a world for us

A WORLD FOR US

This page intentionally left blank

AWorld for UsThe Case for Phenomenalistic Idealism

JOHN FOSTER

1

1Great Clarendon Street, Oxford

Oxford University Press is a department of the University of Oxford.It furthers the Universitys objective of excellence in research, scholarship,

and education by publishing worldwide in

Oxford New York

Auckland Cape Town Dar es Salaam Hong Kong KarachiKuala Lumpur Madrid Melbourne Mexico City Nairobi

New Delhi Shanghai Taipei Toronto

With ofces in

Argentina Austria Brazil Chile Czech Republic France GreeceGuatemala Hungary Italy Japan Poland Portugal SingaporeSouth Korea Switzerland Thailand Turkey Ukraine Vietnam

Oxford is a registered trade mark of Oxford University Pressin the UK and in certain other countries

Published in the United Statesby Oxford University Press Inc., New York

John Foster 2008

The moral rights of the author have been assertedDatabase right Oxford University Press (maker)

First published 2008

All rights reserved. No part of this publication may be reproduced,stored in a retrieval system, or transmitted, in any form or by any means,

without the prior permission in writing of Oxford University Press,or as expressly permitted by law, or under terms agreed with the appropriate

reprographics rights organization. Enquiries concerning reproductionoutside the scope of the above should be sent to the Rights Department,

Oxford University Press, at the address above

You must not circulate this book in any other binding or coverand you must impose the same condition on any acquirer

British Library Cataloguing in Publication Data

Data available

Library of Congress Cataloging in Publication Data

Foster, John, 1941 May 5-A world for us: the case for phenomenalistic idealism/John Foster.

p. cm.Includes index.

ISBN 97801992971391. Idealism. 2. Realism. 3. Phenomenology. I. Title.

B823.F675 2008141dc222008001087

Typeset by Laserwords Private Limited, Chennai, IndiaPrinted in Great Britainon acid-free paper by

Biddles Ltd., Kings Lynn, Norfolk

ISBN 9780199297139

1 3 5 7 9 10 8 6 4 2

ToGeorge Berkeley

Preface

Put at its simplest, the aim of this work is to establish that the existence of thephysical world is logically sustained by the world-suggestive way in which,under Gods ordinance and authority, things are disposed to appear at thehuman empirical viewpoint. This idealist thesisa version of what I speakof as phenomenalistic idealismstands in sharp contrast to the commonlyaccepted realist view, which takes the world to have an existence that is bothlogically independent of the human mind and metaphysically fundamental.I argue that it is only by accepting the idealist thesis that we can represent thephysical world as having the empirical immanence it needs if it is to form aworld for us.

A world whose existence is logically sustained by how things are disposedto appear at the human empirical viewpoint sounds as if it would be, at best, avirtual realitya mere experiential simulation of a world. But in the idealismfor which I argue what logically sustains the existence of the world is not theworld-suggestive system of appearance on its own, but this system as ordainedand authorized by God, and I see this as making a crucial difference. TheGod in question is the God of the Judaeo-Christian tradition, and I conceiveof him as having the full perfection of nature that this tradition accords him.

The framework in which I hold this conception is one which representsGod as a personal beingas a rational mental subjectand so as an entityof the same general category as ourselves. In particular, within the context ofmy argument, I take God to be a being who has knowledge and purposes, andwho brings things about by the exercise of his will. All this, I am condent, isin line with what most ordinary Judaeo-Christian believers implicitly accept.At the same time, I realize that some philosophers and theologians will ndthe approach unacceptably crude. They may say, for instance, that while Godis indeed a being, the sort of being he is cannot be positively specied inany terms that we can understand. Or they may say that God should not bethought of as a being (as an entity among entities), but simply as being itself,or perhaps as active being. Or again, they may insist that the ontological natureof God is beyond our comprehension altogether. In one way or another,these philosophers and theologians will think that my representation of Godas a personal being detracts from his essential mystery, and thereby from the

viii Preface

perfection and unsurpassable greatness that the Judaeo-Christian traditionascribes to him. I do not discuss these delicate issues in the text of the book,nor shall I try to do so here. I am happy to acknowledge that the conceptionof God I employ might turn out to be in certain respects too crude, and standin need of modication. But if it does, this will not, in the end, undermine theforce of my argument. Provided that the modication does not prevent mefrom being able to think of God as ultimately responsible, and in some waypurposively responsible, for the way in which things are disposed to presentthemselves at the human empirical viewpoint, the role that God is requiredto play in my idealist account will not be affected. And a conception of Godthat did not allow me to assign such responsibility to him would not preservetraditional theism in any recognizable sense.

Some readers will be aware that this is not the rst book I have writtenwhich is devoted to a defence of phenomenalistic idealism. There are tworeasons why The Case for Idealism, published in 1982, does not makethe present work redundant. The rst is simply that while the new bookpreserves a signicant resemblance to the earlier one, it also includes extensivechangeschanges in the topics covered, in the nature of the idealist thesis Iendorse, in the arguments I develop in order to establish this thesis, and inthe objections to it that I consider and try to meet. Someone who has readand understood the earlier book should not, on reading the present one, feelthat he has travelled this route before. The second reason is that the wholestyle of the new book is, I hope, much more accessible than the rst: whereverpossible, I have avoided technicalities, and quite generally I have done myutmost to make a complex and difcult topic clear and comprehensible. Thetechnical precision of The Case for Idealism may have a certain virtue inthe abstract, but it made enormous, and arguably intolerable, demands onthe patience of the reader.

In writing the present book I have had to work more or less on myown, partly as a result of serious ill health, partly because it is hard to ndphilosophers who work in this area. So I have not on the whole been ableto gain the benet of friendly criticism and advice. A notable exception isprovided by the many discussions I have had over the years with HowardRobinson. He has helped me greatly to clarify and rene my thinking abouta number of issues that feature in the work, and I would like to record mythanks to him here.

There is one other philosopher I need to acknowledge, and in a sense thank:George Berkeley, the originator of phenomenalistic idealism as I conceive of

Preface ix

it. Berkeleys version of this idealism is not entirely to my liking. In particular,he has what is, from my standpoint, an impoverished view of the sorts ofentity and property that the idealistic world can contain. Sometimes, too, heseems to stray from the path of anything I would count as idealism at all.But it is to his vision of a world that is created by the orderly way in whichGod brings about our sensory experiences that my own approach can beultimately traced. Although I do not discuss Berkeleys idealism in any detail,in dedicating this book to him I am signalling my considerable debt to thatvision.

John FosterBrasenose College, Oxford

Contents

1. The Problem of Perception 1

2. The Inscrutability of Intrinsic Content 42

3. Realism and Phenomenalistic Idealism 83

4. The Refutation of Realism 123

5. The Challenge of Nihilism 164

6. The Issue of Objectivity 199

Bibliography 247Index 249

1The Problem of Perception

I

We ordinarily take it for granted that we have perceptual access to the physicalworld, and that through it we acquire our information about the world andour situation within it. But how should this perceptual access be understood?What is it for someone (a human subject) to perceive (perceive by his senses)a physical item? There are a number of specic theories of the nature of suchperception, but we can classify them all under two rival general views. It ison these rival views, and the issue between them, that I want, in this openingphase of the discussion, to focus. I shall do so, initially, within the frameworkof the common-sense assumption that the physical world is ontologicallyindependent of the human mindthat it is something whose existence islogically independent of facts about human mentality. The adoption of thisframework may seem hardly worth mentioning. How could the world ofspace and material objects be anything other than mind-independent? Butwhat I shall try to show is that, although the two rival views exhaust the rangeof possibilities, or at least do so for any given case of perception, neither ofthem can be made to yield a satisfactory account so long as the assumption ofmind independence is retained. It is this that creates what I see as the problemof perception.

Much of what I shall say, in elaborating this problem, draws on points thatI have developed in more detail in my book The Nature of Perception.

II

Before I can state the rival views, I need to introduce and explain a key concept.The concept is of something I call constitution, and it will play a major role not

John Foster, The Nature of Perception (Oxford: Oxford University Press, 2000).

2 A World for Us

only in the present context, where we are focusing on the topic of perception,but also throughout the rest of the discussion. Indeed, the central topic ofthe whole discussion concerns a certain thesis of constitutiona thesis aboutwhat constitutively sustains the existence of the physical world.

Constitution, in the sense I intend, is a two-place relation that holds withinthe domain of facts (instances of what is the case) and sets of facts. It can beprovisionally dened as follows:

A fact F is constituted by a fact F, or by a set of facts S, if and only if twoconditions are satised, namely:

(1) The obtaining of F is logically due to the obtaining of F (the obtainingof the members of S).

(2) The obtaining of F involves nothing over and above the obtaining ofF (the obtaining of the members of S).

When I speak here of the obtaining of a fact, I do not mean to refer to thehigher-order fact that the relevant fact obtains. I only mean to refer to thefact itself, but in a way that conveys its nature as an instance of what isthe case. So, when I speak of the obtaining of a fact F as logically due to (oras involving nothing over and above) the obtaining of a fact F, this is justanother way of saying that, if F is the fact that p, and if F is the fact that q,its being the case that p is logically due to (involves nothing over and above)its being the case that q.

Thus dened, constitution comes in two forms: a single-fact form, inwhich a fact is constituted by another (single) fact, and a multi-fact form, inwhich a fact is constituted by a set of two or more facts. Strictly speaking, thedenition also allows for the case of a fact being constituted by a set containingjust one fact. But this case can be ignored, since to speak of constitution bythe one-membered set {F} is just a contorted way of speaking of constitutionby its sole member F. Where a fact is constituted by a set of facts (a set oftwo or more facts), I shall also often speak of it as constituted by the factsthemselves, as a collective plurality. And, to emphasize the plural nature ofwhat is doing the constituting, I shall sometimes speak of the constitutedfact as breaking down, or decomposing, into these facts. Any case of multi-factconstitution can, of course, be recast as a case of single-fact constitution, bysimply replacing the relevant set of facts by a fact that conjoins its members;or, at least, this can be done if the number of its members is nite. But all thisshows is that a case of single-fact constitution is only interestingly single-factif the constitutive fact in question is not a conjunctive one.

The Problem of Perception 3

I have described the denition above as provisional. This is because itwill need to be slightly revised if it is precisely to capture the concept ofconstitution I have in mind. But before we can appreciate the need for thisrevision, we have to get clear about the nature of the denition in its presentform, and the content of the two conditions for constitution that it prescribes.

One thing which each of these conditions is to be understood as implyingis that the obtaining of F (the members of S) logically necessitates (logicallyensures, guarantees, sufces for) the obtaining of F. So, in the single-factcase, if F is the fact that p, and if F is the fact that q, each of the conditionsimplies that it is logically necessary that if q, then p; and in the multi-factcase, if F is the fact that p, and if S is the set of facts that q1, that q2, ,each of the conditions implies that it is logically necessary that if q1, q2, ,then p. By logical necessity, I mean, as is standardly meant, strict or absolutenecessity: it is logically necessary that p if and only if there is no possibleworld of any sortnot even a world with different natural lawsin whichit is not the case that p. As Kripke reminded the philosophical world, it isimportant not to confuse the question of what holds as a logical necessityin this sense with the question of what can be established a priori. We needempirical evidence to establish that water, as ordinarily conceived, is the samesubstance as H2O, as chemically dened. But granted that water and H2Oare the same substance, the fact of their identity is logically necessary in therelevant sense, since, if A is the substance water, and B the substance H2O,there is no possible world of any sort in which A and B are numericallydifferent.

Each of the conditions (1) and (2) is to be understood as implying that theobtaining of F (the members of S) logically necessitates the obtaining of F.But there are two respects in which such necessitation does not, on its own,sufce for constitution, as I am conceiving of it, and it is these respects whichbring to light the further implications of the two conditions.

In the rst place, logical necessitation is not, as such, asymmetric. There arecases in which the obtaining of a fact, or set of facts, logically necessitates theobtaining of a fact, or set of facts, and vice versa. The obvious example is thatin which the rst fact, or set of facts, and the second fact, or set of facts, arethe same, since, trivially, the obtaining of any fact logically necessitates itself.In contrast with this, I want the relation of constitution to be necessarilyasymmetric: I want the relation to be such that, where a fact F is constitutedby a fact F (or a set of facts S), F derives its obtaining from (owes its obtainingto) the obtaining of F (the members of S) in a way thaton pain, as it

4 A World for Us

were, of metaphysical circularityprecludes the same relationship holdingin reverse. It is this element of asymmetric dependence which, in additionto mere logical necessitation, is expressed by saying, in condition (1), thatthe obtaining of F is logically due to the obtaining of F (the obtaining of themembers of S).

Second, there are cases where one fact is logically necessitated by anotherfact or set of facts, but where its obtaining has, as it were, an ontologicallife of its own, that lies outside the obtaining of the fact or set of facts thatnecessitates it. For example, suppose that, at a certain time t1, God prescribesthat a certain kind of event will occur at the later time t2, and let F1 be the factthat God issues this prescription at t1, and let F2 be the fact that the relevantkind of event occurs at t2. Granted that God is (in some relevant sense)omnipotent, and is so essentially, the obtaining of F1 logically necessitatesthe obtaining of F2; but, given that the prescription and the subsequentevent are ontologically discrete items, that occur at different times, there isa clear sense in which the obtaining of F2 (its being the case that an eventof the relevant kind occurs at t2) is something separate fromsomethinggenuinely additional tothe obtaining of F1 (its being the case that Godissues the relevant prescription at t1). Now I want the relation of constitutionto exclude this kind of separateness: I want things to be such that, where a factis constituted by another fact or set of facts, its obtaining is wholly includedin the obtaining of this other fact or set of facts. And it is this inclusion that isexpressed by saying, in condition (2), that the obtaining of F involves nothingover and above the obtaining of F (the obtaining of the members of S).

Examples of constitution, in the sense dened, are not hard to nd. Oneclear-cut range of single-fact cases is marked out by the principle that theinstantiation of a generic (determinable) property is always constituted bythe instantiation of a property that is more specic (more determinate). Forinstance, if an object is (generically) red, the fact of its being so is constituted,in the relevant sense, by a more specic fact about its colour, such as its beingscarlet or carmine: the obtaining of the generic colour fact is logically dueto, and involves nothing over and above, the obtaining of the specic fact.A clear-cut range of multi-fact cases is provided by the relationship betweenone objects being heavier than another and their individual weights. Thus,if John weighs twelve stone and Mary weighs ten stone, the fact that Johnis heavier than Mary is constituted by the combination of the facts thatJohn weighs twelve stone and that Mary weighs ten stone: the obtaining oftheir weight relationship is logically due to, and involves nothing over and


above, the obtaining of these separate weight facts about them. These areclear-cut cases in the sense that the relevant claims of constitution are whollyuncontroversial: no one will think of denying that, in the kinds of situationenvisaged, the two prescribed conditions for constitution are satised in thespecied ways. Not surprisingly, the claims of constitution on which we shallbe focusing in our philosophical discussion will not be uncontroversial inthat way.

I have already indicated that the present denition of constitution willneed to be slightly revised. The point of the revision is to allow for cases inwhich the obtaining of a fact has two independent modes of constitution.To focus on a simple example, let us, again, suppose that John weighs twelvestone and Mary weighs ten stone. We can draw up the following list of factswhose obtaining is implicit in this supposition:

F1: the fact that John weighs twelve stone.F2: the fact that Mary weighs ten stone.F3: the fact that either John weighs twelve stone or 2 + 2 = 5.F4: the fact that either Mary weighs ten stone or 2 + 2 = 5.F5: the fact that either John weighs twelve stone or Mary weighs ten stone.

Whenever it is true that p and false that q, the obtaining of the fact that por q is logically due to, and involves nothing over and above, the obtainingof the fact that p. So, the obtaining of F3 is logically due to, and involvesnothing over and above, the obtaining of F1, and the obtaining of F4 islogically due to, and involves nothing over and above, the obtaining of F2.This means that F3 is constituted by F1 and that F4 is constituted by F2. Butnow consider how F5 stands to F1 and F2. Granted that F3 is constitutedby F1, we want to be able to say that F5, too, is constituted by F1, sincethe obtaining of F1 bears on the obtaining of F3 and on the obtaining of F5in exactly the same way. Similarly, granted that F4 is constituted by F2, wewant to be able to say that F5, too, is constituted by F2, since the obtainingof F2 bears on the obtaining of F4 and on the obtaining of F5 in exactlythe same way. But these are not things that we are able to say under ourpresent denition of constitution. Given that it is logically necessitated bythe obtaining of F2, we cannot say, without qualication, that the obtainingof F5 is logically due to, or involves nothing over and above, the obtainingof F1; and, given that it is logically necessitated by the obtaining of F1, wecannot say, without qualication, that the obtaining of F5 is logically due to,or involves nothing over and above, the obtaining of F2. All we can say is

6 A World for Us

that, with respect to its necessitation by F1, the obtaining of F5 is logically dueto, and involves nothing over and above, the obtaining of F1, and that, withrespect to its necessitation by F2, the obtaining of F5 is logically due to, andinvolves nothing over and above, the obtaining of F2.

The solution is to reformulate the two conditions for constitution in away that relativizes what they claim to the context of a single source ofnecessitation. So, the revised denition of constitution will read:

A fact F is constituted by a fact F, or by a set of facts S, if and only if theobtaining of F is logically necessitated by the obtaining of F (the obtainingof the members of S), and, with respect to that source of necessitation:

(1) the obtaining of F is logically due to the obtaining of F (the obtainingof the members of S);

(2) the obtaining of F involves nothing over and above the obtaining ofF (the obtaining of the members of S).

This will allow us to say that F5 is separately constituted by both F1 and F2;and it will accommodate a whole range of cases of a similar kind, in which asingle fact has two or more independent modes of constitution.

Although the new denition is needed for the purposes of precision, theissue of independent modes of constitution will not be relevant to the topics ofour future discussion, and in what follows I shall normally, for convenience,continue to express the conditions for constitution in their simpler originalform, without relativizing their content to a source of necessitation.

III

With the concept of constitution in place, we must now turn to the topicof physical-item perception, and the two rival general views about its nature.In setting out and discussing these views, I shall work on the assumptionthat we can think of human subjects as perceiving items at points (moments)of time, with zero temporal extent. This assumption does not imply, whatwould clearly be incorrect, that a momentary instance of perceiving canoccur in isolation, without being contained within an extended episode ofperceiving, and once that point is understood, the assumption may notseem unreasonable. But, in any case, my reason for adopting it is moreone of expositional convenience than philosophical conviction; and certainlynothing of substance will turn on it. If anyone objects to the assumption,


on the grounds that any genuine instance of perceiving must be temporallyextended, I am happy for him to read what I say with the adjustments that hisposition requires. To harmonize with my recognition of momentary instancesof perceiving, I shall also recognize certain kinds of momentary physical itemto serve as objects, or potential objects, of perception. In particular, amongsuch objects, I shall recognize the momentary stages of persisting physicalitems, where each such stage is something whose existence covers all and onlywhat is covered by the existence of the relevant persisting object at a particularpoint in time. Anyone who wants to make adjustments to what I say in thecase of momentary instances of perceiving will need to make appropriateadjustments here, too. Strictly speaking, some of the things that I represent asmomentary stages of persisting items are really entities whose spatial points,while of zero temporal extent, are spread over a range of moments. But, forsimplicity of exposition, that is something I shall largely ignore.

To enable me to formulate the two rival views clearly and concisely, it willbe helpful if I begin by making terminological provision for certain furtherkey notions, specic to the topic of perception.

First, given a subject S and two items x and y that he simultaneouslyperceives, I shall say that Ss perceiving of x (the fact of his perceiving x) ismediated by his perceiving of y (the fact of his perceiving y) if and only if(1) Ss perceiving of x is constituted by the combination of his perceiving ofy and certain additional facts, and (2), apart from any concern they may havewith Ss perceiving of y, these additional facts do not involve anything aboutSs perceptual condition at the relevant time. Where a subjects perceiving ofone item is mediated, in this way, by his perceiving of another, we can speakof him as perceiving the second item more immediately than he perceives therst. I shall illustrate this notion of perceptual mediation presently.

Second, I shall say that a subject S -terminally perceives an item x at atime t if and only if x is a physical item and S perceives x at t and there isno other physical item y such that Ss perceiving of x at t is mediated byhis perceiving of y at t. So, the perceiving of an item qualies as -terminaljust in case the item is physical and there is no other physical item that is,in the context of that perceiving, perceived more immediately. (This leavesopen the possibility of there being a non-physical item that is perceived moreimmediately.) I shall take it for granted that physical-item perception is notinnitely regressive, and that whenever a physical item is perceived, thereis some physical item which is, in respect of that perception, -terminallyperceived.

8 A World for Us

Finally, when I speak of a psychological state as in itself physically perceptive,I mean that it is logically impossible for someone to be in that state withoutthereby perceiving a physical item, and when I say that a psychological stateis in itself perceptive of an item x, I mean that it is logically impossible forsomeone to be in that state without thereby perceiving x. Obviously, I am hereusing the term state to mean type state, not token state. So, a psychologicalstate is something that is capable of realization in different subjects and ondifferent occasions.

The two rival views of perception can now be set out as follows. Accordingto the rst view, whenever someone perceives a physical item, there is somepsychological state which is not in itself physically perceptive, and the fact ofhis perceiving that item breaks down into (is constituted by the combinationof) two components. One of these components consists in his being in thatstate. The other comprises certain additional facts, but ones that do notinvolve anything further about his psychological condition at the relevanttime. In practice, the advocate of this view will take these additional factsto concern, or concern amongst other things, the qualitative relationshipof the psychological state to the physical item (or to the item -terminallyperceived), and the role of the item in causing the subject to be in that stateat that time. I shall call this rst view the decompositional view. According tothe second view, whenever someone -terminally perceives a physical item,the fact of his perceiving it is something psychologically fundamental. It isnot something that breaks down into (is constituted by the combination of)his being in some further psychological state, which is not in itself physicallyperceptive, and certain additional facts, not involving anything further abouthis current psychological condition. It does not, at the psychological level,break down into further facts at all, except perhaps (if this is possible) in apurely trivial way, where the perceived item is something complex and thesubjects perceiving of it breaks down into the separate perceivings of its parts.In other words, the psychological state that is fundamentally involved in theperceiving of the relevant physical item is one that is in itself perceptive ofthat item. I shall call this second view the fundamentalist view.

It is obvious that, for any case of physical-item perception, the two viewsare, with respect to the -terminal perceiving involved, mutually exclusive:they are explicitly formulated so as to be so. It is also true, though not

The decompositional and fundamentalist views coincide with what, in The Nature of Perception,I respectively labelled the broad representative theory and strong direct realism, except that the latterpositions were dened as ones that explicitly endorsed a realist conception of the physical world.


guaranteed by their mode of formulation, that the views are, case by case,jointly exhaustive. For where an item is -terminally perceived, the onlyway in which we could sensibly think of the psychological state that isfundamentally involved as failing to be in itself perceptive of that item wouldbe by thinking of it as failing to be in itself physically perceptive altogether;and if the state is not in itself physically perceptive, the decompositionalaccount of the perceiving is the only one available.

We can best bring out the nature of the two views by focusing on aparticular case. Suppose, at a certain time t, Ralph sees an apple on the tablein front of him.His seeing of the apple is mediated, in the sense dened, by hisseeing of a certain portion of its surface (it is constituted by the combinationof his seeing this portion and the fact that this latter item is a portion ofthe apples surface), and his seeing of this persisting surface portion is, in anexactly analogous way, mediated by his seeing of a certain momentary stageof ita stage which, depending on the precise temporal length of the causalprocess from the surface portion to its visual registering at t, occurs slightlyearlier than t. This momentary stage of the apples surface portion is thenwhat, in the sense dened, Ralph -terminally sees at twhat, relative to thedomain of physical candidates, he most immediately sees. It is in terms of hisvisual contact with this -terminal object that the two views come into sharpconict. Let us call this -terminal object O. The fundamentalist will saythat, given that O is the -terminal object of perception, Ralphs seeing of Ois a fundamental aspect of his psychological condition at t: his fundamentalpsychological state, though wholly a matter of what is occurring within hismind at that time, inherently involves his standing in this awareness relationto this external physical item. In contrast, the decompositionalist will saythat, instead of being psychologically fundamental, the fact of Ralphs visualcontact with O breaks down into two components. One component willcover all the relevant aspects of Ralphs psychological condition at tall thatobtains or occurs in his mind at that time that in any way logically contributesto the obtaining of the relevant perceptual factand the decompositionalistwill insist that these aspects, on their own, do not secure perceptual contactwith O or sufce for physical-item perception at all. The other componentwill cover the remaining facts that are relevant to the securing of visualcontact with Ofacts that do not involve anything further about the

This is an example of something that I am representing as a momentary stage, but which inreality, because of the slightly varying length of the causal process from different points on thesurface portion, is spread over a range of moments.

10 A World for Us

subjects psychological condition at t. In practice, decompositionalists willtake the rst component to consist in the fact of Ralphs having a certainkind of visual experience at t, and will take the second component to consistin, or centrally involve, facts concerning the qualitative relationship of thisexperience to O and the nature of the causal process from O to the occurrenceof the experience. Fundamentalists too will take the seeing of O to involvethe occurrence of a visual experience and will accept that this experienceoccurs at the end of a causal process starting from O. But they will take thisexperience to be one which, by its intrinsic psychological nature, puts thesubject into perceptual contact with O. They will not think that either thequalitative relationship of the experience to the external item or the nature ofthe causal process from this item to the experience constitutively contributesto the securing of this contact.

I have said that both decompositionalists and fundamentalists will recognizethe existence of a causal process from the perceived physical item to the visualexperience. But the ways in which they will understand the nature of thisprocess are very different. For the decompositionalist, the nature of thisprocess will be, in its general character, relatively straightforward. Lightreected from O enters Ralphs eyes, producing a certain pattern of rings inhis optic nerves, which in turn brings about a further complex event in therelevant part of his brain, which nally results in the occurrence of the visualexperience. At each stage in this process the character of what happens directlycausally depends only on the character of what takes place immediately beforeit and the currently prevailing conditions. So the pattern of rings in the opticnerves directly depends only on the character of the photic input, togetherwith the relevant facts about the structure and prevailing state of the subjectseyes and nervous system; the character of the brain event directly dependsonly on the character of the optic-nerve rings, together with the relevantfacts about the structure and prevailing state of the brain; and, crucially, thecharacter of the visual experience directly depends only on the character ofthe brain event and prevailing brain conditions, together with the relevantfacts about the character of the subjects mind and the form of its cerebralembodiment. In other words, the causal process works in a standard serialway, whereby each stage in the process contributes to the nal outcome onlyin so far as it affects what immediately follows it. But this cannot be how

They could still, of course, think that this psychologically fundamental contact logically couldnot have occurred without the experience being qualitatively and/or causally related to O in acertain way.


the fundamentalist will see the causal process. This becomes clear when weexpand the original example to include a second seeing of the apple. Thus,suppose, after seeing the apple at t, Ralph closes his eyes for a few seconds, andthen, on opening them, has another experience of seeing the apple, this time att. And let us call the -terminal object of this t episode of seeing O , an itemwhich we can assume to be a correspondingly later momentary stage of thesame portion of the apples surface. Let us also suppose that the causal processfrom O to brain is of exactly the same kind as that from O to brain, thatthe resulting brain events are of exactly the same kind, and that the cerebraland mental conditions in which these events occur are relevantly the same.Despite the qualitative identity of these processes, events, and conditions,the fundamentalist will take the resulting psychological eventsthe twovisual experiencesto be, and to be at the fundamental level of description,of different psychological types, since he will say that one of these events(experiences), by its psychological character alone, puts Ralph into visualcontact with O, and that the other, by its psychological character alone, putshim into visual contact with O, a momentary stage that is later than, andso numerically different from, O. Clearly, then, the fundamentalist cannot,like the decompositionalist, think that the psychological outcomes in thetwo cases directly causally depend only on the character of the brain eventsthat immediately precede them and the conditions in which these eventsoccur. Rather, he will have to recognize certain additional factors which, bycombining with these events and conditions, account for why the resultingoutcome is in the one case a seeing of O and in the other case a seeing of O.And it is not difcult to see what these factors will be. Obviously, from thestandpoint of the fundamentalist view, what will combine with the relevantbrain event and prevailing conditions to causally ensure a seeing of O at t willbe the whole causal process from O to brain, and, in particular, the role of O(as distinct from any other physical item) as the initiator of that process; andwhat, subsequently, will combine with the relevant brain event and conditionsto causally ensure a seeing of O at t will be the whole causal process from O

to brain, and, in particular, the role of O (as distinct from any other physicalitem) as the initiator of that process. So, in each case, the causal processfrom the relevant physical item to the brain will play a double causal rolewith respect to the psychological outcome, being both what brings about therelevant brain event immediately preceding that outcome and what combineswith that event, and the other relevant aspects of the prevailing conditions, todetermine the specic perceptive character of the outcome. This may seem,

12 A World for Us

at rst sight, a strange way for causation to operatewhere the nature ofwhat occurs at a given time is affected not just by what occurs and obtainsimmediately prior to that time, but also by the sequence of events leadingup to it. But, on reection, we can see that this kind of causation is preciselywhat ts the distinctive character of the fundamentalist position, in which,on any perceptual occasion, contact with the -terminally perceived item istaken to be a fundamental aspect of the subjects psychological condition,and where that item occurs earlier than the time of the perceiving. Onceit is accepted that the fundamental character of the psychological outcomeincorporates the existence of the perceived item and its perception by thesubject, there is nothing strange about supposing that the process from theitem to the subjects brain has a direct causal inuence on it.

There is one other aspect of the fundamentalist view that needs to beunderlined here. The fundamentalist will insist that the experience of seeingO and the experience of seeing O differ in their fundamental psychologicalcharacter. But he is not obliged to say that they differ in their subjective(phenomenological) characterin how they introspectively appear to thesubject, or would appear if properly scrutinized. Indeed, he is likely toacknowledge that, since the relevant brain events and conditions in the twocases are of exactly the same type, the chances are that the way in whichO sensibly appears to Ralph at t and the way in which O sensibly appearsto him at t will be exactly the same as well, and that, in consequence, thetwo visual experiences will be, in terms of their introspective appearance,indistinguishable. There is nothing paradoxical in this. Once it has beenaccepted that the psychological state that is fundamentally involved in anyperception is in itself perceptive of a particular physical item, it is inevitablethat there is an aspect of what is going on psychologically that cannot revealitself introspectively.

One reason for underlining this point is that it removes the temptationto attempt an instant refutation of the fundamentalist view by appeal tothe phenomenon of hallucinationthe phenomenon of experiences whichsubjectively pose as perceptions of certain kinds of physical item, but are notphysically perceptive at all. Thus, on rst encounter, we might have thoughtthat we could prove the falsity of the view by advancing the following simpleargument:

(1) For any physical-item perception, there could be a subjectively matchinghallucination.


(2) So, for any physical-item perception, there could be an experience whichis exactly the same in its fundamental psychological character, but is notperceptive of any physical item.

(3) So, the psychological states fundamentally involved in physical-itemperception are not in themselves physically perceptive.

(4) So, the fundamentalist view is false.

But, given the nature of his position, the fundamentalist will simply rejectthe move from (1) to (2). And he can do so without embarrassment.For if there is no problem in his holding that the experiences of seeingO and O may be subjectively indistinguishable, while differing in theirfundamental psychological character, there is obviously no problem inhis holding that experiences that are physically perceptive differ in theirfundamental psychological character from those that are not. Nor, as issometimes thought, can we create a problem by focusing on a case where,as well as the perceptive and hallucinatory experiences subjectively matching,the cerebral and mental factors that are causally involved in their occurrenceare of exactly the same type. For, as we have seen, the fundamentalist doesnot suppose that such cerebral and mental factors are the only factors thatcausally contribute to the psychological character of the resulting experience.If, in the case of Ralph, he can say, without embarrassment, that therole of the environmental item in producing the relevant brain event iscausally responsible for ensuring that the resulting visual experience is initself perceptive of that item, he can also say, without embarrassment, thatthe presence of an environmental item that plays the right kind of role with

Distinguishing the fundamental psychological characters of perceptive and hallucinatoryexperiences is an instance of what is often described as a disjunctivist approach to the nature ofperceptual experience. The approach originates with the work of J. M. Hinton; see, in particular,his Experiences (Oxford: Oxford University Press, 1973). Several philosophers have followedHinton in offering disjunctivist accounts of the content of perceptive and hallucinatory experience,including Paul Snowdon (Perception, vision, and causation, Proceedings of the Aristotelian Society,81 (19801), 17592, The objects of perceptual experience, Proceedings of the Aristotelian Society,supp. 64 (1990), 12150), John McDowell (Criteria, defeasibility, and knowledge, Proceedings ofthe British Academy, 68 (1982), 45579), and Michael Martin (The transparency of experience,Mind and Language, 17/4 (2002), 376425, The limits of self-awareness, Philosophical Studies,120 (2004), 3789).

Such a thought is implicit in Howard Robinsons argument against naive realism in his Thegeneral form of the argument for Berkeleian Idealism, in J. Foster and H. Robinson (eds.), Essays onBerkeley (Oxford: Oxford University Press, 1985), 16386, an argument that he further developsin his Perception (London: Routledge, 1994), 15162. I, too, once thought that the fundamentalistview could be undermined by considerations of this sort. Thus, see my Ayer (London: Routledge &Kegan Paul, 1985), 1479, 161.

14 A World for Us

respect to what happens in the subjects brain is causally required if theresulting experience is to be in itself physically perceptive at all.

There would still, of course, be the question of what account the fun-damentalist should give of the psychological character of hallucinationacharacter which he has to take to be fundamentally different from that ofperception, while able to simulate it subjectively. It is possible that therewould be problems for him on that front. But this is not an issue that wecan hope to address until we have considered in more detail his position onthe nature of perception. When we do, we shall see that, quite apart fromany issue about hallucination, his fundamentalist account of the -terminalperceptual relationship is untenable.

IV

I have formulated the two rival general views about the nature of physical-itemperception, and illustrated them by reference to a particular case. It is also clearthat this case is entirely typical, and that the way in which the two views workout in other cases of -terminal perceiving would be, in all basic respects, thesame. The next task must be to try to evaluate the views, and so decide theissue between them. I shall begin by considering the fundamentalist view.This, as we have seen, claims that, in any case of -terminal perceiving, theperceptual relationship between the subject and the perceived physical item issomething psychologically fundamentalsomething which does not, at thepsychological level, decompose into further factors. I have already stressedthat the fundamentalist view cannot be refuted by a simple appeal to thephenomenon of hallucination. But there is a quite different area where itscapacity to provide an adequate account needs to be scrutinized, and it is onthis that I now want to focus.

Whenever someone perceives a physical item, he perceives it, or at leastperceives whatever it is that he -terminally perceives, under a certain sensibleappearance. Thus, when Ralph sees the apple, it (or, at any moment, theportion stage of it he -terminally sees) visually appears to him as a roughlyhemispherical patch of a certain size and colouring, located at a certaindistance in front of him (or, more precisely, whose various parts are locatedat various distances and in various directions from him). Likewise, when I

For a fuller elaboration of this point see my The Nature of Perception, 2343.


take a bottle of wine from the fridge, the bottle (or, at any moment, theportion stage of it I -terminally feel) tactually appears to me as the hard,cold, smooth, curved surface of something I am holding. Quite generally,in any case of -terminal perceiving, the perceived item sensibly appears tothe subject in a certain waya way which represents it as an environmentalitem with a certain sensible character, and, at least typically, assigns to it,in perspective, a more specic environmental location, or set of locations,relative to the subjects own body or current position. The sensible characterthat the perceived item appears to possess draws its elements exclusivelyfrom qualities of spatial and temporal arrangement and qualities that aredistinctively associated with the relevant sense realm. The precise nature ofthis distinctive association need not concern us, but I am thinking of the kindof association which qualities of sensible colour have with the visual realm,qualities of sensible hardness have with the tactual realm, qualities of sensiblesound have with the auditory realm, and so on.

Not every way in which a -terminally perceived item appears to itspercipient counts as part of its sensible appearance in the relevant sense. Ralphmay well see the relevant -terminal object not only as something of a certainshape, size, and colouring, but also as an apple, or as part of the surface ofan apple. But, if so, this further aspect of how the item appears to him doesnot count as an aspect of its sensible appearance. Likewise, I may feel the -terminal object of my tactual perception not just as a surface of a certain shape,hardness, texture, and temperature, but also as the surface of a bottle, and,again, this further aspect of how the item may appear to me would not countas an aspect of its sensible appearance. Setting out precisely what it is thatmakes sensible appearance distinctive is a complicated matter, as I have shownelsewhere; in particular, it cannot be done solely by reference to the kindsof qualitative element that feature in its content and their relationship to therelevant sense realm. For example, whiteness has the right sort of associationwith the visual realm to feature in the content of visual sensible appearance.But if someone familiar with its daytime appearance inspects a white owerin his garden at night, when it sensibly appears to him as something grey, itmay still, in a certain sense, look white to him (the same sense in which a ripestrawberry may look sweet to someone, or a stagnant pool might look smelly),and this looking white is not an instance of sensible appearance. As we shallshortly see, one of the things that distinguishes sensible appearance from other

In The Nature of Perception, 4451.

16 A World for Us

forms of perceptual appearance is its distinctive phenomenological character;but that is not something I can explain until after I have introduced anddened a further key concept. At present, all we need to note is that, in anyinstance of -terminal perceiving, the perceived items sensible appearanceis, in a distinctive way, the basic core of its perceptual appearance, whichis integral to the perceptual relationship itself. A physical item cannot be-terminally perceived except under a sensible appearance. But it is possiblefor there to be instances of -terminal perceiving without any elements ofnon-sensible appearance, and, in any particular case of -terminal perceiving,any elements of non-sensible appearance could be stripped away withoutaffecting perceptual contact with the relevant physical item.

For an item to sensibly appear to a subject in a certain way is for the subjectto perceive that item in a certain experiential way. So, whenever a physicalitem is -terminally perceived, there is something in the content of thesubjects perceptual experience that embodies the items sensible appearancein its mental aspectsomething which forms, as it were, the way in which thesubject is sensibly appeared to. I shall refer to this element in the perceptualexperience as its phenomenal content. Phenomenal content need not form thetotal content of the experience. If Ralph sees the -terminal item as part ofthe surface of an apple, that, too, will be part of the content of his perceptualexperience. But, in the way that the sensible appearance of a perceived itemforms the basic core of its perceptual appearance, the phenomenal content ofa perceptual experience forms the basic core of its experiential content. I haveintroduced this concept of phenomenal content in a context where we arefocusing on the case of physically perceptive experience, where the content ofthe experience forms the experiential manner in which the relevant physicalitem is -terminally perceived. But, although this will be the main area of ourconcern, it will be convenient to allow the concept to apply, additionally, tothe phenomenologically corresponding content of hallucination, where theexperience has the subjective character of physical-item perception, withoutbeing physically perceptive. In other words, it will be convenient to think ofphenomenal content as the way in which the subject is sensibly appeared to,whether or not there is a physical item that is doing the appearing. This, I shouldstress, leaves entirely open the issue of whether the psychological character ofphenomenal content is the same in the two cases. And on this we can expect

I am here diverging from my understanding of this concept in The Nature of Perception, butthe difference is only one of terminology, not one of substance.


the decompositionalist and the fundamentalist to take opposite positions. Adecompositionalist claims that the psychological states that are fundamentallyinvolved in perception are not in themselves physically perceptive, and sois committed to saying that any experiential content that is fundamentallyinvolved in perception could also occur in hallucination. A fundamentalist,in contrast, holds that perceptive and hallucinatory experiences differ in theirfundamental psychological character, and, since the phenomenal contentof a perceptive experience forms the experiential manner in which the-terminal contact is achieved, it is almost inevitable that he will take thepsychological difference between the perceptive and hallucinatory cases toapply, in particular, to the nature of their phenomenal content.

Any adequate theory of perception has to provide an adequate accountof the nature of phenomenal content, as it occurs in perception, and of itsintimate involvement in the obtaining of the perceptual relationship betweenthe subject and what he -terminally perceives. It is the question of whetherthe fundamentalist can provide such an account that I now want to pursue.Is there an acceptable way of understanding what, in the perceptive case,phenomenal content is, and how it is involved in -terminal contact, when,in any case of such contact, the perceptual relationship between the subjectand the relevant physical item is taken to be psychologically fundamental?

The fundamentalists simplest approach would be to adopt what I shallcall the presentational view. Let us say that an item x is presented to a subjectS, or that S is presentationally aware of x, if and only if x is psychologicallyrelated to S in a way that satises three conditions. First, the relationship issuch as to make x available for demonstrative identication by S. In otherwords, it brings x before Ss mind in a way that allows him to pick it outas this item (of which he is now conscious, and on to which he directshis attention), or, at least, in a way that would allow him to do this if hehad the conceptual resources needed for demonstrative thought. Second, therelationship is such as to display, where appropriate in a certain perspective,certain aspects of xs character, or character and location, in a way that makesthem immediately available for cognitive scrutinythough, once again, Sscapacity to take advantage of this availability depends on his having therequisite conceptual resources. Third, and most crucially, the relationship iswholly non-representational. It does not involve the use of concepts, symbols,or mental images, as a psychological means of registering xs presence or therelevant (displayed) aspects of its character (character and location). Rather,the item and, in their concretely realized form, these aspects are before Ss

18 A World for Us

mind in a mode of absolute ontological immediacy, forming, in their ownperson, the very content of his awareness. What the presentational view thenclaims is that -terminal perceiving is a relationship that meets these threeconditions. More precisely, it claims that, whenever a subject -terminallyperceives a physical item, this item is presented to him in the sense that theseconditions dene, and the phenomenal content of the perceptual experiencedraws its qualitative ingredients from the physical features that are thusdisplayed. What particularly needs to be emphasized, here, is that the drawingof these qualitative ingredients is an ontological, not just a causal, matter. It isnot just that the presence of an item with a certain sensible character causesthe subject to have a perceptual experience with a matching content. It isthat the ingredients of the content are themselves the very elements of theexternal situationmade experientially present. The featuring of a quality in thephenomenal content is not something ontologically separate from its externalrealization in the perceived itemsomething that merely serves to representthat realizationbut is that realization itself brought immediately before themind, without the mediating role of concepts, symbols, or mental images.

Adopting the presentational view would be the fundamentalists simplestapproach. It is both the simplest way of representing the perceptual rela-tionship involved in -terminal perceiving as something psychologicallyfundamental, and the simplest way of explaining, within that fundamentalistframework, how the phenomenal content of the perceptual experience formsthe experiential manner in which perceptual contact is achieved. There isalso something else that might make this view initially attractive to him.For it captures, in the most straightforward way, the phenomenological feelof perception. Thus, when a subject -terminally perceives an item, hisexperience carries the subjective impression of being presentational in therelevant sense. It gives him the impression that the perceived item and certainaspects of its character (or character and location) are presentationally beforehis mind in the relevant mode of ontological immediacy, with these displayedaspects exactly covering the content of how the item sensibly appears. It isthis that gives sensible appearance the distinctive phenomenological characterI referred to earlier; for other forms of perceptual appearance do not carrythe subjective impression that the property or character that the perceiveditem appears to possess is, in this way, on presentational display. The pointI am now making is that, to take -terminal perceiving to be genuinelypresentational, and to think of the phenomenal content involved as drawingits qualitative ingredients from what is presentationally displayed, would


be the fundamentalists most straightforward way of doing justice to thephenomenological facts.

The presentational view of perception is the simplest version of thefundamentalist view, and is what the phenomenological facts immediatelysuggest. But it is also, at least when offered as a general theory, open to anobvious and decisive objection. For it cannot accommodate cases of non-veridical perception. I am not thinking, here, of cases of hallucination, whereno physical item is perceived: the presentational view does not even purportto offer an account of these cases. I am thinking, rather, of cases in whicha physical item is -terminally perceived, but under a sensible appearancethat misrepresents its true character. That such cases occur can hardly bedeniedat least on the assumption that we perceive physical items at all.The notorious case of the stick partially immersed in water (in reality straight,but appearing bent) is an obvious examplethough if it were thought(surely implausibly) that the presentationalist could handle this case eitherby assimilating it to the case of seeing veridically but in a special perspective,or by claiming that what is -terminally perceived is not the relevant timeslice of the surface portion of the stick, but the light array it transmits tothe subjects eye, then we could switch our attention to such phenomenaas astigmatism and colour-blindness, where the distorting physical factorslie within the subjects own visual system. In whatever form they arise, thepresentational view cannot accommodate cases of non-veridical perceptionbecause in taking the qualitative ingredients of phenomenal content to bedirectly drawn from the relevant physical item, it excludes the possibility ofthe sensible appearance of this item being at variance with its true character.Sensible appearance just is, for the presentationalist, the direct bringingof certain aspects of the items actual character, sometimes in the relevantperspective, before the subjects mind.

The presentational view cannot deal with the phenomenal content thatoccurs in cases of non-veridical perception, and so cannot be accepted in thegeneral form in which I have formulated it. But this still leaves the optionof retaining the presentationalist approach for cases of veridical perception.So, in the case where a straight stick in water looks bent, we are forcedto say that the featuring of bentness in the phenomenal content is notthe featuring of some physical instance of bentness. But in the case wherea straight stick out of water looks (veridically) straight, we could still saythat the featuring of straightness in the phenomenal content consists in theinstance of straightness in the stick being made present to the mind. But

20 A World for Us

while this mixture of approaches is an option, it is hardly a plausible one.For it is very hard to suppose, in the sorts of case that we are considering,that the veridicality or non-veridicality of an experience correlates with such afundamental difference in its nature. We are not now, as in the earlier context,concerned with the distinction between perception and hallucination, wheredistinguishing the psychological nature of the experiences involved is a seriousoption. In the present context the veridical and non-veridical experiences weare envisaging are alike in being physically perceptive; and, crucially, theyare alike in being perceptive of the concrete physical feature whose sensibleappearance is at issue (so that, in the case of the stick in water, the physicalinstance of straightness is still perceived, though non-veridically). Moreover,they causally originate from the perceived physical item by processes of abroadly similar kind. All this puts us under strong pressure to think of theexperiences as amenable to a unitary account. This pressure becomes, to mymind, irresistible when we focus on a case where a shift from veridical tonon-veridical perception involves only very slight changes to the qualitativecharacter of the phenomenal content and to the details of the causal processfrom the relevant physical item. Think, for example, of a situation in whichsomeone rst looks at an object through plain at glass, seeing its shape asit is, and then looks at it through glass whose very slight degree of curvatureimposes a correspondingly slight distortion on the way that the shape of theobject appears. It would surely be absurd to deny that these two perceptiveexperiencesand experiences that are alike in being perceptive of the objectsshapeare, in their intrinsic character, of the same generic type.

Granted that he needs a unitary account of veridical and non-veridicalperception, the fundamentalist is obliged to conclude that, in all cases ofperception, the qualitative ingredients of phenomenal content are internal to(creatures of) the mind, rather than ontologically drawn from the physicalitems perceived. Thus, if someone -terminally perceives a physical itemx, and if, in the context of that perceiving, x sensibly appears to him asbeing F, then the fundamentalist is obliged to say that, even if x is infact F, and even if its being so is causally responsible for this aspect of itsappearance, the featuring of F-ness in the context of this appearance is notthe featuring of the x-instance of F-ness (or of any other physical instanceof F-ness), but is merely something in the subjects mind that in someway represents that instance. I shall speak of this position as the internalist

I elaborate this argument in detail in The Nature of Perception, 6772.


view. I should stress that I am here using this label to refer exclusively toa version of the fundamentalist view. Decompositionalists, too, of course,will regard the ingredients of phenomenal content as internal to the mind;the decompositional view allows for no other possibility. But, in the sense Ihere intend, the internalist view combines the claim of internality with theacceptance that where there is -terminal perceptual contact with a physicalitem, the fact of this contact is something psychologically fundamental. Thishas to be borne in mind when interpreting the claim of internality itself.Whether a straight stick looks straight or bent, the featuring of this lattershape in the content of the appearance is not, on the internalist view, afeaturing of the instance of straightness in the stick. But the internalist willstill accept that the physical instance of straightness is, in the context ofthis appearance, perceived; and, as a fundamentalist, he will insist that itsperceptual presence (whether it is perceived veridically or non-veridically) isan aspect of what is psychologically fundamental.

By adopting this internalist view, the fundamentalist avoids the problemthat defeated the presentationalist: since the ingredients of phenomenalcontent are not ontologically drawn from the perceived item, there is nodifculty in understanding how phenomenal content can be at variance withthe items true character. But he now faces problems of a different kind.

The basic problem, as I see it, is that, on the internalist view, thefundamentalist cannot make sense of the way in which perceptual contact andphenomenal content t togetherthe way in which phenomenal contentembodies the sensible appearance under which the -terminal object isperceived, and forms the experiential manner in which -terminal contactis achieved. There is no difculty, in this respect, for the presentationalist.As he sees it, phenomenal content is precisely what -terminal perceptualcontact automatically supplies by virtue of its presentational characterbythe way in which it directly brings before the subjects mind certain aspectsof the character and situation of the perceived physical item. There is nodifculty in understanding how these displayed aspects form the qualitativeingredients of the items sensible appearance to the subject, and so of theexperiential manner in which he perceives it. Likewise, there is no difculty,on this front, for the decompositionalist. He takes perceptual contact to besomething that breaks down into the subjects being in a more fundamental(not in itself physically perceptive) psychological state, together with certainadditional factors, and, in the case of -terminal contact, he will take therelevant psychological state to be that of having a perceptual experience with

22 A World for Us

a certain kind of phenomenal content. Phenomenal content will then formthe experiential manner in which -terminal contact is achieved, and embodythe sensible appearance of the perceived item, by being that by which, in thecontext of the other factors, contact is constitutively secured. The difcultyis in seeing what other option is availablein thinking of some way ofmaking sense of the intimate link between contact and content withoutinvoking either a presentational or a decompositional account. In effect, theproblem I see with the internalist view is that its two components pull inopposite directions. Once we take the-terminal relationship to be somethingpsychologically fundamental, I can see no way of tting phenomenal contentinto the picture except as something that draws its qualitative ingredients fromaspects of the perceived physical situation. Once we think of the ingredientsof content as internal to the mind, and ontologically distinct from theirphysical counterparts, I can see no way of making provision for perceptualcontact except in a form to which content constitutively contributes. If,like the internalist, we think of phenomenal content as neither drawingits ingredients from aspects of the physical situation nor as constitutivelycontributing to the securing of perceptual contact, we are, as I see it, left withno way of understanding how it could be anything other than an experientialaccompaniment of such contact, rather than, as it has to be, the manner ofits achievement.

One specic way in which this basic problem manifests itself is withrespect to what I shall speak of as the appropriateness requirement. What Imean by this is the fact that, in order for a perceptual experience to be a-terminal perception of some physical item, its phenomenal content hasto be, to an adequate degree, qualitatively appropriate to that item, relativeto the conditions of observation. The best way to see this is to focus on acase where all the other conditions associated with physical-item perceptionare present, but the factor of appropriateness is conspicuously absent. Thus,suppose I am in my sitting room, with my eyes turned towards the clockon the mantelpiece, with nothing obstructing my line of vision, and with allthe other external factors favouring the achievement of visual contact. Andsuppose that light reected from the clock and its surroundings enters myeyes in the normal way and sets up the appropriate kind of process in myoptic nerves, which in turn transmit the appropriate signals to my brain.But then something peculiar happens. My brain responds to the incomingsignals in a totally bizarre way, producing a visual experience which is notremotely like the sort of experience that is normal for that kind of photic


input. For instance, it might be that the resulting experience is like thatof seeing a football match, or like that of seeing a clear blue sky. Nowit is surely clear that, given the extent of the disparity between the realcharacter of the external environment and the phenomenal content of myexperience, this experience is not physically perceptive. It is true that theclock and its surroundings play a causal role in producing the experience,and, with respect to the photic input, this role is of the normal kind forthe circumstances in question. And we can even suppose that, as in the caseof normal visual perception, the brain response preserves a kind of causalisomorphism between elements of the resulting experience and elementsof the input, so that, relative to a suitably ne-grained division, differentelements in the content of the experience causally trace back to differentelements of the relevant portion of the environment. But it would be absurdto suppose that the experience qualies as an actual seeing of this portion,and that the only way in which its deviant content affects the situationis in making this seeing radically non-veridical. It is just obvious that, inthe context of the conditions envisaged, the extent of the non-veridicalityprecludes visual contact altogether. So, here we have a clear illustration of thepoint at issue, that there can only be -terminal contact in cases where thephenomenal content has an adequate degree of qualitative appropriatenessto the relevant physical item, relative to the conditions of observation; andit is easy to think of a host of other examples that would illustrate the pointin an analogous way. In this kind of example, the failure of the experienceto achieve the requisite degree of appropriateness is revealed by its radicalnon-veridicality. But we should not think of appropriateness as entirely amatter of veridicality. It would, of course, be entirely a matter of veridicalityfrom the standpoint of the presentational view, which excludes non-veridicalperception altogether. But now that we have rejected that viewand rejectedit precisely because it cannot accommodate non-veridical perceptionwemust also accept that appropriateness is partly a matter of conformity to whatis normal, or normative, for the conditions of observation in question. Forexample, if a straight stick is partially immersed at an angle in water, theappropriate way for it to sensibly appear to a visual percipient is as bent,and it would be someone who could only see it as straight whose visionwould be defective.

It is undeniable that the appropriateness requirement holds. But it createsdifculties for the internalist view in two ways, and ones which a defender ofthe view cannot, I think, adequately deal with.

24 A World for Us

In the rst place, the internalist does not, it seems to me, have anyadequate way of accounting for it. Considering the issue of explanation inthe abstract, we can identify two clear-cut ways in which the requirementmight be explained. On the one hand, there is the explanation that wouldbe offered by the decompositionalist. This would be to say that a sufcientdegree of appropriateness is an essential constitutive factor in the securing of-terminal contact. Thus, the decompositionalist holds that, wherever thereis -terminal contact, it is constituted by the combination of the subjectshaving a perceptual experience of a certain phenomenal kind and certainadditional facts, and so he will account for the appropriateness requirementby insisting that, for such a constitution to succeed, the presence of anadequate degree of appropriateness is one of the things that the additionalfacts have to cover. On the other hand, there is the explanation thatwould be offered by the presentationalist. On his account of perception,the qualitative ingredients of phenomenal content are directly drawn fromaspects of the physical environment. There is no room, here, for any degreeof inappropriateness (which, for the presentationalist, would be the same asnon-veridicality), since it is only insofar as there is an accurate display ofthe environmental situation that there is phenomenal content at all. Boththese accounts of perception would, in their contrasting ways, provide acomplete rationale for the appropriateness requirement. But neither of themis available to the internalist, who combines a fundamentalist view of contactwith an internalist understanding of the ingredients of content. Nor, it seemsto me, can he derive a rationale from any other source. Once it has beenaccepted that the qualitative ingredients of phenomenal content are internalto the mind, and not directly drawn from aspects of the environment, theonly way I can see of making sense of there being a limit on the amountof inappropriateness that -terminal perception can tolerate is by supposingthat a sufcient degree of appropriateness is an essential element in the factorsby which -terminal contact is constituted.

The second point involves something more subtle. Although there is alimit on the degree of inappropriateness that -terminal perception cantolerate, our concept of it is, surely, not so precise as to ensure that, inany particular case, there will always be an objective answer to the questionof whether the requirement is satised. As a result, we can envisage caseswhose status, as perceptive or non-perceptive, is inherently borderline. Thus,suppose scientists have constructed a device that can be used to distort thevisual appearance of the physical scene by sending a stream of radiation


through the subjects visual cortex, the amount of the distortion increasingwith the strength of the radiation. And suppose an experimenter is aboutto use this device on someone who is looking at an apple. At one extreme,with very weak radiation, we can envisage the experimenter producing aneffect on phenomenal content so slight that there would be no threat tothe continuation of visual contact: the subject would continue to see theapple (at each moment -terminally seeing a certain momentary stage ofa certain portion of its surface), but its apparent shape would be a littlewarped or its surface colour-pattern look blurred. At the other extreme, withvery strong radiation, we can envisage an effect so great that visual contactwould clearly be severed: how things appear to the subject would bear noresemblance at all to how things are, and the experience could not, by anystretch of the imagination, be construed as perceptive. But between theseextremes we can also, surely, envisage a range of cases that would be inherentlyborderlinecases whose classication as perceptive or non-perceptive wouldbe a matter for stipulation, rather than an issue of objective fact. To reachsuch cases we need only envisage a series, from the rst extreme to thesecond, in which the experimenter very gradually increases the strength ofthe radiation and the resulting degree of the effect on phenomenal content.It is surely clear that, somewhere in the middle, cases would occur where thequestion of whether the extent of the inappropriateness was enough to severvisual contact with the apple would have no denite answer, even from aGods-eye view, and where the status of the relevant experience would be leftindeterminate.

The possibility of these borderline cases is easy enough to explain in theframework of the decompositional view, which takes a subjects -terminalperceiving of a physical item to be partly constituted by the fact that thephenomenal content of his experience stands in the right sort of qualitativerelationship to it. For if a sufcient degree of appropriateness is a constitutiverequirement for the obtaining of -terminal contact, borderline cases couldarise in this area in the way that they are liable to arise in any area wherethe application of a concept constitutively depends on whether the situationachieves a sufcient value along a certain qualitative dimension, but wherethere is no particular point on this dimension that marks a theoreticallycrucial division. So, it would be as easy to understand why there mightbe no objective answer to the question of whether a certain experiencehas enough appropriateness to count as perceptive as it is to understandwhy there is sometimes no objective answer to the question of whether

26 A World for Us

some group of people is sufciently numerous to count as a crowd, or tothe question of whether someone has enough hair on his head to avoidcounting as bald. But the situation for the internalist is quite different. Evenif he could nd some rationale for the appropriateness requirement itself,his commitment to the fundamentalist view would prevent him from evenacknowledging the possibility of borderline cases of the sort envisaged. Afterall, perceptual contact itself, unlike qualitative appropriateness, does notadmit of degrees: it is all or nothing. So, if such contact, in its -terminalform, is taken to be psychologically fundamentalsomething which doesnot, at the psychological level, break down into further factorsthere isnothing at the psychological level of description which could explain howthe question of its obtaining could ever fail to have an objective answer.It is only if the obtaining of contact is constitutively controlled by morefundamental factors, and if the conditions for its obtaining are not fullyprecise in terms of those factors, that there is an opportunity for borderlinecases to occur.

It seems to me, then, that the internalist view does not allow the devel-opment of an adequate account of the relationship between perceptualcontact and phenomenal content. It does not, as I see it, allow any adequateunderstanding of how phenomenal content forms the experiential mannerin which -terminal contact is achieved. Nor does it allow any explana-tion of why, in the case of such contact, there is a limit on the degree towhich phenomenal content can be qualitatively inappropriate to the itemperceived, relative to the conditions of observation. Nor can it accommod-ate the fact that, because the requirements for sufcient appropriatenessare not sharply dened, we have to allow for cases where there is noobjective answer to the question of whether there is perceptual contact ornot. In all these ways, as I see it, the internalist view shows itself to beunsatisfactory.

The internalist view is one version of the fundamentalist view of perception,which holds the -terminal perceptual relationship to be psychologicallyfundamental; it is that version which takes the qualitative ingredients ofphenomenal content to be internal to the mind. The alternative versionis that of the presentational view, which takes -terminal perceiving to bepresentational, and takes the ingredients of content to be ontologically drawnfrom the external physical situation. Since the presentational view has alsoshown itself to be unsatisfactory, and unsatisfactory for cases of veridical andnon-veridical perception alike, and since there are no other fundamentalist


options available, I conclude that the fundamentalist view is mistaken, andmistaken in its application to all cases of perception.

V

The alternative to the fundamentalist view of perception is the decompos-itional view. This holds that perceptual contact with a physical item, evenwhen it is -terminal, breaks down into two components, one of whichconsists in the subjects being in some further psychological state, which isnot in itself physically perceptive, and the other of which comprises certainadditional facts, but ones that do not add anything further to the subjectspsychological condition at the relevant time. Having found fault with the fun-damentalist view, we now need to consider whether this alternative approachfares better. Since there is no other approach to perception now available, itwill be awkward if it does not.

One question that arises with respect to this view is over the nature ofthe relevant psychological states. We are told that they are not in themselvesphysically perceptive; so they are capable of occurrence in both the contextof physical-item perception and the context of hallucination. We also knowthat the states are experientialstates of visual, auditory, tactual, or someother kind of sense-perceptual experienceand that the central componentof their experiential contentthe component we have called phenomenalcontentcovers the way in which things sensibly appear to the subject. Butwe still do not know exactly what psychological character these states aresupposed to have; we do not know how best to understand their characterfrom a decompositional standpoint. This is an issue that I have discussed atconsiderable length in The Nature of Perception (the whole of the third partis devoted to it), and I still accept the conclusions I reached there and thearguments I invoked to establish them. But since the issue is only peripheralto my present concerns, and its outcome does not affect the argument I wantto develop, I shall only briey touch on it here.

Deciding on the psychological nature of the relevant experiential stateshas been a source of a great deal of controversy within the decompositionalcamp. Until about the middle of the last century the standard approach, atleast within the empiricist tradition, was to say that, although these statesare not in themselves physically perceptive, they do involve, as their centralcomponent, the presentational awareness of a non-physical iteman item

28 A World for Us

which is characterized by sensible qualities associated with the sense realm inquestion, but whose existence is internal to the mind. So, in the case of Ralph,the claim would be that, in seeing the apple, the central component of hisvisual experience, at any time, consists in his presentational awareness of acertain spatial array of colours, though an array which, unlike the relevantphysical item, has no existence outside the realm of experience. These putativemental objects of awareness, often labelled by earlier philosophers as ideas orimpressions, came to be known in the twentieth century as sense data, and theposition which postulates them is now usually referred to as the sense-datumtheory. In representing this theory as claiming that the awareness of a sensedatum forms only the central component of the relevant experiential states,I am going beyond the explicit requirements of the tradition. But it seemsto me that the only form of sense-datum theory worth considering is onewhich takes each total perceptual experience to combine the presentationalawareness of a sense datum with an element of interpretation. The reason whythe theory needs to recognize this element of experiential interpretation isthat, without it, it could not do justice to the phenomenological character ofperceptual experience. For such experience not only makes it introspectivelyseem to the subject that there is a sensible item before his mind, but also makesit seem to him that this item is an ingredient of his physical environment;the apparent environmental character of what is before the mind is, indeed,an essential aspect of phenomenal content. The only way in which the sense-datum theorist can account for this phenomenological situation is by takingperceptual experience to include an element of interpretation that is directedon to the presented sense datum and which represents it as environmental inthe appropriate way. So, in the case of the visual realm, he must think of theelement of experiential interpretation as representing the two-dimensionalcolour array which forms the presented sense datum as a three-dimensionalarrangement of colours located at various distances and in various directionsfrom the subject.

In recent years the sense-datum theory has become increasingly unpopular,partly because its ontology of mental objects is seen as inherently problematic,and partly because the acceptance of this ontology is incompatible with thephysicalistic approach to human mentality that modern philosophers tend tofavour. As a result, most current decompositionalists try to understand thenature of perceptual experience without recourse to sense data. One approach,here, is to represent the relevant experiential states as purely conceptual incharacter. For example, it might be claimed that perceptual experience


is simply the conscious acquisition, in a suitably vivid form, of putativeinformation about the external environment. Another approach is to acceptthat the central component of a perceptual experience is something akin to asense datum, in that it involves the experiential realization of certain sensiblequalities, but insist that these qualities are realized not as features of an objectof awareness, but as aspects of a manner of awareness. The main problemfor any alternative approach is in accommodating the phenomenology ofperception. I have already stressed that the sense-datum theorist needs torecognize an element of experiential interpretation in order to explain whythe presented sense datum poses as an environmental object of awareness.These alternative approaches face the more basic challenge of explaining whyperceptual experience carries the subjective impression of being an awarenessof something at all. If such experience is not inherently perceptive of anythingphysicalwhich it cannot be on the decompositionalist viewand if it doesnot involve an ontology of mental objects of awareness either, why should itintrospectively seem to the subjectas it undeniably doesthat there is asensible item before his mind?

Inmy view, this challenge cannot be satisfactorilymet, and the sense-datumapproach is the right one for the decompositionalist to adopt. Admittedly, Ithink that to make this approach intelligible, we need to revise its traditionalform in one crucial respect. Traditionally sense data are taken to be privateand momentary particulars: each sense datum occurs as the object of a singleepisode of awareness, in the mind of a single subject at a single time, andits existence is logically restricted to this unique mindtime location. But itseems to me that, to give sense data the right kind of ontological status toserve as genuine objects of awareness, we need to reconstrue them as sensoryuniversals, capable of presentational occurrence to different subjects and ondifferent occasions. The reasons for this, and the more general reasons forthinking that we need some kind of sense-datum theory at all, are mattersthat I have dealt with in detail in the earlier book, and I shall not pursuethem any further here. They are in any case, as I have indicated, peripheral

This is the approach pioneered byDavid Armstrong andGeorge Pitcher. Thus, see ArmstrongsA Materialist Theory of the Mind (London: Routledge & Kegan Paul, 1968), ch. 10, and PitchersA Theory of Perception (Princeton, NJ: Princeton University Press, 1971).

This is the adverbialist approach. Its early advocates include C. Ducasse, Moores refutation ofidealism, in P. Schilpp (ed.), The Philosophy of G. E. Moore (Chicago, Ill.: Northwestern UniversityPress, 1942), 22351, and R. Chisholm, Perceiving (Ithaca, NY: Cornell University Press, 1957),11525.

The Nature of Perception, 93195.

30 A World for Us

to my present concerns. The issue on which I want to focus, and to which Inow turn, is to do with a different aspect of the decompositional view, andthe conclusions for which I shall argue do not depend on what account ofperceptual experience the decompositionalist adopts.

According to the decompositionalist, perceptual contact with a physicalitem, even when -terminal, is secured by the combination of the subjectsbeing in a certain experiential psychological statea state which is not initself physically perceptiveand certain additional facts, and it is stipulatedthat these additional facts do not involve anything further about the subjectspsychological condition

a world for us

Documents

university of oxford

oxford oxford university

existence of thephysical

worldsuggestive way

rights department

moral rights

phenomenalistic idealismstands

blank aworld