Wittgenstein: Meaning and Judgement

Wittgenstein for Dee Wittgenstein Meaning and Judgement Michael Luntley ß 2003 by Michael Luntley 350 Main Street...

Author: Michael Luntley

31 downloads 929 Views 918KB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form

DOWNLOAD PDF

Wittgenstein

for Dee

Wittgenstein Meaning and Judgement

Michael Luntley

ß 2003 by Michael Luntley 350 Main Street, Malden, MA 02148-5020, USA 108 Cowley Road, Oxford OX4 1JF, UK 550 Swanston Street, Carlton, Victoria 3053, Australia The right of Michael Luntley to be identified as the Author of this Work has been asserted in accordance with the UK Copyright, Designs, and Patents Act 1988. 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, electronic, mechanical, photocopying, recording or otherwise, except as permitted by the UK Copyright, Designs, and Patents Act 1988, without the prior permission of the publisher. First published 2003 by Blackwell Publishing Ltd Library of Congress Cataloging-in-Publication Data Luntley, Michael, 1953– Wittgenstein : meaning and judgement / Michael Luntley. p. cm. Includes bibliographical references and index. ISBN 1-4051-0241-1 (alk. paper) — ISBN 1-4051-0242-X (pbk. : alk. paper) 1. Wittgenstein, Ludwig, 1889–1951. 2. Intentionality (Philosophy) 3. Meaning (Philosophy) 4. Judgment. I. Title. B3376.W564.L895 2003

192—dc21

2003005344

A catalogue record for this title is available from the British Library. Set in 10/12.5pt Minion by Kolam Information Services Pvt. Ltd, Pondicherry, India Printed and bound in the United Kingdom by MPG Books Ltd, Bodmin, Cornwall For further information on Blackwell Publishing, visit our website: http://www.blackwellpublishing.com

Contents

Preface Abbreviations

1 Wittgenstein’s Master Argument 1.1 1.2 1.3 1.4 1.5 1.6

Introduction Animating Signs The Platonist Source of Grammar The Cartesian Source of Grammar The Community Source of Grammar The Negative and Positive Phases – First Statement

2 Realism, Language and Self 2.1 2.2 2.3 2.4 2.5

Introduction Logic Takes Care of Itself The Need for Grammar The Metaphysical Options The Self

3 This is How We Play the Game 3.1 3.2 3.3 3.4 3.5

Introduction Two Kinds of ‘Hidden’ Meaning and Use Use and Self Use and Augustine’s Mistake

vii ix

1 1 2 9 11 15 16

21 21 22 27 34 42

48 48 50 58 64 67

vi

Contents 3.6 3.7 3.8

Is ‘Slab!’ a Shortening of ‘Bring Me a Slab!’ or is the Latter a Lengthening of the Former? This and Similar Things are called ‘Games’ Spontaneity in Particular Circumstances

4 Rules and Other People 4.1 4.2 4.3 4.4 4.5 4.6

Introduction The Structure of an Argument The Bipartite Account of Meaning Practice What You See/Hear is not Normless Seeing the Similarity in Particular Cases

5 Putting Your Self in the Picture 5.1 5.2 5.3 5.4 5.5 5.6 5.7 5.8

Introduction The Standard Treatments What’s Special about Sensations? The Need for Calibration Calibration in Subjectivity Agreement in Forms of Life ‘Inner’ Life Out There Now I Understand

6 Seeing Things Aright 6.1 6.2 6.3 6.4 6.5 6.6

Introduction Two Uses of the Word ‘See’ See What I Mean Paying Attention The Place of Judgement What Comes Natural

Bibliography Index

71 77 83

93 93 94 100 105 115 119

124 124 124 128 133 137 142 146 148

152 152 155 165 168 170 173 177 182

Preface

The will is an attitude of the subject to the world. (NB p. 87) Grammar is perspectival. This is the key to Wittgenstein’s theory of intentionality and the central claim to this book. The main work consists in saying what it means, not least in defending the idea that Wittgenstein had a theory of intentionality. It often seems that Wittgenstein has an anti-philosophy. I think that is wrong. Indeed, I think he has profound insights that illuminate and can be illuminated by contemporary work on intentionality. The Wittgenstein presented herein is no quietist. The point of the claim that grammar is perspectival is that the conditions for the possibility of intentionality consist not in a body of theoretical knowledge, but in perceptual knowledge. The conditions for the possibility of intentionality are that we see things aright. For this claim to make sense, we require not a quietist description of ordinary language use, we need a clarification of the metaphysics of grammar. This turns out to be a metaphysics in which we, as subjects, play a fundamental role. The grammar of language is not a structure that we inhabit, it is a structure that, in part, we sustain and amend in our ongoing activities of judgement. That is why this book is about the conditions for the possibility of judgement. The subject, construed as a self-as-will, has an ineliminable role in the account of grammar. The reading I offer of Wittgenstein is then a reading with substantive metaphysical import. Two of the most central points concern the conception of the subject’s place in the world. I argue in chapter 5 that the private language argument is an argument about retrieving our mindedness as part of the world. It is an argument that both acknowledges that there is such a thing as how things are for me and places how things are for me as part of

viii

Preface

how things are. The perspective of the subject is real and part of the world. Second, the account of how we are in the world has us primarily engaged in the world, not engaged with others. In chapter 4 I argue that the concept of practice that Wittgenstein employs is a non-social concept. Wittgenstein has had an enormous influence across the humanities and social sciences. Much of this has been ill-conceived, for it has been informed, for the main part, by the idea that Wittgenstein’s onslaught on the Cartesian conception of the mind as private is directed towards the advocacy of a social conception of practice, mind and meaning. That is wrong. On the account of practice that I offer, Wittgenstein still provides insights that matter across academia and beyond, but they will be different. I note some of the opportunities for such impacts, but most of that work will have to wait for another occasion. The argument that I provide about the conditions for the possibility of judgement is an extraction from Wittgenstein’s texts. The extraction has been long, but fun and has been aided by the many colleagues and students at Warwick who have contributed to the game of reading Wittgenstein. Being with others is not constitutive of practice, but it surely helps. It scaffolds one’s understanding when, as at Warwick, one reads Wittgenstein in the company of a commonality of understanding and seriousness in approaching Wittgenstein and a general expertise in the philosophy of thought and language. I cannot now identify all the points of contact, but I know that Naomi Eilan, Chistoph Hoerl, Peter Poellner, Johannes Roessler and Tim Thornton have enabled my engagement with Wittgenstein in countless ways. Especial thanks also to my graduate class on Wittgenstein in autumn 2002 as I approached final drafting of some of this material. During April 2002 I was a visiting scholar at the Wittgenstein Archives in Bergen, Norway. While there I delivered four classes at the Institute of Philosophy that drew upon material that now appears in chapters 4–6. I owe a very special debt of gratitude to to Alois Pichler, Harald Johannessen, Ole Martin Skilleas, Simo Saatela, Knut Venneslan and Gerhard Gelbmann for their warm hospitality and hard questions during my stay in Bergen. Material from my paper, ‘Patterns, Particularism and Seeing the Similarity’, Philosophical Papers, 2002, pp. 251–71 reappears in parts of chapter 3 and I am grateful to the editor, Andrew Gleeson, for permission to re-use that material. There is so much that has to be seen and felt about practice in order to understand Wittgenstein aright. It is not a simple matter of academic scholarship getting Wittgenstein into view, let alone the world. The shape of practice and of our attitude to things is revealed in what we do. I cannot begin to estimate what I have learnt about these things from Dee, my teacher in practice.

Abbreviations

References to the Philosophical Investigations are given by paragraph alone, except those to part two which are given by page reference and letter to indicate paragraph. I have used the familiar pagination prior to the compressed version on the new revised 3rd edition 2001. For other Wittgensteinian texts I have used the following abbreviations: MS References to unpublished manuscripts follow the von Wright catalogue: G. H. von Wright, Wittgenstein, Oxford: Blackwell, 1982, pp. 32ff. NB þ page number, Notebooks 1914–1916, ed. G. H. von Wright and G. E. M. Anscombe, trans. by G. E. M. Anscombe, Oxford: Blackwell, 1961. TLP þ proposition number, Tractatus Logico Philosophicus, trans. by D. F. Pears and B. McGuinness, London: Routledge and Kegan Paul, 1961. RFM þ section number, Remarks on the Foundations of Mathematics, ed. G. H. von Wright, R. Rhees and G. E. M. Anscombe, trans. by G. E. M. Anscombe, Oxford: Blackwell, 3rd edn, 1978. RPP II þ section number, Remarks on the Philosophy of Psychology, vol. II, ed. G. H. von Wright and H. Nyman, trans. G. G. Luck, Oxford: Blackwell, 1980. CV þ page number, Culture and Value: A Selection from the Posthumous Remains, rev. 2nd edn with English translation, Oxford : Blackwell, 1998. Z þ section number, Zettel, ed. G. E. M. Anscombe and G. H. von Wright, trans. by G. E. M. Anscombe, Oxford: Blackwell, 1967.

CHAPTER ONE

Wittgenstein’s Master Argument

1.1 Introduction The existence of intentionality is a remarkable fact. It involves the idea of a subject having a point of view. This is not, prima facie, just another fact about how things are in the world. We might try to catalogue all the contingencies of the world, the different species of plant and animal, the physical forces that move mountains and the chemistry of our impact on the environment. The fact that we have a point of view and represent how things are in our thought and talk does not, however, appear to be just another item in the catalogue. It has the appearance of being a radically different sort of fact, a fact of which we have little idea how it could be so. How things are in the world – how volcanoes erupt, how industry pollutes, how birds migrate – with these sorts of things we have some idea of how and why they are so. When it comes to how it is that our thought and talk represents reality, we know that this is the case, but that it is the case is, at first sight, something of a mystery. Aboutness does not look natural. It might be ubiquitous, for us at least, but the aboutness of our thought and talk is not, unlike the colour of our hair, a straightforward natural fact. Wittgenstein’s central preoccupation was with the question of the conditions for the possibility of intentionality. I prefer to put the matter in terms of the conditions for the possibility of judgement. The reason for this concerns the role that the self as agent plays in Wittgenstein’s argument of the conditions for the possibility of intentionality. I shall claim that Wittgenstein’s central insight is that intentionality can only be made sense of from within a conception of a self with an attitude to the world, an attitude which is that of a will – an agent. The subject with intentionality is fundamentally an agent.

2


The notion of agency is required not because the subject of thought acts in and on the world, but because the subject of thought actively configures the attitude of their being a thinker. This involves a wilful organization of systems of representation. Too often systems of representation are thought of as structures (formal, social etc.) which we thinkers inhabit. The view that Wittgenstein presses throughout his work is the opposite of this. The role of the subject is fundamentally that of a judge, putting representations together to make best sense of our ongoing confrontation with things. This confrontation is our basic attitude to things, but it is an attitude that is managed and wilful. It is wilful not just in the sense that things impede our behaviour – that is true of animals and does not capture the real point of the concept of will. The world impedes our behaviour and this is something to which we take an attitude. Being an agent is having the capacity to alter perceptual inputs at will; it is the capacity to organize our engagements with things in order to get what we want. That is the rational attitude, it’s an attitude of will.1 In this opening chapter I want to lay out the foundations for this claim against the backdrop of the problematic for what I shall call ‘Wittgenstein’s master argument’. The argument has two phases, a negative phase and a positive phase. The former is familiar in the literature on Wittgenstein’s work, especially on his later writings. What I call the positive phase is not usually acknowledged. It is the positive phase that introduces the role of the self-as-will. I shall introduce the master argument by putting it in the context of three standard options on the question how intentionality is possible. These options are attempts at what I shall call ‘animatory theories of meaning’. The negative phase of Wittgenstein’s master argument is the rejection of animatory theories of meaning, but it is important to start with some appreciation of why such accounts of meaning can look tempting. At this point I keep references to Wittgenstein and the secondary literature to a minimum, for I want to concentrate on an outline analytical framework for thinking about intentionality. The framework provides a general orientation to much of the literature on Wittgenstein on meaning. It also provides a first account of where my reading of Wittgenstein is located with respect to the usual interpretations.

1.2 Animating Signs The idea of an animatory theory of meaning is a leitmotif throughout modern philosophy. The idea is motivated by the following thought: signs, considered qua inscriptions or as the sounds that articulate them, do not bear meaning.


3

Signs, qua signs, are inert. In order to carry meaning they need to be animated.2 The task of an animatory theory of meaning is to give an account of what brings signs to life. The need for an animatory theory of meaning reflects what I call bipartism about meaning. This is the view that an account of meaning has two components – an account of representations (inert signs) plus, e.g. the rules of use that animate them. Consider the following arrangement: (1)

{¥Sˇ¿¿

1

TM

1zSˇ

If I ask if you see what I mean with (1) you will, of course, be lost for an answer. The signs mean nothing to you. They are raw ink marks. In order for these signs to come alive, in the way that the signs that make up the sentence you are now reading unproblematically possess meaning, it seems plausible to insist that there are two things about these signs that you need to know. You need to know what the signs stand for and you need to know how they should be used. When you know these you know the semantic power of the sign. There are two ingredients to semantic power: the signs’ representational power and their inferential power. It is a question of considerable importance which, if either, of these powers is more primitive. Should we understand the representational power of a sign in terms of its inferential power, the inferential power in terms of the representational power, or are the two ideas mutually dependent? I ignore these questions for the moment. I shall return to them shortly. For now, let us assume that you need to know both the representational and inferential power of the sign.3 It is not enough to know what each sign individually stands for, you need also to know how the signs can be used, how they fit together to frame judgements that can be true or false. Knowing how the signs can be used is a matter of knowing their grammar – the systematic patterns of use that reveal the inferential connections between judgements formed with signs. Forming a judgement with an arrangement of signs is, in so far as it forms something that is either true or false, forming something that stands in inferential relations to other judgements. It does not seem to make sense to suppose that an arrangement of signs could express a judgement that was true or false and yet there was no conception available of how the truth or falsity of the judgement bore upon other judgements. For example, if you judge that a is F it must follow that something is F.

4


More substantially, if I judge that object a is red, then it is at least partly constitutive of the idea that the judgement is a candidate for truth and falsity that its possession of a truth-value bears upon the truth-value of other judgements that can be framed with the name ‘a’ and with the predicate ‘. . . is red’. For example, if it is true that a is red and it is true that b is red then, other things being equal, it must be true that a and b look similar with respect to the way that they fill out visual space. If they do not look similar and we know that b is red, that is prima facie reason to conclude that the judgement that a is red is false. Judgements do not stand alone in their possession of truth-values. Whether or not a judgement is true or false bears systematically on the truth-value of other judgements. To know the truth-conditions of a judgement is to know, inter alia, how its having a particular truth-value affects the distribution of truth-values over the range of potential judgements formed with common components.4 The way that judgements bear upon one another exploits the way they are composed of common components and the patterns of use governing the components. It is these patterns of use that I am calling grammar. Where there is judgement there must be grammar. How this notion of grammar is marked in ordinary language depends on the contingencies of the sign systems we have developed. What is not contingent is that there must be grammar. It is important not to confuse the fundamental idea of grammar with ordinary language grammar. Suppose I said that ‘{¥Sˇ¿¿’ stands for Othello that, ‘1zsˇ’ stands for Desdemona and that, ‘TM1’ stands for the two-place relation ___loves ..... This still does not fix the judgement expressed at (1) until we know, for example, which direction we are to read the signs. This is a superficial point


5

about grammar, it concerns the way that ordinary language grammar marks the connections between one judgement and another. It does not much matter whether our sign systems are read from left to right, or from right to left, or indeed whether they are read in a single linear sequence at all. There is, in principle, no reason why signs could not be read in different directions on alternate days. It would not affect the ability of a sign system to carry meaning if we did this. There is, however, a more fundamental notion of grammar. The claim that ‘TM1’ stands for the two-place relation ___loves ..... already invokes a notion of grammar, for the thing that ‘TM1’ stands for is not an ordinary object. The thing picked out is something with a structure to it, a structure defined by the ordering of the two places that the relation binds. The idea that we can understand the sign string at (1) by being told the things that the various signs stand for already invokes a notion of grammar. It might seem that giving an account of the things the signs stand for would, in showing what (1) meant, reveal that the representational power of signs was primitive. Perhaps then, so the thought might go, the sign string at (1) is a simple string of names. The idea that the string as a whole is a string of names only looks plausible, however, because one of the signs stands for something structured. The supposition that all signs are names and that their semantic power can be accounted for solely in terms of their representational power only looks plausible by admitting an ontology of things that include items with organized ‘gaps’ in them, gaps that can be filled by ordered sequences of objects. It is, then, misleading to suppose that all signs can be treated as names as if their semantic power were exhausted by their representational power to stand for objects, for that requires a heterogeneous category of objects. In particular it includes the need to admit a category of objects with intrinsic structure to them. This structure binds ordinary objects together.5 This suggests that whether or not we take inferential power as the primitive idea, an account of intentionality that dealt only with the representational power of signs would be inadequate. The meaning of ‘____loves .....’ is not really explained by citing its representational power to stand for the relation ____loves ......., for that claim presupposes a grasp of grammatical rules concerning the ordering of the places and the kind of entities that can fill those places. It is only by admitting a heterogeneous conception of objects into our ontology that we can make sense of the idea that representational power is primitive. But the heterogeneity of the conception of objects renders vacuous the defining claim of the idea that representational power is primitive, namely that judgements are formed by strings of names which do no

6


more than stand for their respective objects. That claim is only plausible if the range of things admitted as objects includes items, like relations, that have grammar built into them. But once that is admitted, it is plain that the concept of an object has been detached from its ordinary significance. Indeed, it becomes unclear that there is a real content left to the claim that all signs are names whose semantic power is exhausted by their representational power to stand for objects. There is a deep notion of grammar that we exploit when we say that ‘TM1’ stands for the two-place relation ‘___loves .....’. It is the notion of grammar implicit in our understanding that, on being told this, we know what sorts of things can fill the gaps in the two-place relation and that it matters in which order they fill the gaps. I shall return to consider the relationship between representational power and inferential power later. The above considerations suggest that a theory that dealt only with the former power would be inadequate, but this gives us no reason to assume that inferential power should therefore be taken as primitive. It means, so far, only that we need to find a place both for the notion of the sign’s representational power and its inferential power. Now, the inferential power of a sign is the grammar that governs its patterns of use. The predicate ‘. . . is red’ has a pattern of use that governs its use in inferences that connect judgements involving it with judgements employing other predicates. It is in virtue of this grammar that the connections between a is red and, a is coloured and, a is not green obtain. ‘Grammar’ here means the standard ways of using the expression that govern its contribution to determining the truth-value of judgements in which it figures. These patterns constitute our notion of correct use. There is a normativity built into the notion of grammar. If a subject judges that a is red then, ceteris paribus, they ought not also to judge that it is green, although they ought to judge that it is coloured. There is not an option about whether or not someone follows these normative patterns. One cannot credibly wake up one morning and decide, ‘Today I am going to use ‘‘. . . is red’’ in a totally different way, a way that no one has ever used before and that is unrelated to


7

any previous uses of this expression or any other.’ That sort of thing is not possible. It is, of course, possible to decide to use ‘. . . is red’ as everyone else uses ‘. . . is green’. But the deceptive use of an expression trades on patterns of use that are not questioned. The sense in which it seems true to say that there is no option about whether or not you follow the normative patterns of grammar, is the sense that there is no option but that you acknowledge that there are constraints on your use of signs. You have to use them according to patterns that exist independently of your will. You are not and cannot be utterly free in the way in which you use signs. You might, of course, confound other language users by substituting one pattern for another, but what does not seem possible is that you use a sign meaningfully but without any pattern at all. To accept this is to accept a minimal realism about grammar. One of the things that is distinctive of my reading of Wittgenstein is my account of his realism about grammar and how he manages to be a realist without slipping into Platonism. Wittgenstein is often treated as an anti-realist about grammar in his later writings. I shall argue that his interest in grammar runs throughout his work and, with respect to his realism, there is no significant change in his position. What changes is his account of the metaphysics that supports the realism. To use a sign meaningfully is to use it in a way that it makes a contribution to the determination of the truth-conditions for the judgements in which it figures. The supposition that you could use a sign meaningfully without there being any grammar to its employment amounts to the supposition that the sign could contribute to the truth-conditions of a judgement, and thereby to the judgement’s possession of a truth-value, without its contribution revealing any systematic connection between what it would be for the judgement to possess a truth-value and what it would be for other judgements to be true or false. But that then is the supposition that judgements can stand alone in their possession of truth-values, contrary to our earlier argument. It is extraordinarily difficult to see what this supposition could amount to. You might think, perhaps, that our inability to make sense of the idea of a judgement standing alone in its possession of truth-value is a function of something peculiar to the way we grasp concepts, something peculiar to the human way of understanding meaning. Perhaps other creatures, or perhaps God, could grasp the meaning of a judgement that stood alone? But this hypothesis is purely speculative, for without a positive account of how it could be the case that a judgement could be true or false without there being any conception of how its being so bears upon the truth/falsity of other judgements, we have no grip on what this hypothesis means. It is an empty gesture that provides no reason to suppose that the idea gestured at makes sense.

8


An animatory theory of meaning is a theory that shows how signs come alive. Signs come alive by having both representational and inferential power. We have already seen that there is a prima facie case for not taking representational power as primitive or, at least, insisting that whatever goes into an account of the representational power of a sign it delivers an account of the sign’s inferential power. If this is right, an animatory theory of meaning must have some account of the source of grammar. It will have something to say about the ground of the patterns that constrain our use of signs. Historically, animatory theories have tried to locate the source of grammar in three different locations: Platonic heaven, the individual speaker’s mind and the speaker’s community. The first and third of these locations seem particularly appropriate as candidates for the source of grammar. This is because the notion of grammar is the notion of patterns of use that constrain our employment of signs, so, instinctively, an animatory theory of meaning will provide something independent of individual speakers that can, as it were, police their sign use. Platonic heaven and the community’s behaviour look, to many people, to be appropriate candidates. In contrast, the second source of grammar – the individual speaker’s mind – looks persuasive if we concentrate more on the representational power of signs. Nevertheless, bearing in mind the previous considerations, we know that any account of the sign’s representational power must be capable of delivering from that basis an account of the sign’s inferential power. And so, although the temptation to find the source of grammar in the individual speaker’s mind might arise from the apparent appropriateness of this source as an account of the representational power of signs, it is a condition of adequacy on such accounts that they deliver an account of grammar. None of these locations of the source of grammar is, however, credible. The first and second are generally held to be particularly inane suggestions about the source of grammar. The third still has supporters and, even, supporters who think that Wittgenstein’s contribution to a philosophical theory of meaning was to show how and why it is the only possible source of grammar. Nevertheless, all of the three sources for grammar face immediate and obvious problems. It will be helpful to get a preliminary account of what is problematic with the three putative sources of grammar which I will refer to as the Platonist, Cartesian and community sources of grammar. I do this not because I think that there is a better alternative for an account of the source of grammar, but because I think that one of the most important lessons that we can take from Wittgenstein is the rejection of the very idea of an animatory theory of meaning. If we take the existence of signs and their patterns as the starting point of our enquiry, it looks compulsory to give an account of that


9

which polices the signs’ patterns. This looks compulsory, for the metaphysics of these patterns is prima facie puzzling because they are normative patterns. Animatory theories are motivated by bipartism about meaning in which meaning is characterized in terms of signs plus that which animates them, e.g. rules of grammar. It is the metaphysics of the second component that is puzzling, for it is that component that provides normativity to signs. But Wittgenstein’s insight is that it is not the existence of these normative patterns that matter, it is what they are patterns of that is important. The puzzlement about the metaphysics of normative patterns is resolved if we describe accurately what the patterns are patterns of. In short, the patterns (and they do exist – Wittgenstein is a realist about patterns) are not patterns of signs but of symbols. Symbols are signs-in-use. We cannot start with signs and then add the ingredient to provide them with standards of use, we start with symbols. Wittgenstein always knew this. This is to reject bipartism about meaning and to take the concept of a symbol – sign-in-use – as primitive. The only issue left, and it is complicated, concerns what goes into an account of symbols – signs-in-use. The only thing left concerns our role in the metaphysics of symbols.

1.3

The Platonist Source of Grammar

The Platonist account of the source of grammar holds that what constrains our use of a sign is the existence of abstract patterns of use to which our use must conform. Our use of signs shadows the Platonic patterns of use. The Platonist account is intrinsically obscure. It is a piece of speculative metaphysics and it is unclear that there is any good reason to believe in it. The Platonist conception posits an abstract entity, an abstract pattern of use, in order to explain what it is that polices our use of signs. The account is problematic both with regard to the metaphysics and the epistemology of the Platonic grammar. The epistemology of Platonic grammars is problematic, for without an account of how we are supposed to access a Platonic grammar, there is no reason to believe that the platonic grammar could provide the function required, namely give an account of the source of grammar in use. Indeed, it seems an empty speculation to say that our use of signs has a grammar whose source lies in an abstract Platonic grammar. The point is not, however, merely that without an epistemology of Platonic grammars we are left in the dark as to the mechanism by which the Platonic grammar polices use, although that is a genuine problem. And the point is not just the metaphysical

10


oddity of Platonic realms in general, although that too is a genuine problem. The point against the Platonist conception goes deeper. The main point is that even if we were provided with an epistemology of Platonic grammars, and even if we had some sense of the metaphysical import of saying that such things exist, it would still be totally unclear that we had been offered an account that added anything substantive to the notion of the source of grammar. Suppose we are genuinely puzzled by the source of grammar; that is, suppose we wanted an account that would explain how inert signs became animated with meaning and so came to have patterns of use that determine the truth-conditions of our judgements. The Platonist account says that inert signs come to have grammar (and remember this is the bit that shows that they are ‘alive’) because they partake of the Platonic grammar. But this seems to amount to the claim that inert signs come to have grammar by partaking of the grammar of Platonic signs. If that is what the Platonist conception amounts to, it hardly explains the source of grammar, for it simply posits entities – the Platonic signs – that intrinsically possess grammar. Alternatively, the Platonist might resist the idea of Platonic signs and admit no more than Platonic grammars as abstract structures that real signs exemplify. But this version still takes us no further, for the quest for the source of grammar is met with the idea that real signs partake of grammar (of the Platonic variety) and that explains the patterns in actual use. Either way, the Platonist purports to have an account of the source of grammar that really amounts to no more than a positing of the existence of grammar! Furthermore, the existence of grammar is posited not in actual ordinary sign use, but in an abstract Platonic realm that then, somehow or other, interacts with ordinary sign use. The central problem with the Platonist account of the source of grammar is then that it faces a dilemma: either the account begs the question or it is threatened with a regress. We are puzzled by the existence of grammar and the solution lies in positing another grammar, so the solution reintroduces the original puzzlement about the source of grammar. A regress is threatened if we then repeat our puzzlement about the existence of grammar as applied to the Platonic grammar. A regress of grammars, one behind the other, is only resisted by the claim that the grammar in Platonic heaven has a special status that renders it impervious to the question, ‘What is the source of grammar?’. Such a move begs the question and, furthermore, if that move worked for Platonic grammar it is unclear why, in the absence of a detailed account of the special metaphysical status of Platonic entities, the move should not also apply to ordinary grammar and thus eradicate the original puzzlement about the source of grammar. So either the whole question about the source of grammar is begged, or a regress looms.


11

1.4 The Cartesian Source of Grammar The idea of a Cartesian source of grammar is something to which I shall return repeatedly throughout this book. It is one of the ideas that Wittgenstein spent a great deal of time exploring and criticizing in his later writings. For our first encounter with the idea, it is best to start with the notion of a sign’s representational power, for it is with respect to this that the Cartesian account seems best motivated. Suppose that we are impressed by this sort of question: How does ‘Bill’ stand for Bill? In general, how do signs come to have representational power? It is easy to think that this is the fundamental question about language and thought. We use signs to stand for things, including things that are not there.6 What is the source of this power of a sign to stand for something? An apparently obvious answer to this question, and one that Wittgenstein examined and exposed throughout his career, is that signs come to stand for things because of our determination to mean that with this sign, where the italicized ‘that’ stands for an act of ostensive definition. It is we who give signs their representational power, for it is we who determine, by our focused attention on objects, to mean just that thing there by using such-and-such name. We have the power to baptize objects with names and other signs. Signs qua signs are inert. We supply the extra ingredient that brings them alive. At one level – the empirical level – the Cartesian model is banal. What I mean by the empirical level is simply the account that asks, for example, how my pet cat got to be called ‘Marmite’ and then takes as an adequate answer the fact that my family decided to call her ‘Marmite’. The cat bears that name and my family and I are the ones responsible for this. The empirical level does not, of course, address the issue of how intentionality is possible, for it trades on the existence of intentionality. Indeed, it makes use of our capacity to make sounds and inscriptions stand for things in explaining how that particular object came to have that particular name. So, the empirical question takes for granted that we have the capacity to make signs stand for things. That is okay, provided we do not mistake the empirical question, ‘How did that cat get to be called ‘‘Marmite’’?’ with the transcendental question, ‘How is it possible that the sign ‘‘Marmite’’ stands for a cat?’ It is the transcendental question that is, purportedly, answered by the Cartesian model, the model that says, in general, the capacity for signs to stand for things has its source in the capacity of the mind to make signs stand for things. That is the idea that is being

12


exploited when the representational power of signs is anchored in our determination to mean this by that. But it is clearly not satisfactory to answer the transcendental question in this way and to take for granted our minds’ capacity to make signs stand for things. The transcendental question appears pressing because it seems a remarkable fact about the world that some things, signs, have representational power. The Cartesian response to this is to say the source of this remarkable fact is things, called minds, that have just the right sort of remarkable powers that make representation possible. But that means that, at the end of the day, the Cartesian response has no real answer to the transcendental question. It becomes a brute fact about the world that certain things within it, namely minds, have the capacity to grant representational power to signs. So far, the Cartesian appears subject to the same objection as the Platonist – the account is question-begging. There is also a threat of a regress problem, for one way of taking the Cartesian account is to say that ordinary names, like ‘Marmite’ get their representational power from the representational power of inner mental signs.7 As with the Platonist, we are left with no clear indication why we should be content to accept that inner mental signs can intrinsically represent and yet not accept this of ordinary signs. The problems with the Cartesian model are, however, deeper and more interesting than these similarities with the Platonist might suggest. The very idea of intrinsically representational inner signs is incoherent. The Cartesian response looks its most plausible as a response to a question about the source of signs’ representational power, but I do not think that we can separate the issues about representational power and inferential power. Setting up a sign’s representational power cannot be isolated from setting up its inferential power. For example, determining to let the sign ‘Marmite’ stand for a cat has certain consequences on what will count as proper use of the sign. If the sign is being used for a particular rather than a property, then it is being used for something potentially reidentifiable, a physical object that occupies space and time. If this is right, there is then a degree of stage-setting that constitutes the minimum grammar to which the sign ‘Marmite’ must be responsive. And that means that the Cartesian position, although naturally prompted in response to a question about how a sign gets representational power, must also have an answer to the question about the source of grammar. So although the Cartesian position can look differently motivated to the Platonist and community positions, it shares with them the need to provide an account of the source of grammar, for representational power without grammar does not amount to meaning. It is, however, with regard to the notion of grammar that the Cartesian account comes under severe pressure. With regard to how signs gain representational power, the Cartesian account is as question-begging as the Platonist. But even if we were forced to accept


13

the Cartesian account of the representational power of signs, when it comes to inferential power – grammar – matters are truly difficult for the Cartesian. Grammar concerns the patterns of use for a sign that constitute its role in determining the truth-conditions of judgements in which it figures. This notion of grammar is irreducibly normative, for the patterns of use are patterns that constitute the idea of correctness/incorrectness conditions for judgements in which the sign figures. This conception of correctness/incorrectness conditions must capture the idea of conditions that normatively configure thought and judgement. If I have judged that a is red I ought not also to judge, without further explanation, that it is green. Now, as I have expressed this point so far, in order to capture the normativity of meaning that is constituted by the grammar of our judgements, we have to think of grammar as a pattern of use that exists independently of will. Grammar impinges on us, we are not radically free in our use of signs. That is to say, when an individual thinker or speaker judges that a is red and uses the predicate in a way that conforms to its grammar, their use conforms to a pattern of use that exists independently of will. It is the idea of ‘independence of will’ that captures the normativity to judgement. It expresses a minimal realism about grammar. The idea that grammar impinges on us and is independent of will is, of course, the deep source of the temptation to embrace Platonism about meaning. It introduces a degree of realism about patterns of use that is unavoidable. The degree of realism required to capture the idea that grammar impinges on word use is right, but it does not require the full-blooded Platonism that reifies the patterns of use. It is a confusion to think that Platonism is required to capture this sense of grammar impinging on us, but it is one to which we seem peculiarly prone. I return to this issue at greater length in the following chapters. If the idea of ‘independence of will’ is central to the concept of normativity, this raises a problem for the Cartesian, for the Cartesian has no clear account of that which is independent of will. That is the problem that lies at the heart of the Cartesian conception of mind, a conception of mind in which everything is transparent to the subject of experience, available to the subject in immediate experience. If experience is transparent then what is available within experience cannot bear the weight required to deliver the notion of ‘independence of will’, the notion that is central to the normativity of meaning. For the Cartesian mind, everything is just as it appears to the mind and the properties of what is available to the mind are exhausted by properties that are detectable immediately through and through. For the Cartesian mind nothing is hidden. What you see is what you get. Furthermore, if Hume was right, such transparent experiences must be inert. The contents of experience for the Cartesian mind are exhausted by what is the case, they can never

14


include a notion of what must be the case, what ought to be the case, what should be the case, etc. But these normative ideas are central to some of the most elementary contents we entertain in thought. Having the idea of a physical object, e.g. a table, is an idea of something that is there and remains there, other things being equal, regardless of whether or not we keep on thinking about it. To think of a table behind your back is to think of something that should look a certain way, should feel a certain way, should impede motion, etc., if you were to turn and approach it. These things are part of the grammar of the idea of a table. We are not radically free in how we think of tables. So, if you are thinking that there is a table behind you, you must be committed to ideas about how your experience should develop were you to turn, etc. And it is important here that it is commitments to how your experience should develop that are required, for, from the point of view of the Cartesian mind, there is no such thing as how things should be, there is only the notion of how things are. This is the Cartesian inertness of experience. It flows from the transparency of experience. The latter is an epistemological thesis about experience. Inertness is a constitutive claim about experience that has to be true for experience to enjoy transparency. If experience is inert, there is nothing available within experience to capture the idea of things independent of will. There is nothing in experience that impinges on, or forces itself on the subject. Everything just is as it is. If there is nothing available within experience to capture the idea of things independent of will, there is nothing available within experience to capture the normativity of meaning. This means that, from the point of view of the Cartesian conception of experience, there is no meaning. Nothing from within Cartesian experience captures the robustness of things independent of will, so nothing captures the barest content that our most elementary ideas about physical objects possess. So, from the point of view of the Cartesian mind, meaning is impossible. For the Cartesian, everything is dark, there is nothing that stands for things outside.8 It is important to see that this problem is not an epistemological problem. It is not the familiar problem that the Cartesian mind has no epistemological access to things beyond it. The problem about meaning for the Cartesian is much more fundamental. For the Cartesian mind meaning is not so much as possible. The Cartesian mind is devoid of content, for there is nothing available within the transparent inertness of such a mind that has about it anything that makes it stand for things outside the mind. All that is available from within the Cartesian mind are entities, signs that are characterizable in total in terms of how they are as transparently viewed by the subject of experience. There is nothing available within Cartesian experience that intrinsically points to, or is connected to, things outside, things independent of will.


15

That is to say, what the Cartesian mind contains are signs, things that need animating. The Cartesian mind cannot be the source of the grammar that animates signs, for all it has, all it is constituted by, are collections of dead signs. Descartes was, I believe, aware of this problem, although he doubtless would not recognize the way I express it. For Descartes, there was one item available within experience that provided the intrinsic connection to things without the mind – namely, the idea of God. But that rescue only works if the ontological argument is valid. There are then two problems with the Cartesian account of meaning. The first problem is the more familiar one and is the one that the Cartesian shares with the Platonist. It is the problem that the account of a sign’s representational power is faced with the dilemma of a regress or of begging the question. The first horn of the dilemma arises when our puzzlement about signs’ representational power recurs with respect to the representational power of inner mental signs. In this case a regress threatens, for nothing has been achieved by trying to ground the representational power of ordinary signs in the representational power of inner mental signs. Alternatively, the account begs the question and merely posits that inner mental signs are just the sorts of things to have representational power and there’s the end to it! Once again, if it were okay to say this about inner mental signs, why not say it at the outset about ordinary signs? The second problem with the Cartesian account of meaning only comes into view when we consider the Cartesian account of the source of grammar. It is the problem that concerns the normativity of grammar. The Cartesian account of mind does not supply the resources to deliver the normativity that is central to our notion of content. I have so far characterized the concept of normativity in terms of the idea of things independent of will and it is this idea that is empty from within the Cartesian conception of the transparency and inertness of experience. It is this second problem that has come to dominate much of the literature on Wittgenstein’s later writings about meaning and the possibility of a private language.9

1.5 The Community Source of Grammar One way of taking the problem of the Cartesian account of the source of grammar is to see it as making the claim that the lone language user needs something against which their use of words can be calibrated. That is correct, but many commentators take it that the obvious source of calibration is the community’s use of words. The lone language user gets their use calibrated by setting it against the use of their community. This is the community source of

16


grammar.10 Any assessment of the community source of grammar pivots on what is meant by ‘use’. If ‘use’ picks out a thin conception of use, mere empirical patterns of sign deployment, then it is utterly unclear how sign deployment gets policed by reference to what amounts to no more than more of the same; the sign deployment of the individual gets policed by being measured against the sign deployment of others. If there is a problem with what constitutes the policing of sign deployment as such, for signs are inert and require animation, then simply throwing more signs into the picture offers no policing. Alternatively, if ‘use’ is not understood thinly in terms of sign deployment, then ‘use’ does not necessarily pick out something that stands in need of animation. The community is then redundant as is any other candidate for the source of grammar. That is rather quick as a reply to the community account of the source of grammar, but it will do for now. It is commonly assumed that Wittgenstein adopted some such communitarian account of grammar. The secondary literature, although providing a lot of communitarian readings, provides a good few short sharp rejections of it too.11 The basic flaw with the communitarian account of grammar is, however, as simple as I have suggested.

1.6 The Negative and Positive Phases – First Statement What I am calling the negative phase of Wittgenstein’s master argument is the argument against animatory theories of meaning. The argument has two major strands – the regress threat and the problem of normativity, the problem of capturing the idea of patterns independent of will. Both problems can be found throughout Wittgenstein’s work and he has a common structure to his answers to these problems. The regress problem is familiar from his later writings on rule-following. How does the phrase ‘Add 2’ manage to have a meaning that determines in advance all the results of that operation? Any attempt to answer such a question is threatened with regress, for any answer that was satisfactory would end up presupposing the very possibility that is being probed, namely the possibility that our language carries a meaning that binds us in advance to certain patterns of future behaviour. More fundamentally the negative phase attacks the very possibility of an animatory theory of meaning. The result of the negative phase is the acceptance of the idea of symbols – signs-in-use – as primitive. Signs do not require animating to be supplied with normativity, because we start with signs-in-use. Wittgenstein always thought this. It is the point behind the opening remark in the Notebooks 1914–1916 ‘Logic must take care of itself ’ (NB p. 2a). But is this


17

response lazy? Does it not take for granted the very thing that looks puzzling, namely the fact that signs are used according to normative patterns of correctness and incorrectness? Our response to this challenge depends on the metaphysics of signs-in-use. To take the concept of a symbol as primitive is to put an end to the explanatory aspirations of animatory theories of meaning. It does not follow, however, that such a move amounts to no more than a lazy description of where we started – using words to talk about things. Whether or not the charge of laziness can be met will turn on what goes into the description of symbols that shows that they are not signs in need of animation. Getting the details of such a description right can be revelatory. It might not meet the pseudo-explanatory aspirations of animatory theories, but it can provide innovations in how we see ourselves and our role in the account of meaning. Another label for the target of the negative phase of Wittgenstein’s master argument is bipartism. A bipartite conception of meaning is one in which meaning is a function of two components: (a) a normless bit – signs plus (b) the component that provides normativity, e.g. rules for the use of signs. The latter is normally conceived of as a coherentist model in which the rules of usage are primarily concerned with how signs are used together to make compound signs. If so, then that is an inferentialist conception of meaning as Brandom conceives of it.12 If the rules for the use of signs were rules that applied to signs piecemeal, one sign at a time, that would probably be best construed as a representationalist conception of meaning. It is in contrast to the representationalist conception that a use-theory of meaning looks problematic and anti-realist, for if the use of signs is conceived in terms of how signs are used together, this says nothing about the aboutness of signs. It leaves it unclear how the representational power of signs ever gets off the ground.13 Wittgenstein’s master argument is that there is no possible resolution to the representationalist versus inferentialist issue. Thinking that there is a choice between representationalist and inferentialist is fuelled by an underlying bipartism which is the real focus of Wittgenstein’s critique. Bipartism will always represent the fact of our intentionality as the wrong sort of fact. The negative phase of the master argument works as a regress. There is no resting place for an account of the conditions of the possibility of sense unless we accept that the world is a world of facts, not things; that is, it is structured in such a way as to fit thought. It is here that the charge of metaphysical laziness looks apposite, for it seems appropriate to ask, is the structure real or ideal? However, if you think you can ask that question you have not understood Wittgenstein’s master argument, for that question is rendered superfluous by the positive phase of the argument.

18


The positive phase of Wittgenstein’s master argument concerns what we put in place of the models critiqued in the negative phase. The models critiqued are all options for an animatory theory of meaning and that is what has to be rejected. The rejection is a rejection of bipartite models of meaning in which meaning is a function of two components, a normless bit (the signs) plus that which provides the normativity (which animates the signs). Such theories misconstrue our intentionality. So what, then, is our intentionality like? What is its character that reveals that the project of animatory theories of meaning is redundant? The answer comes in two parts. First, the impossibility of bipartism about meaning amounts to a transcendental argument in favour of a unitary model of meaning, a model in which our account of sign use is not an account that stands in need of supplementation by something extra to animate signs. We start with signs-in-use; that is, we start with symbols. Second, we need an account of use that shows that it is not a component in a bipartite animatory theory of meaning. What goes into use? The answer here is, briefly, ‘quite a lot!’ The basic positive move is to see use not as a pattern of signs, words, or expressions howsoever characterized. The realism about patterns is not about patterns of signs. It cannot be a pattern of signs on a unitary model of meaning. The patterns are patterns of actions. If you think of grammar as a structure, then what we find at the nodes of this structure are not signs, but actions of agents as they use signs. We, as active users, are integral to the positive phase of the master argument. And what we are doing in using signs is, fundamentally, taking an attitude to the world, the attitude of a judge. It is the existence of this attitude that is the main achievement of the positive phase. It is not something capable of full articulation. It is something that has to be seen. The fundamental condition for the possibility of judgement is we see things aright. The fundamental condition for the possibility of judgement is not capable of theoretical articulation. It consists in seeing the world aright, in taking the right attitude to the world. Grammar is perspectival, for the structure of judgement is a structure of acts of judgement, things that cannot be individuated independently of the judge, the subject as agent. In short, the positive phase shows that the subject never drops out of the picture in an account of intentionality. That is why the fundamental question is not: What are the conditions for the possibility of meaning? It is: What are the conditions for the possibility of judgement? The former question invites a perspective in which we forget to tell the story about our own role in intentionality. The thrust of the positive phase of Wittgenstein’s argument is that we cannot and should not forget to give the account of our role in all this, on pain of getting it all wrong. At bottom we see things aright and for that claim not to be a lazy


19

metaphysics, we need to be clear about who and what we are. We have a substantive role to play.

Notes 1 To have the capacity to organize our engagements with things is to operate within the space of reasons, and not merely the space of causes. The idea of the ‘space of reasons’ goes back to W. Sellars, ‘Empiricism and the Philosophy of Mind’ in Minnesota Studies in the Philosophy of Science, vol. 1, eds H. Feigl and M. Scriven, Minneapolis: Minnesota University Press, 1956. The idea is central to McDowell’s avowedly Kantian way of thinking about intentionality: J. McDowell, ‘Having the World in View: Sellars, Kant and Intentionality’, The Journal of Philosophy, XCV, 1998, 431–91; see also his earlier, Mind and World, Cambridge, Mass.: Harvard University Press, 1994. 2 Stroud gives a particularly clear expression of this idea: B. Stroud, ‘Mind, Meaning and Practice’, in The Cambridge Companion to Wittgenstein, eds H. Sluga and D. Stern, Cambridge: Cambridge University Press, 1996, pp. 296–319. In Wittgenstein’s early work he makes a clear distinction between signs and symbols. A symbol is, roughly, an animated sign, for it is a sign with conditions for use. Thanks to Alois Pichler for reminding me of Wittgenstein’s early terminology here. 3 Cf. R. Brandom, Making it Explicit: Meaning, Representing and Discursive Commitment, Cambridge, Mass.: Harvard University Press, 1994, and also Articulating Reasons, Cambridge, Mass.: Harvard University Press, 2000, for the contrast between representationalist and inferentialist account of meaning. 4 The point should be familiar from Davidson’s writings. See also Evans’ idea of the generality constraint: G. Evans, The Varieties of Reference, ed. J. McDowell, Oxford: Clarendon Press, 1982, p. 100 ff. 5 It also binds the proposition together. The present point is central to Wittgenstein’s critique of Russell’s theory of the proposition, cf. chapter 2, §3. 6 The latter is particularly striking. Wittgenstein calls it a paradox, §95. 7 The best example of this is Fodor’s idea of a language of thought, cf. J. Fodor, Psychosemantics, Cambridge, Mass.: MIT Press, 1987, appendix ‘What there still has to be a language of thought’. 8 This point is made in J. McDowell, ‘Singular Thoughts and the Extent of Inner Space’, in Subject, Thought and Context, eds J. McDowell and P. Pettit, pp. 137–68, especially §6. 9 See chapter 5 §5. 10 The clearest defenders of a community interpretation of Wittgenstein are D. Bloor, Wittgenstein, Rules and Institutions, London: Routledge, 1997; M. Williams, Wittgenstein, Mind and Meaning: Towards a Social Conception of Mind, London: Routledge, 1999; S. Glendinning, On Being with Others: Heidegger, Wittgenstein and Derrida, London: Routledge, 1998; N. Malcolm, ‘Wittgenstein on Language

20


and Rules’, Philosophy, 64, 5–28; S. Kripke’s celebrated Wittgenstein on Rules and Private Language, Oxford: Blackwell, 1982, is often read as endorsing a community interpretation. The critique of community interpretations has a long history. Forcritiques specifically focused on Kripke’s interpretation: C. McGinn, Wittgenstein on Meaning: An Interpretation and Evaluation, Oxford: Blackwell, 1984; G. Baker and P. Hacker, Scepticism, Rules and Language, Oxford: Blackwell, 1984. More generally see: J. McDowell, ‘Wittgenstein on Following a Rule’, Synthese, 58, 1984, 325–63. and ‘Meaning and Intentionality in Wittgenstein’s Later Philosophy’, in The Wittgenstein Legacy. Midwest Studies in Philosophy vol. XVII, pp. 40–52, Notre Dame, Ind.: University of Notre Dame Press, 1992, both reprinted in J. McDowell, Mind, Value and Reality, Cambridge, Mass.: Harvard University Press, 1998; D. Pears, The False Prison, vols I and II, Oxford: Clarendon Press. 1987/1988; P. M. S. Hacker, Wittgenstein: Meaning and Mind. Volume 3 of an Analytical Commentary on the ‘Philosophical Investigation’, Oxford: Blackwell, 1990; M. McGinn, Wittgenstein and the Philosophical Investigations, London: Routledge, 1997; S. Schroeder, ‘Private Language and Private Experience’, in Wittgenstein: A Critical Reader, ed. H.-J. Glock, Oxford: Blackwell, 2001, pp. 174–98. Schroeder, like Hacker, is adamant that there is no textual evidence that Wittgenstein ever endorsed a community account of meaning. I think that is right. Pears and McDowell are equally clear that a community account could not be right. 11 For example, Schroeder, op. cit., p. 188. 12 Brandom, op. cit. 13 See Pears, 1988, p. 100 for expression of this very point in response to the idea that Wittgenstein held a use-theory of reference in his early theory of names. I suspect the use-theory interpretation is much closer to the truth than Pears, but only if we see Wittgenstein as rejecting the underlying bipartism that fuels the choice between representationalist versus inferentialist conceptions of meaning. I discuss the issue of a use-theory in the early writings in chapter 2.

CHAPTER TWO

Realism, Language and Self

2.1 Introduction Grammar is perspectival. It is the structure of the subject’s attitude to the world and it is a transcendental requirement that the subject’s attitude have this structure. It is not an empirical fact that the subject’s attitude is structured by grammar and neither is it an empirical achievement by the subject that its attitude is so structured. Although grammar is perspectival, it is a transcendental requirement that this is so and, as such, this is not something achieved by the subject. This means that the existence of grammar is not a fact about the world. It is not something to be discovered. That grammar exists is a condition for the possibility of our discovery of things. That grammar exists is something that is shown, not said, for any attempt to state the existence of grammar would deploy the very grammatical structures whose existence was being described, What can be said can only be said by means of a proposition, and so nothing that is necessary for the understanding of all propositions can be said. (NB p. 25e)

That, I suggest, is why Wittgenstein said that Logic must take care of itself. (NB p. 2a)

The existence and nature of grammar is something that is shown, not said, for there is no statable account of the source of grammar. Indeed, there is no theoretical account of the source of grammar. Grammar consists in the subject seeing the world aright.

22


In this chapter I want to suggest that these claims, that are central to my reading of Wittgenstein, were already implicit within his earliest writings on meaning. It often looks as if Wittgenstein’s early theory of representation presents a formal structure in which everything turns on the isomorphism of systems of representation with the structure of the world. His logical atomism posits a world of facts comprising arrangements of simple objects. The structure of the world is mirrored in the structure of language comprising names that can be organized into structured facts called ‘propositions’. This structure of a proposition is logical form. This is the same structure as the structure of facts in the world. This is the essence of his picture-theory of propositions. Propositions are structured pictures of the structured facts that make up the world. On the face of it, this has nothing to do with the thesis that grammar is perspectival, for the isomorphism is described independently of the subject. I think this is mistaken. The subject appears only fleetingly in Wittgenstein’s early writings (and then as a metaphysical subject, not an empirical subject) because of his essentialism about grammar. The structure of attitude that is the condition for the possibility of meaning is taken to be abstract and eternal and, as such, it is treatable as an enduring structure without giving much attention to that of which it is the structure – the perspective of the subject. Nevertheless, if we acknowledge the perspectival character of grammar in the early writings, at a stroke we make sense of Wittgenstein’s most enigmatic and dark sayings about the self. We provide a reading in which his mysticism about the self and ethics are of a piece with his semantic theory. In this chapter I move into the detail of Wittgenstein’s thought by giving a sketch of a reading of his early work. My choice of detail is selective. It sets the scene for the more detailed examination of his later writings in the following chapters.

2.2 Logic Takes Care of Itself The idea that grammar is perspectival is already glimpsed in his first writings on representation. In the Notebooks he introduces the idea of pictorial representation with the picture of two men fencing.

NB p. 7 This already contains the seeds to his mature theory of representation and the role of the self-as-will within that theory. This is a picture of two men


23

fencing, but ‘fencing’ is not itself an element of the picture. What then makes this a picture of two men fencing as opposed to a string of portraits? The answer that is developed in the picture-theory of propositions is that it is the fact that the two elements are related in a certain way that makes it a picture of two men fencing. The two elements figure in a pictorial form. There is a pictorial grammar. But what does this grammar consist in? The simple answer is that grammar consists in the form of the representation, a form that the representation shares with the world. Now, that is certainly Wittgenstein’s answer, but as it stands, it is not much of an answer. It raises too many further questions, centrally – what is the source of this sharing of form? Another answer is suggested by thinking through this first example. The picture of two men fencing contains just two signs, one for each man. What makes it a picture of two men fencing is the way these signs are arranged. The arrangement is to do with the way the signs are used. That move is of no help, however, unless we say what it is to use the signs in this way. It can look as if the picture-theory of representation invites us to adopt one of the two dominant opposing semantic traditions. To say that the pictorial form consists in the way the pictorial elements are used sounds like an invitation to adopt an inferentialist semantic theory, in which the use of the picture elements is the fundamental concept in a theory of representation. From that perspective, the pressing question is always: How does the use of a sign provide it with representational power? The alternative is to take the representational power of signs as primitive, but then we would need a sign for ‘fencing’ and that would be to treat all signs as names. On the face of it, then, the emphasis on the pictorial form, the arrangement of the elements, suggests that Wittgenstein is proposing an inferentialist theory of representation. I think this is wrong. He is proposing a use-theory of signs, but not one that is usefully labelled ‘inferentialist’. The arrangement of the elements is to do with the way the signs are used, but ‘use’ is not to be understood as something that could be characterized in rules of deployment. Rather, what ‘use’ means for Wittgenstein at this point is how we see the signs. The picture on page 7 of the Notebooks is a picture of two men fencing because that is how we see it. We see that two men are fencing. But the ‘seeing-that’ is not something generated from an account of seeingobjects – the two elements – plus rules for combining elements.1 Such a model would embody a bipartite account of meaning in which meaning is composed of signs plus that which animates the signs. It would be a case in which we see signs, and pictorial grammar is the set of rules that animate the signs and that turns the whole thing into a picture of two men fencing. On that account, grammar would be something that could be studied and stated. In contrast, if we think of grammar as how we see the signs, grammar is perspectival. It is the

24


structure of the way we see things, it is not itself, something we see. Just so with the picture of two men fencing. We see the two elements are representations of men fencing, but we do not have a representation of fencing. That they are fencing is shown by the way they are arranged and the way they are arranged is something we see. What makes this a picture of two men fencing is then the fact that it is composed of symbols – signs-in-use. And signs-in-use are signs from within the perspective of the attitude of the subject. The perspectival character of grammar is implicit in the first entry in the Notebooks. Wittgenstein says, Let us remember the explanation why ‘Socrates is Plato’ is nonsense. That is, because we have not made an arbitrary specification, NOT because a sign is, shall we say, illegitimate in itself! (NB p. 2b)

Two points stand out. The legitimacy of a sign – its possibilities for use – does not inhere in the sign itself but in how we use it. Whether or not a sign makes sense – can be used with others to say things – depends on us, not on the sign in itself. So, the basic unit of meaning is not the sign in itself, it is something to do with what we do with signs. An account of the grammar of signs is not independent of the perspective of the user of signs. Second, although he speaks of an ‘arbitrary specification’, as the discussion continues it is plain that the ‘we’ referred to here is not an empirical subject that animates signs as a Cartesian source of grammar. The arbitrary bit is the point that we have not given an adjectival sense to ‘Plato’.2 That is something that we could do as a matter of arbitrary specification. What is not available is that we create the grammatical category of adjectives. We do not create grammar, for if we did, the existence of grammar would be representable as a further fact about the sign system, a fact about what we have done. Furthermore, if grammar were a set of facts about what we have done, this would threaten the underlying realism about grammar necessary to capture the idea that grammar impinges on us. There are no facts of grammar in the world. Wittgenstein says How is it reconcilable with the task of philosophy, that logic should take care of itself? If, for example, we ask: Is such and such a fact of the subject-predicate form?, we must surely know what we mean by ‘subject-predicate form’. We must know whether there is such a form at all. How can we know this? ‘From the signs’. But how? For we haven’t got any signs of this form. (NB p. 2e)

The point that Wittgenstein is making is that the existence of signs that behave of a subject-predicate form is not and cannot be explained by reference to facts of this form. We have signs, e.g., ‘aRb’ that represent that a stands in a


25

certain relation to b, but that it so stands is not explained by a fact of the form. If so, that would be a special kind of fact. But how would it help if the subjectpredicate form existed? If the subject-predicate form existed, it would serve as an extra fact to be appealed to in understanding the proposition ‘aRb’. This would be like having the pictorial form for fencing as an extra fact to be appealed to in understanding the picture of two men fencing. But such extra facts would explain nothing, for they would introduce a familiar regress. It would mean that in order to understand the picture, or the proposition ‘aRb’, we would have to understand also the further fact that represented the grammar of these signs. This is a bipartite model of meaning in which the second component that is supposed to provide the rules of usage for signs is itself represented by further signs. That is useless. Alternatively, the second component could be supplied by a special experience, but that means that there would be something more fundamental than logic, the experience that supplied the connectivity to sign use. But there is no such experience: ‘If the existence of the subjectpredicate sentence does not show everything needful, then it could surely only be shewn by the existence of some particular fact of that form. And acquaintance with such a fact cannot be essential for logic’ (NB p. 3e), and, ‘If sign and thing signified were not identical in respect of their total logical content then there would have to be something still more fundamental than logic’ (NB p. 4b). Here, the point being made by Wittgenstein is that if there is an issue about how a set of signs represents what it does, then no further facts will resolve that issue. If we start the analysis with bare signs, rather than the fact that signs are arranged according to a grammar, we will need a further fact to represent the grammar of the signs. Such a further fact could only amount to one of two things: either it is something that is itself represented in signs, or it is a special experiencethataccompaniesour useofsigns. Thefirstinvitesaregress,the second Wittgenstein explicitly rejects. What then are we left with? We are left with the thought that the starting point of analysis of representation has to be signs-inuse, signs seen as having a grammar. Their possession of grammar is not, however, a further fact to be represented. That signs have grammar is a condition for the possibility of systems of representation. It is not something to be explained. This is why he says, ‘What can be said can only be said by means of a proposition, and so nothing that is necessary for the understanding of all propositions can be said’ (NB p. 25e). We can, of course, go wrong with signs and think that there is a sense to ‘Plato’ as an adjective, but we cannot go wrong in thinking that there are adjectives, or that there are sentences of subject-predicate form. It must in a certain sense be impossible for us to go wrong in logic. This is already partly expressed by saying: Logic must take care of itself. This is an extremely profound and important insight. (NB p. 2c)

26


The issue about how a set of signs represents is not then an issue to be resolved at all. It is a condition for the possibility of representation that signs have a grammar. As condition for the possibility of representation, this is not something that can be represented. It can only be shown. It is shown by how we use signs and our use of signs is, fundamentally, not how we literally move them around, but how we see them. Our use of signs is fundamentally our attitude to them. What attitude we take to particular signs is arbitrary. We could have used ‘Plato’ as an adjective. What is not arbitrary is that our attitude to signs exhibits the basic grammatical categories. By invoking the concept of attitude I am, in effect, suggesting an alternative way of taking the saying/showing distinction to the normal readings. There are two dominant readings of Wittgenstein’s early writings – the metaphysical and therapeutic readings.3 The metaphysical reading treats the Tractatus as offering substantive claims about the nature of reality and of how language must be to represent this reality.4 On this reading, there is, however, a tension between the claim that the book offers metaphysical claims but also includes philosophical propositions that are nonsense – ineffable truths that cannot be stated but can only be shown.5 The showing/saying distinction has to carry the weight of the idea that philosophical propositions are nonsense. On the therapeutic reading, there are no ineffable truths to be shown rather than stated. Philosophical propositions are literally nonsense and the purpose of the book is to purge us of the impulse to do this nonsense called philosophy.6 This is, however, odd, for it must mean that the philosophical propositions of the work convey nothing and yet lead us to realize that there is nothing to be done in philosophy. How can we climb a non-existent ladder and reach the safety of the view that says we have climbed nothing? My reading is, in some respects, closer to the metaphysical reading, but it is a metaphysics that is seen, clarified rather than theoretically articulated. This is also close to McGinn’s attempt to resolve the debate between the metaphysical and therapeutic readings with the idea that the work offers elucidations.7 One thing that can obstruct the reading I am offering is Wittgenstein’s own attempt in the Tractatus to provide a clear notation, a sign language that excludes the possibility of philosophical errors that arise from having a language in which ‘the same word has different modes of signification – and so belongs to different symbols’ (TLP 3.323). But having a clear notation does not give us a notation for grammar, it gives us a ‘sign-language that is governed by logical grammar – by logical syntax’ (TLP 3.325). Once we have a clear notation, it is not that the signs will then, as it were, all by themselves, articulate thought and exhibit inferential connectedness, it is just that we will see things clearly. We will gain a perspicuous representation that does not mislead, for it will no longer aim for a representation of the conditions for the possibility of meaning as some


27

fact to place alongside the facts of the world: ‘Philosophy does not result in ‘‘philosophical propositions’’, but rather in the clarification of propositions’ (TLP 4.112). If this is right, the metaphysics of the early theory of representation cannot be either a realist or an anti-realist metaphysics. Signs come already structured by grammar. The fundamental unit of semantic analysis is not a bare sign, it is the fact that signs stand in grammatical relations to one another. Similarly, the world is the world of facts, not of things (TLP1.1). The only conception of the world at play in the early theory of representation is a world already semantically organized into facts. It has to be thus in order to share grammatical form with the system of representation. It is tempting to ask if the world thus structured is a realist or anti-realist conception and tempting to think that it must be the latter. I consider this question in section 3, but a preliminary answer can be sketched now. The question whether Wittgenstein’s early semantic theory is realist or not really amounts to an issue about the source of grammar. Does the sharing of grammar by world and representation arise from the structure of the world or from the structure of the system of representation?8 If the former, the theory is realist; if the latter, it is anti-realist. My preferred view is that neither answer is right, for Wittgenstein does not really want an account of the source of grammar. That is the point of insisting that the existence of grammar, and its sharing by world and system of representation, is a condition for the possibility of representation. The matter is transcendental – ‘nothing that is necessary for the understanding of all propositions can be said’ (NB p. 25e). The transcendental requirements for the condition of representation belong in the realm of silence, the mystical about which nothing can be said. Of course, we can say a lot about this. I am doing it now. Wittgenstein’s point is that we are deluded if, in doing this, we think we are solving a theoretical problem. There is no theoretical problem to be solved. In providing these descriptions, in talking about these things, all we are really doing is getting ourselves to see things aright. We are simply reminding ourselves of what strikes us naturally and immediately about the picture on page 7 of the Notebooks – it is a picture of two men fencing. Logic takes care of itself because the existence of grammar is a transcendental condition for the possibility of representation. It does not concern a deep theory that is difficult to articulate. It concerns the way we see things aright. Grammar is the form of our attitude to the world. And that is a metaphysical thesis.

2.3

The Need for Grammar

The picture-theory of propositions says that propositions are essentially composite (TLP 4.032) and that there is a grammar – logical form – to

28


their structure. Wittgenstein also argues that both language and the world have simples. Propositions are complex structures, but there must be atomic propositions the structure of which is a complex of simple names. Simple names are correlated with simple objects. This allows two different accounts of the relation between language and world. Are language and world connected by the correlations between simples, or are they connected by the sharing of form? The former suggests an underlying representationalist semantic theory and a realist metaphysics; the latter suggests an inferentialist semantics and anti-realist metaphysics. Neither of these is correct. Understanding Wittgenstein’s metaphysics in the early theory of representation requires also understanding why he endorsed logical atomism. I turn to these points in the next section. First I want to clarify the need for grammar and its claimed ineffability. The argument for grammar is the argument given in §2 of chapter 1. Propositions are items to be assessed as true or false. They have correctness conditions. Therefore, they must be the sorts of things that can be assessed for whether or not what they say is correct or incorrect. A portrait does not, of itself, say anything. A name alone does not say anything. A picture is not a series of portraits, it is elements ‘related to one another in a determinate way’ (TLP 2.14). ‘A picture is the fact that the elements are so related’ (TLP 2.141 ). Similarly, propositions are structured complexes. A proposition is not a simple string of names (NB p. 8g,h,i; TLP 4.032, also 3.3, 3.141). At TLP 4.22 Wittgenstein says a proposition is a nexus, a concatenation of names. There is then a structure to it. The picture from Notebooks p. 7 is not three objects, but two objects related in a determinate fashion. The way that they are related draws upon a grammar of pictures and it is because the picture has this grammar that it is a picture of two men fencing rather than simply two portraits side by side. It is because propositions are composite that we can distinguish between the possibility that the proposition represents (TLP 2.202) and whether or not that possibility obtains (TLP 2.222). In other words, because of the essential composite character of pictures and propositions we can distinguish between the sense of the proposition and whether or not it is true. With names and simple portraits this is not possible. Portraits do, of course. say something, but that is only because we can discern structure within a portrait. A real portrait always says something about the person represented, e.g., what colour hair they have. But Wittgenstein’s point is that a portrait construed as a simple sign that stands for something does not, of itself, say anything. This is surely right. Uttering or writing a name does not amount to making a claim. The following (1) Russell . . .


29

is not a claim. (1) is the beginning of a claim, but does not yet say anything. The elements within a picture are, of course, correlated with things – this is what Wittgenstein calls the ‘pictorial relationship’ – but a picture is not a series of items that are correlated. A picture is an arrangement of such items. As argued in the previous section, there is no space in Wittgenstein’s thought for there to be an issue about the source of the grammar of this arrangement. The starting point, whether with picture or proposition, is the fact that elements are related according to grammar and the fact that things are related in certain ways – the world is the world of facts. The idea of pictorial form is generalized to that of logical form. The possibility of structure is what Wittgenstein calls pictorial form (TLP 2.15) and this is generalized to logical form which is the structure that a proposition has in common with reality (TLP 4.12). The need for logical form turns on the point made in the previous section. Without the idea of logical form, grammar, the composite character of propositions would not address how the elements of a picture or proposition hung together. There would be no account of the unity of a proposition. In a proposition, such as ‘aRb’, the elements fit together in a way that accounts for the way this proposition bears inferentially on others. We know that if ‘aRb’ is true then ‘Something stands R to b’ is also true as also ‘aR something’, etc. The requirement that propositions are composite does not on its own show how the truth of a proposition inferentially bears on others in virtue of its compositionality. The concept of form or grammar addresses that point. Grammar is the composite structure in virtue of which propositions are true or false and, thereby, in virtue of which their being true or false bears inferentially on the truth-value of other propositions. The points repeats a claim argued for in chapter 1. Russell treated propositions as strings of names. He treated relations as objects, for words for relations were names. In effect, this meant that Russell had no account of the unity of the proposition, for all the weight of the unity of the proposition had to be carried by special objects with gaps in them into which simple objects could fit. Wittgenstein’s position is that relations are not objects any more than fencing is something that needs to figure in the picture in the Notebooks in addition to the arrangement of the elements that stand for the two men. This means that we cannot say how ‘aRb’ represents that aRb, but we can show it by the way we use the signs correctly. We can, of course, say that ‘aRb’ has a certain grammar. It contains a two-place relation plus two names, but to say this is to say something of the form, F(‘aRb’).

30


It is not to state a general account of what grammar is, we can only represent grammar from within the grammatical forms we inhabit.9 It is an enduring theme throughout Wittgenstein’s work that we cannot step outside of our systems of representation and give an account of their grammar: ‘A picture cannot, however, depict its pictorial form: it displays it’ (TLP 2.172), and, ‘A picture cannot . . . place itself outside its representational form’ (TLP 2.174). He says the same about propositions: ‘In order to be able to represent logical form, we should have to be able to station ourselves with propositions somewhere outside logic, that is to say outside the world’ (TLP 4.12), and, What expresses itself in language, we cannot express by means of language. Propositions show the logical form of reality. They display it. (TLP 4.121)

On the face of it, it is not clear that this is right.10 The point about pictures seems reasonable enough. Even when Magritte paints an image of an easel on which there is an image of the same scene that contains the easel, the picture is still not a picture of the picture being painted. We know that the picture is a move from within the language of picture-making, the picture itself is not a picture of a picture depicting its pictorial form. But with the more abstract system of representation that we have in language, why should we not be able to represent logical form? The point is puzzling, for Wittgenstein does, of course, state the general form of the proposition. Now, he will proclaim that the things he states within the Tractatus belong to the realm of nonsense and anyone who understands him aright will come to see this and throw away the ladder of propositions on which they have climbed (TLP 6.54). But this is not an adequate reply. What then are we to make of the supposed ineffability of grammar? My suggestion is that the claimed ineffability is an overstatement on Wittgenstein’s part for the claim that would be better expressed by saying that grammar is perspectival. That is to say, the point at issue is not really whether we can say what logical form is, for we surely can, and Wittgenstein does. The point concerns what is achieved in doing this. Any representation of grammar will employ grammar in making that representation. This, in itself, does not matter. It matters only if we assume that the existence of grammar is a fact, albeit rather abstract, about how things are and that the representation of grammar is an attempt at a theoretical account of what makes representation possible. The point is best seen not in terms of the unity of the proposition, but the unity of inference. Grammar is the structure by which propositions inferentially bear upon one another. If the existence of grammar were a theoretical hypothesis,


31

something to be represented as a piece of theoretical knowledge, then it would have to be a piece of theoretical knowledge that gave an account of how one set of signs entailed another. This is, once more, an account that embraces bipartism about meaning. The suggestion is that there should be a philosophical account of what makes one set of signs entail another. But if we start with mere signs, nothing is going to suffice. We have the signs, p!q p and we want to know why it follows that q. There is, however, no theoretical knowledge required here. For sure, if we know the sense of the signs we will know why it follows that q, but the issue then returns – can the sense of signs be given a fully explicit theoretical articulation? Wittgenstein’s central insight was to see that no such fully explicit theoretical articulation is possible. Understanding the sense of signs is always grounded in our use of signs construed in terms of the way we see them, not in terms of theoretical knowledge we have about them. If someone is puzzled about whether q follows from the first two sign sequences, nothing we can say will get them to see the point. Getting to understand the rules of logic does not require getting into the right relationship with appropriate objects. It is not a matter of learning facts. Frege treated the truths of logic as reflecting the structure of a Platonic realm, but for Wittgenstein there are no truths of logic, despite the fact that he repeatedly represents them! His point, I think, is this. For Wittgenstein, the subject comes to recognize that logic is all right by getting into the right sort of relationship with the world – the relationship not in which logic is a body of truths about the scaffolding of the relationship, but in which logic is a condition of the possibility of standing in that relationship. The relationship is that of holding sentences to be candidates for truth/falsity. Call this the fundamental attitude. The subject has this attitude to the world and it is having this attitude that makes thoughts connected logically. If this attitude is a condition for the possibility of thought and judgement, it is not something about which we can formulate thoughts and judgements in the sense that coming to have this attitude could be seen as a theoretical achievement. Of course we can talk about this attitude – I am doing it now, but talking about it does not change it, nor discover it, nor instruct others into it. Gaining this attitude is not a theoretical achievement, indeed, it is not an achievement at all. Having this attitude is a transcendental condition for the possibility of thought and talk, we do not get it by thinking and talking. So although we can put this attitude into words and talk about

32


the grammar of representation, in so doing we achieve no solution to the problem of how language represents the world, for there is no real problem. The point is analogous to Wittgenstein’s remarks about ethics: ‘The world of the happy man is a different one from that of the unhappy man’ (TLP 6.43), but what is different is the attitude. It is not something that can be put into words. Ethics cannot be put into words (TLP 6.421). Suppose we say to a student of logic, ‘Look, you have assented to p and to p!q, now you have to assent to q and then they say, ‘And what if I don’t?’. We could then, of course, go on by saying, ‘But look, if p and p!q, then q, and you have assented to p and to p!q so don’t you have to assent to q?’ And the student then replies, ‘I don’t see why I have to’. Is there a gap between what they are resisting and what they already know about ‘p’, ‘p!q’, etc., as if we could supply them with some knowledge, some proposition so that then they would respond, ‘Ah, now I see it. You’re right, now I see why I must move on from p and p!q and assent to q’? But there is no gap here. The logical law of modus ponens does not hold the fabric of thought together as if a tear might appear and rip our thoughts asunder. We would say that such a student was not lacking knowledge, they were not waiting to be informed by a course in elementary logic; rather, we would say that they had the wrong attitude to the facts, they did not have the attitude of thinkers. Having the right attitude is a condition for taking part in thought and talk, it is not something that one gets instruction in via thought and talk. A similar point is made in TLP 6.422 where Wittgenstein says When an ethical law of the form. ‘Thou shalt . . .’, is laid down, one’s first thought is, ‘And what if I do not do it?’. . . our question about the consequences of an action must be unimportant. – At least those consequences should not be events . . . There must indeed be some kind of ethical reward and ethical punishment, but they must reside in the action itself.

The person who responds to the ethical law this way shows that they have a different attitude and that therefore no amount of ‘Thou shalt . . .’ is going to impress them. Ethics is not constituted by a body of propositions about the world, or about a way of thinking about the world. Ethics is constituted by our attitude to the world. Similarly, logic is constituted by our attitude to the world. There are no truths of logic or of ethics. In both cases, logic and ethics are to do with how we see the world, the nature of our attitude to the world, not what we discover when we see the world. This is why reward and punishment both lie in the action, the attitude. They are not facts that justify the attitude. Logical form, the grammar of content, is to do with how we take objects, it is not something extra, like an extra realm of objects, to be taken in


33

alongside ordinary objects. Similarly, the picture on page 7 of the Notebooks does not need a third object to represent fencing. It is the arrangement of the two objects that makes it something that exhibits the unity of a proposition, a picture that says that two men are fencing. Logic and ethics do not belong to the world, they belong to the shape of our attitude to the world. I shall speak of character as the shape of our attitude to the world. In the early writings there is such a thing as the essential character, the shape of our attitude that fits the shape of the world. It is because of this essentialism that the subject whose attitude matters is the metaphysical subject. It is not the empirical subject. In the early writings the empirical subject disappears (TLP 5.5421). With the loss of logical essentialism, by the later writings our attitude is more malleable and hence character must include not only the shape of our attitude to the world but also the capacity to amend that shape. This is most easily seen in cases where it is easy to think of the world changing, e.g. the social world, but Wittgenstein also allowed it in the case of mathematics. The character of a mathematician allows the possibility that in doing proofs one changes the world. This is not conventionalism, for it is not an option in a choice between Platonism and conventionalism; neither of the horns of that dichotomy are tenable. It is an acknowledgement of the dynamic nature of attitude. That is the difference between the early and later periods, the difference between an essentialist versus dynamic notion of attitude and concomitant concept of character. The perspectival nature of grammar raises an important metaphysical issue. In order for a proposition to represent reality, proposition and reality must share the same logical form (TLP 4.12). Now, what shares form are facts. Propositional signs are facts (TLP 3.14). They have to be facts for ‘Only facts can express a sense, a set of names cannot’ (TLP 3.142). This, of course, is the point already discussed, that propositions are names arranged according to logical form. It is the fact that ‘a’ leftflanks ‘. . . R . . .’ and ‘b’ rightflanks ‘. . . R . . .’ that says that aRb. A mere string of names cannot do this, only the structured fact can do this and it does it by being structured according to the grammar of logical form. But the world is the world of facts, not things. It too is structured according to logical form. The facts at issue are the facts of grammar. But if grammar is perspectival, these facts too are perspectival. They are not bold facts realistically construed. The world that is the world of facts is, to put the point in Kantian terms, the empirical world. It is the world already configured by grammar, it is the world for the subject whose perspective supplies grammar. The unity of language and world in the sharing of grammar is then a unity for a subject, for it is a unity for the perspective that supplies grammar. The world of facts and the representational systems of facts are united from the perspective whose attitude is the site of grammar. If this is

34


right, Wittgenstein’s metaphysics in the early period must be a version of transcendental idealism. The world is the empirical world for the subject as configured by the grammar of the subject’s attitude. We need to see if this suggestion makes sense. The issue is, at heart, what connects language and world in Wittgenstein’s early theory? Is it the sharing of form or is it the correlations between atomic signs and simples? Furthermore, what is the source of the sharing of form by world and language?

2.4 The Metaphysical Options I want to distinguish four options for the metaphysics of the Tractatus: 1 2 3 4

Uncritical realism Cartesianism Deflationary metaphysics Transcendental idealism.

Position 1 is the reading of Wittgenstein given by Pears.11 It is realist for the source of grammar is the structure of the world and it is uncritical, for it simply posits that language inherits its form from the structure of the world without any account of how this is achieved. I call position 2 ‘Cartesian’, for it is the position that says that the sharing of form by world and language is an achievement due to the working of the empirical subject. The mind has a simple and unexplained power to attach simple signs to simple objects. On position 2, the fundamental connection between language and world is made by correlations, the connections between linguistic and metaphysical simples. Position 2 is attributed to Wittgenstein by a number of authors who claim to find evidence of a phenomenalist account of simples in his early writings.12 Positions 3 and 4 both enjoin a form of silence with respect to the metaphysics of the language/world relationship. Position 3 is the view that the world is no more than that which is revealed by a study of logical form. I call it a ‘deflationary’ metaphysics, for it runs the risk of having no critical leverage to the study of logical form.13 Position 3 can be thought of as the opposite to position 1. Uncritical realism treats the logical form of propositions as arising from the structure of the world; the deflationary option treats the logical form of propositions as revealing the structure of the world (TLP 4.121). Treated this way, the problem for position 3 is obvious, for it is unclear that it offers any conception of a constraint on the investigations into logical form. The most favourable way of interpreting this position is to see it as endorsing a ‘no-priority’ thesis between language and world, but that claim


35

simply labels what needs to be said. It does not really address the problem.14 It does, however, mean that there is nothing substantive to say about the relation between language and world. Position 4 – transcendental idealism – also enjoins silence, but it is a sublime silence. It is the silence that comes from the claim that the relation between language and world is an achievement, but an achievement of the metaphysical subject. It is the attitude of the metaphysical subject that renders grammar perspectival and this is something that, as a transcendental condition for the possibility of meaning, cannot be sensibly discussed as theoretical knowledge. It is something that comes from seeing the world aright, but the seeing is not the seeing of the empirical subject who sees ordinary facts within the world. How things are in the world is unremarkable, ‘It is not how things are in the world that is mystical, but that it exists’ (TLP 6.44), and I suggest that the mystical includes the silence of the attitude of the metaphysical subject from whose perspective world and language are united in a common logical form.15 The key difference between positions 3 and 4 concerns the role of the subject. What differentiates position 4 is precisely that it leaves room for the concept of the metaphysical subject, something that was clearly important to Wittgenstein. I do not propose to examine in detail the issue of which of positions 1–4 provides the correct interpretation of Wittgenstein. Superficially, there are plenty of places in which Wittgenstein speaks of the relationship between language and world in a way that suggests a thoroughly realist conception of the latter. He speaks of comparing pictures and reality (TLP 2.223) and of comparing propositions with reality (TLP 4.05), although such remarks, on their own, say nothing about the nature of comparison, nor of the things being compared. These remarks are compatible with both a realist and a transcendental idealist conception of the world and the nature of comparison at issue. Elsewhere, the evidence within the Tractatus is not straightforward with respect to how the connection between picture and reality is established. At TLP 2.151 he says, ‘Pictorial form is the possibility that things are related to one another in the same way as the elements of the picture’, and then continues, ‘That is how a picture is attached to reality; it reaches right out to it’ (TLP 2.1511); ‘It is laid against reality like a measure’ (TLP 2.1512). This suggests that the connection between picture and reality is established by the sharing of form. And yet only three propositions later he says, ‘The pictorial relationship consists of the correlations of the picture’s elements with things’ (TLP 2.1514), and, ‘These correlations are, as it were, the feelers of the picture’s elements, with which the picture touches reality’ (TLP 2.1515). This strongly suggests that it is the correlations between elements and things that is the dominant ingredient in the connection between picture and reality.

36


The issue of which concept is dominant – correlation or the sharing of form – is not clarified by these remarks, let alone, ‘What a picture must have in common with reality, in order to be able to depict it – correctly or incorrectly – in the way it does, is its pictorial form’ (TLP 2.17). Trying to understand the metaphysics of the picture-theory of propositions turns on whether or not we think that language and reality are conceived by Wittgenstein as structures that are externally related or internally related. Positions 1 and 2 treat the relation between these structures as external. The uncritical realist treats the structure of language as the upshot (most likely causal) of the structure of the world; the Cartesian treats the correlation between name and object as something established by the activity of the empirical mind. In both cases, the connection is contingent. For the transcendental idealist, the connection between language and reality is internal, for it is something achieved by the metaphysical subject in the realm of silence. It is not part of how things are, it is a condition for the possibility of talking about how things are. In short, the issue turns on whether we read Wittgenstein’s remarks on these matters as quasi-empirical remarks about how language and reality are connected, or whether they are treated as transcendental claims.16 I suggest they are always the latter. I shall propose this reading by considering just three related issues concerning objects in the Tractatus. The issues are: What is the argument for simples? What does Wittgenstein mean by objects? Does Wittgenstein hold a use-theory of names? There is a common thread linking these issues, for they all touch on whether or not Wittgenstein’s conception of objects is realist. Wittgenstein says that there must be simple objects, they make up the substance of the world (TLP 2.021). He says the existence of simples is required for otherwise ‘whether a proposition had sense would depend on whether another proposition was true’ (TLP 2.0211). The sense of a proposition cannot depend on the truth of another, for otherwise ‘we could not sketch out any picture of the world (true or false)’ (TLP 2.0212). It is not transparent how the requirement for simples is meant to connect with the prohibition of the sense of a proposition depending on the truth of another. The claim that objects make up the substance of the world sounds a straightforward realist claim, and yet notoriously Wittgenstein gives no clear example of what he means by a simple object.17 Furthermore, when he treats of names for simples he says, The meanings of primitive signs can be explained by means of elucidations. Elucidations are propositions that contain the primitive signs. So they can only be understood if the meanings of those signs are already known. (TLP 3.263)

Black found this passage ‘wilfully obscure’ and Ishiguro took it as clear evidence that Wittgenstein held a use-theory of names in the Tractatus.18 We


37

need a conception of Tractarian objects that makes sense of all these points. I suggest we can get such a conception by concentrating on what seems to be the obvious problem with the Ishiguro interpretation of TLP 3.263. Pears articulates the problem well when he says that without an account of independent reference for names ‘language could never get off the ground’.19 The worry is that a use-theory of names fails to show how names gain independent reference. Pears points the contrast between a use-theory of reference and a theory of independent reference in terms of the idea of ‘direction of fit’. He says that Ishiguro assumes20 ‘that the name in the proposition sets the standards of fit, and its reference is whatever object meets the standards’. In contrast, on the opposite direction of fit, ‘the object is the dominant partner in the relationship, and its inherent possibilities decide whether the name thereafter represents it’. Pears’ distinction between opposing directions of fits expresses concerns that can also be expressed with Brandom’s distinction between representationalist and inferentialist semantics. The use-theorist is an inferentialist, for it is the inferentialist concept of grammar, of the name’s position within language, that sets the standards to which the concept of an object has to fit. In contrast, to take the direction of fit as emanating from objects is, most naturally, to adopt a representationalist semantics in which the representational power of the name is prior to its inferential power. Pears gives clear expression to Brandom’s distinction when he writes, ‘Sentences belong to systems, but their nature is not to be discovered by a lateral investigation of their connections with one another or their place in our lives: what we need to find out is how far their senses reach into the reality which they describe.’21 There is a real danger in the idea of a use-theory and there is something worth worrying about in Pears’ requirement of independent reference. But the points that need making here are not the ones Pears makes. Pears’ discussion of this matter is too epistemological. He treats the ‘direction of fit’ metaphor as if it were a matter of which came first – inferential power or representational power. Pears thinks that TLP 4.0311 supports his preferred direction of fit: One name stands for one thing, another for another thing, and they are combined with one another. In this way the whole group – like a tableau vivant – presents a state of affairs.

He then says, ‘How could this possibly be said by anyone who believed that the references of the names were fixed after this bit of language had been set up?’22 But this makes the whole issue sound like an empirical achievement, as if we had to decide whether to first work out how signs fitted together before

38


attaching them to things, or vice versa. Indeed, Pears’ central objection to Ishiguro that a use-theory leaves it a mystery how ‘language could get off the ground’ again makes the existence of language sound too much like an empirical achievement. If that is what was really going on, one would expect much more epistemological discussion from Wittgenstein in his account of names. It will not do to assume that he simply had not realized the need for a serious epistemological component for either account of the direction of fit.23 In so far as the union of world and language is an achievement, it is a transcendental one and in so far as there is a real issue about securing independent reference, the point again is a transcendental one, not an empirical one. What this means is that the worry is not ‘Do we refer successfully to objects with our names?’, for the only answer available to that question is to look and see! It is an empirical question with a simple empirical answer. The question that needs addressing is whether the account of names is an account that shows that they perform their basic semantic function. That function is to supply an object to the truth-conditions of the propositions in which they occur and ‘supply an object’ means that they make an objective contribution to truth-conditions. That is to say, what matters about names making independent reference is that the truth-conditions to which they contribute genuinely are objective truth-conditions. This is a real constraint on a theory of names. Now, the intuitive purchase on this constraint is got by saying that names contribute to genuine truth-conditions by contributing objects. But that intuitive remark is also puzzling. Again, Pears gives a good expression of what is puzzling about this:24 How does the relation between name and object explain the sense of a proposition in which the name occurs? The object which is the name’s reference simply is what it is, inarticulate and lumpish. So how can it help the name to its miraculously articulate achievement in a proposition? How does this heavier than air machine get off the ground?

Of course, if we treat the remark that names contribute objects to truthconditions as an empirical claim then it is puzzling how lumps of matter get to participate in a name’s having sense. But the world that is spoken of by Wittgenstein is the world of facts. It is the empirical world already semantically organized. Its objects are not ordinary empirical objects, the existence of objects is a logical requirement, not an epistemological discovery. Commenting on Wittgenstein’s claim that the relationship between language and world is unstatable, Engelmann remarked,25


39

Here lies the reason why Wittgenstein’s investigation is a logical one in contrast to the epistemological attempts of traditional philosophy, which are at bottom psychological, whether they admit it or not. For they treat that relationship between language and the world as part of the process of human thought, whereas the impossibility of stating that relationship springs from the logical relation of image to original which obtains between thoughts and propositions on one hand and the facts of reality on the other.

The logical investigation is a transcendental one. The requirement that names contribute to objective truth-conditions is not an empirical requirement; it is a transcendental requirement. The requirement is that for meaning to be possible, propositions must be the sorts of things that can be true or false – ‘To understand a proposition means to know what is the case if it is true’ (TLP 4.024). Whatever else is meant by ‘true’ it must introduce a concept of correctness such that one can be incorrect in uttering a proposition. And the notion of incorrect is a primitive notion of being wrong. It is not being wrong in the sense that this proposition does not fit well with some other proposition; that would be only a deferred sense of error. If all error were deferred, it is unclear that we would have a genuine concept of error. If we have no genuine concept of error, we have no genuine concept of truth. For names to secure independent reference then, it must be the case that not all error is deferred. It must be possible for names to contribute to the truth-conditions of propositions such that genuine error can occur. The account of reference cannot permit the possibility that all error is deferred. Now, the problem with a use-theory of names, as with any inferentialist semantics, is precisely that it looks like it permits the possibility that all error is deferred. If the account of a name’s reference is given in terms of the name’s ‘lateral fit’ with other signs, then there is no guarantee that error is ever genuine. But the problem is not really the point about ‘direction of fit’, for the problem could be solved from within a use-theory point of view. All that is required is the idea that the patterns, or grammar, of use that are taken as primitive are themselves objective patterns. They will be objective patterns if they are patterns that are themselves only capable of individuation in an object-involving manner.26 I suggest then that the requirement for simples is a logical requirement, it serves the transcendental purpose of ensuring that propositions get objective truth-conditions that permit genuine error. If the concept of an object is a logical one, then it is no wonder that Wittgenstein gives no clear examples of simples, nor that he fails to address the epistemology of our contact with such things. These matters are of no concern, for the existence of simples is transcendentally necessary. This is not an empirical claim. This also explains the logical atomism and the prohibition against the sense of a sentence depending on the truth of another.

40


If a proposition has a sense, it has objective truth-conditions. Suppose an account of a proposition having sense required some other proposition to be true. That would mean that the account of sense would not show that genuine error was possible, for until we had a proposition whose sense did not depend on the truth of another, error would remain deferred. That is to say, if the sense-conditions of a proposition required the truth of another proposition there would be no guarantee that sense would reach out and make contact with something objective. The account of sense would be inferentialist. This way of putting the point of the prohibition then is close to Carruthers’ reading of the argument for simples. Carruthers represents the argument for simples in the following way:27 (a) Necessarily, there are genuinely singular propositions that do not presuppose grasp of generality (b) If (a) then the truth-conditions of singular propositions cannot depend on anything empirical, so the truth-conditions depend on the sense of the proposition alone (c) If singular objects are contingent then existence would be involved and that invokes generality and so conflicts with (a).

The key point here is (a). This is the claim that is motivated by the concern to escape an inferentialist account in which error is always deferred. This is the basis for both Pears’ and Carruthers’ criticisms of the Ishiguro/McGuinness reading of Tractatus names. The Carruthers argument does not, however, show that simples are necessary, for that requires (c) and that is false. It is not true that if singular objects are contingent then singular propositions involve generality. That can be avoided not by saying that singular propositions must not entail existential claims, but only that singular propositions are object-involving. Carruthers tries to head this off by saying that an objectinvolving conception of singular propositions does require existential claims. For example, consider a demonstrative proposition. Carruthers claims that one can imagine taking the singular demonstrative proposition ‘This is F’ to a possible world in which there are no Fs. But that is a highly contentious view of singular demonstrative propositions and, I suspect, just wrong. It is wrong because it assumes an object-independent account of singular propositions, but that is the point at issue. On an object-dependent account of singular propositions, the existence of singular objects does not involve generality, it involves a Russellian demonstration of a particular.28 The point here is quite general and applies to understanding contemporary discussion of singular reference. The worry about inferentialism is the worry that an inferentialist account of a singular proposition will be one in which


41

the proposition turns out to entail an existential claim and, on a purely inferentialist account, that claim does not tie us up to the world. The response is to give a use-theory in which singular propositions are not conceived this way, but in which they are object-involving. This is closer to Russell than most commentators normally allow. It also requires that we treat propositions such as ‘This exists’, ‘Objects exist’, etc., as nonsensical, for without the entailment of an existential proposition these are not well-formed. The existence of objects is a condition for the sense of propositions, it is not something expressed by a proposition.29 What the above reconstruction of the argument for simples reveals is that there are two senses in which a theory of reference can be called ‘Russellian’. In one sense, reference requires something like a baptism by acquaintance. This is Russell’s account. The other sense of ‘Russellian’ just means an account of reference that is object-dependent.30 In this sense, reference does not require baptism by acquaintance, it is compatible with a use-theory and also with Frege’s sense/reference distinction. We see our way past the dichotomy of representationalist vs. inferentialist, or Pears’ ‘independent reference vs. use-theory’, when we realize that the harmony between thought and reality consists not in the latter bullying the former nor the former constructing the latter. Rather, it consists in the attitude we take to things. In the early writings, this is the attitude of the metaphysical subject. Once logical essentialism has collapsed and grammar becomes dynamic, it becomes the attitude of the dynamic self with the capacity for creative meaning. To sum up, there are a number of issues about realism in the Tractatus: the issues about objects and the issue about the self and solipsism. The issue about objects comes to this – If meaning is use then the worry is that the whole contribution to an account of objects comes from the lateral connections between sentences in language and that then does not sound like it is an account of objects. The central semantic concept of error remains deferred error and so propositions fail to get objective truth-conditions. This is Pears’ worry when he says that on such an account language will never get off the ground. Pears’ alternative is to say that there is a balance between the contribution of objects and the contribution of language to an account of meaning. But this is a ‘lucky’ balance. It is an uncritical realism that leaves it unclear who we thank for bringing it off. One response to this bears on the issue of realism about the self. What if it were the metaphysical self that brought this off? In that case, objects are empirically real but transcendentally ideal. The account of meaning draws wholly, in the transcendental end, on the contribution of the mind, but only the mind of the metaphysical self. This would be a transcendental idealism that has to be seen, not stated. In contrast

42


to Kant’s transcendental idealism in which the conditions for the possibility of experience are ideal, on this version it is the conditions for the possibility of meaning that are ideal. There is another way of taking all this though.

2.5

The Self

Suppose that the lucky balance between language and world is brought about by a use-theory in which the patterns of use get their objectivity because they are object-involving? This is Russellian, although not in the baptism by acquaintance sense. It allows a no-priority thesis between thought and reality, but not the version of no-priority thesis found in the therapeutic reading of Wittgenstein in which the claims of the Tractatus are literal nonsense. The difference is that the self does not disappear, it remains as the subject with attitude. It almost vanishes. It is left just as an ‘attitude’ that thought has to the world – an attitude of normative engagement. The attitude is still an attitude of a subject, but given the essentialism of the character of attitude, there is no real work left for the subject to perform. Put like this, solipsism and realism are barely separable, for although the union of language and world is perspectival – it is a union for the subject – the subject has so little work to do that its contribution evaporates. The attitude of the metaphysical subject is left as the essential structure of thought and so it is not really the achievement of the metaphysical self, for that vanishes and is in common with the world-self. Hence the heavy dose of mysticism in Wittgenstein’s discussion of the self. What is notable is that he both wants to have the concept of the metaphysical subject and also to have little or nothing to say about it. That the self is left as owner of the essential structure of thought and still has a role to play is due to two points. First, there is a contrast between will and independence of will. Indeed, this distinction is essential for the reading I have suggested. If an object-involving account of use is to mean anything, we need some purchase on the concept of object-involving use that does not beg the question concerning objects. The purchase provided is the notion of that which is independent of will. I argued (chapter 1) that the basic realism about grammar required the idea that patterns of correct/incorrect use exist independent of will. That concept exists in Wittgenstein’s early work: ‘The world is independent of my will’ (TLP 6.373), and can also be seen in the remarks in the Notebooks entry for 8/7/16 p. 74. To get that concept of the world, as that which is independent of will, you need the concept of the subject, even if it is the vanishingly small metaphysical subject. The concept of the subject is needed, for the idea of that which is independent of will is only conceived relative to that which is dependent on will. The two ideas – world


43

and subject – are not separable when conceived this way. That is why grammar – the structure of the world and of thought – is perspectival. The subject never drops out of the picture. Coming to see this is not therapy. It is the metaphysics of attitude, the attitude of will to that which is independent of it. The will cannot alter the facts, it can only alter its attitude to the facts, ‘it can alter only the limits of the world, not the facts – not what can be expressed by means of language’ (TLP 6.43). Second, the concept of attitude is required in order to make sense of a concept that Wittgenstein acknowledges in his early writings although, despite saying that it is of the utmost importance, he barely treats of it. This is the concept of the ‘and so on . . .’. That grammar exists is not a fact in the world, it is the form of our attitude to the world. In the Notebooks Wittgenstein remarks, The fact that it is possible to erect the general form of proposition means nothing but: every possible form of proposition must be f or es e ea bl e. And that, means: We can never come upon a form of proposition of which we could say: it could not have been foreseen that there was such a thing as this. For that would mean that we had had a new experience, and that it took that to make this form of proposition possible. (NB p. 89 e,f,g)

This is one of the clearest early statements of Wittgenstein’s thesis that the conditions for the possibility of meaning are not statable, but must be seen. And this seeing cannot be surprised. The conditions for the possibility of meaning are not further facts, nor experiences that could surprise us. Neither of these would suffice for the former would employ the very logical form that was meant to be explained and the latter would either equally beg the question or still have to be brought within intentionality. It is because we see things aright that we know in advance what form propositions take. There are no surprises and this is why he says, We now need a clarification of the concept of the atomic function and the concept ‘and so on’. The concept ‘and so on’, symbolised by ‘. . . .’ is one of the most important of all and like all the others infinitely fundamental. (NB p. 89 k, l.)

What is striking about these passages is that Wittgenstein is clear that although he is talking of a seeing, rather than speaking, the concept of seeing at play here is not ordinary empirical seeing. The concept of successive applications of an operation is the same as the ‘and so on’ (TLP 5.2523). Without this concept we could not construct logic and mathematics (NB p. 89

44


n). He is clear that there is no experience of logical form. We might see that an operation can be successively applied, but it cannot take an experience of the form of a proposition that underpins the ‘and so on’. These remarks are startling as precursors to the issue that Wittgenstein later explored with such subtlety and vengeance in the rule-following arguments. There is no fact that underpins the ‘and so on’ and neither is there an experience. The argument against the idea of an experience of the ‘and so on’ is presumably something like this. If it took an experience for a form of proposition to be possible, an experience that introduced the form ‘Fa’ as a new form, could this be other than an experience of the ‘and so on . . .’? That is to say, we would have to hit upon the form when experiencing ‘Fa’ and see it as of a new form. But then that would not be an experience of ‘Fa’ but of its form and the idea that it was general and supported an ‘and so on . . .’, but that already requires familiarity with the form. What Wittgenstein is arguing is that familiarity with a form is not an event, something that could happen in experience. It is, if you like, an ‘and so on . . .’ sort of thing. But that is not something that is available within experience. It is not given in experience, for it is a condition for the possibility of meaningful experience. Without the concept of the ‘and so on . . .’, ‘we would be stuck at the primitive signs and could not go ‘‘on’’ ’ (NB p. 90 b). What this concept introduces is not something that could, as it were, be laid against the primitive signs as a further item to be experienced. If that were so, we would never move on from the primitive signs. No addition of further experiences of items on a par with signs will animate them and produce the ‘and so on . . .’. That the form of propositions is foreseeable, that we can see the ‘and so on . . .’, is a condition for the possibility of signs carrying meaning. It comes from the way our attitude is to things. It is not something delivered from within that attitude. What is delivered is the world in all its logical form. What makes it deliverable, and what makes the unity of propositions and the unity of inference, is our attitude to the world. In the early writings, because of the essentialism of this attitude, Wittgenstein provides little further discussion of the concept of the ‘and so on . . .’. The concept is secure in the eternal essential structure of the attitude of the metaphysical subject. Once that essentialism is threatened and the concept of attitude permits a more dynamic reading, the concept of the ‘and so on . . .’ comes to dominate his thinking. His solution to the problem of how to handle that concept is still, as I shall argue in chapter 4, of a piece with the treatment in the early writings. The concept of going on is secured in the attitude of the self-as-will. The concept is, however, in a more precarious position once the concept of attitude is dynamic. In the later writings, the attitude of the subject is more visible because it is dynamic – it is found everywhere in the particularities of language use which


45

now no longer conform to a single crystalline structure, but which comprise continually shifting patterns. The given is the pattern of life, the form of life, although this is not necessarily the pattern/form of a community. The point is that this is still not an empirical form, but it is not static either. The subject, when we get to the later writings, is the empirical subject living in the thick of things, but that it has the fundamental attitude of holding propositions true is still a transcendental requirement for meaning. If we think of grammar as something that stands apart from the world (the world as lived) then we will be inclined to think of it as something an account of the structure of which will always be less than an account of objects, objectivity, truth, etc., and all the other hard concepts. But that is not how Wittgenstein is thinking of it. Forms of life are not empirical structures, they are not patterns an account of which could possibly fall short (leave us short-changed) in getting an account of the world. Thought and reality (form and world) are one. Our thought reaches right up to the facts.31 Seeing how this can be the case is what brings about the sort of philosophical peace for which Wittgenstein strived. In the early period, this seeing was a quiet sublime gaze on the essential form of the unity of language and world. In the later writings, it is the embedded seeing-that is immersed in the hurly-burly of a language use that can shift as our gaze moves. The difference is the difference between an essentialist and a dynamic account of seeing things aright. The philosophical peace issues from the same sort of result – logic takes care of itself because we see things aright.

Notes 1 I return to the contrast between seeing-that and seeing-objects in chapter 6. 2 Compare TLP 5.4733: ‘. . . the reason why ‘Socrates is identical’ says nothing is that we have not given any adjectival meaning to the word ‘identical’. 3 Cf. M. McGinn, ‘Between Metaphysics and Nonsense: Elucidation in Wittgenstein’s Tractatus’, The Philosophical Quarterly, 49, (197), 1999, 491–513, for a discussion of the choice between these readings with which I have much sympathy. 4 The main sources for this reading are: N. Malcolm, Nothing is Hidden: Wittgenstein’s Critique of his Early Thought, Oxford: Blackwell, 1986; D. Pears, The False Prison, vol. I, Oxford: Clarendon Press, 1987; P. M. S. Hacker, Insight and Illusion, Oxford: Clarendon Press, 1972, 2nd edn, 1986, and Wittgenstein’s Place in Twentieth-Century Analytic Philosophy, Oxford: Blackwell, 1996. 5 For a thorough examination of the idea of ineffable but important nonsense, see J. Conant, ‘Must We Show What We Cannot Say?’, in The Senses of Stanley Cavel, eds R. Flemming and M. Payne, Lewisburg: Bucknell University Press, 1989, pp. 242–83.

46 6 7 8 9

10

11 12

13 14

15

16 17

Realism, Language and Self See C. Diamond, ‘Throwing Away the Ladder: How to Read the Tractatus’, in The Realistic Spirit, Cambridge, Mass.: MIT Press, 1995, pp. 179–204. McGinn, op. cit. The metaphysical reading goes for the former, the therapeutic reading holds a nopriority thesis between structure of world and structure of language. Cf. TLP 3.332, 3.333 for this point in response to Russell’s theory of types. Hacker, 1996, op. cit., claims relations were objects (p. 30), but gives as his main source for this Wittgenstein’s later discussion of this in Wittgenstein’s Lectures, Cambridge 1930–32, ed. D. Lee, Oxford: Blackwell, 1979. The remark from the lectures that Hacker seizes on is, at best, obscure; see also T. Ricketts, ‘Pictures, Logic and the Limits of Sense in Wittgenstein’s Tractatus, in The Cambridge Companion to Wittgenstein, eds H. Sluga and D. Stern, Cambridge: Cambridge University Press, 1996, for a similar assessment. See Wittgenstein’s Notebooks 1914–16, p. 110, for a clear rejection of the idea that relations are names. See J. Koethe, The Continuity of Wittgenstein’s Thought, Ithaca, NY: Cornell University Press, 1996, pp. 27–28, for a similar reading of the difference between Russell and Wittgenstein This is the issue between the metaphysical versus therapeutic readings. Are there ineffable truths that can only be shown, or is there nonsense? On my reading, these things fall within the scope of what is seen. This is ineffable only in the sense that it is perspectival and so not capable of articulation independently of the perspective of the subject. Pears, 1987, p. 9ff. Such a view can be found in: A. Kenny, Wittgenstein, London: Allen Lane Press, 1973, and The Legacy of Wittgenstein, Oxford: Blackwell, 1984; P. M. S. Hacker, Insight and Illusion Oxford: Clarendon Press, 1972; J. Hintikka and M. Hintikka, Investigating Wittgenstein, Oxford: Blackwell, 1986. This is the therapeutic reading of Diamond and Conant, op. cit. Deflationary metaphysics is usefully compared to Davidson’s rejection of scheme/ content dualism: D. Davidson, ‘On the Very Idea of a Conceptual Scheme’ in his Essays on Truth and Interpretation, Oxford: Oxford University Press, 1984. Despite his attempts to attribute a Russellian phenomenalism to Wittgenstein in which names are attached to patches of the sensory field, Hacker (1972) also considers a version of transcendental idealism. However, he only used it to interpret Wittgenstein’s remarks on solipsism and ethics at the end of the Tractatus. I think the transcendental idealism runs throughout Wittgenstein’s early work including the semantic theory. I shall not discuss further the therapeutic option (3) which treats these remarks as literal nonsense. First Hacker, 1972, and Kenny, 1973, then Hintikka and Hintikka, 1986, have claimed to find evidence that simples are sensory objects, like the sense-data of classical phenomenalism. I find the evidence shaky and driven by the need to find some account to fit the presumption that the relation between language and world is contingent and in need of establishing by something or someone.


47

18 M. Black, A Companion to Wittgenstein’s Tractatus, Ithaca, NY: Cornell University Press,1964; H. Ishiguro, ‘The Use and Reference of Names’, Studies in the Philosophy of Wittgenstein, ed. P. Winch, London: Routledge, 1969; ‘Can the World Impose Logical Structure on Language?’, Wittgenstein – Towards a Re-evaluation, Proceedings of the 14th International Wittgenstein Symposium, eds R. Haller and H. Brandl, Vienna: Verlag, 1990,pp. 21–34; ‘The So-Called Picture-theory: Language and the World in Tractatus Logico Philosophicus’, in Wittgenstein: A Critical Reader, ed. H.-J. Glock, Oxford: Blackwell 2001, pp. 26–46. 19 Pears, 1987, p. 100. 20 Op. cit., p. 111, he applies the criticism to McGinness too: B. McGuinness, ‘The So-Called Realism of the Tractatus’, in Perspectives on the Philosophy of Wittgenstein, ed. I. Block, Oxford: Blackwell, 1981. 21 Op. cit., p. 115. 22 Op. cit., p. 109. 23 Although this is precisely the oversight Anscombe attributed to Wittgenstein: E. Anscombe, An Introduction to Wittgenstein’s Tractatus, London: Hutchinson, 1971. Anscombe treats the difference between Wittgenstein’s early and later work as, essentially, resulting from his realization that he had left out epistemology from the Tractatus. I find this an incredible interpretation. On my account, the obvious epistemological problems are ignored in the Tractatus because they have been solved, as it were, off the page. The union of world and language is secured transcendentally, for the world is transcendentally ideal and only empirically real. Cf. Ishiguro, 2001, op. cit., for a clear account of Wittgenstein’s low esteem for epistemology during his early period and, hence, the gulf between his methodology and that of his contemporaries. 24 Op. cit., p. 110. 25 P. Engelmann, Letters from Ludwig Wittgenstein with a Memoir, ed. N. Malcom, Oxford: Blackwell, 1987, p. 102. 26 This might sound circular, for what is the ‘object’ in ‘object-involving’? The answer here is that circularity is a problem only on the assumption of a bipartism about meaning in which meaning is use plus that which adds the normativity. To say the account of use is object-involving must be to give a unitary account of use. I return to this in §5 below and chapter 3. 27 P. Carruthers, The Metaphysics of the Tractatus, Cambridge: Cambridge University Press, 1990, p. 118ff. I simplify Carruthers’ account here. 28 Cf. G. Evans, The Varieties of Reference, ed. J. McDowell, Oxford: Clarendon Press, 1982, for this idea and M. Luntley, Contemporary Philosophy of Thought: Truth, World, Content, Oxford: Blackwell, 1999, chapters 11, 12, for discussion. 29 Cf. TLP 4.1272. 30 This is Evans’ use of the adjective ‘Russellian’. 31 Cf. §95. McDowell notes that this reaching right up to the facts is available in both Wittgenstein’s early and later work and amounts to a version of a no-priority thesis regarding the form of the world and the form of language. See J. McDowell, Mind, Value and Reality, Cambridge, Mass.: Harvard University Press, 1998, p. 28.

CHAPTER THREE

This is How We Play the Game

3.1

Introduction

When we use words, we use them in ways that can be measured for their correctness and incorrectness. What we say can be true or false. Regardless of whether or not our word use is rigorous, playful, closely circumscribed or open-ended, in the central case of when we say something with words there is space for the issue of whether we have spoken correctly. Our assertoric word use is subject to a basic semantic constraint, it is subject to correctness/ incorrectness conditions. I approach Wittgenstein’s later work by viewing it as an attempt to deny (1) There is a transcendent source for the correctness/incorrectness of word use without endorsing (2) An account of the standards of correctness/incorrectness of word use is given by a mere description of use. We are often tempted to think that (1) is required for a full-blooded conception of objectivity, for without a transcendent source for the standards of use we will lose the normativity of standards. The standards must be external to use, for otherwise there would be no source to the normativity, the ‘ought’ regarding correct use. Our use of words is calibrated against something. If, however, we deny (1), it looks like we are left with nothing more than a description of use. I call it a ‘mere description’, for the model in (2) is a description that fails to address how language is calibrated against standards of correctness/incorrectness. To endorse (2) would be to relinquish a pre-theoretical full-blooded conception of language use having objective


49

standards of correctness. And all that ‘full-blooded’ means for the moment is whatever is required in order to capture the idea that genuine error is possible in our use of words. This is a notion of error that consists not in using a word at variance with common usage, but in using a word at variance with how things are. The contrast is between error that is merely deferred and genuine error. A full-blooded conception of objectivity is one in which the notion of that against which word use is calibrated does not turn out to be merely further word use. McDowell has done as much as anyone in claiming that there is a position between (1) and (2).1 What I want to do is to say something about what that position is like. What do we learn, including what do we learn about ourselves, from characterizing the position between (1) and (2)? The answer turns out to be that we learn quite a lot. In particular, we learn to replace the idea that the standards of correctness/incorrectness can be articulated with the idea that such standards reside in our capacity to see things aright. A key to understanding this move is Wittgenstein’s anti-intellectualist account of our grasp of meaning.2 It is not possible to articulate fully the conditions for the possibility of language, for the conditions include, critically, our capacity to see things aright. Grammar is perspectival. And from this it follows that we have to uncover a good deal about ourselves, in particular our active role in the account of the standards of language use. At the heart of my reading of Wittgenstein is the idea of a unitary model of meaning. This is a model in which the normative standards of use are immanent to use. This is in contrast to the standard bipartite model of meaning in which that which provides the normative standards of correctness/incorrectness are extrinsic to use. I characterized bipartism as a model comprising two components, the sign plus that which provides the normativity to sign use. The fundamental division is between a normless component and the norm-providing component. The model applies equally to ‘use’. If we conceive of use as normless, then a bipartism about meaning sees meaning as a combination of two ingredients – use plus that which provides the standards of use. The version of this that Wittgenstein opposes is the version that treats use as a pattern of word deployment and in which the standards of use are provided by a Platonic pattern that picks out correct use. That is, however, only one version of a bipartite model of meaning. The lesson that I take from Wittgenstein is more than a response to Platonism. It is the lesson that how we play games with words is not just a matter of uncovering the patterns and rules in our play, the central point is that we are active players and our role is ineliminable from an adequate account of correct use. In this chapter I start the business of characterizing a unitary model of meaning. Sections 2–4 deal mostly with the key conceptual shifts necessary to get a unitary model of

50


meaning in focus. Sections 5–8 illustrate the model with a discussion of key early sections from the Philosophical Investigations.

3.2 Two Kinds of ‘Hidden’ A key to Wittgenstein’s later writings is the rejection of essentialism and his explicit endorsement of the meaning-is-use thesis. He held this thesis in the early period, but only explicitly stated it later, §43, For a large class of cases – though not for all – in which we employ the word ‘meaning’ it can be defined thus: the meaning of a word is its use in the language.

The key difference in the later writings is that there is no essence to use. He rejects the idea that behind use sits the essential structure that drives the application of words in language.3 The search for an essence of language can lead us to ‘feel as if we had to penetrate phenomena’ (§90). But this is not Wittgenstein’s approach in his later work. His investigation is a ‘grammatical one’ that ‘sheds light on our problem by clearing misunderstandings away’ (§90). Clearing away misunderstandings does not amount to coming to a statement of the essence of language, as if the essence is something ‘hidden from us’ (§91, 92). On the contrary, ‘Nothing is hidden’ (§435, cf. also §126, 164). So, clearing away misunderstandings is not achieved by a theoretical articulation of what lies behind language use or practice. Indeed, it is not achieved by any articulation, for there is no theoretical articulation of what makes language work. The conditions for the possibility of judgement do not require stating, they need to be seen.4 There is, for Wittgenstein, nothing behind our use of words. What you see is what you get. Put like that, it might appear that what is left within language use is less than we anticipated. If it is all open to view when we see things aright, then that sounds like it cannot amount to much, for what is open to view is, traditionally, merely the appearances – how things seem. If what you see is what you get, all you are going to get from an account of language use for which there is no essence is an account of how things seem. And that does not sound as if it is a philosophically interesting account of language use. Surely, so an objector might say, language use is the starting point for a theory of language, not the terminus? And if it is the terminus, that just shows that there is, after all, no such thing as theory of language. So, on that way of understanding the thought that what you see is what you get, we get a clear sense of Wittgenstein’s hostility to theorizing. We get a clear model of his antiphilosophical stance (§124, 126, 127, cf. TLP 4.112).


51

On this way of seeing things, doing philosophy is concerned with the hidden essence of language use, the theoretical articulation of that which lies behind the mere appearances found in use and which provides the normative standards to everyday use. In rejecting this notion of the hidden, Wittgenstein appears to be advocating a therapeutic rejection of philosophy. The point to which we return is the point from which we start, our sense of immersion in the ordinary and everyday familiarity of language use. The hidden is contrasted with a flat descriptivism about language use that amounts to no more than homely reminders of a common-sense use of words. In place of philosophical theory we get a limp empiricism that notes who said what, where, how and in what context. The therapeutic turn that rids us of the impulse to philosophize returns us to the quiet of a common-sense use of words that is not changed by the therapy which ‘simply puts everything before us . . . Since everything lies open to view there is nothing to explain’ (§126).5 There is an alternative way of taking the ‘what you see is what you get’ image and thereby generating a more satisfying therapy. Perhaps what we see turns out to be a good deal more complex than we anticipated? Perhaps, in understanding what we see in language use – in getting to command a clear view (§5, 122) – we find that we do not confidently know our way about. Indeed, there can be things hidden from us without being part of an essence, The aspects of things that are most important for us are hidden because of their simplicity and familiarity . . . we fail to be struck by what, once seen, is most striking and most powerful. (§129)

Two things are striking about this passage. First, there is a notion of ‘hidden’ which does not require the idea of the essence that lies behind the surface appearance. What is hidden is better thought of as the ways our use ‘hangs together’, what he also calls the ‘aim and functioning of words’ (§5). What is hidden is the structure of use, not because it is something that lies behind use as an essence to be articulated and stated in a philosophical theory of meaning, but because it is a structure immanent to language use. Given its multi-faceted shape, this structure is difficult to take in in a perspicuous representation. The aim, in removing the obstacles to this hidden, is not, however, to achieve a statement of the hidden. The aim is to achieve a clear sight of this hidden – to see things aright. Any statement of how language hangs together would have to exploit the very phenomena that required explanation. So there could be no general explanation of this. There is no theoretical explanation, there is only a coming to see things aright. We do this not by theory construction, but by ‘throwing light on the facts of our language by way not only of similarities, but also of dissimilarities’ (§130).

52


This ‘coming to see’ can be revelatory in two sorts of ways. First, it reveals that the standards of correct use are immanent to use. We do not need to look anywhere other than in our use of words to find why that use is correct/ incorrect. Although it can sound banal, this revelation constitutes a major divergence from common readings of Wittgenstein. The idea that the standards of use cannot be immanent underpins the thought that what is available within an account of our use of words cannot be enough to justify further use – what has been said and done with words provides no normative ‘should’ to what we do next. That idea is central to Kripke’s celebrated reading of Wittgenstein on rule-following.6 To say that the standards of correct use are immanent to use is to say that an account of what has been said and done with words can on its own, if it is characterized properly, constitute a reason for further use. In other words, the Kripkean reading misdescribes, or better, underdescribes, what goes into use. The second revelation supports the first. The second thing that is revealing about the claim that the norms of use are immanent concerns what this shows us about who we are and how we stand to the world in our language use. To make full and proper sense of the idea that norms are immanent requires a reorientation in our self-conception. We get to have a significant role to play in an account of the source of grammar. Clarifying that role is the second revelation. It is what substantiates the claim that grammar is perspectival. What is hidden are the connections that bind our language use together, not because they exist in a hidden essence behind use, but because although these connections exist in language use it is difficult to achieve a clear sight of them. Nevertheless, a ‘perspicuous representation produces just that understanding which consists in ‘‘seeing connexions’’ ’ (§122).7 What is not clear is why it should be so difficult to see these connections let alone why they should be things that can only be seen, not stated. One temptation needs to be avoided. The difficulty here is not because the demand for a perspicuous representation is the demand for a ‘bird’s eye view’ of language, a view from an outside transcendent perspective and something that is impossible for us to attain. Indeed, it seems unlikely that Wittgenstein meant anything like this with his phrase ‘perspicuous representation’. When given the opportunity, his favoured translation of ‘die ubersichtliche Darstellung’ is a ‘synoptic view’; that is to say, a view that is immanent rather than transcendent.8 The second aspect of §129 goes some way to explaining the difficulty of attaining a synoptic view. The second point about §129 that needs remarking is the idea that when we come to see these connections they are ‘striking and important’. Why should this be so? In rejecting the idea of the hidden that stands in need of theoretical articulation, it is not obvious why there should be anything


53

striking and important in coming to acquire a clear view of the connections that can be immanent to language use. Wittgenstein thinks there is something of great importance in acquiring this clear view and it seems unsatisfactory to suggest that the importance consists simply in the cure for the impulse to philosophize. That might be striking for someone already struck by the philosophical temperament, but it is a rather localized notion of importance. Even from the point of view of the search for a transcendent hidden, it is not clear why the articulation of the hidden should be striking and important. It might be thought that the ‘strikingness’ would be due to the way that an articulation of a transcendent hidden required changes in ordinary use, but this does not follow. If there were a transcendent hidden, there is no particular reason why once articulated into a theory of language it should change anything about actual language use. Of course, if ordinary language use were, in some respects, flawed or damaged, then the provision of a theory of the transcendent hidden would be a tool with which such flaws could be repaired. There is, however, no reason why this model of the hidden should presume that ordinary language use is flawed. Ordinary use might be a poor guide to the character of the transcendent hidden and thus, if taken at face value, it could be misleading. That is the moral one takes from Russell’s development of the idea of logical grammar. Mistaking the superficial structure of language use for something that indicates its logical grammar can lead to unsustainable ontological extravagances – things wear the appearance of singular terms which analysis reveals to be quantificational structures. But even on the basis of that familiar model, nothing is here said about ordinary use being mistaken or in need of reform. Russell may have subscribed to that view, but it does not follow. What follows is simply that one should not read off philosophical results regarding ontology from the structures of ordinary language use. But that does not amount to the idea that ordinary language use stands in need of replacement or repair. It amounts only to a counsel of caution in what one takes the significance of ordinary use to be. In short, unless one assumes that one should always try to employ language the structure of whose use is transparent, then acceptance of the idea of the transcendent hidden offers no intrinsic critique of, or challenge to, ordinary language use. Nevertheless, the very idea of a hidden transcendent structure can itself be striking, but only if one had been labouring under the idea that ordinary language gave an accurate presentation of ontological commitments in the first place. Against that thought, to have a theory of hidden structure articulated can be a striking achievement. We need a more persuasive account than so far considered to support the thought that seeing the connections – the immanent hidden – can be striking and important.

54


I think there are two parts to an account of why seeing the immanent hidden is striking and important. The first part of this account is the easiest to capture. Suppose that the hidden that Wittgenstein seeks to reveal is the immanent hidden, the structure of connections between different parts of language use. The very fact that it is a notion of hidden suggests that coming to acquire a clear sight of this must be revelatory in some sense, for one is coming to acquire a clear sight of something that had previously been obscured. Perhaps it is like this: what one acquires with clear sight is a sense of how language connects together; one learns something about what Ryle would have called the ‘logical geography’ of concepts. One learns something about the grammar of our concepts and about how this is embedded in forms of life – standard ways of behaving with and responding to words in the context of action. Acquiring a clear view of the grammar of language need not amount to a restructuring of ordinary language use, but it can still be revelatory to have things thrown into relief. It is tempting to think that philosophical progress comes from theory construction. Even on the assumption that a theory of language would unearth the transcendent hidden, it does not follow that possession of such a theory would necessitate a restructuring of ordinary use. Ordinary use could be fully in order as it is. The philosophical enterprise would consist not in describing it, but in offering a theory of the underlying structure in virtue of which one’s descriptions of language use were true. So even the philosophical theory-builder can be descriptivist with respect to ordinary language use just so long as that use is correct by the lights of the theory of the structure that underpins use. If philosophical progress comes not from theory construction but from coming to see things aright, descriptivism about ordinary language use is still in order, but the philosophical lesson is also, in an important sense, descriptively couched. If the philosophical lesson is to see things aright, it comes from recognizing how things are, not just within ordinary language use, but recognizing also the structures of connections within that use. That is descriptive in so far as it attempts to catalogue how things are, although it can seem non-descriptive in so far as what gets catalogued includes connections between concepts and ways of using words that cannot be assumed to be transparent to view. Whether such a method is properly called ‘descriptive’ depends in part on the nature of the connections thrown into relief by the investigation into grammar. Clearly, if all that is on offer is an empirical study of how language happens to be connected in various ways, the method is descriptive through and through and amounts to a detailed ethnology of language use. Some commentators have been tempted to read Wittgenstein this way. There is an alternative.


55

Suppose the connections unearthed by a clear view of language use are connections that capture the normativity of language use, the connections in virtue of which language use is measured against standards of correct/ incorrect use. Suppose, that is, that one gets a clear view of the way that language is calibrated against things. This is something that is seen, not described and it is immanent to language use. Getting this idea clear is the major task for this and the next chapter. The idea requires that the calibration of language on things is direct and unmediated, for otherwise it should be possible of description rather than being something that is seen. In other words, what one gets to see is that language can stand in direct calibration with things.9 In that case, in so far as one does describe what one sees, by way of reminders, what gets described are not contingent empirical facts about language use, but a priori connections that make language use the normative practice that it is. Either way, philosophy is still not interfering with language use, it is describing it, but much is at stake in the latter kind of description. It all depends on what goes into ‘use’. If a description of language use is a mere catalogue of contingencies, the enjoinder in §124 amounts to the end of philosophy. This is what I meant by ‘mere description’ in the opening contrast between (1) and (2). Alternatively, if an account of language use is an account of use immanent to which are normative patterns of connections between concepts, utterances, connections between words and things, etc., then pursuing such an account will have to be tentative and focused on the details in order to get things right. For such a description is a description of the immanent hidden, and one reason why it is hidden is now this: It is hidden because it concerns the normative patterns of connections between concept use in virtue of which language is the sort of thing that has standards of correct/incorrect use. But rather than think of these patterns as transcendent structures graspable only by construction of a pseudo-scientific theory, they are structures that come into view. This then makes clear sense of §130, Our clear and simple language-games are not preparatory studies for a future regularisation of language – as it were first approximations, ignoring friction and air-resistance. The language-games are rather set up as objects of comparison which are meant to throw light on the facts of our language by way not only of similarities, but also of dissimilarities.

And I suggest that the ‘facts’ here include the facts about the normative patterns of connection within language use.10 This means that a description of language use that reveals the immanent hidden cannot be transparent, for it is a description not just of how things are, but of how things have to be. And that is the sort of thing that can be difficult to get right, it is contestable. Such

56


descriptions are opaque, not in the sense that we always get them wrong, but simply that we can get them wrong. One sort of revelation that can come from this sort of study concerns the connections between concepts that are foundational for language use. It is unclear that Wittgenstein ever considers the idea that some concepts might have a foundational role in the connections that make language use possible. Notwithstanding Wittgenstein’s apparent silence on this matter, I want to suggest that he is committed to an account of the positioning of certain very basic concepts that is revelatory. Furthermore, the revelation is both radical and in line with developments in contemporary philosophy of thought. In particular, I shall argue that Wittgenstein’s arguments throw into relief the foundational role played by the concept of the self-as-will (agency), the concept of judgement (the action of putting a subject and predicate together to make a claim) and the concept of seeing. In short, the grammar of our ordinary language use typically hides from us a self-revelation – the revelation that as language users we are not pawns in the grip of the transcendent structures of meaning or, for that matter, the empirical structures of the history of words. We are active shapers of the patterns of language use, agents with a capacity to judge and thereby to affect the ongoing patterns of use. Grammar is perspectival. That is the second part of the account of why seeing the immanent hidden can be striking and important. The revelation concerns seeing aright our role as language users – we are judges, agents with a capacity to shape ongoing use. This idea is connected with a key shift between Wittgenstein’s early and later conceptions of philosophy. In the early writings, I argued that Wittgenstein had a conception of the self as judge, but one in which the self ’s role was limited to holding the union of thought, language and world together. Its role was limited because the structure of language use was grounded in a transcendent eternal structure that gave the essence of language. By the later writings, the idea of the essence of language has been abandoned. Now, if the grammar of language has no essential structure and yet still underwrites the normative idea of correct/incorrect use (and that combination is not an easy one to hold onto), then the grammar of language has a dynamic to it that permits change over time. Change may be piecemeal and gradual, but not everything is fully fixed. The grammar of language cannot then be a wholly independent structure that drives our language use. It must, in part, be something that is driven by our use of words. We cannot be wholly passive in using language. We can acknowledge a responsibility for creating and sustaining the grammar of language. Getting the balance right between receptivity and spontaneity here is one of the hardest things in getting to grips with Wittgenstein’s philosophy.11 We are active participants


57

in shaping grammar. We are then judges who not only enact judgement in bringing subject and predicate together to form thoughts, but also in judging the applicability of a concept to something we are not wholly bound by nor driven by rules that could be articulated in advance. Like a judge in the English legal tradition we can make the rules up as we go along. Learning to accommodate that last remark will occupy much of the rest of this book. It is, however, one of the most profound reasons why the revelation of the immanent hidden is striking and important. When language works, and it works well and fully in its ordinary usage, it works not because we as users enact the rules written in some transcendent structure or because we follow the social custom; it works because we as users make it work. It is not the rules, the structure, that is in charge here. We are. That we are in charge is the key point behind acknowledging the notion of the immanent hidden, for there is more going on in ordinary use than immediately meets the eye and gets reported in traditional accounts of Wittgenstein’s meaning-as-use thesis. I want to explore these themes by examining the different ways in which they emerge in the rejection of essentialism and the transcendent hidden in the opening sections of the Philosophical Investigations. A key passage for understanding the account of Wittgenstein that I am offering is §154. This occurs after Wittgenstein has already introduced in §143 his celebrated example of generating an arithmetical series to which he returns in §185 in setting out the rule-following argument. The second paragraph of §154 contains a key ingredient to Wittgenstein’s position on the normativity of rules. He says, If there has to be anything ‘behind the utterance of the formula’ it is particular circumstances, which justify me in saying I can go on . . .

My suggestion for reading this passage will not be justified until the next chapter, but a first account goes like this. There is nothing behind the utterance of a formula that underwrites the normativity of the right way of going on; that is, there is no transcendent hidden in virtue of which going on one way rather than another is correct. The notion of the correctness of going on one way rather than another, for there is such a thing, is available in the particular circumstances. We justify one way rather than another not by subsuming it under a general theoretically articulated pattern, but by seeing the pattern in the particular case. And part of what it is to see the right development consists in the fact that the language user is an agent, a judge with the capacity to contribute to the patterns of right usage. A first sight of this way of understanding §154 comes into view in considering the early passages in which Wittgenstein first expresses his anti-essentialism.

58

This is How We Play the Game 3.3 Meaning and Use

The thesis that the meaning of a word is its use in language is already prefigured in the opening section of the Philosophical Investigations, ‘But what is the meaning of the word ‘‘five’’? – No such thing was in question here, only how the word ‘‘five’’ is used.’ What we need to track down is what the thesis finally formulated in section 43 amounts to. What goes into the concept of use? I think it is clear that the concept has a rich and complex content as used by Wittgenstein. It is less clear that commentators have always been fully sensitive to this content. I want to start by suggesting three different versions of the thesis that the meaning of a word is its use in language. I think it is clear that it is the third notion that Wittgenstein has in mind, but articulating three possible meanings of the thesis helps to clarify how much has to go into the concept of use if it is to play the role Wittgenstein gives it. Making this tripartite division is an analytical tool in clarifying and coming to see aright how Wittgenstein plays the game of using the word ‘use’. The first two concepts of use are thin conceptions of use, for they share a conception that on its own does not provide the normative standards of use. It is a conception that is typically the first ingredient in a bipartite model of meaning in which, because of the paucity of the conception of use, an extra ingredient is required to give an account of the standards of use. I shall freely interchange between talking about three concepts of use and three concepts of grammar. The three concepts that we need to distinguish are: 1 2 3

Syntactic use Transparent ordinary use Use as practice.

1

Syntactic use

The concept of syntactic use might also be called linguistic use, for it picks out an idea of the use of a word in the way in which it fits together with other words in order to form compound expressions. The thesis that the meaning of a word is its syntactic use has much in common with modern inferentialist semantic theories which take the inferential power of a sign as primitive, or at least as prior to the representational power of a sign. If meaning is syntactic use, it is unclear how, for example, ostensive definition could constitute the basic account of how words get meaning.12 In order for an ostensive teaching of the meaning of a word to work, the teacher and pupil would have to have already grasped the word’s syntactic use:


59

So one might say: the ostensive definition explains the use – the meaning – of the word when the overall role of the word in language is clear. (§30) [the] explanation . . . only tells him the use of the piece because, as we might say, the place for it was already prepared . . . We may say: only someone who already knows how to do something with it can significantly ask a name. (§31, cf. also 28, 29)

It is the notion of syntactic use that seems to be implicated in Wittgenstein’s earlier endorsement of a use-theory of meaning. Wittgenstein’s endorsement in TLP 2.673 of Frege’s context principle understood as a thesis about reference has puzzled many commentators.13 I am not suggesting that Wittgenstein has the inferentialist notion of syntactic use in mind in the Investigations, for what his notion of use is has yet to be determined. The above passages do, however, illustrate a line of argument that is compatible with inferentialism. If meaning is understood in terms of syntactic use then the basic semantic concept is that of inferential role. On this interpretation there is room for the idea of a hidden transcendent essence to language, indeed it seems quite plausible to suppose that such an essence must be available. The reason is as follows. If meaning is explained in terms of inferential role then we still have no answer to the question concerning the source of the grammatical patterns that define inferential role. Perhaps there is no such source? On the other hand, if one is making the step of explaining the representational power of signs in terms of their inferential power, that move looks redundant if there is not some account of what that notion of inferential power consists in. It would look odd, although not incoherent, for the inferentialist not to endorse some notion of a transcendent hidden, for why make the claim that representational power is explained in terms of inferential power if the latter provides no authoritative account of use? To make the inferentialist move and then to remark that the notion of inferential power is immanent to ordinary use makes the move between representationalist and inferentialist appear a very local and inconsequential tussle. Why should it matter which is primitive if whichever account is privileged turns out to provide no firmer foothold for an account of what makes language possible than to observe that everything is just as it is in ordinary use? Inferentialism without the idea that there could be a theory of the structural essence of language looks a very shallow idea. If meaning is syntactic use, one would still expect there to be an account of the hidden transcendent essence. This would be an account of the general structure of language. This, of course, is what I argued Wittgenstein did in his early period. He gave the ‘general form of propositions’, the essential structure that all propositions share. It is this that he says he is not doing, he says that

60

This is How We Play the Game these phenomena have no one thing in common which make us use the same word for all, – but that they are related to one another in many different ways. And it is because of this relationship, or these relationships, that we call them all ‘language’. (§65)

Section 65 immediately precedes the celebrated discussion of the concept ‘game’, but before turning to that, it is important to note that this passage, as indeed the discussion of games, can easily be taken to endorse what I take as the second version of the thesis that meaning is use.

2

Transparent ordinary use

The second version of the meaning is that use thesis is a descriptive empirical account of how language functions and of the role of individual words. It is a catalogue of the contingencies of language production. In contrast with the idea of syntactic use, transparent use does not support the idea of a transcendent hidden essence. Indeed, to treat use in this fashion is to deny any concept of a hidden at all. By ‘ transparent ordinary use’ I mean a concept of use in which nothing is hidden, a use of language in which there is nothing to hide and, therefore, nothing to be revealed. It is this concept of use that appears to be at play in the opening sentence of §124: ‘Philosophy may no way interfere with the actual use of language; it can in the end only describe it.’ If we read Wittgenstein as offering a therapeutic philosophy, one that aims to rid us of the impulse to philosophize, it is tempting to think that ‘use’ is transparent ordinary use. There are, on this reading, no revelations to be had in doing philosophy, there is only the regaining of a common-sense orientation within language and a recognition that its familiar contours exhaust all there can be said about how language works. The contours of language are, with respect to transparent ordinary use, familiar. The therapy consists in the regaining of something familiar. It is a return to a familiar orientation in language; it is not a reorientation. And in so far as ‘to imagine a language means to imagine a form of life’ (§19), then the form of life that illustrates use is a description of ordinary living. On this account, the concept of a form of life is an empirical concept. It is descriptive of ways of living and ways of using language and the description is transparent. The point of the adjective ‘transparent’ is that the description has no normative content to it – what you see is what you get and there is no scope for mistaking what you see.14 The trouble with taking ‘use’ in this second sense concerns not just the fate of philosophy, but of the very concepts that make philosophical enquiry about language pressing. The second sense of ‘use’ leaves no work for philosophy to undertake. That in itself is, perhaps, troubling only to philosophers. More


61

seriously, it leaves ungrounded the very idea that in using language we are responsive to normative standards of correctness and incorrectness of use. The big idea that is a catalyst for so much theorizing about language is the idea that in using language we do so in observance of standards of correct/ incorrect use. The core to this idea is the notion of correctness/incorrectness that applies to assertions in virtue of their content; it is the correctness/ incorrectness of truth and falsity. If the existence of such standards is acknowledged, and it is hard to see how it could be avoided, then the key question is: What constitutes the standards of correctness/incorrectness? What are these standards to which we are answerable in using language? What is the source of grammar? Taking ‘use’ as transparent ordinary use appears to render these questions unanswerable, if not incoherent. Once again, the idea that these questions are unanswerable lies at the heart of the sceptical interpretation of Wittgenstein’s rule-following arguments by Kripke and others. For such commentators, an account of how we have used words leaves future use wholly unshaped, for the account of past use contains no contribution to an account of the source of grammar. The core idea that can so easily prompt Platonism is the idea that the standards of use provide some measure of use and it is tempting to construe that in terms of standards that are transcendent of use. In using language in a way that is answerable to notions of correctness/incorrectness we are using language in response to something independent of us and of our use. The standards of use must, it seems, be transcendent, for if they were not then they could not serve their main purpose of providing a measure of language in use. The image that is seductive here is difficult to shake off, but it is the image of the standards of use as some sort of pattern – a transcendent grammar – to which our ordinary use is responsive.15 What does not look coherent is the position that says both that meaning is use and that use is transparent ordinary use. The former claim, if it is to capture the thought that meaning is constituted in part by the idea of normative standards of correctness, requires that use have the resources to say what constitutes these normative standards. But if ‘use’ turns out to be a descriptive account of transparent ordinary use, then we get no further than a description of that which stands in need of an account of normative standards – people saying things with words. Transparent ordinary use is characterized by an empirical catalogue of the contingencies of what people have said, when and where and in what sequence they spoke. As a catalogue of contingencies, there are no normative connections between different language uses and hence no right connections either. Transparent ordinary use invites a bipartite model of meaning, for its resources for characterizing use are so shallow. Hence it seems appropriate to look elsewhere for an account of normativity. At the same time, the

62


proponent of transparent ordinary use as a reading of Wittgenstein denies that there is any such extra ingredient to an account of meaning, all we have is a grammatical investigation. Such a position is unstable, for it provides a norm-less account of use and then cuts off any prospect for recovering norms from a further ingredient.16 There is an epistemological dimension to the concept of normative standards of correct/incorrect use and it concerns the notion of the hidden. If ‘use’ equals transparent ordinary use, then this is a notion of use in which nothing is hidden. The epistemology of norms suggests that this cannot be right. If our use of words is subject to norms, then there must be some capacity for the norms to be hidden, for our use can be mistaken. If the standards of use were transparent, there would be a puzzle about how we could be wrong in our use of words. We need room for the possibility that we get things wrong, whether in incorrectly applying a word or in drawing an inferential connection between one use of words and another. For this to be possible, it cannot be the case that how things seem to us is how things are. There must be the possibility of an epistemological gap here. It must be possible for our notion of how things are in the use of words to be mistaken. We need a right/seems right distinction for our grasp of the use of words, ‘otherwise thinking one was obeying a rule would be the same thing as obeying it’ (§202). The nature of this gap is not straightforward. What is required, minimally, is the possibility that how things seem to us regarding the use of words is different to how things are. We need the possibility of error in our use. This does not entail that that which constitutes the standards of correct use is everywhere distinct from our ordinary use. The requirement of a gap does not require the image that underpins Platonism. This is the image in which that which constitutes correct use is fully transcendent of ordinary use as a structure separate from the structure of ordinary linguistic practice and which the latter attempts to shadow. That image is more than is required to acknowledge the possibility of error. All that is required to acknowledge the possibility of error is that there be the possibility of separation between actual use and correct use. It is compatible with this that on occasion, and perhaps very often, ordinary use does not fall short of correct use and, in not falling short, manifests a direct calibration with things. This provides a conception of correct use that is immanent to ordinary use. Ordinary use and correct use can be fully commensurate with one another. With what frequency this might occur is neither here nor there. All that matters is that there is scope for a critical adjustment to ordinary use in the light of a notion of correct use. It is that scope that disappears on the second version of the ‘meaning is use’ thesis. It does not overstate the point to say that what is threatened by the concept of ‘transparent ordinary use’ is a certain realism about grammar, the patterns


63

of word use that constitute correct use. This is a realism that acknowledges that there is such a thing as the pattern of correct use. The difficulty is to see how to accommodate this much realism without sliding into the imagery that underpins Platonism.

3

Use as practice

The possibility that ordinary use and correct use be congruent and the calibration with things be immanent to use brings the third sense of use into focus. If we resist the Platonist image in which the standards of correct use exist as a transcendent structure that impinges on ordinary use but which is constitutively distinct, then we allow the possibility that norms of correct use are immanent to ordinary use. The third sense of the ‘meaning is use’ thesis treats the notion of ‘use’ as the immanent normative structures of language use. This is, then, an account of use which is not merely descriptively given, for it allows the possibility that coming to describe, and coming to see, ordinary use aright can be revelatory. On this third version of the ‘meaning is use’ thesis, the concept of use is a rich and complex concept I shall say that this concept of use is the concept of use as practice. I use ‘practice’ to signal the differences between this third concept of use and the first two. Syntactic use is capable of being rendered into a transparent description. A syntactic account of ordinary use might not be very transparent, but it seems reasonable to expect that a theory of logical grammar – the model of correct use – could be fully transparently displayed. Whether or not this is ultimately possible is no slight matter, for it would require that our central semantic concepts be capable of definition in formal set-theoretic terms. That is contentious.17 Ordinary use is also transparent. Because of this, both these concepts of use are thin concepts that prompt the need for a bipartite model of meaning in which meaning ¼ use þ that which delivers the norms. Use as practice is opaque, for even if it is describable, what gets described are patterns of correct use immanent in ordinary language. Because practice is opaque it is also contestable and open-ended. It is not capable of complete articulation, for it is not fully transparent. What goes into this third notion of use is something much richer than either of the preceding accounts. This is a thick conception of use. What the word ‘practice’ picks out is that thick use is far more than a conception of words and the way they fit together to form compound symbols. It is also a richer conception than that picked out by transparent ordinary use, a conception that merely catalogues what was said where and when. Transparent ordinary use is, if you like, sociologically rich, it fills in much of the contextual surround of ordinary use. Its descriptions of language use do

64


not, however, connect uses of language together, other than recording the empirical fact that x said this, y said that, etc. Use as practice is, in contrast, a conception of the way in which uses of language are normatively connected. It is a concept of use in which language hangs together. The normative patterns of correct/incorrect use are found within use, they do not lie behind it. It is a way of describing language use such that that which confronts language and renders its use correct or incorrect can be available within the reach of language. It is not a conception in which we describe language and then look for something external to language with which to measure the correctness/incorrectness of use. It is a conception in which use of language is described as not cut off from that which renders it correct/incorrect. That is what makes it a practice. It underpins a unitary model of meaning. It is also what makes this conception of use an account of language that requires what we have come to know from Davidson as the denial of scheme/content dualism.18 There is much at stake in this conception of use. Some of this, in particular the elaboration of the concept of practice will occupy us in the next chapter. In the remainder of this chapter I want to pursue the exploration of use as practice. First, I provide a further set of bearings for distinguishing between the three conceptions of use identified.

3.4 Use and Self The three conceptions of use that I have identified have different implications for the role of the self in an account of language. Suppose essentialism is endorsed regarding the patterns of use, perhaps because it is held that there is such a thing as a transcendent source of the patterns of correct use. This is an idea that is still available to the proponent of use as syntactic use. Such Platonism about the patterns of correct use leaves no clear role for the self in an account of language. With regard to the empirical self, its role is reduced to that of a pawn following the rules of correct use as discovered in Platonic heaven. Of course, if one accepts a conception of a transcendent metaphysical self, as I argued Wittgenstein did in his early writings, then the metaphysical self may have a more interesting role to play in a theory of language. It would, however, be a role that could not be described, for as part of the transcendent conditions for the possibility of language, it would be a role whose fundamental conception would be mystical. If a syntactic conception of use does not endorse essentialism about the patterns of use it still leaves no clear role for the self in an account of language. If use is understood as syntactic use, the dominant image is of the patterns of correct usage existing independently of what speakers and hearers say or do. The thesis of syntactic use endorses a


65

realism about the patterns of correct use in which there is no clear role for the self to play. Grammar is non-perspectival. The position of the self is similar if use is understood as transparent ordinary use. For sure, on this conception, language use is a catalogue of the empirical contingencies of what was said where, when, why and by whom, but there is no especial reason to dwell on the last point. Who made a particular utterance and why they made it is something of no especial concern. The structures of use on this account are the empirical structures of linguistic happenings. They are happenings in which speakers and hearers figure, but it is unclear that they figure as anything other than devices to locate utterances and pin down the structure being described. The speaker figures as site of an utterance.19 Indeed, if they had any more substantive role to play, this would not be a transparent ordinary-use account of use. The point that there is no particular role for the self to play in an account of use as transparent ordinary use is easily overlooked. Surely, so the thought might go, in describing ordinary use one could hardly neglect to mention the speakers and hearers who produce and consume the language being described? It is here that we have to be sensitive to the complexities of language use and to how much is being endorsed when Wittgenstein proposes his meaning is use thesis. If too much is allowed to be absorbed in the concept of ‘use’ without clear checking of its content, we will miss what is doing the work in an account of language. Whenever we describe ordinary use we make reference to speakers and hearers. We refer to shopkeepers and buyers of apples, teachers and young children learning arithmetic, children bouncing a ball against a wall and calling it a game. The issue at present is what goes into our descriptions of such language use. Does the description include the normative connections between these people’s uses of words and between their words and those things and properties that they represent? The problem for the notion of transparent ordinary use is that it describes use with insufficient resources to capture the normativity of meaning and that leaves it bereft of any notion of correct use. There is a danger of moving too quickly from a transparent ordinary description of use in which the patterns of word production are simply logged and the source of normativity is external to use, to a description that picks out much more than the contingencies of the patterns of word production and consumption, but includes the source of grammar too. If that is the case, we need to identify clearly what that source is. If it is something extrinsic to use (bipartite model of meaning) then it is something that is not available within a theory of transparent ordinary use. Alternatively, if it is something intrinsic to use, it is the sort of thing uncovered in a revelatory moment that shows that the sources of normativity are immanent to use. We need to be clear what goes

66


into the conception of use when we have the latter kind of description, for that will be a description that has shifted to an account of use as practice, a description in which the notion of correctness is immanent to use. Suppose then it was suggested that in giving a description of the correct ordinary use of words it was significant to note who made an utterance and why, where the ‘why’ question alluded to more than merely noting the previous utterances to which the target utterance responded. That is to say, suppose that an account of correct use did not merely map a set of patterns of use as a record of patterns but also included reference to the speakers and hearers not in so far as they produced and consumed utterances that satisfied a pattern, but as an element of the description that was in addition to the issue of satisfying a pattern. But that would be to say that the notion of correct use was not simply that of satisfying a pattern, it would include whatever else was alluded to by needing to make reference to the speakers and hearers. That means that such an account would not strictly be an account of use that described a transparent structure of linguistic happenings. It would be an account that had a far greater range of resources with which to characterize the notion of correct use. Once this point is articulated, it can seem obvious that such an account must be preferable to a merely descriptive account of transparent ordinary use. That seems right, but then the attractiveness of the proposed account resides almost wholly in the extra resources appealed to. Simply calling the proposed model an account of meaning as use leaves it unclear what those resources are, and where they spring from. Once we move to consider the concept of use as practice, the role of the self becomes quite different. I have suggested (section 2) that one of the things that is revelatory about the idea of the immanent hidden is the discovery that our role in language use is that of agents, subjects with a capacity for judgement and for shaping the ongoing patterns of correct use. If the norms of correct use are immanent to use, as in the conception of use as practice, then it is natural to ask: ‘How can norms be immanent to practice?’ and ‘What is the source of the normativity of correct use?’. These are awkward questions, but however we answer them it seems difficult to make much of a stab at them without locating a key role for the self. Indeed, I suggest that the idea of the norms of correct use being immanent to practice can only be made sense of within a framework in which the self as agent is in direct confrontation with that against which language is calibrated, namely, how things are. The model of this confrontation that I propose is that of the self-as-will confronted with that which is independent of will. If this is right, and it will take the next couple of chapters to begin to make the case for it being right, then on the conception of use as practice the self has an ineliminable role to play and grammar is perspectival.


67

A central challenge to the position just sketched will be to keep it distinct from the Cartesian account of the source of normativity criticized in chapter 1. On the account that I am proposing, the self as an agent with the capacity for judgement is an integral part of the account of the conditions for the possibility of meaningful language. It is because the self has such a central role to play that I prefer to express the overarching goal as an account of the conditions for the possibility of judgement. Meaning is something that we do and something that resides in our actions. Ourselves and our actions are the loci of interest, not language as such. The patterns of correct use of language are such only because they are strings of actions of agents with a capacity for judgement. It is not the pattern that matters,it is what it is a pattern of.

3.5 Use and Augustine’s Mistake Wittgenstein is not a helpful guide when trying to unpack how much goes into his concept of use. The matter is complicated from the opening section of the Philosophical Investigations where he picks on Augustine as a foil for his philosophical ambitions. Wittgenstein says that Augustine gives voice to a particular picture of the essence of human language and describes this picture as one in which words function as names and sentences are combinations of names. This picture is then criticized by the example of the shopper and the shopkeeper and the meaning of the word ‘five’. It is in discussion of this example in the opening section of the book that Wittgenstein introduces the concept of use when he says, ‘But what is the meaning of the word ‘‘five’’? – No such thing was in question here, only how the word ‘‘five’’ is used.’ Trying to get clear what Wittgenstein means by ‘use’ is thrown by the selection of the Augustine quote as the foil. The simple account of what is going on in the opening section is that Augustine treats meaning as naming, Wittgenstein treats meaning as use.20 Augustine is the representationalist. Wittgenstein is the inferentialist. But this is a parody. On the evidence of the passage that Wittgenstein quotes from Augustine, the latter is clearly a use-theorist of sorts. He is not endorsing a bald representationalism in which the meaning of a word is determined by a simple acquaintance of word with object. Augustine clearly states, ‘as I heard words repeatedly used in their proper places in various sentences, I gradually learnt to understand what objects they signified.’ Russell had held a bare acquaintance theory of meaning in which the meaning of a word was determined by an ostensive baptism of word on thing, but Augustine has nothing like this model. At least part of what constitutes understanding a word for Augustine requires a training in which words are used in their proper places in different sentences. At the very least

68


this amounts to the involvement of the concept of syntactic use in an account of meaning. It is not a position that is in the frame of the criticisms directed against the notion of ostensive definition in sections 28–31. Perhaps the difference between Augustine and Wittgenstein is then that the former is a syntactic use theorist and that Wittgenstein employs a different concept of use? It certainly seems right that Wittgenstein’s concept of use in §1 is more than the concept of syntactic use, for it is reasonable to assume that in the context of the shopper and shopkeeper example ‘use’ includes not just linguistic use but the whole context of actions implicated in the exchange as the shopper buys five red apples. That would suggest at least the concept of transparent ordinary use, the description of the full context of actions and social institutions that fill out the use of a word. But that conception is also available in Augustine. He says the intentions of others was shown by their bodily movements, as it were the natural language of all peoples: the expression of the face, the play of the eyes, the movement of other parts of the body, and the tone of voice which expressed our state of mind in seeking, having, rejecting, or avoiding something.

This sounds like a fifth-century precursor to Wittgenstein’s own remarks about the natural history of mankind. It is tempting to think that the idea of natural language introduces the concept of transparent ordinary use, that it is a descriptive concept. It is far from clear that this can be the case. In getting to grips with what Wittgenstein meant by ‘use’ we are, as he says, often provided with a model of the workings of language with ‘a haze which makes clear vision impossible’ (§5). This is the charge he lays against Augustine, but Augustine’s deployment of the notion of the natural language of peoples in terms of bodily movements, tone of voice and so on is potentially a very rich conception of use. Augustine speaks of the ways in which various bodily movements express our state of mind and there are at least two different ways in which the idea that bodily movements express mental states can be taken. One version of this idea is the conception that Wittgenstein himself struggles to capture throughout the Investigations, the other is the one he repeatedly critiqued. It is likely that it is the latter that Augustine endorsed, although the textual evidence is not straightforward. The problematic way of understanding the expression of mental states in bodily movement is to see the two things as in principle distinct but related such that the latter gives signal of the presence of the former. On this account, the relation between mental state and that which expresses it is contingent. Such a model permits a rigid distinction between inner and outer. For the moment it will suffice to say that the alternative conception of the relation-


69

ship between mental state and bodily expression is one in which the relation between the two is not contingent; we might say that the relation is internal.21 If the relationship between mental state and bodily expression is contingent and if we assume that understanding meaning is grounded in mental states, this has consequences for how we understand the concept of ‘use’. If ‘use’ refers to the bodily behaviour that expresses meaning, and that behaviour is characterizable independently of the mental state that constitutes understanding of meaning, then the characterization of use is independent of that which provides the normative standards to use. The norms, residing in our understanding of meaning, will be extrinsic to the description of use. It seems plausible to think that this is the model that Wittgenstein is responding to in Augustine. Augustine’s conception of understanding is one in which understanding is an inner event, something that is to be translated into and expressed in terms of outer behaviour.22 Indeed, more generally, Augustine has an intellectualist model of the young infant mind. He writes as if the infant mind comes complete with a full range of discrete desires and beliefs that await only the resources for translation into outer behavioural expression. In the passage immediately preceding the part Wittgenstein quotes Augustine says, I was unable to express the thoughts of my heart by cries and inarticulate sounds and gestures in such a way as to gain what I wanted or make my entire meaning clear to everyone as I wished; so I grasped at words with my memory. When people named some object and . . .

In this passage it is clear that Augustine conceives the infant mind with a capacity for thought, grasp of meaning, wishes, memory, etc., all independent of its means for expressing these things to others. The common language learnt from listening to others is a contingent tool, a set of signs that is only contingently related to what is expressed by means of it. Even though this tool has a grammar, and words have ‘proper places’ in sentences, the concept of use that characterizes this grammar is something that must be capable of characterization independently of that which supplies the normative standards to use. Therefore, Augustine’s concept of use must be that of transparent ordinary use. If that is right, in critiquing Augustine, Wittgenstein is critiquing transparent ordinary use and therefore his concept of use must be a richer one. I think this is right. Much goes into Wittgenstein’s concept of use, much more than is often credited. In his celebrated discussion of the rule-following arguments, Kripke develops a model in which an account of how words are used is devoid of normative content. The description of how we have used words cannot, of

70


itself, capture the normative force of how we should use them.23 It might seem that Wittgenstein’s early language games – the shopper and shopkeeper in §1, the builders §2 – are merely descriptive of transparent ordinary use.24 It might seem that they provide accounts of how language is used, but accounts that do not indicate the source of the normative standards of correctness and incorrectness. I think a more sophisticated account is required in order to understand even these fledgling language games. Cavell says,25 ‘we don’t just learn the word ‘‘game’’, but what a game is . . . this involves learning a form of life, words embedded in their standard connection with action’, and to that I would want to add ‘embedded in their standard connection with things’. The general point that Cavell has caught hold of is that we do not just note that a word was used on such-and-such occasion of such-and-such item. If that was all that was happening what we learnt would be a mere catalogue of contingencies. The content of our learning would be transparent ordinary use. But as Cavell remarks,26 What we learn is not just what we have studied; and what we have been taught is not just what we were intended to learn. What we have in our memories is not just what we have memorized.

The point Cavell is making here is central to understanding the complexities to Wittgenstein’s concept of use. Suppose that use were characterizable in syntactic terms. If that were so then what we have in our memories regarding use could be just what we have memorized, for the latter could be thought of in terms of patterns of syntactic use. That is the sort of thing that can be memorized. The notion of ‘memorize’ here is that which can be retrieved and displayed in full. That is the sort of thing that can be captured in terms of symbol patterns, a syntactic conception of use. What else then do we remember if it is more than what we memorize? Surely, the answer is that we remember not just that the word was used at a certain time of a certain thing, we remember also that it was correctly so used. That is something that, contra Kripke and other thin accounts of use, can be cited as reason for why we carry on using words as we do. On a syntactic conception of use, citing previous use gives no reason for future use. There is nothing in accounts of previous use that supports the normative ‘should’s. Thin use, e.g. transparent ordinary use, is a notion of use that leaves it looking like we are short-changed, that we are left without proper resources for carrying on in our deployment of words. Hence the impulse for a bipartite model of meaning. On the contrasting picture, we study patterns of words, but what we learn is more than the exhibited patterns. We remember more than the patterns that we memorize. We learn


71

and remember that words have been and are correctly used in certain ways. What we remember is not something that is detached from its standards of correctness. We remember words in use directly calibrated against things. The calibration must be direct, for otherwise it would be possible to factor out the calibration from what is remembered. Such a factoring out is characteristic of all thin conceptions of use, the sort of conception notorious in Kripke’s reading of Wittgenstein. It is that thin conception that is wrong. There are resources in past use that shape the future, providing we characterize past use correctly. The characterization of past use cannot therefore be simply a catalogue of contingencies. What goes into use has to be richer than most people are prepared to admit. It is, in part, because of our impulse to intellectualize our grasp of language, our sense that we ought to be able to articulate and state the conditions for correct/incorrect use, that it looks like use does not contain the resources to determine a concept of meaning that answers to our pre-theoretic full-blooded conception of meaning and objectivity. We cannot state sufficient facts to determine future use. But it does not follow that we do not have sufficient in our memories to carry forward from our past use and determine correct future use. That would only follow if we thought that what we have in our memories is what we have memorized. But that is what Cavell is denying.

3.6 Is ‘Slab!’ a Shortening of ‘Bring Me a Slab!’ or is the Latter a Lengthening of the Former? Learning a word involves learning a form of life. Like the concept of use, ‘form of life’ can be variously interpreted. One question to consider is whether ‘form of life’ is an empirical concept or a transcendental concept. To say it is an empirical concept is to say that it applies to contingent facts about the social surround of word use. We learn not just the syntactic use of words, we learn also the wider surround of actions and social structure in which word use takes place. To think of these structures empirically is to think of them as further elements that can be added to the catalogue of contingencies. So understood, the idea of a form of life embroiders the account of word use in syntactic use, but does not substantially add to it. We are offered more of the same, not anything fundamentally different. I think that the concept of a form of life functions, for Wittgenstein, as a transcendental concept. It is a concept that does not pick out mere empirical circumstances that surround word use, it picks out circumstances that are necessary conditions for meaningful word use. Taken empirically, a form of life would be concerned with our accommodations with the world, the habits

72


and behaviours by which our language use accommodates with the world. But that is to see such accommodations as extrinsically connected with word use. It is to treat the accommodations as causal encounters that get tacked onto the linguistic use in order to anchor the latter in the regularities of the world’s impinging upon us. If ‘accommodation’ is understood like this, we still have a bipartite conception of meaning. It is a function of ordinary transparent use plus accommodation. The latter, understood as the bodily causal responses to the world’s probing, cannot deliver the normativity. The trick here is to take ‘form of life’ as a transcendental concept that picks out the form of accommodation in which the world normatively impacts upon us. The basic condition is that speakers use words in direct calibration with things. A form of life is a practice of using words calibrated directly against those things that provide the standards of correct/incorrect use. A form of life is, then, another version of use as practice. It is concerned with our accommodations with the world which are conceptual through and through. The bodily accommodations that matter are conceptually structured all the way down, they are not bodily twitches and movements contingently related to transparent word use. They are the direct responses to the world that provide a direct calibration of words on things. At the same time, they are not ‘mere’ bodily impacts; they are the accommodations of a ‘living human being’ (§281, 283).27 Wittgenstein introduces the concept of a form of life in §19 where he raises what appears to be an odd question about the simple language-game from §2. The builder and assistant have been using the simple language comprising ‘block’, ‘slab’, ‘beam’ and ‘pillar’. In §19 he says that to ‘imagine a languagegame is to imagine a form of life’ and then asks whether the word ‘slab’ in the language-game of §2 is a word or a sentence. It seems natural to think that if it is a sentence, it is elliptical for ‘Bring me a slab!’ But then Wittgenstein asks, ‘But why should I not on the contrary have called the sentence ‘‘Bring me a slab!’’ a lengthening of the sentence ‘‘Slab!’’?’ This seems an odd question and, if the point is not understood, Wittgenstein’s use of it can add fuel to the idea that he is an anti-realist about the patterns of word use. Wittgenstein’s question is odd because if there were as much right to treat ‘Bring me a slab!’ as a lengthening of ‘Slab!’ as it is to say the latter is a shortening of ‘Bring me a slab!’ then we seem to lose the very idea of the patterning of language use. Endorsing some form of realism about the patterns of language use is unavoidable. The patterns in which we are interested are the patterns in virtue of which sentences that are true/false bear upon the truth/falsity of other sentences. A form of realism about these patterns is unavoidable because some form of realism about truth and falsity is unavoidable, at least sufficient realism to acknowledge that the standards of correctness/incorrectness obtain independently of will. So, if a sentence is true, it is so independently of our


73

thinking it, wishing it, hoping it to be true, etc. And, if it is true, the bearing its being true has on other sentences will also hold independently of our thinking it so, wishing it so, hoping it so, etc. That’s just the way truth is.28 We account for the way one sentence’s truth-value bears upon those of others in virtue of their common component structure. The idea that seems to be under threat in Wittgenstein’s question is the idea that the notion of patterns to language use exploits the combinatorial structure of language. The normal account of how sentences bear on each other’s truth-value is to say that the inferential connections in which they stand hold in virtue of their combinatorial structure. It may take some time to lay out the details of why connections hold, but if (3) Object 1 is red all over is true then (4) Object 1 is green all over must be false. We know this because we hold, (5) No object can be both red and green all over and the first two sentences have a structure that includes a singular term occupying a position bound by the quantifier in the general sentence. It is because of this common component structure that the sentences are, as it were, wired together. By such means their possession of truth-values bears inferentially upon one another. This much seems straightforward. Similarly, the thought that (6) Bring me the slab! provides the canonical form of what is expressed by (7) Slab! consists in the following obvious points. (6) has a structure that makes transparent the connections between it and sentences such as (8) Bring me a pillar! and,

74

This is How We Play the Game (9) Bring me a block!

or, (10) Take from me a beam! and, (11) Bring me something! The longer form has a three part structure that makes transparent its connectivity with these other sentences. If someone understands (6) and they also understand (10) then we would expect that they would understand, without any further training, (12) Bring me a beam! Such capacities for generative understanding are standardly explained in virtue of the subject’s grasp of the underlying structure of sentences. The four-word sentence makes this structure transparent and hence we hold that the single-word sentence is elliptical for the longer. (6) expresses the canonical form of the logical grammar of (7). The previous paragraph can seem compelling. There is an issue, however, on which Wittgenstein disagrees in the above account. It seems natural to suppose that the longer sentence makes transparent the connections between it and other sentences. The claim about transparency of inference is, however, ambiguous. In one sense it is correct, in the other false. Being clear about the transparency of inference can help provide an account of what Wittgenstein is doing in these passages.29 It clarifies further the thesis that grammar is perspectival. If an inference is valid there must be a pattern to the inference that can be employed on other occasions. Valid inferences are not one-off devices. The sense in which the transparency of inference is correct is that it is possible to provide representations of the pattern of correct inference which, within certain parameters, render it incontestable what the pattern of the inference is. Transparency in this sense is a matter of degree; it is a matter of representing inference in a way that makes it easier to spot the validity/invalidity. The transparency is epistemological. Classifying precisely what the parameters are in rendering the patterns of valid inference transparent is a moot point. Nevertheless given our ordinary way of employing words, (6) makes clearer how the builder’s utterance has the capacity to bear on other utterances than


75

(7). Accordingly we can say that (6) renders more transparent the validity of inference than does (7). The judgement here is, however, relative. It is not absolute, for epistemological transparency is a matter of degree. The sense in which it is false to say that inference is transparent is the sense in which it is held that there is a representation of the validity of inference that obtains independently of the perspective of the thinking subject. The validity of inference is not transparent in the sense that there is a representation of valid patterns of inference such that whether or not something instantiates that pattern can be determined purely syntactically. Whether or not a linguistic sequence instantiates a valid inferential pattern turns on whether or not the expressions employed carry the same sense and sense is not capable of syntactic individuation. The point can be expressed as follows.30 We know that the following is a schema for valid inferences: a is F a is G a is F and G To say that this is a schema for valid inferences is to say that just so long as each variable in the schema is replaced with an expression carrying the same sense, then a valid inference will result. The need for sameness of sense for the expression that substitutes for, e.g., each occurrence of ‘a’ is shown by the fact that mere syntactic identity is not enough due to the ubiquity of ordinary proper names. For example, from (13) George is honest (14) George is President it does not follow that (15) George is an honest president. The point is obvious, but it shows that although schema for valid inferences can be defined syntactically, what it is for something to instantiate such schema cannot be. This means that schema for valid inferences are transparent, but what it is for something to instantiate a valid schema is not transparent. This means there is always a point at which the issue of whether or not a sequence instantiates a valid schema turns on the contribution made by the perspective of the thinking subject to the individuation of

76


sense. This point is often taken for granted, or ignored, in most discussions, but ordinarily it goes unremarked that the identity of sense for ‘George’ in (13)–(15) is traded on in our normal dealings with words. We trade on and take for granted this identity of sense. We have to do this, for it is not something that can itself be stated, it is shown by how we use words. This is not to deny that there is a substantive issue about what it is to trade on the identity of sense; it is just to insist that such an account draws upon how we see the use of words, not on how we state they are used. For example, in the case of ordinary proper names – (13)–(15) – if the uses of the expression are deployed in single ongoing conversation we trade on this fact as constitutive of the transparency of sense. Ordinarily, unless it is explicitly remarked to the contrary, if someone repeats the word ‘George’ within a single conversation they are taken to be using it with the same sense. The point is quite general. There must be cases in which one trades on the identity of sense, for if it always had to be made explicit, all inferences would be enthymematic and we would be faced with an infinite regress. What constitutes the identity of sense is not definable syntactically, for it is constituted by the attention of the subject. This is clear in the case of proper names, for what constitutes the identity of sense through a conversation is the ongoing attention of the participants to the social exchanges going on. The point is undeniable in the case of perceptual demonstratives, for the sense of ‘this’ or ‘that’ is clearly dependent on the ongoing perceptual attention directed towards the object thought about. In short, sense is constituted by attention, the validity of inference is sense-dependent, therefore the validity of inference is attention dependent. How does this connect with the Wittgenstein’s question in §19? The point is that there is no essential and ultimate individuation of logical grammar, for the individuation of structure in virtue of which one sentence bears inferentially on another is attention-dependent. The unity of a thought depends on the point of view of the thinker. What makes predicate and subject fit together is that they do so within the point of view, the attitude of the thinker. Similarly with the transparency of inference. What permits us to trade on the identity of sense is that the individuation is dependent on the ongoing attention, the attitude, of the thinker. One representation of grammar may be more informative than others; there may be an epistemological advantage to one representation over another. There is, however, no such thing as the representation that gives the essential structure, the structure that, as it were, drives the validity of inference. The validity of inference is attentiondependent. The structure in virtue of which one sentence bears upon another is not, ultimately, a linguistic or syntactic structure. It is a structure of actions, including the actions of attending to the ongoing uses of words and the


77

aspects of use to which their inferential role is sensitive. This can include: perceptual coupling with things in the case of demonstrative thoughts; the social conventions governing naming in the case of proper names; information regarding the speaker, time and place of utterance in the case of token reflexives. The grammar of inference is a grammar of action and there is no privileged representation of this. If this is right, although we may find (6) more illuminating than the one word (7), what fundamentally matters is how expressions are used, not how they are composed as sequences of expressions. And that is why the claim that something is red all over rules out the possibility that the same object is green all over. It is not that independent atomic sentences are internally related, sentences do not connect up inferentially all on their own. Our role in the validity of inference is ineliminable. If we use (3) and (4) in such a way that the truth of one bears on that of the other, this is not enthymematic requiring (5) for the validity of the connection to be made good. Validity of inference trades on the ongoing attention-dependency on the subject as revealed in what we do with expressions. There is no account of the validity of inference and the underlying structure of use (logical grammar) to be abstracted from what we do with words and defined in terms of syntactic patterns. The patterns of logical grammar are ineliminably patterns of judgement, attention-dependent activities with words.

3.7 This and Similar Things are called ‘Games’ The thesis that grammar is perspectival has more general consequences. It runs to the heart of Wittgenstein’s mature conception of language and underpins his realism about grammar. This is a realism that embodies a critique of certain forms of generalism about reasons. The relation between our abilities for using words and our capacities for seeing things as similar lies at the core of what it is to be responsive to reasons. Wittgenstein’s account challenges some basic intuitions that underpin generalism about reasons. Classifying things matters. It is, typically, the key move in giving reasons and is often integral to knowing how to judge an event or person. Coming to a view about what to believe and how to behave hinges on how we classify things. If something is a reason it should be possible to defend its status as a reason. This is not to demand that one be able to give a reason for one’s reasons. That is too strong. But if something is a reason we should be able to show how it fits into patterns of reasons and show how what we respond to carries the normativity of reasons. A reason is not something we just happen to respond to, it is something we should respond to.31 The claim that x is a

78


game is a reason for action. Wittgenstein argues that there is no set of necessary and sufficient conditions that define the extension of the concept. No matter how we formulate the antecedent as conjunction or disjunction, there is no general rule, If x satisfies F1, F2, . . . Fn, x is a game. But then how does ‘x is a game’ carry the normativity of a reason? The standard move is to say a reason carries normativity by being subsumable under a rule. But with the concept ‘game’ we hit a dead end. Calling something a game matters, but we have no account in virtue of which it matters. In place of an account of why x is a game that fits the claim into wider patterns, we appear left with a mere assemblage of reminders that people have found various things to be games. Our reasons appear to turn floppy. This worry is unfounded. Wittgenstein’s critique of essentialism does not enjoin a floppy logic of reasons. It provides a rich metaphysics of reasons.32 There are at least three different ways of reading Wittgenstein’s remarks on family resemblance. One reading has Wittgenstein critiquing our ability to articulate the complex commonality of games.33 There is, however, no textual evidence for this. On either of the other readings – the negative and positive readings – Wittgenstein challenges the following intuition: where x and y are seen as similar, there must be some account in virtue of which they are similar.

On the negative reading, Wittgenstein shows that concept possession cannot be theorized such that the patterns of language use are independent of the seeing of similarities. Wittgenstein challenges us to articulate the respect in which games are similar and he offers the idea of family resemblance as an alternative account of concept possession. On the negative reading there is no defence to calling something a game, no normativity to our doing so, we simply do. Our grasp of the concept is merely described by logging the web of our uses of the word. The idea of family resemblance is descriptive. The account might serve a therapeutic role, but it offers no theoretical insight. It offers no account of what constitutes the normativity to the patterns of language use. The positive reading also denies that where we see similarities we owe some account of that in virtue of which things are similar. We cannot articulate the patterns which drive the seeing of similarities. But the critique of patterns that underpin seeing the similarities is not the end of the matter. In addition, Wittgenstein promotes the opposite point of view: the seeing of similarities


79

underpins patterns of language use. Wittgenstein is not merely critiquing conceptual essentialism but asking, ‘Which comes first – seeing the similarities or the patterns?’ His view is that seeing the similarities is primitive. The idea behind family resemblance is not that no pattern can be formulated in advance to underpin the seeing of similarities; neither is he making the negative point that there are no normative patterns. Rather, he advocates a model in which the seeing of similarities is primitive and underpins the normativity of patterns. Claiming that seeing the similarities is primitive is a theoretical result, and not merely descriptive, for it does not leave everything as it is. Treating seeing the similarities as primitive reconfigures our sense of the conceptual connections between our ideas of perception, conceptualization, the role of rules and the place of judgement. It requires a reconceptualization of the role of judgement in concept possession.34 Concept possession requires a grasp of the patterns of correct and incorrect use. Someone who did not grasp these patterns could not be said to understand concepts, for without patterns we would be bereft of any account of what it is to use words to give reason for our beliefs and action. The rational force of our language use consists in its exhibiting a combinatorial structure that patterns the connections between uses. The patterns provide the normative structure to word use. Wittgenstein’s central target is a form of realism about the patterns of concept application. It is the idea that what makes our actual use of a word correct/incorrect is that it conforms to patterns transcendent of use. At its extreme, this is a Platonic model in which the patterns are reified as abstract structures. Wittgenstein’s critique of generalism arises from rejecting the usual forms of realism about patterns of word use. To say that seeing the similarities between things is primitive is to say that the normative patterns of correct use of words emerge from the activity of seeing similarities. This is in contrast to holding that we see similarities in virtue of having grasped the transcendent pattern and applying it. This is not to deny patterns of correct use, it is only to deny that they pre-exist the activity of seeing similarities. Without patterns that stand over and above actual ordinary use, it can seem that we lose the normative idea that use is subject to standards of correctness. We should not let go of that normative idea. The patterns of correct use are not transcendent of actual use. They are immanent. Suppose we can make sense of the idea that the patterns of use emerge from instances of seeing the similarities. One such pattern will be the pattern of correct use of the concept game. There is a pattern, although not only is it not possible to formulate it in general terms, it is not constituted in anything other than our seeing particular things as similar. The pattern consists not in

80


matching an abstract structure, but in our ongoing capacity to see similarities. But now, what makes our use of ‘game’ fit into the patterns of reason-giving language? The answer is not that it fits a rule; it is that we see the similarities. Our sense of reason, as also our grasp of language patterns, is rooted not in general structures, but in our seeing of similarities in particular instances. The groundfloor account of the rational force of citing something as a reason consists not in it being subsumable under a rule, but in our seeing the similarity between the particular circumstances cited in the belief and that for which it is given as reason. As Wittgenstein remarks in a closely related context, ‘If there has to be anything ‘‘behind the utterance of the formula’’ it is particular circumstances, which justify me in saying I can go on’ (§154). I use the word ‘game’ in the context of drawing your attention to various features not to fit a rule, but to get you to see things as I do. The normative force of reason is in the seeing, not in a rule. We do not see general forms, we see things aright.35 Wittgenstein says the phenomena of language ‘have no thing in common which makes us use the same word for all, – but . . . they are related to one another in many different ways’ (§65). This is his rejection of essentialism, the view he had subscribed to in the early writings. He rejects the idea that if we are using one word for many different things there must be that which makes us use it that way. The idea that there might be something that ‘makes’ us use the word as we do acknowledges our sense that our use is subject to standards of correct use, that which cannot be avoided in using words. Wittgenstein is not saying there is no such thing as standards – that which makes us use the same word; he is simply saying that there is no one thing that constitutes the standards. It is tempting to read the game example and the idea of family resemblance in one of two ways. Either it is a homily to the difficulty of expressing the complex commonalities of games or it is offered in lieu of having anything substantive to say about the standards of word use. The first reading does not fit the text. On the latter, it seems that the standards have been rendered floppy. In place of the essentialism of Platonistic standards of word use, we are left with the descriptivism of homely reminders of the complex multi-faceted way in which words are actually used. Such a ‘grammatical investigation’ amounts to little more than a catalogue of the contingencies of word use. What makes it all hang together? What makes word use correct or incorrect? What and where is that which supplies the standards to actual deployment of words?36 In place of demanding Platonistic standards, we get a limp empiricism that simply notes when, where and under what circumstances words get used. Without an account of what makes word use hang together, Wittgenstein’s position would be barely stable. Wittgenstein’s position is stable, for he


81

has an account of what makes word use hang together. In place of the usual suspects, transcendent patterns that invariably lead to Platonism or naturalistic causal patterns, there is an alternative.37 In §69 Wittgenstein considers how we should teach someone the word game: I imagine that we should describe games to him, and we might add: ‘This and similar things are called ‘‘games’’ ’. And do we know any more about it ourselves? Is it only other people whom we cannot tell exactly what a game is? – But this is not ignorance.

The point is familiar from the later discussion of rule-following. It concerns how, in teaching a concept, we reach a stage where the explicit instructions give out and the pupil ‘catches on’. It is the point where we appeal to the idea of ‘similar things’. If you assume that where there are standards to correct use they should be capable of articulation, such points in concept explanation will seem incomplete or indirect gestures. There is something you know about the meaning of ‘game’, but you have exhausted the resources for articulating it and you have to resort to the ‘This and similar things . . .’ locution.38 It can then seem that it is only other people whom we cannot tell exactly what a game is. We know well enough ourselves, but the others get a gesture, a loose exercise in arm-waving that roughly indicates but cannot fully specify what ought to be said. This use of the gestural ‘This and similar things . . .’ indicates then a form of ignorance, for the audience is at risk of not catching on. This is a mistaken interpretation of Wittgenstein. Wittgenstein explicitly says that it is not ignorance if the audience is not given explicit instructions. He does not actually answer his first question ‘do we know any more about it ourselves?’ but the implicit answer has to be ‘No’, for it is not a form of ignorance to fall back on the ‘This and similar things . . .’ construction. The idea of knowing of some activity that it and similar things are called ‘games’ is not a second-best indirect piece of knowledge. It is not something that stands in for the better explicit statement. It does not stand in need of supplementation. The ‘This and similar things . . .’ construction is not a placeholder, it’s as good as it gets. It is not an incomplete articulation of what we know; it is what we know. Being able to see that one activity is similar to another and is therefore a game is not something that stands in need of further analysis. It is primitive. There is no reason behind our use of words for classifying something as a game. The reason lies in our seeing similarities. We cite examples and, as we become competent with the concept, we learn to see activities as similar. We see one activity as like another and call it a game. But the seeing of the similarity cannot be underpinned by an explicit rule, there is none. Seeing the similarity is primitive.

82


This reading of §69 is confirmed directly in §71. Wittgenstein formulates the point slightly differently, but the continuation is clear; One gives examples and intends them to be taken in a particular way. – I do not, however, mean by this that he is supposed to see in those examples that common thing which I – for some reason – was unable to express.

Two things stand out. First, there is the repetition of the claim that there is no common essence that constitutes the meaning of the concept. Second, and more importantly, the meaning being explained is not something that he, the teacher, was unable to express. The idea of the examples being taken in a particular way is not a gestural half-expressed explanation, for the passage continues, ‘Here, giving examples is not an indirect means of explaining – in default of a better’, and he adds, prefiguring ideas that will dominate later passages, ‘For any general definition can be misunderstood too’. We explain meaning directly and clearly without loss by giving examples and using the ‘This and similar things . . .’ locution. Only in the grip of conceptual essentialism does it looks like we are being short-changed when our teacher resorts to examples and says, ‘This and similar things . . .’. But conceptual essentialism has been rejected. The use of ‘This and similar things . . .’ is not a second best, it is the bottom line primitive to concept possession. Wittgenstein still endorses a realism about the patterns of word use. It can look like he is supporting an anti-realism in which the patterns of word use are floppy and from which the standards of correct use evaporate. On such a reading, the normativity of the patterns of word use gets replaced with a recording of the contingencies of speech. My reading is different. It accommodates the normativity to the patterns of word use. There are correct and incorrect ways of using words. What Wittgenstein rejects is the account of what such patterns of use consist in that commonly tempts us when we think about these things. The notion of correct use does not consist in transcendent patterns, things that could in principle be articulated as general rules and in virtue of which our particular seeing of two things as similar could be justified. Instead, our seeing of things on a particular occasion as similar is the source of the normative patterns of word use. The patterns of language use follow the seeing of the similarities, not the other way around. This can look like a refusal to take seriously the issue of what the patterns of correct use of words consists in. We are often tempted to look for something transcendent to use which guides us and which we observe and respect in our linguistic practice. Following that train of thought leads to Platonism and the paradox of the rule-following arguments. It might seem that the option I advise looks equally non-compelling, for it looks like no answer at all; it is


83

just the claim that we see similarities and nothing more can be said. One way of dispersing this unease is to provide a model that captures the structure of my reading of these passages.

3.8 Spontaneity in Particular Circumstances Wittgenstein’s position produces a shift in the resources deployed in giving an account of conceptual content. To get critical leverage to word use, we must acknowledge some form of realism regarding the patterns of word use. It cannot be enough simply to log the facts about how words are used. To see the patterns of word use emerge from particular instances of seeing the similarities between things looks, however, as if it can offer no account of the critical leverage of word use. For what could stand over and above our seeing of similarities that would provide the critical leverage? This question only looks pressing if we assume that critical leverage is forced by subsumption under general rules – an application of a word is criticized just in case it is out of line with the rule. But that cannot be Wittgenstein’s position, for in many cases there is no general rule. The seeing of similarities in particular instances must then be self-authenticating. It must be possible that the particular instance of seeing the similarities can stand in need of no further justification. There are two points here, an epistemological one and a constitutive one. The latter is the more important. I take the epistemological point first. The idea that there is nothing more primitive that could be said in justification of seeing things as similar is not surprising, ‘To use a word without justification does not mean to use it without right’ (§289). Our giving of reasons has to stop somewhere, for we need to take something as given in order to get the giving of reasons off the ground. So to use a word to classify a thing as like another may be to use it without justification in the sense that there is nothing more primitive that one can say other than that the two things look similar. The fact that one has nothing more primitive to say is not to admit that one uses the word without right. Indeed, one’s right to use the word (the fact that one is using it correctly) is shown by the fact that one uses it of two things that are similar. That is to use it with right, but it is not to give a justification, for there may be no further articulation, one has reached the point where the other has to see the similarity that you see. The constitutive point about the self-authenticating character of seeing the similarities in particular instances is harder to grasp. The giving of reasons for what we do in using words can come to rest in a seeing of similarities, rather than in patterns that underpin the seeing of similarities. What is difficult

84


about this idea is not the epistemological concern about whether that can ever be justification enough, the point at issue is, ‘How can such seeing of similarities ever be constitutive of what it is to give a reason?’ This question looks pressing because of the passive role we give ourselves in an account of the rational force of language use. We take the rational force of language to consist in its structure, a structure which we inhabit. The speakers are not central stage. The rules and structure are what matters. Wittgenstein’s argument reconceptualizes our role in the metaphysics of reasons.39 A simple way of expressing the active role for language users is to say they are judges, active users of language with a capacity for judgement. We make mistakes with words and the notion of mistake requires some notion of that against which our use is measured. The model that Wittgenstein rejects sees our mistakes as events to be assessed for how well they match a pattern that could be formulated independently of our word use. Wittgenstein’s critique flows from the observation that even of an explicitly articulated rule there would still be space to interpret the rule, it would still stand in need of further articulation. That move takes away the idea of the rule that our use of words must match in order to be correct. The role for the notion of interpretation in this critique not only provides the argument against the concept of rule-as-measure, it also marks what the alternative account involves. The role of interpretation in the critical argument shows that the language user is not merely a passive follower of rules. The language user brings something to the account of meaning, they bring the activity of interpretation. We only need the idea of the final ultimate interpretation – the fully articulated pattern that covers all eventualities – if we are impotent in our language use, playthings of the structures picked out by the full articulation. The alternative is not that the patterns of meaning are rendered floppy with objective meaning lost, but that the patterns of meaning are real structures in the ongoing practices of judges. The Platonist has rules in charge, the language user follows rules. On my reading, the language user is in charge, for it is the language user with the capacity to see similarities, to make wise discriminations and find saliences in things that is the source of the patterns of language use. This is still realism about patterns, but the condition for the possibility of patterns of correct use is not the existence of transcendent patterns, it is the existence of active language users, judges with a capacity to see similarities in things. An example from a different domain illustrates the logic of the account that I am suggesting. On standard approaches in AI if you want to build a robot that performs to a pattern you have to write an explicit rule in the program that generates the pattern. Behaviour is patterned because it is produced by the pattern in the program for the behaviour. In the new paradigm of situated cognition


85

in AI, building a robot that behaves in patterns does not require formulation of the pattern in the program. Instead, the basic idea is to build a machine with couplings with the environment such that, in the appropriate environmental niche, patterned behaviour results. A simple example illustrates the point.40 Suppose you want to build a wall-following robot. On a traditional AI approach the task is not simple. You need to define what a wall is, give the robot the resources to identify walls and map their position, respond to their corners, curves and doorways. You need to encode everything the robot needs to know about walls in an explicit database and give it a set of explicit instructions for action with respect to walls. On this model, perception produces a discrete cognition sequence the upshot of which is wall-following action. A clear distinction can be drawn between perceptual inputs, their manipulation in cognition and the outcomes in motor control. The environment is the source of the information that gets delivered to cognition and it is the theatre in which the robot acts. The environment is insulated from cognition, for it has no role to play in the robot’s computations. The situated robotics paradigm challenges this environmental insulation. It provides a more efficient computational system that has some powerful conceptual consequences. The situated robotics approach has a simple way of building a wallfollowing robot. First, the robot has motor control – it can move. Second, its motor control gives it a tendency to veer slightly to the right. Third, it has a sensor on its right front bumper which, when activated, produces a momentary steer left. That is a wall-following robot. The key to this device is the following conditional claim: If the environment has a wall in it, the robot will find the wall and follow it. Providing the environment obliges with a wall, a wall-following pattern will emerge in the robot’s behaviour. There is no rule in the robot’s instruction set that mentions walls, no entry defines walls in its databases and no procedure is written for identifying walls. If there is no wall in its environment, the robot will proceed in a circular trajectory. For this robot, the environment is not merely the supplier of information and the theatre for action; it is a constitutive element in cognition. You cannot individuate the machine’s computational states independently of its ongoing adaptation to its environmental niche. Its inputs do not provide information that then enters the cognition box. Perception provides a coupling that enables a continuous dynamic adjustment to its orientation in the environment. Robot and environment

86


are coupled in one seamless dynamic system, the target stable state of which is that the robot follows the wall. The example is simple and the device crude, but it illustrates important ideas. First, patterned behaviour does not have to be driven by a rule for the pattern. Patterns emerge from choice couplings with the environment. Second, the robot exhibits novelty and creativity at the most basic level of cognition. The robot has a capacity to follow walls. The extension of the wall-following capacity is response-dependent and open-ended. No limit is placed on what sorts of things elicit the wall-following response. Response-dependence does not mean that the claim that something is a wall is not objectively true or false, it just means that which claims are true and false cannot be determined in advance of the response. There is, however, a fact of the matter regarding whether something is a wall or not; it is the fact of whether or not the robot’s response is elicited by the environmental condition. There are a number of points of contact between this simple example and the conceptual structure I attributed to Wittgenstein. Let me remark on just one.41 I argued that the notion of seeing the similarity should be taken as a primitive and that the idea of a classification is posterior to, not antecedent to the seeing of similarities. The way I now want to express this idea is to say that the capacity to see similarities is a capacity for coupling with things. The claim that the classification follows the seeing the similarities is then analogous to the claim that the extension of the robot’s wall-following capacity is determined by its responses in coupling with the environment. Things get to be in the extension of ‘. . . is a game’ by eliciting the appropriate response from our couplings when we see similarities between events. It is not that there is a pattern of classification that could be articulated prior to our particular instances of seeing similarities. Such instances are cases in which we perceptually couple with things and respond in a way that finds similarities. Our capacities for seeing similarities give rise to patterns in our responses and a real pattern emerges in what we call games. The point to treating the seeing of similarities as couplings is, as with the robot, to accommodate novelty and creative additions to the extension of the concept at the groundfloor account of concept possession. Possessing the concept ‘game’ requires a capacity for coupling that sees similarities in things and that is open-ended. Until the couplings are achieved there is no prior formulation of what will definitively count or not-count as a game. But this is not because there is a moment of trepidation where we have to decide ab initio whether to count the new event as a game or not.42 The absence of a general rule is not a leap in the dark; it is simply the space in which we have yet to achieve an appropriate coupling with things.43 The analogy with the robot case is not precise, and I do not intend it to suggest any scope for a reductionist account of concept possession. What I am


87

calling ‘coupling’ in the conceptual case is a fully intentional encounter with things, it is not merely a causally triggered response as in the robot case. The point of the analogy is just to press the general claim that patterns emerge from particular couplings with the environment, they do not precede our encounters. That general claim then shows why it is not an empty gesture to say that seeing similarities has priority over classifications, for if the seeing of similarities is treated as a coupling with environmental particulars this can be explanatory of how patterns of classification emerge from such couplings rather than precede them. What this discussion supports is an idea that can otherwise seem quite puzzling; it is the idea that spontaneity – the exercise of creativity – can be an intrinsic core feature of concept possession. Creativity is part of the groundfloor account of concept possession. It is not a later add-on. It is intrinsic. This idea looks difficult against a backdrop in which the realism of grammar – the patterns of correct use – is theorized in terms of the existence of a structure that we inhabit and follow in our actual use. This is the familiar picture in which grammar is thought of as an abstract structure, a rule-like configuration that is capable of individuation independently of what we actually do with words. It is the structure in virtue of which what we do with words is correct or incorrect. But that model is seen time and again to be flawed. It is flawed not because realism about grammar is mistaken, but because the metaphysics of grammar to which we have repeatedly been drawn is a metaphysics that neglects our role in an account of grammar. On the mistaken model freedom can only come after what we know about word use, for our knowledge of word use is conceived as a constraint, something that binds and determines all instances. In contrast, Wittgenstein’s thinking shows how freedom can be intrinsic to what we know about correct word use. The shift is from a model in which freedom only arises outside the boundaries of what we know – we become free in word use through ignorance, to a model in which spontaneity is intrinsic to what we know about word use. Consider the following question, ‘What is the cognitive transfer from one correct use of the word ‘‘game’’ to another?’ The answer that Wittgenstein argues against is the answer that says that the cognitive transfer, what we know in each case of correct use, is what games have in common: our use of the word ‘game’ is correct when it is guided by knowledge of that which they have in common. Getting to be competent with the word is getting to acquire this knowledge, it is what transfers from one use of the word to another and makes them correct uses. What is important about my reading of Wittgenstein is that although he rejects this answer, he does have an answer to the above question. What all games have in common is the way we see them! And because the seeing of similarities is primitive, it cannot be analysed further

88


into a technique of applying a prior rule that articulates the common characteristics of games. Because seeing the similarities is primitive, it is not a technique, it is the exercise of judgement. The patterns of correct use emerge from particular judgements in particular cases: If there has to be anything ‘behind the utterance of the formula’ it is particular circumstances, which justify me in saying I can go on – when the formula occurs to me. (§154)

The rational force of language use, that which makes it inferentially connected, consists in the particular circumstances of use, for it is there that we exercise our capacity to see similarities. Seeing things as similar is what justifies my calling this a game. You do not need to see the overarching rule as an abstract structure linking the one case with another. Grammar is perspectival. You see the similarity in the particular circumstances. Justification of correct use stops here. This is how we play the game.

Notes 1

2

3

4

5

For example, J. McDowell, ‘Wittgenstein on Following a Rule’, Synthese, 58, 1984, 325–63, and, ‘Meaning and Intentionality in Wittgenstein’s Later Philosophy’, in The Wittgenstein Legacy. Midwest Studies in Philosophy, vol. XVII, pp. 40–52, Notre Dame, Ind.: University of Notre Dame Press, 1992. In the later essay (p. 49) McDowell acknowledges that his use of ‘custom’ and ‘practice’ in the 1984 piece did not provide sufficient details on what these things are. Getting clear about practice is one of the central aims of this book. Crary is also clear on this point: A. Crary, ‘Wittgenstein’s Philosophy in Relation to Political Thought’, in The New Wittgenstein, eds A. Crary and R. Read, London: Routledge, 2000, pp. 118–46. Cavell is clear on Wittgenstein’s anti-intellectualist account of meaning, e.g. S. Cavell, ‘Excursus on Wittgenstein’s Vision of Language’, in The New Wittgenstein, eds A. Crary and R. Read, London: Routledge, 2000, pp. 21–37. The rejection of the idea that there is such a thing as the essence of language is not necessarily the rejection of the external view on language, contra Crary, op. cit. Wittgenstein rejects the idea of an external viewpoint in the Tractatus but still endorses essentialism. ‘That the justification and explanations we give of our language and conduct, that our ways of trying to intellectualise our lives, do not really satisfy us, is what, as I read him, Wittgenstein wishes us above all to grasp.’ Cavell, op. cit., p. 26. This is the sort of position that Crary criticizes, one that provides a notion of tradition, custom, etc., which gives no critical leverage on use. All we get are remarks on the ‘place a bit of language has in our lives’. Op. cit., p. 119. See G. P. Baker,


6 7 8 9

10 11

12

13

89

‘Wittgenstein on Metaphysical/Everyday Use’, The Philosophical Quarterly, 52, 2002, 289–302, for a clear account of why ‘ordinary use’ is not a privileged concept for Wittgenstein, it is not a concept that is positively promoted in place of ‘metaphysical use’. Wittgenstein does not offer ordinary use as a preference in contrast to metaphysical use; it is a placeholder for the reaction to the problematic metaphysical use. S. Kripke, Wittgenstein on Rules and Private Language, Oxford: Blackwell, 1982. Seeing connections is a form of seeing-as. See §7 below and chapter 6. I am indebted to Simo Saatela for this point. In §95 Wittgenstein says, ‘When we say, and mean, that such-and-such is the case, we – and our meaning – do not stop anywhere short of the fact; but we mean: this-is-so.’ There is no gap between what we mean and what is the case. Of course, what is puzzling, Wittgenstein calls it paradoxical (ibid.), is that sometimes there is a gap, for we have false thoughts. It is the latter point that has always made it seem that there can be no direct calibration of language on things. That thought is simply wrong. It depends on identifiable and contentious epistemological assumptions about the nature of experience. I consider these points in detail in the next chapter. See J. McDowell, Mind and World, Cambridge, Mass.: Harvard University Press, 1994, p. 27, for discussion of §95 where McDowell points out the similarity between Wittgenstein’s move and the import of Davidson’s denial of the scheme/content distinction. McDowell (p. 28) also attributes the idea of direct calibration to Wittgenstein’s early theory of language in the sense that there is no priority between language and world in giving an account of logical grammar. That, of course, is the position I argued for in the previous chapter. On the idea of direct calibration in the case of singular terms, there is a clear model for this in contemporary philosophy of thought in neo-Fregean theories of singular thoughts, cf. M. Luntley, Contemporary Philosophy of Thought: Truth, World, Content, Oxford: Blackwell, 1999, for details and also how such a theory can accommodate error – the bit that Wittgenstein calls paradoxical. See also M. Luntley, ‘The Practical Turn and the Convergence of Traditions’, Philosophical Explorations, 1, 1998, 10–27, on the latter point. Note the divergence from Kripke’s celebrated reading of Wittgenstein; there is more to the notion of facts than Kripke ever allows. It is also one of the hardest issues in getting to grips with McDowell’s interpretation of Kant: see J. McDowell, ‘Having the World in View: Sellars, Kant and Intentionality’, The Journal of Philosophy, XCV, 1998, 431–91. It is the point at issue when Cavell speaks of the stability and tolerance of concepts, op. cit., p. 34. Compare Pears’ critique of the idea that Wittgenstein had a use-theory of reference in his early period: D. Pears, The False Prison, vol. I, Oxford: Clarendon Press, 1987, p. 100. For example, E. Anscombe, An Introduction to Wittgenstein’s Tractatus, London: Hutchinson,1971, and M. Black, A Companion to Wittgenstein’s Tractatus, Ithaca, NY: Cornell University Press, 1964. In contrast, both Ishiguro and McGuinness take this as part of the evidence for a use-theory of reference: H. Ishiguro, ‘The use

90

14 15 16

17 18

19

20

21 22 23

24

This is How We Play the Game and Reference of Names’, in Studies in the Philosophy of Wittgenstein, ed. P. Winch, London: Routledge, 1969; B. McGuinness, ‘The So-Called Realism of the Tractatus’, in Perspectives on the Philosophy of Wittgenstein, ed. I. Block, Oxford: Blackwell. 1981. As argued in the previous chapter, I do not think Wittgenstein’s early theory was inferentialist. Cf. Crary. op. cit., p. 119. Pears, op. cit., treats Wittgenstein’s early work as a version of this; the structure of language is a response to the structure of the world. It is not clear to me whether McGinn’s reading is open to this critique: M. McGinn, Wittgenstein and the Philosophical Investigations, London: Routledge, 1997. Although McGinn’s emphasis on the idea of a ‘grammatical investigation’, in the absence of a clear account of what that means, leaves her reading vulnerable. See R. Brandom, Making it Explicit: Meaning, Representing and Discursive Commitment, Cambridge, Mass.: Harvard University Press, 1994, chapter 1. See D. Davidson, ‘On the Very Idea of a Conceptual Scheme’, in his Essays on Truth and Interpretation, Oxford: Oxford University Press, 1984; ‘A Coherence Theory of Truth and Knowledge’, in Truth and Intrepretation: Perspectives on the Philosophy of Donald Davidson, ed. E. Lepore, Oxford: Blackwell, 1986, pp. 307–19. For an account of whether Davidson provides an adequate argument for the denial of scheme/content dualism germane to current concerns, see J. McDowell, ‘Scheme-content Dualism and Empiricism’, in The Philosophy of Donald Davidson, ed. L. Hahn, La Salle: Open Court, 1999, pp. 87–104. See also Crary, op. cit., for the idea that Wittgenstein is opposed to the idea of external standards to word use. This is clear in the sociological accounts of ordinary use in which the subject is lost in the social structure, cf. D. Bloor, Wittgenstein, Rules and Institutions, London: Routledge, 1997 For example, M. McGinn, Wittgenstein and the Philosophical Investigations, London: Routledge, 1997. Although she finds, correctly, a number of important themes in Augustine that Wittgenstein critiques, McGinn presents Augustine as holding a naming theory. I shall argue later that this is not the correct way of understanding the alternative, but it suffices for now. See chapter 5. This is then a Cartesian account of the source of grammar. S. Kripke, Wittgenstein on Rules and Private Language, Oxford: Blackwell. 1982. It is acceptance of Kripke’s point that then motivates communitarian accounts of the normativity of use by writers such as Bloor and Williams: D. Bloor, Wittgenstein, Rules and Institutions, London: Routledge, 1997; M. Williams, Wittgenstein, Mind and Meaning: Towards a Social Conception of Mind, London: Routledge, 1999. Cavell raises this thought. He says that in reading this passage we seem to miss ‘a sense of understanding in them’; they seem to be ‘behaving mechanically’. S. Cavell, ‘Notes and Afterthoughts on the Opening of Wittgenstein’s Investigations’, in The Cambridge Companion to Wittgenstein, eds H. Sluga and D. Stern,


25 26 27 28

29

30 31

32

33

34

35 36

91

Cambridge: Cambridge University Press, 1996, at pp. 290–91. Note, Cavell is not endorsing the view that Wittgenstein’s description is behaviourist, he is only noting that it is tempting to take it that way. The temptation should, however, be avoided. Cavell, 2001, op. cit., p. 28. Ibid. Cf. chapter 5 for more on this. The point about truth that I appeal to here is what Wiggins calls the third mark of truth. Cf. D. Wiggins, ‘What Would Be a Substantial Theory of Truth?’, in Philosophical Subjects: Essays Presented to P. F. Strawson, ed. Z. van Straaten, Oxford: Clarendon Press, 1980, pp. 76–116, and ‘Truth as Predicated of Moral Judgements’, in Needs, Values, Truth: Essays in the Philosophy of Value, Oxford: Blackwell, 1987, pp. 139–84. The issue about the transparency of inference was first raised by J. Campbell, ‘Is Sense Transparent?’, Proceedings of the Aristotelian Society, 88, 1988, 273–92. Further discussion extending the point can be found in M. Luntley, 1999, op. cit., and M. Luntley, ‘Agency and our Tacit Sense of Things’, in Library of Living Philosophers volume for Marjorie Grene, ed. L. Hahn, La Salle: Open Court, 2002, pp. 151–71. Here I follow Campbell, op. cit. As such, there is always scope to defend a reason and say more, cf. M. Luntley, ‘Postmodernism and the Education of Character’, in Education in Morality, eds H. Halstead and T. Mclaughlin, London: Routledge, 1999, pp. 185–205. The metaphysics of Wittgenstein’s account of reasons illuminates the contemporary debate about generalism versus particularism regarding moral reasons. I do not think that Wittgenstein’s position is a particularist one, rather it is a position that shows that the terms of that debate are probably ill-defined, even though it perhaps captures much of the spirit of contemporary particularism. See Moral Particularism, eds B. Hooker and M. Little, Oxford: Clarendon Press, 2000, for a useful collection of papers on this, many of which touch on Wittgensteinian themes. Cf. F. Jackson, P. Pettit and M. Smith, ‘Ethical Particularism and Patterns’, in Moral Particularism, eds B. Hooker and M. Little, Oxford: Clarendon Press, 2000, pp. 79–99, at p. 83, for a clear endorsement of this view. Properly understood this bears directly upon the debates about particularism in ethics. See J. Garfield, ‘Particularity and Principle: The Structure of Moral Knowledge’, in Moral Particularism, eds B. Hooker and M. Little, Oxford: Clarendon Press, 2000, pp. 178–204. Garfield says that what is learned by the ethically competent are ‘discriminations based on similarity relations to paradigm cases’ (p. 188). This is right, but without a view on which comes first – seeing the similarities or the patterns, the remark about discriminative capacities cuts no ice. Recall the point from chapter 2 – we do not see the general form, the ‘and so on . . .’. M. McGinn, Wittgenstein’s Philosophical Investigations, London: Routledge, 1997, leaves unclear what the force is of referring to Wittgenstein’s method as a

92

37 38 39

40

41

42

43

This is How We Play the Game ‘grammatical investigation’. Is it just a description of word use? Does such an investigation offer any critical leverage to word use? If so, how? Critiquing Platonism is not a reason for not answering such questions. Jackson et al., op. cit., treat the patterns naturalistically. This way of understanding Wittgenstein is explicit in Jackson et al., op. cit., p. 82. Failure to understand this reconceptualization stands in the way of a proper appreciation of the issues that underlie particularism. The point here underpins the move in ethics when the fundamental question is no longer ‘What should I do?’, but ‘What sort of person should I be?’. Having the virtues that come from being able to discriminate similarities is not necessarily to latch onto what is already there. It is to be, in part, the author of the patterns of similarities. The example is from A. Clark, Being There, Cambridge, Mass.: MIT Press, 1998; see M. Luntley ‘Attention, Time and Purpose’, Philosophical Explorations, VI, 2003, 2–17, for further related discussion. A further key point concerns the account of practice that one ascribes to Wittgenstein. I argue that practice can be treated in a minimal non-social way by treating it in terms of our ongoing engagements with things: see chapters 4 and 5. Crispin Wright’s Wittgenstein is a theorist of this moment of trepidation: C. Wright, Realism, Meaning and Truth, Oxford: Blackwell, 1986. I disagree. Cf. also Jackson et al., op. cit., p. 88, for a similar point against the idea that a new decision is required at each step. Cf. Garfield, op. cit., for an example about the difference in status of the rule for a goal and the rule for a mark in Australian football that would benefit from the analysis I give. Garfield’s example of the latter rule is supposed to be one for which no technique can be defined. This seems right, but the basis for this claim that Garfield gives simply draws upon the difficulty in formulating the rule. My analysis provides a model that explains why such a difficulty exists and, indeed, why the full formulation is impossible. This is one respect in which Garfield’s insistence that particularism is an epistemological thesis and not an ontological one weakens the case against generalism.

CHAPTER FOUR

Rules and Other People

4.1 Introduction Wittgenstein’s central challenge is to replace the impulse to find a theoretical account of what makes judgement possible. In place of a theoretical statement of the possibility of judgement, we must learn to accommodate the idea that judgement is made possible by the subject, an active judge seeing the world aright. This is part of what it means to say that it is practice that makes judgement possible. This move is central to the rule-following arguments in the Philosophical Investigations. These arguments provide a reductio ad absurdum of the bipartite model of meaning. Getting clear about the ideas that underpin the bipartite model enables a distinction between different notions of practice, one of which is, in a useful sense, a collectivist account of practice. I shall argue that the account of practice required to escape the rule-following paradox is not the collectivist concept. The account of practice that I give is a minimalist one. The core thesis to my concept of practice is: Practice ¼ a use of signs in which the speaker is engaged with things. This is a use that cannot be characterized independently of the truth-conditions that provide the normative standards of correct/incorrect use. Practice minimalism gives an account of practice without mentioning other people. I take as a working hypothesis that the role of others is one of enabling or scaffolding for our engagement with things. It is engagement with things that constitutes practice, not engagement with others; the latter supports the former but is not constitutive of practice. One reason for endorsing practice

94


minimalism is simply the methodological prescription to keep our analytical tools as simple as possible. It is always instructive to see how much can be achieved with slight resources. The more important reason for practice minimalism concerns the integrity of the idea of standards of correct/incorrect use. There is a danger that this idea is allowed to elide into a notion of consensus when practice is understood in a social sense. The views of others, in general the community, are useful resources against which to check the correctness/incorrectness of our use of signs. But the reason why checking our sign use against the community’s use of signs is a good thing is because we take it that the community’s use is checked against things and standard objects.1 In short, practice minimalism is recommended to ensure that the calibration of sign use is constitutively a calibration against things, not a calibration against further sign use. Calibration against further sign use would only give a deferred concept of error. It is the way our sign use manifests a being in the world that matters, not a being with others.2 There is a danger that practice minimalism is banal. In place of a theoretical account of the conditions for the possibility of judgement, we get reminders of our common-sense grip on our word use. In place of a theory, it seems that Wittgenstein offers what appears to be little more than a reminder, a description of how things are. But that does not seem to be enough. The key to practice minimalism that avoids this danger is the conception of the subject as judge. The core concept of the subject that is needed here is the self-as-will. Once the reconfiguration of the subject as a self-as-will is acknowledged, the reminders of how things are with word use fall into place as a substantive and complete account of how judgement is possible.

4.2 The Structure of an Argument Much of the literature on rule-following has it that Wittgenstein produced an argument that generates a paradox. The literature has then focused on the issue of whether the way out of the paradox requires an individualist or collectivist account of the norms of meaning.3 I think that this is a diversion from the real issue. Here is a way of representing the outline form of the paradox that Wittgenstein produces in sections 183–242.4 Suppose that the meaning carried by using an expression (saying it, writing it) is a function of the sign plus an interpretation placed upon it. For the meaning to determine a standard of correct/incorrect use, the interpretation that grants the sign meaning must be stable. But any given interpretation can only be supplied by using signs. This means that such further sign deployment also stands in need of an interpretation if the meaning of the first sign is to be


95

stable. The examples of ‘bent’ rules from §185 onwards show that any given use of a sign is potentially unstable. The paradox presents us with a dilemma, neither horn of which is tenable: if there is stable meaning (standards of correctness), then we are presented with an infinite regress of sign deployment in giving an account of meaning. Alternatively, we avoid the regress only on pain of having no account of the normative standards of correctness/incorrectness. On the second horn, all we are left with is the sub-bedrock of normless sign display. The argument that I want to focus on in this can be represented as an argument that threatens to push us into the second horn of the dilemma, the loss of the idea of normatively configured meaning. The argument can be represented thus: A Suppose meaning is a function of sign display (what you say) þ interpretation Nothing you say can fix the interpretation There is no such thing as determinate meaning. The second premise holds because of the threat of regress on the first horn of the dilemma. If that is the structure of the argument leading to the second horn of the dilemma, the conclusion can be avoided by challenging the second premise and thus removing the sting from the first horn of the dilemma. The argument then becomes: B Suppose meaning is a function of sign display þ interpretation Nothing you say can fix the interpretation It is not what you say that determines meaning, it is what you do. B provides a framework in which to position the dominant readings of the rule-following arguments. It provides a natural site for what is seen by many to be the critical issue. If it is what you do that determines meaning, it is natural to focus attention on the nature of the subject, the ‘you’. The whole business of individualist versus collectivist accounts of the rule-following argument can then be captured in this framework. Can an individual do enough without the assistance of collective subjects to sustain the normativity of meaning? For many that has seemed to be the central issue in avoiding the paradox by dodging the regress threatened on the first horn of the dilemma. I think this is a mistaken focus. The focus should be, not on the notion of

96


what you do, but on what you do. That is to say, it is the nature of the ‘doing’ that matters. To put this another way, understanding is a practice, but what matters about the concept of practice is not centrally whether the practitioner is singular or plural, it is the nature of practice itself. In short, on an impoverished conception of ‘doing’, it is tempting to recover normativity by embedding the doer in a community of doers. My account rejects this by focusing on a conception of doing in which normativity is intrinsic. A proper appreciation of this point reveals that the real culprit in the argument is not premise two, but premise one, the idea that meaning is a function of two elements. This is the bipartite conception of meaning. My account is strongly individualist, but this flows from a proper account of the ‘doing’ that goes into the concept of practice. The bipartite conception holds that meaning is a function of two parts – a component that is normless and a component that supplies the normativity. Premise one looks compelling for it gives voice to this underling bipartism about meaning. The use of signs is a display of something that is intrinsically normless. Simply saying, or writing, a sign – such as the string ‘Add 2’ – of itself has no normative force. It is a display that needs to be placed next to that which calibrates it as correct/incorrect. The display of a sign does not of itself provide the normativity of the idea that there are correctness/ incorrectness conditions for use. If this is right, then in order to secure the normativity of meaning something else must be added to sign display. Premise one merely locates interpretation as the extra something required to supply normativity. What argument B suggests, as standardly presented, is that no interpretation can succeed in providing the normativity. The reason is simple. Any interpretation will amount to no more than further sign display. If sign display is, in itself, normless, no amount of further sign display will generate norms. In the face of that sort of consideration, it can then look tempting to suppose that in place of an interpretation, an appropriate account of the ‘you’ in the conclusion to argument B will provide normativity. It has been argued, and the point follows the text closely, that the culprit in all this is the assumption that all grasp of meaning depends on interpretation.5 But that leaves the focus of the argument presupposing the bipartite account of meaning. If it is the idea of interpretation as the source of normativity that is the culprit, that leaves it open to respond by finding some other component as the supplier of norms to add to the normless sign display. And it is here that the issue about singular versus plural subject looks pressing. My suggestion that it is the nature of the ‘doing’ that is the real focus of the conclusion provides an alternative account of the argument structure. It amounts to the thought that it is not that the concept of


97

interpretation is problematic in so far as it is the wrong ‘extra bit’ for the supply of norms; the point against the idea that meaning always depends on interpretation is rather a point against the very notion of the bipartite account of meaning. It is not that something else, other than interpretation, might provide the norms to normless sign display; it is that meaning should not be seen as a function of a normless bit and a norm-adding bit in the first place. But that is to require that the notion of sign use turns out to be not normless after all. That is, I think, a difficult concept to grasp. But it is that issue that comes into view by insisting that the focus in the conclusion to B should be the notion of ‘doing’ not the nature of the subject. What the notion of ‘doing’ has to pick out is a notion of practice in which norms are immanent. If the ‘doing’ is to stop the paradox, then the notion of practice picked out here cannot be something that is merely descriptively characterized; it must be intrinsically normative. It is not an empirical concept; it is a transcendental concept. It is not something that just happens to exist and be contingently related to the existence of intentionality; it is a condition for the possibility of intentionality. That throws the following question into centre stage: How can norms be immanent to practice? That is the question on which I want to concentrate. I believe that the correct reading of the rule-following argument is one in which paradox is avoided only if the idea of norms immanent in practice is accommodated. Furthermore, the concept of practice in which norms are immanent is not a social concept of practice. If we get the individual subject’s doings right, as activities in direct calibration with things, we do not need others to supply the normativity. The normative notion of the conditions for correct and incorrect usage is available from within the point of view of the individual subject using words in direct calibration with things.6 If I am right in taking the point of the conclusion of B in the way indicated, then in so far as it is the bipartite model itself that is under critique, it is premise one that gets rejected. If the idea of practice in which norms are immanent can be accommodated, then it is not the case that meaning is a function of sign display þ something else that supplies the norms. The very notion of practice has to be taken to be something more than the idea of sign display, it has to be more than something devoid of norms. My preferred statement of the structure of the argument is then the following:

98

Rules and Other People C Suppose meaning is a function of sign display þ interpretation No amount of further sign display can fix the interpretation It is not sign display þ ingredient X that determines meaning, it is sign practice.

I shall retain the word ‘display’ for the idea of normless use. It is what I called ‘transparent ordinary use’ in chapter 3. I use ‘practice’ for the concept of ‘use as practice’ first introduced in the last chapter. From another point of view, the above redescription of the structure of the argument is obvious. If we take seriously Wittgenstein’s injunction that meaning is use, then it cannot be the case that the meaning of language is a function of two things – use (considered as normless) plus an extra bit that adds the norms to use. And it is no good to protest that ‘use’ should be retained for the collectively circumscribed activity that gets added to mere display, where the latter is individual use. Clearly, for many commentators, that is the structure of the interpretation they offer.7 Although that view fits the underlying bipartite view of meaning, it cannot be a stable account of what is going on. The account is unstable, for it endorses a model in which full-blooded normative meaning arises out of a normative free notion of use, what I am calling ‘display’, plus something extra that supplies the normativity. And the fundamental problem with all such collectivist accounts is that it remains profoundly unclear why a collective display is any better at delivering normativity than an individual display.8 There is an obvious danger with the idea that it is the bipartite account of meaning that is the culprit here. If we reject the bipartite account do we not run the risk of endorsing the very idea that looks to be under critique in setting up the rule-following examples? The derivation of the paradox starts from the idea that in hearing someone tell us to ‘Add 2’ we have not been constrained in what we must do. That thought comes from accepting the bipartite view in which the normativity of understanding this utterance, if there is any, cannot reside in the display in which the speaker says, ‘Add 2’; it has to come from something extra to that display. To reject the bipartite account is to insist, contrary to this, that what one hears can include the normative correctness condition and that determines what one should do next. You can hear what you have to do, just as you can remember more than you can memorize.9 Now, disregarding for the moment the objection that such a claim is just too counter-intuitive to be considered, the objection I want to remark on first is that such a claim simply amounts to a barely philosophical shrug – a refusal to take seriously the very considerations that start the paradox off.


99

I think this is mistaken. A simpler label for the position that the current objection critiques is ‘quietism’. The objection is that rejection of the bipartite account amounts to endorsing quietism where ‘quietism’ is identified with descriptivism about ordinary use. By ‘quietism’ I mean the view that there is no philosophical account to be had of how meaning is possible, it simply is. All that can be done is to describe the practices by which we mean things with words and how we respond and how we take ourselves in all this to be bound by norms.10 There is, however, no theoretical standpoint from which all this can be legitimized, it has to be accepted at face value as delivered to us in our ordinary common-sense descriptions of what we do with language – ‘Philosophy may in no way interfere with our actual use of language, it can in the end only describe it’ (§124). The idea that the rejection of the bipartite account entails quietism is mistaken. Let me distinguish quietism from something with which it is easily confused. Quietism proper is the view that not only is there no philosophical theory to be given of how meaning is possible, but that in describing our linguistic practices we describe them in familiar terms drawn from our common-sense description of what we are doing with language. As such, quietism is a refusal to move outside of our common-sense account of what we do with language. And this does not sound like a response to a philosophical problem so much as a refusal to engage in philosophy at all. For sure, such an approach often seems close to Wittgenstein’s own conception of philosophy – ‘A philosophical problem has the form: ‘‘I don’t know my way about’’ ’ and ‘It leaves everything as it is’ (§§123, 124). There is, however, another position that it is important to distinguish from quietism. Suppose the upshot of Wittgenstein’s argument was the view that no theoretical account was available of the conditions for the possibility of judgement. Suppose also, that all one could do in place of this was to offer a description of our practices with language. Such a description need not be quietist if its characterization of language use and, importantly, language users changes the structure of our common-sense self-conception – our conception of who and what we are qua language users, of how we are in the world and of how the world relates to us in judgement. That is to say, suppose it is a description that is revelatory (§129), for it reveals the immanent hidden. I suggest that all three items (who we are, how we are in the world and how the world relates to judgement) are up for reconfiguration in the description of our linguistic practices that takes the place of a theoretical response to the rulefollowing paradox. So it is not the case that the description that halts the paradox is no more than a set of homely reminders of our common-sense account of ourselves, our language and our world. Rather, the description that works is one that offers fundamental lessons for our self-image and our image

100


of how we are in the world in our use of language. The description is not philosophically neutral. It might not be an account delivered by a piece of philosophical grand theory, but we are still doing philosophy when we argue that one description of ourselves is to be preferred over another. That is what I propose to do. Another way of putting this would be to say that we stop the paradox with a description, but it is a description of something intrinsically normative. What needs to be in place for this to work is a reconfiguration of a number of key concepts that make such a description possible. That reconfiguration just is the core account of what makes judgement possible, for the description offered is one of judgements being made. As for the objection that the rejection of the bipartite account of meaning is just too counter-intuitive to be considered, it looks a fair point. It is, however, wrong.

4.3

The Bipartite Account of Meaning

The bipartite account of meaning is fuelled by what McDowell calls the ‘intuitive contractual picture of understanding’.11 McDowell claims the contractualism can be purged of the idea that meaning always depends on interpretation. I do not think this can be quite right. Rather, it is the contractualism itself that fuels the bipartite account and that needs rejecting. At §289 Wittgenstein says that using a sign without justification is not to use it without right. This enigmatic section expresses an idea that is fundamental to what is going on. It is tempting to think that to use a sign with right is to use it with justification in the sense that one’s right to use the sign in the way one does consists in what can be said to support the use. This idea reflects an underlying bipartism. For sure, if one is to provide evidence for one’s right to speak one needs to provide a justification, but it does not follow from that that the justification constitutes the right. The idea in §289 is that what is available within the use of a word can be the very idea of correctness, the right with which one uses it. One does not need to look outside that which is available in the use of the word to give an account of one’s right in using it. That would be to conflate using it with right and using it with a justification. A justification is something separate from the use of the word. Wittgenstein’s point is that the right to use a word need not consist in anything separate from the use of the word; one can simply use it with right. What stands in the way of this thought? What stands in the way of accepting §289 at face value is the thesis that norms, and for that matter rights, are always constituted in contracts. The contractualism about norms depends, in turn, on an epistemological thesis about what is available within experience. The specific version of the thesis of immediate concern can be formulated like this:


101

What is available within experience, including our experience of language, is always something less than the norms of correct language use. If this thesis is adhered to, the notion of correctness, like the notion of speaking with right, is something that has to be added onto experience with a contract. The notion of contract is apposite, for what the ideas of norm and of right amount to is that the way experience is at one time bears rationally upon the way it is at another time. So, with respect to a norm, if experience of language commits us to further language production, then that experience carries a notion of what should come next. The epistemological thesis then says that the notion of correctness that supports the ‘should’ cannot be part of the content of the experience, it must be conceptualized in terms of a contract that binds the later experience to the former. Similarly with respect to the idea of speaking with right. If my saying something is justified, I speak with right and that is because it is something I am entitled to say given what has been said before. Once again, the thrust of the epistemological thesis is that this entitlement, the right to speak, cannot be part of the content of the previous utterance; it must be something about how my current saying is related to the previous one. Once again, a contract is the appropriate way of conceptualizing this relationship. Either way, the relationship between experiences over time is something put together by a contract. No one experience can bind others, they only get bound together by contracts. The idea that underpins bipartism is also at work in resisting the claim that seeing the similarities can be primitive. That latter claim amounts to the idea that no bridge is required to identify that in virtue of which two things are games, one can simply see that they are similar. There is no third thing in virtue of which they are similar. Similarly, I am now defending the thought that, on occasion, one can simply see what should follow in sign use without the need of a third item – a contract – to enforce the should. It might be thought that there is another response to the contractualism in the bipartite model. Rather than endorse the concept of practice in the way that I am proposing, why not say that there is an internal relation between rule and action, between sign display and that which constitutes its correctness conditions?12 And practice is then constituted by sign display and action standing in such internal relations. But surely, to say that a rule (sign display) is internally related to action is just to say that on hearing a given sign display I should/ought/must behave in such-and-such a manner. That, of course, is right, but saying that does nothing to address the issue of what the ‘should’ consists in. That is where we started! We think we should continue the series ‘Add 2’ when we get to 1000 with 1002, 1004, 1006 . . . and so on, but what is the source of this ‘should’? To some extent, in answering this question we have

102


no choice but to end up where we started acknowledging the ‘should’ that guides us. We need, however, more than merely a reminder that this is so. We need some grip on what it is to acknowledge that this is so. Otherwise, we get no purchase on the notion of practice and there is a real danger that talk of practices proliferates and all sorts of things that are not really normatively configured get mistaken for things that are. Philosophical discourse is a classic example of this phenomenon as modes of investigation are contested for genuine argumentative force and structure. But the point is quite general. Without more detail and a clearer sight of what constitutes the ‘should’, we have no constraint on when someone is genuinely in the grip of a practice and when they merely think they are. Furthermore, to take the recipe that practice ¼ sign display in internal relation to action, leaves in place the epistemological thesis that underpins bipartism and then tries to bridge the gap of the bipartite model with an internal relation. The concept of practice is central, but if the work this concept does is to be controlled and disciplined, more work needs to be done and detail supplied about what goes into a practice. Invoking internal relations to mend the gap between the components of the bipartite model is just to apply sticky plaster to a problematic metaphysics that makes the epistemology of bipartism seem unavoidable. We need to purge that epistemology and its underlying metaphysics rather than try to plaster over it with internal relations. Wittgenstein allows, at §289, that speaking with right does not require speaking with justification. The epistemological thesis that generates contractualism about norms is a very general thesis. It is worth noting the most abstract formulation of it, for it bears upon work to be covered in later chapters. The thesis in its most abstract is: E: What is available as the content of experience is always less than what is the case. So, when you have a visual perception, what is available is always something that is invariant between the veridical and illusory cases. In hearing a sentence, what you hear is invariant between a correct understanding and a misunderstanding. It is less than the full normative content. The simple visual case bears comparison with the perception of meaning. In most theories of perception, the content of experience (sometimes qualified as the ‘immediate content’ ) is a content that is invariant between veridical and illusory instances. You think you are seeing a table, but the content of the experience is something invariant between the case in which you are really seeing a table and the case in which you are subject to an illusion. The common element between veridical and illusory experience can be factored out and


103

constitutes the content of experience. The reason for endorsing this view goes back to an epistemological thesis first promoted by Descartes – the transparency of experience. The transparency of experience is the thesis that what is available to you in experience is knowable through and through with certainty. It is something that is immediate to the subject and about which error is not possible. Given that error is possible with respect to perception of things, what is available within the transparent content of experience must be less than what is the case. Hence thesis E and the invariance between the veridical and illusory. It does not matter how we theorize this invariant, e.g. as sense data, what matters is the idea of the common factor. The same story lies behind the bipartite model of meaning and contractualism about norms. Suppose you hear someone say ‘Add 2’. You think you are hearing an instruction the content of which determines that you go on in one way rather than another. That is to say, you understand it to mean that you should, after 1000, continue 1002, 1004, etc., rather than any other way. Correct understanding of the instruction is to know that this is how the series should go. Incorrect understanding is to fail to realize this. But if experience is transparent, the content of your experience on hearing ‘Add 2’ cannot include this much. The content must be invariant between the correct understanding (veridical perception) and incorrect understanding (illusory perception). Hence, what is available as the content of your experience on hearing ‘Add 2’ can only amount to sign display, that which is invariant between correct understanding and misunderstanding. The analogy with ordinary perception of physical objects is not perfect, for once one realizes that all one is left with is sign display as the content of our experience of meaning the very notion of a distinction between understanding and misunderstanding is under threat. Nevertheless, the epistemology of the cases is the same, even if the upshot is more dramatic in the meaning case. Thesis E is not compulsory. Perceptual error is commonly held to be the Achilles’ heel for direct realist theories of perception.13 And the existence of perceptual error is often thought to require thesis E. On such an approach error turns out to be inferential. There is no such thing as direct error, nor direct veridical perception. What is direct is the given, the common factor invariant between the veridical and illusory. What makes a perception veridical or illusory is not constitutive of the given content, it concerns what inference is drawn on the basis of that content. The idea that thesis E is required to accommodate perceptual error is deeply ingrained in philosophy. It is, however, mistaken. The move that shows thesis E is not compulsory is simple. Suppose you think that a seen object is a table. There are two options now available: either the perception is veridical and there is a table, or you are having an illusion. On the common factor account, the content of your perceptual

104


experience stays the same either way. Given thesis E, there are two components to a veridical perception – the common factor plus that which makes it veridical or false. But why not say, instead, that these two cases provide different perceptual experiences? Why not say that in one case you have a perceptual experience in which you directly perceive a table and in the other case you do not? In contrast to the bipartism of the standard model, this would be a unitary account of veridical perception. Of course, if you said this, the account of the content of perceptual experience would not be invariant between the veridical and illusory. The two types of perceptual experience would be different unitary experiences, they would not be compound experiences with a common element. Furthermore, given that you cannot tell for certain whether the perception is veridical or illusory, on the suggestion before us, it would follow that you cannot tell for certain which experience you are having. The transparency of experience would be denied. But that’s the very thing that prompts thesis E and leaves us with an account of experience in which we are always, at best, semi-detached from things with experiential contents that never quite match up to what is the case. This transparency and the consequent detachment of experience from how things are is simply not required. There is an alternative.14 The disjunctive account of perception is not well understood, despite the fact that it gives the simplest possible account of what is going on in perception. What stands in the way of a broader acceptance is, I suspect, an underlying affinity with the transparency of experience. It is the move that says, ‘I may not know for sure how things are going in the world, but at least I know for sure how things are for me!’ On the disjunctive account of perception, we lose the gap between how things are for us in perception and how things are. We get a unitary model of experience in contrast to the bipartism of the standard accounts. On occasions, how things are for me and how things are can be the same. On such occasions, how things are for me in perception simply is how things are. The price is the admission that I cannot tell for sure when it is like that and when it is not. The price is the admission that I have lost the transparent Cartesian authority about how my own experience is going. I can, in principle, get that wrong, just as I can get wrong how things are. Indeed, the great virtue of the disjunctive account is that how things are for me becomes, at a stroke, part of how things are. With the transparency of experience and thesis E gone, there is nothing to set the notion of how things are for me over and above, or distinct from, how things are. From a Cartesian perspective, this may look like a gaining of the world at the expense of losing one’s mind. A healthier view is to see the move as a regaining of a sense of how our minds are part of the world and, as such, something that it can take an effort to know. How does this manoeuvre stand in the case of meaning?


105

Well, suppose you hear a sentence, e.g. the instruction ‘Add 2’ and you think you understand it – you’ve got it right. In what does that understanding consist? That, of course, is the taxing question that Wittgenstein asks repeatedly throughout the Investigations. If thesis E is endorsed we get a bipartite model of meaning in which understanding amounts to hearing the sign display plus something else. The rule-following considerations take away the idea that any candidate for the ‘something else’ will work. Looking to mental states, e.g. possession of a mental talisman, falsifies the phenomenology of meaning and, anyway, looks like nothing more than further and spectral sign display. The regress looms. Wittgenstein’s positive suggestions are, however, fully in line with the disjunctive account. Understanding consists in how you go on, it will depend on what you do. And that is something that is not transparent to you, for you cannot know for sure how that will go until you do it. Such thoughts can look like a simple behaviourism. I think they are much more than that. I return to this point in the next section and also chapter 5. For the moment, it is worth noting that Wittgenstein’s positive account squares with the rejection of the transparency of experience and thesis E. In order to fill out this suggestion, we need to know much more about the concept of practice.

4.4 Practice Wittgenstein uses the word ‘practice’ infrequently. After the resolution in §201 in which he says that there is a way of grasping a rule which is not an interpretation, he continues in §202 with, ‘And hence also ‘‘obeying a rule’’ is a practice’. But he also talks of customs, uses, institutions (§199). Elsewhere he says, ‘In order to describe the phenomena of language, we must describe a practice, not something that happens once, no matter of what kind’ (RFM VI §34, p. 335), and compare also RFM III §§66, 67. One thing that the concept of practice clearly involves for Wittgenstein is the idea of a use extended in time. Indeed, the temporal extension of use is something he dwells on even when inviting himself to comment on other dimensions of use. For example, in §199 he asks, ‘Is what we call ‘‘obeying a rule’’ something that it would be possible for only one man to do, and to do only once in his life?’ This is a striking passage, for having raised two separate questions concerning the dimensions of what it is to obey a rule – measured across people and across time, it is only the latter that he explicitly answers: It is not possible that there should have been only one occasion on which only one man obeyed a rule. It is not possible that there should have been only one occasion on which a report was made, an order given or understood; and so on.

106


Until the third revised edition of the Philosophical Investigations published in 2001, the English translation of the first sentence of this response had been It is not possible that there should have been only one occasion on which someone obeyed a rule.

The German has: Es kann nicht ein einziges Mal nur ein Mensch einer Regel gefolgt sein.

On the face of it, the change in translation between the earlier and most recent editions is striking. The question that Wittgenstein asks invites us to consider two different dimensions of generality in concept possession. One concerns repeatability over time, the other concerns repeatability across subjects. The original translation has Wittgenstein only answering the question about the former. This seems striking, for §199 is one of the few places where Wittgenstein explicitly distinguishes these two dimensions. If he fails to even address the latter dimension, then any reading that sees his resolution of the rulefollowing paradox requiring an essentially social concept of practice needs to explain why, having explicitly given himself the opportunity to endorse such a model, he ducks the issue. The new translation, which does more accurately capture the original, might seem to point towards a social reading of practice. But this is not the case. The emphasis in the answer is still on the temporal repeatability that is required for rule-following. Both the question and the reply put the priority on the temporal dimension and not the generality across subjects. It is, after all, compatible with the reply given to allow that only one person could follow a rule just so long as they did it on many occasions. The emphasis is still in the wrong place for this passage to be read as an explicit endorsement of a social concept of practice. Practice requires the repeatability of sign use. For sure, after the quoted passage he continues, ‘To obey a rule, to make a report, to give an order, to play a game of chess, are customs (uses, institutions).’ But even here, it is only the second of the parenthetical words that has a social meaning. The key concept ‘custom’ is not necessarily social and neither is ‘use’. It really is stretching the point to build a social account of rule-following on a brief throwaway word when the question invites an explicit answer about generality across subjects and does not receive it. The only plausible interpretation is that generality across subjects simply is not important, and certainly not a constitutive element to what Wittgenstein is getting at with regard to the generality of concept use. It is generality in use that matters and, minimally, that requires the possibility that a word can be retained and employed in different combinations. For that to be possible, the


107

word must have a use that can be extended in time, the extension of use over persons is not the issue. Compare also RFM §67 where, after acknowledging that a consensus of techniques belongs to the essence of calculation, Wittgenstein says, But what about this consensus – doesn’t it mean that one human being by himself could not calculate? Well, one human being could at any rate not calculate just once in his life.

It is repeatability across time that is at issue, not across subjects. The concentration on repeatability across time, rather than persons, is due to the central need for a concept of grammar: And when we speak of someone’s having given a name to pain, what is presupposed is the existence of the grammar of the word ‘pain’; it shews the post where the new word is stationed. (PI §257)

It is the idea of a word having a ‘post’, a ‘proper place’ (PI §1) in sentences that underpins the need for generality. It is the idea that Evans labelled the generality constraint.15 There is no purchase on this concept of proper place without the capacity for repeatability over time. The issue whether the concept of a proper place in grammar requires repeatability across persons is a driving theme in the passages covering the private language argument. I turn to consider those passages in the next chapter. There is, for sure, a potential problem if the concept of a proper place in grammar is indexed to the use of a single subject. The problem concerns whether the use of a single subject can be seen as a use that is calibrated against that which provides the standards of correctness. This problem turns on how the single subject is characterized. There is, however, independently of the private language argument, no clear evidence that the concept of proper place in grammar ever amounts to more than the general requirement of grammatical position. It is a further issue whether grammatical position has to be understood as something constituted by the social use of words. I shall argue that the private language argument provides no reason whatsoever for constituting grammatical position socially. The thrust of that argument is against a particular conception of the individual subject, the private Cartesian conception. It is not an argument against an individualist conception of the subject in which the subject is conceived as in direct presence of that which calibrates word use, namely, the world. My immediate concern is to characterize what is essential for the concept of practice to capture sufficient to make the notion of grammatical position intelligible. In the previous

108


chapter I argued that ‘use’ should be understood as ‘use in practice’. It is now time to say more about the concept of practice. The key to practice minimalism is the account of the subject. I have said that the subject is a judge, a self-as-will. By ‘self-as-will’ I do not mean an extra entity over and above the empirical self. I mean only to draw attention to the fact that the concept of an empirical self required is that of an active subject. The subject is an agent, something minimally with a capacity to alter its perceptual inputs at will.16 It is having a will that makes us possessors of content and it is by virtue of having a will that we have the world directly in judgement. The basic model of experience, on my account, is of a confrontation between the self-as-will and that which is independent of will. What is independent of will is how things are – the world. Practice minimalism is an account of a use of signs that cannot be individuated independently of the truth-conditions that provide the normative notion of correctness/incorrectness. What that amounts to is a self-as-will engaged with that which is independent of will.17 The self-as-will is the core to the concept of a judge. It is a conception of the subject as essentially a player in the space of reasons, a space that is, because of the operation of will, a space of spontaneity. I argued, in the last chapter, that the thesis that seeing the similarities is primitive places an open-endedness and creativity at the groundlfoor conception of concept possession. That point is now met again. Practice is composed of judges, subjects with a capacity to change perceptual inputs and action outputs at will. This is not wilful in the sense of capricious. It is wilful in the sense that it displays a capacity to alter perception and action so as to meet goals, aims and projects in the light of belief. Given the above, the description of word use can then be taken at face value as a description of intrinsically normative practice. It is not merely a description of common sense, for at its heart lies the above thesis about the self. In short, practice is composed of engagements with things by judges. Anything less than this will not be a conception of practice in which normativity is immanent. Alternative conceptions of practice simply do not provide an account of normativity. The concept of practice that I am proposing we extract from Wittgenstein is Practice is composed of engagements with things by judges. The concept of a judge is implicated in the concept of practice because of the role of attention in characterizing our engagement with things, the engagement in which we see similarities. Attention is purposeful seeing; it is wilful, something you do. It is not something that just happens to you.18 The concept of engagement picks out the directedness of our contact with things, the


109

directness involved in seeing-that a word is used correctly, a seeing-that the relevant truth-conditions obtain. Before adding more detail to this concept of practice, I shall briefly review some alternatives. It is tempting to think an alternative concept of practice might go something like this: Alternative (a): Practice is composed of actions, including our techniques for handling things. This concept of practice subdivides into two variants: (a.i) Practice is composed of the bodily circumstances of sign use and, (a.ii) Practice is composed of actions intentionally described. (a.i) endorses a bipartite model of meaning. It endorse a model in which intentionality cannot quite reach into the world and so our hands have to finish off the connection left dangling by mere thought! That is problematic, for it is unclear how limb movements are to provide the standards of correctness of word display. Alternatively, (a.ii) either does not amount to a concept of practice at all if the intentional is conceived out of contact with things, or it does not differ from my concept of practice minimalism. Another alternative is: Alternative (b): Practice is composed actions informed/shaped/disciplined by those of others, including their linguistic actions. (b) is the communitarian concept of practice. I shall discuss these in turn, starting with (a).

Alternative (a) This is a common way of understanding Wittgenstein and seems to fit with a number of textual sources, e.g. In the beginning was the deed. (CV, p. 31)

In §7 where Wittgenstein first introduces the concept of a language-game, he does so by referring to the ‘whole, consisting of language and the actions into which it is woven’. Again at §23 he emphasizes that the term ‘language-game’is

110


meant to bring into prominence the fact that the speaking of language is part of an activity, or of a life-form. In contrast to the essentialist account of the structure of language that he gave in his early writings, the new account of language is clearly grounded in a more general description of human behaviour. But quite what is meant by behaviour and quite how it grounds the account of language are debatable issues. In MS119 pp. 147–8 Wittgenstein writes ‘The primitive form of the language-game is certainty’, and it is tempting to think that what is important about behaviour is that it picks out the point where justifications and the giving of reasons end, it gives the certainty. In using language we do so subject to normative standards of correctness/incorrectness and those standards structure the way that language is inferentially wired up. The inferential connections manifest the way we justify our use of words and use words to give reasons for our actions, including other word use. But reasons do not go on for ever, they give out in what we do. None of this, however, helps us understand the role of behaviour. Reasons give out in what we say and we have to turn to what we do to stop what would otherwise be a regress of reasons given in words. We use a word in giving a reason for an earlier use, or for a piece of behaviour. If the former gives a reason for the latter, it shows why the latter should be so. We wonder should we believe that P and are told ‘If Q, P’ and are also told ‘Q’. That is why we should believe that P. But what now if someone says they do not see the force of this? Can we appeal to anything more? The options here seem to be either we allow a regress as we try to add an extra premise – ‘If, if Q, P and Q, P’ and so on, or we say, with respect to the original argument, ‘But this is what we do’. We would say to the other that they have to see what to do: All I should say further as a final argument against someone who did not want to go that way, would be: ‘Why, don’t you see . . . !’ – and that is no argument. (RFM I §34)

That is not an argument not because it is something outside the realm of the force of reason, but because it concerns the conditions for the possibility of the operation of reasons. The appeal is not to something outside the space of reason. In contrast, if what we do is simply the bodily circumstances of use for sign display, there is no force to the rational ‘should’; it is only part of the catalogue of contingencies. If (a) is interpreted as (a.i), it provides no basis for the rational should. This is the account in which practice is composed of our accommodations with the world, but where these are treated purely as the causal patterns of our engagements. There is, on such a model, no directedness


111

to our causal engagements. Without such directedness, our causal engagements cannot engage with word use to provide a normative standard. Suppose instead that we interpret ‘what we do’ as characterized in (a.ii). This does not help, for then there is nothing special about doing that is not already available in those special sorts of doings with which we start – speaking and the saying of things of the form, A!B A B And if nothing especial is added by the notion of what we do, the direct entitlement to use a word with right, the direct presence of that which supplies the correctness condition, can be as equally available within our speaking as within our manipulation of things. The involvement of the body and its manipulations of things is important, but this is not something that stands outside our intentionality as the second component required to provide the normativity to symbol display. It is, rather, a manipulation of things that takes place from within the intentional stance, within the space of reasons. Of course, if the space of reasons is understood as disengaged from things, so that even an intentional description of action were detached from an engagement with objects in the world, then nothing would be achieved by invoking the concept of practice. On such a way of understanding practice (a.ii), practice itself gets a bipartite analysis, one part containing the intentional characterization of action and some other component providing the bit that engages with things. To take practice (a.ii) in a unitary sense is to say that there is no gap between exercises of concepts and contact with the world. This is not to say that there is a narrow fringe of bodily twitches that, when added to concept manipulation, provides correctness; it is to say that concept possession is a rationally structured manipulation of things. What is interesting about techniques for manipulating things – the ‘readiness to hand’ of the world, etc., is that they illustrate something that applies to all meaningful language but which we often forget. If such accommodations are to matter, they must be accommodations that are directed. They must be conceptually structured. The choice here is very simple. Either the concept of our bodily accommodations picks out something suited to impact rationally upon word use or it does not. If the former, then accommodations work from within the space of reasons and merely remind us that we are already at home in the world in our word use. We do not need to leave word use and focus merely on the body to make contact

112


with things. If the concept of bodily accommodations picks out something that merely causally impacts, it will fail to touch the normative conceptual patterns within the space of reasons. Despite the voluminous discussions of the way our bodies provide our accommodation with the world in the work of writers such as Heidegger and Merleau-Ponty the critical issue always remains: when the world is ‘to hand’ is that a remark about something beyond the space of reasons, or within it? The latter is the only view that makes sense. And that means that the body does not bring something intrinsically different or extra that was not already available from within word use. It serves only to remind us that throughout the space of reasons we are, in principle, in direct touch with things, for the normative patterns of correct use are immanent. It is not that we need to touch and move things to ground speech, our speech is already something that can be directly grounded.19 Alternative (a) is ambiguous between a version (a.i) in which the role of behaviour stands outside the space of reasons and so unable to provide an account of the rational should, and a version (a.ii) which either reinstates a bipartite model of meaning or collapses into practice minimalism. The second paragraph of §201 says that ‘there is a way of grasping of a rule . . . which is exhibited in what we call ‘‘obeying the rule’’, ‘‘going against it’’ in actual cases’. Note, Wittgenstein does not say that there is a way of grasping a rule that is what we call obeying a rule, etc., he says that it is a way of grasping a rule that is exhibited in these things. So doing certain things, e.g., continuing a series, is not what constitutes grasping a rule; it is what exhibits grasping a rule. What then is exhibited? The answer, I think, has to be our direct engagement with things. Nothing else would do the job. A direct engagement with things in which we see directly a word used correctly because it is calibrated against that which provides the standards of correctness/incorrectness provides the rational should to word use. Anything less provides, at best, the causal shunts and tugs that impinge on us in our use of words, but not the reasons for our use of words nor our rights to use words as we do. The concept of practice is important, but it is confusing to treat it as if it were an empirical concept, something that picked out mere contingencies of bodily behaviour. Ter Hark says, ‘a practice is not justified by anything, since a practice is a standard of justification’.20 That, however, does not help identify a practice. There is a danger that an empirical notion of practice will degenerate into an account of dispositions for bodily movement where these are thought of as the final and certain boundaries in accounts of our reasons, the final arbiters of what we should do, the place where reasons stop – ‘practice as final limit for interpretation and justification’.21 But that is misleading. It is as if we give out words and then, finally, attach them to something secure by touching it. This is, at best, an account in which the words and action are


113

internally related. But that move comes too late. Is the touching an empirically described movement or an intentional action? The former provides no normativity, the latter cannot ground/limit words because it is of the same type of event as the use of words. There are not then two things internally related, there is one thing – an active direct attitude to the world. If this is right, practice is not a fundamental language-game, a fundamental category of human behaviour that grounds the higher level activities of word use. We do not have two sorts of behaviour one of which stands in need of support, the other suited to provide it. ‘Practice’ is not a word for a special extra form of behaviour that ties down verbal behaviour and is the source of its rationality. ‘Practice’ is a word for the complex form of all word use which has, at its heart, subjects with an active direct attitude to things, the things in virtue of which word use is calibrated. ‘Practice’ is a word for drawing attention to the fact that our word use is a use that is directly calibrated with things and does not stand in need of the second ingredient as per the bipartite model that gives word use its point. Wittgenstein says ‘For words have meaning only in the stream of life’ (RPP II §687), and we might wonder whether the stream of life is something to be characterized independently of our use of words in giving reasons; that is, independently of the space of reasons. But Wittgenstein considers this and rejects it: ‘Instincts come first, reasoning second. Not until there is a language-game are there reasons’ (RPP II §689). Language games are normative structures in which our word use is directly calibrated against that which provides the standards of correctness/ incorrectness. All language games have engagement with things, some wear it more obviously than others.

Alternative (b) It is difficult to formulate the communitarian account of practice in a way that does not rapidly reduce to either a banal platitude or clearly untenable position. The communitarian account of practice is a non-starter. The only real issue with this notion of practice is why so many people seem to think it is plausible and, furthermore, attribute it to Wittgenstein. Once the position is formulated in the context of addressing the problems raised by a bipartite model of meaning, its implausibility is apparent. The single most important thing to be understood in seeing what is wrong with alternative (b) is that it fails to ask, let alone answer, the right question. The issue at hand is: what constitutes the normative standards of correctness/incorrectness of word use? All too often this issue is discussed in an epistemological way, it is mistaken for the issue of how we tell that there are normative standards of correctness/ incorrectness for word use. To the latter issue, the fact that others use words in

114


various ways is often a useful clue to what is going on. It can never, however, be an account of what is going on. I formulated alternative (b) as follows: Alternative (b): Practice is composed of actions informed/shaped/disciplined by those of others, including their linguistic actions. This can look truistic. It is, however, flawed. The fundamental flaw with practice (b) is simple. The idea is that the linguistic actions of the lone subject are disciplined by the actions of others, the subject’s community. So, the lone subject’s use of words are calibrated by how they stand against the actions of their community, including the linguistic actions of the community.22 Of course, if the actions of the lone subject are understood as on a unitary model of meaning in which the use of words is not characterizable independently of that which calibrates it, the claim is redundant. The account of the use of words by the lone subject must then be understood as on a bipartite model of meaning. On a bipartite model, meaning is compounded out of word use construed as display plus that which provides the standards against which word use is calibrated. Whatever else that second component is, it must be something that can stand against word display and calibrate it. Now, suppose the actions of others, including their linguistic actions, are further instances of word display. But that version of practice (b) is going nowhere, for all it offers against which to calibrate word display is further word display, the very sort of thing that stands in need of calibration. Perhaps the actions of others, including their linguistic actions, should be construed as cases of word use that satisfy a unitary model of meaning; that is, a use in which words are directly calibrated against things? But that is puzzling, for why should the lone subject be so disabled compared to others? Why should mere numbers provide others with a grasp of meaning that satisfies the unitary model when lone speakers are left with only word display, the first component of a bipartite model? The suggestion is barely intelligible and, at best, simply begs the question against the lone subject. What makes practice (b) look plausible, although also truistic, is if we treat it as an answer to an epistemological issue, namely, how do we tell that word use is disciplined? In answer to that question it is often useful, although not necessary, to show how we uncover calibration by reference to the word use of others. But this is not the constitutive question. It is not an account of what it is for word use to be calibrated. The details of calibration are often given away in the word use of others and this can serve as a useful tool for uncovering that against which word use is calibrated, namely, how things are. From the epistemological point of view, the word use of others is a powerful tool.


115

The reason that it is a powerful tool can only, however, consist in a constitutive account of calibration. The obvious thing to say is that calibrating our word use against that of others is useful just in case the others’ word use is calibrated against things. Our calibrations against things, even if direct, cannot be transparent. The community is useful as a reservoir of standards, a record of contacts with the world. It serves an epistemological purpose only, not a constitutive one.

4.5 What You See/Hear is not Normless The concept of practice that I am extracting from Wittgenstein is the engagement of judges with the world. The subject is a judge, for it is a wilful subject. It is a subject whose attitude to the world is framed by attention. It is an active attitude, it is not the passive receipt of experiences. It is a wilful activity. It is, however, an activity in which the world is directly present in the attention of the subject. This requires the notion of direct apprehension of how things are, a direct presence of truth-conditions. There is, then, a lot that goes into the concept of practice despite its minimalist character. Let me unpack some of the resources being deployed. The model that I am urging is one in which what is available within experience is the world. This availability is a ‘presence’, for it is the notion of experience in which the self-as-will is confronted with that which is independent of will. It is important that this point is seen aright. The idea of an impediment might be thought of as the bare causal resistances and affordances of the world. For sure, the world offers resistances and affordances through its causal powers and capacities relative to the causal powers and capacities of the subject. The mere occasioning of such causal engagements is not, however, sufficient for conceptual experience.23 Sitting still at a desk writing provides me with a location in a complex web of causal resistances and affordances. All manner of causal interactions hold of how I am in the world. Most of these do not, however, enter experience, for most are not experienced as resistances or affordances for my will. The role of the will is that it provides the directedness to the causal structure of my engagement with the world. The world that enters experience is fundamentally the world that has the capacity to thwart and enable my will. Consider very basic experiences in which we learn how to perform simple manual skills. The simplicity of Wittgenstein’s examples can tempt us to think he is endorsing behaviourism and missing the very idea of the way the world is for us in experience. But to see matters thus is to set up a familiar and threatening dichotomy. It is the dichotomy in which our experience

116


of the world only gets lit up if conceived in terms of its inner qualitative blush, otherwise we are left with behaviourism. But whatever Wittgenstein says in critique of the former choice (see chapter 5), he does not leave us with a bare description of the causal structure of our engagements as characterized in behavioural stimuli and response. Consider his first language-game in §2: the builder and his assistant and the language comprising ‘block’, ‘pillar’, ‘slab’, ‘beam’. The language-game provides a use for these words, but what goes into the concept of use? One way of characterizing use would be in terms of a set of causal dispositions, dispositions such as responding to the word ‘block’ by picking up a block and taking it to the builder. If this is thought of merely as a set of causal responses, it is something akin to the sort of training that one can do with animals. Now, when Wittgenstein first introduces this simple language-game, he does it in part to critique the need for an inner image to control the use of words. It was in response to the question ‘What is the meaning of the word ‘‘five’’?’ in §1 that he said, ‘No such thing was in question here, only how the word ‘‘five’’ is used.’ But although explicitly targeting the idea of the inner image as candidate for the meaning, he never slips to describe the use in mere behavioural dispositions. This is clear in §6 where he extends the language-game of §2 to consider it as the language of a whole tribe. He asks what the association is between the word and thing that is established by such training and it is in response to this that he acknowledges our temptation to think of a picture of the object coming before the learner’s mind. His question in response to this temptation is instructive. He says (§6): is it the purpose of the word? – Yes, it can be the purpose. – I can imagine such a use of words (of series of sounds). (Uttering a word is like striking a note on the keyboard of the imagination.) But in the language of §2 it is not the purpose of the words to evoke images.

Even in the §2 language-game there is a purpose. It is a directed language. The use of words is not merely a causal response. You can train animals to acquire causal dispositions – to fetch, move, sort objects on given stimuli. But they never get the point of what they are doing, for they have no will. The dog that runs back and forth for the stick you repeatedly throw doesn’t get it. The dog is a slave to its causal dispositions. In contrast, the builder’s assistant gets the point; he is a subject with purpose. Learning to do the sorts of things that Wittgenstein describes is not just learning causal routines. The will is already there in the builder and assistant scenario and makes this a case in which their behaviour and the experiences on which it rests are conceptual through and through.


117

The sorts of accommodations with the world that Wittgenstein describes are already conceptual because they are cases in which the world is present to the will as impediment. The accommodations are not just reactions, but reactions that provide a capacity for meeting the generality constraint. The builder’s assistant gets the point because there is a directedness to his causal encounters. Learning how to do things, manual skills, is not just a matter of learning causal routines. The will is already there in the bodily control. It is a conceptual control. The role of the will in giving the notion of a directedness to our engagement reinforces the earlier argument that it is generality across time that matters for the concept of practice, not generality across persons. The reason that there cannot be one-off practices is because practices are directed engagements with the world. It is the directedness of the engagement that makes it a component of experience that can be held onto and redeployed in other combinations in the wilful management of our engagements over time. And that is the exercise of concepts. The capacity for redeployment is what makes a directed engagement satisfy the generality constraint. That is why, when he considers the question, Wittgenstein dwells on the temporal repeatability of engagements not their shareability by different subjects. Of course, in so far as an engagement is repeatable and, in so far as there is no special reason why only one subject can attain it (particular circumstances of skill, perceptual sophistication, etc.) then there is no reason why it should not be shareable. Concepts are shareable, but their being so is not what makes them concepts. It is a consequence of their being engagements that satisfy the generality constraint that they are shareable, but the latter point is not constitutive of concepts.24 Practice is a relation of an agent to the world. It is not their action as such, but the way they are related. It is an engagement that gives them the capacity to act, including, centrally, linguistic actions, and these are normatively structured. Practice is not, however, the subject’s non-linguistic actions except in so far as they’re intentionally characterized. Wittgenstein has an overly epistemological way of expressing himself on these points. For example, in §211 he takes the absence of being able to say anything by way of answer to the question ‘How do I know how to continue a pattern?’ as evidence that I act without reasons. But that is contrary to his own advice at §289 that acting without justification does not mean acting without right. It all depends on what Wittgenstein means by acting without reason. If he means that we reach a point in saying what we know at which we simply act without being able to offer any further articulations in justification, then he is treating acting for a reason as acting on the basis of something that can be said in justification for the action. But it does not follow from this that when we so act, we merely behave, or merely indulge a causal response or accommodation with things. When we have exhausted the justifications and reached bedrock and our

118


spade is turned we reach the point at which we say ‘This is simply what I do’ (§217). But my doing here is again not a mere causal accommodation, it is a doing with right. As such, it is something within the space of reasons for it is something that could be placed within a rational framework by being seen in the ongoing directedness of our engagements. We do not need to be always talking to be reasonable and rational. We need to be talking to give reasons, but we can be reasonable without the chatter. The ‘bedrock’ in §217 is not then the basement of causal affordances and accommodations below the level of our conceptually ordered encounters with the world. It is simply the encounters that we do not need to talk about, not because they are ineffable but because they are the sorts of encounters in which we simply see the directedness of what we and others are doing. The difficult thing here is not, to dig down to the ground; no, it is to recognise the ground that lies before us as the ground. For the ground keeps on giving us the illusory image of a greater depth, and when we seek to reach this, we keep on finding ourselves on the old level. Our disease is one of wanting to explain. (RFM VI §31)

We do not need foundations below the conceptual on which to base our use of words, for in using words the conceptual reaches into the world. The giving of reasons does not have to end in non-linguistic action, it can end in speech too. That must be right, for action only works if it manifests the appropriate directedness of engagement, otherwise it is just a non-normative push/shove/ nudge; it is not a full stop. You can, however, stop with a description: the difficulty . . . is not that of finding the solution but rather that of recognising as the solution something that looks as if it were only a preliminary to it . . . whereas the solution of the difficulty is a description, if we give it the right place in our considerations. If we dwell upon it, and do not try to get beyond it. The difficulty here is: to stop. (Z §314)

I have described the idea of a direct presence of truth-conditions in the most abstract way in terms of the self-as-will in direct presence of that which is independent of will. That is the heart of the idea that word use stands in direct calibration with the standards of correctness. The idea of direct presence of truth-conditions is familiar in contemporary philosophy of thought in neoFregean work on singular thoughts.25 In McDowell’s work arguments against epistemological thesis E (§3 above) are clearly deployed.26 McDowell’s metaphor of inner space ‘reaching out’ is not, however, quite right. What provides the friction that shows that experience is directly calibrated against things is not a reaching out, it is the friction between will and not-will.


119

4.6 Seeing the Similarity in Particular Cases It might be thought that I have neglected the hardest part in the whole business of the rule-following argument, for I have said nothing yet about the source of normativity in the general case. What seems puzzling about Wittgenstein’s arithmetical examples is the sense of the ‘logical must’ that reaches to infinity. How can the account given so far accommodate that? Even if a direct confrontation with a particular were possible, surely the model cannot be extended to a direct confrontation with the infinite series.27 The worry is misguided. The worry suggests that, in addition to the idea of a confrontation of will on non-will, rule-following requires a confrontation with a series, or pattern that is independent of will. Although it is tempting to say we experience the ‘because’, there is no such experience (§176, 177). But that is not a problem. Thinking that there could be such an experience as the experience of the ‘because’ would be like thinking that training in arithmetical rules could only be achieved under the guidance of Obi Wan Kenobi! And the problem is that we do not feel the force of the pattern. The concept of friction for which I have argued is found in the impinging of the world in particular circumstances. That is what justifies me in saying I can go on (§154), but how? The answer is that we feel the friction in the particular case and the pattern emerges. The answer to the question of the source of the logical must of the pattern – the resolution to the rulefollowing argument – is that we see the similarities and the pattern emerges. It is the answer already seen at play in the previous chapter on learning the concept game. The arithmetical case is no different, although it has considerably greater strategic impact in Wittgenstein’s argument. But what similarity is primitively seen in the case of the series generated by the instruction, ‘Add 2’? The answer is easy. We mind the gap. We see the series started by the teacher who says, ‘2, 4, 6, 8, . . .’ and, if we are paying attention, we see a similar gap between each numeral. The gap between 6 and 8 is similar to that between 4 and 2. We see that this is like that. There is still room for spontaneity in this. Freedom is intrinsic to what we know. It is not there is what we know and freedom comes after, when the knowledge runs out. That is a bad image of conceptual freedom, it is the image of freedom as absence. Letting the pattern emerge is not an acceptance of a looseness in the patterns. Realism about the patterns of grammar, realism about the idea of grammatical position, is still in play. The point is simply a reflection of the role that active subjects, selves with wills, play in the account of these real patterns. Patterns emerge out of the way we use words and see similarities. Grammar really is perspectival.

120

Rules and Other People Notes

1 2

3 4

5

6

7

8

Pears is very clear on this: D. Pears, The False Prison, vol. II, 1988, p. 369. It remains a total mystery to me why anyone would think that the normativity of meaning could arise out of being with others rather than being with things. Nevertheless, it strikes many people as obvious, e.g. S. Glendinning, On Being with Others: Heidegger, Wittgenstein & Derrida, London: Routledge, 1998. Bloor sets the debate up in this way: D. Bloor, Wittgenstein, Rules & Institutions, London: Routledge, 1997. I am heavily indebted to the work of McDowell and Pears who have pursued the reductio reading of the rule-following argument: J. McDowell, ‘Wittgenstein on following a rule’, Synthese, 58, 1984, 325–63; D. Pears, The False Prison, vol. II, Oxford: Clarendon Press, 1988. Cf. M. Luntley, ‘The Transcendental Grounds of Meaning and the Place of Silence’, in Meaning Scepticism, ed. K. Puhl, Berlin and New York: De Gruyter, 1991, pp. 170–88, for my earlier account of the reductio. McDowell, op. cit., cf. §201, paragraph 2, ‘What this shews is that there is a way of grasping a rule which is not an interpretation, but which is exhibited in what we call ‘‘obeying the rule’’ and ‘‘going against it’’ in actual cases.’ The question how norms can be immanent to practice is difficult, the option for treating norms this way is rarely seen. For example, Baker and Hacker are clear that practice does not need to be read socially – G. Baker and P. Hacker, Scepticism, Rules and Language, Oxford: Blackwell, 1984. Now, they do not say enough to warrant assimilation into the account of practice that I am promoting and would probably not endorse it. Nevertheless, it is instructive to note how Williams responds to their insistence that practice is non-social. She assumes, in effect, that they must have endorsed Platonism. She says that the behaviour of the individual only has a pattern to it because it has ‘a master pattern’. See M. Williams, Wittgenstein, Mind and Meaning: Towards a Social Conception of Mind, London: Routledge, 1999, at p. 172 ff. What Williams does not see is that the notion of ‘master pattern’ might be simply the idea of a normative pattern immanent to the individual’s behaviour. Bloor and Williams, op. cit., are the obvious cases here, but see also Brandom’s theory in which the normativity of meaning arises out of the social scorekeeping on individual’s utterances: R. Brandom, Making it Explicit: Meaning, Representing and Discursive Commitment, Cambridge, Mass.: Harvard University Press, 1994. Bloor is careful to distinguish his reading from a crude collectivism in which normativity consists merely in what the group says. It is institutional facts that are meant to do the work in Bloor’s account. But if the notion of institutional facts is not to collapse into mere group display, he needs some account of where the normativity of institutional facts comes from. No such account is available, nor does it seem possible to supply. It also requires some notion of the limits of sense and that is problematic too: cf. P. M. S. Hacker, ‘Philosophy’, in Wittgenstein: A Critical Reader, Oxford: Blackwell, 2001, pp. 322–47.


121

9 ‘What we learn is not just what we have studied; and what we have been taught is not just what we were intended to learn. What we have in our memories is not just what we have memorized’, S. Cavell, ‘Excursus on Wittgenstein’s Vision of Language’, in The New Wittgenstein, eds A. Crary and R. Read, London: Routledge, 2000, pp. 21–37, at p. 28. Of course, what you memorize is not enough, for what is delivered by what you memorize is just the content of sign display and that is not what you remember; at best, it is a device to help you remember. This is not the content of the memory any more than sign display is ever identifiable with the content carried by sign display. Cf. McDowell: ‘It is really an extraordinary idea that the contents of minds are things that, considered in themselves, just ‘‘stand there’’ ’ – p. 45, J. McDowell, ‘Meaning and Intentionality in Wittgenstein’s Later Philosophy’, in The Wittgenstein Legacy. Midwest Studies in Philosophy, vol. XVII, Notre Dame, Ind.: University of Notre Dame Press, 1992, pp. 40–52. 10 This is how I understand Baker and Hacker, op. cit., and hence have some sympathy with Williams’ critique of their position (op. cit.) even though she fails to see what is possible. It is not clear that Baker would still endorse the earlier view of ordinary use: cf. G. P. Baker, ‘Wittgenstein on Metaphysical/ Everyday Use’, The Philosophical Quarterly, 52, 2002, 289–302. 11 J. McDowell, ‘Wittgenstein on Following a Rule’, Synthese, 58, 1984, 325–63. 12 See M. Ter Hark, Beyond the Inner and the Outer, Dordrecht: Kluwer Academic Press, 1990, p. 47, for this move. 13 Locus classicus for this idea: F. Jackson, Perception: A Representative Theory, Cambridge: Cambridge University Press, 1977, chapter 2, for very clear articulation of this; but see also J. Dancy, An Introduction to Contemporary Epistemology, Oxford: Blackwell, 1985. 14 The alternative is called the disjunctive analysis of experience. This goes back to J. M. Hinton, Experiences: An Enquiry into Some Ambiguities, Oxford: Clarendon Press, 1973; but see P. F. Snowdon, ‘Perception, Vision and Causation’, Proceedings of the Aristotelian Society, 81, 1980–1, 121–50; ‘How to Interpret Direct Perception’, in The Contents of Experience, ed. T. Crane, Cambridge: Cambridge University Press, 1992, pp. 48–78; J. McDowell, ‘Criteria, Defeasibility and Knowledge’, Proceedings of the British Academy, 68, 1982, 455–79, reprinted in J. McDowell, Mind, Value and Reality, Cambridge, Mass.: Harvard University Press, 1998, pp. 369–94. 15 G. Evans, The Varieties of Reference, ed. J. McDowell, Oxford: Clarendon Press, 1982, p. 100 ff. 16 It is this that underpins the spontaneity, the freedom, experienced within the space of reasons. For more on the concept of spontaneity cf. J. McDowell, Mind and World, Cambridge, Mass.: Harvard University Press, 1994; ‘Having the World in View: Sellars, Kant and Intentionality’, The Journal of Philosophy, XCV, 1998, 431–91; and the papers in Reading McDowell – On Mind and World, ed. N. Smith, London: Routledge, 2002. 17 There is no need for the qualifier ‘directly’ to apply to this engagement, for an engagement with that which is independent of will simply is an engagement with nothing less than the world.

122


18 This is not to deny that there is a responsiveness to perception, a sense in which our perceptual experiences can be thrust upon us, their character ‘wrung’ from us. Getting that matter right is not easy. See J. McDowell, op. cit., for discussion of this, although I suspect McDowell’s account cannot be fully right. At p. 444 he raises the question, ‘How is it that sensory relatedness to the environment takes the form of conceptual episodes, episodes that . . . ‘‘contain’’ claims at all?’ The question is apposite, for conceptual claims fall within the space of reason and come under the form of judgement – this is the space of spontaneity. But McDowell, following Sellars, wants to have just the very same experiential contents as things that can impact on us capturing the sense of receptivity to experience. It is not clear to me that McDowell ever satisfactorily accounts for this. At p. 460 he says that intuitions – ‘immediate sensible representations of objects’ – have the ‘same logical togetherness’ that conceptual contents enjoy. But simply saying that the elements of experience in sensory receptivity enjoy the same logical togetherness that the elements deployed in judgement enjoy, does not address how it is that this is so, nor how that squares the tension between experience conceived in terms of receptivity and in terms of the spontaneity of judgement. Part of the problem concerns how we accommodate responsiveness without returning to a concept of nonconceptual content that reintroduces the given. See M. Luntley, ‘Nonconceptual Content and the Sound of Music’, Mind & Language, forthcoming, 2003, for an account of nonconceptual content that does not reintroduce the given. 19 Clarity about the concept of ‘accommodation’ informs our understanding of what Wittgenstein means by ‘natural’ and the ‘natural history of mankind’. From what I know of Heidegger, the primacy of the concept of mood in his account of our fundamental engagement with things leaves the question I am pressing unanswered. Is ‘mood’ a conceptually structured engagement or not? I suspect it is the latter. If so, that makes Heidegger an unreconstructed empiricist with all the traditional problems about the given. See chapter 5, section 5, for more on the notion of ‘accommodation’; also chapter 3, section 6. 20 Ter Hark, op. cit., p. 51. 21 Ibid. 22 There is more than one model of how this can occur. At its simplest, the model is one of simple conformity. Despite all his warnings against a dispositional account of norms, it is difficult to read Kripke’s response to the rule-following sceptic as anything more than a simple communitarian dispositional model. More sophisticated versions of alternative (b) are available, however. For example, Brandom’s inferentialism is a position in which the individual’s encounters with the world (causally conceived) get to enjoy normativity in virtue of the scorekeeping of others on the individual’s utterances, cf. Brandom, op. cit. 23 There might be patterns to causal engagements other than conceptual patterns. If so, this would permit a notion of nonconceptual representational content whose status as nonconceptual was precisely that its grammar was not the perspectival grammar of the willing attitude. Cf. Luntley, ‘Nonconceptual Content and the Sound of Music’ , op. cit.


123

24 Of course, some engagements are social because they are engagements with the social world. Where Wittgenstein has an explicitly social idea of engagement it is because the engagement is with something social. For example, at RPP II §629 he asks how we are to describe human behaviour and says, ‘surely only by sketching the actions of a variety of humans, as they are all mixed up together. What determines our judgement, our concepts and reactions, is not what one man is doing now, an individual action, but the whole hurly-burly of human actions, the background against which we see any action’: cf. also Z §567. The point is only that many if not most of our actions are social actions and thereby involve some engagement with social matters. This does not mean that engagement with the social is constitutive of the very idea of grammar, of the standards of right and wrong use of words. It is constitutive of the grammar of words that speak of the social, but that is a different matter. Note also how in this passage, even when explicitly discussing the social embeddedness of action, Wittgenstein still emphasizes the concept of generality that I have argued is primary – generality across time. 25 Seminal works include J. McDowell, ‘De re senses’, The Philosophical Quarterly, 34, 1984, 283–94; G. Evans, The Varieties of Reference, ed. J. McDowell, Oxford: Clarendon Press, 1982. For a survey, see M. Luntley, Contemporary Philosophy of Thought: Truth, World, Content, Oxford: Blackwell, 1999, esp. chapters 11, 12. 26 Cf. especially §5 of J. McDowell, ‘Singular Thoughts and the Extent of Inner Space’, in Subject, Thought and Context, eds J. McDowell and P. Pettit, Oxford: Clarendon Press, 1986, pp. 137–68. 27 Wright’s work on Wittgenstein focuses almost exclusively on the problem of the infinitary character of the logical must: cf. C. Wright, Realism, Meaning and Truth, Oxford: Blackwell, 1986, and Wittgenstein on the Foundations of Mathematics, Cambridge, Mass.: Harvard University Press, 1987. As I argue below, I do not believe that there is a further problem beyond the one already dealt with. The infinitary character of the arithmetical rules is simply a dramatic way of representing something already present in the case of games and other concepts.

CHAPTER FIVE

Putting Your Self in the Picture

5.1

Introduction

Wittgenstein’s private language argument has probably generated as much secondary literature as any other argument in the history of philosophy. It might seem outrageous to propose yet another account of what is going on in this argument. Nevertheless, that is what I shall do. It is common to treat the private language as an isolated argument that precedes a more general discussion of the inner/outer distinction.1 That is a mistake. The position that I shall extract from Wittgenstein, and it is one that he articulates only with great difficulty, is fundamentally a position that denies the inner/outer distinction. The private language argument is part of a more general attack on the very idea of there being a distinction between inner and outer sense. Wittgenstein often misrepresents the power of the argument that he has, for he is, despite his onslaught on the inner/outer distinction, too often in the grip of an epistemological mode of enquiry that threatens to reinstate that distinction.

5.2 The Standard Treatments The standard treatments of the private language argument take it as a critique of a Cartesian model of meaning. The critique shows that an essentially private language is impossible, for a private linguist would have no resources for calibrating the use of sensation words. On the standard treatment, the calibration of sensation word use is essentially provided by public patterns of use.2 The claimed impossibility of private calibration comes in at least two different versions.


125

For some commentators the impossibility of private calibration arises from the instability of private meanings. The thought is that without public criteria against which to check the use of words, the private linguist would have no account of what made use on one occasion the same as that on a previous occasion. There are passages in the text where Wittgenstein makes this point, e.g. at §271, but this is after the core passages for articulating the private linguist’s problem with calibration.3 On the stronger reading, the private linguist fails to establish any standard of correct use at any time, let alone for that use to be stable over time.4 Wittgenstein says sensations require some outer criterion of correctness – ‘An inner process stands in need of outer criteria’ (§580). But everything hangs on how we read ‘outer’ here and whether we take Wittgenstein to be endorsing the contrast between inner and outer, or whether he is using it as part of a strategy that critiques it. Those who see the private linguist’s problem as one of stability, take stability to be provided by the public outer criteria. Commentators who emphasize the stronger reading tend to represent Wittgenstein’s alternative account of meaning in more abstract terms without a clear endorsement of the idea that the meaning of sensation words is essentially public.5 In either the weaker or stronger versions, the standard reading has it that the Cartesian subject could not have the resources for the idea of standards of correct use for sensation words. Without an account of standards of correct use, there is no account of the meaning of utterances such as, I am in pain. To place the force of the argument on the isolation of the Cartesian subject is doubtfully coherent. Wittgenstein stresses that he is not denying that there is such a thing as the experience of pain – ‘Why should I deny that there is a mental process?’ (§306). And yet, according to most commentators this mental process is tied to outer criteria that are essentially public. This sounds as if he is admitting both the existence of the inner and also insisting that it is constitutively connected to the outer. The puzzle then is this: How can the connection to the outer be constitutive without destroying the acceptance of the inner? Conversely, how can he have a credible notion of the inner if it is constitutively tied to the outer? Clearly, the answer to this has to be that Wittgenstein is, in some useful sense, denying the inner/outer distinction, despite his endorsement of the reality of painful experience. There is a coherent way of accounting for his argument, but simply to say that an essentially private conception of the inner is impossible and to say that the inner is constitutively tied to the outer really amounts to no explanation at all.6 For sure, Wittgenstein’s pronouncements on this puzzle are, at best, opaque:

126


‘ ‘‘Aren’t you at bottom really saying that everything except human behaviour is a fiction?’’ – If I do speak of a fiction, then it is of a grammatical fiction’ (§307). He states the paradox at §304: And yet you again and again reach the conclusion that the sensation itself is a nothing. – Not at all. It is not a something, but not a nothing either! The conclusion was only that a nothing would serve just as well as a something about which nothing could be said. We have only rejected the grammar which tries to force itself on us here.

And then he responds, The paradox disappears only if we make a radical break with the idea that language always functions in one way, always serves the same purpose: to convey thoughts – which may be about houses, pains, good and evil, or anything else you please.

This response tells us nothing and is, furthermore, at odds with most commentators’ accounts of the argument. Wittgenstein advises us to break with the grammar that forces itself on us. It is not clear what he means by ‘grammar’ at this point. The obvious reading is the one suggested forcefully in the second paragraph, namely, the idea that sensation words do not function like names and sentences like ‘I am in pain’ do not serve to express how things are (express a thought). There is, of course, a strong tendency in Wittgenstein’s own writings to endorse a non-cognitive theory of avowals. This is the theory that holds that ‘I am in pain’ is not a possible object of knowledge. It has been thought that Wittgenstein holds this because the sentence does not express a genuine proposition.7 The idea that such a sentence fails to express a proposition fits with those passages where Wittgenstein comes closest to endorsing a behaviourist account of sensations: ‘the verbal expression of pain replaces crying and does not describe it’ (§244). In being instructed in this new expression the child learns ‘new pain-behaviour’ (ibid.). The non-propositional account of avowals expresses the idea that sensations are not properly integrated in our conception of how things are. Sensation words do not report on, or describe, how things are; at best they express our experience of how things are.8 The problem with taking §304(b) as endorsing the non-cognitive theory of avowals is that it takes away the applicability of the critical concept that forces the argument against the private linguist. The central problem for the private linguist is the impossibility of calibrating sensation word use. That critique only works if sensation word use stands in need of calibration, for otherwise they cannot be seen as subject to standards of correctness and incorrectness. The key claim at the end of §258 is that the private linguist


127

lacks the resources to make a right/seems right distinction. For that claim to have force, sensation word use must be the sort of word use for which there is an issue about the source of its calibration. §304(b) cannot, then, be an endorsement of the non-cognitive theory of avowals. Perhaps Wittgenstein’s thought is the simpler point that not every occasion of sensation word use performs the same purpose? That is to say, the claim in §304(a) is exactly the claim that Wittgenstein makes, that a nothing would serve just as well as a something about which nothing could be said. This does not mean that sensations in general are nothing. It means only that the attempt to conceive of sensations as the private linguist does, is to conceive of nothing. The reason for this is that the private linguist’s conception is a conception for which no calibration of sensation words can be provided. The text is problematic, but then so too are the issues with which Wittgenstein is grappling. In the passages around this difficult section, Wittgenstein again and again rejects the accusation of behaviourism. Often, his remarks look like a manifestation of a non-cognitivist expressivist account of sensation words. For example, ‘Pain-behaviour can point to a painful place – but the subject of pain is the person who gives it expression’ (§302(b)). Here we find Wittgenstein acknowledging the experiential reality of pain – a painful place, but also allowing the idea that the person gives pain expression. This latter point could be read along the lines of the non-cognitive theory in which the use of sensation words simply replaces the natural sign, the cry. It could, of course, also be read as endorsing the idea that an inner pain is expressed in outer behaviour, but that would be difficult to attribute to Wittgenstein. There is, however, a much simpler version of what this passage means and it can help clarify the otherwise complex knots in which he appears to tie himself as he tries to affirm the reality of experiential pain and, in denying the Cartesian model, avoid the rebound to behaviourism. The simpler reading is to say that ‘expression’ picks out the common-sense idea that sometimes, how things are for another can be transparent in what they do and say. That is to say, ‘expression’ does not refer to a translating of the felt sensation into a public language; it is the direct revelation of how one’s experiential states are going. Expression is making, or better, allowing oneself to be transparent.9 When Wittgenstein says, ‘The human body is the best picture of the human soul’ (p. 178), he is not saying this because it is the best picture we can get, as if there could be a better one. And he is not saying this because the soul is always hidden deep within the body and at best only poorly translated into the condition of the body. He is saying that how things are for you, how your psychological life is going, can on occasion be just how you show it to be in what you do. ‘If I see someone writhing in pain with evident cause I do not think: all the same, his feelings are hidden from me’ (p. 223).

128


In short, the concept of expression that Wittgenstein needs in §302(b) is the concept that allows that sensation, feelings and other components of our mental life are not necessarily hidden. They are, nevertheless, components of our felt lives. We are tempted to think of this felt life as an inner life. But if it can be expressed in the manner I suggest, then the very distinction between inner and outer lives has been rejected, at least in the terms in which it is normally understood. The rejection of the inner/outer distinction requires the affirmation of the concept of a living human being as a primitive concept. When Wittgenstein says, only of a living human being and what resembles (behaves like) a living human being can one say: it has sensations; it sees; is blind; hears; is deaf; is conscious or unconscious. (§281)

he is not endorsing behaviourism. He is endorsing the idea that our mindedness is expressed in behaviour because it is not hidden. It is not hidden, not because it has been translated into something less than a full-blooded conception of what it is like to have sensations, but because he has rejected the inner/outer distinction that renders such things epistemically privileged. And so, ‘Only of what behaves like a human being can one say that it has pains’ (§283(e)). In order to understand Wittgenstein’s critique of the inner/outer distinction and the role it plays within his philosophy, not to mention how the private language argument functions in this context, we need to unearth the assumptions that make this area so problematic. This is, I suspect, what Wittgenstein was doing, but I want to extract the points in a different order to the way they appear in Wittgenstein’s writings. The starting point is to clarify what, if anything, is special about sensations.

5.3 What’s Special about Sensations? The difficulty with sensation word use is to acknowledge both that there is something special about sensations, in so far as they draw our attention to something characteristic of our mindedness, and yet also to realize that, in another (epistemological) sense there is nothing special about them. I take the latter point first. One reason we are tempted to think sensations are special is the idea of the hidden. We are tempted to think of sensations as things behind the surface behaviour. But why? What is the epistemology that puts sensations out of the epistemic reach of others? It is not obvious, nor is it a simple deliverance of


129

common sense, that they are out of reach. Someone shuts their finger in the door. They scream. Blood drips to the floor. We rush to their aid. Why is that not a simple unproblematic case in which we directly see that they are in pain? Wittgenstein clearly thinks that such cases are unproblematic (p. 223). Why should sensations be epistemologically special? You might think that there is a problem because it is possible to feign pain. But that just shows that we might be mistaken when we see another in pain. But the possibility of error applies to all empirical claims, it is not distinctive of sensations. Again, Wittgenstein makes precisely this point, ‘What is internal is hidden from us.’ – The future is hidden from us. But does the astronomer think like this when he calculates an eclipse of the sun? (p. 223)

It is because there is this possibility of mistake that we are confident that there are standards of correctness/incorrectness for such claims. They are claims to be calibrated. You might think that the difference with sensations is that they are doubly susceptible to sceptical doubt, for even if we closed off the possibility of doubt with respect to empirical claims about the external world, we would still be left with the possibility of doubt with regard to others’ sensations. It is this epistemology that goes unchallenged by Ter Hark, Schroeder and others. That response begs the question, for it only works if the inner/outer distinction is already in place. And why should that be the case? Furthermore, it is not clear what it would mean to say that the possibility of doubt had been closed off in the case of empirical claims about the external world. We need to go back to epistemological basics if we are to get clear about what is special about sensations. The general epistemological thesis that I urged we should reject in the last chapter was: What is available to you as the content of experience is always less than what is the case. Rejection of this amounts to allowing that it is possible for experience to provide us with exactly what is the case. It is, of course, also possible for experience to get things wrong. The fact that both are possible and, furthermore, that as subject of experience you cannot tell for sure which one is which, just shows that as subject you cannot tell for sure which experience you are having. If the content of experience e1 is a direct experience of a cat sitting in front of you, and if the content of experience e2 is of the appearance of a cat sitting in front of you when none is there, then e1 and e2 cannot be the same. This shows that there is a

130


right/seems right distinction that applies to the character of experience. The point about the possibility of doubt is simply that you cannot with certainty tell the difference between these two experiences. You can be fooled into thinking that they are the same. That, of course, is what you have always acknowledged about the world: you cannot with certainty tell the difference between cases when there is a cat in front of you and cases when there is no cat. Rejection of the general epistemological thesis simply pushes this inability back onto the differentiation of experiences. And it also insists that although you will say that normally you have no difficulty in telling whether you are having an experience as of a cat or one that is merely an hallucination, the same point applies to the real state of affairs. You can, most of the time, tell perfectly well that there is a cat in front of you and when there is not. The epistemological fragility of others’ sensations states is not, without already begging the question about the inner/outer distinction, something that shows that sensations are epistemologically special. Without independent reason for drawing an inner/outer distinction, the epistemic fragility of others’ sensations is no more and no less than that of other empirical states. One of the things that Wittgenstein is doing is retrieving the idea of our mindedness. He is retrieving the idea of mindedness from the inner conceived as a place of special privileged access and placing it as part of the empirical. Our mindedness is part of how things are. And to say that mindedness is part of how things are is not necessarily to place it in the ‘outer’, for the outer does not make good sense without the contrast with ‘inner’. It takes a peculiarly Cartesian conception of mindedness to evict it from the regular empirical realm of how things are to the privileged realm of the inner. The regaining of our mindedness is a regaining of the right to count our mindedness as part of the simple basic unadorned realm of how things are. On such an account, being in pain is part of how things are, something that is naturally observable and as naturally doubtable as all cases of how things are. The natural way of taking an experience as of a cat is to take it as just that – an experience of a cat. Most of the time we are right. The natural way of taking an experience as of someone in pain is to take it as just that – an experience of someone in pain. This is a tempting epistemology and one that Wittgenstein explicitly endorses. It can seem, however, that it fails to do justice to our intuitions concerning the reality of felt experience. Disavowing the epistemological sense in which sensations are special can seem to threaten the idea that they are special because characteristic of our felt experience. The disavowal of the epistemological specialness seems to leave out the ‘feels’? There is, of course, need to observe strictures that Wittgenstein clearly articulates: regaining our mindedness as part of how things are is a regaining of the idea that the fact that one is in pain can be a potential item of knowledge


131

for others. It is not the idea that one’s pain can be a potential experience for others, or a potential pain for others. The availability to others is the availability of the fact that you are in pain, it is not the availability of your pain. Others cannot have my pain, they cannot feel my pain and they certainly cannot see my pain. It is not my ‘feels’ that are available for others to inspect. Indeed, no one inspects them, not even me. I have them. What is at issue is the claim that others can know that I am feeling a certain way. The claim is that others can see that I am in pain. Nevertheless, this common-sense epistemology of sensations, endorsed by Wittgenstein at §246, seems to miss out something deep about our mindedness. I think this worry is correct, but the point missed out thus far is not special to sensations. We still need to capture the first sense in which sensations and our mindedness in general appear special. Consider a simple empirical fact, the cat is sitting on the mat. Now, the natural thing for a cat to do is to occupy space and time in just the space and time in which it exists. This observation of what is natural for cats sounds odd and pointless, but this is only because cats cannot fool us with regard to where they are. They cannot lie. There are no cat illusionists. The same applies to all natural objects. With minded things – people/living human beings – it has always seemed tempting to say that what makes them different is that rather than being simply a thing in the world, they are a point of view on the world. Perhaps, in addition, the point of view is also in the world, but the notion of point of view makes it a funny sort of way of being in the world and a way quite unlike that of cats, rocks and trees. Indeed, so difficult is the concept of the point of view being in the world that Wittgenstein, in his early writings, relegated the point of view to the limit of the world. It is the idea of a point of view that captures what is special about our mindedness. Let us take seriously the thought that living human beings enjoy a point of view. This is what makes knowledge of them more problematic than knowledge of things without a point of view. Given this, there is a distinction between two different sorts of mistake that we can make about the world. In the case of finding out about the cat on the mat, there is only one sort of mistake. It is a sort of mistake that is due to us, the perceiver. Cats cannot fool us about where they are. With minded things, there is, however, another sort of mistake that is possible. It is the mistake that comes about when although our perceptual apparatus is working fine, the other person is presenting in a false way. The temptation to which we have often succumbed is to treat this as a hidden, for that is the simplest model for the trick. The notion of ‘hidden’ here is not normal, it has to be learnt (the same point applies to the ‘hidden’ with natural objects when things look other than they are – mirages, etc.). That in itself is a minor point, it provides security but does not address what is really going on. With regard to animals, we do not have space for the

132


second sort of error. Animals cannot lie (§250). But they can have sensations and we have no trouble in counting their having sensations as part of the ordinary realm of the empirical – how things are. With living human subjects, their sensations enjoy a kind of ‘hidden’ which is deliberate, or which takes learning and the capacity for deliberation. But that is not something peculiar to sensations. It is due to the fact that living human beings have a point of view. The point applies also to bodily presentations by others as performed by illusionists. The real state of their body is ‘hidden’ but it is not private. It is not problematically tucked away in an inner realm, even if, in some metaphorical sense, it is out of sight. Being ‘out of sight’ cannot then mean ‘in the mind’. The concept of being ‘out of sight’, of being capable of being epistemically hidden, is a matter of being within a point of view. It applies to bodily properties just as much as psychological properties. It is not the fact that sensations are felt that renders them subject to a sceptical doubt beyond that due to ordinary empirical facts. It is that sensations are experienced by subjects with a point of view that renders them harder to access epistemically than empirical claims about matters not within the purview of a point of view. It is being within another’s point that makes a state epistemically ‘hidden’ and that applies just as much to bodily states as psychological states. There are two sorts of mistake that we can make about empirical states, although the first type can occur in two different ways. The first kind of mistake concerns a deviancy with respect to the way we perceptually connect to what is just as it is. These are cases where the deviancy is either due to something going wrong with us, or it is due to the non-standard environment in which we and the object of perception are placed – e.g. when we see a mirage. Quite rightly, we feel no impulse to account for such a notion of mistake in terms of a hidden. The availability of mistakes of this kind arises from the different ways in which we can lose touch with things. The fact that there are such ways in which we can lose touch with things does not question our fundamental capacity to be in touch with things when, for example, matters are right. Denial of the general epistemological thesis is simply the endorsement of the view that the difference between being in touch and being out of touch turns out to be a difference between two different experiences. Denial of the general epistemological thesis means that experience is not characterized as the invariant between the two cases. The worry that looks to be left unaddressed by Wittgenstein’s endorsement of the comfortable common-sense view of our knowledge of others is that sensations introduce a different kind of error with respect to our perception of how things are with others. But there is no good reason for that thought. The sensations of animals do not present in this way. They do not present in a way that leaves us in principle out of touch with how things are going for


133

them. This is because animals cannot lie. They wear their sensations on their sleeves, not because they are naive and trusting, but because that they are in pain is not something for which there is any reason to treat as part of a realm of the hidden.10 The second kind of mistake available with empirical facts comes about because, with living human beings, our perceptions of how things are with them can be deliberately thrown. In the case of non-human sensations, much as with ordinary natural phenomena (mirages, etc.), we can, as it were, iron out the errors. We can move closer, kick the cat a little harder, move our head to one side to confirm position, use touch to see if the presented object is at the location sight presents it as occupying, etc. It can be difficult to do this, but in principle it is possible. In the case of our perception of how things are with others, we cannot provide this in principle guarantee to ‘iron out’ the distortions, because of the possibility that they are deliberate. The distortions might have been willed by the other. In the case of the first kind of mistake, we can iron out the distortions because they concern things independent of will, brute resistances to our will that remain stable as we bump against them and plot their position, contour, character, etc. But as we bump into and engage with other wills, they can move and wilfully change their presentation. There is a possibility for a concept of hidden here, but it is not a hidden that is special to sensations. If we use the adjective ‘inner’ to label it, the inner that is relevant is the inner of the other’s point of view. But ‘point of view’ is simply the attitude of the will – how the world strikes the will and that includes much more than sensations, it includes bodily states too. Of course, if ‘point of view’ is understood in terms of qualia, and qualia are conceived as private inner ‘feels’ epistemically inaccessible to anyone other than the subject of experience, we reinstate the epistemological sense of specialness alongside that of the notion of a point of view. That is Nagel’s concept of a point of view. I am suggesting that Wittgenstein is critiquing that concept of a point of view. We need to detach the idea of a point of view from the idea of inner qualia. This is not to deny the idea of a point of view, nor to deny that how things feel, are seen, heard, etc. fall within the point of view. It is only to insist that the idea of a point of view is recoverable in terms of the attitude of the will to that which is independent of will.11

5.4 The Need for Calibration If the above is right, what disables the private linguist? Why is it that the private linguist cannot have a right/seems right distinction? The simplest way of taking the disabling effect of the idea of a private language is to say that the

134


very idea of right – the idea of correctness conditions for language use – is social. As it stands that cannot be right, for it begs the question against the Cartesian. For sure, it might turn out as the result of analysis that the Cartesian is mistaken in thinking that a right/seems right distinction can be set up in isolation of the social, but it cannot be that simple that the social just is what provides the notion of right. We need some account here of why the private linguist is disabled. The central claim is: C It is not possible for the private linguist to calibrate words. Suppose a word w is subject to standards of correct/incorrect use; it is calibrated. There have, traditionally, been thought to be two options for an account of calibration, either: (a) w is calibrated on a thing; or, (b) w is calibrated in patterns of use. Option (a) looks to have a robust concept of friction, but it depends on a notion of primitive acquaintance between word and thing and that requires considerable input on the notion of ‘thing’. The point is that mere acquaintance with a thing does not provide the normative engagement that produces standards of correctness. Merely holding a sign against a thing does not calibrate the sign. So the notion of thing has to introduce not only a robust enough concept of object, it has to produce an object already suited to supply the idea of normative standards of correctness. It is a mystery how this could be achieved without already presupposing the concept of grammar.12 The critique of (a) is the critique against ostensive definition: And when we speak of someone’s having given a name to pain, what is presupposed is the existence of the grammar of the word ‘pain’; it shews the post where the new word is stationed. (§257)

But the idea of ‘post’ or grammar throws us back on option (b). On option (b), it is unclear what provides the friction to calibration. If the answer is that the friction is supplied by a use that is essentially social, we are left with no account of why social use stands without the need of calibration that is supposedly problematic for private use. The point is not answered by


135

appealing to the undoubted epistemological stability that comes from checking one’s practices against those of others. The epistemological advantage of checking one’s practices against others depends on the others being taken as already calibrated. It is a move that presupposes calibration is already in place, it cannot be a move that is constitutive of calibration. The claim against the private linguist is not then that they are forgetful and need the help of others to get a sense of what they are doing. The claim against the private linguist is a constitutive claim that it is impossible for them to have a calibration of word use. If a notion of social use is to work as a response to the private linguist’s problem, it must then be a notion of social use that is constitutive of what it is for word use to be calibrated. But that’s the point that looks question begging. The social is no better placed as constitutive of calibration than the private. Neither option (a) nor (b) give adequate outline accounts of calibration. There is, of course, a middle way: (c) w is calibrated in patterns of use that manifest a grip on that which is independent of will. Pears has a version of (c) in which patterns of word use manifest a grip on that which is physical.13 The problem with this is that it looks like it begs the question against the Cartesian, for it is not clear what the status is of the appeal to the physical. I prefer the more abstract formulation at (c) which is enough on its own to produce a version of the private language argument that shows that the private linguist cannot have a concept of calibration for sensation word use. Most commentators treat the private language argument as claiming that calibration by (a) is impossible, it can only be provided by (b) where this is understood as grounded in the social. As noted, there have been two sorts of interpretations of the argument in this vein. The old version is the one in which the key claim is that without (b) type calibration, the private linguist will not be able to remember the meaning of a sensation word. The old version misreads §258. The better version of the standard reading says that calibration by (a) is impossible in the original case; it is not that the private linguist cannot remember the meaning of a sensation word without (b) type calibration, it is that no baptism of meaning ‘in the present case’ is possible without (b) type calibration. Although that gets the text of §258 right, it also puts pressure on the idea that (b) has to be social, for if the social calibration is required ‘in the present case’ what becomes of the possibility of feigning? What becomes of the notion of ‘inner’ that is not an internal theatre, but simply how things are for a point of view? It is possible that how things are for

136


a point of view are presented in a way other than they are (and that means they will be detached from how things appear to others about how things are for that point of view), simply because the notion of a point of view is that of a will, and we may mispresent ourselves. This means that the newer version of the standard private language argument has to rely on what looks like an empirical claim. The empirical claim would be that the private linguist cannot misrepresent without first correctly presenting. This is an empirical claim for it must be distinguished from the stronger conceptual claim that the possibility of mispresenting presupposes presenting correctly. That conceptual claim is not sufficient for the version of the private language argument under consideration. The version under consideration needs the empirical claim that a private linguist could not start by mispresenting and save correct presentation for later. That is empirical, for what is there to stop the private linguist acknowledging that mispresenting presupposes presenting correctly, but they know what presenting correctly is from their own private case? Let us consider the claim in §258 that the private linguist cannot establish standards of correctness in the present case. Why is the private linguist disabled in this way? There is a problem with the language in this area. It is very difficult to avoid language that looks like it reinstates the inner/outer distinction, even when one is clear that the point of the argument is part of a general attack on the very idea of that distinction. Pears gives a middle-way account of why C is true. He says the private linguist needs a concept of ‘discoverable error’.14 But the obvious question is ‘discoverable to whom?’ There is a danger of begging the question, for why should the private linguist not say that error is discoverable, discoverable to him? The role of physical criteria in Pears’ account of the argument raise a similar problem.15 Pears says that the private linguist is trying to set up a description of sensation detached from the external world, but why should the private linguist concede that the external world is the measure of things?16 Pears’ account is instructive, but it raises as many questions as it solves. On one point, however, he is absolutely right. He notes that the deficiency of the private linguist can be seen either as an inability to check judgement against that of others (social (b)), or an inability to check judgement against standard physical objects.17 The latter is a version of option (c) in which the notion of ‘independent of will’ is construed physically. I think the role of the physical here begs an important question against the private linguist. It does not push the argument to its full length. Nevertheless, Pears is right to note that there are two distinct sources of calibration: the community and what the community checks against. The former is only of value on the assumption that the community is itself calibrated against the latter. I disagree with Pears only on the issue of how to characterize the latter.18 That there is a distinction between what provides


137

calibration and that which can be used to test for calibration (the community) is something that Wittgenstein notes on plenty of occasions. There is also clear evidence in the Nachlass that Wittgenstein was aware of this distinction.19 We need then a better account of what the source of calibration is and why the private linguist cannot achieve calibration. McGinn’s treatment is suitably abstract but ultimately unclear. She speaks of ordinary language games as fixing meaning, fixing what kind of thing pain is.20 She says that the private linguist looking for calibration will look ‘at the wider context of the speaker’s actions’, but it is unclear how much goes into the concept of ‘wider context’, nor why it is required. The key concept for McGinn’s interpretation is that of a ‘grammatical investigation’, but McGinn never says clearly what a grammatical investigation is. For sure, a grammatical investigation involves an account of the normative patterns of language use. But the present issue concerns what constitutes having a language use that is normatively configured. Why is the private linguist disabled from having such a language use? That is the question for which there is, as yet, no fully satisfactory answer. McGinn says, correctly I think, that the anti-introspectionism of the private language argument is not necessarily a move that is constitutively pro-public.21 This is important, for it marks the need not to beg the question against the private linguist and merely assume that having language use subject to normative standards of correctness is to have use subject to public standards. McGinn does not, however, say what must be the case for correct use of sensation word. She just says that correct use cannot be conferred by introspection. That leaves us without a clear account of what the disabling effect is of the private linguist’s conception of language. It does not identify in a non-question begging way what the private linguist lacks.

5.5 Calibration in Subjectivity Our central question is this: Why, if it is true, is C true? Why could there not be calibration within subjectivity? Russell seemed to countenance such an idea, for his concept of the atomic elements of language – the genuine proper names ‘this’, ‘that’ and ‘I’ – were only calibrated in subjectivity. And for Russell, subjectivity was a private affair. Pears thinks that detachment from the external makes calibration impossible. His account of the external (and that word already begs questions against the private linguist) is in terms of the physical. So Pears thinks that detachment from the physical makes calibration impossible. But why is this so? Suppose Pears is right, we still need an account of why detachment from the physical is so disabling. It is not obvious that it is disabling. It is especially not obvious to a Cartesian!

138


We need to distinguish two different accounts of why detachment from the physical is disabling of the idea of normative standards of correct word use. The first account assumes: (i) the physical ¼ the objective. If this is right, it gives short shrift to the Cartesian attempt to construct a private language. It begs the question against the possibility of the Cartesian getting started and amounts to little more than a trading of intuitions with the Cartesian concerning what is possible. The assumption of (i) does, however, make plain that there is no constitutive role for the social in an account of the source of the normative standards of word use. The second account assumes: (ii) the physical gives us a basic purchase on the objective. The point to (ii) is that the physical plays an important role not in constituting our concept of the objective, but in giving us our primary grip on, our first purchase on, the concept of things independent of will. The conception of physical things is important but not because it is a conception of things external. That begs the question against the Cartesian project of having a conception of the inner as private. The importance of the physical is that it is our most basic sense of things that are independent of will – the world. The underlying concept of how things are is that of things independent of will and the physical gives us our most basic purchase on that idea. Detachment from the physical is disabling if the above thought is right that the physical, although not constitutive of how things are, provides our basic purchase on the concept of how things are. But why should the private linguist not respond to C and provide calibration from within subjectivity? The answer to this is that calibration within subjectivity is only possible if subjectivity can provide a concept of how things are – things independent of will. It might seem that we now return to our basic and unargued for intuitions. The private linguist will insist that a conception of how things are can be provided from within subjectivity and the critic will insist that the idea of things independent of will is not available on this basis. But we can take the analysis further, for I think the critic is not wholly right. Calibration – a grasp of things independent of will – can be achieved from within subjectivity. But once we are clear what subjectivity is like for this to be possible, then not only do we get a full account of what disables the private linguist, but we reinforce Wittgenstein’s retrieval of our mindedness into a seamless conception of the world – how things are.


139

The private linguist needs to show that calibration can be achieved within subjectivity. The critic needs to show that this is impossible without begging the question and without losing touch with the idea that there is something special about subjectivity. Thus far, I have argued that what is special about subjectivity is the idea of a point of view, where this is understood not in terms of qualia, but in terms of the attitude of the will. The latter way of thinking about subjectivity accommodates the common-sense epistemology of our knowledge of how things are with others, for it leaves the notion of how things are for a subject as part of the normal empirical realm of how things are. This move is often missed, but following the rejection of the general epistemological thesis first identified in the previous chapter, it is a position that is available. Let me clarify this position by distinguishing between two concepts of subjectivity. I shall say that the position implicated in the earlier discussion of the retrieval of our mindedness as part of how things are provides for a concept of simple subjectivity: Simple subjectivity – there are truths about how things are for me; the ‘forme-ness’ is part of how things are. So, when I am in pain, my being in pain is part of what there is. That it is part of what there is, is due to it being something that obtains independent of will. That makes my being in pain part of what is going on in the world. Because my being in pain is something that obtains independent of will, it is in principle simply there in the world as something independent of others’ wills. As such, there is no principled reason why it is not epistemically available to others. I might disguise my pain. I might pretend and feign a pain. I might present the for-me-ness of my experience in a way that misleads. But none of this shows that it cannot also be the case that how things are for me is, as part of how things are, simply available for others to know as well, but also as falteringly, as they know anything. There is nothing special about sensations with respect to their potential for openness. Being in pain is not something I can will. The expression of pain is something that is subject to will with training. Simple subjectivity allows the possibility that the subject can get a conception of things independent of will within subjectivity. Calibration within subjectivity is available for the simple subjectivist. This does not, however, underpin a private language as Wittgenstein sets that up in §243. For the simple subjectivist the calibration of words within subjectivity is available for the simple reason that subjectivity is conceived as part of how things are. Subjectivity, on this conception, is nowhere other than firmly in the world. In

140


order to get the conception of subjectivity that the private linguist requires we need a richer conception of subjectivity. The obvious characterization of this is provided by the analysis of the Cartesian conception of subjectivity provided by McDowell.22 Cartesian subjectivity is more than the simple admission that the notion of how things are for me introduces subject matter that is apt for truth. Simple subjectivity amounts to the acknowledgement that, in addition to the truths to be had about how the world is going, there are also truths to be had about how things are for me. But simple subjectivity does not make the notion of truth that applies to my point of view any different to that which applies to the rest of the world. This is why subjectivity is rendered opaque for the simple subjectivist and the general epistemological thesis is rejected. We retrieve our mindedness as part of how things are and thereby acknowledge that we can, in principle, be mistaken about our own mindedness as we can be mistaken about things outwith our point of view. It is the epistemic opacity of subjectivity that is the key mark of the simple subjectivist conception.23 Replacing that opacity with transparency gives us the full-blooded Cartesian conception of subjectivity. Cartesian subjectivity Simple subjectivity þ transparency. This is the concept of subjectivity at play for the private linguist. It is the conception of subjectivity in which, in addition to there being truths about how things are going for me, these truths are knowable through and through with certainty by me. For the Cartesian conception, how things are for me is delimited as a range of truths knowable transparently. It is this that detaches subjectivity from the rest of the world. The rest of the world is opaque, but for the Cartesian the point of view is transparent and thereby not part of the world. Can the Cartesian subjectivist calibrate words in subjectivity? That is now the key question behind the private language argument. The point against the Cartesian is not that their conception of subjectivity is detached from the rest of the world. That is simply definitive of the Cartesian conception and it is no objection to say that their attempt to calibrate sensation words within subjectivity is problematic because detached from the external physical world. That cannot be an objection, for it simply marks the distinctive two-worlds picture of the Cartesian. The objection formulated in the private language is more abstract, but more powerful, than that. The objection is that the very idea of things independent of will is not available from within Cartesian subjectivity. The key to our conception of the objective world is not that it is external, nor that it is physical, but that it is independent of will. It is the entitlement to that latter idea that is


141

threatened by the private linguist’s attempt to calibrate words within Cartesian subjectivity. If the private linguist is not entitled to the basic notion of things independent of will, then their attempt to calibrate words and construct a private language is disabled. The private linguist fails because of the transparency of subjectivity. It is because this is a conception of subjectivity that is knowable through and through with certainty that the private linguist has no conception of things independent of will. Within Cartesian subjectivity the subject cannot make a mistake. The layout of subjectivity is not a matter on which the subject can get things wrong. There is no notion of what it is to get things right and wrong here. Everything just is exactly as it seems to be. By treating appearances as not merely part of how things are (simple subjectivism), but as things known transparently, the Cartesian treats subjectivity as a realm to retreat to when we get things wrong with other matters. But the retreat is illusory, for the place retreated to is no real place at all, for there is no conception available of what it would be to be mistaken about this place. Without the concept of a mistake, there is no right/seems right distinction. There is no conception of subjectivity as being part of how things are, for there is no conception of how it can be part of things independent of will. That and that alone is Wittgenstein’s point in the celebrated diary example §258. It is not the loss of the external world or the loss of the physical world that disables the private linguist. It is the loss of the conception of things independent of will. The move that repays this loss is not then the move to see subjectivity defined in terms of the outer, the social, or the physical. It is the move that sees subjectivity as part of how things are, things independent of will. And that is the retrieval of our mindedness for which I argued earlier. It is simple subjectivism. The account of the private language argument that I have now defended is not fully complete. The matter is not fully settled against the Cartesian conception of subjectivity. Thus far, I have argued that given the transparency of Cartesian subjectivity, no conception of things independent of will is available to the Cartesian private linguist. It is worth noting that there is a reading of Descartes that has him implicitly aware of this problem. Noting Descartes’ answer secures a firmer sense of the status of what is at issue. The problem for the Cartesian conception of subjectivity is to show how the idea of things objective (independent of will) is so much as possible from the Cartesian subjectivist starting point. Descartes was, I suspect, aware of this problem. He needs to give an account of how the resources available within Cartesian subjectivity can supply a conception of objectivity. He attempts to supply the answer to this with the ontological argument. In effect, the role of God for Descartes is not, I suggest, the role of guarantor or underwriter to our

142


epistemic challenge to regain knowledge of the outer world. God’s role is more fundamental than that. It is to define the very idea of objectivity. The objective (independent of will) for Descartes cannot be identified with the outer external world, for the obvious reason that Descartes thinks that is but half the story about how things are. He needs then a more fundamental concept in terms of which he can define the notion of objectivity that applies to both the inner and outer. I suggest that the objective for Descartes is that which is available to God’s point of view. It does not matter much at present whether we think this idea is intelligible. The point is simply that Descartes requires some such concept, for he must be able to give an embracing conception of objectivity if his metaphysical dualism is to be defensible. On this reading, the ontological argument is required to underwrite Descartes’ conception of objectivity. If the ontological argument worked, then the Cartesian private linguist could, in principle, calibrate words within full-blooded Cartesian subjectivity, for the calibration would be grounded in the conception of things independent of the private linguist’s will but dependent only on the will of God.24 This shows that the private language argument is not, as often interpreted, a straightforward reductio ad absurdum. It is a challenge. It is a challenge to the private linguist to provide a conception of things independent of will from within the resources of the Cartesian conception of subjectivity. If that challenge is not met, we have in place of the Cartesian conception, the simple subjectivism that obliterates the inner/outer distinction. That distinction is only held in place by the Cartesian conception of subjectivity as transparent. With that gone, subjectivity enters the world. But we abandon the Cartesian conception of subjectivity not because it is absurd, but because we do not believe the ontological argument works.

5.6 Agreement in Forms of Life The reading that I have provided of the place of the private language argument in the denial of the inner/outer distinction is at odds with most of the secondary literature on Wittgenstein. My reading is a highly individualistic reading. The conditions for the possibility of judgement are given in terms of the attitude of the self-as-will to that which is independent of will. Our basic purchase on this may be provided by our sense of our attitude to physical things, but there is nothing to preclude that we can take this self-same attitude to our own being in the world. There is a self-reflexive possibility to this attitude. Have I then simply ignored those places where Wittgenstein invokes the public character of the calibration of words? What about those places where Wittgenstein invokes the social?


143

If we look closely and with due consideration of the foregoing analysis I suggest we find that there are no places where Wittgenstein invokes the social.25 We need to observe a distinction between the role of the others as scaffolding for, or enabling of, our engagement with the world, and the role in which others are constitutive of what we engage with. There is nowhere that Wittgenstein endorses the latter view. The temptation to read certain passages as endorsing the latter constitutive role for the social is, I suggest, a function of the underlying epistemology that commentators, and sometimes Wittgenstein, fail to shake off. This is an epistemology that fails to accommodate the dissolution of the inner/outer as traditionally conceived. Very often, the secondary literature is ambiguous. For example, McGinn treats our language-game with ‘pain’ as what fixes the grammar of the word and says there is nothing else that can determine a criterion of identity for pain.26 On my reading of the argument, the concept of simple subjectivity gives an account of the subject’s perspective that allows that there can be something else to fix the grammar of pain, namely the way your perspective is which is, like any other empirical fact, part of how things are. But even disregarding that point for the moment, McGinn’s claim is ambiguous between the language-game determining the thing or property picked out – pain, and determining the way we pick it out. Does grammar determine sense or reference? In addition, it is unclear why grammar should not be available for the individual. Of course, on my reading, that is the case, but for the simple subjectivist the individual’s perspective is not out of touch with things. It is a part of how things are. The real focus here is the rejection of the general epistemological thesis identified earlier. Very few commentators have this in focus, in part, I suspect, because Wittgenstein did not have it clearly in focus. It is clear that the social has a role in helping construct our ways of confronting things. Even so, this not necessarily a conceptual point that is constitutive of engagement, but often an empirical point that assists the focusing of engagement. For example, Wittgenstein says, ‘If humans were not in general agreed about the colours of things, if disagreements were not exceptional, then our concepts could not exist’ No: our concept would not exist. (RPP II §393; also Z §351)

He resists the conceptual claim that, in the face of a lack of agreement our concepts could not exist and affirms only the empirical claim that they would not exist. It is clear also that often enough the social is implicated because what we engage with is the complex filigree of how things are going for others – the patterns we call the social. So, of course, when engaged with the social, the

144


social can be constitutive of engagement, but that is a local point. It is not the general claim made by communitarian interpretations of Wittgenstein. I noted earlier (chapter 4, §5) that Wittgenstein acknowledges the hurlyburly of the background against which we describe human action: ‘Not what one man is doing now, but the whole hurly-burly, . . . background against which we see an action, and it determines our judgement, our concept and our reactions’ (RPP II §629; also Z §567). Even this is not constitutively social. He says, We judge an action according to its background within human life, and this background is not monochrome, but we might picture it as a very complicated filigree pattern, which, to be sure, we cannot copy, but which we can recognise from the general impression it makes. (RPP II §624)

and, The background is the bustle of life. And our concept points to something within this bustle. (RPP II §625)

Furthermore, commenting on the lack of definiteness of this background, he says: And it is the very concept ‘bustle’ that brings about this indefiniteness. For a bustle comes about only through constant repetition. And there is no definite starting point for ‘constant repetition’. (RPP II §662)

But even in these passages, there is no clear endorsement that this bustle, this hurly-burly, is essentially and constitutively social. The bustle of life can still be the individual against the world, and the series of one-on-one confrontations in which an individual will confronts another, changes their attitude in the light of that and then confronts another that requires a further amendment. This is not constitutively social. It is the ongoing individual engagements with the world and one another. Patterns, some of which we call social, emerge from this, but they are not constituted by anything social, let alone social agreement. There is a fallacy all too readily committed when thinking about the bustle of life. It is the fallacy of mistaking the behaviour patterns that emerge as individuals adjust to other individuals as patterns of collective entities. Searle invokes a concept of background in his work on social theory and treats it as a concept descriptive of the social. But many of the examples that he gives of social action simply are not social, they are examples of the way that aggregate behaviour can come to have patterns that look as if they have been collectively established,


145

when the truth is that they emerge simply from the individual’s responses to other individuals with no collective agreement or action at all. When lions hunt in a group there is all the appearance of coordination and collectivity. But we know that coordinated behaviour in such groups can emerge from individuals’ responses to particular circumstances. The behaviour of the lion pack is, arguably, no more collective than the behaviour of a flock of geese. Herding behaviour emerges from discrete individual responses to discrete individuals.27 Group action comprising intentional subjects is doubtless very different, but it can be similar to the lion case in this respect: the pattern of collective action is not constituted by anything collective, it emerges from the intentions, plans and perceptions of individuals as they respond to one another. The bustle and hurly-burly, even when social, might simply be patterns emergent from individual intentions. What the social does, whether conceived collectively or as emergent from individual interactions, is support the focus of individual attitudes. It scaffolds the engagement of the will with that which is independent of will. Similar thoughts apply to ‘forms of life’. A form of life concerns the shape of our attitude. This may be collective or, often, the emergent pattern of the individual’s attitude shaped by its encounters with others. We need to distinguish: (a) agreement in opinions/belief/what we say; and (b) agreement in attitude/form of life. (a) uses ‘agree’ intransitively – we agree in a proposition. In (b) ‘agree’ is intransitive – agreement between accounts. In the former, agreement is something about which we can give reasons. In such cases, it makes sense to say ‘I believe that . . . /I am of the opinion that . . .’. To agree here can be transitive – we need to agree stories, etc. (b) involves agreement in the sense of ‘congruence’, agreement between accounts. Here ‘agree’ is intransitive. Wittgenstein uses ‘Ubereinstimmung’ in §241: ‘So are you saying that human agreement decides what is true and what is false?’ – It is what human beings say that is true and false; and they agree in the language they use. That is not agreement in opinions but in form of life.

It is not agreement in what people say that determines the true and the false, for it is not agreement in opinions. It is agreement (a tally) in attitudes. This is not an issue for true/false. It might be tempting to treat the opinion/attitude

146


distinction as that between beliefs and conceptual schemes. That is not possible given the thesis of direct engagement, for there is no scheme/content distinction. That is why the notion of attitude becomes something that cannot be stated but only shown. If it was statable, then the issue of whether it is correct/incorrect would arise. Similarly, in part 2 of the Investigations Wittgenstein says, ‘My attitude towards him is an attitude towards a soul. I am not of the opinion that he has a soul’ (p. 178). The concept of attitude (einstellung) does not point to a justification, but a condition for justification. The absence of justification does not mean an absence of right (§289). But it would be a mistake to treat attitude as ‘the point at which no (analogical) arguments can be given for the existence of other minds. Such arguments turn out to be based on the presupposition that other people have minds.’28 This is still too much based on what we say, not on what we see. In conclusion, even §241 leaves the individualist reading of Wittgenstein in place. The social is not constitutive of standards of correctness. The contrast with ‘private’ in the private language argument is not public – it is being in the world.

5.7 ‘Inner’ Life Out There Wittgenstein can allow the notion of inner – how things are for the perspective or point of view of the will. He is not fully clear on this, for he too often discusses this in an epistemological manner (his opponents did this too). This explains the awkwardness and air of paradox in those passages where he explicitly grapples with both endorsing the reality of sensations and rejecting the Cartesian conception of subjectivity (§304). If we take the notion of a point of view in terms of the will’s confrontation with things independent of will, then the will is in the world. It is in the world because it is not conceived as something behind the subject. The will is not transcendent. The concept of the self-as-will that I have been urging is the concept of an empirical subject with a directedness to its causal engagements with things. That directedness is not characterized independently of its engagements with things. The will that provides the shape to the concept of a point of view is not understandable independently of its confrontation with that which is independent of it. This is, of course, what Wittgenstein himself struggled to express §§611–629. The will is not a special sort of causal antecedent to action, it is a special kind of action. It is the only kind of action that falls within the province of rational action, for it is action that exhibits directedness in its causal structure.29


147

If my point of view is in the world, then my being in pain can be a state of affairs in the world, so too can my seeing colours and all the other things that have often struck people as things that can be described independently of my direct being in the world. As Wittgenstein remarks, when considering our awareness of the blueness of the sky, When you do it spontaneously . . . the idea never crosses your mind that this impression of colour belongs only to you. And you have no hesitation in exclaiming that to someone else. And if you point at anything as you say the words you point at the sky. (§275)

What is misleading is the treatment of sensation reports like ‘I am in pain’, ‘I am seeing blue’ as if they were reports on an object (§304). There is no inner object, there is, however, how things are for me and the notion of ‘how things are for me’ is to do with how part of the world is going. As such, there is nothing special about it. Ordinarily with our beliefs, when I believe that p what is believed must be capable of obtaining independent of will, otherwise there is no correctness condition for the belief. Where p ¼ I’m in pain then the pain is independent of will, for that is what makes it real and in the world. The pain is to do with how things are for me. Is there now a problem in that ‘how things are for me’ is understood as the way things are for my will, the way my will is orientated. How can I form a proposition about that? Ordinarily, we form propositions about the objects to which I take an attitude. In the present case, the proposition is not about an object to which I take an attitude; it is about the character of my attitude. But the fact that this is not the normal case does not mean that I cannot take an attitude to the character of my attitude to things. It is just that this is a sophisticated move. It is not consciousness that is difficult, it is selfconsciousness. The place of pain here is not, however, different to that of bodily awareness. With the proposition, I am sitting down similar issues arise. For whom do I say this? Not for myself, certainly not in normal circumstances. In certain situations I may say this for the benefit of

148


others, but even with others it is often superfluous. The same points apply to ‘I am in pain’.

5.8

Now I Understand

Attitude is our take on the world. It is not justified, or supported with reasons. It is the point of view that makes reason-giving possible. When I get the point of some activity, my attitude changes. I see the similarity. I get the point and then I know how to go on. I do not always give reasons for this when getting others to see as I see. But the exercise of getting them to see things as I do is still a reasonable activity. The lack of reasons can be the lack of anything to say that could be set out as a formal argument. We can still call it reasonable and something that happens within the space of reasons, if we allow that doing things with right and getting others to see this does not require moving within a formal structure. Giving reasons for what we do can include appealing to our seeing of similarities. There is no substantive account of the necessary and sufficient conditions that make something a game, but that does not mean our calling something a game is unreasonable. All we can appeal to is the game as we see it. It is the game we see, for there is nothing else, no essential commonality to all games. Doing this can be giving a reason for action and belief, but there is nothing required to stand behind our seeing the game that makes it reasonable. It is intrinsically reasonable for, when done correctly, we see with right that it is a game. Of course, it is not always done well and then it can be challenged, defended and discussed. But much of that activity is a matter of further narrative fit, as we see how, if it is a game, this fits in with other things we say. But the ‘fitting in’ does not stand in need of further justification by instantiation of a super abstract grammatical form. It lies in what we do. Grammar is perspectival. What we do includes, simply, seeing things aright. It is then that we say ‘Now I understand!’ and ‘Now I can go on!’ It is a matter of seeing things aright. Wittgenstein provides an extended discussion of this in part 2 of the Investigations, in the section on aspectperception. Attitude is to do with continuous aspect-perception. This is what explains the ‘Now I know how to go on’ move when I suddenly get the point. It is a form of aspect switch when I suddenly see how to play the game. I turn to this in the next and final chapter.


149

Notes 1 For example, M. McGinn, Wittgenstein and the Philosophical Investigations, London: Routledge, 1997. 2 Despite ridiculing, rightly, the communitarian interpretation of the private language argument, Schroeder leaves the innner/outer distinction in place in so far as he treats the private language argument as an argument focused on the epistemology of our knowledge of others and of ourselves. Of course, the argument is concerned with those epistemological issues, but only, I believe, to show that the metaphysical assumptions that hold those problems in place are mistaken. See S. Schroeder, ‘Private Language and Private Experience’, in Wittgenstein, A Critical Reader, ed. H.-J. Glock, Oxford: Blackwell, 2001, pp. 174–98. 3 The idea that the core idea is the need for public verifiability goes back to N. Malcolm, ‘Wittgenstein’s Philosophical Investigations’, in Knowledge and Certainty, Englewood Cliffs: Prentice-Hall, 1963, see p. 113; but for the same point, see: C. McGinn, Wittgenstein on Meaning: An Interpretation and Evaluation, Oxford: Blackwell, 1984, pp. 48–9; A. Grayling, Wittgenstein, Oxford: Oxford University Press, 1988, pp. 86–7; M. Budd, Wittgenstein’s Philosophy of Psychology, London: Routledge, 1989, p. 61; P. M. S. Hacker, Wittgenstein: Meaning and Mind. Volume 3 of an Analytical Commentary on the ‘Philosophical Investigation’, Oxford: Blackwell, 1990, p. 21. 4 Both M. McGinn, Wittgenstein and the Philosophical Investigations, London: Routledge, 1997, and D. Pears, The False Prison, vol. II, Oxford: Clarendon Press, 1988, take the argument as making this central critical point, although they differ in their account of how the argument works. 5 McGinn, op. cit., represents Wittgenstein’s alternative in terms of ‘grammatical investigations’ that provide the standards of use. Pears provides more detail in terms of how the standards of use are defined by calibration against standard objects. As I shall argue below, although I think Pears’ reading gets a lot right, it still begs important questions against the Cartesian and thus still does not quite get to the rejection of the inner/outer right. 6 Ter Hark, Beyond the Inner and the Outer, Dordrecht: Kluwer Academic Press, 1990, and ‘The Inner and the Outer’, in Wittgenstein: A Critical Reader, ed. H.-J. Glock, Oxford: Blackwell, 2001, pp. 199–223, where, at p. 201, he speaks of the ‘rational connection between overt behaviour and inner states’. This leaves the inner and outer as too separate and a problem about how to characterize a notion of rational connection that could bring them together. 7 Cf. P. M. S. Hacker, Insight and Illusion, Oxford: Clarendon Press, 1972 (2nd edn, 1986), for discussion of this and also the idea that the sentence fails to express a proposition because it is not a possible object of knowledge. 8 Although, as Baker observes, the behaviourist reading of §244 simply ignores the qualifier ‘One possibility . . .’ that precedes the passage about the verbal expression replacing the cry. Cf. G. P. Baker, ‘Wittgenstein on Metaphysical/Everyday Use’, The Philosophical Quarterly, 52, 2002, 289–302, at p. 300.

150 9 10

11

12

13 14 15 16 17 18

19 20 21 22

23

Putting Your Self in the Picture This is in contrast to Augustine’s model of expression in §1. Cf. chapter 3, §5, above. If this is right, Nagel’s account of the qualitative content of experience is wrong. For a criticism of Nagel that makes this very point, cf. N. Eilan, ‘The Reality of Consciousness’, in Wittgensteinian Themes: Essays in Honour of David Pears, eds D. Charles and W. Child, Oxford: Clarendon Press, 2001, pp. 163–94. A corollary of this is that the differentiation between the veridical and illusory experience for any one case will not be achieved in terms of their qualitative feels – the things that, on a bipartite model of experience, are assumed to be invariant between the two cases. On a unitary model of experience the veridical and illusory cases are distinct experiences differentiated by the different impact they have on the will. The veridical case of seeing a cat impacts on the will differently to the illusory case. Recall the passage from Pears quoted in chapter 2: ‘How does the relation between name and object explain the sense of a proposition in which the name occurs? The object which is the name’s reference simply is what it is, inarticulate and lumpish. So how can it help the name to its miraculously articulate achievement in a proposition? How does this heavier than air machine get off the ground?’, D. Pears, The False Prison, vol. I, Oxford: Clarendon Press, 1987, p. 110. For this lumpish thing to be implicated in language use it needs to be brought within grammar, the space of reasons. D. Pears, The False Prison, vol. II, Oxford: Clarendon Press, 1988, p. 350. Ibid., p. 344. Ibid., p. 350. Ibid., p. 357. Ibid., p. 362. Indeed, it is the role of the physical in Pears’ account that makes the social look attractive, for it offers some resources for allowing that mindedness is not straightforwardly physical. Some might worry that Pears’ account leaves too little flavour to the notion of mindedness. See §199; RFM III §§66, 67; MS 165; MS 129, p. 89; MS 124, pp. 213, 221; also discussion of Robinson Crusoe in MS 116. McGinn, op. cit., pp. 125, 157–8. McGinn, p. cit., p. 130. J. McDowell, ‘Singular Thoughts and the Extent of Inner Space’, in Subject, Thought and Context, eds J. McDowell and P. Pettit, Oxford: Clarendon Press, 1986, pp. 137–68, §5. This explains, in part, Wittgenstein’s hostility to speaking of knowledge in the case of first person avowals, for part of the conception he is rejecting is the conception in which such knowledge has a special authoritative transparency. This, I suggest, is the proper target of the objections regarding the apparent vacuity of saying ‘only I can know I am in pain’ etc. Wittgenstein’s own treatment of this is, however, incomplete. I suspect his discussion is too influenced by an epistemological tradition that places special emphasis on the need for certainty. Adoption


24

25 26 27

28 29

151

of simple subjectivism amounts to a radical break with that tradition. For more on the general epistemological issues, see J. McDowell, ‘Criteria, Defeasibility and Knowledge’, Proceedings of the British Academy, 68, 1982, 455–79, reprinted in J. McDowell, Mind, Value and Reality, Cambridge, Mass.: Harvard University Press, 1998, pp. 369–94. It is curious to note that Davidson is closer to the communitarian conception of the standards of use, despite, officially, denying the scheme/content distinction, cf. D. Davidson, ‘A Coherence Theory of Truth and Knowledge’, in Truth and Interpretation: Perspectives on the Philosophy of Donald Davidson, ed. E. Lepore, Oxford: Blackwell, 1986, pp. 307–19. His coherentism would appear to amount to a conception in which the concept of things independent of will gets treated in terms of things dependent on the will of others. This is, in effect, the pre-modern conception of objectivity in terms of authority, in which we measure our attempts at utterance against those of the authority. That idea has echoes in contemporary accounts of Wittgenstein’s argument that see the objective as constitutively social. Wittgenstein is, I think, still struggling with the effort to shake off our premodern conception of the objective while observing the reality of subjectivity. The story has got around that there is no place for subjectivity in our modern conception of the world. If my reading of Wittgenstein is right, that story is simply wrong. Wittgenstein’s genius lies in the way in which he resolves what had looked to be a fundamental flaw in modernity – the dismissal of subjectivity from our conception of how things are. That the retrieval he effects is monumental is shown by the long shadows of pre-modern thought that can still be found in the fashionable communitarian readings of Wittgenstein’s work. Schroeder, op. cit., is clear on this too, as is Pears, op. cit. Op. cit., p. 125. See J. Searle, The Construction of Social Reality, London: Penguin, 1996, p. 38, for the example of hyenas hunting. See P. Ormerod, The Death of Economics, London: Faber, 1994, for an accessible discussion of the way in which many instances of ‘social collective’ behaviour in animals can be explained as emergent from individual responses. Ter Hark, 1990, op. cit., p. 61. See S. Candlish, ‘The Will’, in Wittgenstein: A Critical Reader, ed. H.-J. Glock, Oxford: Blackwell, 2001, pp. 156–73, for an account of the will in Wittgenstein’s work with which I have considerable sympathy.

CHAPTER SIX

Seeing Things Aright

6.1

Introduction

I have argued that Wittgenstein’s central insight is to make a shift to a perceptual account of the conditions for the possibility of judgement. The conditions for judgement consist not in a body of theoretical knowledge possession of which grants the subject with the capacity for using language, it consists in the subject seeing the world aright. For the idea of seeing things aright to take on the role I have given it in my account of Wittgenstein’s thought, a number of things have to hold. First, the subject that sees the world aright must be an active subject, an agent with an attitude to the world. The fundamental form of this attitude is captured by saying that the subject stands to the world as a self-as-will engaged with that which is independent of will. This permits a concept of inner, but not the inner theatre of sensations, images and ideas familiar from the Cartesian conception of the inner. The inner is the way we stand to the world. It is understood relationally, for it is the special kind of attitude that gives us the capacity for concept use and judgement. The inner is simply the notion of a point of view where that is understood as the directedness of the subject’s engagement. And ‘directedness’ is no more than the subject’s capacity to alter its engagement at will. We start with the conception of a purposive subject, but rather than relegate the subject’s purposiveness to something outside of its engagement with things and look for the will as a special causal antecedent to action; the will is what is manifest in a special kind of action. The kind of action at issue is the kind of action found only in creatures with conceptual experiences, rational action.


153

Second, the subject’s attitude to the world consists in a repertoire of capacities for seeing similarities, where these are not reducible to lists of properties in virtue of which one thing is similar to another. Learning concepts requires experience and training, but the point of the training is to get you to see the correct use. Training in this sense cannot be conceived as simply acquiring a set of responses – causal dispositions to respond to suchand-such stimuli with a word. The goal of training is the point when the subject knows how to go on because they see the point. Training consists in learning more discriminative forms of awareness; it consists in sharpening one’s attitude to things, focusing the directedness of our causal engagements with things.1 Third, the subject’s attitude is an attitude of judgement. It is a matter of judgement that one goes on correctly. If the seeing of similarities is primitive, there can be no explicit technique for testing whether the execution on any one occasion is correct or not. This will always be a matter for judgement. But the role of judgement is not to label a looseness of fit in esoteric and complex uses of words, its role is fundamental to concept possession. The basic model of understanding is one of the exercise of judgement, not of technique. Fourth, the role of judgement, in contrast to technique, points up the fundamental creativity of language use. Educated competent language use is not something modelled in terms of the application of techniques, procedures that could be articulated in explicit rules. Educated competent language use is fundamentally open-ended. There are real patterns, but our role as judges is ineliminable from the metaphysics of those patterns. Language is a game. It is something we play, freely and creatively. It is not a structure that impedes or constrains us, and it is not a structure that exists wholly independent of us that we have to inhabit with care lest we chip the corners by taking short cuts. The structures and patterns of language use are dynamic. Fifth, the whole point to seeing the conditions for the possibility of judgement in this way is not to offer a new theory of judgement. It is not to offer something that provides a novel theoretical articulation or in any way a reductionist description of the core cognitive skills that lie at the heart of our ability to use language. The point of the reading of Wittgenstein that I have been outlining is simply to provide a picture, a synoptic representation of how we are and how we stand in the world. If this picture is right, it legitimizes concepts that might otherwise seem problematic. It is a picture that reinstates a self-conception that has often seemed eroded. It reinstates a conception of our role in things that is not only too frequently overlooked, but often enough thought impossible from within the compass of theories that attempt theoretical articulations of the conditions for the possibility of meaning.

154


There is, then, much at stake in this reading of Wittgenstein. If my reading is right, there is no reductionist analysis available of what constitutes seeing things aright. This means that the typical phenomenology of language learning would be one in which the learner expresses their coming to understand, not by articulating a piece of theoretical knowledge, but by saying ‘Now I know how to go on’. They will say that they see what comes next, they see how to play the game. Whether or not they see things aright will not be transparent to them. It certainly will not be evident from the occurrent characteristics of their experience when they feel they understand – understanding is not a feeling (§169, and see throughout §151–184). What will show that they see things aright will be in their practice, their going on right. But what constitutes their understanding is not simply the going on (as if that were simply a piece of behaviour that could be logged as part of the catalogue of contingencies); it is their seeing aright how to go on and that they are going on right. Their correctness consists in their taking the right attitude to things, of setting their face to the world in the appropriate way. This is to take part in an engagement with things in which word use is directly calibrated against that which provides the standards of correctness; it is a practice. When the learner acquires the concept game, they do not acquire a body of theory, they acquire an ability to face things with the right attitude, the attitude that enables them to judge that such-and-such is a game. This attitude is an attitude to the world, anything less and it would not be an attitude of seeing things aright. If this is right, then one ought to be able to have experiences in which the meaning of a word suddenly shifts, as one suddenly realizes that the practice can go in another altogether different way. It ought to be possible, for a word, construed simply as a sign, could be used in two altogether separate practices and the switch between them would be something instantaneous. It would not be a matter of moving from one interpretation to another, it would be a matter in which one’s attitude switched like a gestalt switch. One would suddenly see the word and its use differently. Not only ought this to be possible, but on the account I have extracted from Wittgenstein, there must be such experiences. If, in many cases, seeing the world aright is constitutive of understanding, then there must be cases in which we grasp in a flash. And if there are cases in which we grasp in a flash what a word means and words construed as bare signs can bear different meanings, then the switches between those meanings must be made in a flash. Of course, Wittgenstein’s writings are replete with remarks about the phenomenology of understanding: Can what we grasp in a flash accord with a use, fit or fail to fit it? (§139) But there is also this use of the word ‘to know’: we say ‘Now I know!’ – and similarly ‘Now I can do it!’ and ‘Now I understand!’ (§151)


155

In addition, he not only allows, but describes the phenomena of meaning switches in considerable detail in §11 of part 2 of the Philosophical Investigations. The discussion of aspect-perception in this section fits squarely with the account extracted thus far. The concept of seeing the world aright is what Wittgenstein means by continuous aspect-perception. Furthermore, in the closing pages of this section, Wittgenstein explicitly endorses the very model of judgement contra technique that I have ascribed to him.

6.2 Two Uses of the Word ‘See’ The treatment of aspect-perception in part 2 of the Philosophical Investigations is still not much discussed in the secondary literature. When it is discussed it is mostly to draw out a negative point that Wittgenstein makes about the content of perceptual experience.2 The negative point is a critique of a bipartite model of experiential content. This negative point does not, of itself, explain why Wittgenstein included this discussion of aspect-dawning in part 2 of the Investigations. There is, however, a positive point that Wittgenstein is making and this is, I believe, the main import of the section. Broadly speaking, the positive point is the endorsement of a unitary model of content for perceptual experience, for only such a model can explain the puzzling phenomenology regarding our perception of ambiguous figures.3 In part 2 §11 Wittgenstein is ostensibly discussing what seem to be exceptional perceptual experiences in which a seen object switches from being seen in one way to an altogether different way. Figures such as the duck-rabbit figure (p. 194) can be experienced in two quite different ways and yet leave the subject with a clear sense that nothing has changed. It switches from looking like a duck to looking like a rabbit and yet we know at the same time that nothing has changed. Wittgenstein calls such cases ‘aspect-switches’ or cases of aspectdawning. The negative point that Wittgenstein makes is familiar. It is the critique of a model of perception in which there are two components to perceptual experience – a given element and a component that provides the meaning to the given. Wittgenstein discusses two possibilities for the second component – a visual impression (inner picture) and an interpretation or organization of the given element. As Mulhall observes, if this negative argument were the sole purpose of part 2 §11, it is unclear that it would need to be there at all, let alone that the section be so long. The positive point of the section is seen when we recall that the point is to account for the paradoxical character of aspectswitches. These are experiences in which we want to say both that nothing has changed and that something has. The negative critique of bipartite models of perceptual content does nothing to account for that paradox. The positive

156


point of this section is the claim that only a unitary model of perceptual content can account for the paradoxical character of aspect-switches. Wittgenstein introduces the material for the concept of seeing-as in the opening paragraph of part 2 §11: Two uses of the word ‘see’. The one: ‘What do you see there?’ – ‘I see this’ (and then a description, a drawing, a copy). The other: ‘I see a likeness between these two faces’ – let the man I tell this to be seeing the faces as clearly as I do myself. The importance of this is the categorical difference between the two ‘objects’ of sight. (p. 193a)

The contrast made in this opening paragraph is between a concept of the object of perception as something exhausted by an account of the spatial arrangement of lines and the object of perception when characterized by seeing likenesses. It is the former concept of the object of perception that gives us the notion of the constant unchanging character of our perception when the duck-rabbit figure switches from being seen as a duck to being seen as a rabbit. It is the notion of seeing likenesses that marks the difference. The paradox is that both seem viable objects of perception. There is a view that holds that only the former kind of object is the legitimate object of perception and that seeing likenesses is not really a form of seeing at all; it is an interpretation or organization on the basic object of perception proper. Wittgenstein spends little time dismissing that view in part 2 §11, for the negative point is not his real purpose.4 He starts the section by speaking of two uses of the word ‘see’ and there is no suggestion that they are not proper uses. Indeed, he later explicitly rejects any attempt to suggest that one of the uses of ‘see’ might be improper, ‘The concept of ‘‘seeing’’ makes a tangled impression’ (p. 200a), and, Here we are in enormous danger of wanting to make fine distinctions . . . What we have rather to do is to accept the everyday language-game, and to note false accounts of the matter as false. The primitive language-game which children are taught needs no justification; attempts at justification need to be rejected. (p. 200b)

Although Wittgenstein also notes that ‘the flashing of an aspect on us seems half visual experience, half thought’ (p. 197d), it is clear that he thinks this latter view the result of a false account. We might sum up what is going on in the negative thesis in the following way. We can distinguish between the following three concepts of seeing which we could call seeing-objects, seeing-as and seeing-that. The latter two are instances of seeing likenesses.


157

By ‘seeing-objects’ I mean the technical philosophical concept of seeing something described minimally as a spatial arrangement of lines and shades. I do not mean the ordinary sense of seeing material objects. It is a specialized form of seeing that takes training to acquire, for it is the kind of seeing that artists need to acquire in order to represent three-dimensionality in twodimensional patterns. Seeing-as is the concept implicated in aspect-dawning, when the ambiguous figure changes visually for us and it moves from being seen as a duck to being seen as a rabbit. Now, one reason for holding that seeing-objects is the only legitimate model of genuine perceptual experience is that it offers to provide an account of perceptual experience as pure receptivity, for it attempts to strip the object of perception of conceptual content. Of course, this attempt fails, for there is still a minimal conceptual content to seeing-objects. The thought behind the concept is simply that the content has been stripped down to the sort of content that, it is assumed, can be carried unproblematically by visual experience conceived as a sort of passive imprinting. The focus, however, of Wittgenstein’s investigations in this section is seeing-that. By ‘seeing-that’ I mean what Wittgenstein calls ‘continuous aspectperception’. ‘Seeing-as’ is aspect-perception as it figures in the dawning of an aspect. But that concept, and its distinction from seeing-objects, although important for the negative thesis, does nothing to resolve the air of paradox that aspect-dawning experiences produce. What needs to be resolved is the sense of paradox in the duck-rabbit both looking the same and looking different when the aspects switch. Simply availing ourselves of both the concepts of seeing-objects and seeing-aspects to describe such experiences does nothing to say why such experiences seem paradoxical to us. The solution is surely this. Seeing-that is primitive. The ordinary way of seeing things consists in seeing-that: seeing-that x is a rabbit, seeing-that y is a duck. Having introduced the concept of a picture-object (p. 194d) where pictureobjects are things I stand towards as I stand toward real objects, Wittgenstein says, if I had always seen the Jastrow figure as a picture-rabbit, ‘I should not have answered the question ‘‘What do you see here?’’ by saying: ‘‘Now I am seeing it as a picture-rabbit’’. I should simply have described my perception’ (p. 194f), and, It would have made as little sense for me to say ‘Now I am seeing it as . . .’ as to say at the sight of a knife and fork ‘Now I am seeing this as a knife and fork’. (p. 195a) One doesn’t ‘take’ what one knows as the cutlery at a meal for cutlery; any more than one ordinarily tries to moves one’s mouth as one eats, or aims at moving it. (p. 195b)

158


In other words, seeing-that is natural. It is second nature to us.5 If seeing-that is primitive and the natural form of seeing, then that would explain the sense of paradox with the Jastrow figure, for it disrupts our ordinary sense of seeing-that and makes us aware of what is normally and naturally taken for granted – our perceptual experience has a conceptual content all the way down. Once that is seen, the stripped down account of perception as seeingobjects is seen for what it really is: a crude empiricism in which perception is a passive imprinting construed in a literal way that, once articulated, is seen to be unwarranted. The claim that seeing-that is primitive is clear in the opening pages of this section. Immediately on introducing Jastrow’s figure Wittgenstein says, ‘And I must distinguish between the ‘‘continuous seeing’’ of an aspect and the ‘‘dawning’’ of an aspect’ (p. 194a). Indeed, the opening paragraph in which he first introduces two uses of ‘see’ does not mention seeing-as. He speaks of seeing ‘a likeness between . . . two faces.’ At this stage it is not clear that this is a seeing-as, rather than a seeing-that, e.g. seeing-that the faces express the same emotion, or look similar in some other way. It is only in the third paragraph of the section that Wittgenstein introduces the idea of suddenly noticing likenesses. If the point of part 2 §11 is to explore the phenomenology of perception that reveals that seeing-that is primitive, the section fits squarely with the reading that I gave in chapters 3 and 4 of earlier passages. I argued in chapter 3 that Wittgenstein takes the concept of seeing the similarity as primitive. Seeing similarities are cases of seeing-that. It is because seeing the similarity is primitive that the ambiguous cases now discussed in part 2 §11 have the air of paradox. And the point is not just a phenomenological one in the sense that it is to do with the perceived primitiveness of certain sorts of experiences. The point is that this primitiveness of seeing-that is part of the general picture of a unitary model of content in contrast to a bipartite model of content. That is to say, cases of seeing-that, of primitive seeing the similarities, are cases in which our word use is directly calibrated against that which provides the normative standards of correctness. Cases of seeing-that are cases in which it is not just that we take for granted how our experience is going, but they are cases in which we are normally entitled to take them so, for they are cases in which we use a word with right. These are the sorts of cases that make up the direct engagement with things that contribute to the concept of minimal practice – chapter 4. The concept of seeing-that is a rich concept. It is tempting to say that this is not really seeing at all, but a form of knowing. It is tempting to say that seeing-that is really just a case of seeing-objects + interpretation: ‘the flashing of an aspect on us seems half visual experience, half thought’ (p. 197d), and,


159

‘how is it possible to see an object according to an interpretation? – The question represents it as a queer fact; as if something were being forced into a form it did not really fit’ (p. 200e). The negative thesis, the denial that all seeing is really seeing-objects, represents our ordinary phenomenology as a queer fact. Wittgenstein does not spend much time in part 2 §11 on the argument for the negative thesis. The idea of treating the case of ambiguous figures in terms of an imprinting seeing-object plus organization or interpretation is already complete by p. 196. The negative thesis draws too much on previous arguments for Wittgenstein to need to give it much attention by the time of part 2. What Wittgenstein dwells on is getting the phenomenology right once it is accepted that seeing-objects is not the end of the story about perception. He is concerned to iron out our temptation to find a uniform simple model of what counts as seeing and what does not: ‘We find certain things about seeing puzzling, because we do not find the whole business of seeing puzzling enough’ (p. 212f). The uniform simple model of seeing in terms of seeing-objects leaves cases like the Jastrow figure paradoxical, but that is only because we have not done justice to the richness available within the concept of seeing in the first place: ‘Is being struck looking plus thinking? No. Many of our concepts cross here’ (p. 211e). At p. 203c he introduces the following case, I see that an animal in a picture is transfixed by an arrow. It has struck it in the throat and sticks out at the back of the neck. Let the picture be a silhouette. – Do you see the arrow – or do you merely know that these two bits are supposed to represent part of an arrow?

His first response is to say, ‘ ‘‘But this isn’t seeing!’’ – ‘‘But this is seeing!’’ – It must be possible to give both remarks a conceptual justification’ (p. 203d). If we endorse the negative thesis and reject the assimilation of all cases of seeing to seeing-objects, we need to find space to accept the primitiveness of seeing-that. What Wittgenstein is struggling with in part 2 is the resistance to taking our phenomenology at face value. To take seeing-that as primitive is to allow a richness to the phenomenology of experience that is striking: ‘what I perceive in the dawning of an aspect is not a property of an object, but an internal relation between it and other objects’ (p. 212a). No wonder then that we are tempted to treat the arrow as something not seen, but inferred, for taking seeing-that as primitive allows us to perceive internal relations. What can this mean? Let us return to the simple case of the duck-rabbit. Wittgenstein’s point is to draw our attention to seeing-that (continuous aspect-perception). So when I see a picture of a rabbit, my seeing a rabbit is not just a seeing of lines and patterns of shade (seeing-objects). I see the likeness between this rabbit and

160


others. I see that they are related. But that they are related is not due to their propensity to produce a common visual-object, a pattern of lines and shade, any more than that two events are games is due their satisfying a common set of jointly necessary and sufficient conditions. So when it dawns on me that a figure I had previously seen as a duck looks like a rabbit, the change is not in the picture-object construed as an arrangement of lines and shades. In terms of seeing-objects, it is the very same object. What I come to see is that this picture-object is like others. That is a case of seeing-that. Because it is a case of seeing with conceptual content, what I see is something normatively configured with respect to other contents. If it is a rabbit, certain things follow about what is right and what is wrong to say of the picture-object. Given the primitiveness of the seeing-that, the primitiveness of seeing the similarity, there is no further account in virtue of which the normative configurations hold. What I see is rationally related to further claims, including perceptual claims, but there is no account in virtue of which such rational relations hold. What I see in such seeing-that is not a property of an object, I see how it is related to other objects. The relation is one of similarity, but the concept of similarity cannot be understood in terms of similarity with respect to arrangements of line and shade. The similarity concerns the common position within the space of reasons, the common position these objects have in the rational linkage of conceptual contents. Perception can directly deliver us such rational linkages. It produces just that ‘understanding which consists in ‘‘seeing connexions’’’ (§122). That, I suggest, is what Wittgenstein means when he says I perceive an internal relation between the picture-object and other objects. It will be objected that such a thing cannot be delivered in perception. Perception delivers only the objects of seeing-objects – patterns of line and shade. Despite the argument against the simple empiricism that has perception as a bare imprinting (the negative thesis), there appears to remain a problem with the concept of seeing-that. The problem that appears to remain is, however, illusory. The problem might be expressed like this. In so far as perception is a matter of receptivity, of things striking us, how can we be struck by the rational linkages between conceptual contents? In short, how can we be struck by an internal relation?6 There is a mistake here in thinking that this is the problem, for it looks as if we are asking how the picture-object can strike us and force a linkage upon us. That cannot be right, for the picture-object remains the same when the aspect-switches. So, of course, the picture-object itself, construed as an arrangement of lines and shade, cannot force a linkage upon us. If we are looking for a source of the rational linkage in what is available within the picture-object, then no source will be forthcoming. That must be right, for otherwise the rational linkage could


161

be forced by a case of seeing-objects and we know that that is not right. The present issue about the source of how we can be struck by the rational linkage between concepts is another version of the opening problem in chapter 1 – the problem that arises from supposing that we need an animatory theory of meaning. In the current context, the present objection is asking for an account that shows how seeing-objects gets animated. The short answer is that it doesn’t. Instead, Wittgenstein introduces a second concept of seeing that serves the purpose. This second concept – seeing-that – works not because when added to seeing-objects we can construct an account of how we are struck by rational linkages, rather, the concept of seeing-that introduces a form of seeing that takes the notion of being struck as primitive. Wittgenstein expresses the same point much earlier in the Investigations, when discussing the possibility that my understanding of the word ‘cube’ might consist in having an internal picture, Is there such a thing as a picture, or something like a picture, that forces a particular application on us; so that my mistake lay in confusing one picture with another? (§140a)

No addition to the picture of the cube, e.g. having the method of projection come before my mind, will do. Indeed, in this earlier discussion in critique of the role of pictures in concept mastery, Wittgenstein already foreshadows the moves he makes in part 2. He says, What is essential is to see that the same thing can come before our minds when we hear the word and the application still be different. Has it the same meaning both times? I think we shall say not. (§140c)

In other words, he has already allowed that, if you like, the ‘reading’ of a picture cannot be supplied by adding further ingredients to the picture. The extensive discussion of reading, §§156–178, that interpolates between the first discussion of understanding following the introduction of arithmetical rules in §143 until just before the return of that topic at §185, receives too little attention in the secondary literature. The points Wittgenstein makes in the discussion of reading are of a piece with the aspect-perception discussion in part 2. He says, for example, It would never have occurred to us to think that we felt the influence of the letters on us when reading, if we had not compared the case of letters with that of arbitrary marks. And here we are indeed noticing a difference. And we interpret it as the difference between being influenced and not being influenced. (§170)

162


The puzzle is to locate the sense of influence in something that cannot possibly deliver it – seeing-objects. But this does not mean that there is not a difference between reading words with the familiar fluency of normal readers and struggling to interpret them. The latter case is seeing-objects + interpretation, the former is seeing-that. The difference lies not in the words, they remain the same. The difference lies in our attitude to the words. The same point is repeated in the discussion of perception in part 2 §11. The point is, once more, the critique of a bipartite model of meaning. If we start with the concept of perception as imprinting (seeing-objects) we need to add a further ingredient in order to obtain the idea that a picture-object can strike us and force a rational linkage upon us, so that what is delivered in perception turns out to be something that is rationally linked to other concepts. We need a second ingredient to add to the conception of perception as an event within the space of causes and transform it into something that occupies a position within the space of reasons. The puzzle is that no such second ingredient can be found within the material delivered by seeing-objects. Hence the thought that the internal relation cannot be perceived. But that is to ignore Wittgenstein’s opening claim in part 2 §11 that there are two uses of the word ‘see’. One of those uses is the notion of seeing-objects. That use of ‘see’ cannot deliver the notion of seeing-that. It cannot deliver it, because seeing-objects is a matter of passive imprinting and as long as we take our role in perception to be that of passive subjects being imprinted on, then there is no room for the idea that we can simply see the likeness between things. The whole point to the reading of Wittgenstein that I have been giving is, however, that in giving an account of content our role is not passive. The role of the subject as active agent, a self-as-will, is integral in getting a unitary, rather than a bipartite account of meaning. The point applies just as much to an account of perceptual content. The ‘reading’ of a picture must lie in what we do. We will have no account of how we see the similarities unless we change focus and think of perception not as a passive imprinting, but as a conceptually structured attitude to the world. We do not see the similarities because the similarity is, as it were, another property of the object. That would be to think that in seeing the picture-object as a rabbit I was seeing, in addition to the lines and shades, some further property, the similarity that connects this pictureobject to others. That, of course, is ridiculous. The similarity is not a property of the picture-object. It is not and cannot be seen in the way that the picture object can be seen – seeing-objects. To return to Wittgenstein’s example of a picture of an animal pierced with an arrow his conclusion is, ‘To me it is an animal pierced by an arrow.’ That is what I treat it as; this is my attitude to the figure. This is one meaning in calling it a case of ‘seeing’. (p. 205a)


163

Seeing-that consists in taking the right attitude to things. It is not seeingobjects plus a second ingredient that supplies the account of the rational linkage between what is delivered by perception and what follows. Seeing-that carries a unitary content, something that cannot be factored into component ingredients. It is different to seeing-objects, it has to be. But it is, for all that, still a form of seeing and, once again, it has to be, for otherwise we would have a bipartite model of meaning once more composed of seeing-objects plus interpretation. The key concept is that of attitude. The key to seeing-that is the notion of a subject with an attitude to the world, an engaged attitude in which the self-aswill confronts that which is independent of will. As long as we take the subject as a passive receptacle onto which the world imprints we will struggle to make sense of what Wittgenstein is exploring in part 2 §11. We will also, as I have argued, struggle to make sense of his core insights about meaning. In teaching concepts we instinctively employ active vocabulary, we treat the learning subject as an agent taking from the world, not passively being imprinted upon: How does one teach a child (say in arithmetic) ‘Now take these things together!’ or ‘Now these go together’? Clearly ‘taking together’ and ‘going together’ must originally have had another meaning for him than that of seeing in this way or that. (p. 208c)

In learning concepts the child has to learn to see things in certain ways, but the seeing is not imprinting, it is a taking. It is learning the ‘phrasing’ of things, and by this I do not mean learning to spot common features, for there need not be any. Learning the phrasing of things is a matter of learning how to discriminate within one’s fundamental attitude, but having an attitude is not something that gets learnt, that is given. The concept of attitude is primitive. It has to be, otherwise there would not be two uses of the word ‘see’. That it is primitive turns on the point that Augustine got right. The human infant may not have anything like the sophisticated catalogue of beliefs and desires that Augustine permits, but it is a self-aswill. It has a directedness. What it lacks are the capacities to structure its directedness with the discriminations that educated subjects possess. Education is a matter of training the basic attitude, it’s a fine-tuning of the directedness. It is not a matter of acquiring that directedness. To think that that was something that had to be acquired would be to conceive of learning as a transition from seeing-objects to seeing-that. There is no such transition. What is the description of an ‘attitude’ like? One says, e.g. ‘Disregard these spots and this little irregularity, and look at it as a picture of a . . .’ ‘Think that away. Would you dislike the thing even without this . . .’ (Z §204)

164


We describe an attitude only from within, by pointing out what we regard and what we disregard. We teach the learner by getting them to see this and disregard that. These are primitive moves. There is nothing simpler. For how do we arrive at the concept ‘seeing this as that’ at all? On what occasions does it get formed, when is there need of it? (Very frequently in art.) Where, for example there is a question of phrasing by eye or ear. We say ‘You have to hear these bars as an introduction,’ ‘You must hear it as in this key.’ . . . Now, is it a real case of seeing or hearing? Well, we call it that; we react with these words in particular situations. And we react to these words in turn by particular actions. (Z §208d)

Teaching concepts to the learner is to teach them how to phrase their attitude to the world. We illustrate or describe the attitude with examples of phrasing. We do not construct an attitude from something that lacks phrasing. The moves in education, in teaching the human infant, are moves to show them how to phrase their basic attitude to the world. Augustine thought these moves were moves to label a phrasing that was already in place. That was his mistake. But what he got right was that the human infant is, in the beginning, a self-as-will. It has attitude, but not phrasing. The phrasing of attitude is the directedness of our causal engagements, it is the structure of our attention. Our attention can get a new direction, we can be shown how to attend by our teacher, but this is still a form of seeing. ‘The genuineness of an expression cannot be proved.’ ‘One has to feel it.’ But what does one go on to do with this now? If someone says ‘Voila, comment s’exprime un coeur vraiment e´pris’, and it also converts someone to his viewpoint, – what are the further consequences? In a vague way consequences can be imagined. The other one’s attention gets a new direction. (RPP II §697, cp. part 2 §11 p. 228a)

But to have a direction at all requires the capacity for directedness, something suitable for phrasing, and that comes from the fundamental attitude of the self-as-will. Wittgenstein expresses the gap involved here as ‘Instinct comes first, reasoning second. Not until there is a language-game are there reasons’ (RPP II §688). On the account I have been urging, the language-game comes from having the attitude of the self-as-will. Paraphrasing RPP II §688, we might say, seeing-objects comes first, seeing-that second, but not until there is a self-as-will do we move from the first to the second. The point is that this transition is not a move we ever make, for it is not a move that we could


165

negotiate. Negotiation and reason-giving only start once seeing-that is in place. Once seeing-that is in place, it is something that can come in a flash. Now I suddenly see it as a rabbit. This is not and cannot be incremental. I do not move by rational steps from seeing the arrangement of lines and shades to seeing the rabbit. It strikes me and I see the rabbit. As when I look in the direction of an old acquaintance for some time and I suddenly recognize the face I had failed to know – ‘Suddenly I know him, I see the old face in the altered one’ (p. 197 g) and ‘The very expression which is also a report of what is seen, is here a cry of recognition’ (p. 198a). The immediacy of seeing-that, the way that the phenomenology can be accurately reported as happening in a flash, should not be surprising. It was bound to be thus, for if there are no incremental steps, no movements of reason from seeing-objects to seeing-that, the latter must be the sort of thing that can come in a flash. But what comes in a flash is not the receipt of an imprint, it is not a rapid case of seeing-objects. What comes in a flash is the new direction of attention, the new phrasing of attitude: ‘ ‘‘Now I understand’’ is not a description, it’s a signal. It directs attention to the origin of such a signal’ (RPP I §691). The origin of the signal is the phrasing, the similarity that I grasp by changing my attitude to things. When suddenly I see, I present to things in a different way. I start a new continuous aspect-perception, I start seeing new similarities.

6.3

See What I Mean

The reading I have given of the discussion of family resemblance and seeing the similarity (chapter 3) and the rule-following argument (chapter 4) are squarely of a piece with what I have argued Wittgenstein is doing in part 2 §11. He makes the comparison himself: Aspect blindness will be akin to the lack of a ‘musical ear’. The importance of this concept lies in the connexion between the concepts of ‘seeing an aspect’ and ‘experiencing the meaning of a word’. For we want to ask ‘What would you be missing if you did not experience the meaning of a word?’ What would you be missing, for instance, if you did not understand the request to pronounce the word ‘till’ and to mean it as a verb, – or if you did not feel that a word lost its meaning and became a mere sound if it was repeated ten times over? (p. 214c, d)

The connection is clear: experiencing the meaning of a word is a case of seeingthat, continuous aspect-perception. And the points made in the previous section apply in just the same manner to experiencing the meaning. Experiencing

166


a word as meaningful is not a matter of having an extra ingredient enter into the content of seeing-objects, in addition to the seeing-object perception of being imprinted by sounds or ink marks. Experiencing a word as meaningful belongs to an altogether different category of seeing – two uses of ‘see’! It consists in taking the right attitude to a word. When I suddenly cotton on and understand a word, this is not because I have seen, in the sense of seeing-objects, an extra something or other that animates the word. I cotton on when I have an aspect switch, a seeing-as and start a new form of continuous aspect-perception – seeing-that. ‘Now I understand’ is not a description, it’s a signal. It directs attention to the origin of such a signal’ (RPP I §691). The ‘Now I understand’ cannot be a description, for it does not pick out something available from within my attitude to the word, it signals that I am taking a certain attitude to the word. In the earlier discussion of the meaning of ‘cube’ Wittgenstein says, When someone says the word ‘cube’ to me, for example, I know what it means. But can the whole use of the word of the word come before my mind, when I understand it in this way? . . . Can what we grasp in a flash accord with a use, fit or fail to fit it? (§139a, b)

The puzzle here arises from assuming that understanding must consist in something capable of full articulation. If that were so, then in principle one would expect that articulation to be something that could be expressed and, in turn, be the object of a seeing-objects. That, of course, is the move that Wittgenstein repeatedly critiques and subjects to the regress challenge from this section and on through the rule-following argument. The error lies in thinking that understanding is something that should be deliverable in terms of seeing-objects, for if it were so deliverable it would be something that could be treated as theoretical knowledge. To return to a point I took from Cavell, on such a model understanding would be something that could be memorized as opposed to what we remember.7 But understanding is not like that. I have argued that Wittgenstein’s central innovation is to take the conditions for the possibility of judgement to consist not in theoretical knowledge, but in perceptual knowledge. The conditions for the possibility of judgement consist in our seeing things aright. The notion of ‘seeing things aright’ cannot be modelled in terms of seeing-objects, for otherwise it would collapse back into theoretical knowledge. There is a sense in which we can grasp the meaning of a word in a flash. In one of his many examples of concept learning in which the pupil is trying to find the law for a number sequence Wittgenstein allows that If he succeeds he exclaims: ‘Now I can go on!’ – So this capacity, this understanding, is something that makes its appearance in a moment. (§151b)


167

We can grasp the meaning in a flash. But what then was the problem at §139? The problem was that if we take this grasping to be a perceptual seeingobjects, such an appearance in a moment is impossible. It would require the whole potentially infinite sequence of applications to be available in an instant. ‘And how can what is present to us in an instant, what comes before our mind in an instant, fit a use?’ (§139b). The answer, of course, is that it cannot. Nothing that comes before the mind in an instant can fit a use nor constitute understanding. All that comes before the mind in an instant is what is provided by seeing-objects. Seeing-that, coming to grasp in a flash the meaning of a word is not a case of that sort of seeing. What makes its appearance in an instant is an attitude, a way in which one attends and presents oneself to that which is independent of will. That is something that can change in an instant (aspect-dawning) and it is also something that can be recalled in an instant (continuous aspect-perception). This is not something that is assembled out of seeing-objects + ingredient X, it is primitive. It can, however, be degraded or lost. Repeating a word many times over can degrade our experience of the word so that we hear it just as a sequence of sounds, not as bearing meaning. What is this transition? Hearing the word meaningfully is to be prepared to deploy it according to its grammar, its pattern of use. This grasp of grammar is not delivered by the hearingobjects perception (the sound of the word), it makes its appearance in the hearing-that, the attitude that we take as active judges with a capacity for using words in normative patterns. The repetition does not and cannot affect the hearing-objects perception. In terms of hearing-objects, the content of the perception stays just the same. The word ‘till’ sounds, in terms of hearingobjects, exactly the same as we repeat it, otherwise we would not be repeating it! The sense in which it sounds exactly the same is simply in terms of patterns of sound waves, the causal impact of the air-carried vibrations as they strike the ear. The repetition cannot then affect that. If anything, the repetition reinforces the hearing-object perception. What the repetition does affect, of course, is the subject who hears. The repetition forces attention away from grasp of the potential for deployment within grammar to attention to the bare sound. The change is in my attitude. I change my way of listening to things. The extraordinary character of such experiences is testimony to the commonality of hearing-that with respect to words. They acquire a familiarity. They have a readiness-to-hand for us and we come to feel and appreciate their physiognomy (cf. p. 218g). Wittgenstein’s discussion of the phenomenology of our attitude to words is rich and fascinating. There is much that could be said in pursuit of this topic. For now, it suffices to note the central point that flows throughout part 2 §11: the fundamental feature being picked out in this rich phenomenology is the

168


idea of our attitude to things, including pictures and words. The concept of attitude is, as I have argued in earlier chapters, the key to understanding Wittgenstein’s shift from a theoretical to a perceptual model of the conditions for the possibility of judgement.

6.4 Paying Attention I have used the word ‘attitude’ to pick out the basic confrontation of the selfas-will with that which is independent of will. What makes us distinctive and language-using creatures is that we are creatures with attitude. We are agents, creatures with wills. I have introduced also the idea that language learning is a process of discriminating the phrasings within this attitude. It is a tuning of the will to that which is independent of it. Language learning cannot happen if there is not that basic attitude. This is why the condition for the possibility of judgement is, on my reading, fundamentally individualistic. The community might assist by enabling the tuning of the will to the world, it provides scaffolding for this tuning, but it is not constitutive of the tuning nor of that to which the will tunes. Of course, as we get sophisticated, some of the things to which we tune our attitude are themselves social. There is a social world. But it plays no constitutive role in an account of the conditions for the possibility of judgement. The basic attitude is one of presenting to the world – how I present to the world. I do this without justification, but with right. The entitlement to speak of ‘right’ here comes from the fact that the attitude is fundamentally configured with respect to truth. Our basic concept of truth, that which figures in the concept of normative standards of correctness/incorrectness, simply is the concept of that which is independent of will. The starting point that provides the conditions for the possibility of judgement is the will presented to that which provides the standards for word use. The engagement is direct. It is because this attitude is the point of view of an agent, that the form of seeing that characterizes it – seeing-that – is a matter of attending. The perception that fills out the space of our fundamental attitude is attention, not passive imprinting.8 Attention is something we do, it is not something that happens to us. The attentional field is the field of seeing-that, it is a field of opportunities for action. The attentional field comprises the sorts of things and sorts of ways in which my basic attitude can be configured. Concept learning is a tuning of the attentional field. The basic move in concept learning is then coming to see things differently, for it is coming to acquire skills for attending to things that had previously not been differentiated within one’s attitude. We can speak of ‘character’ as the set of attentional skills by which I flex my attitude.9


169

Because attention is an active perceiving it requires an act of will. The attending subject is already a judge, for judgement is involved. Seeing-objects is static. It is a passive imprinting and is something that simply happens at a point in time. It might be repeated and might continue for some time. But if it does, there is no respect in which it endures that amounts to a connection between its repeated occurrences. The imprinting just happens, and happens again, and again, and again, and so on. In contrast, attention requires a persistence of attitude by the subject. It involves an effort and a purpose or, at the very least, a lack of purposelessness. This is, in part, why Wittgenstein says that attention is dynamic, Attention is dynamic, not static – one would like to say. I begin by comparing attention to gazing but that is not what I call attention; and now I want to say that I find it is impossible that one should attend statically. (Z §673)

and, If in a particular case I say: attention consists in preparedness to follow each smallest movement that may appear – that is enough to shew you that attention is not a fixed gaze: no, this is a concept of a different kind. (Z §674)10

Here Wittgenstein contrasts attention with gazing and the point of that must be to treat gazing as too much like a seeing-objects. But why is it impossible for attending to be static? The answer is implicit in the passage from §674 and the concept of preparedness. If you are attending to something you are engaging with it from a rational perspective. It is part of the point of view of the self-aswill. The concept of activity here is not an actual activity, but a potential for redeployment of what you are attending to within the space of reasons. Attending to something is a form of seeing-that. It is conceptual and thereby satisfies the generality constraint. Another way of putting this is to say that attending manifests a basic restlessness of the attitude of the will. This is not a restlessness that comes from fidgeting, it is the restlessness of being prepared to consider alternatives, of being ready to adjust your perceptions and your thoughts in order to keep the will’s engagement with the world stable. It is the restlessness that lies at the heart of agency, of being a will with the potential to be in tension with that which is independent of will. This is not a model in which attitude is something that lies behind attention. I am not suggesting that there is a simple pre-conceptual given that lies behind the conceptual structures of attention. Rather, the basic structure of experience is of the self-as-will engaged with the world. It is possible that there are some aspects of this that the self-as-will cannot handle. If so, that would introduce a notion of nonconceptual content to experience.

170


But that would not be something that comes before concepts, nor that stands between concepts and world. It would simply be a concept of content that fell between the phrasings available to one’s attention.11

6.5 The Place of Judgement In the closing pages of part 2 §11 Wittgenstein includes a passage that sums up nearly everything that I have been pressing throughout this book. After the discussion of the phenomenology of words – their physiognomy – Wittgenstein continues with the phenomenology of our experience of others. He notes that we say of some people that they are transparent to us (p. 223f) and the discussion develops to include further reflections on knowledge, certainty and belief. At p. 227 the discussion returns, ostensibly, to the topic of our knowledge of how things are for others, in particular our knowledge of whether or not expressions of feeling are genuine. He then raises the following question, Is there such a thing as ‘expert judgement’ about the genuineness of feeling? – Even here, there are those whose judgement is ‘better’ and those whose judgement is ‘worse’. (p. 227h)

His first response is simply to say ‘Correcter prognoses will generally issue from the judgements of those with better knowledge of mankind’ (ibid.). As it stands this response is barely significant, it all depends on what constitutes ‘better knowledge of mankind’. When Wittgenstein elucidates this idea he finally articulates exactly the view that I have been extracting from his work. The passage warrants quotation in full: Can one learn this knowledge? Yes; some can. Not, however by taking a course in it, but through ‘experience’. – Can someone else be a man’s teacher in this? Certainly. From time to time he gives him the right tip. – This is what ‘learning’ and ‘teaching’ are like here. – What one acquires here is not a technique; one learns correct judgements. There are also rules, but they do not form a system, and only experienced people can apply them right. Unlike calculation rules. (ibid.)

In one of the few discussions of this passage in the secondary literature, McGinn treats it as if it is merely a reflection on the indefiniteness of others’ psychological states. Her thought is that due to the indefiniteness of such states, our knowledge must exhibit a similar indefiniteness, hence the contrast between judgement and technique.12 This, I think, misses the point altogether.


171

The point goes right to the heart of the interpretation of Wittgenstein that I have been urging.13 I argued in chapter 5 that Wittgenstein’s attack on the inner/outer distinction leaves the mind as part of how things are. The minds of others are, therefore, as much part of how things are as anything else. As such, they are open to inspection. One can learn to see how things are going for others. As part of how things are, the fact that there is such a thing as the way things are going for others is not something to be tested, cautiously, by hypothesis against a backdrop of how things are. With an inner/outer distinction in place, our knowledge that there is such a thing as how things are going for another will always be precarious. It will be something to be hypothesized on the basis of knowledge about the outer realm. But with the inner/outer distinction gone, it will make no more sense to wonder whether another human being has a way things are for them as it will to wonder whether there is an external world. These are not things on which we form opinions. It is part of our basic attitude that experience presents us with how things are, and that can include how things are going for the other. ‘My attitude towards him is an attitude towards a soul. I am not on the opinion that he has a soul’ (p. 178d). We might add, our attitude to things is that they exist unperceived, we are not of the opinion that they do so. But what we accept here is the fundamental form of the attitude of judgement, that things are thus-and-so, not how they are thus-and-so: ‘What has to be accepted, the given, is – so one could say – forms of life’ (p. 226d). In learning to see how things are going for others this will, as ever, be a case of seeing-that. It will involve attention and therefore judgement. It will involve an ability to see phrasings in the fine shades of behaviour, in order to discriminate how things are for the other and to distinguish the genuine from the feigned emotion. As with all learning in seeing-that, there is no theoretical foundation to this knowledge. One has to be shown, in experience, and learn to tune one’s attitude to what the teacher points out. The aim is learning to see the other aright, it is not learning a theory about the other. That is why what you learn is judgement, not technique. The distinction between judgement and technique as Wittgenstein employs it on p. 228 is the distinction between what is delivered by seeing the world aright, and what is delivered by a theoretical articulation of how things are. The latter is capable of full and explicit articulation, the former is not. The former makes sense only within a form of life of judges with a capacity for paying attention and seeing similarities and, on the basis of such couplings with things, forming judgements that shape the ongoing use of words. Wittgenstein continues, What is most difficult here is to put this indefiniteness, correctly and unfalsified, into words. (p. 228i)

172


and of course this is difficult, for we can only put it into words for those who have learnt to judge in the way that we judge – those who share our abilities for attending to things and properties. What we see and what we judge is something we can teach others only through experience. Learning how to make judgements about the emotions of others is not a special case of language learning, it is characteristic of all language learning. We teach others language not theoretically, not by giving them the rules, but by giving them the right experiences and getting them to see things the way we do. We play shops with them and introduce them to the idea of buying five apples. We engage them in exercises of joint attention and get them to pass us building blocks and other toys. The activity here is just that – an activity, but it is a shared activity in which, by securing joint attention, we slowly help the learner see, hear and feel the phrasings in experience that we find salient.14 On this view, the primary motive for the learner in acquiring language is not to obtain a theory that explains their experience. The primary motive is to join in a shared attention with the things that matter – how things are, including how things are for others. There is a shift in imagery here with regard to the infant mind. It is the shift from the infant as proto-theorist looking for explanations of their experience to a view of the infant as an engaged will looking to join in shared forms of engagement. Of course, this means that the social dimension of our attention matters terribly, but it still matters as a resource for scaffolding, for what the infant engages with is the world. Others help it phrase its engagement, but they are not the object of engagement except in so far as others are, of course, part of the world. In continuing his discussion of our knowledge of others’ emotional states Wittgenstein says, It is certainly possible to be convinced by evidence that someone is in such-andsuch state of mind, that, for instance, he is not pretending. But ‘evidence’ here includes ‘imponderable’ evidence. (p. 228c)

Why does Wittgenstein say that the evidence is imponderable? I think the answer goes like this. Let us distinguish between evidence that it is not possible to assess by rules and evidence that it is not possible to assess by reason. Now, if evidence is imponderable, it is not possible to assess by rules. There is no rule book against which the evidence can be weighed. But that need not mean that it is evidence that it is not possible to assess by reason. It need only mean that giving reasons does not equal giving rules, employing techniques. Giving reasons can include giving judgement and offering an attitude to the world. There are many ways of offering an attitude to the world by sharing, showing and stating things. But offering an attitude is a


173

move in getting others to see as we do. It is not a ‘trump’ move, for it can be challenged, but then if evidence is imponderable, there are no ‘trump’ moves. Training in language is a training to be a judge. This comes from experience, not from the statement of rules. There are consequences of our adopting an attitude but of a diffuse kind. Experience . . . can inform us of them, and they too are incapable of general formulation; only in scattered cases can one arrive at a correct and fruitful judgement, establish a connexion. (p. 228a, cf. also RPP II §§694, 695, 699)

Training to be a judge includes, centrally, coming to see similarities (§69) and coming to speak with right (§289). This is a coming to see justification in the particularities (§154), it is seeing things aright. Because there are no trump moves in judgement, there is a dynamism to the concept of attitude in Wittgenstein’s later writings. This is the dealing with the ‘bustle’.15 In many respects the bustle is not necessarily social, although it does include the meeting of wills, things that move. When this occurs it is dynamic, it is the meeting of will on will. When wills meet, patterns emerge – economic patterns provide the most obvious case of this. In such cases wills agree, they tally, or they fail to agree. And some of those agreements and tallies become, in turn, objects of our attitude – we talk about them. Their constitution is dynamic because they are constituted by a meeting of wills. It takes wills to tango. The involvement of will means that it is a case of judgement all the way down. This has all been about the conditions for the possibility of judgement. The recipe for that is the will in direct engagement with that which is independent of will, and the latter can include the attitude of another will.

6.6 What Comes Natural Wittgenstein had a concept of second nature. He does not always clearly distinguish it from what is natural in the sense of our causal engagements with things. Second nature requires a directedness to our causal engagements. It is what is at stake when we get the point and know what’s going on. It is what is involved when we see the ‘and so on . . .’, when we take the right attitude to things. Once this concept of second nature is in view, it seems to me likely that most, if not all, of Wittgenstein’s examples of the ‘natural’ are already second nature. The point certainly bears closer detailed investigation. Let me close with a few simple reminders. The natural cry of pain (§244) is the cry of the human subject and this, unlike the animal subject, is already a subject with a directedness to its

174


engagements. The temptation to read Wittgenstein as a behaviourist comes from failing to see that the ‘natural’ is often described within the context of such directedness. As I remarked earlier, even in the most primitive language-game – the builders of §2 – we have directed engagements in block moving.16 They are not fetching and carrying as an animal does, they are acting with purpose. In §6 Wittgenstein imagines the language of §2 is the language of a whole tribe: The children are brought up to perform these actions, to use these words as they do so, and to react in this way to the words of others.

This sounds behaviourist, but note what comes next: ‘An important part of the training will consist in the teacher’s pointing to the objects, directing the child’s attention to them.’ The infant has the capacity to attend. It has more than mere causal accommodations with things, it has a capacity for directedness. Wittgenstein has retained the Augustinian insight that the subject is a self-as-will. Even in the middle of the criticism of ostensive definition Wittgenstein freely talks of the structure of attention: ‘You attend to the shape, sometimes by tracing it, sometimes by screwing up your eyes so as not to see the colour clearly, and in many other ways’ (§33). But as he carries on, ‘it isn’t these things by themselves that make us say someone is attending to the shape, the colour, and so on’ (ibid.). In other words, ‘attention’ is the right word for what happens, but attending is not just tracing the shape, screwing up your eyes, etc. These are things associated with seeing-objects. They bear upon seeing-objects, but attention is much richer. It is purposeful seeing-that – a seeing with attitude.17 And recall, in §258 the private linguist attempts to ‘concentrate my attention on the sensation’. What puzzles the Cartesian about the thrust of the private language argument is they think that they can see and feel their sensations. Their vocabulary is wrong, for they do not perceive sensations at all – they have them. But if language use is to get going here, we need more than brute seeing-object perception. We need seeing-that, we need attention. It is not the brute passivity of seeing-objects that brings the world into view in language, it is seeing-that. And it is that kind of seeing that cannot be calibrated in Cartesian subjectivity. It requires purposeful seeing, attention to what which is independent of will. That is what the private linguist cannot substantiate. They lose their grip on the idea that when we see things aright, we enjoy a directed engagement with that which is independent of will. So natural is this view of ourselves that we can forget to remark on it for what it is. But it is not just breathing out and breathing in, it is setting your will against the world and testing and probing to achieve best fit between will and non-will.


175

Notes 1 This has obvious bearing on issues in the philosophy of education. For application of this to an account of expertise, see M. Luntley, ‘Knowing How to Manage: Expertise and Embedded Knowledge’, Reason in Practice, 2, 2002, 3–14, and ‘Growing Awareness’, plenary address to the Philosophy of Education Society of Great Britain, annual conference, New College, Oxford, April 2003. There is also a tradition of scholarship, mainly Scandinavian, on Wittgenstein that explores the themes of expertise and work-based learning and judgement: cf. B. Goranzan and M. Florin (eds), A.I., Culture and Language: On Education and Work, Berlin: Springer-Verlag, 1990; B. Goranzan and M. Florin (eds), Skills and Education, Berlin: Springer-Verlag, 1992; A. Janik, Style, Politics and the Future of Philosophy, Dordrecht: Kluwer Academic Publishers, 1989, esp. chapter XIV. Thanks to Alois Pichler for pointing me in the direction of this work. 2 The obvious exception is S. Mulhall, On Being in the World, London: Routledge, 1990, and ‘Seeing Aspects’, in Wittgenstein: A Critical Reader, ed. H.-J. Glock, Oxford: Blackwell, 2001, pp. 246–67; although see also R. Fogelin, Wittgenstein, London: Routledge, 1987; M. Budd, Wittgenstein’s Philosophy of Psychology, London: Routledge, 1989; and M. McGinn, Wittgenstein and the Philosophical Investigations, London: Routledge, 1997, for further useful discussion. Mulhall’s is the most developed discussion thus far of the material from part 2 §11. 3 This is to agree, broadly, with Mulhall’s interpretation of this section, although even Mulhall fails to draw the full import of what Wittgenstein is doing here. 4 McGinn, op. cit., chapter 6, mostly treats this section as presenting the negative thesis. I think this underestimates the interest and resources available within the section. 5 The concept of second nature is slippery. In the song ‘I’ve Grown Accustomed to Her Face’ (My Fair Lady), Henry Higgins sings ‘It’s second nature to me now, like breathing out and breathing in.’ This is surely wrong. Breathing, much like Wittgenstein’s example of chewing, is second nature in the sense of being a basic causal habit, one that we do not need to think about to indulge. It is a causal response that exhibits none of the directedness of causal dispositions that a conceptually structured second nature possesses. McDowell’s concept of second nature is, I suggest, that of directed causal responses; responses that are conceptually structured all the way down. Wittgenstein, like Henry Higgins, often runs the two together. 6 Recall §176, ‘I should like to say that I had experienced the ‘‘because’’, and yet I do not want to call any phenomenon the ‘‘experience of the because’’.’ Cf. also §177, and the discussion of this point in chapter 4 above. 7 ‘What we learn is not just what we have studied; and what we have been taught is not just what we were intended to learn. What we have in our memories is not just what we have memorized.’ S. Cavell, ‘Excursus on Wittgenstein’s Vision of Language’, in The New Wittgenstein, eds A. Crary and R. Read, London: Routledge, 2000, pp. 21–37, esp. p. 28.

176 8 9 10 11

12 13

14

15 16 17

Seeing Things Aright The space of our fundamental attitude is what McDowell, following Sellars, calls the space of reasons. I explore this remark further in my ‘Growing Awareness’, op. cit., and also in An Education of Character (in preparation). See also RPP II §512, §520. D. Bell, ‘The Art of Judgement’, Mind, XCVI, 1987, 221–44, tried to introduce a notion of nonconceptual content in order to capture the idea that aesthetic experience can fit together in a way that is beyond the reach of attention. Bell took this as a preconceptual notion of nonconceptual content, a level of content out of which an account of conceptual content might be given. See also S. Saatela, ‘Wittgenstein on ‘‘Aesthetic Reactions’’ ’, Revue Internationale de Philosophie, 2002, 49–72, for a similar view, in which aesthetic experience is treated as composed of preconceptual ‘reactions’. My own view is that there is space for a notion of nonconceptual content in characterizing aesthetic experience, but that this is not a level of preconceptual content. See M. Luntley, ‘Nonconceptual Content and the Sound of Music’, Mind and Language, forthcoming, 2003. The account of nonconceptual content that I advocate is, in many respects, a notion of content that is very sophisticated in that it becomes available after one has stripped away the conceptual content. Apart from the apparent reintroduction of the notion of the given, both Bell and Saatela defend a view in which aesthetic content is a lower level of content than the conceptual. My suspicion is that this might, in some cases at least, have matters the wrong way around. The matter deserves much further discussion, for which there is not space here. But treating the aesthetic content of experience as a ‘higher’ although nonconceptual content might even help make sense of some of the most puzzling passages in part II §11 – the passages where Wittgenstein introduces secondary meanings in the discussion of ‘Wednesday is fat’ and ‘Tuesday is lean’. I leave these matters for another occasion. McGinn, op. cit., p. 169. The heart of the matter concerns the concept of judgement. In the literature on education and expertise, there have been frequent references to the role of judgement: see n. 1 above, and also P. Hager, ‘Know-How and Workplace Practical Judgement’, Journal of Philosophy of Education, 34, 2000, 281–96. Judgement is central, but most of the extant literature simply invokes the concept without showing how it embeds in other cognitive skills and, therefore, provides no real understanding. For more on joint attention, cf. Joint Attention, Communication and Other Minds: Issues in Philosophy and Psychology, eds N. Eilan, C. Hoerl, T. McCormack and J. Roessler, Oxford: Oxford University Press, forthcoming. Cf. chapter 5, §6. Cf. chapter 4, §5. Similar remarks apply to other passages, e.g. see the role of attention throughout the discussion of ostensive definition, and later, e.g. in §156.

Bibliography

Anscombe, E., An Introduction to Wittgenstein’s Tractatus, London: Hutchinson, 1971. Baker, G. P., ‘Wittgenstein on Metaphysical/Everyday Use’, The Philosophical Quarterly, 52, 2002, 289–302. Baker, G. and Hacker, P., Scepticism, Rules and Language, Oxford: Blackwell, 1984. Bell, D., ‘The Art of Judgement’, Mind, XCVI, 1987, 221–44. Black, M., A Companion to Wittgenstein’s Tractatus, Ithaca, NY: Cornell University Press, 1964. Bloor, D., Wittgenstein, Rules and Institutions, London: Routledge, 1997. Brandom, R., Making it Explicit: Meaning, Representing and Discursive Commitment, Cambridge, Mass.: Harvard University Press, 1994. Brandom, R., Articulating Reasons, Cambridge, Mass.: Harvard University Press, 2000. Budd, M., Wittgenstein’s Philosophy of Psychology, London: Routledge, 1989. Campbell, J., ‘Is Sense Transparent?’, Proceedings of the Aristotelian Society, 88, 1988, 273–92. Candlish, S., ‘The Will’, in Wittgenstein: A Critical Reader, ed. H.-J. Glock, Oxford: Blackwell, 2001, pp. 156–73. Carruthers, P., The Metaphysics of the Tractatus, Cambridge: Cambridge University Press, 1990. Cavell, S., ‘Notes and Afterthoughts on the Opening of Wittgenstein’s Investigations’, in The Cambridge Companion to Wittgenstein, eds H. Sluga and D. Stern, Cambridge: Cambridge University Press, 1996. Cavell, S., ‘Excursus on Wittgenstein’s Vision of Language’, in The New Wittgenstein, eds A. Crary and R. Read, London: Routledge, 2000, pp. 21–37. Clark, A., Being There, Cambridge, Mass.: MIT Press, 1998. Conant, J., ‘Must We Show What We Cannot Say’, in The Senses of Stanley Cavell, eds R. Flemming and M. Payne, Lewisburg: Bucknell University Press, 1989, pp. 242–83. Crary, A., ‘Introduction’, in The New Wittgenstein, eds A. Crary and R. Read, London: Routledge, 2000, pp. 1–18.

178

Bibliography

Dancy, J., An Introduction to Contemporary Epistemology, Oxford: Blackwell, 1985. Davidson, D., ‘On the Very Idea of a Conceptual Scheme’, in his Essays on Truth and Interpretation, Oxford: Oxford University Press, 1984. Davidson, D., ‘A Coherence Theory of Truth and Knowledge’, in Truth and Intrepretation: Perspectives on the Philosophy of Donald Davidson, ed. E. Lepore, Oxford: Blackwell, 1986, pp. 307–19. Diamond, C., ‘Throwing Away the Ladder: How to Read the Tractatus’, in The Realistic Spirit, Cambridge, Mass.: MIT Press, 1995, pp. 179–204. Eilan, N., ‘The Reality of Consciousness’, in Wittgensteinian Themes: Essays in Honour of David Pears, eds D. Charles and W. Child, Oxford: Clarendon Press, 2001, pp. 163–94. Eilan, N., Hoerl, C., McCormack, T. and Roessler, J. (eds), Joint Attention, Communication and Other Minds: Issues in Philosophy and Psychology, Oxford: Oxford University Press, forthcoming. Engelmann, P., Letters from Ludwig Wittgenstein with a Memoir, ed. N. Malcom, Oxford: Blackwell, 1987. Evans, G., The Varieties of Reference, ed. J. McDowell, Oxford: Clarendon Press, 1982. Fodor, J., Psychosemantics, Cambridge, Mass.: MIT Press, 1987. Fogelin, R., Wittgenstein, London: Routledge, 1987. Garfield, J., ‘Particularity and Principle: The Structure of Moral Knowledge’, in Moral Particularism, eds B. Hooker and M. Little, Oxford: Clarendon Press, 2000, pp. 178–204. Glendinning, S., On Being with Others: Heidegger, Wittgenstein and Derrida, London: Routledge, 1998. Goranzan, B. and Florin, M. (eds), A.I., Culture and Language: On Education and Work, Berlin: Springer-Verlag, 1990. Goranzan, B. and Florin, M. (eds), Skills and Education, Berlin: Springer–Verlag, 1992. Grayling, A., Wittgenstein, Oxford: Oxford University Press, 1988. Hacker, P. M. S., Insight and Illusion, Oxford: Clarendon Press, 1972, 2nd edn, 1986. Hacker, P. M. S., Wittgenstein: Meaning and Mind. Volume 3 of an Analytical Commentary on the ‘Philosophical Investigation’, Oxford: Blackwell, 1990. Hacker, P. M. S., Wittgenstein’s Place in Twentieth-Century Analytic Philosophy, Oxford: Blackwell, 1996. Hacker, P. M. S., ‘Philosophy’, in Wittgenstein: A Critical Reader, ed. H.-J. Glock, Oxford: Blackwell, 2001, pp. 322–47. Hager, P., ‘Know-How and Workplace Practical Judgement’, Journal of Philosophy of Education, 34, 2000, 281–96. Hark, M. Ter, Beyond the Inner and the Outer, Dordrecht: Kluwer Academic Press, 1990. Hark, M. Ter, ‘The Inner and the Outer’, in Wittgenstein: A Critical Reader, ed. H.-J. Glock, Oxford: Blackwell, 2001, pp. 199–223. Hintikka, J. and Hintikka, M., Investigating Wittgenstein, Oxford: Blackwell, 1986. Hinton, J. M., Experiences: An Enquiry into Some Ambiguities, Oxford: Clarendon Press, 1973.

Bibliography

179

Hooker, B. and Little, M. (eds), Moral Particularism, Oxford: Clarendon Press, 2000. Ishiguro, H., ‘The Use and Reference of Names’, in Studies in the Philosophy of Wittgenstein, ed. P. Winch, London: Routledge, 1969. Ishiguro, H., ‘Can the World Impose Logical Structure on Language?’, in Wittgenstein – Towards a Re–evaluation, Proceedings of the 14th International Wittgenstein Symposium, eds R. Haller and H. Brandl, Vienna: Verlag, 1990, pp. 21–34. Ishiguro, H., ‘The So–Called Picture Theory: Language and the World in Tractatus Logico Philosophicus’, in Wittgenstein: A Critical Reader, ed. H.-J. Glock, Oxford: Blackwell, 2001, pp. 26–46. Jackson, F., Perception: A Representative Theory, Cambridge: Cambridge University Press, 1977. Jackson, F., Pettit, P. and Smith, M., ‘Ethical Particularism and Patterns’, in Moral Particularism, eds B. Hooker and M. Little, Oxford: Clarendon Press, 2000, pp. 79–99. Janik, A., Style, Politics and the Future of Philosophy, Dordrecht: Kluwer Academic Publishers, 1989. Kenny, A., Wittgenstein, London: Allen Lane Press, 1973. Kenny, A., The Legacy of Wittgenstein, Oxford: Blackwell, 1984. Koethe, J., The Continuity of Wittgenstein’s Thought, Ithaca, NY: Cornell University Press, 1996. Kripke, S., Wittgenstein on Rules and Private Language, Oxford: Blackwell, 1982. Luntley, M., ‘The Transcendental Grounds of Meaning and the Place of Silence’, in Meaning Scepticism, ed. K. Puhl, Berlin and New York: De Gruyter, 1991, pp. 170–88. Luntley, M., ‘The Practical Turn and the Convergence of Traditions’, Philosophical Explorations, 1, 1998, 10–27. Luntley, M., ‘Postmodernism and the Education of Character’, in Education in Morality, eds H. Halstead and T. Mclaughlin, London: Routledge, 1999, pp. 185–205. Luntley, M., Contemporary Philosophy of Thought: Truth, World, Content, Oxford: Blackwell, 1999. Luntley, M., ‘Agency and our Tacit Sense of Things’, in Library of Living Philosophers volume for Marjorie Grene, ed. L. Hahn, La Salle: Open Court, 2002, pp. 151–71. Luntley, M., ‘Knowing How to Manage: Expertise and Embedded Knowledge’, Reason in Practice, 2, 2002, 3–14. Luntley, M., ‘Patterns, Particularism and Seeing the Similarity’, in Philosophical Papers, Grahamstown, SA: Rhodes University Press, 2002. Luntley, M., ‘Attention, Time and Purpose’, Philosophical Explorations, VI, 2003, 2–17. Luntley, M., ‘Nonconceptual Content and the Sound of Music’, Mind & Language, forthcoming, 2003. Luntley, M., ‘Growing Awareness’, plenary address to the Philosophy of Education Society of Great Britain, annual conference, New College, Oxford, April 2003. McDowell, J., ‘Criteria, Defeasibility and Knowledge’, Proceedings of the British Academy, 68, 1982, 455–79, reprinted in J. McDowell, Mind, Value and Reality, Cambridge, Mass.: Harvard University Press, 1998, pp. 369–94.

180

Bibliography

McDowell, J., ‘De re senses’, The Philosophical Quarterly, 34, 1984, 283–94. McDowell, J., ‘Wittgenstein on Following a Rule’, Synthese, 58, 1984, 325–63. McDowell, J., ‘Singular Thoughts and the Extent of Inner Space’, in Subject, Thought and Context, eds J. McDowell and P. Pettit, Oxford: Clarendon Press, 1986, pp. 137–68. McDowell, J., ‘Meaning and Intentionality in Wittgenstein’s Later Philosophy’, in The Wittgenstein Legacy. Midwest Studies in Philosophy, vol. XVII, Notre Dame, Ind.: University of Notre Dame Press, 1992, pp. 40–52. McDowell, J., Mind and World, Cambridge, Mass.: Harvard University Press, 1994. McDowell, J., ‘Having the World in View: Sellars, Kant and Intentionality’, The Journal of Philosophy, XCV, 1998, 431–91. McDowell, J., Mind, Value and Reality, Cambridge, Mass.: Harvard University Press, 1998. McDowell, J., ‘Scheme–content Dualism and Empiricism’, in The Philosophy of Donald Davidson, ed. L. Hahn, La Salle: Open Court, 1999, pp. 87–104. McGinn, C., Wittgenstein on Meaning: An Interpretation and Evaluation, Oxford: Blackwell, 1984. McGinn, M., Wittgenstein and the Philosophical Investigations, London: Routledge, 1997. McGinn, M., ‘Between Metaphysics and Nonsense: Elucidation in Wittgenstein’s Tractatus’, The Philosophical Quarterly, 49 (197), 1999, 491–513. McGuinness, B., ‘The So-Called Realism of the Tractatus’, in Perspectives on the Philosophy of Wittgenstein, ed. I. Block, Oxford: Blackwell, 1981. Malcolm, N., ‘Wittgenstein’s Philosophical Investigations’, in Knowledge and Certainty, Englewood Cliffs: Prentice-Hall, 1963. Malcolm, N., Nothing is Hidden: Wittgenstein’s Critique of his Early Thought, Oxford: Blackwell, 1986. Malcolm, N., ‘Wittgenstein on Language and Rules’, Philosophy, 64, 1989, 5–28. Mulhall, S., On Being in the World, London: Routledge, 1990. Mulhall, S., ‘Seeing Aspects’, in Wittgenstein: A Critical Reader, ed. H.-J. Glock, Oxford: Blackwell, 2001, pp. 246–67. Ormerod, P., The Death of Economics, London: Faber, 1994. Pears, D., The False Prison, vols I and II, Oxford: Clarendon Press, 1987/1988. Ricketts, T., ‘Pictures, Logic and the Limits of Sense in Wittgenstein’s Tractatus’, in The Cambridge Companion to Wittgenstein, eds H. Sluga and D. Stern, Cambridge: Cambridge University Press, 1996. Saatela, S., ‘Wittgenstein on ‘‘Aesthetic Reactions’’ ’, Revue Internationale de Philosophie, 56 (219), 2002, 49–72. Schroeder, S., ‘Private Language and Private Experience’, in Wittgenstein: A Critical Reader, ed. H.-J. Glock, Oxford: Blackwell, 2001, pp. 174–98. Searle, J., The Construction of Social Reality, London: Penguin, 1996. Sellars, W., ‘Empiricism and the Philosophy of Mind’, in Minnesota Studies in the Philosophy of Science, vol. 1, eds H. Feigl and M. Scriven, Minneapolis: Minnesota University Press, 1956.

Bibliography

181

Smith, N. (ed.), Reading McDowell – On Mind and World, London: Routledge, 2002. Snowdon, P. F., ‘Perception, Vision and Causation’, Proceedings of the Aristotelian Society, 81, 1980–1, 121–50. Snowdon, P. F., ‘How to Interpret Direct Perception’, in The Contents of Experience, ed. T. Crane, Cambridge: Cambridge University Press, 1992, pp. 48–78. Stroud, B., ‘Mind, Meaning and Practice’, in The Cambridge Companion to Wittgenstein, eds H. Sluga and D. Stern, Cambridge: Cambridge University Press, 1996, pp. 296–319. Wiggins, D., ‘What Would Be a Substantial Theory of Truth?’, in Philosophical Subjects: Essays Presented to P. F. Strawson, ed. Z. van Straaten, Oxford: Clarendon Press, 1980, pp. 76–117. Wiggins, D., ‘Truth as Predicated of Moral Judgements’, in Needs, Values, Truth: Essays in the Philosophy of Value, Oxford: Blackwell, 1987, pp. 139–84. Williams, M., Wittgenstein, Mind and Meaning: Towards a Social Conception of Mind, London: Routledge, 1999. Wright, C., Realism, Meaning and Truth, Oxford: Blackwell, 1986. Wright, C., Wittgenstein on the Foundations of Mathematics, Cambridge, Mass.: Harvard University Press, 1987.

Index

aboutness, 1, 17 action, 32, 54, 56, 70, 77–9, 85, 101–2, 108, 111–13, 117–18, 144–6, 148, 152, 168 agency, 2, 56, 169 agent, 1–2, 18, 57, 66–7, 108, 117, 152, 162–3, 168 agreement, 142–5 in attitude, 145 in forms of life, 142 in opinion, 145 and so on . . . (concept of), 43–4, 173 animatory theory of meaning, 2–3, 8, 16, 18, 161 Anscombe E., 47 n.23, 89 n.13 anti-realism, 82 aspect-dawning, 155, 157 aspect perception, 148, 155, 157, 161 attention, 11, 22, 76, 80, 95, 108, 113, 115, 119, 128, 159, 161, 164–6, 167–72, 174 attention-dependent, 76–7 attitude, 1–2, 18, 21–2, 24, 26–7, 31–3, 35, 41–5, 76, 115, 133, 139, 142, 144–8, 152–4, 162–9, 171–4 basic, 2, 163–4, 168, 171 direct, 113

Augustine, 67–9, 150 n.9, 163–4 avowals, 126 non-cognitive theory of, 126–7 Baker G. P., 20 n.10, 88 n.5, 120 n.6, 121 n.10, 149 n.8 behaviourism, 105, 115–16, 127–8 behaviourist, 126, 174 Bell D., 176 n.11 bipartite, 17–18, 23, 25, 49, 58, 61, 63, 65, 70, 72, 93, 96–103, 105, 109, 111–14, 155, 158, 162–3 bipartite account of content, 23, 25, 49, 58, 61, 63, 65, 70, 93, 96, 97–8, 100, 103, 105, 109, 112–13, 114, 162–3 bipartite account of meaning, 3, 9, 17–18, 31, 49, 96, 100–2, 104 Black M., 36, 47 n.18, 89 n.13 Bloor D., 19 n.10, 90 n.19, 90 n.23, 120 n.3, 120 n.7, 120 n.8 Brandom R., 17, 19 n.3, 20 n.11, 37, 90 n.17, 120 n.7, 122 n.22 Budd M., 149 n.3, 175 n.2 bustle (see also hurly-burly), 144–5, 173 calibration, 15, 48–9, 55, 63, 66, 71–2, 94, 96, 97, 107, 112–15, 118, 124–9, 133–9, 142, 174

Index direct, 55, 62, 71–2, 97, 113–14, 118, 154, 158 Campbell J., 91 n.29 Candlish S., 151 n.29 canonical form, 73–4 Carruthers P., 40, 47 n.27 Cartesian, 8, 11–15, 24, 34, 36, 67, 104, 107, 124–5, 127, 130, 134–5, 137, 138, 140–2, 146, 152, 174 causal engagement, 111, 115, 146, 153, 164, 173 Cavell S., 70–1, 88 n.2, 90 n.24, 121 n.9, 166, 175 n.7 character, 18, 22, 24, 28–9, 33, 42, 53, 83, 115, 130, 133, 142, 147, 155–6, 167, 168 Clark A., 92 n.40 cognitive transfer, 87 collectivist, 93–5, 98 combinatorial structure, 73, 79 common factor account, 103 communitarian, 16, 109, 113, 144 community, 8, 12, 15–16, 45, 94, 96, 114–15, 136–7, 168 Conant J., 45 n.5, 46 n.13 concept possession, 78–9, 82, 86–7, 106, 108, 111, 153 conceptual content, 83, 157–8, 160 content of experience, 102–3, 129 content, invariant, 102–4, 132 contingencies, catalogue of, 51, 55, 61, 70–1, 80, 110, 154 continuous aspect perception (see also seeing-that), 148, 155, 157, 159, 165–7 contract, 101 contractualism, 100–3 conventionalism, 33 correctness conditions, 13, 17, 28, 39, 48–9, 57, 61, 64, 66, 70–2, 79, 94–6, 98, 100–1, 107–13, 118, 125–6, 129, 134, 136, 137, 146–7, 154, 158, 168

183

correctness/incorrectness, 48–9, 61, 71–2, 79, 95, 107, 109, 112–13, 118, 126, 129, 134, 136, 146, 154 conditions, 13, 48, 96 correlation, 36 coupling, 77, 85–7, 171 Crary A., 88 n.1, 88 n.4, 88 n.5, 90 n.14 creativity, 86–7, 108, 153 Dancy J., 121 n.13 Davidson D., 46 n.14, 64, 90 n.18 demonstrative, 40, 77 Descartes 15, 103, 141–2 Diamond C., 46 n.6, 46 n.13 direct presence, 107, 111, 115, 118 directedness, 108, 110–11, 115, 117–18, 146, 152–3, 163–4, 173–4 direction (of fit), 37–9 disjunctive account (of perception), 104–5 display, 30, 95–8, 101–3, 105, 109–11, 114 dispositions, 112, 116, 153 Eilan N., 150 n.10 empirical self, 22, 24, 33–5, 45, 64, 108, 146 engagement with others, 93 with things, 93, 108, 112–13, 152, 154, 158 Engelmann, 38, 47 n.25 enthymematic, 76–7 epistemic fragility, 130 error, 39–41, 49, 62, 94, 103, 129, 132, 136, 166 deferred, 41 direct, 103 genuine, 39–40, 49 perceptual, 103 essentialism, 22, 33, 41–2, 44, 50, 51–2, 56–7, 59–60, 64, 67, 78, 79–80, 82, 107 Evans G., 19 n.4, 47 n.8, 107, 121 n.15, 123 n.25

184

Index

experience, 13–15, 25, 42–4, 100–4, 108, 115–19, 125–6, 129–32, 139, 152–9, 165, 167, 169–73 illusory, 102 inertness of, 14–15 transparency, 14, 103–5 family resemblance, 78–80, 165 Fodor J., 19 n.7 Fogelin R., 175 n.2 form of life, 45, 60, 70–2, 145, 171 Frege G., 31, 41, 59 friction, 55, 118–19, 134

hidden, 50–5, 57, 59–60, 62, 127–9, 131–3 immanent, 53–7, 66, 99 nothing is, 13, 60, 62 transcendent, 53–4, 57, 59–60 Hintikka J. and M., 46 n.12, 46 n.17 Hinton J., 121 n.14 Hume D., 13 hurly-burly, 45, 144–5

Garfield J., 91 n.34 generality, 40, 106–7, 117, 169 generality constraint, 107, 117, 169 Glendinning S., 19 n.10, 120 n.2 grammar, 3–10, 12–16, 18, 21–30, 32–3, 35, 37, 39, 41–3, 45, 52–4, 56–8, 61–2, 66, 69, 74, 76, 77, 87, 107, 119, 126, 134, 143, 167 Cartesian source of, 11, 24 community source of, 15–16 platonist source of, 9 source of, 8–10, 12, 15–16, 21, 27, 34, 52, 61, 65 grammatical investigation, 62, 80, 137 grammatical position, 107, 119 grasp in a flash, 154, 165–7 Grayling A., 149 n.3

imponderable evidence, 172–3 individualist, 94–6, 107, 146 individualistic, 142, 168 inference transparency of, 74, 76 unity of, 30, 44 inferential power, 3, 5–6, 8, 12, 13, 37, 58–9 inferentialism, 17, 23, 28, 37, 39–41, 58–9, 67 inner, 12, 15, 68–9, 116, 118, 124–5, 127–30, 132–3, 135–6, 138, 142–3, 146–7, 152, 155, 171 inner/outer distinction, 124–5, 128–30, 136, 142–3, 171 inner process, 125 intentionality, 1–2, 5, 11, 17, 18, 43, 97, 109, 111 internal relation, 101–2, 159–60, 162 interpretation, 35, 37, 59, 61, 81, 84, 94–8, 100, 105–6, 112, 137, 154–6, 158–9, 162–3, 171 Ishiguro H., 36–8, 40, 47 n.18, 89 n.13

Hacker P., 20 n.10, 45 n.4, 46 n.9, 46 n.12, 46 n.15, 46 n.17, 120 n.6, 120 n.8, 121 n.10, 149 n.3, 149 n.7 Hager P., 176 n.3 Hark M. Ter, 112, 121 n.12, 122 n.20, 129, 149 n.6, 151 n.28 hearing-objects, 167 hearing-that, 167 Heidegger M., 112, 122 n.19

Jackson F., 91 n.33, 92 n.37, 121 n.13 judge, 2–4, 6, 13, 18, 56–7, 77, 84, 93–4, 108, 115, 144, 153–4, 167, 169, 171–3 judgement, 1, 3–5, 7, 13, 18, 31, 50, 56–7, 66–7, 75, 77, 79, 84, 88, 93–4, 99–100, 108, 136, 142, 144, 152–3, 155, 166, 168–73 justification, 83–4, 100, 102, 112, 117, 146, 148, 156, 159, 168, 173

Index Kenny A., 46 n.12, 46 n.17 Koethe J., 46 n.9 Kripke S., 20 n.10, 52, 61, 69–71, 89 n.6, 89 n.10, 90 n.23, 122 n.22 language game, 70, 72, 110, 113, 116, 137, 143 living human being, 72, 128, 131–3 logical atomism, 22, 28, 39 logical form, 22, 27, 29–30, 33–5, 43–4 logical grammar, 26, 53, 63, 74, 76–7 logical must, 119 Luntley M., 47 n.28, 89 n.9, 90 n.29, 91 n.31, 92 n.40, 120 n.4, 122 n.18, 122 n.23, 123 n.25, 175 n.1, 176 n.11 Magritte R., 30 Malcolm N., 19 n.10, 45 n.4, 149 n.3 McDowell J., 19 n.1, 19 n.8, 20 n.10, 47 n.31, 49, 88 n.1, 89 n.9, 89 n.11, 90 n.18, 118, 120 n.4, 120 n.5, 121 n.9, 121 n.14, 121 n.16, 122 n.18, 123 n.25, 123 n.26, 140, 150 n.22, 150 n.23 McGinn C., 20 n.10, 149 n.3 McGinn M., 20 n.10, 26, 45 n.3, 46 n.7, 90 n.16, 90 n.20, 91 n.36, 137, 143, 149 n.1, 149 n.4, 149 n.5, 150 n.21, 170, 175 n.2, 175 n.4 McGuinness B., 40, 47 n.20, 89 n.13 Merleau-Ponty M., 112 metaphysical subject, 22, 33, 35–6, 41–2, 44 mindedness, idea of, 128, 130–1, 138–41 Mulhall S., 155, 175 n.2 Nagel T., 150 n.10 names, 5–6, 11–12, 22–3, 28–9, 33, 36–40, 67, 75–7, 126, 137 theory of, 38 natural history (of mankind), 68 norm, 52, 62–3, 66, 69, 94, 96–103 normative engagement, 42, 134

185

patterns, 6, 7, 9, 17, 55, 64, 79, 82, 112, 137, 167 standards, 49, 51, 58, 61–2, 69–70, 93, 95, 110, 113, 134, 137–8, 158, 168 normativity, 6, 9, 13–18, 48–9, 55, 57, 61, 65–7, 72, 77–9, 82, 95–8, 108, 111, 113, 119 normless sign display, 17–18, 95–7 ‘now I can go on’, 148, 154, 165–6 object-dependent, 40–1 object-involving, 39–42 objectivity, 42, 45, 48–9, 71, 141–2 objects, 5–6, 11, 14, 22–3, 28–9, 31–4, 36–8, 40–2, 45, 55, 67, 94, 103, 111, 116, 131, 136, 147, 156–7, 159, 160, 173–4 ontological argument, 142 Ormerod P., 151 n.27 ostensive definition, 11, 58–9, 68, 134, 174 outer criteria, 125 pain, 18, 95, 107, 125–7, 129–31, 133–4, 137, 139, 143, 147–8, 173 particular circumstances, 57, 80, 83, 88, 117, 119, 145 patterns (of word use), 72, 79, 82–3, 135 patterns, emergent, 87, 145, 173 Pears D., 20 n.10, 20 n.13, 34, 37–8, 40–1, 45 n.4, 46 n.11, 47 n.19, 89 n.12, 90 n.15, 120 n.1, 120 n.4, 136–7, 149 n.4, 149 n.5, 150 nn.12–18, 151 n.25 perception, 79, 85, 102–4, 108, 132–3, 155–60, 162–3, 166–8, 174 perceptual experience, 104, 155, 157–8 perspective (point of view), 1, 14, 18, 22–4, 33, 35, 39, 52–3, 75–6, 78, 97, 98, 104, 114, 131–3, 135–6, 139–40, 142–3, 146–8, 152, 168–9 perspicuous representation (clear view), 26, 51–2 phenomenology, 105, 154–5, 158–9, 165, 167, 170

186

Index

phrasing, 163–5, 168, 170–2 Pichler A., 19 n.2, 175 n.1 pictorial form, 23, 25, 29–30, 36 pictorial relationship, 29, 35 picture theory (of propositions), 22–3, 27, 36 platonism, 33 platonist, 8–10, 12, 15, 63, 84 practice, 50, 55, 62–4, 66, 72, 82, 93–4, 96–8, 101–2, 105–15, 117, 154, 158 private, 15, 107, 124–8, 132–42, 146, 174 calibration, 124–5 language, 15, 107, 124, 128, 133, 135–42, 146, 174 linguist, 124–7, 133–42, 174 proposition, unity of, 44 public standards of use, 125, 137 purpose, 26, 39, 61, 115–16, 126–7, 155–6, 161, 169, 174 quietism, 99 quietist, 99 rational force (of language use), 79–80, 84, 88 realism, 7, 13, 18, 21, 24, 34, 41–2, 62–3, 65, 72, 77, 79, 82–4, 87, 119 realist, 7, 9, 27, 28, 34–6, 103 reasons, generalism about, 77, 79 receptivity, 56, 157, 160 reference, 16, 24, 37, 38–41, 59, 65–6, 114, 143 representational power, 3, 5–6, 8, 11–13, 15, 17, 23, 37, 58–9 representationalism, 17, 28, 37, 41, 59, 67 revelatory (of description), 17, 52, 54, 56, 63, 65, 66, 99 Ricketts T., 46 n.9 right/seems right distinction, 62, 127, 130, 133–4, 141 rule-following, 16, 44, 52, 57, 61, 69, 81–2, 93–5, 97–9, 105–6, 119, 165–6 rules, 3, 5, 9, 17, 23, 25, 31, 49, 57, 64, 77, 79, 82–4, 95, 119, 153, 161, 170–3

Russell B., 29, 41, 53, 67, 137 Ryle G., 54 Saatela S., 89 n.8, 176 n.11 Schroeder S., 20 n.10, 20 n.11, 129, 149 n.2, 151 n.25 Searle J., 144, 151 n.27 second nature, 158, 173 seeing similarities, 78–81, 83–4, 86–8, 101, 108, 148, 153, 158, 171 seeing things aright, 18, 21, 27, 30, 35, 43, 45, 49–51, 54, 56, 58, 63, 80, 93, 115, 148, 152, 154, 155, 166, 171, 173–4 seeing-as, 156, 158, 166 seeing-objects, 156–60, 162–7, 169, 174 seeing-that, 23, 156–60, 162–3, 165–6, 168–9, 171, 174 self, 1–2, 21–2, 41–2, 44, 56, 64–7, 108, 115, 124, 142, 146, 152, 162–4, 168–9 self-as-will, 94, 108, 118, 174 self-authenticating, 83 Sellars W., 19 n.1 semantic power, 3, 5–6 sensation(s), 124–8, 135–7, 139, 140, 146, 147, 152, 174 sensation word, 124–8, 135, 137, 140 sense, transparency of, 76 showing/saying distinction, 26 signs, 2–18, 23–7, 29, 31, 33–4, 36–7, 39, 44, 59, 69, 93, 94, 96, 108, 154 inert, 3, 10 similarities, 12, 51, 55, 78–84, 86–8, 108, 119, 158, 162, 165, 173 simples, 28, 34, 36, 39–41 Snowdon P., 121 n.14 social, 2, 33, 57, 68, 71, 76–7, 94, 97, 106–7, 134–6, 138, 141–6, 168, 172–3 solipsism, 41–2 space of causes, 162 space of reasons, 108, 111–13, 118, 148, 160, 162, 169

Index speaking with right, 101–2 Stroud B., 19 n.2 subject of experience, 13–14, 129, 133 subjectivity, 137–42, 146 Cartesian, 140–2, 174 simple, 139–40, 143 symbol (sign in use), 9, 16–18, 24, 26, 63 technique, 88, 153, 155, 170–1 transcendental argument, 18 transcendental idealism, 34–5, 41–2 truth, 3, 4, 6–7, 10, 29, 31, 36, 39–40, 45, 61, 72–3, 77, 140, 145, 168 truth conditions, 4, 7, 10 uncritical realism, 41 unitary account of content, 18, 49, 64, 114 of experience, 104 use, 3, 4, 6–11, 13, 15–18, 23–6, 29, 31, 36, 39, 41–2, 44–5, 48–56, 58–73, 76–84, 87–8, 93–101, 105–16, 118–19, 124–8, 133–5, 137–8, 145, 152–4, 158, 162, 166–8, 171, 174 as practice, 58, 63–4, 66, 72, 98

187 correct, 6, 12, 48–9, 52, 62–7, 79, 81–2, 84, 87–8, 112, 125, 137, 153, 156 ordinary, 53–4, 57, 59–63, 65–6, 79, 99 standards of, 9, 48–9, 52, 58, 61–3, 80, 82, 125 syntactic, 58–9, 60, 64, 68, 70–1 theory of meaning, 59 theory of names, 36 transparent ordinary, 58, 60–3, 65–6, 68–70, 98

valid inference, 74–5 validity, 74–7 Wiggins D., 91 n.28 will, 1–3, 7, 8, 12–14, 16–17, 22, 25–6, 30–1, 38–45, 48, 55–7, 64–9, 72–3, 75, 79, 81–2, 85–6, 96, 105, 108, 112, 115–19, 130, 133, 135–42, 144–7, 152–4, 160–5, 168–74 will, independent of, 13–16, 42, 66, 108, 115, 118–19, 133, 135–6, 138–42, 145–7, 152, 163, 167–9, 173–4 Williams M., 19 n.10, 90 n.23, 120 n.6, 120 n.7, 121 n.10 Wright C., 92 n.42, 123 n.27

Wittgenstein: Meaning and Judgement

Wittgenstein: Meaning and Judgement

Wittgenstein: Meaning and Judgement

Judgement

Judgement

Judgement

Judgement

Judgement

Judgement

Judgement

Judgement

Art As Language: Wittgenstein, Meaning, and Aesthetic Theory

Wittgenstein, Mind and Meaning: Towards a Social Conception of Mind

Judgement and Wrath

Death and Judgement

Sceptical Guide to Meaning and Rules: Defending Kripke’s Wittgenstein

Judgement and Wrath

Wittgenstein

Wittgenstein

Death and Judgement

Judgement and Wrath

Wittgenstein

Wittgenstein

Wittgenstein

Wittgenstein

Wittgenstein

Orbitsville Judgement

Judgement Day

Blind Judgement

III - Judgement

Wittgenstein, Empiricism, and Language

Wittgenstein: Meaning and Judgement