The Nature of Consciousness: Philosophical Debates

THE NATURE OF CONSCIOUSNESS Philosophical Debates

edited by Ned Block, Owen Flanagan, and Guven Giizeldere

A Bradford Book The MIT Press Cambridge, Massachusetts London, England

© 1997 Massachusetts Institute of Technology All rights reserved. No part of this book may

be repro­

duced in any fonn by any electronic or mechanical means (including photocopying, recording, or infonna­ tion storage and retrieval) without pennission in writing from the publisher. This book was set in Times Roman by Asco Trade Typesetting Ltd, Hong Kong. Printed and bound in the United States of America. Library of Congress Cataloging-in-Publication Data The nature of consciousness: philosophical debates I edited by Ned Block, Owen Flanagan, and Giiven Giizeldere. p.

em.

"A Bradford book." Includes bibliographical references (p.

) and index.

ISBN 0-262-02399-7 (hb: alk. paper). ISBN 0-262-52210-1 (pb: alk. paper) 1. Consciousness. 2. Philosophy of mind. I. Block, Ned Joel, 1942-

. II. Flanagan, Owen J.

III. Giizeldere, Giiven. B808.9N37

I 26-dc20

1997 96-17500 CIP

10 9 8 7 6

In memory of Harry Bradford Stanton

Contents

Preface and Acknowledgments Sources

ix

9

XVII

INTRODUcnON

Guven Giizeldere

I

STREAM OF CONSCIOUSNESS

69

1

The Stream of Consciousness

71

10

The Cartesian Theater and "Filling In" the Stream of Consciousness

THE

PSYCHOLOGY AND NEUROPSYCHOLOGY OF CONSCIOUSNESS

185

Contrastive Phenomenology: A Thoroughly Empirical Approach to Consciousness

187

Bernard 1. Baars

William lames

2

181

Robert Van Gulick

ill

The Many Faces of ConsciotmleSS: A Field Guide

Time for More Alternatives

11 83

Daniel C. Dennett

Visual Perception and Visual Awareness after Brain Damage: A Tutorial Overview

203

Martha 1. Farah

3

The Robust Phenomenology of the Stream of Consciousness

89

12

Owen Flanagan

II

CONSCIOUSNESS, SCIENCE, AND METHODOLOGY Prospects for a Unified Tbeory of Consciousness or, What Dreams Are Made Of

237

Edoardo Bisiach

95

13 4

Understanding Com;ciousness: Oues from Unilateral Neglect and Related Disorders

Modularity and Consciousness

255

Tim Shallice

97

14

Owen Flanagan

Towards a Neurobiological Theory of Consciousness

277

Francis Crick and Christof Koch

5

Consciousness, Folk Psychology, and Cognitive Science

111

IV

Alvin I. Goldman

6

Can Neurobiology Teach Us Anything about Consciousness?

15

CONSCIOUSNESS AND CONTENT

293

Consciousness and Content

295

Colin McGinn

127

Patricia Smith Churchland

16 7

8

Time and the Observer: The Where and When of Consciousness in the Brain

Externalism and Experience

141

17

Daniel C. Dennett and

A Representational Tbeory of Pains and Their Phenomenal Character

Marcel Kinsbourne

Michael Tye

Begging the Question against Phenomenal Com;ciousness Ned Block

309

Martin Davies

18 175

Sensation and the Content of Experience: A Distinction Christopher Peacocke

329

341

Contents

V

viii

29

FUNcnON OF CONSCIOUSNESS

355

Finding the Mind in the Natural World

483

Frank Jackson

19

Conscious InessentiaIism and the EpiphenomenaIist Suspicion

357

30

Owen Flanagan

Breaking the Hold: Silicon Brains, Conscious Robots, and Other Minds

493

John R. Searle

20

On a Confusion about a Function of Consciousness

375

31

Ned Block

21

503

Sydney Shoemaker

The Path Not Taken

417

VB

Daniel C. Dennett

22

The First-Person Perspective

EXPLANATORY GAP

32

Availability: The Cognitive Basis

421

of Experience?

SUBJECTIVITY AND

What

Is It Like to Be a Bat?

517 519

Thomas Nagel

David J. Chalmers

33 23

Fallacies or Analyses?

425

Problem?

Jennifer Church

24

Two Kinds of Consciousness

427

34

On Leaving Out What It's Like

543

Joseph Levine

vm

Understanding the Phenomenal Mind: Are We

529

Colin McGinn

Tyler Burge

25

Can We Solve the Mind-Body

All Just Armadillos?

THE KNOWLEDGE ARGUMENT

557

Part ll: The Absent Qualia Argument

435

35

Understanding the Phenomenal

Robert Van Gulick

Mind: Are We All Just Armadillos?

METAPHYSICS OF

Explanatory Gaps

Part I: Phenomenal Knowledge and

VI

26

CONSCIOUSNESS

443

The Identity Thesis

445

36

Saul A. Kripke

27

37 451

A Question about Consciousness Georges Rey

567

Knowing Qualia: A Reply to Jackson

John R. Searle

28

What Mary Didn't Know Frank Jackson

Reductionism and the Irreducibility of Consciousness

559

Robert Van Gulick

571

Paul M. Churchland

461

38

What Experience Teaches David Lewis

579

IX

Contents

39

Phenomenal States

597

Brian Loar

IX

QUALIA

617

40

Quining Qualia

619

References to Introduction

807

Suggested Readings

817

compiled by Guven Giizeldere Index

Daniel C. Dennett

41

The Inverted Spectnun

643

Sydney Shoemaker

42

The Intrinsic Quality of Experience

663

Gilbert Harman

43

Inverted Earth

677

Ned Block

44

Curse of the Qualia

695

Stephen L. White

X

IDGHER-ORDER MONITORING CONCEPTIONS

45

OF CONSCIOUSNESS

719

What Is Consciousness?

721

David Armstrong 46

A Theory of Consciousness

729

David M. Rosenthal

47

Consciousness as Internal Monitoring

755

William G. Lycan

48

Conscious Experience

773

Fred Dretske

49

Is Consciousness the Perception of What Passes in One's Own Mind? Guven Giizeldere

789

825

Preface and Acknowledgments

These are exciting times for thinking about con­ sciousness, and this book contains a coUection of essays that represents the cornerstones of contemporary philosophical thinking on the sub­ ject. Also included is a small corpus of articles representative of current psychology and neuro­ psychology research on consciousness that has given rise to fruitful discussions in the philosophy of mind. We hope that this book will be useful for philosophers in presenting a structured overview of the relevant literature on consciousness, bring philosophical problems of consciousness to bear on empirical issues being pursued in other fields, and be of interest to psychologists, neuroscien­ tists, and other mind-scientists. With the exception of chapter I on the stream of consciousness, a tribute to the remarkable body of William James's foundational work on consciousness and psychology in general, all the essays in this book belong to philosophers, psy­ chologists, and neuropsychologists of our own day. Many of these essays revolve around ongoing debates concerning the nature of consciousness, often cross-cutting disciplinary boundaries. Hence the title, "The Nature of Consciousness: Philo­ sophical Debates." Consciousness, as a subject matter of research in philosophy and in science, has a fascinating history. Perhaps there is no other phenomenon for which the pendulum of intellectual credibility swung to such extremes as it did for consciousness within roughly a matter of a single century, steering it through times of both exaltation and of taboo treatment. Consciousness was taken to be the "starting point of all psychology," only a few decades be­ fore it was condemned as being part and parcel of "superstition and magic." It was characterized as "the only true reality" as well as "a nonentity with no right place among first principles." It was regarded as "inert, uninftuential, a sim­ ple passenger in the voyage of life," as well as "an organ added for the sake of steering a nervous system grown too complex to regulate itself."

It was considered "something the meaning of which is known to everyone beyond doubt" as well as "something impossible to specify what it is, what it does, or why it evolved." (The quo­ tations belong to William James, John Watson, George Miller, Thomas Huxley,-as paraphrased by James-Sigmond Freud, and Stuart Suther­ land, respectively.) Perhaps it is now time to ask, borrowing Julian Jaynes's words: "This consciousness that is my­ self of selves, that is everything, and yet nothing at all-what is it?" The answer to this question, or attempts to formulate parts of such an answer, lie among the fifty essays in this book. But it is also important to ask: Where do we currently stand in the study of consciousness? We are witnessing an upsurge of interest in consciousness concurrently in several disciplines -most notably, philosophy, psychology, and neuroscience. We try to give an explanation of this trend within its historical context in the in­ troduction to this book. Today. it is not difficult to see the signs of the new "consciousness wave." In philosophy in the 1990s, several influential books have taken on the issue of consciousness: Daniel Dennett's Consciousness Explained (I 99 I ), Colin McGinn's The Problem of Consciousness (1991), Owen Flanagan's Consciousness Recon­ sidered (1992), John Searle's The Rediscovery of the Mind (1992), Paul Churchland's The Engine of Reason, The Seat of the Soul (1995), Fred Dretske's Naturalizing the Mind (1995), Michael Tye's Ten Problems of Consciousness (1995), David Chalmers's The Conscious Mind (1996), and William G. Lycan's Consciousness and Ex­ perience (1996). At present, an overwhelming number of new books exclusively on conscious­ ness are waiting in the pipelines of respected publishers. In contrast, there were fewer such books published between the 19608 and I 980s. Consciousness seems to have become the topic of the decade for philosophy of mind. In neuroscience and psychology, perhaps the biggest positive impact in favor of consciousness

Preface and Acknowledgments

research was due to Francis Crick and Christoph Koch. Their 1990 article, "Towards a Neuro­ biological Theory of Consciousness," carrying the secure stamp of approval by a Nobel laureate, helped shatter the current psychological and pro­ fessional barrier against the credibility of study­ ing consciousness in science. Around the same time, many studies on neuropsychology, espe­ cially those on blindsight, generated interesting and approximately concurrent results, which re­ lied on consciousness as an essential theoretial construct. Larry Weiskrantz's Blindsight (1986) and Bernard Baars's A Cognitive Theory of Con­ sciousness (1988), although they were published in the 1980s, have gained more popularity in the 1990s. There has also started a new industry of popu­ lar and semi popular books on consciousness. Dennett's Consciousness Explained, pitched at a level between laypeople and professional philos­ ophers, became a best-seller. In 1994, Crick put together his ideas on consciousness in the form of a popular book, The Astonishing Hypothesis: The Scientific Search for the Soul. Later in the same year, Roger Penrose followed up on his ideas from his earlier book on artificial intelligence, this time on consciousness: Shadows of the Mind:

Search for the Missing Science of Consciousness. Others are on their way. It is also worth men­ tioning the increased interest in consciousness in the scientific and semiscientific journals. See, for instance, the special "Mind and Brain" issue of Scientific American, which included conscious­ ness as one of the eight topics covered (September 1 992); the "10 Great Unanswered Questions of Science" of Discover magazine, which promoted the question, "What is consciousness?" (No­ vember 1992); and the special issue of Omni, "Science and the Soul" (October 1 993). (What is dubbed "soul" in the titles tum out to be "con­ sciousness" between the covers.) Finally, a number of new edited volumes and journals on consciousness are appearing. In ad­ dition, the recent frequency of major conferences specifically on consciousness (for example, the

Xll

success and popularity of the 1994 and 1996 Tucson conferences) is yet another indication of the new positive academic attitude toward con­ sciousness. Among the anthologies, Davies and Humphreys's Consciousness (1993), Metzinger's Conscious Experience (1995), and Velman's The Science of Consciousness (1996) stand out. Among the volumes that are worthy outcomes of major conferences are Marcel and Bisiach's Consciousness in Contemporary Science (1 988), Villanueva's Philosophical Issues: Consciousness (1991), Milner and Rugg's The Neuropsychology of Consciousness (1992), Bock and Marsh's Ex­

perimental and Theoretical Studies of Conscious­ ness, Revonsuo and Kamppinen's Consciousness in Philosophy and Cognitive Neuroscience (1 994), Cohen and Schooler's Carnegie Symposium Proceedings on Consciousness (in press), and Hameroff, Kaszhiak, and Scott's Towards a Scientific Basis for Consciousness (1 996). Among the journals, Consciousness and Cognition (since 1 992), Psyche (since 1993, first electronic journal on consciousness), and most recently Journal of Consciousness Studies (since 1994) are of spe­ cial interest. All this booming activity belongs to the recent time frame of less than a decade, and it marks a novel period for the study of consciousness. Now you may ask, among all the cited jour­ nals, conference proceedings, books, and anthol­ ogies, why should one acquire this book? Who would need it, and for what purposes can it be used? These are fair questions, and they have straightforward answers. Despite the rapid growing interest in con­ sciousness, no book has brought together all or most of the principal texts on the subject. This volume fills the gap, at least for philosophy, with a supplementary component for psychology and neuropsychology. The Nature of Consciousness contains fifty essays, including the Introduction, categorized under ten headings. Typically, each of these essays deals with a multitude of problems that are raised and debated in the literature about consciousness, and as such, would fall under

Preface and Acknowledgments

more than a single heading. This is because al­ most any single problem addressed in this volume is theoretically interrelated to a number of others, and proposals for resolution do not carry much weight unless their scope simultaneously spans over a whole set of connected issues. On the other hand, the intricacy of these inter­ relations sometimes results in theoretical tangles. For instance, as Giizeldere notes in the Intro­ duction, different issues about qualia, subjec­ tivity, the explanatory gap, and the knowledge argument recurrently get conflated and run to­ gether, at times making the controversies in the literature appear greater than they really are. Thus it is also important to try to "divide and conquer" the consciousness corpus. In the face of these opposing factors, the cate­ gorization of the articles in this volume into the present sections aims at both distinguishing dis­ crete areas of the relevant literature on the basis of distinct issues, as well as acknowledging the theoretical interconnections between them. Our goal is not only to bring together important es­ says on consciousness, but to present them within a superimposed structure so that this volume can be a road map for students of consciousness with which to explore and find major trails in the vast literature. To that end, we have also prepared a structured bibliography of additional readings on consciousness to serve as pointers for further study, and appended it to the end of the volume. In putting this collection together, we have tried to strike a balance between composing it as a classic reference book, and making sure that it presents the leading edge of philosophical re­ search on consciousness. Consequently, the pres­ ent volume contains seminal works of the past few decades, as well as new essays. Naturally, given the enormous size of the relevant literature, we had to restrict our choices. Thus, this book is mostly a philosophy anthology. Nonetheless, many of the entries in it do explain, rely on, or argue about current scientific thinking on consciousness (touching on, e.g., works of neuro­ scientists, cognitive neuropsychologists, AI re-

xiii

searchers, psychologists). Finally, this volume had to leave out entirely the works of continen­ tal philosophers (e.g., Husserl, Merleau-Ponty, Sartre); each of these intellectual lineages we chose not to trace here could easily form main themes for separate books. We hope that our anthology will serve many as a reference on debates on consciousness, a useful pedagogical tool in teaching medium- to advanced­ level undergraduate courses and graduate semi­ nars, and a guidebook for anyone trying to find his or her way through the large and confusingly interwoven literature on consciousness, for years to come.

The Essays The volume opens with an Introduction by Guven Giizeldere in which he attempts to lay out con­ ceptual foundations for the study of conscious­ ness in a broader intellectual context. Giizeldere traces many of the contemporary problems and controversies about consciousness back to their historical antecedents, demonstrating the recur­ rent nature of these issues in a wide variety of schools of thought in philosophy and psychology. Among them are the scope and reliability of the introspectionist method, the nature of the (Freudian versus cognitive) unconscious, the dis­ tinction between first-person versus third-person perspectives, the explanatory gap, the knowledge argument, the controversy of "consciousness mysterianism," the question of "zombies" and epiphenomenalism, and the "hard problem." Giizeldere then distinguishes among four "W­ questions" (What are the media and mechanisms of consciousness? Where, if anywhere, is the locus of consciousness? Who can be said to be a con­ scious being? Why is there consciousness at all?) and the "further-how" question (How does con­ sciousness arise in, or emerge from, its underlying substance, structure, and mechanism, in the way it does?) in an attempt to probe into the inner structure of a body of interrelated problems

Preface and Acknowledgments

about phenomenal consciousness. To that end, he further outlines a schema to tease out tangled issues about qualia, subjectivity, the knowledge argument, and "what it is like to be a certain organism." Finally, he presents an analysis of the current debates on consciousness in terms of a fundamental opposition between "essentialist" and "causal" intuitions, which he characterizes, respectively, by the mottos "consciousness is as consciousness feels" and "consciousness is as consciousness does." Giizeldere's overarching diagnosis is that what lies at the bottom of some of the most persistent problems about consciousness has to do with the explanatory gap between the characterization of conscious experiences as they are experienced by their subjects versus as they are investigated from a third-person perspective, but the epistemological nature of this gap, to the extent that epistemology and ontology are separable, does not necessa­ rily warrant ontological conclusions. Therefore, Giizeldere concludes, what seems the most prom­ ising direction in re-approaching consciousness and pursuing its deep-rooted· problems in the present era involves rethinking epistemology and conceptual schemes (as opposed to a priori postulation of new ontology) to yield a cross-fer­ tilization of the first-person and third-person per­ spectives, which would allow theorizing about the causal efficacy of how consciousness feels and the phenomenal quality of what consciousness does. Section I of this volume, Stream of Conscious­ ness, explores the nature and phenomenology of the inner goings-on of consciousness. In "The Stream of Consciousness," an all-time psychol­ ogy classic, William James (chap. I) starts out with the assertion that "the first and foremost concrete fact which everyone will affirm to belong to his inner experience is the fact that conscious­ ness of some sort goes on." He then explores the appearance and characteristics of the stream of consciousness: that it is continuous but in con­ stant change, that every object of consciousness has a "fringe," and that consciousness always functions as a selective agency.

XIV

Daniel Dennett (chap. 2) and Oweo Flanagan (chap. 3), in the two subsequent chapters, provide contemporary perspectives on James's idea. Ac­ cording to Dennett, the continuity of conscious­ ness is, contrary to James and the received view at present, only apparent. Furthermore, Dennett claims, this appearance of continuity stems from the ability of the brain to not "notice" gaps, rather than "fill them in" in some suitable fash­ ion. Flanagan stresses the singularity and defini­ tiveness of the stream of consciousness as it is phenomenally experienced, while acknowledging the discontinuities at the level of underlying neural processes. So long as we don't think: of consciousness as diaphanous, and expect to see through its causal structure, Flanagan concludes, the discontinuous nature of these underlying pro­ cesses should neither undermine nor compete with its description as a continuous stream. Section II, Consciousness, Science, and Meth­ odology, brings together a number of essays on the nature of consciousness, its place within phi­ losophy and science, and methodologies involved in its study. Owen Flanagan (chap. 4) gives a profile of prospects for a unified theory of con­ sciousness, with a positive outlook. He argues that the best strategy to study consciousness is what he calls the "natural method," a triangu­ lation of perspectives and data from phe­ nomenology, psychology/cognitive science, and neuroscience. As a demonstration of the natural method in action, Flanagan presents a multi­ faceted account of the phenomenon of dreaming, drawing on a wide array of philosophical, psy­ chological, and neuroscientific resources and combining them into a coherent explanatory framework. Alvin Goldman (chap. 5) explores the role of (phenomenal) consciousness in cognitive science, and argues that the folk-theoretic notion of qualia should not be hastily dismissed. Gold­ man suggests that qualia are the properties that uniquely allow us to recognize and classify our mental states, and, as such, they should be dis­ tinguished and credited for their significant role in scientific psychology and cognitive science.

xv

Preface and Acknowledgments

Patricia Churchland (chap. 6) outlines a full-scale,

Drafts model does not deal adequately with the

multilayered, reductionist research strategy for

Kantian view that consciousness involves con­

consciousness, locating it in the general context of

tinuous sensuous manifolds of space and time,

a reductionist scientific program. She argues that

and concludes, in accord with Block, that it

research on consciousness should allow for and

remains inadequate as long as it eschews the dis­

welcome simultaneous generation and co-evolu­

tinction between phenomenal and non-phenom­

tion of hypotheses at several levels of description,

enal representations in the brain that can play

from neurobiological to psychological, for an in­

different roles in the subject's mental life. Section

tegrated understanding of the phenomenon.

Daniel Dennett

and

Marcel Kinsboume

(chap.

III,

The

Psychology

and

Neuro­

psychology of Consciousness, consists exclusively

7) present their "Multiple Drafts" model of con­

of essays by psychologists and neuroscientists,

sciousness as an alternative to what they call the

who examine and explicate various lines of em­

"Cartesian Theater" model, a model based on the

pirical research that bear on philosophical prob­

idea that there is single place, a "finish line" in

lems about consciousness.

the brain where different components of experi­

�mard Baars (chap. 10) suggests "treating consciousness as a vari­

ences all come together for subjective judgment.

able" in empirical research, and delineates the

According to the Multiple Drafts model, aU there

differences between conscious and unconscious

is instead is a parallel stream of conflicting and

mental states and processes as constraints on any

continuously revised contents that can be probed

theory of consciousness.

at different times to elicit different narrative

gives a tutorial overview

Martha Farah (chap. 11)

accounts from the subject. However, there is no

psychological disorders of consciousness that

of

various neuro­

single, canonical narrative that, independently

result from brain damage, such as blindsight,

of probing, represents the actual stream

prosopagnosia, and neglect, and attempts to

of

locate them in a general theoretical framework.

consciousness. On the basis of this model, Dennett and Kins­

Edoardo Bisiach (chap. 1 2)

focuses on the phe­

bourne diagnose errors that result from an equiv­

nomenon of unilateral neglect as it reveals clues

ocation between representational properties of

about the spatiotemporal structure of the content

time and temporal properties of representations

of conscious visual experience, and he lends sup­

in the brain in such previously puzzling phenom­

port to Dennett and Kinsbourne's critique of the

ena. In particular, they criticize Benjamin Libel's

"Cartesian Theater" model. Tim

work on "backward referral in time," and pro­

13) gives an assessment of the modularity hy­

Shallice

(chap.

vide explanations for a number of previously

pothesis of mental function, especially with re­

puzzling temporal anomalies of conscious expe­

gard to conscious mental states, and provides

rience, such as the "color phi" and the "cuta­

Block (chap. 8) 9) present critical

motivation for studying consciousness as part of a

neous rabbit" phenomena. Ned

modular information-processing system.

and

Crick

Robert Van Gulick

commentaries on

(chap.

Dennett and

Kinsboume's

and

Christof Koch

(chap.

Francis 14) argue that

the unity of conscious visual experience (under­

work. Block accuses Dennett and Kinsbourne of

written by the "binding phenomenon") crucially

begging the question against phenomenal con­

depends on the semisynchronous firing of cortical

40-70 Hz. frequency range, and

sciousness by starting with the assumption that

neurons in the

there is no such thing as phenomenal conscious­

they present reasons and further pointers for the

ness, and dismissing the question of how to dis­

pursuit of consciousness research at the neural

tinguish between those revisions of contents that

level.

are phenomenally conscious, and those that are

In Section IV, Consciousness and Content,

not. Van Gulick points out that the Multiple

the relation between intentional (contentful) and

XVi

Preface and Acknowledgments

phenomenal (qualitative) properties of mental states is taken up.

Colin McGinn

(chap.

15) ex­

science and evolutionary biology, why epiphe­ nomenalism

(the

view

that

consciousness

is

plores aspects of the problem of intentionality

inconsequential to mental life in general) should

and the problem of consciousness that overlap

be rejected. In conclusion, Flanagan gives a cri­ tique of a distinction that Ned

seems configured by content." Nonetheless, he

draws between "phenomenal" and "access" con­

Block

(chap.

20)

with each other, observing that "phenomenology argues, the pessimism surrounding the problem

sciousness in the following chapter, arguing in

of consciousness does not have to spread to

favor of an alternative distinction between "in­

the problem of intentionality, insofar as the in­

formational sensitivity" (access to information

dividuation of content, unlike the mind-body

without phenomenal awareness) and "experien­

problem, is not inherently mysterious.

Davies

(chap.

Martin

16) argues that perceptual experi­

tial sensitivity" (access to information as well as phenomenal awareness).

ences, like beliefs, have representational proper­

In contrast, Block's distinction is based on the

ties, but perceptual content is non-conceptual and

thesis that phenomenal (experiential) properties

different in kind from belief content. He notes

of consciousness are different in kind from cogni­

that, as such, arguments for externalism for belief

tive, intentional, or: functional properties, and as

content do not readily apply to externalism for

such they constitute two distinct types of con­

perceptual content, and proceeds to give a sepa­

sciousness. According to Block, while the func­

rate line of argumentation for the latter.

Michael Tye

(chap.

17) also adopts the view

tional role of access consciousness in the guidance of rational thought and action is obvious, the

that experiences are representational but non­

function of phenomenal

conceptual. He argues, in particular, that the

properly conceptualized and not confused with

consciousness, when

phenomenon of pain is a proper object of study

its cognitive counterpart (as it often is in the

for cognitive psychology because pains have rep­

literature), is far from clear. Block points out

resentational content, and that their phenomenal

instances of this kind of confusion, which he

character is utilized in the operation of the cogni­

claims is prevalent among philosophers and

tive (belief/desire) systems.

neuropsychologists who write on consciousness,

(chap.

in the works of John Searle and Daniel Dennett,

Christopher Peacocke 18) draws a distinction between repre­

sentational and sensational properties of experi­

among others. He also presents a· number of

ences. In accordance with the historical distinction

thought experiments, as well as examples from

between perception and sensation, he then argues

the current neuropsychology literature, to sup­

that the latter, while indispensable for experi­

port his distinction.

ential content,

The four successive chapters, by Daniel Den­ nett (chap. 21), David Chalmers (chap. 22), Jennifer Church (chap. 23), and Tyler Burge (chap. 24), provide different perspectives on Block's

are

non-representational, and

demonstrates his claim via examples from visual perception. Section

V deals with the question of the Func­ Owen Flanagan (chap. 19)

tion of Consciousness.

essay.

defends the view that phenomenal consciousness

Block's distinction, arguing that it is untenable.

(he also calls it "subjective awareness") plays an

In contrast, Chalmers and Burge promote some

Dennett and Church refuse to accept

important role in our mental lives. He argues in

version of the distinction, even though they criti­

favor of teleological functionalism, a view that

cize Block in either not drawing it in the right

conceives of most mental capacities, conscious­

way, or stopping short of drawing its full con­

ness in particular, as typically having an adaptive

sequences. Dennett argues that Block's phenom­

role for their subjects. Flanagan then presents

enal

reasons, based on current research in neuro-

points to, in effect, a difference in degree among

versus

access

consciousness

distinction

Preface and Acknowledgments

xvii

conscious states of a single kind, with respect to

sciousness may have the unique functional role of

two dimensions: "richness of content" and "de­

"broadcasting information within the nervous

gree of influence." According to Dennett, Block

system."

inflates this difference in degree into an imaginary

In Section VI, Metaphysics of Consciousness,

difference in kind, and mislocates the question of

various metaphysical questions about the nature

the function of consciousness. Similarly, Church

and ontological status of consciousness are ex­

criticizes Block for not providing sufficiently good

plored. The well-known attack on identity theory

reasons to support the most fundamental ele­

presented by Saul Kripke (chap. 26) contends that

ments of his thesis, that the concept of phenom­

the identification of certain types of physical and

enal consciousness is coherent, and furthermore,

mental states involves unique difficulties that do

that the concepts of phenomenal and access con­

not plague other theoretical reductions in science,

sciousness are actually distinct. Chalmers, on the

such as the identification of temperature and

other hand, not only is sympathetic to Block's

mean kinetic energy of gas molecules. Kripke's

distinction, but also claims that Block stops short

argument is given in the broader framework of

of carrying his own line of reasoning to a further

rigid designation and necessity: both the term

conclusion-that, on the basis of this distinction,

"pain" and its counterpart "C-fiber firing" are

there simply is no function or explanatory role

rigid designators, and thus, if their referents are

that can be attributed to phenomenal conscious­

identical, they are necessarily identical. But they

ness within our mental lives. Burge suggests that

are not necessarily identical, Kripke maintains,

although there is a conceptual distinction between

given the possibility (based on the conceivability)

phenomenal and access consciousness, it must be

of either one existing in the absence of the other.

acknowledged that any notion of consciousness in

Therefore the thesis of identity between the men­

an individual, including access consciousness, has

tal and the physical fails.

to presuppose phenomenal consciousness. He

draws distinctions among ontological, theoret­

John

Searle (chap. 27)

also draws a further distinction between phe­

ical, logical, and causal kinds of reduction, and

nomenal consciousness and phenomenal qual­

argues in favor of causal reduction of conscious­

ities, where the latter can be instantiated as being

ness to its neurobiological underpinnings, while

either "felt" or "unfelt," and argues that for a

at the same time rejecting ontological reduction.

subject to be phenomenally conscious, the subject

Even though Searle agrees with Kripke on the

has to not only possess phenomenal qualities, but

failure of ontologically identifying the mental and

feel (or sense) them as well. Finally, Van

the physical, unlike Kripke he claims that this fact

(chap. 25) approaches the

stems from the pragmatics of our definitional

question of the function of consciousness by way

practices vis-a-vis consciousness and has no deep

of discussing the well-known "absent qualia"

metaphysical consequences for the unity of the

Gulick

objection to functionalism (see also Section

IX).

overall "scientific worldview" he independently

He argues that the reasoning that qualia have no

defends.

functional role, since for any system that pos­

role of conceptual analysis in philosophy of mind,

sesses qualia it is always possible to imagine

and argues that any credible materialist theory

Frank Jackson

(chap. 29) examines the

functional isomorphs of that system that lack

has to engage in some a priori reasoning in order

qualia, is question begging. Van Gulick maintains

to ontologically locate mental properties in the

that it is an open empirical question whether it

natural world, even in dealing with identifications

really makes no difference for a system to have

that involve a posteriori necessity. According to

phenomenal, as opposed to nonphenomenal,

Jackson, materialism cannot make a case for it­

mental representations, and he suggests in ac­

self unless it provides an account of a priori rela­

cord with Baars (chap.

10) that phenomenal con-

tions between the physical and the mental.

XV1l1

Preface and Acknowledgments

Georges Rey (chap.

28) adopts an eliminativist

losophy of mind classic, "What is it like to be a

perspective on phenomenal consciousness and

bat?" Thomas

claims that there is no place or need for our folk­

ment with Kripke (chap. 26), that the mind-body

theoretic notion of consciousness in the most

relation is unique in nature and doesn't yield to

Nagel (chap.

32) argues, in agree­

plausible accounts of mental life. He argues

reduction as other phenomena, e.g., lightning and

that this notion, which likens consciousness to an

electrical discharge, typically do. He character­

"inner light," is underwritten by the "Cartesian

izes an organism's possession of conscious states

Intuition" about the infallibility of first-person

as there being "something it is like to be" that

beliefs about consciousness. Once the Cartesian

organism, and claims that this "subjective char­

Intuition is dispelled, Rey concludes, we will see

acter" of experience is not capturable by any

good reasons for "doubting there really is such a

functional or causal analysis. For Nagel, physi­

John

calist accounts of any sort have to be objective,

30) argues for the vindication of the

and as such they necessarily remain incomplete,

common-sense notion of consciousness, and the

leaving out the ineliminable SUbjectivity of the

ontological thesis that "behavior, functional role,

mental.

and causal relations are irrelevant to the existence

mately arrives at is agnostic: he maintains not

of conscious mental phenomena." Searle grounds

that physicalism is false, but that "physicalism is

his argument on an epistemic distinction between

a position we cannot understand because we do

thing as consciousness at all." In contrast,

Searle (chap.

However, the conclusion Nagel ulti­

first-person and third-person points of view, and

not at present have any conception of how it

various thought experiments in which this dis­

might be true."

tinction plays a central role, all the while recom­ mending that regarding "possibility scenarios"

Both Colin McGinn (chap. 33) and Josepb Levine (chap. 34) take up the issues Nagel raises,

involving consciousness, one should always ask

but carry them in different directions to different

oneself: "What would it be like for me?" In ac­

conclusions. McGinn argues not only that we

cord with Searle's suggestion,

have no conception of how the mind-body

Sydney Sboemaker

(chap. 31) turns to epistemology of modality, and

problem can be solved (e.g., how physicalism

explores the role and scope of the first-person

can explain consciousness), but that we should

perspective in assessing metaphysical claims about

never expect to acquire such a conception. Ac­

the mind. In particular, he examines Kripke's

cording to McGinn, an adequate explanation of

possibility argument against the identity of the

the mind-body problem will forever elude human

mental and the physical, and Searle's thought

understanding, not because it is miraculous or

experiments designed to establish the ontological

involves

independence of consciousness from behavioral,

transcends our cognitive abilities, just as the

divine

intervention,

but

because

it

functional, and causal elements. Shoemaker's

understanding of quantum mechanics forever

general approach is sympathetic to the first­

transcends the intellectual abilities of monkeys.

person perspective, but he presents reasons that

This thesis makes McGinn a nonconstructive

undermine the kind of absolute authority Searle

naturalist.

seems

Kripke's work, that there is an "explanatory gap"

to be suggesting. The moral Shoemaker ul­

Levine

also

thinks,

drawing

on

timately draws is also very different from Searle's:

between theories that employ solely physical

whenever the first-person perspective seems to

properties, on one hand, and phenomena that in­

present a metaphysical possibility, "one should

volve qualitative character on the other. How­

always ask whether the seeming discovery could

ever, unlike McGinn, he doesn't go on to make

be conflnne d from the third-person point of view."

the further claim that this gap will remain forever

Section

VII,

Subjectivity

and

Explanatory

Gap, begins with an essay that has become a phi-

for

us

human beings. Moreover, unlike Kripke,

Levine argues that the explanatory gap points

Preface and Acknowledgments

not to an ontological, but to a merely epistemo­ logical problem for physicalism. He concludes that as we improve our concept of qualitative character so as to include causal elements in it, the explanatory gap is likely to close. Section VIII comprises five articles that focus on yet another form of argument designed to show that physicalism fails to account for con­ sciousness. The Knowledge Argument is based on a thought experiment that involves a vision­ scientist, Mary, who learns "everything physical there is to know" about color experiences but never experiences colors herself due to her special circumstances (being raised in a monochrome environment). The point of the argument is that Mary would learn something new upon her first experience of colors, and since she was supposed to already know, ex hypothesi, everything she could know in a physicalist framework, physi­ calism is false, or at best incomplete. This argu­ ment has generated a number of different but interrelated responses, and Robert Van Gulick (chap. 35) gives a schematic summary of various objections raised against it in the literature, in­ cluding his own. (See Loar [chap. 39] below for a line similar to Van Gulick's response to the knowledge argument.) Van Gulick also examines the issue of the explanatory gap in the second half of his essay, and concludes that physicalism is not refuted by either family of arguments. Frank Jackson (chap. 36) and Paul Cburchland (chap. 37) present a continuation of an exchange between the authors that started with Jackson's "Epiphenomenal Qualia" (Philosophical Quar­ terly, 1982, 32:127- 36), where the present form of the knowledge argument first appeared. Church­ land's "Reduction, Qualia and the Direct Intro­ spection of Brain States" (Journal of Philosophy, 1985, 82:8-28) presented a critique of the argu­ ment and charged Jackson with committing an intensional fallacy on the basis of an equivocation in his use of "knowing about" (brain states versus experiences); Churchland also argued there that Mary could very well be able to imagine what it would be like to have color experiences. In that

xix

case, Mary would not learn anything new upon actually having a color experience, and therefore physicalism would remain unrefuted. In the present part of the exchange, Jackson restates the argument, and emphasizes that it doesn't rest on the claim that Mary cannot imag­ ine what having color experiences is like unless she actually has them. Jackson's claim is that Mary, as a matter of fact, would not know the experiential facts, and this in itself is sufficient to show that physicalism is false. Jackson further adds that the intensionality of "knowing about," just like the powers of Mary's imagination, is or­ thogonal to the validity of his argument. Finally, he argues that the kind of knowledge that is to the point is Mary's knowledge about the experiences of others, not about her own. Churchland, in his rejoinder, insists that the equivocation of "knowing about" indeed invalidates Jackson's argument, and he presents an entirely physicalist alternative model of visual experience in which the premises of Jackson's argument come out true but the conclusion false. Hence, Churchland concludes, Jackson's argument is invalid, and physicalism survives his attack unscathed. Churchland, while discussing possible re­ sponses to Jackson's argument, considers an­ other approach but states that it is, although well-motivated, scientifically unsatisfactory. This approach is based on a distinction between "knowing how" and "knowing that," due to Laurence Nemirow and David Lewis, and is commonly known as the "ability analysis" of the knowledge argument. It maintains that what Mary gains upon actually having a color experi­ ence is not knowledge of new facts or proposi­ tions, which would undermine physicalism, but a new recognitional ability that involves color sen­ sations. David Lewis (chap. 38) advances a de­ tailed defense against Jackson on the basis of this line of reasoning: Learning what a particular experience is like amounts to gaining certain new abilities to remember, imagine, and recognize, and if gaining new abilities means gaining new information, there is good reason to think that

xx

Preface and Acknowledgments

this new information is neither a special kind of

and "inverted spectrum" arguments. In the first

irreducibly non-physical, "phenomenal informa­

piece of the section,

tion," nor propositional information represented

starts out with a common-sense description of

in the form of knowledge or belief. Therefore,

qualia as "the ways things seem to us," but

Daniel Dennett

(chap.

40)

Lewis concludes, the fact that one cannot learn

quicldy moves on to a more specific set of defini­

what a new experience is like unless one actually

tional criteria that he thinks

has it, is not to be taken as a measure of the limits

terizes the philosophically problematic notion of

of science or the metaphysics of mind.

essentially

charac­

qualia. According to Dennett, what makes qualia

In the last article of this section, Brian Loar

lie at the bottom of notoriously vexing problems

39) defends physicalism against a number

in philosophy of mind and metaphysics is the way

(chap.

of objections, among them SUbjectivity, the ex­

qualia are traditionally taken to be "ineffable, in­

planatory gap, and the knowledge argument. His

trinsic, private, and directly or immediately ap­

overarching solution is based on a distinction be­

prehensible in consciousness," supported by a

as they figure in

belief in the infallibility and incorrigibility of first­

discussions about the mental versus the physical.

person epistemic authority. Dennett's aim is not

tween properties and concepts,

Loar argues that our phenomenal and functional

only to show that, under these criteria, there are

( physical) concepts are distinct, and that while

no properties that fit the bill, but also to "make it

this accounts for the presence of the explanatory

prima facie

just as uncomfortable for anyone to talk of qualia

intuitive appeal of the

[ ... ] with the standard presumption that they,

knowledge argument, it doesn't imply that the

and everyone else, knows what on earth they are

gap and the

phenomenal and functional properties are not

talking about." To this end, he presents fifteen

identical. In accord with the general line of the

"intuition pumps" designed to demonstrate the

ability analysis, Loar characterizes phenomenal

conceptual confusions involving the notion of

as recognitional concepts, grounded in

qualia. Ultimately, Dennett turns his strategical

discriminatory and recognitional dispositions of

choice that it is "far better, tactically, to declare

the subject, but he differs from Lewis in accepting

that there simply are no qualia at all" into an

that Mary comes to know new facts or proposi­

ontological resolution, and he concludes that

concepts

tions in having color experiences for the first time.

"contrary to what seems obvious at first blush,

Loar does not have to accept Jackson's verdict

there simply are no qualia at all."

that physicalism is false, however, for the mode of

The next three articles center around the possi­

individuation on which Loar takes Mary to ac­

bility of a systematic difference between color

quire new propositions is laden with conceptual

experiences of subjects (or, the same subject at

constituent structure. Thus what Mary learns

different times) that yields a complete inversion in

makes reference not to new properties, but to

their experienced "hue circle" which, nonetheless,

properties she has already referred to in the past,

does not manifest itself in language, or, more

albeit by means of different concepts. The con­

generally, in behavior. This possibility of an in­

clusion Loar arrives at at the end of his essay

verted spectrum (together with the different pos­

expresses the gist of the physicalist approaches

sibility of absent qualia, the complete absence of

represented in this section: "Nothing in philoso­

qualitative aspects of experience) is regarded as

phy should make one doubt either that (phenom­

one of the most tenacious obstacles to function­

enal] qualities are real or that they are physical."

alism, as well as to physicalism in general. The

Section

IX

is a continuation of the debates on

difficulty lies in the implication that qualia can

the nature and ontological status of phenomenal

neither be defined in terms of their functional role

(qualitative) properties of experience, Qualia,

or physical make-up, nor be captured in a func­

with a particular focus on the "absent qualia"

tionalist or physicalist explanatory framework.

Preface and Acknowledgments

The disagreements in the qualia debates are often underwritten by fundamental differences regarding the ontology of qualitative properties of experience, for instance differences over whether qualia form a class of properties meta­ physically autonomous from intentional/repre­ sentational, functional, or causal properties. This issue also figures as a substratal element in the debate in the present section. Drawing a distinction between "intentional" versus "qual­ itative" senses in which experiences may resemble or differ from one another, Sydney Shoemaker (chap. 4 1 ) opts for characterizing qualia as dis­ tinct from intentional properties. He maintains, however, that neither the absent qualia nor the inverted spectrum argument constitutes a refu­ tation of functionalism, even though the latter presents a more veritable difficulty, since it in­ volves, for Shoemaker, a genuine possibility, whereas the former does not. Shoemaker bases the plausibility of the full-scale intersubjective in­ verted spectrum scenario on the epistemically less problematic possibility of intrasubjective spec­ trum inversion, and he argues that while the functional/behavioral evidence may establish in­ tentional similarities and differences between the (mental states of) subjects, it cannot capture qualitative similarities and differences. Shoe­ maker concludes his essay with a discussion of varieties of functionalist approaches to qualia. Gilbert Hannan (chap. 42) presents a defense of functionalism against three objections commonly raised in the literature: the claim that functional­ ism cannot account for the "intrinsic qualities" of experience, the knowledge argument, and the inverted spectrum. Against the first objection, Harman argues that the claim that we are ordi­ narily aware of the intrinsic properties of our experiences, as defended by sense-data theorists, involves a fallacy that stems from a confusion between "the properties of a represented object and the properties of a representation of that ob­ ject." According to Hannan, in perceiving an object we are aware of not the intrinsic properties of our experience (the representation, or the rep-

xxi

resenting-vehicle of the object), but the properties of what we experience (the represented object). Thus, Harman concludes, the claim that func­ tionalism fails to capture a psychologically rele­ vant aspect of experiences because it takes into account solely non-intrinsic properties has no theoretical force. Harman's defense against the knowledge argument differs from those of Van Gulick (chap. 35), Churchland (chap. 37), Lewis (chap. 38), and Loar (chap. 39) in that he accepts that Mary comes to learn a new fact upon having color experiences, at the same time rejecting Jackson's premise that "Mary could know every­ thing physical there is to know" while confined to her monolithic environment. Harman claims that Mary could not have a proper concept of redness (or any other color) since having concepts, being functionally determined, is dependent on the presence of perceptual inputs from red objects­ something Mary, ex hypothesi, lacks. But then Mary could not possibly acquire aU the facts about color experiences prior to having those ex­ periences, Harman concludes, because "mental representations are constructed from concepts," and "to know what it is like to see something red is to be capable of representing [to oneself] something's being red." Finally, Hannan maintains that functionalism can plausibly reject the possibility of the inverted spectrum, unless the sense-data theorist's thesis that perception involves the awareness of intrinsic features of experience is held as a prior assump­ tion. Having already argued against this thesis in the first objection, he states that in normal per­ ception "there can be no distinction between how things look and how they are believed to be," thereby rejecting the distinction Shoemaker offers between intentional and qualitative properties of mental states. Consequently, Harman concludes, spectrum inversion is a possibility only when it comes simultaneously with inversion of belief and meaning. Since the latter can be captured in a functionalist framework, the inverted spec­ trum argument presents no genuine difficulty for functionalism.

Preface and Acknowledgments

xxii

Ned Block (chap. 43), in response to Harman,

objects, as well as the color vocabulary of its in­

presents the "inverted earth" argument, while also

habitants, are inverted. In other words, the expe­

challenging Harman's critique of the knowledge

riences of the natives of inverted earth have

argument. Regarding the latter, Block maintains

inverted qualitative as well as intentional content

that Harman's rejection of the first premise of

with respect to ours. Block's contention is that on

the knowledge argument, that Mary can learn

inverted earth, while the qualitative contents of

"everything physical there is to know" without

our visual experiences would remain the same

having color experiences herself, at best estab­

(the two inversions cancelling each other out),

lishes that "the knowledge argument does not

after a period of adaptation into our new physical

serve very well as a locus of controversy" between

and linguistic environment (but still without

those who take qualia to be identical to inten­

the knowledge of the fate that has befallen us),

tional (broadly construed) properties and those

the "intentional contents would shift so as to be

who do not. Block chooses to bring out this con­

the same as those of the natives." The result is an

troversy in his inverted earth thought experiment,

exemplification of a case where intentional con­

a variation of the inverted spectrum scenario that,

tent is inverted while qualitative content remains

he maintains, avoids the complications of the

invariant. This, Block concludes, is a vindication

original formulation. He argues both in favor of

of the inversion objection against functionalism,

the existence of intrinsic mental features of expe­

without recourse to the original inverted spec­

rience, and against the thesis that functionalism

trum argument.

can unproblematically reject the possibility of the

In summary, while in this debate Harman takes

inverted spectrum, and he grounds these two ar­

the thesis that in perception we become aware of

guments on the kind of distinction Shoemaker

intrinsic properties of experience as a prior as­

defends and Harman rejects: a distinction be­

sumption that is required by the possibility of

tween intentional and qualitative properties (or

spectrum inversion, Block claims that it is the

content) of experience. According to Block, in­

possibility of spectrum inversion that will, once

tentional content is functionally characterizable

convincingly established (as he tries to do with

but qualitative content is not, and Harman's

the inverted earth argument), "settle the issue of

identification of the two involves a fallacy, "the

whether we are aware of intrinsic features of

fallacy of intentionalizing qualia." Block asserts

experience" (by showing that there indeed are

that the inverted spectrum scenario is designed

intrinsic

precisely to drive a wedge between these two

such, Harman and Block seem to agree, in effect,

mental

features of our experience). As

types of content, for it demonstrates the pos­

on the following conditional as a premise in their

sibility of the invariance of intentional content

respective arguments: if spectrum inversion is a

(belief and meaning) in the case of inversion of

genuine possibility,

qualitative content (qualia).

awareness of intrinsic properties of experience.

Block's inverted earth scenario is constructed

then

perception involves the

Harman then goes on to argue that the idea that

to yield the same theoretical result, though it is

perception involves the awareness of intrinsic

the converse of the inverted spectrum case, as its

properties of our experience must be rejected, for

aim is to demonstrate the invariance of qual­

it involves a fallacy based on "confusing intrinsic

itative content in the case of inversion of inten­

features of the intentional object of experience

tional content. In this scenario, we are asked

with intrinsic features of the experiences," and he

to imagine ourselves being fitted with color­

concludes that the inverted spectrum argument as

inverting lenses, without our knowledge, and

an objection against functionalism is untenable.

transported to "inverted earth," a duplicate of

In contrast, Block argues that inversion argu­

our earth except for the fact that the colors of all

ments against functionalism are tenable on the

xxiii

Preface and Acknowledgments

basis of his improved version of the inversion ar­

there is an a priori conceptual connection be­

gument, and he concludes that this establishes

tween qualitative and functional properties in

what Harman rejects on the basis of his analysis

virtue ofthe meaning of terms such as "pain" and

of intentionality of experiences, namely that

expressions that describe the typical correspond­

experiences have intrinsic mental properties of

ing causal/functional dispositions.

which we become aware in perception. ( Harman

For White, the criteria that distinguish the two

also attempts to support his analysis of inten­

versions of physicalist-functionalism are the po­

tionality by introspective evidence but Block dis­

sitions each takes with respect to absent qualia

misses this effort as "an error in philosophical

and the inverted spectrum. Block differs from

method. ")

Shoemaker in that he believes in the possibility of

If this characterization of the controversy is

both (and so holds that functionalism cannot dis­

correct, then an interesting feature in the dialectic

tinguish among creatures with normal, absent, or

of the disagreement between Harman and Block

inverted qualia) whereas Shoemaker rejects the

is worth pointing out: while Harman rejects the

former while accepting the latter (thereby up­

consequent of the conditional premise they both

holding that functionalism can characterize what

accept, which leads him to reject the antecedent,

it is like to have experiences with qualitative con­

Block argues in favor of the antecedent, which

tent in general, but not what it is like to have

leads him to accept the consequent. It seems, in

experiences with particular qualitative content).

modus

White argues that while Shoemaker's position

tol/ens becomes Block's modus ponens. In the last article of this section, Stephen White

tionalism or Block's position, Block's brand of

other words, that in this debate Harman's

(chap.

ultimately collapses into either orthodox func­

44) presents a detailed analysis of qualia

physicalist-functionalism ends up with the un­

arguments and their consequences for physi­

happy consequence of "saddling itself with all the

calism and (orthodox) functionalism, as well

problems of pre-functionalist physicalist theo­

as what he calls "physicalist-functionalism" and

ries," while being deprived of "the only plausible

"transcendentalism." Physicalist-functionalism is

solution those theories had generated," namely

the view that certain aspects of a subjecCs mate­

topic-neutral translation of mental terms. White's

rial constitution, even if they can vary without

conclusion is that inversion arguments fail to

varying functional constitution, are relevant to

provide any theoretical support for physicalist­

the nature of the subject's qualia; as such it is a

functionalism over orthodox functionalism, and

compromise between physicalism and function­

although orthodox functionalism "fails to cap­

alism. Transcendentalism differs from all varieties

ture

of physicalism and functionalism in maintaining

ing transcendentalism aside, it is the best theory

that the qualitative character of mental states

of the mental at hand.

all of our intuitions

regarding qualia," leav­

cannot be captured in either physical or func­

Section X, the last section of the volume, brings

tional terms, and that experiential facts are es­

together five articles on the Higher-Order Mon­

sentially irreducible "first-person" facts. White

itoring Conceptions of Consciousness. The de­

distinguishes two versions of physicalist-func­

bate in this section centers around the idea of

tionalism and attributes them to Ned Block and

construing consciousness in terms of the aware­

Sydney Shoemaker, respectively, while reserving

ness (or, more generally, some form of higher­

transcendentalism for Thomas Nagel. The view

order representation) of first-order mental states.

he ultimately defends against the alternatives is one he attributes to Gilbert Harman, orthodox

David Armstrong (chap. 45), David Rosenthal (chap. 46), and William Lycan (chap. 47) all

functionalism (akin to what is elsewhere called

defend this view,

"analytic functionalism"), according to which

Lycan characterize the higher-order representa-

but while Armstrong and

Preface and Acknowledgments

tion as "perception-like," Rosenthal espouses the "higher-order-thought" characterization of con­ sciousness. According to these "higher-order monitoring conceptions of consciousness" (so­ called because they sometimes liken conscious­ ness, in Armstrong's words, to "a self-scanning mechanism in the central nervous system"), what makes a belief, a perceptual state, or a particular pain conscious is the subject's having an accom­ panying higher-order awareness, in the form of perception or thought, of that first-order mental state. David Armstrong, in an attempt to develop an "anti-Cartesian" account that regards conscious­ ness as not the essence but a special and sophisti­ cated development of mentality, distinguishes three kinds of consciousness and identifies con­ sciousness "in the most interesting sense of the word" with a "perception-like awareness of cur­ rent states and activities in our own mind." This is what Armstrong calls "introspective conscious­ ness" (unlike Lycan and Rosenthal, who call it simply "consciousness"), and he regards it a late evolutionary development and a faculty highly specific to human beings, in contrast with "min­ imal consciousness" (the state of a creature who is normally conscious but at the time sound asleep or knocked out) and "perceptual consciousness" (first-order awareness one has of one's environ­ ment and bodily states). For Armstrong, intro­ spective consciousness (which he essentially identifies with Kantian "inner-sense") is theoret­ ically on a par with ordinary sense perception ("outer-sense"), in that they involve partial awareness of current states and activities of one's mind, and one's body and environment, respec­ tively. Furthermore, both are selective, fallible, and causally characterizable. Armstrong con­ cludes that the significance of introspective con­ sciousness lies in the fact that it provides us with the awareness of the past history and present status of ourselves: "Without introspective conscious­ ness, we would not be aware that we existed." William Lycan also advocates a charac­ terization of consciousness, which he calls the

xxiv

"Inner Sense" view, as a "perception-like second­ order representation of our own psychological states and events." He defends this view against a number of questions and objections raised by various philosophers, including Daniel Dennett (who argues that this model allows the meta­ physically dubious possibility of consciousness of a stimulus in the absence of the subject's belief in that consciousness), Fred Dretske (who ques­ tions why consciousness of certain physical states is enough to make those states themselves con­ scious), David Rosenthal (who argues the pos­ sibility of a regress threatens if second-order perception involves, as perceiving always does, some sensory quality), Christopher Hill (who charges that second-order attention to a mental state may transform that state into another), and Georges Rey (who objects that if all it takes to make a first-order state a conscious state is that it be monitored by a scanner that makes integrative use of the relevant information, then any note­ book computer can be said to have conscious states). Having also noted similarities and differ­ ences between his, Armstrong's, and Rosenthal's views, Lycan concludes that the Inner Sense view of consciousness surVives all criticisms unscathed. David Rosenthal differs from Armstrong and Lycan in relying on thinking, rather than per­ ceiving, as the form of higher-order awareness in his model of consciousness. For Rosenthal, "a mental state is a conscious state when, and only when, it is accompanied by a suitable higher-or­ der thought." As such, the question of what it is for a mental state to be conscious can be sepa­ rated and pursued independently from the ques­ tion of what it is for mental states to have intentional or sensory properties. This is so, Rosenthal argues, because a state's being con­ scious is not intrinsic to its having intentional or sensory character, and further that if it were so intrinsic, "no theoretical understanding of what it is to be a conscious state would be possible at all." Drawing on a distinction between transitive and intransitive consciousness (in the sense of our being conscious of something versus our mental

xxv

Preface and Acknowledgments

states' being of the conscious or unconscious

ceiving or thinking; he in fact speculates that in­

kind), Rosenthal goes on to develop the details of

trospection is probably akin to coming to have

his "higher-order thought" theory of conscious­

thoughts that one has mental states such as be­

ness and argues that the explanatory scope and

liefs, desires, and experiences of a certain kind.

power of the theory are enhanced, as it success­

Rather, the thesis he argues against is that what

fully predicts certain important aspects of our

makes mental states conscious consists in the

mental lives (e.g., that reporting one's mental

subject's awareness of those states. According to

states is distinct from verbally expressing them,

Dretske, what makes our internal states conscious

and that we become aware of more fine-grained

is, instead, "the way

differences among sensory qualities as we acquire

something else-the world we live in and (in pro­

more fine-grained relevant conceptual distinc­

prioception) the condition of our own bodies. "

tions). After criticizing the "higher-order per­

Finally,

they

make us conscious of

GuveD Giizeldere

(chap.

49) distin­

ception" theories of consciousness, Rosenthal

guishes between the higher-order perception and

concludes that the higher-order thought theory

higher-order thought theories of consciousness by

constitutes our best hope of providing a credible

specifying their underlying assumptions in the

account of consciousness while doing justice to

writings of Armstrong, Lycan, Churchland, and

our pretheoretic intuitions about the subject

Rosenthal, and argues that these two varieties of

matter.

higher-order representation theories, upon close

48), in contrast to Arm­

examination, do not survive as distinct compet­

strong, Lycan, and Rosenthal, argues that a sub­

ing models. In particular, he demonstrates that

Fred Dretske

(chap.

ject can have conscious experiences without

the higher-order perception accounts face a tri­

necessarily being aware of those experiences, or

lemma, if they are literally taken to be modeled

aware of having them, and he thereby attempts to

after perception: they either must abandon their

sever the conceptual link between consciousness

characteristic two-tiered structure, or commit

of mental states and our higher-order awareness

themselves to what he calls "the fallacy of the

of those states. Dretske demonstrates his point by

representational divide," or tum into a species of

a number of examples, after having drawn a dis­

their competitor, higher-order thought theories.

tinction between "consciousness of things" (ex­

Taking the first hom of this trilemma means sim­

periences that do not necessarily involve any

ply giving up the essential introspective element

concepts about what is being experienced) and

in the theory, Giizeldere maintains, while taking

"consciousness of facts" (beliefs that necessarily

the second hom involves a confusion between

deploy such concepts). Dretske's examples rely on

properties of what is represented in consciousness

two theses which he tries to motivate indepen­

and properties of that which represents what is

dently: that consciousness of a particular object

represented. Giizeldere concludes that if Locke's

does not necessarily entail consciousness of any

dictum that "consciousness is the perception of

facts about that object, and that a subject's being

what passes in a man's own mind" cannot there­

conscious of an object or a fact implies that that

fore be taken literally, to the extent that higher­

subject is in (or has) a conscious state of some

order thought theories prove unsatisfactory, we

sort. Having introduced this conceptual appara­

should look in the direction of developing uni­

tus, he goes on to show that we can sometimes

level, nonintrospective accounts of consciousness.

have experiences that are conscious in the absence of our consciousness of them, or our conscious­ ness of the fact that we are having them. Dretske's objection is not that we never form a higher-order awareness of our own mental states, or that such awareness is necessarily unlike per-

Acknowledgments We thank the three anonymous reviewers who provided several useful suggestions on the content

xxvi

Preface and Acknowledgments

and fonn of this collection. We express our thanks to Teri Mendelsohn, fonnerly of Brad­ ford Books/MIT Press, who patiently and pro­ fessionally accommodated our endless requests, alterations, and disagreements over the elongated period of the preparation of this book. Thanks also to Larry Russell, Murat Aydede, Altug Giizeldere, and Tomiko Yoda for editorial help, and to Judith Feldmann, Pam Quick, Sandra Minkkinen, and Betty Stanton

of The

MIT

Press for overseeing the final editing and pro­ duction of this book and bringing our efforts to completion.

�ven �zeldere

Sources

Giiven Giizeldere Introduction Adapted from "Consciousness: What it Is, How to Study it, What to Learn from its History" and "Prob­ lems of Consciousness: A Perspective on Contemporary Issues, Current Debates" in JOU17/Q1 of Consciousness Studies, 2: 1 , 30-51, and 2:2, 1 12-143, (1995), by per­ mission of the author and publisher. William James "The Stream of Consciousness" From Psychology, Chap XI (New York: Henry Holt and Co., 1910), 1 5 1 - 1 75. Daniel C. Dennett "The Cartesian Theater and 'Filling In' the Stream of Consciousness" Adapted from Consciousness Explained (Boston: Little, Brown, and Co., 1991), by permission of the author and publisher. Owen Flanagan "The Robust Phenomenology of the Stream of Con­ sciousness" From Consciousness Reconsidered (Cambridge: MIT Press, 1992), 1 69-1 75, by permission of the author and publisher.

Ned Block "Begging the Question Against Phenomenal Con­ sciousness" From Behavioral and Brain Sciences, 15 (1992), 205206, by permission of the author and publisher. Robert Van Gulick "Time for More Alternatives" From Behavioral and Brain Sciences, 1 5 (I992), 228229, by permission of the author and publisher. Bernard J. Baars "Contrastive Phenomenology: A Thoroughly Empirical Approach to Consciousness" Adapted from a previous version that appeared in Psy­ che, 1 :2 (1 994), 32-55, by permission of the author and publisher. Martha Farah "Visual Perception and Visual Awareness after Brain Damage: A Tutorial Overview" From A ttention and Performance xv: Conscious and Nonconscious Information Processing, eds. Carlo Umilta and Moris Moscovitch (Cambridge: MIT Press, 1 995), 37-75, by permission of the author and pub­ lisher.

Owen Flanagan "Prospects for a Unified Theory of Consciousness" From Scientific Approaches to the Study of Conscious­ ness: 25th Carnegie Symposium, eds. J. Cohen and J. Schooler (Hillsdale: J. Erlbaum Associates, 1996), by permission of the author and publisher.

Edoardo Bisiach "Understanding Consciousness: Clues from Unilateral Neglect and Related Disorders" From The Neuropsychology of Consciousness, eds. A. D. Milner and M. D. Rugg (London: Academic Press, 1992), 1 1 3-1 37, by permission of the author and pub­ lisher.

Alvin Goldman "Consciousness, Folk Psychology, and Cognitive Sci­ ence" From Consciousness and Cognition, 2 (1993), 364-382, by permission of the author and publisher.

Tim Shallice "Modularity and Consciousness" From From Neuropsychology to Mental Structure (Cambridge: Cambridge University Press, 1988), 381404 , by pennission o f the author and publisher.

Patricia ChurcWand "Can Neurobiology Teach Us Anything About Con­ sciousness?" From the APA Proceedings, 67(4): 23-40, by permis­ sion of the author and publisher.

Francis Crick and Christof Koch "Towards a Neurobiological Theory of Consciousness" From Seminars in the Neurosciences, 2 (1990), 263-275, by permission of the author and publisher.

Daniel C. Dennett and Marcel Kinsboume "Time and the Observer: The Where and When of Consciousness in the Brain" From Behavioral and Brain Sciences, 1 5 (1 992), 1 83247, by permission of the author and publisher.

Colin McGinn "Consciousness and Content" From The Problem of Consciousness (Oxford: Black­ well, 1991), 23-43, by permission of the author and publisher.

xxviii

Sources

Martin Davies

John R. Searle

"Externalism and Experience"

"Reductionism and the Irreducibility of Consciousness"

Philosophy And Cognitive Science: Categories, Consciousness, and Reasoning, eds. A. Clark, J. Ez­

From

From

querro, and J. M. Larrazabal (Dordrecht: Kluwer Aca­ demic Publishers,

1996), 1 -33,

by permission of the

author and publisher.

The Rediscovery of the Mind, Chap. 5 (Cam­ 1 992), 1 1 1 - 126, by permission of the

bridge: MIT Press,

author and publisher. Georges Rey "A Question about Consciousness"

Perspectives on Mind, eds. H. Otto and J. Tueidio 1988), 5-24, by

Michael Tye

From

"A Representational Theory of Pains and Their Phe­

( Norwell: Kluwer Academic Publishers,

nomenal Character"

permission of the author and publisher.

From:

Philosophical Perspectives,

Vol.

9,

ed. J. Tom­

berlin (Atascadero: Ridgeview Publishing Co.,

223-239, by permission of the

1990),

author and publisher.

Christopher Peacocke

Frank Jackson "Finding the Mind in the Natural World"

Philosophy and the Cognitive Sciences: Proceed­ ings of the 16th International Wit/genstein Symposium, From

"Sensation and Content of Experience: A Distinction"

eds. R. Casati, B. Smith, and G. White (Vienna: Verlag

Sense and Content (Oxford: Clarendon Press, 1983), 4-26, by permission of the author and publisher.

Holder-Pichler-Tempsky,

From

1 994), 101-1 12,

by permis­

sion of the author and publisher.

Owen Flanagan

John R. Searle

"Conscious Inessentialism and the Epiphenomenalist

"Breaking the Hold: Silicon Brains, Conscious Robots,

Suspicion" From: Press,

and Other Minds"

Consciousness Reconsidered (Cambridge: MIT 1992), 1 29-152, by permission of the author and

From

The Rediscovery of the Mind, Chap. 3 (Cam­ 1992), 65-82, by permission of the

bridge: MIT Press,

publisher.

author and publisher.

Ned Block

Sydney Shoemaker

"On a Confusion About a Function of Consciousness"

"The First-Person Perspective"

From

Behavioral and Brain Sciences, 1 8 (1995), 227-

From the APA Proceedings,

247, by permission of the author and publisher.

of the author and publisher.

Daniel C. Dennett

Thomas Nagel

68(2): 7-22, by permission

"What Is it Like to Be a Bat?"

"The Path Not Taken"

253, by permission of the author and publisher.

From Philosophical Review, 83:4 (1974), permission of the author and publisher.

Jennifer Church

Colin McGinn

From Behavioral and Brain Sciences,

18 (1995), 252-

by

"Can We Solve the Mind-Body Problem?"

"Fallacies or Analyses?" From

435-450,

Behavioral and Brain Sciences, 1 8 (1995), 251 -

From

Mind, 98:891 (1989), 349-366,

252, b y permission o f the author and publisher.

the author and publisher.

Robert van Gulick

Joseph Levine

"Understanding the Phenomenal Mind: Are We All

"On Leaving Out What It's Like"

Just Armadillos?", Part II

From

Consciousness, eds. M. (Oxford: Blackwell, 1993),

Consciousness, eds.

M. Davies and G. Hum­

1993), 1 2 1 - 1 36,

Adapted from a version in

phreys (Oxford: Blackwell,

Davies and G. Humphreys

sion of the author and publisher.

149-1 54, by permission of the author and publisher.

by permission of

by permis­

Robert van Gulick

Saul Kripke

"Understanding the Phenomenal Mind: Are We All

"The Identity Thesis"

Just Armadillos?", Part I

From

Naming and Necessity, Lecture III (Cambridge: 1980), 144-1 55, by permis­

Adapted from a version in

Consciousness, eds. M. 1993),

Harvard University Press,

Davies and G. Humphreys (Oxford: Blackwell,

sion of the author and publisher.

pp.

1 37-149, by permission of the author and publisher.

xxix

Sources

Frank Jackson

David Rosenthal

"What Mary Didn't Know" From Journal of Philosophy,

"A Theory of Consciousness"

83:5 (1986),

pp.

291 -295,

Adapted from ZiF Technical Report, Bielefeld, Ger­

1990,

by permission of the author and publisher.

many,

Paul M. Churchland

William G. Lycan

by permission of the author and publisher.

"Knowing Qualia: A Reply to Jackson"

"Consciousness as Internal Monitoring"

A Neurocomputational Perspective (Cambridge: MIT Press, 1989), 67-76, by permission of the author

Adapted from

From

Philosophical Perspectives,

Vol.

9,

ed.

J. Tomberlin (Atascadero: Ridgeview Publishing Co.,

and publisher.

1 990), 1 - 1 4, by permission of the author and publisher.

David Lewis

Fred Dretske

"What Experience Teaches"

"Conscious Experience"

Mind and Cognition, ed. W. Lycan (Oxford: Blackwell, 1990), 499-519, by permission of the author

From

the author and publisher.

and publisher.

Guven Giizeldere

From

Mind, 1 02:406 (1993), 263-283, by permission

of

Brian Loar

"Is Consciousness the Perception of What Passes in

"Phenomenal States"

One's Own Mind?"

Conscious Experience, ed. T. Metzinger 1 995), 335-357, by

Adapted from a version in Philosophical Perspectives,

From

ed. J. Tomberlin, Vol. 4 (Atascadero: Ridgeview Pub­

born: Schoeningh-Verlag,

lishing Co.,

1990), 81- 108,

by permission of the author

and publisher. Daniel C. Dennett "Quining Qualia" From

ConsciJJusness in Contemporary Science, eds.

A.

Marcel and E. Bisiach (Oxford: Oxford University Press,

1988), 43-77, by permission of the author and publisher. Sydney Shoemaker "The Inverted Spectrum" From

Journal of Philosophy, 74:7 (1981), 357�38 1,

by

permission of the author and publisher. Gilbert Harman "The Intrinsic Quality of Experience"

Philosophical Perspectives, ed. J. Tomberlin, Vol. 4 (Atascadero: Ridgeview Publishing Co., 1990), 3 1 -52, From

by permission of the author and publisher. Ned Block "Inverted Earth" From Philosophical Perspectives, ed. J. Tomberlin, Vol.

4 (Atascadero: Ridgeview Publishing Co., 1 990), 52-79, by pennission of the author and publisher. Stephen L. White "The Curse of the Qualia" From

Synthese, 68 (1986), 333-368,

by pennission of

the author and publisher. David Armstrong "What is Consciousness" From Press,

The Nature of Mind (Ithaca: 1 98 1), 55-67, by pennission

publisher.

Cornell University of the author and

sion of the author and publisher.

(Pader­ pennis­

INTRODUCTION

The Many Faces of Consciousness: A Field Guide

Giiveo Giizeldere There is perhaps no other phenomenon besides consciousness that is so familiar to each of us and yet has been so elusive to any systematic study, philosophical or scientific. In thinking about con­ sciousness, the puzzlement one often finds oneself in is rather like St. Augustine's riddle in his con­ templations about the nature of time: When no one asked him, he knew what it was; being asked, however, he no longer did. (Augustine of Hippo 1 96 1 : Book I I .) What is at the heart of this puzzlement? Is there a genuine difficulty that underlies it? What are the specific issues that comprise the problem of con­ sciousness? ( Is there really a "the problem of consciousness"?) And are we facing a phenom­ enon the understanding of which lies forever be­ yond our intellectual capacities? These are the questions that I will pursue below. The overarching goal of this introduction is to provide a field guide (with a particular perspec­ tive) for anyone interested in the history and present status of philosophical issues in the study of consciousness. Part One is a preliminary over­ view of the current philosophical positions in the literature, as well as a discussion of the unique difficulties inherent in the concept and nature of consciousness. Part Two is an account of the study of consciousness in the history of modem psychology. Finally, Part Three is an exposition of the mosaic of philosophical puzzles of con­ sciousness, as well as an exploration of their in­ terrelations. 1

PART ONE CONCEPTUAL FOUNDATIONS The feeling ofhelplessness was terrifying. I tried to let the staff know I was conscious but I couldn't move even a finger or eyelid. II was like being held in a vice and grad­ ually I realized that I was in a situation from which there was no way out. I began to feel that breathing was im­ possible, and Ijust resigned myself to dying. -Patient: Male, aged fifty-rour, bronchoscopy, 1978.

This testimonial was one of the many unsettling personal accounts of "becoming conscious" under general anesthesia, gathered in response to the following advertisement, which appeared in four national newspapers in Great Britain in 1 984: SURGERY: Have you ever been conscious dur­ ing a surgical operation when you were supposed to be anaesthesized? A medical research team would like an account of your experiences. Write in confidence. The goal of this advertisement was to gather firsthand accounts of gaining consciousness under general anesthesia, in order to investigate the truth of a number of patients' discomforting post­ surgery reports and to provide legal guidance for the accumulating court cases. 2 Whatever philosophical problems may be as­ sociated with the term consciousness, it might be thought that it would be a straightforward matter to specify an operational definition of being con­ scious for anesthesiologists to work with. Can consciousness simply not be detected on the basis of the patient's being alert and responsive? What is in question, after all, is neither the notoriously elusive problem of phenomenal experience nor the concept of the evanescent Humean self. A brief look at the anesthesiology literature, brimming with terms like real awareness and in­ cipient consciousness, quickly proves otherwise. (Cf. Rosen and Lunn 1 987.) If anything, the con­ sensus is that "with the spectral edge of the EEG [electroencephalogram] or median frequency, or any other processed EEG signal, there does not seem to be a clear cut-off, without overlap, between consciousness and unconsciousness" (Vickers 1 987, p. 1 8 2). The phenomenon of con­ sciousness does not have clear-cut boundaries, and its complex structure does not admit any easy formulations. (See, for instance, the Roche Handbook of Differential Diagnosis on "Coma" [1 979] and on "Transient Loss of Consciousness" [1 989].) Even if it is in principle possible to invent

2

Guven Giizeldere

a "consciousness monitor," a device that would detect the physical signs of the presence of

co�­

such technology IS sight, because it is not even known what exactly is to be measured. The root of the problem lies deeper than the inadequacy of the technology or the lack of suffi­ cient data, however. What seems to be critically lacking is a solid theoretical framework to ground and facilitate the experimental research. For ex­ ample, there is no established consensus, even in the medical field, as to what should count as the criteria of consciousness, so as to demarcate the domain of the conscious from that of the uncon­ scious or the nonconscious. The problem with building a consciousness monitor is not confined to a lack of sufficiently fine-grained measuring instruments; it ultimately has to do with not knowing where to begin measuring and where to . . end up with the measured quantities. 3 Worse, it is not clear whether everyone means the same thing by the term consciousness, even within the bounds of a single discipline. There is considerable variation in people's pre­ theoretic intuitions, for instance, regarding the kinds of creatures to which consciousness can be attributed.4 And in the absence of well-grounded theories, the lack of robust pretheoretical intuitions be­ comes even more importunate. Consider again the case of anesthesia. A person who is totally unresponsive to stimuli can, in one very impor­ tant sense, be said to have lost consciousness. Nonetheless, can she still be said to be conscious in another sense-in the sense of passively expe­ riencing the sensations caused by the stimuli, for instance? Similarly, are we to grant consciousness to a patient in a vegetative state, even when she lacks a well-functioning brain stem? Or what would justify granting consciousness to the patient if she did have a functioning brain stem that maintained the autonomic functions of her body?5 How many senses of consciousness are there anyway, and how are we to taxonomize them? sciousness in a patient, no

anywhere in

I

The Puzzle of ConsciolBless

These questions do not have any easy, obvious answers. Nor is there at present anything that could be regarded as a received view on problems of consciousness in the scientific and philosoph­ ical community. Furthermore, it is common to find serious doubts expressed in the literature about whether there can ever be a complete un­ derstanding of the phenomenon of consciousness. The gloomy opening lines of Thomas Nagel's famous essay, "What Is It Like to Be a Bat?" have become formative for many in thinking about consciousness: "Consciousness is what makes the mind-body problem really intract­ able. . . . Without consciousness the mind-body problem would be much less interesting. With consciousness, it seem s hopeless" (Nagel 1974, pp. 1 65-1 66). The puzzle of consciousness can be regarded in various ways, all the way from a supernatural mystery that will forever elude naturalist expla­ nations to a natural but extremely complicated pheno enon about which we know very little. And some of the time, the blue line that lies in between becomes very thin. There are also those who express skepticism about the existence of consciousness as a real phenomenon or about the coherence of its conceptual grounding, as well as others with a much more positive outlook, busily constructing their own accounts of consciousness to solve the puzzle. A brief look at some of the representatives of these different positions is in order.

�

The Mystery of ConsciolNless and the Explanatory Gap In the opening pages of Consciousness Explained, Daniel Dennett (1991) remarks: Human consciousness is just about the last surviving mystery . . . . There have been other great mysteries: t�e mystery of the origin of the universe, the mystery of hfe and reproduction, the mystery of the design to be found

Approaching Consciousness

in nature, the mysteries of time, space, and gravity. . . . We do not yet have the final answers to any of the questions of cosmology and particle physics, molecular genetics, and evolutionary theory, but we do know how to think about them. The mysteries haven't vanished, but they have been tamed. . . . With consciousness, however, we are still in a terrible muddle. Consciousness stands alone today as a topic that often leaves even the most sophisticated thinkers tongue-tied and confused. (pp. 21-22)

Dennett should not be taken as promoting the sense of mystery, however. After all, his book is entitled Consciousness Explained. Of course, it is hard to say that everyone (or even many) agrees with Dennett's conviction. In fact, the general sentiment among those who work on conscious­ ness (including philosophers, psychologists, and neuroscientists) seems to be on the "puzzled" side. Moreover, in the wake of a recent rise in inter­ est in the study of consciousness, almost each ap­ pearance of consciousness as a subject matter in the popular press has been tagged with some ele­ ment of mystery. For instance, Francis Crick and Christof Koch called consciousness the "most mysterious aspect of the mind-body problem" in their article that appeared in a special issue of Scientific American titled Mind and Brain (Sep­ tember 1 992). Discover magazine enlisted con­ sciousness as one of the "ten great unanswered questions of science" (November 1 992), and Omni published a special issue on consciousness but titled it "Science and the Soul" (October 1993). (Perhaps the rather unusual title Francis Crick chose for his book that appeared shortly afterward was a response to Omni's inquiry: The

Scientific Hypothesis: The Scientific Search/or the Soul.) Finally, Time magazine, in an issue that featured consciousness research, put the words "that evanescent thing called consciousness" on its cover (July 1 7, 1 995). Now, no one refers to other biological or psy­ chological phenomena in such terms. There is never a special magazine issue that pronounces the problem of cell mutation with the question of

3

the soul in the same breath, nor does anyone refer to language as "that evanescent thing." So it seems obvious that consciousness is perceived as special, possibly unique, and not readily ame­ nable to ordinary scientific or philosophical ex­ planation. Some take this sense of mystery even further, and this attitude is not at all restricted to the popular press. It is in fact possible to find the same sentiments expressed in philosophical and scientific circles, by those whom Owen Flanagan (199 1 ) calls the "New Mysterians." For instance, Colin McGinn ( 1 989) finds it humanly impossible ever to understand "how technicolor phenomen­ ology can arise from grey soggy matter," and ap­ provingly quotes the English biologist Huxley who famously stated: "How it is that anything so remarkable as a state of consciousness comes about as a result of irritating nervous tissue is just as unaccountable as the appearance of Djin when Aladdin rubbed his lamp" (p. 349).6 The expression of this sort of puzzlement is hardly new. Similar perplexity has been expressed by a number of people over the years, especially since the mid-nineteenth century, with the ad­ vancement of neurology and neuropsychology and the consequently well-grounded conviction that facts about consciousness must have some explanatory basis in the facts about the brain. For instance, in 1 874, physicist John Tyndall made the following remark: "We can trace the devel­ opment of a nervous system, and correlate with it the parallel phenomena of sensation and thought. We see with undoubting certainty that they go hand in hand. But we try to soar in a vacuum the moment we seek to comprehend the connection between them. An Archimedean fulcrum is here required which the human mind cannot com­ mand; and the effort to solve the problem . . . is like that of a man trying to lift himself by his own waistband" (p. 1 95). 7 In contemporary literature, Karl Popper, in a similar vein, finds "the emergence of full consciousness . . . which seems to be linked to the human brain . . . one of the greatest

Guven Giizeldere

miracles" (Popper and Eccles 1 993, p. 1 29). And most recently, McGinn ( 1 989) delivers what he considers to be the final verdict on the mind-body problem: "We have been trying for a long time to solve the mind-body problem. It has stubbornly resisted our best efforts. The mystery persists. I think the time has come to admit candidly that we cannot solve the mystery" (p. 349). There is, however, more than one way to read assertions about the mystery of consciousness. Accordingly, it is important not to lump together everyone who expresses puzzlement about con­ sciousness into the same category. In particular, it is important to pay attention to the following two questions: Is the mystery essentially a result of a commitment to a materialist framework? Is the mystery essentially inherent in our (lack of) cog­ nitive capacities? As such, these questions constitute an onto­ logical and an epistemic axis, respectively, that cross-cut each other. Not every combination in the matrix receives equal philosophical attention. A negative answer to both questions essentially leaves one out of the circle of those who find something mysterious in consciousness, and not many defend a view that find consciousness mys­ terious in both aspects. Rather, the focus is on views based on an exclusively positive answer to one or the other question. Those who think that consciousness will re­ main a mystery in a materialist ontology suggest that the proper place to pursue investigation is instead an immaterial realm-such as the realm of the res cogitans for Descartes, or World 2 of mental entities for Karl Popper. This move brings with it a problem perhaps larger than that it was presumed to solve: how to account for the link between consciousness in the immaterial realm and brains (and bodies) in the material realm. Descartes's notorious solution was to postulate the pineal gland as the gatekeeper of interac­ tion between the two essentially different kinds of substances. Alternatively, Leibniz chose to rely on divine intervention to secure a "pre-

4

established harmony" between the events of the two realms. More than three centuries after Descartes, John Eccles ( 1 99 1 , pp. 1 90- 191) makes a repeat attempt, although in neurologically sophisticated dress, by postulating psychons (mental units) as counterparts of dendrites in brains, to connect Popper's mental World 2 to physical World 1 . Another contemporary expression of the view that defends an antimaterialist framework for consciousness is given by Robert Adams (1 987), who finds theism theoretically advantageous to materialism in explaining the relation of con­ sciousness to bodily physical states. 8 On the other hand are those who do not cut the bill of the mystery of consciousness to the pre­ sumed immaterial ontology of consciousness but rather to our lack of cognitive capacities that would enable us to understand the nature of the "psychophysical link" between brains and minds. Tyndall seems to be in this group, and so is McGinn, who states that although "we know that brains are the de facto causal basis of consciousness," we have no idea about how "the water of the physical brain is turned into the wine of consciousness" (McGinn 1989, p. 349). This is an epistemic rather than ontological problem. For McGinn, "there is, in reality, nothing mys­ terious about how the brain generates con­ sciousness," but we human beings are forever "cognitively closed" to understanding the nature of this process, much the same way the under­ standing of quantum mechanics lies beyond the cognitive capacities of monkeys. 9 The general difficulty involving consciousness forms a basis for what Joseph Levine (1983, 1 993) called the "problem of the explanatory gap." Al­ most everyone agrees that there is indeed some explanatory gap in this area; what is controversial is, as I will note later, whether there is just an epistemic or also an ontological lesson that needs to be drawn from it. In any case, these positions constitute only a fraction of the whole dialectical space. There are

5

Approaching Consciousness

also the skeptics and the naturalists, and to them I now tum.

SkepticWn About Conscio�ess The skeptics among philosophers fundamentally doubt the coherence of the very concept of con­ sciousness, and the merits of consciousness itself as a phenomenon fit for scientific or philosophical investigation. Patricia Churchland, in one of her early papers, compares the concept of con­ sciousness (under a certain reading that she ex­ plicates) to such now-defunct concepts as ether, phlogiston, and demonic possession-concepts that "under the suasion of a variety of empirical­ cum-theoretical forces . . . lose their integrity and fall apart" (Churchland 1983, p. 80). In a similar vein, Kathleen Wilkes claims not only that "science can dispense with the concept of consciousness and lose thereby none of its comprehensiveness and explanatory power," but "so too could ordinary language." She then sug­ gests that "perhaps 'conscious' is best seen as a sort of dummy-term like 'thing', useful [only) for the flexibility that is assured by its lack of specific content" (Wilkes 1 984, pp. 241 -242). Along the same lines Georges Rey goes a step further and suggests that there are "reasons for doubting that oneself is conscious and . . . thinking that nothing is conscious." Consciousness, Rey suggests, "may be no more real than the simple soul exorcised by Hume" (Rey 1 988, p. 6). Notice that this sort of skepticism about con­ sciousness is a very different attitude from any form of "mysterianism." In particular, those who think that consciousness is mysterious are com­ mitted to the existence of some significant phe­ nomenon, however elusive it may be in relation to scientific investigation or philosophical analysis. Consciousness skepticism, on the other hand, embraces an eliminativist stance: the concept of consciousness is defunct, and the phenomenon it­ self may actually be inexistent, at least so far as it is construed in the literature that the skeptics are attacking.

The Consciousness Naturalists Finally, there are those who believe that con­ sciousness is a real and perfectly natural phenom­ enon and that there will remain no mysterious unexplained residue about consciousness in a completely naturalist, but surely more advanced and mature, theoretical framework. Among the naturalists, however, there is a wide spectrum of positions representing different levels of con­ fidence in the success of a naturalist program. For instance, one can straightforwardly distinguish between full-blown naturalists and naturalists-at­ heart. 1 0 Into the first group fall a number of philos­ ophers who have explicitly defended a naturalist framework to explain consciousness, without theoretical reservations. However, some of the full naturalists, most significant among them Paul Churchland (1 988) and Daniel Dennett ( 1 99 1 ), have been charged with trying to do away with consciousness for the sake of explaining it. Some others have been more careful not to fall under this decree. For instance Owen Flanagan (1 992), who proposes what he calls the natural method­ a triangulated approach for studying conscious­ ness that combines phenomenology, psychology, and neuroscience-states: "Consciousness exists, and it would be a mistake to eliminate talk of it because it names such a multiplicity of things. The right attitude is to deliver the concept from its ghostly past and provide it with a credible naturalistic analysis. . . . It will be our proudest achievement if we can demystify consciousness" (Flanagan 1 995, p. 20). Among the naturalists who give accounts of consciousness in terms of causal and functional roles (broadly construed) are Armstrong (1 980b, 1 993), Lewis (1 966, 1 972, 1 980, 1 995), Shoe­ maker (1975, 1 99 1 , 1 994), Lycan ( 1 987, 1 997b), Van Gulick (1 988, 1 989, 1 993), and Rosenthal ( 1 986, 1 997). Most recently, Fred Dretske ( 1 995) and Mi­ chael Tye ( 1 995) came up with naturalist ac­ counts that explain consciousness in entirely

Guven Giizeldere

6

representational terms. Finally, one can add John

I am persuaded of its truth" (Nagel

Searle to this group, who dubs his view "bio­

Later he moves to a position he calls "dual aspect

logical naturalism":

1 965, p. 1 10).

theory" -a position that lies between asserting

The "mystery" of consciousness today is in roughly the same shape that the mystery of life was before the de­ velopment of molecular biology or the mystery of elec­ tromagnetism was before Clerk-Maxwell equations. It seems mysterious because we do not know how the sys­ tem of neurophysiology/consciousness works, and an

falsity of physicalism but remaining short of pos­ tulating nonphysical substances for accounting for the ontology of the mind-while admitting that "to talk about a dual aspect theory is largely hand waving" (Nagel

1 986, pp. 29-30). So per­

haps the early Nagel was a naturalist-at-heart,

adequate knowledge of how it works would remove the

and now it is more accurate to characterize him as

mystery . . . . [11here has been no question of "natural­

a half-hearted-naturalist. In either case, Nagel's

izing consciousness"; it is already completely natural.

position seems, at least fundamentally, some­

(Searle 1 992, p. 1 02, 93, respectively)

where in the naturalist camp, despite the skeptical

Despite Searle's fully naturalist convictions, his views have, as I will point out later, significant disagreements with those of the above on many other (relevant) points. Nonetheless, his position also differs from a position that I call naturalism­ at-heart. Naturalists-at-heart are those who openly feel the pull of naturalism, while not quite being able to find a satisfactory place for consciousness in a naturalist framework. For example, Levine

( 1983, 1 993), who thinks that the problem of ex­ planatory gap poses ultimately no ontological problems for materialism, nonetheless expresses the following troubled sentiments: "The absent

and pessimistic undertones that make him some­ times look closer to the mysterians or skeptics. To recapitulate: It is noteworthy that the spec­ trum of disagreements ranges over not only particular accounts of consciousness but, more fundamentally, whether any satisfactory natural­ istic explanation of consciousness can

in principle

be given. Part of this disagreement owes, no doubt, to the difficulty in the

nature

of the phe­

nomenon of consciousness. But there is also a part that stems from a conceptual disarray surrounding the

notion

of consciousness. It is thus instructive

to examine these two dimensions that contribute to the puzzle of consciousness separately.

and inverted qualia hypotheses are thought ex­ periments which give concrete expression to what I will call, following the Churchlands, the 'pro­ qualia' intuition. This is the intuition that there is

n

Approaching Consciousness: A Multitude of Difficulties

something special about conscious mental life

What we

that makes it inexplicable within the theoretical

what we are when we sink into a profound and perfectly

are

when we

are

awake, as contrasted with

framework of functionalism, and, more generally,

dreamless sleep or receive an overpowering blow upon

materialism" (Levine

1 988, p. 272). Short of dis­

the head-that it is to be conscious. What we are less

believing materialism, Levine finds it difficult to

and less, as we sink gradually down into dreamless

place consciousness in a naturalist framework.

sleep, or as we swoon slowly away: and what we are

Perhaps the most eloquent proponent of this position is Thomas Nagel, possibly more so in his earlier writings than his later work. In one of his earlier works, Nagel examines and rejects "the reasons for believing that physicalism cannot

more and more, as the noise of the crowd outside tardily arouses us from our after-dinner nap, or as we come out of the midnight darkness of the typhoid fever crisis­

that it is to become conscious. This is how George Trumbull Ladd

( 1909),

possibly be true" and concludes: "My attitude

noted American psychologist (whose definition of

toward physicalism [is that it] repels me although

psychology as the "description and explanation

Approaching Consciousness

of states of consciousness as such" was adopted and promoted by William James), characterized the phenomenon of consciousness (p. 30). This characterization seems straightforward, common­ sensical, and familiar to everyone. So familiar that perhaps, as George Stout (1 899), another psychologist of the same era, declares in the opening pages of his Manual of Psychology, no precise definition is necessary, or even possible: "What is consciousness? Properly speaking, definition is impossible. Everybody knows what consciousness is because everybody is conscious" (p. 7). Similarly, William James never attempts to give a definition of consciousness anywhere in his two-volume work, Principles of Psychology (1950a, 1 950b). This is not because James had no interest in, or nothing to say on, consciousness; on the contrary, many of the chapters in his two volumes are about consciousness-its under­ pinnings in the nervous system, its function in evolution, its streamIike phenomenology, and so on. Rather, according to James, consciousness was a phenomenon too familiar to be given a def­ inition. James (1950a) was convinced that every­ one took themselves to be possessors of conscious states that were accessible by introspection, and he regarded this belief as "the most fundamental of all the postulates of Psychology." In his refusal to discuss this postulate any further, James adds that he would "discard all curious inquiries about its certainty as too metaphysical" for the scope of his book (p. 185). This somewhat peculiar "all-too-familiar a phenomenon" attitude toward consciousness has indeed been quite common among many other prominent investigators of consciousness. Sig­ mund Freud (1 964), for instance, supports Stout's and James's convictions in his introductory lec­ tures on psychoanalysis: "What is meant by con­ sciousness we need not discuss; it is beyond all doubt" (p. 70). Closer to our times, neuroscientist Francis Crick and Christof Koch ( 1990) endorse the same line in the opening paragraphs of their article, "Towards a Neurobiological Theory of Consciousness," even though they are attempting

7

to lay out the foundations of a new theory of consciousness: they need not provide a precise definition of consciousness since "everyone has a rough idea of what is meant by consciousness" (p. 263). On the other hand, it is not uncommon to come across statements about consciousness that con­ vey a conviction opposite to those mentioned above: that not only is there no clear and gen­ erally accepted definition, but we are not even in possession of a stable pre theoretical conception of consciousness. And this view, too, has been around for quite a while, as expressed by Edward Titchener (1915) who cites two British psycholo­ gists, Alexander Bain and James Ward, ofthe late nineteenth century: " 'Consciousness,' says Pro­ fessor Ward, 'is the vaguest, most protean, and most treacherous of psychological terms'; and Bain, writing in 1 880; distinguished no less than thirteen meanings of the word; he could find more today" (pp. 323-324). Unfortunately, there is RO hope of receiving help from antonyms, either. Here is the entry for "unconscious" in a psychology dictionary: "It is said that there are no less than 39 distinct mean­ ings of 'unconscious'; it is certain that no author limits himself consistently to one. And nearly all meanings are closely linked to debatable theories. Any user of the term therefore risks suggesting agreement with theories he may deplore" (En­ glish and English 1958). Finally, Julian Jaynes (1 976) rhetorically asks: "This consciousness that is myself of selves, that is everything, and yet nothing at all-what is it?" (p. I). Moreover, not only is there no consensus on what the term consciousness denotes, but neither is it immediately clear if there actually is a single, well-defined "the problem of consciousness" within disciplinary (let alone across disciplinary) boundaries. Perhaps the trouble lies not so much in the ill definition of the question, but in the fact that what passes under the term consciousness as an all too familiar, single, unified notion may be a tangled amalgam of several different concepts, each inflicted with its own separate problems.

Guven Giizeldere

What exactly, for example, is the problem of consciousness in philosophy, in psychology, and in the neurosciences? Are philosophers concerned with the same problem, or set of problems, as psychologists and neuroscientists who work on consciousness? Whereas Thomas Natsoulas ( 1992), a psychologist, questions: "Is conscious­ ness what psychologists actually examine?" (p. 363), Kathleen Wilkes (1988), a philosopher, gives the following advice: "Just as psychologists do not study 'mind' per se, so they need not bother with consciousness (because] in all the contexts in which it tends to be deployed, the tenn 'conscious' and its cognates are, for scientific purposes, both unhelpful and unnecessary" (pp. 38-9). An even more pessimistic view is enunciated by Stuart Sutherland (1 995) under the entry "con­ sciousness" in his International Dictionary of

Psychology : CONSCIOUSNESS: The having of perceptions, thoughts, and feelings; awareness. The term is impossible to define except in terms that are unintelligible without a grasp of what consciousness means . . . . Consciousness is a fasci­ nating but elusive phenomenon: it is impossible to specify what it is, what it does, or why it evolved. Nothing worth reading has been written about it. (p. 95)

Why are there such glaring polarities? Why is consciousness characterized both as a phenom­ enon too familiar to require further explanation, as well as a source of obscurity that remains typi­ cally recalcitrant to systematic investigation by those who work largely within the same para­ digm? There is something uniquely peculiar here. Is it the phenomenon of consciousness that is more puzzling, one sometimes wonders, or the magnitude of the puzzlement itself and the theo­ retical dissonance surrounding consciousness? Could it perhaps be as R. J. Joynt (1981) pre­ dicted: "Consciousness is like the Trinity; if it is explained so that you understand it, it hasn't been explained correctly" (p. 108)? George Miller (1962), faced with these dif­ ficulties, tentatively entertains the interesting

8

proposition that "maybe we should ban the word for a decade or two until we can develop more precise terms for the several uses which 'con­ sciousness' now obscures." Nonetheless, he ulti­ mately decides against it: "Despite all its faults, however, the term would be sorely missed; it refers to something immediately obvious and familiar to anyone capable of understanding a ban against it" { p. 25). 1 1 In the end, I find myself in agreement with Miller's positive conclusion. It is historically ac­ curate to note that consciousness as a phenom­ enon in need of not only explanation, but also definition, has persistently kept resurfacing. It also seems reasonable to think that further attempts to provide carefully constructed con­ ceptual tools could only help the situation, pro­ viding a common platform of interaction among all who choose consciousness as their object of study. This is not to say that it is not crucial to proceed cautiously to steer clear of conceptual dead ends, as well as to make sure that one does not fall into the trap of reinventing the wheel of consciousness over and over. After all, if we hope that anything toward a better understanding of consciousness will come out of the joint efforts of different disciplines, it is of utmost importance to minimize crosstalk and make sure that common terms actually point to the same referents. As a result, it seems even more imperative to look for and try to delineate the specific conceptions and aspects of consciousness under which the different problems arise. The next section is a brief attempt to address conceptual issues along these lines, before turning to the epistemological and onto­ logical difficulties that arise from the nature of consciousness.

Difficulties with the Concept of Consciousness Because consciousness is a word whose semantics has shifted over time, a brief lexical and etymo­ logical exposition may be of some service as a preliminary step. Let us start with present definitions. The Oxford English Dictionary {OED:

9

Approaching Consciousness

unabridged second edition), for example, gives

transitive

eight definitions of "consciousness," and twelve

sciousness. He calls these two senses "creature

definitions of "conscious." For the purposes of

consciousness" and "state consciousness," re­

OED

and

intransitive

senses

of

con­

entries for "con­

spectively. In taking consciousness as a property

sciousness" can be divided roughly into two

of organisms one can talk about a person's being

this chapter, the eight

groups. On the one hand, there is a largely social

consciousness: joint

conscious in the sense of being awake and alert, as

or mutual

opposed to being in a transient state of no con­

knowledge shared by a community of people.

sciousness, in a deep coma, or in non-REM sleep.

aspect of the term

This was indeed the earliest sense of conscious­

In another vein, one can ask of a bat, or a spider,

ness, derived from the Latin term

or a stickleback, or perhaps a robot if it is

conscius. 1 2

It is

this sense of consciousness that is used in talking

conscious. All of this has to do with

about "class consciousness" in Marxist thought,

consciousness.

creature

In addition, it makes sense to talk

or that appears in titles like Gerda Lerner's recent

about whether a particular mental state is con­

The Creation of Feminist Consciousness

scious. This is not quite the same as someone's

book,

being conscious. The creature sense of conscious­

(1993). On the other hand, consciousness has a largely

ness denotes an overall state one is in; the other

psychological (mental) sense which relates to in­

classifies one's (mental) states as of one type or

dividuals, rather than groups, with no particu­ larly ethical or political overtones. 1 3 This sense of

another. Further, a state that is not conscious can be among those that in principle cannot become

consciousness, too, can be subdivided into two

conscious

meanings-it means either "the state or faculty of

postulated by cognitive psychology), or those that

(e.g.,

certain

computational

states

being conscious, as a condition and concomitant

can be made conscious only by such specific

of all thought, feeling, and volition" or "the state

methods as Freudian psychoanalysis. In any case,

of being conscious, regarded as the normal con­

this sense of consciousness which functions as a

(OED). The former

type-identifier for mental states is what I have in

dition of healthy waking life"

state consciousness. I S

sense is in accord with Descartes's usage of

mind by

"consciousness," but it is more closely associated

tant distinction is that of characterizing creature

An

or state consciousness in causal versus phenom­

with, if it does not originate from, Locke's

Essay Concerning Human Understanding.

(Another impor­

It is

enal terms. This distinction will be introduced in

also this sense of consciousness that bears an in­

the next section, and used recurrently throughout

tentional component, inherent in the intentional

the chapter.)

transitive usage: is the intransitive

aside the social conception of consciousness and

usage of "consciousness," something more basic

proceed with the psychological conception. This

states it subserves, leading to the "consciousness of." The latter

In the rest of this discussion, I will largely put

than, or perhaps a necessary constituent of the

move approximately halves the size of the liter­

former, unless one wants to insist that aU con­

ature to be examined; yet there is still plenty that

sciousness is consciousness of. The distinction

needs teasing apart and sorting out.

16

between the transitive and the intransitive senses of the psychological concept of consciousness is

Difficulties witb tbe Nature of Consciousness

alluded to numerous times in contemporary phi­ losophy and psychology, and I will accept it here as such. 14

Among the ophers, David

It is often remarked that conscious experiences are "immediately familiar" to any subject of such

philos­

experiences. It is also argued that they are "so

( 1 997) has written

immediately close" that it is at times difficult, if

contemporary Rosenthal

analytic

substantially about the distinction between the

not impossible, to separate their

appearance from

Guven Giizeldere

their reality. This ambiguity between the appear­ ance versus the reality of consciousness is also re­ garded as unexampled; presumably nothing else in the world suffers from it. As such, it tends to uniquely blur the line between the epistemology and ontology of consciousness: if all there is to (the reality of ) conscious states is their appearing in a certain way to subjects, and if they have no existential status independent of their so appear­ ing, the ontology of consciousness seems to col­ lapse into its epistemology. 1 7 Ironically, o n the other hand, the problem ap­ pears to be just the opposite from a different angle: the appearance of consciousness seems so different from its "physical reality" that a comprehensive theory that bridges this gap is regarded as a near impossibility. In Sellarsian terms, the scientific and manifest images of con­ sciousness are considered to embody a theoretical gap perhaps greater than in any other subject matter. (Cf. Sellars 1 99 1 .) The view that consciousness (or, in general, the mind) and its physical basis (or, in general, the body) seem essentially so different from one an­ other that they must have distinct existences is based on a deep-rooted idea in the history of philosophy. This idea and its variants were" con­ stitutive of arguments for the metaphysical in­ dependence of mind and body throughout early modem philosophy of the seventeenth and eigh­ teenth centuries, perhaps most notably exempli­ fied in the work of Descartes. The essential and complete nature of mind, generally speaking, seems to consist solely in thinking, and, as such, it must be unextended, simple (with no parts), and essentially different from the body, and therefore immaterial. This was Descartes's argument in a nutshell, ultimately drawing a strong ontological conclusion (regarding the distinctness of mind and body) from a starting point constituted by epistemic considerations (regarding the distinct­ ness of their appearances). As Ben Mijuskovic (1974) observes, in this type of argumentation, "the sword that severs the Gordian knot is the principle that what is conceptually distinct is on-

10

to logically separable and therefore independent" ( p. 1 23). Mijuskovic, in locating this form of reasoning in its historical context, also notes the presence of the converse of its inference: "If one begins with the notion, explicit or implicit, that thoughts or minds are simple, unextended, indivisible, then it seems to be an inevitable step before thinkers connect the principle of an unextended, immate­ rial soul with the impossibility of any knowledge of an extended, material, external world" and consequently, of the nature of the relation be­ tween them (p. 121). That is, this time an epis­ temological conclusion (regarding an epistemic gap between mind and body) is reached from a starting point constituted by ontological con­ siderations (regarding the distinctness of their natures). The difficulties inherent in the nature of con­ sciousness constitute many of the philosophical problems that will be discussed in depth in the rest of this chapter. In particular, the nature and validity of inferences between matters epistemo­ logical and matters ontological, especially those that go from the former to the latter, will con­ tinually appear as a leitmotif. Of course, some of these difficulties may be overcome quite rapidly as the study of consciousness advances; others may prove more obstinate. But it may also tum out that certain problems that seemed unsolvable had appeared that way because of the specific ways in which they were formulated, or the im­ plicit assumptions they rested on. This possibility, too, will emerge as a relevant concern later in the discussion. The best way to gain insight into this sort of a difficulty may very well be through locating the analysis in a broad historical perspective. The long history of consciousneS$ research no doubt contains-hints that can be parlayed not only to draw methodological lessons for further study, but also to reveal the constituent fibers of the past paradigms which couched persistent core prob­ lems under different guises over and over. It is in fact the most striking feature of the consciousness

Approaching Consciousness

literature, as I try to exemplify numerous times in this exposition, that the very same problems, analyses, and suggested solutions repeatedly ap­ pear, and the very same theoretical moves re­ peatedly get introduced at different times in the history of philosophy and psychology with little (if any) acknowledgement of past attempts and failures. In this regard, a historical approach that exposes the misleading implicit assumptions common to the past failures should prove useful in illustrating, at minimum, which steps not to take in approaching consciousness at present. m

Looking Abead: The Two Faces of Coosciousoess Before proceeding further, I will offer a brief first pass at a diagnosis that will be made at the end of this chapter: a principal reason underlying the confusion and seeming mystery surrounding the concept and phenomenon of consciousness lies in the presence of two influential, equally attractive, pretheoretic characterizations. These two charac­ terizations not only shape the methods with which consciousness is studied, but more fun­ damentally, shape the way the problems to be studied are defined and delineated. They can be summarized in the following mottos: "Con­ sciousness is as consciousness does" versus "Consciousness is as consciousness seems. " The former is the causal characterization: it takes the causal role consciousness plays in the general economy of our mental lives as basic. The latter, in contrast, is the phenomenal characterization: it takes as fundamental the way our mental livel> seem (or "feel," for lack of a better term) to us­ that is, the phenomenal qualities that we typically associate with our perceptions, pains, tickles, and other mental states. Most of the time, these two characterizations are taken to be mutually exclusive for explana­ tory purposes, to the extent that accounts of con­ sciousness built around one characterization are typically accused of failing to capture the other. I believe that this undesirable consequence, which

11

often seems to deadlock debates on conscious­ ness, stems from a fundamental and ultimately misleading intuition that I will call the segrega­ tionist intuition: if the characterization of con­ sciousness is causal, then it has to be essentially nonphenomenal, and if it is phenomenal, then it is essentially noncausa1. (I call this formulation an "intuition" rather than a "thesis" due to its widely diffused, often implicit and unarticulated, but highly influential nature.) In contrast with the segregationist intuition is what I call the integrationist intuition: what con­ sciousness does, qua consciousness, cannot be characterized in the absence of how consciousness seems, but more importantly, that how con­ sciousness seems cannot be conceptualized in the absence of what consciousness does. This counter intuition underwrites the project of trying to dissolve the stalemate between accounts of consciousness respectively based on the causal and the phenomenal characterizations, and marry them into a single unified account. This introduction is not an attempt towards accomplishing such a project. Nevertheless, while presenting a conceptual mapping of the territory and locating in it contemporary problems and debates that center around consciousness, I hope to provide support for the integrationist intuition that motivates it. In doing so, I also aim to sub­ stantiate an antiskeptical position with respect to consciousness: there is a a deep-rooted and con­ tinuous theoretical thread connecting a set of re­ current problems in the history of philosophy and psychology typically associated with conscious­ ness, indicating the presence of a persistent, sig­ nificant, and challenging object of study.

PART lWO A BRIEF IDSTORY OF CONSCIOUSNESS: PHILOSOPHY AND PSYCHOLOGY There is possibly no other subject matter in the history of philosophy and science with as fasci­ nating a historical record as consciousness. Even

Guven Giizeldere

within the past one hundred years, consciousness has more than once been crowned as the most significant aspect of human mentality, to be fol­ lowed by periods of scapegoat treatment for the failures of philosophy and science (in particular, psychology) to give a satisfactory account of the mind. In either case, consciousness was hardly ever ignored. Explicitly or implicitly, it was an ever-present concern for everyone thinking about the human mind. Following is a brief journey through the his­ torical path that consciousness research has trav­ ersed in approximately the last hundred years, in particular via the schools of introspectionism, behaviorism, and cognitivism in psychology, with early modem philosophy taken as a starting point.

IV

Consciousness in Early Modem Philosophy

In accord with the fact that the origins of the word consciousness go back to early modem phi­ losophy, it is generally agreed that Descartes gave the mind-body problem its modem formulation. Descartes's own account respectively charac­ terized mind and body as thinking versus ex­ tended substances, and postulated that the nature of their relation was that of interaction between the res cogitans and the res extensa. But how did his notion of consciousness compare with his no­ tion of mind? And to what extent does his notion of consciousness capture the notion that presently figures in contemporary debates? 1 8 Descartes claimed that consciousness was an essential component of everything that was men­ tal, and by "consciousness" he meant something akin to one's awareness of one's own mental states: "As to the fact that there can be nothing in the mind, in so far as it is a thinking thing, of which it is not aware, this seems to me to be self­ evident. For there is nothing that we can under­ stand to be in the mind, regarded in this way, that is not a thought or dependent on a thought" ( Descartes 1 993b, p. 1 7 1 , fourth set of replies to

12

Arnauld). By "thought" Descartes must have had in mind something very similar to one of the contemporary usages of "consciousness," or "awareness," especially given his definition in the Principles ofPhilosophy: "By the term 'thought' I understand everything which we are aware of as happening within us, in so far as we have aware­ ness of it" ( Descartes 1992, p. 1 74). Locke, coming after Descartes's rationalism from a distinctively empiricist tradition, was nonetheless largely in agreement with his prede­ cessor with respect to the nature of the relation between what was mental and what was con­ scious: they were conceptually tied. In Locke's words, "thinking consists in being conscious that one thinks," and "the idea of thinking in the ab­ sence of consciousness is as unintelligible as the idea of a body which is extended without having parts" ( Locke 1 959, bk. 2, chap. 1 , p. 1 38). There is also another sense in which Descartes and Locke seem to be in agreement: the idea of construing consciousness in roughly some­ thing like higher-order awareness. For Descartes, proper sensations in adults exist only insofar as they are accompanied by a second-order reflective awareness: "When an adult feels something, and simultaneously perceives that he has not felt it before, I call this second perception reflection, and attribute it to the intellect alone, in spite of its being so linked to sensation that the two occur together and appear to be indistinguishable from each other" ( Descartes 1 99 1 , p. 357: letter to Ar­ nauld, 29 July 1 648, AT V, 221 ). Along this the­ oretical line, Descartes concludes, for instance, that "pain exists only in the understanding" ( Descartes 1 99 1 , p. 148: letter to Mersenne, 1 1 June 1 640, AT III, 85). In a somewhat similar vein, Locke famously stated that "consciousness is the perception of what passes in a man's own mind" (Locke 1 959, bk. 2, chap. 1 , §19, p. 1 38). However, it may be unfair to read too much into the "higher-order awareness" construal and make Descartes's and Locke's views seem more similar than they ac­ tually are. For instance, it is not altogether ob-

13

Approaching Consciousness

vious that Locke's second-order "perception" is as cognitively loaded as Descartes's "reflective perception," although they seem to serve the same purpose in being responsible for conscious­ ness of "first-order" mental goings on. 1 9 There are also contemporaries of Descartes and Locke, who located the epistemic locus of mind in qual­ itative conscious states rather than thoughts or reflective perceptions. Most notably, Male­ branche holds this view. 2 0 This brief characterization of the early modem philosophical thought on consciousness no doubt fails to do justice to the subtleties involved. But for the sake of finding a starting point common to both philosophy and "scientific psychology" in the study of consciousness, and tracing the issues in double-track to the contemporary debates, I will leap ahead to the late nineteenth century and, skipping over the problem of unity of conscious­ ness and Kant's treatment of "unity of appercep­ tion," continue with the work of William James.

V The Last Hundred Years: William James's Puzzle William James may be the philosopher and psy­ chologist who thought and wrote more about consciousness than anyone else in history. Inter­ estingly enough. the record of his stance(s) to­ ward consciousness is also the most curious one. James allots a great deal of space to discussing the neural underpinnings, the evolutionary function, and the phenomenal nature of consciousness in his monumental work, Principles of Psychology. According to James of this book, consciousness is the starting place of all psychology, the most cru­ cial aspect of human mentality. In a chapter on the methodology of psychology, he states: "In­

trospective Observation is what we have to rely on first and foremost and always. The word in­ trospection need hardly be defined-it means, of course, the looking into our own minds and re­ porting what we there discover. Every one agrees

that we there discover states of consciousness"

(James 1 950a, p. 1 85; originally published in 1 890). But only fourteen years later, James would bit­ terly denounce consciousness in an article titled, "Does Consciousness Exist?" with the following verdict: For twenty years past I have mistrusted "conscious­ ness" as an entity; for seven or eight years past I have suggested its non-existence to my students, and tried to give them its pragmatic equivalent in realities of experi­ ence. It seems to me that the hour is ripe for it to be openly and universally discarded . . . . [Consciousness] is the name of a non-entity, and has no right place among first principles. Those who still cling to it are clinging to a mere echo, the faint rumor left behind by the disappearing "soul," upon the air of philosophy. (James 1971, p. 4; originally published in 1 904)

The reasons for this remarkable change of mind may partly lie deep in James's personal his­ tory, but they also have to do with the unique place of consciousness as a subject matter in phi­ losophy and psychology. 2 1 A somewhat similar, almost neurotic shift of attitude, though in a much larger scale, spanning the whole discipline of psychology and, to some extent, philosophy, occurred in a rel atively short period of transition, early in this century. This transition involved the collapse of the then very established school of introspectionism and the subsequent rise of behaviorism. 22

VI

Introspectionism

Introspectionism can be regarded as the first off­ spring of the effort of pulling psychology apart from philosophy and establishing it as an in­ dependent, "scientific" discipline on its own. Ironically, behaviorism would later denounce in­ trospectionism as having tangled with meta­ physics and present itself as the true, alternative scientific school of psychology. In actuality, in their struggle for identity. both schools borrowed a great deal from the scientific methodology of their times, and neither one's approach was in­ trinsically more "scientific" than the other.

Guven Giizeldere

Introspectionism's fundamental assumption was that psychology was the study of the "phe­ nomenology" of the human mind; it attempted to give a full description of the mental landscape as it appeared to the subject. The data points con­ sisted of discriminations in subjects' sensations of colors, sounds, sinens, and the like. In doing so, introspectionism largely modeled its method­ ology after the modern chemistry of the day, which was enjoying a high reputation due to its successes in having put together the atomic table. The fundamental belief underlying most of in­ trospectionist research was that a full under­ standing of the mind was possible only after completing an exhaustive inventory of its "atomic units," most elemental sensory impressions one can discriminate. Introspectionism, in other words, was in the business of constructing an atomic table of the human mind.23 (Cf. Killpe 1 90 1 ; Titchener 191 5.) Giving a full inventory of anything is no easy feat, and attempting this for the totality of the human "sensory space" was a daunting task, even in the hands of scrupulous researchers and metic­ ulously trained subjects.24 However, the fall of introspectionism did not result from depleted patience or the lack of a sufficient number of experiments. The failure had deeper reasons, both external and internal. The external reason for failure was the overall changing intellectual climate in Europe and the United States, especially the rising influence of positivism in all sectors of science, as well as the humanities. The general positivist attitude constituted a significant motivation for psycho­ logists, who had been trying hard to sever their professional ties with philosophers and to move away from anything "mental" in an attempt to relocate psychology among natural sciences. Consciousness was the subject matter of no natural science, so it could not be the subject matter of psychology either. 25 The positivist atmosphere further provided a context in which it was easier for behaviorists to

14

make introspectionism appear as a scientifically baseless enterprise, further burdened with the metaphysically dubious cargo of consciousness­ so much so that John Watson, in the opening pages of his book that served as the behavior­ ist manifesto, belittled introspectionists' concern with consciousness by likening it to witchcraft: "Behaviorism claims that consciousness is neither a definite nor a usable concept. The behaviorist, who has been trained always as an experimenta­ list, holds, further, that belief in the existence of consciousness goes back to the ancient days of superstition and magic" (Watson 1970, p. 2). In reality, introspectionism was as much an at­ tempt to bring psychology up to par with natural sciences-to make it a "science of the mental," with ideas and methods inspired largely by chem­ istry. Although the founding adherents of be­ haviorism wholeheartedly denied any intellectual debt to their predecessors and did a good job of making themselves appear completely detached from their past, they were very similar to the in­ trospectionists in aspiration and professional policy with regard to methodology. 2 6 In any case, it would not be fair to place on solely external causes all responsibility for in­ trospectionism's formidable downfall, which brought with it the downfall and disgrace of both consciousness (as a subject matter for research) and introspection (as a method for studying the mind) for several decades. There were serious reasons internal to the paradigm as well. Most important was the apparently irreconcilable conflict between results coming out of different laboratories. The most significant polarity was constituted by two main streams of research pur­ sued in two different continents: the W1lrzburg school, represented by Killpe and his students in Leipzig, versus the Cornell school, lead by Titch­ ener and his associates in Ithaca, New York. For instance, Titchener's laboratory reported that they discovered a total of "more than 44,435" discriminably different sensations, largely con­ sisting of visual and auditory elements. In con-

15

Approaching Consciousness

trast, Killpe's published results pointed to a total of fewer than 1 2,000 (Boring 1 942, p. 10). Who was telling the truth? Conflicting results are no surprise in any ex­ perimental discipline. What led introspectionism to a dead end was an additional methodological shortcoming: the lack of a generally agreed-upon method offalsifying any of the results. The nature of introspective reports constituting the core of the data in the introspectionist paradigm was colored by the subjects' previous training. Titch­ enerian introspectionists were very careful not to work with "naive subjects," enforcing strict procedures to avoid "stimulus errors," but this policy worked against them in the end. Although a rigorous and careful training program and meticulous repetition of the experiments provided an acceptable degree of statistical consistency within individual laboratories, the results across different laboratories were sometimes highly contradictory. Unfortunately, when individual "seemings" (of colors, sounds, tactile sensations) were what counted as the sole data, each "phe­ nomenal" report had to be taken at face value. And within such a framework, the degree and nature of previous training, which apparently was not standardized, made all the difference. Edwin Boring recounts an anecdote in which Titchener and Edwin Holt debated, in front of an audience of other psychologists, whether green was an "atomic color" or a combination of blue and yellow. Each side insisted on his own judg­ ment, and there was no means to settle the issue. One of the most serious conflicts, somewhat sim­ ilar to this but larger in scope, involved a staunch disagreement between the followers of Titchener and those of Killpe on the existence of "imageless thought." Titchener was convinced that all con­ scious thought involved some form of imagery, at least some sensory elements. However, subjects from Killpe's laboratory came up with reports of having experienced thoughts with no associated imagery whatsoever. The debate came to a stale­ mate of, "You cannot experience X," of Titch­ enerians versus "Yes, we can!" of Killpeians, and

remained the theoretical knot that i t was until introspectionism, as a whole, eventually disap­ peared against the rising tide of behaviorism. 2 7 In sum, the fact that introspectionism ulti­ mately located the locus of authority with regard to the data in the word of the subject, while training procedures for subjects were not stand­ arclized across laboratories to immunize against "stimulus-error," brought the death sentence to the movement. When the subjects' reports showed statistical inconsistencies, the whole in­ trospectionist community found itself up against a theoretical wall. This impasse, which surfaced as a result of several years of careful laboratory work, brought with it a sad ending to a research paradigm of hundreds of experiments and thou­ sands of subjects. 2 8 VB

Behaviorism

In contrast to introspectionism, behaviorism ar­ rived with an extremely straightforward method­ ology (that would ultimately cut, rather than try to untangle, the knot of consciousness), and it appeared as a fresh alternative in the troubling times of introspectionism. There was one and . only one element in its research agenda: publicly observable behavior. In the natural sciences, be­ haviorists argued, all phenomena under scrutiny were open to third-party observation. Behavior was a perfect candidate as a subject matter of this sort. Moreover, behaviorism was able to avoid introspectionism's fatal problem ofirreconcilably conflicting subjective reports by, in Boring's words, shifting "the locus of scientific responsi­ bility from an observing subject to the experi­ menter who becomes the observer of the subject" (Boring 1 953, p. 1 84). Watson championed this shift of locus and the change in the subject matter of the new psychol­ ogy from "facts of the internal" to "facts of the external" in a rather rallying manner in the fol­ lowing advice to his colleagues:

16

Giiven Giizeldere

Psychology as the behaviorist views it is a purely ob­ jective experimental branch of natural science. Its theo­ retical goal is the prediction and control of behavior. Introspection forms no essential part of its method nor is the scientific value of its data dependent upon the readiness with which they lend themselves to inter­ pretation in terms of consciousness. (Watson 1 9 1 3 , p. 1 58) You, as a psychologist, if you are to remain scientific, must describe the behavior of man in no other terms than those you would use in describing the behavior of the ox you slaughter. (Watson 1970, p. ix)

The term consciousness had never figured in the vocabulary of any natural science, and it had to leave the vocabulary of the scientific psychology as well. Watson was confident that behaviorism marked the beginning of an era that was also the point of no return for consciousness: The time seems to have come when psychology must discard all reference to consciousness; when it need no longer delude itself into thinking that it is making men­ tal states the object of observation . . . . This suggested elimination of states of consciousness as proper objects of investigation in themselves will remove the barrier from psychology which exists between it and the other sciences. (Watson 1 9 1 3, p. 1 63, 1 77)

Behaviorism remained a very influential para­ digm for psychology for over half a century and managed to have the words consciousness and in­ trospection disappear from the face of the Anglo­ American world.29 There were obvious reasons for the enthusiastic acceptance of behaviorism by psychologists, motivated by its promising, "trouble-free" methodology. However, behav­ iorism became influential as a doctrine not only of methodology but also one of ontology. The behaviorist line turned into a fundamental belief not only that whatever psychology-the disci­ pline-could study could be studied by observ­ ing behavior, but also that aU there was to psychology-the phenomenon-was observable behavior. This was what made (ontological) behaviorism both very strong and very weak: strong as a doc-

trine in its metaphysical claims and weak in grounding the strong claims it was making on what it took the world's constituents to be. The ultimate expression of the extreme view behav­ iorism came to hold about the ontology of con­ sciousness is reflected in the formula Karl Lashley used to characterize "strict Behaviorism": "Con­ sciousness is the particular laryngeal gesture we have come to use to stand for the rest" (Lashley 1 923, p. 240). However, the metaphysical foun­ dations of behaviorism, what it so passionately tried to detach itself from, turned out to be its own Achilles' heel, the cracked brick in the edifice. Even during the heyday of behaviorism, when all talk about consciousness was strictly taboo, consciousness was always present as a hidden variable in the minds and research agen­ das of psychologists. Boring was cognizant of this fact as early as the 1930s, when he declared: "Behaviorism owes its ism to consciousness. And what would it be without its ism? Well, it would be physiology" ( Boring 1963, p. 275). Much later, Julian Jaynes would retrospectively note that "off the printed page, behaviorism was only a refusal to talk about consciousness" (Jaynes 1 976, p. 1 5). This make-believe attitude about the absence of anything, let alone consciousness, occurring somewhere between the input impinging on the subject and the subject's subsequent behavior was also precisely what provided cognitive psychol­ ogy the fulcrum it needed to topple behaviorism. In Neisser's words, "the basic reason for studying cognitive processes has become clear as the rea­ son for studying anything else: because they are there. Our knowledge of the world must be some­ how developed from the stimulus input" (Neisser 1 967, p. 5).30

vm

Cogoitivism (and Beyond)

Ulric Neisser's Cognitive Psychology became a mark of a new era in psychology and proclaimed the name of the new game in its title. Neisser, in

Approaching Consciousness

the introduction to his book, mentions the change in the intellectual atmosphere among psycholo­ gists in a wry tone: "A generation ago, a book like this one would have needed at least a chapter of self-defense against the behaviorist position. To­ day, happily, the climate of opinion has changed, and little or no defense is necessary. Indeed, stimulus-response theorists themselves are in­ venting hypothetical mechanisms with vigor and enthusiasm and only faint twinges of conscience" (Neisser 1 967, p. 5). For consciousness research, the era of cognitive psychology was marked with a few timid over­ tures. With the advent of cognitive psychology, whose fundamental ideas were largely inspired by computational models, consciousness found a new niche, though in terms completely foreign to its past: it became a kind of component or aspect of information-processing models. Although only a small percentage of the models developed at the time secured a role for consciousness, cognitivism brought about the first signs of the dissolution of a taboo. Nonetheless, even these cautious begin­ nings were not easy; consciousness would have to wait until the current ongoing ascent of neuro­ psychology research to come back under the spotlight. In cognitivism, cognition needed de­ fense over behavior no more, but consciousness over cognition still did. In this context, George Mandler's manifesto "Consciousness: Respectable, Useful, and Prob­ ably Necessary," even though it was not the first article that came out of the cognitivist literature on consciousness, and despite being written in a somewhat gingerly manner, stands out as a cornerstone. J 1 Mandler opens his article with the following historical remarks: "I welcome this opportunity to act as amicus curiae on behalf of one of the central concepts of cognitive theory­ consciousness. Another statement, however im­ perfect, may be useful to undo the harm that consciousness suffered during fifty years (ap­ proximately 1910 to 1 960) in the oubliettes of be­ haviorism. It is additionally needed because so many of us have a history of collaboration with

17

the keepers o f the jail and t o speak freely o f the need for a concept of consciousness still ties the tongues of not a few cognitive psychologists" (Mandler 1 975, p. 229). Of course, Mandler was not alone in pointing to the importance of con­ sciousness in cognitive psychology. Tim Shallice, for instance, had observed a few years earlier that "theoretical developments in cognitive psychol­ ogy and the increasing use of introspective reports require a rationale, and that this should involve consideration of consciousness" (Shallice 1 972, p. 383). Interestingly, it was the success, not the failure, of information-processing models in explain­ ing learning, memory, problem solving, and the like-actually almost everything except con­ sciousness-that brought some attention to consciousness itself. The fact that consciousness seemed to be the last remaining unexplained phe­ nomenon in an otherwise successful new research paradigm helped highlight old questions about consciousness buried during the behaviorist era. Furthermore, similar developments were taking place in philosophy. Functionalist accounts, largely inspired by computational ideas, were being met with noticeable success in explaining propositional attitudes, whereas consciousness (in the sense of the subjective character of experience, or qualia) was largely being regarded as the only aspect of mind escaping the net of functionalist explanation. (See, for instance, Ned Block's in­ fluential article, "Troubles with Functionalism" 1 978, as well as Block and Fodor 1 980.) Shallice was one of the first to point out the special place consciousness occupied in the prob­ lem space of cognitive psychology: "The problem of consciousness occupies an analogous position for cognitive psychology as the problem of lan­ guage behavior does for behaviorism, namely, an unsolved anomaly within the domain of the ap­ proach" (Shallice 1 972, p. 383). Attempts to find some role for consciousness in a cognitive econ­ omy turned up results that at times exceeded expectations. Mandler, for instance, pointed to the possibility that consciousness might be the

18

Guven Gllzeldere

missing central element in a cognitivist frame­ work, able to tie together several separate lines of cognitive research: "I hope to show that con­ sciousness is . . . probably necessary because it serves to tie together many disparate but ob­ viously related mental concepts, including atten­ tion, perceptual elaboration, and limited capacity notions" (Mandler 1 975, p. 229).32 Of course, there were others on whose work Mandler was basing his claim. Most notably, Norman (1 968) and Atkinson and Shiffrin (1 968) had used consciousness as a property demarcat­ ing processes of different kinds (conscious versus unconscious processes) in their respective unistore and multistore models of memory. Treisman (1 969) and Posner and Boies (197 1 ), among others, talked about consciousness as a limited capacity processing mechanism. Shallice's idea was to equate consciousness as selector input in his cognitive model of the dominant action sys­ tem. Johnson-Laird characterized the "contents of consciousness" as the "current values of pa­ rameters governing the high-level computations of the operating system" ( 1983a, p. 465; 1 983b). All in all, the common presupposition driving the cognitivist research on consciousness was that "the basic phenomenological concept-con­ sciousness-can be mapped onto an information­ processing concept" (Shallice 1 972, p. 383). 3 3 Most of these models came complete with their flowcharts, with each functionally defined ele­ ment confined to its own black box and arrows indicating the direction of information flow among them. Consciousness, then, became a box among boxes-a module connected to various other modules of processing in which input was registered, intermediate results were transmitted, and output was delivered. This approach to con­ sciousness has, according to Neisser, a special strategical advantage: "It represents a theoretical coup: not only are the facts of attention appar­ ently explained, but psychology's most elusive target is finally nailed down to a box in a flow chart" (Neisser 1 976, p. 103).

A prominent account of consciousness in re­ cent cognitive psychology, Bernard Baars's (1988) A Cognitive Theory of Consciousness, is similarly given in information-theoretic terms with substantial use of functional diagrams. This trend of diagramming in cognitive psychology, inspired largely by flowcharts of computational models in computer science, also got imported into philosophy of mind by empirically minded philosophers. A primary example is Daniel Den­ nett's model in his "Toward a Cognitive Theory of Consciousness" (1 986). Similar functional flowchart models are also being used in some of the present day neuropsychological accounts. (Cf. Schacter 1 988 and Shallice 1 988, especially chap. 1 6.) IX

The Study of tbe UncomciolM

One important line of thought in the study of consciousness that has not yet been addressed in this chapter is the foundation of the crucial dis­ tinction between the conscious and unconscious aspects of mentality. According to Johnson­ Laird, "The division between conscious and un­ conscious processes is the best available clue to the structure of the mind" (Johnson-Laird 1983a, p. 466). Freud would probably agree. None­ theless, conceptions of the unconscious have changed from their Freudian origins to their cog­ nitivist incarnations. Following is a brief histor­ ical account of the unconscious.

The Freudian Unconscious Until the time of Freud, there was no proper the­ oretical framework in which to reject the Carte­ sian idea of equating the mind with whatever lay within the scope of one's consciousness. In other words, consciousness was generally taken to be "the point of division between miftd and not mind" (Baldwin 1 90 1 , p. 21 6)-the mark of the mental.

Approaching Consciousness

The received conception of the transparency of the mind to one's consciousness, found in Des­ cartes and Locke, was not without exceptions, however.34 Most notably, Leibniz, in his vision­ ary reply to Locke in New Essays, can be said to have anticipated some very important develop­ ments to come in psychology two centuries ahead of their time, especially those with regard to the nature and role of the unconscious: "There are a thousand indications which lead us to think that there are at every moment numberless perceptions in us, but without apperception and without re­ flection . . . . In a word, insensible [unconscious] perceptions are of as great use in psychology as insensible corpuscles are in physics, and it is equally as unreasonable to reject the one as the other under the pretext that they are beyond the reach of our senses" (Leibniz 1 9 5 1 , pp. 374378). Nonetheless, taking consciousness as marking the boundaries of mind by and large remained an influential maxim until the time of Freud. For instance, the ·entry for "consciousness" in the 1901 edition of the Encyclopedia of Philosophy and Psychology reads as follows: "[Conscious­ ness] is the distinctive character of whatever may be called mental life" (Baldwin 1 90 1 , p. 2 1 6). Within this context, the introspectionist convic­ tion of the time-that psychology is the "science of the mental"-provided an especially strong basis for rejecting the unconscious as part of the mental, and hence as a subject matter for psy­ chology. Titchener, for example, was resistant to the idea of the unconscious, to the extent of de­ claring it a theoretically dangerous construct for psychology: "The subconscious may be defined as

an extension of the conscious beyond the limits of observation. . . . [T]he subconscious is not a part of the subject-matter of psychology. . . . In the first place, the construction of a subconscious is un­ necessary. . . . Secondly, the introduction of a sub­ conscious is dangerous" (Titchener 1 9 1 5, pp. 326327, emphases in the original text). 3 5 None of this should be taken as claiming that the concept of the unconscious as a part or aspect

19

o f the mental was completely unheard o f o r un­ acknowledged, however. In other words, Freud was not really the inventor (or discoverer) of the concept of the unconscious in any way. On the contrary, the general intellectual atmosphere of the times preceding Freud's appearance allowed talk about mental activity of various sorts that occurred without the subject's awareness, at least in any direct way. For instance, the well-known metaphor of the mind as an iceberg, consisting of consciousness as the tip above the surface and of a subsurface unconscious component, constituted by hidden currents but nonetheless effective on one's conscious mental life, was generally recog­ nized and used. In particular, toward the end of the nineteenth century, the idea of the unconscious mind had become operative among many scientists, philos­ ophers, and literary scholars, in a lineage trace­ able back from Rousseau to Goethe, to Fichte, and to Nietzsche (Whyte 1960). Freud apparently acknowledged this, as reported by Ernest Jones, one of the most prominent Freud scholars, in the following statement he made at his seventeenth birthday celebrations: "The poets and philos­ ophers before me discovered the unconscious. What I discovered was the scientific method by which the unconscious can be studied" (quoted in MacIntyre 1 958, p. 6). There were also attempts to study the unconscious empirically. For in­ stance, Henri Ellenberger credits Gustav Fechner, a pioneer of psychophysics research, as the first person who tried to reveal the nature of the un­ conscious by experimental methods, though his work did not prove fruitful (Ellenberger 1 970, chap. 5). However, none of these ideas about mental processes going on in one's mind without being conscious were well formulated: there was no co­ herent account to explain the structure, func­ tional role, or operation of the unconscious, or the modality of its relation to consciousness in the general scheme of an individual's mental life. There was consensus regarding neither the nature of the unconscious, nor its place in regard to

Guven Giizeldere

consciousness, in the intellectual community. To this situation Freud brought a steadily evolving theoretical framework in which, for the first time, construction of hypotheses to answer each of these questions became possible. This is the sense in which Freud can be said to be the pioneer of the unconscious. 36 In Freudian theory, the unconscious proper consists of repressed processes, exerting stress on the conscious component of the subject's mind and shaping his or her daily life in substantial ways. This is in contrast to the preconscious, which includes those processes that only con­ tingently happen to lie outside awareness. What is preconscious can easily become conscious with­ out special techniques or effort; what is uncon­ scious has to be "brought to the surface" through the psychoanalytic technique with the help of an analyst. The Freudian unconscious, although related, is not the same sense of unconsciousness employed in the current cognitive psychology research re­ garding unconscious processes-the "cognitive unconscious." Unconscious processes of both kinds are opaque to introspection, but there is a difference between them. The Freudian uncon­ scious exists because of past events, explainable by repression mechanisms and the like, and is not in principle inaccessible. The cognitive uncon­ scious, on the other hand, exists due to the way our perceptual-cognitive system is constituted and lies in principle outside our access. The mechanisms that subserve depth perception, for instance, are taken to be hard-wired: they are not there because of repression, and they can never become conscious through any method, psycho­ analytic or otherwise. The recognition and study of the cognitive un­ conscious goes even further back than Freud, at least to von Helmholtz's work on perceptual constancy, and spans a substantial period, all the way up to the thesis of "unconscious perceptual inference" by Rock ( 1 983). Despite these differ­ ences, however, Freud's approach to the uncon­ scious was very modem and in anticipation of the "cognitive revolution."

20

An encompassing account of the Freudian un­ conscious, including its structure and dynamics, is given in Erdelyi (1 985). Erdelyi also makes a strong case that Freudian psychology was indeed very close, in essence, to the cognitive psychology of our day-especially in terms of its approach to understanding mental phenomena, and research methodology. He even goes on to reconstruct Freudian schemas of the structure of conscious­ ness, quite plausibly, in modem flowchart style. Neisser ( 1 976) also refers to Freud's diagrams depicting the structure of the tripartite division of consciousness, preconsciousness, and uncon­ sciousness as "flowcharts" (See Freud 1 950, p. 394).37 It is true not only that Freud antici­ pated some of the developments in cognitive psy­ chology but also that the Freudian unconscious, even if under different names, has played a sig­ nificant role as an influential construct in cogni­ tive psychology.

The Cognitive Unconscious There has also been a whole research industry in contemporary cognitive psychology involved in investigating the nature of the unconscious: men­ tal processes that underlie cognition but are them­ selves not conscious. 3 8 Over the past few decades, there has been an enormous wealth of data accumulated, operative in current psychological theory, in this area -from rules of Chomskian universal grammar, to computational mecha­ nisms underlying vision and the 2 1D sketch in­ spired by the work of the late David Marr (1 982), and to Newell and Simon's work on cognitive constraints in planning, problem solving, and game playing (1 972). Consequently, the classi­ fication of mental processes as conscious versus nonconscious is useful and not unusual (though controversial) in psychological practice, espe­ cially in research on psycho linguistics, attention, and perception. 3 9 Furthermore, as evidenced from contemporary psychology literature, re­ search on type identifying mental states as con­ scious versus nonconscious, and research on the nature of consciousness of the subjects who have

Approaching Consciousness

such states is being pursued on independent con­ ceptual grounds. In fact, the dichotomy of con­ scious versus nonconscious processes is not the only such ground on which current research in cognitive psychology rests. There are several other such distinctions, all overlapping i� various ways in their function to distinguish mental pro­ cesses that are directly available to the subject ("introspectable," reportable, etc.) and those that are opaque and unavailable, as reflected in a re­ cent note by cognitive psychologists Holyoak and Spellman: Theorists of diverse persuasions have been led to pro­ pose cognitive dichotomies, which have been given a rather bewildering array of labels: unconscious vs. con­ scious, procedural vs. declarative, automatic vs. con­ trolled, reflexive vs. reflective, and many others. These distinctions do not always divide cognition along the same lines . . . [but] there are tantalizing sim­ ilarities among the proposed dichotomies. In particular, the first member of each pair is generally viewed as in­ volving unconscious mental processes, a topic that has seen a recent resurgence of interest among experimental psychologists. (Holyoak and Spellman 1993, p. 265)

In sum, whether or not lohnson-Laird is right in his claim about the distinction between the conscious and the unconscious being the most important theoretical tool to study the mind, one can easily say that the investigation of the un­ conscious in cognitive psychology has proved to be at least as fruitful as the investigation of the conscious.40 x Status Report: From Information Processing to Qualia

What is the current status of cognitive psychol­ ogy? Information-processing models in psychol­ ogy are still popular, but they do not constitute the sole dominant paradigm any more. This is also reflected in models of consciousness. But there are other reasons, too, for the shift in the research paradigm with respect to consciousness. One of them is the recognition that the functional diagrammatic depictions of consciousness seem

21

to leave out something important: the subjective, experiential aspect ofconsciousness. Perhaps there is something about consciousness that makes its identification with specific modules of isolated functions fundamentally inadequate. Interestingly, it was Neisser who registered such a concern about the information-processing models of consciousness during the heyday of cognitivism: The treatment of consciousness as a processing stage is unsatisfactory in a still more fundamental way. It does justice neither to the usages of the word "consciousness" in ordinary discourse nor to the subtleties of experience. A better conception of consciousness, which has been suggested many times in the history of psychology, would recognize it as an aspect of activity rather than as an independently definable mechanism . . . . Conscious­ ness is an aspect of mental activity, not a switching center on the intrapsychic railway. (Neisser 1 976, pp. 104-105)

Many people, including philosophers, pro­ ceeded with Neisser's intuitions in the past few decades. Something essential to (at least our commonsense conception of) consciousness; it was largely believed, was necessarily left out in characterizing consciousness only by specifying its functional role in the cognitive economy of human mentation and behavior. This some­ thing-the phenomenal face of consciousness­ brings us back full circle to the problem of the two faces of consciousness. In the last part of this chapter, I will examine the dialectic of the opposition between the segre­ gationist and the integrationist intuitions, in the context of the causal and phenomenal character­ izations of consciousness.

PART THREE PROBLEMS OF CONSCIOUSNESS: A PERSPECTIVE ON CONTEMPORARY ISSUES AND CURRENT DEBATES "Consciousness is a word worn smooth by a mil­ lion tongues," George Miller once said. "De­ pending upon the figure of speech chosen it is a

Giiven Giizeldere

state of being, a substance, a process, a place, an epiphenomenon, an emergent aspect of matter, or the only true reality" (Miller 1 962, p. 25). The conceptual and historical analyses I have presented are in agreement with Miller: it is probably best to regard and treat consciousness as a cluster concept. There are simply too many connotations that go under the term, and it seems futile to try to specify a single concept that would cover all aspects of consciousness or a single "the problem of consciousness." Nonetheless, I have tried to illustrate, there is a coherent theoretical thread constituted by certain problems and not others, that one can trace from texts in early modem (if not ancient Greek) philosophy to the emergence of scientific psychology in the nine­ teenth century, to the present. The most troublesome feature of this thread is what has been most difficult to explain, and it is the topic I arrived at by the end of the historical analysis: the qualitative, or phenomenal aspects of consciousness, or qualia. Of course, the notion of consciousness theoretically outstrips the notion of qualia, and there are many fascinating aspects to consciousness that do not necessarily have a qualitative component (e.g., its representational aspect, its attentive and control components, and mechanisms of the unconscious). But it is also questionable whether qualitative and nonqualita­ tive aspects of consciousness can really be under­ stood or explained independent of one another. These are the questions I will focus on and pursue below. XI Consciousness and Intentionality: Two Dimensions of Mind

Jerry Fodor once remarked: "There are, I think, three great metaphysical puzzles about the mind. How could anything material have conscious states? How could anything material have se­ mantical properties? How could anything mate­ rial be rational?" (Fodor 1991a, reply to Devitt, p. 285). Having enumerated these three questions,

22

Fodor chooses to stay away from the first, despite his career-long devotion to the latter two (which he takes to be closely related). This attitude is not at all uncommon. It is generally accepted as a re­ ceived view that the two fundamental aspects of mind, consciousness and intentionality, can be studied in the absence of one another-at least that intentionality can be so studied with no ref­ erence to consciousness. Here, "consciousness" typically refers to the qualitative aspects of consciousness, and "intentionality" is taken sufficiently broadly to embrace questions about semantics, as well as rationality. Fodor's justifi­ cation is the following: It used to be universally taken for granted that the problem about consciousness and the problem about intentionality are intrinsically linked: that thought is ipso facto conscious, and that consciousness is ipso facto consciousness of some or other intentional ob­ ject. . . . Freud changed all that. He made it seem plau­ sible that explaining behavior might require the postulation of intentional but unconscious states. Over the last century, and most especially in Chomskian linguistics and in cognitive psychology, Freud's idea appears to have been amply vindicated . . . . Dividing and conquering-concentrating on intentionality and ignoring consciousness-has proved a remarkably suc­ cessful research strategy so far. ( Fodor 1991b, p. 1 2)

Not everyone agrees, however. In particular, John Searle recently argued for what he called the "connection principle": the thesis that conscious­ ness and intentionality are immanently linked, and, contra Fodor's thesis, any research strategy that tries to explain the latter without recourse to the former is doomed to failure. Searle states the connection principle as follows: "Only a being that could have conscious intentional states could have intentional states at all, and every uncon­ scious intentional state is at least potentially conscious . . . . [T]here is a conceptual connection between consciousness and intentionality that has the consequence that a complete theory of intentionality requires an account of conscious­ ness" (Searle 1 992, p. 1 32). 4 1 More recently,

23

Approaching Consciousness

Strawson (1 994) takes the similar position that consciousness is the only distinctive characteristic of the mind. The beginnings of this line of thought can be traced back to Brentano's discussion of the rela­ tion between mental and physical phenomena in Psychology from an Empirical Standpoint. Bren­ tano is acknowledged as being first to postulate intentionality as the mark of the mental in mod­ em terms. For instance, he says: "Every mental phenomenon is characterized by . . . the inten­ tional inexistence of an object, and what we might call, though not wholly unambiguously, reference to a content, direction toward an object (which is not to be understood here as meaning a thing) . . . . This intentional inexistence is characteristic ex­ clusively of mental phenomena. No physical phe­ nomenon exhibits anything like it" ( Brentano 1 874, pp. 88-89). And regarding consciousness, Brentano states, for instance, that "no mental phenomenon is possible without a correlative consciousness" (p. 1 2 1 ). Although the connota­ tions of the terms consciousness and intentionality have somewhat shifted from Brentano's time to the present, I think it is fair to note that there is a great deal of theorizing in his work that lay the foundations of an account of the mental that attempts to incorporate these two dimensions of the mind in a principled way.

Another attempt to characterize (phenomenal) consciousness and intentionality as the two hall­ marks of the mind, though for the purposes of a critique, is given by Richard Rorty (1 979, p. 24) in terms of the diagram in figure 1. 1 . There are in fact various ways to fill in such a diagram, and which elements of the mental are to occupy which cells is a matter of controversy. For instance, not everyone thinks that beliefs and desires are with­ out a phenomenal component (Searle 1 992), or that pains have no representational or intentional aspects (Dretske 1 995, Tye 1 995). And the ad­ herents of panpsychism would probably maintain that the cell that holds what Rorty labels "the merely physical" is bound to remain nil.42 None­ theless, this way of depicting the · dimensions of the mental is useful in terms of illustrating what has been the most problematic aspect of the study of consciousness. In this diagram, it is what Rorty calls "raw feels."43 Most of the current debates involving con­ sciousness revolve around the (possible) inhabi­ tants of this particular cell and their nature. Are there really such things as nonrepresentational but phenomenal properties? If there are, what is their ontological nature, what kinds of special epistemological problems do they present, and how can their semantics be given? Can they ever be captured in naturalistic explanatory scheme,

With phenomenal properties

Without phenomenal properties

Intentional, representational

Occurrent thoughts, mental images

Beliefs, desires, intentions

Noni ntentional, nonrepresentational

Raw feels (e.g., pains and what babies have when they see colored objects)

"The merely physical"

Figure I.I

Two dimensions of mind

Guven Giizeldere

or are they inherently bound to remain mysteri­ ous? These are the questions that constitute the consciousness debates today. xn

Perspectivity and Episternic Asymmetry

Naturally, there can be several different entry points to the kind of exposition I aim to present here. The investigation of consciousness is as fas­ cinating as it is difficult, and it presents unique epistemological and ontological difficulties. Al­ though my overarching goal is to provide an overview of the contemporary problems of con­ sciousness rather than try to present my own sol­ utions to them, I will start by presenting a brief profile of what I take to be the primarily respon­ sible component in the consciousness puzzle: the epistemic element of perspectivity. Perspectivity, or the fact that consciousness is a phenomenon that admits a distinction between "perspectives," or "points of view" in its explication, lies deep at the roots of the common understanding of con­ sciousness, as well as the attitude of puzzlement. Furthermore, epistemically based theses about consciousness seem much less controversial than ontologically based theses. I start by sketching a commonsense conception of consciousness and try to reveal just how perspectivity figures in it. Then I will proceed to examine its possible onto­ logical ramifications. Why does consciousness keep appearing as an unsolved puzzle for philosophy, psychology, and neuroscience? There do not seem to be similar puzzles associated with the study of, say, mem­ ory or learning, or biological development and growth. What is so special about consciousness? George Miller thinks that perhaps the unique difficulty involved in the understanding of con­ sciousness stems from the fact that consciousness is both the phenomenon we try to investigate and the very tool we need to use to pursue this inves­ tigation. "Turning a tool on itself," he says, "may be as futile as trying to soar off the ground by a tug at one's bootstraps." He continues: "Perhaps

24

we become confused because whenever we are thinking about consciousness, we are surrounded by it, and can only imagine what consciousness is not. The fish, someone has said, will be the last to discover water" (Miller 1 962, p. 25). Miller's observation is intriguing. One cannot, in principle, study the minute details of a micro­ scope's outer surface, for instance, by using the very same microscope. This would be impossible simply because of the way the microscope, as a tool, is designed and used. Neither can one di­ rectly take the picture of a camera by using the camera itself. But why should these considera­ tions apply to the study of consciousness using consciousness itself? One can certainly lay the body of a microscope under another microscope for examination or take pictures of one camera with another. It may be that the sort of recursive impossibility involved in the self-study of tools applies to the phenomenon of self-conscious­ ness-for example, one's study of one's own consciousness by introspection. But Miller is concerned with the study of consciousness in general here, not only self-consciousness, and it is not clear why the analogy should hold. 44 Nonetheless, Miller's point is related to what I see as the source of what makes consciousness pUZZling. The difficulty lies in the curious duality inherent in the (epistemic) study of the phenom­ enon. This duality does not need to be inherent in the (ontological) nature of the phenomenon of consciousness itself or its properties. In fact, as I have mentioned, the ontology of consciousness is an issue open to current debates. But as far as the epistemology of the matter goes, there appears to be a genuine asymmetry between the mode of access to facts of one's own consciousness and the mode of access to facts about others' conscious states. This asymmetry is what grounds the important distinction between systematic ap­ proaches to consciousness from the first-person perspective versus the third-person perspective. On the one hand, nothing is more intimately known by conscious human beings than the way the world (including themselves) appears to them.

Approaching Consciousness

We are all subjects of a variety of perceptual ex­ periences, thoughts and ideas, pains and tickles, joys and sorrows. Under normal circumstances, there is nothing more familiar with the way the face of one's spouse looks, the way a favorite drink tastes, the way the chronic heartburn starts to make itself felt. We all have, it seems, firsthand, immediate, direct knowledge of the rich phenom­ enology of colors, sounds, tastes, aromas, and tactile sensations that embellish our experiences­ the qualia. All these are constituents of a specific mode of being for every individual; they de­ termine, in Thomas Nagel's famous phrase, what it is like to be that individual (Nagel 1 974). Moreover, we all seem to have a "privileged" way of knowing about our own thoughts, feel­ ings, and sensations. Epistemological problems about knowledge notwithstanding, and even the question of the incorrigibility of the mental put aside, it seems that there is at least a special mode in which one's own experiences are present to one, in an immediate, direct way, not available for anyone else.45 Further, the common wisdom goes, we cannot genuinely entertain the possibility that we may be lacking consciousness; the very fact that we are questioning our own consciousness renders the possibility of our not being the entertainer of some occurrent thoughts logically contradictory. If in nothing else, Descartes was perhaps right in this regard: The mere fact of being the bearer of (these) thoughts is, in the Cartesian sense, un­ mistakable evidence, for oneself, that one is con­ scious. This is the characteristic of the first-person perspective; from the inside, consciousness seems all-pervasive, self-evident, and undeniable. On the other hand, contemporary science tells us that the world is made up of nothing over and above "physical" elements, whatever their nature (waves, particles, etc.) may be. Where does this leave us with respect to the place of consciousness in an entirely physical world? "How can tech­ nicolor phenomenology arise from the soggy grey matter of brains?" as Colin McGinn asks (McGinn 1989, p. 349). Can one accommodate a

25

respectable account of consciousness that does justice to the richness of our conscious experi­ ences of sights and sounds within a framework based on a monistic materialist ontology? Con­ sciousness just does not seem to be the kind of phenomenon that is amenable to the sort of sci­ entific explanation that works so well with all other biological phenomena, such as digestion or reproduction. The facts that would settle the question of whether some organism-an animal or a fellow human being-is digesting do not seem to be available in the same way when it comes to the question of consciousness in others, especially in the case of organisms phylogeneti­ cally distant from ourselves. There seems to be no ordinary way to peek into the inner lives of oth­ ers-to feel their pains, go through their sensa­ tions, or directly observe their consciousness.46 That is, there seems to be an epistemic impossi­ bility for anyone to have direct access to the qualia of others-literally share their first-person perspective, in short, to partake in the mode of what it is like to be them. These are the limitations of the third-person perspective: from the outside, firsthand exploration of the consciousness of others just seems to be out of the reach of ordi­ nary scientific methods, others' experiences being neither directly observable nor noninferentially verifiable. And therein this asymmetry between the first- and the third-person perspectives lies the epistemic duality in the study consciousness. But what exactly follows from this asymmetry? What do the limitations of the third-person approach entail, for example? Are there any insurmountable problems for a systematic study of consciousness-its nature, underlying mecha­ nism, evolutionary function, ontological status? After all, the third-person perspective is what is and has successfully been operant in the scientific practice of the past several centuries, and no one doubts that it can provide valuable advances in the understanding of consciousness. But the issue is whether such an approach is always doomed to leave something essential to consciousness out of its explanatory scope. In short, is there an

26

Guven Giizeldere

unbridgeable "explanatory gap" inherent in the

to describe an experience by stating its qualitative

third-person approaches to consciousness, and if

aspects as they seem to the experiencer. The fruity

so, can it be remedied by the deployment of a

character is directly experienced

crossbred conceptual scheme that embodies a

son whose gustatory and olfactory nerves are ex­

first-person approach in the investigation of con­

cited by the wine. As such, the experiencer has a

sciousness?

privileged status; she gets to have the experience,

only

by the per­

whereas the listener only gets to hear the descrip­

xm

First-Person versus Third-Person Approaches to Consciousness The epistemic asymmetry inherent in the study of consciousness can be found as manifested under different names, roughly as variations of one an­ other and as occupying critical roles in theoretical junctions. The notion of the first-person versus third-person perspectives is one such contrastive pair. Yet another similar distinction is that of the subjective versus the objective, or the "phenom­ enal" versus "physical." One way of describing a particular experience, say, of tasting a particular vintage of a certain kind of wine, would be to try to state how the wine tastes to me-that is, what it is like for me to have that particular gustatory experience. This is indeed the ordinary, even if not so easy, way. It involves the usage of qualitative terms such as "fruity," "with a hint of tobacco," or "full­ bodied," and the hope of conveying some sense of what the tasting of that wine would be like had it been experienced by the listener. The other way would be to proceed by way of giving a description of the specific ways in which my tastebuds are excited, my olfactory nerves are activated, my blood chemistry has changed, and so forth. This would not be the most ordinary way of describing one's gustatory experiences, but perhaps one can overhear two devoted neu­ rologists talking this way to one another at a conference reception. In any case, in the right context, it is clear that this alternative method would also be informative in conveying something about the nature of one's experience. There is clearly an important difference be­ tween the two methods.47 The first one attempts

tion. As far as the second method is concerned, however, the experiencer and the listener are epistemically on a par. The description of the perturbations in the experiencer's nervous system is open to public observation and verification, and ordinarily, no qualitative terms about the experience (how it feels) need to be involved. In other words, whereas the instantiation of the phenomenal properties of an experience is di­ rectly accessible only to the experiencer, the in­ stantiation of its intrinsic neurophysiological properties can be equally observed by many.48 In the latter case, what is at issue are the publicly observable aspects of the experience-not how it feels but what it does. The

important

question

is

to

determine

whether these two methods have distinct scopes of explanation and whether they are necessarily committed to distinct ontologies.

I have outlined

two general approaches to consciousness, each of which respectively takes one of the two above methods as primary. The first of these approaches takes consciousness "as consciousness seems" and, in accord with the phenomenal character­ ization, regards its qualitative aspects as the pri­ mary components of any explanatory scheme. The second one takes consciousness "as con­ sciousness does," and, in accord with the causal characterization, tries to account for conscious­ ness in terms of what it does and the role it plays in one's cognitive economy. Put in different terms, one can call these the "first-person-per­ spectival" versus the "third-person-perspectival" approaches to consciousness.49 The dichotomy of the "two faces of consciousness" manifests itself in yet other distinctions and under other names, to which I now tum.

27

Approaching Consciousness

XIV

The Two Faces of Consciousness Revisited

The phenomenal and the causal characterizations are merely expressions of what seems most im­ portant, or primary, in the understanding of the nature of consciousness. They are not, in them­ selves, in opposition with one another. It is only under the dictum of the segregationist intuition that they are considered essentially antipodal and mutually exclusive. The issue of how to locate the phenomenal and the causal characterizations with regard to each other is central to the dialectic of certain ongoing debates surrounding the "phenomenal" versus "access" senses and the "easy" versus "hard" problems of consciousness. By "phenomenal characterization of con­ sciousness" I mean a characterization given fu�damentally in first-person terms, describing epIsodes of inner life in terms of how they feel or seem to the subject who experiences them. William James was interested in both the tempo­ ral and the spatial structure of consciousness and his chapter "Stream of Thought" (James 1 50a, pp. 224-290) provides an excellent example of such a characterization. 5 0 Apart from James and the continental phenomenologist philosophers introspectionist psychologists were paradigmati � cally interested in the phenomenal aspects of ex­ perience, and they relentlessly pursued the project of "�apping the boundaries of the inner space of conscIousness." Along those lines, Titchener de­ fined consciousness as the occurrent parts of one's mind, accessible by introspection, at any given moment: "My 'consciousness' is the sum of men­ �al . processes which make up my experience now; It IS the mind of any given 'present' time. We might, perhaps, consider it as a cross-section of mind" (Titchener 1 902, p. 1 3). But there is more to the phenomenal charac­ terization in the way "problem of phenomenal consciousness" is understood today. In particu­ lar, the problem has now transformed into the exploration of explanatory laws that would ac­ count for how particular phenomenal aspects of

9

consciousness could arise from their physiological underpinnings. 5 1 Deep down, this problem is a manifestation of the gap that separates our direct understanding of consciousness in first-person terms, versus the objective, physicalist accounts of consciousness given in third-person terms. The roots of this problem are indeed unique; no other phe�omenon �resents us with two distinct epis­ terruc perspecttves from which it can be inves­ tigated. Given this duality, how does the "causal characterization" of consciousness fare against its phenomenal counterpart? A causal characterization of consciousness can be given in many dimensions. One can try to ac­ count for consciousness in terms of behavioral manifestations, or of its role and place in the general mental economy. The former approach was behaviorists' failed solution to account for (or, rather, do away with) consciousness. The former became a canonical characterization in behaviorism's successor, cognitive psychology and in functionalist schools of philosophy. Behaviorism, in its explicit form, is no longer around. But it is worth mentioning again how the most obvious difficulty in relying entirely on ex­ rnal criteria gave way to information-proces­ smg accounts of consciousness in particular, and of mental phenomena in general. Behaviorism left no room for the possibility of the presence of consciousness in the absence of external behavior. Put differently, the absence of evidence from the third-person perspective implied the theoretical rejection of all experience that is generally char­ acterized in first-person terms. Given that the most familiar aspects of consciousness have to do with its phenomenology (think of James' stream), . thIS result stood out as the most difficult one to accept. Even many behaviorists balked at biting the bullet and claiming that a person who is sitting perfectly still with no vocal cord activity whatsoever (behaviorists' characterization of "thinking") would ipso facto be unconscious. 52 This claim, it seemed, was readily refutable in one's own everyday phenomenology. There was, after all, an epistemic component to the

�

28

Guven Giizeldere

phenomenon of consciousness that cried for a characterization in first-person terms. However, behaviorism, in its attempt to associate con­ sciousness and behavior conceptually, and thereby fully externalize consciousness, left no room for talking about consciousness as it is experienced by the subject. Everyone smiled at the joke about one behaviorist's asking another, after having made love, "It was great for you; how was it for me?" but (fortunately) not many took the scenario as a serious possibility. 53 A logical next step in trying to account for consciousness in causal terms was to reverse the behaviorist direction and, to some extent, re­ internalize the causal criteria of consciousness. This provided a groundwork for functionalist philosophy of mind, and cognitive psychology. Under such a more relaxed framework, con­ sciousness was allowed to be individualized by the role it played, as an integral component of the larger network of mental states and processes. With the promising application of computational ideas and information-processing models in psy­ chology, it was canonically characterized as a process accomplishing a specific task, a module with a specific function in a cognitive diagram, or an abstract property of the overall system. Here is a paradigmatic characterization in the cognitivist framework: Consciousness is a process in which infonnation about multiple individual modalities of sensation and percep­ tion is combined into a unified multidimensional repre­ sentation of the state of the system and its environment, and integrated with infonnation about memories and the needs of the organism, generating emotional re­ actions and programs of behavior to adjust the organ­ ism to its environment. . . . The content of consciousness is the momentary constellation of these different types of infonnation. (Thatcher and John 1977, p. 294)

What is important to note here is that this char­ acterization is given largely in a third-person per­ spective. Consciousness is identified with what it does but not necessarily how it feels to the experi­ encing subject.

Once again, it is essential to ask at this stage whether these two perspectives are mutually ex­ clusive regarding their explanatory roles-that is, whether an account of causal consciousness provides us with no understanding of how con­ sciousness seems to the first-person subject, and whether an account of phenomenal consciousness has no elements that figure in the understanding of what consciousness does. As I will suggest be­ low, my answer is "not necessarily." A further question along this line would be what onto­ logical consequences the perspectival asymmetry in the epistemology of consciousness entails. There, my answer will be "not very much." But first let me bring into the picture another dis­ tinction, proposed by Ned Block, which aligns well with the distinction between the causal and phenomenal conceptions of consciousness.

Access Versus Phenomenal Consciousness Block ( 1995) distinguishes between "access con­ sciousness" and "phenomenal consciousness" as follows: Access (A) consciousness: A state is access-conscious if, in virtue of one's having the state, a representation of its content is (I) inferentially promiscuous, that is, poised for use as a premise in reasoning, (2) poised for rational control of action, and (3) poised for rational control of speech . . . . These three conditions are together suffi­ cient, but not all necessary. Phenomenal (P) consciousness: P-consciousness is ex­ perience. P-conscious properties are experiential ones. P-conscious states are experiential, that is, a state is P­ conscious if it has experiential properties. The totality of the experiential properties of a state are "what it is like" to have it. (pp. 230-23 1)

Defining A-consciousness is a straightforward matter. In the case of human beings, A-con­ sciousness is a cognitively interwoven aspect of mental life, underlain by three crucial capacities centered around rationality: rational cogitation, speech, and action. (For the general case, not all three conditions are necessary for A-conscious-

29

Approaching Consciousness

ness because, Block maintains, animals without speech can have mental states of the A-conscious type.) Construed as such, A-consciousness fits well in the domain of propositional attitudes in philosophy of mind, and it is just the perfect sort of subject matter for cognitive psychology. 54 P-consciousness is more problematic. Block starts out his analysis of P-consciousness by stating the difficulty particular to it: "Let me acknowledge at the outset that I cannot define P-consciousness in any remotely noncircular way. . . . The best one can do for P-consciousness is . . point to the phenomenon" ( Block 1 995, p. 230). The way Block himself goes about char­ acterizing P-consciousness is either "via rough synonyms" or by examples. P-consciousness, as expected, is what I have been referring to as the phenomenal aspect of consciousness. Among the P-conscious properties that endow a mental state with P-consciousness in virtu� of its having them are, for instance, the way it feels to "see, hear, smell, taste, and have pains" and more generally, "the experiential properties of sensations, feel­ ings, and perceptions". Furthermore, Block maintains, P-conscious properties are "distinct from any cognitive, intentional, or functional property" (p. 230). Block thinks that it is not an embarrassment that he cannot provide a noncircular definition of P-consciousness. But why is it difficult to provide a straightforward definition of P-consciousness, and why should this not be considered a cause of disconcertment? According to Block, that there is no way to give a reductive definition of P-con­ sciousness is not embarrassing given the "history of reductive definitions in philosophy," presum­ ably full offailures. It is still not clear, however, whether the in­ ability to define P-consciousness reductively is sufficient reason to think that the only other al­ ternative must be an ostensive definition. Even if it is, it may be useful to ask why the definition of P-consciousness is, unlike other definitions, thus obliged to ostention. In the case of A-conscious­ ness, there seems to be no such problem. Block's .

definition of A-consciousness is not, strictly speaking, reductive, and it serves its purpose well with no need for ostension. Could there be some­ thing inherent in the pretheoretical construal of P-consciousness such that it does not allow a nonreductive but also nonostensive definition? More importantly, could it be that the particular way Block's distinction carves out phenomenal consciousness, separating it completely from its causal and functional aspects in accord with the "segregationist intuition," renders its inves­ tigation by means of scientific methods theoret­ ically impossible? Put differently, could we be painting ourselves into a corner by a conceptual commitment to Block's distinction such that we end up with a number of straightforward prob­ lems about A-consciousness and a conjured-up "hard problem" of P-consciousness that in prin­ ciple admits no solution?55 This last question leads directly into a related debate that has its roots in the "explanatory gap" problem, re­ cently dubbed by David Chalmers the "easy and hard problems of consciousness."

The "Easy Problems of Consciousness" and the "Hard Problem" Chalmers ( 1 995) characterizes the "easy prob­ lems" as those concerning the explanation of various cognitive functions: discriminatory abil­ ities, reportability of mental states, the focus of attention, the control of behavior. Of course these are not trivial problems at all, and labeling them "easy problems" should not be taken as down­ playing their complicated nature. Rather, Chal­ mers's point is that "there is no real issue about whether these phenomena can be explained scientifically." They can be. "All of them are straightforwardly vulnerable to explanation in terms of computational or neural mechanisms" (p. 20 1). What makes "the hard problem" of consciousness a different kind of problem is, Chalmers maintains, its resistance to all the methods that explain, or have the potential to explain, the rest of the problems. Put differently,

Guven Giizeldere

there is a different kind of problem about con­ sciousness that may evade the successes of all standard scientific advances. Such a problem would be a hard problem indeed. What is it? According to Chalmers, "The really hard problem is the problem of experience." More specifically, it is the "subjective" aspect of every experience that resists explanation. The notion of "subjective aspect" is given, as Block does, in Nagelian terms: There is something it is like to be a conscious organism and have experiences. In other words, "what it is like to be" constitutes the subjective character of the experiences of the organism in question. This much is also in line with Block's characterization of phenomenal consciousness. I will give a sorted-out schema of the theoret­ ically interwoven notions of "phenomenal as­ pect," "subjective character," and "what it is like to be" in section XVI below. For now, it is useful to observe that the line that separates Chalmers's "easy" and "hard" problems is the counterpart of the line that separates "access" and "phenom­ enal" consciousness in Block, which also mirrors the distinction between the causal versus phe­ nomenal characterizations of consciousness out­ lined earlier. Given these distinctions Chalmers states the "hard problem" as the problem of bridging the explanatory gap between accounts of the causal-functional (physical) kind and the oc­ currence of specific phenomenal aspects. He asks: Even when we have explained the performance of all the cognitive and behavioral functions in the vicinity of experience-perceptual discrimination, categorization, internal access, verbal report-there may still remain a further unanswered question: Why is the performance of these functions accompanied by experience? . . . This fur­ ther question is the key question in the problem of con­ sciousness. Why doesn't all this information-processing go on "in the dark", free of any inner feel? (Chalmers 1995, p. 203).

There are two related questions Chalmers raises, and the roots of both go back at least a hundred years in the history of psychology and

30

philosophy. The first has to do with how to bridge the explanatory gap between physical mechanism and phenomenal appearance, or brain and mind, as discussed in the "Mystery of Consciousness" section in section I. The second asks whether all the activity on the "physical" side could go on as usual in the total absence of any counterpart phenomenology. The former question is based on Levine's (1983, 1 993) original formulation of the problem of the "explanatory gap," which has an­ tecedents in considerations raised by Saul Kripke (1980) and Thomas Nagel (1974, 1 986) (though reaching different conclusions). The latter ques­ tion is a version of the so-called absent qualia problem. 56 Similar considerations also underlie what William Seager ( 1 99 1 ) calls the ultimate problem of consciousness. He asks: "Why is it so hard to think about consciousness, to formulate reasonable models of the relation of particular modes of consciousness to their physical bases?" The answer Seager offers, in agreement with Nagel, has to do with the uniqueness of conscious­ ness as a phenomenon: "There is no model by which we can satisfactorily understand the rela­ tion between conscious experience and subvening physical state since this relation is absolutely unique in nature" (pp. 223-234). Before discussing the status of the hard prob­ lem further, let me first sketch a larger framework in which a number of relevant questions can be located and pursued. This framework will also be useful in revealing just how much explanation along one question-path can be useful in explain­ ing issues in neighboring problems. XV The Four W Questions and the Further-How Question It is true that the enterprise of approaching con­ sciousness within a scientific discipline has tradi­ tionally been very problematic, largely due to the inadequacy of the scientific third-person perspec­ tive all by itself as a penetrating tool for the study of the phenomenal character of consciousness.

Approaching Consciousness

The most influential assumption about conscious­ ness, as evidenced by the diversity ofthe literature, is that what makes it a tough nut to crack is in some crucial way related to those properties of consciousness that have to do with its phenomenal aspect. Problems such as the irreducibility of con­ sciousness, its imminent subjectivity, the status of its relation to its physical underpinnings, and so on all relate to the phenomenal side of consciousness. In order to get into some of the inner structure of this difficulty, consider the following five ques­ tions, which I will call the four W questions and thefurther-How question of consciousness: 1 . What are the media and mechanisms of con­ sciousness? Can consciousness occur in any type of material substance, or does it have to have a specific kind of underpinning (e.g., a carbon­ based molecular structure)? And what are the underlying mechanisms that facilitate conscious­ ness? 2. Where is, if anywhere, the locus of conscious­ ness? Can consciousness be localized in a specific organ, the brain (or a module in the brain), or is it endemic to the whole of the nervous system? Where is the seat of consciousness? 3. Who can be said to be a conscious being? Using consciousness as a type-identifying pred­ icate, one can ask: Is a chimp, a spider, a pro­ tozoan, or a robot conscious or nonconscious? (In a slightly different sense of consciousness, one can also ask of a person in a coma, or in sleep. or in a petit mal seizure whether she is conscious or unconscious.) 4. Why is there consciousness at all, and what is the role it plays in the general scheme of mental life and behavior of an organism? To put it in evolutionary terms, which function does con­ sciousness serve such that it was selected as a trait in the phylogeny of certain classes of living things? 5. How does consciousness arise in, or emerge from, its underlying substance, structure, and mechanism, in the way it does?

31

Notice that the How question seems to be a further question, the answer to which may not be completely revealed even if all the previous four W questions are already adequately answered. The answer to the How question may involve the postulation of, in Chalmers's terms, an extra in­ gredient, which makes the question difficult in a unique way. Even when all the underpinnings of consciousness, including its medium, locus, and mechanism, are revealed, and conscious and nonconscious things are, at least according to some operational definition, properly categorized and explained, a further question may remain: Just how is it that one experiences the particular sort of phenomenal quality that one does, rather than a different quality, or even none at all? Or, more generally, how does any physical mechanism give rise to any kind of phenomenal experience? Because of this extra ingredient seemingly inherent in the How question, I call it the further-How question. As must be clear from the formulation of the further-How question, the difficulty surrounding the extra ingredient, the gap that remains not bridged, owes its difficulty to the phenomenal as­ pect of consciousness. The further-How question is generally considered to be categorically more difficult compared to the other four-hence the dubbing: "the hard problem of consciousness" versus the rest, the "easy problems." It is important to notice that the term extra ingredient can carry greatly different theoretical weights. For instance, the missing extra in­ gredient may be merely explanatory, due to an undeveloped concept, or some other theoretical tool. That would raise only an epistemological problem. But it may also mean a missing in­ gredient in the part and parcel of the world, in its ontology. It has been suggested that the missing ingredient is indeed ontological, and conscious­ ness should be added to the list of fundamental physical elements of the universe. For instance, Chalmers ( 1 995) claims that "a theory of con­ sciousness requires the addition of something fundamental to our ontology" and suggests we

Giiven Giizeldere

take experience as fundamental "along-side mass, charge, and space-time" (p. 2 1 0; see also Chalm­ ers 1 996). Nagel and Searle have respectively made the same point in terms of the subjective properties of consciousness, in, for instance, the following passages: The subjectivity of consciousness is an irreducible fea­ ture of reality-without which we couldn't do physics or anything else-and it must occupy as fundamental a place in any credible world view as matter, energy, space, time, and numbers. (Nagel 1986, pp. 7-8) Conscious mental states and processes have a special feature not possessed by other natural phenomena, namely SUbjectivity. It is this feature of consciousness that makes its study so recalcitrant to the conventional methods of biological and psychological research, and most puzzling to philosophical analysis. . . . The world . . . contains subjectivity as a rock-bottom element. . . . In the sense in which I am here using the tern, "sub­ jective" refers to an ontological category, not to an epistemic mode. (Searle 1992, pp. 93, 95)57

The same idea has also been favored among those who try to find fundamental theoretical connections between consciousness and quantum physics, as well as those who popularize on this theme. For instance, an interview with Nick Herbert, the author of Elemental Mind: Human Consciousness and the New Physics, outlines his position as arguing that "consciousness itself must be considered a 'fundamental force' of the universe, 'elemental', on a par with such irredu­ cible phenomena as gravity, light, mass, and electrical charge" (quoted in "The Consciousness Wars," Omni, October 1993, p. 56; see also Her­ bert 1 993). In a somewhat similar spirit, theoret­ ical physicist Henry Stapp claims that "an analysis of the measurement problem of quantum theory points to the need to introduce conscious­ ness, per se, to physics," stressing as well that a complete account of consciousness can be given not in an "ontologically and dynamically mon­ istic conceptualization of the world provided by classical-mechanics" but only "within a dualistic

32

quantum-mechanical conceptualization of na­ ture" (Stapp 1 996, p. i). The study of consciousness can take any one of the above five questions as its entry point to investigation. Indeed, various people have made attempts to approach the phenomenon of con­ sciousness by respecfive\y addressing each of these issues. But the further-How question has typically generated less success than others. As a matter of fact, it led to grim diagnoses about the "explanatory gap," thought by some, such as Colin McGinn ( 1989), to lie possibly forever be­ yond the grasp of human understanding. The seeming uniqueness of the further-How question, given the lack of apparent promising directions to pursue it in any existing method­ ology, led McGinn ( 1989) to take its conclusions perhaps too seriously. The very same consid­ erations, on the other hand, can lead one to think that there is perhaps something fishy about the whole setup. The way the problem is presented relies obviously on a set of presumptions about the metaphysics of phenomenal consciousness, as well as the nature of scientific explanation. Could it be that the reason we seem to have no clue about how to explain the further-How question is that there is really nothing there to explain?58 This brings up a metalevel issue: whether a com­ plete explanation of the four W questions will in fact leave some further aspect of consciousness unexplained, such that the further-How question will remain untouched, unscathed, and in need of explanation as ever? To take up a favorite exam­ ple of the Churchlands from the history of scienti­ fic explanation, what can assure us that the further­ How question will not evaporate in time just as did questions about elan vital and phlogiston? Note that this sort of skepticism against the further-How question need not entail a defla­ tionary attitude toward consciousness in general. One can remain convinced that consciousness presents fascinating and real problems for philos­ ophy and science and that this is already justified in the history of its study, while not believing that

Approaching Consciousness

there is a further-How question in the way it is formulated, isolable from the four W questions such that no degree of understanding there will shed any light on it. I do not know if there is a decisive way to settle the metaissue at this stage of our understanding of consciousness, and thereby decide the fate of the further-How question. I do not know if it is useful, or even yet possible, to settle it at present. It seems that the opposing attitudes toward consciousness stem largely from pretheoretical, though (or perhaps, hence) deep-rooted and very strongly held, intuitions. Of course, it is crucial to try to systematically examine and uncover the often implicit presumptions that these intuitions embody, but doing that also requires under­ standing what is currently known and accepted about consciousness at the present theoretical level-that is, understanding what is known about the four W questions. Each of these W questions is interesting in its own way, and each has generated some fruitful thinking in­ dependently in different fields. Thus, I now tum to a brief exposition of their current status, pri­ marily the What and the Where questions.

The What Question With regard to functionally characterized vari­ eties of access consciousness, there is hardly any suspicion that consciousness is medium indepen­ dent. But regarding phenomenal consciousness, this question is open to speculation. The func­ tionalist intuitions suggest that if the existence of all mental phenomena, including P-conscious­ ness, is a matter of the functional organization of the elements in the nervous system, then the pos­ sibility that consciousness is a trait that is not re­ stricted to carbon-based animal brains of this planet should be allowed. Denying this possibility would be "neural chauvinism." Perhaps Searle comes closest to claiming that consciousness, and actually the mind in general, can occur only in human and animal brains,

33

or their causal (but not necessarily functional) equivalents, because of the "special powers of the brain," which cannot be matched by, for exam­ ple, digital computers. 59 Notice that the question here is different from those in the various absent qualia arguments. The possibility being ques­ tioned is not one of non-emergence (i.e., absence) of consciousness in functional equivalents of hu­ man brains or in human brains themselves. Rather, somewhat symmetrically, it is the possi­ bility of the emergence of consciousness in non­ brains. What about mechanism? Regarding the un­ derlying mechanism of a very important compo­ nent of consciousness, the binding of the various sensory features into a coherent whole in experi­ ence, the most promising recent results come from the work of Christof Koch and Francis Crick. In "Towards a Neurobiological Theory of Consciousness" (1 990), they hypothesize that what underlies the phenomenon of binding is the pattern of synchronous oscillations in the brain within the 40 to 70 Hz range during visual experience. 60 (See also Llimis and Ribary 1 994, in support of the 40 Hz hypothesis in the context of dream experiences. Metzinger 1 99 Sa explores how the binding problem relates to the integra­ tion of phenomenal content.) Now, let's examine this hypothesis in light of the distinctions introduced so far. Does it, for instance, explain the access or the phenomenal senses of consciousness (or both)? Since Crick and Koch do not have such a distinction, it is hard to know what they think. According to Block, the hypothesis is designed to explain P-consciousness; failing that, it can explain, if anything, only A­ consciousness. A true explanation of P-con­ sciousness, Block maintains, has to explain fur­ ther questions about why, for instance, it is the 40 to 70 Hz range and not some other. The discovery of an empirical correlation does not suffice to bridge the explanatory gap between the phenom­ enon as it appears to the subject and what its un­ derlying mechanism does.

Guven Giizeldere

This formulation is just another expression of the "hard problem" and, as such, falls in the pur­ view of the further-How question. Thus, while the Crick-Koch hypothesis (so far as it is correct) can be considered to explain successfully the What question of consciousness for some, it remains es­ sentially incomplete for the defenders of a Block­ ean conception of phenomenal consciousness.

The Where Question Is there a seat of consciousness? This question in its various incarnations has been discussed from the time of the ancient Greeks. What was once the question of the organ of reason in humans (e.g., the brain versus the heart) has now transformed into the question of the whole brain or a module in it, and if the latter, which? As early as the late nineteenth century, lames had discussed the question of the seat of con­ sciousness and declared that the cortex, and not the rest of the brain, is what is responsible for consciousness: For practical purposes, nevertheless, and limiting the meaning of the word consciousness to the personal self of the individual, we can pretty confidently answer the question prefixed to this paragraph by saying that the cortex is the sole organ of consciousness in man. If there be any consciousness pertaining to the lower centres, it is a consciousness of which the self knows nothing. (James 1 950, pp. 66-67)

lames's view was based on the experimental results of his day, which showed a significant correlation between cerebral processes and sub­ jective reports of conscious experience. Note that lames does not attempt to give an explanation of how the brain can possibly subserve conscious experience any further than outlining the relevant mechanism. In other words, lames does not seem to be after anything beyond the ordinary W questions. Clearly this sort of explanation does not satisfy those who are after the further-How question.

34

But regardless of whether the further-How question is a well-formed formulation of inquiry, there is a lot of work to be done in explairiing the mechanism of how and where consciousness emerges in a given organism. lames was perhaps one of the first "consciousness modularists" by proposing that it was only a certain component of the brain that subserved consciousness. 6 1 Al­ though it has always been in the scientific agenda, the belief in modularity in brain function has gained particular popularity over the last decade, especially due to the results coming from neuro­ psychology. Recent discoveries involving certain types of brain damage, such that the subjects be­ come deprived of only very specific, encapsulated perceptual or cognitive abilities (e.g., prosopag­ nosia--the deficit of recognizing faces while almost all other visual capabilities remain intact), have provided support for theses of modular architecture. 6 2 Extending this idea, one can transform the question of the modularity of mental function in general into the question of whether phenomenal consciousness in particular may be subserved by a module of some sort. Tim Shall ice (1988) puts forth such a view, and a modularity hypothesis seems to lie behind Daniel Schacter's DICE model, where consciousness is depicted as a sepa­ rate, functionally individuated box in the wiring diagram sketch (roughly speaking) of a nervous system (Schacter 1 988). Block is also sympathetic to these models and calls the view "that treats consciousness as something that could be accom­ plished by a distinct system in the brain" Carte­ sian moduiarism, in contrast to Dennett and Kinsbourne's Cartesian materialism. Cartesian materialism is the name Daniel Den­ nett and Marcel Kinsbourne (1 992) give to the general belief that there is literally a place in the brain "where it all comes together" -something like a spatial or at least a temporal finish line that determines the outcome of various brain pro­ cesses as a coherent, unitary, single experience. Dennett ( 1 99 1 ) calls this the "Cartesian Theater"

35

Approaching Consciousness

model of consciousness.63 The idea of such a

could be, in nomological or just logical possi­

logical line in the brain makes it possible to ask

bility, human replicas who nonetheless lack phe­

questions about the temporality of certain events

nomenal consciousness. I come back to this issue

that take place inside the brain against the mile­

in the discussion of epiphenomenalism.

post of the phenomenology of experience. Deny­ ing that such a line exists makes it logically

Regarding the

Why/Which

question, the liter­

ature is somewhat barren. Perhaps this is partly

impossible to impose a fine-grained order on brain

as a result of the fact that it is nearly impossible

processes as having occurred prior to or folIowing

to find any evolutionary role for phenomenal con­

a particular experience. Dennett and Kinsboume

sciousness to play under the decree of the segre­

present a forceful argument against Cartesian

gationist intuition,

materialism; for them it is the whole brain, if any­

contribution of consciousness, when it is taken as

thing, that is in some sense the seat of conscious­

causally efficacious in accord with the causal

ness. Today, the Where question, just like the

characterization, is just too obvious. In other

What question, remains a hotly debated issue.64

whereas the

evolutionary

words, so long as consciousness is characterized as essentialIy noncausal and nonfunctional, ren­ dered an epiphenomenon that makes no difference

The Who and the Why/Which Questions

in the world, it drops out of the pool of factors The question of who can be classified as a con­

that have survival value, and thus becomes ex­

scious being is largely subordinate to the question

planatorily irrelevant to evolutionary theory.

of what the underlying medium and mechanism

This is the conclusion Frank Jackson ( 1 982) de­

of consciousness are, at least in a materialist

fends (using the term "qualia" for phenomenal

framework.

consciousness):

Roughly

speaking,

those

beings

"[Qualia) are an excressence.

whose physical constitution (medium) allows the

They do nothing, they

instantiation of those properties that indicate

merely to soothe the intuitions of dualists, and it

explain nothing, they serve

the working mechanism of consciousness can be

is left a total mystery how they fit into the world

safely allowed into the "charmed circle" of

view of science. . . . Epiphenomenal qualia are

consciousness (barring difficulties inherent in the

totally irrelevant to survival" (p. 1 35)65 On the

What question itself). It is also common practice

other hand, if consciousness is taken as a

in medicine to have a more or less circumscribed

for different modalities of perceptual awareness

genus

set of behavioral and psychological criteria to

under a causal-representational characterization,

determine the occurrent presence or absence

pace Dretske (forthcoming), there remains no

of consciousness in patients (e.g., see the Roche Handbook ofDifferential Diagnosis on "Transient Loss of Consciousness," 1 989). Of course, the issue is not

so

straightforward,

philosophically puzzling question about its evo­ lutionary role. It would clearly be somewhat dif­ ficult for any creature to survive without sight, hearing, touch, smell, and so forth. (See also

especially when it comes to phenomenal con­

Dretske 1 996 for a discussion of what kind of

sciousness. Is there anything it is like to be a bat

differences qualia make vis-a.-vis judgments and

catching prey with its sonar system, or a dogfish

beliefs.) Similarly, Armstrong ( 1 980) attributes to

detecting electromagnetic fields in the ocean, or

introspective consciousness the biological func­

a robot clumsily walking about in an artificial

tion of making us aware of current mental states

intelligence laboratory? The answers and, more

and activities of our own mind, such that it be­

important, the advice on how to obtain these

comes "much easier to achieve

answers greatly vary. This question also leads

states and activities, to get them working together

to the discussion on "zombies" -whether there

in the complex and sophisticated ways necessary

integration of the

Guven Giizeldere

to achieve complex and sophisticated ends" (p. 65). See also Van Gulick ( 1 988, 1 989) for at­ tempts to locate a functional role for phenomenal consciousness, and Dennett ( 1 99 1 ), Flanagan (1 992), Dretske (1 995), and Flanagan and Polger ( 1 995) for further evolutionary considerations. 66 Finally, Jaynes (1 976) and Crook (1 980) take en­ tirely different approaches to the idea of the evo­ lution of consciousness (characterized in terms closer to what I called the social sense of con­ sciousness). Searle (1 992) tries to strike a balance between defending a version of the essentialist intuition while assigning an evolutionary role to consciousness. Finally, a number of neuropsy­ chological accounts identify consciousness with a specific information-processing module, in terms of a specific function it serves in the whole system. Although the concern is almost never evolu­ tionary in such accounts, they can be mentioned here for their effort to find a specific function for consciousness (see, for example, Schacter 1 988 and Shallice 1 988). But on the whole, there is much about the Why/Which question that re­ mains to be written than what is already there. Having considered the various characteriza­ tions of consciousness and the various questions one can ask about them, I now turn to the examination of questions about phenomenal consciousness in the landscape of current philo­ sophical debates.

XVI A Road Map for Phenomenal Consciousness and the Unbearable Lightness of Wluztitisliketobe

As we have seen, the concept of consciousness is a hybrid that lends itself to several different char­ acterizations. Part of my goal in this chapter was to tease them apart and treat them separately. Having done so, however, one sees that the prob­ lem of consciousness is like a Chinese box puzzle; for every distinction made, one discovers that further embedded distinctions are required.

36

In any case, given that the philosophical prob­ lems all revolve around the phenomenal charac­ terization of consciousness, it is reasonable to focus discussion there. Doing so actually reveals that what is commonly referred to as "phenom­ enal consciousness" is also itself a hybrid. Con­ sequently, it becomes imperative to bring the analytical microscope over there and to dissect the different elements in the tangle of phenomenal consciousness. Here I present a conceptual road map for locating various different philosophical problems, each associated with phenomenal con­ sciousness in one way or another. The term phenomenal consciousness is often used interchangeably with a variety of others, such as qualitative character, qualia, phenomenal properties, subjective awareness, experience, and what it is like to be a certain organism. (See, for instance, Block 1 994, pp. 2 1 0-2 1 1 .) This is a bunch. And to make matters worse, each of these concepts is known for its notorious elusiveness. Traditionally, the properties that go under the various names of "raw feels," "qualia," "qual­ itative character of experience," "phenomenal aspect of consciousness," and so on have all proved to be recalcitrant to systematic explana­ tions. Dennett points out that attempts to give a straightforward account of phenomenal proper­ ties have typically been frustrating; "no sooner does [the concept of quaJia) retreat in the face of one argument that 'it' reappears, apparently in­ nocent of all charges, in a new guise" (Dennett 1 988, p. 42).67 This elusiveness actually goes to the heart of the particular and long-standing problem of phe­ nomenal consciousness, which is often labeled a mystery. Elusiveness by itself is not what makes the problem persistent, however. Other­ wise, eliminativism could appear as a more appealing option. Rather, it is our unique epis­ temic relation to consciousness: phenomenal consciousness is perhaps the most difficult aspect of the mind to give up. An eliminativist stance toward the phenomenal aspect of mental life

Approaching Consciousness

seems the most counter-intuitive of all eliminati­ vist attitudes. That is why the question of phe­ nomenal consciousness does not just disappear out of the philosophical and, in other guises, psy­ chological and scientific landscapes. I thus find it important to lay out properly each conceptual component that contributes to the puzzle. It is worth noting, however, that among all notions that are associated with phenomenal consciousness, one has particularly captured philosophical intuitions more than any other-so much so that it has become the central notion underlying almost any discussion about con­ sciousness during the two decades since its pub­ lication. Unfortunately, it is also the most difficult to pin down or muster theoretical agreement upon. I have in mind Nagel's (1974) notion of "what it is like to be" a certain creature, or sub­ ject of experience. 6 8 Nagel's notion of "what it is like to be" has been so influential that it seems to have an omni­ presence in several distinct (even if related) prob­ lems with regard to consciousness. In particular, it gets pronounced in an intertwined way with the problem clusters that can be grouped under the headings of qualia, subjectivity, and the knowl­ edge argument. Nagel himself presents the issue of what it is like to be a certain creature as a theoretical basis for establishing the claim about that creature's having a certain ontologically irreducible point of view, which furnishes certain facts about the creature with subjectivity. For others, however, the notion of "what it is like to be" is taken to lay the ground for arguing for the reality of qualia, and for others, for the persuasiveness of the knowledge argument which claims that physi­ calism, as an ontological doctrine, is false. But the nature of the relations among each one of these problems is hardly ever spelled out in any detail. In fact, it seems that the notion of "what it is like to be" has become the wild card of consciousness problems. I will henceforth refer to it simply as the notion of whatitisliketobe.

37

Given this tangle, let me present the following schema as a conceptual road map to distinguish problems typically associated with phenomenal consciousness: 1 . Qualia: Experiences have phenomenal and thus noncausal, nonrepresentational, nonfunc­ tional, and perhaps nonphysical properties. 2. Subjectivity: Certain facts about experiences are subjective, that is, they cannot be completely understood except from a single kind of point of view. 3. Knowledge Argument: Certain facts about experiences are nonphysical. To this, one can add the "base element" in the formula: • Whatitisliketobe: There is something it is like to have experiences for a certain organism (or, simply, something it is like to be that organism).

I call whatitisliketobe a wild card, because it gets alluded to in discussions concerning any of the three problems mentioned above. To have certain qualia, it is generally presumed, is what­ itisliketobe an organism undergoing a certain experience; certain facts about an experience are subjective because there is somethingitisliketobe having that experience; and finally, whatitislike­ tobe having a certain experience constitutes non­ physical facts about that experience. I think, however, that it can be questioned whether this common denominator is not in fact theoretically vacuous. Perhaps whatitisliketobe has turned into nothing but a wild card-a convenient way of talking about any one of the three problems of phenomenal consciousness, without, due to its intuitive charm, having to specify anything fur­ ther. Then there would be no reason to look for a shared ingredient in need of explanation, above and beyond the explanation of these three prob­ lems.69 Nagel's original intention in introducing the notion of whatitisliketobe was, I think, to use it as an "intuition pump" for instating subjectivity

38

Giiven Giizeldere

rather than as a tool to talk about qualia. Fur­ the�, for Nagel, the scope of applicability of the notIon of subjectivity greatly transcends the problem of qualia, or consciousness in general; it �lso �der1ies problems about free will, personal Iden ty, and the self, as well as the ontological doctnne of physicalism. 7 0 The knowledge argument, formulated in con­ temporary literature by Frank Jackson, is also �u�h. closer in nature to issues surrounding sub­ JectIVIty than to the problem of qualia. In fact, it can be seen as a logical conclusion of the diffi­ culties Nagel raises about accommodating sub­ jectivity in a physicalist ontology. In a nutshell, the knowledge argument is based on the claim that certain facts about experiences evade all physicalist accounts, and no matter how much one learns about the physical (causal, functional, representational, and so on) aspects of an ex­ perience, some facts about how the experience fe�ls (to oneself, but more important, to others) WIll remain in the dark until one actually has that experience. Jackson attempts to establish this claim by means of a thought experiment that involves an imaginary vision scientist, Mary, who learns "everything physical there is to know" about color experiences without ever having color ex­ periences herself. Jackson's contention is that upon seeing a colored object for the first time in her l fe, Mary will learn something new, belying . phYSicalism. The pivotal issue here is whether the having of an experience constitutes a special class of irreducible "first-person facts" or whether what is lacking in Mary has to do with her ex­ periential "mode of access" to facts that she is �lready acquainted with (in the form of proposi­ tIOnal knowledge); on this point of contention the knowledge argument has generated a fair amount of literature.71 The interrelations between these problems need to be pursued further, but I will stop and opt for focusing on the most central player of the phenomenal consciousness debate: qualia. (I will henceforth use "phenomenal con­ sciousness" and "qualia" interchangeably.)

�

�

xvn

The Qualia Battles

The problem of qualia is one that surfaced under d tIerent gui�s in the philosophy literature during dltIerent penods. It is probably fair to state that qualia was the single most recalcitrant notion that resisted the rising wave of materialists in their program of giving an account of the mental by means of identity theory. For example, J. J. C. Smart mentions in his now-classic "Sensations an B�ain Processes" (1959) that among the eight objectIOns he considers, he feels the least con­ fident in his answer to the one about phenomenal properties. (This is Objection 3, attributed to Max Black.) Both U. T. Place ( 1956) and B. A. Farrell ( 1950), philosophers of the same era, note that the identification of the so-called raw feels with the straightforwardly physical properties of the nervous system has been the most elusive component of the overall program of identity theory in "Is Consciousness a Brain Process?" and "Experience," respectively. Herbert Feigl also wrestles with the same problem in his lengthy manuscript, The "Mental" and the "Physical" ( 1 967).72 A second wave in philosophy of mind came about, this time that of functionalism, in the 1 970s. The problem of qualia was again on stage; the phenomenal feels were considered the "Achilles' heel of functionalism" (Shoemaker 198 1 a)-the only aspect of mentality that es­ caped the net of functional explanations.73 It is during this period that the problems of absent qualia and a reincarnation of Locke's puzzle of inverted spectrum reached celebrity status. Critics of functionalism argued that a functionalist framework can provide an account of all compo­ nents of mental life but cannot capture its qualia, . lacki�g the theoretical tools to settle decisively questIOns about whether any two functionally equivalent systems ditIer (e.g., can be inverted) in their phenomenal aspects, or even whether a given system has any qualia at all. Thomas Nagel gives a concise characterization of the problem of

�

�

Approaching Consciousness

absent qualia as follows: "The subjective charac­ ter of experience . . . is not captured by any of the familiar, recently devised reductive analyses of the mental, for all of them are logically compat­ ible with its absence. [E.g.,) It is not analyzable in terms of any explanatory system of functional states, or intentional states, since these could be ascribed to robots or automata that behaved like people though they experienced nothing" (Nagel 1974, pp. 1 66- 1 67; my emphasis). 74 In a footnote to this passage, Nagel also enter­ tains the possibility of the impossibility of absent qualia, but rejects it: "Perhaps there could not ac­ tually be such robots. Perhaps anything complex enough to behave like a person would have expe­ riences. But that, if true, is a fact which cannot be discovered merely by analyzing the concept of ex­ perience" (fn. 2, p. 1 67; my emphasis). But what Nagel merely asserts as true has no argumentative force against certain causal-state identity theo­ rists and some functionalists. For they take exactly the opposite of Nagel's assertion (broadly construed to include not only behavior, but also causal, functional, and intentional character­ ization) as a fundamental assumption. For instance, David Lewis ( 1 966) states: "The definitive characteristic of any (sort of) experi­ ence as such is [by analytic necessity) its causal role, its syndrome of most typical causes and effects" (p. 1 7). Similarly, the concept of a men­ tal state for David Armstrong (1 993) is that of a "state of the person apt for bringing about certain sorts of physical behavior," where he regards the mind as "an inner arena identified by its causal relations to outward act" (p. 1 29). As such, the relation between experiences and causal (and/or functional, intentional, etc.) characteristics is taken to be, contra Nagel's assumption, inherent in the concept of experience. This kind of funda­ mental disagreement where each side is vulner­ able to the charge of question-begging against the other is a typical syndrome of the "qualia battles." The possibility of absent qualia is closely re­ lated to the doctrine of epiphenomenalism (that

39

phenomenal consciousness has no causal powers) and thus to the possibility of "zombies" (human replicas with all mental and behavioral attributes present save for phenomenal consciousness), as I argue below. The possibility of inverted spec­ trum, on the other hand, simply requires an in­ version of a particular set ofphenomenal qualities in some sensory domain, such as the hues in one's color space. 7 5 But let us pause and ask the same question al­ ready posed about consciousness: When friends and foes of qualia disagree about whether qualia exist, are they really talking about the same thing? The ontologically rather ordinary fact that phe­ nomenal properties of an experience exist only insofar as they belong to someone's experience (compare: geometric properties of a shadow exist only insofar as they belong to someone's sha­ dow), when combined with the epistemologi­ cally rather extraordinary fact that experiences cannot epistemically be shared, and hence every­ one can have "direct access" to only his or her qualia, seem to make it uniquely, even surpris­ ingly difficult to investigate the ontological nature of qualia. As such, it gives rise to a wide variety of positions regarding what qualia are. In "Quining Qualia," Dennett, one of the staunchest critics of the notion of qualia, tries to establish that "conscious experience has no prop­ erties that are special in any of the ways qualia have been supposed to be special." He attempts to show this by laying out what exactly it is that he wants to deny in denying the existence of qualia and sets up his target by identifying qualia with the "properties of a subject's mental states that are: 1 . ineffable, 2. intrinsic, 3. private, and 4. directly or immediately apprehensible in con­ sciousness" (Dennett 1 988, pp. 43, 47). The final verdict Dennett arrives at, after an elaborate chain of "intuition pumps" designed to show that the very concept of qualia is inherently confused, is an e1iminativist one: "There simply are no qualia at all" (p. 74). In contrast to Dennett's e1iminativist stance, the spectrum of other positions with respect to

Guven Giizeldere

qualia extends from taking qualia to be non­ physical properties that require a new ontology to reductively identifying qualia with neurophysio­ logical properties. There are also midway, con­ ciliatory positions. Paul and Patricia Churchland, for example, agree that when qualia are construed in the way Dennett does, the situation is indeed hopeless: "So long as introspectible qualia were thought to be ineffable, or epiphenomenal . . . one can understand the functionalist's reluctance to have anything to do with them" (Churchland and Churchland 1 982, p. 34). While promoting a realist attitude toward qualia, they claim that qualia will tum out to be properties intrinsic to the nervous system, such as spiking frequencies in the brain. Construed as such, qualia cease to be elusive, but their investigation also falls into the scope of disciplines other than philosophy or psychology. In the Churchlands' words: "The functionalist need not, and perhaps should not, attempt to deny the existence of qualia. Rather, he should be a realist about qualia. . . . [But, at the end], the nature of specific qualia will be revealed by neurophysiology, neurochemistry, and neurophysics" (Churchland and Churchland 1 982, p. 3 1 ). Owen Flanagan, who believes that an effort of triangulation involving phenomenology, psy­ chology, and neuroscience, which he calls the "natural method," can penetrate the mystery of qualia and help dispel it, follows suit in promot­ ing a more positive characterization of qualia: "Those who would quine qualia are bothered by the fact that they seem mysterious-essentially private, ineffable, and not subject to third-person evaluation. Qualia are none of these things." Al­ though Flanagan does not necessarily share the Churchlands' conviction that qualia will tum out to be properties in the domain of neuroscience, he too concludes that "there are no qualia in Den­ nett's contentious sense, but there are qualia" ( Flanagan 1 992, p. 85). A recent proposal in accounting for qualia comes from Fred Dretske's representational nat­ uralism ( Dretske 1 995). According to this view,

40

"all mental facts are representational facts" and hence, a fortiori, all facts about qualia are also representational. Dretske identifies qualia as properties that one's experience represents objects (or whatever the experience is about) as having. As such, qualia do not have to be given a func­ tional characterization or identified with neuro­ physiological properties. Rather, Dretske locates qualia outside the mind, in accordance with his externalist theory of the mind. This view has the advantage of maintaining a realist stance toward qualia while remaining in a perfectly naturalistic framework. 76 Finally, Ned Block brings the qualia issue back to the problem of "explanatory gap" and raises suspicions about the conceptual machinery of cognitive psychology to deal with qualia: "On the basis of the kind of conceptual apparatus now available in psychology, I do not know how psy­ chology in anything like its present incarnation could explain qualia" (Block, 1 978, p. 289). Block is neither as sure as the Churchlands about whether the answer to the nature of qualia will tum out to be in the domain of neuroscience, nor is he as optimistic as Flanagan in the promise of interdisciplinary methods to deliver a successful account of phenomenal properties. Nor is he convinced that qualia can be accounted for in a Dretskean representational framework. On the contrary, Block actually wants to raise more general doubts about the explanatory power of any mechanistic, functionalist, or in general phys­ icalistic schemes to account for the presence or emergence of qualia. His worry, in other words, is about how qualia can be accounted for as part and parcel of any physical system, including (or rather, especially) a brain, even if one thinks that it must be so accounted. Block states: "No physi­ cal mechanism seems very intuitively plausible as a seat of qualia, least of all a brain . . . . Since we know that we are brain-headed systems, and that we have qualia, we know that brain-headed sys­ tems can have qualia. So even though we have no theory ofqualia which explains how this ispossible, we have overwhelming reason to disregard what-

Approaching Consciousness

ever prima facie doubt there is about the qualia of brain-headed systems" ( Block 1 978, p. 28 1 ). As a consequence of this kind of general doubt about physicalism, the scenario involving beings physiologically and behaviorally similar to us, perhaps even identical down to the last molecular structure and behavioral trait, who nonetheless lack qualia altogether, is considered a genuine theoretical possibility. This step brings us to the debate on the notion of zombies and the doctrine of epiphenomenalism. xvm

Epiphenomenalism and the POSliibility of Zombies

Consciousness epiphenomenalism is the view that (phenomenal) consciousness has no causal powers and hence exhibits no effects in the world, though it may be the effect of some other cause itself. This doctrine and the possibility of zombies are closely related. If consciousness is an epi­ phenomenon, that is, not essentially linked to causal processes, or is only a recipient of but not a contributor to effects in a causal network, then there exists the possibility that the same organism that is taken to possess consciousness could be going through the very same mentations and be­ havior even if it had no phenomenal conscious­ ness at all. Subtract away the consciousness, and you still get the same beliefs, desires, motives, preferences, reasoning capacities, and behavior in the organism. But what you get is a zombie. Its pains, tickles, and itches are all "ersatz." The zombie does not feel anything, even if it thinks and acts as if it does. Its experiences lack the qualitative feels altogether. There is nothing it is like to be it.77 Put differently, zombiehood becomes a pos­ sibility only under a view that accords with epiphenomenalism. If we maintain that con­ sciousness has causal powers, then the absence of consciousness in my zombie twin, which is iden­ tical to me in every other respect, would make some difference. But by stipulation, there is no

41

difference whatsoever between persons and their zombie twins except the fact that the latter lack consciousness. Hence, denying epiphenomenal­ ism would also block the possibility of zombie­ hood. That is, if we accept that consciousness has causal powers, then my zombie twin cannot exist, even as a genuine theoretical creature. The doctrine of epiphenomenalism has a deep­ rooted history. The philosophers and the psy­ chologists of the nineteenth century hotly debated whether consciousness was part and parcel of the causal network that was responsible for the deci­ sions we make, actions we take, and so forth or whether it was just an idle spectator, riding along the causal processes, perhaps being caused by them, but without exerting any causal effect on those processes itself. Perhaps, the idea was, we are all automata, since all of our mental life and behavior seem to be determined by our nervous systems, in a purely mechanical framework, with no respectable place in it for consciousness. 78 Thomas Huxley was one of the most influential advocates of such a thesis, known as the autom­ aton theory of consciousness. The thesis was first formulated to apply to animals, in perfect agree­ ment with Cartesian intuitions. Huxley put the matter as follows: "The consciousness of brutes would appear to be related to the mechanism of their body simply as a collateral product of its working, and to be as completely without any power of modifying that working as the steam whistle which accompanies the work of a loco­ motive engine is without influence upon its ma­ chinery" (Huxley 1 90 1 , p. 240). But, of course, the real target was human beings and the nature of human consciousness. This is where Huxley's automaton theory differed with Descartes's in­ teractionist dualism. Huxley's account of the "brutes" was just a lead to make the same point for humans: "The argumentation which applies to brutes holds equally good of men. . . . It seems to me that in men, as in brutes, there is no proof that any state of consciousness is the cause of change in the motion of the matter of the organ­ ism" (pp. 243-244). 79

42

Guven Giizeldere

In Huxley, consciousness plays no contributory

a "friend of phenomenal consciousness" to a

role in the causal chains that take place in the

reconciliatory middle ground with a "qualia

nervous systems that totally determine the be­

skeptic." Most often, the disagreement between

havior of an organism; it only gets affected by the

the two parties comes down, for each side, to the

neural interactions. In contrast, Descartes's idea

charge of begging the question against the other.

of consciousness was one of an equally causally

The eliminativists charge the defenders of phe­

efficacious parameter in the formula of mind­

nomenal consciousness with believing in a fiction

Descartes is

and creating a philosophical problem out of it. In

thought to be the founder of interactionism,

return, the eliminativists get charged with holding

body

interaction.

As much

Huxley can be thought of

as

as

having laid out

the most preposterous philosophical fancy for

a clear foundation for epiphenomenalism with

denying their

respect to the mind.

qualia.

The fundamental

idea

opponents'

characterization of

about epiphenomenalism remained intact until

As an example, consider Daniel Dennett, who

the present day, but what was then dubbed the

is convinced that the notion of qualia "fosters

"Automaton Theory" has been transformed into

nothing but confusion, and refers in the end to

the "Problem of Zombies" in contemporary lit­

no properties or features at all" ( Dennett

erature. Of course, it is important to note that

p. 49). Ned Block, as a representative of the other

even if we establish that it is the truth of some

side of the spectrum, accuses Dennett of begging

1988,

version of epiphenomenalism that makes zom­

the question against (the existence of) phenom­

biehood a possibility, there remain important is­

enal consciousness ( Block

sues about what the nature of this possibility is,

ingly, the dialectic of the debate seems to be at

for example, whether it is empirical, metaphy­

an impossible impasse: the contention is at the

1 993, 1 995). Interest­

I

fundamental level of taking for granted versus

cannot do justice to in the limited space here.

denying the existence of a feature of mentality

sical, or conceptual. These are subtle issues that

ostensively.

Hence, rather than pursuing this line further, I

that can at best be defined

will step back once again and examine how the

of qualia, as exemplified by Block, claim that

something

Friends

background conditions for bringing the meta­

there is obviously

physical disagreement on the possibility of zom­

that can be theorized about under the name

biehood (just like the disagreement on the status

"phenomenal consciousness," while the qualia

of phenomenal consciousness) can be brought to a settlement. 80

there is

skeptics,

as

no

in their mental life

exemplified by Dennett, state that such thing to point at in their own

experience. 8 1

XIX

Stalemate: How to Settle the Phenomenal Comciousoess Dispute? There are a variety of positions on the ontological status of phenomenal consciousness in the liter­ ature, all the way from substance dualism to property dualism, to reductionism (via some form of identity thesis), to eliminativism (usually coupled with some kind of anti realist stance), to representationalism (maintaining a naturalized realism). However, the literature does not contain any knock-down argument that would convince

This is unfortunately the kind of philosophical junction at which most worthy disagreements hit rock bottom. Neither side is willing to concede their own point, and moreover neither side seems to have any way of demonstrating the validity of their claim. In another statement on the side of the friends of phenomenal consciousness, John Searle satirically asks: "How, for example, would one go about refuting the view that consciousness does not exist? Should I pinch its adherents to re­ mind them that they are conscious? Should I pinch myself and report the results in the Journal

of Philosophy?"

(Searle

1 992, p. 8). In contrast,

Approaching Consciousness

Dennett declares: "I cannot prove that no such sort of consciousness exists. I also cannot prove that gremlins don't exist. The best I can do is to show that there is no respectable motivation for believing in it" (Dennett 1 99 1 , p. 406). Of course, the situation on the whole (and the particular state-of-the-art philosophical under­ standing of the mind-body problem we have arrived at after twenty-five hundred years of pondering) is more nuanced than I have just sketched. For instance, the eliminativist position has more resourceful ways of undermining belief in qualia, and the "friends of qualia" have in­ tuitively appealing conceptual tools on their side, such as the absent and inverted qualia puzzles and theknowledge argument. Nonetheless, neither side can help finding the other's theoretical maneuvers equally unconvincing. The eliminativist strategy largely depends on the deconstruction of the concept of phenomenal consciousness, thus revealing theoretical tensions internal to it. In different ways, both Daniel Dennett and Richard Rorty take this approach (Dennett 1 988; Rorty 1 979). Dennett does this by providing a number of "intuition pumps," de­ signed to slww that our pretheoretical intuitions about phenomenal consciousness are far from being reliable and sound. On the contrary, as Dennett attempts to show, our commonsense grasp of the facts about phenomenal conscious­ ness can result in such conceptual dilemmas that it might be a better strategy to abandon any talk about phenomenal properties altogether. Dennett is quite straightforward in this approach; he says: "I want to make it just as uncomfortable for anyone to talk of qualia-or 'raw feels' or 'phe­ nomenal properties' or 'subjective and intrinsic properties' or 'the qualitative character' or expe­ rience-with the standard presumption that they, and everyone else, knows what on earth they are talking about" (Dennett 1 988, p. 43). If there indeed is a conceptual disarray sur­ rounding the notion of phenomenal conscious­ ness, it seems only fair to demand from those who

43

take the idea of phenomenal consciousness seri­ ously and use it as a fundamental theoretical tenet to come up with a clarified conceptual network of terms that all go along with the umbrella term of phenomenal consciousness. On the other hand, there may be good reasons to respect the words of the supporters of phenomenal consciousness that the only definitional way open to them is by os­ tension. Being unable to provide a nonostensive definition is not, by itself, sufficient reason to pronounce the notion of phenomenal conscious­ ness as theoretically illegitimate, and thereby promote its complete abandonment. The merits or shortcomings of an ostensive definition in re­ vealing the essence of a phenomenon have to be judged on its own ground, in virtue of its success in providing conceptual clarity and theoretical agreement in the relevant discussions. It should be acknowledged, however, that the strategy of revealing essences by means of "pointing" has not delivered any kind of agree­ ment with respect to phenomenal consciousness thus far. 82 The same problem appears even more acutely in thinking about the possibility of zom­ bies. How can you tell a zombie from a non­ zombie, someone who has absent qualia from someone whose qualia are intact? If zombiehood is a possibility, not only could your closest friend tum out to be a zombie, without anyone's knowledge or awareness, so could you, and not know it yourself. Zombiehood brings with it not only the problem of other minds, and thus third­ person skepticism, but first-person skepticism as well. If you, the reader of these lines, suddenly turned into a zombie, no one would notice any difference, and in a significant sense of "notic­ ing," neither would you. Remember that know­ ing, judging, thinking, and being aware of-in a nonphenomenal sense-are all capabilities granted to a zombie, and furthermore, "there is no need to invoke qualia in the explanation of how we ascribe mental states to ourselves [be­ cause a zombie] after all, ascribes himself the same qualia; it's just that he's wrong about it"

Giiven Giizeldere

(Chalmers 1 993; Chalmers 1 996 embraces the consequences of this result under the title "the paradox of phenomenal judgement"). Thus, to the extent that "seemings" of your own phenomenal states are constituted by self­ ascriptive judgments, beliefs, thoughts, memo­ ries, expectations, and so forth about those states (and no doubt there is a significant extent to which such seemings are so constituted), it would be warranted to say that your inner life would continue to seem the same to you, despite the fact that you would cease to have any genuine phe­ nomenal states once you turned into a zombie. Put differently, according to the zombie hypoth­ esis, you could now be "hallucinating" your own phenomenology. You would, ex hypothesi, be confidently judging that nothing changed in your inner life, and be mistaken about it, but you would never be able to find this out. Indeed, for all you know, your present existence on earth could be continuing in alternating phases of hwnanhood versus zombiehood, switching every other minute. Hmm. . . . Coming back to a distinction I introduced at the beginning of this chapter, it is also important to note that the segregationist intuition plays into the hands of epiphenomenalism and the possi­ bility of zombiehood. Characterizing conscious­ ness in essentially noncausal (nonfunctional, nonrepresentational) terms leaves no epistemic hook for making it possible to detect the presence or absence of phenomenal consciousness, even from a first-person perspective. But if we are to accept the possibility that any one of us can be a zombie and not know it, that is, if any one of us can be totally lacking phenomenal consciousness while not being able to find out about it, how can we possibly expect a stalemate over the ontology of phenomenon consciousness to be resolved, while fundamentally relying on ostention for its presence? The stalemate seems unresolvable under the proposed terms. Perhaps, then, there is something fundamentally misleading here, and it is time to start looking for ways of building .an alternative

44

conception of phenomenal consciousness based on the integrationist intuition-not one that elim­ inates phenomenal consciousness but not one that renders it completely inefficacious, or opaque even from the first-person perspective either. Rather, the conception should take the first-per­ son characterization of experience seriously and support the commonsense understanding of phe­ nomenal consciousness. The bottom line of what seems most unaccep­ table here is the fact that under a framework that allows for the possibility of zombies, phenom­ enal consciousness is to be regarded as making no difference, in an epistemically significant sense, even in the first person. That is, a well­ intended effort to promote phenomenal con­ sciousness by conceptually separating it from all causal and representational properties actually yields a position with the opposite theoretical consequence: the demotion of phenomenal con­ sciousness to a ghostly existence. If it is this sort of a property that we talk about when we con­ sider phenomenal consciousness, would we really lose much (anything) by doing away with it?83 And if we are committed to (internal) "pointing" as the only reliable way to verify the existence of phenomenal consciousness, the knowledge of the absence or presence of which is hidden even from the first-person perspective, that is, to the person who has it, should we perhaps not reconsider our very concept of phenomenal consciousness? XX

In Place of a Conclusion

I would like to leave the reader with the two questions I just posed above. But let me also give a brief recapitulation and try to tie some of the loose ends. I started by noting an epistemological asym­ metry in the way one has access to (the facts about) one's own experiences versus those of others. This asymmetry leads us to the notion of perspectivity, something quite unique to (the

45

Approaching Consciousness

study of) consciousness, and to the distinction

tology to include consciousness as a fundamental

between first-person and third-person points of

element could indeed relieve one of the nagging

view. This duality between points of view with

problem of having to bridge mechanism and ex­

respect to accessing facts about experiences also

perience (by emergence, reduction, elimination,

manifests itself in a duality in characterizing con­

and so forth) or vice versa. Others think that the

sciousness, in causal versus phenomenal terms.

epistemological nature of the explanatory gap

Taking these characterizations as mutually ex­

does not warrant ontological conclusions. Al­

clusive, based on the presumption that phenom­

though I cannot go into this debate in any further

enal consciousness is essentially phenomenal and

detail here, I too would like to lend my support to

essentially noncausal, yields what I called the

this latter position. True, in the presence of the

segregationist intuition. Opposing it is the in­ tegrationist intuition, which maintains that phe­

explanatory gap, the link between experience and

nomenal consciousness can only be characterized

but I think that the decision to introduce a new

its physical underpinnings may seem arbitrary,

by means of all causal, functional, or repre­

fundamental element into the ontology, based

sentational elements. Given these two intuitions,

on the explanatory gap, seems equally arbitrary

I briefly argued that the former plays into the

as well. At least I fail to see how the most

hands of the doctrine of epiphenomenalism,

steadfast belief in a thus-expanded new ontol­

less puzzled

which, when combined with considerations from

ogy would leave one

the possibility of absent qualia and zombiehood,

how consciousness relates to its physical under­

about just

leads us into untenable and noncommonsensical

pinnings, hence diminishing the explanatory gap

conceptions of phenomenal consciousness. This is

and explaining away the further-How question.

good evidence, on the other hand, to take

What seems the most promising direction in

the latter seriously and use it as the pretheore­

reapproaching consciousness and pursuing its

tical basis in reexamining our notion of phenom­ enal consciousness. 84

deep-rooted problems in the present era involves rethinking epistemology and conceptual schemes

Another domain where the epistemic element

(as opposed to a priori postulation of new ontol­

of perspectivity figures in is the problem of the

ogy) to yield a cross-fertilization of the first­

explanatory gap and the question of the "hard

person and

problem" of consciousness. There seems to be

would allow theorizing about how causal efficacy

third-person

perspectives,

which

how consciousness feels, and how phe­ what consciousness

an unbridged gap in the explanation of how

figures in

physical embodiment and conscious experience

nomenal quality relates to

are linked. The former is in general given a causal

does.

characterization from a third-person perspective,

In any case, at present it just does not seem as if

the latter a phenomenal characterization in first­

there is a way to settle the dispute decisively about

person terms. It seems that under our existing

the "hard problem" or the consequences of the

conceptual scheme, bolstered by the segregation­ ist intuition, the "hard problem" just does not,

explanatory gap. And given the troublesome . stalemate over the ontological nature of phe­

and cannot, lend itself to a solution.

nomenal consciousness, we seem to be not quite

What is important to note here is that the ex­

near a satisfactory understanding of the phenom­

planatory gap, in the way it is set up, stems from

enon. If anything, the survey of the contemporary

an epistemological issue. The further question

issues and current debates surrounding con­

that remains is whether its persistence is good

sciousness points to a need for a careful re­

enough evidence to yield ontological conclusions.

examination of our pretheoretical intuitions and

Some think yes; introducing an "extra ingredient"

conceptual foundations on which to build better

into the picture and thus augmenting one's on-

accounts of consciousness. It also seems probable

Giiven Giizeldere

that an entirely satisfactory understanding of con­ sciousness will be possible, if at all, only when the constitutive elements of a more comprehensive framework, in which consciousness needs to be theoretically situated, are themselves better un­ derstood. And these elements include nothing less than causality, representation, indexicality, and personhood, and especially the deep-rooted di­ chotomies between mental and physical, and sub­ jective and objective. As such, it is probably reasonable to assume, as Jerry Fodor likes to prognosticate regarding a complete account of rationality, that "no such theory will be available by this time next week." B s This being said, I conclude o n a more positive note. Presently, there is an impressive rising tide of interest in the study of consciousness, and thanks to recent advances in interdisciplinary research, we are now in a better position to penetrate the mysteries of this great intellectual frontier. By integrating methodologies and per­ spectives from psychology, philosophy, neuro­ science, cognitive science, and other disciplines and by keeping a mindful eye on the successes and failures of the past, we should be able to reach a higher vantage point and to see more broadly and more deeply than has ever before been ·possible. These are very exciting times for thinking about consciousness.

46

tellectual history and, more specifically, in the analytic tradition. Obviously, there is a wealth of fascinating is­ sues, questions, and approaches concerning conscious­ ness that lie outside this limited scope. This chapter should not be taken as an attempt to give an exhaustive survey of all aspects of consciousness even within its own scope, and certainly not as embodying the grander ambition of covering all paradigms of the study of consciousness.

2.

The results of this effort were discussed as part of

a workshop

attended

by

anesthesiologists,

neuro­

physiologists, psychologists, and medicolegal experts in Cardiff, United Kingdom, in 1986. Rosen and Lunn (1987) is an outcome of this workshop.

3.

Michenfelder makes the same point in another way,

in concluding, "Thus there are a variety of end points one might choose to answer the question 'When is the brain anesthetized?' and there is no obvious basis for selecting one over another" (Michenfe1der 1988, p. 36). However, cf. Nikolinakos ( 1 994) for an opti­ mistic outlook on the role of consciousness in anes­ thesia

research.

See,

also,

Flohr

( 1 995)

for

an

information-theoretic model of anesthesia where the "threshold of consciousness" is determined in terms of the brain's representational activity. 4. Ifwe go down the phylogenetic ladder-for instance, from humans all the way to amoebae-where are we to cut the line and determine the bounds of the

circle

charmed

to which only those who possess consciousness

can belong? (The metaphor of the "charmed circle" is from Dennett 1987, p.

1 6 1 .)

Chimps, dogs, spiders?

What about infants, fetuses, or comatose patients? On the other hand, if we insist on experience of sensations,

Acknowledgments

itches, and tingles as necessary components of con­

I would like to thank my coeditors, Ned Block and Owen Flanagan, for many helpful sugges­ tions. Special thanks to Fred Dretske, John Perry, Roger Shepard, Brian C. Smith, and again to Owen Flanagan, for teaching me what I needed to know to write this chapter, and for their sup­ port throughout the preparation of this book.

short of requiring something further, such as a con­

sciousness, is there any principled reason for stoping ceptual overlay that makes possible one's situated awareness of one's own place and relations with others (not to talk of the Cartesian

res cogitans), as essential to

the nature of consciousness? The answers to these ques­ tions are all up for theoretical grabs. It is also sociologically interesting to look at patterns in the common sense attributions of various mental abilities to various organisms. In contrast to widely

Notes 1 . The scope of my work has been limited to the philo­ sophical and scientific paradigms rooted in Western in-

dissenting opinions on the attribution of consciousness to others, there does not seem to be such a significant variation in pretheoretic intuitions with regard to attri­ butions of intelligence, or perceptual capabilities.

A

preliminary survey study conducted on approximately

Approaching Consciousness

47

one hundred Stanford students and faculty, based on a

comes about as the result of irritating nervous tissue, is

two-dimensional matrix of mental attributes (e.g., the

just as unaccountable as the appearance of the Djin

ability to perceive, the ability to learn, intelligence,

when Aladdin rubbed his lamp in the story, or as any

1866, 193).

consciousness) versus kinds of organisms (e.g., proto­

other ultimate fact of nature" ( Huxley

zoa, spiders, chimpanzees, humans) seemed to indicate

terestingly, Huxley seems to have removed the reference

a bias in our attributions toward reserving conscious­

to the Djin (as well as Aladdin's lamp) in the later edi­

1 876 edition,

In­

ness most exclusively for ourselves, while being more

tions of this book. For instance, in the

generous with the attributions of other mental abilities.

same passage appears as: "But what consciousness is,

(Giizeldere

1 993).

the

we know not; and how it is that anything so remarkable

This is, it turns out, a very tricky question. On the

as a state of consciousness comes about as the result of

one hand, patients with only cortical brain damage

irritating nervous tissue, is just as unaccountable as any

make a striking contrast with those who further lack a

other ultimate fact of nature" (p.

functional brainstem. For instance, a report on the di­

is no longer around, for we could perhaps have wished

5.

1 88). Too bad the Djin

agnosis of death, prepared by the President's Commis­

from it to tell us if we would ever be able to solve the

sion, makes the following statement:

mind-body problem.

The startling contrast between bodies lacking

all

brain functions and patients with intact brain stems (despite severe neocortical damage) manifests [a tremendous difference with respect to responsive­ ness, and hence the attribute of life]. The former lie with fixed pupils, motionless except for the chest movements produced by their respirators. The latter can not only breathe, metabolize, maintain temper­ ature and blood pressure, and so forth,

on their own,

but also sigh, yawn, track light with their eyes, and react to pain or reflex stimulation.

7.

There is a longer passage in Tyndall's "Scientific

Realism" where he addresses the mind-body problem in length, and concludes, in agreement with McGinn (though in

1868) that it

is "as insoluble, in its modern

form, as it was in the prescientific ages." Because the points Tyndall touches upon are so remarkably close to the contemporary formulations of the issues debated in the literature under the labels "explanatory gap" and the "hard problem" (e.g., the nature of the explanation between mind and body as opposed to other physical phenomena; the prospects for the mind-body problem upon reaching a fully advanced neuroscientific under­

On the other hand, the commission shies away from

standing of the brain; and the status of possible corre­

reaching any conclusion with respect to the absence or

lation-based accounts of consciousness), I quote this

presence of consciousness in patients of either kind: "It is not known which portions of the brain are responsible for cognition and consciousness; what little is known points to substantial interconnection among the brain stem, subcortical structures, and the neocortex" (Presi­

198 1 , 161, 160, respectively).

dent's Commission

quoted in Capron

1988,

pp.

Perhaps it is altogether misleading to think of the presence of consciousness in a binary fashion. It might be necessary to talk about degrees of consciousness, which could allow one to say that normal human beings are "more conscious" than those with brain damage, the brain-damaged patients more than those without a brainstem, and so forth.

6.

Even though McGinn

passage here in its entirety.

(A more detailed discussion

follows in section XIV.) The relation of physics to consciousness being thus invariable, it follows that, given the state of the brain, the corresponding thought or feeling might be inferred: or, given the thought or feeling, the corre­ sponding state of the brain might be inferred. But how inferred? It would be at bottom not a case of logical inference at all, but of empirical association. You may reply, that many of the inferences of sci­ ence are of this character-the inference, for exam­ ple, that an electric current, of a given direction, will deflect a magnetic needle in a definite way. But the cases differ in this, that the passage from the current

(1989)

cites Julian Huxley as

to the needle, if not demonstrable, is conceivable,

the author of this by now very popular, colorful quote

and that we entertain no doubt as to the final me­

(with no source), the credit belongs to T. H. Huxley

chanical solution of the problem. But the passage

(Julian Huxley's grandfather). In full, it reads: "But

from the physics of the brain to the corresponding

what consciousness is, we know not; and how it is that

facts of consciousness is inconceivable as a result of

anything so remarkable as a state of consciousness

mechanics.

Guven Giizeldere

48

Granted that a definite thought, and a definite mo­

the connection becomes clear only when we descend

lecular action in the brain, occur simultaneously; we

to the level of molecular description: till then it ap­

do not possess the intellectual organ, nor apparently

pears as a contingent correlation. In the psycho­

any rudiment of the organ, which would enable us

physical case we have no idea whether there is such a

to pass, by a process of reasoning, from the one

deeper level or what it could be; but even if there is,

to the other. They appear together, but we do not

the possibility that pain might be necessarily con­

know why. Were our minds and senses so expanded,

nected with a brain state at this deeper level does

strengthened, and illuminated, as to enable us to

not permit us to conclude that pain might be di­

see and feel the very molecules of the brain; were

rectly analyzable in physical or even topic-neutral

we capable of following all their motions, all their there be; and were we intimately acquainted with

terms. . . . Even if such a deeper level existed, we might be permanently blocked from a general under­ standing of it. (Nagel 1986, pp. 48-49; my em­

the corresponding states of thOUght and feeling, we

phasis).

groupings, all their electrical discharges, if such

should be as far as ever from the solution of the problem, 'How are these physical processes con­ nected with the facts of consciousness?'. The chasm between the two classes of phenomena would still remain intellectually impassable.

The possibility of a permanent cognitive closure in humans with regard to the understanding of the mind­ body relation is an intriguing idea. But when it comes to taking this possibility as a statement of certainty, as McGinn does, it seems fair to question what warrants

Let the consciousness of love, for example, be asso­

this conviction, especially in the absence of empirically

ciated with a right-handed spiral motion of the mole­

grounded reasons. In particular, there is a curious ten­

cules of the brain, and the consciousness of hate with

sion between McGinn's confidence, on one hand, of his

a left-handed spiral motion. We should then know,

own cognitive ability to

when we love, that the motion is in one direction,

on behalf of all human beings (at present as well as in

assert such

a sweeping verdict

and, when we hate, that the motion is in the other;

the foreseeable future), and the aim of his argument, on

but the 'WHY?' would remain as unanswerable as

the other hand, which ultimately strives to

before. (Tyndall

1 868, pp. 86-87)

It is worth noting that equally dissenting opinions were also available at the time. For instance, Huxley

(1901)

states: "I hold, with the Materialist, that the

human body, like all living bodies, is a machine, all op­ erations of which, sooner or later, be explained on physical principles. I believe that we shall, sooner or later, arrive at a mechanical equivalent of conscious­ ness, just as we have arrived at a mechanical equivalent of heat" (p.

8. For (1995).

191).

a materialist response to Adams, see Lewis

attenuate

confidence in the powers of human cognitive abilities to solve the mind-body problem. Put differently, McGinn wants his readers to simul­ taneously believe, as a result of his largely a priori rea­ soning, both that our cognitive abilities are limited to forever fall short of bringing a solution to the mind­ body problem,

and that they are nonetheless sufficiently

powerful to foresee the exploratory limits of the human mind vis-a-vis the very same issue. By the same kind of reasoning, isn't there equally good reason to think, one wonders, that the opposite claim might rather be true­ that it is more likely (or, at least not less likely) that we will someday come to a satisfactory understanding of

Antecedents of McGinn's view can be found in

the nature of the mind-body relation than it is that

Nagel's work in several places, although Nagel only

we will ever be able to determine how far the human

points out the possibility of McGinn's position with­

understanding will extend? (For a thorough critique

out committing himself to it. For instance, he says:

of McGinn's position, see Flanagan

9.

We cannot directly see a necessary connection, if there is one, between phenomenological pain and a physiologically described brain state any more than we can directly see the necessary connection between increase in temperature and pressure of a gas at a constant volume. In the latter case the necessity of

10.

(1992), chap. 6.)

It is actually not a straightforward matter to give a

precise definition of naturalism, and I will not attempt one here. Roughly speaking, I take "naturalism" to de­ note the view that everything is ' composed of fun­ damental entities recognized by the natural sciences (ontological dimension), and possibly that the accept-

49

Approaching Consciousness

able methods of theorizing about these entities are those

and Italian

commensurable with methods employed in the natural

side the boundaries of Indo-European languages, the

(coscienza) that carries both meanings.

Out­

sciences (methodological dimension). For a compre­

term corresponding to consciousness in, say, Turkish­

hensive overview of naturalistic approaches in epis­

an A1taic language--carries a political, but not reall y

temology and philosophy of science in the twentieth

ethical, connotation, i n addition to the common psy­

century, see Kitcher

chological usage

(1992);

for a recent analysis of the

present status of naturalism, see Stroud

In philosophy mind, naturalism is often regarded as a close ally of two related but not identical views, materi­ alism and physicalism, but there are exceptions. See, for instance, Post

(1987) for a detailed attempt to lay out a

fully naturalist but nonreductive metaphysics that is committed to a monism of entities with a pluralism of irreducible "emergent" properties. See also Chalmers

I I.

1 3.

When you speak of "consciousness," you do not re­ fer to the moral conscious: the very rigor of your methods ensures that you do not leave the strictly scientific domain which belong to you. What you have in mind exclusively is the faculty of perceiving and of reacting to perception, that is to say, the psy­ chological concept which constitutes one of the

sort out the tangles in related terminologies. Here is a

accepted meanings of the word "consciousness. "

quote from the turn of the century, by philosopher and

(Pope Paul VI, addressing a gathering o f scientists

(1904):

for the conference

"Were the use of the term 'consciousness' to be for­

in Rome in

bidden for a season, contemporary thought would be set

1 974, p.

for the wholesome task of discovering more definite terms with which to replace it, and a very considerable amount of convenient mystery would be dissipated. There is no philosophical term at once so popular and so devoid of standard meaning. . . . Consciousness com­ prises everything that is, and indefinitely much more. It is small wonder that the definition of it is little at­ Hence, according to Perry, it is not

(only) the ubiquitous familiarity with consciousness that renders attempts to give it a precise characterization or definition unnecessary; the reason is rather the difficulty of the analytic task involved in doing so. 12. A related term in Latin was

The following quotation, taken from the circles

marily with the social rather than the psychological

about placing a ban on "consciousness" in order to help

282).

vicdan).

sense highlights this distinction quite eloquently:

As it happens, Miller is not the first one to think

tempted" (p.

as opposed to

where the concern with consciousness has to do pri­

(1994); for contemporary Robinson (1993).

William James scholar Ralph Barton Perry

f/Jur,

lates to one based on language families.

(1996) for a somewhat similar view he calls "naturalistic

objections to physicalism, see

or

dence of a semantic taxonomy that systematically re­

dualism." For a thorough examination of the physical­ ist program, compare Poland

(bilinr

In any case, this little linguistic excursion gives no evi­

(1996).

14.

1964;

Brain and Conscious Experience

quoted in Kanellakos and Lukas

i)

This distinction is not uncontested, however. Some,

especially in continental philosophy, think that there is no intransitive sense of consciousness: all conscious­ ness is consciousness of. Jean-Paul Sartre, for example, is a typical representative of this view: "We establish the necessity for consciousness to be consciousness

of

something. In fact it is by means of that of which it is conscious that consciousness distinguishes itself in its own eyes and that it can be self-consciousness; a con­ sciousness which would not be consciousness (of) something would be consciousness (of) nothing. (Sartre

conscientia,

which can

1956,

p.

1 73).

The origins of this kind of an essentially

literally be translated as "knowledge with," which ap­

intentional construal of consciousness goes back to Ed­

peared in English and in French as "conscience."

mund Husserl's work from which Sartre adopted his

"Conscience" also had, and still has, a significant eth­

view, most likely

ical aspect, which is reflected in another

Brady for this pointer.)

OED

entry:

"Internal knowledge, especially of one's own innocence, guilt, deficiencies, etc." (Cf. Baldwin less,

consciousness

and

conscience

1 901).

Nonethe­

have been separate

words with quite distinct meanings in English, at least since the time of Locke. In contrast, there is only one word in Romance languages like French

(conscience)

I S.

Ideas

See Giizeldere

( Husser!

(1996)

1913).

(I thank Ron

for an analysis of how the

creature and state senses of consciousness can be con­ nected by means of the "Introspective Link Principle," yielding various "higher-order monitoring" concep­ tions of consciousness. For other recent attempts to distinguish different senses of consciousness and sort

Guven Giizeldere

out some definitional issues, see Lycan (1987, preface; 1 997b, chap. I), Goldman (1993), and Natsoulas (1 983, 1986). 16. My decision to address problems that bear only on the psychological sense of consciousness should not be taken to imply that the two SUbcategories are not related in interesting ways. In fact, it seems a philosophically significant task to investigate the nature of the relation between the social and the psychological senses of con­ sciousness-is it something more like a genus-species relation, or one of family resemblance, or something completely unique? It can also be questioned whether one can fruitfully give an analysis of one of these halves while eschewing the other. Nonetheless, for the pur­ poses of this chapter, I opt for focusing solely on the psychological sense of consciousness. Even if it may hi: impossible to fully understand the social sense of con­ sciousness without referring to the psychological sense, or vice versa, due to the conceptual disarray surround­ ing the term "consciousness" such an analytic strategy seems essential as a first step. 1 7 . For an exploration of the distinction between epis­ temological and ontological considerations, as well as the question of whether a set of criteria to distinguish the mental in general from the physical can be coher­ ently formulated, see among others Rorty (1 970a, I 970b) and Kim (1 972). 18. The status of the mind-body problem in ancient Greek philosophy is also worth a visit. It is generally argued that there is no single term in ancient Greek that reflects the counterpart of the Cartesian/Lockean con­ ception of consciousness, and that nothing like the con­ temporary debates on the mind-body problem or the problem of consciousness was ever in their horizon. For instance, Matson (1 966) claims that "the Greeks had no mind-body problem" (p. 1 01), and Wilkes (1 995) argues that "[Aristotle) paid absolutely no attention to con­ sciousness per se" (p. 1 22). Similarly, Hamlyn (1 968a) states: "There is an almost total neglect of any problem arising from psycho-physical dualism and the facts of consciousness. Such problems do not seem to arise for him. The reason appears to be that concepts like that of consciousness do not figure in his conceptual scheme at all; they play no part in his analysis of perception, thought, etc. (Nor do they play any significant role in Greek thought in general.) It is this perhaps that gives his definition of the soul itself a certain inadequacy for the modem reader" (p. xiii). See also Kahn (1 966), Harnlyn (1 968b), and Wilkes (1 988) for similar views.

50

However, this view is not uncontested. For instance, Alastair Hannay ( 1 990) suggests that in Greek philoso­ phy one can find, contrary to the skepticism expressed above, something like a distinction between the social and psychological senses of consciousness. According to Hannay, Greeks distinguished between syneidesis (pri­ marily ethical individual or shared knowledge) and synaesthesia (Aristotle's variation of the "unity of ap­ perception"), in much the same way as modem philos­ ophy proceeded in the seventeenth and eighteenth centuries. In alliance, Hardie (1 976) argues against Matson, Kahn, and Hamlyn, and states that "it is . . . paradoxical to suggest that Aristotle was unaware of the mind-body problem" (p. 410). According to Hardie, Aristotle was "the first psychologist, and for him psy­ chology without the conscious psuche would have been Hamlet without the Prince" (p. 405). Ostenfeld (1 987) goes a step further, and claims that both Plato and Aristotle were dealing with the mind-body problem in much the same sense Descartes did and we are. This debate, so far as I can see, is far from resolved at present. 19. The debate about whether consciousness consists in the higher-order awareness of first order mental states is very much alive in the contemporary literature. Among those who defend this view, some take the higher-order representation to be some form of perception (for ex­ ample, Armstrong 1980, Churchland 1988, Lycan 1 997), and others as some form of thought (for example, Rosenthal 1986, 1997; Carruthers 1989, 1 996). For cri­ tiques, see Dretske (1993, 1995) and Shoemaker (1 994). Despite the fact that such higher-order awareness ac­ counts of consciousness have many promising aspects, I have to 'stop short of giving a proper exposition here. (A more detailed treatment of this approach can be found in Giizeldere 1995b.) I will also leave the dis­ cussion of "self-consciousness" (which is sometimes underwritten by such higher-order accounts) out of the scope of this chapter. 20. For Malebranche, although we can have a "clear idea" of our bodies, we cannot, unlike what Descartes believed, have a clear idea of our souls or minds. Put differently, we cannot know our minds through a clear idea; rather we know them "only through consciousness or inner sensation" such as "pain, heat, color, and all other sensible qualities" (Malebranche 1923, Elucida­ tion I I : Knowledge of the Soul, pp. 86-87). As such, Malebranche gives qualitative aspects of the mind a much more central place in his theory, in contrast to the

Approaching Consciousness

51

Cartesian view. See Schmaltz (1 996) for a thorough ac­

James's attitude toward the metaphysics of conscious­

count of Malebranche's philosophy of mind. (I thank

ness is reflected most accurately in Gerald Myers's fol­

Tad Schmaltz for the relevant material and helpful dis­

lowing remark, from his extensive study of James's life

cussion on this issue.)

and thought: "James wanted to hold that in one way

2 1 . Labeling the seemingly opposing views of James of

consciousness does not exist, but that in another way it

1890 and James of 1904 simply a "change of mind" is

does; yet he was never able, even to his own satisfaction,

probably too superficial a conclusion in terms of histor­ ical scholarship, and not quite fair to James either. It is

to define the two ways clearly enough to show that they are consistent rather than contradictory" (Myers 1986,

important to note that James's denouncement in the

p.

latter work is of "consciousness as an entity" rather

( 1 940), Lovejoy ( 1 963), and Reck ( 1972). (I thank Denis

than the reality of "conscious states." Regarding con­ sciousness as an "entity" has close connotations to Cartesian substance dualism. Even though there is no straightforward advancement of such a metaphysical position in the

Principles of Psychology,

64).

For related work, see among others Dewey

Phillips, Imants Baruss, and Eugene Taylor for helpful pointers and discussion on William James's views on consciousness. ) 22. The claim of introspectionism's being well estab­

James's posi­

lished here refers not as much to the soundness of its

tion with respect to the ontology of consciousness is not

methodology and theoretical grounding as to its perva­

entirely clear there. Hence it might be better to charac­

siveness and preeminence in the field of psychology as a

terize his 1904 article as marking merely the abandon­

whole. To see this, one only needs to survey the mono­

ment of consciousness as a nonmaterial entity, not

lithic psychology literature of the few decades roughly

consciousness per se as a subject matter. This inter­

between the end of nineteenth and the beginning of the

pretation is supported by James's own remark that he

twentieth centuries: all major psychology journals are

means to "deny that the word [consciousness] stands for

edited by the protagonists of the introspectionist school,

an entity, but to insist most emphatically that it does

all articles report studies involving introspection as their

stand for a function" (James 1 97 1 , p. 4).

primary method, and so on. Ironically, the same ob­

Even though this much seems quite straightforward,

servation holds of the period that immediately follows

we are by no means faced with an unproblematic ac­

(roughly from late 1 9 1 0s to early 1 960s), except with

count of consciousness. In fact, James never quite works

behaviorism substituted for introspectionism.

out the metaphysical presuppositions and consequences

It would be interesting to pursue the question of

of his view of characterizing consciousness as a func­

whether the ftuctuation in James's life with respect to

tion, as opposed to an entity. Moreover, the ontological

consciousness occurred as a result of, or was inftuenced

tum he takes toward "radical empiricism" at around the

by, the general air of dissatisfaction with the internal

same period as the pUblication of his "Does Conscious­

conflicts of the introspectionist school toward the end of

ness Exist?" complicates matters. It is probably well

its tenure, which led to behaviorism's rapid rise and

warranted to remark that William James never held

takeover of the intellectual landscape. Or was the influ­

a long-standing metaphysical position with respect

ence in the opposite direction? These are aU intriguing

to consciousness void of internal tensions. At a cer­

questions, but unfortunately they lie outside the scope of

tain stage in his life, roughly midway between the

this chapter.

publication of the two above mentioned works, he

23. This is only one (as it happens, also historically the

went so far as to defend the plausibility of the immor­

most significant) use of the word

tality of consciousness in an article titled "Human

ber of different phenomena have passed under the same

introspection.

A num­

Immortality," in the following words: "And when

name. For instance, toward the end of the last century,

finally a brain stops acting altogether, or decays, that

Brentano and Comte argued that introspection, as a

special stream of consciousness which it subserved will

second-order mental act that gathers information about

vanish entirely from the natural world. But the sphere of

first-order sensations, was misconstrued.

being that supplied the consciousness would still be in­

James agreed and proposed a model of introspection as

tact; and in that more real world with which, even whilst

Mill and

retrospection: the examination of one's own mental

here, it was continuous, the consciousness might, in

happenings retrospectively, through the medium of

ways unknown to us, continue still" (James 1 956, pp. 1 7-

memory of the immediate past. (For details, see Lyons

18). Perhaps the historical fact of the matter regarding

1986, chap \ .) A second, separate phenomenon that

Guven Giizeldere

52

made its way to the cognitive psychology literature in

produce my findings, it is not due to some fault in your

the 1970s under the name

apparatus or in the control of your stimulus, but it is due

introspection

was the phe­

nomenon of reasoning about the causes of one's own

to the fact that your introspection is untrained. The at­

behavior, in terms of one's beliefs, desires, motivations,

tack is made upon the observer and not upon the ex­

and so forth. (For a seminal article that piqued most of

perimental setting" (Watson 1 9 1 3 , p. 163).

the initial interest in this literature, see Nisbett and

Interestingly, approximately two hundred years ear­

Wilson 1977.) In any case, my analysis deals with in­

lier, a similar debate had taken place between two em­

trospection only in the former sense.

piricist philosophers, Locke and Berkeley, on almost

24. Edwin Boring notes that no subject left Wilhelm

exactly the same issue. The question was whether there

Wundt's laboratory without having provided 10,000 data points (Boring 1 953, p. 1 72). William James hu­ morously observes that if it had not been for the sus­ tained patience and the inability to get bored of the leaders of introspectionism who came from the Ger­ manic part of the continental Europe, the enterprise of introspectionism could have never endured. "They mean business," James remarks, "not chivalry" (James 1950, pp. 192-1 93). 25. The nature of the relation and the degree of influ­ ence between positivism and behaviorism are not un­ controversial Even though it is generally taken for granted that the two movements enjoyed a genuine ally status, the details of this received view have recently been contested by Laurence Smith. Smith claims, "With their common intellectual background and orientation, behaviorism and logical positivism were naturally dis­ posed to form some sort of alliance. But only after both movements were well under way was there any signif­ icant interaction between them" (Smith 1986, p. 5; cf. the rest of his book for further details). 26. In all fairness I should add that behaviorism did manage to bring in fresh air to psychology of the late nineteenth and early twentieth centuries at a time when an uncomfortable sense of containment within the rigid introspectionist paradigm was rapidly growing. The realization that psychology could employ nonhuman subjects and pursue research without being solely de­ pendent on the linguistic data to be provided by trained introspectionists seemed, rightly, to open up new hori­ zons. This should also explain, in

part,

the rather im­

mediate success and popularity of behaviorism and the symmetrically rapid fall of introspectionism. Unfor- . tunately, as I will detail below, behaviorism turned out to constrict psychology into an even more rigid cast in comparison to its predecessor. 27. John Watson would not miss the chance to put a nail in introspectionism's coffin by alluding to this con­ troversy: "Psychology, as it is generally thought of, has something esoteric in its methods. If you fail to re-

were any "abstract ideas" : ideas that are not of partic­ ular things but of universals-"types" of particular things. In the following quotation, notice that Berkeley rests his challenge of Locke's position on this question on exactly the same grounds that Titchener challenged Kiilpe: personal experience based on introspection. If any man has the faculty of framing in his mind such an idea of a triangle as is here described, it is in vain to pretend to dispute him out of it, nor would

I

go about it. All I desire is that the reader would fully and certainly inform himself whether he has such an idea or no. And this, methinks, can be no hard task for anyone to perform. What more easy than for anyone to look a little into his own thoughts, and there try whether he has, or can attain to have, an idea that shall correspond with the description that is here given of the general idea of a triangle, which is "neither oblique nor rectangle, equilateral, equi­ crural or sca1enon, but all and none of these at once"? ( Berkeley 1977, pp. 1 3- 14) 28. There is of course a third, unmentioned but impor­ tant school of psychology that emerged during the pe­ riod of transition from introspectionism to behaviorism: Gestalt psychology. The fact that a separate account of Gestalt psychology is not being provided here is cer­ tainly not because it is intellectually unworthy of con­ sideration. Quite the opposite, Gestaltists were very keen about the reasons for introspectionism's failure, and they brought a fresh new perspective on the basis of which a large number of facts in the psychophysics of perception could be fruitfully reinterpreted. Nonethe­ less, Gestalt psychology shared many of the same onto­ logical assumptions with respect to consciousness and the role of phenomenology in studying the mind with introspectionism. As a result, as far as the history of consciousness in psychology research is concerned, it does not constitute the sort of sharp contrast that be­ haviorism provides. Hence, the brief treatment. Finally, the emerging clinical wing of psychology, the psychoanalytic school, also had its disagreements with

Approaching Consciousness

53

introspectionism, and it constituted the third distinct

In contrast, the ontological behaviorists were in favor

angle of attack alongside with behaviorism and Gestalt

of doing away with consciousness, or any aspect of the

psychology. Unfortunately, I cannot go into a detailed

mind, by

analysis concerning these three movements here. For a

example, Lashley ( 1 923) maintained the following the­

itkntifying it with some piece of behavior.

For

well-documented historical account of introspectionism

sis: "The conception of consciousness here advanced is,

and the debates and movements that surrounded it, cf.

then, that of a complex integration and succession of

the section on "Modem Experimental Psychology" in

bodily activities which are closely related to or involve

Boring ( 1 929), as well as chapters I and 2 of Lyons

the verbal and gestural mechanisms and hence most

(1 986).

frequently come to social expression" (p. 34 1 ).

29. So much so that it is very rare, even today, to come across

"consciousness"

or

"introspection"

in

any

psychology or cognitive science textbook, or even psy­

Although the assumptions of these three schools are, by and large, logically independent of one another, Watson ( 1 9 1 3 ,

1 970), an indoctrinated behaviorist,

chology dictionaries. See, for example, Corsini ( 1 984)

seems to have believed in all of them, arguing that the

or Stillings et aJ. (1 987), which contain no entries for

time was ripe for psychology to discard all reference to

"consciousness," "awareness," or "introspection."

consciousness. It is no doubt that a Watsonian universe

30. Naturally, there were internal disagreements, and thus different schools, within Behaviorism, and not each brand of the doctrine was as hardheaded. Most notably, the

analytical

(logical) behaviorists (who were mostly

philosophers, e.g., Hempel 1 949) were interested in an­ alyzing meanings ofmental terms in a purely behavioral vocabulary, whereas the

methodological

behaviorists

(who were mostly psychologists) wanted merely to re­ strict their research to the study of publicly observable behavior without having to attempt any conceptual analysis or even deny the reality of the publicly un­ observable mental phenomena. For example, according to Edwin Holt ( 1 9 1 4), "the true criterion of consciousness is not introspection, but specific responsiveness" (p. 206). Since making be­ havior the

criterion

of consciousness is not quite the

same as itkntifying the two, consciousness thus becomes "externalized" by means of a publicly observable mea­ sure, but the metaphysical question of identity is left open. As such, the two phenomena could be said to be coexistent, as Holt (19 1 5) acknowledges in a later work: "When one is conscious of a thing, one's movements are readjusted to it, and to precisely those features of it of which one is conscious. The two domains are cotermi­ nous" (p. 172). Edward Tolman's position in his "A Behaviorist's Definition of Consciousness" ( 1 927) is also similar to Holt's in stopping short of advancing a metaphysical

claim: "Whenever an organism at a given

moment of stimulation shifts and there from being ready to respond in some relatively differentiated way to being ready to respond in some relatively more dif­

IS

would make life much easier for philosophers and psy­ chologists. It would, for instance, remove the epistemic duality in the study of consciousness by collapsing the distinction between the first-person and third-person perspectives. Furthermore, by making consciousness ultimately an operationalized parameter in the domain of behavior, it would allow a set of behavioral criteria to settle questions about who or what possesses creature consciousness. But, as is evident from the history of psychology, life is never easy in the domain of mind. Questions about consciousness remained a nagging issue during behaviorism's tenure, and they eventually led its prominent figures like B. F. Skinner to not only acknowledge the existence of the phenomenon, but also adopt a conciliatory position in his later works. For ex­ ample, Skinner ( 1 974), after stating that the common conception of behaviorism as a school of thOUght that "ignored consciousness, feelings, and states of mind" was all wrong, concedes that the "early behaviorists wasted a good deal of time, and confused an important central issue, by attacking the introspective study of mental life" (pp. 3-5). Undoubtedly, consciousness was not the only factor that brought the demise of behaviorism. A different line of attack, for example, came from the quarters of new­ born modem linguistics on the issue of explanation of verbal behavior. In particular, Noam Chomsky's famous review ( 1 959) of Skinner ( 1 957) is a milestone that shook behaviorism (in psychology) in its foundations. For an influential critique of logical behaviorism (in philosophy), see Putnam ( 1963).

consciousness" (p. 435). (See

3 1 . Of course, there were a few exceptions who spoke

also Tolman 1967 and note 43 for his position with re­

up while the reign of behaviorism was still tight and

spect to the study of "raw feels" in psychology.)

proved to be visionaries. Worth mentioning here is a

ferentiated way, there

54

Giiven Giizeldere

lengthy discussion Miller gave on consciousness in his

as the "Chinese room argument," Searle argued that no

excellent survey of psychology as early as 1962. It is

amount of information processing could alone provide

possible to recount even earlier attempts to break the

a system with original (as opposed to derivative, as­

silence, and directly or indirectly talk about conscious­

signed, etc.) semantics (Searle 1980). Dreyfus's critique

ness, especially in the fields of attention, learning, and

never focused on consciousness per se, but Searle, in a

cybernetics. Cf. Hebb (1 949), Abramson (1951 -55),

newer work, deals exclusively with the problem of con­

Hilgard ( 1956), and Broadbent ( 1958). See also Hil­

sciousness in cognitive science, and in general compu­

gard's remarks on this issue in his lucid survey, "Con­

tational paradigms (Searle 1 992). In contrast, a rival

sciousness in Contemporary Psychology" (1 980). For

account of consciousness built entirely on computa­

Consciousness

an account of the "cognitive revolution" in psychology,

tional ideas can be found in Jackendoff's

see Baars (1 985) and Hilgard (1 987), chap. 7.

and the Computational Mind (1 987).

32. Years later, Alan Baddeley ( 1993), a prominent

of the information-processing accounts of conscious­

psychologist who has devoted his career to the inves­ tigation of memory, validates Mandler's insight in the following words:

"I am rather surprised to find myself

writing about consciousness . . . . There are very good reasons why the study of consciousness has been dis­ cretely ignored by cognitive psychology during its early years of development. . . . Why, then, have I changed my mind? In my own case, the strongest reason has corne

ness, see Donald Hebb's

For a predecessor

The Organization of Behavior

( 1949), a work that came out of the behaviorist era but anticipated what was ahead with foresight: Hebb argues to identify consciousness "theoretically with a certain degree of complexity of phase sequence in which both central and sensory facilitations merge, the central act­ ing to reinforce now one class of sensory stimulations, now another" (p. 1 45).

from the pressure of empirical evidence; I am an ex­

34. Note, however, that a curious passage in Des­

perimental psychologist who uses empirical data to

cartes's

drive theory, and it has become increasingly difficult to

commitment to something very much like Freud's un­

Principles of Philosophy

suggests a theoretical

have a model of memory that is at all complete, without

conscious, which does not sit squarely with his explicit

directly or indirectly including assumptions about con­

commitment to the transparency of the mind:

sciousness" (pp. 1 1 - 1 3). (Note, by the way, that Bad­ deley

I 99O-in many respects a very thorough book on

memory-contains no references to consciousness.) 33. The information-processing models of conscious­ ness, although not the only game in town, are still very much alive today, in psychology as well as in philoso­ phy. For instance, Dennett's central claim in his most recent

Consciousness Explained

is

that

"conscious

human minds are more-or-Iess serial virtual machines implemented-inefficiently-on the parallel hardware that evolution has provided for us" (Dennett 1 99 1 , p . 2 1 8). See also Hofstadfer (1 979), Hamad ( 1 982), and Sornmerhof (1 990,

1 996) for theorizing about con­

sciousness in computational and systems-analysis terms.

The strange aversions of certain people that make them unable to bear the smell of roses, the presence of a cat, or the like, can readily be recognized as re­ sulting simply from their having been greatly upset by some such object in the early years of their life. . . . And the smell of roses may have caused severe headache in a child when he was still in the cradle, or a cat may have terrified him without anyone notic­ ing and without any memory of it remaining after­ wards; and yet the idea of an aversion he then felt for the roses or for the cat will remain imprinted on his brain till the end of his life (Descartes 1 992), p. 195:

Principles of Philosophy, pI. I, §9, AT, 429)

However, information-processing models of the mind

Unfortunately, I have not been able to find any fur­

(and, a fortiori, of consciousness in particular) have not

ther elaboration of this idea in Descartes's writings,

always been everyone's favorite. For example Hubert

which would surely be relevant in better understanding

Dreyfus, in his well-known critique of the research pro­

the nature of what seems to be an apparent theoretical

gram and methodology of artificial intelligence, brought

tension.

the whole information-processing approach under se­

35. The controversy over the status of unconscious

vere criticism (Dreyfus 1979, 1992, esp. chap. 4, "The

mental states is multifaceted. Another staunch critic

Psychological Assumption"). Another line of attack

of the unconscious, though for reasons different from

was developed from neighboring quarters by philos­

Titcheners' (that have to do with his stance against

opher John Searle. In an essay that later became known

panpsychic

views

of consciousness),

was

William

Approaching Consciousness

James (1950a). He fretfully remarks: "The distinction . . . between the unconscious and the conscious being of the mental state . is the sovereign means for be­ lieving what one likes in psychology, and of turning what might become a science into a twnbling-ground for whimsies" (p. 1 63). As discussed in section XI, a different line of objection is also raised, this time in a Cartesian spirit, by Searle (1992) and Strawson (1994). Note also that in Titchener, talk about the uncon­ scious has switched to talk about the subconscious, but there is enough reason to think that nothing theoret­ ically significant hangs on this implicit substitution. This terminological variation stems from the fact that Freud and his contemporary, Pierre Janet, had an initial dis­ agreement that left them with two different terms (un­ conscious and subconscious), and each one adopted and perpetually owned his own term with a vengeance. But this was more a result of personal quarrels between the two personalities than a genuine theoretical dissonance on the nature and structure of that which is not con­ scious. And so far as I can tell, there is no evidence that Titchener's use of Janet's term, subconscious, rather than Freud's unconscious is the result of a "conscious decision" and a theoretical commitment. For an illumi­ nating account of the relation between Freud and Janet ' see Perry and Laurence (1984). .

.

36. For an interesting discussion of the question of whether the Freudian unconscious is a "theoretical construct" on a par with scientific theoretical entities' see Oilman (1972). 37. Note that over time, Freud grew dissatisfied with his tripartite structure and eventually introduced the new elements of the id, the ego, and the superego into the picture: In the further course of psycho-analytic work, how­ ever, these distinctions (i.e., conscious, preconscious, and unconscious) have proved to be inadequate, and for practical purposes, insufficient. This has been clear in more ways than one; but the decisive in­ stance is as follows. We have formed the idea that in each individual there is a coherent organization of mental processes; and we call this ego. (Freud 1962, p. 7) Later Freud (1964) gives a schematic depiction of the structure of consciousness, with the id, the ego, and the superego being "superimposed" on the classical tri-

55

partite division of the conscious, the preconscious, and the unconscious. 38. Cf. John Kihlstrom's work for a cognitivist over­ view of the various forms of the unconscious (e.g., Kihlstrom 1 984, 1 987). 39. For a thoughtful discussion of the theoretical issues involved, see Reingold and Merilde (1990). 40. Another paradigm in contemporary psychology that makes use of the conscious-unconscious distinction is that of implicit learning and implicit memory, as well as implicit perception. The focus of interest in this paradigm is on measuring the amount of learning and memory possible in the absence of subjects' awareness of the stimuli presented to them. A certain branch of this work became sensationalized in the media under the ti­ tle "subliminal perception" in the 1 970s. For a thor­ ough and sympathetic account of the nature of this phenomenon, as well as the history of related research, see Dixon's Subliminal Perception: The Nature of a Controversy (1971) and his later Preconscious Proces­ sing (1981). For possibly the most influential recent work in this area (especially in masking studies), see Marcel (1983a, 1 983b). Naturally, there are also skeptics. For instance, Erik­ sen stated quite early on, "At present there is no convincing evidence that the hwnan organism can dis­ criminate or differentially respond to external stimuli that are at an intensity level too low to elicit dis­ criminated verbal report. In other words, a verbal report is as sensitive an indicator of perception as any other response that has been studied" (Eriksen 1960, p. 298). More recently, Holender (1 986) presented a negative and rather controversial statement on subliminal per­ ception, which also included a comprehensive survey of the field. For a collection of contemporary position papers in this paradigm, see the special issue of Mind and Language on "Approaches to Consciousness and Intention" (Spring 1990). 4 1 . Searle also has an explanation to offer regarding the motivations underlying the sort of separationist view that Fodor promotes with respect to consciousness and intentionality: There has been in recent decades a fairly systematic effort to separate consciousness from intentionality. The connection between the two is being gradually lost, not only in cognitive science, but in linguistics and philosophy as well. I think the underlying-and perhaps unconscious-motivation for this urge to

Giiven Giizeldere

56

in the visual field allows

separate intentionality from consciousness is that we

And nothing

do not know how to explain consciousness, and we

that it is seen by an eye.

would like to get a theory of the mind that will not

5.6331

be discredited by the fact that it lacks a theory of

like this:

you to infer

For the form of the visual field is surely not

consciousness. (Searle 1 992, p. 1 53) Perhaps a piece of careful Freudian psychoanalysis would resolve this issue for good. Lacking such ex­ pertise, I choose to leave the question open. 42. Panpsychism is a deep-rooted idea that can prob­ ably be traced, in one form or another, back to Thales

EYO-O

and other ancient Greek philosophers. Nagel (1 988) presents a contemporary discussion of panpsychism, characterizing it as the view that "the basic physical constituents of the universe have mental properties, whether or not they are parts of living organisms" (p. 1 8 1). Panpsychism was quite popular as a metaphy­ sical doctrine among the psychologists (in particular, the psychophysicists) of the nineteenth century, includ­ ing such prominent figures as Gustav Fechner and Her­ mann Lotze. William James, in contrast, was never sympathetic to this view; chapter VI of James (1 950a) contains a cogent critique of panpsychism (under the title "Mind-Stuff Theory"). For recent discussions of panpsychism in the context of ongoing consciousness debates, see Seager (1 995) and Hut and Shepard (1 996). 43. The origins of the term

raw feels goes

back, so far

as I can trace, to the work of behaviorist psychologist Edward Tolman. In outlining what falls outside the scope of "scientific psychology," Tolman (1 967) char­ acterizes raw feels (from his opponents' perspective) as follows: "Sensations, says the orthodox mentalist, are more than discrirninanda-expectations, whether indicated by verbal introspection or by discrimination­ box experiments. They are in addition immediate mental givens, 'raw feels'. They are unique subjective suffusions in the mind" (Tolman 1967, pp. 250-25 1). But it is probably Herbert Feigl (1 967) who is responsi­ ble for the introduction and wide acceptance of rawfeels In a short passage in the

Trae/a/us,

Wittgenstein

makes a similar point regarding the self (or subject­ hood):

5.633

Keith Gunderson (1 970) also discusses this issue under the title, "The Investigational Asymmetries Problem" and makes the similar point that "just as the eye does not, cannot, see itself in its own visual field, so too, the self will never, in its inventory-taking of the world, find itself in the world in the manner in which it finds other people and things" (p. 1 27). Again, the point raised is well taken for consciousness so far as one's own selfuood is involved in it, but it is not obvious just how it generalizes into a difficulty (much less an impossibility) with the study of consciousness in general by (other) conscious beings. 45. This observation is intended to be ontologically neutral. The emphasis here is on the "mode of access" part and not on the "facts" themselves. In particular, it does not entail the existence of a special class of facts, "first-person facts," on the basis of an assumption of ontological difference between facts of one's own con­ sciousness and those of others. 46. Of course, technically speaking, it is not possible to digest food in someone else's stomach either, but "digestive epistemology" just does not seem to be a fashionable topic these days. 47. Perhaps the most succinct expression of this dif­ ference is given in Sydney Shoemaker's question: "If

in the philosophical terminology.

44.

(Wittgenstein 1 974, p. 57)

what I want when I drink fine wine is information about its chemical properties, why don't I just read the label?" (quoted in Dennett 1 99 1 , p. 383). There is a ready­ made answer to this question: It indeed

Where in the world is a metaphysical subject

is information

about the chemical properties of the wine that a con­

to be found?

noisseur is interested in, but only if that information can

You will say that this is exactly like the case of the

be accessed in a certain sensory modality-gustatorily,

eye and the visual field. But really you do eye.

not see the

not visually. Put in Fregean (1 892) terms, reading the label and sipping the wine would provide access to the

Approaching Consciousness

57

same referent via different "modes of presentation."

sciousness, continental Europe was certainly more of a

Note, however, that Shoemaker's question remains not

center than James's Cambridge. See , for instance,

fully addressed until this answer is supplemented by a

Brentano's chapter, "On the Unity of Consciousness, " in

satisfactory account of something akin to modes of

Psychology from an Empirical Standpoint (Brentano Principles of Psychology ( l 890). A more detailed analysis of this

presentation regarding qualia. 48. Of course, not all publicly observable properties of an experience are intrinsic. There are often a great many extrinsic properties that detennine what the experience is about that are equally accessible to the experiencer and the observer. Some think that all important prop­

his

1 874), a work that slightly precedes James's

sort was later given by Husserl. See, among others, his

Ideas: General Introduction to Pure Phenomenology ( 1 9 1 3) and The Phenomenology of Internal Time-Con­ sciousness (I 928}. A more recent attempt along these

erties of experiences, including those that detennine an

lines, which comes from the analytic tradition, can be

experience's phenomenal character, are extrinsic. See

found in Searle (1 992, chap. 6).

Dretske (1995).

5 1 . The idea of finding systematical bridging relations

49. There is a spectrum of positions with respect to

between the "mental" and the "physical" in order to

these dichotomies that yield deep differences in the

establish explanatory hooks on consciousness was also

metaphysics of consciousness. Let me mention a few

the driving factor behind the emergence of psychophy­

exemplary positions. Nagel (1 979, 1986) takes the dis­

sics as a research program in the nineteenth century.

tinction between subjective and objective points of view

This is exemplified in, for instance, Gustav Fechner's

as fundamental to important philosophical problems,

work where he sought ways offormalizing a logarithmic

such as personal identity, free will, and the mind-body

relation between the intensity of physical stimuli (mea­

problem. Velmans (1991) posits that first-person and

sured in "physical units"), and the magnitude of felt

third-person accounts of consciousness are comple­

sensory experience (measured in "psychological units")

mentary, but not reducible, to one another. In contrast,

as reported by the subject. (Cf. Fechner 1966, see also

Dretske (1 995) argues that as a "result of thinking about

Boring 1 942 and Hilgard 1 987, chap. 4.) The same idea

the mind in naturalistic terms, subjectivity becomes

was also operative in Gestalt psychology in the hy­

part of the objective order. For materialists, this is as it

pothesizing of a relation of isomorphism between

should be" (p. 65). This is in accord with an earlier

"the structural characteristics of brain processes and

statement by Lashley ( 1 923), who claims that "the sub­

of related phenomenal events" (Kohler 1 97 1 , p. 8 1 ; see

jective and objective descriptions are not descriptions

also Kohler 1980 and Boring 1 929, chap. 22).

from two essentially different points of view, or de­

52. John Searle presents a major attack on behaviorist

scriptions of two different aspects, but simply descrip­

theories of consciousness in his Rediscovery of the

tions of the same thing with different degrees of

arguing for what he dubs "the principle of the in­

accuracy and detail" (p. 338). Papineau (1 993) argues

dependence of consciousness and behavior." His thesis

that it is a mistake to think that first-person and third­

is that "the capacity of the brain to cause consciousness

Mind,

person thoughts refer to different entities on the basis of

is conceptually distinct from its capacity to cause motor

an epistemic difference, and calls it the "antipathetic

behavior." One consequence he draws from this thesis is

fallacy." Finally, Perry (1 979, 1 993) examines the status

that "a system could have consciousness without be­

of the first-person in relation to the role of indexicality

havior." Under certain qualifications,

in mind and language. Also, for two alternative ap­

plausible. However, Searle goes further to claim that

proaches, see Hut and Shepard (1 996) for a prioritiza­

"ontologically speaking, behavior, functional role, and causal relations are irrelevant to the existence of con­ scious mental phenomena" (Searle 1992, p. 69; emphasis

tion of the first-person over the third-person, and Smith (forthcoming) on how to get to the third-person from

I find this view

the first-person. For a scrutiny of the metaphysical

in original). This further and more encompassing claim

foundations of such dichotomies as objectivity versus

does not directly follow from the weaker one. More­

subjectivity, see among others Goodman (1 978), Rorty

over, it opens up a path the logical conclusion of which

(l 979), Putnam (1981), and Smith ( l996).

may turn into "epiphenomenalism": the view that con­

50. For a lucid analysis of James's account of the

sciousness plays no causal role itself, though it may be

structure of "fringe consciousness," see Mangan (l 993).

the causal effect of other phenomena. I find Searle's

Regarding works on the structure of phenomenal con-

second thesis untenable and its consequence very

Giiven Giizeldere

undesirable. I will come back to this issue in my dis­ cussion of epiphenomenalism and the possibility of zombies. 53. There are also scientifically documented cases where the relation between consciousness and externally ob­ servable behavior breaks down. I briefly discussed the phenomenon of "gaining consciousness' while under general anesthesia as one example. There are also several diseases of the nervous system that fall in the category of demyelinating neuropathies (diseases that result from loss of conduction of nerve impulses due to the lack of formation of myelin, a fatty substance es­ sential to the insulation of axons in neurons), which re­ sult in the patient's gradual loss of reflexes and muscular strength, and hence behavior, while not resulting in substantial sensory changes (e.g., the Guillain-Barre syndrome). The ultimate state of such a patient involves very little outward behavior with no loss of conscious­ ness, defying the behaviorist dogma. (Cf. Reeves 1 98 1 .) A moving account of a somewhat related nervous sys­ tem disorder, encephalitis lethargica (commonly known as the sleeping sickness) was given in Oliver Sacks's popular book, Awakenings. Sacks's description of the victims of encephalitis lethargica is worth quoting at least for its literary value: Patients who suffered but survived an extremely severe somnolent/insomniac attack of [encephalitis lethargica] often failed to recover their original aliveness. They would be conscious and aware-yet not fully awake; they would sit motionless and speechless all day in their chairs, totally lacking en­ ergy, impetus, initiative, motive, appetite, affect or desire; they registered what went on about them with profound indifference. They neither conveyed nor felt the feeling of life; they were as insubstantial as ghosts, and as passive as zombies: von Economo compared them to extinct volcanoes. [However, . . . ] one thing, and one alone, was (usually) spared amid the ravages of this otherwise engulfing disease: the "higher faculties"-intelligence, imagination, judgement, and humour. These were exempted-for better or worse. Thus these patients, some of whom had been thrust into the remotest or strangest ex­ tremities of human possibility, experienced their states with unsparing perspicacity, and retained the power to remember, to compare, to dissect, and to testify. Their fate, so to speak, was to become unique witnesses to a unique catastrophe. (Sacks 1974, pp. 9, 12)

58

54. Let me also mention that some philosophers think that the only legitimate sense of consciousness is phe­ nomenal consciousness (e.g., Searle 1992, Flanagan 1992), while others believe only in access consciousness (e.g., Dennett 1991), and still others believe in phenom­ enal consciousness but try to account for it in causal, functional, or representational (i.e. "access-related") terms (e.g., Van Gulick 1988, Tye 1 992, Dretske 1995). 55. Some of these ideas are briefly explored in Gii­ zeldere and Aydede (forthcoming). 56. I will discuss absent qualia, an offspring of philo­ sophical imagination that was conceived as a result of taking the "hard problem" (perhaps too) seriously, in discussing zombies below. Regarding the explanatory gap, here is a surprisingly contemporary expression of the problem from the nineteenth-century philosopher­ psychologist Charles Mercier: "The change of con­ sciousness never takes place without the change in the brain; the change in the brain never . . . without the change in consciousness. But why the two occur to­ gether, or what the link is which connects them, we do not know, and most authorities believe that we never shall and can never know" (Mercier 1 888, p. I I). Note that the point Mercier is raising is very similar to the one expressed by John Tyndall in section I. A similar but more recent statement, though in a more determinedly pessimist tone, can be found in a rather unlikely source. Here is Freud on the "hard problem": We know two things concerning what we call our psyche or mental life: firstly, its bodily organ and scene of action, the brain (or nervous system), and secondly, our acts of consciousness, which are im­ mediate data and cannot be more fully explained by any kind of description. Everything that lies between these two terminal points is unknown to us and, so far as we are aware, there is no direct relation be­ tween them. If it existed, it would at the most afford an exact localization of the processes of conscious­ ness and would give us no help toward under­ standing them. (Freud 1 949, pp. 1 3-14) In contemporary philosophy of mind, Nagel's for­ mulation of this problem has been most influential. The difficulties Nagel raises with respect to "bridging the explanatory gap" between things physiological and things phenomenal are also reflected in Kripke's attack against identity theory. Even though the latter follows a different path, using tools from philosophy of lan­ guage, they arrive at very similar conclusions. (Cf.

Approaching Consciousness

59

Nagel 1974, 1979, 1986, and Kripke 1980.) The follow­

59. Searle says: "For any artefact that we might build

ing quotation eloquently summarizes Nagel's (and pre­

which had mental states equivalent to human mental

sumably, Kripke's) position:

states, the implementation of a computer program

We cannot directly see a necessary connection, if there is one, between phenomenological pain and a physiologically described brain state any more than we can directly see the necessary connection between increase in temperature and pressure of a gas at a constant volume. In the latter case the necessity of the connection becomes clear only when we descend to the level of molecular description: till then it ap­ pears as a contingent correlation. In the psycho­ physical case we have no idea whether there is such a deeper level or what it could be; but even if there is, the possibility that pain might be necessarily con­ nected with a brain state at this deeper level does not permit us to conclude that pain might be directly analyzable in physical or even topic-neutral terms. (Nagel 1986, pp. 48-49) 57. Searle's position is not as

would not by itself be sufficient. Rather, the artefact would have to have powers equivalent to the powers of the human brain" (Searle 1 984, p. 41). Searle's argument is against functionalist accounts of consciousness. Even though he should not be taken to commit himself to a single specific underlying substance (i.e., neuronal structures), he nonetheless seems to think that consciousness is not medium-independent, at least so far as the causal powers of the medium go. This view seems to be based on the implicit assumption that there are causal powers that cannot be captured by functional organization, but it unfortunately leaves the central no­ tion of causal power unexplicated.

60. Here is a noteworthy historical fact: Over a hundred years before Crick and Koch presented their findings on the 40-70 Hz phenomena, in 1 879, Payton Spence pub­ lished an essay in which he argued, on purely metaphy­

straightforward as

sical grounds, that the basic form of consciousness

Nagel's, however. Although Searle talks about sub­

consists of a constant alteration of conscious and un­

jectivity as an irreducible ontological property unique to

conscious states. But the alteration is so rapid that the

consciousness, he also maintains the following position,

subject never becomes aware of the discrete nature of

in a somewhat puzzling way, in the same book: "Con­

her consciousness; she is under the illusion of having

sciousness is, thus, a biological feature of certain or­ ganisms, in

exactly the same sense

a continuous stream.

Spence then speculated that

of 'biological' in

there must be an underlying mechanism in the brain

which photosynthesis, mitosis, digestion, and repro­

that is responsiole for this alteration-something like

duction are biological features of organisms . . . . One of

a very rapid oscillation of neural tissue. In his own

the main aims of this book is to try to remove that

words:

obstacle, to bring consciousness back into the subject matter of science as a biological phenomenon

other" (Searle

like any

1992, pp. 93, 95; emphasis added).

58. As Wittgenstein somewhat sarcastically remarks: The feeling of an unbridgeable gulf between con­

The simplest form of consciousness, or mental life, must consist in an alteration of a state of conscious­ ness with a state of unconsciousness-a regular rhythmical

revelation

of the

Affi rmation,

con­

sciousness, by its Negation, unconsciousness, and

sciousness and brain-process: how does it come

vice versa. . . .

about that this does not come into the consid­

present, to call it a pulsation, or an undulation in the

Perhaps it would be safer, for the

erations of our ordinary life? This idea of a differ­

brain, or a vibration of the molecules of the brain,

ence in kind is accompanied by slight giddiness. . . .

paralleled in consciousness. This pulsation or vi­

When does this feeling occur in the present case? It is

bration is, of course, very rapid; otherwise, we

when I, for example, tum my attention in a particu­

would not have to infer its existence, but would

lar way on to my own consciousness, and, aston­

know it by perceiving the alterations of one state

ished, say to myself: TInS is supposed to be produced

with another. (Spence 1 879, p. 345)

by a process in the brain!-as it were clutching my forehead. But what can it mean to speak of "turning my attention on to my own consciousness"? This is surely the queerest thing there could be! (Wittgen­ stein 1958, §4 1 2, p. l 24e)

The interesting part comes when

M. M. Garver, a

neurophysiologist of the same era, finds the idea plau­ sible and follows up on it on experimental grounds. In particular, he investigates the neural basis of voluntary action (often associated with or regarded as an aspect of

Guven Giizeldere

60

consciousness in those times, by psychologists including William James) and publishes his results in the

can Journal of Science

Ameri­

in 1 880. According to Garver,

of blindsight against the background of Block's access versus phenomenal consciousness distinction in Gii­ zeldere (1 995e).

mental activity is subserved by a cerebral oscillatory

63. The precursors to Dennett's

mechanism with a frequency range of 36-60 Hz. Garver

metaphor can be found in the writings of Gilbert Ryle

Cartesian Theater

hypothesizes that the change in the frequency of the

and U. T. Place. Ryle ( 1 949) characterized and criti­

neural oscillations correlates with minimum and max­

cized the Cartesian notion of the mind as "a secondary

imum levels of mentation, which results in voluntary

theater in which the episodes enacted enjoy the sup­

action. Garver formulates his hypothesis as follows:

posed status of 'the mental' " (p. 1 58). Similarly, Place

"The cerebral portion of the nervous system is con­

( 1 956) called it a mistake to suppose that "when the

tinually varying in its activity, waxing and waning

subject describes his experience, when he describes how

between certain limits, periods of maximum activity

things look, sound, smell, taste, or feel to him, he is de­

following periods of minimum activity at the rate of 36

scribing the literal properties of objects and events on a

to 60 times per second" (Garver 1 880, p. 1 90).

peculiar sort of internal cinema or television screen,

At the end of his article, Garver claims that this

usually referred to in the modem psychological liter­

pattern can be extended to accommodate Spence's

ature as the 'phenomenal field' " (p. 1 07). (The view that

hypothesis of alternating states of consciousness and

Ryle and Place criticize also constitutes a particular

suggests that the lower and upper limits of the oscil­

family of

lation frequency can be taken as the correlates of con­

quite popular at the time.)

sciousness and unconsciousness, respectively.

sense-data

theories of perception that were

Interestingly, while the metaphor of mind as an inner

While Spence and Garver cannot perhaps be said to

theater never occurs explicitly in the writings of Des­

be in pursuit of a solution to the binding problem, and

cartes (so far as I could tell), it can be found in a vivid

thus have anticipated the Crick and Koch hypothesis

passage in Hume (in his discussion of personal identity):

over a hundred years ahead of its time, I find the sim­

"The mind is a kind of theatre, where several percep­

ilarity in the basic idea of seeking a neural oscillatory

tions successively make their appearance; pass, re-pass,

basis for consciousness fascinating. Except a passing

glide away, and mingle in an infinite variety of postures

remark by William James in his discussion of the con­

and situations." But Hume is also careful not to en­

tinuity of consciousness (James 1950, p. 220, footnote),

dorse, in virtue of using this metaphor, the kind of on­

Spence and Garver's work seems to have gone, so far as

tological conclusion Descartes is criticized as holding:

I could trace, unnoticed to date.

"The comparison of the theatre must not mislead us.

6 1 . The idea of localized functions in the brain precedes

They are the successive perceptions only, that constitute

James's work, and goes back at least to the once-too­

the mind; nor have we the most distant notion of the

popular phrenology of Franz Gall in early 1 800s (see,

place, where these scenes are represented, or of the

for instance, Ackerknecht and Vallois 1 956). Jean Bap­

materials, of which it is compos'd." (Hume 1955, Book

tiste Bouillaud, in 1 825, proposed a hemispheric asym­

I, IV:V, 85).

metry in brain function, but it was Carl Wernicke and

64.

Paul Broca who made the greatest contribution to the

do with lateralization of brain function in light of the

Another issue related to the Where question has to

idea of modularity in the brain. Wernicke hypothesized

"split-brain" research of the last few decades. The per­

that two particular areas in the left hemisphere of the

formance of commissurotomy on humans (proposed

brain (roughly, the left frontal lobe and the posterior

and initiated by surgeon Joseph Bogen in 1 960) to con­

cortex), which later became known as Broca's and

trol interhemispheric spread of epilepsy produced a

Wernicke's areas, were responsible for language pro­

number of patients in whom individual investigation

duction and language understanding, respectively. For

of specialized hemispheric capabilities became possible.

an elemental neuropsychological account of aphasias

The research on such patients (initially pursued by psy­

that result from damage to these areas and other related

chobiologist Roger Sperry and his collaborators) re­

matters, see Kolb and Whishaw (1 990).

vealed a number of interesting facts about hemispheric

62. For a comprehensive survey of similar neuro­

specialization and resulted in a sizeable scientific liter­

I discuss

ature, as well as a huge corpus of popular psychology

some philosophical issues involved in the phenomenon

writing on the so-called left-brain versus right-brain

psychological disorders, see Farah ( 1995).

Approaching Consciousness

distinction with regard to personality types, social be­ havior, and so on. One of the major results that came out of the commissurotomy research is the hypothesis proposed and defended by (among others) Michael Gazzaniga (1993), that human cognition as well as consciousness (in the sense of awareness of experience) are subserved by special brain circuitry normally lo­ cated in the left hemisphere. For an account of the early work on commissurotomy, see Gazzaniga (1970). Galin (1974) explores the implications of hemispheric special­ ization for psychiatry. For a comprehensive collection of current research results in human neuropsychology, including articles on modularity of mental function and hemispheric specialization, see Gazzaniga et al. (1 995). 65. It is interesting to note that this is a junction at which some upholders of the "pro-qualia intuition" meet on common ground with the most indoctrinated qualia skeptics, such as the behaviorist psychologists for whom exorcising qualia out of the scope of psychology was a primary goal. Notice, for instance, the similarity between Jackson's ( 1982) position (who characterizes himself as a "qualia freak," p. 1 27), and the position defended by Edward Tolman: "[Regarding visual per­ ception in others) we never learn whether it 'feels' like our 'red' or our 'green' or our 'gray', or whether, indeed, its 'feel' is perhaps sui generis and unlike any of our own . . . . Whether your 'raw feels' are or are not like mine, you and I shall never discover. Your color 'feels' may be the exact complementaries of mine, but, if so, neither of us will ever find it out, provided only that your discriminations and my discriminations agree . . . . If there be 'raw feels' correlated with such discriminanda-expectations, these 'raw feels' are by very definition 'private' and not capable of scientific treat­ ment. And we may leave the question as to whether they exist and what to do about them, if they do exist, to othe disciplines than psychology-for example, to logic, epistemology, and metaphysics. And whatever the answers of these other disciplines, we, as mere psychol­ ogists, need not be concerned" (Tolman 1967, pp. 252253). Whether this surprising "meeting of minds" between such arch-opponents as behaviorist psychologists and a certain brand of qualia defenders on presumably the very point of contention between them speaks in favor of the former or the latter party (if either), I leave open to the judgment of the reader.

;

66. William James (I 950a), who characterizes con­ sciousness as a "fighter for ends," makes the following

61

statement (largely in agreement with Herbert Spencer 1891, 1 898) in support of the causal construal and the evolutionary relevance of consciousness:

It is a well-known fact that pleasures are generally associated with beneficial, pains with detrimental ex­ periences. . . . These coincidences are due, not to any pre-established harmony, but to the mere action of natural selection which would certainly kill off in the long-run any breed of creatures to whom the funda­ mentally noxious experience seemed enjoyable. An animal that should take pleasure in a feeling of suf­ focation would, if that pleasure were efficacious enough to make him immerse his head in water, en­ joy a longevity of four or five minutes. But if the pleasures and pains have no efficacy, one does not see why the most noxious acts, such as burning, might not give thrills of delight, and the most necessary ones, such as breathing, cause agony . . . . The conclusion that [consciousness) is useful is . . . quite justifiable. But, if it is useful, it must be so through its causal efficaciousness. James's conclusion is that "the study a posteriori of the distribution of consciousness shows it to be exactly such as we might expect in an organ added for the sake of steering a nervous system grown too complex to reg­ ulate itself" (pp. 1 4 1 , 143-144). 67. See also Kitcher (1979) and Revonsuo (1 994) for related points. 68. Precursors to Nagel's thinking on this issue can be found in the writings of B. A. Farrell and Timothy Sprigge. Even though it was made famous by Nagel, the original formulation of the question "what is it like to be a bat?" goes back to Farrell's somewhat neglected essay, "Experience" (1950). In discussing the issue of experi­ ential knowledge, Farrell imagines a Martian visitor about whose sensory capacities we obtain all the in­ formation there is. According to Farrell, "We would probably still want to say: 'I wonder what it would be like to be a Martian.' " He continues: "There is some­ thing more to be learned about the Martian, and that is what his experience is like." Farrell then extends the question to babies and mice, as well as an opium smoker, and finally a bat: "I wonder what it would be like to be, and hear like, a bat" (pp. 34-35). The lessons Farrell draws out of his ruminations are quite the opposite of Nagel's conclusions, however. It is rather Sprigge (1971) who makes the connection be­ tween the "what it is like" aspect of experience and

Guven Giizeldere

62

physicalism's difficulty with accommodating it in the

of Nitrogen; it is a gas readily soluble in water, and

particular way Nagel problematizes the issue: "When

possessed of a pungent and characteristic smell." If

one imagines another's conscious state, there is no con­

the mechanistic theory be true the archangel could

clusive way of checking up whether one has done so

deduce from his knowledge of the microscopic

correctly or not. . . . Presuming that the object (that is, at

structure of atoms all these facts but the last. He

least normally, the organism) with which one is con­

would know exactly what the microscopic structure

being that organism

of ammonia must be; but he would be totally unable

will have a certain definite complex quality at every

to predict that a substance with this structure must

cerned, is indeed conscious, then

waking moment. . . . Physical science makes no refer­

smell as ammonia does when it gets into the human

ence to qualities of this kind. Thus consciousness is that

nose. The utmost that he could predict on this

which one characterises when one tries to answer the

subject would be that certain changes would take

question what it is or might be like to be a certain object

place in the mucous membrane, the olfactory nerves

in a certain situation" (p. 1 68).

and so on. But he could not possibly know that these

69. Lycan (1 997b) also makes a similar point: "The

changes would be accompanied by the appearance

phrase 'what it's like' is more sinning than sinned

of a smell in general or of the peculiar smell of am­ so

against; nothing whatever is clarified or explained by

monia in particular, unless someone told him

reference to it, and it itself is not only badly in need of

he had smelled it for himself. (Broad 1 962, p. 71)

explanation, in general, but at least three-ways ambig­ uous in particular" (p. 1 76).

or

Similarly, an early formulation of the knowledge ar­ gument, as well as an antecedent of the Nemirow-Lewis

70. See Nagel (1 974, 1979, 1 986) for a full range of

critique, appears in Feigl's discussion of "cognitive roles

problems that involve SUbjectivity. See Lycan (1 990),

of acquaintance." Feigl asks: "What is it that the blind

Biro

man cannot know concerning color qualities?" and

(1991),

Akins

(1 993),

and

Dretske

(1 995),

among others, for deflationary responses. See also

proposes the following answer:

Nagel (1983) for a discussion of how subjectivity figures in the problem of self, without ever touching on qualia,

If we assume complete physical predictability of hu­

and Perry (forthcoming) for a penetrating analysis and

man behavior, i.e., as much predictability as the best

critique of Nagel's account.

developed physical science of the future could con­

7 1 . For Jackson's formulation, see Jackson (1982, 1 986).

For various critiques, see Nemirow (1 980),

Churchland (1 989), Lewis (1 990), Dennett (199 1 , chap. 10), Van Gulick (1 993), Loar ( 1 990), Harman (1 993b), Dretske (1 995, chap. 3), and Perry ( 1 995). For a related empirical study on the conceptual representation of colors in the blind and the color-blind, see Shepard and Cooper (1 992). One of the earlier formulations of the knowledge argument can be found in C. D. Broad's thought ex­ periment about the archangel who knows all about chemistry but lacks the sense of smell. Broad sets up the problem as follows: Would there be any theoretical limit to the de­ duction of the properties of chemical elements and compounds if a mechanistic theory of chemistry were true? Yes. Take any ordinary statement, such as we find in chemistry books; e.g., "Nitrogen and Hydrogen combine when an electric discharge is passed through a mixture of the two. The resulting compound contains three atoms of Hydrogen to one

ceivably provide, then it is clear that the blind man or the Martian would lack only

knowledge by acquaintance

acquaintance

and

in certain areas of the

realm of qualia. Lacking acquaintance means not having those experiential qualia; and the consequent lack of knowledge by acquaintance simply amounts to being unable to label the qualia with terms used previously by the subject (or by some other subject) when confronted with their occurrence in direct ex­ perience. Now, mere

knowledge

having

or

living through is

not

in any sense. "Knowledge by acquaint­

ance," however, as we understand it here, is propo­ sitional, it does make truth claims. Feigi then goes on to suggest, anticipating many of the present-day critiques of the knowledge argument, that the blind person (or Jackson's fictional color scien­ tist, Mary) does not lack any knowledge per se; all he or she lacks is a particular mode of knowing the same facts as do normally sighted people: "What one person and

knows by acquaintance may

has

be identical with what

someone else knows by description. The color experi-

Approaching Consciousness

ences of the man who can see are known to him by acquaintance, but the blind man can have inferential knowledge, or knowledge by description about those same experiences" (Feigl I967, p. 68). A related puzzle, based again on a thought-experiment, was posed as a "jocose problem" to John Locke by William Molyneux, an amateur philosopher, in a letter dated 1 693: "Sup­ pose a man born blind, and now adult, and taught by his touch to distinguish between a cube and a sphere of the same metal, and nighly of the same bigness, so as to tell, when he felt one and the other, which is the cube, which the sphere. Suppose then the cube and sphere placed on a table, and the blind man be made to see: quaere, whether by his sight. before he touched them, he could now distinguish and tell which is the globe, which the cube?" (Quoted in Locke (1959), Book II, Chapter IX, pp. 186-1 87.) As such, Molyneux's question transfonns the inquiry of whether non-experientiaI facts can yield knowledge of experiential facts (in the knowledge argu­ ment) into a puzzle about intersensory translation­ whether tactile facts can yield knowledge of visual facts ("facts" taken broadly). Locke's negative answer to this question was in agreement with Molyneux's opinion. But others disagreed, and Molyneux's question became one of the central topics of contention among such phi­ losophers as Berkeley, Leibniz, Reid, Diderot, and Voltaire, in the greater context of the controversies over innateness and abstract ideas. See Morgan (1977) and Sanford (1983) for further exposition and discussion of the related issues. 72. For a classical treatment of the identity theory, see Sellars (1964). For recent arguments in its defense, see En� (1983) and Hill (1991). A thus far unmentioned but related concept that has played a significant role in the philosophy of mind over the past few decades is super­ venience (presumably imported from Rom Hare's work in ethics [1952] by Donald Davidson 1 970). The thesis that the mental supervenes on the physical is put forth as a better materialist solution to the mind-body problem than the identity thesis. The supervenience thesis is roughly that the mental character of a state or event is wholly determined by its physical profile, such that there cannot be a change in the fonner without a change in the latter. Put differently, sameness in the (subvening) physical properties is hypothesized to guarantee same­ ness in the (supervening) mental properties. But the status of the supervenience thesis vis-a-vis the mind-body problem, especially with regard to phenomenal consciousness, remains controversial. Are

63

supervenient properties ontologically distinct (auton­ omously emergent) from physical properties, for in­ stance? Or (otherwise) does supervenience boil down to old-fashioned identity? How are we to explain the nature of the supervenience relation itself? Given the wide variety of possible supervenience relations (for ex­ ample, weak versus strong, local versus global), there is a vast and technically complicated literature in this area, but there exist no clear-cut received views that are taken to unanimously answer all these questions. For a sys­ tematic and thorough exploration of the supervenience thesis in all its different characterizations, see Jaegwon Kim's essays collected in Kim (1993). Two other useful collections that contain representatives of contem­ porary theorizing on supervenience are Beckerrnann, Flohr, and Kim (1992), and Savellos and Yal�m (1995). See also Mclaughlin ( 1989) for a discussion that relates supervenience to the question of epiphenomenalism, and McLaughlin (1 992) for a thorough exposition of the thought of British emergentists in which one can find numerous clues for the present supervenience debates. In addition to philosophy of mind, notions of super­ venience, emergence, and different kinds of reduction have been central to discussions in philosophy of biol­ ogy, particularly during the first half of this century. See Brandon ( 1996) for an exposition of these issues that also ties them to debates in contemporary biology, and Harris ( 1 993) for an exploration of the relations be­ tween the natures and the study of mind and life in phi­ losophy, psychology, and biology. 73. It is interesting to note, by the way, that the word

consciousness was hardly ever present in the philosoph­ ical literature around the time of functionalism, which was instead brewing with the tenn qualia and, to a lesser extent, tenns like raw feels and phenomenal aspect of consciousness. In the 19905, in contrast, many of the same old problems have gained new interest and im­ petus (perhaps from slightly but essentially similar perspectives) within a terminology populated with the magic word consciousness and its derivatives. The rea­ sons for this terminological change probably lie partIy outside philosophy, for example in the wide acceptance of the tenn consciousness into other fields with which philosophy interacts. 74. There is a sizeable literature regarding the absent qualia argument. Some of the most influential thought experiments that support the possibility of absent qualia are due to Block (1978, 1980b) and Block and Fodor (1980). See Shepard (1993) for a recent discussion in the

Giiven Giizeldere

context of color vision and evolutionary theory. Shoe­ maker (1975, 198 1 a) presents an eloquent defense of functionalism against the absent qualia arguments. See also Dretske (1995) who raises the possibility of absent qualia in an extemalist context in relation to Donald Davidson's (1987) "swampman" argument, and Tye (1995) for a representationalist critique of absent qualia. For further discussion, see Lycan (1981), Levin (1985), Graham and Stephens ( 1985), White ( 1986), Fox (1989), Levine (1989), Horgan ( 1987), and Hardcastle (1996). One of the most commonly cited absent qualia argu­ ments is based on Block's ( 1978) "Chinese Nation" scenario, designed to "embarrass all versions of func­ tionalism" by showing that functionalism is guilty of "classifying systems that lack mentality as having men­ tality" (p. 275). Block asks us to imagine the functional simulation of a human brain by the Chinese nation by connecting each of the billion inhabitants of China in appropriate ways through radio links, and having them communicate from a distance like neurons in a brain and thereby animate an artificial body for a certain pe­ riod of time. According to Block, while this China-body system is "nomologically possible" and "it could be functionally equivalent to [a human being] for a short time," it is doubtful "whether it has any mental states at all-especially whether it has . . . 'qualitative states', 'raw feels', or 'immediate phenomenological qualities' " (pp. 276-278). Block's point is to establish the short­ comings of functional characterizl!tions of qualia, by appealing to intuitions that he takes as common sens­ ical, such as the intuition that such "distributed minds" are absurd. A similar intuition was commonly employed in dis­ cussions regarding the unity of mind versus the divisi­ bility of matter with the aim of embarrassing all forms of materialism in early modem philosophy. For exam­ ple, the eighteenth century English theologian Samuel Clarke appeals to the absurdity of the distributed-minds intuition in a piece of hypothetical reasoning, similar to Block's, to make a case for the "immateriality and nat­ ural immortality of the soul" as follows: "That the soul cannot possibly be material is moreover demonstrable from the single consideration even of bare sense and consciousness itself. For suppose three, or three hun­ dred, particles of matter, at a mile, or at any given dis­ tance, one from another; is it possible that all those separate parts should in that state be one individual conscious being?" But Clarke then takes his argument a step further to apply it to human beings: "Suppose then

64

all these particles brought together into one system, so as to touch one another; will they thereby, or by any motion or composition whatsoever, become any whit less truly distinct Beings, than they were at the greatest distance? How then can their being disposed in any possible system, make them one individual conscious being?" (Clarke 1 707, p. 82) 75. The origins of this problem go back indeed to a puzzle about visual experience, described by John Locke a few centuries ago: Neither would it carry any imputation of falsehood to our simple ideas, if by the different structure of our organs it were so ordered, that the same object

should produce in several men's minds different ideas at the same time; e.g. if the idea that a violet pro­ duced in one man's mind by his eyes were the same that a marigold produced in another man's, and vice versa. For, since this could never be known, because one man's mind could not pass into another man's body, to perceive what appearances were produced by those organs: neither the ideas hereby, nor the names, would be at all confounded, or any falsehood be in either. For all things that had the texture of a violet, producing constantly the idea that he called blue, and those which had the texture of a marigold, producing constantly the idea which he constantly called yellow, whatever those appearances were in his mind; he would be able as regularly to distin­ guish things for his use by those appearances, and understand and signify those distinctions marked by the name blue and yellow, as if the appearances or ideas in his mind received from those two flowers were exactly the same with the ideas in other men's minds (Locke 1959, bk. II, chap. 32, §15, p. 520). There are also several positions with respect to the inverted spectrum argument. Here is a simple set: Block (1978) and Block and Fodor (1980) raise the possibility of inverted spectrum against functionalism, and Block (1990) presents an original twist on the same problem as a reply to Harman (1990), who argues against inverted spectrum on extemalist grounds. Shoemaker (1975, 1 98 1 b, 1991) is more lenient toward accepting the pos­ sibility of inverted spectrum compared to his rejection of the possibility of absent qualia, but he presents an argument on how to accommodate qualia inversion within a broadly functionalist framework. However, see also Levine ( 1988), who argues that absent qualia and inverted spectrum stand or fall together as logical

Approaching Consciousness

65

possibilities against functionalism. Dretske (1995) re­ gards inverted spectrum as a problem for functionalism but not for representationalism (and hence, not for materialism in general). Tye (1995) argues for the conclusion that is somewhat similar to Shoemaker's: spectrum inversion is possible in narrow functional duplicates, but this does not constitute a problem for wide functionalism (that Tye defends). 76. A similarly externalist theory is put forth by Gilbert Harman (1990) where qualia are identified with inten­ tional properties, as well as by Michael Tye (1995). David Armstrong had earlier suggested identifying qualia (what he called "secondary qualities") with properties of physical objects in his discussion of "Re­ alist Reductionism" (Armstrong 1 993, chap. 12, 270290). Clues for these positions can be found in Elizabeth Anscombe's discussion of the intentional nature of sen­ sations (Anscombe 1965). 77. Perhaps a caveat about the particular brand of epiphenomenalism and zombiehood I am referring to is in order here. Epiphenomenalism about phenom­ enal consciousness is, of course, different from epi­ phenomenalism about the mind in general. Discussions about epiphenomenalism earlier this century generally assumed the latter kind (see, for instance, Broad 1 962). In the contemporary literature, however, the focus has somewhat shifted. Probably largely due to the advent of functionalist and computational-representational theo­ ries of mind in a materialist framework, many today take intentional states, such as beliefs and judgments, as contentful internal structures in the brain (see, for ex­ ample, Fodor 1987). As such, no one thinks that beliefs qua such physical structures, lack causal properties. Th controversy is rather on whether their content (seman­ tics) has a causal role in the explanation of behavior (cf. Dretske 1988). But the real hot spot of the epiphenomenalism debate has to do with phenomenal consciousness-whether qualia play any role in the otherwise causally charac­ terizable economy of our mental lives. Note, after all, that while there is a vast literature on the possibility of absent qualia, no one seems to be worrying about the possibility of "absent beliefs" or "absent judgments." There seems to be a crucial difference between beliefs and pains: while it is considered legitimate to attribute beliefs to someone who behaves in ways that can be ex­ plained by belief-attributions of the relevant sort, it is considered very problematic to so attribute pains, be-

�

cause the essence of pains, the reasoning goes, is not at­ tributable (by a third party) but rather accessible in a privileged way (through the first-person perspective). There is something it is like to have pains, but there is nothing it is like to believe that there is no greatest prime number (or even that one is in pain). It is this difference that warrants epiphenomenalism, in the present liter­ ature, as a possibility with respect to pains, but not be­ liefs (about prime numbers, one's pains, or anything else). Accordingly, under the stipulation of appropriate environmental and historical conditions, it is generally regarded as a possibility that a physical replica of a hu­ man being can lack all qualia, while not lacking beliefs or judgments (or other such intentional states). This is the possibility of the modem zombie that has center stage in debates about phenomenal consciousness. Ned Block (1995) calls such replicas phenomenal zombies: "the familiar . . . robots that think but don't feel" (p. 234). Or, as David Chalmers (1 996) describes: "My zombie twin . . . will be psychologically identical to me. He will be perceiving the trees outside, in the functional sense, and tasting the chocolate, in the psychological sense [similar to its ordinary twin modulo qualia] . . . . He will be awake, able to report the contents of his internal states, able to focus attention in various places, and so on. It is just tliat none of this functioning will be ac­ companied by any real conscious experience. There will be no phenomenal feel. There is nothing it is like to be a zombie" (p. 95). It is important to notice that zombies, construed as such, are taken to be in possession of all sorts of beliefs, thoughts, and judgments that their human twins typi­ cally have, including the self-ascribed ones. As Chalm­ ers (1993) states: "Zombie Dave's beliefs may not be colored by the usual phenomenological tinges, but it seems reasonable to say that they are nevertheless beliefs. Beliefs, unlike qualia, seem to be characterized primarily by the role that they play in the mind's causal economy." Accordingly, the zombie twin, too, takes aspirin because he thinks he has a headache, wants an­ esthetics at the dentist chair because he believes the root canal will hurt, and so on. It is just that all his beliefs and judgments about his own qualia are systematically false. Nothing hurts in him, even ifhe sincerely believes he has a splitting headache. Accordingly, we should not be motivated to put him under anaesthesia when his tooth is being drilled, despite all his screams, if it is the pain quale that matters.

Giiven Giizeldere

66

Put differently, what distinguishes this kind of a zombie from its human twin is the

stipulated

ever­

presence of a gap between the "appearance" and "re­ ality" of the zombie's qualitative states-what qualia he judges himself to have versus what qualia he really

JOUTfUlI of Consciousness Studies (2:4, 1995) for a sym­ posium on zombies based on Moody 1994). 8 1 . Here is Dennett ( 1979) on describing the phenom­ enology of one's own experience: "We are all, I take it, unshakably sure that we are each in a special position to

has-and nothing much else. (Note that something like

report, or to know, or to witness or experience a set of

an "appearance-reality" distinction is required in order

something-or-others we may call, as neutrally as pos­

to coherently conceptualize the possibility of a zombie.) As such, there is a psychologically significant and ex­ planatorily important sense in which things

seem

(i.e.,

are judged, thought, believed, expected, noticed, . . . to be) a certain way to the zombie twin.

elements of our own conscious experience. . . .

sible,

Propositional episodes . . . comprise our streams of consciousness by embodying our semantic intentions of the moment, by being the standards against which we correct, or would correct, any failures of execution were

Consequently, to the extent that nonqualitative in­

we to utter anything at the time . . . . These are . . . think­

tentional states, (including self-ascribed beliefs about

ings that p. . . . I call them judgments. . . . Such judg­

one's own qualitative states) are constitutive of a first­

ments

person perspective (cf. Chisholm 1 98 1 ), the zombie can

individual streams of consciousness consist of nothing

exhaust

our immediate consciousness, that our

be said to have such a perspective (albeit a systemati­

but such propositional episodes. My view, put bluntly,

cally misguided one). A zombie's life is not, after all,

is that there is no phenomenological manifold in any

completely devoid of all mental elements. He only lacks an important component of an otherwise intact epis­

such relation to our reports. There are the public reports we issue, and then there are the episodes of our propo­

temic perspective. After all, it is in virtue of having such

sitional awareness, our judgments, and then there is-so

a perspective that receiving anaesthetics at the dentist

far as

chair

seems

to matter to the zombie twin, even if his

introspection

is concemed-darkness" (pp. 93-

95).

preferences are entirely on the basis of false self-ascribed

This quotation probably represents a position at a

beliefs (about his own non-existent pain qualia). It is this

far end of the spectrum of views on the nature of inner

characterization of zombiehood that is typically in­

experience. There are, of course, many midway views

voked in contemporary debates, and I will confine my

between the positions that occupy the two endpoints.

discussion accordingly throughout the rest of this

For instance, one can agree with "friends ofqualia" that

chapter.

there

78. Notice that this view, as such, does not deny that we

term "phenomenal consciousness" distinct from "epi­

are conscious. It comes close, however, in positing that

sodes of propositional awareness and judgments," but

is

something crucial to theorize about under the

our being conscious, in itself, makes no difference­

maintain that the way they choose to theorize about it is

hence the path to zombiehood.

misguided.

79. James (1 879) is a vigorous response to Huxley. See

82. The disagreement is not only between those who

also Capek (1 954) for a commentary on this exchange.

believe in phenomenal consciousness and those who

80. As far as I could trace, the term

zombie

enters the

philosophical vocabulary with Kirk ( 1974) in an argu­ ment against materialism. For other arguments that defend the intuition in favor of the possibility of zom­

deny it; it pervades the community of the supporters of this very notion. For instance, Block claims that Searle, in an attempt to point to consciousness, confounds too many senses of the term. As Block says, "It is important

the omniscient pointer?

biehood, see Block (1 978), Block and Fodor (1 980),

to point properly. " But who has

Searle ( 1992), and Chalmers (1996). For counterargu­

83. Jaegwon Kim ( 1 996) makes a similar point with re­

ments, see Lewis ( 1972) and Shoemaker ( 1975), as well

spect to mind in general: "Saving mentality while losing

as Kirk ( 1 994). For an expression of philosophical in­

causality doesn't seem to amount to saving anything

tolerance for epiphenomenalism and the possibility of

worth saving. For what good is the mind if it has no

zombies, see Dennett ( 1 99 1 , chp. 10). Giizeldere ( 1995d)

causal powers?" (p. 237). Or, to transform a point Fred

and

Dretske (1 988) makes with respect to the explanatory

merely behavioral zombies, and briefly examines their

role of content into one about phenomenal conscious­

respective underlying metaphysical assumptions. (See

ness (with apologies), "if having a mind is having

distinguishes among physiological,

functional,

this

67

Approaching Consciousness

kind of qualia (which don't do anything), one may as well not have a mind" (p. 80). However, this position should be distinguished from eliminativism about consciousness, defended, for in­ stance, by Rey ( 1 988). Rey's suggestion is to do away with the largely folk-theoretic notion of consciousness because it contains Cartesian elements and plays no useful causal-explanatory role. I am suggesting (and probably Kim and Dretske would agree), in contrast, that because we do not want to do away with our com­ mon sense notion of consciousness (or, at least, a sig­ nificant part of it), we need to seek to secure a genuine role for it in the causal web of the world. 84. Put differently, I am essentially in agreement with David Lewis (1 980) on a point he makes regarding the status of pain that I take as an objection to the segrega­ tionist intuition: "Only if you believe on independent grounds that considerations of causal role and physical realization have no bearing on whether a state is pain should you say that they have no bearing on how a state feels" (p. 222). Along these lines, see Humphrey (1 992); see also Hardin (1987, 1988) for an important attempt to deftate the explanatory gap between the causal and phenomenal aspects of consciousness in the case of vis­ ual perception and color qualia. 85. The Fodor quote is from Fodor 1987, p. 1 56. Fodor holds a much more pessimistic opinion regarding the prospects for a theory of consciousness in comparison to a theory of rationality, however. Regarding rationality, he thinks that "certain residual technical difficulties" notwithstanding, "we are (maybe) on the verge of solv­ ing a great mystery about the mind: How is rationality mechanically possible?" (Fodor 1987, pp. 1 56, 21). Re­ garding consciousness, here is what he says: "Nobody has the slightest idea how anything material could be conscious. Nobody even knows what it would be like to have the slightest idea about how anything material could be conscious. So much for the philosophy of con­ sciousness" (Fodor 1 992, p. 5). Perhaps then, one thinks, Fodor should consider starting to think about consciousness and give those working in the field a helping hand. But similarly gloomy sentiments are expressed by Block (1 994), a prominent figure in the consciousness literature, as well: "The notable fact is that in the case of thought, we ac­ tually have more than one substantive research pro­ gramme, and their proponents are busy fighting it out, comparing which research program handles which phe­ nomena best. But in the case of consciousness, we have

nothing-zilch-worthy of being called a research pro­ gram, nor are there any substantive proposals about how to go about starting one. . . . Researchers are stumped. . . No one has yet come up with a theoretical . perspective that uses these data to narrow the explana­ tory gap, even a little bit" (p. 2 1 1). My ultimate conclusion, expressed in the last para­ graph of this chapter and hopefully substantiated by the exposition presented thus far, differs from both. .

References to the Introduction appear on pages 8078 1 6.

I

STREAM OF CONSCIOUSNESS

1

The Stream of Consciousness

WUliam James The order of our study must be analytic. We are now prepared to begin the introspective study of the adult consciousness itself. Most books adopt the sO �

- .8

.;! « � 31 U il: C O �

-

1: 5 E O:: �� Z ij

«

2SO 225

Same Occupation

- - - 0- - -

Different Occupation

-,. -

Name Only

200 175

f---- f -

1SO 125 1 00

--

-

t

- -

- -

�

-__

__

_

75

so 25 0 0.0

12.5

25.0

37.5

SO.O

62.5

Percent Units Removed

75.0

87.5

100.0

22 1

Martha J. Farah

222

, '.

, "

f�

,

,

· · · ,

I

"Apple "

• " " . . . .

' . " .

. , ,

, •

ffi "Apple "

I

,

I

I

"

" .

',

"Comb"

,

,

, • I

:

,

. I.

W

Patient:

"Different"

Doctor:

"What exactly?"

Patient:

"A comb and I don 't know what the other was, "

Figure 11.10 Typical trials from the experiment of Yolpe, LeDoux. and Gazzaniga ( 1979), showing extinction of the left stimulus with preserved same/different matching.

Farah, Monheit, and Wallace ( 1 99 1 ) noted that the dissociation between naming and same/ different matching could, in principle, be ex­ plained by the differing demands these tasks make on the quality of the representation of the left stimulus. Consider the example of the comb and apple, shown in figure 1 1 . 10. If only partial stimulus information were picked up on the left­ for example, the perception that there is some­ thing roundish and light-colored-this would be sufficient to enable fairly accurate same/different judgments. However, there are so many roundish, Iight-colored objects in the world that this partial perception would be of no help in naming the stimulus on the left. We performed two experi­ ments to test the hypothesis that the differential amounts of stimulus information required for same/different matching and overt identification is what causes the dissociation between them. First, we degraded the left side of the display with a translucent mask and repeated Volpe, LeDoux, and Gazzaniga's experiment with normal sub-

jects. This manipUlation was not intended as a simulation of extinction but rather as test of our hypothesis concemin& the quality of visual in­ formation needed in the two kinds of task. Merely depriving normal subjects of some information from the left stimulus produced the same dis­ sociation in normal subjects that Volpe, LeDoux, and Gazzaniga observed in patients with ex­ tinction. Second, we repeated the paradigm with extinction patients in its original form, and re­ plicated the original finding, and also in an al­ tered from, in which the overt identification task was administered in a forced-choice format. The purpose of this alteration was to enable us to equate the same/different trials and identification trials for the amount of information needed from the left stimulus. We did this by yoking trials from the two conditions, such that if there was a same/ different trial with a triangle and a square, there was an identification trial in which one of these stimuli was presented on the left and the subject was asked, "Did you see a triangle or a square?"

223

Visual Perception and Visual Awareness after Brain Damage

When same/different matching and identification were equated for their demands on the quality of the subjects' representations ofthe left stimuli, the

Primlns

RVF

LVF

dissociation vanished. We concluded that the task

........; ,--.:' _ · IIE

of same/different matching and identification dif­

-

of Volpe, LeDoux, and Gazzaniga are consistent

200 msec

with extinction's affecting perception per se and do not require us to conclude that perception is normal and only postperceptual access to con­ scious awareness is impaired by extinction. Spe­

Z

- � . ..:;

fer significantly in their demands on the quality of visual representation. This implies that the results

Tarpt

Stimuli

400 msec lSI

BALL Unlimited

Figure 1 1.11 Sequence of stimuli from a typial trial in the experiment of McGlinchy-Berroth et al. ( 1993).

cifically, our results suggest that extinction results in partial, low-quality perceptual representations

ceive perceptual and even conceptual processing,

of stimuli.

despite subjects' inability to report specifically

Berti et al. ( 1 992) extended these findings in

what they saw.

a different direction, with a study of one patient

A recent study by McGlinchey-Berroth et al.

who showed extinction following a temporal

(1 993) showed semantic priming by neglected

lobectomy. The goal of their study was to de­

pictures in a lexical decision task. On each trial of

termine the level of processing to which extin­

this experiment, subjects with left neglect viewed

guished stimuli were encoded. In Volpe, LeDoux,

a picture in one hemifield, followed by a letter

and Gazzaniga's ( 1 979) study, stimuli were either

string in central vision to which they made a

physically identical or entirely different, and sub­

"word" /"nonword" response (fig. 1 1 . 1 1). When

jects could therefore make accurate responses

the picture was semantically related to the word,

on the basis of relatively early perceptual repre­

"word" responses were faster, even when the pic­

sentations, were they of high or low quality. Berti

ture had been presented on the left side of the

et al. included pairs of pictures that were physi­

display. Perhaps most striking about this result is

cally different but depicted either the same object

that the amount of priming did not differ signif­

from a different view or different-looking exem­

icantly between left- and right-field stimuli. Al­

plars of the same type of object, such as two dif­

though this might appear to imply that the amount

ferent cameras, and they instructed their subject

of semantic priming from neglected stimuli is

to say "same" if the two stimuli had the same

quantitatively normal, such a conclusion is not

name. If extinguished stimuli are encoded to early

warranted because normal subjects show greater

levels of visual perception prior to object con­

left- than right-field semantic priming (Chiarello

stancy but do not make any contact with stored

1 988). In fact, this asymmetry was visible in the

object representations or meaning, then the sub­

control subjects in McGlinchy-Berroth et al.'s

ject should perform well only when the two stimuli

experiment.

are physically identical. In contrast, if extin­

The finding of preserved semantic priming by

guished stimuli are encoded to the level of mean­

left-field stimuli must also be interpreted in the

ing, then the subject should be able to say, for

context of the degree of neglect for left-field stim­

example, that two different cameras are the same.

uli. McGlinchy-Berroth et al. conducted a second

In fact, this is what was found, implying that ex­

experiment in which the degree of neglect for left

tinguished stimuli are not only registered at some

stimuli was assessed in a delayed matching task.

early level of visual processing but that they re-

Patients were indeed poor at matching a left

Martha 1. Farah

224

stimulus picture with a central one presented

In further research, by Bisiach and Rusconi

56 per­

(1 990), who tested four patients with four pairs of

cent correct. However, they were also surprisingly

drawings, the striking finding of Marshall and

poor at performing the same task with stimuli

Halligan ( 1 988) was not replicated. Often patients

shortly after, achieving on average only

presented on the right, achieving on average only

displayed no consistent preference. Perhaps more

74 percent correct, raising the worry that perfor­

revealing, when patients did have a preference, it

mance in this matching task was limited by diffi­

was not necessarily in the normal, rational direc­

culties other than neglect and that not all left

tion. For example, the only patients who showed

stimuli would therefore necessarily have been ne­

a preference with the house stimuli preferred the burning house! Bisiach and Rusconi queried their

glected in the priming task. Unconscious perception of neglected stimuli

subjects as to the reasons for their preferences and

has also been investigated under conditions of

found that the reasons fell into two categories:

nontachistoscopic viewing in central vision. Sier­

real but minor differences on the right sides of the

off, Pollatsek, and Posner ( 1 988) analyzed the

stimulus materials, and confabulated differences,

reading capabilities of patients with neglect and

such as alleged differences in the layouts of the

found that letters falling in the neglected regions

houses or the numbers of rooms. These confabu­

of the visual field (as assessed by poor naming of

lations are not uninteresting; they suggest that

letters in nonword letter strings) could never­

patients with neglect may detect differences on

theless be used to identify familiar words. For

the left that they cannot describe verbally, instead

table correctly, but

evoking confabulations. However, the results of

example, patients might read

read taspi as "f-o-s-p-i" or "s-p-i," misperceiving

Bisiach and Rusconi do suggest that Marshall

or omitting the letters on the left. Sieroff, Posner,

and Halligan's case was probably responding to

and Pollatsek argue against the possibility that

some minor difference on the right side of the

patients were simply guessing the letters on the

stimulus pair or to some dimly perceived differ­

left sides of words by pointing out that words and

ence on the left, rather than to the unconscious

nonwords were intermixed in their experiments,

realization that one of the houses was on fire and

and patients did not often guess words when

would therefore be a dangerous place to live.

shown the nonword letter strings. Although these authors were not specifically concerned with the

Awareness of Perception

issue of conscious awareness in neglect, their findings seem to imply that the leftmost letters

Patients with neglect and extinction generally re­

were unconsciously perceived and therefore avail­

port no awareness of having seen stimuli that they

able to the word recognition system, despite pa­

fail to detect or of having recognized stimuli that

tients' inability to report the letters.

they fail to identify. One measure of the extent to

Marshall and Halligan ( 1988) described a pa­

which the subjects were not aware of their own

tient with neglect who evinced a consistent pref­

perceptions in the studies reviewed here is their

erence for a picture of a normal house over a

tendency to remark on the silliness or absurdity of

picture of the same house with flames coming

answering questions about stimuli on the left, or

from the left side, despite her inability to say how

choosing between a pair of stimuli identical ex­

(or even whether) the two stimuli differed. This

cept for a feature on the left (Marshall and Halli­

pattern of performance is consistent with the hy­

gan 1 988; Volpe, LeDoux, and Gazzaniga 1979).

pothesis that neglected stimuli are perceived and

Such tasks would indeed seem silly if one were not

understood, at least to the degree that they can

aware of a stimulus on the left or of the difference

evoke an affective response, without being acces­

between the paired stimuli . It should be noted,

sible to the mechanisms of conscious verbal report.

however, that the ways in which awareness of

Visual Perception and Visual Awareness after Brain Damage

perception was fonnally operationalized in these studies was not to ask the subjects whether they detected a stimulus. Rather, the most common means was verbal identification of the stimulus. This was used in the studies of same/different matching ability and in the studies of letter per­ ception within word and non-word letter strings. Although Volpe, LeDoux, and Gazzaniga stated that their subjects sometimes denied seeing any stimulus at all on the left, no systematic data are reported on detection; they contrasted same/dif­ ferent matching perfonnance only with naming. Other tasks intended to assess conscious aware­ ness of stimulus identity included the delayed matching task of McGlinchy-Berroth et al. and the same/different judgment of Marshall and Halligan. In contrast to the research on covert recognition in prosopagnosia, research on un­ conscious perception in neglect and extinction has not drawn a clear and principled line between tasks that require overt conscious awareness of perception and tasks considered to be indirect tests of unconscious perception. The fact that same/different matching was considered an in­ direct test of unconscious perception by Volpe, LeDoux, and Gazzaniga and a test of overt con­ scious perception by Marshaal and Halligan is symptomatic of this problem.

Explanations of Perceptual Dissociations in Patients with Neglect and Extinction The earliest and most straightforward interpreta­ tion of the dissociation between perception and awareness in patients with neglect and/or extinc­ tion was offered by Volpe, LeDoux, and Gaz­ zaniga. They suggested that extinction in their subjects consisted of "a breakdown in the flow of infonnation between conscious and nonconscious mental systems. The stimulus comparison task in our study appears to have been carried out at a . postperceptual, preverbal level, with only the re­ sultant comparison entering consciousness." This account clearly falls into the privileged rote cat­ egory, because consciousness is being attributed

225

to certain systems and not others. Although the authors are not very explicit on the point, it ap­ pears from the quotation just cited and another reference to "some level of neuronal processing which allows for verbal description, if not con­ scious awareness" that the system required for consciousness is closely related to the language system (Gazzaniga 1 988). An alternative interpretation of the kinds of dissociations reviewed here has been offered by Kinsbourne (1988). Rather than viewing con­ sciousness as a property of some neural systems and not others, he considers it a state of in­ tegration among different neural systems. Ac­ cording to this view, neglect does not divert percepts from conscious awareness by somehow preventing their transmission to another system that is required for consciousness. Rather, it weakens or degrades the representation of the stimulus, such that the representation does not have sufficient influence over the other, con­ current, patterns of activity in the brain to create a new global brain state into which the stimulus representation is integrated. This contrasts with the integration account of covert face recogni­ tion by Damasio (1 990), according to which the percept is nonnal but disconnected from other systems. A related interpretation seems to be ad­ vocated by Bisiach ( 1992). He rejects the notion of a single central locus of consciousness, which he identifies with Cartesianism and the homun­ culus, and calls our attention to the possibility of numerous parallel mechanisms linking stimuli and responses. The dissociations reviewed earlier can also be explained by a quality of representation account. According to this type of account, extinction and neglect result in poor-quality perceptual repre­ sentations, which supply input, albeit degraded input, to higher levels of conceptual and linguistic processing. This type of account is similar to Kinsbourne's integration account; both em­ phasize the degradation of perception and con­ sequent weakening of the influence of neglected and extinguished percepts on other parts of the

Martha J. Farah

226

system. The accounts differ in emphasis in that

fect of limiting the exposure duration of a word

the integration account stresses the lessening of

and following it with a mask is that the perceptual

influence on the rest of the system, whereas the

processing of the word is impaired. 1 Therefore,

quality of representation account stresses the ex­

we should not infer that if a word or picture can

istence of the residual influence.

semantically prime subsequent stimulus process­

The available data provide some constraints on

ing, that it must have been perceived normally.

possible mechanisms but do not discriminate de­

A mechanistic explanation of how poor-quality

cisively among them. The privileged role account

perceptual representations could produce pri­

of Volpe, LeDoux, and Gazzaniga can be dis­

ming at semantic levels is suggested by the covert

tinguished from the others in that it hypothesizes

face recognition model of Farah, O'Reilly, and

normal perception, with the impairment occur­

Vecera ( 1 993). In our model the locus of damage

ring downstream, whereas the other accounts

was visual, and the quality of the visual repre­

hinge on an impairment in perception per se. The

sentations was such that multiple-choice naming

finding of Farah, Monheit, and Wallace ( 1 99 1 )

was poor or at chance. However, the model

that same/different matching was dissociated

showed semantic priming. This was because the

from identification only by virtue of the demands

patterns of activation reaching semantic levels of

on visual perception is more consistent with the

representation contained partial, noisy, and sub­

integration and quality of representation ac­

threshold information about the semantic identity

counts than with the privileged role account.

ofthe stimulus. The resultant pattern in semantics

It might appear that several of the other find­

would, on average, be more consistent with the

ings described earlier are inconsistent with the

semantics of that stimulus, or a semantically re­

hypothesis example,

that perception McGinchy-Berroth

is

impaired.

et al.'s

For

lated stimulus, than with an unrelated stimulus,

subjects

and hence tended to have a net facilatory effect on

showed semantic priming from left field stimuli;

the semantic representation of identical or related

Sieroff, Pollatsek, and Posner's subjects could use

stimuli, relative to unrelated stimuli. However,

letters from neglected regions of the field to read

because much of this priming effect was caused by

words; and Berti et al. 's subject could match

subthreshold activation in the semantics units, the

extinguished stimuli on the basis of conceptual

semantic representations engendered by the poor­

rather than structural equivalence. However, as

quality visual input were not, themselves, able to

noted earlier, neural representation is not all or

activate response representations such as names.

none, and it is probably not necessary for the vis­

This suggests a reason why indirect tests may

ual system to have completed its processing of a

be more sensitive to the residual capabilities of

stimulus and to have derived a high-quality rep­

damaged systems: such tests generally require

resentation in order for it to pass some infor­

that residual knowledge affect the processing of a

A

probe stimulus within the perceptual and seman­

mation on to subsequent levels of processing.

degraded or incomplete visual representation will

tic layers, as opposed to requiring that knowledge

therefore presumably activate, to some degree,

to be propagated through additional levels of

conceptual and semantic representations. The possibility of semantic priming by words

representation for an overt response. The model of Farah, O'Reilly, and Vecera is sufficiently

that are poorly perceived is well established in the

simple and generic that it is equally relevant to

cognitive psychology literature. Marcel ( 1 983)

priming in neglect as in prosopagnosia.

and others have shown that subliminal tachisto­

The ability of neglect patients to use letters on

scopic presentations of words can prime judg­

the left sides of words to read, despite their in­

ments about subsequent supraliminal words. The

ability to identify letters in equivalent spatial

most natural and direct interpretation of the ef-

positions in nonword letter strings, is another

Visual Perception and Visual Awareness after Brain Damage

phenomenon that might at first seem to imply that perception has been spared. However, like semantic priming, it has been explained in terms of impaired perception and interactions among graded representations. Mozer and Behrmann (1 990) present a computer simulation in which an impairment of spatial attention results in im­ poverished stimulus information reaching stored word knowledge. By a completion-like process characteristic of distributed neural networks, the good-quality information from the right side of the word and the stored word structure combine with the poor-quality information from the left side of the word to "complete" or "clean up" the perceptual representation of the left side. Behr­ mann et al. ( 1990) present additional data from patients that support this account. Brunn and Farah ( 1 99 1 ) demonstrated that the top-down ef­ fects of stored word representations on perception include an attenuation of the neglect during word reading. For example, patients named more col­ ors, of different colored letters, when the letters made words than when they made nonwords. This implies that the preserved letter perception ofSieroff, Pollatsek, and Posner's neglect patients with words may not have been a case of preserved perception despite neglect but of a lessening of the neglect. There are no independent demonstrations that conceptual matching of the kind documented by Berti et al. ( 1 992) could be accomplished on the basis of impaired visual percepts. However, the same general principles invoked by Farah, O'Reilly, and Vecera ( 1 992) and Mozer and Behrmann ( 1 990) suggest that this finding, too, can be explained without hypothesizing normal visual perception and semantic representation, deprived of access to other brain systems neces­ sary for conscious awareness. With the relatively light demands on representational quality made by the same/different matching task ( Farah, Monheit, and Wallace 1 99 1 ) and the additional constraints provided by a limited stimulus set, it seems at least possible that a poor-quality visual representation could be used to make reasonably

227

accurate judgments of the kind made by Berti et al.'s subject. In addition, Berti et al. point out that perception of the extinguished stimulus on "same" trials may have been semantically primed by the ipsilesional stimulus. In sum, we find ourselves in roughly the same position with respect to explanations of uncon­ scious perception in neglect and extinction as we did with explanations of covert face recognition. The current body of empirical knowledge is of great interest and utility in establishing certain important qualitative conclusions: that patients can be unable to report the identity of a neglected or extinguished stimulus but still manifest a fair degree of perceptual and even semantic knowl­ edge about it. Yet these results are ambiguous with respect to the mechanisms by which neglect and extinction have these effects. The hypothesis that neglect and extinction spare perceptual and semantic processing, and affect only subsequent processing by some other system that plays a privileged role in mediating conscious awareness, seems inconsistent with the finding that per­ ception is not normal in neglect and extinction ( Farah, Monheit, and Wallace 1 99 1 ). Demon­ strations of semantic processing in neglect and extinction have not provided the kind of careful quantitation of effect sizes needed to determine whether semantic processing is normal or merely partially preserved. Thus, there is no decisive evi­ dence in favor of the privileged role account and a small amount of evidence against it. The available data appear compatible with the view of Kinsbourne ( 1988), Bisiach ( 1992), and Farah, Monheit, and Wallace ( 1 99 1 ), that the effect of neglect and extinction is to degrade per­ ceptual representation and thus attenuate its in­ fluence on the integrated state of the cognitive system as a whole. This conclusion could be strengthened, or disconfirmed, by further quanti­ tative empirical studies of the degree of preserva­ tion of implicit processing and by computational simulations of a fuller range of the experimental paradigms used to demonstrate unconscious per­ ception in neglect and extinction.

Martha J. Farah

6 Implicit Reading in Pure Alexia Patients with pure alexia are impaired at reading, despite being able to write normally and under­ stand spoken words. To the extent that they can read at all, they appear to do so in a letter-by­ letter fashion, spelling the word to themselves before they can recognize it. There are three main hypotheses about the cause of pure alexia and letter-by-Ietter reading: 1 . Disconnection account: Visual input from the occipital cortices is unable to access left hemi­ sphere language representations because of some combination of visual cortical damage and dis­ connection, and at most one letter at a time can be transmitted by alternative routes (e.g., Gesch­ wind, 1 965). 2. Word form hypothesis: Reading-specific knowledge of visual word structure has been de­ stroyed, or is partly disconnected from earlier levels of visual representation such as letter representations, so that word recognition must proceed letter by letter, without the benefit of larger-scale word structure (e.g., Warrington and Shallice 1 980). 3. Visual impairment hypothesis: The ability to encode multiple visual shapes of any sort rapidly and accurately, including but not limited to letters in words, is impaired, so that the most reliable strategy is to read letter by letter (e.g., Farah and Wallace 1 99 1 ). What the three hypotheses have in common is that there is some capacity essential for reading­ linguistic knowledge in general, word forms in particular, or the ability to encode multiple visual shapes-that incoming visual information cannot gain access to because of either the loss of that capacity or disconnection from it. These analyses suggest that patients' oral reading abilities reflect their comprehension abilities, and indeed the everyday behavior and clinical test performance of pure alexics suggest that they cannot under-

228

stand words that they fail to decipher by letter-by­ letter reading. This belief was first called into question by Landis, Regard, and Serrat (1980) in a case study of a patient recovering from surgery for a left oc­ cipital tumor. The patient was a pure alexic, and when tested within a week of the surgery, was unable to read words that were flashed for only 30 ms in a tachistoscope. However, he was able to point to objects in the room whose names had been presented in this manner. Unfortunately, he was tested with relatively few words on that day, and when retested a week later, he had lost his ability to point to the words' referents, even though his explicit reading performance had improved.

Representative Findings Subsequent studies have examined the implicit reading of pure alexics more thoroughly. Patter­ son and Kay ( I 982) reported several unsuccessful attempts to elicit evidence of comprehension of briefly presented unread words in their pure alexic subjects. Shortly after, Grossi et al. ( 1984) pre­ sented evidence of disproportionately preserved picture-word matching in one case of pure alexia, and Shallice and Saffran ( 1986) described a com­ prehensive investigation of implicit word recog­ nition and comprehension in another pure alexic. Shallice and Saffran's subject was able to perform lexical decision with relatively high accuracy on letter strings that were presented for 2 sec, which was too quickly for him reliably to identify the words explicitly. The subject was best at recog­ nizing high-frequency words, and the closer the resemblance was between non-words and words the harder it was for him to reject the non-words For high-frequency words, he was able to classify 90 percent as words, and false-alarmed to only 38 percent of the pseudowords derived from the high-frequency words by changing one or two letters. An interesting exception to his generally good lexical decision ability was that he was unable to discriminate appropriately and in-

:

Visual Perception and Visual Awareness after Brain Damage

appropriately affixed words, for example, calling

229

Awareness of Perception

applaudly a word. Shallice and Saffran also dem­ onstrated that their subject was able to make reasonably accurate semantic categorizations of words presented too briefly to be read. For ex­ ample, he correctly classified 94 percent of unread place names as in or out of Europe, 93 percent of unread people's names as authors or politicians, and 87 percent of unread concrete nouns as living or nonliving. Coslett and Saffran (1 989) replicated and ex­ tended these findings with four new cases of pure alexia. Like Shallice and Saffran (1 986), they found effects or word frequency on lexical deci­ sion and an insensitivity to affixes. They also found better lexical decision performance for concrete and imageable words and for content words, in general, relative to functors. In a sub­ sequent study, Coslett and Saffran (1 994) report that they tested one of their subjects on rhyme! nonrhyme judgments with written words and found that he performed at chance. Perhaps their most striking finding is that their subjects per­ formed the implicit reading tasks more accurately with extremely brief exposures of the words, such as 250 ms, than with exposures of 2 sec. They in­ terpret this in terms of the different strategies needed for implicit and explicit reading. Explicit letter-by-letter reading is incompatible with the strategy needed for implicit reading. Coslett and Saffran were able to foil subjects' attempts at letter-by-letter reading by using extremely short exposure durations and thereby enabled the al­ ternative strategy to be expressed. This is con­ sistent with Landis, Regard, and Serrat's original case study, in which implicit reading was lost as explicit reading improved. It also suggests a rea­ son that Patterson and Kay not have detected any preserved implicit reading in their SUbjects: in attempting to maximize the chances of eliciting implicit reading in their subjects, they used words that were most likely to be recognized explicitly­ that is, very short words and words that had been successfully read explicitly.

Although patients' reports of their subjective ex­ periences are not reported in great detail in most cases, there is a suggestion of some variability among cases. Whereas Landis, Regard, and Ser­ rat and Coslett and Saffran's subjects generally complained of not even being able to see the stimuli clearly when they were manifesting im­ plicit reading, Shallice and Saffran's subject was able to give rather detailed descriptions of how he performed the lexical decision and semantic categorization tasks. For lexical decision, he de­ scribed "a process of looking at the whole length of the word and finding a combination of letters that can't be right, or looking at the whole word and seeing that it looked sensible." For semantic categorization, he reported getting a "first im­ pression" of each word and, with a particular category in mind, deciding whether he "feels the first flash is right." Karen Klein and I have stud­ ied a similar patient, R.H., and with the help of Coslett and Saffran replicated all of their findings with him. In terms of his implicit reading perfor­ mance, he seems to differ from their subjects only in that he is able to perform implicit reading at any exposure duration and never needs prompt­ ing to abandon the letter-by-Ietter reading strat­ egy (at which he is extremely poor). Like Shallice and Saffran's subject, R.H. seems quite aware of the information he gleans during implicit reading and is reasonably confident of his responses. From this very small and sketchy empirical base, I would tentatively suggest that for pure alexics who are able to abandon the letter-by­ letter strategy and manifest implicit reading at relatively long exposure durations, there is no dissociation between the information that has been processed in implicit reading and awareness of having processed that information. The strik­ ing dissociation in these cases is between knowing the specific word presented and knowing its lexical status and semantic category. However, this dissociation does not have any direct implications

Martha J. Farah

230

for the relation between word perception and

conscious awareness (Schacter, McAndrews, and

awareness thereof. Only when exposure durations

Moscovitch 1 988). This is an example of a priv­

must be extremely short, for subjects who would

ileged role account.

otherwise persist in letter-by-Ietter reading, do

Implicit reading has also been explained by a

subjects manifest knowledge in their implicit

quality of representation account. Shal1ice and

reading performance that they themselves are not

Saffran (1 986) suggested that "lexical decision

aware of possessing. Under these circumstances,

above chance but well below normal levels; rec­

however, the dissociation between performance

ognition of morphemes but insensitivity to the

and awareness may be attributable to the same

appropriateness of their combination; limited ac­

mechanisms as subthreshold perception in nor­

cess to semantic information [and] the failure to

mal subjects, with thresholds being higher for

identify the stimulus explicitly, could conceivably

brain-damaged subjects.

be explained in terms of decreased levels of acti­

Explanations of Implicit Reading

explicit identification."

Because it is not clear whether implicit reading

possibility that implicit reading reflects the oper­

vation within the system that normally subserves Finally, Shallice and Saffran also discuss the necessarily involves a dissociation between sub­

ation of right hemisphere reading mechanisms,

jects' experience of reading and their perfor­

a view that Coslett and Saffran (1 989) later

mance, it is equally unclear whether explanations

endorsed. This constitutes a different type of

of making reference to the neural mechanisms of

privileged role account according to which the

conscious awareness will have any relevance to

normal left hemisphere component of the reading

the phenomena of implicit reading. Nevertheless,

system is uniquely endowed with the ability to

some of the proposed explanations can be applied

mediate conscious awareness of reading (if the

either to petformance/awareness dissociations or

account is to be applied to awareness) and with

simply to accounting for the surprising pattern of

the more fine-grained semantic distinctions and

performance in these patients.

syntactic and morphological capabilities found to

Recall that several account of pure alexia exist

be lacking in implicit reading.

and that they have in com mon some hypothesized

How does the available evidence adjudicate

capacity that is necessary for reading and is either

among these alternatives? The DICE model does

damaged or disconnected from early visual rep­

not provide an explanation of the peculiarities of

resentations of words. The capacity in question

implicit reading with regard to part of speech,

has been hypothesized to be language in general,

word from knowledge in particular, or the ability

morphology, and concreteness, and so on, and is therefore not a very satisfactory account of the

to encode multiple visual shapes rapidly, and in

phenomenon. In contrast, the right hemisphere

all cases is assumed to be localized to the left

hypothesis seems particularly well-suited to ex­

hemisphere. The different accounts of implicit

plaining these findings. On the basis of indepen­

reading can be differentiated from one another

dent evidence with split-brain patients and with

with respect to the role played by these left hemi­

normal subjects, the right hemisphere appears to

sphere mechanisms.

be better at reading high-frequency and concrete

The DICE model has been applied to implicit

words, and to be ignorant of morphology (see,

reading and suggests that implicit reading is

e.g., Baynes 1 990 for a review). The right hemi­

simply normal reading, drawing as usual on the

sphere is also believed to be deficient at deriving

critical left hemisphere-mediated processing, de­

phonology from print, consistent with the inabil­

prived of access to other systems necessary for

ity of one of Coslett and Saffran's subjects to

Visual Perception and Visual Awareness after Brain Damage

make rhyme judgments on pairs of printed words, which Klein and I also obselVed with our implicit reader. H is not clear exactly how well the hypothesis of a degraded reading system accounts for these phenomena. On the basis of general properties of neural network models, one might well expect higher-frequency words to be better preselVed, on the basis of more initial learning. One might also predict greater preselVation of concrete and m­ ageable words, especially with the assumptlon that such words have more associated semantic attributes ( Plaut and Shallice 1 993). On these points, the quality of representation approach does well in accounting for some of the features of implicit reading. In contrast, the more regular mapping between orthography and phonology, relative to the mapping between orthography and semantics, would seem to predict the relative preselVation of phonological judgments relative to semantic ones, just the opposite of what is ob­ selVed, on grounds similar to the prediction con­ cerning frequency effects: to the extent that sets of similar-looking words have similar-sounding pronunciations, the learning accomplished for each one will also generalize toward the others. Finally, the finding that shorter exposure dura­ tions lead to better implicit reading in at least some patients is awkward for a quality of repre­ sentation account. In sum, there is a good deal of relatively fine­ grained information about implicit reading c?n­ cerning the effects of different stimulus propertIes, instructions, and tasks, and these enable us to evaluate the different explanations of the phe­ nomena that have been put forward. Pending some counterintuitive results concerning the modes of failure in damaged reading networks, the available evidence seems most consistent with right hemisphere mediation of implicit reading. However, dissociated awareness is not an essen­ tial aspect of implicit reading, and the mechani�m of implicit reading is therefore not necessanly relevant to the neural bases of visual awareness.

�

7

23 1

General Conclusions

The goal of this final section is to review and compare the different syndromes in terms of the ways in which conscious awareness has been op­ 'erationalized and the mechanisms that seem most likely to account for dissociated awareness, and to consider what general conclusions might fol­ low about the relations between perception and conscious awareness.

Awareness and the Likely Mechanisms ofIts Dissociation from Perceptual Performance in the Four Syndromes There is a range of ways that the absence of patients' conscious awareness , has been d�mon­ strated in the four different syndromes reVIewed here. Within the literature on blindsight, there are two common types of introspective report. Sub­ jects often claim that they are gue�sing on he basis of no subjective experience, WIth resultmg very low confidence. In other instances, subjec s say that they are answering based on an expen­ ence but not one that they would classify as vis­ ual. Occasionally a specifically visual experience is reported. Prosopagnosic subjects are, of course, aware of when they are viewing a face. But they typically report no sense of familiarit� when they . view a face, perform poorly when trymg to Iden­ tify faces, and have low confidence in their iden­ tifications. In neglect and extinction, subjects may fail even to report detecting a stimulus. However, the only systematically collected data on aware­ ness in the experiments demonstrating uncon­ scious perception measure the ability to make various explicit judgments of such stimuli. Never­ theless, these data suggest a dissociation between awareness and performance, as explicit perfor­ mance is poor, and confidence is generally low. Finally, in implicit reading, the dissociation that holds for all patients is between the ability to re­ port a specific word and the ability to I?ake judg­ ments about the lexicality and semantic category

�

�

232

Martha J. Farah

of the word, not the dissociation between word perception and awareness of that perception. For two cases, at least, subjects are aware of the in­ formation they are using to make thesejudgments, and they have confidence in their answers. In sum, the four syndromes constitute a heteroge­ neous group from the point of view of awareness. The four syndromes also appear to be hetero­ geneous from the point of view of mechanism. I will tentatively suggest that the groupings of the syndromes according to the status of awareness in each syndrome correspond with the groupings one might impose on the basis of the currently most likely mechanisms. For example, in blind­ sight, there appear to be two main types of sub­ jective report, depending on the patient and task. There are also two main types of mechanism that have been proposed, without any intended mu­ tual exclusivity. There is evidence consistent with both subcortical mediation of some blindsight abilities, and extrastriate cortical mediation of others, interpretable with privileged role and quality of representation accounts, respectively. Whether these two mechanisms can, in fact, be placed in correspondence with the two types of subjective experience is an empirical question for which I know of no direct evidence. Covert face recognition in prosopagnosia and unconscious perception of stimuli in neglect and extinction are similar in that subjects are gen­ erally unaware of those aspects of the stimulus for which they manifest unconscious knowledge. The mechanisms responsible for the dissociations ob­ served in these two syndromes are not well estab­ lished because of a dearth of evidence capable of discriminating among them. However, both syndromes currently seem explainable in terms of a quality of representation account or, equiv­ alently, in terms of the version of Kinsboume's ( 1 988) integration account that postulates a fail­ ure of integration due to low-quality representa­ tion. In prosopagnosia the poor quality of the representations is due to a loss of stored percep­ tual knowledge of faces, whereas in neglect and

extinction it is due to a more dynamic processing failure of the attention system. Implicit reading seems distinct from all of the foregoing dissociations, both in terms of subjects' conscious awareness of the information they are able to extract from stimuli and in terms of mechanism. Unless word presentations are so brief that subjects report not seeing the words, implicit reading is accompanied by awareness of the information gleaned from the words. Al­ though the issue of mechanism is far from settled, currently the right hemisphere reading hypothe­ sis, a type of privileged role hypothesis, appears more plausible than a quality of representation hypothesis.

General Implications One general conclusion that should be abun­ dantly clear is that the four different syndromes reviewed here are unlikely to share a common explanation. Although there is a family resem­ blance among them, a close inspection reveals that the kinds of perceptual abilities and forms of experience that are preserved and impaired are not uniform. Ifwe are still inclined to characterize them as demonstrating perception in the absence of conscious awareness (which seems reasonable to me, at least for the first three), then we may need to consider the possibility that the relation between conscious awareness and neural systems is itself not explicable by just one type of account. For example, we might conclude that enabling conscious awareness is a privileged role of certain cortical regions and that within these regions, a certain quality of representation is also necessary (thus denying consciousness to both the normal­ quality functioning of the superior co11iculus and the impaired functioning of cortical face repre­ sentations). Among all of the syndromes, there is none for which visual perception, in its totality, has been convincingly demonstrated to be normal or near normal. Therefore, there is no reason to view

Visual Perception and Visual Awareness after Brain Damage

these syndromes as consisting of normal percep­ tion with conscious awareness merely stripped away. This, in tum, should deprive theories like DICE of their basic motivation: to explain a straightforward dissociation between perception and awareness of perception. Although the results of future research could reverse this situation, at present we lack evidence of such a straightfor­ ward dissociation. There is currently no evidence for a dedicated conscious awareness system, dis­ tinct from the systems that perform specific per­ ceptual or cognitive functions. The study of perception and awareness in brain-damaged patients is fairly new, particularly for syndromes other than blindsight, and our empirical knowledge of them is still rudimentary. Alan Allport urged in his 1 988 article, " What Concept of Consciousness?," that researchers in this area do more conceptual analysis of their subject matter. This is certainly a highly desirable goal, especially with a subject matter that so tempts us to cross and confuse explanatory do­ mains, such as the personal and subpersonal and the empirical and metaphysical. To this wise ad­ vice I would like to add a plea for more humble old empirical analysis as well. Are levels of per­ formance on implicit perceptual tasks. in these various syndromes normal? Is performance on explicit tasks poor but above chance, or truly at chance by a reasonably powerful test? How do these measures relate to the likelihood of report­ ing some awareness of the visual information be­ ing tested? What additionaIly can we discover about the more fine-grained charateristics of un­ conscious visual performance that might provide further constraints on mechanistic explanations (e.g., the part-of-speech effects in implicit reading, the relative sizes of interference and facilitation in priming by faces in prosopagnosia and neglected stimuli in parietal-damaged patients, or the nasal! temporal asymmetries in blindsight)? These ques­ tions must be answered if we are to distinguish among the different accounts of the syndromes reviewed here and thereby distinguish their asso-

233

ciated views of the neural correlates of conscious awareness.

Notes The writing of this chapter was supported by ONR grant NOOOI4-93-I0621 , NIMH grant ROI MH48274, NIH career development award K04-NSOI405, and a grant from the McDonnell-Pew Program in Cognitive Neuroscience. I thank several colleagues who read an early draft of this chapter and made many helpful sug­ gestions: Tom Carr, Max Coltheart, David Finkelstein, Morris Moscovitch, and Larry Weiskrantz. I . This interpretation has been questioned by Marcel, among others, because of the preservation of semantic processing. This reasoning appears to be based on the assumption that impairment in visual processing is in­ consistent with evidence of semantic processing. The simulation results described in the next paragraph dem­ onstrate that this assumption need not be true.

References Allport, A. (1988). What concept of consciousness? In A. 1. Marcel and E. Bisiach (Eds.), Consciousness in COl1lemporary science. Oxford: CI,uendon Press. Bauer, R. M. (1984). Autonomic recognition of names and faces in prosopagnosia: A neuropsychological ap­ plication of the guilty knowledge test. Neuropsycho­ logia, 22, 457-469. Bauer, R. M. (1986). The cognitive psychophysiology of prosopagnosia. In H. D. Ellis, M. A. Ieeves, F. New­ combe, and A. Young (Eds.), Aspects offace processing. Dordrecht: Martinus Nijholf. Bauer, R. M., and Trobe, I. D. (1984). Visual memory and perceptual impairments in prosopagnosia. Journal of Clinical Neuro-ophthalmology, 4, 39-46. Baynes, K. (1990). Language and reading in the right hemisphere: Highways or byways of the brain? Journal of Cognitive Neuroscience, 2, 1 59-179. Behrmann, M., Moscovitch, M., Black, S. E., and Mozer, M. (1990). Perceptual and conceptual mecha­ nisms in neglect dyslexia. Brain, I J3, 1 1 63- 1 1 83. Berti, A., Allport, A., Driver, 1 . , Dienes, Z . , Oxbury, I., and Oxbury, S. (1992). Levels of processing for visual

Martha J. Farah

234

stimuli in an "extinguished" field. Neuropsychologia, 30, 403-41 5.

ert recognition in a prosopagnosic patient. Cortex, 28, 77-95.

Bisiach, E. (1992). Understanding consciousness: Clues from unilateral neglect and related disorders. In A. D. Milner and M. D. Rugg (Eds.), The neuropsychology of consciousness. San Diego: Academic Press.

De Haan, E. H. F., Young, A. W., and Newcombe, F. (I 987a). Faces interfere with name classification in a prosopagnosic patient. Cortex, 23, 309-316.

Bisiach, E., and Rusconi, L. L. (1990). Break-down of perceptual awareness in unilateral neglect. Cortex, 26, 643-649. Brunn, J. L., and Farah, M. 1. (199 1 ). The relation be­ tween spatial attention and reading: Evidence from the neglect syndrome. Cognitive Neuropsychology, 8, 5975. Bruyer, R., Laterre, C., Seron, X., Feyereisne, P., Strypstein, E., Pierrard, E., and Rectem, D. (1983). A case of prosopagnosia with some preserved covert re­ membrance of familiar faces. Brain and Cognition, 2, 257-284. Burton, A. M., Young, A. W., Bruce, V., Johnston, R. A., and Ellis, A. W. (1991). Understanding covert rec­ ognition. Cognition, 39, 1 29- 1 66. Campion, 1., Latto, R., and Smith, Y. M. (1 983). Is blindsight an effect of scattered light, spared cortex, and near-threshold vision? Behavioral and Brain Sciences, 3, 423-447 . Chiarello, C. (1 988). Right hemisphere contributions to lexical semantics. Berlin: Springer-Verlag. Coslett, H. 8., and Saffran, E. M. (1989). Evidence for preserved reading in "pure alexia." Brain, 112, 327359. Coslett, H. D., and Saffran, E. M. (1994). Mechanisms of implicit reading in alexia. In M. J. Farah and G. Ratcliff (eds.), The neuropsychology of high-level vision: collected tutorial essay. Hillsdale, NJ: Erlbaurn Asso­ ciates. Cowey, A., and Stoerig, P. (1 989). Projection patterns of surviving neurons in the dorsal lateral geniculate nucleus following discrete lesions of striate cortex: Implications for residual vision. Experimental Brain Research, 75, 631 -638. Crick, F., and Koch, C. (1990). Function of the thala­ mic reticular complex: The searchlight hypothesis. Seminars in the Neurosciences, 2, 263-275. Damasio, A. R. (1 990). Synchronous activation in multiple cortical regions: A mechanism for recall. Sem­ inars in the Neurosciences, 2, 287-296. De Haan, E. H. F., Bauer, R. M., and Greve, K. W. (1992). Behavioral and physiological evidence for cov-

De Haan, E. H. F., Young, A. W., and Newcombe, F. (1987b). Face recognition without awareness. Cognitive

Neuropsychology, 4, 385-4 1 5. Farah, M. J. ( 1990). Visual agnosia: Disorders of object recognition and what they tell us about normal vision. Cambridge, MA: MIT Press/Bradford Books. Farah, M. 1., Monheit, M. A., and Wallace, M. A. (199 1). Unconscious perception of "extinguished" vis­ ual stimuli: Reassessing the evidence. Neuropsychologia, 29, 949-958. Farah, M. J., O'Reilly, R. c., and Vecera, S. P. (1993). Dissociated overt and covert recognition as on emergent property of lesioned attractor networks. Pychological Review, 100, 571 -588. Farah, M. l., and Wallace, M. A. ( 1 991). Pure alexia as a visual impairment: A reconsideration. Cognitive Neu­ ropsychology, 8, 31 3-334. Fendrich, R., Wessinger, C. M., and Gazzaniga, M. S. (1992). Residual vision in a scotoma: Implications for blindsight. Science, 258, 1489-149 1 . Gazzaniga, M. S . (1988). Brain modularity: Towards a philosphy of conscious experience. In A. 1. Marcel and E. Bisiach (Eds.), Consciousness in contemporary sci­ ence, Oxford: Clarendon Press. Geschwind, N. (1965). Disconnexion syndromes in ani­ mals and man. Part II. Brain, 88, 585-645. Gray, C. M., and Singer, W. (1989). Stimulus specific neuronal oscillations in the orientation columns of cat visual cortex. Proceedings of the National Academy of Science, 86, 1698- 1702. Greve, K. W., and Bauer, R. M. (1990). Implicit learn­ ing of new faces in prosopagnosia: An application of the mere exposure paradigm. Neuropsychologia, 28, 10351041 . Grossi, D., Fragassi, N. A., Orsini, A., De Falco, F. A., and Sepe, O. (1984). Residual reading capacity in a pa­ tient with alexia without agraphia. Brain and Language, 23, 337-348. Hinton, G., and Plaut, D. C. (1987). Using fast weights to beblur old memories. Proceedings of the Ninth An­ nual Meeting of the Cognitive Science Society, Hillsdale, NJ: Erlbaurn Associates.

235

Visual Perception and Visual Awareness after Brain Damage

Hinton, G. E., and Sejnowski, T. 1.

(1 986).

Learning

Rafal, R., Smith, 1.,

Krantz, 1., Cohen, A., and

and relearning in Boltzmann machines. In D. E. Ru­

Brennan, C

melhart and 1. L. McClelland (Eds.),

anopic humans: Saccade inhibition by signals in the

Parallel dis­ tributed processing: explorations in the microstructure of cognition. Cambridge, MA: MIT Press. Karnath, H. O.

(1988). Deficits of attention in acute and Neuropsychologia, 26,

recovered visual hemineglect.

27-43.

blind field.

(1 990).

Extrageniculate vision in hemi­

Science, 250, 1 I 8-1 2 1 .

Schacter, D. L., McAndrews, M . P., and Moscovitch, M.

(1988).

Access to consciousness: Dissociations be­

tween implicit and explicit knowledge in neuropsycho­ logical syndromes. In L. Weiskrantz (Ed.),

( 1987).

Karanath, H.-O., and Hartje, W.

Residual in­

without language.

Thought

Oxford: Oxford University Press.

(197 1 ).

formation processing in the neglected visual half-field.

Schiller, P. H., and Koerner, F.

Journal of Neurology, 234, 180- 1 84.

characteristics of single units in superior coIlicus of the

Kinsbourne, M.

(1988).

Integrated field theory of con­

sciousness. In A. 1. Marcel and E. Bisiach ( Eds.),

sciousness in contemporary science.

Con­

Oxford: Clarendon

Press. Landis, T., Regard, M., and Serrat, A.

(1 980).

Iconic

Brain and

brain tumor: A tachistoscopic study.

Lan­

guage, 11, 45-53. Marcel, A. 1.

(1983).

Conscious and unconscious per­

Cognitive Psychology, 15, 1 97-237.

( 1986).

Marzi, C A., Tassinari, C., Aglioti, S., and Lutzem­

(1986).

from a letter-by-Ietter reader.

Lexical processing

Cognitive Neuropsycho­

logy, 3, 429-458. Sieroff, E., Pollatsek, A., and Posner, M.

Spatial summation across the vertical

meridian in hemianopics: A test of blindsight.

Neuro­

psychologia, 24, 749-758.

I. (1 988).

M., Alexander, M., and Kilduff, P. T.

(1993).

Semantic

processing in the neglected visual field: Evidence from a

Cognitive Neuropsychology, 10,

79- 108.

(1 992).

Memory and the hippocampus: A

synthesis of findings from rats, monkeys and humans.

Psychological Review, 99, 1 95-23 1 . (1 990). Wavelength sensi­ Nature, 342, 916-918. Damasio, A. R. (1985). Knowledge

Stoerig, P., and Cowey, A. tivity in blindsight. Tranel, D., and

ognition by prosopagnosics.

Science, 228, 1453- 1454. (1988).

Tranel, D., Damasio, A . R., and Damasio, H .

Intact recognition of facial expression, gender, and age in patients with impaired recognition of face identity.

Neurology, 38, 690-696.

Mozer, M. C, and Behrmann, M.

(1990).

On the inter­

action of selective attention and lexical knowledge: A connectionist account of neglect dyslexia.

Journal of

Cognitive Neuroscience, 2, 96- 1 23. Patterson, K. E., and Kay, 1.

(1982).

Letter-by-Ietter

reading: Psychological descriptions of a neurological

Quarterly Journal of Experimental Psychol­ ogy, 34A, 41 1-441. Plaut, D. C . , and Shall ice, T .

(1993).

Deep dyslexia: A

study of connectionist neuropsychology.

Cognitive

Neuropsychology, 10, 377-500. Poppel, E., Held, R., and Frost, D.

Volpe, B. T., LeDoux, 1. E., and Gazzaniga, M. S.

(1979).

Information processing of visual stimuli in an

"extinguished" field. Wallace, M. A.

syndrome.

Nature, 282, 722-724.

( 1994).

glect and extinction.

Unconscious perception in ne­

In M . 1. Farah and G. Ratcliff

( Eds.), The neuropsychology of high-level vision: Col­ lected tutorial essays. Hillsdale, Nl: Erlbaum Asso­ ciates. Wallace, M. A., and Farah, M. 1.

Residual

Nature, 243, 2295-2296.

(1992).

Savings in re­

learning face-name associations as evidence for "covert recognition" in prosopagnosia.

(1 973).

visual functions after brain wounds involving the cen­ tral visual pathways in man.

Cognitive

Neuropsychology, 5, 427-449.

without awareness: An autonomic index of facial rec­

McGlinchey-8erroth, R., Milberg, W. P., Verfaellie,

lexical decision task.

the posterior visual spatial attention system. Squire, L.

(1988). Blindsight and insight in visuospatial neglect. Nature, 336, 766-767. Marshall, 1. C, and Halligan, P. W.

case

Shallice, T., and Saffran, E.

Recognition of visual letter strings following injury to

ception: Experiments on visual masking and word rec­

berger, L.

Journal of Neurophysiology, 36,

920-936. in the absence of explicit word identification: Evidence

reading in a case of alexia without agraphia caused by

ognition.

alert rhesus monkey.

Discharge

Journal of Cognitive

Neuroscience, 4, 1 50- 1 54. Warrington, E. K., and Shallice, T. dyslexia.

Brain, 103, 99- 1 1 2.

( 1980).

Word-form

Martha J. Farah

Weiskrantz, L. (1986). Blindsight: A case study and im­ plications. Oxford: Oxford University Press. Weiskrantz, L. (1990). Outlooks for blindsight: Explicit methodologies for implicit processes. Proceedings of the Royal Society of London, 8239, 247-278. Weiskrantz, L., Sanders, M. D., and Marshall, J. (1974). Visual capacity in the hemianopic visual field following a restricted occipital ablation. Brain, 97, 709728. Young, A. W., Hellawell, D., and De Haan, E. H. F. (1988). Cross-domain semantic priming in normal sub­ jects and a prosopagnosic patient. Quarterly Journal of Experimental Psychology, 38A, 297-3 18.

236

12

Understanding Consciousness: Clues from Unilateral Neglect and Related Disorders

Edoardo Bisiach . . . if we did not have phenomenal experience we would not have a concept ofconsciousness at all. -Anthony J. Marcel

Suspicion about consciousness as an object of scientific inquiry is rooted in the cultural preju­ dice due to which we regard physics as the para­ digm of science par excellence. This prejudice may lead us to legislate away subjective experi­ ence from the scientific domain-a solution with which I have formerly sympathized ( Bisiach 1 988)-or vice versa (e.g., Nagel 1 974), depend­ ing on where the emphasis is laid. Alternatively, it may induce outright denial of the subjectivity of consciousness, as exemplified by Rorty's state­ ment, "The problem is not to take account of the special difficul ty raised by the 'inside' -character of consciousness, but to convince people that there is no reason to grant that persons have more of an inside than particles" ( Rorty 1 982, p. 1 8 3). Not many people, I surmise, are likely to be persuaded by Rorty. In fact, although I would hardly describe myself as having inside experience of the particles that constitute the complex physi­ cal system I am, I strongly oppose defining as "fictional" or "eliminable" the inside experience inherent in some states and processes of that sys­ tem. This does not imply regarding that experience as a mysterious non-physical accompaniment of those states and processes. That experience is my "point of view." That's the way it is. The bat, I presume has its own (Nagel 1 974; chapter 32). Why on earth should they be "eliminated?" People who resist programmatic repudiation of their own consciousness may legitimately develop a specific scientific interest in it. They may enquire as to how phenomenal experience is structured; what distinguishes conscious from nonconscious states and processes besides subjective experience, that is, in neural and/or information processing terms; what cf''1ditions determine access to, and recess from, that experience; how conscious and nonconscious states and processes interact; and

so forth. The case of blindsight (Weiskrantz 1 986) has been very instrumental in disciplining the urge and clarifying the scope of scientific inquiry on consciousness. Recognizing consciousness as a proper object of scientific inquiry entails extending the para­ digm of science beyond the standards of tradi­ tional physics. So what? It has often been claimed that the study of the mind is an "immature" dis­ cipline, but it would be more accurate to view it as the most problematic and least explored frontier of science, and it is probably our general con­ ception of science that needs further maturation.

1

Methodological Issues

After having sanctioned, at least in principle, the project of studying one's own consciousness em­ pirically, something should be said about tactics, expectations, and boundary conditions, while bearing in mind that it would be a mistake to be too prescriptive and exacting at the start. The only method I can envisage is the ascription of phenomenal experience to an observed individual, by analogy with the phenomenal experience of the observer. Commonplace as it might appear at first sight, this move is exceptional in science: what it affords, indeed, is neither an anthropo­ morphic metaphor nor a theoretical construct. It is not a metaphor because what is ascribed is meant to apply literally to the recipient. It is not a theoretical construct because the observer's con­ sciousness is not such, and to the extent that the ascription of phenomenal experience to the ob­ served individual is not simply intended to explain his or her behavior. What the move consists in, is the hypothetical ascription to another individual of a property, that the observer asserts to be real of him- or herself on the basis of inside infor­ mation. This is the first step one must make in an attempt to gain insight into one's own conscious mentality and, secondarily, into that of other

Edoardo Bisiach

238

people. The transaction may be reversed, as it

perimental manipulations of its own brain pro­

were, in the sense of the observer's hypothetical

cesses a gifted bat could devise, they would not

self-ascription of the behavior displayed by the

tell it the whole story about why being a bat is like

observed individual, so that the question becomes

what it is. There cannot be a science of those

what would it be like to be me, if I displayed such

aspects of phenomenal experience that are ma­

behavior. An advantage of this reversal of per­

nipulated in thought experiments such as qualia

spective is that it tallies with the condition in

inversion. Our problem is to understand what,

which I would directly learn about my conscious­

besides

ness if I were a subject, for example, of Marcel's

tivities of a system, in themselves and with respect

(1 983) experiments on non-conscious priming.

to any other activity of the same system.

such aspects, distinguishes conscious ac­

While this condition can be realized, granting

A more serious disadvantage of the move at

phenomenal experience to other people is the only

issue is its being based, as it were, upon intro­

possibility we are given to learn

spection by proxy. This is a severe handicap,

indirectly

about

consciousness when we study the consequences of

considering the halo of disrepute surrounding

brain lesions (unless we were

direct introspection itself. For those who are not

ourselves

suffering

from dysfunctions such as blindsight or unilateral

frightened away from the outset, this means in­

neglect). Trying to figure out what it would be like

vesting a great deal of labor in an intricate trig­

to be the person whose behavior we are studying,

onometry of the mind while still not being in a

is not the "first step in the operationalist sleight

position to form a correct estimate of the validity

1 980, p. 45 1 ). On the contrary,

of its results; it also requires the utmost caution in

it prevents that step and its unavoidable con­

of hand" (Searle

assessing such results. 'Introspection', to make

sequence, namely, the replacement of conscious­

things worse, is a potentially misleading term. On

ness by a conventional, impoverished, and more

the one hand, it may induce questionable linguistic

or less rigid substitute such as a piece of behavior

constructs (such as 'consciousness of conscious­

or some putative functional components of the

ness' or 'reflexive consciousness') that suggest

cognitive machinery (see Shallice

1 988, p. 40 I , for

an enigmatic higher order consciousness rather

a catalogue of proposals for such components, as

than the mundane act of conscious thinking

well as for his own account). On the other hand,

about conscious experience. On the other hand, it

no matter how problematic and tricky it might be,

may be misused or misunderstood as meaning

reference to the observer's phenomenal experi­

(fallible) self-interpretation. I use the term 'in­

ence obviates the need for the "objective" criteria

trospection' in the sense of bona fide self-wit­

in the absence of which sceptical conclusions are

nessing and I think that what can first of all be

drawn about consciousness as a proper object of

expected from a scientific approach to conscious­

empirical investigation (e.g., by Allport,

ness is a closer view of what is being witnessed

1 988).

The move I am advocating has obvious dis­ advantages. First of all, it could be claimed that

and a better estimate of the veridicality of its report.

it misses "the real thing," that is, what it would actually be like to be the observed individual (or what it would actually be like to be oneself as subject of the observed behavior). To the extent to which it is qualia-reiated, this claim is a truism with which it would be otiose to bother, because it misses the point altogether. There is no way in which a natural science of consciousness could have anything to do with qualia. Whatever ex-

2 The General Content of Comciousness and Its Spatiotemporal Structure As regards what is self-witnessed, we must make a distinction between externally driven (i.e., per­ ceptual) and internally driven phenomenal expe­ rience, such as mental images and thought. The

239

Understanding Consciousness

first kind of experience is strictly constrained by,

Both kinds of phenomenal experience, percep­

and anchored to, the properties of the perceived

tual and internally driven, are spatio-temporal.

environment. The second is conceivable inherent

They are spatial in two senses:

in the largely parallel activities of a network of

and

"fuzzy-jittery" neuronal assemblies whose sur­ facing in consciousness is

enthymematic,

information processing.

phenomenological

In the phenomeno­

logical sense, their sensory or sensory-like content

in the

extends in an egocentric space; a space that is

sense that it leaves out a great deal of what gives

likely to be represented by the brain in an ana­

meaning to what we are aware of. This is what

logue medium ( Bisiach and Berti, 1 987). Within

makes internally driven experience almost im­

this space, consciousness may either be focal

possible to describe except when we endeavor

or (borrowing the term from Kinsbourne 1 988)

to entertain a relatively stable mental image or

panoramic, or focal with a panoramic periphery.

crystallize an idea into the format of inner speech.

In the information processing sense, phenomenal

The notion of Juzzy-jittery

experience has spatial structure insofar as it is

cell assemblies

is an

admittedly vague expansion of Hebb's ( 1 949)

most likely to be assembled from activities dis­

seminal proposal. Roughly, the term is intended

tributed among centralized processors. These

to capture and combine four ideas;

processors work more or less in parallel, follow­

I . The huge expanse of the busy neural network in which representational activity is implemented and which is beyond any realistic hope of pan­ oramic assessment by neurophysiological means. 2. The imprecise correspondence of a cell assem­

ing a complex interactive schedule by which, no matter how stabilized it might appear in a short interval of time, the content of consciousness is continually restructured. In saying that conscious experience is the result of processing distributed in neural space, I do not

bly involved in thinking about a certain object to

simply refer to the horizontal separation of the

the cell assembly activated during perception of

various sensory modalities, nor to the fact that,

that object.

within single modalities, there are spatially orga­

3. The co-occurrence of different cell assemblies,

nized neural maps conforming to spatial and even

some of which may have segments of the relevant circuitry in common. 4. The instability of endogenous clusters of neural representational activity.

to non-spatial features of environmental stimuli (see, e.g., Merzenich and Kaas 1 980, as regards the distribution of intensity- and pitch-sensitive cells in the auditory cortex). I refer in particular to the vertical organization along the input-output

1-4 are sug­

axis that offers the possibility of sensory-premotor

gested to be the sketchiness and possible overlap

interaction in the generation of conscious experi­

of the fleeting conscious experiences resulting

ence. The dependence of consciousness on such

from endogenous neural activity at any given

an interaction is theoretically most significant and

Phenomenal correlates of points

instant. As an explanatory construct, the notion

is suggested, by way of example, by the effect of

of fuzzy-jittery cell assemblies has the obvious

response modality on awareness of sensory stim­

problem of being too versatile. It is hoped that its

uli. This effect may give rise to phenomena that,

main virtue may be that of warning against some

unlike other phenomena of interference such as

models of the brain's representational activities,

backward masking, are not explicable in terms

offered by AI-inspired theorists, that are often too

of a modulation of incoming information at

clear-cut and rationalistic inasmuch as they relate

a response-independent preconscious level, and

more to the final product than to the process of

therefore question any pre-theoretical intuition of

thought.

consciousness as a flat interface dividing efferent from afferent processes.

Edoardo Bisiach

Neuropsychological investigation offers several examples of such an effect. In right brain-dam­ aged patients, awareness of contralesional visual stimuli may to a variable extent be impaired by damage to visual pathways or due to the dys­ function(s) underlying unilateral neglect of space. So far, it is often impossible to assess the contri­ bution of each of these factors to the disorder. However, it has been found that awareness of contralesional stimuli may appear more impaired if manual choice-reactions, rather than verbal responses, are used to monitor detection (Bisiach et al. 1 989). In our tachistoscopic experiment pa­ tients had to react to elementary visual stimuli200-ms illumination of a single diode in either hemifield, or of two diodes, one in each hemi­ field-either by verbally reporting the number of stimuli, or by pressing a key if they had perceived a single stimulus, witholding their response to double stimuli and on blank trials. Whatever may be the correct explanation of the reduced rate of detection of contralesional stimuli we recorded in right brain-damaged patients-both in condi­ tions of single and in conditions of double simul­ taneous stimulation-when a manual as opposed to verbal response was required, this result shows that the different ways in which sensory infor­ mation has to be processed in order to result in different kinds of intentional activity may differ­ ently affect access to consciousness. A further example is provided by another ex­ periment in right brain-damaged subjects suffer­ ing from unilateral neglect (Bisiach et al. I 990a). It is well known that such patients usually bisect horizontal lines to the right of the objective mid­ point. In our experiment, we asked patients to execute the task by means of a pulley device op­ erating in two different conditions: a congruent (canonical) condition in which a pointer was directly moved from either end of the line to its subjective midpoint, and a non-congruent con­ dition in which the pointer was indirectly moved along the line by a hand movement in the opposite direction. The amount of the rightward bisection error was found to decrease, to an extent varying

240

among patients, in the non-congruent as opposed to the congruent condition. We interpreted the result as revealing two different factors of neglect, respectively related to input analysis and pre­ motor programming. The latter corresponds to Watson et al.'s ( 1978) concept of 'directional hy­ pokinesia', that is, reluctance to carry out actions toward the contralesional side of egocentric space. More precisely, we concluded that in the congruent condition of the task perceptual and premotor factors contributed to displacing the SUbjective midpoint in the same direction, that is, rightward, whereas in the non-congruent con­ dition the rightward displacement of the sub­ jective midpoint resulting from faulty perceptual analysis was to a variable degree counteracted by the inertness of the leftward hand-movement re­ quired to move the pointer rightward. Here too, phenomenal experience appears to be influenced by variables that relate to premotor programming of intentional actions, that is, to a stage tradi­ tionally conceived to be later than the repre­ sentation in consciousness of sensory information on the basis of which actions are initiated. We found the premotor component of the dysfunc­ tion to be relatively more pronounced in patients whose lesions extended to the frontal lobe, or to subcortical structures held to have motor func­ tions, than in patients with exclusively retro­ rolandic lesions. As regards neglect, this is in agreement with a distinction conjectured by Watson et al. (1 978) and Mesulam ( 1 98 1). As regards consciousness, this differentiation affords a glimpse of the anatomical arrangement of the distributed mechanism by which input and output stages of information processing interactively frame perceptual judgments. A dependence of perceptual awareness on response mechanisms is also suggested by the results of a recent investigation by Tegner and Levander ( 1 99 1). Eighteen patients with left ne­ glect were asked to execute a variant of Albert's (1 973) cancellati�n task in which the subject has to cross out with a pencil a number of short lines drawn on a sheet of paper. Patients had to per-

Understanding Consciousness

form first in the canonical condition, and then in a condition in which the visual array was left-to­ right inverted by means of a 90° mirror device, while direct view of their hand and of the sheet of paper was prevented by a bench. In this con­ dition, the subject's hand must operate on the left side of the sheet of paper to cross out the lines he sees on the right side of the visual display, and vice versa. In the canonical condition, as usual, the patients only crossed out lines lying on the right side. The inverted condition revealed three different patterns of behavior. Ten patients only crossed out the lines they could see on their right; to do that their (right) hand had to cross the midline and work on the left side of the sheet of paper. Their neglect was therefore interpreted as being due to perceptual factors. Four patients, on the contrary, only crossed out lines lying on the right of the sheet of paper (as they had done in the canonical condition), although those lines were seen through the mirror as lying on the left side. In these patients neglect appeared thus to be due to a (broadly conceived) premotor dysfunc­ tion. The remaining four patients, in the mirror­ inverted condition, only crossed out the lines in the middle. Their performance was interpreted by the authors as being caused both by perceptual and premotor factors of neglect. In agreement with our results, Tegner and Levander found evi­ dence of 'directional hypokinesia' only in pa­ tients whose lesion involved the frontal lobe or the basal ganglia. Perhaps "directional hypokinesia" is also re­ sponsible for the fact, observed by Halligan and Marshall (unpublished; cited by Allport 1 988), that a neglect patient may seem unaware of the full extent of the horizontal line he is asked to bi­ sect, whereas he seems to be aware of it when he has to use that same line as the base of an imagi­ nary square of which he is asked to draw the right-hand (vertical) side. There are still other examples of the role of response mechanisms in gating access to con­ sciousness in patients with unilateral neglect. Joanette et al. (1 986) and Halligan and Marshall

241

( 1 989) described patients whose ability to detect left-side stimuli was more impaired when they had to respond with their right hand than when they had to respond with their left. Finally, an impressive demonstration of the influence of response factors on reports of visual stimuli in blindsight has been given by Zihl and Von Cramon ( 1 980). Their study was replicated and expanded both in blindsight and normal subjects by Marcel ( 1990). In his Experiment 2, Marcel asked his patient GY to react either by blinking or by pressing a button with his right forefinger, or by saying "yes" when he "felt" that a light had come on in his blind field. Blink responses were found to be more sensitive than finger responses; verbal responses were least sensitive. Similar results were obtained in two normal subjects by using near-threshold visual stimuli. 1 The data so far reviewed suggest that conscious experience of external stimuli cannot be separated from, and is affected by, the whole situation in which they are perceived, including any response to such stimuli (no matter whether spontaneous or, as in most experimental conditions, comply­ ing with instructions). If so, we must conclude that the mechanisms by which conscious experi­ ence of a relatively complex situation is achieved are indeed widely distributed in neural space and subdivided into several processors. None of these, so far as we know from observation of brain­ damaged subjects, centralizes in itself the role of conscious homunculus. The division of preconscious labor among sev­ eral processors with no sole gateway to con­ sciousness entails a relativity of the timing of consciousness as well as the possibility of ongoing rearrangement of what is being experienced. In normal people, a rearrangement of this kind is suggested by the results of an unpublished ex­ periment by Cumming, cited by Allport (1 988). In that experiment, a row of five letters was pre­ sented to the SUbjects. The letters flashed one after another in rapid sequence and in a spatio-tempo­ ral order such as to produce masking of the letters

Edoardo Bisiach

in the second and fourth locations in the row. Subjects had to press one key if a target letter was present and another key if it was not. Under time pressure, when the target letter was in a masked position subjects on some occasions correctly re­ sponded by pressing the "present" key and then apologized for what they believed was an error. The phenomenon is similar to the abortive movement which, under conditions of double simultaneous stimulation, I observed some brain­ damaged patients to make toward the stimulus undergoing extinction before indicating the loca­ tion of the antagonist stimulus. In both cases, however, the problem arises as to whether the subjects had responded on the basis of a fleeting sensory experience or had produced a non-con­ sciously triggered reaction. The kind of problems arising from the dia­ chronic structure of conscious experience is illus­ trated by the following example. In an experiment done some years ago (Bisiach et al. 1985), pa­ tients with left hemineglect had to react to single 200-ms illuminations of red or green diodes in either visual hemifield by pressing a lit response­ key of the same color with their right forefinger. There were four response keys, one red and one green on each side of the apparatus which lay in front of the patient. On each trial, two keys of different colors, one on the left and one on the right, came on simultaneously with the flash of light and remained lit until one of them was pressed. Stimuli were flashed during visual fixa­ tion on the center of the frontal panel of the ap­ paratus; after their appearance, the patient was free to move his eyes in any direction. There were four stimulus-response conditions, two spatially uncrossed and two crossed, with 1 6 trials for each of them (left-side stimuli requiring left-side re­ sponse; left-side stimuli requiring right-side re­ sponse; etc.). Catch trials in which the response keys were turned on in the absence of any stim­ ulus were interspersed among regular trials. The rate of correct reactions to left (contralesional) stimuli was found to be more severely reduced when the response was required on the left-side

242

key than when required on the right-side key. Reactions to stimuli appearing in the right visual hemifield (i.e., in the hemifield unaffected by the brain lesion) were almost errorless when the re­ sponse was required on the right side; when it was required on the left side, however, the patients made many errors. The latter were mainly due to an erroneous ipsilateral response, either to the unlit key of the correct color or to the lit key of the incorrect color, but in several instances the patients gave no response at all, as if they had not perceived the stimulus in the intact visual field. This paradoxical phenomenon was most evi­ dent in patient FS, who had a lesion involving the frontal lobe and anterior subcortical structures. Whereas his performance was errorless on trials in which right-side stimuli required responses to the right, on the 1 6 trials on which he should have reacted to the same type of stimuli by pressing the left key, FS gave only one correct response. On several such occasions he asserted, as he did on catch trials, that no stimulus had appeared. Given his correct performance on trials with right-hand stimuli requiring responses to the right side of the panel, it is plausible to assert that he had been phenomenally aware of those stimuli, at least for a brief interval, also when he denied the occur­ rence of some of them in trials requiring responses to the left side. What is not possible to say is whether any trace of phenomenal experience per­ sisted at the moment when he uttered his denial. Perhaps, all conscious memory of the stimulus had been suppressed by inhibitory processes caused by not having found the appropriate response-key on the preferred side. Alternatively, his denial could have been confabulated, as sug­ gested by Shallice (1 988, p. 398) over a preserved phenomenal experience of the stimulus. It is worth noting that a paradoxical denial of stimuli present on the ipsilesional side of egocentric space has also been found by Tegner and Levander in the above mentioned study. Those of their pa­ tients who showed 'directional hypokinesia', after having crossed out in the mirror-inverted condition only the lines which lay on their right

Understanding Consciousness

but were seen on their left, denied the presence of further lines, thus ignoring those which were present on the right side of the visual array. If one considers these puzzling data, Allport's (1988) doubts about the possibility of giving a full answer to some questions about consciousness appear justified. Nevertheless, it is evident that this sceptical conclusion is drawn from data that tell us a great deal about what phenomenal expe­ rience consists of.

3 Unity and Disunity of Consciousness One may still feel inclined to play down the im­ port of the distributed origin of phenomenal ex­ perience suggested by the data reviewed above and claim that what really matters is the coher­ ence of that experience. However, the unity that at first glance distinguishes most of everyday phenomenal experience breaks down progres­ sively if we take into consideration the transition from illusions to delusions and other more dra­ matic dissociations, some of which, though of marginal occurrence, are not of a pathological nature. A touch of dissociation is already present in common perceptual illusions, where phenom­ enal experience may peacefully coexist with the conscious belief in its lack of veridicality. The dissociation becomes much more impressive when an illusion overwhelms a belief of which the ra­ tionality is still acknowledged by the subject. I have two favourite examples. One is the patient mentioned by Hecaen (1 972), who remarked on her inability to be convinced by her own sight that the upper limb joined to the left side of her body belonged to her. The other is LA-O, a patient of mine. Shortly after the onset of a right hemi­ sphere lesion, while taking for granted the own­ ership of her left shoulder, she inferred that the left upper arm was also her (because it was at­ tached to her left shoulder), but was unsure about her elbow and lower arm and firmly refused to admit that the left hand was part of her body. During recovery, she seemed to entertain con-

243

temporaneously opposing doxastic attitudes to­ ward the same proposition, as can be observed in patients recovering from schizophrenia: aver­ ments of true and false beliefs about the owner­ ship of her left upper limb appeared in close temporal proximity and were equally suffixed, as it were, with a short-lived look of tense puzzle­ ment. It might be objected that what occurs in such instances is not a split of phenomenal awareness but an alternation of normal and abnormal expe­ riences which does not undermine the notion of the unity of consciousness. There are, however, situations in which the dissociability of phenom­ enal experience seems undeniable. Dreams, in­ deed, are sometimes experienced as real and unreal at the same time, as happens when one finds solace in the arriere-pensee that the terrify­ ing event one is involved in is in fact a dream. A double awareness, in dreaming, may sometimes be likened to two different characters on the stage. This is exemplified by the following per­ sonal anecdote. I once dreamt that I was being told by my teacher Alexsandr Romanovich Luria that a certain patient had undergone an intra­ carotid sodium amytal test a few days earlier at the Burdenko Neurosurgical Institute. "Which side?", I asked. Alexsandr Romanovich hesitated an instant and then answered, somewhat dryly: "Both sides, of course." On which I felt ashamed of my silly question: of course, I "remembered" patients always undergo sodium amytal injection on both sides when a test is required. The above example, however, demonstrated that, even when under particular circumstances the content of awareness reveals the disunity of its source, a great deal of integration may be achieved through mechanisms we are still very far from knowing. In my dream, a mechanism of some sort had done the job of the playwright, giving coherence to incompatible pieces of cogni­ tive activity through the expedient of dramatiza­ tion, that is, by meaningfully dividing between two individuals what would have been incoherent in one. In an old-time party game, a haphazard

Edoardo Bisiach

cluster of words may directly combine or be edited into surrealist prose such as "Le cadavre exquis goutera Ie vin nouveau." Likewise, when the elements of what is going to be the content of phenomenal experience swarm from the endoge­ nous activity of the nervous system in a disorderly way, they might be integrated by the intervention, as it were, of the brain's editorial staff. In the awareness of the external world, the coherence of phenomenal experience largely reflects the coher­ ence of the environment; a slight amount of re­ touching is evident in phenomena such as the visual completion occurring in the blind spot, the phoneme restoration effect (Warren 1 970), or intriguing illusions (reviewed by Dennett and Kinsbourne 1 99 1 ) such as the color changes dur­ ing illusory movement or the "cutaneous rabbit" phenomenon. In the case of internally generated experiences, on the other hand, coherence is probably achieved with reference to a model set by previous exposure to the environment and through induction; this leaves much room for in­ ventiveness but imposes constraints, vestiges of which are retained even in dreaming and schizo­ phrenia. Conformity with such a model is likely to require a great deal of pruning and readjusting of preconscious activity, whereas it may admit nonveridical intrusions in the recollection of previous experiences. At any rate, the process of coherent cross-integration of phenomenal experi­ ence does not require a mind within the mind­ let alone a mind above the mind-any more than a resident architect is required for the shaping of a cluster of H20 molecules into a snow crystal. The result of this process may be a more or less complete temporary relaxation achieved through a distributed counterbalance of different compet­ ing representations arising within the expanse of the neural network. The fact that in wakefulness thought processes are radically different from those occurring in dreaming suggests that there are progressive levels of cross-integration. This, however, does not imply the endorsement of an intelligent homunculus responsible for the highest level: (relative) rationality may still be conceived

244

to result from a virtual machine distributed across a vast expanse of the cerebral cortex.

4 Levels of Processing of Conscious Representations The foregoing attempt to clarify what it means for phenomenal experience to have spatio-tem­ poral structure and to dispel homuncular pre­ conceptions still leaves unanswered the question as to what level of processing characterizes the representation of individual contents of con­ sciousness. This aspect of the problem has been discussed at length by lackendoff (1 987). If I understand it correctly, his Intermediate-Level Theory of Consciousness states that: (a) the con­ tents of phenomenal experience are sensory­ specific; (b) as regards perceptual awareness, their structure is neither an unelaborated neural re­ production of the most peripheral stimulation pattern, nor a "central" (categorical and amodal) interpretation of that same pattern. In other words, the contents of phenomenal experience consist of parsed and selected information pre­ serving the instant, modality-specific properties of objects in a singular, definite orientation with re­ spect to the sensory surface. Thus, for example, phenomenal awareness of a word is held by lackendoff to correspond to the phonological structure of that word, rather than to its acoustic spectrum or to its meaning, and phenomenal awareness of a visual object to Marr's concept of "2�-D sketch" of that object, rather than to a "3-D model" of the latter. I largely subscribe to lackendoff's views, to which, however, I will add a few comments. It is true that, in the normal case, we are not separately conscious of, for example, the inter­ mediate product of the dedicated processors that in the case of vision analyze in parallel single fea­ tures of objects such as shape, location, color, and motion. However, when damage to all but one of such processors leaves a single feature being processed, awareness of that isolated feature is

Understanding Consciousness

245

still possible, as for example in patient BRA,

qualify such activities as 'abstracting' (rather than

studied by Warrington ( 1 985, and personal com­

'abstract') and 'generative', and call them 'pro­

munication). While being severely impaired as

cedures' involved in the construction of the con­

regards visual acuity, shape perception, and vis­

tents of phenomenal experience, rather than

ual location due to brain infarctions incurred

'structures' (in order to avoid confusion with that

during open-heart surgery, BRA was fully aware

which is being constructed).

of the color of an object though unable to attrib­

Finally, granted that what lackendoff calls

ute definite shape or location to that "feeling" of

"intermediate-level structure" is the skeleton of

color.

what we experience, the question remains open as

As regards the lack of any phenomenal experi­

to the extent to which such a structure can be

ence of more abstract structures such as types or

teased apart from what goes beyond it, that is,

categories, even within single sensory domains,

from the unavoidable experiential halo aroused

this might be due to the fact that such conceptual

by associations of any kind. As soon as I happen

entities, as opposed to tokens, have no existence

to hear "If" played by Art Tatum, my experience

in the brain, except as procedures available to

differs in important respects from what it would

catalogue

be if I were born in the 1 960s.

or

generate,

when required, modality­

specific instances of any category through activa­

Within the specific scope of this chapter, how­

tion of determinate cell assemblies. The applica­

ever, the main problem is whether reaching an

tion of procedures of this kind to incoming

"intermediate-level structure" is a sufficient con­

sensory information might underlie capabilities

dition for access to consciousness or, if it is not,

such as perceptual constancies and intersensory

what else is needed. Do "nonconscious" masked

matching. Their endogenous activation might

primes in experiments such as those of Marcel

give rise to mental images, dreams, and halluci­

(1 983) and Allport et al. ( 1 985) fail to access con­

nations. lackendoff argues that if phenomenal

sciousness despite having been processed up to an

experience of a pereceived token went beyond the

intermediate-level structure or do they immedi­

intermediate-level structure (which, in the case of

ately recede from it, without leaving any trace, as

t

vision, would be something like Marr's "2 D

a consequence of backward masking? To what

sketch") we would not be conscious of a seen cube

extent are neglect patients nonconscious of stim­

as something different from a felt cube. But this

uli lying on the side of space contralateral to their

might be due to the fact that, whatever might be

lesioned hemisphere when their behavior shows

the ultimate structural level of conscious percepts,

that such stimuli are processed up to the level at

our being conscious of a seen cube as something

issue?

different from a felt cube depends on lower-level

My patient EB, who suffered from right brain

but not super­

damage, was apparently unaware of the left side

higher-level structures. On the other

of written words and pronounceable non-words,

structures being organized into,

seded by,

hand, there is no earmark left by earlier sensory

so that, for example, she read

processing upon the contents of phenomenal

or

RAGIONE

(reason) or

STAGIONE

CARNAGIONE

(season)

(complex­

awareness, on the basis of which to explain why

ion); yet she would correctly name the letters

an

composing the left side of unpronounceable non­

imagined or hallucinated cube is experienced

XRTMNGIONE,

as seen or felt rather than amodal. Activities of

words such as

the mind underlying uninstantiated concepts such

longing to the three categories were randomly

per se,

although stimuli be­

among the

interleaved ( Bisiach et ai, 1 990b). Did the left

contents of phenomenal experience. We may

side of words and pronounceable nonwords have

as 'cubeness' do not figure, know them, as it were, by

our own

description,

short-lived permanence in consciousness, or did

rather than by acquaintance. This is why I would

they have no access to it at all? These questions

Edoardo Bisiach

seem to take for granted that access to con­ sciousness is an all-or-nothing step. Probably it is not. In the case ofEB we had an observable input, that is, the written word, and an observable out­ put, that is, the spoken word. What was in be­ tween is conjectural. Earlier stages of visual processing were evidently unimpaired, as demon­ strated by her ability to react differently to pronounceable and unpronounceable strings of letters. Yet, when she read real words and pro­ nounceable nonwords the output did not match the input. A possible explanation of EB's be­ havior is that what occurred in her head was the wrong construction of an incremental hypothesis about the input. Due to the dysfunction under­ lying her heminegiect, that hypothesis was even­ tually dominated by the right side of the input, thus giving rise to phenomena of neglect or com­ pletion, unless an interrupt was brought about by the strangeness of an unpronounceable string of letters. However, given the data, we cannot say anything definite about EB's phenomenal experi­ ence during the process lead,ing up to the output, all the more so because there is no reason to be­ lieve that processes of that kind are linear. All we can say is that, as in the case of nonconscious priming, infonnation that has reached a relatively advanced stage of cognitive processing may not have access to (or pennanence in) consciousness. In unilateral neglect a dissociation between graphemic and word awareness may even be found. Whereas we have no definite evidence about conscious experience of the content of the graphemic buffer as regards EB's neglect and completion dyslexia, RB, a neglect patient de­ scribed by Hillis and Caramazza ( 1 990), was at least temporarily aware of graphemes in the word-segment undergoing negiect or completion; invited to read words (e.g., VILLAGE), he would first spell them correctly and, immediately after, make a neglect or completion error in reporting the whole word (MILEAGE) . Patients suffering from unilateral neglect may provide further evidence about the possible lack of consciousness of visual infonnation the pro­ cessing of which has reached what lackendoff

246

calls "intennediate level." This evidence comes from simple experiments in which two drawings are shown, one above the other, differing in a conspicuous detail on the left (neglected) side. When it is established that the patient is unable to detect that difference, he is invited to indicate his preference for one of the two drawings. If he ob­ jects to the silliness of being asked to opt for one oftwo identical drawings, it is necessary to induce him to choose. The preference test is repeated a number of times, with one of the two drawings placed at random above or below the other. In the original experiment by Marshall and Halligan (1 988), the drawings were outlines of two houses, of which one had bright red flames on the left side; asked which house she would prefer to live in, the patient, despite having found the two pic­ tures identical, consistently chose the non-burn­ ing house. We attempted to replicate Marshall and Halligan's results in four patients, with three additional pairs of drawings (Bisiach and Rus­ coni 1 990; see fig. 12. 1). The preference test was not always possible, either because some patients were able to indicate the relevant difference in a pair, or because they were not able to identify some of the drawings. When it was possible, the result was not always in agreement with that ob­ tained by Marshall and Halligan. Of the three patients who were asked to show their preference for one of the two houses, one gave inconsistent responses and two consistently preferred the burning house. On the contrary, the unbroken glass was consistently preferred by the two pa­ tients who could be tested with that pair. The pot containing flowers was also preferred in all trials by the only patient to whom that test could be given. Our patients offered spontaneous com­ ments on their choices, which appeared in some instances to be made on the basis of real (though negligible) differences on the right side of two paired drawings. In other instances, however, the alleged difference was clearly confabulated. The extent of the confabulation was impressive when it referred, for example, to the different number of windows or the different layout of the two houses. Taken together, the results suggest that infor-

Understanding Consciousness

Figure 12.1 From Bisiach and Rusconi

1 990.

mation about a prominent detail, the presence of which is not acknowledged by the patient, may influence his appraisal of the configuration in which that detail is located. In some instances, information about that detail appears to be processed up to the level at which meaning is captured.

5 Breakdown of Awareness of Stimuli in Central Vision Our replication of Marshall and Halligan's (1988) study was supplemented by a very elementary task from which we obtained unexpected results.

247

In an attempt to force the patients to find the rel­ evant difference existing between the two draw­ ings of each pair, we asked them to trace the contour of each drawing with the tip of their right forefinger. Altogether, we were able to collect the results of 1 0 administrations of this tracing task. In one instance, the procedure was effective in revealing the differing detail to the patient; its effectiveness was dubious in two, but in the re­ maining seven the correct execution of the task was of no help. Hard to believe as it might ap­ pear, a patient could slide his fingertip through the flames without noticing them, or trace the contour of the left side of the broken glass with­ out realizing that it was broken. The phenomenon we observed on the tracing task is most significant because it reveals a severe disorder in the very area to which the patient's gaze is directed. The patients never lost sight of the path their finger had to follow; not even when that path was half concealed, as was the case with the drawing in which flames were superimposed on the outline of the house. Nevertheless, they seemed unaware of the left-side differences they came across in tracing the contour of the draw­ ings, even when-as in the case of the two houses or of the two flowerpots-such differences lay, for an interval of time, to the right of the moving fingertip. This finding suggests that neglect, in some instances, may be related to the (observer­ relative) left side of a visual object, rather than to retinotopic co-ordinates, or other self-centered co-ordinates, independent of the content of the visual field. It is well known that people can engender vari­ able gradients of spatial attention independent of the line of their gaze. Therefore, and in agreement with Kinsboume's theory ( 1987), neglect patients' covert attention is likely to be polarized toward the ipsilesional side of the visual field and to decrease-with a more or less continuous gra­ dient-toward the fixation point and beyond, reaching its minimum in the most contralesional area. Some experiments by Posner and associated (Posner et al. 1 982) indeed suggest that in uni­ lateral neglect covert attention cannot easily be

Edoardo Bisiach

248

disengaged from the ipsilesional side when con­

suggested by what Crowne et al. ( 1 98 1 ) found

ative attention should be addressed to a more

in the monkey after unilateral frontal eye-field

contralesional area. In left hemineglect, this may

ablation. Before fully recovering from the effects

explain the paradoxical extinction of a parafoveal

of the operation, the monkeys reacted to menac­

stimulus in the right (unaffected) half-field, when

ing stimuli approaching from either side by re­

a more peripheral (i.e., rightward) stimulus is si­

treating to the opposite side of the cage. However,

that half-field ( Bisiach

if two such stimuli were simultaneously pre­

and Geminiani 1 99 1 ). Accordingly, it may be

sented, one on each side, the avoidance reaction

multaneously presented in

conjectured that an inability to disengage covert

appeared only to the ipsilesional stimulus, so that

attention from the right side of a drawing might

the animal retreated near to the contralesional.

cause extinction-like phenomena as regards left­

(Preoperatively,

side details even when such details come into

double avoidance and the monkey would retreat

central vision. Obviously, however, an extinction

to the middle of the cage.) "Frequently," the au­

double

stimulation

caused

hypothesis cannot explain the patients' failure

thors reported, "the operated monkeys would

to discover any left-side difference in a pair of

turn their heads briefly to look at the proximate

drawings even when the difference lies in a con­

threat object, but immediately returned to regard

tour they are able to trace. On the other hand, the

the distal threat. These contralateral-gaze glances

failure can hardly be interpreted as being due to a

were not accompanied by shifts in position in the

faulty comparison depending on a disorder of

cage" (p. 1 79). This is all the more significant be­

visual memory restricted to contralesional space:

cause when the monkey looked at the contrale­

in the pictures of the burning house, the broken

sional threat the ipsilesional was out of sight, thus

glass, and the tom banknote, indeed, the differing

ruling out any explanation in terms of

details were anomalies worthy of attention in

competition. Evidently, however, the ipsilesional

their own right.

threat was not out of mind, otherwise the mon­

Although what we have learned from blind­

sensory

keys would have reacted to the contralesional as

sight warns against forming rash opinions as re­

they would if it had been presented alone. Full

gards the amount of phenomenal experience of

awareness of the contralesional threat thus ap­

sensory information guiding willed movements, it

peared to be barred by competition at a higher

would seem preposterous to assert that an un­

representational level .

familiar outline could be traced by the patient's finger with no perceptual awareness of the edge along which the finger has to proceed. It seems more reasonable to suggest that

mindful proces­

sing of the right side of a visual object shaded, in our patients, into

absent-minded processing of the

left side, where sensory information would fail to achieve integration into a fully fledged conscious representation. This might also be the case when the object is the patient's own body. A common explanation would thus apply to our findings relative to the tracing task and to the puzzling instances in which neglect patients deny their hemiplegia even when they look at the examiner's demonstration of its presence. A similar phenomenon is likely to affect per­ ception of the whole environment. This is strongly

6 Unilateral Distortion of the Content of Subjective Experience The contralesional disorder of representational activity appearing as a consequence of damage to crucial areas of one side of the brain may not only manifest itself in the form of the more or less complete eclipse from consciousness of one side of egocentric space (within one or more domains of representational content) that characterizes unilateral neglect in the strict acceptance of the term. Productive phenomena such as pathologi­ cal completion, confabulation, and somatopar­ aphrenia (Gerstmann 1 942) may also emerge. For reasons that are still unknown, such pro-

249

Understanding Consciousness

ductive phenomena are most infrequent as re­ gards the content of far extrapersonal space, al­ though some patients claim, for example, that the side opposite to the hospital environment they correctly perceive on their right belongs to their horne and behave in accordance with this claim (e.g., by calling for members of the family they assume to be there). In near extrapersonal space, pathological completion of one side of perceived words and drawings is much more frequent, but the imported portion is, as a rule, totally com­ patible with what is correctly reported from the other side. It is worth noting that in this case the phenomenon seems to be the result, as it were, of conceptual confabulation rather than of sensory completion; this, at least, is the impression one gets from the reading disorders of my patient EB and patient RB of Hillis and Caramazza ( 1 990). Processes of a different kind are likely to underlie the paradoxical leftward displacement of the subjective midpoint observed in neglect patients when they bisect very short lines ( Halligan and Marshall 1 988; Tegner et al. 1 990). One of these patients reported seeing a "shadow-line" extend­ ing leftward from the line he had to bisect (Tegner, personal communication). Productive misrepresentation of the contrale­ sional side is much more frequent in bodily space. Its implications for the study of neural correlates of consciousness have been discussed elsewhere (Bisiach and Geminiani 1991) and I will not deal with them here, except to attenuate my former opposition to a psychodynamic interpretation of unawareness and denial of neurological disorders such as hemiplegia and hemianopia, and to make a cursory remark on the impetus, as it were, of contents that are gated from consciousness in these pathological conditions. I still believe that at the root of unawareness or denial of hemiplegia and hemianopia lies a dys­ function of the brain's representational activity circumscribed to one side of the body, rather than the concealment of extremely disturbing evi­ dence. However, I agree that, at least as regards denial of hemiplegia and disavowal of the para-

Iyzed limbs, part of the patient's behavior reflects the need to corne to terms with the pathological representation of one side of his or her body. That is to say that the contents of the patient's con­ sciousness corresponding to manifestations rang­ ing from "personification" of the paralyzed limbs (Critchley 1 955) to the most abstruse forms of somatoparaphrenia are likely to combine active and reactive components; the former being the direct product of the lesion of a critical area of the neural network, the latter issuing from the spared areas of that same network. An issue that has so far been neglected is whether and to what extent the phenomenal ex­ perience and the behavior of the patients who are oblivious of neurological disorders, or utterly deny them, is influenced by contents that find no access to conscious representation. Although its assessment requires further data and reflection, this issue is worth attention, given the results of investigations in normal subjects that demon­ strate different utilization of sensory input de­ pending on whether or not its content reaches consciousness (Marcel 1 980). Contents held in incommunicado might still influence phenomenal experience in patients showing unilateral neglect and related disorders. Such nonconscious con­ tents, rather than an (originally) content-free emotional attitude caused by the lesion, might be responsible for the phenomena oflevophobia that can sometimes be observed in patients with left hemineglect and/or misrepresentaton. Misople­ gia (Critchley 1 953), that is, hatred for the para­ lyzed limbs, is a well-known instance of levophobia. The latter, however, can also mani­ fest itself in regard to the extrapersonal envi­ ronment, for example, when the patients assert a dislike for things on their left, or simply behave as if they felt such a dislike.

7

Conclusions

In this chapter, I have argued-against the Cartesian tradition ( Fig. 1 2.2)-that there is no

Edoardo Bisiach

Figure 12.2 The pineal gland

( L) in Vesalius's De Humani Corporis

Fabrica (1 543).

definite brain center for conscious experience (sections 2 and 3). Clinical and experimental data collected in patients suffering from unilateral spatial neglect and related disorders, indeed, sug­ gest that conscious experience rests entirely on a virtual mechanism distributed over brain circuits that are still very poorly understood. This distri­ bution, however, seems to be both horizontal (across and within sensory modalities) and verti­ cal insofar as conscious experience appears to be sh ped not only by afferent occurrences but also by the kinds of actions that happen to be under­ taken on the basis of such occurrences. The spread of such mechanisms in neural space entails the possibility of local failures and the dissoci­ ability of conscious experience into contrasting states. Clinical neurology provides dramatic ex­ amplex of both kinds of disorder: the former is evident in the (cognitively irremediable) lack of awareness of one side of space or of the con­ sequences of a brain lesion; among those most

�

250

suggestive of the latter are the mirror-reversal of the lack of awareness of one side of visual space found by Tegner and Levander and the doxastic doubleness that may affect particular beliefs on a background of normal cognitive activities ( Bisiach and Geminiani 1 99 1 ). In section 4, I touched on the issue of the level of processing and representational contents (as regards the conscious/nonconscious dichot­ omy) and argued that patients may be apparently unaware of what lies on the side opposite to the brain lesion despite its having been processed up to a high level, even to the level, perhaps, at which meaning is captured. It might be objected that neglect and pathological completion of the contralesional side of visual stimuli could be the result of a false revision of earlier conscious ver­ idical percepts, as suggested by the correct read­ ing of the left side of unpronounceable nonwords by my patient EB; an "Orwellian" revision, as Dennett ( 1 99 1 ) would put it. The validity of this conjecture, however, is seriously questioned by the finding that neglect patients may adequately react to features of stimuli of which they seem to be totally unaware even while they focus attention on those features: such patients may in fact be quite consistent in "liking" the drawing of an unbroken glass better than that of a broken one while being unable to tell the difference between the two of them even when they trace the contour of the broken edge (section 5). All in all, there seem to be at present no available data about the precise way in which conscious and nonconscious representational structures might differ. Finally, I raised (without even attempting to answer it) the question as to whether anomalous experiences about the contents of the contrale­ sional side of personal and extrapersonal space might be an effect of nonconsciously processed features of such contents (section 6). This, I believe, is one of the main tasks left for future investigation. Another fundamental task for future inves­ tigation is to ascertain to what extent the failure underlying unilateral neglect and misrepresenta-

Understanding Consciousness

tion is due to structural unavailability (or sub­ version, or disconnection) of the representational medium, to withdrawal of processing resources from it, or to its inhibition by competing brain processes. Firmer grounds will be available for comprehensive theories of the brain mechanisms of conscious representation, in proportion to the development of our knowledge of the mecha­ nisms underlying neglect and related disorders, as well as those that underlie their compensation. Let us now turn to the crucial issue (section 1). To what extent do the data reviewed in this chapter contribute to an understanding of con­ sciousness? If we select each time just one of the possible operational definitions of consciousness (e.g., the ability to report verbally the contents of our mental processes), then it is clear that such data tell us a great deal. Conflicting conclusions, however, seem to be reached if we refer to dif­ ferent operational definitions, since we have confronted instances in which what is verbally reported fails to be witnessed by nonverbal re­ sponses (Bisiach et al. 1 989) and instances in which the opposite is true (e.g., Marcel 1 990). But here indeed seems to lie the main moral we can draw from the corpus of empirical data so far collected: any intuition to the contrary notwith­ standing, consciousness is far from being unitary. The aim stated at the beginning of the chapter, however, was somewhat more ambitious. It made explicit reference to the understanding of phe­ nomenal experience, rather than of an operation­ alized consciousness. Now there are, as we have seen, extreme instances in which trying to make inferences about the phenomenal experience of a patient may seem to be hopeless, given the degree of incoherence of the observed behavior. Such is the case, for example, as regards the delayed de­ nial of visual stimuli in the unaffected field by patient FS (Bisiach et al. 1985), or the denial of lines lying in the unaffected visual field by Tegm!r and Levander's patients showing "directional hypokinesia." There is much less difficulty as re­ gards the different rates of perceptual report that characterize different kinds of response. Here, the

25 1

observed behavior may be quite similar to that of normal subjects tested with near-threshold stim­ uli. In this case, the evaluation of data in terms of phenomenal experience is much safer and any re­ sidual doubt could be cleared up by first-person experience. Between these two extremes lies an area of uncertainty where reference to phenom­ enal experience, no matter how utopian in its aims, might at least guide retreat to an intelligent, non-eliminative operationalism.

Acknowledgments The preparation of this chapter has greatly bene­ fited from my participation in the workshop on Human Consciousness and the Brain. I gratefully acknowledge my debt to Dan Dennett who orga­ nized the workshop, the small group of partic­ ipants who contributed in making it productive and pleasurable, and the Rockefeller Foundation, who gave us magnificent hospitality at the Villa Serbelloni in Bellagio at the end of March, 1 990. David Milner, Mick Rugg, and the Russell Trust! Wellcome Trust Symposium gave me the incen­ tive to lay down, for what they are worth, some of the ideas that have grown out of my work with brain-damaged patients. Joe Bogen, Ray Jacken­ doff, and Tony Marcel made very valuable com­ ments on a preliminary draft; I hope I have not misinterpreted them in the final version.

Note I . It must be noted that Marcel used simple nonverbal reactions, whereas we ( Bisiach et al. 1989) used choice nonverbal responses. TIlls may explain why nonverbal indicators of perceptual awareness were more sensitive than verbal indicators in his experiments and less sensi­ tive in ours (cr. the interpretation we gave in our paper).

References Albert, M. L. ( 1 973). A simple test of visual neglect. Neurology, 23, 658-664.

Edoardo Bisiach

Allport, D. A. ( 1 988). What concept of consciousness? In A. J. Marcel and E. Bisiach (eds), Consciousness in Contemporary Science. Oxford: Clarendon Press. AIlport, D. A., Tipper, S. P., and Chmiel, N. R. J. (1985). Perceptual integration and postcategorical filtering. In M. I. Posner and O. S. M. Marin (eds), A ttention and Performance XI. Hillsdale NJ: Erlbaum. Bisiach, E. (1 988). The (haunted) brain and conscious­ ness. In A. J. Marcel and E. Bisiach (eds), Consciousness in Contemporary Science. Oxford: Clarendon Press. Bisiach, E., and Berti, A. (1987). Dyschiria. An attempt at its systemic explanation. In M. Jeannerod (ed.), Neu­

rophysiological and Neuropsychological Aspects of Spa­ tial Neglect, Amsterdam: North-HoIland. Bisiach, E., and Geminiani, G. ( 1 991). Anosognosia re­ lated to hemiplegia and hemianopia. In G. P. Prigatano and D. L. Schacter (eds), A wareness of Deficit after Brain Injury, pp. 1 7-39. New York: Oxford University Press. Bisiach, E., and Rusconi, M. L. (1990). Break-down of perceptual awareness in unilateral neglect. Cortex, 26, 643-649. Bisiach, E., Berti, A., and VaIlar, G. ( 1 985). Analogical and logical disorders underlying unilateral neglect of space. In M. I. Posner and O. S. M. Marin (eds), A tten­ tion and Performance XI. Hillsdale NJ: Erlbaum. Bisiach, E., Vallar, G., and Geminiani, G. ( I 989). In­ fluence of response modality on perceptual awareness of contralesional visual stimuli. Brain, 1 12, 1 627-1636. Bisiach, E., Geminiani, G., Berti, A., and Rusconi, M. L. (1990a). Perceptual and premotor factors of uni­ lateral neglect. Neurology, 40, 1278 - 1 28 1 . Bisiach, E., Meregalli, S., and Berti, A. ( l990b). Mech­ anisms of production control and belief fixation in human visuospatiai processing: clinical evidence from unilateral neglect and misrepresentation. In M. L. Commons, R. J. Hermstein, S. M. Kosslyn and D. B. Mumford (eds), Quantitative Analyses of Behavior. Vol.

IX· Computational and Clinical Approaches to Pattern Recognition and Concept Formation. Hillsdale NJ: Erlbaum. Critchley, M. ( 1 953). The Parietal Lobes. London: Hafner Press. Critchley, M. ( 1 955). Personification of paralyzed limbs in hemiplegics. British Medical Journal, 2, 284-286. Crowne, D. P., Yeo, C. H., and Steele RusseIl, I. (1981). The effects of unilateral frontal eye field lesions in the

252

monkey: Visual-motor guidance and avoidance behav­ iour. Behavioural Brain Research, 2, 1 65-1 87. Dennett, D. ( 1 991). Consciousness Explained. Boston: Little, Brown. Dennett, D. and Kinsboume, M. ( 1 991). Time and the observer. Behavioral and Brain Sciences, I S, 183-247. Gerstmann, J. (1 942). Problems of imperception of disease and of impaired body territories with organic lesions. Archives of Neurology and Psychiatry, 48, 8909 1 3. Halligan, P. W. and MarshaIl, J. C. (1988). How long is a piece of string? A study of line bisection in a case of neglect. Cortex, 24, 321-328. Halligan, P. W. and Marshall, J. C. (1989). Laterality of motor response in visuospatial neglect: A case study. Neuropsychologia, 27, 1 301 -1307. Hebb, D. O. (I 949}. Organization of Behavior. New York: Wilcy. Hecaen, H. (1972). Introduction a la Neuropsychologie. Paris: Larousse. Hillis, A. and Caramazza , A. (1 990). The effects of attentional deficits on reading and spelling. In A. Car­ amazza (ed.), Cognitive Neuropsychology and Neuro­

linguistics: Advances in Models of Cognitive Function and Impairment. Hillsdale NJ: Erlbaum. Jackendoff, R. (1987). Consciousness and the Computa­ tional Mind. Cambridge MA: MIT Press. Joanette, Y., Brouchon, M., Gauthier, L. and Samson, M. (1 986). Pointing with left vs right hand in left visual field neglect. Neuropsychologia, 24, 391 -396. Kinsboume, M. (1987). Mechanisms of unilateral ne­ glect. In M. Jeannerod (ed.), Neurophysiological and Neuropsychological Aspects of Spatial Neglect. Am­ sterdam: North-Holland. Kinsboume, M. (1988). Integrated field theory of con­ sciousness. In A. J. Marcel and E. Bisiach (eds), Con­ sciousness in Contemporary Science. Oxford: Clarendon Press. Marcel, A. J. (1980). Conscious and preconscious rec­ ognition of polysemous words: Locating the selective effects of prior verbal context. In R. S. Nickerson (cd.), Attention and Performance VIII. Hillsdale NJ: Erl­ baum. Marcel, A. 1. ( 1 983). Conscious and unconscious per­ ception: Experiments on visual masking and word rec­ ognition. Cognitive Psychology, I S, 197-237.

Understanding Consciousness

Marcel, A. J. (1 988). Phenomenal experience and func­ tonalism. In A. J. Marcel and E. Bisiach (eds), Con­ sciousness in Contemporary Science. Oxford: Clarendon Press. Marcel, A. J. (1 990). Slippage in the unity of con­ sciousness, or, to what do perceptual speech acts refer? Poster presented at Russell Trust Symposium, St An­ drews, Scotland, September 1990. Marshall, J. C. and Halligan, P. W. (1 988). Blindsight and insight in visuo-spatial neglect. Nature, 336, 766767. Mesulam, M.- M. (1981). A cortical network for directed attenton and unilateral neglect. Annals of Neurology, 10, 309-325. Merzenich, M. M. and Kaas, J. H. ( 1980). Principles of organization of sensoriperceptual systems in mammals.

Progress in Psychobiology and Physiological Psychol­ ogy, 9, 1 -4 1 . Nagel, T . (1 974). What i s i t like t o be a bat? Philosoph­ ical Review, 83, 435-450. Posner, N. I., Cohen, Y. and Rafal, R. D. (1 982). Neu­ ral systems control of spatial orienting. Philosophical Transactions of the Royal Society. B, 298, 1 87-198. Rorty, R. (1982). Comments on Dennett. Synthese, 53, 1 8 1 -187. Searle, J. R. ( 1980). Minds, brains and programs. Be­ havioral and Brain Sciences, 3, 41 7-457. Shallice, T. (1988). From Neuropsychology to Mental Structure. New York: Cambridge University Press. Tegner, R. and Levander, M. (1991). Through a look­ ing glass. A new technique to demonstrate directional hypokinesia in unilateral neglect. Brain, in press. Tegner, R., Levander, M. and Caneman, G. (1 990). Apparent right neglect in patients with left visual ne­ glect. Cortex, 26, 455-458. Vesalius, A. (1 543). De Humani Corporis Fabrica. Basel: 1. Oporinus. Warren, R. M. (1 970). Perceptual restoration of missing speech sounds. Science, 1 67, 392-393. Warrington, E. K. ( 1 985). Visual deficits associated with occipital lobe lesions in man. Pontificiae Academ­ iae Scientiarum Scripta Varia, 54, 247-261 . Watson, R . T., Miller, 8. D. and Heilman, K. M . (1978). Nonsensory neglect. Annals of Neurology, 3, 505-508. Weiskrantz, L. ( 1986). Blindsight. A Case Study and Implications. Oxford: Clarendon Press.

253

Zihl, J. and Von Cramon, D. ( 1980). Registration of light stimuli in the cortically blind hemifield and its effect on localization. Behavioural Brain Research, I , 287-298.

13

Modularity and Consciousness

Tim Shallice

1

The Relevance of Modeling Consciousness

There is an even more severe conceptual dif­ ficulty for the modular view of mind than how it can operate efficiently. The efficiency problem can at least be posed in a "mechanistic" frame­ work closely related to the framework in which "modularity" itself is expressed. A more complex issue is why the human cognitive-processing system, which is apparently modular, should have the property of being conscious, unlike most modular systems-for example, present­ day complex machines. The obvious strategy within a modular ap­ proach is to identify some aspects of the operation of some particular module-say, its input-as conscious experience. However, one is then faced with the question of what could be so special about the processing in that module as to give its input such exceptional status. No real progress appears to have been made. One appears �er�ly to be taking the first step on the road to mfinite regress. The situation is worse than it appears. Not only is there no apparent line of attack on how and why a modular system might be conscious, but an explanation of consciousness within the con­ ceptual framework ofmodularity would probably need to be "functionalist" (Putnam 1 960). In other words, consciousness would correspond to some "system-level" property (i.e., information­ processing characteristic), of the brain and not t� some aspect of its material constituents. In walt for any explanation of this form that might be produced, there already lie the prepare tank traps of philosophical criticism. Attemptmg to refute a functionalist explanation is at present a popular enterprise for philosophers (e.g., Block 1980; Searle 1 980). 1 There is a simple and basically behaviorist response to these types of conceptual difficulty.

�

Worry about consciousness has historically bee� a snare and a delusion in psychology. What, It might be asked, is so special about the syste�­ level of explanation of human thought that It needs to incorporate consciousness? After all, a number of philosophical positions on con­ sciousness are in conflict with the idea that consciousness has any particular relation to an information-processing level of explanation of cognition. Consciousness has been conceived of as a property of non-material entities, as in clas­ sical dualism, of more molecular constituents of the brain, such as matter itself, and also of per­ sons (e.g., Strawson, 1959), all in conflict with any priority for an information-processing explana­ tion. Why should cognitive psychology and neu­ ropsychology not progress as nicely as, say, neurophysiology without being concerned about consciousness? The obvious response is that consciousness exists. Historically, there have been other phe­ nomena as basic and obvious that in their time have been as inexplicable as the existence of con­ sciousness or were explained in a way antithetical to modem scientific thought. Life is an obvious example, but planetary motions or fire would be others. It seems ahistorical to assume that con­ sciousness is some freak phenomenon totally re­ sistant to a scientific account. Yet for such an explanation to be achieved, the existence of awareness must be interpretable in terms of some scientific level of explanation of human cognition or of the more basic processes that are required for it to operate. No progress has been made in linking consciousness with concepts relev�nt to these more basic processes, those of phYSICS, biology, or physiology. 2 The conceptual frame­ work that cognitive psychology offers is tantaliz­ ingly different. The names of its subject areas­ memory, attention, perception, and so on-come from terms denoting mental states. In its practice, cognitive psychology even requires that subjects

Tim Shallice

understand such terms in order to carry out the instructions ofexperiments, and it uses procedures that require subjects to comment on their mental states (see, e.g., Ericcson and Simon 1 985). The most convincing concrete support for the need to understand consciousness in system terms would be provided by the existence of unexpected phenomena within the domain of system-level explanations of human cognition that are best describable in terms of consciousness. Conscious­ ness would then become part of the scientific phenomena requiring explanation and not just a philosophical luxury. In fact, two of the neuro­ psychological phenomena that have created gen­ eral interest in recent years-blindsight and the split-brain-have been compelling for this very reason.3 They will be considered in the follow­ ing sections, together with lesser known neuro­ psychological phenomena relating to knowledge without awareness. This argument has an Achilles heel. It can be reversed. Bisiach (1 988) has recently claimed that there are neuropsychological phenomena that show the limitations of consciousness as a scien­ tific concept. These phenomena have led him to argue that "we may have to learn to live together with the idea that some of the questions set by commisurotomy [the split-brain operation], blind­ sight, unilateral neglect of space etc. will remain forever unanswerable." His arguments will be considered in section 5.

2

Blindsight

One of the most commonplace observations in neurology is the existence of a scotoma, a loss of visual capacity over a certain region of the visual field. It is known that it can arise from lesions to the striate cortex (area 1 7) (Holmes 1 9 1 8; Teuber, Battersby, and Bender, 1 960). In its most extreme form, the patient has no visual experience of any­ thing in that area of the field; in other cases, the patient can be aware of stimuli, but only when they move or flicker (Riddoch 1 9 1 7). One of the

256

most interesting developments in modem neuro­ psychology has been the discovery of an un­ expected potential in some patients for the visual processing of stimuli presented in a scotoma. The initial investigation was made by Poppel, Held, and Frost (1 973). They studied four pa­ tients, each of whom had a visual-field defect in­ volving at least one quadrant of the visual field. The patients were tested along an axis in one of the quadrants in which they were totally blind. A brief flash was presented, together with a tone. When the patients heard the tone, their task was to move their eyes to the position in the visual field where they guessed the flash had been pre­ sented. The patients could make a voluntary eye movement in the general direction of a stimulus projected into the scotoma, even though they de­ nied seeing the stimulus. All four subjects showed a slight tendency to move the eyes farther when the flash was presented 25° eccentrically than when it was presented 1 0° eccentrically. Poppel et suIts across subjects in a rather arbitrary fashion. Although the correspondence between stimulus and response positions was above chance, there was only a weak correlation in three of the four patients. A more extensive study of an individual patient, DB, was made the following year by Weiskrantz, Warrington, Sanders, and Marshall ( 1 974). As blindsight is a condition that does not manifest itself in an identical way in different patients, it is worth considering one prototypic patient in more detail. DB is by general consent the prototypic patient. DB had had an arteriovenous malformation in the right occipital pole removed at the age of 26. The initial experiments on his blindsight were conducted in the year following the operation. At the time they were carried out, he had no visual experience in the left visual field, except for a crescent of preserved vision at the periphery of the upper quadrant. All experiments at that time were performed on the horizontal meridian in order to remain in the entirely blind part of the field.

Modularity and Consciousness

257

eo

75

18

eo

;;;

... ... It: 45 .., ...

75

�

z 0 i= 30 iii 0

tiD

...

It: ... .., Z ...

15

45

' ./ ... ...

�

,..",

30

15 ... ..

.. .

...

o

. ..

.. . '

t

,

15

30

TARGET

I

45

POSI T I O N

,

ao

n

80

( DEGRE E S )

Figure 13.1 Mean finger-reaching responses to stimuli flashed to different positions in the sighted (right) and blind (left) half-fields of DB. The left results are plotted against the upper of the two scales. Reprinted from Weiskrantz, Warrington, Sanders, and Marshall

(1974); Brain 97: 7 14; by permission or Oxford University Press.

The procedure adopted by Poppel et al. (1973) was also used with DB. In locating a light turned on in different positions in his blind field by means of a voluntary eye movement, DB showed only a non-significant trend towards matching his eye movement to the light position. However, when the response of making a voluntary eye movement was replaced by one of the reaching to the position of the light with a finger, he was re­ markably accurate (see figure 1 3 . 1 ).4 In addition, DB could discriminate a long horizontal from a

long vertical line exposed in the lower quadrant of the blind half-field at over.97% accuracy, even for an exposure duration of less than 1 00 msec. Any unreliability of fixation is not therefore a relevant factor. DB could also discriminate a circle from a cross at 90% accuracy. Moreover, he showed an acuity threshold using gratings in his blind half­ field that was no more than double that obtained in his normal sighted field. When the stimulus crossed the border of his scotoma, he was very reliable about reporting it.

Tim Shallice

Otherwise, throughout these experiments, DB re­ ported having no visual experience. When re­ peatedly pressed, he said at times that he had a "feeling" about a stimulus pointing in a particu­ lar direction or that it was "smooth" (the circle) or "jagged" (the cross). He repeatedly, though, stressed that he had no sensation of seeing and that he was always guessing. Since these experiments were done, a number of related studies have been carried out (e.g., Perenin and Jeannerod 1 978; Barbur, Ruddock, and Waterfield 1 980; Zihl 1 980). 5 None of the patients reported in these studies is as clear a case of blindsight as DB. These authors do, though, agree on two conclusions. The first is that there are patients who can use visual information for a number of purposes, such as reaching to a point, without any visual experience being available to guide the act. Second, it is believed that these functions depend not on information transmitted from the retina via the lateral geniculate to the visual cortex, but on information transmitted by other pathways, probably including one that involves a subcortical structure--the superior colliculus. In a critical review article, Campion, Latto, and Smith ( 1 983) have challenged both these conclusions. They have argued that all the phe­ nomena claimed for blindsight can be attributed to one or other of two processes. The first is that light from the target might have spread to por­ tions of the retina unrelated to the scotoma as a result of reflection off structures external to the eye or off ones within the eye, such as the retina itself (the "scattering" hypothesis). The second is that there may be spared, if abnormally func­ tioning, tissue within the cortical region me­ diating visual analysis of the area of the field covered by the scotoma (the "residual striate vision" hypothesis). In addition, they argued that the phenomenological accounts given by the patients are inconsistent and irrelevant. Light scattering within the eye is a far from negligible phenomenon (see, e.g., Bartley 1 941). The weight that Campion et al. gave to this effect

258

as an explanation of blindsight is based on two types of experiments they carried out on normal subjects. 6 One experiment attempted to mimic the reaching results shown in figure 1 3.2 using the blind field of a simulated hemianopia in a normal subject. Both the left eye and the right visual field of the right eye of the subject were occluded. Targets were then presented to the subject's right visual field. This meant that the subject could not see the targets directly, so that any detection would have to depend on scattered light. Local­ ization ability under these conditions was de­ scribed as "very accurate," but it is clearly less impressive than that of DB. Campion et al. also reported experiments on normal subjects on the detection of stimuli projected onto the blind spot using this as an analogue of a scotoma. They held that any intact capacity in this situation must also depend on the use of scattered light. However, as Barbur and Ruddock ( 1 983) pointed out, there are great differences in degree of scatter between the blind-spot situation, in which stimuli must be very close to the intact field, and those in which stimuli are projected well into a hemianopic field. 7 A variety of additional arguments have been given to counter the scattered-light claims. 8 Two particularly compelling ones are available with DB. First, Weiskrantz (1986) pointed out that DB was able to detect black stimuli on a white ground-for instance, in the X versus 0 discrim­ ination-and that the testing of acuity used a grating with light intensity modulated sinus­ oidally. Indeed, with the latter technique, acuity could be better in part of the blind field than in more peripheral parts of the sighted field, al­ though this was never the case for form discrim­ ination. It is very hard to see how these findings could be explained by scattered light. A second procedure devised by Weiskrantz (1 988) used the fact that a stimulus projected onto the blind spot in the blind field would not produce a neural response to light. However, light would be reflected to other parts of the retina. Therefore, if DB were using scattered light, he should have

259

Modularity and Consciousness

Early V i s u a l Proces s i n g

(Altentional Control)

System

- - - - /' (H ig h e r Systems)

J' .

Visuo-Spatial

- Att en t io n a l Cont ro l

( M o d u l a t i o n of S e le c t i o n of Ta rget Pos i t i o n )

//

//

Object

/

I de ntification

" -- - - ' \- - - - - - /

/

Lower- Level Motor- C o n t ro l Sy s t e m

Action Selection & Initiation)

,: Systems which relate � to awareness

/

Figure 13.2

Functionalist model of subsystems involved in reaching to objects detected by normal or blindsight vision_ It is presupposed that the visuo-spatial attentional control system is also linked to the system controlling the direction of action.

been able to detect the stimulus. DB could no\ detect a stimulus projected onto the blind spot in the blind field at above chance level when it could be detected well in other parts of the field defect. DB's blindsight would appear to be a real phenomenon. The other line of criticism made by Campion et al. (1 983) remains more problematic. They claimed that there is no strong evidence that the blindsight phenomenon can be attributed to the operation of operation of subsystems not in­ volved in conscious perception. Thus they point out that over time, DB's scotoma reduced in size (see Weiskrantz 1 980) and that parts of the field previously used for blindsight testing functioned

normally later. This, they argued, suggests recov­ ery of function of parts of the visual cortex and hence is compatible with blindsight having been mediated earlier by the visual cortex operating in a damaged but not totally impaired fashion. What this account fails to explain is the close match between what DB can and, more critically, cannot do and how monkeys perform after striate cortex lesions. DB, in particular, could not per­ form above chance on any pattern-discrimination test that could not be mediated by a simple sen­ sory cue, such as orientation (Weiskrantz 1 988). Thus he could not carry out the simple discrim­ ination of a square from a rectangle. . . . More­ over, the functions that . have been shown to

Tim Shallice

return in the monkey after visual cortex lesions, although not to the same level as before the op­ eration-spatial localization, orientation detec­ tion, spatial frequency discrimination, and very simple shape discrimination (see Pasik and Pasik 1 982)-are those that can be carried out by DB in a scotoma (see Weiskrantz 1 980). Overall, the idea that "blindsight" depends on a processing route that involves structures other than the visual cortex seems most plausible. Finally, how do these behavioral and anatom­ ical issues relate to questions of awareness? Campion et al. ( 1 983) advocated the standard behaviorist position that as there is no way of verifying the fidelity of the report of personal experiences, reports of phenomenal experience should not be a part of scientific psychology. The inappropriateness of using reports of experience as scientific evidence was, they claimed, shown by the way that the reports given by . blindsight patients in test sessions are very heterogenous. The reports given by these patients, when light is presented to their impaired field, do differ. They appear to fall into three classes. DB, for instance, showed a variation in what he experienced within his scotoma (Warrington, personal communica­ tion). In one part, which he called the "dead" part of his field, he never had any form of experience. His responses were complete guesses. A second type of experience, that mentioned earlier, oc­ curred in another part of the blind field. When pressed, he admitted some form of awareness of the stimulus, but it was not described as a visual awareness. Thus he reported feelings like a "feel­ ing of smoothness" for the 0 in the X versus 0 discrimination. Similarly, two of Zihl's ( 1 980) subjects reported that they could "feel" the cor­ respondence between target and eye position. It is noteworthy that in the scattered-light ex­ periments performed by Campion et al. ( 1 983) on the blind spot of normal subjects, the subjective reports were not of these two types. Subjects re­ ported instead some form of visual experience. A "halo of light" was perceived, and the experience of the light "waxed and waned." Reports of

260

visual experience do also occur with blindsight subjects; such reports constitute the third type of experience described by the patients. Often, they arise when intense light sources like flash bulbs are used (e.g., Perenin and Jeannerod 1 978), where light scatter might indeed be producing the expe­ rience. Occasionally, visual experiences can occur in these patients with lower intensity stimuli. For instance, DB's scotoma reduced in size in the years after he was initially tested, and within the reduced scotoma, he now acknowledged seeing something under certain condition-namely, when moving stimuli were used. It is now clear that the difference in visual ex­ perience between normal subjects and blindsight patients is not an absolute one. Nor is it predict­ able under what conditions any given subject will have a visual experience (see Weiskrantz 1980), although moving stimuli and, particularly, the beginning or end of movement seem to be the most likely conditions to produce it. Yet good performance on tasks that use visual input in the absence of visual experience occurs reliably across tasks and across blindsight patients. It is not re­ ported in scattered-light experiments in normal subjects. The heterogeneity of experiential reports of blindsight patients is therefore no grounds for ignoring their most surprising characteristic. The absence of visual experience when a response can be made to a visually presented stimulus is a phe­ nomenon that requires explanation. 9 It is clear that blindsight is not an artifact. Cer­ tain types of judgments or actions can be based on input processes that differ from those involved in normal object identification. Most critically for the argument of this chapter, these judgments are not associated with a visual experience. Our everyday intuition is that a reaching action requires an awareness of what one is trying to touch or grasp. The study of blindsight shows that this is not necessarily so. Moreover, this un­ expected fact about awareness is tightly linked to one about processing systems; the absence of vis­ ual awareness seems to occur when the standard input to the visual cortex is not available. It seems

261

Modularity and Consciousness

anomalous to ban such phenomena concerned with awareness from science, while the rest of the patient's behavior is admitted. If one does at­ tempt to link accounts of awareness and scientific conceptual systems, a functionalist framework­ a system-level one--provides a natural starting place for a phenomenon of this sort. In the case of blindsight, it seems most plausible that informa­ tion transmitted from the input is failing to arrive at some higher level subsystem, but the sub­ systems to which it can arrive are sufficient to ef­ fect appropriate reaching behavior. It is simple to assume that it is the failure of the input to arrive at these higher level subsystems that is responsible for the patient's lack of visual awareness of the stimulus (figure 1 3 .2).

3 Knowledge without Awareness An issue that has been vigorously debated over the past 30 years in cognitive psychology is whether semantic processes can take place for stimuli of which the subject does not become aware. During the 1 950s, there was great interest in whether it was more difficult to recognize a taboo word than a control word-so-called per­ ceptual defense. By the 1 960s, workers in the field had split into two non-communicating camps. One group believed that the effects are artifacts (e.g., Eriksen 1 960). The other argued that the effects are real (e.g., Dixon 1 97 1 ) and arise from subconscious identification of the taboo word before it can be consciously identified. During the 1 970s, new paradigms were developed. The most popular were concerned with whether there could be any effect on the processing of a subsequent word from the rapid presentation of a word that was followed virtually immediately by a pattern mask, so that the subject had no awareness of it. The most widely supported position is that the masked word can facilitate recognition of an as­ sociate that follows it (e.g., Marcel 1 983a, 1 983b). However, again, as with perceptual defense, crit­ ics have attacked the phenomenon on methodo­ logical grouds, claiming it to be an artifact (see,

e.g., Holender 1 986; see also Cheesman and Merikle 1 985). Can neuropsychology shed any light on the controversy? Recently, Trane1 and Damasio ( 1 985) reported a study of autonomic responses (galvanic skin response, or GSR) in two patients who did not recognize the stimuli presented to them. The patients were "prosopagnosics"; in other words, they had a specific difficulty in rec­ ognizing faces. 10 Tranel and Damasio inves­ tigated the GSRs to faces that would have been familiar to the patients if they had not been pro­ sopagnosic. The faces of three types of people were used in the study: family members of the patients, famous people, and hospital personnel who under normal circumstances the patients would have learned to recognize while in hospital. These faces were interspersed with unknown faces and shown one at a time to the patients at inter­ vals of 20 sec. The GSR response was recorded. Mterwards, they were shown the faces again and had to rate them for familiarity under a scale of 1 (familiar) to 5 (unfamiliar). Subject I recognized no faces and yet showed a significantly higher GSR reading to those that, but for her proso­ pagnosia, would have been familiar. Subject 2 failed to recognize only the faces of people she had met since her illness. However, for these faces, she, too, showed a significantly greater GSR than for faces of people she had not met. A closely related study was carried out by Bauer ( 1 984). De Haan, Young, and Newcombe ( 1 987) have shown analogous effects using en­ tirely psychological procedures. They found that semantic-category judgments about printed names showed interference from distracting faces in a different category in a prosopagnosic patient who could not explicitly identify the faces. Tranel and Damasio (1 985) explained their findings in terms of a four-stage theory of face perception: I . Early visual processes. 2. Matching the face to a template; the function of this stage is similar to the perceptual-classi­ fication stage of Warrington and Taylor ( 1 978)

Tim Shallice

262

or the pictogens of Seymour ( 1 979) in object perception. 3. The activating of multi-modal associations. 4. Read-out of the evoked associations, which permits a conscious experience of familiarity and verbal report or non-verbal matching of the stimulus.

�,

They held that the impairment lies in stage and on their account, awareness of whose face It is requires stage 3 so the impairment leads to a loss of awareness. Stages 1 and 2 were held to be relatively intact in their patient, so that the GSR response can be triggered. They were not specific about how it is triggered, but Bauer ( 1 984) argued that this is because GSR activation requires a different route from stage 2 onwards. This is a parallel account to that given for blindsight (see figure 1 3.2). Awareness depends on input reach­ ing certain higher level systems, and some ot er responses are based on a diverging proc essmg route. AU these findings on prosopagnosia have a family resemblance to semantic access phenom­ ena. . . . Thus in semantic access dyslexia, the patient has no explicit knowledge of what the stimulus is and yet can make certain correct se­ mantic decisions about it. An account that has been given for these phenomena is that the p�e­ served function requires a weaker or more nOISY input reaching critical systems than does explicit identification; explicit identification requires in­ hibition of competing possibilities as well as acti­ vation of the representation of the stimulus itself (Shallice and Saffran 1 986; see also Shallice and McGill 1 978). A similar explanation could be given for the prosopagnosia phenomenon. Re­ duced activation at Tranel and Damasio's ( 1985) Stage 3 would be sufficient to produce a GSR response, but insufficient for conscious identi­ fication of the stimulus. Whichever interpreta­ tion is accepted, the phenomenon itself offers a neuropsychological parallel to experiments on normal subjects that are held to show uncon-

?

scious semantic pnmmg (e.g., Marcel 1 983a, 1 983b; Cheeseman and Merikle 1 985). Moreover, both of the two explanations are basically system­ level ones. They fit well with a functionalist perspective.

4 The Split-Brain PatieDt aDd Dual CODSCiollSDess One of the most widely known neuropsycho­ logical syndromes is that of the split-brain pa­ tient. It is the condition produced by a rarely carried out operation in which the corpus cal­ losum and the anterior commissure-the fiber tracts that link the hemispheres-are sectioned for the relief of intractable epilepsy (see, e.g., Gazzaniga and Le Doux 1 978). This syndrome has become famous because its implications seem so dramatic. The operation was widely inter­ preted in the 1 960s and I 970s as resulting in two relatively normal hemispheres being present in the same skull but each disconnected from the other. It was generally believed that by using ap­ propriate input and output processes, one coul . study the operation of the normal nght hemt­ sphere in isolation from its partner. Moreover, it was widely believed that the operation resulted in two conscious entities being present within the same body. The first of these beliefs no longer seems plau­ sible according to Gazzaniga (1 983a), the inves­ tigator with probably the greatest experience of these patients. By 1 983, at least four series of split-brain patients had been described. In the original series, studied by Akelaitis and his col­ leagues (see Geschwind 1 965), no remarkable be­ havioral problems or capacities were noted. The second series of 1 1 patients, that of Bogen and Vogel ( 1 962)-the California series-was dra­ matically different. At least two patients, NG and LB showed evidence of considerable language abi ity in their right hemispheres, a finding at variance with earlier views of right hemisphere function. It therefore became widely believed that

�

i

Modularity and Consciousness

the normal right hemisphere has considerable language capacities. Evidence for the existence of language abilities in the right hemisphere had been obtained in a variety of ways. The most elegant and reliable procedure used a device developed by Zaidel (1975), which occludes one-half of the visual field while allowing prolonged viewing of the display with the other half-field. The Peabody test-a word-picture matching test-was carried out with the word presented auditorily and the four pic­ tures presented to the patient's left visual field and so to the right hemisphere. NG had a raw score of 82 and LB, of 103 (Zaidel 1 976). These scores in­ dicate that the right hemispheres of these patients had comprehension skills equivalent to those of children of mental ages of I I and 1 6. The other two series, though, show that these results are not typical of split-brain patients. A later East Coast series, carried out by Wilson, consisted of28 patients, and an additional patient has been operated on by Rayport in the Mid­ west (see Gazzaniga 1 983a). Of these 29 patients, Gazzaniga stated that only 3 (PS2 , VP, and JW) show clear evidence of language processing in the right hemisphere. He concluded that right hemi­ sphere language in split-brain patients can be at­ tributed in almost every case to left hemisphere brain damage occurring at an early age, at which time language capacities can transfer better to the right hemisphere. This was disputed by Zaidel (l 983b), who ar­ gued that LB and NG are not atypical as far as the California series is concerned. He said that "all of these patients have by now shown evidence ofright hemisphere language" (p. 544), but quoted no specific findings. Gazzaniga (1983b) assumed that Zaidel was referring to the six most studied patients in the California series. He retorted skeptically, "We look forward to examining the data finally acquired from patients in whom the existence of right hemisphere language resisted empirical demonstration for over 1 5 years" (p. 548). If someone who worked extensively with both the major series of split-brain patients-the

263

California one in the 1 960s and the East Coast one in the 1 970s and 1980s-considers that right hemisphere language is a rare abnormality in such patients and objective evidence is not avail­ able to disprove him, then the safest conclusion is to accept the conservative position. This implies that the most widely studied split-brain patients can tell us little about the capacities of the normal right hemisphere. The split-brain syndrome bears on a second issue. Indeed, it is for this second reason that the syndrome has been so widely discussed. The ana­ tomical separation of the two hemispheres has been held to result in the presence of two con­ scious entities in the same body (Sperry 1 968, 1 984). As far as the wider implications of this claim are concerned, it is not critical whether the right hemisphere cognitive systems in the most widely studied split-brain patients are or are not normal. For there is no reason to assume that before their operation, such patients differed from normal subjects in how unitary was their experi­ ence of the world. What evidence, though, is there that the "right hemisphere" of a split-brain patient like NG is conscious? As the right hemisphere of the classi­ cal split-brain patients produces no language, no evidence can be obtained through report about whether it experiences the world. The standard argument is to list a set of capacities for the right hemisphere and to infer from them that the right hemisphere must be aware. So Sperry (1 984) stated: The mental performance of this hemisphere after com­ missurotomy has been found repeatedly to be superior and dominant to that of the speaking hemisphere in a series of nonverbal, largely spatial tests. . . . Examples include the copying of designs, reading faces [sic], fitting forms into molds, discrimination and recall of nonde­ script tactual and visual forms, spatial transformations and transpositions, judging whole circle size from a small arc, grouping series of different sized and shaped blocks into categories, perceiving whole plane forms from a collection of parts and intuitive apprehension of geometrical properties. (p. 666)

11

264

Tim Shallice

He concluded that "after watching repeatedly the superior performance of the right hemisphere in tests like the above, one finds it most difficult to think of this half of the brain as being only an automaton lacking in conscious awareness" (p. 666). This is undoubtedly the natural assumption of those who work with split-brain patients, but one can hardly take what observers assume as a strong form of evidence. An observer, watching a mute blindsight patient reaching towards a spot of light would, for instance, if ignorant of the literature, assume that the patient could see the light. The work discussed in section 1 3.3 and that on action lapses in normals indicate that our in­ tuitive assumptions of what cognitive processing is possible without awareness is not solidly based. Does this mean that one can hold a position like that advocated by Eccles ( 1 965) and MacKay (I 966b) that the right hemisphere in split­ brain subjects is some form of non-conscious automaton? Consider, as one example, the study of Zaidel, Zaidel, and Sperry ( 1 9 8 1 ) on the performance of the right and left hemispheres of two classic split­ brain patients, NG and LB, on a standardly used non-verbal reasoning test-Progressive Matrices. The test was carried out using the device devel­ oped by Zaidel ( 1 975). The task was done first with the left visual field and left hand, to assess the capacity of the right hemisphere. A week later, it was performed with the right visual field and right hand, to assess the capacities of the left hemisphere. Finally, a week after that, it was carried out with free vision . Both patients did as well or nearly as well with their right hemisphere as with their left. Indeed, with his right hemi­ sphere, LB scored at a level equivalent to an 1 1 year-old child. To perform correctly on this test, the relation between two items has to be ab­ stracted and then extrapolated so as to infer the third item in a progression; finally, the result must be matched to one of a set of possible answers. The processes involved are far more demanding than, say, those used in picture-word matching

Table 13.1 Mental-age equivalents of perfonnance on Progressive Matrices of two split-brain patients given input to right visual field, to left visual field, and bilaterally Input Right field NG

LB

7.9 >14

Left field 7.9 1 1 .3

Bilateral 8.2 > 14

From Zaidel, Zaidel, and Sperry (1981)

in the number of components, the level of ab­ straction, and the involvement of more than the operation of a routine schema. If this level of performance could be obtained unconsciously, then it would be really difficult to argue that con­ sciousness is not an epiphenomenon. Given that it is not, it is therefore very likely, if not un­ equivocally established, that the split-brain right hemisphere is aware. If one does accept that the split-brain right hemisphere can be conscious, then this has cre­ ated much difficulty for some philosophical posi­ tions . 1 2 Thus the philosopher Nagel ( 1 97 1 ) has argued, referring to split-brain patients, If we decide that they definitely had two minds, then it would be problematical why we didn't conclude on anatomical grounds that everyone has two minds, but that we don't notice it except in these odd cases because most pairs of minds in a single body run in perfect par­ allel due to the direct communication between the hemispheres which provide their anatomical base. (p. 409)

This, he pointed out, would be a reductio ad ab­ surdum position because the unitariness of mind derives directly from our own experience. An­ other philosopher-Puccetti ( 1973)-has grasped the nettle and claimed that normal subjects are indeed two distinct conscious entities. On Puc­ cetti's approach, the left hemisphere self is un­ aware of the other and so considers mind to be unitary. The right hemisphere self, however, "has known the true state of affairs from a very tender

Modularity and Consciousness

265

age. It has known this because beginning at age

Le Doux, Wilson, and Gazzaniga were inter­

two or three it heard speech emanating from

ested in responses made by the left hemisphere to

the common body" (Puccetti 1 98 1 , p. 97).

a stimulus presented to the left visual field and so

That two fairly self-contained processing sys­

to the right hemisphere. In the test situation used,

tems inhabit the same body in normal people is

two stimuli were presented simultaneously, one to

a common, but quite unjustified, extrapolation

each visual field. The subject's task was to select

from split-brain research. In the normal person,

those two pictures out of an array of eight that

processing in the two hemispheres is frequently

were semantically related to the two stimuli pre­

complementary . . . . The different processes in­

sented. One picture had to be selected with either

volved in simple tasks like reading a word in

hand. Thus if a snow scene was presented to the

script or recognizing a picture taken from an

right visual field and a chicken claw to the left, the

unusual angle or copying a drawing can involve

left hand is supposed to choose a shovel and

different hemispheres. Any conscious experience

the right hand, a chicken. Split-brain patients PS2

limited to one hemisphere would be quite unlike

and VP are said to be able to perform this task.

the one we have. A system-level, or functionalist, approach to

Of particular relevance to Le Doux, Wilson, and Gazzaniga's theory is what the left hemi­

consciousness provides a very simple answer

sphere said about responses presumed to be

to these philosophical puzzles. As Kinsboume

generated by the right hemisphere. When in the

( l 974a) pointed out, the conceptual difficulties

example just given PS2 was asked what he saw, he

over split-brain conscious experience are not ba­

replied, "I saw a claw and I picked the chicken

sically different from those arising from any other

and you have to clean out the chicken shed with a

major disconnection syndrome or from what

shovel. " From examples such as this, Gazzaniga

normal subjects experience over short intervals of

and Le Doux ( 1 978) have developed the following

time when carrying out two fairly demanding

"metaphor":

tasks simultaneously. 1 3 If consciousness is some, as yet unspecified, property of the system-level organization of cognition, then the split-brain operation has the effect of creating two separate, if peculiarly functioning, systems where origi­ nally the subsystems operated as a co-ordinated whole. To flesh out this argument, one would need to specify what these system-level properties might be so that it is clear why before the separation, they apply to one system and after the operation, to two. Before discussing this issue, it is necessary to consider certain views on consciousness that derive directly from neuropsychological research. The first--that of Le Doux, Wilson, and Gaz­ zaniga (1 979)-derives directly from split-brain research, particularly from work with three East Coast patients with some capacity both to com­ prehend and to produce language in both hemi­ spheres (see Gazzaniga, Holtzman, Deck, arid

Lee 1 985).

Our sense of subjective awareness arises out of our dominant left hemisphere's unrelenting need to explain actions taken from any one of a multiplicity of mental systems that dwell within us . . . . These systems, which coexist with the language system, are not necessarily in touch with language processes prior to a behavior. Once actions are taken, the left, observing these behaviors, constructs a story as to the meaning, and this in turn becomes part of the language system's understanding of the person. (Gazzaniga, 1983a, p. 536) The argument is inadequate. Assume that the evidence obtained by Gazzaniga and Le Doux does indeed support the idea that if an action is produced and the left hemisphere systems re­ sponsible for language production have no in­ formation about the causal antecedents of the action, then a rationalization will tend to be pro­ duced to account for the action. To generalize from the split-brain patient to the normal subject, one has to assume that for the normal subject,

266

Tim Shallice

too, there is usually no information about the causal antecedents of an action available to the cognitive system. The idea that one acted from an "intention" would need to be a delusion. To make this position plausible, the normal cognitive system must contain some functional analogue of the disconnection of the split-brain system or some other means by which part of the overall system might operate independently of the rest to produce actions that have causes "unknown" to the rest of the system. Gazzaniga offers no evidence that trus is generally the case or that "a cognitive system that strives for con­ sistency and order in the buzzing chaos of behav­ iors that are constantly being produced by the total organism" (p. 536) is indeed what is respon­ sible for our sense of subjective awareness. For he does not show that a "buzzing chaos of behav­ iors" is an appropriate way to characterize human action. Indeed, it is a classical position about consciousness (James 1 8 90) that it is the product of a mechanism that has come about in evolution to prevent just such a "buzzing chaos" from ex­ isting. The behavior of split-brain patients can be a misleading guide to normal function in more ways than one!

5

Bisiacb's Critique

The three types of phenomena discussed so far in this chapter are paradoxical from a lay view of consciousness. They are, though, easily explicable on functionalist theories of consciousness. Two other types of neuropsychological observation­ also paradoxical to the lay perspective-have been used by Bisiach et al. ( 1 985) to argue more subtly against attempting to explain conscious experience in scientific terms. Bisiach et al. begin their argument with an as­ pect of unilateral neglect . . . -the patient's denial of any deficit. Denial of deficit, anosognosia, can also occur for other syndromes, such as cortical blindness or aphasia. It is far from universal in the syndromes in wruch neglect is observed, but when

it occurs, it can be dramatic. Take, for instance, the hemiplegic patient described by Bisiach, Meregalli, and Berti ( 1 985): E

(Holding one of his fingers in the patient's right vis­ ual field)

Seize my finger with your left hand. . . .

Well? Can't you move your left hand at all?

p

. . . . Just give me time to proceed from thought to action.

E

Why don't you need time to proceed from thought to action when you use your right hand? Maybe you

can't move your left hand? P I

can

move it perfectly. Only, there are sometimes

illogical reactions in behavior; some positive and some negative . . . .

E

(Placing the patient's left hand between his own hands) Whose hands are these?

P

Your hands.

E

How many of them?

P Three.

E

Ever seen a man with three hands?

P

A hand is the extremity of an arm. Since you have three arms you must have three hands. (Bisiach, Meregalli, and Berti,

1 985, unpublished)

Bisiach, Luzzatti, and Perani ( 1 979) have ar­ gued that dementia is not a sufficient explanation for such denial of symptoms. It would not ac­ count, they say, for phenomena like the classic one of Anton (1 899), where the patient denied being cortically blind even though she was vividly aware of much milder dysphasic difficulties. In­ stead, Bisiach and rus colleagues argue that all expectations concerning the damaged function can be lost as a result of relatively local damage as well as the perceptual processing itself. Why else would the evidence of other senses have so little effect? Bisiach and Berti (1 987) say, In severe cases, any effort to beset the patient and force him to admit and critically evaluate his pathological conditions is doomed to failure; either the patient eludes the problem altogether, or he cuts short and shelters his cognitive disorder by arguments of which a confronta­ tion would be in vain . . . . This seems to entail that con­ sciousness is

inherent

in the representational activity of

Modularity and Consciousness

o o

267

o o

o o

o

o

Figure 13.3 The apparatus used with patient FS. There are pairs oflights (one red, one green) at four positions in the upper half of the apparatus. The stimulus flash was the illumination of one of these eight lights. The response was to press one of the four response panels in the lower half, two of which lighted up on a trial. Subjects always used their unimpaired hand (i.e., the one ipsilateral to the lesion). Reprinted from Bisiach, Berti, and Vallar (1985); Attention and Performance J J: 240; by pennission of Erlbaum. these analog structures, both as referring to the mon­ itoring of these activities and as referring to their control . . . no further mechanisms for consciousness seem to exist, either in the fonn of a unitised hierarchically su­ perimposed component of the cognitive machinery or emerging from the whole of cognitive activities of the brain.

Here Bisiach and Berti are attacking the idea of consciousness as an all-seeking homunculus. The critique of functionalist theories of consciousness is taken further in Bisiach ( 1988), where he uses some observations of Bisiach, Berti, and Vallar (1 985). In the experiment described in this study, subjects were asked to respond manually to the red or green colour of a 200-msec light flash pre­ sented to the left or right half-field. The subjects had four panels to press, two for each side (figure 1 3 .3). On any trial, two would light up on differ­ ent sides-one red and one green-and the sub­ ject had to press the lighted panel of the same color as the flash. In his interpretation of the results, Bisiach ( 1988) discussed the observations made on a pa­ tient, FS, who had an extensive right frontal lesion. Of particular interest were the responses FS made when the flash was to the right half-field, which projects to the undamaged left hemisphere. In a preliminary experiment in which he just had to name the color, FS was completely accurate

with this stimulus in this situation. This was also the case in the main experiment, when FS had to respond to the color by pressing one of the right­ side keys. If, however, the response to the right­ field stimulus had to be with a left-side key, FS made no response on 8 out of 1 6 trials. More critically, on several occasions, he spontaneously stated that no stimulus had occurred. Bisiach (1 988) considered the possibility that FS did not become aware of the stimulus until he made a response, but argued that this suggestion is implausible because we are normally aware of many stimuli to which we make no response. So he continued that it seems more sensible to hold that "denial of supra-threshold stimuli . . . would be better interpreted as an inhibitory effect of the action made which suppresses or overrides any experience of the stimulus after its fleeting ap­ pearance in consciousness, and prevents recovery of episodic memory of it." Given the other anosognosia phenomena he had investigated, denial is a natural explanation for Bisiach of why his right frontal patient FS said that no stimulus had occurred on some of the trials on which he failed to respond. 14 The purpose of Bisiach's argument is not, though, to establish a denial explanation. The conclusion he draws "is that in trying to explain neuro­ psychological disorders we cannot turn from the Spinozian 'inner perspective' of the mental events

268

Tim Shallice

ture it and when the complex mechanisms of ab­ straction, categorization, memory, and language required to articulate it are intact. These mecha­ nisms function effectively in many psychological experiments on normal subjects (see Ericcson and Simon

1 985). They can be presumed to be intact

in, say, blindsight patients and in Tranel and Damasio's

( 1 985) prosopagnosic patient. They

may well, however, be impaired in FS. Thus Bisiach's argument about FS could be valid and

�

the phenomenal experience of other subject still be of value for science. This defense does not, however, help as far as

ITl Figure 13.4 A purposefully tendentious and highly simplified model of the organization of the cognitive system developed to illustrate a theoretical argument about consciousness. T are sensory transducers, S, higher level sensory pro­ cesses, R, response organization systems. From Bisiach, Mergalli, and Berti (1985).

Bisiach's deeper objections about anosognosia are concerned. Denial phenomena have led Bisiach, MeregalIi, and Berti

(1985) to propose the fol­

lowing "purposefully tendentious" model (figure

1 3 .4). In the model, the T components are sensory transducers, the S are multi-level sensory pro­ cessors that carry out both perceptual and repre­ sentational activities, and the R are response systems. The essence of the model is its modu­ larity, the relative segregation of S-R paths, and

lived by our subjects to the functionalist level. " 15 Instead, he believes that we should be concerned with the information-processing properties that allow such processes as reflection or intersubjec­ tive communication to occur, but not with their relation to "inner experiences." "raw feels," or the like. Bisiach's argument that it makes little sense to speculate on the inner experience of a patient like FS seems apt. Indeed, in the spirit of his argu­ ment, one might even say that "inner experience" is not an appropriate phrase to use of FS's state during the episodes that he discussed. However, it is not valid to extrapolate and assume that all phenomenal experience is irrelevant for science or that it is invalid to attempt to explain an ab­ straction based on it-consciousness-in scien­ tific terms. As far as the reporting of phenomenal experi­ ence is concerned, it is useful for science only if its subject has concepts that, at least roughly, cap-

the absence of a superordinate general-purpose module for detecting and controlling failures in the system. Neglect arises on this view because a certain level of operation of S is responsible not only for the processing of the stimulus, but also expectations concerning it. Such a model provides no reasonable correspondence for consciousness, for there is no element that controls and no uni­ tariness in its processing. How, then, can anosognosic denial be ex­ plained? The position developed by Bisiach and his colleagues is that the locus of the impair­ ment is close to that of the sensory processors themselves. To make this type of position more concrete, it is useful to consider a phenomenon discovered by Gregory (personal communica­ tion). He constructed a self-Iurninant outline cube. It is possible to invert the cube as one would a Neckar cube so that in the dark what appears to be its near face is actually the far one. Gregory then asks the subject to rotate the cube. The visual

and kinaesthetic cues are now in con:flict because

Modularity and Consciousness

269

they suggest that the cube is rotating in diametri­ cally opposed ways. Either one observes a strange visual distortion of the cube or, alternatively, the object remains a cube but one's wrist feels as though it were breaking. The power of "un­ conscious inference" is so great that one or the other percept, which would otherwise be clear, cannot survive. The outputs of the two percep­ tual modules-visual and tactile-have to pro­ duce representations of the world, which are consistent. An analogous explanation can be given for denial in anosognosia. If the mechanisms that mediate perceptual "unconscious inference" are intact, but part of the perceptual subsystems­ because they are damaged-give a grossly dis­ torted reading, the preceptual-inference system has to resolve the discrepancy. Knowledge is no help. In Gregory's example, one knows that one is seeing a real cube inverted like a Neckar cube but one still feels one's wrist breaking. So for Bis ach, Meregalli, and Berti's patient, it may be that the knowledge that the doctor attempts to induce in the patient cannot shift the operation of the "unconscious inference" mechanism. Its output continues to register the position of the arm as elsewhere than in the doctor's hands. How do these positions on anosognosic denial relate to the conceptual framework of contention­ scheduling and Supervisory Systems? On the sur­ face the model and the conceptual framework seem diametrically opposed. Yet the opposition is more apparent than real. The distributed pro­ cesing that characterizes the model of Bisiach and his colleagues, is, in fact, present in the schemata that are selected in contention sche­ duling and in the modules on which they are implemented. The denial phenomenon, itself, is compatible with the idea that perceptual subsystems, includ­ ing spatial representations, are "impenetrable"­ to use Pylyshyn's (1 980) term-to higher systems. It does not, however, rule out the possibility that a general-purpose Supervisory System exists, es­ pecially one that activates schemata not modules.

i

The way that schemata utilize modules would not be known to the Supervisory System; it does not need to know because if one schema is activated any incompatible one will be automatically in� hibited. So why should the Supervisory System be in a position to know what is happening to modules when they are damaged? As the use of Hehnholtz's views on unconscious inference should make clear, the position being developed is entirely compatible with consciousness corre­ sponding to the operation of some higher level processing system that has access only to the out­ puts of the perceptual system.

6

Conclusion

Neuropsychological research is not, then, in conflict with the need to explain consciousness scientifically. It is entirely compatible with a functionalist solution of the body-mind dilemma. Over the past 1 5 years, a wide variety of theories for an information-processing correspondence for consciousness have been offered by cognitive psychologists. It has been identified with the op­ eration of a high-level processing system, as in my early idea that it corresponded to the input that selects which action-system (action or thought schema) is dominant and sets its goal for that particular operation (Shallice 1 972). Related po­ sitions are Posner and Klein's ( 1 973) view that it corresponds to a limited-capacity processing sys­ tem and Iohnson-Laird's (1983) suggestion that it corresponds to the operating system of a com­ puter. It has been identified with a system anal­ ogous to the Supervisory System by MacKay (I 966a), Luria ( 1969), Marshall and Morton (1 978), and Mandler ( 1 975, 1 985). It has been held to correspond to the contents of a short-term store by Atkinson and Shiffrin ( 1 97 1 ), the input to the speech system by Dennett ( 1 969), the con­ tents of a globally distributed data-base by Baars (1983), the recovery of synthesized percepts by Marcel ( 1983b), and the functions carried out by the hippocampus by O'Keefe (1 985)!

Tim Shallice

It is natural to view these as competing hy­ potheses, although their very variety suggests that little progress is being made. One could, however, ask whether neuropsychological research can help to choose between them. At present, it seems to offer relatively little in this respect. From a neuropsychological perspective, it can, for in­ stance, be argued that short-term stores are vari­ ous and specific and lowly in function, and so their contents are not satisfactory candidates for a correspondence with consciousness. One can point out that no syndrome has yet been charac­ terized as damage to a globally distributed data­ base. Yet this still leaves many alternatives. However, to a considerable extent, it can con­ tribute little because the theories themselves are too vaguely specified and so difficult to falsify. I will take a different tack, . . . [making] two basic assumptions. First-and here there are echoes of Fodor's (1 983) position-there are many subsystems with internal mechanisms and even outputs that are "unknown" to higher sys­ tems. Second--and here there is a resemblance to Bisiach's view-it seems likely there is no single higher-level system that directly controls the op­ eration of the lower level modules. There are a variety of higher level systems, and control is too strong a term to describe their relation to lower level subsystems. The higher level systems influ­ ence the activation of schemata, each of which, in its specific way, will, if selected, determine which modules will be operative and modulate their activities. On the present approach, these con­ trol systems include contention scheduling, the Supervisory System, episodic memory process­ ing, and the language system. My position is to reject a common denomi­ nator in the theories listed earlier, the attempt to identify any particular subsystem-or, better, its output or some other aspect of its processing­ with consciousness. Instead, I would argue that another aspect of the operation of our infor­ mation-processing system holds the key. The processes of willed action, of reflection, of re-

270

membering, and of speaking about something do not involve identical subsystems or even the same control system, but it is normal for these control systems to operate in an integrated way. The output of episodic memory, say, feeds the Super­ visory System, which feeds the language system or contention scheduling; alternatively, the activ­ ities of contention scheduling are fed to episodic memory and the Supervisory System (figure 13.5). Consider the information that is successively rep­ resented, in different forms, in different control systems in such a process. By comparison with the other types of information being processed at the same time in various special-purpose modules, this information would be both much more widely accessible and especially important for present and future activities. If the organism also has sufficient powers of abstraction to attempt to cat­ egorize its internal states, the distinction between the two types of information would be a key aspect of the concepts that it employs to under­ stand itself. The conscious/non-conscious con­ trast would correspond to the distinction between the two types of information. 1 6 If the process that corresponds to having a conscious experience is one that necessarily in­ volves more than one control system, then it is apparent why each of two theorists could take a different control system and produce a plausible argument about why the properties of conscious experience map with what is happening in the particular control system that he or she chose. Also, there will be rare occasions in normal sub­ jects or frequent occasions in rare patients when the control systems operate in an uncoordinated fashion. Patient FS of Bisiach. Berti, and Vallar ( 1 985) may well be an example. In these cases, the adjective conscious could not be unambiguously applied. For instance, it might happen that the language system could, say, be controlled by the Supervisory System, but a strong input to epi­ sodic memory could be coming from contention scheduling. The person would then be unable to remember what had been said because of what

27 1

Modularity and Consciousness

SUPERVISORY SYST EM

LANGUAGE SYSTEM

M E M ORY

SPECIAL- PURPOSE PROCESSI N G SYSTEM

Figure 13.5

The four subsystems or processes hypothesized to have a relation with awareness (those extensively interconnected by arrows). Solid arrows represent obligatory transmission operations; dashed arrows are optional (i.e., under the control of the Supervisory System). else was being done. One such occasions, there

psychological research thus poses a deep con­

would be no clear answer to the question of

ceptual problem in an acute fashion.

whether one is conscious of what one was saying

The discussion in this chapter has strayed

or doing. As an example, in the course of an

increasingly far from the more solid aspects of

action-lapse, the operations of contention sched­

cognitive neuropsychological research. In a way,

uling and the Supervisory System are in conflict.

though, these issues concerning consciousness are

It would not be appropriate to say of people in

typical of modem cognitive neuropsychology.

this situation either that they are conscious of

The phenomena that have been uncovered are

how they are acting or that they are not. 1 7

stunningly counterintuitive, challenging many of

Whatever the merit of this particular approach

our most standard assumptions about the mech­

to consciousness, neuropsychology's most funda­

anisms underlying cognition. Thirty years ago,

mental contribution to our understanding of

except for certain seemingly outdated schools of

human cognition, however, is to provide by far

neurologists, the modular view of the cognitive

the strongest evidence for its modular basis. The

system that cognitive neuropsychology offers

conception of the organization of the cognitive

would have seemed as implausible as that pro­

system as composed of specialized subsystems

vided by Gall. The answers that have been given

in tum offers few openings for a concept like

for the variety of phenomena discovered and

consciousness. Yet neuropsychology, with a phe­

documented over many aspects of perception,

nomenon like blindsight, seem s to require that

language, memory, and cognition may not sur­

consciousness be included within science. Neuro-

vive. The range of conceptual problems that these

Tim Shallice

phenomena pose will, however, remain. When they can be adequately answered, psychology will have become a science.

Notes I . The most critical issue is how functionalism accounts for so-called qualia, or "raw-feels" -that is, the redness of red, the experience of pain and, so on. I prefer to think of the relationship between information process­ ing explanations and accounts of experience as analo­ gous to the relation between two different maps of the same part of the world; neither map need be totally deducible from the other, but there will be many corre­ spondences (see Shallice I 972). In my view, the best strategy as far as the philosophical problems are con­ cerned is to assume, hopefully, that any paradoxes rep­ resent an intellectual eddy that can be safely neglected, if positive evidence for a functionalist position can be obtained. There are precedents for such intellectually psychopathic behavior. Darwin, in his theorizing on evolution, was quite at a loss to account for the sim­ ilarity among the flora of South America, Africa, and Australia, which should, on the theory, mean that the three regions were close together. The phenomenon be­ came comprehensible only with the development of the theory of continental drift. 2. Indeed, one might almost say that consciousness viewed from the perspective of those levels of explana­ tion is becoming more paradoxical! Thus Libet, Wright, Feinstein, and Pearl (1 979) and Libet, Gleason, Wright, and Pearl (1983) have discovered counterintuitive phe­ nomena concerning correlations between evoked po­ tential recordings and reported experience. This led them to the paradoxical conclusion that the subjective timing of experiences is "retroactively antedated." (For criticisms of this claim, see Bisiach 1988) For discussion of the more general issue of whether the existence of consciousness should be interpretable in terms of a sci­ entific conceptual framework, see the articles in Marcel and Bisiach (1 988), which take conflicting po�itions. 3. For related phenomena in cognitive psychology, see Marcel ( 1983a, 1 983b) and Groeger (\984, 1986). 4. This was so for all but the smallest diameter light spot used. 5. The patients reported in the studies are often collec­ tively referred to as having "blindsight." However, the

272

tests that have been carried out by the different authors vary considerably, and the patients themselves are also rather heterogeneous. Thus the patients of Perenin and Jeannerod ( 1 978) had had a whole hemisphere removed at operation, while DB had most of his right hemisphere including post-striate visual regions intact. This hetero­ geneity means that it is premature to treat their visual impairments as exemplars of the same underlying func­ tional syndrome. The argument will therefore primarily be based on DB, with other patients referred to in a secondary fashion. 6. Campion, Latto, and Smith ( 1 983) also carried out two experiments on three hemianopic patients. How­ ever, the results, even given their assumptions, did not clearly support their position that "blindsight" results are artifacts. 7. These differences are not just quantitative. Light from stimuli projected onto the blind spot can be de­ tected, whichever way it scatters. When a patient has a hemianopia, only scatter towards the intact field is rele­ vant. The findings of experiments on stimuli projected onto the blind spot are to be treated cautiously, espe­ cially if accurate measurement of fixation is not made. Thus Bartley ( 1 941) found that in this situation, "a great variety of results . . . occurred, and that by deliberately shifting slightly from the original good fixation in which the direct illumination undoubtedly fell within the bounds of the optic disc [i.e., the blind spot], the variety of results reported by different observers will be re­ peated" (p. 1 06). In the experiments conducted by Campion et aI. (1 983), deviations of fixation of 1 °, suf­ ficient to bring the stimulus to the edge of the sighted field for two of the three subjects, could not be reliably detected by the procedure adopted to check fixation position. 8. Barbur, Ruddock, and Waterfield (1980) attempted to calculate the amount of scattering of light from the blind to the sighted field that would be needed to pro­ duce the "blindsight" results they obtained from a hemianopic patient, G. They deduced that the stimulus light would need to scatter 1 5 times its own energy into the sighted field to be detected, which is obviously impossible. 9. To argue that the phenomenon is only analogous to effects that occur in normal subjects with near-threshold stimuli only substitutes one experiential phenomenon that requires explanation for another. The two situa­ tions are probably qualitatively different.

273

Modularity and Consciousness

10. For a review of this syndrome, see Damasio, Ya­ mada, Damasio, Corbett, and McKee ( 1 980) and Bauer and Rubens (1985). In one case, the impairment oc­ curred from bilateral occipital lobe strokes; in the other, from herpes simplex encephalitis. It is stated that patient I was normal on tests of language, intellect, and visual perception. For patient 2, it is rather obscurely stated that "language abilities are intact and her visual per­ ception is compatible with normal recognition of faces learned before the onset of her illness" (Tranel and Damasio, 1985, p. 1 454). Unfortunately, no quantita­ tive data are given on these points. However, it does not seem too likely that other deficits unrelated to the pro­ sopagnosia influence the results. I I . Sperry (1 984) continues at this point, "the literature is still scattered but see reviews" (p. 666), and then gives five references, four of which are to reviews of the older split-brain literature of the 1 960s. In fact, descriptions of methodologically adequate quantitative experiments are very rare in the split-brain literature until the 1970s. 12. One way of avoiding the difficulties is to adopt a suggestion of Kinsboume (1 974b), that a function of the corpus callosum is to maintain the activation level of both hemispheres at reasonable levels. Without it, he argued, only one hemisphere, at most, can be reason­ ably activated by mid-brain structures. It seems unlikely that this is always true. It would not, for instance, seem possible on this account to explain "cross-cueing," where one hemisphere detee? What possibly could be its source? How is it "perceived," necessarily each in his own case, not possibly by any other? What is its relation to attention, reasoning, nested intentions, problem solving, decision making, memory? Somehow, these detailed questions seem inappropriate, a little like asking of a Funda­ mentalist, "Just how did God go about creating the world in six days?" "How did His saying 'Let there be light!' bring about there being light?" Indeed, just as the Fundamentalist seems to be­ lieve in his account of the world independently of scientific research, so do we seem to believe in our consciousness and the machine's lack of it in­ dependently of any reasonable arguments. 3 7 Perhaps the problem is better seen the other way around: once we have accounts of the vari­ ous processes I have mentioned, what is added by consciousess? What further light does this inner light shed upon our minds? What phenomena is unexplained without it? Perhaps there is some­ thing. But, perhaps too, as Hume found in the case of personal identity, there is nothing more, and it would be wrong-headed to identify con­ sciousness with any of these actual processes, sin­ gly or together. None of them play particularly the moral role that consciousness traditionally is supposed to play. There would seem to be no actual thing or process (or even "function" 38) that our past usages have been "getting at" (cf. Put­ nam 1975). That would seem to afford at least one reason for doubting that the term refers, that is, it would give a reason for doubting there really is such a thing as consciousness at all. This doubt, however, is pernicious. Once we allow it, it would seem that the concept of consciousness no longer has a hold. Although arguments about necessary conditions for the application of a concept are difficult in general to defend, it would seem that, if we abandon the Cartesian Intuition, we've lost what little hold on the notion of consciousness that we have. But if the truth of the Cartesian Intuition is a necessary condition on the applicability of the notion of consciousness, then the mere possibility of a

473

machine of the sort I have described not being conscious entails that the there is no such thing as consciousness. We should be clear about precisely what the consequences would be were we to give up our belief in consciousness. It would by no means en­ tail any extravagant Behavioristic or Eliminitivist claim that no mental terms at all are scientifically respectable, much less that none of them suc­ ceeded in referring. I have used mental terms throughout my descriptions of the capacities of people and machines, and I doubt very much that they could ever be reasonably eliminated. But one needn't be committed thereby to every pre-scien­ tific mentalistic term, or to finding for every such term some post-scientific equivalent. Some terms may simply have to go, as "angels" did in an account of planetary motion, and as "the soul" does in our account of our personal identities. Nor would one be committed to abandoning the term in ordinary talk. If the term "conscious" is only meant to indicate that a person is, say, awake and capable of intentional, attended ac­ tivity on which she might be able to report, then the term is clearly harmless enough. I think it is often used in this way: it would seem to be the usage underlying such claims as those of Moruzzi (1 966) and Penfield (1 975) that locate conscious­ ness merely in the activation of the reticular for­ mation. We need only notice that, according to such usage, a computer programmed with just clauses (1)-(5), if my earlier arguments are cor­ rect, would qualify as conscious too. In view of the doubts that I have raised here, what are we to make of our beliefs about con­ sciousness? In a famous passage, Wittgenstein (1953/1 967, p. 97e) writes: "only of a living human being can one say: it has sensations . . . is conscious or unconscious," and, more recently, Karl Pribram (1976, p. 298) innocently remarks, "I tend to view animals, especially furry animals, as conscious-not plants, not inanimate crystals, not computers. This might be called the "cudd­ liness criterion." I don't see any justification in these claims; they seem, in Block's (1978) phrase,

Georges Rey

arbitrarily chauvanistic, "speciesist." But they may accurately describe the pattern of our as­ criptions, unjustifiable though that pattern may be. We may be strongly inclined to think of our­ selves and our biological kin in a special way that we are not disposed to think of machines, or at least not machines that don't look and act like us. Should a machine look and act as human beings normally do-indeed, should we discover that one of us is a machine-then we would think of it in the same way. We might, of course, try to justify this disposition by behaviorism (and I think this accounts for much of the attraction of that otherwise bankrupt theory), or, failing that, we might try to find some inner condition that would mark the distinction that we want. We are tempted, I think, to try to ground the difference that is so vivid to us in the kind of metaphysical difference that the traditional picture of con­ sciousness suggests, and to claim for it a certainty to which we feel introspection naturally entitles us. (On the extent to which introspection may be susceptible to this sort of imposition, see Nisbett and Wilson 1 977.) But then we find, as I have found in this chapter, that no such inner con­ dition exists. In all theoretically significant ways we seem to be indistinguishable from the "mere machines" from which we nevertheless insist upon distinguishing ourselves. If some story like this were true, perhaps all we could do is acquiesce to our apparently arbitrary biases. We would need to abandon the attempt to find for them any false metaphysical buttressing in some special condition of consciousness, just as we need to abandon the attempt to find such but­ tressing for our personal identities in same special soul. In both cases, of course, the consequences for moral theory would be disappointing: we would have trouble justifying the special concern we feel towards people and animals, just as we have trouble justifying the special concern we feel toward our futures and our pasts (cf. Parfit 1 97 1 a, 1971b). Human reason would tum out to have the peculiar fate, that in one species of its beliefs it would be burdened by questions that it would not

474

be able to ignore, but which it would never be able satisfactorily to answer.

Postscript Since the appearance of this essay eight years ago, I have expanded on a number of its ideas in articles that the interested reader may want to consult (See Rey 1 997 for an expansion of the argument as a whole). In section III, I make a passing analogy be­ tween sensation experiences and first-person re­ flexive thoughts that I have since developed at length (Rey 1 992, 1993, 1 997, chap. I I). Similar suggestions have been independently developed by Lycan ( 1990) and Leeds (1 992). In section IV, I failed to emphasize what I briefly mention in its antepenultimate paragraph: that I have no argument against what might be called a "weak" notion of consciousness, which involves, for example, merely wakefulness, atten­ tion, or reportability, and is of obvious utility to psychology. The only problem with such a weak notion is that if the rest of my argument is correct, it applies equally well to the suitably programmed computer. The notion whose appli­ cation I am doubting is a "stronger" one, in­ tended to distinguish us from such machines. Brian Loar has complained to me that the ar­ gument in section IV seems excessively epistemic, in ways reminiscent of positivism and the pas­ sages of the later Wittgenstein to which I allude. Although the possibility of first-person skepticism that I emphasize is a disturbing consequence of the considerations I marshal, the ultimate argu­ ment against (strong) consciousness that I am raising does not rest on it alone. Rather, my point is that such a property is explanatorily otiose: it plays no more role in any warranted account of the world than do angels or absolute space. We have no non-question-begging reason to postu­ late such properties. As I argue in Rey 1 994, 1996, 1 997, chap. 1 1 , it is this metaphysical insight­ that such consciousness would be like "a wheel

A Question about Consciousness

that can be turned though nothing else moves with it" (Wittgenstein 1 953: §271 )-that lies bur­ ied amid Wittgenstein's admittedly far too epis­ temic discussions of "private languages" and the like. Elsewhere (Rey 1993, 1994, 1 996, and 1 997: chap. 1 1 ), I develop the suggestions of the last two paragraphs of section IV: that our notion of (strong) consciousness involves a projection of our ineluctable (perhaps weakly conscious) re­ actions to the appearances of our conspecifics and things that look like them. I argue that it is this projection that underlies our insistence on what Wittgenstein rightly noted was the poor "picture" of the "inner" modeled on the "outer" that has so many of us in its grip. Notes This chapter is a slightly revised version of Rey (1 986), which was a substantial revision of (1 983a). Although details could be changed and updated, none wC\uld affect the general strategy or conclusions. Intermediate drafts have served as bases for talks at the Universitil's of Arizona, Illinois. Belgrade, Ljubljana, Graz, and Vienna, at Bates College, at the 1984 Dubrovnik Con­ ference in the Philosophy of Science, and at the filo­ zofski fakultet, Zadar (Croatia). I am indebted to these audiences for their hospitality and stimulating re­ sponses. I am also grateful to the National Endowment for the Humanities and the Fulbright Commission for financial support, and to Louise Antony, Ned Block, Richard Davidson, Joe Levine, Gary Matthews, Eliz­ abeth Robertson, Michael Siote, and Eleanor Saunders for helpful discussions and suggestions.

I. For more recent developments of the Humean argu­ ment, see Parfit (197 Ia), and for replies, Lewis (1 976) and Rey (1 976). 2. See many of the essays in, e.g., Eccles (1 966); Globus, Maxwell, and Savodnik ( 1976); Schwartz and Shapiro (1 976); Davidson and Davidson (1 980); and Davidson, Schwartz, and Shapiro (1 983). 3. See , e.g., Wigner ( 1 967, 1 982) and Wheeler (1 982). 4. The Oxford English Dictionary lists ten definitions of "conscious," seven of "consciousness," on the latter of which Natsoulas (1 978) expanded, although Ryle (1 949) considered only seven for both.

475

5. That the business of providing definitions of even commonly used words is not an entirely a priori activity, but may go hand in hand with the development of a theory of their referents seems a natural corollary of re­ cent discussions of reference in Putnam (I 970, 1 975) and Kripke ( 1 972/1980). I develop this in relation to concept identity in Rey (1 983b, 1985, 1997: 1 .5). 6. I shall use "inductive" here very broadly so as to in­ clude not only principles of enumeration, but also any "reasonable" relations between evidence and explana­ tory hypotheses, especially those associated with "ab­ duction," "analysis-by-synthesis," and "inference to the best explanation"; see, e.g., Peirce (1901 / 1 955). MacKay ( 1 95 1), Harman ( 1 965), Chomsky (1 972). 7. Behaviorists thought they had a better one that didn't advert to any mental terms at all. I take for granted the standard criticisms of that view, as found in, e.g., Chomsky (1 959), Taylor (1 963), Dennett (1 978, Pt. 2, sec. 4, pp. 53-70), Gleitman (1 981 : Ch. 5), many of which I summarize in Rey (1 984). 8. I discuss practical reason and its status as a law of thought in Rey (1 987). See also, e.g., Horgan and Woodward (1 985); and Elster (1 986). 9. The properties of a sentential model that seem to me to recommend it are its capabilities of (a) expressing structured propositions; (b) capturing rational relations into which attitudes enter; (c) individuating attitudes sufficiently finely to distinguish ones containing, e.g., synonymous descriptions, co-referential names, and in­ dexicals (cf. Burge 1978; Kripke 1 979; Bealer 1 982; Ch. 3); (d) explaining how attitudes can be causally effica­ cious; and (e) allowing different roles and access rela­ tions for different attitudes. It is difficult to imagine any sort of representation other than a sentence that can perform all these roles (e.g., images certainly COUldn't). 10. This seems to be the main issue bothering Searle (1 980) in his well-known example of the "Chinese Room." In Rey (1 986), I discuss this and related issues that I think are confounded in Searle's discussion. The Stampe (1 977) proposal that I advocate here and there has also been advanced in slightly different form by Dretske (1981), Stalnaker (1 984), and Fodor (1 987). I I . This is not the place to defend such a theory of meaning in detail. Suffice it to say that (a) I intend it only as a sufficient condition of meaning: there may be other ways that meaning may arise, but should some­ thing be able to operate as described, its states would seem to play the role that meaningful states seem ordi­ narily to play and (b) such a basis figures in a variety of

Georges Rey

theories of meaning, from truth-conditional proposals of Wittgenstein (1 920) and Davidson (1 967), to possi­ ble world proposals of Stalnaker (1 984), to discrim­ ination proposals one finds in the work of Behaviorists like Quine (1 960: Ch. 2) and even in Searle ( 1 983: 1 77). The proposal is plausibly confined to the primitive predicates of a language, the semantics of logical par­ ticles relying on their inferential roles, and that of com­ plex expressions on standard recursions (capturing Searle's (1 984) "aspectual shape"). 1 2 . I have in mind here the "frame problem" currently vexing work in artificial intelligence (see McCarthy and Hayes 1 969, and Dennett 1 984 for useful discussion). Although I disagree with many of the ways Dreyfus (1 972) proposes for thinking about the problem, I do think it a merit of his book that it anticipated many as­ pects of it, particularly what Fodor (1 983) has called the "globality" and "Quinity" of central belief fixation. In distinguishing artificial intentionality from artificial in­ telligence. I hope I've made .it plain that none of this discussion depends on a solution to these problems. 1 3 . Some, e.g., Lucas (1961, 1968), have argued that it is a consequence of the Gooel Incompleteness theorem that the human mind is no machine; so, the machine I am imagining would necessarily fall short of human mentation. The short answer to this argument is that it presumes human minds are consistent. Aside from the vast implausibility of that claim generally, the mecha­ nist may also reply that, if indeed we can decide the formally undecidable sentence, we do it only on pain of inconsistency. See Lewis (1 969, 1 979) for longer an­ swers, and Chemiak (1 986) for an excellent discussion of the "minimal rationality" that needs to be required of us, and consequently of any machine. 14. In considering here and below the consciousness of machines programmed in this and other ways, I shall be citing what I take to be ordinary pre-theoretic intuitions about the notion of consciousness. I S. For similar, but less complex proposals by leading neurophysiologists, see Moruzzi (1 966), Knapp (1 976), John ( 1 976). Wilks ( 1 984: 1 23) makes the criticism I make here of such proposals in computer science of, e.g., Sayre (1 976).

16. One might argue that genuine thoughts, intentions, perceptions presuppose consciousness at least as a background condition. But they would need to account for the explanatory power, adumbrated above, that such ascription seems to possess without it. In any case

476

we would still be entitled to an answer to the question being addressed in this paper, that of precisely wherein this presupposition consists. 17. See also Casteneda (1 968), Anscombe (1 975), Perry (1 979). 1 8 . Where being "self-aware" or "self-conscious" doesn't merely verbally entail consciousness. That is, to avoid begging the question, these expressions must be used merely for emphatic re/lexive propositional attitudes. 19. What is wanted here to clinch the case are experi­ mentally controlled examples a la Nisbett and Wilson (1 977) of the effects of specifically second-order em­ phatic reflexive thoughts that the agent is unable to ex­ press. Unfortunately, Laing et al. (1 966) don't provide any, and I've been so far unable to find any studies that do (Gur and Sackheim (1 979) in their interesting work on self-deception provide some, but inconclusive evi­ dence in this regard). However, it's hard to believe that this isn't simply an accident due to the disinterest of psychologists in specific philosophical theses. It's not at all difficult to imagine plausible psychoanalytical ex­ planations adverting to second-order thoughts that the agent is repressing: e.g., someone might unconsciously think that he wants to murder his father, and because he can't bear to admit this, neither the thought nor the want is permitted to enter consciousness; the uncon­ scious thought explains, though, his "free /loating" feelings of guilt, which the unconscious want by itself would fail to do. Another objection to second-order views like Rosen­ thal's and Smith's is that they conflate an agent's being conscious of a thought with that thought's being con­ scious: an "educated neurotic" might eagerly accept Freudian theory, and ascribe to himself all manner of unconscious motives that he becomes "conscious of", without them becoming thereby conscious. The "buffer memory" I propose below, and discuss in detail in Rey (1 987) affords a way to deal with this problem in a way that is not committed to second-order attitudes. 20. See also Wilks (1 984, 1 22), who endorses a similar criterion, citing Danto (1 960). 2 1 . Dennett (I 978a, 1 56) himself makes such a pro­ posal, one that, incidentally, permits ascribing con­ sciousness to animals and to the humanly inarticulate. In Rey (1 987) I develop this proposal in some detail, distinguishing "avowed" from "central" attitudes, and arguing that the distinction is important to us in a

477

A Question about Consciousness

number of ways, for example, as a basis for an account

plausible account of mental states that preserves first­

of akrasia and self-deception. I am inclined to think that

person privileges. Unfortunately it seems to do so at the

the proposal does capture a "weak" notion of con­

cost of third person attributions of such states to a

sciousness, i.e., consciousness as we apply it to human

computer of the sort I am describing.

beings and perhaps some animals (see below); but it

27. See, e.g., Descartes

would seem to fail to capture the full, ordinary notion

IV, 2), Hume (Treatise, I, iv, 2), Ayer ( 1 956, 55), Mal­

insofar as it is equally applicable to existing desk top

colm (1963, 85), Chishoim ( 1 98 1 , 79-8 1). For a useful

(MeditatiOns, II),

Locke

(Essay,

computers which few people would seriously regard as

survey of these and other versions of privileged access,

conscious for a moment.

see Alston ( 1 97 1 ). For interesting experimental data

22. The problem here is

not

that raised by Searle's

undermining the privileges even in the human case, see

"Chinese Room," concerning whether anyone follow­

Hilgard (1 977).

any program could be said to understand Chinese. That problem is solved simply by considering richer

It now seems to me that all that is necessary are all the

(particularly recursive semantic) programs than Searle

probably very subtle cognitive effects of the hormones

considers, as well as by avoiding fallacies of division.

and neuroregulators (e.g., beliefs about how one's body

ing

See Rey (1986, sec. I).

28. I am no longer confident that that claim is plausible.

feels, the timing of cognitive processes). If this were true,

23. So that the machine won't spend all of its time

then the appeal to physiology here would be even less

merely brooding over increasingly nested beliefs, a limit

plausible than I argue it to be below.

would need to be placed on the nestings generated by

29. Lycan (1981) has pointed out that in general one

(6), particularly as it might interact with (7).

would expect our various psychological faculties to be

24. Bealer (1 982, 247) claims that "attending, concen­

characterizable on many different levels of functional

trating, meditating, contemplating" do not have "me­

abstraction from the basic physical properties of the

chanically recognizable functions." I fail to see why the

brain.

mechanical operations described here wouldn't essen­ tialIy suffice.

30. That is, it is possible within the world(s) described by the psychological laws being considered here. It is

25. Searle ( 1 98 1 ) argues that, since someone could fol­

important that the possibility here is not merely epis­

low a computer program like the one I am sketching

temic but metaphysical, Were it only epistemic, it could

without feeling pain, a machine running on such a pro­

be dismissed as due merely to ignorance, as many pre­

gram wouldn't feel pain either. This, of course, is simply

vious arguments, e.g., against materialism reasonably

another version of his "Chinese Room" argument, and

have been, cf. Maloncy ( 1 985: 32).

would seem to be guilty of the same fallacy of division

3 1 . I say "unacceptable," not "false," I leave open as

that advocates of the "systems" reply have deplored,

part of the puzzle I am raising whether one is entitled to

and which I have further criticized in Rey (1 986).

move from the former to the latter.

Briefly, the person following a program corresponds to a computer's central processing unit, which no more needs to share properties of the entire system than Ronald Reagan needs to share the properties of the United States.

32. By 'x is infallible about p' here I intend no claim about knowledge, but only 'it is impossible for x to think or believe that p when not-p': this latter is closed under logical implication even if 'x knows p' isn't. For pur­ poses of contrast with "arcane" conditions. I'm also

26. The important (im)possibility of scepticism about

presuming here that the agent is

one's own present sensations that I shall be considering

mal, thinking that (s)he/it is conscious clearly and con­

in what follows was, I believe, first raised by Wittgen­

fidently, not in the way one might do when asleep or in

stein ( 1 953/ 1 967, 243-3 1 5) in his famous "private lan­

some other cognitively marginal state.

guage argument." Shoemaker ( 1 975/ 1 984, 1 89- 1 90) in a passage that was the inspiration for much that I have written here, takes the impossibility of such scepticism to be an argument for functionalism, a specific version of which I am developing in this paper. I think Shoe­ maker is right in regarding functionalism as the only

cognitively

fully nor­

33. Recent proponents of life as a condition on con­ sciousness include Wittgenstein ( 1 953/ 1 967, p. 97e), Ziff ( 1 959), Cook ( 1 969), Gareth Matthews ( 1 977).

34. Why shouldn't the computer I am imagining have the status of an angel, a god, a ghost, or some other non-biological, but still psychological agent?

Georges Rey

35. It is some indictment of our ordinary notion of consciousness that we seem to have contradictory in­ tuitions about its application, our judgments about something's consciousness being sensitive to the way in which we encounter it. Described from the "bottom up," i.e., imagining first a non-biological machine, and then a program that would realize various human men­ tal abilities, we are disinclined to regard it as conscious; but proceeding from the "top down" and discovering that close friends, or we ourselves, are just such ma­ chines seems to elicit the opposite verdict. I suspect this has to do with the ways in which our understanding of people and animals are "triggered" by certain patterns of stimulation (cf. the penultimate paragraph of this essay). 36. The image of consciousness as an inner light is ad­ vanced in many places, perhaps most recently in Smith (1986). Even Wittgenstein, so critical of traditional pictures of the mind, compares "turning my attention to my own consciousness" to the glance of someone "admiring the illumination of the sky and drinking in the light" (1953/1 967, 1 24). Other passages in Witt­ genstein, e.g., sec. 308, do suggest, however, a version of the e1iminativist proposal being avocated here. The peculiar amalgamation of epistemology and ethics that the traditional image involves is explored at some length by Richard Rorty (1979:42 .. 69), when he discusses what he regards as the traditional belief in "Our Glassy Es­ sence," a term he draws from Shakespeare and Peirce. The view I am entertaining in this chapter might be called "eliminativist" with respect to consciousness, although (token ontologica1ly) "reductionist," with regard to the propositional attitudes. 37. Cf. Wittgenstein (1953/1 967: p. 178e): "My attitude towards him is an attitude towards a soul. I am not of the opinion that he has a soul." 38. James (1912) also denied the existence of con­ sciousness, although he went on to explain that he meant "only to deny that the word stands for an entity, but to insist most emphatically that it does stand for a function" (p. 4). When I say there may be no such thing, I mean no such thing wluJtsoever.

References Alston, W. (1971). "Varieties of Privileged Access." American Philosophical Quarterly 8(3), pp. 223-24 1 .

478

Anscombe, G. (1975). "The First Person." In Mind and Language, Edited by S. Guttenplan, Oxford: Oxford Univ. Press, pp. 45-66. Ayer, A. J. (1956). The Problem of Knowledge. London: Macmillan. Bealer, G. (1982). Quality and Concept. New York: Oxford University Press. Berliner, H. (1 980). "Computer Backgammon." Scien­ tific American 242(6), pp. 64-85. Block, N. J. (1 978). "Troubles with Functionalism." In

Readings in the Philosophy of Psychology, vol. 2. Edited by N. Block. Cambridge: Harvard University Press. Brown, G. (1974). "The Believer System." Technical Report R UCBM-TR-34. Department of Computer Science, Rutgers University. Burge, T. (1978). "Belief and Synonymy." Joumal of

Philosophy 75(3), March, pp. 1 19-1 38. Castenada, H. (1968). "On the Logic of Attributions of Self-Knowledge to Others." Joumal of Philosophy 65, pp. 439-456. Chemiak, C. (1 986). Minimal Rationality. Cambridge: MIT Press/Bradford Books. Chisholm, R. M. (1981). The First Person. Minneapolis: University of Minnesota Press. Chomsky, N. (1959). Review of Skinner's Verbal Be­

havior. Language 35, pp. 26-58. . ( 1972). Language and Mind (enlarged edition). New York: Harcourt Brace Jovanovich. --

Cook, J. (1969). "Human Beings." In Studies in the Philosophy of Wiltgenstein, pp. 1 17-1 5 1 . Edited by P. Winch. New York: Humanities Press. Danto, A. (1961). "On Consciousness in Machines." In

Determinism and Freedom in the Age of Modern Science. Edited by S. Hook. New York: Collier. Darley, J. M., and E. Berscheid (1967). "Increased Liking as a Result of the Anticipation of Personal Contact." Human Relations 20, pp. 29-40. Davidson, D. (1 967). "Truth and Meaning." Synthese, 17, pp. 304-323. Davidson, J., and R. Davidson, eds. (1980). The Psy­ chobiology of Consciousness. New York: Plenum Press. Davidson, R., G. Schwartz, and D. Shapiro, eds. (1 983). Consciousness and Self-Regulation vol. 3. New York: Plenum Press. Dennett, D. (1969). Content and Consciousness. Lon­ don: Routledge & Kegan Paul.

A Question about Consciousness

. ( 1 97 1 / 1 978). "Intentional Systems." Journal of

479

--

Hempel, C. (1 965). Aspects ofScientific Explanation and Other Essays in the Philosophy of Science. New York: Free Press.

Books.

Hess, R. ( 1 975). "The Role of Pupil Size in Communi­ cation." Scientific American 233(5), pp. 1 10-1 1 8.

--

Philosophy 68, pp. 87- 1 06. . ( 1 978). Brainstorms: Philosophical Essays on Mind and Psychology. Cambridge: MIT Press/ Bradford

. ( 1 984). "Cognitive Wheels: The Frame Prob­ lem of AI." In Minds, Machines and Evolution. Edited by C. Hookway. Cambridge: Cambridge University Press. --

Descartes, R. ( 1 64 1 / 1 9 1 1). "Meditations on First Phi­ losophy (with Replies and Objections)." In The Philo­ sophical Works of Descartes. Edited by E. S. Haldane and G. R. T. Ross. London: Cambridge University Press. Dretske, F. (1981). Knowledge and the Flow of Infor­ mation. Cambridge: MIT Press/Bradford Books. Dreyfus, H. (1 972). What Computers Can't Do, revised edition. New York: Haqler and Row. Eccles, 1. c., ed. (1 966). Brain and Conscious Experi­ ence. Berlin: Springer Verlag. Elster, J., ed. (1 986). Rational Choice. New York: New York University Press. Festinger, L. (1 957). Cognitive Dissonance. Stanford: Stanford University Press. Field, H. ( 1 978). "Mental Repressentation." In Read­ ings in the Philosophy of Psychology, vol. 2. Edited by N. Block. Cambridge: Harvard University Press. Fodor, J. A. ( 1 975). The Language of Thought. New York: Crowell. . ( 1 983). The Modularity of Mind. Cambridge: MIT Press.

--

'--. (1 987). Psychosemantics. Cambridge: MIT Press. Gleitman, H. (1981). Psychology. New York: Norton & Co.

Globus, G., G. Maxwell, and I. Savodnik, eds. ( 1 976). Consciousness and the Brain. New York: Plenum Press. Grice, H. P. (1 957). "Meaning." Philosophical Review 66, pp. 337-388. Gur, R., and H. Sackheim ( 1 979). "Self-Deception: A Concept in Search of a Phenomenon." Journal of Per­ sonality and Social Psychology 37(2). Harman, G. (1 965). "Inference to the Best Explana­ tion." Philosophical Review 74, pp. 88-95. . ( 1 973). Thought. Princeton: Princeton Uni­ versity Press. --

Hilgard, E. ( 1 977). Divided Consciousness: Multiple Controls in Human Thought and Action. New York: Wiley & Sons. Horgan, T. and Woodward, 1. ( 1 985), "Folk Psychol­ ogy is Here to Stay." Philosophical Review 94, 2, pp. 1 97-225. Hume, D. ( 1 739/1 965). A Treatise of Human Nature, Edited by L. A. Selby-Bigge. Oxford: Clarendon. Jackson, F. (1 982). "Epiphenomenal Qualia." Philo­ sophical Quarterly 32, pp. 1 27-1 36. James, W. ( 1 9 1 2). "Does Consciousness Exist?" In

Essays in Radical Empiricism and Pluralistic Universe. Edited by R. B. Perry. New York: Dutton. John, E. (1 976). "A Model of Consciousness." In Con­ sciousness and Self-Regulation, Vol. I . Edited by G. Schwartz and D. Shapiro. New York: Plenum. Kahneman, D. (1 973). A ttention and Effort. Englewood Cliffs: Prentice-Hall. Knapp, P. ( 1 976). "The Mysterious 'Split': A Clinical Inquiry into Problem of Consciousness and Brain." In Consciousness and the Brain. Edited by G. qlobus, G. Maxwell, and I. Savodnik. New York: Plenum Press. Kripke, S. A. ( 1972/1 980). Naming and Necessity. Cambridge: Harvard University Press. -- ' . (1 979). "A Puzzle about Belief." In Meaning and Use, pp. 239-283. Edited by A. Margalit. Dor­ drecht: Reidel. Lackner, J. R., and M. Garrett ( 1 972). "Resolving Ambiguity: Effects of Biasing Context in the Un­ attended Ear." Cognition I . Laing, R., H . Phillipson, and A . Lee ( 1 966). Inter­ personal Perception: A Theory and Method of Research. London: Ravistock. Latane, B., and J. M. Darley (1 970). The Unresponsive Bystander: Why Doesn't He Help? New York: Apple­ ton-Century-Crofts. Leeds, S. (1 992). "Qualia, Awareness and Sellars." Nous, pp. 303-329. Lewis, D. (1 969). Convention: A Philosophic Study. Cambridge: Harvard University Press.

480

Georges Rey

. (1 976). "SUIvivaI and Identity." In The Identi­ ties of Persons, pp. 1--40. Edited by A. Rorty. Berkeley:

--

University of California Press. -

--

. (1979). "Lucas against Mechanism II." Cana­

Nisbett, R., and T. Wilson (1977). "Telling More Than We Can Know." Psychological Review, pp. 23 1259. Parfit, D. (197 I a). "Personal Identity." Philosophical

dian Journal of Philosophy 9, pp. 373-376.

Review 80(1), pp. 3-28. . (197 I b). "On the Importance of Self-Identity."

Lucas, J. (1961). "Minds, Machines, and GOdel." Phi­

--

losophy 36, pp. 1 12-137.

Journal of Philosophy, pp. 683-690.

. (1968). "Satan Stultified: A Rejoiner to Paul Beneceraff." Monist 52, pp. 45-48.

Peirce, C. S. (1901/1955). "Abduction and Induction." In Philosophical Writings ofPeirce, pp. 1 50-1 56. Edited by J. Buchler. New York: Dover.

--

Luria, A. (1976). "The Human Brain and Conscious Activity." In Consciousness and Self-Regulation, vol. I . Edited by F . Schwartz and D. Shapiro. New York: Plenum Press. Lycan, W. (198 1). "Form, Function, and Feel." Journal of Philosophy 78(1), pp. 24-50. . (1990). "What Is the 'SUbjectivity of the Men­ tal'?" In J. Tomberlin, ed., Philosophical Perspectives, Vol. 4: Action Theory and the Philosophy of Mind. Atascadero, CA: Ridgview. --

McCarthy, J., and P. Hayes (1969). "Some Philosoph­ ical Problems from the Standpoint of Artificial In­ telligence." In Machine Intelligence, Vol. 4, pp. 463502. Edited by B. Meltzer and D. Michie. Edinburgh: Edinburgh University Press. MacKay, D. (195 1). "Mind-Like Behavior in Arti­ facts." British Journal for the Philosophy of Science 2, pp. 1 05- 1 2 1 . Maier, N . (1931). "Reasoning i n Humans: II. The Solution of a Problem and Its Appearance in Con­ sciousness." Journal of Comparative Psychology 12, pp. 1 8 1 -1 94. Malcolm, N. (1 963). Knowledge and Certainty. Engle­ wood Cliffs: Prentice-Hall. Maloney, J. (1985). "About Being a Bat." Australasian

Journal of Philosophy, 63(1), pp. 26--47. Matthews, G. (1 977). "Consciousness and Life." Phi­ losophy 52, pp. 1 3-26. Miller, G. (1 956). "The Magic Number Seven Plus or Minus Two: Some Limits on Our Capacity for Proces­ sing Information." Psychological Review 63, pp. 8 1 -97.

Penfield, W. (1 975). The Mystery of the Mind. Prince­ ton: Princeton University Press. Perry, J. (1979). "The Problem of the Essential Index­ ical." Nous 1 3, pp. 3-2 1 . Piaget, 1 . ( 1976). The Child and Reality: Principles of Genetic Epistemology. New York: Penguin. Pribram, K. (1 976). "Problems Concerning the Struc­ ture of Consciousness." In G. Globus, G. Maxwell, and I. Savodnik, eds., Consciousness and the Brain. New York: Plenum Press. Putnam, H. (1970/ 1975). "Is Semantics Possible?" In his Collected Papers, Vol. 3, pp. 1 39- 152. Cambridge: Cambridge University Press. --

. (1975). "The Meaning of Meaning." In in his

Collected Papers, Vol. 3, pp. 21 5-27 1 . Cambridge: Cambridge University Press. Quine, W. V. (1960). Word and Object. Cambridge: MIT Press. Rey, G. (1976). "Survival." In The Identities of Persons, pp. 4 1 -66. Edited by A. Rorty. Berkeley: University of California Press. ---

. (1 980). "Functionalism and the Emotion." In

Explaining Emotions. Edited by A. Rorty. Berkeley: University of California Press. --. (1983a). "A Reason for Doubting the Existence of Consciousness." In Consciousness and Self-Regu­ lation, Vol. 3. Edited by R. Davidson, G. Schwartz, and D. Shapiro. New York: Plenum. --. ( 1983b). "Concepts and Stereotypes." Cogni­ tion 1 5, pp. 237-262.

Moruzzi , G. (1 966). "Functional Significance of Sleep with Particular Regard to the Brain Mechanisms Un­ derlying Consciousness." In J. C. Eccles, ed., Brain and Conscious Experience. Berlin: Springer Verlag.

--. (1984). "Ontology and Ideology of Behavior­ ism and Mentalism." Commentary on B. F. Skinner, "Behaviorism at Fifty." Behavioral and Brain Sciences 7(4), pp. 640-641 .

Natsoulas, T. (1978). "Consciousness." American Psy­

--. (1985). "Concepts and Conceptions: A Reply to Smith, Medin, and Rips." Cognition 19.

chologist 1 2, pp. 906-914.

481

A Question about Consciousness

. . (1 986). "What's Really Going On is Searle's Chinese Room." Philosophical Studies. --

-- . ( 1 98 1). "Analytic Philosophy and Mental Phe­ nomena." In Midwest Studies in Philosophy, Vol. 6. Edited by P. French, T. Uehling, and H . Wettstein. Minneapolis: University of Minnesota Press.

. (1 987). "Beliefs, Avowals, Akrasia, and Self­ Deception." In Perspectives on Self-Deception. Edited by A. Rorty and B. MacLaughlin. Berkeley: University of California Press.

. ( 1983). Intentionality: An Essay in the Philoso­ phy of Mind. Cambridge: Cambridge University Press.

. (1 992). "Sensational Sentences." In Conscious­ ness. Edited by M . Davies and G. Humphysies. Oxford:

--. (1 984). "Intentionality and Its Place in Na­ ture." Synthese 6 1 , pp. 3-16.

Blackwell.

Shoemaker, S. (1 975). "Functionalism and Qualia." Philosophical Studies 27, pp. 29 1 -3 1 5.

--

--

. (1 993). "Sensational Sentences Switched." Philosophical Studies. Special issue. --

... (1 994). "Wittgenstein, Computationalism and Qualia." In Philosophy and the Cognitive Sciences, pp. 61 -74. Edited by R. Casati and G. White. Vienna: Holder-Pichler-Tempsky. --

.--. (1 996). "Towards a Projectivist Account of Conscious Experience." Ferdinand-Schoningh-Verlag. (German version "Anniiherung an eine projectivistische Theorie bewuBten Erlebens" published by same, 1995.) --. (1 997). Contemporary Philosophy of Mind: A Contentiously Classical Approach. Oxford: Blackwell. Rorty, T. (1 979). Philosophy and the Mirror of Nature. Princeton: Princeton University Press. Rosenthal, D. (1 984). "Armstrong's Casual Theory of the Mind." In D. M. Armstrong, pp. 79- 1 20. Edited by R. Bogdan. Dordrecht: Reidel. Ryle, G. (1949). The Concept of Mind. London: Hutch­ inson.

---

Smith, D. W. (1 986). "The Structure of (Self-)Con­ sciousness." Topoi. Stalnaker, R. (1 984). Inquiry. Cambridge: MIT Press/ Bradford Books. Stampe, D. (1 977). "Towards a Causal Theory of Lin­ guistic Representation." In Midwest Studies in Philoso­ phy II: Studies in the Philosophy of Language, pp. 4263. Edited by H. Wettstein and T. Uehling. Morris: University of Minnesota Press. Stich, S. P. (1 983). From Folk Psychology to Cognitive Science: The Case Against Belief Cambridge: MIT Press/Bradford Books. Taylor, C. (1 963). The Explanation of Behavior. Lon­ don: Routledge & Kegan Paul. Weber, M. (1 922/1 980). "The Nature of Social Action. " In Weber: Selections in Translation. Edited by W. G. Runciman. Cambridge: Cambridge University Press.

Schiffer, S. (1 972). Meaning. Oxford: Oxford University Press.

Weizenbaum, J. (1 965). "ELIZA-A Computer Pro­ gram for the Study of Natural Language Communica­ tion Between Men and Machine." Communication of the Association for Computing Machinery 9(1), pp. 3645.

--. (1 980). "Truth and the Theory of Content." In Meaning and Understanding. Edited by H. Parrett and J. Bourveresse. Berlin: de Gruyter.

Wheeler, E. (1 982). "Bohr, Einstein, and the Strange Lesson of the Quantum." In Mind in Nature, pp. 1 -30. Edited by R. Elvee. San Francisco: Harper and Row.

Schmidt, C., and G. D'Addami (1 973). "A Model of the Common Sense Theory of Intention and Personal Causation. " In Proceedings of the Third International Joint Conference on Artificial Intelligence. Standford: Stanford University Press.

Wiener, N. (1954). The Human Use of Human Beings. Garden City, NY: Doubleday.

Schwartz, G., and D. Shapiro, eds. (1976). Conscious­ ness and Self-Regulation, Vol. 1 . New York: Plenum

--. (1 982). "The Limitation of the Validity of Present Day Physics. " In Mind in Nature, pp. 1 1 81 33. Edited by R. Elvee. San Francisco: Harper and Row.

Sayre, K. ( 1 976). Cybernetics and the Philosophy of Mind. Atlantic Highlands, NJ: Humanities Press.

Press. Searle, J. R. (1 980). "Minds, Brains, and Programs." Behavioral and Brain Sciences 3, pp. 41 7--457.

Wigner, E. (1 967). "Remarks on the Mind-Body Ques­ tion." In Symmetries and Reflections, pp. 1 7 1 - 1 84. Bloomington: Indiana University Press.

Wilks, Y. (1 984). "Machines and Consciousness." In Minds, Machines and Evolution. Edited by C. Hookway. Cambridge: Cambridge University Press.

Georges Rey

482

Wittgenstein, L. (1 920/1967). Tractatus Logico-Philo­ sophicus. Translated by D. Pears and B. McGuiness. London: Routledge & Kegan Paul. ---

.

(1953/1 967).

Philosophical Investigations.

Translated by G. E. M. Anscombe. 3d edition. Oxford: Blackwell. Zajonc, R. (1968). "The Attitudinal Effects of Mere Exposure." Journal of PersofUJlity and Social Psychol­

ogy 8. Ziff, P. (1964). "The Feelings of Robots." In Minds and Machines, pp. 98-103. Edited by A. R. Anderson. Englewood Cliffs, NJ: Prentice-Hall.

29

Finding the Mind in the Natural World

Frank Jackson Conceptual analysis played a prominent role

the world is, at bottom, like. The second reason is

in the defense of materialism mOWlted by the

that the history of conceptual analysis is the his­

Australian materialists and their American ally

tory of failure. For any proffered analysis some­

David Lewis. It was how they fOWld a place for

one clever always finds a cOWlterexample. The

the mind within the material world. The leading

final reason turns on the claim that we have

idea is encapsulated in the following argument

learned from Hilary Putnam and Saul Kripke

schema:

1. F

2.

about the necessary

Mental state M = occupant of functional role (By conceptual analysis) Occupant of role

F = brain

state

B

(By

science) Therefore, M = B

a posteriori,

and that tells us

that there can be necessary connections that, pre­ cisely by virtue of being

a posteriori,

are not re­

vealed by or answerable to conceptual analysis. The materialist should, according to this line of thought, hold that the connection between the

(By Transitivity)

This schema gives the role of conceptual analysis in the Australian defense. But it does not tell us why conceptual analysis

had to have a role in the

defense. Indeed, the schema positively invites the thought that conceptual analysis was not needed. For to get the conclusion that M = B, all that is needed is the truth of the two premises. It is not necessary that one of them be a conceptual truth. And I think, speaking more generally, that the Australian materialists left it Wlclear why mate­

mental and the material or physical is a necessary

a posteriori one, and so not a matter accessible via conceptual analysis. During the course of the dis­ cussion we will see how to reply to each of these objections to the need for conceptual analysis in the defense of materialism. The first step in our defense of the materialists' need for conceptual analysis is to note that mate­ rialism is a piece of what I will call serious metaphysics, and that, like any piece of serious metaphysics, it faces the location problem.

rialists need to do some conceptual analysis. Nevertheless, I think that they were right that materialists need to do some conceptual analysis. This paper is a defense of this view. In a nutshell my argument will be that only by doing some conceptual analysis can materialists find a place for the mind in their naturalistic picture of the world. In a final section we will note the im­ plications of our discussion for the knowledge argument. In arguing for the necessity of conceptual analysis I am swimming against the tide. Current orthodoxy repudiates the role of conceptual analysis in the defense of materialism for at least three reasons. First, materialism is a doctrine in speculative metaphysics. And, runs the first rea­ son, though conceptual analysis has a role in the philosophy oflanguage and the study of concepts, it has no essential role when our subject is what

The Location Problem Metaphysics is about what there is and what it is like. But it is concerned not with any old shopping list of what there is and what it is like. Meta­ physicians seek a comprehensive account of some subject matter-the mind, the semantic, or, most ambitiously, everything-in terms of a limited number of more or less fWldamental notions. Some who discuss the debate in the philosophy of mind between dualism and monism complain that each position is equally absurd. We should be pluralists. Of course we should be pluralists in some sense or other. However, if the thought is that any attempt to explain it all, or to explain it all as far as the mind is concerned, in terms of some limited set of fundamental ingredients is mistaken in principle, then it seems to me that we

Frank Jackson

484

are being, in effect, invited to abandon serious

in addition to resisting encroachment. But resist­

metaphysics in favor of drawing up big lists. And

ing encroachment explains the stubbing of toes

we know we can do better than that. At least

quite well enough for it to be pedantic to insist on

some of the diversity in our world conceals an

anything more in order to be solid. Hence, solid­

underlying identity of ingredients. The diversity is

ity gets a location or place in the molecular story

a matter of the same elements differently selected

about our world by being entailed by that story,

and arranged. But if metaphysics seeks compre­

and we see this by asking ourselves about our

hension in terms of limited ingredients, it is con­

concept of solidity in the sense of asking what it

tinually going to be faced with the problem of

takes to be solid.

location. Because the ingredients are limited, some

Thus, one way materialists can show that the

putative features of the world are not going to

psychological has a place in their world view is by

appear explicitly in the story told in the favored

showing that the psychological story is entailed

terms. The question then will be whether the fea­

by the story about the world told in the materi­

tures nevertheless figure

alists' favored terms. We will see, however, that it

implicitly

in the story.

Serious metaphysics is simultaneously discrim­

is not just one way; it is the one and only way.

inatory and putatively complete, and the combi­ nation of these two facts means that there is bound to be a whole range of putative features of our world up for either elimination or location. What then is it for some putative feature to have a place in the story some metaphysic tells in its favored terms? One answer is for the feature to be

entailed by the story

told in the favored terms.

Perhaps the story includes information about mass and volume in so many words, but nowhere mentions density by name. No matter-density facts are entailed. by mass and volume facts. Or perhaps the story in the favored terms says that many of the objects around us are nothing but aggregations of molecules held in a lattice-like array by various intermolecular forces. Nowhere in the story in the favored terms is there any mention of solidity. Should we then infer that nothing is solid, or at any rate that this particular metaphysic is committed to nothing being solid? Obviously not. The story in the favored terms will, we may suppose, tell us that these lattice-like arrays of molecules

exclude each other,

the inter­

molecular forces being such as to prevent the lattices encroaching on each others' spaces. And

that

is what we understand by solidity. That is

what it takes, according to our concept, to be solid. Or at least it is near enough. Perhaps pre­ scientifically we might have been tempted to insist that being solid required being everywhere dense

Completeness and Supervenience Materialism is the very opposite of a "big list" metaphysics. It is highly discriminatory, operat­ ing in terms of a small set of favored partic­ ulars, properties, and relations, typically dubbed "physical" -hence, its other name, "physical­ ism"; and it claims that a complete story, or any­ way a complete story of everything contingent, including everything psychological, about our world can in principle be told in terms of these physical particulars, properties, and relations alone. Only then is materialism interestingly dif­ ferent from dual attribute theories of mind. Now what, precisely, is a complete story? We can make a start by noting that one particularly clear way of showing

incompleteness is by appeal

to independent variation. What shows that three coordinates do not provide a complete account of location in space-time is that we can vary position in space-time while keeping any three coordinates constant. Hence, an obvious way to approach completeness is in terms of the lack of indepen­ dent variation. But, of course, lack of independent variation is supervenience: position in space-time supervenes on the four coordinates. So the place to look when looking for illumination regarding the sense in which materialism claim to be com-

Finding the Mind in the Natural World

plete, and, in particular, to be complete with respect to the psychological, is at various super­ venience theses. 1 Now materialism is not just a claim about the completeness of the physical story concerning certain individuals or particulars in our world. It claims completeness concerning the world itself, concerning, that is, the total way things are. Ac­ cordingly, we need to think of the supervenience base as consisting of possible worlds-complete ways things might be. We need, accordingly, to look to global superevenience theses, an example of which is (I) Any two possible worlds that are physical duplicates (physical property, particular and relation for physical property, particular and relation identical) are duplicates simpliciter. But (I) does not capture what the materialists have in mind. Materialism is a claim about our world, the actual world, to the effect that its physical nature exhausts all its nature, whereas (I) is a claim about worlds in general. A more re­ stricted supervenience thesis in which our world is explicitly mentioned is: (II) Any world that is a physical duplicate of our world is a duplicate simpliciter of our world. However, materialists can surely grant that there is a possible world physically exactly like ours but which contains as an addition a lot of mental life sustained in non-physical stuff, as long as they insist that this world is not our world. Consider the view of those theists that hold that materi­ alism is the correct account of earthly existence but it leaves out of account the afterlife. When we die our purely material psychology is reinstated in purely non-physical stuff. Surely materialists can grant that these theists are right about some world, some way things might be, as long as they insist that it is not our world, not the way things actually are. Hence, materialists are not com­ mitted to (II). The trouble with (II) is that it represents mate­ rialists' claims as more wide-ranging than they in

485

fact are. What we need is something like (II) but that limits itself to worlds more nearly like ours, or at least more nearly like ours on the materialists' conception of what our world is like. I suggest (III) Any world that is a minimal physical duplicate of our world is a duplicate simpliciter of our world. What is a minimal physical duplicate? Think of a recipe for making scones. It tells you what to do, but not what not to do. It tells you to add butter to the flour but does not tell you not to add whole peppercorns to the flour. Why not? Part of the reason is that no one would think to add them unless explicitly told to. But part of the reason is logical. It is impossible to list all the things not to do. There are indefinitely many of them. Of ne­ cessity the writers of recipes rely on an intuitive understanding of an implicitly included "stop" clause in their recipes. A minimal physical dupli­ cate of our world is what you would get if you used the physical nature of our world (including of course its physical laws) as a recipe in this sense for making a world. We arrived at (III) by eliminating alternatives. But we can give a positive argument for the con­ clusion that the materialist is committed to ( III). Suppose that (III) is false; then there is a dif­ ference in nature between our world and some minimal physical duplicate of it. But then either our world contains some nature that the minimal physical duplicate does not, or the minimal phys­ ical duplicate contains some nature that our world does not. The second is impossible because the extra nature would have to be non-physical (as our world and the duplicate are physically identi­ cal), and the minimal physical duplicate contains no non-physical nature by definition. But if our world contains some nature that the duplicate does not, this nature must be non-physical (as our world and the duplicate are physically identical). But then materialism would be false, for our world would contain some non-physical nature. Hence, if ( III) is false, materialism is false-that is to say, materialism is committed to (III).

Frank Jackson

486

From (TIl) to Entry by Entailment Given that (III) follows from materialism, there is a straightforward and familiar argument to show that if materialism is true, then the psycho­ logical story about our world is entailed by the physical story about our world. We can think of a statement as telling a story about the way the world is, and as being true in­ asmuch as the world is the way the story says it is. Let 0 be the statement which tells the rich com­ plex, and detailed physical story that is tru at the actual world and all and only the minimal physi­ cal duplicates of the actual world, and false else­ where. Let n be any true statement entirely about the psychological nature of our world: n is true at our world, and every world at which n is false differs in some psychological way from our world. If (III) is true, every world at which 0 is true is a duplicate simpliciter of our world, and so a fortiori a psychological duplicate of our world. But then every world at which 0 is true is a world at which n is true-that is, 0 entails n . We have thus derived what we might call the entry by entailment thesis. A putative psycho­ logical fact has a place in the materialists's world view if and only if it is entailed by the physical story about the world. The one and only way of getting a place is by entailment.

:

From Entry by Entailment to ConceptuaJ AnaJysis How does entry by entailment show the iIppor­ tance of conceptual analysis? If 0 entails n, what makes 0 true also makes n true (at least when 0 and n are contingent). But what makes 0 true is the physical way our world is. Hence, the materi­ alist is committed to each and every psychological statement being made true by a purely physical way our world is. But it is the very business of conceptual analysis to address which matters framed in terms of one set of terms and concepts are made true by which matters framed in a dif-

ferent set of terms and concepts. For instance, when we seek an analysis of knowledge in terms of truth, belief, justification, causation, and so on, we seek an account of how matters described in terms of the latter notions make true matters de­ scribed in terms of the former. When we seek an account of reference, we seek an account of the kinds of causal and descriptive facts which make it true that a term names an object. When and if we succeed, we will have an account of what makes it true that "Moses" names in terms of among other things, causal links between uses o the word and Moses himself. And so on and so forth. How could the a priori reflections on, and intuitions about, possible cases so distinctive of conceptual analysis be relevant to, for instance, the causal theory of reference? Well, the causal theory of reference is a theory about the con­ ditions under which, say, "Moses" refers to a certain person. But that is nothing other than a theory about the possible situations in which "Moses" refers to that person, and the possible situations in which "Moses" does not refer to that person. Hence, intuitions about various possible situations-- --the meat and potatoes of conceptual analysis-are bound to hold center stage. (This is particularly true when the test situations cannot be realized. We cannot, for instance, make Twin Earth to check empirically what we would say about whether XYZ is water.) The alternative is to invent our answers. Faced with the question, say, of whether the physical way things are makes true the belief way things are, we could stipulate the conditions under which something counts as a belief in such a way as to ensure that there are beliefs, or, if we preferred, that there are no beliefs. But that would not bear on whether beliefs according to our concept have a place in the materialists' picture of things, only on whether beliefs according to the stipu­ lated concept have a place. In order to address the question of whether beliefs as we understand them have a place, what else can we do but con-

f

Finding the Mind in the Natural World

suit and be guided by our honed intuitions about what counts as a belief'? Would it be better to in­ vent, or to go by what seems counter intuitive? I should emphasize, though, that a sensible use of conceptual analysis will allow a limited but significant place for

a posteriori

stipulation. We

487

solidity in the molecular account of the objects around

us.

Reflection on our concept of solidity

tells us that the molecular account includes solid­ ity, but it does not tell us that solidity is an

tion to

addi­

what appears in the molecular account of

objects, let alone an infallible one.

mentioned earlier the example of finding a place for solidity in the molecular picture of our world, and the fact that what the molecular picture vin­ dicates is the existence of solid bodies according to a conception of solidity cashed out in terms of mutual exclusion rather than in terms of the conjunction of mutual exclusion and being everywhere dense. For our day-to-day traffic with objects, it is the mutual exclusion that matters, and accordingly it is entirely reasonable to rule that mutual exclusion is enough for solidity. The role of conceptual analysis of K-hood is not al­ ways to settle on a nice, neat,

totally a priori list of

necessary and sufficient conditions for being a K-indeed, that is the task that has so often been beyond us. It is rather to guide us in dividing up the cases that clearly are not cases of a K, from the cases that a principle of charity might lead us to allow as cases of a K. Then, armed with this in­ formation, we are in a position to address the question of whether some inventory of funda­ mental ingredients, does, . or does not, have a place for Ks. I should also emphasize that the contention is not that a priori reflection on possible cases gives us new information, let alone some sort of in­ fallible new information, about what the world is like. The reflection is

a priori

in the sense that

we are not consulting our intuitions about what would

happen

in certain possible cases-it is not

like the famous thought experiments in science­ rather we are consulting our intuitions about how to

describe certain

possible cases. And what we

learn (in the sense of making explicit) is not something new about what the world is like, but something about how, given what the world is like as described in one set of terms, it should be described in some other set of terms. Perhaps the point is clearest in the example about finding

The Objection from the Necessary a Posteriori It might well be urged that the argument given above from (III ) to the conclusion that 0 entails

II is undermined by the existence of necessary a posteriori truths. The objection can be put in two different ways. Consider Over

60 percent of the Earth is covered by H20.

Therefore, over 60 percent of the Earth is covered by water. One way of putting the objection is that although every world where the premise is true is a world where the conclusion is true, the argument is not valid because the premise does not entail the conclusion in the relevant sense. It is not possible to move

a priori

from the premise to the con­

clusion. The premise fixes the conclusion without entailing it, as it is sometimes put. Likewise, for all we have shown by the considerations based on (III), 0 fixes II but does not entail it. This way of putting the objection makes it sound like a quarrel over terminology. It invites the response of distinguishing entailment

pliciter,

sim­

the notion cashed out simply in terms of

being necessarily truth-preserving, from

a priori conceptual, entail­ notion tied to a priori

or, as it is sometimes called, ment, the latter being the

deducibility. But the real objection, of course, is that the necessarily truth-preserving nature of the passage from "Over

60

percent of the Earth

is covered by H20" to "Over

60

percent of the

Earth is covered by water" is not one that can in principle be revealed by conceptual analysis. Re­ flection on, and intuitions about, possible cases and concepts, unless supplemented by the

a pos­ teriori information that water is H20, will get you

488

Frank Jackson

nowhere. Materialists, it seems, can allow that

the very same sense, I think that we should insist

(III) forces them to admit a necessarily truth­

that water's being H20 and water's being water

preserving passage from 0 to II, without allowing

are necessary in the very same sense.

a role for conceptual analysis. They can simply insist that the entailment from 0 to II is an

a

My reason for holding that there is one sense of necessity here relates to what it was that con­ vinced us that "Water = H20" is necessarily true.

posteriori one. We will see, however, that acknowledging the

What convinced us were the arguments of Saul

does not alter matters in

Kripke and Hilary Putnam about how to describe

any essential respects as far as the importance of

certain possibilities, rather than arguments about

necessary

a posteriori

conceptual analysis goes. The argument to this

what is possible per

conclusion turns on a negative claim about the

vinced us that a world where XYZ plays the water

se.

Kripke and Putnam con­

nature of the necessity possessed by the necessary

role-that is, satisfies enough of (but how much is

a posteriori, and a consequent view about the role

enough is vague): filling the oceans, being neces­

of conceptual analysis, in the sense of intuitions

sary for life, being colorless, being called "water"

about possibilities, in the detection of the neces­

by experts, being of a kind with the exemplars we are acquainted with, and so on-did not warrant

sary a posteriori.

the description "world where water is XYZ," and the stuff correctly described as water in a coun­

The Necessity of the Necessary a Posteriori

terfactual world is the stuff-H20-which fills

There are two different ways of looking at the distinction between necessary

a posteriori

state­

ments like "Water = H20" and necessary a priori ones like " H20 = H20" (all necessary modulo worlds where there is no water, of course). You might say that the latter are analytically or con­ ceptually or logically (in some wide sense not tied to provability in a formal system) necessary, whereas the former are

metaphysically necessary,

meaning by the terminology that we are dealing with two senses of "necessary" in somewhat the way that we are when we contrast logical neces­ sity with nomic necessity. One this approach, the reason the necessity of water's being H20 is not available

a priori

is that its necessity is not the

kind that is available

a priori.

I think, as against this view, that it is a mistake to hold that the necessity possessed by "Water = H20" and "If over

60 percent of the Earth is covered by H20, then over 60 percent of the Earth is covered by water" is different from that possessed by "Water = water" and "If over

60

percent of the Earth is covered by H20, then over

60 percent of the Earth is covered by H20." Just as Quine insists that numbers and tables exist in

the water role in the actual world. The key point is that the right way to describe a counterfactuaI world sometimes depends in part on how the ac­ tual world is, and not solely on how the counter­ factual world is in itself. The point is not one about the space of possible worlds in some newly recognized sense of "possible," but instead one about the role of the actual possible world in determining the correct way to describe certain counterfactual possible worlds----in the sense of "possible" already recognized. All this was, it seems to me, an exercise in con­ ceptual analysis. We had an old theory about the meaning of "water," namely, that it meant "that which fills the water role," a theory that was re­ futed by appealing to our intuitions about how to describe possible worlds in which something dif­ ferent from that which actually fills the water role fills the water role. We became convinced of a new theory-again by reflection on possible cases, the meat and potatoes of conceptual analysis-ac­ cording to which "water" is a rigid designator of the stuff that fills the water role in the actual world. At no time did we have to recognize a new sort of possibility, only a new way for something in some counterfactual situation to count as a

K,

489

Finding the Mind in the Natural World

namely, by virtue not solely of how things are in

at least not propositions thought of as sets of

that counterfactual situation, but in part in virtue

possible worlds. For, by the conclusion that we

of how things actually are.

are not dealing with a new sort of necessity, the set of worlds where water is water is the very same

If this is right, the inference

60 percent of the Earth is covered by H20. Therefore, over 60 percent of the Earth is covered Over

by water.

set as the set where water is

osition that water is water differing from the proposition that water is

is not an example of an

a posteriori

entailment

that shows the irrelevance of conceptual analysis to the question of whether an

a posteriori entail­

ment holds. For it is conceptual analysis that tells us, in light of the fact that

H20 fills the water role,

that the entailment holds.

H20, and so, by

Leibniz's Law, there is no question of the prop­

one is not, necessary

H20 in that one is, and

a posteriori.

Their con­

tention is that there are sentences such that the proposition expressed by them depends on the context of utterance. 2 We understand them in that we know how the proposition expressed de­ pends on the context, but if we do not know the relevant fact about the context, we will not know the proposition expressed. ( In Robert Stalnaker's terminology, we know the propositional concept

Two-Dimensionalism and the Knowledge Argument

but not the proposition; in David Kaplan's, we

I have argued that materialists must hold that the

sider "Over

complete story about the physical nature of our

by water." Because "water" is a rigid designator

world given by 0 entails everything about our

whose reference is fixed by "the stuff that fills the

psychology, and that such a position cannot be

water role," someone who does not know what

maintained independently of the results of con­

that stuff is does not know which proposition the

know the character but not the content. 3) Con­

60

percent of the Earth is covered

ceptual analysis. But it is quite another question

sentence expresses, but they understand the sen­

whether they must hold that 0

entails

tence by virtue of knowing how the proposition

everything about our psychology, including its

expressed depends on how things actually are,

a priori

phenomenal side, and so quite another question

and, in particular, this being the relevant con­

whether they must hold that it is in principle pos­

textual matter in this case, on what actually fills

sible to deduce from the full physical story alone

the water role. The explanation of the necessary a

what it is like to see red or smell a rose-the key

posteriori

assumption in the knowledge argument that

Earth is covered by

materialism leaves out

I will conclude by

the Earth is covered by water" then runs as fol­

noting how the two dimensional treatment of the

lows. The proposition expressed by the sentence

qualia.

necessary

a posteriori---the obvious treatment of the necessary a posteriori for anyone sympathetic

"Over

60

status of "If over

60

percent of the

H20, then over 60 percent of

percent of the Earth is covered by

to the view that such necessity is not a new sort of

H20," is the same as the proposition expressed by "Over 60 percent of the Earth is covered by

necessity-means that materialists are committed

water," and so the

to the

conditional sentence is

a priori

deducibility of the phenomenal

from the physical. If the explanation of the the necessary

proposition expressed by the a priori and necessary.

But consistent with what is required to count as

a posteriori

nature of

a posteriori does not lie in the spe­

cial necessity possessed, where does it lie? Two­

understanding the conditional tingent and

sentence, it is con­ a posteriori that it expresses a neces­

sary a priori proposition.

dimensionalists insist that the issue is an issue

I should emphasize that this does not mean that

about sentences, and not about propositions, or

people who fully understand a sentence like

490

Frank Jackson

"Over 60 percent of the Earth is covered by water" but do not know that water is H20 do not, in some perfectly natural sense, know the con­ ditions under which what they are saying is true.4 True, full understanding of the sentence does not in itself yield which proposition is expressed by the sentence, but knowledge of the way in which the proposition expressed depends on context, combined with knowledge of the truth conditions of the various propositions, does enable them to say when the sentence they produce is true. For their knowledge about how the proposition ex­ pressed depends on context together with the conditions under which the various propositions are true is given in the following array: If H20 fills the water role, then "Over 60 percent of the Earth is covered by water" expresses a proposition that is true iff over 60 percent of the Earth is covered by H20. If XYZ fills the water role, then "Over 60 percent of the Earth is covered by water" expresses a proposition that is true iff over 60 percent of the Earth is covered by XYZ. If---fills the water role, then "Over 60 per­ cent of the Earth is covered by water" expresses a proposition that is true iff over 60 percent of the Earth is covered by . ---

For each distinct, context-giving, antecedent, a distinct proposition is expressed by the sentence. Nevertheless, simple inspection ofthe array shows that the sentence is true iff over 60 percent of the Earth is covered by the stuff that fills the water role. That is the sense in which the fully under­ standing producer of the sentence knows when the sentence is true. 5 Now, to return to the main plot, although un­ derstanding alone does not necessarily give the proposition expressed by certain sentences-that is, how they can be necessary and yet this fact be in principle not accessible to understanding plus acumen alone, that is, how they can be necessary a posteriori-understanding alone does give us the way the proposition expressed depends on

context; and that fact is enough for us to move a priori from, for example, sentences about the distribution of H20 combined with the right context-giving sentences, to information about the distribution of water. Consider, for instance, a supplementation of our earlier inference: 1 . Over 60 percent of the Earth is covered by H20.

2. H20 fills the water role. 3. Therefore, over 60 percent of the Earth is cov­ ered by water. Although, as noted earlier, the passage from (1) to (3) is necessarily truth-preserving but a poste­ riori, being an a posteriori entailment, the passage from (1) and (2), to (3) is a priori. And it is so be­ cause, although our understanding of "Over 60 percent of the Earth is covered by H20" does not in itself yield the proposition expressed by the sentence, it yields how the proposition depends on context, and (2) gives that context. Sentence (2) gives the relevant fact about how things are "outside the head." We did not know that (1) en­ tailed (3) until we learned (2), because we did not, and could not, have known that (1) and (3) ex­ press the same proposition until we learned (2). But .as soon as we learn (2), we have the where­ withal, if we are smart enough, to move a priori to (3). The point, then, is that the necessary a posteri­ ori nature of "Water H20" does not mean that the fact that the H20 way things are entails the water way things are is not answerable to our grasp of the relevant concepts plus acumen. It means, rather, that we need to tell a rich enough story about the H20 way things are, a story that includes the crucial contextual information, be­ fore we can move from the H20 way things are to the water way they are using our grasp of the concepts alone. More generally, the two-dimensional way of looking at the necessary a posteriori means that even ifthe entailment the materialist is committed to from some physical story about the world to =

491

Finding the Mind in the Natural World

the full psychological story is a posteriori, there is still an

a priori story tellable about how the story

in physical terms about our world makes true the story in psychological terms about our world. Although understanding may not, even in princi­ ple, be enough to yield the proposition expressed by the physical story, understanding and logical acumen are enough to yield how the proposition expressed depends on context. But, of course, the context is, according to the materialist, entirely physical . The context concerns various matters about the nature of the actual world, and that nature is capturable in entirely physical terms ac­ cording to the materialist. Hence, the materialist is committed to there being an a priori story to tell about how the physical way things are makes true the psychological way things are. But the story may come in two parts. It may be that one part of the story says which physical way things are, 0 ] , makes some psychological statement true, and the other part of the story, the part that tells the context, says which different physical way things are,

O2, makes it the case that it is 01

that makes

a O2 together make 6 the psychological statement true.

the psychological statement true. What will be

priori accessible

is that 01 and

Notes I.

What follows is one version of a familiar story, see,

e.g., Terence Horgan, "Supervenience and Microphy­ sics,"

Pacific Philosophical Quarterly, 63, 1 982, 29-43

and David Lewis, "New Work for a Theory of Univer­ sals,"

Australasian Journal of Philosophy, 6 1 , 1983,

343-377. 2. I take it that what follows is a sketch of the approach suggested by the version of two-dimensionalism in Robert C. Stalnaker. "Assertion." In P. Cole (Ed.),

Syntax and Semantics, Press,

Vol.

9. New York, Academic

1978, pp. 31 5-332.

3. Stalnaker, op. cit., and David Kaplan, "Dthat." In Syntax and Semantics, Vol. 9, New York, Academic Press, 1978.

P. Cole ( Ed.),

4. I am indebted here to David Lewis and David Chalmers.

5.

This observation bears on the dispute about whether

Earthians and Twin Earthians believe alike. Although the sentence "Water is plentiful" expresses different propositions in the mouths of the Earthians and the Twin Earthians, they agree about when the sentence is true, and so, in

that sense agree in belief.

6. I am indebted to Lloyd Humberstone, David Chalmers, David Lewis, Michael Smith, and Philip Pettit.

30

Breaking the Hold: Silicon Brains, Conscious Robots, and Other Minds

John R. Searle The view ofthe world as completely objective has

deteriorate and the doctors continue to implant

a very powerful hold on us, though it is incon­

more silicon chips. You can see where the thought

sistent with the most obvious facts of our experi­

experiment is going already: in the end, we imag­

ences. As the picture is false, we ought to be able

ine that your brain is entirely replaced by silicon

to break the hold. I don't know any simple way to

chips; that as you shake your head, you can hear

do that. . . . In this chapter I want to describe some

the chips rattling around inside your skull. In

thought experiments that will challenge the accu­

such a situation there would be various possibil­

racy of the picture. Initially the aim of the thought

ities. One logical possibility, not to be excluded

experiments is to challenge the conception of the

on any a priori grounds alone, is surely this:

mental as having some important internal con­

you continue to have all of the sorts of thoughts,

nection to behavior.

experiences, memories, etc. , that you had pre­

To begin undennining the foundations of this

viously; the sequence of your mental life remains

whole way of thinking, I want to consider some

unaffected. In this case, we are imagining that the

of the relationships between consciousness, be­

silicon chips have the power not only to duplicate

havior, and the brain. Most of the discussion will

your input-output functions, but also to duplicate

concern conscious mental phenomena; but leav­

the mental phenomena, conscious and otherwise,

ing out the unconscious at this point is not such a

that are normally responsible for your input­

great limitation, because . . . we have no notion of

output functions.

an unconscious mental state except in terms de­

I hasten to add that I don't for a moment think

rived from conscious states. To begin the argu­

that such a thing is even remotely empirically

ment, I will employ a thought experiment that I

possible. I think it is empirically absurd to sup­

Gedan­

pose that we could duplicate the causal powers of

is something of an old chestnut in

neurons entirely in silicon. But that is an empiri­

philosophy, and I do not know who was the first

cal claim on my part. It is not something that we

have used elsewhere (Searle 1 982). This

kenexperiment

to use it. I have been using it in lectures for years,

could establish a priori. So the thought experi­

and I assume that anybody who thinks about

ment remains valid as a statement of logical or

these topics is bound to have something like these

conceptual possibility.

ideas occur to him or her eventually.

But now let us imagine some variations on the thought experiment. A second possibility, also

Silicon Brains Here is how it goes. Imagine that your brain starts to deteriorate in such a way that you are slowly going blind. Imagine that the desperate doctors, anxious to alleviate your condition, try any method to restore your vision. As a last resort, they try plugging silicon chips into your visual cortex. Imagine that to your amazement and theirs, it turns out that the silicon chips restore your vision to its normal state. Now, imagine further that your brain, depressingly, continues to

not to be excluded on any a priori grounds, is this: as the silicon is progressively implanted into your dwindling brain, you find that the area of your conscious experience is shrinking, but that this shows no effect on your external behavior. You find, to your total amazement, that you are in­ deed losing control of your external behavior. You find, for example, that when the doctors test your vision, you hear them say, "We are holding up a red object in front of you; please tell us what you see." You want to cry out, "I can't see any­ thing. I'm going totally blind. " But you hear your voice saying in a way that is completely out of

494

John Searle

your control, "I see a red object in front of me." If

your mental life is unchanged. So in this case, you

we carry this thought experiment out to the limit,

might hear the doctors saying,

we get a much more depressing result than last time. We imagine that your conscious experience slowly shrinks to nothing, while your externally observable behavior remains the same. It is important in these thought experiments

The silicon chips are able to maintain heartbeat, respi­ ration, and other vital processes, but the patient is ob­ viously brain dead. We might as well unplug the system, because the patient has no mental life at all.

that you should always think of it from the first­

Now in this case, you would know that they are

person point of view. Ask yourself, "What would

totally mistaken. That is, you want to shout out,

it be like for me?" and you will see that it is per­ fectly conceivable for you to imagine that your external behavior remains the same, but that your internal conscious thought processes gradually shrink to zero. From the outside, it seems to ob­

No, I'm still conscious! I perceive everything going on around me. It's just that I can't make any physical movement. I've become totally paralyzed. The point of these three variations on the

servers that you are just fine, but from the inside

thought experiment is to illustrate the

you are gradually dying. In this case, we are

lationships between brain processes, mental pro­

imagining a situation where you are eventually

cesses, and externally observable behavior. In the

mentally dead, where you have no conscious

first case, we imagined that the silicon chips had

mental life whatever, but your externally ob­

causal powers equivalent to the powers of the

servable behavior remains the same. It is also important in this thought experiment

causal re­

brain, and thus we imagined that they caused both the mental states and the behavior that brain

to remember our stipUlation that you are becom­

processes normally cause. In the normal case,

ing unconscious but that your behavior remains

such mental states mediate the relationship be­

unaffected. To those who are puzzled how such a

tween input stimuli and output behavior.

thing is possible, let us simply remind them: As

In the second case, we imagined that the medi­

far as we know, the basis of consciousness is in

ating relationship between the mind and the

certain specific regions of the brain, such as, per­

behavior patterns was broken. In this case, the

haps, the reticular formation. And we may sup­

silicon chips did not duplicate the causal powers

pose in this case that these regions are gradually

of the brain to produce conscious mental states,

deteriorating to the point where there is no con­ sciousness in the system. But we further suppose

they only duplicated certain input-output func­ tions of the brain. The underlying conscious

that the silicon chips are able to duplicate the in­

mental life was left out.

put-output functions of the whole central nervous

In the third case, we imagined a situation where

system, even though there is no consciousness left

the agent had the same mental life as before, but

in the remnants of the system.

in this case, the mental phenomena had no be­

Now consider a third variation. In this case, we

havioral expression. Actually, to imagine this

imagine that the progressive implantation of the

case we need not even have imagined the silicon

silicon chips produces no change in your mental

chips. It would have been very easy to imagine

life, but you are progressively more and more

a person with the motor nerves cut in such a

unable to put your thoughts, feelings and inten­

way that he or she was totally paralyzed, while

tions into action. In this case, we imagine that

consciousness and other mental phenomena re­

your thoughts, feelings, experiences, memories,

mained unaffected . Something like these cases

etc., remain intact, but your observable external

exists in clinical reality. Patients who suffer from

behavior slowly reduces to total paralysis. Even­

the Guillain-Barre syndrome are completely par­

tually you suffer from total paralysis, even though

alyzed, but also fully conscious.

Breaking the Hold

495

What is the philosophical significance of these

the circumstance in which mental phenomena

three thought experiments? It seems to me there is

were present, but the behavior disappeared, so

a number of lessons to be learned. The most im­

behavior is not a necessary condition for the

portant is that they illustrate something about the

presence of the mental either.

relationship between mind and behavior. What

Two other points are illustrated by the thought

exactly is the importance of behavior for the con­

experiments. First, the ontology of the mental is

cept of mind? Ontologically speaking, behavior, functional role, and causal relations are irrelevant to the existence ofconscious mentalphenomemena. Epistemically, we do learn about other people's conscious mental states in part from their be­ havior. Causally, consciousness serves to mediate

essentially a first-person ontology. That is just a fancy way of saying that every mental state has to be

somebody's

mental state. Mental states only

exist as subjective, first-person phenomena. And the other point related to this is that, epistemically speaking, the first-person point of view is quite

the causal relations between input stimuli and

different from the third-person point of view. It is

output behavior; and from an

easy enough to imagine cases, such as those illus­

evolutionary

point

of view, the conscious mind functions causally to

trated by our thought experiments, where from a

control behavior. But

the

third-person point of view, somebody might not

phenomena in question can exist completely and

be able to tell whether I had any mental states at

have all of their essential properties independent

all. He might even think I was unconscious, and

ontologically speaking,

of any behavioral output. Most of the philosophers I have been criticizing would accept the following two propositions:

I . Brains cause conscious mental phenomena.

it might still be the

case

that I was completely

conscious. From the first-person point of view, there is no question that I am conscious, even if it turned out that third-person tests were not available.

2. There is some sort of conceptual or logical connection between conscious mental phenom­ ena and external behavior.

Conscious Robots

But what the thought experiments illustrate is

I want to introduce a second thought experiment

that these two cannot be held consistently with a

to buttress the conclusions provided by the first.

third:

The aim of this one, as with the first, is to use our

3.

The capacity of the brain to cause conscious­

ness is conceptually distinct from its capacity to cause motor behavior. A system could have con­ sciousness without behavior and behavior with­ out consciousness.

thought experiments is what we might call "the principle of the independence of consciousness and behavior." In case number two, we imagined the circumstance in which the behavior was un­ affected, but the mental states disappeared, so behavior is not a sufficient condition for mental case

ing robots to work on a production line. Imagine that our robots are really too crude and tend to make a mess of the more refined elements of their task. But imagine that we know enough about the

But given the truth of I and 3, we have to give up 2. So the first point to be derived from our

phenomena. In

intuitions to try to drive a wedge between mental states and behavior. Imagine that we are design­

number three, we imagined

electrochemical features of human consciousness to know how to produce robots that have a rather low level of consciousness, and so we can design and manufacture conscious robots. Imagine fur­ ther that these conscious robots are able to make discriminations that unconscious robots could not make, and so they do a better job on the production line. Is there anything incoherent in the above? I have to say that according to my

John Searle

"intuitions," it is perfectly coherent. Of course, it is science fiction, but then, many of the most im­ portant thought experiments in philosophy and science are precisely science fiction. But now imagine an unfortunate further fea­ ture of our conscious robots: Suppose that they are absolutely miserable. Again, we can suppose that our neurophysiology is sufficient for us to establish that they are extremely unhappy. Now imagine we give our robotics research group the following task: Design a robot that will have the capacity to make the same discriminations as the conscious robots, but which will be totally unconscious. We can then allow the unhappy robots to retire to a more hedonically satisfying old age. This seems to me a well-defined research project; and we may suppose that, operationally speaking, our scientists try to design a robot with a "hardware" that they know will not cause or sustain consciousness, but that will have the same input-output functions as the robot that has a "hardware" that does cause and sustain consciousness. We might suppose then that they succeed, that they build a robot that is totally unconscious, but that has behavioral powers and abilities that are absolutely identical with those of the conscious robot. The point of this experiment, as with the earlier ones, is to show that as far as the ontology of consciousness is concerned, behavior is simply ir­ relevant. We could have identical behavior in two different systems, one of which is conscious and the other totally unconscious.

Empiricism and the "Other Minds Problem" Many empirically minded philosophers will be distressed by these two thought experiments, es­ pecially the first. It will seem to them that I am alleging the existence of empirical facts about the mental states of a system that are not ascertain­ able by any empirical means. Their conception of the empirical means for ascertaining the existence of mental facts rests entirely on the presupposi-

496

tion of behavioral evidence. They believe that the only evidence we have for attributing mental states to other systems is the behavior of those systems. In this section . . . part of my aim will be to show that there is nothing incoherent or ob­ jectionable in the epistemic implications of the two thought experiments I just described, but my primary aim will be to give an account of the "empirical" basis we have for supposing that other people and higher animals have conscious mental phenomena more or less like our own. It is worth emphasizing at the beginning of the discussion that in the history of empirical philos­ ophy and of the philosophy of mind, there is a systematic ambiguity in the use of the word "empirical", an ambiguity between an onto­ logical sense and an epistemic sense. When people speak of empirical facts, they sometimes mean actual, contingent facts in the world as opposed to, say, facts of mathematics or facts of logic. But sometimes when people speak of empirical facts, they mean facts that are testable by third-person means, that is, by "empirical facts" and "em­ pirical methods," they mean facts and methods that are accessible to all competent observers. Now this systematic ambiguity in the use of the word "empirical" suggests something that is cer­ tainly false: that all empirical facts, in the onto­ logical sense of being facts in the world, are equally accessible epistemically to all competent observers. We know independently that this is false. There are lots of empirical facts that are not equally accessible to all observers. The previous sections gave us some thought experiments de­ signed to show this, but we actually have empiri­ cal data that suggest exactly the same result. Consider the following example. 1 We can with some difficulty imagine what it would be like to be a bird flying. I say "with some difficulty" because, of course, the temptation is always to imagine what it would be like for us if we were fiying, and not, strictly speaking, what it is like for a bird to be flying. But now some recent research tells us that there are some birds that navigate by detect-

Breaking the Hold

ing the earth's magnetic field. Let us suppose that just as the bird has a conscious experience of flapping its wings or feeling the wind pressing against its head and body, so it also has a con­ scious experience of a feeling of magnetism surg­ ing through its body. Now, what is it like to feel a surge of magnetism? In this case, I do not have the faintest idea what it feels like for a bird, or for that matter, for a human to feel a surge of magnetism from the earth's magnetic field. It is, I take it, an empirical fact whether or not birds that navigate by detecting the magnetic field actually have a conscious experience of the detection of the mag­ netic field. But the exact qualitative character of this empirical fact is not accessible to standard forms of empirical tests. And indeed, why should it be? Why should we assume that all the facts in the world are equally accessible to standard, ob­ jective, third-person tests? If you think about it, the assumption is obviously false. I said that this result is not as depressing as it might seem. And the reason is simple. Although in some cases we do not have equal access to cer­ tain empirical facts because of their intrinsic sub­ jectivity, in general we have indirect methods of getting at the same empirical facts. Consider the following example. I am completely convinced that my dog, as well as other higher animals, has conscious mental states, such as visual experi­ ences, feelings of pain, and sensations of thirst and hunger, and of cold and heat. Now why am I so convinced of that? The standard answer is be­ cause of the dog's behavior, beca\lse by observing his behavior I infer that he has mental states like my own. I think this answer is mistatken. It isn't just because the dog behaves in a way that is ap­ propriate to having conscious mental states, but also because I can see that the causal basis of the behavior in the dog's physiology is relevantly like my own. It isn't just that the dog has a structure like my own and that he has behavior that is in­ terpretable in ways analogous to the way that I interpret my own. But rather, it is in the combi­ nation of these two facts that I can see that the behavior is appropriate and that it has the ap-

497

propriate causation in the underlying physiology. I can see, for example, that these are the dog's ears; this is his skin; these are his eyes; that if you pinch his skin, you get behavior appropriate to pinching skin; if you shout in his ear, you get be­ havior appropriate to shouting in ears. It is important to emphasize that I don't need to have a fancy or sophisticated anatomical and physiological theory of dog structure, but simple, so to speak, "folk" anatomy and physiology­ the ability to recognize the structure of skin, eyes, teeth, hair, nose, etc., and the ability to suppose that the causal role that these play in his experi­ ences is relevantly like the causal role that such features play in one's own experiences. Indeed, even describing certain structures as "eyes" or "ears" already implies that we are attributing to them functions and causal powers similar to our own eyes and ears. In short, though I don't have direct access to the dog's consciousness, nonethe­ less it seems to me a well-attested empirical fact that dogs are conscious, and it is attested by evi­ dence that is quite compelling. I do not have anything like this degree of confidence when it comes to animals much lower on the phylogenetic scale. I have no idea whether fleas, grasshoppers, crabs, or snails are conscious. It �ems to me that I can reasonably leave such questions to neuro­ physiologists. But what sort of evidence would the neurophysiologist look for? Here, it seems to me, is another thought experiment that we might well imagine. Suppose that we had an account of the neuro­ physiological basis of consciousness in human beings. Suppose that we had quite precise, neuro­ physiologically isolable causes of consciousness in human beings, such that the presence of the relevant neurophysiological phenomena was both necessary and sufficient for consciouness. If you had it, you were conscious; if you lost it, you be­ came unconscious. Now imagine that some ani­ mals have this phenomenon, call it "x" for short, and others lack it. Suppose that x was found to occur in all those animals, such as ourselves, monkeys, dogs, and so on, of which we feel quite

John Searle

498

confident that they are conscious on the basis of their gross physiology, and that

with us, but they are in certain important respects

was totally ab­

similar. They have eyes, ears, nose, mouth, and so

sent from animals, such as amoebas, to which

forth. For this reason we do not really doubt that

x

we do not feel inclined to ascribe any conscious­

they have the experiences that go with these vari­

ness. Suppose further that the removal of x from

ous sorts of apparatus. So far, all these consid­

any human being's neurophysiology immediately

erations are prescientific. But let us suppose that

produced unconsciousness, and its reintroduction

we could identify for the human cases exact

produced consciousness. In such a case, it seems

causes of consciousness, and then could discover

to me we might reasonably assume that the pres­

precisely the same causes in other animals. If so,

ence of x played a crucial causal role in the pro­

it seems to me we would have established quite

duction of consciousness, and this discovery

conclusively that other species have exactly the

would enable us to settle doubtful cases of ani­

same sort of consciousness that we have, because

mals either having or lacking conscious states. If

we can presume that the same causes produce the

and mites lacked it, then we might

same effects. This would not be just a wild spec­

reasonably infer that mites were operating on

ulation, because we would have very good reason

simple tropisms and snakes had consciousness in

to suppose that those causes would produce the

the same sense that we, dogs, and baboons do.

same effects in other species.

snakes had

x,

I don't for a moment suppose that the neuro­

In actual practice, neurophysiology textbooks

physiology of consciousness will be as simple

routinely report, for example, how the cat's per­

as this. It seems to me much more likely that we

ception of color is similar to and different from

will find a great variety of forms of neurophysio­

the human's and even other animals. What breath­

logies of consciousness, and that in any real ex­

taking irresponsibility! How could the authors

perimental situation we would seek independent

pretend to have solved the other cat's mind prob­

evidence for the existence of mechanical-like

lem so easily? The answer is that the problem is

tropisms to account for apparently goal-directed

solved for cats' vision once we know exactly how

behavior in organisms that lacked consciousness.

the cat's visual apparatus is similar to and differ­

The point of the example is simply to show that

ent from our own and other species'. 2

we can have indirect means of an objective, third­

Once we understand the causal basis of the

person, empirical kind for getting at empirical

ascription of mental states to other animals, then

phenomena that are intrinsically subjective and

several

therefore inaccessible to direct third-person tests.

"other minds" have an easy solution. Consider

traditional

skeptical

problems

about

It shouldn't be thought, however, that there is

the famous problem of spectrum inversion . . . . It

something second rate or imperfect about the

is often said that, for all we know, one section of

third-person empirical methods for discovering

the population might have a red/green inversion

these first-person subjective empirical facts. The

such that though they make the same behavioral

methods rest on a rough-and-ready principle that

discriminations as the rest of us, the actual expe­

we use elsewhere in science and in daily life:

riences they have when they see green, and which

causes-same effects, and effects. We can readily see

they call "seeing green", are experiences that we

same similar causes-similar in the case of other

would, if we had them, call "seeing red, " and vice

human beings that the causal bases of their expe­

versa. But now consider: Suppose we actually

riences are virtually identical with the causal

found that a section of the popUlation actually

bases of our experiences. This is why in real life

did have the red and green receptors reversed in

there is no "problem of other minds. " Animals

such a way, and so connected with the rest of their

provide a good test case for this principle because,

visual

of course, they are not physiologically identical

neurophysiological evidenced that though their

apparatus,

that we had overwhelming

Breaking the Hold

molar discriminations were the same as ours, they actually had different experiences underlying them. This would not be a problem in philosoph­ ical skepticism, but a well-defined neurophysio­ logical hypothesis. But then if there is no such section of the population, if all of the non-color­ blind people have the same red/green perceptual pathways, we have solid empirical evidence that things look to other people the way they look to us. A cloud of philosophical skepticism condenses into a drop of neurocscience. Notice that this solution to "the other minds problem," one that we use in science and in daily life, gives us sufficient but not necessary con­ ditions for the correct ascription of mental phe­ nomena to other beings. We WOUld, as I suggested earlier in this chapter, need a much richer neuro­ biological theory of consciousness than anything we can now imagine to suppose that we could isolate necessary conditions of consciousness. I am quite confident that the table in front of me, the computer I use daily, the fountain pen I write with, and the tape-recorder I dictate into are quite unconscious, but, of course, I cannot prove that they are unconscious and neither can anyone else.

Summary In this chapter I have so far had two objectives: First, I have tried to argue that as far as the on­ tology of the mind is concerned, behavior is sim­ ply irrelevant. Of course in real life our behavior is crucial to our very existence, but when we are examining the existence of our mental states as mental states, the correlated behavior is neither necessary nor sufficient for their existence. Sec­ ond, I have tried to begin to break the hold of three hundred years of epistemological discus­ sions of "the other minds problem," according to which behavior is the sole basis on which we know of the existence of other minds. This seems to me obviously false. It is only because of the connection between behavior and the causal structure of other organisms that behavior is at

499

all relevant to the discovery of mental states in others. A final point is equally important: except when doing philosophy, there really is no "problem" about other minds, because we do not hold a "hypothesis," "belief," or "supposition" that other people are conscious, and that chairs, tables, computers, and cars are not conscious. Rather, we have certain Background ways of be­ having, certain Background capacities, and these are constitutive of our relations to the conscious­ ness of other people. It is typical of philosophy that skeptical problems often arise when elements of the Background are treated as if they were hypotheses that have to be justified. I don't hold a "hypothesis" that my dog or my department chairman is conscious, and consequently the question doesn't arise except in philosophical debate.

Intrinsic, As-H, and Derived Intentionality Before proceeding further, I need to introduce some simple distinctions that have been implicit in what I have said so far, but will need to be made explicit for what follows. To introduce these distinctions, let us consider the similarities and differences among the various sorts of truth­ conditions of sentences that we use to ascribe intentional mental phenomena. Consider the sim­ ilarities and differences among the following:

1 . I am now thirsty, really thirsty, because haven't had anything to drink all day. 2. My lawn is thirsty, really thirsty, because it has not been watered in a week.

3 . In French, "j'ai grand soif" means "I am very thirsty." The first of these sentences is used literally to ascribe a real, intentional mental state to oneself. If I utter that sentence, making a true statement, then there is in me a conscious feeling of thirst that makes that statement true. That feeling

John Searle

has intentionality because it involves a desire to drink. But the second sentence is quite different. Sentence 2 is used only metaphorically, or figu­ ratively, to ascribe thirst to my lawn. My lawn, lacking water, is in a situation in which I would be thirsty, so I figuratively describe it as if it were thirsty. I can, by analogy, quite harmlessly say that the lawn is thirsty even though I do not sup­ pose for a moment that it is literally thirsty. The third sentence is like the first in that it literally ascribes intentionality, but it is like the second and unlike the first in that the intentionality de­ scribed is not intrinsic to the system. The first sort of ascription ascribes intrinsic in­ tentionality. Ifsuch a statement is true, there must really be an intentional state in the object of the ascription. The second sentence does not ascribe any intentionality at all, intrinsic or otherwise; it is merely used to speak figuratively or meta­ phorically. Therefore, I will say that the "in­ tentionality" in the ascription is merely as-if, and not intrinsic. To avoid confusion, it is important to emphasize that as-ifintentionality is not a kind of intentionality, rather a system that has as-if intentionality is as-if-it-had-intentionality. In the third case I literally ascribe intentionality to the French sentence, that is, the French sentence lit­ erally means what I say it does. But the inten­ tionality in the French sentence is not intrinsic to that particular sentence construed just as a syn­ tactical object. That very sequence might have meant something very different or nothing at all. Speakers of French can use it to express their in­ tentionality. Linguistic meaning is a real form of intentionality, but it is not intrinsic intentionality. It is derived from the intrinsic intentionality of the users of the language. We can summarize these points as follows: intrinsic intentionality is a phenomenon that humans and certain other animals have as part of their biological nature. It is not a matter of how they are used or how they think of themselves or how they choose to describe themselves. It is just a plain fact about such beasts that, for example, sometimes they get thirsty or hungry, they see

500

things, fear things, etc. All of the italicized ex­ pressions in the previous sentence are used to refer to intrinsic intentional state. It is very con­ venient to use the jargon of intentionality for talking about systems that do not have it, but that behave they did. I say about my thermostat that it perceives changes in the temperature; I say of my carburetor that it knows when to enrich the mix­ ture; and I say of my computer that its memory is bigger than the memory ofthe computer I had last year. All of these attributions are perfectly harm­ less and no doubt they will eventually produce new literal meanings as the metaphors become dead. But it is important to emphasize that these attributions are psychologically irrelevant, be­ cause they do not imply the presence of any men­ tal phenomena. The intentionality described in all of these cases is purely as-if Cases of the third sort are rendered interesting by the fact that that we often do literally endow nonmental phenomena with intentional proper­ ties. There is nothing metaphorical or as-if about saying that certain sentences mean certain things or certain maps are correct representations of the state of California or that certain pictures are pictures of Winston Churchill. These forms of in­ tentionality are real, but they are derived from the intentionality of human agents. I have been using the tenninology of "in­ trinsic" for over a decade (see Searle 1980), but it is subject to certain persistent misunderstandings. In common speech "intrinsic" is often opposed to "relational." Thus the moon intrinsically has a mass, but is not intrinsically a satellite. It is only a satellite relative to the earth. In this sense of in­ trinsic, people who believe inrintentional states with "wide content," that is content essentially involving relations to objects outside the mind, would be forced to deny that such intentional states are intrinsic, because they are relational. I don't believe in the existence of wide content (see Searle 1 983, ch. 7), so the problem does not arise for me. The distinctions I am making now are in­ dependent of the dispute about wide and narrow content. So I am just stipulating that by "intrinsic

Breaking the Hold

intentionality" I mean the real thing as opposed to the mere appearance of the thing (as-if), and as opposed to derived forms of intentionality such sentences, pictures, etc. You do not have to ac­ cept my objections to wide content to accept the distinctions I am trying to make. Another-amazing to me--misunderstanding is to suppose that by calling cases of the real thing "intrinsic" I am implying that they are somehow mysterious, ineffable, and beyond the reach of philosophical explanation or scientific study. But this is nonsensc. I have right now many intrinsic intentional states, for example, and urge to go to the bathroom, a strong desire for a cold beer, and a visual experience of a lot of boats on the lake. All of these are intrinsic intentional states, in my sense, which just means they are the real thing and not just something more or less like the real thing (as-if), or something that is the result of somebody else's uses of or attitudes toward the thing (derived). 3 I have seen efforts to deny these distinctions, but it is very hard to take the denials seriously. If you think that there are no principled differences, you might consider the following from the journal

Pharmacology. Once the food is past the chrico-pharyngus sphincter, its movement is almost entirely involuntary except for the final expUlsion of feces during defecation. The gastro­ intestinal traet is a highly intelligent organ that senses not only the presence of food in the lumen but also its chemical composition, quantity, viscosity and adjusts the rate of propulsion and mixing by producing appro­ priate patterns of contractions. Due to its highly devel­ oped decision making ability the gut wall comprised of the smooth muscle layers, the neuronal structures and paracrine-endocrine cells is often called rhe gur brain. (Sarna and Otterson 1988, my italics)4

This is clearly a case of as-if intentionality in the "gut brain." Does anyone think there is no principled difference between the gut brain and the brain brain? I have heard it said that both sorts of cases are the same; that it is all a matter of taking an "intentional stance" toward a system.

501

But just try in real life to suppose that the "per­ ception" and the "decision making" of the gut brain are no different from that of the real brain. This example reveals, among other things, that any attempt to deny the distinction between in­ trinsic and as-if intentionality faces a general re­ ductio ad absurdum. If you deny the distinction, it turns out that everything in the universe has in­ tentionality. Everything in the universe follows laws of nature, and for that reason everything behaves with a certain degree of regularity, and for that reason everything behaves as if it were following a rule, trying to carry out a certain project, acting in accordance with certain desires, etc. For example, suppose I drop a stone. The stone tries to reach the center of the earth, be­ cause it wants to reach the center of the earth, and in so doing it i t /ollows the rule S = 1/2 gt2• The price of denying the distinction between intrinsic and as-ifintentionality, in short, is absurdity, be­ cause it makes everything in the universe mental. No doubt there are marginal cases. About grasshoppers or fleas, for example, we may not be quite sure what to say. And no doubt, even in some human cases we might be puzzled as to whether we should take the ascription of inten­ tionality literally or metaphorically. But marginal cases do not alter the distinction between the sort of facts corresponding to ascriptions of intrinsic intentionality and those corresponding to as-if metaphorical ascriptions of intentionality. There is nothing harmful, misleading, or philosoph­ ically mistaken about as-if metaphorical ascrip­ tions. The only mistake is to take them literally. I hope the distinctions I have been making are painfully obvious. However, I have to report, from the battlefronts as it were, that the neglect of these simple distinctions underlies some of the biggest mistakes in contemporary intellectual life. A common pattern of mistake is to suppose that because we can make as-if ascriptions of in­ tentionality to systems that have no intrinsic intentionality, that somehow or other we have discovered the nature of intentionality. 5

John Searle

Notes I . In the style of Thomas Nagel's article, "What Is It Like to Be a Bat?"(1974). 2. For example, "As one might expect, cells whose receptive fields are specifically color-coded have been noted in various animals, including the monkey, the ground squirrel, and some fishes. These animals, in con­ tradistinction to the cat, possess excellent color vision and an intricate neural mechanism for processing color" (Kulfter and Nicholls 1976, p. 25, my italics). 3. For an example of this misunderstanding, see P. M. and P. S. Churchland (1983). 4. I am indebted to Dan Rubermann for calling my at­ tention to this article. 5. See, for example, Dennett (1987).

References Churchland, P. M., and Churchland, P. S. (1983). "Stalking the Wild Epistemic Engine." Nous 17:5-18. Dennett, D. C. (1987). Intentional Stance. Cambridge, MA: MIT Press.

Kulfter, S. W., and Nicholls, 1. G. (1976). From Neuron to Brain. Sunderland, MA: Sinauer Associates. Nagel, T. (1 974.) "What Is It Like to Be a Bat?" Philo­ sophical Review 4, 83:435-450. Sarna, S. K., and Otterson, M. F. (1 988). "Gastro­ intestinal Motility: Some Basic Concepts." Phannacol­ ogy: Supplement 36:7-14. Searle, 1. R . (1 980). "Intrinsic Intentionality: Reply to Criticisms of Minds, Brains, and Programs." Behavioral and Brain Sciences 3:450-456. . (1982). "The Chinese Room Revisited: Re­ sponse to Further Commentaries on 'Minds, Brains, and Programs." Behavioral and Brain Sciences 5, 2: 345-348.

--

. (1983). Intentionality: An Essay in the Philoso­ phy of Mind. Cambridge: Cambridge University Press. ---

502

31

The First-Person Perspective

Sydney Shoemaker Some would say that the philosophy of mind without the first-person perspective, or the first­ person point of view, is like Hamlet without the Prince of Denmark. Others would say that it is like Hamlet without the King of Denmark, or like Othello without lago. I say both. I think of myself as a friend of the first-person perspective. Some would say that I am too friendly to it, for I hold views about first-person access and first-person authority that many would regard as unaccept­ ably "Cartesian." I certainly think that it is es­ sential to a philosophical understanding of the mental that we appreciate that there is a first­ person perspective on it, a distinctive way mental states present themselves to the subjects whose states they are, and that an essential part of the philosophical task is to give an account of mind which makes intelligible the perspective mental subjects have on their own mental lives. And I do not think, as I think some do, that the right theory about all this will be primarily an "error theory." But I also think that the first-person perspective is sometimes rightly cast as the villain in the piece. It is not only the denigrators of introspection that assign it this role. Kant did so in the Paralogisms, seeing our vantage on our selves as the source of transcendental illusions about the substantiality of the self. And Wittgenstein's "private language argument" can be seen as another attempt to show how the first-person perspective can mislead us about the nature of mind. My concern here is with the role of the first­ person perspective in the distinctively philosoph­ ical activity of conducting thought experiments designed to test metaphysical and conceptual claims about the mind. In conducting such a thought experiment one envisages a putatively possible situation and inquires whether it really is possible and, if so, what its possibility shows about the nature of mind or the nature of mental concepts. Such envisaging can be done either from the "third-person point of view" or the "first-person point of view." In the one case, one

imagines seeing someone doing, saying, and un­ dergoing certain things, and one asks whether this would be a case of something which has been thought to be philosophically problematic-for example, someone's having an unconscious pain. In the other case, one imagines being oneself the subject of certain mental states-· imagines feel­ ing, thinking, and so on, certain things-in a case in which certain other things are true, for exam­ ple, one's body is in a certain condition, and asks what this shows about some philosophical claim about the relation of mind to body. The question I want to pursue is whether there is anything that can be established by such first-person envisag­ ings that cannot be revealed just as effectively by third-person envisagings. It is not difficult to see why first-person thought experiments have often been thought to be more revealing than third-person thought experiments. In a broad range of cases, first-person ascriptions of mental states are not grounded on evidence of any sort. It is natural to move from this to the claim that they are grounded on the mental states themselves, or on "direct acquaintance" with the mental states themselves. One can, apparently, have this knowledge without presuming anything about the connections between the mental states and the bodily states of affairs, behavioral or physiological, which serve as the evidence for our ascription of these same mental states to other persons. Thus the first-person perspective appa­ rently gives one a freer rein than the third-person perspective in investigating, empirically, the con­ nections between mental states of affairs and bodily ones. And so if we are concerned with what the possibilities are, with respect to these connections, imagining what we could discover from the first-person perspective seems poten­ tially more revealing than imagining what we could discover from the third-person perspective. There are a number of areas in the philoso­ phy of mind in which first-person imaginings have played an important role in philosophical

Sydney Shoemaker

reflection. These include the issue of whether the identity over time of a person involves the identity of a body or brain, the issue of whether disem­ bodied existence of persons is a possibility, the issue of whether "spectrum inversion" is a possi­ bility, and the issue of whether mental states can be identical with physical states of bodies or with functional states realized in physical states of bodies. In many of these cases I agree with the possibility claims that first-person imaginings have been used to support. I think that personal identity does not require bodily identity or brain identity, and I think that spectrum inversion is a possibility. But I think that in these cases the possible states of affairs in question are ones that could be known to obtain from the third-person perspective. In any event, it is not these cases I shaH be discussing here. My focus in the re­ mainder of this talk will be on the bearing of first­ person and third-person imaginings on physicalist views of mind, in particular the identity theory and functionalism. My thoughts about this were partly inspired, or perhaps I should say provoked, by a recent argu­ ment of John Searle's, and my discussion will be in large part about that. But I will lead up to this by considering briefly what is perhaps the best known piece of philosophical imagining in recent times, namely Saul Kripke's assault, over twenty years ago, on the psychophysical identity theory. 1 Kripke claimed that for any given brain state that is a candidate for being identical with pain, one can imagine both being in pain without one's brain being in that state, and also not being in pain when one's brain is in that state. What was novel about Kripke's argument was of course not the claim that these states of affairs are imagi­ nable, or the claim that they are possible, but certain other claims that licensed the inference from these imaginability and possibility claims to the conclusion that pain cannot be identical with any such brain state-most importantly, the claim that "pain" is a rigid designator, and the point that identity judgments involving rigid des­ ignators are necessarily true if true at all. But it

504

did seem central to his case, as he presented it, that the imagining was from the first-person point of view. When he speaks of the "epistemic situa­ tion" vis-a-vis pain, he is plainly speaking of the epistemic situation of the putative subject of pain. The claim that it is possible that there should be pain without C-fiber stimulation, or C-fiber stim� ulation without pain, seems to be grounded on the claim that one can imagine being in pain without there being any C-fiber stimulation oc­ curring in one, and can imagine not being in pain when there is C-fiber stimulation occurring in one. Kripke was of course well aware that the in­ ference from imaginability to possibility could be chaIlenged, and he had very interesting things to say about this; but these do not bear on the issue of first-person versus third-person imagining that is our concern here. Around the time Kripke presented this argu­ ment, other philosophers, most notably Hilary Putnam, were challenging the psychophysical identity theory in a way that also depended on the claim that for any given sort of brain state there could be pain in the absence of that brain state. 2 The arguments of these philosophers were in sup­ port of the view that pain is a functional state that is "multiply realizable." As originally presented, these arguments involved claims of nomological possibility rather than claims of metaphysical or logical possibility, and appealed to actual phys­ iological differences between different species, for example, humans and mollusks, rather than to imaginings of purely hypothetical situations. But it is easy enough to convert them into arguments from imaginings that differ from Kripke's only in that the imaginings are from the third-person rather than the first-person perspective. One imagines finding creatures that manifestly experi­ ence pain, as is shown by their behavior and cir­ cumstances, but lack whatever brain state is the candidate for being identical with pain. One goes from there to the claim that it is possible for there to be pain unaccompanied by that brain state, and uses Kripke's claims about the necessity of identity and the rigidity of the concept of pain to

The First-Person Perspective

argue from this that pain cannot be identical with that brain state. There is certainly a difference in spirit between Kripke's argument and its third-person counter­ part. The arguments appeal to different groups of philosophers, and annoy different groups of philosophers. But it is far from obvious, to say the least, that the first-person imaginings carry any more evidential weight, vis-a-vis the issue of psychophysical identity, than the third-person imaginings. 3 But now recall that there were two parts to Kripke's claim. For any brain state that is a can­ didate for being pain, one can imagine being in pain without being in that brain state, and one can imagine being in that brain state without being in pain. If the second half of the claim can be made out, and if the inference from imaginability to possibility is accepted, we will have more than an argument against the identity theory-we will have an argument against the view that pain can be realized in, or implemented by, brain states, and against the view that pain supervenes on states of the brain. For the latter views, while al­ lowing that pain, being "multiply realizable," can occur without any given brain state occurring, will hold that there are brain states, perhaps a large number of them, each of which is such that its occurrence is necessarily sufficient, although not necessary, for the occurrence of pain. If every brain state is such that one can imagine it occur­ ring without pain occurring, and if imaginability here implies possibility, then all such views topple. Can the imagining that leads to this result be done from the third-person point of view? Well, it's easy enough to imagine a case in which C­ fiber stimulation is going on in someone and that person is not in pain. But C-fiber stimulation never was a very good candidate for being pain. The question should be whether it is true of each and every brain state that it can be imagined to occur without the subject being in pain. This will have to include brain states that are good candi­ dates for being pain. What will make a brain

505

state a good candidate? Well, the ideal candi­ date would be one that satisfies some description which we have reason to think only pain satisfie s. If the first part of the Kripkean argument is suc­ cessful, and it is established that for any given brain state it is possible for pain to occur without that brain state occurring, then no brain state is an ideal candidate in this sense. No brain state is an ideal candidate for being identical with pain­ and here nothing less than an ideal candidate will do. But it is compatible with this that there are brain states that are good candidates for "being pain" in the sense of being realizations of pain. And presumably what this requires is that they play the causal role of pain-that they make the contribution to causing other things, including other mental states as well as behavior, that we believe pain to make, and are caused by the things that we take pain to be caused by. There are of course ways and ways in which causal or functional roles can be described. Some ways make explicit reference to particular mental states- for example, part of the causal role of pain is causing the belief that one is in pain. A state having a causal role thus described cannot of course belong to something devoid of mental states. But I will assume here that the roles are described in "topic neutral" terms; this will per­ mit us to consider the idea, rejected by function­ alists but affirmed by philosophers such as John Searle, that for any causal or functional role there could be something that has a state playing that role without having any mental states at all. Even if C-fiber stimulation did play the causal role of pain, it might do so only contingently. That is, it might do so in virtue of the fact that the brain is "wired" in a certain way, a way in which it could fail to be wired and still have C-fiber stimulation occur in it. It would then be only contingently an optimal candidate for being a re­ alization of pain. In that case we could imagine discovering from the third-person perspective someone who was not in pain but in whom C­ fiber stimulation was occurring-this would be a case in which the brain was wired up differently.

506

Sydney Shoemaker

But now consider the state, call it "C-fiber-stim­ u1ation-plus," which consists in the brain's having C-fiber stimulation occuning in it

and its

being

normal effect; and then we can no longer believe on the basis of the behavior that the person is not in pain.

wired in such a way that C-fiber stimulation plays

So it seems that nothing we can imagine from

the causal role of pain (or what we believe to be

the third-person perspective would entitle us to

the causal role of pain). Let's say here that, on the

say that someone is not in pain despite instan­

supposition we are making, C-fiber stimulation is

tiating what we acknowledge to be an optimal

core realization

candidate for being a total realization of pain.

of pain, and C-fiber-stimulation-plus is an opti­

And now we may seem to have a case in which a

an optimal candidate for being a mal candidate for being a

total

realization of

first-person imagining can achieve something no

pain. One can easily enough imagine from the

third-person imagining can achieve. For can't I

third-person perspective a case in which someone

imagine feeling no pain and yet finding, with the

is not in pain despite having in his brain what is

help of an autocerebroscope, that I am in state C­

an optimal candidate for being a core-realization

fiber-stimulation-plus, a state I know to be an

of pain. But can one imagine from the third-per­

optimal candidate for being a realization of pain?

in

One physicalist response to this would be to say

pain despite having in his brain an optimal can­

that faced with such a case I ought to conclude

didate for being a total realization of pain?

that while it

son perspective a case in which someone is not

I think that the answer is no. The reason is that

seems to me

that I do not feel pain,

the evidence of the autocerebroscope should per­

I do. This is not my re­

playing the causal role of pain or at any rate

suade me that after all

playing what we have good reason to think is the

sponse, and I count it as not sufficiently respecting

causal role of pain will essentially involve pro­

the first-person perspective. While I am willing to

ducing precisely the kinds of behavior that serve

allow that there are circumstances in which a sin­

as our-third-person basis for ascribing pain. We

cere self-ascription of pain can be mistaken,

cannot be in a position to judge about someone

not wiJ1ing to allow that someone might be in ex­

I am

both that she is not in pain and that she is in a state

cruciating pain and yet that it might seem to him,

that influences her behavior in just the ways we

when he reflects in a calm and unflustered way on

think pain influences behavior. Of course, some­

his state, that he feels no pain at all, and that it

one can be in pain when there is no behavioral

might continue to seem that way to him through­

evidence that she is, and when there is behav­

out the extended period during which the ex­

ioral evidence that she is not-she may be suc­

cruciating pain is supposed to last. And yet that is

cessfully suppressing the manifestations of pain.

what would have to be possible if the seeming

And in such a case we will normally believe, mis­

evidence of the autocerebroscope were oveniding

takenly, that the person is not in pain. We might

in such a case.

in such a case know that the person's brain is

But we need to appreciate how bizarre our lat­

in the state C-fiber-stimulation-plus-and if we

est version of Kripke's example is. You are to

don't realize that this is an optimal candidate for

imagine feeling no pain while having very good

being a total realization of pain, we may continue

evidence that you are in a state that plays the

to believe that the person is not in pain while be­

causal role of pain. If you have such a state, you

lieving that she has in her brain what is in fact an

ought to be behaving, or disposed to behave, like

optimal candidate for being a total realization of

someone who is in pain. Suppose you are. Then

pain. But if we realize that it is an optimal candi­

you should reply affirmatively if asked whether

date, we will have to believe that something, such

you are in pain. Suppose you do. How does this

as an effort to suppress tendencies to manifest

seem to you, from the inside? Does it seem like

pain behavior, is preventing it from having its

your own action, something you are intentionally

The First-Person Perspective

doing? Will it seem to you that you are lying? But if you can imagine what we have already envis­ aged, surely you can also imagine that in addition to feeling no pain you have your normal desire to tell the truth. So is it instead that you try to say that you feel no pain, but hear coming from your mouth an avowal of excruciating pain? But if you are alienated from your verbal behavior in this way, presumably you will be similarly alienated from other kinds of behavior as well. You will have no intention of taking aspirin; but you will see your hand reaching for the medicine cabinet and removing the aspirin bottle. You will have no intention of seeing the doctor; but you will see your hand reaching for the telephone, and hear your voice making an appointment. And so on. Before I discuss the implications of this feature of the case, I want to switch to a somewhat dif­ ferent example -�a first-person thought experi­ ment that is presented by John Searle in The Rediscovery ofMind.4 1 think that Searle's though experiment can usefully be viewed as a version of Kripke's, or rather, an elaboration of the ex­ tended version of the second half of Kripke's that we have just been considering. Searle is, of course, an outspoken and eloquent advocate of the first­ person point of view. Searle's example is a variation on the familiar one in which, in a series of operations, the parts of someone's brain are progressively replaced by silicon chips, until eventually the brain is entirely composed of silicon. The replacements are always such as to preserve the behavioral dispositions of the person, and the functional organization needed to sustain these. Searle's variation on the example is to invite the reader to imagine being the subject of this procedure, and to imagine the results from the inside. In his presentation of the case the procedure starts as a treatment for blindness due to deterioration of the brain, and is successful as long as the replacements are limited to the visual cortex. For my own expository pur­ poses, I prefer to have the subject be someone who has bravely volunteered to be the subject ofa

507

philosophical experiment-one designed to test the hypothesis that a creature with a brain having a certain functional organization, one that un­ derwrites behavioral dispositions that enable it to pass the most stringent Turing test, will be con­ scious, no matter what the material composition of that brain. One possible outcome of the experiment is that one finds that each successive replacement of gray stuff with silicon makes no difference to one's conscious life: "you continue to have all of the sorts of thoughts, experiences, memories, etc., that you had previously; the sequence of your mental life remains unaffected" (p. 66). Searle thinks that in fact it is "empirically absurd to suppose that we could duplicate the causal posers of neurons entirely in silicon," but says that this cannot be ruled out a priori. He goes on to describe two other ways the experiment might tum out. One is that "as the silicon is progressively implanted into your dwin­ dling brain, you find that the area of your con­ scious experience is shrinking, but that this shows no effect on your external behavior. You find, to your total amazement, that you are indeed losing control of your external behavior. You find, for example, that when the doctors test your vision, you hear them say, 'We are holding up a red ob­ ject in front of you; please tell me what you see.' You want to cry out, 'I can't see anything. I'm going totally blind.' But you hear your voice say­ ing in a way that is completely out of your con­ trol, 'I see a red object in front of me' " (pp. 667). Here, he says, "we are imagining a situation in which you are eventually mentally dead, where you have no conscious mental life whatever, but your externally observable behavior remains the same." This second case is the main one I want to ex­ amine. But Searle also mentions a third possible outcome. This is that "the progressive implanta­ tion of the silicon chips produces no change in your mental life, but you are progressively more and more unable to put your thoughts, feelings,

508

Sydney Shoemaker

and intentions into action" (p.

67). Here your

my point of view, for by hypothesis I have no

external behavior eventually ceases, and the doc­

voluntary control over it, and it moves about, and

tors think you are dead. But you know better.

spouts utterances, in defiance of my will. I do

My focus, as I said, will be on the second case.

seem to see through its eyes and hear through its

Searle takes this to show how it could be known

ears. But given my alienation from it, or rather,

that a certain physical makeup, one consisting in

given that I am alienated from it

assemblies of silicon chips, fails to support men­

ences are veridical, shouldn't I be wondering

tality and consciousness? even if it is the case that

whether these experiences

something having this makeup would be behav­

that I lack the normal ways of checking to see

are

if

my experi­

veridical? Notice

iorally and functionally indistinguishable from

whether things are as they appear-I cannot

a normal human being. This is not explicitly

initiate tests of any sort, and I cannot consult

presented as a possibility argument of the sort

with others. My situation seems rather like a bad

Kripke gives. But, plainly, ifhe can show what he

dream. And of course, if it is a bad dream, or if I

thinks he shows, he thereby establishes at least the

am not entitled to think it isn't, then I am not es­

conceptual possibility of a mindless creature that

tablishing what Searle has me establishing-that

passes the most stringent Turing test imaginable,

the behavior of my body is independent of such

and does so in virtue of having physical states that

consciousness there is in it, and that very soon,

are from a functionalist point of view optimal

when my consciousness has vanished completely,

candidates for being realizations of mental states.

this body will be behaving as it is, passing the

Let the sad character in Searle's story, the one on the verge of extinction, be me. My situation,

most stringent Turing tests, without there being any consciousness in it at all.

as imagined here, is much the same as it is if I

Admittedly, I would be rash to conclude, just

imagine myself as the subject of the most recently

on the basis of my alienation from my behavior,

considered version ofthe Kripke example. In both

that I cannot trust my senses. What I am calling

cases I am alienated from my behavior. In the

alienation from behavior occurs, although in a

Searle example I hear my voice engaged in con­

less dramatic way, in actual cases of paralysis,

versations with others, conversations to which

and we do not think that people in that condition

I am not myself a party, and-here I extend the

should doubt their senses. But it would seem that

example-I see my hand writing answers to

I have reasons for doubting my senses over and

questions on an IQ test, questions that I am, in

above the fact that I have no voluntary control

my weakened state of mind, unable even to un­

over the body I seem to be perceiving from . For if

derstand. The main difference between the two

what my senses tell me is right, people are ascrib­

cases, besides the fact that in Searle's case my

ing to me, on the basis of my behavior, mental

consciousness is waning, is that Searle's story

states that I know I don't have. And these ascrip­

contains an account of how I got into this mess­

tions are regularly being confirmed by my sub­

it is the result of my volunteering to be the subject

sequent behavior. What is at stake here is the

of the philosophical experiment, and the infusions

reliability of a well established practice of mental

of silicon I subsequently underwent. From now

state ascription, one we rely on in all of our deal­

on I will concentrate on Searle's version of the

ings with other people. Can I justifiably take it

case, but my main points will apply to the earlier

that my perceptual experiences are veridical in

version as well.

this instance? .{fthey are then, it seems,

Let me focus on the alienation from my be­

both a well

established practice is systematically issuing in

havior that this case involves. This amounts to

mistaken mental state ascriptions in this case,

a kind of alienation from my body. Indeed, its

I am alienated from by behavior and body. That

status as my body should seem problematic, from

seems to me a reason for saying that I can't be

and

509

The First-Person Perspective

justified in taking my perceptual experiences to be

could reasonably be held to tip the balance

veridical. And if I can't, then in the imagined sit­

against the veridicality assumption.

uation I do not establish what Searle thinks I do.

To see that the veridicality of experiences

can

I am not going to rest my case on this point.

be an issue in such cases, consider a modification

But I think that it has some force. When first­

of Searle's example in which the brain operations

person thought experiments seem to have philo­

are replaced by something less invasive. For ex­

sophically interesting results, the content of the

ample, instead of a series of operations in which

first-person imagining always has an "objective"

brain matter is replaced by silicon, they give me a

as well as a "subjective" component. In the pres­

series of shampoos! As before, w,e will try to sup­

ent case, the subjective component is my being

pose, each item in the series is followed by a

in a certain mental condition, which includes my

diminution of consciousness, and I end up radi­

having certain sense-experiences-my seeming to

cally alienated from my body. But in this case, the

see and hear certain things. The objective com­

behavior of my body in the final stages of the

ponent is my body's being in a certain condition,

procedure stems from just the sorts of neural

and my being surrounded by people and instru­

goings-on that such behavior stems from in nor­

ments of which certain things are true. Normally,

mal cases. At any rate, this is how things seem to

when the veridicality of sense experiences is in no

me. I doubt if anyone will want to maintain that

way in question, it is unproblematic to move from

this thought experiment shows, or even provides

seeming to see such and such to saying that one imagines seeing such and

prima facie evidence, that the actual processes

saying that one imagines

going on in our brains are not metaphysically

such, and from there to saying that one imagines

sufficient for the mental states we take to be

such and such being the case.

But when what is in

manifested by the behavior they produce. And I

question is the relation between the mental states

think that if things did seem to me as just imag­

of a creature and the creature's bodily condition

ined, it would be more reasonable for me to doubt

and situation, then the veridicality or otherwise of

my senses or memories than to conclude that a

the creature's sense experiences is part of what is

series of shampoos could destroy my mentality

in question. Obviously, that my sense-experiences

without affecting my brain or its influence on my

are veridical is not something I know "from the

behavior.

inside," in the way I know that I have them. This

But let's return to Searle's example, with me

is not something about which I have "first-person

again as the subject. As it happens, we do not

authority." There is, to be sure, a presumption in

have to choose between overriding the presump­

favor of the assumption of veridicality. But there

tion that my sense experiences are veridical and

is also a presumption in favor of the assumption

overriding the presumption that our ordinary

that our ordinary third-person ways of ascribing

practice of third-person mental state ascription is

mental states are reliable. If, holding fixed the

reliable. For there is a way of honoring both pre­

nature of my mental states over some interval,

sumptions. Assume that my sense experiences are

these two assumptions come into conflict, nothing

veridical. Then I am radically alienated from the

that I know from the inside, nothing about which

body I am perceiving from. Given its indepen­

I have first-person authority, tips the balance in

dence of my will, my claim that it is my body is a

favor of the assumption that my perceptual expe­

bit shaky. So maybe someone else has a less shaky

riences are veridical. Perhaps what we have is

claim on it. That is precisely what the third­

a standoff between the two assumptions. But I

person evidence indicates to others. The behavior

think that the fact that on the veridicality as­

of the body is such as to lead them to take it to be

sumption I am alienated from the body from

the body of a person having certain mental states,

which I am supposed to be doing the perceiving

mental states that as a matter of fact are, except

Sydney Shoemaker

for the perceptual experiences among them, utterly different from my own. Nothing I know from the inside, nothing about which I have first­ person authority, gives me any reason to reject this possibility. Is this a possibility? Well, to begin with, there would seem to be no conceptual incoherence in the idea that two minds, or persons, or "con­ sciousnessess," might simultaneously animate a single body, or at any rate have that body as the point of view from which they experience the world. This is what some have thought happens in split-brain cases, and what others have thought happens in cases of multiple personality; and while that view of those cases appears not to fit the actual facts, there are possible facts that it does fit. If we think about the features of these cases that make it tempting to speak of there being multiple persons, or multiple minds, in a single body, it is not difficult to envisage cases of which this would be the literally correct description. But how do we apply this to Searle's example? As an example of an unsatisfactory application of it, let me quote my own response to Searle when he presented this example at a conference a year or two ago: "you seem to imply that . . . just after your consciousness fades out you have all the functional organization there without any men­ tality. Surely that's not warranted, because it's perfectly compatible with this that as you fade out someone else is coming in." 5 Daniel Dennett subsequently put the point in a similar way in his review of Searle's book. 6 Supposing that I am the subject of the series of operations, what this for­ mulation suggests is that what the scientists were perhaps unwittingly doing in their series of oper­ ations was building in my skull a new person, one with a silicon brain, while gradually destroying the brain and person, namely me, that was there originally. As appropriate connections were es­ tablished between the silicon brain and the ner­ vous and motor systems, the new brain took over control of the body. The trouble with this version

510

o f the story i s that i t i s not compatible with a central feature of it, namely that throughout the series of replacements the behavior of my body was such that from a third-person point of view it appeared to be animated by a single person with normal mental abilities and normal conscious­ ness, and exhibiting normal mental continuity over time. Yoking a waning mind to a waxing mind could hardly be expected to produce this result. What it should produce, instead. is con­ flicted behavior, dominated in the early parts of the interval by behavior showing mental decline, dominated during the later parts of it by behavior showing the opposite, and perhaps manifesting at various times behavior analogous to that of the split-brain monkey, reported by Tom Nagel, whose right and left hands had a tug of war over a nut. 7 But there is a version of the story that is com­ patible with what we are supposed to imagine about the behavior of the body. It says that a single person animated the body throughout the interval, and that during the interval that person's brain was gradually reconstituted-it began as a normal human brain, and ended as a functionally equivalent silicon brain. It follows that if at the end of the story there was "in " the body a feeble mind that was not in control of it, then that was not the mind that was there at the outset. Or, putting it in terms of persons, the person who near the end found himself a mentally enfeebled prisoner in the body is not, although he thinks he is, the same as the person who at the beginning volunteered to undergo the series of operations. Perhaps the case could be construed as a case of "fission," in which both of the inhabitants of the body, the mentally enfeebled prisoner as well as the person who controls the body's behavior, have veridical "quasi-memories" corresponding to the life of the original person. But if so, it is unequal fission, and the "closest continuer," the person with the best claim to be the original possessor of the body, is the one who controls the body's behavior, not the mentally enfeebled prisoner.

The First-Person Perspective

511

Someone might object that I have here fallen

nability of the series of events is unproblematic.

into skepticism about memory, and about per­

And if each event in the series is described in the

sonal identity. Surely, it will be said, if someone

first-person, it may seem that one has imagined

remembers being the person who did such and

oneself undergoing such a series of events. But if

such, that person has every reason to think that he

imaginability is to bear on possibility, we need to

is the person who did such and such. And if I

go slow here. First of all, we need to distinguish

imagine remembering being the person who did

two senses in which one can imagine something

such and such,

from the "first-person point of view." This might

I have every reason to describe the

imagined situation as one in which I am the per­

mean simply imagining it "from the inside"­

son who did such and such.

imagining some aspect of the life of a person as it

But the point here is much the same as the point

might be experienced by that person. Or it might

made earlier about sense perception. Normally

mean imagining oneselfdoing or undergoing such

one has no reason to distrust one's senses, and

and such, where the imagining is again from the

normally one has no reason to distrust one's

inside. Imagining from the first-person point of

memories. But in the thought experiment now

view in the first sense needn't involve imagining

under consideration we have mover very far from

from the first-person point of view in the second

normal circumstances. And the content of the

sense. If! imagine the battle of Cannae as it might

memory, insofar as it is first-person content, ac­

have been experienced by Hannibal, I do not

tually conflicts with the judgments of personal

thereby imagine

identity others would make on the basis of third­

rate, in a sense in which the imaginability of

being

Hannibal�not, at any

person evidence. Assuming that our subject can

something is at least prima facie evidence of its

trust his senses to the extent of being entitled to

possibility. And if I imagine a series of personal

think that the people about him really are saying

episodes, imagining each from the point of view

the things they seem to be saying, this provides

of the person involved and so imagining each

him with positive reasons for distrusting his

from the inside, I do not thereby imagine myself,

memory. Conversely, if he trusts his memory this

or any single person, being the subject of all of

gives him a reason for distrusting his senses. What

those episodes. It is tempting to say that I can

he cannot have good reason for believing is that

simply stipulate that the subject of one of my

he is the person who volunteered to be the

imaginings is myself. And so I can, up to a point.

both

subject of the series of operations

and

he is in

But there are limits to what one can coherently

the envisaged situation, that is, is alienated from

stipulate. If I successively imagine how the Presi­

the behavior of "his" body in the way described.

dent's State of the Union Address is being re­

So the imaginability of being in the situation he

ceived by the different members of the Congress,

knows he is in (certain things

in each case imagining the reception from the in­

seeming to

be the

case) is not evidence for the possibility of some­

side, I cannot

one's being in the situation whose possibility is in

a single series of events and stipulate that in each

question (those things really being the case).

case the person imagined from the inside is

both regard this as an imagining of

One upshot of this discussion is that some cases

myself-not, at least, if imaginability is to be

of imagining from the first-person perspective are

evidence of possibility. Returning to the Searle

problematic in a way that is not initially appar­

example, I can imagine from the inside first the

ent. One imagines what purports to be a succes­

agreeing to be the subject of the experiment, and

sion of events involving oneself. Each event in

then the somewhat diminished mental condition

the succession is imagined "from the inside." If

of someone after the first operation, and then the

the imaginability of each event in the succession

somewhat more diminished mental condition of

is unproblematic, it may seem that the imagi-

someone after the second operation, and so on. It

Sydney Shoemaker

is natural to describe this by saying that I imagine agreeing to be the subject, and then experiencing diminished consciousness after the first operation, and then experiencing a further diminished con­ sciousness after the second operation, and then . . . and so on. But this implies, illegitimately, that it is one and the same person who is the subject of all of these imagined episodes. If I stipulate that I am the imagined person who initially agrees to submit to the series of operations, then I cannot, without begging the question, stipulate also that I am the imagined person who experiences the final stages of the extinction of his consciousness while observing the external behavior of the body to go on as before. And if I stipulate that the latter person is myself, then I cannot stipulate that the former person is. That is, I cannot make these stipulations if the description of what is imagined is not to beg the question by assuming the truth of the possibility claim it is supposed to support. But let me stipulate now that I am the person with radically diminished consciousness who is totally without control of the body from which he experiences the world. And suppose that I am right in claiming that I would not, in these cir­ cumstances, be entitled to say that there is no consciousness behind the silicon-driven behavior of the body I am imprisoned in. So we agree that for all we know there is another mind in there who is running the show. If we let "Sydney" be the name of the person who agreed to undergo the series of operations, we agree that while it seems to me that I am Sydney, since I remember Syd­ ney's life from the inside, there is good third­ person evidence that Sydney is the man whose behavior is generated in the now largely silicon brain that now inhabits my skull. Still, it might be thought that I know at least that there is something wrong with the view that a person's mentality is determined by the functional organi­ zation of the person's brain or body. By hypoth­ esis, the brain part replacements were all such as to preserve the relevant sorts of functional orga­ nization. And let's take this to mean that the functional organization throughout is such that

512

according to functionalism there is just one mind realized in the brain. It is ruled out, in other words, that there are two functional organiza­ tions there, one superimposed on the other, cor­ responding to the total mental states of two different persons. So, assuming functionalism, the mentality associated with this body ought to be that which others ascribe, on the basis of be­ havior, to the man they call "Sydney." And while I am not in a position to know directly that there is not this sort of mentality associated with the body, I am in a position to know that there is an­ other sort of mentality, quite different from this, which is associated with it-namely the dimin­ ished consciousness, frustration and despair that I am now experiencing. This might seem to under­ mine the functionalist view, even if it does not undermine the reasons others have for thinking that Sydney, as they conceive him, exists. Like the third outcome Searle imagines for his thought experiment, this seems to support the claim, not that that there is no functional organization that is sufficient for the possession of the mental states in question (which is what the possibility of the second outcome was alleged to show), but that there is none that is necessary for this. But I do not think that this argument fares any better than the earlier one. First of all, once we are this far into the realm of fantasy, multiple bodies for one mind are as much a possibility to be reckoned with as multiple minds for one body. Supposing that the autocerebroscope indicates that the functional states in question are not real­ ized in the body from which I see, it remains a possibility that they are realized elsewhere. I do not think that there is anything I could observe that could assure me that nowhere. and in no way, are there instantiated functional states that could underwrite the mental states I know myself to have. Moreover-and here I revert to the "bad dream point" invoked earlier-if what I seem to observe did seem to assure me of this, I would have at least as much reason to doubt the verid­ icality of my sense experiences and/or memories as I did in the previous case in which a series of

The First-Person Perspective

513

shampoos seemed to produce alienation from my

in the negative. I t i s of course an empirical

own body. In fact, of course, it is unrealistic to

question whether creatures with such a physical

suppose that I would have to choose between

makeup could be behaviorally like us, to the

wholesale distrust of my senses and the rejection

extent of being able to pass the most stringent

of functionalism; I might more reasonably con­

Turing tests. And it seems entirely possible to me,

clude that the cerebroscope is on the blink. But there is more to be said than this. For I

as I am sure it does to Searle, that the answer to this question is no. To put the point in a func­

submit that it is out and out incoherent, and not

tionalist way, it may well be that the evolutionary

just highly implausible, to suppose that the testi­

process that resulted in us came up with the only

mony of my senses could establish the negative

possible implementation of the functional orga­

existential that nowhere, neither in the body I see

nization that bestows our sorts of mental states,

from nor anywhere else, does there exist a real­

and the only possible physical organization that

ization of the appropriate functional states that

bestows our behavioral dispositions. But Searle

could be a realization of my current mental states.

thinks that even supposing that there are possible

For the testimony of my senses to establish that

creatures whose physical makeup is very different

negative existential would be for me to come to

from ours but who are behaviorally just like us,

know on the basis of my senses that my current

and have whatever functional makeup you like,

mental states are not playing the functional roles

is an empirical question whether such creatures

it

that according to functionalism are constitutive

could have a conscious mental life like ours. Now

of such states. But if I

it is very difficult to see how this empirical ques­

know

something on the

basis of my senses, this requires, surely, that my

tion, supposing that it is one, could be settled in

beliefs be modified by my sensory inputs, acting

the negative by an investigation from the third­

in concert with my background beliefs, in a way

person point of view. Obviously it is not to be

that conforms to certain principles of rationality.

settled by observing the behavior of the creatures,

sense-expe­

since the creatures whose mentality is in question

riences, and beliefs, to play functional roles that

are precisely those which can be counted on to

And that

is for certain mental states,

according to functionalism are constitutive of

pass every behavioral test with flying colors. Nor

them. To the extent that knowledge requires ra­

does it seem that it could be settled in the negative

tionality, and that what are held to be the defining

by investigating their internal makeup; the most

functional roles of mental states are the causal

that could be established in that way is that cer­

roles constitutive of rationality, it is incoherent to

tain functional state� are realized differently in

claim that one could know that one's own mental

them than in us, and not that those functional

states are not playing the defining causal roles.

states do not bestow mentality. Searle thinks that

Admittedly, this argument by itself shows only

for any functional organization, it is an empirical

that one could not know that there is a total lack

question whether creatures that have it have

of realization of the relevant functional states, not

genuine mental states. It does appear that if this

that one could not know that there is a partial

is an empirical question, and one that could in

such lack. But we have seen other reasons for

principle be settled in the negative, it is one that

denying that one could know even that.

could only be so settled from the first-person

Let's take stock. Searle has claimed that it is an

point of view. No wonder, then, that Searle de­

empirical question whether beings with a physical

vised his first-person thought experiment. What I

makeup different from ours, for example, beings

have shown, however, is that where the question

made of silicon, can have conscious mental lives 8 of the sort we have. I take it this means a ques­

cannot be settled in the negative from the third­

tion that could in principle be settled, empirically,

the negative from the first-person point of view.

person point of view, it also cannot be settled in

Sydney Shoemaker

5 14

I suspect that some people will be unmoved by

the sufficiency of bodily (or functional) states of

all this because they think that the third-person

affairs for mental states of affairs depends on the

point of view is, right from the start, parasitic on

assumption that one is the sole inhabitant of one's

the first-person point of view. This would be true

body, and that assumption is not one whose truth

if we know what the various mental states are

one can discover empirically from the first-person

"from our own case," and if our entitlement to

point of view without relying on other assump­

ascribe them to others on the basis of behavior

tions about how the mental and physical realms

rests on something like the argument from anal­

are related.

ogy. This view, call it the analogical position, is

The epistemology of modality is a large topic,

one of the hardiest weeds in the philosophical

and I have barely scratched the surface. But here,

garden. It has been sprayed, in this century,

for whatever it is worth, is the moral I am inclined

with everything from the verificationist theory of

to draw. Where it seems that one can imagine

meaning to Wittgenstein's private language ar­

discovering the realization of a putative possi­

gument, and it keeps coming back, usually under

bility

assumed names. A wholesale assault on it would

should always ask whether the seeming discovery

from

the

first-person perspective, one

be a task for another occasion. But I think that

could be confirmed from the third-person point of

the considerations I have raised provide part of

view. If one finds that it is impossible in principle

see whether

the case against it. They bring out that the first­

that it should be, one should look to

person point of view does not provide a perspec­

the first-person thought experiment can be faulted

tive from which, starting with no assumptions

in the ways I have tried to fault the first-person thought experiments here . I have not proven that

about the relations between mental and physical states of affairs, and about the causal roles of mental states, one can proceed to investigate em­

the latter will always be the case. But I suspect

that it will be . 1 0

pirically what these relations and roles are, first discovering what they are in one's own case and then extrapolating inductively to the case of others. This is for three related reasons. First, one can discover nothing at all about bodies, or about physical states of affairs, without assuming the veridicality of one's sense experiences, and to assume that is to assume something about the relations between mental and physical states of affairs-relations that are constitutively bound up in what it is for a body to

be

the body of a

particular person. 9 The truth of this assumption is certainly not something one can straightfor­ wardly discover empirically from the first-person point of view. Second, some of the causal rela­ tions amongst mental states involved in ration­ ality must hold as a condition of one's corning to know anything at all, and there is no sense to the idea that one might investigate empirically, by introspection, whether these hold in one's own case. Finally, the use of the first-person point of view to discover counterexamples to claims about

Notes This chapter was delivered as the Presidential Address before the Ninetieth Annual Eastern Division Meeting of The American Philosophical Association in Atlanta, Georgia, December 29, 1 993.

I. Naming and Necessity, Harvard University Press, 1972, pp. 144-1 55. 2. See, e.g., Putnam's "The Nature of Mental States," in his Philosophical Paper. Volume 2 (Cambridge: Cam­ bridge University Press, 1975). 3. Someone might try to ground a difference in the fact that judgments about the pains of others are inferential and subject to error in ways in which judgments about one's own pains are not. This might seem a reason for saying that in the first-person thought experiment what one imagines knowing is that someone is in pain without there being any C-fiber stimulation going on, while in the third-person thought experiment what one imagines knowing is only that someone is manifesting pain be­ havior without there being any C-fiber stimulation going

The First-Person Perspective

on. But if one insists that the only imaginable states of affairs are ones to which one has an access that is non­ inferential and not subject to error, then the first-person thought experiment is no better off than the third-person thought experiment; both require that one have access to whether there is C-fiber firing going on in one, and any access one has to that will be subject to error and, arguably, inferential in whatever sense one's access to the pains of others is inferential. 4. MIT Press, Cambridge, Mass., 1992.

5. See Experimental and Theoretical Studies of Con­ sciousness, Ciba Foundation Symposium 1 74 (John Wiley & Sons, 1 993), p. 73. Unaccountably, the pub­ lished transcript has the word "representation" where I have put "organization." 6. The Journl of Philosophy, XC, 4, April 1 993, p. 1 98. 7. See Nagel's "Brain Bisection and the Unity of Con­ sciousness," Synthese, vol. 22, 1977. 8. See his "Minds, Brains and Programs," The Behav­ ioral and Brain Sciences III, 3 (September 1 980), 4 1 7 24. 9. See my "Embodiment and Behavior," in A. Rorty (cd.), The Identities of Persons ( Berkeley: University of California Press, 1 976), reprinted in my Identity, Cause, and Mind, Cambridge, 1984. 10. My thanks to Ned Block, Mark Crimmins, Carl Ginet, Chris Hill, Norman Kretzmann, Dick Moran, and Bob Stalnaker for their very helpful comments on earlier versions of this. All gave me good advice, not all of which I took.

515

VII

SUBJECTIVITY AND EXPLANATORY GAP

32

What Is It Like to Be a Bat?

Thomas Nagel

Consciousness is what makes the mind-body

Conscious experience is a widespread phenom­

problem really intractable. Perhaps that is why

enon. It occurs at many levels of animal life,

current discussions of the problem give it little

though we cannot be sure of its presence in the

attention or get it obviously wrong. The recent

simpler organisms, and it is very difficult to say in

wave of reductionist euphoria has produced

general what provides evidence of it. (Some ex­

several analyses of mental phenomena and men­

tremists have been prepared to deny it even of

tal concepts designed to explain the possibility

mammals other than man.) No doubt it occurs

of some variety of materialism, psychophysical

in countless forms totally unimaginable to us, on

identification, or reduction. 1 But the problems

other planets in other solar systems throughout

dealt with are those common to this type of

the universe. But no matter how the form may

reduction and other types, and what makes

vary, the fact that an organism has conscious

the mind-body problem unique, and unlike the

experience

water-H20 problem or the Turing machine-IBM

something it is like to

machine problem or the lightning-electrical dis­

may be further implications about the form of the

at all

means, basically, that there is

be

that organism. There

charge problem or the gene-DNA problem or the

experience; there may even (though I doubt it) be

oak tree-hydrocarbon problem, is ignored.

implications about the behavior of the organism.

Every reductionist has his favorite analogy

But fundamentally an organism has conscious

from modem science. It is most unlikely that any

mental states if and only if there is something that

of these unrelated examples of successful reduc­

it is like to be that organism-something it is like for the organism.

tion will shed light on the relation of mind to brain. But philosophers share the general human

We may call this the subjective character of

weakness for explanations of what is incompre­

experience. It is not captured by any of the famil­

hensible in terms suited for what is familiar and

iar, recently devised reductive analyses of the

well understood, though entirely different. This

mental, for all of them are logically compatible

has led to the acceptance of implausible accounts

with its absence. It is not analyzable in terms of

of the mental largely because they would permit

any explanatory system of functional states, or

familiar kinds of reduction. I shall try to explain

intentional states, since these could be ascribed to

why the usual examples do not help us to under­

robots or automata that behaved like people

stand the relation between mind and body-why,

though they experienced nothing. 2 It is not ana­

indeed, we have at present no conception of what

lyzable in terms of the causal role of experiences

dn explanation of the physical nature of a mental

in relation to typical human behavior-.for sim­

phenomenon would be. Without consciousness

ilar reasons. 3 I do not deny that conscious mental

the mind-body problem would be much less in­

states and events cause behavior, nor that they

teresting. With consciousness it seems hopeless.

may be given functional characterizations. I deny

The most important and characteristic feature of

only that this kind of thing exhausts their analy­

conscious mental phenomena is very poorly un­

. sis. Any reductionist program has to to be based

derstood. Most reductionist theories do not even

on an analysis of what is to be reduced. If the

try to explain it. And careful examination will

analysis leaves something out, the problem will

show that no currently available concept of re­

be falsely posed. It is useless to base the defense

duction is applicable to it. Perhaps a new theo­

of materialism on any analysis of mental phe­

retical form can be devised for the purpose, but

nomena that fails to deal explicitly with their

such a solution, if it exists, lies in the distant in­

subjective character. For there is no reason to

tellectual future.

suppose that a reduction which seems plausible

Thomas Nagel

520

when no attempt is made" to account for con­

related to us than those other species, nevertheless

sciousness can be extended to include conscious­

present a range of activity and a sensory appara­

ness. Without some idea, therefore, of what the

tus so different from ours that the problem I want

subjective character of experience is, we cannot

to pose is exceptionally vivid (thOUgh it certainly

know what is required of a physicalist theory.

could be raised with other species). Even without

While an account of the physical basis of mind

the benefit of philosophical reflection, anyone who

must explain many things, this appears to be the

has spent some time in an enclosed space with an

most difficult. It is impossible to exclude the phe­

excited bat knows what it is to encounter a fun­

nomenological features of experience from a re­

damentally alien form of life.

duction in the same way that one excludes the

I have said that the essence of the belief that

phenomenal features of an ordinary substance

bats have experience is that there is something

from a physical or chemical reduction of it­

that it is like to be a bat. Now we know that most

namely, by explaining them as effects on the

bats (the Microchiroptera, to be precise) perceive

minds of human observers.4. If physicalism is to

the external world primarily by sonar, or echo­

be defended, the phenomenological features must

location, detecting the reflections, from objects

themselves be given a physical account. But when

within range, of their own rapid, subtly modu­

we examine their subjective character it seems

lated, high-frequency shrieks. Their brains are

that such a result is impossible. The reason is that

designed to correlate the outgoing impulses with

every subjective phenomenon is essentially con­

the subsequent echoes, and the information thus

nected with a single point of view, and it seems

acquired enables bats to make precise discrim­

inevitable that an objective, physical theory will

inations of distance, size, shape, motion, and tex­

abandon that point of view.

ture comparable to those we make by vision. But

Let me first try to state the issue somewhat

bat sonar, though clearly a form of perception, is

more fully than by referring to the relation be­

not similar in its operation to any sense that we

tween the subjective and the objective, or between

possess, and there is no reason to suppose that it is

the pour-soi and the

This is far from easy.

subjectively like anything we can experience or

Facts about what it is like to be an X are very

imagine. This appears to create difficulties for

en-soi.

peculiar, so peculiar that some may be inclined

the notion of what it is like to be a bat. We must

to doubt their reality, or the significance of claims

consider whether any method will permit us to

about them. To illustrate the connection between

extrapolate to the inner life of the bat from our

subjectivity and a point of view, and to make

own case, s and if not, what alternative methods

evident the importance of subjective features, it

there may be for understanding the notion.

will help to explore the matter in relation to an

Our own experience provides the basic material

example that brings out clearly the divergence

for our imagination, whose range is therefore

between the two types of conception, subjective

limited. It will not help to try to imagine that one

and objective.

has webbing on one's arms, which enables one to

I assume we all believe that bats have experi­

fly around at dusk and dawn catching insects in

ence. After all, they are mammals, and there is no

one's mouth; that one has very poor vision, and

more doubt that they have experience than that

perceives the surrounding world by a system of

mice or pigeons or whales have experience. I have

reflected high-frequency sound signals; and that

chosen bats instead of wasps or flounders because

one spends the day hanging upside down by one's

if one travels too far down the phylogenetic tree,

feet in an attic. Insofar as I can imagine this

people gradually shed their faith that there is ex­

(which is not very far), it tells me only what it

perience there at all. Bats, although more closely

would be like for

me to behave as a bat behaves.

What Is It Like to Be a Bat?

But that is not the question. I want to know what it is like for a bat to be a bat. Yet if I try to imag­ ine this, I am restricted to the resources of my own mind, and those resources are inadequate to the task. I cannot perform it either by imagining additions to my present experience, or by imag­ ining segments gradually subtracted from it, or by imagining some combination of additions, sub­ tractions, and modifications. To the extent that I could look and behave like a wasp or a bat without changing my funda­ mental structure, my experiences would not be anything like the experiences of those animals. On the other hand, it is doubtful that any mean­ ing can be attached to the supposition that I should possess the internal neurophysiological constitution of a bat. Even if I could by gradual degrees be transformed into a bat, nothing in my present constitution enables me to imagine what the experiences of such a future stage of myself thus metamorphosed would be like. The best evi­ dence would come form the experiences of bats, if we only knew what they were like. So if extrapolation from our own case is in­ volved in the idea of what it is like to be a bat, the extrapolation must be incompletable. We cannot form more than a schematic conception of what it is like. For example, we may ascribe general types of experience on the basis of the animal's struc­ ture and behavior. Thus we describe bat sonar as a form of three-dimensional forward perception; we believe that bats feel some versions of pain, fear, hunger, and lust, and that they have other, more familiar types of perception besides sonar. But we believe that these experiences also have in each case a specific subjective character, which it is beyond our ability to conceive. And if there is conscious life elsewhere in the universe, it is likely that some of it will not be describable even in the most general experiential terms available to us. 6 (The problem is not confined to exotic cases, however, for it exists between one person and an­ other. The subjective character of the experience of a person deaf and blind from birth is not ac­ cessible to me, for example, nor presumably is

521

mine to him. This does not prevent us each from believing that the other's experience has such a subjective character.) If anyone is inclined to deny that we can believe in the existence of facts like this whose exact nature we cannot possibly conceive, he should reflect that in contemplating the bats we are in much the same position that intelligent bats or Martians 7 would occupy if they tried to form a conception of what is was like to be us. The structure of their own minds might make it im­ possible for them to succeed, but we know they would be wrong to conclude that there is not anything precise that it is like to be us: that only certain general types of mental state could be as­ cribed to us (perhaps perception and appetite would be concepts common to us both; perhaps not). We know they would be wrong to draw such a skeptical conclusion because we know what it is like to be us. And we know that while it includes an enormous amount of variation and complex­ ity, and while we do not possess the vocabulary to describe it adequately, its subjective character is highly specific, and in some respects describable in terms that can be understood only by creatures like us. The fact that we cannot expect ever to accommodate in our language a detailed descrip­ tion of Martian or bat phenomenology should not lead us to dismiss as meaningless the claim that bats and Martians have experiences fully comparable in richness of detail to our own. It would be fine if someone were to develop con­ cepts and a theory that enabled us to think about those things; but such an understanding may be permanently denied to us by the limits of our nature. And to deny the reality or logical signif­ icance of what we can never describe or under­ stand is the crudest form of cognitive dissonance. This brings us to the edge of a topic that re­ quires much more discussion than I can give it here: namely, the relation between facts on the one hand and conceptual schemes or systems of representation on the other. My realism about the subjective domain in all its forms implies a beliefin the existence of facts beyond the reach of hwnan

Thomas Nagel

concepts. Certainly it is possible for a human being to believe that there are facts which humans never will possess the requisite concepts to repre­ sent or comprehend. Indeed, it would be foolish to doubt this, given the finiteness of humanity's expectations. After all, there would have been transfinite numbers even if everyone had been wiped out by the Black Death before Cantor dis­ covered them. But one might also believe that there are facts which could not ever be represented or comprehended by human beings, even if the species lasted forever-simply because our struc­ ture does not permit us to operate with concepts of the requisite type. This impossibility might even be observed by other beings, but it is not clear that the existence of such beings, or the possibility of their existence, is a precondition of the significance of the hypothesis that there are humanly inaccessible facts. (After all, the nature of beings with access to humanly inaccessible facts is presumably itself a humanly inaccessible fact.) R:fl.ection on what it is like to be a bat seems to lead us, therefore, to the conclusion that there are facts that do not consist in the truth of propositions expressible in a human language. We can be compelled to recognize the eXistence of such facts without being able to state or com­ prehend them. I shall not pursue this subject, however. Its bearing on the topic before us (namely, the mind­ body problem) is that it enables us to make a general observation about the subjective charac­ ter of experience. Whatever may be the status of facts about what it is like to be a human being, or a bat, or a Martian, these appear to be facts that embody a particular point of view. I am not adverting here to the alleged privacy of experience to its possessor. The point of view in question is not one accessible only to a single in­ dividual. Rather it is a type. It is often possible to take up a point of view other than one's own, so the comprehension of such facts is not limited to one's own case. There is a sense in which phe­ nomenological facts are perfectly objective: one person can know or say of another what the

522

quality of the other's experience is. They are sub­ jective, however, in the sense that even this ob­ jective ascription of experience is possible only for someone sufficiently similar to the object of ascription to be able to adopt his point of view­ to understand the ascription in the first person as well as in the third, so to speak. The more differ­ ent from oneself the other experiencer is, the less success one can expect with this enterprise. In our own case we occupy the relevant point of view, but we will have as much difficulty understanding our own experience properly if we approach it from another point of view as we would if we tried to understand the experience of another species without taking up its point of view. 8 This bears directly on the mind-body problem. For if the facts of experience-facts about what it is like for the experiencing organism-are acces­ sible only from one point of view, then it is a mystery how the true character of experiences could be revealed in the physical operation of that organism. The latter is a domain of objective facts par excellence-the kind that can be observed and understood from many points of view and by individuals with differing perceptual systems. There are no comparable imaginative obstacles to the acquisition of knowledge about bat neuro­ physiology by human scientists, and intelligent bats or Martians might learn more about the human brain than we ever will. This is not by itself an argument against reduction. A Martian scientist with no under­ standing of visual perception could understand the rainbow, or lightning, or clouds as physical phenomena, though he would never be able to understand the human concepts of rainbow, lightning, or cloud, or the place these things occupy in our phenomenal world. The objective nature of the things picked out by these concepts could be apprehended by him because, although the concepts themselves are connected with a particular point of view and a particular visual phenomenology, the things apprehended from that point of view are not: they are observable from the point of view but external to it; hence

What Is It Like to Be a Bat?

523

they can be comprehended from other points of

terms of its more general effects and of properties

view also, either by the same organisms or by

detectable by means other than the human senses.

others. Lightning has an objective character that

The less it depends on a specifically human view­

is not exhausted by its visual appearance, and this

point, the more objective is our description. It is

can be investigated by a Martian without vision.

possible to follow this path because although the

To be precise, it has a

concepts and ideas we employ in thinking about

more

objective character

than is revealed in its visual appearance. In

the external world are initially applied from a

speaking of the move from subjective to objective

point

characterization, I wish to remain noncommittal

apparatus, they are used by us to refer to things

of view

that

involves

our

perceptual

about the existence of an end point, the com­

beyond themselves-toward which we

pletely objective intrinsic nature of the thing,

phenomenal point of view. Therefore we can

have

the

which one might or might not be able to reach. It

abandon it in favor of another, and still be think­

may be more accurate to think of objectivity as a

ing about the same things.

direction in which the understanding can travel.

Experience itself, however, does not seem to fit

And in understanding a phenomenon like light­

the pattern. The idea of moving from appearance

ning, it is legitimate to go as far away as one can

to reality seems to make no sense here. What is

from a strictly human viewpoint.9

the analogue in this case to pursuing a more ob­

In the case of experience, on the other hand, the

jective understanding of the same phenomena by

connection with a particular point of view seems

abandoning the initial subjective viewpoint to­

much closer. It is difficult to understand what

ward them in favor of another that is more ob­

could be meant by the

jective but concerns the same thing? Certainly it

objective

character of an

experience, apart from the particular point of

appears unlikely that we will

view from which its subject apprehends it. After

nature of human experience by leaving behind

get closer to the real

all, what would be left of what it was like to be a

the particularity of our human point of view and

bat if one removed the viewpoint of the bat? But if

striving for a description in terms accessible to

experience does not have, in addition to its sub­

beings that could not imagine what it was like to

jective character, an objective nature that can be

be us. If the subjective character of experience is

apprehended from many different points of view,

fully comprehensible only from one point of view,

then how can it be supposed that a Martian in­

then any shift to greater objectivity-that is, less

vestigating my brain might be observing physical

attachment to a specific viewpoint-does not

processes which were my mental processes (as he

take us nearer to the real nature of the phenom­

might observe physical processes which were

enon: it takes us farther away from it.

bolts of lightning), only from a different point

In a sense, the seeds of this objection to the

of view? How, for that matter, could a human

reducibility of experience are already detectable

physiologist observe them from another point of

in successful cases of reduction; for in discovering

view?

sound to be, in reality, a wave phenomenon in air

10

We appear to be faced with a general difficulty

or other media, we leave behind one viewpoint to

about psychophysical reduction. In other areas

take up another, and the auditory, human, or

the process of reduction is a move in the direction

animal viewpoint that we leave behind remains

of greater objectivity, toward a more accurate

unreduced. Members of radically different species

view of the real nature of things. This is accom­

may both understand the same physical events

plished by reducing our dependence on individual

in objective terms, and this does not require that

or species-specific points of view toward the ob­

they understand the phenomenal forms in which

ject of investigation. We describe it not in terms

those events appear to the senses of members of

of the impressions it makes on our senses, but in

the other species. Thus it is a condition of their

Thomas Nagel

referring to a common reality that their more

524

the "is" alone. We know how both "X" and " Y"

particular viewpoints are not part of the common

refer, and the kinds of things to which they refer,

reality that they both apprehend. The reduction

and we have a rough idea how the two referential

can succeed only if the species-specific viewpoint

paths might converge on a single thing, be it an

is omitted from what is to be reduced.

object, a person, a process, an event, or whatever.

But while we are right to leave this point of

But when the two terms of the identification are

view aside in seeking a fuller understanding of the

very disparate it may not be so clear how it could

external world, we cannot ignore it permanently,

be true. We may not have even a rough idea of

since it is the essence of the internal world, and

how the two referential paths could converge, or

not merely a point of view on it. Most of the neo­

what kind of things they might converge on, and

behaviorism of recent philosophical psychology

a theoretical framework may have to be supplied

results from the effort to substitute an objective

to enable us to understand this. Without the

concept of mind for the real thing, in order to

framework, an air of mysticism surrounds the

have nothing left over which cannot be reduced.

identification.

If we acknowledge that a physical theory of mind

This explains the magical flavor of popular

must account for the subjective character of ex­

presentations of fundamental scientific discov­

perience, we must admit that no presently avail­

eries, given out as propositions to which one must

able conception gives us a clue how this could be

subscribe without really understanding them. For

done. The problem is unique. If mental processes

example, people are now told at an early age

are indeed physical processes, then there is some­

that all matter is really energy. But despite the

thing it is like, intrinsically, l 1 to undergo certain

fact that they know what "is" means, most of

physical processes. What it is for such a thing to

them never form a conception of what makes this

be the case remains a mystery. What moral should be drawn from these re­

claim true, because they lack the theoretical background.

flections, and what should be done next? It would

At the present time the status of physicalism is

be a mistake to conclude that physicalism must

similar to that which the hypothesis that matter

be false. Nothing is proved by the inadequacy of

is energy would have had if uttered by a pre­

physicalist hypotheses that assume a faulty ob­

Socratic philosopher. We do not have the begin­

jective analysis of mind. It would be truer to say

nings of a conception of how it might be true. In

that physicalism is a position we cannot under­

order to understand the hypothesis that a mental

stand because we do not at present have any con­

event is a physical event, we require more than an

ception of how it might be true. Perhaps it will

understanding of the word "is." The idea of how

be thought unreasonable to require such a con­

a mental and a physical term might refer to the

ception as a condition of understanding. After all,

same thing is lacking, and the usual analogies

it might be said, the meaning of physicalism is

with theoretical identification in other fields fail

clear enough: mental states are states of the body;

to supply it. They fail because if we construe the

mental events are physical events. We do not

reference of mental terms to physical events on

know

which

physical states and events they are,

the usual model, we either get a reappearance of

but that should not prevent us from understanding

separate subjective events

the hypothesis. What could be clearer than the

which mental reference to physical events is

words "is" and "are?" But I believe it is precisely this apparent clarity of the word "is" that is deceptive. Usually, when

as the effects through

secured, or else we get a false account of how mental terms refer (for example, a causal behav­ iorist one).

we are told that X is Y we know how it is supposed

Strangely enough, we may have evidence for

be true, but that depends on a conceptual or

the truth of something we cannot really under­

theoretical background and is not conveyed by

stand. Suppose a caterpillar is locked in a sterile

to

525

What Is It Like to Be a Bat?

safe by someone unfamiliar with insect meta­

equipped to think about the SUbjective character

morphosis, and weeks later the safe is reopened,

of experience without relying on the imagina­

revealing a butterfly. If the person knows that the

tion-without taking up the point of view of the

safe has been shut the whole time, he has reason

experiential subject. This should be regarded as a

to believe that the butterfly is or was once the

challenge to form new concepts and devise a new

caterpillar, without having any idea in what sense

method-an objective phenomenology not de­

this might be so. (One possibility is that the

pendent on empathy or the imagination. Though

caterpillar contained a tiny winged parasite that

presumably it would not capture everything, its

devoured it and grew into the butterfly.)

goal would be to describe, at least in part, the

It is conceivable that we are in such a position

subjective character of experiences in a form

with regard to physicalism. Donald Davidson has

comprehensible to beings incapable of having

argued that if mental events have physical causes

those experiences.

and effects, they must have physical descriptions.

We would have to develop such a phenomen­

He holds that we have reason to believe this even

ology to describe the sonar experiences of bats;

though we do not-and in fact could not-have a

but it would also be possible to begin with

general psychophysical theory. 12 His argument

humans. One might try, for example, to develop

applies to intentional mental events, but I think

concepts that could be used to explain to a person

we also have some reason to believe that sensa­

blind from birth what it was like to see. One

tions are physical processes, without being in a

would reach a blank wall eventually, but it should

position to understand how. Davidson's position

be possible to devise a method of expressing in

is that certain physical events have irreducibly

objective terms much more than we can at pres­

mental properties, and perhaps some view de­

ent, and with much greater precision. The loose

scribable in this way is correct. But nothing of

intermodal analogies-for example, "Red is like

which we can now form a conception corresponds

the sound of a trumpet" -which crop up in dis­

to it; nor have we any idea what a theory would

cussions of this subject are of little use. That

be like that enabled us to conceive of it. 1 3

should be clear to anyone who has both heard a

Very little work has been done on the basic

trumpet and seen red. But structural features of

question (from which mention of the brain can be

perception might be more accessible to objective

entirely omitted) whether any sense can be made

description, even though something would be left

of experiences' having an objective character at

out. And concepts alternative to those we learn in

all. Does it make sense, in other words, to ask

the first person may enable us to arrive at a kind

what my experiences are really like, as opposed to

of understanding even of our own experience

how they appear to me? We cannot genuinely

which is denied us by the very ease of description

understand the hypothesis that their nature is

and lack of distance that subjective concepts

captured in a physical description unless we un­

afford.

derstand the more fundamental idea that they

Apart from its own interest, a phenomenology

an objective nature (or that objective pro­

that is in this sense objective may permit ques­

have

cesses can have a subjective nature). 1 4

tions about the physical 1 5 basis of experience to

I should like to close with a speculative pro­

assume a more intelligible form. Aspects of sub­

posal. It may be possible to approach the gap

jective experience that admitted this kind of ob­

between subjective and objective from another

jective description might be better candidates for

direction. Setting aside temporarily the relation

objective explanations of a more familiar sort.

between the mind and the brain, we can pursue a

But whether or not this guess is correct, it seems

more objective understanding of the mental in

unlikely that any physical theory of mind can be

its own right. At present we are completely un-

contemplated until more thOUght has been given

Thomas Nagel

526

Any intelligent extraterrestrial beings totally differ­

to the general problem of subjective and ob­ jective. Otherwise we cannot even pose the mind­ body problem without sidestepping it. 1 6

7.

Notes

by a form of sonar, using vocal clicks or taps of a cane.

ent from us.

8.

It may be easier than I suppose to transcend inter­

species barriers with the aid of the imagination. For ex­ ample, blind people are able to detect objects near them

\.

Examples are J. J. C. Smart,

tific Realism

(London,

1963);

Perhaps if one knew what that was like, one could by

Philosophy and Scien­

David

Argument for the Identity Theory,"

K. Lewis, "An

Journal of Philoso­

(I 966), reprinted with addenda in David M. Rosenthal, Materialism & the Mind-Body Problem (Englewood Cliffs, N.J., 1 971); Hilary Putnam, "Psy­ chological Predicates" in Capitan and Merrill, Art. Mind. & Religion (Pittsburgh, 1967), reprinted in Rosenthal, op. cit. as "The Nature of Mental States"; D. M . Armstrong, A Materialist Theory of the Mind (London, 1968); D. C. Dennett, Content and Con­ sciousness (London, 1 969). I have expressed earlier doubt in "Armstrong on the Mind," Philosophical Re­ view LXXIX (1 970), 394-403; "Brain Bisection and the Unity of Consciousness," Synthese, 22 (1971); and a review of Dennett, Journal of Philosophy, LXIX (I 972).

phy,

LXIII

.

See also Saul Kripke, "Naming and Necessity" in Da­ vidson and Harman, (Dordrecht,

Semantics of Natural Language 334-342; and M . T. Thorn­ Terms and Materialism," The Monist,

1 972),

ton, "Ostensive

esp. pp.

56 (1972). 2.

son would have experiences. But that, if true, is a fact which cannot be discovered merely be analyzing the concept of experience. It is not equivalent to that about which we are in­

corrigible, both because we are not incorrigible about experience and because experience is present in animals lacking language and thought, who have no beliefs at all about their experiences.

4.

The Review of Metaphysics,

XIX

(1965), esp. 37-38. By "our own

tween oneself and other persons and other species can fall anywhere on a continuum. Even for other persons the understanding of what it is like to be them is only partial, and when one moves to species very different from oneself, a lesser degree of partial understanding may stilI be available. The imagination is remarkably flexible. My point, however, is not that we cannot what it is like to be a bat.

know

I am not raising that epis­

temological problem. My point is rather that even to form a

conception

of what it is like to be a bat (and a

fortiori to know what it is like to be a bat) one must take up the bat's point of view. If one can take it up roughly, or partially, then one's conception will also be rough or partial . Or so it

seems

in our present state of under­

standing.

9.

The problem I am going to raise can therefore be

posed even if the distinction between more subjective and more objective descriptions or viewpoints can itself

case"

not accept this kind of conceptual relativism, but it need not be refuted to make the point that psychophysical reduction cannot be accommodated by the subjective­ to-objective model familiar from other cases.

10. The problem is not just that when I look at the 'Mona Lisa,' my visual experience has a certain quality, no trace of which is to be found by someone looking into my brain. For even if he did observe there a tiny image of the

'Mona Lisa,'

he would have no reason to

identify it with the experience.

Cf. Richard Rorty, "Mind-Body Identity, Privacy,

and Categories,"

5.

the much more refined sonar of a bat. The distance be­

be made only within a larger human point of view. I do

Perhaps there could not actually be such robots.

Perhaps anything complex enough to behave like a per­

3.

extension imagine roughly what it was like to possess

I \ . The relation would therefore not be a contingent one, like that of a cause and its distinct effect. It would be necessarily true that a certain physical state felt a

I do not mean just "my own

case," but rather the mentalistic ideas that we apply

certain way. Saul 'Kripke

(op. cit.)

argues that causal

behaviorist and related analyses of the mental fail be­

unproblematically to ourselves and other human beings.

cause they construe, e.g., "pain" as a merely contingent

Therefore the analogical form of the English ex­

name of pains. The subjective character of an experience

6.

pression "what it is like" is misleading. It does not mean "what (in our experience) it resembles," but rather "how it is for the subject himself."

("its immediate phenomenological quality" calls it

Kripke

[po 340)) is the essential property left out by such

analyses, and the one in virtue of which it is, necessarily,

What Is It Like to Be a Bat?

527

the experience it is. My view is closely related to his.

14. This question also lies at the heart of the problem of

Like Kripke, I find the hypothesis that a certain brain

other minds, whose close connection with the mind­

necessarily

have a certain subjective char­

body problem is often overlooked. If one understood

acter incomprehensible without further explanation. No

how subjective experience could have an objective na­

such explanation emerges from theories which view the

ture, one would understand the existence of subjects

mind-brain relation as contingent, but perhaps there are

other than oneself.

state should

other alternatives, not yet discovered.

1 5 . I have not defined the term "physical. " Obviously

A theory that explained how the mind-brain relation

it does not apply just to what can be described by the

was necessary would still leave us with Kripke's prob­

concepts of contemporary physics, since we expect fur­

lem of explaining why it nevertheless appears con­

ther developments. Some may think there is nothing to

tingent. That difficulty seems to me surmountable, in

prevent mental phenomena from eventually being rec­

the following way. We may imagine something by

ognized as physical in their own right. But whatever else

representing it to ourselves either perceptually , sym­

may be said of the physical, it has to be objective. So if

pathetically, or symbolically. I shall not try to say how symbolic imagination works, but part of what happens in the other two cases is this. To imagine something

our idea of the physical ever expands to include mental phenomena, it will have to assign them an objective character-whether or not this is done by analyzing

perceptually, we put ourselves in a conscious state re­

them in terms of other phenomena already regarded as

sembling the state we would be in if we perceived it. To

physical. It seems to me more likely, however, that

imagine something sympathetically, we put ourselves

mental-physical relations will eventually be expressed in

in a conscious state resembling the thing itself. (This

a theory whose fundamental terms cannot be placed

method can be used only to imagine mental events and

clearly in either category.

states-our own or another's.) When we try to imagine a mental state occurring without its associated brain state, we first sympathetically imagine the occurrence of the mental state: that is, we put ourselves into a state that resembles it mentally. At the same time, we attempt to perceptually imagine the non-occurrence of the asso­ ciated physical state, by putting ourselves into another state unconnected with the first: one resembling that which we would be in if we perceived the non-occur­ rence of the physical state. Where the imagination of physical features is perceptual and the imagination of mental features is sympathetic, it appears to us that we can imagine any experience occurring without its asso­ ciated brain state, and vice versa. The relation between them will appear contingent even if it is necessary, be­ cause of the independence of the disparate types of imagination. (Solipsism, incidentally, results if one misinterprets sympathetic imagination as if it worked like perceptual imagination: it then seems impossible to imagine any experience that is not one's own.) 12. See "Mental Events" in Foster and Swanson,

perience and Theory

(Amherst, 1 970); though

Ex­

I don't

understand the argument against psychophysical laws. 1 3 . Similar remarks apply to my paper "Physicalism,"

Philosophical Review

LXXIV ( 1 965),

339-356,

printed with postscript in John O'Connor,

terialism (New York,

1 969).

re­

Modern Ma­

16. I have read versions of this paper to a number of audiences, and am indebted to many people for their comments.

33

Can We Solve the Mind-Body Problem?

Colin McGinn "How it is that anything so remarkable as a state of con­ sciousness comes about as a result of initiating nerve tis­ sue, is just as unaccountable as the appearance of the Djin, where Aladdin rubbed his lamp in the story " . . .

-Thomas Huxley

We have been trying for a long time to solve the mind-body problem. It has stubbornly resisted our best efforts. The mystery persists. I think the time has come to admit candidly that we cannot resolve the mystery. But I also think that this very insolubility-or the reason for it-removes the philosophical problem. In this chapter I explain why I say these outrageous things. The specific problem I want to discuss concerns consciousness, the hard nut of the mind-body problem. How is it possible for conscious states to depend upon brain states? How can technicolor phenomenology arise from soggy gray matter? What makes the bodily organ we call the brain so radically different from other bodily organs, say the kidneys-the body parts without a trace of consciousness? How could the aggregation of millions of individually insentient neurons gen­ erate subjective awareness? We know that brains are the de facto causal basis of consciousness, but we have, it seems, no understanding whatever of how this can be so. It strikes us as miraculous, eerie, even faintly comic. Somehow, we feel, the water of the physical brain is turned into the wine of consciousness, but we draw a total blank on the nature of this conversion. Neural transmissions just seem like the wrong kind of materials with which to bring consciousness into the world, but it appears that in some way they perform this mysterious feat. The mind-body problem is the problem of understanding how the miracle is wrought, thus removing the sense of deep mys­ tery. We want to take the magic out of the link between consciousness and the brain. 1 Purported solutions to the problem have tended to assume one of two forms. One from, which we may call constructive, attempts to

specify some natural property of the brain (or body) which explains how consciousness can be elicited from it. Thus functionalism, for example, suggests a property-namely, causal role-whh... is held to be satisfied by both brain states and mental states; this property is supposed to explain how conscious states can come from brain states. 2 The other form, which has been historically domi­ nant, frankly admits that nothing merely natural could do the job, and suggests instead that we in­ voke supernatural entities or divine interventions. Thus we have Cartesian dualism and Leibnizian pre-established harmony. These "solutions" at least recognize that something pretty remarkable is needed if the mind-body relation is to be made sense of; they are as extreme as the problem. The approach I favor is naturalistic but not construc­ tive: I do not believe we can ever specify what it is about the brain that is responsible for conscious­ ness, but I am sure that whatever it is it is not in­ herently miraculous. The problem arises, I want to suggest, because we are cut offby our very cog­ nitive constitution from achieving a conception of that natural property of the brain (or of con­ sciousness) that accounts for the psychophysical link. This is a kind of causal nexus that we are precluded from ever understanding, given the way we have to form our concepts and develop our theories. No wonder we find the problem so difficult! Before I can hope to make this view plausible, I need to sketch the general conception of cognitive competence that underlies my position. Let me introduce the idea of cognitive closure. A type of mind M is cognitively closed with respect to a property P (or theory T) if and only if the con­ cept-forming procedures at M's disposal cannot extend to a grasp of P (or an understanding of T). Conceiving minds come in different kinds, equip­ ped with varying powers and limitations, biases and blindspots, so that properties (or theories) may be accessible to some minds but not to others.

Colin McGinn

530

What is closed to the mind of a rat may be open

sions";

to the mind of a monkey, and what is open to us

cannot transcend what can be perceptually pre­

therefore the concept-forming system

may be closed to the monkey. Representational

sented to the subject. Such a mind will be closed

power is not all or nothing. Minds are biological

with respect to unobservables; the properties of

products like bodies, and like bodies they come in

atoms, say, will not be representable by a mind

different shapes and sizes, more or less capacious,

constructed in this way. This implies that ex­

more or less suited to certain cognitive tasks.3

planatory theories in which these properties are

This is particularly clear for perceptual faculties,

essentially mentioned will not be accessible to a

of course: perceptual closure is hardly to be de­

Humean mind.5 And hence the observable phe­

nied. Different species are capable of perceiving

nomena that are explained by allusion to un­

different properties of the world, and no species

observables will be inexplicable by a mind thus

can perceive every property things may instantiate

limited. But notice: the incapacity to explain cer­

(without artificial instrumentation anyway). But

tain phenomena does not carry with it a lack of

such closure does not reflect adversely on the re­

recognition of the theoretical problems the phe­

ality of the properties that lie outside the repre­

nomena pose. You might be able to appreciate

sentational capacities in question; a property is no

problem without being able to formulate (even in

less real for not being reachable from a certain

principle) the solution to that problem (I suppose

kind of perceiving and conceiving mind. The

human children are often in this position, at least

invisible parts of the electromagnetic spectrum

for a while). A Humean mind cannot solve the

are j ust as real as the visible parts, and whether a

problems that our physics solves, yet it might be

specific kind of creature can form conceptual

able to have an inkling of what needs to be

representations of these imperceptible parts does

explained. We would expect, then, that a mod­

not determine whether they exist. Thus cognitive

erately intelligent enquiring Humean mind will

closure with respect to

feel permanently perplexed any mystified by the

about

P does not imply irrealism noumenal with respect to M does not show that P does not occur in some naturalistic scientific theory T-it shows only that T is not cognitively accessible to

physical world, since the correct science is forever

P.

Thilt

P

is (as we might say)

beyond its cognitive reach. Indeed, something like this was precisely the view of Locke. He thought that our ideas of matter are quite sharply con­

M. Presumably monkey minds and the property

strained by our perceptions and so concluded that

of being an electron illustrate this possibility. And

the true science of matter is eternally beyond us­

the question must arise as to whether human

that we could never remove our perplexities about

minds are closed with respect to certain true ex­

(say) what solidity ultimately is.6 But it does not

planatory theories. Nothing, at least, in the con­

follow for Locke that nature is itself inherently

cept of reality shows that everything real is open

mysterious; the felt mystery comes from our own

to the human concept-forming faculty-if, that

cognitive limitations, not from any objective eeri­

is, we are realists about reality.4

ness in the world. It looks today as if Locke was

Consider a mind constructed according to

wrong about our capacity to fathom the nature

the principles of classical empiricism, a Humean

of the physical world, but we can still learn from

mind. Hume mistakenly thought that human

his fundamental thought-the insistence that our

minds were Humean, but we can at least conceive

cognitive faculties may not be up to solving

of such a mind (perhaps dogs and monkeys have

every problem that confronts us. To put the point

Humean minds). A Humean mind is such that

more generally: the human mind may not con­

perceptual closure determines cognitive closure,

form to empiricist principles, but it must conform

since "ideas" must always be copies of "impres-

to

some principles-and

it is a substantive claim

Can We Solve the Mind-Body Problem?

that these principles permit the solution of every problem we can formulate or sense. Total cogni­ tive openness is not guaranteed for human beings and it should not be expected. Yet what is nou­ menal for us may not be miraculous in itself. We should therefore be alert to the possibility that a problem that strikes us as deeply intractable, as utterly batHing, may arise from an area of cogni­ tive closure in our ways of representing the world. 7 That is what I now want to argue is the case with our sense of the mysterious nature of the con­ nection between consciousness and the brain. We are biased away from arriving at the correct explanatory theory of the psychophysical nexus. And this makes us prone to an illusion of ob­ jective mystery. Appreciating this should remove the philosophical problem: consciousness does not, in reality, arise from the brain in the miracu­ lous way in which the Djin arises from the lamp. I now need to establish three things: (i) there exists some property of the brain that accounts naturalistically for consciousness; (ii) we are cog­ nitively closed with respect to that property; but (iii) there is no philosophical (as opposed to sci­ entific) mind-body problem. Most of the work will go into establishing (ii). Resolutely shunning the supernatural, I think it is undeniable that it must be in virtue of some natural property of the brain that organisms are conscious. There just has to be some explanation for how brains subserve minds. Ifwe are not to be eliminativists about consciousness, then some theory must exist which accounts for the psycho­ physical correlations we observe. It is implausible to take these correlations as ultimate and inex­ plicable facts, as simply brute. And we do not want to acknowledge radical emergence of the conscious with respect to the cerebral: that is too much like accepting miracles de reo Brain states cause conscious states, we know, and this causal nexus must proceed through necessary connec­ tions of some kind-the kind that would make the nexus intelligible if they were understood. 8 Consciousness is like life in this respect. We know that life evolved from inorganic matter, so we ex-

531

pect there to be some explanation o f this process. We cannot plausibly take the arrival of life as a primitive brute fact, nor can we accept that life arose by some form of miraculous emergence. Rather, there must be some natural account of how life comes from matter, whether or not we can know it. Eschewing vitalism and the magic touch of God's finger, we rightly insist that it must be in virtue of some natural property of (organized) matter that parcels of it get to be alive. But consciousness itself is just a further bi­ ological development, and so it too must be sus­ ceptible of some natural explanation-whether or not human beings are capable of arriving at this explanation. Presumably there exist objective natural laws that somehow account for the up­ surge of consciousness. Consciousness, in short, must be a natural phenomenon, naturally arising from certain organizations of matter. Let us then say that there exists some property P, instantiated by the brain, in virtue of which the brain is the basis of consciousness. Equivalently, there exists some theo.ry T, referring to P, which fully ex­ plains the dependence of conscious states on brain states. If we knew T, then we would have a constructive solution to the mind-body problem. The question then is whether we can ever come to know T and grasp the nature of P. Let me first observe that it is surely possible that we could never arrive at a grasp of P; there is, as I said, no guarantee that our cognitive powers per­ mit the solution of every problem we can recog­ nize. Only a misplaced idealism about the natural world could warrant the dogmatic claim that everything is knowable by the human species at this stage of its evolutionary development (con­ sider the same claim made on behalf of the in­ tellect of Cro-Magnon man). It may be that every property for which we can form a concept is such that it could never solve the mind-body problem. We could be like five-year-old children trying to understand Relativity Theory. Still, so far this is just a possibility claim: what reason do we have for asserting, positively, that our minds are closed with respect to P?

Colin McGinn

Longstanding historical failure is suggestive, but scarcely conclusive. Maybe, it will be said, the solution is just around the corner, or it has to wait upon the completion of the physical sciences. Perhaps we simply have yet to produce the Ein­ stein-like genius who will restructure the problem in some clever way and then present an aston­ ished world with the solution. 9 However, I think that our deep bafflement about the problem, amounting to a vertiginous sense of ultimate mystery, which resists even articulate formula­ tion, should at least encourage us to explore the idea that there is something terminal about our perplexity. Rather as traditional theologians found themselves conceding cognitive closure with respect to certain of the properties of God, so we should look seriously at the idea that the mind-body problem brings us bang up against the limits of our capacity to understand the world. That is what I shall do now. There seem to be two possible avenues open to us in our aspiration to identify P: we could try to get to P by investigating consciousness directly; or we could look to the study of the brain for P. Let us consider these in turn, starting with con­ sciousness. Our acquaintance with consciousness could hardly be more direct; phenomenological description thus comes (relatively) easily. "In­ trospection" is the name of the faculty through which we catch consciousness in all its vivid nakedness. By virtue of possessing this cognitive faculty we ascribe concepts of consciousness to ourselves; we thus have "immediate access" to the properties of consciousness. But does the in­ trospective faculty reveal property P? Can we tell just by introspecting what the solution to the mind-body problem is? Clearly not. We have di­ rect cognitive access to one term of the mind­ brain relation, but we do not have such access to the nature of the link. Introspection does not present conscious states as depending upon the brain in some intelligible way. We cannot there­ fore introspect P. Moreover, it seems impossible that we should ever augment our stock of in­ trospectively ascribed concepts with the concept

532

P-that is, we could not acquire this concept simply on the basis of sustained and careful introspection. Pure phenomenology wiII never provide the solution to the mind-body problem. Neither does it seem feasible to try to extract P from the concepts of consciousness we now have by some procedure of conceptual analysis-any more than we could solve the life-matter problem simply by reflecting on the concept life. 1 0 P has to lie outside the field of the introspectable, and it is not implicitly contained in the concepts we bring to bear in our first-person ascriptions. Thus the faculty of introspection, as a concept-forming capacity, is cognitively closed with respect to P; which is not surprising in view of its highly limited domain of operation (most properties of the world are closed to introspection). But there is a further point to be made about P and consciousness, which concerns our restricted access to the concepts of consciousness them­ selves. It is a familiar point that the range of con­ cepts of consciousness attainable by a mind M is constrained by the specific forms of consciousness possessed by M. Crudely, you cannot form con­ cepts of conscious properties unless you yourself instantiate those properties. The man born blind cannot grasp the concept of a visual experience of red, and human beings cannot conceive of the echolocatory experiences of bats. I I These are cases of cognitive closure within the class of con­ scious properties. But now this kind of closure will, it seems, affect our hopes of access to P. For suppose that we were cognitively open with respect to P; suppose, that is, that we had the solution to the problem of how specific forms of consciousness depend upon different kinds of physiological structure. Then, of course, we would understand how the brain of a bat sub­ serves the subjective experiences of bats. Call this type of experience B, and call the explanatory property that links B to the bat's brain PI . By grasping PI it would be perfectly intelligible to us how the bat's brain generates B-experiences; we would have an explanatory theory of the causal nexus in question. We would be in possession of

Can We Solve the Mind-Body Problem?

the same kind of understanding we would have of our own experiences if we had the correct psy­ chophysical theory of them. But then it seems to follow that grasp of the theory that explains B­ experiences would confer a grasp of the nature of those experiences: for how could we understand that theory without understanding the concept B that occurs in it? How could we grasp the nature of B-experiences without grasping the character of those experiences? The true psychophysical theory would seem to provide a route to a grasp of the subjective form of the bat's experiences. But now we face a dilemma, a dilemma which threatens to become a reduction: either we can grasp this theory, in which case the property B becomes open to us; or we cannot grasp the theory, simply because property B is not open to us. It seems to me that the looming reduction here is compelling: our concepts of consciousness just are inherently constrained by our own from of consciousness, so that any theory the under­ standing of which required us to transcend these constraints would ipso facto be inaccessible to us. Similarly, I think, any theory that required us to transcend the finiteness of our cognitive capacities would ipsofacto be a theory we could not grasp­ and this despite the fact that it might be needed to explain something we can see needs explaining. We cannot simply stipulate that our concept­ forming abilities are indefinitely plastic and unlimited just because they would have to be to enable us to grasp the truth about the world. We constitutionally lack the concept-forming capacity to encompass all possible types of conscious state, and this obstructs our path to a general solution to the mind-body problem. Even if we could solve it for our own case, we could not solve it for bats and Martians. P is, as it were, too close to the different forms of sUbjectivity for it to be accessible to all such forms, given that one's form of subjectivity restricts one's concepts of subjectivity Y I suspect that most optimists about construc­ tively solving the mind-body problem will prefer to place their bets on the brain side of the relation.

533

Neuroscience is the place to look for property P, they will say. My question then is whether there is any conceivable way in which we might come to introduce P in the course of our empirical inves­ tigations of the brain. New concepts have been introduced in the effort to understand the work­ ings of the brain, certainly: could not P then occur in conceivable extensions of this manner of in­ troduction? So far, indeed, the theoretical con­ cepts we ascribe to the brain seem as remote from consciousness as any ordinary physical properties are, but perhaps we might reach P by diligent application of essentially the same procedures: so it is tempting to think. I want to suggest, to the contrary, that such procedures are inherently closed with respect to P. The fundamental reason for this, I think, is the role ofperception in shap­ ing our understanding of the brain-the way that our perception of the brain constrains the con­ cepts we can apply to it. A point whose signif­ icance it would be hard to overstress here is this: the property of consciousness itself (or specific conscious states) is not an observable or percep­ tible property of the brain. You can stare into a living conscious brain, your own or someone else's, and see there a wide variety of instantiated properties-its shape, color, texture, and so on­ but you will not thereby see what the subject is experiencing, the conscious state itself. Conscious states are simply not, qua conscious states, po­ tential objects of perception: they depend upon the brain but they cannot be observed by direct­ ing the senses onto the brain. You cannot see a brain state as a conscious state. In other words, consciounsess is noumenal with respect to per­ ception of the brain. 1 J I take it this is obvious. So we know there are properties of the brain that are necessarily closed to perception of the brain; the question now is whether P is likewise closed to perception. My argument will proceed as follows. I shall first argue that P is indeed perceptually closed; then I shall complete the argument to full cogni­ tive closure by insisting that no form of inference from what is perceived can lead us to P. The

Colin McGinn

534

argument for perceptual closure starts from the

consider quantum theory and cosmology. Unre­

thought that nothing we can imagine perceiving

stricted perceptual openness is a dogma of em­

in the brain would ever convince us that we have

piricism if ever there was one. And there is no

located the intelligible nexus we seek. No matter

compelling reason to suppose that the property

what recondite property we could see to be in­

needed to explain the mind-brain relation should

stantiated in the brain we should always be baf­

be in principle perceptible; it might be essentially

fled about how it could give rise to consciousness.

"theoretical", an object of thought, not sen­

I hereby invite you to try to conceive of a percep­

sory experience. Looking harder at nature is

tible property of the brain that might allay the

not the only (or the best) way of discovering its

feeling of mystery that attends our contemplation

theoretically significant properties.

of the brain-mind link: I do not think you will be

closure does not entail cognitive closure, since

Perceptual

able to do it. It is like trying to conceive of a per­

we have available the procedure of hypothesis

ceptible property of a rock that would render it

formation, in which unobservables come to be

perspicuous that the rock was conscious. In fact, I

conceptualized.

think it is the very impossibility of this that lies at

I readily agree with these sentiments, but I

the root of the felt mind-body problem. But why

think there are reasons for believing that no co­

is this? Basically, I think, it is because the senses

herent method of concept introduction will ever

P. This is because a certain principle of

are geared to representing a spatial world; they

lead us to

essentially present things in space with spatially

homogeneity

defined properties. But it is precisely such proper­

oretical concepts on the basis of observation. Let

operates in our introduction of the­

ties that seem inherently incapable of resolving

me first note that consciousness itself could not be

the mind-body problem: we cannot link con­

introduced simply on the basis of what we ob­

sciousness to the brain in virtue of spatial prop­

serve about the brain and its physical effects. If

erties of the brain. There the brain is, an object of

our data, arrived at by perception of the brain, do

perception, laid out in space, containing spatially

not include anything that brings in conscious

distributed processes; but consciousness defies

states, then the theoretical properties we need

explanation in such terms. Consciousness does

to explain these data will not include conscious

not seem made up out of smaller spatial pro­

states either. Inference to the best explanation of

cesses; yet perception of the brain seems limited

purely physical data will never take us outside the

to revealing such processes. 1 4 The senses are

realm of the physical, forcing us to introduce

of properties-those

concepts of consciousness. 1 5 Everything physical

that are essentially bound up with space--but

has a purely physical explanation. So the prop­

these properties are of the wrong sort (the wrong

erty of consciousness is cognitively closed with

P. Kant was right, the

respect to the introduction of concepts by means

responsive to certain

category)

kinds

to constitute

form of outer sensibility is spatial; but if so, then

of inference to the best explanation of perceptual

P will be noumenal with respect to the senses,

data about the brain.

since no spatial property will ever deliver a sat­

Now the question is whether

P could ever be

isfying answer to the mind-body problem. We

arrived at by this kind of inference. Here we must

simply do not understand the idea that conscious

be careful to guard against a form of magical

states might intelligibly arise from spatial config­

emergentism with respect to concept formation.

urations of the kind disclosed by perception of the

Suppose we try out a relatively clear theory of

world.

how theoretical concepts are formed: we get them

I take it this claim will not seem terribly con­

by a sort of analogical extension of what we ob­

troversial. After all, we do not generally expect

serve. Thus, for example, we arrive at the concept

that every property referred to in our theories

of a molecule by taking our perceptual repre­

should be a potential object of human perception:

sentations of macroscopic objects and conceiving

Can We Solve the Mind-Body Problem?

of smaller scale objects of the same general kind. This method seems to work well enough for un­ observable material objects, but it will not help in arriving at P, since analogical extensions of the entities we observe in the brain are precisely as hopeless as the original entities were as solutions to the mind-body problem. We would need a method that left the base of observational prop­ erties behind in a much more radical way. But it seems to me that even a more unconstrained con­ ception of inference to the best explanation would still not do what is required: it would no more serve to introduce P than it serves to introduce the property of consciousness itself. To explain the observed physical data we need only such theo­ retical properties as bear upon those data, not the property that explains consciousness, which does not occur in the data. Since we do not need con­ sciousness to explain those data, we do not need the property that explains consciousness. We shall never get as far away from the perceptual data in our explanations of those data as we need to get in order to connect up explanatorily with consciousness. This is, indeed, why it seems that consciousness is theoretically epiphenomenal in the task of accounting for physical events. No concept needed to explain the workings of the physical world will suffice to explain how the physical world produces consciousness. So if P is perceptually nournenal, then it will be nournenal with respect to perception-based explanatory in­ ferences. Accordingly, I do not think that P could be arrived at by empirical studies of the brain alone. Nevertheless, the brain has this property, as it has the property of consciousenss. Only a magical idea of how we come by concepts could lead one to think that we can reach P by first perceiving the brain and then asking what is needed to explain what we perceive. 16 (The mind­ body problem tempts us to magic in more ways than one.) It will help elucidate the position I am driving toward if I contrast it with another view of the source of the perplexity we feel about the mind­ brain nexus. I have argued that we cannot know which property of the brain accounts for con-

535

sciousness, and so we find the mind-brain link unintelligible. But, it may be said, there is another account of our sense of irremediable mystery, which does not require positing properties our minds cannot represent. This alternative view claims that, even if we now had a grasp of P, we would still feel that there is something mysterious about the link, because of a special epistemo­ logical feature of the situation. Namely this: our acquaintance with the brain and our acquaint­ ance with consciousness are necessarily mediated by distinct cognitive faculties, namely perception and introspection. Thus the faculty through which we apprehend one term of the relation is necessarily distinct from the faculty through which we apprehend the other. In consequence, it is not possible for us to use one of these faculties to apprehend the nature of the psychophysical nexus. No single faculty will enable us ever to apprehend the fact that consciousness depends upon the brain in virtue of property P. Neither perception alone nor introspection alone will ever enable us to witness the dependence. And this, my objector insists, is the real reason we find the link baffling: we cannot make sense of it in terms of the deliverances of a single cognitive faculty. So, even if we now had concepts for the properties of the brain that explain consciousness, we would still feel a residual sense of unintelligibility; we should still take there to be something mysterious going on. The necessity to shift from one faculty to the other produces in us an illusion of inex­ plicability. We might in fact have the explanation right now but be under the illusion that we do not. The right diagnosis, then, is that we should rec­ ognize the peculiarity of the epistemological situation and stop trying to make sense of the psychophysical nexus in the way we make sense of other sorts of nexus. It only seems to us that we can never discover a property that will render the nexus intelligible. I think this line of thought deserves to be taken seriously, but I doubt that it correctly diagnoses our predicament. It is true enough that the prob­ lematic nexus is essentially apprehended by distinct faculties, so that it will never reveal its

Colin McGinn

secrets to a single faculty; but I doubt that our intuitive sense of intelligibility is so rigidly gov­ erned by the "single-faculty condition". Why should facts only seem intelligible to us if we can conceive of apprehending them by one (sort of) cognitive faculty? Why not allow that we can recognize intelligible connections between con­ cepts (or properties) even when those concepts (or properties) are necessarily ascribed using different faculties? Is it not suspiciously empiricist to insist that a causal nexus can only be made sense of by us if we can conceive of its being an object of a single faculty of apprehension? Would we think: this of a nexus that called for touch and sight to apprehend each term of the relation? Suppose (per impossible) that we were offered P on a plate, as a gift from God: would we still shake our heads and wonder how that could resolve the mystery, being still the victims of the illusion of mystery generated by the epistemological duality in ques­ tion? No, I think this suggestion is not enough to accounts for the miraculous appearance of the link:: it is better to suppose that we are perma­ nently blocked from forming a concept of what accounts for that link:. How strong is the thesis I am urging? Let me distinguish absolute from relative claims of cog­ nitive closure. A problem is absolutely cognitively closed if no possible mind could resolve it; a problem is relatively closed if minds of some sorts can in principle solve it while minds of other sorts cannot. Most problems, we may safely suppose, are only relatively closed: armadillo minds can­ not solve problems of elementary arithmetic but human minds can. Should we say that the mind­ body problem is only relatively closed or is the closure absolute? This depends on what we allow as a possible concept-forming mind, which is not an easy question. Ifwe allow for minds that form their concepts of the brain and consciousness in ways that are quite independent of perception and introspection, then there may be room for the idea that there are possible minds for which the mind-body problem is soluble, and easily so. But if we suppose that all concept formation is tied to

536

perception and introspection, however loosely, then no mind will be capable of understanding how it relates to its own body-the insolubility will be absolute. I think we can just about make sense of the former kind of mind, by exploiting our own faculty of a priori reasoning. Our mathe­ matical concepts (say) do not seem tied either to perception or to introspection, so there does seem to be a mode of concept formation that operates without the constraints I identified earlier. The suggestion might then be that a mind that formed all of its concepts in this way-including its con­ cepts of the brain and consciousness-would be free of the biases that prevent us from coming up with the right theory of how the two connect. Such a mind would have to be able to think: of the brain and consciousness in ways that utterly prescind from the perceptual and the introspec­ tive-in somewhat the way we now (it seems) think: about numbers. This mind would conceive of the psychophysical link in totally a priori terms. Perhaps this is how we should think: of God's mind, and God's understanding of the mind-body relation. At any rate, something pretty radical is going to be needed if we are to devise a mind that can escape the kinds of closure that make the problem insoluble for us-if I am right in my diagnosis of our difficulty. If the problem is only relatively insoluble, then the type of mind that can solve it is going to be very dif­ ferent from ours and the kinds of mind we can readily make sense of (there may, of course, be cognitive closure here too). It certainly seems to me to be at least an open question whether the problem is absolutely insoluble; I would not be surprised if it were. 1 7 My position is both pessimistic and optimistic at the same time. It is pessimistic about the pros­ pects for arriving at a constructive solution to the mind-body problem, but it is optimistic about our hopes of removing the philosophical perplexity. The central point here is that I do not think: we need to do the former in order to achieve the latter. This depends on a rather special under­ standing of what the philosophical problem con-

Can We Solve the Mind-Body Problem?

537

sists in. What I want to suggest is that the nature

sciousness. But we do not need to accept this:

of the psychophysical connection has a full and

we can rest secure in the knowledge that some

non-mysterious explanation in a certain science,

(unknowable) property of the brain makes every­

but that this science is inaccessible to us as a

thing fall into place. What creates the philosoph­

matter of principle. Call this explanatory scien­

ical puzzlement is the assumption that the problem

tific theory

must somehow be scientific but that any science

T: T

is as natural and prosaic and

devoid of miracle as any theory of nature; it de­

we can

scribes the link between consciousness and the

terly miraculous. And the solution is to recognize

come up with will represent things as ut­

brain in a way that is no more remarkable (or

that the sense of miracle comes from us and not

alarming) than the way we now describe the link

from the world. There is, in reality, nothing mys­

between the liver and bile. 1 8 According to

T,

there is nothing eerie going on in the world when

terious about how the brain generates conscious­ metaphysical problem. 2 0

ness. There is no

an event in my visual cortex causes me to have

So far that deflationary claim has been justified

an experience of yellow- -however much it seems

by a general naturalism and certain considera­

to

that there is. In other words, there is no

tions about cognitive closure and the illusions it

intrinsic conceptual or metaphysical difficulty

can give rise to. Now I want to marshall some

about how consciousness depends on the brain. It

reasons for thinking that consciousness is actually

is not that the correct science is compelled to

a rather simple natural fact; objectively, con­

us

postulate miracles

it is rather that the cor­

sciousness is nothing very special. We should now

rect science lies in the dark part of the world for

be comfortable with the idea that our own sense

us. We confuse our own cognitive limitations with

of difficulty is a fallible guide to objective com­

de re;

objective eeriness. We are like a Humean mind

plexity: what is hard for us to grasp may not be

trying to understand the physical world, or a creature without spatial concepts trying to un­

very fancy in itself. The grain of our thinking is not a mirror held up to the facts of nature. 2 1 In

derstand the possibility of motion. This removes

particular, it may be that the extc:nt of our un­

the philosophical problem because it assures

derstanding of facts about the mind is not com­

us that the entities

pose no inherent

mensurate with some objective estimate of their

philosophical difficulty. The case is unlike, for

intrinsic complexity: we may be good at under­

example, the problem of how the abstract world

standing the mind in some of its aspects but

themselves

of numbers might be intelligibly related to the

hopeless with respect to others, in a way that cuts

world of concrete knowing subjects: here the

across objective differences in what the aspects

mystery seems intrinsic to the entities, not a mere

involve. Thus we are adept at understanding ac­

artefact of our cognitive limitations or biases

tion in terms of the folk psychology of belief and

in trying to understand the relation. 1 9 It would

desire, and we seem not entirely out of our depth

not be plausible to suggest that there exists a sci­

when it comes to devising theories of language.

ence, whose theoretical concepts we cannot grasp,

But our understanding of how consciousness

which completely resolves any sense of mystery

develops from the organization of matter is non­

that surrounds the question how the abstract

existent. But now, think of these various aspects

becomes an object of knowledge for us. In this

of mind from the point of view of evolutionary

case, then, eliminativism seems a live option. The

biology. Surely language and the propositional

problem about consciousness and

attitudes are more complex and advanced evolu­

philosophical

the brain arises from a sense that we are com­

tionary achievements than the mere possession of

pelled to accept that nature contains miracles-as

consciousness by a physical organism. Thus it

of the brain

seems that we are better at understanding some

could really spirit into existence the Djin of con-

if the merely metallic lamp

of the more complex aspects of mind than the

(lump)

Colin McGinn

simpler ones. Consciousness arises early in evo­ lutionary history and is found right across the animal kingdom. In some respects it seems that the biological engineering required for conscious­ ness is less fancy than that needed for certain kinds of complex motor behavior. Yet we can come to understand the latter while drawing a total blank with respect to the former. Conscious states seem biologically quite primitive, com­ paratively speaking. So the theory T that explains the occurrence of consciousness in a physical world is very probably less objectively complex (by some standard) than a range of other theories that do not defy our intellects. If only we could know the psychophysical mechanism it might surprise us with its simplicity, its utter natural­ ness. In the manual that God consulted when He made the earth and all the beasts that dwell thereon the chapter about how to engineer con­ sciousness from matter occurs fairly early on, well before the really difficult later chapters on mam­ malian reproduction and speech. It is not the size of the problem but its type that makes the mind­ body problem so hard for us. This reflection should make us receptive to the idea that it is something about the tracks of our thought that prevents us from achieving a science that relates consciousness to its physical basis: the enemy lies within the gates.22 The position I have reached has implications for a tangle of intuitions it is natural to have re­ garding the mind-body relation. On the one hand, there are intuitions, pressed from Descartes to Kripke, to the effect that the relation between conscious states and bodily states is funda­ mentally contingent. 23 It can easily seem to us that there is no necessitation involved in the de­ pendence of the mind on the brain. But, on the other hand, it looks absurd to try to dissociate the two entirely, to let the mind float completely free of the body. Disembodiment is a dubious possi­ bility at best, and some kind of necessary super­ venience of the mental on the physical has seemed undeniable to many. It is not my aim here to ad­ judicate this longstanding dispute; I want simply

538

to offer a diagnosis of what is going on when one finds oneself assailed with this flurry of conflicting intuitions. The reason we feel the tug of contin­ gency, pulling consciousness loose from its physi­ cal moorings, may be that we do not and cannot grasp the nature of the property that intelligibly links them. The brain has physical properties we can grasp, and variations in these correlate with changes in consciousness, but we cannot draw the veil that conceals the manner of their connection. Not grasping the nature of the connection, it strikes us as deeply contingent; we cannot make the assertion of a necessary connection intelligible to ourselves. There may then be a real necessary connection; it is just that it will always strike us as curiously brute and unperspicuous. We may thus, as upholders of intrinsic contingency, be the dupes of our own cognitive blindness. On the other hand, we are scarcely in a position to assert that there is a necessary connection between the properties of the brain we can grasp and states of consciousness, since we are so ignorant (and irre­ mediably so) about the character of the con­ nection. For all we know, the connection may be contingent, as access to P would reveal if we could have such access. The link between con­ sciousness and property P is not, to be sure, con­ tingent-virtually by definition-but we are not in a position to say exactly how P is related to the "ordinary" properties of the brain. It may be necessary or it may be contingent. Thus it is that we tend to vacillate between contingency and ne­ cessity; for we lack the conceptual resources to decide the question definitively-or to under­ stand the answer we are inclined to give. The indicated conclusion appears to be that we can never really know whether disembodiment is metaphysically possible, or whether necessary supervenience is the case, or whether spectrum inversion could occur. For these all involve claims about the modal connections between properties of consciousness and the ordinary properties of the body and brain that we can conceptualize; and the real nature of these con­ nections is not accessible to us. Perhaps P makes

Can We Solve the Mind-Body Problem?

the relation between C-fiber firing and pain nec­ essary or perhaps it does not: we are simply not equipped to know. We are like a Humean mind wondering whether the observed link between the temperature of a gas and its pressure (at a constant volume) is necessary or contingent. To know the answer to that you need to grasp atomic (or molecular) theory, and a Humean mind just is not up to attaining the requisite theoretical understanding. Similarly, we are constitutionally ignorant at precisely the spot where the answer exists. I predict that many readers of this chapter will find its main thesis utterly incredible, even ludi­ crous. Let me remark that I sympathize with such readers: the thesis is not easily digestible. But I would say this: if the thesis is actually true, it will still strike us as hard to believe. For the idea of an explanatory property (or set of properties) that is noumenal for us, yet is essential for the (con­ structive) solution of a problem we face, offends a kind of natural idealism that tends to dominate our thinking. We find it taxing to conceive of the existence of a real property, under our noses as it were, which we are built not to grasp-a property that is responsible for phenomena that we observe in the most direct way possible. This kind of re­ alism, which brings cognitive closure so close to home, is apt to seem both an affront to our in­ tellects and impossible to get our minds around. We try to think of this unthinkable property and understandably fail in the effort; so we rush to infer that the very supposition of such a property is nonsensical. Realism of the kind I am pre­ supposing thus seems difficult to hold in focus, and any philosophical theory that depends upon it will also seem to rest on something systemati­ cally elusive.24 My response to such misgivings, however, is unconcessive: the limits of our minds are just not the limits of reality. It is deplorably anthropocentric to insist that reality be con­ strained by what the human mind can conceive. We need to cultivate a vision of reality (a meta­ physics) that makes it truly independent of our given cognitive powers, a conception that in-

539

c1udes these powers as a proper part. It is just that, in the case of the mind-body problem, the bit of reality that systematically eludes our cognitive grasp is an aspect of our own nature. Indeed, it is an aspect that makes it possible for us to have minds at all and to think about how they are re­ lated to our bodies. This particular transcendent tract of reality happens to lie within our own heads. A deep fact about our own nature as a form of embodied consciousness is thus neces­ sarily hidden from us. Yet there is nothing in­ herently eerie or bizarre about this embodiment. We are much more straightforward than we seem. Our weirdness lies in the eye of the beholder. The answer to the question that forms my title ,, is therefore "No and Yes. 25

Notes Previously published in Mind, Vol. xcviii, no. 891 , July 1989, this article appears by permission of the pub­ lishers, Oxford University Press. I. One of the peculiarities of the mind-body problem is the difficulty of formulating it in a rigorous way. We have a sense of the problem that outruns our capacity to articulate it clearly. Thus we quickly find ourselves re­ sorting to invitations to look inward, instead of specify­ ing precisely what it is about consciousness that makes it inexplicable in terms of ordinary physical properties. And this can make it seem that the problem is spurious. A creature without consciousness would not properly appreciate the problem (assuming such a creature could appreciate other problems). I think an adequate treat­ ment of the mind-body problem should explain why it is so hard to state the problem explicitly. My treatment locates our difficulty in our inadequate conceptions of the nature of the brain and consciousness. In fact, if we knew their natures fully we would already have solved the problem. This should become clear later. 2. I would also classify panpsychism as a constructive solution, since it attempts to explain consciousness in terms of properties of the brain that are as natural as consciousness itself. Attributing specks of proto-con­ sciousness to the constituents of matter is not super­ natural in the way postulating immaterial substances or divine interventions is; it is merely extravagant. I shall here be assuming that panpsychism, like all other extant

Colin McGinn

constructive solutions, is inadequate as an · answer to the mind-body problem-as (of course) are the super­ natural "solutions." I am speaking to those who still feel perplexed (almost everyone, I would think, at least in their heart). 3. This kind of view of cognitive capacity is forcefully advocated by Noam Chomsky in Reflections on Lan­ guage, Pantheon Books, 1 975, and by Jerry Fodor in The Modularity of Mind, (The MIT Press: Cambridge, Mass., 1 983.) Chomsky distinguishes between "prob­ lems," which human minds are in principle equipped to solve, and "mysteries," which systematically elude our understanding; and he envisages a study of our cogni­ tive systems that would chart these powers and limi­ tations. I am here engaged in such a study, citing the mind-body problem as falling on the side of the mys­ teries. 4. See Thomas Nagel's discussion of realism in The View Form Nowhere (Oxford University Press: Oxford, (986), ch. 6. He argues there for the possibility of prop­ erties we can never grasp. Combining Nagel's realism with Chomsky-Fodor cognitive closure gives a position looking very much like Locke's in the Essay Concerning Human Understanding: the idea that our God-given faculties do not equip us to fathom the deep truth about reality. In fact, Locke held precisely this about the rela­ tion between mind and brain: only divine revelation could enable us to understand how "perceptions" are produced in our minds by material objects. 5. Hume, of course, argued, in effect, that no theory essentially employing a notion of objective causal ne­ cessitation could be grasped by our minds-and like­ wise for the notion of objective persistence. We might compare the frustrations of the Humean mind to the conceptual travails of the pure sound beings discussed in ch. 2 of P. F. Strawson's Individuals, (Methuen: Lon­ don, 1 959); both are types of mind whose constitution puts various concepts beyond them. We can do a lot better than these truncated minds, but we also have our constitutional limitations. 6. See the Essay, Book II, ch. 4. Locke compares the project of saying what solidity ultimately is to trying to clear up a blind man's vision by talking to him. 7. Some of the more arcane aspects of cosmology and quantum theory might be thought to lie just within the bounds of human intelligibility. Chomsky suggests that the causation of behavior might be necessarily mysteri­ ous to human investigators: see Reflections on Lan-

540

guage, p. 1 56. I myself believe that the mind-body problem exhibits a qualitatively different, and higher, level of mystery from this case (unless it is taken as an aspect of that problem). 8. Cf. Nagel's discussion of emergence in "Panpsy­ chism," in Mortal Questions (Cambridge University Press: Cambridge, 1979). I agree with him that the ap­ parent radical emergence of mind from matter has to be epistemic only, on pain of accepting inexplicable mira­ cles in the world. 9. Despite his reputation for pessimism over the mind­ body problem, a careful reading of Nagel reveals an optimistic strain in his thought (by the standards of the present chapter): see , in particular, the closing remarks of (see chapter 32) "What Is It Like to be a BatT' Nagel speculates that we might be able to devise an "objective phenomenology" that made conscious states more amenable to physical analysis. Unlike me, he does not regard the problem as inherently beyond us. 1 0. This is perhaps the most remarkably optimistic view of all-the expectation that reflecting on the ordi­ nary concept of pin (say) will reveal the manner of pain's dependence on the brain. If I am not mistaken, this is in effect the view of common-sense functionalists: they think that P consists in causal role, and that this can be inferred analytically from the concept of con­ scious states. This would make it truly amazing that we should ever have felt there to a mind-body problem at all, since the solution is already contained in our mental concepts. What optimism! I I . See Nagel, "What is it Like to be a BatT' (chapter 32). Notice that the fugitive character of such properties with respect to our concepts has nothing to do with their "complexity"; like fugitive color properties, such expe­ riential properties are "simple". Note too that such properties provide counter-examples to the claim that (somehow) rationality is a faculty that, once possessed, can be extended to encompass all concepts, so that if any concept can be possessed then every concept can.

12. It might be suggested that we borrow Nagel's idea of "objective phenomenology" in order to get around this problem. Instead of representing experiences under subjective descriptions, we should describe them in en­ tirely objective terms, thus bringing them within our conceptual ken. My problem with this is that, even al­ lowing that there could be such a form of description, it would not permit us to understand how the subjective aspects of experience depend upon the brain--which is

Can We Solve the Mind-Body Problem?

really the problem we are trying to solve. In fact, I doubt that the notion of objective phenomenology is any more coherent than the notion of subjective physiology. Both involve trying to bridge the psychophysical gap by a sort of stipulation. The lesson here is that the gap cannot be bridged just by applying concepts drawn from one side to items that belong on the other side; and this is be­ cause neither sort of concept could ever do what is needed. 1 3 . We should distinguish two claims about the im­ perceptibility of consciousness: (i) consciousness is not perceivable by directing the senses onto the brain; (ii) consciousness is not perceivable by directing the senses anywhere, even toward the behavior that "expresses" conscious state. I believe both theses, but my present point requires only (i). I am assuming, of course, that perception cannot be unrestrictedly theory-laden; or that if it can, the infusions of theory cannot have been originally derived simply by looking at things or tasting them or touching them or . . . 14. Nagel discusses the difficulty of thinking of con­ scious processes in the spatial terms that apply to the brain in The View From Nowhere, pp. 50- I , but he does not draw my despairing conclusion. The case is exactly unlike (say) the dependence of liquidity on the proper­ ties of molecules, since here we do think of both terms of the relation as spatial in character; so we can simply employ the idea of spatial composition. 1 5. Cf. Nagel: "it will never be legitimate to infer, as a theoretical explanation of physical phenomena alone, a property that includes or implies the consciousness of its subject," "Panpsychism," p. 1 83. 1 6. It is surely a striking fact that the microprocesses that have been discovered in the brain by the usual methods seem no nearer to consciousness than the gross properties of the brain open to casual inspection. Nei­ ther do more abstract "holistic" features of brain func­ tion seem to be on the right lines to tell us the nature of consciousness. The deeper science probes into the brain the more remote it seems to get from consciousness. Greater knowledge of the brain thus destroys our illu­ sions about the kinds of properties that might be dis­ covered by traveling along this path. Advanced neuro­ physiological theory seems only to deepen the miracle. 17. The kind of limitation I have identified is therefore not the kind that could be remedied simply by a large increase in general intelligence. No matter how large the frontal lobes of our biological descendants may become,

541

they will still be stumped by the mind-body problem, so long as they form their (empirical) concepts on the basis of perception and introspection. 1 8. Or again, no more miraculous than the theory of evolution. Creationism is an understandable response to the theoretical problem posed by the existence of com­ plex organisms; fortunately, we now have a theory that renders this response unnecessary, and so undermines the theism required by the creationist thesis. In the case of consciousness, the appearance of miracle might also tempt us in a "creationist" direction, with God required to perform the alchemy necessary to transform matter into experience. Thus the mind-body problem might similarly be used to prove the existence of God (no miracle without a miracle-maker). We cannot, I think, refute this argument in the way we can the original cre­ ationist argument, namely by actually producing a non­ miraculous explanatory theory, but we can refute it by arguing that such a naturalistic theory must exist. (It is a condition of adequacy upon any account of the mind­ body relation that it avoid assuming theism.) 70, 8:66 1 679. 19. See Paul Benacerraf, "Mathematical Truth," Jour­ nal ofPhilosophy, 1973, for a statement of this problem about abstract entities. Another problem that seems to me to differ from the mind-body problem is the problem of free will. I do not believe that there is some un­ knowable property Q which reconciles free will with determinism (or indeterminism); rather, the concept of free will contains internal incoherencies-as the concept of consciousness does not. This is why it is much more reasonable to be an e1iminativist about free will than about consciounsess. 20. A test of whether a proposed solution to the mind­ body problem is adequate is whether it relieves the pressure toward eliminativism. If the data can only be explained by postulating a miracle (i.e., not explained), then we must repudiate the data-this is the principle behind the impulse to deny that conscious states exist. My proposal passes this test because it allows us to resist the postulation of miracles; it interprets the eeriness as merely epistemic, thought deeply so. Constructive solu­ tions are not the only way to relieve the pressure. 2 1 . Chomsky suggests that the very faculties of mind that make us good at some cognitive tasks may make us poor at others; see Reflections on Language, pp. 1 55-6. It seems to me possible that what makes us good at the science of the purely physical world is what skews us

542

Colin McGinn

away from developing a science of consciousness. Our faculties bias us toward understanding matter in mo­

tion, but it is precisely this kind of understanding that is inapplicable to the mind-body problem. Perhaps, then,

the price of being good at understanding matter is that we cannot understand mind. Certainly our notorious tendency to think of everything in spatial terms does not help us in understanding the mind.

22. I get this phrase from Fodor,

Mind,

The Modularity of

p. 1 2 1 . The intended contrast is with kinds of

cognitive closure that stem from exogenous factors-as, say, in astronomy. Our problem with

P is not that it

is

too distant or too small or too large or too complex; rather, the very structure of our concept-forming appa­ ratus points us away from 23. Saul Kripke,

P.

Naming and Necessity

( Basil Black­

well: Oxford, 1 980). Of course, Descartes explicitly

argued from (what he took to be) the essential natures of the body and mind to the contingency of their con­ nection. If we abandon the assumption that we know these natures, then agnosticism about the modality of the connection seems the indicated conclusion.

24. This is the kind of realism defended by Nagel in ch.

6 of The View From Nowhere: to be is not to be con­ ceivable by us. I would say that the mind-body problem provides a demonstration that there

are

such concept­

transcending properties-not merely that there

could

be.

I should also say that realism of this kind should be accepted precisely because it helps solve the mind-body problem; it is a metaphysical thesis that pulls its weight in coping with a problem that looks hopeless otherwise.

There is thus nothing "epiphenomenal" about such radical realism: the existence of a reality we cannot know can yet have intellectual significance for us. Anti­ realists are unable to solve the mind-body problem, in addition to their other troubles. 25. Discussions with the following people have helped me work out the ideas of this paper: Anita Avramides, Jerry Katz, Ernie Lepore, Michael Levin, Thomas Nagel, Galen Strawson, Peter Unger. My large debt to Nagel's work should be obvious throughout the paper: I would not have tried to face down the mind-body

problem had he not first faced up to it.

34

On Leaving Out What It's Like

Joseph Levine Among the reasons for doubting the adequacy of physicalist theories of the mind is the charge that such theories must "leave out" the qualitative, conscious side of mental life. One problem with evaluating this objection to physicalism is that it is not clear just what physicalist theories are being charged with. What is it for a theory to "leave out" a phenomenon? My project in this chapter is threefold: First, I want to clarify the anti-physi­ calist charge of "leaving out" qualia, distinguish­ ing between a metaphysical and an epistemological reading of the objection. Second, I will argue that standard anti-physicalist conceivability argu­ ments fail to show that physicalist theories "leave out" qualia in the metaphysical sense. But, third, I will also argue that these conceivability argu­ ments do serve to establish that physicalist theo­ ries "leave out" qualia in the epistemological sense, because they reveal our inability to explain qualitative character in terms of the physical properties of sensory states. The existence of this "explanatory gap" constitutes a deep inadequacy in physicalist theories of the mind. 1

The Metaphysical Reading To begin, let us focus on the metaphysical reading of the phrase "leave out." In this sense, to say that a theory leaves out a certain phenomenon is to say that there are objects, events or properties to which the descriptive apparatus of the theory cannot refer. For instance, on Descartes's view, since the mind is composed of a non-physical, unextended substance, there is no way to use the predicates that apply to extended objects to refer to the mind. Property dualist views are similar in this respect. For a property dualist, there is no way of constructing descriptions using physical predicates2 that apply to mental properties. At least since Descartes, anti-physicalist argu­ ments have taken roughly the following form. It

is alleged that certain situations are imaginable, conceivable etc., and then a metaphysical con­ clusion is drawn. So, Descartes claims that from the fact that he can coherently conceive of the situation in which his body does not exist-for example, he may be deceived by an evil demon­ and from the fact that he cannot conceive of the situation in which his mind does not exist (i.e., consistent with his having his current experi­ ences), it follows that his body and his mind are not identical. A look at the current state of the debate shows that anti-physicalist arguments have not ad­ vanced significantly beyond Descartes's. In par­ ticular, I want to focus on the two most prominent contemporary anti-physicalist argu­ ments, those of Saul Kripke (1980) and Frank Jackson (1 982).

Kripke's Argument Kripke argues that there is an important asym­ metry between purported mental-physical iden­ tity statements and those that derive from other scientific reductions. In both cases, if the identity statements are true, they are necessarily true. Also, in both cases, the identity statements in­ volved appear contingent. 3 The asymmetry arises when we attempt to explain away their apparent contingency. Whereas the apparent contingency of other scientific identity statements can be ex­ plained away adequately, this cannot be done for mental-physical identity statements. Suppose we compare a standard scientific identity statement like ( I ) below to a mental­ physical identity statement like (2) below: l . Water

2. Pain

=

=

H20

the firing of C-fibers

Since neither statement is known a priori, they are both imaginably false. Yet, if they are true, they

Joseph Levine

are necessarily true-they are not even possibly false. How do we reconcile the apparent con­ tingency with the actual necessity? According to Kripke, this is easy to do in the case of (I). When we think we are imagining a situation in which water is not H20, in fact we are imagining a sit­ uation in which some substance which behaves superficially like water-but is not water-is not H20. On the other hand, a similar account will not work to explain the apparent contingency of (2), for to imagine a situation in which one is ex­ periencing a state superficially like pain just is to imagine a situation in which one is experiencing pain. Conscious mental states are unlike external objects in that the standard distinction between how they appear and how they really are does not apply. Many responses to Kripke's argument have appeared over the years. Early on it was pointed out that materialism does not entail the sort of type-type reductionism of the mental to the phys­ ical that is manifested in statements like (2). Rather, mental states are higher-order functional states, which can be realized, at least in principle, in a wide variety of physical systems. Hence, it is quite consistent with materialism that it is possi­ ble for one to experience pain and yet have no C-fibers whatever to fire. However, functionalism itself has come under attack from Cartesian-style objections, particu­ larly the inverted and absent qualia hypotheses. 4 The essence of these objections is that it seems perfectly imaginable that there could be creatures functionally alike who nevertheless differed in the qualitative character of their experiences; or, even worse, that there could be a creature functionally like ourselves who had no qualitative experiences at all. One line of response to either or both of these objections is to retreat to a physiological reductionist view with respect to qualia. That is, instead of identifying qualia with functional states, we identify them with the neurophysio­ logical states that play the relevant functional roles in human beings, which would explain the pos­ sibility of both inverted and absent qualia. S Of

544

course, this just brings us back to where we started. I favor another strategy in response to Kripke's argument. Suppose he's right that we can coher­ ently imagine feeling pain without having C­ fibers firing. What's more, suppose he's right that this coherently imagined scenario cannot be ex­ plained away in the manner in which we explain away imagining that water is not H20. Still, what is imaginable is an epistemological matter, and therefore what imagining pain without C-fibers does is establish the epistemological possibility that pain is not identical with the firing of C­ fibers. It takes another argument to get from the epistemological possibility that pain is not the firing of C-fibers to the metaphysical possi­ bility, which is what you need to show that pain isn't infact identical to the firing of C-fibers. 6 Kripke, following Descartes, seems to rely on the idea that when you have a really "clear and distinct" idea you have access to how things are, metaphysically speaking. If one believes in this sort of access to metaphysical facts, it then makes sense to use the Kripke test, by which I mean the test that determines whether the imagined sce­ nario can be explained away appropriately, to determine whether one has hold of a genuine metaphysical possibility or not. So, in the water/ H20 case, Kripke shows that, as it were, when your idea is made properly clear and distinct, you see that what you are really entertaining is the thought that something that behaves like water is not H20. Notice that the situation satisfying this description is indeed metaphysically possible. Since the same move doesn't work for the paint C-fibers case, we conclude that there is a meta­ physically possible world in which pain isn't the firing of C-fibers. But suppose we reject the Cartesian model of epistemic access to metaphysical reality alto­ gether. One's ideas can be as clear and distinct as you like, and nevertheless not correspond to what is in fact possible. The world is structured in a certain way, and there is no guarantee that our ideas will correspond appropriately. If one fol-

On Leaving Out What It's Like

lows this line of thought, then the distinction Kripke points out between the pain/C-fibers case and the water/H20 case turns out to be irrelevant to the question of what is or is not metaphysically possible. Thus, for all we know, pain just is the firing of C-fibers or, if functionalism is right, the realization of a certain functional state. Early identity theorists, in their response to Cartesian conceivability arguments, protested that they only intended their theory to be empiri­ cal, and therefore it was not subject to objec­ tions from what was conceivable or not. 7 Kripke correctly pointed out the error of that sort of response. Empirical or not, if they were making identity claims, then a consequence of their theory is that it is not possible for some mental state not to be identical to its physical or functional correlate. But the basis of Kripke's objection lies in a strict distinction between meta­ physical and epistemological possibility. Once we appreciate that distinction, the physicalist can return to her original ploy, i.e., to say that meta­ physical consequences cannot be drawn from considerations of what is merely conceivable. Thus, without an argument to the effect that what is metaphysically possible is epistemologically accessible, the Cartesian argument fails.

Jackson's Argument A similar problem�that is, a reliance on the Cartesian model of epistemic access to meta­ physical reality or, in other words, using epis­ temoloical premises to support a metaphysical conclusion-seems to infect Frank Jackson's well-known "knowledge argument" against ma­ terialism. Jackson takes the thesis of physicalism to be the claim that "all (correct) information is physical information" (Jackson 1 982, p. 1 27). of course, his use of the notion of information here is already fraught with ambiguity as between mat­ ters epistemological and metaphysical, a point to which I will return shortly. His argument against physicalism revolves around examples like the following:

545

Mary is a brilliant scientist who is . . . forced to inves­ tigate the world from a black and white room via a black and white television monitor. She specializes in the neurophysiology of vision and acquires . . . all the physical information there is to obtain about what goes on when we see ripe tomatoes, or the sky, and use terms like "red," "blue," and so on . . . . What will happen when Mary is released from her black and white room or is given a color television monitor? Will she learn anything or not? It seems just obvious that she will learn something about the world and our visual experience of it. But then it is inescapable that her previous knowledge was incomplete. But she had all the physical information. Ergo there is more to have than that, and Physicalism is false. (Jackson 1982, p. 1 30)

There have been a number of replies to Jackson in the literature, and the sort of reply I am most interested in is exemplified by Horgan ( 1 984). Horgan argues that Jackson is equivocating on the notion of "physical information." In one sense this might mean information expressed in terms used in the physical sciences. In another sense it might mean information about physical facts, processes, etc. It is only in the second sense that any reasonable physicalist is committed to the claim that all information is physical in­ formation. Of course, in this sense, the thesis could be better put by just saying that all token events and processes are physical events and pro­ cesses�by which one means something like, they have a true description in the terms of the physical sciences. (Actually, I think any interesting doc­ trine of physicalism is committed to more than this, though it's difficult to pin down exactly how much more. At any rate, it doesn't affect the present point.) But no plausible version of physi­ calism is committed to the claim that all in­ formation is physical information in the first sense: in the sense that it is expressed in (or trans­ latable into) the terms of the physical sciences. What Mary's case shows, argues Horgan, is that there is information Mary acquires after leaving the room that isn't physical information in the first sense, but not that it isn't physical information in the second sense. Certainly, she

Joseph Levine

may think something like, "Oh, so

546

this is what red

something out, it can't be in the sense that there

looks like." Her experience of learning something

are facts that physicalistic descriptions fail to pick

new shows that she now knows something she

out, since we have no argument to show that the

didn't know before. She now knows what it's like

two sorts of descriptions just cited do not refer to

to see red, which she didn't know before. But it

the same facts.

doesn't follow that her new information isn't

I will briefly consider two replies on behalf of

physical information in the second sense: that is,

the metaphysical anti-physicalist. First, perhaps

that it isn't information about a physical event

Cartesian conceivability arguments can't demon­

or process. On the contrary, the case of Mary

strate that qualia aren't physical states or pro­

typifies the phenomenon of there being several

cesses, but they at least throw the burden of

distinguishable ways to gain epistemic access

argument back onto the physicalist to show why

to the same fact. One cannot infer from a variety

we should think they are physical states or pro­

of modes of access to a variety of facts being

cesses. The physicalist strategy presented above

accessed. A similar emphasis on the distinction between

only opens a space for the physicalist hypothesis, but it doesn't give us any reason to believe it.

the epistemology and the metaphysics of the

Fair enough. That's all it was intended to ac­

matter underlies the following sort of reply to

complish. The main burden of the physicalist

Jackson. What the case of Mary shows is that one

argument is borne by considerations of causal

can know which physical (or functional) descrip­

interaction. If qualia aren't physical processes (or

tion a mental state satisfies without knowing what

realized in physical processes), then it becomes

it's like to occupy that state. But of course! After

very difficult to understand how they can play a

all, in order to know what it's like to occupy a

causal role in both the production of behavior

state one has actually to occupy it! All Mary's

and the fixation of perceptual belief. Jackson

newly acquired knowledge amounts to is her new

himself admits the cogency of this argument,

experience, which is indeed new, since she didn't

and therefore

have those experiences until leaving the room.

epiphenomenalism. Those who don't find that

bites

the

bullet

by

endorsing

So it remains perfectly possible that what she

bullet particularly appetizing must either show

learns is what it's like to occupy a certain physico­

how the requisite mental-physical causal relations

functional state. There is no threat to physicalism

are possible on a dualist account or endorse

here.

physicalism. The second reply on behalf of the metaphysical

Two Metaphysical Anti-Physicalist Replies The common thread in the responses to both Kripke and Jackson is that their thought experi­ ments demonstrate only an epistemological di­ vide between different modes of access to what may, for all we know, be the very same phenom­ enon. On the one hand, we have certain physico­ functional descriptions of certain states occupied by psychological subjects. On the other hand, we have whatever descriptions are derived from one's first-hand experience of these states. If these thought experiments show that physicalism leaves

anti-physicalist goes like this. 8 Take some iden­ tity statement that is not epistemologically neces­ sary, like (3) below: 3. The Morning Star = the Evening Star. Though one might accept Kripke's claim that (3) is necessarily true if true at all, still one has to ex­ plain its apparent contingency. The way we do this is to say that what is contingent is that the very same heavenly body should appear where Venus does in the morning and also where it does in the evening. Notice that our explanation of the apparent contingency of (3) adverted to a real

On Leaving Out What It's Like

distinction between two of Venus's properties: namely, appearing at a certain heavenly location in the morning and appearing at a certain heav­ enly location in the evening. That is, we can ex­ plain the epistemological state of conceiving of the Morning Star and the Evening Star as two distinct objects, despite their identity, by refer­ ence to two distinct properties through which we have epistemic access to the one object. Suppose one grants that the absent qualia ar­ gument does indeed establish at least the epis­ temological possibility that a qualitative state and a functional state are distinct, even though they are in fact identical. In order to explain how it is possible to conceive of this one state as two dis­ tinct states, we must assume that there are (at least) two "modes of presentation" under which we apprehend this one state. Let us call them the "first-person mode of presentation" and the "third-person mode of presentation. " But now we seem committed to the claim that there are at least two distinct properties of the state corre­ sponding to the two modes of access, akin to the two spatiotemporal properties of Venus by which we gain epistemic access to it in the morning and in the evening. If so, this shows that qualitative character, the property by which we identify a conscious state in the first-person mode of access, is distinct from the property of playing a certain functional role, the property by which we identify that conscious state in the third-person mode of access. So, we seem to be back to deriving a metaphysical conclusion from an epistemological premise, namely, that the property of having a certain qualitative character is distinct from the property of playing a certain functional role (or being in a certain neurophysiological state). 9 The physicalist, however, can reply as follows. Certainly whenever we conceive of a single object in two distinct ways-sufficiently distinct ways, in fact, that we believe we are conceiving of two distinct objects-the object in question must possess (at least) two distinct properties that cor­ respond to these different modes of presentation.

547

But whether or not we now have a problem for physicalism depends on which two district prop­ erties we find ourselves committed to. This re­ quires some elaboration. What the physicalist needs to maintain is that having a certain qualitative character is a physical or functional property. This reduction of qual­ itative character is necessary in order to account for the causal role that qualia play in the fixation of perceptual belief and the production of be­ havior. So, if the argument above could establish that having a certain qualitative character is a property distinct from a mental state's physical and functional properties, that would be the sort of metaphysical conclusion the anti-physicalist is after. However, the argument above does not in fact establish the non-identity of having a certain qualitative character and any of a state's physical or functional properties. The argument begins with the premise that there must be two properties of the one state, providing two epistemic paths by which the subject conceives of that state, in order to account for the fact that it is epistemologically possible for someone to experience qualitative character without occupying the relevant phys­ ico-functional state. We can accept this premise and yet refuse to grant the conclusion--that hav­ ing a certain qualitative character is irreducible to a state's physico-functional properties-by find­ ing two other properties to provide the requisite epistemic paths. For instance, we can account for the conceivability of experiencing a certain quale without occupying the relevant physico-func­ tional state by noting that the two relational properties, being thought of under the description "what I am now consciously experiencing" and being thought of under the description "the state that normally causes [such-and-such behavioral effects]," are not identical. However, there is no reason for the physicalist to claim that these two properties are identical, and therefore the argument above fails to mount a challenge to physicalism.

Joseph Levine

The Epistemological Reading I have argued that on a metaphysical reading of "leave something out," Cartesian conceivability arguments cannot establish that physicalist theo­ ries of mind leave something out. However, there is also an epistemological sense of "leave some­ thing out," and it is in this sense that con­ ceivability arguments, being epistemological in nature, can reveal a deep inadequacy in physical­ ist theories of mind. For a physicalist theory to be successful, it is not only necessary that it provide a physical descrip­ tion for mental states and properties, but also that it provide an explanation of these states and properties. In particular, we want an explanation of why when we occupy certain physico-func­ tional states we experience qualitative character of the sort we do. It's not enough for these pur­ poses to explain the contribution of qualitative states to the production of behavior, or the fixation of perceptual belief; this is a job that a physicalist theory can presumably accomplish. (At least there is no reason stemming from con­ ceivability arguments to suppose that it cannot.) Rather, what is at issue is the ability to explain qualitative character itself; why it is like what it is like to see red or feel pain. Conceivability arguments serve to demonstrate the inability of physicalist theories to provide just this sort of explanation of qualitative character. To see this, consider again the disanalogy Kripke draws between statements ( 1 ) and (2) above. Kripke bases his argument on the fact that both statements appear contingent, and then distin­ guishes between them by pointing out that the apparent contingency of (1), but not of (2), can be explained away. My strategy is quite different. I see the disanalogy between the water/H20 case and the pain/C-fibers case in the fact that there is an apparent necessity that flows from the reduc­ tion of water to H20, a kind of necessity that is missing from the reduction of pain to the firing of C-fibers.

548

The necessity I have in mind is best exemplified by considering statement (1 '): 1 '. The substance that manifests [such-and-such macro properties of water] is H20. On Kripke's view, ( 1 ') is in fact contingent, and it is the contingency of ( 1 ') that explains the appar­ ent contingency of (1). So, on his view, (1') and (2) are on a par. Yet, it seems to me that there is an important difference between them. If we con­ sider the apparent contingency that attaches to (2), we notice that it works in both directions: it is equally conceivable that there should exist a pain without the firing of C-fibers, and the firing of C-fibers without pain. However, the apparent contingency of ( 1 ') only works in one direction. While it is conceivable that something other than H20 should manifest the superficial macro prop­ erties of water, as Kripke suggests, it is not con­ ceivable, I contend, that H20 should fail to manifest these properties (assuming, of course, that we keep the rest of chemistry constant). This difference between the two cases reflects an important epistemological difference between the purported reductions of water to H20 and pain to the firing of C-fibers: namely, that the chemical theory of water explains what needs to be explained, whereas a physicalist theory of qualia still "leaves something out." It is because the qualitative character itself is left unexplained by the physicalist or functionalist theory that it remains conceivable that a creature should occupy the relevant physical or functional state and yet not experience qualitative character. The basic idea is that a reduction should ex­ plain what is reduced, and the way we tell whether this ' has been accomplished is to see whether the phenomenon to be reduced is epis­ temologically necessitated by the reducing phe­ nomenon, that is, whether we can see why, given the facts cited in the reduction, things must be the way they seem on the surface. I claim that we have this with the chemical theory of water but not with a physical or functional theory of qualia.

On Leaving Out What It's Like

The robustness of the absent and inverted qualia intuitions is testimony to this lack of explanatory import. Let me make the contrast between the reduc­ tion of water to H20 and a physico-functional reduction of qualia more vivid. What is explained by the theory that water is H20? Well, as an in­ stance of something that's explained by the re­ duction of water to H20, let's take its boiling point at sea level. The story goes something like this. Molecules of H20 move about at various speeds. Some fast-moving molecules that happen to be near the surface of the liquid have sufficient kinetic energy to escape the intermolecular at­ tractive forces that keep the liquid intact. These molecules enter the atmosphere. That's evapo­ ration. The precise value of the intermolecular attractive forces of H20 molecules determines the vapor pressure of liquid masses of H20, the pres­ sure exerted by molecules attempting to escape into saturated air. As the average kinetic energy of the molecules increases, so does the vapor pressure. When the vapor pressure reaches the point where it is equal to atmospheric pressure, large bubbles form within the liquid and burst forth at the liquid's surface. The water boils. I claim that given a sufficiently rich elaboration of the story above, it is inconceivable that H20 should not boil at 2 1 2°F at sea level (assuming, again, that we keep the rest of the chemical world constant). But now contrast this situation with a physical or functional reduction of some con­ scious sensory state. No matter how rich the in­ formation processing or the neurophysiological story gets, it still seems quite coherent to imagine that all that should be going on without there being anything it's like to undergo the states in question. Yet, if the physical or functional story really explained the qualitative character, it would not be so clearly imaginable that the qualia should be missing. For, we would say to ourselves something like the following: Suppose creature description

F. I

X

satisfies functional (or physical)

understand---from my functional (or

549

physical) theory of consciousness-what it is about instantiating

F

that is responsible for its being a

conscious experience. So how could with those very features and yet

X occupy a state not be having a

conscious experience?

One might object at this point that my position presumes something like the deductive-nomo­ logical account of explanation, an account that is certainly controversial. 1 0 In fact, I quite openly endorse the view that explanations involve show­ ing how the explanandum follows from the ex­ planans. I believe that the deductive-nomological model, in analyzing explanation in terms of exhibiting a necessary connection between ex­ planans and explanandum, is certainly on the right track. I am not committed, however, to the view that all explanations take the form of the "covering law" model described by Hempel ( 1 965) in his classic account of explanation. For instance, Robert Cummins ( 1983) has argued that some explanations take the form of "property theo­ ries," in which the instantiation of one sort of property is explained by reference to the in­ stantiation of some other properties. So we might, for example, explain a certain psychological capacity by reference to the physico-functional mechanisms that underlie it. In such cases we are not explaining one event by citing initial con­ ditions and subsuming it under a law, so it does not quite fit the traditional deductive-nomo­ logical model. I have no problem with Cummins's objection to the covering law model. Yet even in his exam­ ple-explaining how a psychological capacity is instantiated by reference to the underlying mech­ anisms-the element of necessity is there, even if there is no subsumption under laws. For it is clear that if citing the relevant underlying mechanisms really does explain how the psychological capacity in question is instantiated, then it would be in­ conceivable that some creature should possess these mechanisms and yet lack the capacity. If not, if we could conceive of a situation in which a creature possessed the relevant underlying

Joseph Levine

mechanisms and yet didn't possess the capacity in question, then I would claim that we haven't adequately explained the presence of the capacity by reference to those mechanisms. For we are still left wondering what distinguishes the actual situation, in which the creature possesses the capacity, from those conceivable situations in which it (he/she) does not. The Conceptual Basis of the Explanatory Gap I have argued that there is an important difference between the identification of water with H20, on the one hand, and the identification of qualitative character with a physico-functional property on the other. In the former case the identification affords a deeper understanding of what water is by explaining its behavior. Whereas, in the case of qualia, the subjective character of qualitative experience is left unexplained, and therefore we are left with an incomplete understanding of that experience. The basis of my argument for the existence of this explanatory gap was the conceiv­ ability of a creature's instantiating the physico­ functional property in question while not under­ going an experience with the qualitative character in question, or any qualitative character at all. In order fully to appreciate the nature and scope of the problem, however, it is necessary to explore in more detail the basis of the explanatory adequacy of theoretical reductions such as that of water . to H20, as well as the difference between these cases and the case of qualitative character. I can only being that project here, with the follow­ ing admittedly sketchy account. We will see that an adequate account must confront deep prob­ lems in the theory of conceptual content, thus drawing a connection between the issue of inten­ tionality and the issue of consciousness. Explanation and Reduction To begin with, it seems clear that theoretical re­ duction is justified principally on the basis of its

550

explanatory power. For instance, what justifies the claim that water is H20 anyway? Well, we might say that we find a preponderance of H20 molecules in our lakes and oceans, but of course that can't be the whole story. First of all, given all the impurities in most samples of water, this may not be true. Second, if we found that everything in the world had a lot of H20 in it-suppose H20 were as ubiquitous as protons-we wouldn't identify water with H20. Rather, we justify the claim that water is H20 by tracing the causal re­ sponsibility for, and the explicability of, the vari­ ous superficial properties by which we identify water-its liquidity at room temperature, its freezing and boiling points, and so forth-to H20. But suppose someone pressed further, asking why being causally responsible for this particular syndrome of superficial properties should be so crucial. 1 1 Well, we would say, what else could it take to count as water? But the source of this "what else" is obscure. In fact, I think we have to recognize an a priori element in our justification. That is, what justifies us in basing the identi­ fication of water with H20 on the causal respon­ sibility of H20 for the typical behavior of water is the fact that our very concept of water is of a substance that plays such-and-such a causal role. To adopt Kripke's terminology, we might say that our pretheoretic concept of water is charac­ terizable in terms of a "reference-fixing" descrip­ tion that roughly carves out a causal role. When we find the structure that in this world occupies that role, then we have the referent our concept. But now how is it that we get an explanation of these superficial properties from the chemical theory? Remember, explanation is supposed to involve a deductive relation between explanans and explanandum. The problem is that chemical theory and folk theory don't have an identical vocabulary, so somewhere one is going to have to introduce bridge principles. For instance, suppose I want to explain why water boils, or freezes, at the temperatures it does. In order to get an ex­ planation of these facts, we need a definition of "boiling" and "freezing" that brings these terms

On Leaving Out What It's Like

into the proprietary vocabularies of the theories appealed to in the explanation. . . . Well, the obvious way to obtam the reqwslte bridge principles is to provide theoretical reduc­ tions of these properties as well. 1 2 To take another example, we say that one of water's superficial properties is that it is colorless. But being colorless is not a chemical property, so be­ fore we can explain why water is colorless in terms of the molecular structure of water and the way that such structures interact with light waves, we need to reduce colorlessness to a property like having a particular spectral reflectance function. Of course, the justification for this reduction will, like the reduction of water to H20, have to be justified on grounds of explanatory enrichment as well. That is, there are certain central phenomena we associate with color, by means of which we pick it out, such that explaining those phenomena is a principal criterion for our acceptance of a theoretical reduction of color. The picture of theoretical reduction and ex­ planation that emerges is of roughly the followi�g form. Our concepts of substances and properties like water and liquidity can be thought of as representations of nodes in a network of causal relations, each node itself capable of further re­ duction to yet another network, until we get down to the fundamental causal determinants of na­ ture. We get bottom-up necessity, and thereby explanatory force, from the identification of the macroproperties with the microproperties be­ cause the network of causal relations constitutive of the micro level realizes the network of causal relations constitutive of the macro level. Any concept that can be analyzed in this way will yield to explanatory reduction. . Notice that on this view explanatory reductIOn is, in a way, a two-stage process. Stage I invol �es the (relatively? quasi?) a priori process of workmg the concept of the property to be reduced "into shape" for reduction by identifying the ca�sal role for which we are seeking the underlymg mechanisms. Stage 2 involves the empirical work of discovering just what those underlying mecha­ nisms are. 1 3

551

A Digression about Concepts In order to clarify the sense in which it is incon­ ceivable that something should be H20 and not be water I have had to slip into talking about concepts even worse, talking about analyzing the contents of concepts. This is unfortunate for my position, since the whole topic of concepts is filled with controversy, and I do not yet see how to construct a theory of conceptual content that will do the work, briefly outlined above, that needs to be done. Let me briefly indicate where the prob­ lems lie. In the literature on concepts and contents, var­ ious distinctions have emerged which are useful to our concerns here. First of all, we can distinguish between a concept's l4 "narrow content" and its "broad content." 1 5 A concept's broad content is its satisfaction conditions. This is the referential component of its content. The notion of �arrow content is meant to capture that aspect of Its con­ tent that is psychologically significant and inde­ pendent of facts external to the subject. With regard to the famous Twin Earth example, nar­ row content is what my concept of water and my twin's concept of water have in common. It is, of course, controversial whether or not it is narrow content that is relevant to the individuation of psychological states, or even whether there is such a thing as narrow content. However, I believe there is something psycholog­ ically significant that I and my twin have in com­ mon when we entertain the concept of water and, moreover, it is this aspect of our concept that seems relevant to the question of explanatory re­ duction. In will not defend this claim here, I will just presume it. So, how do we characterize the narrow content of our concept of water? On one view, the "functional role" view, 1 6 narrow content is de­ termined by the cluster of beliefs involving the concept of water that determine the inferential relations among them. On this view, to analyze the concept of water is just to present those cen­ tral beliefs. Our concept of water is the concept of

;

552

Joseph Levine

a substance that. . . , where statements about the typical behavior of water fill in the blank. On a functional role view, then, to say that our concept of water can be analyzed as the concept of a causal niche is just to say that the beliefs which go into the blank all involve the causal role of water. Thus a functional role view of narrow content seems quite amenable to my needs. However, there are real problems with this view. In partic­ ular, there is the problem of holism. That is, if you change any element of the description of causal relations definitive of the concept, you change the concept. Now, in the course of scientific inves­ tigation, we can expect to revise our beliefs about these causal connections as we learn more about the phenomenon under investigation. If such changes counted as changing the concept, then it wouldn't be water we were learning about when we discovered that water was H20. This would seem to be an intolerable consequence. There are three ways one might deal with this problem. First, just bite the bullet and admit that our concepts are hopelessly holistic. Second, attempt to distinguish those elements of the func­ tional role that are essential to the concept from those that are accidental, so that only changes in the essential elements constitute changes in concept. Third, find a different theory of narrow content. One might argue that biting the bullet is not as bad as it seems on the grounds that we are only talking about narrow content. So long as we are atomistic about reference, we can still make sense of the claim that two theories contain conflicting claims about water, since they are talking about the very same thing. However, what we are look­ ing for here is a notion of conceptual content suitable for grounding the explanatory relation, and this is clearly a matter of narrow content. Unless we can build stability into the notions, it is unclear how to make sense of the idea that the chemical theory of water explains why water be­ haves the way it does. The second sort of strategy has a sad history, and I do not see how to make it work. For any

element of the functional role you pick as essen­ tial, there always seems to be a story you can tell in which that element is missing and yet it seems intuitively right to claim that the subject still has the concept in question. As for a different theory of narrow content, the only one I know, that de­ parts radically from the functional role theory, is Fodor's ( 1 987) theory of narrow content as a function from contexts to broad contents. It is not at all clear to me how the notion of a conceptual content as the specification of a causal niche could be made to work on this view. To sum up, there seems to be a need for a theory of conceptual content that both grounds explanatory reductions on the basis of some sort of functional/causal analysis of the requisite con­ cepts, and yet does not entail holism. I do not have such a theory, and so must content myself with merely characterizing this desideratum on a theory yet to be developed.

Qualia Again If we apply the same model of explanatory value to the theoretical reduction of qualia as we used for the reduction of water, then we need to look for a property that is being reduced and then a property, or set of properties, by which the to­ be-reduced property is normally picked out. Of course, this raises a problem. When it comes to something like the qualitative character of a sen­ sation of red, what other property could we point to to play the role of the reference-fixer? We seem to pick out this property by itself. The distinction between the property to be reduced and the properties by which we normally pick it out, or its superficial manifestation, seems to collapse. (Obviously this is connected to Kripke's point about the appearance/reality distinction not get­ ting a hold in this case.) There are, of course, other properties of qualia that we can expect a theoretical reduction to explain; namely, those properties associated with their causal role in mediating environmen-

On Leaving Out What It's Like

tal stimuli and behavior. It is precisely on the grounds that a particular physico-functional property can explain the "behavior" of qualita­ tive states that we would endorse an identification between a particular quale and that property. Furthermore, if that were all there were to our concept of qualitative character-as the ana­ lytical functionalist maintains-then there would be no difference between the theoretical reduction of water and of qualia with respect to explanatory success. But the very fact that one can conceive of a state playing that role and yet not constituting a qualitative experience shows, or at least so I have argued, that causal role is not all there is to our concept of qualitative character. What seems to be responsible for the explana­ tory gap , then, is the fact that our concepts of qualitative character do not represent, at least in terms of their psychological contents, causal roles. Reduction is explanatory when by reducing an object or property we reveal the mechanisms by which the causal role constitutive of that ob­ ject or property is realized. Moreover, this seems to be the only way that a reduction could be ex­ planatory. Thus, to the extent that there is an element in our concept of qualitative character that is not captured by features of its causal role, to that extent it will escape the explanatory net of a physicalistic reduction.

Conclusion I will conclude by drawing out another conse­ quence of this discussion of the explanatory gap. It is customary to attack the mind-body problem by a divide-and-conquer strategy. On the one hand, there is the problem of intentionality. How can mere matter support meaning; how can a bit of matter be about something? On the other hand, there is the problem of consciousness, or, to be more specific, the problem of qualitative charac­ ter. How can there be something it is like to be a mere physical system? By separating the two questions, it is hoped that significant progress can be made on both.

553

Certainly in recent years we have come to have a deeper understanding of the issues surrounding the question of intentionality, and this progress has been largely the result of divorcing the ques­ tion of intentionality from the question of con­ sciousness. 1 7 However, if I am right, there may be more of a connection between the problem of qualitative character and the problem of in­ tentionality than it is fashionable now to suppose. It is not that one needs to be capable of experi­ encing qualia in order to bear intentional states, as Searle would have it. Rather, since the problem of qualitative character turns out to be primarily epistemological, the source of which is to be found in the peculiar nature of our cognitive rep­ resentations of qualitative character, a theory of intentional content ought to explain what makes these representations so uniquely resistant to in­ corporation into the explanatory net of physical science. 1 8 Thus the problem of qualia threatens to enlarge into the problem of the mind generally.

Acknowledgments An earlier version of this chapter was delivered at the conference on Mind, Meaning, and Nature, at Wesleyan University , March 3 1 , 1 989. The chap­ ter was completed while I was holding a fellow­ ship from the National Endowment for the Humanitites. I would also like to thank Louise Antony, David Auerbach, Martin Davies, and Georges Rey for helpful discussions and critical comments on earlier drafts.

Notes I . See Levine (1983) where I first argued for the ex­ istence of an explanatory gap. 2. Of course, it is a non-trivial question to decide which predicates count as "physical" predicates, but for pres­ ent purposes we need not attempt a precise explication of the notion.

3. As Kripke (1 980, p. 1 54) puts it, there is a "certain obvious element of contingency" about such theoretical identity statements.

Joseph Levine

4. For extensive discussion of the absent and inverted gualia hypotheses, see Block and Fodor (1972); Block (1978, 1 980); Horgan (1984); Shoemaker (1984, chs. 9, 14, and 1 5); Conee (1985); Levine (1989). 5. For various versions of this position, see Block (1978, 1980), Horgan (1984), and Shoemarker (1984). 6. Given that Kripke is largely responsible for drawing the philosophical world's attention to the distinction between epistemological possibility and metaphysical possibility, it might seem odd to accuse him of confusing the two in this case. I diagnose his mistake as follows. Since he believes that any state which appears painful is thereby a pain, he infers that there is no appearance/ reality distinction with respect to pain, and therefore epistemological and metaphysical possibility collapse in this case. But even if he's right that any state that appears to be a pain is a pain, he still has to justify the premise that it's possible for one to suffer even apparent pain without having one's C-fibers firing, and he can't do that, I contend, merely by noting that it seems possible. 7. See, for instance, Smart's (1959) reply to his "Ob­ jection 2." 8. This objection was suggested to me by a discussion in White (1986). The analogy to the Morning Star­ Evening Star case is his. 9. As White expicitly acknowledges, a precursor of this objection can be found in Smart's (1959) famous "Objection 3." 10. For the classic presentation of the deductive­ nomological model of explanation, see Hempel (1965, ch. 12). I I . Of course, it's possible to imagine situations in which we would accept a theory o[ water that never­ theless left many of its superficial properties un­ explained. However, unless the theory explained at least some of these properties, it would be hard to say why we consider this a theory of water.

12. In some cases, for instance, with properties such as liquidity and mass, it might be better to think of their theoretical articulations in physical and chemical theory more as a matter of incorporating and refining folk the­ oretic concepts than as a matter of reducing them. But this is not an idea I can pursue here. 13. To a certain extent my argument here is similar to Alan Sidelle's (1989) defense of conventionalism, though I don't believe our positions coincide completely.

554

14. Some readers might find my speaking of a concept's content, as opposed to a term's content confusing. I am interested in the nature of our thoughts, not with their expression in natural language. For present purposes, we can think of a concept as a term in whatever internal, mental language is employed in our cognitive processing. 1 5 . For the source of this distinction, see Putnam (1975). For further discussion of its significance for psychology, see Fodor (1987, ch. 3) and Burge (1986). 16. See Block (1986) for a defense of the functional role view. 1 7. For a dissenting opinion on the question of divorc­ ing intentionality from consciousness, see Searle (1989). 18. See Rey (chapter 12) and Van Gulick (chapter 7) in Davies and Humphries [or suggestive approaches to just this problem.

References Block, N. (1978). Troubles with functionalism. In C. Wade Savage (ed.), Perception and Cognition: Issues in the Foundations of Psychology. Vol. 9, Minnesota Studies in the Philosophy of Science. Minneapolis: Uni­ versity of Minnesota Press. ---

. (1 980). Are absent qualia impossible? Philo­

sophical Review 89, 257-·74. . (1986). Advertisement for a semantics for psy­ chology. In P. A. French, T. E. Uehling, Jr., and H. K. Wettstein (eds.), Studies in the Philosophy of Mind. Vol. 10, Midwest Studies in Philosophy. Minneapolis: Uni­ versity of Minnesota Press, 615-78. ---

Block, N., and Fodor, J. A. (1972). What psychological states are not. Philosophical Review, 8 1 , 1 59-8 1 . Burge, T . (1986). Individualism and psychology. Philo­

sophical Review 95, 3-46. Conee, E. (1985). The possibility of absent qualia, Philosophical Review 94, 345-66. Cummins, R. (1983). The Nature of Psychological Ex­ planation. Cambridge, MA: MIT Press. Davies, M., and Humphreys, G. (eds.). (1 993). Con­

sciousness. Oxford: Blackwell. Fordor, J. A. (1 987). Psychosemantics: The Problem of Meaning in the Philosophy of Mind. Cambridge, MA: MIT Press

On Leaving Out What It's Like

Hempel, C. G. ( 1 965). Aspects of Scientific Explanation. New York: Free Press. Horgan, T. (1984). Functionalism, qualia, and the in­ verted spectrum. Philosophy and Phenomenological Re­ search 44, 453-69. Jackson, F. (1982). Epiphenomenal qualia. Philosoph­

ical Quarterly 32, 1 27-36. Kripke, S. (1 980). Naming and Necessity. Oxford: Blackwell. Levine, J. (1 983). Materialism and qualia: The explana­ tory gap. Pacific Philosophical Quarterly 64, 354-61 . --

. ( 1989). Absent and inverted qualia revisited.

Mind and Language 3, 27 1 -87. Putnam, H . ( 1 975). The meaning of meaning. In K. Gunderson (ed.), Language, Mind. and Knowledge, 1 3 1 -93. Minneapolis: University of Minnesota Press. Rey, Georges ( 1 993). Sensational sentences. In Davies, M., and Humphreys, G. (eds.) Consciousness, 240-57. Oxford: Blackwell. Searle, 1. (1 989). Consciousness, unconsciousness, and intentionality. Philosophical Topics 1 7, 1 93-209. Shoemaker, S. ( 1984). Identity. CalLfe. and Mind. Cam­ bridge: Cambridge University Press. Sidelle, A. (1989). Necessity. Essence, and Individuation: A Defense of Conventionalism. Ithaca: Cornell Univer­ sity Press. Smart, 1. 1. C. (1959). Sensations and brain processes.

Philosophical Review 68, 1 4 1 -56. Van Gulick, Robert ( 1 993). Understanding the phe­ nomenal mind: Are we all just armadillos? In Davies, M., and Humphreys, G. (eds.), Consciousness, 1 34-54. Oxford: Blackwell. White, S. L. (1986). Curse of the qualia. Synthese 68.

555

VIII

THE KNOWLEDGE ARGUMENT

35

Understanding the Phenomenal Mind: Are We AU Just Armadillos? Part I: Phenomenal Knowledge and Explanatory Gaps

Robert Van Gulick Do phenomenal mental states pose a special ob­ stacle to materialism or functionalism? Three main families of arguments in the recent philo­ sophical literature may seem to show that they do: the Knowledge Argument, the Explanatory Gap Argument, and the various versions of the Inverted and Absent Qualia Argument. I shall here discuss the first two of these and show that neither presents an insurmountable barrier to materialistic functionalism. Elsewhere, in part II, (chap. 25, this volume) I show the same for the third group of arguments. Before turning to the arguments, let me issue one caveat about the use of the word "phenom­ enal." In much of the recent philosophical liter­ ature it has been used to refer exclusively to sensory qualia or so-called raw feels, such as the redness of which one is immediately aware when viewing a ripe tomato, or the taste of a fresh mango. Although I will myself spend some time discussing just such sensory qualia,

I

nonetheless

believe it is a serious mistake to equate the phe­ nm.lenal aspect of mind solely with such prop­ erties. We should not forget that the idea of

�

the phenomenal structure of experience ente ed philosophical thought through Kant, who m­ troduced it in the context of rejecting the sen­ sational theory of experience associated with traditional empiricism. Phenomenal experience is not merely a succession of qualitatively dis­ tinguished sensory ideas, but rather the organized cognitive experience of a world of objects and of ourselves as subjects within that world. Any adequate theory of the phenomenal aspect of mentality should take this richer Kantian con­ cept into account. To focus exclusively on raw feels would be a mistake in at least two respects. First, it would provide too narrow a definition of what needs to be explained, and, secondly, I doubt that qualia and raw feels can themselves be understood in isolation from the roles they play within the richer Kantian structure of phenom­ enal experience.

The Knowledge Argument Let us tum then to the Knowledge Argument. Its basic underlying assumption is that there is some knowledge about experience that can be acquired

only

by undergoing the relevant experience one­

self. In the paradigm case, one can come to know what the character of phenomenal red is only by having a red experience. No physical knowl­ edge of what goes on in the brain when one has a red experience would suffice. It is for this reason that Thomas Nagel believes that no human can ever know what it's like to be a bat

(see chapter

�­

32). Given our human inability to undergo ex

riences of the sort the bat has when sensmg its surroundings by echo location, the relevant knowledge about bat-type experience is forever beyond us. The relevant facts about what its like to be a bat are cognitively inaccessible to us in the sense that we are incapable of even understanding them.

. To see how this is supposed to lead to an antI­

materialist conclusion, let us consider the knowl­ edge argument as presented by Frank Jackson ( 1 982, 1 986), certainly the most widel

�

disc�d

anti-physicalist argument in the Amencan phIlo­

�

sophical world during the 1 980's. Jackson offe s the hypothetical case of Mary the super color SCI­ entist who has spent her entire life within a strictly black and white (and gray) environment (see chapter 36). Mary nonetheless (via television) has become the world's greatest expert on color perception and

ex hypothesi she is said to know everything physical there is to know about ,,:hat

goes on in a normal perceiver when he perceives something red. Yet, Jackson argues, Mary does not know

everything there is to know about hav­

ing a red experience, a fact that he believes is ob­ vious if we consider what would happen if Mary were released from her achromatic isolation and shown a ripe tomato for the first time. Jackson's claim is that Mary would come to know some-

Robert Van Gulick

560

thing that she didn't know before. But since

ex everything physical

when they experience red. Since our present

there was to know about seeing red, the knowl­

knowledge of the brain is so far short of what

hypothesi

she already knew

to know about what goes on in peoples' brains

edge or information she gains must be non­

Mary would know, it is difficult to say what she

physical phenomenal information.

would or would not be able to understand or an­

AI

ticipate. It seems at least possible that when Mary

The Knowledge Argument

sees her first tomato (rather than expressing sur­

P. I . Mary (before her release) knows

physical there is to know

everything

about seeing red.

P.2. Mary (before her release) does not know

everything

prise) she might remark, "Ah yes, it is just

as

I

expected it would be." Thus there is the possi­

there is to know about seeing red

because she learns something about it on her release.

bility of undercutting the argument right at the start by simply rejecting its first premise. But let us follow most philosophers in conceding that Mary gains at least some knowledge and push on to question 2.

Therefore C.3. There are some truths about seeing red that

Question 2:

escape the physicalist story.

gain? Is it strictly know-how or does it include new

C.4. Physicalism is false and phenomenal prop­ erties cannot be explained as (or identified with) physical properties. This argument has been regarded

knowledge of/acts, propositions or in/ormation? One reply to the knowledge argument, origi­ nally proposed by Lawrence Nemirow (1 980,

as

a serious

threat to physicalism. Indeed so arde.nt a physi­ calist and formidable a philosopher

as

David

Lewis (see chapter 38) has held that if we admit that Mary really gains information when she first experiences red, then physicalism must be false. However, being regarded as a serious threat is not the same as being regarded as a sound argument, and a wide variety of critical objections have been raised against the knowledge argument. We

can

classify them into groups by use of a few diag­ nostic questions about Mary (figure 3 5. 1 ). They are all variants of the question that riveted the Watergate investigations of Richard Nixon:

What

did she know and when did she know it? Question 1 :

What sort of knowledge does Mary

gains is know-how; she gains no new knowledge of facts or propositions. According to this so-called

ability reply, she gains only new practical abilities to recognize and imagine the relevant phenom­ enal properites. If the argument's second premise P. 1 . is thus read as saying only that Mary gains new abilities (new know-how) then its conclusion C. 3 . no longer follows. There need be no truths or information left out of the physicalist story or out of Mary's prior knowledge. The ability reply thus promises a quick and clean solution to the knowledge argument. But its viability depends on the plausibility of the claim that Mary gains no new knowledge of facts or propositions, and like

Does Mary in fact learn anything

or gain any knowledge when she first experiences red?

many other philosophers I find that claim not very plausible. Part of what Mary gains is know­ how, but that does not seem to be all she gains.

Most philosophers have been willing to con­ cede that Mary does learn something, but Paul Churchland (1 985) has argued that the claim is open to reasonable doubt. Remember that

1 990) and championed by David Lewis ( 1 983, 1 988) is to hold that the only knowledge Mary

ex­

hypothesi Mary knows everything physical there is

There seems to be a fact about how phenomenal red appears that she apprehends only after her release. Let us thus tum to a similar portantly different question.

-

sounding but

im­

01

Does Mary learn anything or gain any knowledge when she first experiences red1

(Possibly) Not Churchland

P.

Y'

Yes

•

I

02 What sort of knowledge does Mary gain1 1s it strictly know-how or does it include new knowledge of facts, propositions or information1

Strictly know-how. No new k �,:>wledge of facts or propositions. She acquires only new abilities to recognize and imagine. L. Nemirow D. Lewis

V�

She gains new knowledge of facts and propositions.

•

J

Does Mary come to know new facts new propositions, or come to possess new information1

03

�� No. she only comes to know in a new way facts or propositions she already knew. P. Churchland M. Tye T. Horgan Physical ism not refuted.

I

Yes she does.

..t. 04

On what rnode of individuating propositions does Ma;'y learn a new proposition1

¥" Only on a fine-grained mode of individuating propositions that takes into account conceptual constituent structure. W. Lycan B. Loar R. Van Gulick Physicalism not refuted.

Figure 35.1 Classification of critical objections raised against the knowledge argument

I �

On a coarse-grained mode of individuation. F. Jackson Physicalism refuted.

Robert Van Gulick

Question 3:

562

Does Mary come to know

facts, new propositions, information?

new new

or come to possess

linguistic and biological nonlinguistic systems for representing experiential states seem far greater

The difference between questions 2 and 3 is a change in the scope of the word "new." In 2 it

knowledge), but in 3 it qualifies the object of knowledge (knowledge of new propositions or new facts). The difference qualifies knowledge (new

can be significant since there are philosophers who hold that Mary gains new knowledge only in the sense that she comes to know

unfair insofar as the differences between Mary's

in a new way

facts and propositions that she already knew. (Churchland 1 985; Tye 1 986) There are various options for unpacking the notion of knowing an

directly introspection what · she knew before only in­ directly by inference. Mary is now able to repre­ sent such facts to herself using a basic biological and probably pre-linguistic system of representa­ tion quite distinct from the linguistic representa­ tions she had to use to represent such facts in the past. Such differences in mode of access or system of representation might justify us in saying that old fact in a new way. Mary now knows

by

Mary was now in a new epistemic state and might suffice to explain our inclination to accept the claim that Mary gains new knowledge. Read in this way premise 2 again fails to support the ar­ gument's conclusion; there need be no facts or propositions that Mary failed to know before her release and thus none that were left out by the physicalist story. David Lewis ( 1 988) has recently criticized such approaches as inadequate to account for the in­ tuitive appeal of the knowledge argument. He argues that they make the sense in which Mary gains new knowledge from experiencing red no different from that in which she gains new knowledge about her brain when she learns Rus­ sian or Urdu and thus acquires a new system of representation. He finds it uncharitable to sup­ pose the proponents of the knowledge argument have confused so innocuous a sense of "new knowledge" with any that would support their conclusion. Lewis' criticism strikes me as a bit

than those between English and Russian and thus better able to account for our sense that Mary has gained new knowledge,

even if she

has not come

to know a new proposition. There is however one last question we need to ask, one which can allow us to concede even the strong claim that Mary learns a new proposi­ tion without being forced to an anti-physicalist conclusion. Question 4: On what mode of individuating propositions does Mary learn a new proposition? (That is what counts here as a new proposition?) Propositions, like beliefs, can be individuated in a variety of fine- or coarse-grained ways. Coarse-grained propositions might be taken as functions from possible worlds to truth-values. On such a mode of individuation, the proposition that 5 +

7

=

1 2 is the same proposition as the

proposition that 38 is the square root of 1 ,444 , they are both true in every possible world. And the proposition that water freezes at 32 degrees Fahrenheit is the same proposition as the propo­ sition that H20 freezes at 32 degrees Fahrenheit. However, one can use a more fine-grained scheme of individuation which treats propositions as having

constituent structure composed of such concepts that must also match up if two

things as

propositions are to be identified. Thus

even if Mary

has come to know a new

proposition, that in itself need not undercut materialism

as

long

as propositions are being

individuated in a sufficiently fine-grained way. ( Lycan 1 9 90; Loar 1 990) I find this reply to the knowledge argument quite attractive especially as developed in a recent paper by Brian Loar ( 1 990). Indeed I have made a similar argument in the past in a reply to Thomas Nagel's argument regarding the individuation of facts ( 1 985). Loar argues that Mary acquires a new concept, a concept that en­ ters her cognitive repertoire in part on the basis of

Understanding the Phenomenal Mind, Part I

her newly acquired discriminative abilities. Thus using this new concept she is able to apprehend the truth of new propositions. Yet the addition of such a new proposition to her store of knowledge need not cause any concern to the physicalist insofar as the properly to which her new concept refers can be just some property she referred to in the past by use of a purely physical

concept,

that

563

fiber firing that makes it naturally "fit" the phenomenal properties of pain, any more than it would fit some other set of phenomenal properties. Unlike its functional role, the identification of the qualitative side of pain with C­ fiber firing (or some property ofC-fiber firing) leaves the connection between it and what we identify it with completely mysterious. One might say it makes the way pain feels into merely a brute fact. (p. 358) Levine maintains that psychophysical state­

is, a concept constructed within the resources of

ments asserting such brute fact identities are un­

the physical sciences. What then is the bottom line on the knowledge argument? Despite its widespread intuitive appeal

intelligible; they leave an explanatory gap that we have no idea how to fill.

and the air of mystification it produces about the

In chapter 33 Colin McGinn argues that mak­

explanatory elusiveness of phenomenal qualities,

ing the psychophysical link intelligible may be

I think it's pretty clearly a

loser as

an argument

beyond our conceptual capacities. We humans

against the possibility of giving a materialist ex­

should not with hubris assume that every fact

planation of phenomenal mentality. There are

about the natural world is within our cognitive

any number of points at which one can cut off the

capacity to comprehend. Just as monkeys are un­

argument. Though I favor a Loar-type solution, I

able to comprehend the concept of an electron,

think each of the other replies provides a plausible

and armadillos (those ugly armored ant-eating

place at which to draw the line against the anti­

survivors of an earlier biological era that still

materialist attack. Please choose your favorite.

inhabit Texas and points south) are not even up to the task of doing elementary arithmetic. So too we may all be armadillos when it comes to

The Explanatory Gap Argument

understanding the link between brain and phe­

Let us tum then to our second argument, the Explanatory Gap Argument. It aims at a more modest though still substantiative result. It does not try to show that materialism

isfalse, but only

that with respect to the phenomenal aspect of

unin­ incapable of being adequately com­

mind, materialism is in an important sense

telligible or prehended, at

least by us humans. Thus Thomas

Nagel writes, "We have at present no conception of how a single event or thing could have both physical and phenomenological aspects or how if it did they might be related" ( 1 986, p. 47). Joseph Levine (1 983) puts the point like this:

. . . there is more to our concept of pain than its causal role, there is how it feels; and what is left unexplained by the discovery of C-fiber firing (the standard philosoph­ ical candidate for the neural basis of pain, despite its total empirical implausibility) is why pain should feel the way it does! For there seems to be nothing about C-

nomenal mentality.

Or so McGinn suggests.

The explanatory gap argument is not an argu­ ment in the sense that the knowledge argument is. Its strength derives not from any path of de­ ductive reasoning but rather from the intuitive appeal of its conclusion, which speaks to our be­ wilderment about how any physical story about the brain could ever explain phenomenal con­ sciousness. Recall Wittgenstein holding his head in the

"

THIS

Philosophical Investigations

and saying,

is supposed to be produced by a process in

the brain!" (1 953, p. 4 1 2). Nonetheless, I want to consider one deductivt; reconstruction of the argument because I think it will help us see what needs to be done to defuse the intuitive appeal of the position. At one point, Joe Levine makes an approving reference to John Locke's seventeenth-century claim that sensory qualia are arbitrary.

Robert Van Gulick

564

The point I am trying to make was captured by Locke in his discussion of the relation between primary and sec­ ondary qualities. He states that the simple ideas which we experience in response to impingements from the external world bear no intelligible relation to the cor­ puscular processes underlying impingement and re­ sponse. Rather the two sets of phenomena are stuck together in an arbitrary manner. (p. 359). Immediately following this passage Levine at­ tempts to support his point by appeal to a stan­ dard philosophical case of hypothetical spectrum inversion with red and green qualia switching causal roles in an otherwise normal subject. What is important here is the suggestion that basic ideas, such as color qualia, are simples. They have no structure, and since each one is what it is sui generis, it is hard to see how their connection to anything could fail to be anything but arbitrary. We might reconstruct the gap argument as A2. A2

The Explanatory Gap Argument

qualia provides the basis for establishing

planatory connections

ex­

between them and their

neural substrates. The method is the familiar one of explaining higher-order organization in terms of underlying structure. Consider a few brief examples. 1 . Some color qualia are phenomenally expe­ rienced as binary (e.g., orange and purple) and others as unary (e.g., red and blue) while other combinations are phenomenally impossible, (e.g. , a color that is both red and green in the way pur­ ple is both red and blue.) This phenomenal orga­ nization is explained by appeal to the existence of two underlying opponent color channels, one subserving red/green discrimination and the other yellow/blue. Since red and green are mutually exclusive extreme outputs of the red/green chan­ nel they cannot be combined; binaries are always combinations of the outputs of the two distinct channels.

simples; they have no structure.

2. Some hues are experienced as warm, pos­ itive, and advancing (red and yellow) and others as cool, negative, and receding (green and blue).

Thus

They have an affective dimension. The warm hues

P. l . Qualia such as phenomenal hues are basic

C.2. Any links between such qualia and the or­ ganizational structure of their neural substrates must be arbitrary.

their respective channel, while the cool hues result from decreased stimulation. 3. Phenomenal yellow tends to be "captured"

Thus C . 3 . The links between qualia and their neural bases are unintelligible and present us with an unfil lable explanatory gap.

by phenomenal green in the sense that hues inter­ mediate between yellow and green tend to be perceived as shades of green (rather than of yellow) far more quickly as one moves away from

Formulating the argument in this way suggests a possible line of reply, one that has in fact been made by Larry Hardin ( 1 988). It is the first premise of A2 that we must reject. Hardin argues the phenomenal hues that philosophers such as Locke and Levine regard as

sui generis simples

are not in fact such, but rather elements within a highly organized and structured color space. Any attempt to invert them or interchange them

in undetectable ways would

are those that result from increased stimulation in

have to preserve that

structural organization. More impotant the artic­ ulation of an organized structure among color

unary yellow than when one moves away from any other unary hue. This phenomenon has as yet no neural explanation. (All examples from Hardin 1 988, pp. 1 34-42.) The three examples allow us to make our two points. First, most (perhaps even all) of the hypothetical spectrum inversions considered by philosophers would disrupt the organization of the phenomenal color space in detectable ways. One cannot preserve structure while interchang­ ing red with orange or with any other binary hue, nor by interchanging red with the cool hues of green or blue, nor red with yellow given the

Understanding the Phenomenal Mind, Part I

phenomenon of capture. Secondly and of more immediate relevance to the explanatory gap ar­ gument, the phenomenal color space is revealed to have a complex organizational structure that allows us to establish explanatory rather than simply brute fact connections between it and underlying neural processes. Our critic may still reply that any explanations we produce will still leave something essentially qualitative unexplained. No matter how much structural organization we can find in the phe­ nomenal realm and explain neurophysiologically, she will insist that the distinct redness of phe­ nomenal red will not have been captured or ex­ plained by our theory. For though it may be impossible to map the hues of our color space in a way that rearranges them while making the change behaviorally undetectable by preserving all their organizational relations, it seems there could be other nonhuman creatures who had alien qualia quite unlike our own, but whose in­ terrelations exactly mirrored those among our phenomenal hues (Shoemaker 1981). Red"', their correlate of red, would bear to green'" and or­ ange'" the same relations that red bears to our phenomenal greens and oranges; yet red'" would not be the same as red. If so, then something es­ sentially qualitative would seem to have escaped our explanation (Shoemaker 1 975, 198 I a). We might have explained why a given brain process is identical with experiencing red rather than with experiencing orange but not why it is a case of experiencing red rather than red"'. This countercharge needs to be taken seriously, but I think we have altered the nature of the debate. The more one can articulate structure within the phenomenal realm, the greater the chances for physical explanation; without struc­ ture we have no place to attach our explanatory "hooks." There is indeed a residue that continues to escape explanation, but the more we can ex­ plain relationally about the phenomenal realm, the more the leftover residue shrinks toward zero. Though I admit that we are as yet a long way from that.

565

Conclusions Neither of the arguments we have considered succeeds in giving a sound a priori reason for concluding that the phenomenal aspect of mind cannot be adequately explained in physical or functional terms. First the knowledge argument turns out to be vulnerable in a wide variety of ways. Second, as long as phenomenal properties are not thought of as basic simples, filling the explanatory gap between the phenonemal and the physical (or functional) remains an option by continuing the step-by-step process of articulating struCture within the phenomenal realm and map­ ping it onto the structure of underlying non­ phenomenal processes. Notes This chapter is the first half of "Understanding the Phenomenal Mind: Are We All Just Armadillos?" which was originally published in Consciousness-A Mind and Language Reader, edited by Martin Davies and Glynn Humphries, Basil Blackwell, Oxford, 1992. That chapter was divided and slightly revised to form two independent parts at the request of the editors of this volume. An earlier version of the complete paper was presented at the conference on The Phenomenal Mind: How Is It Possible? Why Is It Necessary? held at the Center for Interdisciplinary Studies (ZIF), Bielefeld University, Bielefeld, Germany, May 1990.

References Block, Ned (1980) "Troubles with functionalism" in

Readings in the Philosophy of Psychology, Vol. I , N. Block (ed.) Harvard University Press, Cambridge, MA, 268-305. Churchland, Paul (1985) "Reduction, qualia and the direct introspection of brain states," Journal of Philoso­

phy, 2-28. Flohr, Hans (in press) "Brain processes and conscious­ ness: a new and specific hypothesis." Philosophical Psy­

chology.

Robert Van Gulick

Hardin, C. L. ( 1 988) Color for Philosophers, Hackett Publishers, Indianapolis, In. Jackson, Frank (1982) "Epiphenomenal qua1ia," Amer­ ican Philosophical Quarterly, 127-36. Johnson-Laird, Philip (1 983) Mental Models, Cam­ bridge University Press, Cambridge. Johnson-Laird, Philip (1988) "A computational analy­ sis of consciousness" in Consciousness in Contemporary Science, A. Marcel and E. Bisiach (eds.), Oxford Uni­ versity Press, Oxford, 357-68. Levine, Joseph (1983) "Materialism and qualia: the ex­ planatory gap," Pacific Philosophical Quarterly, 35461. Lewis, David (1983) "Postscript t o 'Mad pain and martian pain' " in D. Lewis Philosophical Papers, Vol. I, Oxford University Press, Oxford, Lorenz, Konrad (1973) Behind the Mirror, Harcourt Brace Jovanovich, New York. Lycan, William (1 990) "What is the subjectivity of the mental?" in Philosopical Perspectives, Vol. 4. Action Theory and Philosophy of Mind, Ridgeview Publishing, Atascadero, CA, 1 09-30. Marcel, Anthony (1988) "Phenomenal experience and functionalism" in Consciousness in Contemporary Sci­ ence, A. Marcel and E. Bisiach (eds.), Oxford Uni­ versity Press, Oxford, 1 2 1 -58. Nernirow, Lawrence (1980) Review ofT. Nagel Mortal Questions, Philosophical Review, 475-76. Nemirow, Lawrence (1990) "Physicalism and the cog­ nitive role of acquaintance" in Mind and Cognition, W. Lycan (ed.) Basil Blackwell Publishers, Oxford, 490-99. Tye, Michael (1 986) "The subjectivity of experience,"

Mind, I -17. Sacks, Oliver (1984) A Leg To Stand On, Summit Books, New York. Shoemaker, Sidney (1975) "Functionalism and qualia," Philosophical Studies, 291 - 3 1 5. Shoemaker, Sidney (198Ia) "Absent qualia are impos­ sible·-a reply to Block," Philosophical Review, 581 -99. Van Gulick, Robert (1985) "Physicalism and the sub­ jectivity of the mental," Philosophical Topics, 5 1 -70. Van Gulick, Robert (1988a) "A functionalist plea for self-consciousness," Philosophical Review, 149-8 1 . Van Gulick, Robert ( l988b) "Consciousness, intrinsic intentionality and self-understanding machines" in

566

Consciousness in Contemporary Science, E. Bisiach & A. Marcel (eds.), Oxford University Press, Oxford, 78100. Van Gulick, Robert (1 989) "What difference does con­ sciousness make?," Philosophical Topics, 2 1 1 -30.

36

What Mary Didn't Know

Frank Jackson Mary is confined to a black-and-white room, is

what sensing red is like unless you have sensed

educated through black-and-white books and

red. Powers of imagination are not to the point.

through lectures relayed on black-and-white tele­

The contention about Mary is not that, despite

vision. In this way she learns everything there is

her fantastic grasp of neurophysiology and every­

to know about the physical nature of the world.

thing else physical, she could not imagine what it is

She knows all the physical facts about us and our

like to sense red; it is that, as a matter of fact, she

environment, in a wide sense of 'physical' which

would not know.

completed physics,

But if physicalism is true, she

chem­

would know; and no great powers of imagination

istry, and neurophysiology, and all there is to

would be called for. Imagination is a faculty that

know about the causal and relational facts con­

those who

includes everything in

lack knowledge need to fall back on.

sequent upon all this, including of course func­

Second, the intensionality of knowledge is not

tional roles. If physicalism is true, she knows all

to the point. The argument does not rest on as­

there is to know. For to suppose otherwise is to

suming falsely that, if S knows that

suppose that there is more to know than every

a

=

b,

then

S knows that b

is

F.

a

is

F and

if

It is concerned

physical fact, and that is just what physicalism

with the nature of Mary's total body of knowl­

denies.

edge before she is released: is it complete, or do

Physicalism is not the noncontroversial thesis

S F and know that a = b, yet know that b is F, by virtue of

some facts escape it? What is to the point is that

that the actual world is largely physical, but the

may know that a is

challenging thesis that it is entirely physical. This

arguably not

is why physicalists must hold that complete

not being sufficiently logically alert to follow the

physical knowledge is complete knowledge sim­

consequences through. If Mary's lack of knowl­

pliciter. For suppose it is not complete: then our

edge were at all like this, there would be no threat

world must differ from a world, W(P), for which

to physicalism in it. But it is very hard to believe

it is complete, and the difference must be in non­

that her lack of knowledge could be remedied

physical facts; for our world and W(P) agree in

merely

all matters physical. Hence, physicalism would be

enough logical consequences of her vast physical

false at our world [though contingently so, for it

would be true at W(p)]. l

It seems, however, that Mary does not know all

by

her

explicitly

following

through

knowledge. Endowing her with great logical acu­ men and persistence is not in itself enough to fill in the gaps in her knowledge. On being let out,

"I could have worked all this

there is to know. For when she is let out of the

she will not say

black-and-white room or given a color television,

out before by making some more purely logical

she will learn what it is like to see something red,

inferences. "

say. This is rightly described as learning-she will

Third, the knowledge Mary lacked which is of

not say "ho, hum." Hence, physicalism is false.

particular point for the knowledge argument

This is the knowledge argument against physi­ calism in one of its manifestations. 2 This note is

against physicalism is knowledge about the expe­ riences of others, not about her own. When she is

a reply to three objections to it mounted by Paul

let out, she has new experiences, color experiences ' she has never had before. It is not, therefore, an

M. Churchland.3

objection to physicalism that she learns something

I Three Clarifications The knowledge argument does not rest on the dubious claim that logically you cannot imagine

on being let out. Before she was let out, she could not have known facts about her experience of red, for there were no such facts to know. That physi­ calist and nonphysicalist alike can agree on. After

Frank Jackson

568

she is let out, things change; and physicalism can

ever the tags, he is right that the displayed argu­

happily admit that she learns this; after all, some

ment involves a highly dubious use of Leibniz's

physical things will change, for instance, her

law.

brain states and their functional roles. The trou­

My reply is that the displayed argument may

ble for physicalism is that, after Mary sees her

be convenient, but it is not accurate. It is not the

first ripe tomato, she will realize how impov­

knowledge argument. Take, for instance, premise

erished her conception of the mental life of others has been

all along.

She will realize that there was,

1 . The whole thrust of the knowledge argument is that Mary (before her release) does

not

know

all the time she was carrying out her laborious

everything there is to know about brain states

investigations

and their properties, because she does not know

into

the

neurophysiologies

of

others and into the functional roles of their in­

about certain qualia associated with them. What

ternal states, something about these people she

is complete, according to the argument, is her

was quite unaware of. All along their experiences

knowledge of matters physical. A convenient and

(or many of them, those got from tomatoes, the

accurate way of displaying the argument is:

sky, . . . ) had a feature conspicuous to them but until now hidden from her (in fact, not in logic). But she knew all the physical facts about them all along; hence, what she did not know until her re­ lease is not a physical fact about their experiences. But it is a fact about them. That is the trouble for physicalism.

II

(\)' Mary (before her release) knows everything physi­ cal there is to know about other people. (2)' Mary (before her release) does not know everything there is to know about other people (because she learns something about them on her release). Therefore,

(3)' There are truths about other people (and herself) which escape the physicalist story.

Churchland's Three Objections

What is immediately to the point is not the

Churchland's first objection is that the knowledge argument contains a defect that "is simplicity itself" (p. 23). The argument equivocates on the sense of 'knows about' . How so? Churchland suggests that the following is "a conveniently tightened version" of the knowledge argument:

kind, manner, or type of knowledge Mary has, but

what she knows. What she knows beforehand

is ex hypothesi everything physical there is to know, but is it everything there is to know? That is the crucial question. There is, though, a relevant challenge involving questions about kinds of knOWledge. It concerns the support for premise 2'. The case for premise 2'

(1) Mary knows everything there is to know about brain states and their properties.

is that Mary learns something on her release,

(2) It is not the case that Mary knows everything there is to know about sensations and their properties.

her knowledge beforehand

Therefore, by Leibniz's law,

mind

(3) Sensations and their properties i- brain states and their properties. (p. 23) Churchland observes, plausibly enough, that the type or kind of knowledge involved in premise

I is

distinct from the kind of knowledge involved in premise

2. We might follow his lead and tag

the first 'knowledge by description', and the second 'knowledge by acquaintance'; but, what-

she acquires knowledge, and that entails that whether

by

(what she knew, never

description,

acquaintance,

or whatever) was incomplete. The challenge, mounted by David Lewis and Laurence Nem­ irow, is that on her release Mary does

not

learn

something or acquire knowledge in the relevant sense. What Mary acquires when she is released is a certain representational or imaginative ability; it is knowledge how rather than knowledge that. Hence, a physicalist can admit that Mary ac­ quires something very significant of a knowledge

What Mary Didn't Know

kind-which can hardly be denied-without admitting that this shows that her earlier factual knowledge is defective. She knew all that there was to know about the experiences of others beforehand, but lacked an ability until after her release.4 Now it is certainly true that Mary will acquire abilities of various kinds after her release. She will, for instance, be able to imagine what seeing red is like, be able to remember what it is like, and be able to understand why her friends regarded her as so deprived (something which, until her release, had always mystified her). But is it plausible that that is all she will acquire? Suppose she received a lecture on skepticism about other minds while she was incarcerated. On her release she sees a ripe tomato in normal conditions, and so has a sensation of red. Her first reaction is to say that she now knows more about the kind of experiences others have when looking at ripe tomatoes. She then remembers the lecture and starts to worry. Does she really know more about what their experiences are like, or is she indulging in a wild generalization from one case? In the end she decides she does know, and that skepticism is mistaken (even if, like so many of us, she is not sure how to demonstrate its errors). What was she to-ing and fro-ing about-her abilities? Surely not; her representational abilities were a known constant throughout. What else then was she ag­ onizing about than whether or not she had gained factual knowledge of others? There would be nothing to agonize about if ability was all she ac­ quired on her release. I grant that I have no proD/that Mary acquires on her release, as well as abilities, factual knowl­ edge about the experiences of others-and not just because I have no disproof of skepticism. My claim is that the knowledge argument is a valid argument from highly plausible, though admit­ tedly not demonstrable, premises to the conclu­ sion that physicalism is false. And that, after all, is about as good an objection as one could expect in this area of philosophy.

569

Churchland's second objection (pp. 24/5) is that there must be something wrong with the ar­ gument, for it proves too much. Suppose Mary received a special series of lectures over her black­ and-white television from a full-blown dualist, explaining the "laws" governing the behavior of "ectoplasm" and telling her about qualia. This would not affect the plausibility of the claim that on her release she learns something. So if the argument works against physicalism, it works against dualism too. My reply is that lectures about qualia over black-and-white television do not tell Mary all there is to know about qualia. They may tell her some things about qualia, for instance, that they do not appear in the physicalist's story, and that the quale we use 'yellow' for is nearly as different from the one we use 'blue' for as is white from black. But why should it be supposed that they tell her everything about qualia? On the other hand, it is plausible that lectures over black-and­ white television might in principle tell Mary everything in the physicalist's story. You do not need color television to learn physics or function­ alist psychology. To obtain a good argument against dualism (attribute dualism; ectoplasm is a bit of fun), the premise in the knowledge argu­ ment that Mary has the full story according to physicalism before her release, has to be replaced by a premise that she has the full story according to dualism. The former is plausible; the latter is not. Hence, there is no "parity of reasons" trou­ ble for dualists who use the knowledge argument. Churchland's third objection is that the knowl­ edge argument claims "that Mary could not even imagine what the relevant experience would be like, despite her exhaustive neuroscientific knowl­ edge, and hence must still be missing certain cru­ cial information" (p. 25), a claim he goes on to argue against. But, as we emphasized earlier, the knowledge argument claims that Mary would not know what the relevant experience is like. What she could imagine is another matter. If her knowledge is

Frank Jackson

defective, despite being all there is to know ac­ cording to physicalism, then physicalism is false, whatever her powers of imagination. Notes I am much indebted to discussions with David Lewis and with Robert Pargetter.

I . The claim here is not that, if physicalism is true, only what is expressed in explicitly physical language is an item of knowledge. It is that, if physicalism is true, then if you know everything expressed or expressible in ex­ plicitly physical language, you know everything. Pace Terence Horgan, "Jackson on Physical Information and Qualia," Philosophical Quarterly, 34, 135 (April 1 984): 147- 1 52. 2. Namely, that in my "Epiphenomenal Qualia," ibid., XXXI I, 1 27 (April 1 982): 1 27-1 36. See also Thomas Nagel, "What Is It Like to Be a Bat?," Philosophical Review, 83, 4 (October 1 974): 435-450, and Howard Robinson, Matter and Sense (New York: Cambridge, 1 982). 3. "Reduction, Qualia, and the Direct Introspection of Brain States." The JOUI7UJI of Philosophy, (January 1 985): 8-28. Future page references are to this paper. 4. See Laurence Nemirow, review of Thomas Nagel,

Mortal Questions, Philosophical Review 89, 3 (July 1 980): 473-477, and David Lewis, "Postscript to 'Mad Pain and Martian Pain'," Philosophical Papers, vol. 1 (New York: Oxford, 1983). Churchland mentions both Nemirow and Lewis, and it may be that he intended his objection to be essentially the one I have just given. However, he says quite explicitly (bottom of p. 23) that his objection does not need an "ability" analysis of the relevant knowledge.

570

37

Knowing Qualia: A Reply to Jackson

Paul M. Churchland

m = Mary; Kyx = y knows about x; Hx x Px x is about something physical in character; and x ranges over "know­

In a paper concerning the direct introspection

Here

of brain states (1 985) I leveled three criticisms

is about persons;

against Frank Jackson's "knowledge argument."

=

=

At stake was his bold claim that no materialist

abies," generously construed so as not to beg any

account of mind can possibly account for all

questions about whether they are propositional or

mental phenomena. Jackson has replied to those

otherwise in nature.

criticisms in his 1 986. It is to those replies, and to

Thus expressed, the argument is formally valid:

the issues that prompted them, that the present

the salient move is a modus tollens that applies the

chapter is directed. [Jackson's essay is chapter 36

second conjunct of premise (2),

in this book.]

waiting consequent of premise ( I ),

" � Kmx," to the "Kmx." The

questions now are whether the premises are jointly true, and whether the crucial notion

The Persistent Equivocation

"Kmx"

Jackson concedes the criticism I leveled at my own statement of his argument-·-specifically, that it involves an equivocation on "knows about"

-­

but he insists that my reconstruction does not represent the argument he wishes to defend. I ac­ cept his instruction, and tum my attention to the summary of the argument he provides in section II of his paper (page 568, this volume). Mary, you will recall, has been raised in innocence of any color experience, but has an exhaustive command of neuroscience.

is univocal in both of its appearances.

Here I am surprised that Jackson sees any prog­ ress at all with the above formulation, since I continue to see the same equivocation found in my earlier casting of his argument. Specifically,

premise

( I ) is plausibly

true,

within Jackson's story about Mary's color-free upbringing, only on the interpretation of "knows about" that casts the object of knowledge as something propositional, as something adequately expressible in an English sentence. Mary, to put it briefly, gets 1 00 percent on every written and oral exam; she can pronounce on the truth of any

( I ) Mary (before her release) knows everything physical there is to know about other people. (2) Mary (before her release) does not know everything there is to know about other people (because she learns somerhing about them on her release).

given sentence about the physical characteristics of persons, especially the states of their brains. Her "knowledge by description" of physical facts about persons is without lacunae. Premise (2), however, is plausibly true only on

Therefore

the interpretation of "knows about" that casts

(3) There are truths about other people (and herself) which escape the physicalist story.

positional, as something inarticulable, as some­

Regimenting further, for clarity's sake, yields the

missing is some form of "knowledge by ac­

following. (I)

thing that is nontruth-valuable. What Mary is quaintance, " acquaintance with a sensory char­ acter, prototype, or universal, perhaps.

(x) [(Hx & Px) :::) Kmx]

(2) (3x) [Hx &

the object of knowledge as something nonpro­

�

Kmx] (viz. ,

"what it is like to see

Given this prima facie difference in the sense of "knows about," or the kind of knowledge

red")

appearing in each premise, we are still looking at

Therefore

a prima facie case of an argument invalid by rea­

(3)

(3x)[Hx & � Px]

son of equivocation on a critical term. Replace

Paul M. Churchland

572

either of the "K"s above with a distinct letter, as

A

contrast between

"knowing how"

and

acknowledgment of the ambiguity demands, and

"knowing that" is one already acknowledged in

the inference to (3) evaporates. The burden of

common sense, and thus it is not surprising that

articulating some specific and unitary sense of

some of the earliest replies to Jackson's argument

"knows about," and of argwng that both prem­

(Nerriirow 1 980; Lewis 1 983) tried to portray its

ises are true under that interpretation of the epis­

equivocation in these familiar terms, and tried to

temic operator, is an undischarged burden that

explicate Mary's missing knowledge solely in

still belongs to Jackson.

terms of her missing some one or more

It is also a heavy burden, since the resources of

abilities

(to recognize red, to imagine red, etc.). While the

modem cognitive neurobiology already provide

approach is well motivated,

us with a plausible account of what the difference

tinction in types of knowledge barely begins to

this binary dis­

in the two kinds of knowledge amounts to,

suggest the range and variety of different sites and

and of how it is possible to have the one kind

types of internal representation to be found in a

without the other. Let me illustrate with a case

normal brain. There is no reason why we must be

distinct from that at issue, so as not to beg any

bound by the crude divisions of our prescientific

questions.

idioms when we attempt to give a precise and

Any competent golfer has a detailed repre­

positive explication of the equivocation displayed

sentation (perhaps in his cerebellum, perhaps in

in Jackson's argument. And there are substantial

motor

grounds for telling a somewhat different story

his motor cortex) of a golf swing. It is a

representation, and it constitutes his "knowing

concerning the sort of nondiscursive knowledge

how" to execute a proper swing. The same golfer

at issue. Putting caution and qualification mo­

will also have a discursive representation of a golf

mentarily aside, I shall tell such a story.

swing (perhaps in his language cortex, or in

In creatures with trichromatic vision (i.e., with

the neighboring temporal and parietal regions),

three types of retinal cone), color information is

which allows him to describe a golf swing or

coded as a pattern of spiking frequencies across

perhaps draw it on paper. The motor and the

the axonal fibers of the parvocellular subsystem

discursive

distinct.

of the optic nerve. That massive cable of axons

Localized brain trauma, or surgery, could remove

leads to a second population of cells in a central

representations

are

qwte

either one while sparing the other. Short of that,

body called the lateral geniculate nucleus (LGN),

an inarticulate golf champion might have a

whose axonal projections lead in tum to the sev­

superb representation of the former kind, but

eral areas of the visual cortex at the rear of the

a feeble representation of the latter kind. And a

brain's cerebral hemispheres, to V I , V2, and ulti­

physicist or sports physiologist might have a de­

mately to V4, which area appears to be especially

tailed and penetrating representation of the me­

devoted to the processing and representation of

chanics of a good swing, and yet be unable to duff

color information (Zeki 1 980; Van Essen and

the ball more than ten feet because he lacks an

Maunsell 1 983; Hubel and Livingstone 1 987).

motor representation, of the desired be­

Human cognition divides a smooth continuum of

havioral sequence, in the brain areas that control

color inputs into a finite number of prototypical

his limbs. Indeed, if our physicist is chronically

categories. The laminar structure at V4 is perhaps

adequate

disabled in his motor capacities, he may have no

the earliest place in the processing hierarchy to

motor representation of a golf swing whatsoever.

which we might ascribe that familiar taxonomy.

In one medium of representation, his representa­

A creature competent to make reliable color dis­

tional achievements on the topic may be com­

criminations has there developed a representation

plete; while in another medium of representation,

of the range of familiar colors, a representation

he has nothing.

that appears to consist in a specific configuration

Knowing Qualia: A Reply to Jackson

573

of weighted synaptic connections meeting the

haustive set of discursive, propositional, truth­

millions of neurons that make up area V4.

valuable representations of what goes on in peo­

That configuration of synaptic weights parti­

ple's brains during the experience of color, a set

tions the "activation space" of the neurons in

she has brought into being by the exhaustive

V4: it partitions that abstract space into a

reading of authoritative texts in a completed cog­

structured set of subspaces, one for each proto­

nitive neuroscience. She may even be able to ex­

area

typical color. Inputs from the eye will each

occa

­

plain her own representational deficit, as sketched

sion a specific pattern of activity across these

above, in complete neurophysical detail. Jack­

cortical neurons, a pattern or vector that falls

son's premise

within one of those subspaces. In such a pigeon

true on these wholly materialist assumptions.

( 1), we may thus assume, is also

holing, it now appears, does visual recognition of

The view sketched above is a live candidate for

a color consist. . . . This recognition depends upon

the correct story of sensory coding and sensory

the creature possessing a prior representation­

recognition. But whether or not it is true, it is at

a learned configuration of synapses meeting the

least a logical possibility. Accordingly, what we

relevant popUlation of cells -that antecedently

have sketched here is a consistent but entirely

partitions the creature's visual taxonomy so it can

physical model

respond selectively and appropriately to the flux

conclusion is false) in which both of Jackson's

of visual stimulation arriving from the retina and

premises are true under the appropriate inter­

LGN. This distributed representation is not remotely

(i.e., a model in which Jackson's

pretation. They can hardly entail a conclusion, then, that is inconsistent with physicalism. Their

propositional or discursive, but it is entirely real.

compossibility, on purely physicalist assump­

All trichromatic animals have one, even those

tions, resides in the different character and the

without any linguistic capacity. It apparently

numerically different medium of representation at

makes possible the many abilities we expect from

issue in each of the two premises. Jackson's argu­

color-competent creatures: discrimination, rec­

ment, to refile the charge, equivocates on "knows

ognition, imagination, and so on. Such a repre­

about."

sentation is presumably what a person with Mary's upbringing would lack, or possess only in stunted or incomplete form. Her representational space within the relevant area of neurons would contain only the subspace for black, white, and the intervening shades of gray, for the visual ex­ amples that have shaped her synaptic config­ uration were limited to these. There is thus more than just a clutch of abilities missing in Mary: there is a complex representation-a processing framework that deserves to be called "cogni­ tive"-that she either lacks or has in reduced form. There is indeed something she "does not know." Jackson's premise (2), we may assume, is thus true on these wholly materialist assumptions. These same assumptions are entirely consistent with the further aSsumption that elsewhere in Mary's brain-in the language areas, for exam­ ple-she has stored a detailed and even ex-

Other Invalid Instances An argument form with one invalid instance can be expected to have others. This was the point of a subsidiary objection in my

1 985 paper: if valid,

Jackson's argument, or one formally parallel, would also serve to refute the possibility of sub­

stance dualism. I

did not there express my point

with notable clarity, however, and I accept re­ sponsibility for Jackson's quite missing my in­ tention. Let me try again. The basic point is that the canonical presen­ tation of the knowledge argument, as outlined above, would be just as valid if the predicate term

"P" were everywhere replaced by "E." And the resulting premises would be just as plausibly true if

574

Paul M. Churchland

( I ) "E"

stood for "is about something ecto­

plasmic in character" (where "ectoplasm" is an arbitrary name for the dualist's nonphysical sub­ stance), and

(2) the story is altered so that Mary becomes an ectoplasmic

exhaustive expert on a completed science of human nature.

The plausibility would be comparable, I sub­ mit, because a long discursive lecture on the

cause no amount of discursive knowledge, on any topic, will constitute the nondiscursive form of knowledge that Mary lacks. Jackson's argument is one instance of an indiscriminately antireduc­ tionist form of argument. If it works at all, an analogue will work against any proposed reduc­ tive, discursive, objective account of the nature of our subjective experience, no matter what the re­ ducing theory might happen to be. I

see this as

a further symptom of the logical pathology de­

objective, statable, law-governed properties of

scribed earlier. Since the argument "works" for

ectoplasm, whatever they might be, would be

reasons that have nothing essential to do with

exactly as useful, or

useless, in helping Mary to know-by-acquaintance "what it is like to see red,"

physicalism, it should "work" against the ex­

as would a long discursive lecture on the ob­

And so it "does." The price of embracing Jack­

planatory aspirations of other ontologies as well.

jective, statable, law-governed properties of the

son's argument is thus dramatically higher than

physical matter of the brain. Even if substance

first appears. For it makes any scientific account

dualism were true, therefore, and ectoplasm were

of our sensory experience entirely impossible, no

its heroic principal, an exactly parallel "knowl­

matter what the ontology employed.

edge argument" would "show" that there are some aspects of consciousness that must forever escape the

ectoplasmic

story. Given Jackson's

antiphysicalist intentions, it is at least an irony that the same form of argument should inci­ dentally serve to blow substance dualism out of the water. Though I am hardly a substance dualist (and neither is Jackson), I do regard substance dualism as a theoretical possibility, one that might con­ ceivably succeed in explicating the psychological ontology of common sense in terms of the under­ lying properties and law-governed behavior ofthe nonmaterial substance it postulates. And I must protest that the parallel knowledge argument against substance dualism would be wildly un­ fair, and for the very same reason that its analogue against physicalism is unfair: it would equivocate on "knows about." It would be no more effective against dualism than it is against materialism. The parallel under examination contains a fur­

A Genuinely Nonequivocal Knowledge Argument We can appreciate the equivocation more deeply it we explore a version of Jackson's argument that does

not equivocate on "knows about." The

equivocation can quickly be closed, if we are de­ termined to do so, and the results are revealing. Given that the problem is a variety in the possible forms of knowing, let us simply rewrite the argu­ ment with suitable quantification over the rele­ vant forms of knowing. The first premise must of knowledge, if

x

x.

The second premise is modified in the

same modest fashion, and the conclusion is iden­ tical. Canonically,

(I ' ) (x)(f) [(Hx & Px)

(2') (3x)(3f) [Hx &

-

Therefore

works against physicalism not because of some

(3') (3x) [Hx & - Px]

it works be-

is about humans and

is physical in character, about

ther lesson. If it works at all, Jackson's argument defect that is unique to physicalism;

x, and for any form x then Mary knows (f)

assert that, for any knowable

f

�

K(f)mx]

K(f)mx]

Knowing Qualia: A Reply to Jackson

This argument is also formally valid, and its premises explicitly encompass whatever variety there may be in forms of knowing. What can we say about its soundness? Assume that Mary has had the upbringing de­ scribed in Jackson's story, and thus lacks any knowledge-by-acquaintance with "what it is like to see red." Premise (2') will then be true, as and for the reasons that Jackson's story requires. What will be the truth value of premise ( l ') on these assumptions? Premise ( l ') is now a very strong claim indeed, much stronger than the old premise (I), and a materialist will be sure to insist that it is false. The reason offered will be that, because of her de­ prived upbringing, Mary quite clearly lacks one form of knowledge of a certain physical aspect of people. Specifically, she lacks a proper config­ uration of synaptic connections meeting the neu­ rons in the appropriate area of her visual cortex. She thus lacks an appropriately partitioned acti­ vation vector space across those neurons, and therefore has no representation, at that site, of the full range of sensory coding vectors that might someday come from the retina and the LON. In other words, there is something physical about persons (their color sensations, or identically, their coding vectors in their visual pathways), and there is some form of knowledge (an antecedently partitioned prelinguistic taxonomy), such that Mary lacks that form of knowledge of that aspect of persons. Accordingly, premise ( l ') is false and the conclusion (3') is not sustained. From a materialist's point of view, it is obvious that (t') will be false on the assumptions of Jack­ son's story. For that story denies her the up­ bringing that normally provokes and shapes the development of the relevant representation across the appropriate population of cortical neurons. And so, of course, there is a fonn of knowledge, of a physical aspect of persons, that Mary does not have. As just illustrated, the materialist can even specify that form of knowledge, and its ob­ jects, in neural tenns. But this means that premise (I '), as properly quantified at last, is false. Mary

575

does not have knowledge of everything physical about persons, in every way that is possible for her. (That is why premise (2') is true.) There is, of course, no guarantee that the ma­ terialist's account of sensations and sensory rec­ ognition is correct (although the experimental and theoretical evidence for a view of this general kind continues to accumulate). But neither is Jackson in a position to insist that it must be mistaken. That would beg the very question at issue: whether sensory qualia fonn a metaphysi­ cally distinct class of phenomena beyond the scope of physical science. To summarize, if we write a deliberately non­ equivocal fonn of Jackson's argument, one that quantifies appropriately over all of the relevant forms of knowledge, then the first premise must almost certainly be false under the conditions of his own story. So, at any rate, is the materialist in a strong position to argue. Jackson's expressed hope for "highly plausible premises" is not real­ ized in ( I ' ). The original premise ( I ) was of course much more plausible. But it failed to sustain a valid argument, and it was plausible only because it failed to address all the relevant forms of knowledge.

Converting a Third-Person Account into a First­ Person Account My final objection to Jackson was aimed more at breaking the grip of the ideology behind his ar­ gument than at the argument itself. That ideology includes a domain of properties-the qualia of subjective experience-that are held to be meta­ physically distinct from the objective physical properties addressed by orthodox science. It is not a surprise, then, on this view, that one might know all physical facts, and yet be ignor­ ant of some domain of these nonphysical qua­ lia. The contrast between what is known and what is not known simply reflects an antece­ dent metaphysical division in the furniture of the world.

Paul M. Churchland

576

But there is another way to look at the sit­

picture of human consciousness, one blind to the

uation, one that finds no such division. Our

subjective dimension of self, as Jackson's argu­

capacity for recognizing a range of (currently)

ment suggests. Rather, it would be the vehicle of a

inarticulable features in our subjective experience

grand reconstruction and expansion of our sub­

is easily explained on materialist principles; the

jective consciousness, since it would provide us

relevant sketch appears earlier in this chapter . . . .

with a conceptual framework that, unlike folk

Our discursive inarticulation of those features is

psychology, is at last equal to the kinematical and

no surprise either, and signifies nothing about

dynamical intricacies of the world within . . . .

their metaphysical status. . . . Indeed, that veil of

Real precedents for such a reformation can be

inarticulation may itself be swept aside by suit­

drawn from our own history. We did not lose

able learning. What we are now able sponta­

contact with a metaphysically distinct dimension

neously to report about our internal states and

of reality when we stopped seeing an immutable,

cognitive activities need not define the limit on

sparkle-strewn quintessential crystal sphere each

what we might be able to report, spontaneously

time we looked to the heavens, and began to see

and accurately, if we were taught a more appro­

instead an infinite space of gas and dust and giant

priate conceptual scheme in which to express our

stars structured by gravitational attractions and

discriminations. In closing, let me again urge on

violent nuclear processes. On the contrary, we

Jackson this exciting possibility. The intricacies of brain function may be sub­

now see far more than we used to, even with the unaided eye. The diverse "colors" of the stars

jectively opaque to us now, but they need not re­

allow us to see directly their absolute temper­

main that way forever. Neuroscience may appear

atures. Stellar temperature is a function of stellar

to be defective in providing a purely "third­

mass, so we are just as reliably seeing stellar

person account" of mind, but only familiarity of

masses. The intrinsic luminosity or brightness of a

idiom and spontaneity of conceptual response are

star is tightly tied to these same features, and thus

required to make it a "first-person account" as

is also visually available, no matter how bright or

What makes an account a ''first-person ac­ count" is not the content of that account, but the fact that one has learned to use it as the vehicle of spontaneous conceptualization in introspection and self-description. We all of us, as children, learned

faint the star may appear from Earth. Apparent the contrast between the apparent and the in­

to use the framework of current folk psychology

three-dimensional distribution of complex stellar

well.

brightness is visually obvious also, of course, and trinsic brightnesses gives you the star's rough dis­ tance from Earth. In this way is the character and

in this role. But it is entirely possible for a person

objects in a volume of interstellar space hundreds

or culture to learn and use some other framework

of light years on a side made visually available to

in that role, the framework of cognitive neuro­

your unaided eyes from your own back yard,

science, perhaps. Given a deep and practiced

given only the right conceptual framework for

familiarity with the developing idioms of cogni­

grasping it, and observational practice in using

tive neurobiology, we might learn to discriminate

that framework. From within the new frame­

by introspection the coding vectors in our internal

work, one finds a systematic significance in ex­

axonal pathways, the activation patterns across

periential details that hitherto went largely or

salient neural populations, and myriad other

entirely unnoticed (compare Feyerabend 1 963).

things besides. Should that ever happen, it would then be ob­

The case of inner space is potentially the same. We will not lose contact with a metaphysically

vious to everyone who had made the conceptual

distinct dimension of self when we stop intro­

shift that a completed cognitive neuroscience

specting inarticulable qualia, and start intro­

would constitute not a pinched and exclusionary

specting "instead" sensory coding vectors and

Knowing Qualia: A Reply to Jackson

sundry activation patterns within the vector spaces of our accessible cortical areas. As with the revolution in astronomy, the prospect is one we should welcome as metaphysically liberating, rather than deride as metaphysically irrelevant or metaphysically impossible. References Churchland, P. M . (1 979). Scientific Realism and the Plasticity of Mind. Cambridge: Cambridge University Press. . (198 1 ). "Eliminative Materialism and the Propositional Attitudes." Journal of Philosophy 78, no. 2:67-90. --

--. (1985). "Reduction, Qualia, and the Direct In­ trospection of Brain States." Journal of Philosophy 82, no. 1 :8-28. Feyerabend, P. K. (1 963). "How to Be a Good Empiri­ cist." In B. Baumrin, ed., Philosophy of Science, vol. 2, pp. 3-19. New York: Interscience Publications. Hubel, D. H., and Livingstone, M. S. ( 1 987). "Segre­ gation of Form, Color, and Stereopsis in Primate Area 1 8." Journal of Neuroscience 7:3378-3415. Jackson, F. ( 1 986). "What Mary Didn't Know." Jour­ nal of Philosophy 83, no. 5:291 -295. Lewis, D. (1983). "Postscript to 'Mad Pain and Martian Pain.' " Philosophical Papers, vol. I . New York: Oxford. Nemirow, L. (1 980). Review of Thomas Nagel, Mortal Questions. Philosophical Review 89, no. 3:473-477. Van Essen, D. c., and Maunsell, J. (1 983). "Hier­ archical Organization and Functional Streams in the Visual Cortex." Trends in Neuroscience 6:370-375. Zeki, S. (1 980). "The Representation of Colours in the

Cerebral Cortex." Nature 284:4 1 2-4 1 8 .

577

38

What Experience Teaches

David Lewis

Example: The Bat2

Experience is the Best Teacber They say that experience is the best teacher, and the classroom is no substitute for Real Life. There's truth to this. If you want to know what some new and different experience is like, you can learn it by going out and really

The bat is an alien crea­

ture, with a sonar sense quite unlike any sense of

having that expe­

rience. You can't learn it by being told about the experience, however thorough your lessons may be. Does this prove much of anything about the metaphysics of mind and the limits of science? I think not.

Example: Skunks and Vegemite

ours. We can never have the experiences of a bat; because we could not become bat-like enough to have those experiences and still be ourselves. We will never know what it's like to be a bat. Not even if we come to know all the facts there are about the bat's behavior and behavioral dis­ positions, about the bat's physical structure and processes, about the bat's functional organiza­ tion. Not even if we come to know all the same sort of physical facts about all the other bats, or about other creatures, or about ourselves. Not even if we come to possess all physical facts

I have smelled

skunks, so I know what it's like to smell skunks. But skunks live only in some parts of the world, so you may never have smelled a skunk. If you haven't smelled a skunk, then you don't know what it's like. You never will, unless someday you smell a skunk for yourself. On the other hand, you may have tasted Vegemite, that famous Aus­ tralian substance; and I never have. So you may know what it's like to taste Vegemite. I don't, and unless I taste Vegemite (what, and spoil a good example!), I never will. It won't help at all to take lessons on the chemical composition of skunk

whatever. Not even if we become able to recog­ nize all the mathematical and logical implications of all these facts, no matter how complicated and how far beyond the reach of finite deduction. Experience is the best teacher, in this sense: having an experience is the best way or perhaps the only way, of coming to know what that expe­ rience is like. No amount of scientific information about the stimuli that produce that experience and the process that goes on in you when you have that experience will enable you to know what it's like to have the experience.

scent or Vegemite, the physiology of the nostrils or the taste-buds, and the neurophysiology of the

. . . But Not Necessarily

sensory nerves and the brain.

Example: The Captive Scientist!

Mary, a bril­

liant scientist, has lived from birth in a cell where everything is black or white. (Even she herself is painted all over.) She views the world on black­ and-white television. By television she reads books, she joins in discussion, she watches the re­ sults of experiments done under her direction. In this way she becomes the world's leading expert on color and color vision and the brain states produced by exposure to colors. But she doesn't know what it's like to see color. And she never will, unless she escapes from her cell.

Having an experience is surely one good way, and surely the only practical way, of coming to know what that experience is like. Can we say, flatly, that it is the only

possible

way? Probably not.

There is a change that takes place in you when you have the experience and thereby come to know what it's like. Perhaps the exact same change could in principle be produced in you by precise neurosurgery, very far beyond the limits of present-day technique. Or it could possibly be produced in you by magic. If we ignore the laws of nature, which are after all contingent, then there is no necessary connection between cause

David Lewis

and effect: anything could cause anything. For instance, the casting of a spell could do to you exactly what your first smell of skunk would do. We might quibble about whether a state pro­ duced in this artificial fashion would deserve the name "knowing what it's like to smell a skunk," but we can imagine that so far as what goes on within you is concerned, it would differ not at all. 3 Just as we can imagine that a spell might pro­ duce the same change as a smell, so likewise we can imagine that science lessons might cause that same change. Even that is possible, in the broad­ est sense of the word. If we ignored all we know about how the world really works, we could not say what might happen to someone if he were taught about the chemistry of scent and the physiology of the nose. There might have been a causal mechanism that transforms science lessons into whatever it is that experience gives us. But there isn't. It is not an absolutely necessary truth that experience is the best teacher about what a new experience is like. It's a contingent truth. But we have good reason to think it's true. We have good reason to think that something of this kind is true, anyway, but less reason to be sure exactly what. Maybe some way of giving the lessons that hasn't yet been invented, and some way of taking them in that hasn't yet been prac­ ticed, could give us a big surprise. Consider sight­ reading: a trained musician can read the score and know what it would be like to hear the music. If I'd never heard that some people can sight­ read I would never have thought it humanly poss ble. Of course the moral is that n�w m�ic isn't altogether new--the big new expenence IS a rearrangement of lots of little old experiences. It just might tum out the same for new smells �nd . tastes vis-a-vis old ones; or even for color VISIon vis-a-vis black and white4; or even for sonar sense experience vis-a-vis the sort we enjoy. The thing we can say with some confidence is that we have no faculty for knowing on the basis of mere sci­ ence lessons what some new enough experience would be like. Bilt how new is "new enough?"­ There, we just might be in for surprises.

i

580

Three Ways to Miss the Point

The First Way A literalist might see the phrase "know what it's like" and take that to mean: "know what it resembles." Then he might ask: what's so hard about that? Why can't you just be told which experiences resemble one another? You needn't have had the experiences-all you need, to be taught your lessons, is some way of referring to them. You could be told: the smell of skunk somewhat resembles the smell of burning rubber. I have been told: the taste of Vegemite somewhat resembles that of Marmite. Black-and­ white Mary might know more than most of us about the resemblances among color-experiences. She might know which ones are spontaneously called "similar" by subjects who have them; which gradual changes from one to another tend to escape notice; which ones get conflated with which in memory; which ones involve roughly the same neurons firing in similar rhythms; and so forth. We could even know what the bat's sonar experiences resemble just by knowing that they do not at all resemble any experiences of humans, but do resemble--as it might be-certain experi­ ences that occur in certain fish. This misses the point. Pace the literalist, "know what it's like" does not mean "know what it resembles." The most that's true is that knowing what it resembles may help you to know what it's like. If you are taught that experience A resembles B and C closely, D less, E not at all, that will help you know what A is like-ifyou know already what B and C and D and E are like. Otherwise, it helps you not at all. I don't know any better what it's like to taste Vegemite when I'm told that it tastes like Marmite, because I don't know what Mar­ mite tastes like either. (Nor do I know any better what Marmite tastes like for being told it tastes like Vegemite.) Maybe Mary knows enough to triangulate each color experience exactly in a network of resemblances, or in many networks of resemblance in different respects, while never knowing what any node of any network is like.

581

What Experience Teaches

Maybe we could do the same for bat experiences.

i n all these ways. Therefore you should have no

But no amount of information about resem­

problem in telling her exactly what experiences

blances, just by itself, does anything to help us

one has upon seeing the colors. Or rather, your

know what an experience is like.

only problem is that you'd be telling her what she

The Second Way

what is the X is to know that the X is the Y, where

knows very well already! In general, to know Insofar as I don't know what

it would be like to drive a steam locomotive fast

it's not too obvious that the X is the Y. (Just

on a cold, stormy night, part of my problem is just

knowing that the X is the X won't do, of course,

that I don't know what experiences I would have.

because it is too obvious.) If Mary knows that the

The firebox puts out a lot of heat, especially when

experience of seeing green is the experience asso­

the fireman opens the door to throw on more

ciated with such-and-such pattern of nerve firings,

coal; on the other hand, the cab is drafty and

then she knows the right sort of unobvious iden­

gives poor protection from the weather. Would I

tity. So she knows what experience one has upon

be too hot or too cold? Or both by turns? Or

seeing green.

would it be chilled face and scorched legs? If

(Sometimes it's suggested that you need a

I knew the answers to such questions, I'd know

"rigid designator": you know what is the X by

much better what it would be like to drive the

knowing that the X is the Y only if "the Y" is a

locomotive. So maybe "know what it's like" just

term whose referent does not depend on any con­

means "know what experiences one has." Then

tingent matter of fact. In the first place, this sug­

again: what's the problem? Why can't you just be

gestion is false. You can know who is the man on

told what experiences you would have if, say,

the balcony by knowing that the man on the bal­

you tasted Vegemite? Again, you needn't have

cony is the Prime Minister even if neither "the

had the experiences -all you need, to be taught

Prime Minister" nor any other phrase available

your lessons, is some way of referring to them.

to you rigidly designates the man who is, in fact,

We have ways to refer to experiences we haven't

the Prime Minister. In the second place, accord­

had. We can refer to them in terms of their causes:

ing to one version of Materialism (the one

the experience one has upon tasting Vegemite, the

accept) a description of the form "the state of

experience one has upon tasting a substance of

having nerves firing in such-and-such a pattern"

such-and-such chemical composition. Or we can

is a rigid designator,

refer to them in terms of their effects: the experi­

fact an experience; and according to another

ence that just caused Fred to say "Yeeuch! " Or

version of Materialism, a description of the form

I

and what it designates is in

we can refer to them in terms of the physical states

"having some or other state which occupies so­

of the nervous system that mediate between those

and-so functional role" is a rigid designator of an

causes and effects: the experience one has when

experience. So even if the false suggestion were

one's nerves are firing in such-and-such pattern.

granted, still it hasn't been shown, without beg­

I myse1f for one,

ging the question against Materialism, that Mary

this means the experience which is identical with

could not know what experience one has upon

such-and-such firing pattern. According to other

seeing red.)

(According to some materialists,

materialists it means the experience which is real­

Since Mary

does

know what experiences she

doesn't

ized by such-and-such firing pattern. According

would have if she saw the colors, but she

to many dualists, it means the experience which is

know what it would be like to see the colors, we'd

merely the lawful companion of such-and-such

better conclude that "know what it's like" does

firing pattern. But whichever it is, we get a way

not after all mean "know what experiences one

of referring to the experience.) Black-and-white

has." The locomotive example was misleading.

Mary is in a position to refer to color-experiences

Yes, by learning what experiences the driver

David Lewis

would have, I can know what driving the loco­ motive would be like; but only because I already know what those experiences are like. (It matters that I know what they're like under the appro­ priate descriptions-as it might be, the descrip­ tion "chilled face and scorched legs." This is something we'll return to later.) Mary may know as well as I do that when the driver leans out into the storm to watch the signals, he will have the experience of seeing sometimes green lights and sometimes red. She knows better than I what ex­ periences he has when signals come into view. She can give many more unobviously equivalent de­ scriptions of those experiences than I can. But knowing what color-experiences the driver has won't help Mary to know what his job is like. It will help me.

The Third Way Until Mary sees green, here is one thing she will never know: she will never know that she is seeing green. The reason why is just that until she sees green, it will never be true that she is seeing green. Some knowledge is irre­ ducibly egocentric, or de se. 5 It is not just knowl­ edge about what goes on in the world; it is knowledge of who and when in the world one is. Knowledge of what goes on in the world will be true alike for all who live in that world; whereas egocentric knowledge may be true for one and false for another, or true for one at one time and false for the same one at another time. Maybe Mary knows in advance, as she plots her escape, that 9 a.m. on the 1 3th of May, 1 997, is the mo­ ment when someone previously confined in a black-and-white cell sees color for the first time. But until that moment comes, she will never know that she herself is then seeing color-because she isn't. What isn't true isn't knowledge. This goes as much for egocentric knowledge as for the rest. So only those of whom an egocentric proposition is true can know it, and only at times when it is true of them can they know it. That one is then seeing color is an egocentric proposition. So we've found a proposition which Mary can never know until she sees color-which, as it happens, is the very

582

moment when she will first know what it's like to see color! Have we discovered the reason why ex­ perience is the best teacher? And not contingently after all, but as a necessary consequence of the logic of egocentric knowledge? No; we have two separate phenomena here, and only some bewitchment about the "first-per­ son perspective" could make us miss the differ­ ence. In the first place, Mary will probably go on knowing what it's like to see green after she stops knowing the egocentric proposition that she's then seeing green. Since what isn't true isn't known she must stop knowing that proposition the moment she stops seeing green. (Does that only mean that we should have taken a different egocentric proposition: that one has seen green? No; for in that case Mary could go on knowing the proposition even after she forgets what it's like to see green, as might happen if she were soon recaptured.) In the second place, Mary might come to know what it's like to see green even if she didn't know the egocentric proposition. She might not have known in advance that her escape route would take her across a green meadow, and it might take her a little while to recognize grass by its shape. So at first she might know only that she was seeing some colors or other, and thereby finding out what some color-experiences or other were like, without being able to put a name either to the colors or to the experiences. She would then know what it was like to see green, though not under that description, indeed not under any de­ scription more useful than "the color-experience I'm having now"; but she would not know the egocentric proposition that she is then seeing green, since she wouldn't know which color she was seeing. In the third place, the gaining of ego­ centric knowledge may have prerequisites that have nothing to do with experience. Just as Mary can't know she's seeing green until she does see green, she can't know she's turning 50 until she does turn 50. But-I hopeI-turning 50 does not involve some special experience. In short, though indeed one can gain egocentric knowledge that one is in some situation only when one is in it, that

583

What Experience Teaches

is not the same as finding out what an experience is like only when one has that experience.

the physical information that we might gather about bats could help us know what it's like

We've just rejected two suggestions that don't

to have their experiences; and likewise in other

work separately, and we may note that they don't

cases. There is a natural and tempting explana­

work any better when put together. One knows

tion of why physical information does not help.

X by knowing that the X is the Y,

That is the hypothesis that besides physical in­

where the identity is not too obvious; and "the Y"

formation there is an irreducibly different kind of

what is the

phenomenal information.

might be an egocentric description. So knowledge

information to be had:

that the X is the Y might be irreducibly egocentric

The two are independent. Two possible cases

knowledge, therefore knowledge that cannot be

might be exactly alike physically, yet differ phe­

X is the Y. So one way of knowing what is the X will remain unavailable until it comes true of one that the X is

narrow down the physical possibilities, and per­

had until it is true of one that the

nomenally. When we get physical information we haps we narrow them down all the way to one,

the Y. One way that I could gain an unobvious

but we leave open a range of phenomenal possi­

identity concerning the taste of Vegemite would

bilities. When we have an experience, on the other

be for it to come true that the taste of Vegemite

hand, we acquire phenomenal information; pos­

was the taste I was having at that very moment­

sibilities previously open are eliminated; and that

and that would come true at the very moment I

is what it is to learn what the experience is like.

tasted Vegemite and found out what it was like!

(Analogy. Suppose the question concerned the

Is this why experience is the best teacher?-No;

location of a point within a certain region of the

cases of gaining an unobvious egocentric identity

x-y plane. We might be told that its x-coordinate

are a dime a dozen, and most of them do not re­

lies in certain intervals,

sult in finding out what an experience is like.

others. We might even get enough of this in­

fix

and outside certain

Suppose I plan ahead that I will finally break

formation to

down and taste Vegemite next Thursday noon.

no amount of x-information would tell us any­

the x-coordinate exactly. But

Then on Wednesday noon, if I watch the clock, I

thing about the y-coordinate; any amount of x­

first gain the unobvious egocentric knowledge

information leaves open all the y-possibilities.

that the taste of Vegemite is the taste I shall be having in exactly

24

hours, and thereby I have a

But when at last we make a y-measurement, we acquire a new kind of information; possibilities

new way of knowing what is the taste of Vege­

previously open are eliminated; and that is how

mite. But on Wednesday noon I don't yet know

we learn where the point is in the y-direction.)

what it's like. Another example: from time to time

1

What might the subject matter of phenomenal

find myself next to a Vegemite-taster. On those

information be? lfthe Hypothesis of Phenomenal

occasions, and only those, I know what is the

Information is true, then you have an easy an­

taste of Vegemite by knowing that it is the taste

swer: it is information about experience. More

being had by the person next to me. But on no

specifically, it is information about a certain part

such occasion has it ever yet happened that I

or aspect or feature of experience. But if the

knew what it was like to taste Vegemite.

Hypothesis is false, then there is still experi­ ence (complete with all its parts and aspects and

The Hypothesis of Phenomenal Information No amount of the physical information that black-and-white Mary gathers could help her know what it was like to see colors; no amount of

features) and yet no information about experi­ ence is phenomenal information. So it cannot be said in a neutral way, without presupposing the Hypothesis, that information about experience is phenomenal information. For if the Hypothesis is false and Materialism is true, it may be that all

David Lewis

the information there is about experience is phys­ ical information, and can very well be presented in lessons for the inexperienced. It makes no difference to put some fashionable new phrase in place of "experience. " If instead of "experience" you say "raw feel" (or just "feel­ ing"), or "way it feels," or "what it's like," then I submit that you mean nothing different. Is there anything it's like to be this robot? Does this robot have experiences?-I can tell no difference be­ tween the new question and the old. Does sun­ burn feel the same way to you that it does to me? Do we have the same raw feel? Do we have the same experience when sunburned?-Again, same question. "Know the feeling," "know what it's like" -interchangeable. (Except that the former may hint at an alternative to the Hypothesis of Phenomenal Information.) So if the friend of phenomenal information says that its subject matter is raw feels, or ways to feel, or what it's like, then I respond just as I do if he says that the subject matter is experience. Maybe so, if the Hypothesis of Phenomenal Information is true; but if the Hypothesis is false and Materialism is true, nevertheless there is still information about raw feels, ways to feel or what it's like; but in that case it is physical information and can be con­ veyed in lessons. We might get a candidate for the subject matter of phenomenal information that is not just ex­ perience renamed, but is still tendentious. For instance, we might be told that phenomenal information concerns the intrinsic character of experience. A friend of phenomenal information might indeed believe that it reveals certain spe­ cial, non-physical intrinsic properties of experi­ ence. He might even believe that it reveals the existence of some special non-physical thing or process, all of whose intrinsic properties are non­ physical. But he is by no means alone in saying that experience has an intrinsic character. Plenty of us materialists say so too. We say that a certain color-experience is whatever state occupies a cer­ tain functional role. So if the occupant of that role (universally, or in the case of humans, or in the

584

case of certain humans) is a certain pattern of neural firing, then that pattern of firing is the ex­ perience) (in the case in question). Therefore the intrinsic character of the experience is the in­ trinsic character of the firing pattern. For in­ stance, a frequency of firing is part of the intrinsic character of the experience. If we materialists are right about what experience is, then black-and­ white Mary knows all about the intrinsic charac­ ter of color-experience; whereas most people who know what color-experience is like remain totally ignorant about its intrinsic character. 6 To say that phenomenal information concerns "qualia" would be tendentious in much the same way. For how was this notion introduced? Often thus. We are told to imagine someone who, when he sees red things, has just the sort of experiences that we have when we see green things, and vice versa; and we are told to call this a case of "in­ verted qualia." And then we are told to imagine someone queerer still, who sees red and responds to it appropriately, and indeed has entirely the same functional organization of inner states as we do and yet has no experiences at all; and we are told to call this a case of "absent qualia." Now a friend of phenomenal information might well think that these deficiencies have something to do with the non-physical subject matter of phenom­ enal information. But others can understand them otherwise. Some materialists will reject the cases outright, but others, and I for one, will make sense of them as best we can. Maybe the point is that the states that occupy the roles of experiences, and therefore are the experiences, in normal people are inverted or absent in victims of inverted or absent qualia. (This presupposes, what might be false, that most people are enough alike.) Experience of red-the state that occupies that role in normal people---occurs also in the victim of "inverted qualia," but in him it occupies the role of experience of green; whereas the state that occupies in him the role of experience of red is the state that occupies in normal people the role of experience of green. Experience of red and of green-that is, the occupants of those roles for

What Experience Teaches

normal people-do not occur at all in the victim of "absent qualia"; the occupants of those roles for him are states that don't occur at all in the normal. Thus we make good sense of inverted and absent qualia; but in such a way that "qualia" is just the word for role-occupying states taken per se rather than qua occupants of roles. Qualia, so understood, could not be the subject matter of phenomenal information. Mary knows all about them. We who have them mostly don't. 7 It is best to rest content with an unhelpful name and a via negativa. StipUlate that "the phenom­ enal aspect of the world" is to name whatever is the subject matter of phenomenal information, if there is any such thing; the phenomenal aspect, if such there be, is that which we can become informed about by having new experiences but never by taking lessons. Having said this, it will be safe to say that information about the phenomenal aspect of the world can only be phenomenal information. But all we really know, after thus closing the circle, is that phenomenal information is supposed to reveal the presence of some sort of non-physical things or processes within experience, or else it is supposed to reveal that certain physical things or processes within experience have some sort of non-physical properties.

The Knowledge Argument If we invoke the Hypothesis of Phenomenal In­ formation to explain why no amount of physical information suffices to teach us what a new expe­ rience is like, then we have a powerful argument to refute any materialist theory of the mind. Frank Jackson (see note I) calls it the "Knowl­ edge Argument." Arguments against one materi­ alist theory or another are never very conclusive. It is always possible to adjust the details. But the Knowledge Argument, if it worked, would directly refute the bare minimum that is common to all materialist theories.

585

It goes as follows. First in a simplified form; afterward we'll do it properly. Minimal Materi­ alism is a supervenience thesis: no difference without physical difference. That is: any two pos­ sibilities that are just alike physically are just alike simpliciter. If two possibilities are just alike phys­ ically, then no physical information can eliminate one but not both of them. If two possibilities are just alike simplicite� (if that is possible) then no information whatsoever can eliminate one but not both of them. So if there is a kind of in­ formation-namely, phenomenal information­ that can eliminate possibilities that any amount of physical information leaves open, then there must be possibilities that are just alike physically, but not just alike simpliciter. That is just what minimal Materialism denies. (Analogy. If two possible locations in our re­ gion agree in their x-coordinate, then no amount of x-information can eliminate one but not both. If, per impossibile, two possible locations agreed in all their coordinates, then no information whatsoever could eliminate one but not both. So if there is a kind of information-namely, y-in­ formation-that can eliminate locations that any amount of x-information leaves open, then there must be locations in the region that agree in their x-coordinate but not in all their coordinates.) Now to remove the simplification. What we saw so far was the Knowledge Argument against Materialism taken as a necessary truth, applying unrestrictedly to all possible worlds. But we ma­ terialists usually think that Materialism is'a con­ tingent truth. We grant that there are spooky possible worlds where Materialism is false, but we insist that our actual world isn't one of them. If so, then there might after all be two possibilities that are alike physically but not alike simpliciter, but one or both of the two would have to be pos­ sibilities where Materialism was false. Spooky worlds could differ with respect to their spooks without differing physically. Our minimal Mate­ rialism must be a restricted supervenience thesis: within a certain class of worlds, which includes

David Lewis

our actual world, there is no difference without physical difference. Within that class, any two possibilities just alike physically are just alike simpliciter. But what delineates the relevant class? (It is trivial that our world belongs to some class wherein there is no difference without physical difference. That will be so however spooky our world may be. The unit class of our world is one such class, for instance. And so is any class that contains our world, and contains no two physical duplicates.) I think the relevant class should con­ sist of the worlds that have nothing wholly alien to this world. The inhabitants of such a non-alien world could be made from the inhabitants of ours, so to speak, by a process of division and re­ combination. That will make no wholly different kinds of things, and no wholly different funda­ mental properties of things.8 Our restricted ma­ terialist supervenience thesis should go as follows: throughout the non-alien words, there is no dif­ ference without physical difference. If the Hypothesis of Phenomenal Information be granted, then the Knowledge Argument re­ futes this restricted supervenience nearly as decisively as it refutes the unrestricted version. Consider a possibility that is eliminated by phe­ nomenal information, but not by any amount of physical information. There are two cases. Maybe this possibility has nothing that is alien to our world. In that case the argument goes as before: actuality and the eliminated possibility are just alike physically, they are not just alike simpliciter; furthermore, both of them fall within the restriction to non-alien worlds, so we have a counterexample even to restricted supervenience. Or maybe instead the eliminated possibility does have something X which is alien to this world­ an alien kind of thing, or maybe an alien funda­ mental property of non-alien things. Then the phenomenal information gained by having a new experience has revealed something negative: at least in part, it is the information that X is not present. How can that be? If there is such a thing as phenomenal information, presumably what it reveals is positive: the presence of something

586

hitherto unknown. Not, of course, something alien from actuality itself; but something alien from actuality as it is inadequately represented by the inexperienced and by the materialists. If Mary learns something when she finds out what it's like to see the colors, presumably she learns that there's more to the world than she knew before­ not less. It's easy to think that phenomenal information might eliminate possibilities that are impoverished by comparison with actuality, but that would make a counterexample to the restricted supervenience thesis. To eliminate possibilities without making a counterexample, phenomenal information would have to eliminate possibilities less impoverished than actuality. And how can phenomenal information do that? Com­ pare ordinary perceptual information. Maybe Jean-Paul can just see that Pierre is absent from the cafe, at least if it's a small cafe. But how can he just see that Pierre is absent from Paris, let alone from the whole of actuality? ( Is there a third case? What if the eliminated possibility is in one respect richer than actuality, in another respect poorer? Suppose the eliminated possibility has X, which is alien from actuality, but also it lacks Y. Then phenomenal informa­ tion might eliminate it by revealing the actual presence of Y, without having to reveal the actual absence of X-but then I say there ought to be a third possibility, one with neither X nor Y, poorer and in no respect richer than actuality, and again without any physical difference from actuality. For why should taking away X automatically re­ store Y? Why can't they vary independentIy?9 But this third possibility differs simpliciter from ac­ tuality without differing physically. Further, it has nothing alien from actuality. So we regain a counterexample to the restricted supervenience thesis.) The Knowledge Argument works. There is no way to grant the Hypothesis of Phenomenal In­ formation and still uphold Materialism. There­ fore I deny the Hypothesis. I cannot refute it outright. But later I shall argue, first, that it is more peculiar, and therefore less tempting, that it

587

What Experience Teaches

may at first seem; and, second, that we are not

with the special role of Russian! I will have to say

forced to accept it, since an alternative hypothesis

before I'm done that phenomenal information is

does justice to the way experience best teaches us

an illusion, but I think I must look elsewhere for a credible hypothesis about what sort of illusion it

what it's like.

might be.

Three More Ways to Miss the Point The

Hypothesis of Phenomenal

The Fourth Way Information

characterizes information in terms of eliminated possibilities. But there are other conceptions of "information." Therefore the Hypothesis has look-alikes: hypotheses which say that experience produces "information" which could not be gained otherwise, but do not characterize this "information" in terms of eliminated possibil­ ities. These look-alikes do not work as premises for the Knowledge Argument. They do not say that phenomenal information eliminates possi­ bilities that differ, but do not differ physically, from uneliminated possibilities. The look-alike hypotheses of phenomenal "information" are consistent with Materialism, and may very well be true. But they don't make the Knowledge Argument go away. Whatever harmless look­ alikes may or may not be true, and whatever conception may or may not deserve the name "information," the only way to save Materialism is

fix our attention squarely on the genuine Hy­

pothesis of Phenomenal Information, and deny it. To avert our eyes, and attend to something else, is no substitute for that denial. Might a look-alike help at least to this extent: by giving us something true that well might have been confused with the genuine Hypothesis, thereby explaining how we might have believed the Hypothesis although it was false? I think not. Each of the look-alikes turns out to imply not only that experience can give us "information" that no amount of lessons can give, but also that lessons in Russian can give us "information" that no amount of lessons in English can give (and vice versa). I doubt that any friend of phenomenal information ever thought that the special role of experience in teaching what it's like was on a par

If a hidden camera takes photo­

graphs of a room, the film ends up bearing traces of what went on in the room. The traces are dis­ tinctive: that is, the details of the traces depend on the details of what went on, and if what went on had been different in any of many ways, the traces would have been correspondingly different. So we can say that the traces bear information, and that he who has the film has the information. That might be said because the traces, plus the way they depend on what went on, suffice to eliminate possibilities; but instead we might say "infor­ mation" and just mean "distinctive traces." If so, it's certainly true that new experience imparts "information" unlike any that can be gained from lessons. Experience and lessons leave differ­ ent kinds of traces. That is so whether or not the experience eliminates possibilities that the lessons leave open. It is equally true, of course, that les­ sons in Russian leave traces unlike any that are left by lessons in English, regardless of whether the lessons cover the same ground and eliminate the same possibilities.

The Fifth Way

When we speak of transmission

of "information," we often mean transmission of text. Repositories of "information," such as libraries, are storehouses of text. Whether the text is empty verbiage or highly informative is beside the point. Maybe we too contain information by being storehouses of text. Maybe there is a lan­ guage of thought, and maybe the way we believe things is to store sentences of this language in some special way, or in some special part of our brains. In that case, we could say that storing away a new sentence was storing away a new piece of "information," whether or not that new piece eliminated any possibilities not already eliminated by the sentences stored previously.

David Lewis

Maybe, also, the language of thought is not fixed once and for all, but can gain new words. Maybe, for instance, it borrows words from public lan­ guage. And maybe, when one has a new experi­ ence, that causes one's language of thought to gain a new word which denotes that experience­ a word which could not have been added to the language by any other means. If all this is so, then when Mary sees colors, her language of thought gains new words, allowing her to store away new sentences and thereby gain "infonnation." All this about the language of thought, the storing of sentences, and the gaining of words is spec­ ulation. But it is plausible speculation, even if no longer the only game in town. If it is all true, then we have another look-alike hypothesis of phe­ nomenal "infonnation." When Mary gains new words and stores new sentences, that is "infor­ mation" that she never had before, regardless of whether it eliminates any possibilities that she had not eliminated already. But again, the special role of experience turns out to be on a par with the special role of Russian. If the language of thought picks up new words by borrowing from public language, then lessons in Russian add new words, and result in the storing of new sentences, and thereby impart "infor­ mation" that never could have been had from lessons in English. (You might say that the new Russian words are mere synonyms of old words, or at least old phrases, that were there already; and synonyms don't count. But no reason has been given why the new inner words created by experience may not also be synonyms of old phrases, perhaps of long descriptions in the lan­ guage of neurophysiology.)

The Sixth Way A philosopher who is skeptical about possibility, as so many are, may wish to replace possibilities themselves with linguistic ersatz possibilities: maximal consistent sets of sentences. And he may be content to take "con­ sistent" in a narrowly logical sense, so that a set with "Fred is married" and "Fred is a bachelor" may count as consistent, and only an overt con-

588

tradiction like "Fred is married" and "Fred is not married" will be ruled out. I O The ersatz possibil­ ities might also be taken as sets of sentences of the language of thought, if the philosopher believes in it. Then if someone's language of thought gains new words, whether as a result of new experience or as a result of being taught in Russian, the ersatz possibilities become richer and more nu­ merous. The sets of sentences that were maximal before are no longer maximal after new words are added. So when Mary sees colors and her lan­ guage of thought gains new words, there are new ersatz possibilities; and she can straightway elim­ inate some of them. Suppose she knows before­ hand that she is about to see green, and that the experience of seeing green is associated with neu­ ral firing pattern F. So when she sees green and gains the new word G for her experience, then straightway there are new, enriched ersatz pos­ sibilities with sentences saying that she has G without F, and straightway she knows enough to eliminate these ersatz possibilities. (Even if she does not know beforehand what she is about to see, straightway she can eliminate at least those of her new-found ersatz possibilities with sentences denying that she then has G.) Just as we can characterize infonnation in tenns of elimination of possibilities, so we can characterize ersatz "infonnation" in tenns of elimination of ersatz "possibilities." So here we have the closest look­ alike hypothesis of all, provided that language-of­ thoughtism is true. But we still do not have the genuine Hypothesis of Phenomenal Infonnation, since the eliminated ersatz possibility of G with­ out F may not have been a genuine possibility at all. It may have been like the ersatz possibility of married bachelors.

Curiouser and Curiouser The Hypothesis of Phenomenal Infonnation is more peculiar than it may at first seem. For one thing, because it is opposed to more than just Materialism. Some of you may have welcomed

What Experience Teaches

the Knowledge Argument because you thought all along that physical information was inade­ quate to explain the phenomena of mind. You may have been convinced all along that the mind could do things that no physical system could do: bend spoons, invent new jokes, demonstrate the consistency of arithmetic, reduce the wave packet, or what have you. You may have been convinced that the full causal story of how the deeds of mind are accomplished involves the causal interactions not only of material bodies but also of astral bodies; not only the vibrations of the electromagnetic field but also the good or bad vibes of the psionic field; not only protoplasm but ectoplasm. I doubt it, but never mind. It's ir­ relevant to our topic. The Knowledge Argument is targeted against you .no less than it is against Materialism itself. Let parapsychology be the science of all the non-physical things, properties, causal processes, laws of nature, and so forth that may be required to explain the things we do. Let us suppose that we learn ever so much parapsychology. It will make no difference. Black-and-white Mary may study all the parapsychology as well as all the psychophysics of color vision, but she still won't know what it's like. Lessons on the aura of Vege­ mite will do no more for us than lessons on its chemical composition. And so it goes. Our in­ tuitive starting point wasn't just that physics les­ sons couldn't help the inexperienced to know what it's like. It was that lessons couldn't help. If there is such a thing as phenomenal information, it isn't just independent of physical informa­ tion. It's independent of every sort of information that could be served up in lessons for the inex­ perienced. For it is supposed to eliminate possi­ bilities that any amount of lessons leave open. Therefore phenomenal information is not just parapsychological information, if such there be. It's something very much stranger. The genuine Hypothesis of Phenomenal In­ formation, as distinguished from its look-alikes, treats information in terms of the elimination of

589

possibilities. When we lack information, several alternative possibilities are open, when we get the information some ofthe alternatives are excluded. But a second peculiar thing about phenomenal information is that it resists this treatment. (So does logical or mathematical "information." However, phenomenal information cannot be logical or mathematical, because lessons in logic and mathematics no more teach us what a new experience is like than lessons in physics or para­ psychology do.) When someone doesn't know what it's like to have an experience, where are the alternative open possibilities? I cannot present to myself in thought a range of alternative possibil­ ities about what it might be like to taste Vegemite. That is because I cannot imagine either what it is like to taste Vegemite, or any alternative way that it might be like but in fact isn't. ( I could perfectly well imagine that Vegemite tastes just like peanut butter, or something else familiar to me, but let's suppose I've been told authoritatively that this isn't so.) I can't even pose the question that phe­ nomenal information is supposed to answer: is it this way or that? It seems that the alternative possibilities must be unthinkable beforehand; and afterward too, except for the one that turns out to be actualized. I don't say there's anything alto­ gether impossible about a range of unthinkable alternatives; only something peculiar. But it's peculiar enough to suggest that we may somehow have gone astray. From Pbenomenal to Epiphenomenal A third peculiar thing about phenomenal in­ formation is that it is strangely isolated from all other sorts of information; and this is so regardless of whether the mind works on physical or parapsychological principles. The phenomenal aspect of the world has nothing to do with ex­ plaining why people seemingly talk about the phenomenal aspect of the world. For instance, it plays no part in explaining the movements of

David Lewis

the pens of philosophers writing treatises about phenomenal information and the way experience has provided them with it. When Mary gets out of her black-and-white cell, her jaw drops. She says "At last! So this is what it's like to see colors!" Afterward she does things she couldn't do before, such as recognizing a new sample of the first color she ever saw. She may also do other things she didn't do before: unfortunate things, like writing about phenom­ enal information and the poverty of Materialism. One might think she said what she said and did what she did because she came to know what it's like to see colors. Not so, if the Hypothesis of Phenomenal Information is right. For suppose the phenomenal aspect of the world had been otherwise, so that she gained different phenom­ enal information. Or suppose the phenomenal aspect of the world had been absent altogether, as we materialists think it is. Would that have made the slightest difference to what she did or said then or later? I think not. Making a difference to what she does or says means, at least in part, making a difference to the motions of the particles of which she is composed. (Or better: making a difference to the spatiotemporal shape of the wave-function of those particles. But let that pass.) For how could she do or say anything dif­ ferent, if none of her particles moved any differ­ ently? But if something non-physical sometimes makes a difference to the motions of physical particles, then physics as we know it is wrong. Not just silent, not just incomplete-wrong. Either the particles are caused to change their motion without benefit of any force, or else there is some extra force that works very differently from the usual four. To believe in the phenomenal aspect of the world, but deny that it is epiphe­ nomenal, is to bet against the truth of physics. Given the success of physics hitherto, and even with due allowance for the foundational ailments of quantum mechanics, such betting is rash! A friend of the phenomenal aspect would be safer to join Jackson in defense of epiphenomenal qualia. But there is more to the case than just an em-

590

pirical bet in favor of physics. Suppose there is a phenomenal aspect of the world, and suppose it does make some difference to the motions of Mary's jaw or the noises out of her mouth. Then we can describe the phenomenal aspect, if we know enough, in terms of its physical effects. It is that on which physical phenomena depend in such-and-such way. This descriptive handle will enable us to give lessons on it to the inex­ perienced. But insofar as we can give lessons on it, what we have is just parapsychology. That whereof we cannot learn except by having the experience still eludes us. I do not argue that everything about the alleged distinctive subject matter of phenomenal information must be epi­ phenomenal. Part of it may be parapsychological instead. But I insist that some aspect of it must be epiphenomenal. Suppose that the Hypothesis of Phenomenal Information is true and suppose that V. and V2 are all of the maximally specific phenomenal possibilities concerning what it's like to taste Vegernite; anyone who tastes Vegemite will find out which one obtains, and no one else can. And suppose that PI and P2 are all the maximally spe­ cific physical possibilities. (Of course we really need far more than two Ps, and maybe a friend of phenomenal information would want more than two Vs, but absurdly small numbers will do for an example.) Then we have four alternative hypotheses about the causal independence or dependence of the Ps on the Vs. Each one can be expressed as a pair of counterfactual conditionals. Two hypotheses are patterns of dependence. K I : ifV. then p . , ifV2 then P2 K2: ifV. then P2, ifV2 then p. The other two are patterns of independence. K3: ifV. then p . , ifV2 then p. �: if V. then P2, if V2 then P2 These dependency hypotheses are, I take it, con­ tingent propositions. They are made true, if

What Experience Teaches

591

they are, by some contingent feature of the

sibilities have been swapped. Whether it's in­

world, though it's indeed a vexed question what

dependence or whether it's dependence, therefore,

sort of feature it is. 1 1 Now we have eight joint

we have found an epiphenomenal part of the

possibilities.

phenomenal aspect of the world. It is the residue left behind when we remove the parapsycho­ logical part. Suppose that someday I taste Vegemite, and hold forth about how I know at last what it's like.

Between the four on the top row and the four on

The sound of my holding forth is a physical effect,

the bottom row, there is the physical difference

part of the realized physical possibility PI . This

between PI and P2. Between the four on the left

physical effect is exactly the same whether it's

and the four on the right, there is the phenomenal

part of the joint possibility KI V I PI or part of its

difference between VI and V2. And between the

alternative K2 V2PI . It may be caused by V I in

four on the edges and the four in the middle there

accordance with K I , or it may instead be caused

is a parapsychological difference. It is the differ­

by V2 in accordance with K2, but it's the same

ence between dependence and independence of

either way. So it does not occur because we have

the physical on the phenomenal; between efficacy

KI VI rather than K2 V2, or vice versa. The alleged

and epiphenomenalism, so far as this one exam­

difference between these two possibilities does

ple is concerned. There's nothing ineffable about

nothing to explain the alleged physical manifes­

that. Whether or not you've tasted Vegemite, and

tation of my finding out which one of them is

whether or not you can conceive of the alleged

realized. It is in that way that the difference is

difference between VI and V2, you can still be told

epiphenomenal. That makes it very queer, and

whether the physical difference between PI and P2

repugnant to good sense.

does or doesn't depend on some part of the phe­ nomenal aspect of the world. Lessons can teach the inexperienced which parapsychological possibility obtains, dependence or independence. Let it be dependence: we have either KI or K2. For if we had independence, then already we would have found our epiphenomenal difference: namely, the difference between VI and V2. And lessons can teach the inexperienced which of the two physical possibilities obtains. Without loss of generality let it be PI . Now two of our original eight joint possibilities remain open: K I V I PI and K2V2PI . The difference between those is not at all physical, and not at all para­ psychological: it's PI , and it's dependence, in both cases. The difference is entirely phenomenal. And also it is entirely epiphenomenal. Nothing physi­ cal, and nothing parapsychological, depends on the difference between KI V I PI and K2 V2PI . We have the same sort of pattern of dependence either way; it's just that the phenomenal pos-

The Ability Hypotbesis So the Hypothesis of Phenomenal Information turns out to be very peculiar indeed. It would be nice, and not only for materialists, if we could re­ ject it. For materialists, it is essential to reject it. And we can. There is an alternative hypothesis about what it is to learn what an experience is like: the

Ability Hypothesis.

Laurence Nemirow

summ arizes it thus:

some modes of understanding consist, not in the grasping of facts, but in the acquisition of abilities . . . . As for understanding an experience, we may construe that as an ability to place oneself, at will, in a state rep­ resentative of the experience. I understand the experi­ ence of seeing red if I can at will visualize red. Now it is perfectly clear why there must be a special connection between the ability to place oneself in a state represen­ tative of a given experience and the point of view of ex­ periencer: exercising the ability just is what we call

David Lewis

"adopting the point of view of experiencer." . . . We can, then, come to terms with the subjectivity of our under­ standing of experience without positing subjective facts as the objects of our understanding. This account ex­ plains, incidentally, the linguistic incommunicability of our subjective understanding of experience (a phenom­ enon which might seem to support the hypothesis of subjective facts). The latter is explained as a special case of the linguistic incommunicability of abilities to place oneself at will in a given state, such as the state of having lowered blood pressure, and the state of having wiggling ears. 1 2

I f you have a new experience, you gain abilities to remember and to imagine. After you taste Vegemite, and you learn what it's like, you can afterward remember the experience you had. By remembering how it once was, you can afterward imagine such an experience. Indeed, even if you eventually forget the occasion itself, you will very likely retain your ability to imagine such an experience. Further, you gain an ability to recognize the same experience if it comes again. If you taste Vegemite on another day, you will probably know that you have met the taste once before. And if, while tasting Vegemite, you know that it is Vegemite you are tasting, then you will be able to put the name to the experience if you have it again. Or if you are told nothing at the time, but later you somehow know that it is Vegernite that you are then remembering or imagining tasting, again you can put the name to the experience, or to the memory, or to the experience of imagining, if it comes again. Here, the ability you gain is an ability to gain information if given other in­ formation. Nevertheless, the information gained is not phenomenal, and the ability to gain informa­ tion is not the same thing as information itself. Earlier, I mentioned "knowing what an expe­ rience is like under a description." Now I can say that what I meant by this was having the ability to remember or imagine an experience while also knowing the egocentric proposition that what one is then imagining is the experience of such-and-

592

such description. One might well know what an experience is like under one description, but not under another. One might even know what some experience is like, but not under any description whatever-unless it be some rather trivial de­ scription like "that queer taste that I'm imagining right now." That is what would happen if you slipped a dab of Vegemite into my food without telling me what it was: afterward, I would know what it was like to taste Vegemite, but not under that description, and not under any other non­ trivial description. It might be suggested that "knowing what it's like to taste Vegemite" really means what I'd call " knowing what it's like to taste Vegemite under the description 'tasting Vegemite' "; and if so, knowing what it's like would involve both ability and information. I disagree. For surely it would make sense to say: "I know this experience well, I've long known what it's like, but only today have I found out that it's the experience of tasting Vegemite." But this verbal question is unimportant. For the in­ formation involved in knowing what it's like under a description, and allegedly involved in knowing what it's like, is anyhow not the queer phenomenal information that needs rejecting. ( Is there a problem here for the friend of phenomenal information? Suppose he says that knowing what it's like to taste Vegemite means knowing that the taste of Vegemite has a certain "phenomenal character." This requires putting the name to the taste, so clearly it corresponds to our notion of knowing what it's like to taste Vegemite under the description "tasting Vege­ mite." But we also have our notion of knowing what it's like simpliciter, and what can he offer that corresponds to that? Perhaps he should an­ swer by appeal to a trivial description, as follows: knowing what it's like simpliciter means knowing what it's like under the trivial description "taste I'm imagining now," and that means knowing that the taste one is imagining now has a certain phenomenal character.)

593

What Experience Teaches

As well as gaining the ability to remember and

contributes to know-how, but often it doesn't

imagine the experience you had, you also gain the

contribute enough. That's why music students

ability to imagine related experiences that you

have to practice.

never had. After tasting Vegemite, you might for

Know-how is ability.

But of course some

instance become able to imagine tasting Vege­

aspects of ability are in no sense knowledge:

mite ice cream. By performing imaginative ex­

strength, sufficient funds. Other aspects of ability

periments, you can predict with some confidence

are, purely and simply, a matter of information. If

what you would do in circumstances that have

you want to know how to open the combination

never arisen-whether you'd ask for a second

lock on the bank vault, information is all you

helping of Vegemite ice cream, for example.

need. It remains that there are aspects of ability

These abilities to remember and imagine and

that do

not

consist simply of possession of in­

do

recognize are abilities you cannot gain (unless by

formation, and that we

super-neurosurgery, or by magic) except by tast­

Ability Hypothesis holds that knowing what an

call knowledge. The

ing Vegemite and learning what it's like. You

experience is like is that sort of knowledge.

can't get them by taking lessons on the physics

If the Ability Hypothesis is the correct analysis

or the parapsychology of the experience, or even

of knowing what an experience is like, then phe­

by taking comprehensive lessons that cover the

nomenal information is an illusion. We ought to

whole of physics and parapsychology. The Abil­

explain that illusion. It would be feeble, I think,

ity Hypothesis says that knowing what an expe­

just to say that we're fooled by the ambiguity of

rience is like just is the possession of these abilities

the word "know": we confuse ability with in­

to remember, imagine, and recognize. It isn't the

formation because we confuse knowledge in the

possession of any kind of information, ordinary

sense of knowing-how with knowledge in the

or peculiar. It isn't knowing that certain possibil­

sense of knowing-that. There may be two senses

ities aren't actualized. It isn't knowing-that. It's

of the word "know," but they are well and truly

knowing-how. Therefore it should be no surprise

entangled. They mark the two pure endpoints of a

that lessons won't teach you what an experience is

range of mixed cases. The usual thing is that we

like. Lessons impart information; ability is some­

gain information and ability together. If so, it

thing else. Knowledge-that does not automati­

should be no surprise if we apply to pure cases

cally provide know-how. There are parallel cases. Some know how to wiggle their ears; others don't. If you can't do

of gaining ability, or to pure cases of gaining in­ formation, the same word "know" that we apply to all the mixed cases.

it, no amount of information will help. Some

Along with information and ability, acquain­

know how to eat with chopsticks, others don't.

tance is a third element of the mixture. If Lloyd

Information will help up to a point-for instance,

George died too soon, there's a sense in which

if your trouble is that you hold one chopstick in

Father never can know him. Information won't

each hand-but no amount of information, by

do it, even if Father is a most thorough biogra­

itself, will bring you to a very high level of know­

pher and the archives are very complete . (And the

how. Some know how to recognize a C-38 loco­

trouble isn't that there's some very special in­

motive by sight, others don't. If you don't, it

formation about someone that you can only get

won't much help if you memorize a detailed geo­

by being in his presence.) Know-how won't do it

metrical description of its shape, even though that

either, no matter how good Father may be at

does all the eliminating of possibilities that there

imagining Lloyd George, seemingly remember­

is to be done. (Conversely, knowing the shape

ing him, and recognizing him. ( Father may be

by sight doesn't enable you to write down the

able to recognize Lloyd George even if there's no

geometrical description.) Information very often

longer any Lloyd George to recognize-if

per

David Lewis

irnpossibile

594

he did tum up, Father could tell it

for those abilities turns out also to be a special

was him.) Again, what we have is not just a third

kind of representation of some sort of infor­

separate sense of "know." Meeting someone,

mation, so be it. We need only deny that it repre­

gaining a lot of information about him that would

sents a special kind of information about a special

be hard to gain otherwise, and gaining abil­

subject matter. Apart from that it's up for grabs

ities regarding him usually go together. The pure

what, if anything, it may represent. The details of

cases are exceptions.

stimuli: the chemical composition of Vegemite,

A friend of phenomenal information will agree,

reflectances of surfaces, the motions of well­

of course, that when we learn what an experience

handled chopsticks or of ears? The details of

is like, we gain abilities to remember, imagine,

inner states produced by those stimuli: patterns of

and recognize. But he will say that it is because we

firings of nerves? We could agree to either, so long

gain phenomenal information that we gain the

as we did not confuse 'having information' rep­

abilities. He might even say the same about other

resented in this special way with having the same

cases of gaining know-how: you can recognize the

information in the form of knowledge or, belief.

C-38 when you have phenomenal information

Or we could disagree. Treating the ability-con­

about what it's like to see that shape, you can eat

ferring trace as a representation is optional.

with chopsticks or wiggle your ears when you

What's essential is that when we learn what an

gain phenomenal information about the experi­

experience is like by having it, we gain abilities to

ence of doing so, and so on. What should friends

remember, imagine, and recognize.

of the Ability Hypothesis make of this? Is he offering a conjecture, which we must reject, about the causal origin of abilities? I think not. He thinks, as we do, that experiences leave distinctive traces in people, and that these traces enable us to do things. Likewise being taught to recognize a

C-38 or to eat with chopsticks, or whatever

happens on first wiggling the ears, leave traces that enable

us

to do things afterward. That much

Acknowledgment Part of this paper derives from a lecture at La­

198 1 . I thank LaTrobe for 1 98 1 , Harvard University for sup­ port under a Santayana Fellowship in 1988, and Trobe University in

support in

Frank Jackson for very helpful discussion.

is common ground. He also interprets these en­ abling traces as representations that bear in­ formation about their causes. ( If the same traces had been caused in some deviant way they might perhaps have carried misinformation.) We might even be able to accept that too. The time for us to quarrel comes only when he says that these traces

Notes This chapter was collected in Proceedings of the Russel­ !ian Society, University of Sydney, 1988.

I. See Frank Jackson, "Epiphenomenal qualia," Philo­ sophical Quarterly 32 (1 982), pp. 1 27-36; "What Mary

represent special phenomenal facts, facts which

didn't know," chapter 36.

cannot be represented in any other way, and

2. See B. A. Farrell, "Experience," Mind 59 (1 950), pp. 1 70-98; and Thomas Nagel, "What is it like to be a batT' chapter 32.

therefore which cannot be taught in physics les­ sons or even in parapsychology lessons. That is the part, and the

only part, which we must reject.

But that is no part of · his psychological story about how we gain abilities. It is just a gratuitous metaphysical gloss on that story. We say that learning what an experience is like means gaining certain abilities. If the causal basis

3. See Peter Unger, "On experience and the develop­ ment of the understanding," American Philosophical Quarterly 3 (1 966), pp. 1 -9. 4. For such speCUlation, see Paul M. Churchland, "Reduction, qua lia and the direct introspection of brain states," Journal of Philosophy 82 (1 985), pp. 8-28. ,

What Experience Teaches

5. See my "Attitudes de dido and de se," Philosophical Review 88 (1979), pp. 5 1 3-43, also in my Philosophical Papers, vol. I (New York: Oxford University Press, 1983); and Roderick Chisholm, The First Person: An Essay on ReJerence and Intentionality (Minneapolis: University of Minnesota Press, 1 98 1 ). 6. See Gilbert Harman, "The intrinsic quality of expe­ rience," chapter 43. 7. See Ned Block and Jerry A. Fodor, "What psycho­ logical states are not," Philosophical Review 8 1 ( 1 972), pp. 1 59-81 , also in Ned Block (ed.), Readings in Phi­ losophy oj Psychology, vol. I (Cambridge, MA: Harvard University Press, 1980); and my "Mad Pain and Mar­ tian Pain", in Readings in Philosophy oj Psychology, vol. I, and in my Philosophical Papers, vol. I. 8. See my "New work for a theory of universals," Aus­ tralasian Journal oj Philosophy 61 (1 983), pp. 343-77, especially pp. 361-4. For a different view about how to state minimal Materialism, see Terence Horgan, "Supervenience and microphysics," Pacific Philosoph­ ical Quarterly 63 (1 982), pp. 29-43.

9. On recombination of possibilities, see my On the Plurality oj Worlds (Oxford: Blackwell, 1986), pp� 8792. The present argument may call for a principle that also allows recombination of properties; I now think that would not necessarily require treating properties as non-spatiotemporal parts of their instances. On re­ combination of properties, see also D. M. Armstrong, A Combinatorial Theory oj Possibility (Cambridge: Cam­ bridge University Press 1 989).

10.

See

On the Plurality oj Worlds, pp. 142-65, on lin­

guistic ersatz possibilities.

I I . On dependency hypotheses, see my "Causal deci­ sion theory," Australasian Journal oj Philosophy 59 (1981), pp. 5-30, reprinted in my Philosophical Papers, vol. II (New York: Oxford University Press, 1986).

1 2. Laurence Nemirow, review of Nagel's Mortal Questions, Philosophical Review 89 (1 980), pp. 475-6. For a fuller statement, see Nemirow, "Physicalism and the cognitive role of acquaintance," in W. G. Lycan (ed.), Mind and Cognition: A Reader (Oxford: Basil Blackwell, 1989) and Functionalism and the Subjective Quality oj Experience (doctoral dissertation, Stanford, 1979). See also Michael Tye, "The subjective qualities of experience," Mind 95 (1986), pp. 1 - 1 7. I should record a disagreement with Nemirow on one very small point. We agree that the phrase "what expe­ rience E is like" does not denote some "subjective qual-

595

ity" of E, something which supposedly would be part of the subject matter of the phenomenal information gained by having E. But whereas I have taken the phrase to denote E itself, Nemirow takes it to be a syn­ categorematic part of the expression "know what expe­ rience E is like." See "Physicalism and the cognitive role of acquaintance", section III.

39

Phenomenal States

Brian Loar On a natural view of ourselves, we introspectively

to deny because, as regards ordinary visual expe­

discriminate our own experiences and thereby

riences, we cannot apparently conceive them

fonn conceptions of their qualities, both salient

phenomenally in a way that abstracts from their

and subtle. These discriminations are of various

purporting

degrees of generality, from small differences in

The antiphysicalist intuition is compatible with

tactual and color experience to broad differences

visual experiences' having (some sort of) inter­

to represent things in a certain way.

of sensory modality, for example, those among

nally detennined intentional structure, so that it is

smell, hearing, and pain. What we apparently

an introspectable and nonrelational feature of a

discern are ways experiences differ and resemble

visual experience that it represents things visually

what it is like to have

as being thus and so. Antiphysicalists suppose

each other with respect to

them.

Following common usage, I will call these

that we have conceptions of how visual experi­

phenomenal qualities, and the conceptions we have of them, phenomenal concepts. Phenomenal concepts are fonned "from one's own case. " They are type-demonstratives

ences differ and resemble each other with respect

experiential resemblances

to what it is like to have those experiences. These conceptions then are of qualities of experiences, whatever allowances one may also make for the

that derive their reference from a first-person

apparent qualities of the intrinsic objects of those

perspective: 'that type of sensation', 'that feature

experiences. I will assume that the antiphysicalists'

of visual experience'. And so third-person ascrip­

phenomenological and internalist intuitions are

tions of phenomenal qualities are projective as­

correct. The idea is to engage them over the cen­

criptions of what one has grasped in one's own

tral point, that is, whether those aspects of the

case: 'she has an experience of that type.'

mental that we both count as phenomenologically

'Phenomenal quality' can have a different sense, namely, how the

object of a

compelling raise substantive difficulties for the

perceptual ex­

thesis that phenomenal qualities (thus under­

perience appears. In this sense, a phenomenal

stood) are physical properties of the brain that lie

quality is ascribed to an object and not directly

within the scope of current science.

to an experience. Some have argued that all we discern phenomenologically are phenomenal

We have to distinguish between

properties,

concepts

and

and this chapter turns on that distinc­

qualities in this sense; they deny that experiences

tion. Antiphysicalist arguments and intuitions

themselves have introspectible qualities that are

take off from a sound intuition about concepts.

not ascribed primarily to their objects ( Hannan

Phenomenal concepts are conceptually irredu­

1 990; Block 1 990). I will not pursue the issue here',

cible in this sense: they neither a priori imply, nor

but will assume a certain view of it. For the pres­

are implied by, physical-functional concepts. Al­

ent objective is to engage antiphysicalist argu­

though that is denied by analytical functionalists

ments and entrenched intuitions to the effect that

( Levin 1 983,

conscious mental qualities cannot be identical

including me, find it intuitively appealing. The

1 986), many other physicalists,

with ordinary physical properties, or at least that

antiphysicalist takes this conceptual intuition a

it is problematic to suppose that they are so. Anti­

good deal further, to the conclusion that phe­

physicalists typically suppose that such mental

nomenal qualities are themselves irreducible, are

properties are not relational-that is, that they

not physical-functional properties, at least not of

present themselves as not intrinsically involving

the ordinary sort. The upshot is a range of anti­

relations to things outside the mind. They may

reductionist views: that consciousness and phe­

allow that, say, visual experiences are in some

nomenal qualities are unreal because irreducible; 1

sense intrinsically representational. That is hard

that they are irreducibly non-physical-functional

Brian Loar

facts;2 that they are forever mysterious, or pose an intellectual problem different from other em­ pirical problems, or require new conceptions of the physical. 3 It is my view that we can have it both ways. We may take the phenomenological intuition at face value, accepting introspective concepts and their conceptual irreducibility, and at the same time take phenomenal qualities to be identical with physical-functional properties of the sort envis­ aged by contemporary brain science. As I see it, there is no persuasive philosophically articulated argument to the contrary. This is not to deny the power of raw meta­ physical intuition. Thoughtful people compare phenomenal qualities and kinds of physical­ functional property, say the activation of neural assemblies. It appears to them to be an evident and unmediated truth, independent of further premises, that phenomenal qualities cannot be identical with properties of those types or perhaps of any physical-functional type. This intuition is so compelling that it is tempting to regard anti­ physicalist arguments as rationalizations of an intuition whose independent force masks their tendentiousness. It is the point of this chapter to consider the arguments. But I will also present a positive account of the relation between phe­ nomenal concepts and physical properties that may provide some relief, or at least some dis­ tance, from the illusory metaphysical intuition. In recent years the central problem with physi­ calism has been thought by many to be "the ex­ planatory gap." This is the idea that we cannot explain, in terms of physical-functional proper­ ties, what makes a certain experience 'feel like this', in the way we can explain what makes a certain substance a liquid, say. It is concluded that physicalism is defective in some respect, that there cannot be a (proper) reduction of the mental to the physical. Before we consider this explana­ tory gap, we must first examine, in some detail, a more basic antiphysicalist line of reasoning that goes back to Leibniz and beyond, a leading ver­ sion of which is now called the knowledge argu-

598

ment. Answering this argument will generate a framework in which to address antiphysicalist concerns in general.

1 The Knowledge Argument and Its Semantic Premise The knowledge argument is straightforward on the face of it. Consider any phenomenal quality and any physical property however complex. We c�n know that a person has the physical property Without knowing that she experiences the phe­ nomenal quality. And no amount of a priori rea­ soning or construction can bridge this conceptual gap. That is the intuitive premise. The conclusion is drawn that the phenomenal quality cannot be identical with the physical property. The argu­ ment is equivalent to this: since physical and phenomenal conceptions can be connected only a posteriori, physical properties must be distinct from phenomenal properties. The best known and liveliest version of the knowledge argument is Frank Jackson's, which features the physiologically omniscient Mary, who has never seen color and so does not know what it is like for us to see red, despite her know­ ing all the physical-functional facts about us.4 She later sees colors, and thus learns what it has been like all along for us to see red. She learns a new fact about us. Jackson concludes that this fact is not among the physical facts, since Mary already knew them. It is not difficult to see that this argu­ ment depends on a more or less technical premise. In my view, the physicalist should accept Jack­ son's intuitive description of Mary: she fails to know that we have certain color experiences even though she knows all relevant physical facts about us. And when she acquires color experi­ ence, she does learn something new about us-if you like, learns a new fact or truth. But this is to be granted, of course, only on an opaque reading of 'Mary learns that we have such and such color experiences', and on corresponding readings of 'learns a new fact or truth about us'. For as

Phenomenal States

regards the transparent versions of those ascrip­ tions of what Mary did not know and then learned, they would beg the question, amounting to this: 'as for the property of having such and such color experiences, Mary did not know, but then learned, of that property that we have it'. Physicalists reject this, for according to us those experiential properties are physical properties, and Mary already knew of all our physical prop­ erties that we have them-under their physical descriptions. What she lacked and then acquired, rather, was knowledge of certain such properties couched in experiential terms. Drawing metaphysical conclusions from opaque contexts is risky. And in fact inferences of Jackson's form, without additional premises, are open to straightforward counterexamples of a familiar sort. Let me describe two cases. (1) Max learns that the bottle before him con­ tains CH3CH20H. But he does not know that the bottle contains alcohol. This holds on an opaque reading: he would not assert that there's stutT called alcohol in the bottle, or that the bottle contains the intoxicating component of beer and wine. Let sheltered Max even lack the ordinary concept 'alcohol'. After he acquires that ordinary concept, he learns something new-that the bottle contained alcohol. If the knowledge argument has a generally valid form, we could then infer from Max's epistemic situation that alcohol is not identical with CH3CH20H. Evidently this does not follow. (2) Margot learns about the element Au and reads that people decorate themselves with alloys of Au. But she has never seen gold and cannot visually identify it: she lacks an adequate vis­ ual conception. She later is shown some gold and forms a visual conception of it, "that stuff," and she acquires a new piece of information­ individuated opaquely-to the effect that those previously read about embellishments are made of that stutT. Again, if the knowledge argument were unrestrictedly valid, it would follow that that stutTis not identical with Au. This case differs from the case of Max by involving not a descrip-

599

tive mode of presentation but (as we might say) a perceptual mode of presentation. It is not difficult to find a difference between both these cases and the case of Mary. Max lacks knowledge of the bottle's contents under a con­ tingent description of it-"ingredient of wine and beer that makes you intoxicated." What Margot lacks is a certain visual conception of Au, which is to say gold. This typically would not be a de­ scriptive conception; it would not self-consciously take the form "the stutTthat occasions this type of visual experience." Still on the face of it such a concept implicates a visual-experience type. For it picks out the kind it picks out by virtue of that kind's occasioning experiences of that type. And that is a crucial contingency in how the concept that Margot lacks is related to its reference. I hope I will be understood, then, if I say that the visual take on Au that Margot lacks would have conceived Au 'under a contingent mode of presentation' . This brings us back to Mary, whose acquired conception of what it is like to see red does not conceive it under a contingent mode of presenta­ tion. She is not conceiving of a property that presents itself contingently thus: it is like such and such to experience P. Being experienced like that is essential to the property Mary conceives. She conceives it directly. When Mary later acquires new information about us (construed opaquely), the novelty of this information cannot be ex­ plained-as in the case of Margot-as her ac­ quiring a new contingent mode of presentation of something she has otherwise known of all along. She has a direct grasp of the property involved in the new information; she conceives of it somehow, but not under a contingent mode of presentation. Proponents of the knowledge argument will say that is why it is valid on an opaque reading: there is no contingency in Mary's conception of the new phenomenal information that explains it as a novel take on old facts. She learns new facts sim­ pliciter and not new conceptions of old facts. Notice how close this comes to Saul Kripke's well-known antiphysicalist argument (1 980).

600

Brian Loar

Kripke assumes that a phenomenal concept such 'pain' cannot be a priori linked with a physical concept such as that of the stimulation of C­ fibers. The case of Mary is a vivid way of making the same point. Kripke points out that property identities can be true even if not a priori, for example, 'heat = such and such molecular prop­ erty'. It seems fair to represent the next step in his argument as follows. 'Heat' has a contingent higher-order mode of presentation that connotes the property 'feeling like this'. That is what ac­ counts for the a posteriori status of the identity. But, as Kripke points out, this cannot be how 'pain' works: the phenomenal concept 'pain' does not pick out its referent via a contingent mode of presentation; it conceives pain directly and essen­ tially. Kripke concludes that pain is not identical with a physical property. The two arguments then tum on the same im­ plicit assumption. The only way to account for the a posteriori status of a true property identity is this: one of the terms expresses a contingent mode of presentation. This ought to be giv.en a place of prominence. as

(Semantic premise) A statement of property identity that links conceptually independent concepts is true only if at least one concept picks out the property it refers to by connoting a contingent property of that property. The knowledge argument and Kripke's argument then depend on two assumptions: the conceptual independence of phenomenal concepts and phys­ ical-functional concepts, which I accept, and the semantic premise, which I deny. The antiphysicalist intuition that links concept­ individuation and property-individuation (more closely than is in my view correct) is perhaps this. Phenomenal concepts and theoretical expressions of physical properties both conceive their refer­ ences essentially. But if two concepts conceive a given property essentially, neither mediated by contingent modes of presentation, one ought to be able to see a priori-at least after optimal re-

flection-that they pick out the same property. Such concepts' connections cannot be a posteri­ ori; that they pick out the same property would have to be transparent. But as against this, if a phenomenal concept can pick out a physical property directly or es­ sentially, not via a contingent mode of presenta­ tion, and yet be conceptually independent of all physical-functional concepts, so that Mary's his­ tory is coherent, then Jackson's and Kripke's arguments are ineffectual. We could have two conceptually independent conceptions of a prop­ erty, neither of which connote contingent modes of presentation, such that substituting one for the other in an opaquely interpreted epistemic con­ text does not preserve truth. Even granting that our conception of phenomenal qualities is direct, physicalism would not entail that knowing the physical-functional facts implies knowing, on an opaque construal, the phenomenal facts; and so the failure of this implication would be quite compatible with physicalism. The next few sec­ tions give an account of phenomenal concepts and properties that would justify this claim. 2 Recognitional Concepts Phenomenal concepts belong to a wide class of concepts that I will call recognitional concepts. They have the form 'x is one of that kind'; they are type-demonstratives. These type-demonstra­ tives are grounded in dispositions to classify, by way of perceptual discriminations, certain ob­ jects, events, situations. Suppose you go into the California desert and spot a succulent never seen before. You become adept at recognizing instances, and gain a recognitional command of their kind, without a name for it; you are disposed to identify positive and negative instances and thereby pick out a kind. These dispositions are typically linked with capacities to form images, whose conceptual role seems to be to focus thoughts about an identifiable kind in the absence of currently perceived instances. An image is pre-

60\

Phenomenal States

sumably 'of' a given kind by virtue of both past recognitions and current dispositions.

a key role below in the account of third-person

Recognitional concepts are generally formed against a further conceptual background.

In

identifying a thing as of a recognized kind. we almost always presuppose a more general type to which the kind belongs: four-legged animal, plant. physical thing, perceptible event.

A

recog­

nitional concept will then have the form 'physical thing of that (perceived) kind' or 'internal state of that kind',

and so forth. 5

Here are some basic features of recognitional concepts that it will help to have in mind in con­ nection with the account of phenomenal concepts that follows.

I.

up close)' makes perfectly good sense. This plays ascriptions of phenomenal concepts. (This casual invoking of reference-determining dispositions will be a red flag for many who are aware of the vexing foundations of the theory of reference. Problems about referential scrutability, rule-following, naturalizing intentionality-how­ ever one wishes to put it-are as frustrating as any in contemporary philosophy. I do not pro­ pose to address them here. The idea rather is to appeal to unanalyzed common sense concerning a natural group of concepts and apparent con­ ceptual abilities. The apparent irreducibility of phenomenal qualities itself arises from appeal to

You can understand 'porcelain' from a tech­

intuitions independent of the theory of reference;

nical description and only later learn visually,

and it seems reasonable that we should, in re­

tactually, and aurally to recognize instances. By

solving that issue, appeal to notions that arise at

contrast, in the phenomenon I mean the concept

the same intuitive level. That we

is recognitional at its core; the original concept is

recognitional concepts and identifying disposi­

appear to

have

recognitional.

tions that are more or less determinate in their

2. A

reference is hard to deny. My conception of 'those

recognitional concept need involve no ref­

erence to a past instance, or have the form 'is of the same type as that (remembered) one' . You can forget particular instances and still judge 'another one of those'.

3.

Recognitional abilities depend on no con­

sciously accessible analysis into component fea­ tures; they can be irreducibly gestalt.

hedges' (seen around the neighborhood) may un­ ambiguously pick out a variety of eugenia. An example closer to the present topic is this. We can imagine an experiment in which the experimenter tries to determine which internal property is the focus of her subject's identifications: 'again', . . . 'there it is again' . There seems no commonsens­ ical implausibility-putting aside foundational

4. Recognitional concepts are perspectival. Sup­

worries about the inscrutability of reference-in

pose you see certain creatures up close and form

the idea that there is a best possible answer to

a recognitional concept- 'those creatures) '; and suppose you see others at a distance, not being

the experimenter's question, in the scientific long run. 6

able to tell that they are of the same kind (even when they are), and form another recognitional concept -'those creatures2 '. These concepts will be a priori independent. Now the respect in which

3 Phenomenal Concepts as Recognitional Concepts

they differ is perspectival, in some intuitive sense.

A recognitional concept is in part individuated by

Here is the view to be defended. Phenomenal

its constitutive perspective. Here is the important

concepts are recognitional concepts that pick out

point: a recognitional concept can be ascribed

certain internal propertie�; these are physical­

outside its constitutive p:rspective; 'that thing

functional properties of the brain. They are the

(seen at distance) is one of those creatures) (seen

concepts we deploy in our phenomenological reflections; and there is no good philosophical

Brian Loar

reason to deny that, odd though it may sound, the properties these conceptions phenomenologically reveal are physical-functional properties-but not of course under physical-functional descrip­ tions. Granted that brain research might discover that (what we take to be) our phenomenal con­ cepts do not in fact discriminate unified physical­ functional properties. Failing that, it is quite coherent for a physicalist to take the phenomen­ ology at face value: the property of its being like this to have a certain experience is nothing over and above a certain physical-functional property of the brain. Phenomenal concepts are conceptually inde­ pendent of physical-functional descriptions, and yet pairs of such concepts may converge on, pick out, the same properties. Rebutting the semantic premise of the knowledge argument requires making sense of the idea that phenomenal con­ cepts conceive physical-functional properties 'di­ rectly', that is, not by way of contingent modes of presentation. The objective is to show that the knowledge argument fails for the same reason in the case of Mary as in the case of Max: both ar­ guments require substitution in opaque contexts of terms that are conceptually independent. In the case of Max, the conceptual independence appears to derive from 'aIcohol"s connoting a contingent mode of presentation that is metaphysically inde­ pendent of the property referred to by the chem­ ical concept. In the case of Mary it has a different source. What then accounts for the conceptual inde­ pendence of phenomenal and physical-functional concepts? The simple answer is that recognitional concepts and theoretical concepts are in general conceptually independent. It is true that reeogni­ tional concepts other than phenomenal concepts connote contingent modes of presentation that are metaphysically independent of the natural kinds they pick out, and hence independent of the kind referred to by the theoretical term of the pair. But we need not count this metaphysical independence as essential to the conceptual in­ dependence of coreferring recognitional and

602

theoretical concepts. Concepts of the two sorts have quite different conceptual roles. It is hardly surprising that a recognitional conception of a physical property should discriminate it without analyzing it in scientific terms. Nor should it be surprising that, if there are recognitional concepts that pick out physical properties not via con­ tingent modes of presentation, they do not discriminate their references by analyzing them (even implicitly) in scientific terms. Basic reeogni­ tional abilities do not depend on or get triggered by conscious scientific analysis. If phenomenal con­ cepts reflect basic recognitions of internal physi­ cal-functional states, they should be conceptually independent of theoretical physical-functional descriptions. That is what you expect quite apart from issues concerning physicalism. An antireductionist may reply that the physi­ calist view depends on an ad hoc assumption and that it is tendentious to suppose that phenomenal concepts differ from all other reeognitional con­ cepts in not having contingent modes of pre­ sentation. But this is not fair. Even on the antiphysicalist view, phenomenal concepts are recognitional concepts, and we have 'direct' recognitional con­ ceptions of phenomenal qualities, that is, con­ ceptions unrnediated by contingent modes of presentation. Evidently it would be absurd to in­ sist that the antiphysicalist hold that we conceive of a phenomenal quality of one kind via a phe­ nomenal mode of presentation of a distinct kind. And why should the physicalist not agree that phenomenal reeognitional concepts are struc­ tured in whatever simple way the antiphysicalist requires? That is after all the intuitive situation, and the physicalist simply claims that the intuitive facts about phenomenal qualities are compatible with physicalism. The physicalist makes the ad­ ditional claim that the phenomenal quality thus directly conceived is a physical-functional prop­ erty. On both metaphysical views, phenomenal concepts differ from other reeognitional concepts; phenomenal concepts are a peculiar sort of ree­ ognitional concept on any account, and that can

603

Phenomenal States

hardly count against physicalism. The two views agree about conceptual structure and disagree about the nature of phenomenal qualities. To insist that physicalism implies, absurdly, that phenomenal concepts could pick out physical properties only via metaphysically distinct phe­ nomenal modes of presentation is unmotivated. There is, though, still more to be said about whether phenomenal concepts should be regarded as having modes of presentation of some sort, and we continue the account in section 5. Suppose this account of how phenomenal con­ cepts refer is true. Here is a semantic conse­ quence. The physicalist thesis implies that the judgments "the state a feels like that" and "the state a has physical-functional property P" can have the same truth condition even though their joint truth or falsity can be known only a posteri­ ori. I mean, same condition of truth in a possible world. For truth conditions are determined in part by the possible world satisfaction conditions of predicates; and if a phenomenal predicate di­ rectly refers to a physical property, that property constitutes its satisfaction condition. n this account, a phenomenal concept rigidly desIgnates the property it picks out. But then it rigidly designates the same property that some theoretical physical concept rigidly designates. . co�ld seem problematic, for if a concept rig­ Idly desIgnates a property not via a contingent mode of presentation, must that concept not cap­ ture the essence of the designated property? And if two concepts capture the essence of the same pr?perty, must we not be able to know this a pri­ on? These are equivocating uses of 'capture the essence of'. On one use, it expresses a referential notion that comes to no more than 'directly rigidly designate'. On the other, it means some­ thing like 'be conceptually interderivable with some theoretical predicate that reveals the in­ ternal structure of' the designated property. But the first does not imply the second. What is cor­ rect in the observation about rigid designation has no tendency to imply that the two concepts must be a priori interderivable.