers to explain only what makes it true of reflective creatures that they are conscious of their every conscious state. And dogs aren’t reflective. So your interpretation of CO is still inadequate.” I am perfectly happy to say that dogs enjoy phenomenally conscious vision, even if they are unreflecting beasts. I think one can sensibly (perhaps futilely) wonder just what it’s like for something to look as it does to a dog. And it sounds right to say that in some sense they are conscious of feeling pain, when they do. However, I have to say it does not sound to me clearly right to say they are, for example, “conscious of seeing the ball roll under the sofa.” For this reason I am reluctant to grant that the CO principle really does hold of unreflecting creatures’ every conscious state. Still, one might wonder how to make sense of their being “conscious of feeling” without imputing reflection to them. And I think once we see how to do this, we will see how we could also, if someone insisted, give CO broader scope, without assuming that all who have conscious states possess self-representing minds. Suppose Pickles the unreflecting dog has, unfortunately, a feeling of pain. We need not take this to mean that pain is some object of her feeling, or that her feeling constitutes a representation of her pain. To say she has a feeling of pain may be taken to mean that she has a feeling of a certain sort—namely, a painful sort. Now do we also wish to say that Pickles is conscious of feeling pain? We could say this, but if we do, nothing compels us to take the “of” in this “conscious of” to introduce an intentional object, or object of representation. We may treat it more or less as I just proposed we treat the “of” in “feeling of pain.” To say she is conscious of feeling is only to say that she has a consciousness of a certain sort: a “feeling” sort. And to say she has a certain sort of consciousness may be taken to mean she has a certain sort of conscious state. Now, likewise,

252

phenomenal intentionality

if we insist (strangely, to my mind) that it’s right to say when Pickles sees a ball, and is conscious of a ball, she is, in some sense conscious of seeing,4 then we may take this just to mean she has consciousness of a certain sort—a visual sort. The point is that if you think there is some sense in which the principle “if a state is phenomenally conscious, then one is conscious of it” should be taken to apply universally, even to all the states of creatures that are incapable of thinking about their own minds, there is a way one can understand this, while again avoiding the regress worries. And that is to suppose that here the “of” in the phrases “conscious of feeling,” “conscious of seeing,” and so on, does not introduce an intentional object. Rather it indicates what sort of conscious state the creature is in. Since “conscious of,” understood in this way, does not mean conscious of an object, I would label this interpretation the “non-objectual” proposal. I have offered three ways of understanding what might make the CO principle sound right without running afoul of the regress problem. Either we discern an ambiguity in “conscious state” as it is found in CO and IC; or we understand CO to indicate nothing more than a relaxed availability principle; or we adopt a non-objectual interpretation of “conscious of.” We do not need to say that only one of them is correct. It could be that they all help to explain how the idea that “conscious states are states one is conscious of” could sound plausible or intuitive to us in various contexts. My claim is that once we recognize such construals of that slogan (and perhaps some others), there is nothing more we need take it to express that we should accept that supports SRM—the idea that having a self-representing mind is essential to phenomenality. If one is determined to treat CO as a necessary or constitutive truth that supports SRM, one encounters a dilemma. To escape the regress problem, either one denies that states of being conscious/aware of something are conscious, or else one posits an unbroken activity of conscious representation fused with the ground-floor experience it represents. Going the first route would be to forsake the sort of pretheoretical understanding of consciousness supposedly captured in CO—which would deprive one of the right to appeal to this principle in motivating one’s theory. And notice now—and this is a new point—the problem persists even if you stoutly refuse to grant that IC is “commonsensical” or “intuitive,” or confine this status to an “awareness only” version of CO. Let’s set aside the Kriegel-style concern that you would be introspectively oblivious to your alleged unconscious awareness of your experience. Just consider which of the following yields a more commonsensical or intuitive claim. Whenever we are not consciously thinking about our own conscious thoughts and sensory states, we are (a) necessarily somehow unconsciously thinking about them, or (b) usually as a matter of fact capable of thinking consciously of them. Would anyone maintain (a) rather than (b) is the voice of “commonsense” or “intuition”? But then CO does nothing to support the higher-order theory that would urge

Phenomenality and Self-Consciousness

253

us to endorse (a) over (b). Similar remarks would apply to other “unconscious self-representation” views. When we compare the interpretations they would have us give of CO to non-SRM-supporting alternatives like those I’ve suggested, their claim to speak for commonsense or intuition evaporates. Suppose then we take the second horn of the dilemma I raised and affirm that conscious experience is constantly consciously representing itself. This is also undesirable. For if this constant self-representing is either thought or sense-like, it is—though purportedly conscious—strangely indiscernible to introspective reflection. And if the sort of self-representation alleged is neither of these, but something special and sui generis, then its postulation is still vulnerable to a response to the regress problem like that I’ve suggested, which also retains both CO and IC, but without having to posit some new, otherwise unheard-of form of conscious mental representation. I conclude that, despite its popular use as the launching pad for SRM theorizing about consciousness, the “Conscious-Of” principle really provides it with no justification. And, reasonably interpreted, it is no “datum” that “fi rst-order theorists” must embarrass themselves by denying. Of course, other considerations are adduced in favor of higher-order representation theories of consciousness that I have not addressed here (though I have discussed some elsewhere, in Siewert 1998: 201, and 2012). However, the interpretation of CO I have just criticized plays, it seems to me, a fundamental role in getting such accounts off the ground and making them seem plausible, and it shows up repeatedly in some form in discussions of consciousness generally. If we should (as I have argued) give the principle so interpreted no weight at all in our theorizing, that will significantly affect its course.

5. Non-Representational Self-Consciousness, “Self-Presence,” and “For-Me-ness” I have focused here on theories that make a self-representing mind essential to phenomenality. But there are other, rather different accounts of consciousness that apparently overlap partly with these in maintaining that, necessarily, one is conscious of one’s every conscious state, in some sense that constitutes a sort of self-consciousness. However, these accounts should not be seen as maintaining that this “consciousness of consciousness” consists in a kind of representation of one’s own state of mind. So consider first Husserl’s (1991) view that the proper understanding of the “consciousness of internal time” demands that we recognize, as a part of all experience, a form of intentionality (exemplified in our ongoing “retention” of what we just experienced) that continually points to one’s own consciousness without “making it into an object.” And more recently, Dan Zahavi (2005, 2006) defends a Sartre-influenced, neo-Husserlian view that

254

phenomenal intentionality

all consciousness includes a “non-reflective,” “non-objectifying” consciousness of one’s own experience. Evan Thompson (2011) also argues for a position of this sort, building not only on Husserl and Sartre, but on the “memory argument” for reflexive awareness discussed in Indo-Tibetan Buddhist philosophy. Such views deserve separate treatment. But they also draw on the “Conscious-Of” Principle or near relatives, interpreted in ways to which I aim to present an alternative. So I think that reasons similar to those I’ve given for preferring my interpretation to option (c) also apply to them. The main point is that there is a gap between saying (as I have) that we can interpret “conscious of feeling (seeing, thinking, etc.)” in a “non-objectual” manner, and saying (as some would) that we are dealing here with some non-objectual, non-reflective, or non-representational form of reflexivity or intentionality. This further step would need to be both clarified and justified. Unless that gap is bridged, we should stick with the more modest interpretations of the CO principle I have proposed. Moves to fi ll the gap can be found in claims (such as we see in Zahavi) that our own phenomenal states are essentially “present to us” (or “given” to us) in a special first-person way, and that it’s distinctive of my phenomenally conscious experience that it is “for me” or “experienced as mine”—in a manner that nothing unconscious is. So much so that a rejection of this experiential “for-me-ness” or “self-presence” can seem tantamount to a blindness to consciousness itself. (It should be noted that this talk of “for-me-ness” also figures importantly in Kriegel’s arguments.) I certainly do not want to dismiss these claims. But I also do not think we are entitled, without further examination, to treat them as phenomenological data that reveal that, in some sense, all consciousness is essentially conscious of itself—in some sense, self-conscious or reflexive. For we may be able to do justice to much or all of what is rightly invoked by the talk of “presence,” “for-me-ness,” and so on, without this. I will explain provisionally how this might be so, again with reference to visual appearance. Regarding the peculiar “presence” of my own experience to myself, the suggestion is this. There is a form of thought, encountered earlier in my remarks on “noticing” or being struck by how something looks, whereby I can attend to how something appears to me, which enables me to identify for myself this manner of appearing, so as to go on to recognize possible ways of articulating this as correct or incorrect. Such thoughts are expressible by the use of phrases such as “the way this looks to me now” (where “look” is interpreted phenomenally, as earlier explained). It seems (though this needs further discussion) this form of thought cannot even arise in the absence of the manner of appearance it identifies. If there just is in fact no way anything now appears to me that I successfully identify in a thought expressible in some such manner, then I fail altogether to express and form a thought about a specific way of

Phenomenality and Self-Consciousness

255

appearing. In that sense, this sort of thought depends on my having the very sort of appearance it is about. Further, this type of first-person reflection cannot occur, we may say, without a phenomenal target. (If I try to think about my unconscious visual processing via some thought I might express by saying “the way this is visually represented by me,” I find I don’t thereby understand myself to have identified a specific manner of visual representation, whose classification I can then recognize as correct or incorrect.) Thus for two reasons my own phenomenal experience is peculiarly “present” to me to think about. First, it is and must be “actually there” (hence “present”) for me to think about, if I am even to think such thoughts. Second, when I do think them, the experience cannot fail to “present itself” in the thought, it cannot “hide”—that is, it cannot escape identification. For the same reasons we may want to say such thoughts are peculiarly closely or “intimately” related to the phenomenal states they are about. But notice: none of this entails that whenever something does look to me somehow, some (perhaps diminished) form of self-consciousness must already be part of the visual appearance, or that the experience must somehow already be “pointing at itself,” prior to reflection. So this is how I propose to understand the first-person “presence” or “givenness” of experience. I do not claim here to have fully explained and justified my account of this (though I do discuss these matters in more detail in Siewert 2012). Here I claim only to have presented an intelligible and plausible way to flesh out the notion of the peculiar “presence” of one’s experience to oneself, without building some sort of self-consciousness into every moment of phenomenality. But now, what about the idea that my conscious experience is always “for me” in some distinctively phenomenal way? I would first distinguish between saying that, just in virtue of being conscious, all my experience involves a consciousness of itself—thus a self-consciousness—and saying that, in virtue of being conscious, it is possessed by a self. And I propose that it is in light of the latter notion that we may often plausibly regard as true the claim that my phenomenally conscious state is “for me” or “experienced as mine.” One way to see how this could work: think about the way the viewpoint of the looker is implicit in how things look. One might say that when I look at things, the perspective from which I am looking is implied by how things that do look somehow to me appear to be situated—even though this is not itself something that looks some way to me. To put it differently: it belongs to the manner in which I experience the visual appearance of things to me that they appear from here—which is not to say that they appear to me to be appearing to me from here. The appearance of what is seen can entail where one is viewing from, even if this viewing and the viewpoint are not themselves seen, nor apparent, nor the accusative of any form of self-consciousness. So, in a sense the viewpoint of the viewer—the one to whom things appear—is implicit in the manner of their appearing. And

256

phenomenal intentionality

if a phenomenal point of view (unlike a pictorial perspective) implies an actual viewer, the fact that an appearance is an appearance to someone is included in its phenomenal character, no less than the fact that it is an appearance of (for example) color or shape. There is more implied in how something appears to me than what appears to me—and its being an “appearance to” is part of this “how” of appearance. Finally, the fact that the appearance is an appearance to someone is all there is to its being an appearance “for” someone.5 This last point suggests that I can recognize (compatibly with my earlier criticisms of higher-order representational theories) a sense in which a special form of self-consciousness is built into the character of experience, after all. To see the sort of concession I have in mind, consider this potential support for it. Suppose, for example, we accept that your activity of looking at things belongs to how you experience things looking to you as they do. An activity of looking at them is implied by what it’s like for them to look to you as they do, as your attention darts around, alighting and lingering here and there. Now, putting this together with the earlier point, we may say: our experience has a phenomenal character that implies it is a perspectivally constrained looking at things. Then we might say, “looking at” involves a “looker” and so some (perhaps “marginal” or “peripheral”) awareness of oneself as a looker looking from somewhere is involved in ordinary visual experience. Hence some sort of peripheral consciousness of oneself as a situated, active perceiver—as a “perspective-taker (or maker)”—a special form of self-consciousness, is essentially involved in the ordinary experience of looking. But we shouldn’t want to say this looker or this looking are, for this reason, themselves among what is “visually represented.” (Again, we shouldn’t reduce the “how” of appearance to its “what.”) And we needn’t even say that one’s awareness of oneself as a looker involves in some other way “representing oneself as a looker.” For to be aware of oneself as a looker may be simply to be aware of oneself (specifically, of one’s own body) in the manner in which a looker is self-aware. This doesn’t entail that one thereby somehow attributes to oneself, truly or falsely, the property of looking at things from a certain place. Consider, we can distinguish: “He is conscious of himself as an adult” (that is, he takes himself to be an adult), and “He is conscious of himself as an adult is” (that is, conscious of himself in an adult manner). Similarly, we can recognize a sense in which we are nearly always conscious of ourselves as active situated perceivers are. But this isn’t to say we are nearly constantly representing ourselves to be active situated perceivers. Th is is how I would make sense of the notion that I am, in visually consciousness, conscious of myself “as subject” even if I do not thereby represent myself as a visible object. We may be tempted to say just here that visual experience “refers to” or “points back to” its subject as well as at its object—though with a distinctively subjective kind of (“pre-reflective,” “non-objectifying,” “non-representational”) “reflexivity” or self-reference. Borrowing Colin McGinn’s (1991, 34) image, we might want to

Phenomenality and Self-Consciousness

257

say that consciousness, “Janus-faced,” both looks out at things seen and back at the see-er, though the pointing/facing “inward” is fundamentally distinct in kind from the pointing/facing “outward.” But I cannot endorse this sort of talk, as it remains unclear that there is some genus of “pointing” (or “directedness”) of which both “subject-pointing” and “object-pointing” should be seen as species.6 However, I do find it plausible that consciousness is somehow necessarily tied to self-consciousness in the way I’ve suggested. What I doubt here (partly on the basis of regress worries) is that we can establish this link by intuitive appeal to principles alleged to apply univocally to all forms of phenomenal consciousness—such as the idea that all consciousness is essentially conscious of itself, or that all experience has the feature of “for-me-ness” or “self-presence.” We will, in my view, better understand phenomenal consciousness and its connection with self-consciousness by asking how various specific forms of phenomenality may either partly constitute or implicitly include various kinds of self-consciousness—as in the ways just proposed. If I am right about this, then phenomenological examination of the relationship of consciousness and self-consciousness ultimately speaks against explaining consciousness in terms of self-directed intentionality—seeing it as nothing but a certain way the mind represents itself to itself. What we should say, first, rather, is that we may be conscious of ourselves in various ways, in some of which we occasionally represent our own conscious states to ourselves. And, when this happens, the phenomenality of what is represented helps constitute—it is not explained by—the special reflective form of self-consciousness in which it is involved. Second, if there is some form of self-consciousness ubiquitously (or nearly ubiquitously) involved in consciousness, this is to be sought in our being conscious of ourselves in the mode of active perspective-takers, a mode whose most basic form lies in ordinary perceptual experience.

Acknowledgments I would like to thank Steve Crowell and Uriah Kriegel for their comments and questions on drafts of this chapter. I also would like to thank members of the audience at the University of Fribourg who heard an early ancestor of this chapter for their feedback—I am particularly indebted to Martine Nida-Rümelin and Gianfranco Soldati. And I wish to express my gratitude to Declan Smithies and numerous commenters on his Facebook thread for their helpful and stimulating remarks concerning the durian example (prompted by its use in previous work of mine). Finally, I am grateful to audiences at the SMU workshop “Consciousness and the Self” and the NEH Summer Institute “Investigating Consciousness.”

258

phenomenal intentionality

Notes 1. It can be doubted (and is doubted, by Benj Hellie) that “conscious state” is the appropriate notion here—we sholdn’t think that “states” (as opposed to activities) are what’s properly regarded as phenomenally conscious. But I would like to use the notion of “state” as broadly as possible here to include any particular instance of a repeatable feature—not so as to contrast it with the category “activity.” 2. Kriegel (2004) makes a similar point in criticism of Rosenthal. 3. Kriegel (2009) appeals to the idea that one is always at least marginally/peripherally conscious of one’s conscious states in defense of a self-representational view, and he defends the notion that we peripherally think certain thoughts, but he doesn’t commit to saying that the marginal self-consciousness typically involved in consciousness is a form of thought. 4. Th is relative strangeness indicates, I think, that this proposed “non-objectual” “feeling of pain” interpretation of “conscious of . . . ” is more natural in the “conscious of feeling . . . ” case than in the “conscious of seeing . . . ” case. 5. Levine (2012) would object that a visual appearance being an appearance “to me” is no more “innocent of reflexivist connotations” than its being “for me.” My response is that once the active, perspectival phenomenal character of looking (described here and in the following paragraph) is taken into account, there seems to me no further potential “reflexivist connotation” in “to me” we need to honor. 6. Th is is part of why I am unpersuaded by the sorts of “retention” and “memory” arguments put forward by Zahavi (2005) and Thompson (2011). Perhaps it is true that I need to (non-objectually) experience my own experiences “as mine” in order to retain or recall them “as mine,” but I don’t see that this shows they need to be non-objectually “pointed at,” or what that would mean.

References Block , N. (2002). “Concepts of Consciousness.” In David J. Chalmers (ed.), Philosophy of Mind: Classical and Contemporary Readings. Oxford: Oxford University Press. Brentano, F. (1972). Psychology from an Empirical Standpoint, trans. T. Rancurello, D. Terrell, and L. McAlister. London: Routledge. Carruthers P. (2000). Phenomenal Consciousness. Cambridge: Cambridge University Press. Carruthers P. (2004). “HOP over FOR, HOT Theory.” In R. Gennaro (ed.), Higher-Order Theories of Consciousness. Philadelphia: John Benjamins. Carruthers P. (2007). “Higher-Order Theories of Consciousness.” In M. Velman and S. Schneider (eds.), The Blackwell Companion to Consciousness. Oxford: Blackwell. Chalmers, D. (2002). Philosophy of Mind: Classical and Contemporary Readings. Oxford: Oxford University Press. Gennaro R . (ed.). (2004). Higher-Order Theories of Consciousness. Philadelphia: John Benjamins. Husserl, E . (1991). On the Phenomenology of the Consciousness of Internal Time (1893–1917), trans. J. Brough. Dordrecht: Kluwer. Levine, J. (2012). “On Phenomenal Access.” In Themes From Block, A. Pautz and D. Stoljar (eds.). Cambridge, MA: MIT Press. Lycan, W. (2004). “The Superiority of HOP to HOT.” In R. Gennaro (ed.), Higher-Order Theories of Consciousness. Philadelphia: John Benjamins. McGinn, C . (1991). The Problem of Consciousness. Oxford: Blackwell. Merikle P., Smilek, D., and Eastwood, J.D. (2001) “Perception Without Awareness: Perspectives from Cognitive Psychology.” In S. Dehaene (ed.), The Cognitive Neuroscience of Consciousness. Cambridge, MA: MIT Press.

Phenomenality and Self-Consciousness

259

Nagel, T. (1974). “What Is It Like to Be a Bat?” Philosophical Review 4:435–50. Kriegel, U. (2004). “Consciousness and Self-Consciousness.” Monist 87: 185–209. Kriegel, U. (2009). Subjective Consciousness: A Self-Representational Theory. Oxford: Oxford University Press. Rosenthal, David. (2002). “Explaining Consciousness.” In David J. Chalmers (ed.), Philosophy of Mind: Classical and Contemporary Readings. Oxford: Oxford University Press. Siewert, C . (1998). The Significance of Consciousness. Princeton: Princeton University Press. Siewert, C. (2011). “Phenomenal Thought.” In T. Bayne and M. Montague (eds.), Cognitive Phenomenology. Oxford: Oxford University Press. Siewert, C. (2012). “On the Phenomenology of Introspection.” In D. Smithies and D. Stoljar (eds.), Consciousness and Introspection. Oxford: Oxford University Press. Smith, D. W. (2005). “Consciousness with Reflexive Content.” In D. W. Smith and A. Thomasson (eds.), Phenomenology and the Philosophy of Mind. Oxford: Oxford University Press. Thompson, E . (2011). “Self-No-Self? Memory and Reflexive Awareness.” In Mark Siderits, Evan Thompson, and Dan Zahavi (eds.), Self, No Self? Perspectives from Analytical, Phenomenological, and Indian Traditions. Oxford: Oxford University Press. Williford, K . (2006). “The Self-Representational Structure of Consciousness.” In U. Kriegel and K. Williford (eds.), Self-Representational Approaches to Consciousness. Cambridge, MA : MIT Press. Zahavi, D. (2005). Subjectivity and Selfhood. Cambridge MA : MIT Press. Zahavi, D. (2006). “Th inking about Self-Consciousness: Phenomenological Perspectives.” In U. Kriegel and K. Williford (eds.), Self-Representational Approaches to Consciousness. Cambridge MA : MIT Press.

This page intentionally left blank

I N DE X

Belief 14, 90–20, 22, 30, 63, 71, 76, 80–1, 88, 96, 126–8, 132, 134–5, 153, 156–8, 160, 162–72, 183, 192, 194, 196–9, 202–7, 209, 212–5, 218–32 Brentano 2, 21, 24–5, 84, 98, 202, 204, 246 Content conscious see content, phenomenal descriptive 50–1, 55, 62, 68 determinate 10–3, 15, 17, 22, 194–7, 199, 201–4, 206, 209–11, 217–8, 222–3, 226, 229, 231 indeterminate see content, determinate indexical 49–52, 54, 56, 59, 61–2, 64–7 intentional 14–6, 19, 63, 67, 138–9, 155, 157, 174, 194–7, 205, 207, 209, 222, 224, 226, 228 mental 2, 27, 68, 141, 162, 194–6, 220, 230–1 narrow 12, 16, 52, 144–50, 152, 154, 198–9, 205–6, 209–10, 212–3, 223, 228–9 perceptual 9, 34, 37, 41, 123, 127, 129, 134 phenomenal 11, 19, 42, 49, 86, 89–90, 138, 141, 144, 146, 162 representational 28, 79–80, 82, 85–6, 112, 152–3 wide see content, narrow Chalmers 22, 67, 69, 80, 82, 133–4, 153, 201, 210, 212, 222, 224, 227, 229, 232 Consciousness phenomenal 1, 4, 6, 21–2, 89, 118, 156–62, 164, 166, 169, 235–8, 240–4, 246, 250–4, 257–8 stream of 7, 17, 157, 168, 171, 190, 192 Experience 6–9, 17–8, 22–3, 2–34, 36–7, 39, 41–7, 52,71–90, 99–114, 116–20, 122–35, 137–54, 156–8, 160, 164,

170–1, 174–92, 194–6, 198, 201–4, 206–7, 211–2, 214–5, 217–26, 228–32, 239–42, 247–58 Harman 18, 73, 86, 140, 153 Horgan 2, 5, 8, 10–5, 19, 21–2, 47, 67, 79, 90, 100, 112, 134–5, 138, 143–6, 148–9, 152–4, 191–2, 194–202, 205, 207–9, 211–4, 217–9, 223, 227–31 Husserl 2, 137–8, 140, 142, 147, 152–4, 192, 253–4 Intentional nonexistence 90, 138, 140–1, 143, 147–8, 151, 153 Intentional objects 30, 93, 106, 137–45, 147–54, 251–3 Intentionality conscious see intentionality, phenomenal experiential see intentionality, phenomenal phenomenal 1–24, 52, 67, 72, 78, 83, 85–6, 89–90, 93, 95–8, 100, 112, 116–8, 133–4, 137–8, 140–1, 144, 146–53, 157–8, 171–2, 174, 191, 194–5, 199–200, 204–5, 208, 221, 223–8 Loar 2, 10, 12, 14–5, 19, 22–3, 88, 90, 112, 135, 137–41, 146–9, 151–4, 230 Meinong 2, 87, 141, 153–4 Nonexistent objects see intentional nonexistence Perception 6, 8–9, 17–8, 23, 27–31, 33–35, 37, 40–4, 46–8, 68, 71, 74–5, 77–8, 81, 84, 87–90, 97, 99–114, 116–8, 122–8, 132–5, 137, 140–1, 153, 156–7, 164, 170, 188, 196, 198, 202–3, 211, 222–3, 236–7, 242–5, 247, 257 261

262

Index

Perceptual experience see perception Pitt 9, 22, 24, 67–8, 157–8, 192, 197, 199, 202–3, 208, 214, 228 Phenomenality see phenomenology Phenomenology cognitive 6, 9, 17, 21–2, 49, 52, 62, 142, 153, 174, 187–9, 191, 195–6, 198–211, 213, 215–20, 222, 225, 228–9 of thought see phenomenology, cognitive perceptual 18, 23, 46, 109, 196, 203, 211 sensory 8–9, 195–7, 201, 204, 207, 210–1, 215, 217–21, 229 Presence 11, 71–2, 75–83, 85–6, 89, 99–100, 104–6, 253–5 Qualia 89, 158–9, 201–3, 212–3, 216–20, 225, 229–30

Representationalism 72, 75, 78–83, 86–7, 89, 109, 140–1, 147, 153, 160 Russell (Bertrand) 29–30, 35, 38, 45–6, 68, 90, 94–5, 153, 172, 203 Searle 10, 13–5, 19, 22, 30, 40, 47, 88–9, 154, 166 Subjectivity 5–6, 11–3, 15–6, 77, 83–4, 89–90, 120, 137, 238, 240–2, 256 Siewert 7–8, 67, 90, 153, 196, 214, 220, 228–9, 240, 249, 253, 255 Strawson (Galen) 9–10, 13–5, 17, 20, 22, 24, 30, 46, 89, 152–3, 158, 196, 198, 200, 209–10, 213–4, 216–9, 227–9 Tienson 2, 5, 10, 12, 14–5, 22, 47, 90, 100, 112, 135, 138, 143–4, 148, 152–4, 192, 194–202, 205, 207, 209, 212–4, 217–9, 224, 227–9 Tracking 1–3, 11, 20–2, 35–6, 44–6, 118, 135, 194–5, 198, 206–10, 223–4, 227–8 Transparency (of experience) 18, 20, 23