Truth and Speech Acts: Studies in the Philosophy of Language 041540651X, 9780415406512

Citation preview

Truth and Speech Acts

Whereas the relationship between truth and propositional content has already been intensively investigated, there are only very few studies devoted to the task of illuminating the relationship between truth and illocutionary acts. This collection aims to fill that gap. It focuses on such themes as the general significance of the concept of truth for the analysis of speech acts, the connections between the concept of truth and the concept of assertion, and the normative role that this concept seems to play in cognitive speech acts. Dirk Greimann is Professor of Philosophy at the Federal University of Santa Maria, Brazil. Geo Siegwart is Professor of Philosophy at the University of Greifswald, Germany.

Copyrighted

Material

Routledge Studies in Contemporary Philosophy

1 Email and Ethics Style and ethical relations in computer-mediated communication Emma Rooksby 2 Internalism and Epistemology The architecture of reason Timothy McGrew and Lydia McGrew 3 Einstein, Relativity and Absolute Simultaneity Edited by William Lane Craig and Quentin Smith 4 Epistemology Modalized Kelly Becker 5 Truth and Speech Acts Studies in the philosophy of language Edited by Dirk Greimann and Geo Siegwart 6 Fiction, Narrative and Knowledge A sense of the world Edited by John Gibson, Wolfgang Huemer and Luca Pocci 7 A Pragmatic Philosophy of Democracy Communities of inquiry Robert B. Talisse

Copyrighted

Material

Truth and Speech Acts Studies in the philosophy of language

Edited by Dirk Greimann and Geo Siegwart

Routledge

Taylor & Francis Group NEW YORK AND LONDON

Copyrighted

Material

First published 2007 by Routledge 270 Madison Avenue, New York. N Y 10016 Simultaneously published in the U K by Routledge 2 Park Square, Milton Park, Abingdon. Oxon OX 14 4RN Routledge is an imprint of the Taylor & Francis Group, an informa

business

(' 2007 Dirk Greimann and Geo Siegwart. selection and editorial matter; the contributors, their chapters Typeset in Times New Roman by Taylor & Francis Books Printed and bound in Great Britain by M P G Books Ltd, Bodmin All rights reserved. No part of this book may be reprinted or reproduced or utilized in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. Library of Congress Cataloging in Publication Data A catalogue record for the book has been request British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library ISBN 978-0415-40651-2

Copyrighted

Material

Contents

List of

contributors

Introduction DIRK G R E I M A N N A N D G E O SIEGWART

PART I

The illocutionary significance of the concept of truth 1

Illocutionary acts and truth WILLIAM P. ALSTON

2

Illocutionary acts and the concept of truth JOHN R. SEARLE

3

Alethic acts and alethiological reflection. An outline of a constructive philosophy of truth G E O SIEGWART

PART II

Truth and assertion 4

The use of force against deflationism: assertion and truth D O R I T BAR-ON A N D KEITH SIMMONS

5

The assertoric use of the concept of truth DIRK G R E I M A N N

6

Assertion as a practice GARY K E M P

7

Truth, assertion and the sentence WOLFRAM HINZEN

Copyrighted

Material

vi

Contents

8 Declarative thought, deflationism and metarepresentation

157

J O H N COLLINS

9

The concept of truth and multiple facets of the speech-act equivalence thesis concerning "true"

178

BO MOU

PART III The normativity of truth

199

10 Inflating truth: Wright's argument from normativity to propertyhood

201

ADAM KOVACH

11 Norms of assertion

226

G R A M A M OPPY

12 A linguistic reason for truthfulness

250

M I C H A E L RESCORLA

13 Truth as a normative modality of cognitive acts

280

GILA SHER A N D CORY D. W R I G H T

14 Why truth is not an epistemic concept

307

RICHARD SCHANTZ

15 Truth as perfect belief. On the Peircean conception of truth

321

U L R I C H METSCHL

PART IV Truth and propositional meaning

347

16 Meaning and truth-conditions

349

R I C H A R D HECK. JR

17 Meaning, truth and normativity

377

M I C H A E L WILLIAMS

Index of names

396

Copyrighted

Material

Contributors

William P. Alston is Professor Emeritus of Philosophy at Syracuse University and a past President of the Central Division of the American Philosophical Association. His most recent books include A Realist Conception of Truth, Illocutionary Acts and Sentence Meaning, and Beyond "Justification": Dimensions of Epistemic Evaluation. Dorit Bar-On is Professor of Philosophy at the University of North Carolina and teaches and does research in philosophy of mind, philosophy of language, and epistemology. She has recently published a book on selfknowledge entitled Speaking My Mind: Expression and Self-Knowledge. John Collins is a philosophy lecturer at the University of East Anglia. He researches on the syntax-semantics interface and the philosophical reception of the work of Chomsky. He has a book forthcoming on Chomsky. Dirk Greimann is Professor of Philosophy at the Federal University of Santa Maria, Brazil. His areas of specialization are philosophy of language, ontology and the history of Analytic Philosophy. He is author of Freges Konzeption der Wahrheit. Richard Heck, Jr. is Professor of Philosophy at Brown University. His main interests are in the areas of philosophy of language, logic, and mathematics, and he has worked extensively on the philosophy of Gottlob Frege. He is presently working on books on Frege's philosophy of arithmetic and on the nature of semantic competence. His recent papers include "Ramified Frege Arithmetic," "Truth and Disquotation" and " D o Demonstratives Have Senses?" Wolfram Hinzen is Professor of Philosophy at the University of Durham. He specializes in the philosophy of language and mind. His publications include the book Mind Design and Minimal Syntax. Gary Kemp is a Senior Lecturer in the Department of Philosophy at the University of Glasgow. He has published on truth and meaning and related topics, on aesthetics, and on Frege, Russell, Quine and Davidson.

Copyrighted

Material

viii

Contributors

Adam Kovach is Assistant Professor of Philosophy at Marymount University. He is interested in metaphysics and the philosophy of mind, and has published articles on theories of truth, on the emotions and the ethics of belief. Ulrich Metschl is Privatdozent at the University of Munich and is currently Visiting Professor at the University of Innsbruck. He specializes in the philosophy of logic and the theory of rational choice. He published Über einige verwandle Möglichkeiten der Behandhing des Wahrheitsbegriffs and a number of papers on the philosophy of logic, epistemology, and social choice theory. Bo Mou is Associate Professor of Philosophy at San Jose State University in California, USA. He has published in Analytic Philosophy, Chinese philosophy and comparative philosophy, concerning philosophy of language, metaphysics, philosophical methodology and ethics. Graham Oppy is Professor of Philosophy at Monash University. His areas of specialization are philosophy of religion and metaphysics, and he also has longstanding interests in philosophy of science and philosophy of language. He has recently published two books: Philosophical Perspectives on Infinity and Arguing About Gods. Michael Rescorla is Assistant Professor of Philosophy at the University of California, Santa Barbara. He specializes in philosophy of language, mind, and logic. His current research focuses on the nature of assertion and on non-propositional varieties of representational content. Richard Schantz is Professor of Philosophy at the University of Siegen, Germany. He is working in the philosophy of language, metaphysics and epistemology. He is author of Der sinnliche Gehalt der Wahrnehmung and Wahrheit, Referenzund Realismus. John R. Searle is Mills Professor of the philosophy of mind and language at the University of California, Berkeley. His publications include the books Speech Acts, Expression and Meaning, and The Foundations of Illocutionary Logic (with Daniel Vanderveken). Gila Sher is a Professor of Philosophy at the University of California, San Diego. Her areas of specialization are epistemology and metaphysics, philosophy of logic, and philosophy of language. Her publications include: The Bounds of Logic. "In Search of a Substantive Theory of Truth," "Logical Consequence: An Epistemic Outlook," "The FormalStructural View of Logical Consequence," "The Logical Roots of Indeterminacy." and "Is There a Place for Philosophy in Q u i n e s Theory?" Geo Siegwart is Professor of Philosophy at the University of Greifswald, Germany. He specializes in logic, epistemology and philosophy of language and is author of Vorfragen zur Wahrheit.

Copyrighted

Material

Contributors

ix

Keith Simmons is Professor of Philosophy at U N C Chapel Hill. He has special interests in logic, philosophy of logic and philosophy of language. He is the a u t h o r of Universality and the Liar and the co-editor (with Simon Blackburn) of Truth. Michael Williams is Krieger-Eisenhower Professor of Philosophy at the Johns H o p k i n s University. He works in the philosophy of language and epistemology. His publications include the b o o k s Groundless Belief, Unnatural Doubts and Problems of Knowledge. Cory D. Wright is in the Philosophy D e p a r t m e n t at the University of California, San Diego. He mainly works in epistemology and philosophy of cognitive science, with particular emphasis on veridical representation, j u d g m e n t , grounding, and certainty. Recent chapters and articles are on truth and on mechanistic explanation in psychology.

Copyrighted

Material

Copyrighted

Material

Introduction Dirk Greimann and Geo Siegwart

A speech act normally has a basic structure that may be represented as "F(p)", where " F " stands for the illocutionary force (or type) and " p " for the propositional content of the act. The propositional content of the assertion "Snow is white", for instance, is the component it shares with the question "Is snow white?", and its illocutionary force is the component it shares with other assertions like "Grass is green" and "Sea-water is salty." What is the relationship between the concept of truth on the one hand and propositional content and illocutionary force on the other? With regard to propositional content, this question has already been intensively studied. Its answer is controversial, of course. One party claims that truth is the master concept of semantics in terms of which the propositional meaning of sentences must be explained. According to this approach, which is usually called "truth-conditional semantics", the propositional content of, say, the sentence "Snow is white" must be explained by means of sentences like (T) "Snow is white" is true if and only if snow is white. The other party denies that the concept of truth has any essential connections with the concept of propositional content. The basis of this "deflationary approach", as it is now often called, is the linguistic observation first made by Frege and usually ascribed to Ramsey that the sentence '"Snow is white' is true" says nothing more than the simple sentence "Snow is white." Given this analysis, (T) does not contain any information about the propositional content of the sentence "Snow is white", because it says nothing more than the mere tautology "Snow is white if and only if snow is white." Whereas the relationship between truth and propositional content has been a main focus of current discussions about truth, there have been only very few studies devoted to the task of illuminating the relationship between truth and illocutionary force. 1 This comes as a surprise, considering that traditional theories of judgment attribute to the concept of truth an important role in speech acts like assertion. Thus, in Kant's theory, truth is considered to be a "modality" whose application is an essential constituent of "assertoric judgment." In the Critique of Pure Reason, he writes:

Copyrighted

Material

2

Dirk Greimann and Geo

Siegwart

The modality of j u d g m e n t s is a quite peculiar function. Its distinguishing characteristic is that it contributes nothing to the content of the j u d g m e n t ( . . . ) , but concerns only the value of the copula in relation to t h o u g h t in general. Problematic j u d g m e n t s are those in which affirmation or negation is taken as merely possible (optional). In assertoric j u d g m e n t s affirmation or negation is viewed as real (true), and in apodeictic j u d g m e n t s as necessary. 2 Anachronistically speaking, K a n t ' s point is that the function of the concepts of possibility, truth and necessity is not to make a contribution to the propositional content of a j u d g m e n t , but to fix its mode, that is, their function is to determine whether the propositional content is judged as possible or as true or as necessary. Thus, the propositional content of the problematic j u d g m e n t "Snow is presumably white" coincides with the propositional content of the assertoric j u d g m e n t "Snow is really white", and these j u d g m e n t s differ only with regard to their mode: while in the assertoric j u d g m e n t the propositional content is judged as true, in the problematic j u d g m e n t it is j u d g e d as possible. A similar account of the relationship between t r u t h and assertion (or "assertoric j u d g m e n t " ) is to be f o u n d in Frege's theory of j u d g m e n t . He also holds that the application of the concept of truth does not m a k e an essential contribution to the propositional content of judgments. A n d , like K a n t , he does not infer f r o m this that truth is a r e d u n d a n t concept; on the contrary, he also considers truth to be a key concept of logic whose application is an essential c o m p o n e n t of cognitive a n d illocutionary acts such as judgment and assertion. 3 Frege's conception of the expressive function of the concept of truth is based on the following analysis of the linguistic devices of expressing truth: [I]t is by using the f o r m of the assertoric sentence that we express truth [womit wir die Wahrheit aussagen], and to do this we d o not need the word "true." Indeed, we can say that even where we use the locution "it is true t h a t . . . " the essential thing is really the form of the assertoric sentence. 4 When, for instance, we want to express that the proposition that snow is white is true, we d o not need to apply the word "true"; it is already sufficient to assert the sentence "Snow is white." For, in virtue of its assertoric form, this sentence expresses that snow's being white is true. The clause "that snow is white" expresses that snow is white, too, but it does not express that this is also true. This shows that, in natural language, assertoric sentences have a twofold linguistic function: they serve to express a propositional content and, at the same time, to express the truth of this content. As Frege complains, this duality is hidden by the syntactic structure of assertoric sentences: In an assertoric sentence two different kinds of things are usually intimately bound up with one another: the thought expressed a n d the

Copyrighted

Material

Introduction

3

assertion of its truth. A n d this is why these are often not clearly distinguished. However, one can express a thought without at the same time putting it forward as true. 5 O n Frege's and Kant's approach, the concept of t r u t h plays an i m p o r t a n t expressive role at the level of illocutionary force, and it is r e d u n d a n t only with regard to the level of propositional content. This view implies that the current discussion a b o u t the expressive role of the concept of truth is unduly restrictive, because the illocutionary level is ignored. In order to determine the expressive role of the concept of truth, we must not only analyze the content a n d the function of the word "true", but also the structure of speech acts like assertion. With regard to the level of illocutionary force, the concept of truth seems to play also an i m p o r t a n t explanatory role. In order to explain the notion of assertion, for instance, it seems to be necessary that we make use of the concept of truth, because, at least prima facie, the conditions for assertion are to be explained as follows: The speaker x asserts that p iff (i) x expresses the proposition that p; (ii) x presents the proposition that p as true; (iii) x tries to convince the hearer of the truth of the proposition that p; (iv) x expresses the conviction that the proposition that p is true; (v) x commits himself to the truth of the proposition that p; (vi) x implicitly claims to have convincing arguments for the truth of the proposition that p. In order to explain the notions of question and order adequately, it seems also to be necessary to m a k e use of the concept of truth. For, to pose the question whether p is to request the hearer to determine the truth-value of the proposition that p, and to order that p is to request the hearer to make the proposition that p true. Finally, it appears that the concept of truth plays also an i m p o r t a n t normative role at the illocutionary level insofar as truth is a n o r m of assertion: when we are m a k i n g an assertion, we normally aim to say something true. 6 If this view is correct, then there are i m p o r t a n t connections between the success-conditions of assertions and kindred speech acts on the one h a n d and the concept of truth (and related concepts such as proof) on the other. The intention of the present anthology is to stimulate a m o r e intensive and systematic discussion about the illocutionary side of the concept of truth. To realize the present volume, we invited several a u t h o r s specialized in the philosophy of language a n d the theory of truth to write a paper tackling one or more of the following questions: •

Is Frege's analysis of assertion correct according to which "to assert that p " means " t o present the thought that p as a truth (or as a fact)"?

Copyrighted

Material

4

Dirk Greimann and Geo

•

What is the relation between the concept of truth and the success conditions of assertions and kindred speech acts? Is truth a n o r m of assertion and, if so, which one? What are the connections between the notion of truth and the notion of "rational assertibility"? What are the linguistic devices of expressing the truth of a proposition? Is t r u t h expressed in natural language by the " f o r m of the assertoric sentence", as Frege thought? What kinds of speech acts are performed by uttering the word "true"? Is this word (sometimes) used to express agreement, as the early Strawson maintained? W h a t is the relation between predication and truth? Is, for example, the predication in "Is snow white?" identical with the predication in "Snow is white", although only the latter sentence claims that the predicate is true of the object? Are there other speech acts beside assertion in which the concept of truth plays an essential role? If so, how are they interrelated? Are utterances of assertoric sentences that lack assertoric force proper truth-bearers? Can, for instance, the utterances made by an actor on stage be said to be true or false?

• •

•

•

• •

Siegwart

T h e papers collected here roughly divide into four groups. The first deals primarily with the general significance of the concept of truth for the analysis of speech acts. To this g r o u p belong the contributions of William P. Alston, John Searle and G e o Siegwart. The second g r o u p focuses on the connections between truth and assertion. It consists of the contributions of Dorit Bar-On a n d Keith Simmons, Dirk G r e i m a n n , G a r y Kemp. Wolfram Hinzen, John Collins, and Bo M o u . The papers of the third group arc primarily concerned with the normative role of truth. This g r o u p includes the papers by A d a m Kovach, G r a h a m Oppy, Michael Rescorla, Gila Sher a n d C o r y D. Wright, Richard Schantz, and Ulrich Metschl. The f o u r t h g r o u p finally deals with the connection between truth and propositional content. It consists of the papers contributed by Richard Heck and Michael Williams. Since the questions listed above are systematically interrelated in various ways, almost all papers naturally deal with a multiplicity of different questions. T h e division into four groups is therefore to be taken cum grano salis. Several people have contributed to the realization of this book. All of them we t h a n k warmly. Special t h a n k s are due to Sebastian Paasch for preparing the typescript and to Boguslawa G r e i m a n n for preparing the notes on contributors. Last but not least, we would also like to t h a n k the editorial staff at Routledge for their expert assistance. Dirk G r e i m a n n G e o Siegwart May 2006

Copyrighted

Material

Introduction

5

Notes 1 For an overview over the current debate about truth, see the collection edited by R. Schantz (2002) What is Truth? (Current Issues in Theoretical Philosophy 1), Berlin: Walter de Gruyter. 2 I. Kant (1781) Critique of Pure Reason, 2nd edn 1787, translated by Norman Kemp Smith, electronic edition available at http://www.arts.cuhk.edu.hk/Philosophy/Kant/cpr, B 99 ff. 3 G. Frege (1997) The Frege-Reader, ed. M. Beaney, Oxford: Blackwell, pp. 323, 325. 4 G. Frege (1983) Nachgelassene Schriften und Wissenschaftlicher Briefwechsel, Vol. 1, ed. H. Hermes, F. Kambartel and F. Kaulbach, 2nd edn, Hamburg: Meiner, p. 140 (our translation). 5 G. Frege (1997) The Frege-Reader, ed. M. Beaney, Oxford: Blackwell, p. 239. 6 This role has first been stressed by Michael Dummett in (1978) Truth and other Enigmas, Cambridge: Harvard University Press, pp. 2 ff.

Copyrighted

Material

Copyrighted

Material

Part I

The illocutionary significance of the concept of truth

Copyrighted

Material

Copyrighted

Material

1

Illocutionary acts and truth William P. Alston

This paper will be devoted to considering certain issues about the relation, or lack thereof, of truth-values and illocutionary acts. At the outset I need to stipulate certain parameters of the discussion. First I am concerned specifically with illocutionary acts (IAs) rather than more generally with the genus of speech acts of which illocutionary acts constitute one species, to be distinguished from, for example, locutionary acts (uttering sentences or surrogates thereof) and perlocutionary acts (producing effects on audiences by utterances). The category of illocu-tionary acts is not as easy or unproblematical to demarcate. My way of doing this is somewhat different from those of J. L. Austin and of John Searle, though closer to the latter. For a quick indication I could say that to perform an IA is to issue an utterance with a certain "content." But what do I mean by "content" here? The shortest useful answer involves following Searle in taking the content to consist of one or more propositions and an "illocutionary force." Thus different IA-types can differ in propositional content, illocutionary force, or both. Telling X that the door is open, requesting X to open the door, and asking X whether the door is open, all express the same proposition but with different illocutionary forces. Whereas the first of the above examples shares illocutionary force with telling X that the mail has come, and telling X that dinner is ready, they differ in propositional content. This way of identifying something as an illocutionary act could be put by saying that it is a speech act a fully explicit sentential vehicle for which would involve one or more propositional clauses and a term for, or other linguistic indication of, illocutionary force. 1 This should give an adequate preliminary indication of how I am thinking of illocutionary acts except for a type-token complexity. The above explanation was in terms of maximally specific illocutionary act types, but what about tokens of such types? In particular, suppose that S tells X the mail has come by uttering the sentence "The mail has come" and thereby gets X to realize that the mail has come. S has, in one breath, performed acts of locutionary, illocutionary, and perlocutionary types. Are we to say, as my last sentence would have it, that S has performed three different act tokens, or should we say that there is only one act token that belongs to

Copyrighted

Material

10

William P. Alston

three different types? This is a special case of the much discussed problem of act identity, and, more generally, of event identity. For my purposes here I don't care which position is taken. But since my concern will be primarily with illocutionary act types, it will be most convenient for me to speak of the locutionary. illocutionary. and perlocutionary acts performed simultaneously by the same speaker as different acts. Let it be so ruled. T h e other preliminary parameter-setting has to do with truth. I will be employing what I call in Alston (1996) a "realist conception" of truth. Before explaining what that a m o u n t s to let me say that I take the primary bearers of truth-value to be propositions. Beliefs and assertions that are true are so by virtue of their propositional content, and if we want to think of (some) sentences as true, as I don't, it is because of the propositions they express. If any reader is worried about the ontology of propositions, let me say that all it takes to deal with propositions as I construe them is to be able to handle that-clauses. If you can understand, for example, "that it is snowing," you have all you need to understand talk about propositions. To get back to the realist conception of truth, the basic idea is that the concept of truth is uniquely identified by what I call the " t r u t h schema": The proposition that p is true (It is true that p) i f f p. If you recognize the truth of any substitution instance of this schema, that suffices for you to have grasped the concept of propositional truth. This does not provide for even a contextual definition of " t r u e " unless we allow for substitutional quantification, but that doesn't prevent the above explanation from providing a unique identification of the concept of truth. There are many other ways of formulating the same basic idea. For example: A belief is true i f f what is believed is the ease. Again: An assertion is true i f f what the assertion is about is as it is being asserted to be.2 This is a minimalist account of the concept of truth, for it makes no attempt to go into details as to what it is by virtue of which the t r u t h - m a k e r for a given proposition does that job. T h u s it makes no pretense at being a theory of the property of truth. But it is distinct from a "deflationary" account of truth that attempts to dispose of the claim that there is a property of truth whose exemplification distinguishes true from false propositions. An extreme version of this is the "disappearance" view of truth according to which asserting that it is true that grass is green a m o u n t s to no more than asserting that grass is green. My account of the concept is not that minimalist. T h o u g h it is not an account of the nature of the property of truth, unlike a correspondence theory with which it is naturally associated, it is open to the possibility of such a theory's being developed, though it doesn't imply that this can be successfully done. 3 It is not crucial for my purposes in this essay to explain why I call this a "realist conception" of truth. But in case anyone is wondering, the point is that this view makes the truth of a proposition dependent on what the proposition is about, which is usually something that is external to the proposition and its role in thought and discourse, unlike epistemic conceptions

Copyrighted

Material

Illocutionaryacts and truth

11

of truth according to which for a proposition to be true is for it to have some epistemic status such as being what P u t n a m calls "ideally justified." But my account of the concept of truth does not carry with it a commitment to a metaphysical realism, the view that what (many) propositions arc about are items that are what they are independently of our thought and discourse about them. That is a further issue. With this background I can begin to consider various alleged ways in which IAs and truth are related. The most obvious suggestions have to do with assertoric illocutionary acts. That is a large family of acts rather than a single homogeneous type, for it embraces a large variety of distinct illocu-tionary forces, such as replying that p, admitting that p, insisting that p, remarking that p, objecting that p, and so on. What these all have in c o m m o n is that they all involve asserting that p. They could all be construed as "merely asserting" that p + some additional feature that distinguishes them from each other. I think of (merely) asserting (stating) that p as a maximally simple assertoric IA, one that exhibits the c o m m o n feature of assertoric IAs in its purity without any additional feature. Alternatively we could think of assertion as a genus of which the various "special" modes of asserting are species. Dealing with the issues about assertoric IAs I will be concerned with here does not require separate treatment of the different special assertoric IAs. It is their c o m m o n or generic nature that is relevant to those issues. Hence in discussing them I will restrict myself to speaking in terms of assertions and thereby implying that the points made apply to all the "special" assertoric .IAs as well. Now I can proceed to a fairly popular view of what distinguishes assertoric IA-types as such. It was classically formulated by Frege as follows: To assert a proposition is to present it as true.4 But despite its appeal, it is fairly easy to see that this view will not hold up under scrutiny. For one thing, if this position means what it seems to mean, it represents all assertions as having the same main predicate. For if to present a proposition as true is to assert that proposition to be true, then every assertion is a predication of truth to some proposition. But this is a Procrustean bed indeed. We certainly mean to be predicating different properties of different things (usually not propositions), depending on the constitution of the propositional content. And this view would, so to say, prevent our doing what we mean to be doing, and seem to be doing, in making assertions. It takes the propositional content of an assertion and makes that what the assertion is about, thereby foisting a role on it that distorts its proper function in the proceedings. I assert that M t . Shasta is impressive. Unless I am completely out of touch with what I a m doing, then what my assertion is about is Mt. Shasta, and what I a m asserting of it is that it is impressive. I could also make an assertion about the propositional content of that first assertion, and assert that it is true. But that would be a different assertion, one of a sort that most speakers seldom make. And so this suggestion is on the wrong track altogether.

Copyrighted

Material

12

William P. Alston

But perhaps this is the wrong way to construe Frege and others who take his kind of position. 5 Perhaps he didn't mean to imply that all assertions are about propositions, predicating truth of them, but instead to assert some different connection between asserting that p and presenting p as a truth. Here are some possibilities. (1) To make an assertion is to claim that the propositional content of one's utterance is true. (2) In making an assertion one represents oneself as presenting a true proposition. (3) When one makes an assertion, one would be prepared to claim that one was presenting a true proposition if the question were raised as to what one takes oneself to be doing. (4) If one asserts that p but does not suppose that it is true that p, one is guilty of deceit of some sort misleading the hearer, misrepresenting one's beliefs, or the like. N o n e of these connections involve altering propositional contents of assertions to make them predicating truth of a proposition, the defect that was seen to discredit the earlier suggestion. But each of them claims a significant relationship between assertion and truth. I have arranged these claims in an order of decreasing similarity to the discredited thesis. (1) is just like that thesis except for avoiding the claim that the propositional content of every assertion is of the form " T h e proposition that p is true." (2) differs from (1) only in taking the same claim to be implicit rather t h a n explicit. (3) puts the claim to the truth of the proposition f u r t h e r away by placing it in how one would answer a question or challenge. Whereas (4) is still f u r t h e r away from an explicit claim to truth by only specifying a result of not supposing the claim to be true. The same order exhibits a decreasing plausibility of construing the thesis as a statement of what it is to make an assertion. (1). I would say. has the same title to that as the original thesis. (2). t h o u g h less explicit, could with some plausibility claim to be spelling out what it is to make an assertion. It would seem to be less plausible to give (3) that role, since making an assertion seems to be something more overt, more out in the open, than merely a disposition to respond to questions or challenges in a certain way. A n d (4) would seem to be even f u r t h e r away from an account of the nature of assertion for the same kind of reason. Although those are my intuitions, I don't want to hang a great deal on their correctness. A n d since I will go on to reject all four of these suggestions, both as accounts of what making an assertion is, and even as something that is universally true of assertion making, I take the question of which of them is a correct account of what it is to make an assertion as a question that does not arise. So why d o I think that none of the four theses holds universally, much less necessarily, of all assertions. The reason is that they all require the asserter to have c o m m a n d of the concept of propositional truth, and that is an untenable requirement to make of all asserters. The crucial point here is that taking a proposition to have a certain truth-value is a more sophisticated cognitive p e r f o r m a n c e than m a k i n g the most rudimentary or primitive sort of assertion. In arguing for this I will leave aside the question of what it takes to have a working mastery of a concept of a proposition, since earlier

Copyrighted

Material

Illocutionary

acts and truth

13

I maintained that being able to deal with that-clauses is sufficient for that. A n d though it could be contended that even that is something more sophisticated t h a n what is required for the most primitive sorts of assertions, I will not press that point here. Instead I will concentrate on what it is to have and use the concept of truth. In the previous section I maintained that one has the ordinary concept of propositional truth provided one recognizes that for a proposition, p, to be true it is both necessary and sufficient that it is the case that p. But to be in a position to recognize that one must have at one's disposal the concepts of necessary a n d sufficient conditions, as well as the concept of something's being the case. I a m not suggesting that the cognitive subject must be capable of spelling all this out explicitly. Again, a working mastery in practice of the relevant similarities and distinctions is all that is required. But if either (1) or (2) is true of a cognitive subject, S, S must be capable of claiming, explicitly or implicitly, that the propositional content of S's utterance is true. A n d this requires S to predicate the property of truth to a proposition, and in addition to be able to recognize that proposition as the propositional content of a particular utterance. T h u s it requires mastery of a fairly complex conceptual scheme. A n d think of the most rudimentary example you can of an utterance that would count as an assertion and ask yourself whether one could not be credited with making an assertion unless one is able to wield that complex a conceptual scheme. Take as your rudimentary assertion something like "It h u r t s " or " I ' m tired" or "It's cold." D o we really want to say that S could not count as making assertions in these cases if S were incapable of predicating truth of the propositional content of those utterances? Suppose that S could pass the following tests of making a genuine assertion when saying " I ' m tired," an assertion of the type normally and conventionally m a d e by uttering those words, (a) S, on being questioned as to what m a d e him/her so tired, would say that he had been doing a lot of running outside, (b) W h e n asked whether s/he would like to go for a walk, replies "Of course not; I ' m tired." (c) When the m o t h e r says "You always say you're tired when you want to get out of picking u p your toys," S replies "But I often pick up my toys." This list could be extended indefinitely. Isn't it clear that a toddler could carry on a conversation like this without being able to claim, as implicitly as you like provided it is a genuine claim, that the propositional content of the original utterance is true, or even that any proposition is true, lacking at that point the conceptual development necessary for m a k i n g such a claim? T h e same decision is to be m a d e with respect to (3). For if one lacks the conceptual equipment needed to claim that a certain proposition is true, one obviously c a n n o t be prepared to claim that one is presenting a true proposition in response to certain questions or challenges. But what about (4)? This differs from the other theses on the list by not representing the asserter as making, or as capable of making, claims of the sort featured in (l)-(3). Instead it rules that the asserter is guilty of deceit if he is not prepared to make such a claim. It doesn't require the asserter to recognize that or

Copyrighted

Material

14

William P. Alston

engage in any other cognitive p e r f o r m a n c e that would require the conceptual repertoire I have been suggesting that need not be possessed in order to make the most rudimentary assertions. That is certainly correct. But by the same token (4) as stated is not a good candidate for an account of what it is to make an assertion. If (4) were strengthened into (4A) - If one apparently asserts that p but does not suppose that it is true that p, one recognizes that one is guilty of deceit of some sort - that would be a different matter. T h a t could be a claim as to what it is to make an assertion. But, again by the same token, it would suffer discredit in the same way as (1)(3), by requiring conceptual powers that are not always possessed at the level at which a subject is capable of the most rudimentary assertions. For to recognize that one is guilty of deceit in an utterance, one must have a conception of what it would be like to avoid that kind of deceit. And this in turn would require that one is capable of recognizing that the propositional content of (the proposition presented by) one's utterance is true. A n d so we would again be impaled on the point that genuine assertions can be made at a stage of development at which the requisite conceptual equipment has not been acquired. T h u s the attempt to develop an account of what it is to m a k e an assertion that features the concept of propositional truth, an account that avoids m a k i n g every propositional content of the form "Proposition p is true," does not pan out. This does not imply that there are no significant connections between assertion and truth, t h o u g h they are not of the maximally intimate sort exemplified by an account of what it is to make an assertion in terms of t r u t h . Most speakers of natural languages are masters of the cluster of concepts that are needed to attribute truth-values to propositions, including those that furnish propositional content to assertions. It would be unrealistic to suppose that such asserters always either consciously take the propositions they assert to be true, or consciously recognize that they are at fault for not doing so when they assert the proposition, or even that this is a frequent occurrence. Nevertheless something along the line of (3) and (4A) is frequently true of them, and this constitutes a significant connection between assertion and truth-value, t h o u g h for the reasons given above, we can't validly suppose that it is always true of ...asserters, much less that it is necessarily true of them so as to constitute what making an assertion is. T h u s far the discussion has been restricted to suggestions as to how assertive IAs are related to truth. M u c h of the rest of this essay will be an exploration of possibilities for i m p o r t a n t connections of IAs generally with truth. There have been a n u m b e r of attempts to generalize (supposed) connections of assertive IAs with truth to IAs of other sorts. T h e most c o m m o n move of this kind is to look for items that are related to other categories of IAs in something like the way truth-value is (actually or supposedly) related to assertive IAs. A minimal suggestion along this line would be that all

Copyrighted

Material

Illocutionary

acts and truth

15

(almost all?) IAs have propositional content. This is something that is generally accepted by everyone who is prepared to speak in terms of propositions at all. T h e only plausible exceptions are IAs (if there are such) performed by interjections: " D a m n ! , " "Wow!," "Bollocks!," and others not suitable for mention in polite company. It is controversial both whether such utterances count as IAs, and, if they do, whether they have propositional content. In any event, apart f r o m such possible counterexamples as these, it is uncontroversial that all IAs have propositional content. A n d the propositions that furnish this content for non-assertive IAs have truth-value just as much as those that play this role for assertive IAs. This is a genuine relation of IAs generally with truth-value, albeit a relatively modest and unexciting one. However, there are other less obvious suggestions for a parallel between the way an assertion is true or false and alternatives that play an analogous role for IAs of other sorts. In Searle (1983) and Vanderveken (1990) we find a notion of "conditions of satisfaction" for IAs. Without attempting to go into their rationale for picking these out in one way rather t h a n others, Vanderveken, to follow his version, suggests that an assertion is satisfied i f f its propositional content is true, a "commissive" IA, like a promise is satisfied i f f it commits the speaker to a f u t u r e course of action, and "directive" IA, like a request or an order, is satisfied iff he addressee does what s/he is directed to do, and so on. Clearly what we have here is not a way in which all IAs are related to the truth-value of their propositional content, but ways in which non-assertive IAs have the same kind of relation to other conditions that assertive IAs have to the truth of their propositional content. We find a different development of this sort in, for example. Stenius (1967) and Pollock (1982), in which the notion of rules for linguistic communication are featured. T h e basic rule for assertions requires of the speaker that the proposition asserted be true, whereas the basic rule for orders requires of the addressee that s/he does what s/he was ordered to do. Again these views d o not represent truth-value as playing the same role for nonassertoric IAs that they play for assertions. Rather they find something to which a non-assertoric IA is related to something like the same way in which truth-value is related to assertions. A n d so we look in vain in these positions for some way in which truth-value is related to all, or almost all, IAs. A more determined effort to treat all IAs as related to truth-value in the same way is found in Lewis (1972), reprinted in Lewis (1983). The view in question is presented by Lewis not in terms of the relation of IAs of various sorts to truth-value, but in terms of the relation of sentences of various sorts to truth-value. M o r e specifically it is an attempt to treat all sentences, not just declarative ones, as susceptible to truth or falsity. But it is clear that this position could be translated into the way I have been posing the problem. For if non-assertive sentences like "Please pass the salt" could be regarded as true or false, then by the same token so could the IA of requesting someone to pass the salt.

Copyrighted

Material

16

William P. Alston

Lewis proposes to pull a rabbit out of this hat by the simple device of construing all sentences as declarative, or at least as paraphrases of, or surface structure t r a n s f o r m a t i o n s of, declarative sentences in the deep structure of the language. This is d o n e by construing each apparent nonassertive sentence as a p a r a p h r a s e of a corresponding "explicit performa-tive." T h u s "Please pass the salt" is construed as a p a r a p h r a s e of "I request you to pass the salt," and "Is your daughter at home?" a p a r a p h r a s e of "I ask you whether your d a u g h t e r is at home." And these explicit performatives are treated, contrary to J. L. Austin and most other IA theorists including myself, as assertions. So construed, when I say "I request you to pass the salt" or its surface level derivative, what I say is true i f f I do request you to pass the salt. A n d when I say "I ask you whether your daughter is at h o m e " or its surface level derivative, what I say is true i f f I d o ask you that question. Of course there are plenty of explicit performatives at the surface level already, such as "I promise to give you a ride to the meeting" and "I declare the meeting a d j o u r n e d . " And these will be treated as assertions as well. But unfortunately for this project of pan-assertivism, this won't work. Even if we bypass objections to treating explicit performatives as assertions, and there are serious objections to it. there are two fatal defects. (1) N o t all non-assertive sentences have corresponding explicit performatives. Consider "That's disgusting!" Saying "I express disgust at t h a t " is not a way of expressing disgust at that, m u c h less a way of asserting that I am expressing disgust at it. The scheme does not cover the whole territory. (2) A more crucial objection is this. Lewis acknowledges (how could he not?) that "I order you to close the d o o r " is usable to order someone to close the door, as well as to assert that I am ordering you to close the door. But does he or does he not hold that the sentence is fitted by its meaning to d o the former as well as the latter? If he does not hold that it is fitted by its meaning to issue the order but that it is an extra-semantic fact that it is so usable (and this is his explicit view), then he is flying in the face of facts. Surely it is part of grasping the semantics of the sentence to realize that "I order you to close the d o o r " is usable to order someone to close the door. But then if he comes to terms with the obvious facts, from a semantic point of view we have two different sentences with the same phonology and grammar, one assertive and one non-assertive. A n d so he has not after all succeeded in treating all sentences as declarative and all IAs as assertive, thereby treating them all as related to truth-values in the same way. U p to this point the search for i m p o r t a n t connections between IAs and truth has not turned up a great deal. The connections we have discovered have been minor and fairly obvious. The more exciting suggestions have not survived critical scrutiny. But we have by no means exhausted the territory. In particular we have not explored ways of analyzing IA concepts to see if some mention of truth-value plays an essential role there. That is what I will now proceed to do.

Copyrighted

Material

Illocutionary

acts and truth

17

I cannot within the limits of this essay d o a complete critical survey of all the attempts to analyze IA concepts. I shall restrict myself to a discussion of two prominent approaches - the Perlocutionary Intention analysis pioneered by H. P. Grice and developed somewhat differently by Stephen Schiffer, and the Normative Stance analysis hinted at by J. L. Austin and developed in Searle (1969), and by myself in a number of works culminating in Alston (2000). The basic idea of what I call the "Perlocutionary Intention" account of IAs is that to p e r f o r m an IA of a certain type is to perform a locutionary act 6 with the intention of producing a certain effect in one's audience. Grice gave the initial impetus to this way of construing IAs in his seminal essay " M e a n i n g " (1957), t h o u g h his specific target there was not IAs but "speaker meaning," a speaker's meaning something by what s/he said. It was left to Schiffer (1972) to use Grice's idea to provide a pattern of analysis for IA concepts. After an incredibly complicated version of a Gricean account of speaker meaning Schiffer approaches IAs with the suggestion that a speaker, U, performed an IA in uttering x i f f U meant something by uttering x. There are two broad categories he recognizes for what U might mean. (1) He meant that p, for some proposition p, which is analyzed, following Grice, in terms of uttering x with the intention of bringing H to believe that p, or (2) he meant that H (the hearer) was to perform an action D, which is analyzed as uttering x with the intention of getting H to p e r f o r m D. The first alternative is realized in assertive IAs, the second in "directive" IAs. Then each of these categories is divided into more specific types both by the p or D involved a n d / o r by the reason U meant H to have for believing that p or for performing D. For example, U reports that p in uttering x i f f U intended H to believe that p on the basis of H's belief that U believed that p on the basis of U's observations or investigations. A n d U ordered H to do D only if U intended part of H's reason for doing D is this: there is a certain legal or conventional or institutional relation, R, between U and H, such that by virtue of their being related in that way U has the right to expect H to d o an act of kind K if in circumstances of kind C U means that H is to do an act of kind K; that D is an act of kind K; and that the circumstances in which U uttered x were of kind C. 7 I have two main objections to this account. (1) Not all illocutionary act types can be construed as ways of meaning that p or meaning that H is to do D. Suppose that I promised to read your paper by Friday. Is this a case of meaning that p or meaning that H is to d o D. You might think that it a m o u n t s to meaning that I will read your paper by Friday. Perhaps it involves that, but that does not distinguish it from other IAs, such as expressing an intention to read your paper by Friday or predicting that I will read your paper by Friday. So that can't be the whole of what it is to make that promise. O r consider declaring the meeting a d j o u r n e d , which I d o in my function as chairman. Again you might think that this is meaning that the meeting is a d j o u r n e d . But again that does not distinguish it f r o m

Copyrighted

Material

18

William P. Alston

various other IAs such as complaining that the meeting is a d j o u r n e d , reporting that the meeting is a d j o u r n e d , or expressing appreciation that the meaning is a d j o u r n e d . T h u s the conceptual f r a m e w o r k Schiffer employs is much t o o narrow to cover the territory for which he claims to provide a pattern of analysis. (2) Even if we restrict ourselves to assertives and directives for which Schiffer's patterns are specifically designed, it is possible to p e r f o r m IAs of these sorts without the audience-directed intentions that Schiffer supposes to be necessary for that. Here are some examples, all of which Schiffer considers and tries to square with his theory. 8 1

Countersuggestible cases: " T h e Berkeley police thinking that the c a m p u s would be the best place to contain a riot, a n n o u n c e to the Berkeley radicals that under no circumstances are they, the radicals, to hold their rally on the campus. It is the intention of the police that this a n n o u n c e ment should cause the rally to be held on c a m p u s " (Schiffer 1972: 69). Here U m e a n s that H is not to do D with the intention of getting H to d o D. 2 Examination case Teacher: "Tell me, if you can, when the Battle of Hastings was fought." Student: "1066" (Schiffer 1972: 70). Obviously the student was not asserting that the Battle of Hastings was fought in 1066 in order to get the teacher to believe this. The student was, no d o u b t , assuming that the teacher already knew that the battle was fought in that year. 3 Confession case Police: "O.K. Capone, the jig's up. We know you stole the bubble gum, so you'd better confess." C a p o n e : "I confess, I stole the bubble g u m " (Schiffer 1972: 70). Again, C a p o n e could not be uttering x in order to get the police to believe that he stole the bubble gum, since they have m a d e it perfectly clear that they already know this. 4 Accusation case Mr. Smith: "I was working at the office all evening." Mrs. Smith: "You're lying" (Schiffer 1972: 71). Here, again, Mrs. Smith is already quite sure that Mr. Smith knows that he is lying and hence does n o t say what she says with the intention of getting him to believe this. We should note that Grice's later version of meaning that p and meaning that H is to d o D, in Grice (1969). is not subject to the above counter-examples. He substitutes for "getting H to believe that p," "getting H to believe that U believes that p," and for "getting H to d o D," "getting H to believe that U intends (wants) H to d o D." This will handle all the above examples nicely. But there are other cases that cause trouble for Grice as well as for Schiffer. Here are two. 5

Don't care cases: Consider situations in which U makes a report or a n n o u n c e m e n t because it is his j o b to do so but doesn't care in the least what H's reaction is, in particular doesn't care whether H either believes

Copyrighted

Material

Illocutionaryacts and truth

19

what he asserts or even whether H believes that U believes it. This description often applies to persons m a n n i n g information booths at railway stations and those announcing flights in an airport. They are performing various assertive IAs but without the perlocutionary intentions Grice and Schiffer take to be required for that. 6 Making conversation'. We are thrown together at a party. I say: "Rotten weather we've been having lately." You reply: "Still it could be worse." "Yes, there could have been freezing rain." A n d so on in the same vein. It is clear to both of us that we b o t h know the weather has been rotten lately, but not as bad as it could be, etc. Moreover, neither of us cares what beliefs his or her utterances leads the others to form. Again IAs of the assertive sort are being performed without the Grice-Schiffer requirements being satisfied. So the Perlocutionary Intention account of IA concepts simply doesn't d o the job. Shortly I will be presenting an account that, I claim, does the job. But first I want to ask whether the Perlocutionary Intention account would reveal an i m p o r t a n t connection of IAs and truth if it were an adequate theory of the former. This a m o u n t s to asking whether on the pattern of analysis of IA concepts proffered by the Perlocutionary Intention view the concept of truth puts in an appearance anywhere. It seems clear that the answer to that is in the negative. The analysis is purely in terms of intentions, beliefs, and other intentional psychological states. There is no requirement that the beliefs be true (or false). To be sure, intentions and beliefs, like all other intentional states have propositional content. They are types of propositional attitudes. A n d those propositions, of course, have truth-values. But I can't see that this a m o u n t s to a distinctive i m p o r t a n t connection of IAs and truth. And this is for two reasons. First, the same can be said of every other psychological state that has propositional content. Being afraid of burglars has a propositional content, something like burglars being dangerous. Being grateful to X involves the propositional content that X has done something favorable to one. Whenever there are propositions involved, truth-values are automatically in the picture. But this is not in any way distinctive of IAs. Second, and as a consequence of the first point, this doesn't tell us anything important about IAs. It doesn't provide a connection with truth that helps us to u n d e r s t a n d IAs better t h a n we would otherwise. At most it tells us that they involve intentional psychological states, and, as just pointed out, this is a connection with truth only in the most minimal way, in the psychological states involved having propositional contents that themselves are susceptible of truth-values. This is, at most, a relation to truth at a second remove and hardly the thing that would be featured in an account of what IAs are like. In a series of articles including Alston (1964), (1977), and (1994) I developed the idea that the crucial move in analyzing IA concepts is to recognize that in performing an IA U takes responsibility for the holding of certain

Copyrighted

Material

20

William P. Alston

conditions. A n d different IA-types are most fundamentally distinguished in terms of the conditions in question. This is an account that applies in the same way to all IAs, both assertive and non-assertive. Here are some examples of IA-types with the most obvious associated conditions for each, conditions for which U takes responsibility in performing an IA of that type. I

Asking H to lend me his power saw. A H has a power saw. B It is possible for H to lend me his power saw. C I have some interest in H's lending me his power saw. II Telling (ordering) my son to clean up his room. A M y son has a room. B It is possible for him to clean up his room. C I have the authority to lay on him an obligation to d o so. III T h a n k i n g H for recommending me for a job. A H has recommended me for a job. B I am grateful to H for having done so. IV A d j o u r n i n g the fall meeting of the University Senate. A T h e fall meeting of the University Senate is in session. B I have the authority to terminate that meeting. C Conditions are appropriate for the exercise of that authority. V Replying to V that my car is in the garage. A V has asked me where my car is. B M y car is in the garage. In Alston (2000) I explained the sense of "take responsibility f o r " that is employed in this account as follows. The idea is not that U took responsibility for state of affairs C in the sense that he was prepared to acknowledge that he brought C into existence. It is, rather, like the way in which when I become the head of a d e p a r t m e n t or agency, I take responsibility for the efficient and orderly conduct of its affairs, including the work done by my subordinates. I am responsible for all that work, not in the sense that I have done it all myself, but in the sense that I am rightly held to blame if the work is not done properly. I a m the one who must " r e s p o n d " to complaints about that work. (Alston 2000: 54) It is in a like sense that U takes responsibility for the holding of the conditions listed for each IA-type in performing an IA of that type. In uttering x U knowingly lays himself open to complaints, objections, correction, blame, or the like in case the relevant conditions are not satisfied. So (using " R " as an abbreviation of "take responsibility for"), we can explain " R " as follows.

Copyrighted

Material

Illocutionary acts and truth

21

D. 1 In uttering x, U R'd that p - In uttering x U knowingly took on a liability to blame (censure, reproach, being taken to task, being called to account) in case of not-p. (Alston 2000: 55) However this is not the most f u n d a m e n t a l way of bringing out the sense of " R . " T h e reason is that the liability one is said to take on in performing the IA in question is only prima facie. It can be overridden or cancelled if one has a sufficient excuse. For example, if one had "every reason" to think that a certain relevant condition is satisfied but it isn't, that would presumably excuse one f r o m being blamed for the failure of that condition. C a n we find something that is true of U no matter what when U asks H to lend him H's power saw? Well, a definiens that strictly holds no matter what else is perhaps a bit much to ask of any philosophically interesting definition. But we can at least come m u c h closer to that t h a n we do with D. 1. To d o so we must introduce rules for the permissibility of utterances. If we suppose that the conditions R ' d in p e r f o r m i n g IA of type T are involved by way of being required by a rule for the permissibility of uttering x in a t t e m p t i n g to perf o r m an IA of type T, that will give us the following definition of " R . " D.2 U R'd that p in uttering x - In uttering x, U subjects the utterance to a rule that, in application to this case, implies that it is permissible for U to utter x only if p. (Alston 2000: 60) I will call such rules Illocutionary Rules ("I-rules" for short). D.2 enables us to give a satisfactory treatment of the possible reactions to the utterance of one who R's that p. For example, if it is false that p, but U was justified in supposing p to be true, then U could not be held to blame for uttering x with the intention of performing an IA of the type in question. In C h a p t e r 3, Sections ii-v of Alston (2000) I mention various other ways of explaining R , but the above will suffice to give an idea of this that is sufficient for the purposes of this essay. The aspect of this I-rule account of IA-types that will be the basis for considering whether the account reveals some important connection with truth is simply this. To p e r f o r m an IA of a certain type is to subject one's utterance to one or m o r e I-rules, in the simplest case an I-rule for each of the "associated" conditions, exemplified by the above list of IA-types with associated conditions for each. Formulating I-rules is a complicated matter. The general f o r m of such a rule is that x may (permissibly) be uttered only if C. For one thing qualifications are needed if conforming to the rule is actually required for permissible utterance. As an example, building on a item f r o m the above list of IA act types with conditions R ' d for each, consider this simple formulation of an I-rule.

Copyrighted

Material

22

William

P. Alston

1.2 "Thanks for recommending me for the j o b " may be permissibly uttered by U only if the addressee, H. recommended U for the j o b in question. This is much too simple. T h a t sentence, like any sentence can be permissibly uttered without intending to be performing an IA of the type most commonly associated with the sentence. Indeed it can be permissibly uttered without intending to p e r f o r m any IA at all, instead of just practicing pronunciation or testing a microphone, etc. This is why D.2 includes the qualification "in application to this case." That is put there to signal that the rule is not designed to apply to all cases of the utterance of the sentence in question. Moreover there are sticky problems as to how to handle reference in the rule formulations. Unless we ensure identity of reference, in this case for "the j o b " in the IA specification and in the rule, then conformity to the rule would not be required for the p e r f o r m a n c e of the IA. If there are different j o b s referred to in those two contexts, then conformity to the rule would be neither necessary n o r sufficient for performing the IA. There are also problems a b o u t handling elliptical ways of performing an IA. These matters are gone into at length in C h a p t e r 7 of Alston (2000). Finally there are problems about the status of these rules, what sort of existence they have, what their existence depends on. and so on. 9 Fortunately it is not necessary for present purposes to work out the details of how to formulate I-rules to give them the role they have in this account of IA-types. To set out how the account reveals a connection with truth, it is sufficient to give the general idea of I-rules and their place in the account. I should also mention that there is no suggestion that subjecting one's utterance to the appropriate I-rules is sufficient for performing an IA of the type in question. In the last section of the essay we will see a crucial example of a n o t h e r kind of condition that is required for the sufficiency of the account for assertive IAs. The prospects look favorable for deriving an important connection between IAs and truth from this I-rule account of the latter. It is not as simple or sweeping a connection as one might hope for. The account does not, for example, imply that a necessary condition for U's ordering U's son to clean up his room is that the conditions listed above for that IA-type are true. The connection is more indirect t h a n that. It is that a necessary condition for performing that IA is that U subjects his utterance to a rule that requires, in this case, the truth of those conditions for a permissible utterance of the sentential vehicle employed. Moreover. as hinted above, this is not a complete account of all the necessary conditions for permissible utterance, even in this case. Nevertheless, it would seem that this is a genuinely important connection with truth, however it might fall short of antecedent expectations. One thing in favor of its being worthy of notice is that, according to this account of IAs, the basis of the connection is something that is true of IAs generally, rather than a distinctive feature of some particular sort of IAs, such as

Copyrighted

Material

Illocutionary

acts and truth

23

assertions. Moreover, even though the basis of the connection is not the full story of what it is to p e r f o r m an IA of a certain type, that basis is still at the heart of what distinguishes a given IA-type from other types. T h u s this is a connection with truth that is a matter of some considerable importance. Still, there is a n o t h e r consideration that belongs on the other side of the ledger. I-rules as I formulate them do not explicitly mention truth. They simply specify the conditions for permissible utterance. Well, you might say, what about these conditions? T h e abstract existence of the conditions in logical or conceptual space is not what the I-rule takes to be required for permissible utterance; at least that is not all that is required. Then what else? Isn't it the truth of the conditions (or, if you prefer, of the formulation or the statement of the conditions)? That is certainly one way to put it. But, and this is the present point, it's not the only way to put it. We could just as well say that what is required is that the conditions obtain or, in a given case that it is a fact that H has a power saw or that it is possible for U's son to clean up his room. And so a devil's advocate might contend that we still have failed to uncover an i m p o r t a n t connection of IAs with truth and only with truth, to p a r a p h r a s e the court room formula. T h o u g h this last point is well taken and must be built into any accurate formulation of what this I-rule account of IAs reveals, it still remains true that the indirect, partial connection in question is a connection with either truth or something logically equivalent to truth. For although what is required by I-rules could be put in terms of facts or of conditions obtaining, it still remains that this is something that is necessarily equivalent to truth. It is logically necessary that if it is a fact that H has a power saw, then it is true that H has a power saw. At least that holds on the "realist" conception of truth in terms of which, as I say in Section 1, this discussion is being conducted. The final issue I will discuss is why, given that all IAs on my view involve the R'ing of propositions, which can be assessed as true or false, only assertives are normally so evaluated. If I assert that Jim was at the party last night, it is quite in order to respond with "That's true" or "That's false." But if I ask you to help me carry this box, it would not be in order to respond "That's false" on the grounds that the proposition It is possible for you to help me carry this box is false, even though that is one of the propositions that I am R'ing to be true in making that assertion. Why is this? H o w can we understand this discrimination? 1 will consider two answers to this question, one based on the account of assertions presented in C h a p t e r 5 of Alston (2000), and one based on the account of assertions given by Tanesini (2005). Assertions, and hence assertives in general, pose several difficulties for my position on IAs. The most prominent one is this. If we ask what is U R'ing to be the case (or to be true) in asserting that lemons are sour, the obvious answer is that U is R'ing that lemons are sour, just what is being asserted. But if we were to take R'ing that p as b o t h necessary and sufficient for asserting that p,

Copyrighted

Material

24

William P. Alston

it would turn out that all IAs are assertions (as well, perhaps, of being, in each case, some other type of IA as well). G o back to the list of IA-types and the conditions R'd in performing each. Consider the conditions R'd by U in ordering U's son to clean up his room. If R'ing a certain condition, C, is sufficient for asserting that C. then in ordering my son to clean up his room I thereby assert the conjunction of those conditions, as well as giving the order. But this seems obviously mistaken. In ordering my son to clean up his room I don't assert anything. Perhaps I could be said to be implying or presupposing that my son has a room, that it is possible for him to clean it, etc. But we must distinguish that from asserting these things. (Moreover there would also be the problem of spelling out what makes my utterance an ordering my son to clean up his room as well as an assertion of the conjunction, something over and above the propositions R'd. But I won't get into that part of the problem here.) To make R'ing that p sufficient for asserting that p lands us into supposing that the category of assertives embraces all IAs, and that seems just plainly wrong. There are two less central but still serious difficulties involved. First, as the above paragraph suggests, we must be able to distinguish what is asserted in an utterance from what is presupposed or implied. R'ing propositions in non-assertive IAs all belong to the latter side of the distinction, as was illustrated in the above example of an order. But we also have the distinction within assertives. Both asserting "There is a chair in the c o r n e r " and asserting " T h e chair in the corner is broken" involve R'ing that there is a chair in the corner. But the first asserts that, whereas the second only presupposes it and does not assert it. Again this shows that R'ing that p is not sufficient for asserting that p, and it raises the question of what else is needed to get a sufficient condition. Second, there is the distinction between asserting and expressing. If I say '"Wow!" or "Superb!" u p o n tasting a dish at dinner I express my pleasure with it. whereas if I say "That gives me a lot of pleasure" I assert the same thing. T h e proposition that this pleases me a great deal is R'd in both cases, but in only the second case is it asserted. Again we have the same problem. R'ing that p is not by itself sufficient for asserting that p, and we need to find what else is needed for sufficiency. A n d , again, we have to be able to distinguish expressing one's pleasure or satisfaction from asserting that one is pleased or satisfied. In reflecting on this nest of problems I came to the position that what we have in asserting that p in addition to what that shares with other cases of R'ing that p is that the sentence (or sentence surrogate) that is uttered in the underlying locutionary act explicitly presents the proposition that p. An explicit presentation of a proposition by a sentence is a matter of isom o r p h i s m between the sentence and the proposition. For every c o m p o n e n t of a proposition there is a corresponding element of the sentence, and the sentence and proposition have the same form that binds these c o m p o n e n t s into a whole. We can get an intuitive notion of this isomorphism by adverting to the fact that the n o r m a l way to refer to a proposition is by a

Copyrighted

Material

Illocutionary acts and truth

25

sentence or a propositional clause. We can take the latter as, so to say, standing in both worlds. As a sentence or a clause it is m a d e up of words or m o r p h e m e s and has a certain grammatical structure, and as specifying a proposition it enumerates the c o m p o n e n t s of the proposition and "shows," as we might say, the logical f o r m thereof. So just by specifying a certain proposition as the proposition that lemons are sour, we, without setting out to d o so, exhibit the isomorphism that, according to this view, is involved in explicitly presenting a proposition in a locutionary act. Before continuing I had better acknowledge objections to this view. In addition to objections to any talk of propositions, which puts one outside this field of play altogether, some will object to the idea that propositions have components that can be lined u p with sentential elements in this way or can share structures with sentences. Those who think of propositions as sets of possible worlds would lodge such an objection. An objection that falls between the two just mentioned is that propositions don't enjoy the kind of independent existence that would be required for this view. To move from straightforward objections to difficulties, there are serious and widely debated problems about how to construe propositions about individuals. Does the individual itself figure as a c o m p o n e n t of the proposition, or is it rather something like an individual essence or a uniquely exemplified property? A n d that is just the tip of the iceberg. W h a t about general propositions? What are the propositional correlates of quantifiers and variables in the sentence? A n d so on. Then there are objections and problems that have to d o with the distinctive feature of my view that an explicit presentation of a proposition is required for assertion. First, there is the point that the simple account applies only to maximally explicit assertions. This has to be supplemented with a way to handle ellipses for explicit assertions. To avoid bursting the bonds of this essay, I will simplify the discussion by restricting it to maximally explicit assertions. Then there is the feeling that this kind of additional requirement is ad hoc, that it constitutes a foreign element in an otherwise neat scheme that differentiates different IA-types solely in terms of what propositions are R'd. There is no d o u b t that it complicates the simplicity of such an account. But such is the price of adequacy, here and elsewhere. Moreover this is not the only point at which such simplicity is sacrificed. Treating singular reference is another one. For examples of such "foreign bodies" in the account consult the list of analyses of various IAtypes in Alston (2000: 239-47). Pleading restrictions of space for not dealing with these objections, problems, and difficulties, I will move on to how this view solves the problems a b o u t assertions that launched the discussion. Since it should be obvious f r o m the above, the exposition can be brief. By requiring an explicit presentation of the proposition in question by the sentential vehicle this goes beyond simply R'ing the proposition in just the right way for constituting an explicit assertion. What better way to "come right o u t " and assert it, in addition to merely R'ing it, t h a n to explicitly present the proposition in

Copyrighted

Material

26

William P. Alston

one's sentential vehicle? In the same way the explicit presentation gives an intuitively obvious way of distinguishing between explicitly asserting that p and merely presupposing or implying it. Finally, I explicitly assert that I am pleased with the dish, rather than just expressing my pleasure, by explicitly presenting the proposition that I am pleased with it. Clearly the explicit presentation is tailor m a d e to solve these problems. That will be taken by some as an objection to it. but be that as it may. it is clearly a strength. N o w for the explanation of the fact that assertive IAs are alone given the distinction of being subject to truth and falsity. The explanation that is based on this account of assertives is very simple. We regard assertives and no other categories of IAs as true or false because the proposition R's by an assertive is explicitly presented in the utterance. The proposition R's stares one in the face. It imposes itself on H's attention. Whereas the propositions R'd in a request or a promise or an exercitive like declaring the meeting a d j o u r n e d are hidden f r o m public view. It takes analytical theorizing to dig them out. Hence it seems much less natural, to put it mildly, to regard a non-assertive IA itself as true or false because of the truth or falsity of some or all of the propositions R'd therein. That is a very simple explanation of the difference between assertions and all non-assertives with respect to the applicability or non-applicability of truth-values to the IAs themselves. If you don't like my account of assertives. whether for some of the reasons I mentioned or otherwise, there is a n o t h e r promising account of the nature of assertives in Tanesini (2005). She presents some objections of her own to my account in Section (i) of that essay. I will not go into that here. If you want to see why I find her objections misguided, you can find it in my response to her in that volume (see Alston 2005). In Section (ii) of her essay she presents her own account of assertives and her explanation of the m o n o p o l y on truth-values enjoyed by assertives that is based on that account. T h a t is what I will discuss here. Her basic idea is this. Two speakers can perform assertive IAs of exactly the same type by R'ing exactly the same conditions, whereas this is not possible for any non-assertive IAs. The reason for this difference is that at least some of the conditions R'd in performing a non-assertive IA have to do with the speaker, so that if the speaker is changed some of the conditions R'd are changed, and hence the IA is not of exactly the same type. Whereas this is not the case for assertives. If I assert that arsenic is poisonous you can make exactly the same assertion by R'ing exactly what I did, naemely that arsenic is poisonous. Whereas for a commissive like a promise it is an essential part of what is R'd that the speaker takes on a certain obligation. With an exercitive like hiring an employer it is an essential part of what is R'd that the speaker has a certain social position. And for a directive like ordering H to do D, it is an essential part of what is R'd that the speaker has authority over H. Here is how Tanesini uses this account to explain why only assertive IAs arc taken to have truth-values. Since you can m a k e the same assertion I was

Copyrighted

Material

Illocutionaryacts and truth

27

making by R'ing just the conditions I was R'ing, I can make the same assertion by responding to you by saying "That's true." That is, I can make the same assertion by riding piggy-back on your R'ing. responding to it by a " C o u n t me in on that too," or in a vernacular phrase, " W h a t he said." But for the reasons given above this is not possible with non-assertive IAs. And this, Tanesini thinks, is the source of our willingness to count assertives as true or false but not to treat non-assertives in the same way. This is an ingenious suggestion and deserves to be taken seriously. Let me begin to oppose it by pointing out a feature of the use of her account to explain the exclusive right of assertives to have truth-values. That feature is a certain superficiality. It consists of a reason why we talk that way about others' assertive IAs. Tanesini is undoubtedly right about that, but this doesn't go very deeply into why assertive IAs, unlike IAs of other categories, are correctly termed true or false. It confines itself to a function that treating them as true or false plays in our conversations. But a deeper question concerns what there is a b o u t assertives that makes it correct to apply truthvalues to them. Interestingly enough, the very feature that is missing from her account is the one I fasten on to explain what there is distinctive about assertives that makes it correct to attribute truth-values to them and only to them a m o n g IAs, namely the explicit presentation of what condition is being R'd. This latter is not itself a reason to prefer my account to hers, only one way of seeing their opposition. But there are more serious difficulties with Tanesini's position. I could question her unrestricted claim that all non-assertive IAs involve R'ing one or more conditions about the speaker. But I take that to be close enough to the truth of the matter to let it pass. But the most obvious defect is this. It is not universally true of assertives that performing them does not involve R'ing something about the speaker, something that would prevent a n o t h e r speaker from performing an IA of exactly the same type by R'ing exactly the same conditions. This does not hold of all assertives because of a very f u n d a m e n t a l feature of assertives, namely that no restriction can be put on the kinds of propositions that can be asserted. This being the case, it is obviously possible for a speaker to assert something about him/herself, such as "I am hungry," "I am a better tennis player than Robinson," "I am the very model of a m o d e r n m a j o r general," and so on. For such assertives it is not possible, on Tanesini's own showing, for a n o t h e r speaker to make exactly the same assertion by R ' i n g exactly the same conditions R'd by the original speaker. And hence her claim to have given a general way of distinguishing assertives f r o m other kinds of IAs does not hold up. With its demise her explanation of why assertives alone a m o n g IAs have the distinction of being true or false suffers defeat. The same over-generalization over assertives makes itself felt there as well. For it is just as undeniable that assertions about oneself are true or false as assertions about anything else. Tanesini might seek to salvage something from her account of assertives by claiming that at least there is this difference between assertives and

Copyrighted

Material

28

William P. Alston

non-assertives. It is impossible to perform a non-assertive IA without R ' i n g something that prevents a n o t h e r speaker from performing exactly the same IA by R ' i n g that condition. But this is obviously possible for assertives. Just making the assertion about something other than the speaker is all it takes. This would give a weakened form of the original view of how to distinguish between assertives and non-assertives, just how seriously weakened I will leave it to the reader to judge. But it will not save her explanation of why truth-values are correctly applied to only assertives a m o n g IAs. For the fact remains that truth-values are equally applied to assertions about oneself, where this explanation is not available. Rather t h a n being focused on a single problem, this essay has dealt with a n u m b e r of different problems. Such unity as it has comes f r o m the fact that all these problems have to do with ways in which it has, or might have, been t h o u g h t that IAs are importantly connected with truth. A reader would be excused for coming away with the impression that the overall message of the discussion is negative, that such allegations properly evoke a "Bah! H u m b u g " response. T h o u g h such an impression would be oversimplified, I will not contest that the weight of emphasis is on that side of the ledger. Here is a brief s u m m a r y of the treatment of each suggestion. 1 It c a n n o t be maintained that assertions "present what is asserted as t r u e " if that means that the property of t r u t h is ascribed to the proposition, even if only implicitly. That would seem to imply the mistaken view, endorsed by Frege, that all assertions predicate truth (or factuality) to propositions. Weaker readings can be given, but they fail to be serious candidates for specifying what it is to make an assertion. 2 There are various attempts to get all IAs into connection with truthvalues by taking assertions as basic for IAs generally. T h e boldest proposal is that of David Lewis. He forthrightly claims that all IAs are assertions. Putatively non-assertive IAs are given that status by taking their deep structure to be "explicit performatives," like "I order you to eat your spinach." which in turn are construed as assertions that one is performing the IA specified. In addition to problems about this last construal, the account runs into the fatal difficulty that it c a n n o t explain the obvious fact that sentences like "Eat your spinach" and "The meeting stands a d j o u r n e d " are fitted by their meanings to perform nonassertive IAs (even if they were also fitted by their meanings to make assertions, as Lewis supposes). This means that not all sentences can be treated as purely assertive. as Lewis claims. There are also various attempts to find something that plays roles for non-assertive IAs that truth and falsity play for assertions. But even if such projects were extended to IAs of all kinds, as is generally not the case, they would still fall short of showing that truth is importantly connected to IAs of all kinds.

Copyrighted

Material

Illocutionary

acts and truth

29

3 The most promising suggestion for an important connection between IAs generally and truth comes f r o m a view of mine as to how to analyze IAtypes. The basic idea is that a large part, and the central part, of what it is to perform an IA of a certain type is to "take responsibility" for the holding of certain conditions. Different IA-types are primarily, though not solely, distinguished by different conditions that have this role. This implies that there is something out of order, something amiss if one sets out to, for example, ask someone to pass the salt and one or more of the conditions associated with that IA-type is false. This provides an i m p o r t a n t connection with IAs and truth. It is not the simplest kind of connection as it would be if the truth of the propositions in the content of the conditions (satisfaction of the conditions) were necessary for an IA of that type being performed. But still, since taking responsibility for the truth of the propositions in question is a necessary condition for IA performance, it is a significant, and indeed important, connection with truth. 4

The final problem treated is this. Given that any IA involves the speaker's taking responsibility for the truth (satisfaction) of certain conditions, why is it only assertive IAs to which we attribute truth-values? I consider two theories of assertion and the explanation of this derived from each. I defend one of those theories (not surprisingly my own) according to which what is required for an assertion of p over and above taking responsibility for the truth of p is that the sentence used for the assertion "explicitly presents" the proposition that p. This is, of course, a connection with assertive IAs only, but since it is in contrast to non-assertive IAs, there is a kind of indirect relationship with IAs generally.

I would like to find a way to summarize the conclusions of the essay in a short formula. But I must confess that the plurality of problems discussed is such that I am at a loss for any such formula that is more substantive t h a n saying that I take my discussion to show that there are some significant connections between IAs and truth, but that, in general, they are less direct and less exciting than is often thought.

Notes 1 For a much fuller discussion of how to demarcate the illocutionary act category see Alston (2000: Ch. 1). 2 For more details see Alston (1996: Ch. 1. esp. Sees vi and vii). 3 For a treatment of the concept-property relationship with respect to truth see Alston (2002). 4 See, e.g.. Frege (1970: esp. Sections 2 and 3). Frege sometimes puts his point with respect to the judgment expressed by an assertion, and he represents the speaker as "expressing his recognition of the truth" of a proposition. And he sometimes speaks in terms of presenting the proposition as a fact. But I take these locutions as equivalent to the one I used. 5 But in (Frege 1970: Ch. 1, Sec. 3) Frege embraces such an understanding of his view with respect to the formal language he is developing. "You may if you like

Copyrighted

Material

30

William

P. Alston

distinguish subject and predicate even here; but the subject contains the whole content, and the only purpose of the predicate is to present this in the form of a judgment. Such a language would have only a single predicate for all judgments, viz., 'is a fact.'" 6 "Locutionary act" is understood broadly here to include not only the utterance of sentences and linguistic sentence surrogates, but the production of anything publicly observable that could be a vehicle of meaning something. 7 This is a typical level of complexity for Schiffer. For further details of his view see Schiffer (1972: Ch. IV, esp. Sec. 2). 8 These cases arc introduced by Schiffer in relation to his account of S's meaning something by an utterance rather than in connection with his account of IAs. But since the former is an essential part of the latter, the problems carry over to the latter. 9 This is treated in Chapter 8 of Alston (2000).

References Alston. William P. (1964) Philosophy of Language. Englewood Cliffs, NJ: Prentice-Hall. (1994) "Illocutionary Acts and Linguistic Meaning," in L. Tsohatzidis Savas (ed.) Foundations of Speech Act Theory. London: Routledge. (1996) A Realist Conception of Truth. Ithaca, NY: Cornell University Press. (2000) Illocutionary Acts and Sentence Meaning. Ithaca, NY: Cornell University Press. —(2002) "Truth." in R. Schantz (ed.) What is Truth?. Berlin and New York: de Gruyter. —(2005) "Response to Tanesini," in H. D. Battaly and M. P. Lynch (eds) Perspectives on the Philosophy of William P. Alston. Lanham. MD: Rowman & Littlefield, pp. 251-55. Frege, G. (1970) Begriffssehrift (1879). in P. Geach and M. Black (eds) Translations from the Philosophical Writings of Gottlob Frege. Oxford: Basil Blackwell. Grice, H. P. (1957) "Meaning." Philosophical Review. 66: reprinted in H. P. Grice, Studies in the Way of Words. Cambridge. MA: Harvard University Press, pp. 213 -23. (1969) "Utterer's Meaning and Intentions." Philosophical Review. 78; reprinted in H. P. Grice, Studies in the Way of Words. Cambridge. MA: Harvard University Press, pp. 86-116. Lewis, D. (1972) "General Semantics," in D. Davidson and G. Harman (eds) Semantics of Natural Language, Dordrecht: Reidel; reprinted in D. Lewis, Philosophical Papers, vol. I. New York: Oxford University Press. (1983) Philosophical Papers, vol. 1. New York: Oxford University Press. Pollock, J. (1982) Language and Thought, Princeton: Princeton University Press. Schiffer, S. R. (1972) Meaning, Oxford: Clarendon Press Searle, J. (1969) Speech Acts. Cambridge: Cambridge University Press. (1983) Intentionality, Cambridge: Cambridge University Press. Stenius, E. (1967) "Mood and Language Game." Synthese, 17: 254-74. Tanesini, A. (2005) "A Theory of Assertives." in H. D. Battaly and M. P. Lynch (eds) Perspectives on the Philosophy of William P. Alston, Lanham. MD: Rowman & Littlefield, pp. 239-49. Vanderveken, D. (1990) Meaning and Speech Acts. Cambridge: Cambridge University Press.

Copyrighted

Material

2

Illocutionary acts and the concept of truth John R. Searle

What exactly is the role of the truth in a theory of speech acts? The theory of speech acts tells us about the structure and functioning of illocutionary acts. A theory of truth will give us an analysis of the concept of truth. What is the relationship between these? Notice that this question is not quite the same, though closely related to, the question, What is the role of truth in the actual performance of speech acts? The reason they are not exactly the same is that a phenomenon can be important in the actual performance of speech acts without being important in the theory. For example, if you ask. What is the role of phonetics and phonology in the theory of speech acts? The answer is: Not much. But if you ask. What is the role of phonetic and phonological phenomena in the actual performance of speech acts? The answer is: They are very important. I will concentrate on the first question before addressing the second. In order to address the question, I need to say something about the theory of speech acts and about the concept of truth. The basic structure of the illocutionary act can be represented as "F(p)," where " F " marks the illocutionary force, or type of the speech act and " p " marks the propositional content of that speech act. Thus the two illocutionary acts "Please leave the room," and "Will you leave the room?" have, at least in part, the same propositional content, "that you will leave the room," but it is presented with two different illocutionary forces, and thus in two different illocutionary acts. There are five basic types of illocutionary acts: Assertives, whose illocutionary point is to represent how things are in the world; Directives, whose illocutionary point is to get the hearer to do something; Commissives, whose illocutionary point is to commit the speaker to some course of action; Expressives, whose illocutionary point is to express some psychological state of the speaker, and Declarations, whose illocutionary point is to bring about a correspondence between the propositional content and reality by declaring the correspondence to exist. The task of Assertives is to fit the world, and thus they have the word-to-world direction of fit. The task of Directives and Commissives is to get the world to change to match the content of the speech act, and thus they have the world-to-word direction of fit. The task of Expressives is, typically, to express

Copyrighted

Material

32

John R. Searle

the speaker's feelings about a state of affairs that is already presupposed to exist, and thus they have no special direction of fit. I call this the null direction of fit. The task of Declarations is to bring about changes in the world by representing the world as being so changed, and thus they have b o t h directions of fit simultaneously. I am leaving out a very large n u m b e r of details, exceptions, qualifications and special cases in order to give you the bare bones of the theory of speech acts. N o w let us turn to truth. There is no generally accepted analysis of truth in c o n t e m p o r a r y philosophy, but there are certain general tendencies and constraints that are widely recognized. I will concentrate on two of these. First, disquotation. It is a constraint on any theory of truth that it satisfies the following condition. In the open sentence S is true if and only if p for " S " in any given case we substitute the specification of a statement, usually by putting quotation m a r k s a r o u n d the associated sentence, and for " p " we put in an expression of the propositional content of that very statement, usually by putting in the words of the original sentence, without the q u o t a t i o n marks. T h u s to take an example m a d e f a m o u s by Tarski. "Snow is white" is true if and only if snow is white. Tarski called this requirement Convention T. The condition is sometimes called disquotation, because the same sequence of words occurs on both the left and the right hand side, but on the right hand side it is without quotes, i.e. disquoted. A second constraint on the analysis of truth is that it should capture our intuition that when a statement is true, there must be something in virtue of which it is true. There is something that makes it true. This second constraint is expressed in a variety of ways, but the intuitive idea behind it appears to be that if a statement is true, it must satisfy a certain condition, its t r u t h condition. A n o t h e r way to put the point is to say that if a statement is true, it must somehow m a t c h some feature of the world that makes it true. I call this second condition the " m a t c h i n g c o n d i t i o n . " Some of the most f a m o u s formulations of this requirement involve the notion of facts. T h u s a statement is said to be true if and only if it corresponds to the facts, fits the facts, states the facts, matches the facts, or describes the facts. These two conditions tend to lead to different theories of truth. Disquotation is the inspiration behind various minimalist or deflationary or even redundancy theories of truth that say in general that there is n o such thing as a separate property of truth, that the expression "is t r u e " really a d d s nothing except perhaps emphasis. To say that "snow is white" is true is just to say that snow is white.

Copyrighted

Material

Illocutionary

acts and the concept of truth

33

The matching condition inspires other theories of truth; the m o s t f a m o u s of these is the correspondence theory, the theory that a statement is true if and only if it corresponds to some features of the world, a n d these features are typically called "facts." With these preliminaries taken care of, let us now turn to o u r main question, W h a t is the role of truth in the theory of illocutionary acts? Let us begin by asking, first, what exactly is the entity that is literally and strictly characterized as true or false. It has become respectable, or at least acceptable, in philosophy once again, to describe sentences as true or false. I think this is usually harmless, but it is not, strictly speaking, accurate. A sentence is a syntactical entity a n d as such has syntactical properties. Of sentences as syntactical objects we can say such things as that they have ten words, contain a transitive verb and are in English. A sentence as a syntactical entity has semantic properties attached to it, its meaning, and in virtue of these semantic properties it can be used to talk with, to perform speech acts. For sentences in the indicative, we can often say of these semantic p h e n o m e n a associated with sentences that they are true or false, but truth and falsity are not syntactical properties. If we arbitrarily decide, as Tarski did, to take sentences as the entities to which truth and falsity are ascribed, we get the absurd result that t r u t h is always relative to a specific language. "Snow is white" is true, but only in English. We can see that this is unsatisfactory, because what we say in English when we say "Snow is white" is exactly the same as what we say in French when we say " L a neige est blanche" and in G e r m a n when we say "Schnee ist weiss," not three different things. It is that very same thing that is said in all three languages and it is that single thing which is true, not the three different things. Why am I so confident that it is a mistake to think truth is essentially a property of sentences? Because on the language side it is the same thing that is said when one utters any of these three sentences and it is the same thing in the world, the same truth maker, that satisfies the matching condition. It is a condition of adequacy on our analysis of truth that what is true, when said in one language, is the very same thing which is true, when said in another language. To say that the sentence is true is a kind of metonymy. We are ascribing properties of the thing contained, the semantic element, to the container, the syntactical sentence. But literally speaking, sentences are not the bearers of truth. The answer to our question about the entity which can be literally true or false seems obvious. It is the statement, along with other members of the assertive class of speech acts, such as assertion, description and characterization, which can be said to be true or false, and it is i m p o r t a n t in this connection to emphasize that it is the statement and not the act of stating which is true or false. It is, rather, the statement m a d e in the act of stating which can be said to be true or false, not the act of stating itself. The act of stating, of course, admits of such adverbial characterizations as " H e stated truly t h a t " but the adverbial characterization will be true if and only if the

Copyrighted

Material

34

John R. Searle

corresponding statement is true. So let us suppose, then, that our primary objects of assessment with the expressions " t r u e " and "false" are those illocutions of the assertive class that have the word-to-world direction of fit. Implicit in the account so far is a rather simple answer to the question concerning the role of truth in a theory of illocutions. "True" and "false" are the favorite terms of assessment for achieving success of iillocutions which have the word-to-world direction of fit. Thus, we assess statements, assertions, descriptions, and explanations as literally true or false, but we do not, in that sense, assess promises, vows, threats, questions, c o m m a n d s and requests as true or false. Furthermore, those speech acts that have the null direction of fit, such as apologies, t h a n k s and congratulations, again are not assessed as either true or false. It would look, then, as if truth and falsity have no special role to play in a theory of illocutionary acts. Some speech acts have the word-to-world direction of fit, some do not. Those that do, are assessed as true and false. Those that d o not. are not. On this account " t r u e " and "false" are terms used to assess a restricted class of speech acts, an i m p o r t a n t class, of course, but nonetheless just one class of speech acts a m o n g many others. Orders and c o m m a n d s are assessed as obeyed or disobeyed, requests as granted or denied, promises as kept or broken and statements as true and false. In the theory of speech acts there is nothing special about truth and falsity. There is something intuitively appealing about this account, but I think in the end it is mistaken, and if we can see exactly how it is mistaken, the analysis that shows it is mistaken may give us some i m p o r t a n t insights. We need first to remark that there appears to be an asymmetry between statements and many other sorts of speech acts in the following respect. If we ask the question. In virtue of what is the order said to be obeyed? the answer is, in virtue of the fact that the person who was ordered to do something did the thing he was ordered to do a n d did it because he was ordered to d o it. And if we ask, In virtue of what is the promise said to be kept? the answer is, in virtue of the fact that the person who m a d e the promise did the thing he promised to d o and did it because he promised to do it. And if we ask. In virtue of what is the statement true? the answer is, in virtue of the fact that the world is the way that the statement says it is. So far the cases seem to be perfectly parallel. But the asymmetry appears as follows. If we are asked to generalize the previous account that gave the conditions of satisfaction of the various types of speech act. we would have to say, in every case, the speech act is satisfied if and only if the propositional content is true. T h u s the statement is true when the proposition expressed is true, but the order is obeyed when the propositional content is true (specifically when the hearer acts so as to make the rest of the propositional content true), and the promise is kept when the propositional content is true (specifically when the speaker acts so as to make the rest of the propositional content true). Notice that we do not say of orders and promises that the propositional content is obeyed, or that the propositional content is

Copyrighted

Material

Illocutionary

acts and the concept of truth

35

kept. Rather, the obedience and being-kept attach to the whole speech act in virtue of the fact that the propositional content is true. So the asymmetry comes out as follows: in the case of statements, you get true statements in virtue of true propositions. But in the case of orders, you do not get an obeyed order in virtue of an obeyed proposition, but rather in virtue of a true proposition. T h e propositional content determines what constitutes obedience, hence the truth of the proposition constitutes an obeyed order. And in the case of promises, you do not get a kept promise in virtue of a kept proposition, but in virtue of a true proposition. The propositional content determines what constitutes keeping the promise, hence the truth of the propositional content of a promise constitutes a kept promise. If we remind ourselves that all of the speech acts we are talking about have the structure "F(p)," and we are looking at the illocution which is the result of the performance of a speech act that has this structure, we find this asymmetry. Obedience attaches only to the whole speech act, not to the propositional content, and being kept attaches to the whole speech act and not to the propositional content. But in each case the order is obeyed or the promise kept only if the propositional content is true. To put the point more precisely, the actions which constitute the obedience to the order, or the keeping of the promise, will constitute the truth conditions of the corresponding proposition. The propositional content is not that which is obeyed or disobeyed, rather the propositional content represents that which constitutes obedience or disobedience. The propositional content can be true in Directives and Commissives as much as in Assertives. Furthermore, though the order is not said to be true or false, but rather obeyed or disobeyed, when the order is obeyed, the propositional content is true. A n d similarly for promises; though a promise is not said to be true or false but kept or broken, the promise is kept when the propositional content is true. So, what is going on here? The hypothesis that is suggested, though of course certainly not demonstrated by these few reflections, is that truth plays a more basic role in describing the relationship between language and reality t h a n other terms of assessment. The other terms seem to presuppose truth, but truth does not presuppose them. Let us explore this idea further. If I say, "Please leave the room," and thus make a request, I seem to presuppose that you are at that time in the room. What is the status of that presupposition? Well, it seems natural to say something like the following: Unless the presupposition is true, there is something improper about the request. The request seems to require that the presupposition be true in order that the request can, so to speak, come into effect, that it can have any force. And notice that when we said the presupposition is true, we were not suggesting that someone might have m a d e a true assertion by asserting the presupposition, we were not appealing to a possible assertion. Rather, we were just considering whether or not the proposition was true or false quite apart from whether or not it ever had been or ever would be asserted.

Copyrighted

Material

36

John R. Searle

Let us expand the hypothesis a little bit further. The basic function of language is to represent reality in the various illocutionary modes, the Assertive, Directive, Commissive, Expressive and Declarational. The words true a n d false have evolved to assess success in the word-to-world direction of fit. However, they seem to be more f u n d a m e n t a l than the other terms of assessment for the matching relation between language and reality because the notion of representing how things are (or will be, or have been) seems to me the most f u n d a m e n t a l f o r m of representation. Now, exactly in what sense is it more f u n d a m e n t a l than the other forms? To put the point in its crude form, the other illocutionary forms presuppose the true-false forms, but the true false forms do not presuppose the other forms. You might have a language just consisting of statements where it would seem mysterious or unintelligible if you tried to give an order or make a promise, but you could not have a language containing only orders and promises. Why not? Because in such a language, you still need to describe the world in which the order is obeyed or disobeyed and the promise kept or broken; and furthermore, you still need to be able to say that the presuppositions are, or are not. true. And "kept," "broken," "obeyed" and "disobeyed" are not sufficient to do that because they only apply to the order or promise itself and not to the subsequent descriptions of reality in which the order is obeyed or disobeyed. nor to the presuppositions of the order and promise. So the first upshot of our discussion is that the word "true" is more f u n d a mental than other terms of appraisal of speech acts and does more for us than just set the criteria of success of a certain Assertive class of speech acts, because we need it t h r o u g h o u t our theory for all classes of speech acts, including those that do not even have a truth value such as Directives and Commissives. Truth is different from other terms of assessments of speech acts because the f u n d a m e n t a l linguistic relation is representation, in the way that a propositional content sets truth conditions, and in so doing represents the state of affairs expressed by those truth conditions. And this relation is essential to understanding all speech acts that have the structure " F ( p ) . " I and it is now time to explore that relation in more detail. The s t a n d a r d way in which the matching relation is described in the history of philosophy is with the word "correspondence." and the idea that statements, if true, match reality is sometimes called the correspondence theory of truth. This particular theory has such a checkered and, sometimes even sordid, history; and the debates are so lengthy and confused, that one is reluctant to raise the issue again, but I am going to do so anyhow. If stated correctly, the correspondence theory is not subject to the standard objections m a d e to it. Indeed, if stated correctly, the correspondence theory is a logical consequence of the disquotational principle, together with certain very natural assumptions. One way to get at these points is to see how the disquotational principle, a principle that is taken to be antithetical to the correspondence theory, in

Copyrighted

Material

Illocutionary

acts and the concept of truth

37

fact, understood in a very plausible way, implies the correspondence theory. Let us go through the steps. 1 Step 1: for any statement S, the ascription of truth to S must satisfy the disquotation condition, S is true iff p where for S we substitute a specification of a statement and for p we substitute the expression of the propositional content of that statement. This gives us a criterion for the analysis of the concept of truth because any analysis must be consistent with it. What motivates the disquotation principle (convention T)? Why should it be like that a n d not some other way? The answer, I think, is clear. Any statement sets a truth condition and the statement will be true i f f that condition is satisfied. T h e disquotational principle is an expression of this feature because it illustrates the fact that the statement specified on the lefth a n d side, "Snow is white," sets a certain condition expressed in its propositional content, that snow is white, and the statement will be true iff the condition is satisfied. Or, to put it briefly, "Snow is white" is true iff snow is white. And so on for every statement (I am leaving out cases of indirect speech acts, indexicality, etc. for the purposes of this discussion). T h e notion of a condition has the usual process-product ambiguity between a requirement and the thing required. The statement specifies the requirement, the fact is the thing required. Step 2: But, now, it would be useful to have a name for those satisfied truth conditions and, indeed, there are various names that have evolved, most famously, the word "fact." A fact is a condition (thing required) in the world that satisfies the truth condition (requirement) of the corresponding statement. Other expressions are "states of affairs," "conditions in the world," etc. Step 3: We have so far identified statements and facts as the elements of a theory of truth. Notice that this follows trivially from disquotation, given only the natural assumptions that we need a general term for the entities specified on the left hand side, statements, and a general term for those specified on the right hand side, satisfied truth conditions, i.e. facts. But now it seems equally natural to ask, What is the relation between them? What fills the gap in: A statement is true iff it

a fact.

And now we need a verb for the blank. A number of words have evolved. Most f a m o u s is "corresponds t o " but others are "states," "describes," "accurately represents," etc. So now we can say a statement is true iff it corresponds to a fact, (or fits the facts, or states a fact, etc.). Properly understood, the correspondence theory is a generalization of the disquotation principle. Disquotation gives us for each statement a criterion

Copyrighted

Material

38

John R. Searle

of truth and if we get a general statement of disquotation, with n a m e s for the two elements, the correspondence theory follows. But, as always, language can be a source of confusion to us because it becomes tempting to think that facts must be a special kind of complex objects in the world, and that there must be a special kind of relation in which statements stand to these entities when they are true, a special kind of picturing or isomorphism relation. A n d of course, all of that is wrong. This seems to raise traditional issues about negative facts, hypothetical facts, etc. But the correspondence theory must be as trivial as the disquotation criterion which it generalizes. T h u s the negative statement "There are no horses in the r o o m " is true precisely because it corresponds to the fact that there are no horses in the room, and that is just a long-winded way of saying what convention T says in a briefer way. A fact, to repeat, is just a satisfied truth condition. W h a t shall we say then about the apparent dispute between the minimalist or disquotationalist view, and the correspondence or substantive view? As so often happens in philosophy the appearance of disagreement is entirely an illusion generated by a failure to understand the logic of our language. The principles that are apparently in conflict are two tautologies. A n d two tautologies c a n n o t be in any logical conflict. The first tautology is that any account of statements and truth must satisfy the disquotation principle S is true if and only if p, where we substitute a specification of a statement for " S " and the expression of the propositional content of the statement itself for "p." T h e second tautology is that a statement is true if and only if it corresponds to the facts. Both tautologies are analytic truths. There c a n n o t be any conflict between them - we simply have to remove the appearance of conflict by showing how language works in this case. So why was there such a longstanding problem? Well the problem is that the articulation of the correspondence theory, "a statement is true if and only if it corresponds to the facts," gives us the illusion that there are two quite distinct, separately identifiable entities like the picture of a man and the m a n himself - and that we arrive at an assessment of truth by scrutinizing the two entities and seeing if they really match. But of course you cannot do that with statements and facts, because once you have identified the fact, you have already stated a true statement. You cannot answer the question, Which fact does the statement that snow is white correspond to? without expressing the propositional content, that snow is white, because of course the fact that snow is white is simply the satisfied condition set by the propositional content of the statement that snow is white. So it might look like the correspondence theory is saying something false, but it is not, if it is properly construed. It is a trivial consequence of the equally trivial disquotational principle.

Copyrighted

Material

Illocutionary

acts and the concept of truth

39

Several i m p o r t a n t philosophers make a very c o m m o n mistake about the relationship between the assertive class of speech acts and truth. They have the idea that in the relationship between a statement and truth, the commitment to truth is somehow external to the act of stating. I am thinking of Williams (2002), Lewis (1970) and Grice (1975: 44-58). Williams asks, how could our interest in truth have evolved genealogically assuming the existence of statements. Paul Grice says that it is a maxim of conversation that when we are making statements we should make statements we believe to be true, (or at least not make statements we believe to be false) and David Lewis says that we have a rule, or convention, to the effect that when we are making statements, we should only make true ones. I believe that all three of these approaches involve a f u n d a m e n t a l error, but it is not obvious what the error is, so let me spell it out. Ask yourself: What is the definition of a statement? What is constitutive of m a k i n g a statement? The answer is that a statement is defined as a commitment to truth. There are not two things, the statement and then the c o m m i t m e n t to truth. To make a statement is already to be committed to truth. Grice, Williams and Lewis see correctly that it is possible to make false statements, and so they think that somehow or other the rule Make statements you believe to be true is an external rule, like the rule Be relevant or Do not talk too loud. But the cases are quite different. The relationship of truth to statement making is internal, not in the sense that every statement is true but, rather, that every statement is a commitment to truth. I put this many years ago by saying that it is a constitutive rule of statement making that a statement is a c o m m i t m e n t to truth. But now I want to develop this point a little bit further. Just as it is internal to the notion of a statement that it is a commitment to truth, so it is internal to the notion of a proposition that it can be either true or false. I want to pursue this analogy and also point out its limitations. The concept of a statement is not a philosopher's creation. Statements are p h e n o m e n a in the real social world, whereas "proposition" is a technical term. However, it is a powerfully motivated technical term because there is no way we can account for what is c o m m o n to all sorts of different speech acts without the notion of a proposition. At the barest minimum, we need to be able to state what is c o m m o n in arguments, for example, m o d u s ponens arguments: If p then q, and p, therefore q We cannot understand this formula unless we take p and q as m a r k i n g separate contents, but each separate content cannot be an assertion or a statement because, of course, in the hypothetical form it is not asserted. So the notion of a proposition is one we c a n n o t do without. T h e p h e n o m e n a of propositionhood we discover, but the definition of a proposition is up to us, because "proposition" is a technical term. There is no way we could give an intelligible definition of a proposition without m a k i n g the notion of t r u t h

Copyrighted

Material

40

John R. Searle

once again internal to p r o p o s i t i o n h o o d because to be a proposition is already to be in the dimension marked by " t r u e " and "false." So, just as truth is internal to statements, truth is internal to propositions. But the relations are different. Truth is internal to statements because every statement is a commitment to truth. But truth is also internal to propositions, because the propositional content of the statement determines the truth condition to the existence of which the maker of the statement is committed, and therefore because - with the usual qualifications, about vagueness, presupposition failure, etc. - every proposition is either true or false. And similarly for orders, promises and rest. The truth of the propositional content constitutes the obedience condition of the order, and the being-kept condition of the promise. In my discussions of statements over the years, there is a peculiar oscillation between defining a statement as a c o m m i t m e n t to truth and defining a statement as a representation of how things are. But if you think about it a m o m e n t , those are exactly the same. Why? Because the notion of truth is precisely the notion of the assessment of a statement that states how things are. If it really states how things are then it is true. This is the deeper point behind the disquotation-matching condition. Every proposition sets a truth condition, and when that condition is satisfied, it is true.

Conclusion The f u n d a m e n t a l use of the word " t r u e " in assessing speech acts is to appraise members of the assertive class of speech acts. But the f u n d a m e n t a l concept of truth is defined not in terms of statements but in terms of propositions. A n d a proposition is true if and only if it satisfies the disquotation-matching condition. A n d these two apparently different conditions are exactly the same.

Note 1 This argument is a summary of material presented in more detail in Searle (1995: Ch. 9)."

References Grice, P. (1975) "Logic and Conversation." in P. Cole and J. Morgan (eds) Syntax and Semantics, vol. 3, New York: Academic Press. Lewis, D. (1970) "General Semantics," Symthese. 22: 18 67: reprinted in D. Davidson and G. Harman (eds) (1972) Semantics of Saturai Language, Dordrecht: Reidel: also in D. Lewis, (1983) Philosophical Papers 1. Oxford: Oxford University Press. Searle. John R. (1995) The Construction of Social Reality. New York: The Free Press. Williams, B. (2002) Truth and Truthfulness, Princeton, NJ: Princeton University Press.

Copyrighted

Material

3

Alethic acts and alethiological reflection1 An outline of a constructive philosophy of truth Geo Siegwart

In this paper, I would like to give an overall picture of the topic of truth, in which the following theses will gain some evidence. (1) The cognitive business is part and parcel of our coping with life. Acts of truth or alethic acts, as I will call them, are cognitive acts, whereas the reverse does not hold. (2) To assert a proposition is to present it as true. But asserting is just one way, among others, of presenting something as true. (3) Alethic acts as well as cognitive acts in general sometimes fail, as do all our acts. The reflection on cognitive acts in general and the specific reflection on alethic acts arc ultimately enterprises seeking to manage this failure. (4) We all perform alethic acts every day. To do this, we do not need a truth-predicate, but an illocutionary-truth-operator. Truth-predicates are only used by the few people reflecting on truth. (5) Cognitive acts are illocutionary acts. So the idea of a rule helps to distinguish between correct and incorrect performances of a cognitive act. It also provides the starting point for developing the connection between (illocutionary and propositional) meaning on the one hand and alethic acts and truth on the other in a natural way. (6) To give the meaning of a word is. roughly speaking, to establish the rules of cognitive, alethic as well as non-alethic, acts in which this expression may play a role or to take an equivalent measure. So, the truth or falsity of a proposition results from the meanings of its parts. The paper is divided into a short introduction and three main parts. The introductory remarks point to the place the alethic business has with respect to the whole of human acts (A). Distinguishing between alethic acts and alethiological reflection in the first main part, I also establish the concepts that will be used in the following sections (B). Placing alethic acts within the totality of cognitive acts, I will also introduce the rule-idea (C). Well-equipped with the mentioned tools, I am able to discuss the essential question of the connection between meaning and truth in the third main part (D). A Locating the alethic business in the totality of human acts Instrumental rationality is a characteristic of our coping with life as human beings: we try to achieve certain ends by means of appropriate actions. In

Copyrighted

Material

42

Geo

Siegwart

order to ensure that we take the right means to given ends, we need knowledge, preferably truths, providing us with adequate orientation. Usually drastic examples are presented to illustrate this insight: Someone who wants to prepare the picked m u s h r o o m s also wants to make sure in advance that they belong to an edible sort. One does not provide someone with a driving-license who does not have the perception that is necessary to drive a car without causing damage. In order to prepare a well-tasting risotto of nettles, one must be able to distinguish reliably between young and old nettles. The first two examples compel approval because of the drastic consequences, but they fail to d e m o n s t r a t e the close and continuous connection and interplay of cognitive particularly alethic - acts and non-cognitive acts in all the sequences of our acts. The reader may develop the mentioned example of preparing and enjoying a risotto of nettles to consider the way in which cognitive acts and non-cognitive acts interplay. We are used to carrying out the alethic business: we qualify propositions as true, and we fall back on the stock of true propositions and integrate such propositions with our acting. Often we do not get the needed truths and have to be content with presumptions and hypotheses, with estimates, predictions and retrodictions, but we would always prefer cognition in its perfect f o r m , i.e. true cognitive results, to these mentioned kinds of imperfect cognition. The need to act purposefully induces efforts to qualify entities as true or to get respective information, which, in turn. is the result of qualifications-as-true that have already been made and have been incorporated into a body of knowledge. There is an indissoluble connection between the c o m m o n and familiar practice of performing cognitive and alethic acts and the performance of other (non-cognitive) acts. For someone who bears this state of affairs in mind, it does not matter if cognitive or alethic acts are performed using an everyday language or a particular language, for example a scientific language. A n o t h e r aspect has to be considered: all acts are subject to failure. M o r e precisely: the ends we aim at by performing certain acts are not always achieved - that would be like heaven on earth! On the other hand, we do not always fail to achieve the aims we try to achieve by acting - that would be like hell on earth. To give the right description, we should say that it is rather like earth on earth: Acting purposefully runs smoothly in many cases, but sometimes - or perhaps often it is disturbed by some trouble. The interest in assuring success, in managing trouble, in preventing difficulties and in optimizing action generally is due to the sketched basic h u m a n condition. Interest in the support of actions is guided by reflection on acting. Reflection, no matter to what kind it belongs. is a business of action support. Cognitive and alethic acts are part and parcel of the stream of action. So, being acts themselves, they are subject to failure as all actions are: Trouble in the cognitive d o m a i n a p p e a r s in the form of fallacies, observation mistakes, inadmissible presuppositions, pseudo-problems, errors, delusions, specious consent, dissensions, miscalculations, incorrect definitions, paradoxes, anomalies, antinomies, pseudo-explanations, and so on.

Copyrighted

Material

Alethic acts and alethiological reflection

43

Epistemological reflection in general and alethiological reflection aim at overcoming cognitive and alethic disturbances respectively. Speaking of disciplines, one should say that epistemology (or gnoseology) and its branch philosophy of truth (or alethiology) serve to support the cognitive and the particular alethic business. Three specifications are advisable: (i) Gnoseology and alethiology are not concerned with all features of cognition. They are concerned with those aspects of cognition that are more than familiar and that play a role in all domains or nearly all domains, no matter if they are scientific ones or part of everyday life. (ii) Reflections in epistemology and philosophy of truth are part of the cognitive and alethic practice. So, they belong themselves to the subjects of consideration because they are themselves subject to trouble and therefore in need of support. (iii) Originating in trouble and disturbance, epistemology and philosophy of truth are definitely normative projects. The distinction between successful and failing cognitive acts of all kinds constitutes both starting point and end of the epistemological project.

B Basic distinction: acts of truth and reflection on truth In a first step, I present a simple proof serving as an illustration of the first intuition to be presented and further developed (1). By delivering a comment on this proof, essential parts of the terminology of the underlying conception are introduced (2). Then, I establish the difference and the connection between alethic acts and alethiological reflection. Against this background, it makes good sense to analyze some features of the literal truth-talk and to correct some widespread errors (3). 1 An example: the

asymmetry-proof

Let us suppose a fragment of a physical language which is first order and which, besides the usual logical operators, contains some two-place predicates like ".. is heavier than ..". We may then, at a certain stage of the development of our physical knowledge, assert that if a body is heavier than another, the reverse does not hold. We establish this claim by delivering the argument under [I], which is, for ease of reference, provided with an obvious label. [1] The 0

asymmetry-proof

It holds

1 Suppose 2 Suppose 2 3 Since

For all x, y than x) u is heavier w is heavier For all x, y,

z(If x is heavier than y then y is not heavier than w than u z (If x is heavier than and y is heavier than

z, then x is heavier than z)

Copyrighted

Material

44 4 5 6 7 8 10 11 13 14

Geo

Siegwart

Therefore

For all y,z (If u is heavier than y and y is heavier than z, then u is heavier t h a n z) Therefore For all z (If u is heavier than w and w is heavier than z, then u is heavier t h a n z) Therefore If u is heavier than w and w is heavier than u, then u is heavier than u Therefore) u is heavier than w and w is heavier than u u is heavier Therefore Since than u For all z (z is not heavier than z) Therefore u is not heavier than u Therefore, w is not heavier than u I2 ThereforeIf u is heavier than w, then w is not heavier than u For all y (If u is heavier than y, then y is not heavier than u) Therefore For all x. y (If .x is heavier than y, then y is not heavier Therefore than .x)

Using a suitable meta-language, one can say that the asymmetry is d e m o n s t r a t e d by recourse to two other structural properties of the heavierrelation, transitivity and irreflexivity. Being well known and trivial, the proof itself does not need any particular attention: hence it may serve as an ideal example for several purposes - The speech acts that are combined in the proof under [1] are performed in an elementary physical language. The following considerations always presuppose a reference to a language: All remarks I will make are to be completed by the phrase "in such-and-such a language."

2 A conceptual framework

of the philosophy

of language and

knowledge

The initial move in line 0 consists in the assertion that if an arbitrary entity is heavier than another arbitrary entity, then the reverse does not hold. The illocutionary operator, in short, the performator, "It holds " indicates the illocutionary force of asserting: the remaining sequence of symbols forms the proposition for which truth is claimed. the thesis. The unit consisting of the assertion-performator and the asserted proposition is the assertion-sentence. By uttering an assertion-sentence. one performs an assertion. T h e truth claim of line 0 is redeemed by the performance of 14 illocutionary acts: Two of these are assumptions. two acts are adducing-as-reasons (the truth of the adduced propositions has already been established) and ten acts are inferences. Continuing the presented way of speaking, one has to say: One performs acts of assumption by uttering assumption-sentences; one p e r f o r m s the a d d u c i n g of reasons by uttering a d d u c i n g - a s - r e a s o n sentences; one performs inferences by uttering inference-sentences. The sentences consist of an a s s u m p t i o n - p e r f o r m a t o r , an adducing-as-reasonp e r f o r m a t o r , an inference-performator, and of the propositions that are assumed, adduced as reason and inferred respectively.

Copyrighted

Material

A ler hie acts and alethiological

reflection

45

Acts of a s s u m p t i o n , indicated by " S u p p o s e ", are to be f o u n d in lines 1 and 2. Acts of adducing-as-reason, indicated by "Since ", are performed in lines 3 and 9. The other acts are acts of inference, indicated by "Therefore ". The propositions that are adduced as reasons say something for the thesis, they support the thesis, as one can draw conclusions f r o m them in favor of the thesis; eventually one can infer the thesis itself. The thesis appears twice: it opens the argument as a claim and finishes the reasoning as a conclusion. Because grounds serve to draw conclusions from them, they are taken as premises. But not all premises are reasons because one can also draw conclusions from propositions that have been inferred or assumed. 1 generalize and complete the sketched conceptual framework: We perform illocutionary acts by uttering suitable sentences consisting of a p e r f o r m a t o r E and a proposition Á. Seen f r o m the perspective of logical grammar, a sentence is built by applying the p e r f o r m a t o r S to the proposition Á. Acts of cognition, and therefore also acts of truth, can be depicted as illocutionary acts. Reversely, it is obvious that not every illocutionary act is an act of cognition. Performing an illocutionary act, we use all subexpressions n of the uttered sentence I , i.e. the p e r f o r m a t o r E and the subexpressions of the proposition A. It is to be recognized that partial acts are embedded in the illocutionary act; they consist in using certain subexpressions. Such partial acts are, a m o n g others, reference, i.e. the use of closed terms, and predication, i.e. the use of predicates. Uttering the sentence "I d o u b t that Socrates is cool", I carry out the act of doubting. Doing this, I use the expressions "I doubt ", ".. is cool" and "Socrates". Using the predicate ".. is cool" I carry out the act of predication; using the proper name "Socrates", I carry out the act of reference. Cognitive acts can be combined into sequences of such acts, into discourses. Arguments, proofs, making something plausible, descriptions, hypothetical considerations, refutations, falsifications, disputes, and explanations are types of discourses; in our case the kinds of discourses which deserve special attention are proofs and arguments, i.e. the kinds of discourses establishing truth.

.? Aletliic acts and the accompanying

alethiological

reflection

Equipped with the example given under [1] and the sketched conceptual framework, I take up the issue of alethic action and alethiological reflection in order to illustrate their differences and to show how they are connected. First: the members of communities in which a cognitive language is used constantly perform acts of truth by advancing assertions, denying something, stating something, setting principles and so on. A truth-predicate is neither necessary nor suitable to answer these purposes. A truth-performator, however, is necessary and suitable to accomplish the task of alethic action.

Copyrighted

Material

46

Geo

Siegwart

In the case of our example, the p e r f o r m a t o r "it holds " fulfils this task. Second: the example also shows that one does not need literal truth-talk in order to perform alethic acts. Although one could perform the act of assertion using synonym p e r f o r m a t o r s such as "I assert as true___ " or simply "It is true that ___", it is to be recorded that literal truth-talk is not necessary for this purpose. Inspecting natural languages, we see that speakers not only d o not use literal truth-talk, but do often not indicate the illocutionary force by explicit p e r f o r m a t o r s either, so that we have to reconstruct the performative status f r o m the context in which a sentence is uttered. Third: stating that uttering "Socrates is cool" with assertoric force and "It is true that Socrates is cool" are equivalent with respect to their communicative status, we may not infer the general redundancy of truthpredicates and of the problem of truth. We have to remember that some x is always r e d u n d a n t relative to some purpose v. The correct observation that we can p e r f o r m alethic acts without using a literal t r u t h - p e r f o r m a t o r alone does not permit any inference concerning the purposes that are pursued using truth-predicates and cognates! Fourth: one does not need a truthpredicate before one wants to make alethic action the subject of reflection. Whereas all a u t h o r s of a language perform alethic acts, the reflection on the conditions, criteria, etc. of this alethic practice is a matter for experts, who concentrate on providing and using means of speech that realize that aim; predicates like ".. is t r u e " and ".. is false" belong to this vocabulary. The focus on alethiological reflection does not exclude alethiologists from performing alethic acts - quite on the contrary! T h e difference between alethic acts and alethiological reflection and their connection have both caused much confusion. A list of examples will illustrate differences and relations. [2] Alethic acts and alethiological a) Presumably b) It holds c) Is it true that d) It is false that

reflection:

examples

Socrates is cool for all x, y (If x is heavier than y then r is not heavier than x) some consequences of false propositions are true all true propositions are refutable

Someone who utters a) expresses a presumption; t h u s he p e r f o r m s a cognitive act. He does not perform an alethic act nor does he contribute to alethiological reflection. Someone who utters b) p e r f o r m s an alethic act without using literal truth-talk. However. he does not contribute to the reflection on truth, but to the inquiry of the heavier-relation. Uttering c), one poses a question concerning matters of truth. Still, it is not an alethic, but an interrogative act. On the other hand, it is a contribution to the reflection on truth. Notice a particularity: The posing of the question contains literal truth-talk. Finally. by uttering d), one performs a negative

Copyrighted

Material

Alethic acts and alethiological

reflection

47

alethic act, a denial; a predication of truth is embedded. By presenting a proposition as false, a contribution to the inquiry of truth is made, articulating a relation between true and refutable propositions. Besides, the insight that literal truth-talk is not necessary to perform alethic acts and the possibility to perform other cognitive acts using literal truth-talk both point to the fact that focusing only on literal truth-talk, one does not take into account all alethic acts nor are all accounted acts alethic acts. In short, the focus on literal truth-talk yields an incorrect and incomplete result - if one wants to capture the alethic acts. The relation between alethic acts and alethiological reflection is completely analogous to the relation between the practice of inferring and the theory of inferring or the accompanying logical reflection: We constantly draw conclusions, indicated by words like "therefore", "hence", "so", "thus", "consequently", etc.; in order to perform these acts, consequencepredicates are neither necessary nor suitable. However, when a logician, following his or her special interest, takes the practice of inference as the subject of reflection, he or she must have recourse to the use of consequence-predicates. In both cases the purpose is the same: T h e alethiologist or the logician wants to m a p the possibilities of alethic and inferential acts as such and in their relation to other cognitive acts. 2

C Territory: gnoseology und alethiology Alethic acts belong to the cognitive acts, as has already been said. In order to understand their function in the whole business of cognition, it is indispensable to obtain a general view of cognitive acts (4). Cognitive acts can be performed correctly or incorrectly. To establish this distinction, it is helpful to refer to the concept of rules (5). Finally, one has to look at the interplay of cognitive acts (6).

4 Acts of cognition and acts of truth: a map Alethic acts in their affirmative and negative forms constitute the perfect types of cognition; asserting, stating and setting-as-postulate are examples of affirmative alethic acts. Stronger qualifications t h a n these mentioned ones are not intended: Due to the connection of truth and meaning, stronger alethic acts are not construable (see D) - The perfect forms of cognition are to be distinguished from the imperfect forms. Examples of affirmative imperfect forms of cognitive acts are presumptions, estimations, forecasts and retrodictions. D o u b t i n g and hypothetically rejecting are negative imperfect forms of cognition. The imperfect and perfect forms together constitute the qualifying or substantia! acts of cognition. Apart from these are the subsidiary acts, cognitive auxiliary acts, such as, for example, inferring, assuming and a d d u cing-as-reason. Substantial and subsidiary cognitive acts together f o r m the

Copyrighted

Material

48

Geo

Siegwart

non-interrogative cognitive acts; they have to be distinguished f r o m the interrogative cognitive acts, i.e. the different sorts of questions. The survey under [3] maps the cognitive landscape. [3] The cognitive

landscape

cognitive acts

non-interrogative

interrogative

subsidiary

affirmative

negative

asserting

denying

doubting

inferring

stating

denying-bypostulate

hypotheticatly

assuming

rejecting

adducing-asreason

setting-aspostulate

wetherquestions whichquestions

retracting

defining consentingfirmly

hypothesis

5 Cognitive acts and alethic acts: the normative

aspect

Cognitive acts and alethic acts are to be considered concerning the aspect of success and failure, i.e. the aspect of (in)correctness. Let us not forget the birth of epistemological reflection out of the failure of cognitive acts (see A)! Someone who poses the question why the earth is a square-stone asks incorrectly, as he violates a prerequisite of succeeding why-questions. Someone who simply d o u b t s and is unable to present any indication just performs an incorrect attempt to d o u b t . Someone who asserts without having reasons just p e r f o r m s a simulatio affirmationis. At this point, it is advisable to make use of the concept of rules. G e n erally, rules specify roughly speaking - on what conditions which actions are permitted, required or forbidden. Rules of cognition specify on what conditions we may or should perform what kind of cognitive act. They characterize the respective procedure, m e t h o d or n o r m . Some examples may serve as an illustration. [4] Examples

of rules of

cognition:

Copyrighted

Material

Alethic acts and alethiological

reflection

49

a) If a proposition Г is the consequence of a set of propositions which have been accepted by a disputant Si, and the co-disputant S 2 asks for a concession of Г, then Sj is b o u n d to concede �. b) If there is evidence for/against the plausibility of a proposition �, then � may be accepted as plausible/� may be doubted. c) If a universal proposition has been derived in dependence on a class of propositions, then each of its instances may be inferred in dependence on this class. d) If there is a proof for a proposition A/for the negation of A, then A may be asserted/A may be denied. e) If a formula-schema serves to achieve the aims that we pursue a n d does not generate inconsistency, then every instance � of this schema may be posited as an axiom. f) If one places two objects a and b each upon a pan of an evenly balanced beam scale, and the pan carrying a sinks lower than the pan carrying b, then an atomic proposition of the kind "a is heavier than b " may be stated. g) If a person c is able to solve such-and-such a test within a certain period, then one may state an atomic proposition of the kind "c is an intelligent person". Example a) presents a rule for disputing. The rules under b) guide an imperfect affirmative or negative cognitive act respectively. Rule c) gives guidance for inferring; therefore it belongs to the rules for subsidiary cognitive acts. The remaining rules are alethic rules: d) regulates asserting or denying respectively, e) guides the positing-as-axiom, whereas f) and g) direct stating. T h e alethic rules represent exactly the so-called "truth criteria" that have been discussed within a broad and respected tradition of m o d e r n epistemology. Four facts support this thesis. First: many alethic rules d e m o n strate the plurality and variety of the cognitive business. T h o s e belonging to arithmetic may and should be different f r o m those ruling vampirology, theology, mineralogy and measuring: the rules for a special kind of propositions have to account for its particularity. Second: alethic rules direct the actions of an alethically acting author. Nevertheless, it is not necessary for an a u t h o r to k n o w the rules; he or she must just be able to d o what the criterion demands. A n a u t h o r has to be able to act in accordance with the rules, but he or she does not have to be able to formulate or even justify the respective rules. Third: alethic rules serve to enable the ex-post j u d g m e n t of p e r f o r m e d acts or achieved results. Ex-post j u d g m e n t is particularly advisable when d o u b t s whether or not a p e r f o r m a n c e is in accordance with the rules exist. Fourth: alethic rules and the establishment of a t r u t h predicate in m a n y cases it will be a definition - have to be arranged in such a way that the following holds: If a proposition A fulfils the antecedent of an alethic rule, then A is true. Speaking in traditional speech:

Copyrighted

Material

50

Geo

Siegwart

propositions that are true according to a criterion are also true according to the definition. 3

6 Orchestra of cognition Inspecting the cognitive acts further, one has to state that they function cooperatively t h r o u g h o u t . In particular. the alethic acts work together with the other cognitive acts. An image may illustrate the cooperative interplay: the various acts form an ensemble, an orchestra, while the rules represent the score, providing the guidance for the orchestra's performances. The cooperative relations in the cognitive d o m a i n become particularly obvious in disputes, sequences of questions and answers, and in processes of confirmation and corroboration. The asymmetry-proof under [1] will now be used to give a representative account of the way in which the cognitive score can be played. The rule of assertion d e m a n d s a proof, i.e. the p e r f o r m a n c e of assumptions, adducingas-reasons and inferences. The rules of inference are formulated in such a way that one can make inferences from propositions that have been assumed, adduced as reasons, or concluded. The rule for adducing-asreason permits the recourse to alethic acts that have been executed before and finally to start-setting truths. Start-setting acts are, for example, settingas-axiom or acts of stating. They are performed in accordance with a rule for setting-as-axiom or with rules for stating (see section 5, [4]e)-g), 9f). Definitions may also be adduced as reasons in proofs, as definitory acts comply with rules that guarantee the eliminability of the defined term. T h e sketched m a p and the remarks concerning the score, the space of cognitive rules, give at least a first impression of the whole landscape of cognition. They may prevent one-sided views; in particular they should prevent unjustified emphasizing of particular cognitive acts or sequences of acts. Having pointed to the interplay of cognitive acts, I can now correct two errors concerning assertion that can often be found: (i) Assertion represents one alethic act, more precisely, one affirmative act, but it is not the alethic act pure and simple! (ii) Besides, assertion is in a sense a secondary act, as its p e r f o r m a n c e d e m a n d s the execution of other acts: in order to assert a proposition correctly, one has to p e r f o r m correct acts of inference and of adducing-as-reason; in order to perform correct adducing-asreasons, one has ultimately to execute correct start-setting alethic acts. Formulated for the rule side this means: the concept of proof, which plays an essential role in the rule of assertion, is defined by the rules of inference and the rules of start-setting. These rules are, in this respect, more f u n d a m e n t a l than the rule of assertion. Viewing the cognitive landscape in its whole fosters or facilitates the localization and organization of some epistemological offers. Again, an example may serve as an illustration: the substantial cognitive acts are performances which qualify the respective propositions epistemically. By imperfect

Copyrighted

Material

Alethic acts and alethiological

reflection

51

cognitive acts, propositions are given some (positive or negative) epistemic weight, whereas they are qualified as definitely true or false by perfect cognitive acts. Inspecting the offers that carry the label of a "coherence theory of t r u t h " in this respect, one has to state that they often contribute to the extensive territory of the plausible, hypothetical, presumed, and estimated, but not to the perfect forms of understanding. 4

D T h e nexus: truth and meaning I will now come to the core of my consideration, the nexus of truth and meaning. I begin with a further comment on the asymmetry-proof, this time from an alethic-semantical perspective (7). Then, I discuss three kinds of interplay between truth and meaning (8-10). Finally, their general connection is formulated and the central task of a philosophy of truth is pointed out (11).

7 An alethic—semantical reading of the

asymmetry-proof

The proof establishes the truth of the asserted proposition - and it does this along the meaning of the proposition. Generally, it holds: The alethic status of a proposition results, along its grammatical structure, from the meaning of its subexpressions; that programmatic statement has to be demonstrated in detail on the well-known example. It is a universal proposition that has to be proven. According to the meaning of "all", i.e. the universal quantifier, which is the main o p e r a t o r of the proposition, a proposition of this kind is proven when the asserted conditions have been proven for arbitrary but fixed objects. With regard to the asymmetry-proof, this means that the proposition in line 12 is to be proven: If u is heavier than w, then w is not heavier than u. The conclusions in lines 13 and 14 are then justified by the rules for the introduction of the universal quantifier. The new candidate for proof is just the conditional. In accordance with the underlying meaning of the "if-then", being the m a i n operator, the proposition is proven if one has derived the consequent in dependence on the antecedent. If this has been done, one can infer the conditional in question, as the corresponding rule for the introduction of the conditional horizontalizes the vertical dependence. With regard to this step of the proof, which is to be found in line 12, one has to assume the antecedent, which is done in line 1. The rule of assumption is, with good reason, very liberal: one may assume any proposition! The proposition that has to be derived in dependence on the assumption in line 1 is a negation. According to the meaning of the negator, which is the main operator, one is allowed to infer the negation of a proposition in a situation of discursive trouble, i.e. f r o m a contradiction, provided the proposition itself is involved in this trouble. To prepare this step, the proposition we have to negate is assumed in line 2: suppose that w is heavier than u. The remaining task is to generate the trouble.

Copyrighted

Material

52

Geo

Siegwart

T h e considerations presented up to this point result exclusively f r o m the m e a n i n g of the logical o p e r a t o r s that play a role in the thesis. T h e starting-point of the proof consists of two atomic propositions with the same main operator: the heavier-predicate. Its meaning is two-fold: The synthetic-operational part is fixed by the rules of stating, its analytic-structural part is fixed by meaning-postulates. In the correlation that is examined here, only the meaning-postulates are relevant. We adduce them as reasons in line 3 and line 9. T h e proposition adduced in line 9 can be instantiated, according to the meaning of the universal quantifier. In the context of this proof, it is sensible to instantiate with respect to one of the objects considered, i.e. u. T h a t happens in line 10. As we want to derive a contradiction involving the instantiated proposition "u is not heavier than u", we now have to infer "u is heavier than u " in dependence on the appropriate propositions. In accordance with the meaning of " a n d " , i.e. the conjunctor, we join the results of the two assumptions in line 7: This step is legitimated by the rule for the introduction of the conjunctor. Again in accordance with the meaning of the universal quantifier, suitable instantiations of the proposition adduced in line 3 are made in lines 4, 5 and 6. According to the meaning of the "if-then", m o r e specifically, by means of the rule for the elimination of the conditional (i.e. m o d u s ponens), one can then infer f r o m lines 6 and 7 that u is heavier than u. Lines 8 and 10 present the m e m b e r s of the contradiction. In accordance with the meaning of the negator, more specifically with the rule for the introduction of the negator (i.e. the rule of indirect proof), one can infer the negation of the assumption of the indirect proof (noted in line 2) in line 11. T h e last three lines, which present the introduction of the conditional and of the universal quantifier respectively, have already been commented. T h e following connection is crucial: the truth-qualifying procedure, which, in the case under consideration, is a proof, is carried out in accordance with the meaning of the proposition that is to be shown to be true, i.e. in accordance with the meanings of the subexpressions of the proposition as they are combined in a specific structure. T h e examined assertion presents a correct performance of an alethic act; with respect to alethiological comments, truth-predicates have to be established in such a way that they apply to the examined proposition. F u r t h e r more, every expression of the assertion sentence is used correctly. This also applies to the sentences of the 14 steps of the proof.

8 Inferential truth and meaning The assertion rule d e m a n d s proofs or arguments. As has already been stated in Section 7 these receive their structure by the rules for inference and by means of the rule for adducing-as-reason. and finally by the start-setting rules. Each of the steps of the discussed example is correct because a corresponding rule applies to it.

Copyrighted

Material

Alethic acts and alethiological

reflection

53

The rules for inferring (examined here) fix the meaning of the logical operators together with the meaning of the inference-performator. They d o this by specifying in what kind of discursive state what inference is allowed, i.e. they specify on what conditions one may use the inference-performator and the respective logical operators. Formulating in a n o t h e r terminology, one should say that the illocutionary meaning, i.e. the usage of the performator, and the meaning of the logical operators, i.e. the propositional meaning, are fixed eodem actu. The establishment of meaning follows a well-known material idea of systematization: a rule of introduction and a rule of elimination are formulated for each logical operator. The first determines the discursive circumstances in which one is allowed to infer a proposition that contains the respective logical operator as its main operator. The latter specifies which propositions one is allowed to infer f r o m a proposition containing the respective operator as its main operator. The most simple (but not totally harmless) case is that f r o m two propositions, one is allowed to infer their conjunction; this is the introduction rule for the conjunctor. F r o m a given conjunction, one may infer each of its conjuncts; this is the elimination rule for the conjunctor. The example shows in an elementary fashion that the meanings of the logical constants, i.e. the conclusion paths given with the respective rules, are also ways to establish the truth of a proposition. The logical operators contained in a proposition determine the path of a proof according to their place in the structure of the proposition up to the point at which meanings of non-logical terms come into play. Logically true or false are those propositions whose alethic status depends on the meaning of the logical operators alone. The meaning of the inference-performator a n d the logical operators is established by the rules of inference. As a consequence, the alethic status of logically determined propositions depends exclusively on the rules of inference, whereas the alethic status of logically indetermined molecular propositions results only partly f r o m the logical rules. Some historical and systematic remarks may be advisable in order to prevent misunderstandings. (i) The offered approach connects three wellknown philosophical conceptions: the theory of speech acts, the use-theory of meaning and the conception of natural deduction. The rules for the speech act of inference are also rules for the use of the logical operators, and for each logical operator there is exactly one introduction and one elimination rule. 5 (ii) However, the connection between speech act theory and the use-theory of meaning which is characteristic of the whole presented a p p r o a c h is not necessarily b o u n d to the idea of introduction and elimination in natural deduction, but can also be realized in axiomatic conceptions of logic. Further, the a p p r o a c h is extendable by rules of material inferences. (iii) The presented understanding of inference sets a wide f r a m e w o r k . It is left completely open in which way one fills this framework in a special case, i.e. it is open if (for these or those purposes) a minimal, intuitionistic, classical, relevant, paraconsistent etc. logic is preferred. This

Copyrighted

Material

54

Geo

Siegwart

will be a result of the specific material principles accepted for the justification of meaning-design. 6

9 Analytic-structural

truth and meaning

In the exemplary proof we refer to two axioms or meaning-postulates that fix the heavier-predicate in its analytic-structural meaning. In line 3 the transitivity of the heavier-relation is adduced as reason, in line 9 its irreflexivity. Within the supposed structure of a physical language, they are set as categorically true and represent bases of proofs, insofar as they are propositions that need not to be proven. As the handling of this kind of propositions is well regulated, the f u r t h e r use of the heavier-predicate, a basic predicate of our physical language, is also well regulated. M o r e precisely: the analytic-structural part of the meaning of the predicate is fixed. T h e establishment of a beginning by m e a n s of the propositions coincides with the establishment of meaning for the basic concepts. Besides, the language of o u r life-world is - adequately reconstructed full of such postulates! It has to be stressed that these analytic-structural beginnings are neither dogmatic nor arbitrary. First, they have to satisfy the postulate of consistency and more generally, they have to fulfill the d e m a n d of non-trivializability. Second, we aim to realize our purposes of speech: so the beginnings should also be suitable. The formulation of the rule of setting-as-axiom presented as an example above expresses this idea (see section 5, [4]e). Any more specific formulation results from the particular aims of speech and cognition. The idea that guides the justification of the axiomatic setting may be illustrated by a f a m o u s - perhaps the most f a m o u s example in the history of mathematics. First of all, we notice that many relations, apart f r o m the heavier-relation, possess the attributes of transitivity, irreflexivity, and asymmetry. As there exist general correlations between these attributes, and as one does not want to prove these in each single case, the following procedure is advisable: one proves once that irreflexive and transitive relations are asymmetric; having d o n e this proof, one does not have to prove this correlation again for the heavier-relation, the more-intelligent-relation, the greater-relation, etc. In order to realize this general idea adequate possibilities of speech have to be created first. This aim can be achieved in different ways. One can switch to a secondorder-language, or one can provide possibilities of speaking a meta-language. A particularly convenient method is the widening of a first-order-language by two axioms. The first axiom, the f a m o u s comprehension schema, can be roughly formulated in the following way: for each predicate, there is a set that contains exactly those entities as elements to which the predicate applies. T h e axiom of extensionality d e m a n d s the identity of entities that have exactly the same elements; the basic concept fixed by these two axioms is the predicate ".. is an element o f . . Both postulates assure that there is

Copyrighted

Material

Alethic acts and alethiological

reflection

55

exactly one entity that contains all and only those elements to which a predicate applies. T h a t allows us - after having provided variable-binding and term-generating operators to speak, for example, of the class of all red entities or, again with suitable amendments, to speak of the relation of all x, y so that .x is heavier than y. Taking these postulates as a basis, one could construct mathematics in most of its parts. Since the unrestricted schema of comprehension is trivializable a well-known fact - the setting of the schema violates a d e m a n d guiding the setting of axioms. T h e mostly used set-theoretical systems as ZF, N B G or N B G U are justified until f u r t h e r notice: they block the known ways to trivialization by weakening the postulate of comprehension, a n d they preserve the efficiency of the system by adding f u r t h e r axioms. Because we can achieve our aims by speaking these languages, it would not be sensible to d e m a n d more justification!

10 Synthetic-operational

truth and meaning

The rule concerning the heavier-predicate (see section 5, [4]f) connects the statement that one of two bodies is heavier than the other to the use of a balanced beam scale. The rule for the intelligent-predicate (see section 5, [4]g) d e m a n d s a corresponding test before one is allowed to make a statement about the intelligence of a person. These rules represent rules of cognition. M o r e specifically: the rules guide an affirmative form of perfect cognition, they represent rules for acts of truth, criteria of truth or, the last specification, rules of stating. The if-part of the rules describes what one has to do in order to state propositions that contain the predicates ".. is heavier than .." or ".. is intelligent" as their main operators. Therefore they specify a truth-qualifying procedure for propositions containing these predicates as their main operators. Since they specify at the same time on what conditions one is allowed to apply these predicates, one can also designate these rules as rules of predication. Such rules partly determine both the use of the stating-performator and the use of such predicates. Again, illocutionary and propositional meanings are fixed at once; at the same time the conditions for the truth of the propositions become obvious. For the examples already mentioned and for all similar ones it holds that the fixing of meaning and the establishment of the truth-qualifying procedure are two sides of one and the same coin, two ways of grasping a rule, two ways to illustrate its function. With respect to the execution holds: An alethic act is performed correctly, the execution of a truth-qualifying procedure is done correctly, if it is performed in accordance with the meaning, i.e. in accordance with the respective rule. Stating is the prototype of a non-discursive, of an empirical speech act. Propositions that can be stated are empirems. By the act of adducing-as-reason, they could be incorporated in any discourse in the same way as axioms and propositions that have already been proved. Stating demands further feeding

Copyrighted

Material

56

Geo

Siegwart

acts that are different from speech acts in most cases; the act of weighing in the mentioned rule illustrates this fact. Insofar as these acts lead to a speech act, the regulated interplay of hand and m o u t h is essential. In the case of empirical cognition, meaning and truth are linked to the actions of a u t h o r s "in the world." Such acts can be of any kind, for example they could, in the simplest cases, even consist in acts of perception. Subsequently, the use of measuring instruments and other instruments of cognition will be integrated. Such instruments - for example the b e a m scale present themselves as instruments of truth and meaning. So, reflection on m a n u f a c t u r i n g and structuring instruments belongs to the core of that d o m a i n of alethiology which refers to the empirical. Likewise, someone who plans, produces or maintains such an aletho-semantical instrument, enables alethic acts. 7 Someone who wants to elaborate on the deeply correspondistic or realistic idea that o u r understanding is "committed to reality" and that reality is "authoritative" for our alethic acts takes the right direction by constructing, reconstructing and analyzing the rules of stating. They are the means by which we help the world - in our own interest to determine the alethic status of propositions. In that field, friends of correspondistic truth conceptions may watch how a d e q u a t e regulations of use establish the "connection between language and reality." 8

11 The shared origin of truth and meaning Boldly extrapolating from my sketched considerations, I formulate in general, at first negatively: one c a n n o t characterize the truth of a proposition without referring to the meaning of the expressions that are combined in this proposition; reversely, one c a n n o t establish meaning without determining the alethic status of the propositions that contain the respective terms. Positively: by establishing the meaning of words, one determines, directly or indirectly, modes of alethic performance, establishes truth criteria, truthqualifying procedures. By organizing the alethic performance, one determines the meaning of words. Only this connection of truth and meaning provides answers to three why-questions. First: why is it required, particularly in case of dissension, to eliminate ambiguities, to explicate, to explain, to make the respective terms m o r e precise, i.e. why is it required to fix the meaning of a term or to improve a determination that has already been given? - Answer: Because only the fixing of a term gives the o p p o r t u n i t y to perform alethic acts correctly and therefore enables the creation of an approved store of knowledge. Reversely, if the establishment of meaning were without consequences for the alethic status, it could be omitted without loss. Second: wanting to determine the alethic status of a proposition, for example if a is heavier than b, why does one not ask the stars, or go to a specialist for the reading of coffee grounds or for the flight of the birds or

Copyrighted

Material

Alethic acts and alethiological

reflection

57

look for a n o t h e r Dadaistic option in the widespread d o m a i n where anything goes? - Answer: Because the offered procedures are in no way connected with the meaning and the truth-qualifying procedure of the respective proposition. To decide if a is heavier than b, one has to handle a balanced beam scale - other m e t h o d s of weighing are obviously possible but not regarded here. To ask if a is heavier than b just means to ask for the result of the weighing procedure. Third: why perform imperfect forms of cognition, for example forecasts, retrodictions, presumptions and hypotheses at all? - Answer: in response to the question what the temperature will be in Munich in four days' time, we can only give a forecast now, but we cannot state it. This perfect form requires - in accordance with the meaning of the proposition - an act of measurement that we cannot perform today, but only in four days. To answer the question how many blossoms my cherry-tree had last year, we can only give a retrodiction, but no stating. This would require - in accordance with the meaning of the proposition in question - the correct counting of the blossoms of my cherrytree last year. The impossibility of always performing the truth-qualifying procedures required by the meanings leads to the establishment of forecast, retrodiction, and m a n y other forms of imperfect cognition. Against the b a c k g r o u n d of the outlined organization of truth and meaning, the central task of alethiology and thereby theoretical philosophy can finally be formulated: How should we judge, reconstruct and, if necessary, revise the determinations of meaning and, ipso facto, the truth-qualifying procedures that already exist? How should we organize and justify new determinations?

Notes 1 Dedicated to Peter Hinst, teacher and friend, on the occasion of his 70th birthday. 2 For details cf. Siegwart (1997: Chapter F); concerning the truth-talk in our everyday language and the superfluity theories of truth cf. Siegwart (1997: especially §28). 3 Siegwart (1997: Part G) offers a comprehensive approach to the connections between alethic rules, truth criteria, truth conditions and definitions of truth. 4 Rescher (1973) presents good evidence for that presumption; for a detailed argumentation see Siegwart (1997: 490-502). Epistemological conceptions that favor a decision-theoretic approach can usually be interpreted as contributions to the imperfect forms of cognition as well. 5 Hinst (1982) presents this synthesis in detail and formulates the logical rules of a classical calculus. As far as I can see, the relation between truth and meaning in terms of speech act theory and use-theory of meaning is developed systematically for the first time in Hinst (1974). 6 See for the issues touched upon in (i) and (ii) Siegwart (1997: 174-84). Tennant (1997: especially Chapter 10) is an example of a well worked out justification of a particular logic. 7 The role of aletho-semantical instruments is stressed by constructivists of the "Erlangen School" ("Erlanger Schule"); see, for example, Janich (1997: especially Parts 2, 6 and 11).

Copyrighted

Material

58

Geo

Siegwart

8 This is not the only intuition found in theories of correspondence that is to be accounted for in every comprehensive truth conception. Another one consists in the (unjustly) so called platitudes of correspondence, e.g. "LA proposition is true if and only if the described state of affairs is a fact" or "A predicate applies to an object if and only if the object possesses the property expressed by the predicate". Within the framework of a constructive alethiology such statements should be theorems. Demanding them is due to the interest of speaking invariantly with respect to some equivalences: hence the so called definition by abstraction is the appropriate strategy of introducing expressions like "state of affairs", "property", etc.; see Siegwart (1997: 328-32), Meggle and Siegwart (1996: 984ff.).

References Hinst, P. (1974) Wahrheit unci Bedeutung. Vorschliige zu einem fundamentalsemantischen Aufbau von Wissenschaftsspraehen. unpublished habilitation. Munich. (1982) "Pragmatische Regeln des logischen Argumentierens," in C. F. Gethmann (ed.) Logik und Pragmatik, Frankfurt/Main: Suhrkamp, pp. 199 215. Janich, P. (1997) Das Maβ der Dinge. Protophysik von Raum. Zeit und Materie, Frankfurt/Main: Suhrkamp. Meggle, G. and Siegwart, G. (1996) "Der Streit um Bedeutungstheorien." in M. Dascal, and K. Lorenz (eds) Philosophy of Language: An International Handbook of Contemporary Research, Berlin: De Gruyter. pp. 964-88. Rescher, N. (1973) The Coherence Theory of Truth. Washington DC: University Press of America. Siegwart, G. (1997) Vorfragen zur Wahrheit. Ein Traktat üher kognitive Sprachen, Munich: Oldenbourg. Tennant, N. (1997) The Taming of the True. Oxford: Clarendon Press.

Copyrighted

Material

Part II

Truth and assertion

Copyrighted Material

Copyrighted Material

4

The use of force against deflationism Assertion and truth Dorit Bar-On and Keith Simmons

Deflationists share a core negative claim that truth is not a genuine, substantive property. Deflationism can be seen in p a r t as a form of eliminativism: we can eliminate the property of truth f r o m o u r ontological inventory. This is the distinctive claim of what we will call metaphysical deflationism. But a n y o n e w h o accepts metaphysical deflationism must still m a k e sense of o u r pervasive truth talk. W h a t is it we are d o i n g when we call something true, if we are not ascribing a genuine property? W h a t is the m e a n i n g of the word " t r u e " ? W h a t are its main uses or functions? A n d how should we u n d e r s t a n d the concept of t r u t h a n d the role it plays in b o t h o r d i n a r y and philosophical discourse? An acceptable deflationism must supplement the negative metaphysical claim with an a c c o u n t of the word " t r u e " as well as an account of our concept of t r u t h . In what follows, we wish to keep separate three deflationary claims that, it seems t o us, have been run together in the literature. T h e r e is first the metaphysical claim c o m m o n to all deflationists that t r u t h is not a genuine property. Second, there is the particular account of various everyday uses of the word " t r u e " (or the phrase "is true"). Deflationary a c c o u n t s of " t r u e " vary widely: the c o m m o n metaphysical core fans out into disquotationalism, minimalism, the r e d u n d a n c y theory, the prosentential theory, a n d more. Call this linguistic deflationism. T h i r d , there is what we will call conceptual deflationism: the claim that an acceptable deflationary a c c o u n t of " t r u e " will give us all there is to k n o w or u n d e r s t a n d a b o u t the concept of t r u t h a n d its potential e x p l a n a t o r y role. A t h o r o u g h g o i n g deflationary a c c o u n t of t r u t h will go beyond the negative metaphysical claim a n d the positive linguistic a c c o u n t of " t r u e " : it will also maintain that the concept of t r u t h is a " t h i n " concept that can play no substantive explanatory role in o u r conceptual scheme. As recently presented by Williams (1999), deflationists share the idea that the f u n c t i o n of truth talk is "wholly expressive, never explanatory": t r u t h talk only serves to allow us " t o endorse or reject sentences (or p r o p o sitions) t h a t we c a n n o t simply assert." "[W]hat makes deflationary views deflationary," Williams claims, "is their insistence that the i m p o r t a n c e of t r u t h talk is exhausted by its expressive f u n c t i o n " . 1

Copyrighted Material

62

Dor it Bar-On and Keith

Simmons

Deflationists a n d their o p p o n e n t s often seem to take it that conceptual deflationism falls out of metaphysical a n d linguistic deflationism. If metaphysical deflationism is right, then we c a n n o t assign the property of being true as the meaning of "true," and we c a n n o t explain our understanding of " t r u e " as consisting in the apprehension of this property. N o w suppose we accept this or that deflationary account of "true." Shouldn't we further accept that the account exhausts our understanding of the concept of truth? But if so, then in explicating any seemingly truth-related concept, we can appeal to nothing m o r e t h a n the preferred deflationary account of "true." T h e concept of truth can thus bear no substantive conceptual connections to other concepts of interest to us, such as meaning, validity, belief, assertion, verification, explanation, practical success, and so on. We will thus be left with a deflated concept of t r u t h - a " t h i n " concept whose u n d e r s t a n d i n g is exhausted by the deflationary account of "true," a concept that is isolated from all other concepts of interest to us a n d can play n o substantive explanatory role with respect to them. In this paper, we d o not plan to take metaphysical deflationism to task. We put to one side the deflationist claim that we stand to gain n o t h i n g f r o m appealing to a substantive property of t r u t h . We also d o not plan to challenge linguistic deflationism by criticizing the success of particular deflationist treatments of "true." O u r main concern is with conceptual deflationism: the claim that our understanding of truth is fully exhausted by this or that particular deflationary account of "true," so that a deflated, "thin" concept of truth is all that we need in our conceptual scheme. In o u r view, even if it is granted that paradigmatic uses of "is t r u e " can be treated in a deflationary way, this does not show that we have no need for a richer concept of truth than is allowed by deflationary accounts. As against the conceptual deflationist, it can be argued that there are ineliminable connections between truth a n d other concepts. For example, it might be argued that an account of linguistic meaning c a n n o t be provided without invoking the notion of truth. 2 O u r main interest in this paper is in the illocutionary notion of assertion. We will argue that there is a conceptual link between truth a n d assertion that c a n n o t be broken, and that o u r u n d e r s t a n d i n g of the concept of assertion requires m o r e t h a n a deflated concept of truth. So m u c h the worse, then, for conceptual deflationism.

I Deflating truth As we have presented conceptual deflationism, it is committed to the claim that t r u t h cannot play any substantive explanatory role. If this is so, the deflationist ought to tell us what to m a k e of various apparent connections between truth and other concepts of interest to us. It has been claimed that the deflationist can readily endorse such connections. 3 To set the scene, let us consider two apparent platitudes that tie together the notions of truth, truth-aptness, assertion and belief. First, if a sentence is t r u t h - a p t - that is,

Copyrighted Material

Assertion

and truth

63

either true or false and 1 utter it sincerely, then I have m a d e an assertion. Second, "assertion has the following analytical tie to belief: if someone makes an assertion, and is supposed sincere, it follows that she has a belief whose content can be captured by means of the sentence used" (Wright 1992: 14). Consequently, there is an apparently platitudinous tie (perhaps an analytic one) between truth-aptness and belief: a sentence is truth-apt only if it can be used to give the content of a belief. F r o m a commonsensical point of view, this is just what we would expect. If I sincerely utter the truth-apt sentence "Aardvarks amble," it seems that I have asserted it; and it f u r t h e r seems that I have said what I believe about the world, where the content of my belief is given by the sentence "Aardvarks amble." According to Jackson et al. (1994), platitudinous conceptual connections must be preserved by any good analysis: when we d o conceptual analysis we have to respect platitudinous connections. They are the very stuff that conceptual analyses are m a d e from. In analyzing a concept our aim is to capture the network of platitudes that surround that concept, and so capture its meaning. (Jackson et al. : 294) Jackson et al. (1994: 296-7) think we can a d o p t a certain view of conceptual analysis that they regard as "minimal." The view is this: when we analyze a concept, the analysis should comprise all the platitudes about the concept and nothing more. Call this view platitude-respecting minimalism. The view is minimal in the sense that it supposedly requires us to make no controversial assumptions. In particular, a platitude-respecting minimalist about truth will embrace platitudinous conceptual ties between truth, truth-aptness, assertion and belief. These ties, though platitudinous, may be quite substantial, according to Jackson et al. (1994). For example, to show that a sentence is truth-apt, it needs to be shown "that the state an agent is in when she is disposed to utter a sentence . . . bears the relations to information, action and rationality required for the state to count as a belief. This is a substantial m a t t e r " (Jackson et al. 1994: 296). So the platitude-respecting minimalist will endorse substantive connections between truth and other concepts. If, as Jackson et al. (1994) assume, platitude-respecting minimalism is a version of deflationism, then we c a n n o t maintain the general claim that deflationists isolate truth f r o m other concepts. But platitude-respecting minimalism should not count as a genuine form of deflationism. Deflationists d o not take platitudes about truth as their starting point. Rather, deflationists are motivated by the thought that much of what is said about truth, about its nature and role and connections to other concepts, is radically misguided. Apparent platitudes about truth cannot be taken at face value. Instead, they are to be regarded with some suspicion - it may be these very platitudes that encourage the thought that truth is a substantive

Copyrighted Material

64

Dor it Bar-On and Keith

Simmons

notion. Indeed, it seems that the deflationist would reject the basic methodology of platitude-respecting minimalism, given its c o m m i t m e n t to incorporating everyday platitudes into the theoretical analysis of truth. The c o m m o n s e n s e platitudes can mislead us into thinking that truth is a substantive, explanatory notion. Yet, if conceptual deflationism is right, there is nothing m o r e to our understanding of truth than what is captured by a preferred deflationist account of "true." For a test case, consider the following apparent platitude: true beliefs engender successful action.4 On its face, this "fact a b o u t t r u t h " 5 seems to forge substantial links between truth, belief, and action. But according to the deflationist, this appearance is misleading: we need only a deflationary account of truth to explain the role of truth in this thesis. Consider H o r wich's (1990) account of such "facts a b o u t t r u t h . " The axioms of Horwich's minimal theory of truth are all and only the instances of the sentence schema: T h e proposition that p is true if and only if p, instances like " T h e proposition that penguins waddle is true if and only if penguins waddle." The denominalizing function of " t r u e " embodied in these axioms exhausts what there is to be said by way of explaining truth n o other notions enter into the theory. The minimal theory of truth is "a theory of truth that is a theory of nothing else."6 Moreover, it is a complete theory of truth - we are not to gain further understanding of truth by appeal to anything other t h a n the equivalences. In particular, alleged platitudes that make use of truth locutions c a n n o t be regarded as in any way enhancing our understanding of truth. By the same token, if we resort to truth-talk in our explication of other concepts, we c a n n o t expect the notion of truth to contribute to o u r understanding of these concepts beyond what is afforded by the minimal theory, since: "all of the facts whose expression involves the truth predicate may be explained . . . by assuming no m o r e a b o u t truth than instances of the equivalence schema" (Horwich 1990: 24). In particular, Horwich argues that this is so for the thesis that true beliefs engender successful action. Horwich considers the following instance: If all Bill wants is to have a beer, and he thinks that merely by n o d d i n g he will get one, then, if his belief is true, he will get what he wants. At one point in his explanation, Horwich makes "the familiar psychological a s s u m p t i o n " that if one has a desire, and believes that a certain action will satisfy that desire, one will p e r f o r m the action. 7 That is, conceptual connections are assumed between belief, desire, and action. But all that is assumed about truth in Horwich's explanation is its denominalizing role. In the course of the explanation, we move f r o m " T h e proposition that if Bill n o d s then Bill has a beer is t r u e " to "If Bill nods then Bill has a beer"; a n d a

Copyrighted Material

Assertion

and truth

65

little later we move f r o m "Bill has a beer" to "The proposition that Bill has a beer is true." These are the only steps where truth has a role to play, and it is the role given to it by the equivalence schema. This style of explanation, says Horwich, may be universalized to show how in general true beliefs lead to successful action. And beyond that, it extends to all other facts involving "true." The explanation of all these facts will, according to Horwich's minimalism, appeal to no more about truth than is given by the instances of the T-schema. In explaining these facts, we will not improve our grasp of truth, or deepen our understanding of it. N o r does the explanation of other concepts invoke anything m o r e than what we've described as a "thin" concept of truth - what is captured by the minimalist definition. We presumably learn more about, for example, the concepts of belief, desire and action by an improved u n d e r s t a n d i n g of their interrelations. But there will be no such improvement in the case of truth: the equivalence schema tells us all there is to know about truth, and it exhausts all that the notion of truth can c o n t r i b u t e to our understanding of any other concept. In this sense, truth is isolated from other concepts. This isolationism is not peculiar to Horwich's minimalism. Consider disquotationalism. According to the disquotationalist. there is no more to the truth of, say, the sentence "Aardvarks amble" than is given by the disquotation of its q u o t e name. One can think of the so-called T-sentence: "Aardvarks a m b l e " is true if and only if aardvarks amble as a partial definition of "true": the biconditional defines " t r u e " with respect to the sentence "Aardvarks amble." And all such T-sentences together constitute an exhaustive and complete definition of "true." The idea behind the disquotational view is sometimes put this way: to say that a sentence is true is really just an indirect way of saying the sentence itself. To say that the sentence "Penguins w a d d l e " is true is just an indirect way of saying that penguins waddle. This p r o m p t s the question: Why not dispense with the truth-predicate in favor of direct talk about the world? The disquotationalist will respond by pointing to generalizations such as "Every sentence of the form ' p or not p' is t r u e " and to truthascriptions such as " W h a t Joe said is true." In the case of the generalization, we could dispense with the truth-predicate here if we could produce an infinite conjunction of sentences of the form " p or not p": aardvarks amble or a a r d v a r k s do not amble, and bison bathe or bison don't bathe or . . . .But we cannot produce such an infinite conjunction, and instead we achieve the desired effect by generalizing over sentences, and then bringing those sentences back down to earth by means of the truth predicate. 8 In cases like " W h a t Joe said is true," where the target utterance is picked out by means other than a quote-name, " t r u e " serves to express an infinite disjunction:

Copyrighted Material

66

Dor it Bar-On and Keith

Simmons

What Joe said = "S1" and S1, or W h a t Joe said = "S 2 " and S2, or

where "s 1 ", "s 2 ", ... are quote-names of the sentences of Joe's language. The truth-predicate, says the disquotationalist, is a logical device: a device for disquotation, and for expressing infinite conjunctions and disjunctions. 9 There is no more to the meaning of " t r u e " than its disquotational role. A n d moreover there is no more to our understanding of the concept of truth than an understanding of the disquotational role of the truth-predicate. If the concept of truth is a "thin" concept in this sense, then it can make no substantive contribution to our understanding of assertion, meaning, belief, or any other concept in this cluster. Explanations of these notions that make use of the truth-predicate can avail themselves only of its role as a logical device of disquotation. In this vein. Field, a leading disquotationalist, observes that it may seem as t h o u g h we need to appeal to truth to characterize the realist doctrine that "there might be . . . sentences of our languages that are true that we will never have reason to believe" (Field 1999: 369) (where the realist is contrasted with the anti-realist, who identifies truth with some notion of justifiability). However, Field claims that the role of truth in such a characterization is "purely logical" (1999: 369). But for our finite limitations, the realist doctrine could be expressed without use of a truth-predicate via an infinite disjunction (where each disjunct is of the form "p and we will never have reason to believe p"). And Field thinks that appeal to truth in general claims such as that there is "a ' n o r m ' of asserting and believing the t r u t h " is merely disquotational. 1 0 T h e idea is that such general claims are in effect abbreviations for infinite conjunctions. The disquotationalist denies the concept of truth any substantive explanatory role. O r consider Ramsey's redundancy theory. According to Ramsey, truth and falsity are ascribed to propositions, where the propositions may be given explicitly (as in "it is true that Caesar was m u r d e r e d " ) or indirectly (as in " H e is always right"). In the first case, the word " t r u e " is readily eliminated: "it is evident that i t is true that Caesar was murdered' means no more than that Caesar was m u r d e r e d " (Ramsey 1927: 106). "True" and "false" are merely terms "which we sometimes use for emphasis or for stylistic reasons, or to indicate the position occupied by the statement in our a r g u m e n t " (Ramsey 1927: 106). Beyond this, " t r u e " is r e d u n d a n t , and clearly has no substantive connections to other concepts. In the second case, it is not so obvious that " t r u e " is eliminable: T h u s if I say " H e is always right". I mean that the propositions he asserts are always true, and there does not seem to be any way of expressing this without using the word "true". But suppose we put it thus " F o r all p, if he asserts p. p is true", then we see that the propositional

Copyrighted Material

Assertion

and truth

67

function p is true is simply the same as p, as e.g. its value "Caesar was murdered is t r u e " is the same as "Caesar was m u r d e r e d " . (Ramsey 1927: 106) So in the second case too, Ramsey says, " t r u e " is eliminable. Ramsey dismisses any problem about what it is for a proposition or j u d g m e n t to be true just make the j u d g m e n t . For Ramsey, the real question is what is involved in making a j u d g m e n t in the first place. Ramsey's essentially behavioristic approach to belief and j u d g m e n t makes connections to various concepts, including use and commitment; but as far as truth is concerned, there is no place in this account for anything but a "thin" concept of truth. For Ramsey, " t r u e " is an eliminable predicate. For the prosententialist, it is not even a predicate at a l l , " but rather a component of prosentences. In the discourse: Mary: Chicago is large. John: If that is true, it probably has a large airport. the expression "that is t r u e " is a prosentence, which shares its content with its antecedent, namely "Chicago is large." On the prosentential view, " t r u e " serves as a purely grammatical prosentence-forming operator. Under the prosententialist analysis, it does not survive as a discrete term expressing a separable concept that could stand in relations to other concepts. These various deflationary accounts minimalism, disquotationalism, the redundancy theory, and prosententialism differ in various ways. They differ, for example, over the utility of the truth predicate, the bearers of truth, and the g r a m m a r of "true." But their proponents share an isolationist view of truth: the claim that a complete account of truth will proceed independently of the concepts to which truth is traditionally tied - meaning, belief, assertion, and the rest. As Hill (2002: 4) puts it, an acceptable form of deflationism will maintain that "there is no particular set of concepts that one must acquire prior to acquiring the concept of t r u t h " ; it will present " t r u t h as a u t o n o m o u s and presuppositionless." 1 2 Moreover, to explain facts whose statements involve truth-locutions all we need to assume about truth, according to the deflationist, is what falls out of the preferred deflationist account of "true." Isolationism cuts both ways: our understanding of truth is exhausted by the deflationary account, and truth in turn has no role in explaining other concepts, beyond the purely logical or grammatical role assigned to it by the deflationary account. Any appeal to substantive connections between truth and other concepts is based on a mistaken, inflated idea of truth. We take isolationism to be a c o m p o n e n t of any genuine form of deflationism. It is telling that leading deflationists such as Horwich and Field explicitly spell out the isolationist consequences of their deflationary theories. But "deflationism" is a slippery term, with no agreed fixed extension -

Copyrighted Material

68

Dor it Bar-On and Keith

Simmons

perhaps there are those who would still want to count platitude-respecting minimalism as a kind of deflationism. To avoid terminological distractions, we describe our target as conceptual deflationism: the view that a deflationary account of "is true" (be it minimalist, disquotationalist, prosententialist, or some other) will give us all there is to know or understand about the concept of truth and its explanatory role. As explained at the outset, we think that conceptual deflationism should be kept distinct from both metaphysical deflationism (which eliminates the property of truth from our ontological inventory) and linguistic deflationism (which offers this or that deflationary account of various everyday uses of the word "true"). To defend conceptual deflationism it is not enough to give arguments for the futility of appealing to a metaphysically robust relation obtaining between items we call " t r u e " a n d something else (reality, facts, the way things are). N o r is it enough to provide a recipe for eliminating "is t r u e " from our everyday discourse in which we apply it to sentences or other items. We need to see how to deflate the explanatory role apparently played by truth in elucidations of various concepts of interest to us. In what follows, our principal focus will be assertion. We will be arguing that understanding assertion requires more that the "thin" concept of truth afforded by deflationary accounts. Before turning to assertion, however, it may be useful to clarify further what we take to be at stake regarding conceptual deflationism. An analogy might help. Consider for example the predicate "is g o o d . " Many philosophers would deny that there is a sui generis non-natural property denoted by this predicate. Some of them call them "ethical reductionists" - propose to identify the property of being good with some natural property (though they disagree a m o n g s t themselves on which natural property it is). Other philosophers, however, might deny that "is g o o d " denotes any property, natural or not. These philosophers call them "ethical eliminativists" might suggest that we can eliminate the property of being good from our ontological inventory altogether: for there is no single feature that is shared by all a n d only things we call good. The eliminativist owes us a story about ethical discourse. What is it that we're doing when we call something " g o o d " ? What is the meaning of "is g o o d " as it is used in everyday discourse? And so on. An ethical expressivist, for example, might suggest that "is g o o d " serves an expressive function - perhaps it is used to express approval, or some other pro-attitude. C r u d e as this m a p may be, we hope it can serve to recall familiar debates in metaethics. What is relevant about the ethical case for our purposes is the fact that it would seem implausible to claim that goodness is a "thin" concept - that it serves no explanatory role save what can be captured by some purely formal specification of the things that "is g o o d " applies to, and that it bears no rich connections to other concepts. C o n t e m p o r a r y ethical noncognitivists, it seems, want to preserve the explanatory role of ethical concepts and to exhibit the place each of them occupies in a complex and rich network of concepts that hold interest for us. 13

Copyrighted Material

Assertion

and truth

69

As we see it, deflationists tend to overlook the possibility of an analogous position regarding truth. Eager to eliminate truth from our ontology and impressed by this or that deflationary treatment of "is true," they have lost sight of the place the concept of truth occupies in our conceptual scheme. Yet it seems to us that metaphysical deflationism and linguistic deflationism, taken either separately or together, d o not entail conceptual deflationism. This is what we hope to show below.

II Truth and the force of assertion Frege writes: "When we inwardly recognize that a thought is true, we are making a judgement: when we communicate this recognition, we are making an assertion" 1 4 (Frege 1979: 139). M a k i n g a j u d g m e n t , Frege emphasized, must be sharply distinguished from the entertaining of a thought - we must not confuse merely predicating with judging. And, in parallel, assertion must be sharply distinguished from the mere expression or articulation of a thought. This is reflected in F r e g e s judgement sign - from the Begriffsschrift. The horizontal stroke - the so-called "content-stroke" - combines the symbols following it into a whole thought; the vertical stroke the "judgment-stroke" expresses the recognition or affirmation that this thought is true. If we omit the little vertical stroke at the left end of the horizontal stroke, then the judgement is to be transformed into a mere complex of ideas; the a u t h o r is not expressing his recognition or non-recognition of the truth of this. (Frege 1879: 1-2) Sometimes the mere expression of a thought is all that matters. In Section 2 of the Begriffsschrift, Frege points out that one might present the thought that unlike magnetic poles attract one a n o t h e r merely for hypothetical consideration one's intention is: "just to produce in the reader the idea of the mutual attraction of unlike magnetic poles - so that, e.g.. he may make inferences from this t h o u g h t and test its correctness on the basis of these" (Frege 1879: 2). O r the thought might be the antecedent of a conditional: as Frege (1979: 185-6) puts it, "even if the whole c o m p o u n d sentence is uttered with assertoric force, one is still asserting neither the truth of the t h o u g h t in the antecedent nor that of the thought in the consequent." O r again, it might be what Frege (1979: 130) calls a "mock t h o u g h t " of fiction. But, says Frege, the logician has no interest in mock t h o u g h t s 1 5 or, more generally, in the mere presentations of thoughts: "the thing that indicates most clearly the essence of logic is the assertoric force with which a sentence is uttered" (Frege 1979: 252). And, according to Frege, assertoric force is to be understood in terms of truth: to assert that p is to present p as true (or. as Frege sometimes puts it, to express one's acknowledgement of p as true, or to express one's affirmation of p as true).

Copyrighted Material

70

Dor it Bar-On and Keith

Simmons

Frege's view of assertion is a natural one. When 1 assert that penguins waddle, I am not merely predicating waddling of penguins (where, for Frege, "merely predicating waddling of penguins" is that c o m p o n e n t c o m m o n to stating that penguins waddle, asking or wondering whether penguins waddle, promising to make it so that penguins waddle, and so on). N o r am I merely presenting the thought that penguins waddle for consideration. There are m a n y speech-acts I can perform that involve a given thought: I can suppose it, propose it, float it, question it. I can also express a t h o u g h t in the course of asserting (or questioning, or supposing, etc.) a c o m p o u n d proposition, such as a conditional or disjunction. Frege plausibly claims that the distinguishing m a r k of assertion what sets it apart f r o m other speech-acts - is the fact that when I assert something, I present a certain t h o u g h t as true. C o m m e n t i n g on this aspect of Frege's view. D u m m e t t says: When we make an assertion we are not merely uttering a sentence with determinate truth-conditions u n d e r s t o o d by the hearer, and hence with a particular truth-value; that, after all. we should do if the sentence expressed only part of what we were asserting for instance, if it were the antecedent of a conditional. We are also, rightly or wrongly, saying that the sentence is true. This activity of asserting that the thought we are expressing is true is sui generis: it is not a f u r t h e r determination of the truth-conditions of the sentence, which remain unchanged whether we are asserting it to be true or not, but rather something which we do with a sentence whose truth-conditions have already been fixed. This is not the only thing we can do with a sentence: we can use it in giving a definition, in asking a question (of the kind requiring the answer "Yes" or " N o " ) , or in the course of telling a story. ( D u m m e t t : 1978: 106) The kind of " d o i n g " D u m m e t t is speaking of here - which he associates with Frege's notion of the force of an utterance, as opposed to its sense is not restricted to speech-acts. It has its analogue in thought. Frege also wanted to distinguish the mere entertaining of a thought from the m a k i n g of a judgment. And here again, the difference is to be captured in terms of truth. To judge that p. as opposed to merely entertaining the thought that p. or considering whether p, is to take p to be true. Moreover, judging, entertaining, considering, and so on, are attitudes or mental acts involving a single kind of entity a thought. A n d . as D u m m e t t points out, Frege's characterization of t h o u g h t s itself also appeals to the notion of truth: What, then, distinguishes t h o u g h t s from other constituents of our mental life, from mental images, ideas, feelings, desires, impulses, and the rest? That was the question Frege asked, and was the first to strive to answer:

Copyrighted Material

Assertion

and truth

71

and his first step toward an answer was to say that thoughts, and only they, are apt to being characterized as being true or as being false. ( D u m m e t t 1993: 154) The notion of truth, then, has a central role to play in Frege's account of assertion, j u d g m e n t , and thought. 1 6 T h e foregoing serves to usher in the following challenge for the conceptual deflationist: to explain how to achieve a proper theoretical understanding of what it is to assert or judge that p without help f r o m the concept of truth. What is it, for the deflationist, to put forward a thought with the force of an assertion or j u d g m e n t ? What is it that one is doing when one judges or asserts as opposed to merely supposing it. or considering it, and so on? And, in the mental realm specifically, what is it that distinguishes having a thought from other kinds of mental episode? Frege suggests that in answering these questions we ourselves, as theorists, must appeal to a notion of truth. How might the deflationist respond? Consider disquotationalism or Horwich's minimalism. According to this kind of deflationary view, the function of " t r u e " is exhausted by its disquotational or denominalizing role. For example, " t r u e " contributes no more than its denominalizing role to an explanation of why true beliefs engender successful action. Now consider the thesis that to assert is to present as true. The thesis invokes a certain concept - that of presenting as true in a natural explanation of assertion. It also involves the use of the truth-predicate; in Horwich's terms, it is a fact about " t r u e " that needs to be explained. With the denominalizing role of " t r u e " in mind, a deflationist might claim that the thesis that to assert that p is to present p as true is equivalent to the thesis that to assert that p is to present p. This commits us to the claim that to present p as true is just to present p; for example, to present as true the thought that aardvarks amble is just to present the t h o u g h t that aardvarks amble. 1 7 But this claim is false, for there are m a n y ways to present a thought. I can present a thought as worthy of your consideration, or as a conjecture, or as a remote possibility, or as outrageous and I can also present it as true. Presenting as true is just one way of presenting. So this cannot be the right way to denominalize away " t r u e " in the locution "present as true." It might be suggested that the correct denominalizing move is a wholesale "semantic descent": to present the proposition that aardvarks amble as true is just to present aardvarks as ambling. Here not only truth drops out, but so does the proposition (or thought, or sentence) that is said to be presented as true. 1 8 T h e claim "to assert that aardvarks amble is to present the proposition that aardvarks amble as t r u e " is just a r o u n d a b o u t way of saying "to assert that aardvarks amble is to present aardvarks as ambling." But what is it to present aardvarks as ambling? One way of understanding this claim is as saying that we present the worldly mammals, aardvarks, as engaging in a certain kind of activity, as when a zoo-keeper gestures toward

Copyrighted Material

72

Dor it Bar-On and Keith

Simmons

ambling aardvarks. But this clearly will not d o as an explanation of assertion. (Asserting that aardvarks amble does not require the presence of ambling aardvarks; that is part of the point of assertion.) 1 9 A m o r e plausible way of understanding the present suggestion is as saying that presenting aardvarks as ambling is a matter of representing aardvarks a certain way - as, well, ambling. But not any old form of representing would do the trick of capturing what is distinctive about asserting (or judging) that aardvarks amble, as opposed to, say, merely pointing to or drawing a picture of (or forming a mental image of) ambling aardvarks. T h e kind of representing that is relevant to assertion is surely " f a c t u a l " representation: representing things as being so, or describing things as they are. But it seems that the very same task will face the deflationist, this time with respect to representation: how to understand what it is to represent as being so, which is a special kind of representing, just as presenting as true is a special kind of presenting. It does not seem that a deflationist would want to trade truth-talk in favor of unreconstructed talk of "the way things are" or "its being thus and so", etc. So the detour via representation will be of no help to the deflationist. We cannot, then, denominalize away " t r u e " as it appears in "present as true." The point applies equally well to the other locutions that Frege employs. It is implausible to claim that Frege's uses of " t r u e " in "express one's recognition that p is t r u e " or "express one's acknowledgement of p as true," or "express one's affirmation of p as t r u e " are r e d u n d a n t ; one can recognize or acknowledge or affirm a thought in various ways. And if it is insisted that to recognize (acknowledge, affirm) is just the same as to recognize as true (acknowledge as true, affirm as true), then the involvement of truth should instead be regarded as implicit. For now the notions recognize, acknowledge, and affirm must be understood respectively along the lines of apprehend as true, register as true, and put forward as true. "True" in the theorist's m o u t h still cannot be disquoted or denominalized away, and the deflationist must find another tack. According to the present objection, we are to explain what it is to assert in terms of presenting as true, and the deflationist does not have the resources to a c c o m m o d a t e this explanation. But a deflationist might respond as follows: I accept that there is an undeniable connection between assertion and truth. But it is misleading to present the connection in terms of the slogan to assert is to present as Hue. Better to reverse the order: to present as true is to assert. Assertion is not to be characterized in terms of truth; rather, our use of the predicate "true" is to be characterized in terms of assertion. The objection gets the direction of explanation the wrong way a r o u n d . The idea is this: to predicate " t r u e " of a sentence (or a thought, or a proposition) is just to assert the sentence (thought, proposition). To say:

Copyrighted Material

Assertion

and truth

73

"Penguins waddle" is true is just to assert the sentence "Penguins waddle." We assert the sentence, we present it as true, by predicating " t r u e " of it. Ayer puts it this way: to say that a proposition is true is just to assert it, and to say that it is false is just to assert its contradictory. And this indicates that the terms " t r u e " and "false" connote nothing, but function in the sentence simply as m a r k s of assertion and denial. And in that case there can be no sense in asking us to analyze the concept of " t r u t h " . 2 0 (Ayer 1936: 88-9) This is what we may call an illocutionary form of deflationism. We use " t r u e " not to describe sentences or propositions, but rather to p e r f o r m the speech-act of assertion. The term " t r u e " is not really a property-denoting predicate; nor does it express a concept that stands in need of analysis, or can play an explanatory role in the explanation of assertion. T h e illocutionary deflationist will take on board the equivalence thesis, and agree that the content of " A a r d v a r k s amble' is t r u e " is no different f r o m that of "Aardvarks amble." But though " t r u e " does not add content, it does introduce assertoric force: to say '"Aardvarks amble' is t r u e " is to produce an assertion with the content that aardvarks amble. We can understand Ayer as suggesting that " t r u e " and "false" play a role very similar to that played by explicit performatives. Saying "p is t r u e " is equivalent to saying "I assert p." But on this reading of Ayer's proposal, his remark that "there can be no sense in asking us to analyze the concept of ' t r u t h ' " seems wrong. For we (theorists of language) do seek analyses of explicit performatives such as "assert." Perhaps Ayer would say that what makes no sense is to seek a particular kind of analysis: one which pairs up with "is t r u e " a special worldly property - the property of being true. It is a mistake a kind of category mistake to engage in that kind of search, since calling something " t r u e " is not predicating a property of it; rather it is articulating a kind of act one is performing in putting p forward. If so, Ayer is also advocating what we earlier called metaphysical deflationism: the denial that there is any need to populate our ontology with a substantive property of truth to serve as the extension of our truth predicate. Like Ayer, Frege emphasizes the illocutionary aspect or role of truth. Frege regards truth as belonging to the same family of concepts as assertion and j u d g m e n t . Truth is associated with what we do when we use language and when we think, rather t h a n with the content of what we say or think. Moreover, Frege famously endorses the equivalence thesis, that " p " and ""p' is t r u e " are equivalent in content. 2 1 So is Frege a deflationist (as is often supposed), perhaps an illocutionary deflationist? No: it is striking that Frege is fundamentally opposed to illocutionary deflationism, and deflationism more generally.

Copyrighted Material

74

Dorit Bar-On and Keith

Simmons

At the core of disquotationalism, minimalism and the redundancy theory is the familiar equivalence thesis for instance, "Sea-water is salt" is equivalent to "'Sea-water is salt' is t r u e " (or "The proposition that seawater is salt is true"). And Frege certainly endorses this equivalence. In his terms, both sentences express the same thought, and "[t]he word 'true' is not an adjective in the ordinary sense" (Frege 1979: 251). Frege writes: If I attach the word "salt" to the word "sea-water" as a predicate, I form a sentence that expresses a thought. To make it clearer that we have only the expression of a thought, but that nothing is m e a n t to be asserted, I put the sentence in the dependent form "that sea-water is salt". . . . With the word " t r u e " the matter is quite different. If I attach this to the words "that sea-water is salt" as a predicate. I likewise form a sentence that expresses a t h o u g h t . For the same reason as before I put this also in the dependent form "that it is true that sea-water is salt". The thought expressed in these words coincides with the sense of the sentence "that sea-water is salt". So the sense of the word " t r u e " is such that it does not m a k e any essential contribution to the thought. (Frege" 1979: 251) Predicating " t r u e " of a sentence makes no difference to the thought expressed: it is the same thought. But according to Frege, " t r u e " also makes no difference to the force with which the thought is expressed. Frege immediately goes on to say: "If I assert "it is true that sea-water is salt', I assert the same thing as if I assert 'sea-water is salt'. This enables us to recognize that the assertion is not to be found in the word ' t r u e ' " (Frege 1979: 251). If one's deflationary view of " t r u e " is based on the equivalence thesis, then according to Frege " t r u e " c a n n o t be the m a r k of assertion. Indeed, Frege says that "there is no word or sign in language whose function is simply to assert something" (Frege 1979: 185). Frege is explicitly opposed to illocutionary deflationism, and for good reason. If one accepts the equivalence thesis, there is no difference between asserting that p and asserting that p is true. Further, the locution "p is true" can occur as the antecedent of a conditional, where it c a n n o t be produced with assertoric force as in "If it is true that aardvarks amble, then. . . . " F u r t h e r still, I can say "It is true that aardvarks amble" with a variety of different illocutionary forces - I can be supposing, conjecturing, pretending, or acting. Frege writes: O n e can. indeed, say: "The thought, that 5 is a prime number, is true." But closer examination shows that nothing more has been said than in the simple sentence "5 is a prime number." The truth claim arises in each case f r o m the f o r m of the declarative sentence, and when the latter lacks its usual force, e.g. in the m o u t h of an actor upon the stage, even the sentence "The thought that 5 is a prime n u m b e r is t r u e " contains

Copyrighted Material

Assertion

and truth

75

only a thought, and indeed the same thought as the simple "5 is a prime number". (Frege 1892: 64) The use of " t r u e " doesn't produce assertoric force; rather, "even where we use the form of expression 'it is true that . . . ' the essential thing is really the assertoric form of the sentence" (Frege 1979: 129). So although Frege appears to give a deflationary account of the word "true", he explicitly rejects illocutionary deflationism. Moreover Frege's remarks about truth seem inhospitable to conceptual deflationism. For example: Truth is obviously something so primitive and simple that it is not possible to reduce it to anything still simpler. (Frege 1979: 129) The goal of scientific endeavour is truth. (Frege 1979: 2; italics in the original). Logic is the science of the most general laws of truth. (Frege 1979: 128) It is the striving for truth that drives us always to advance f r o m the sense to the reference. (Frege 1892: 63) Clearly we must distinguish what Frege says about the word "true," and what he says about truth. Frege writes: If I assert that the sum of 2 and 3 is 5, then I thereby assert that it is true that 2 and 3 make 5. So I assert that it is true that my idea of Cologne Cathedral agrees with reality, if I assert that it agrees with reality. Therefore it is really by using the form of an assertoric sentence that we assert truth, and to do this we do not need the word "true." (Frege 1979: 129) We do not "assert t r u t h " (as Frege puts it) by predicating " t r u e " or by prefixing the operator "It is true that." Rather, we assert truth by using a sentence with assertoric force: "In order to put something forward as true, we do not need a special predicate: we need only the assertoric force with which the sentence is uttered" (Frege 1979: 233). Science aims at the truth, and "logic is the science of the most general laws of t r u t h " but it does not follow that science or logic is concerned with the word "true": "what logic is really concerned with is not contained in the word 'true' at all but in the assertoric force with which a sentence is uttered" (Frege 1979: 252).

Copyrighted Material

76

Dorit Bar-On and Keith

Simmons

How should we understand Frege on truth? One way is that truth is a simple, unanalyzable yet substantive primitive that is implicated ("asserted") not by the use of any special word or phrase that denotes it, but only by utterances m a d e with the appropriate kind of illocutionary force assertoric force. A second way is suggested by G r e i m a n n (2004: 431): " t r u t h is a constituent of the assertoric force with which assertoric sentences are normally uttered." 2 2 In either case, truth is bound up with assertion in a way that vitiates conceptual deflationism. For. on either way of reading Frege, grasp of the concept of truth c a n n o t be exhausted by any deflationary account of "true"; it requires understanding the f u n d a m e n t a l connection between truth and assertion. So. while Frege explicitly embraces the equivalence thesis, which is c o m m o n to all linguistic deflationists, he should be seen as denying conceptual deflationism, along with its isolationist commitments. Far from assigning no explanatory role to the concept of truth, or isolating it from other theoretical concepts of interest, he takes it to be a central notion in his philosophy of language and thought, intimately interwoven with other central notions. We think that Frege teaches us a very important lesson. Deflationism about the word "true" is one thing, deflationism about the concept of truth quite another. Frege endorses the equivalence thesis about "true." He denies that " t r u e " introduces descriptive content when appended to sentences or thoughts. He also denies that the use of " t r u e " introduces any illocutionary force, or plays the role of an explicit performative. But for all that Frege is not a conceptual deflationist. To fully appreciate Frege's lesson, it is helpful to distinguish uses of " t r u e " in a language we theorize about from uses of " t r u e " in the language in which we d o our theorizing. Deflationists typically focus on uses of " t r u e " in locutions such as '"Aardvarks amble' is t r u e " (where " t r u e " applies to the q u o t e - n a m e of sentences). "Socrates' last utterance was t r u e " (where " t r u e " is applied to a sentence picked out by means other t h a n its q u o t e name), and "Everything Plato said was t r u e " (where " t r u e " is applied to a d o m a i n of sentences). Call these fust-order uses of "true." In these uses, " t r u e " is applied to sentences (or utterances, or propositions). Disquotationalists, minimalists and redundancy theorists aim to provide accounts of all first-order uses of "true." accounts which attribute to " t r u e " a disquotational or denominalizing or prosentence-forming role. It may even be claimed that " t r u e " can in principle be extruded altogether f r o m the language; for example, as we saw in connection with disquotationalism, it might be claimed that if only we could handle infinite lists, " t r u e " could be replaced by infinite disjunctions and conjunctions. Given a language f r o m which " t r u e " is in this way eliminable, it is tempting to conclude that there is no substantive role for the concept of truth to play regarding this language. After all. with the word " t r u e " eliminated, we do not even have a term to express the concept of truth. If "true" is to be regarded merely as a device for forming infinite disjunctions and conjunctions, then it

Copyrighted Material

Assertion

and truth

77

can easily seem that a deflated, "thin" concept of truth is all we need. It might seem that linguistic deflationism brings conceptual deflationism in its train. Accept for the m o m e n t that first-order uses of " t r u e " can be handled in a deflationary way. From this it does not follow that we have d o n e away with the concept of truth. As we continue to reflect on or theorize about a language and its practitioners, we may turn to the speech-act of assertion. We may say, following Frege, that to assert is to put forward as true. Here is the word " t r u e " again, appearing in the language in which we theorize. 2 3 In our m o u t h , the word " t r u e " is not used as a disquotational or denominalizing or prosentential device. We are not even p u r p o r t i n g to describe some sentence or thought. This is not a first-order use, and it c a n n o t be disquoted away. O u r use of the word is what we may call a reflective or explanatory use; it is m a d e in the course of offering a general explanation of what speakers are doing when they use language in certain ways. We are not calling any specific sentence true, nor are we making oblique reference to some set of sentences and saying of its members that they are true. Rather, we are trying to identify a distinguishing feature of a class of acts - assertions (or in the mental case, judgments). We should emphasize that the role we are assigning to the concept of truth is reserved for reflective, explanatory uses of "true." We d o not take a stand on first-order uses of "true"; for the sake of argument, we are willing to accept a deflationary account of first-order uses. At the same time, we are not claiming that explanatory uses of " t r u e " are the sole province of highlevel theorizing about language. Reflective or explanatory uses of " t r u e " need not be technical or recherché. For example, the claim that to assert is to present as true seems intuitive and natural enough. 2 4 O u r present focus is on uses of the concept of truth in contexts which do not involve calling something true. If Frege is right, truth is implicated in the assertoric force with which a sentence is uttered. For what is distinctly characteristic of acts of assertion is that they present a thought as true. So when we explain assertion, we ourselves use a truth-locution and employ the concept of truth. Thus, even if we grant, as does Frege, that first-order uses of " t r u e " submit to the equivalence thesis, we may need to employ the concept of truth for explanatory purposes. As we have seen. Frege is not at all shy about using truth-locutions in an explanatory way in connection with assertion, logic and science. He does not accept a deflationary view of the concept of truth. We can get at the same point by imagining a language which has no semantic vocabulary at all. A Fregean explanation of key p h e n o m e n a of this language in terms of truth may still be in order. Speakers of the language, we may suppose, will make assertions, and we will want to explain what distinguishes a speaker's assertion that p from other speech-acts wondering whether p, j o k i n g that p, and the rest. For Frege, as long as we have assertoric force, we have truth. An act of assertion puts forward a thought as true; assertions advance us from sense to reference, f r o m the

Copyrighted Material

78 Dorit Bar-On and Keith

Simmons

thought to the True. There is no understanding of what it is to assert without the notion of truth. 2 5 It is worth emphasizing that the involvement of truth with assertion, on the Fregean story, does not emerge through the use of "true" to ascribe a property to a sentence, or a thought. If Frege is right, when we say that assertion is presenting as true, we are not ourselves describing a sentence or a thought a certain way by ascribing truth to it. The Fregean point is precisely that presenting as true (that is. asserting) is not a matter of ascribing a property to a sentence or thought, but rather is a special kind of doing or act. different f r o m conjecturing, or surmising, or wondering, etc. T h u s contrast: (i)

presenting the thought that p as true

with (ii) presenting the thought that p is true. The deflationist can claim that (ii) is equivalent to: (ii') presenting the thought that p since in (ii) " t r u e " is applied to a particular thought. But it is not plausible to equate (i) with (ii'), for the familiar reason that presenting as true is just one way of presenting. In (i), " t r u e " qualifies the kind of presenting at issue, not the thought presented. Frege's "presenting as true" is intended to characterize a class of acts that more often than not do not themselves involve calling something true; that is. F r e g e s e x p l a n a n d u m is not primarily acts that involve first-order uses of "true." And explanatory uses of t r u t h locutions themselves do not involve appending " t r u e " to a citation (direct or indirect) of a sentence or a thought. What implicates the concept of truth is our use of "presenting as true," which does not involve calling anything true, either. 2 6 To sum up: even if " t r u e " in its first-order uses is correctly treated by the linguistic deflationist, truth may nevertheless be a substantive concept, one that we invoke as we reflect on the nature of language use. If, like Frege, you are deflationist about first-order uses of "true" but you think asserting is presenting as true, you have not yet combined a truth-based account of assertion with a genuine form of conceptual deflationism.

Ill Assertion as taking-true Frege's lesson is this: you can be a deflationist about " t r u e " (in its first-order uses) but not a deflationist about the concept of truth. But it is not clear where Frege stands on the metaphysical issue regarding truth. On the one

Copyrighted Material

Assertion

and truth

79

hand, many of his remarks suggest that he thinks we get nowhere in our understanding of truth by pairing with the predicate "is t r u e " some property that all and only true items share. On the other hand, Frege's talk of truth as "something primitive and simple" may suggest that he is reifying truth as a special, irreducible property. So some may argue that F r e g e s rejection of conceptual deflationism goes hand in hand with his attraction to metaphysical inflationism. However, next we will consider the view of a self-proclaimed deflationist. Robert B r a n d o m , who is in our sense both a linguistic and a metaphysical deflationist, 2 7 but who is not we will argue - a conceptual deflationist. Brandom's case will show us that F r e g e s lesson can be strengthened: you can be a deflationist about " t r u e " and deny the existence of a property of truth, and still fail to be a deflationist about the concept of truth. T h r o u g h o u t Making It Explicit, Brandom equates asserting with true or putting forward as true. For example:

taking-

Everyone ought to agree that asserting is putting forward a sentence as true. (Brandom 1994: 231) The attitude of taking-true is just that of acknowledging an assertional commitment. . . . A theory of asserting and assertional c o m m i t m e n t is a theory of taking-true. ( B r a n d o m 1994: 202) Of the claim or principle that a theory of asserting is a theory of takingtrue, Brandom writes: this principle can be exploited according to two different orders of explanation: moving f r o m a prior notion of truth to an understanding of asserting (or judging) as taking, treating, or putting forward as true, or moving from a notion of asserting to a notion of truth as what one is taking, treating, or putting forward a claim as. (Brandom 1994: 202) Brandom rejects the former order of explanation, since he thinks it accords truth a m o r e basic explanatory role than it deserves. Instead, B r a n d o m endorses the latter way of exploiting the principle, that of "starting with an antecedent notion of assertional significance and then moving via that principle to an understanding of what is involved in talk of t r u t h " (Brand o m 1994: 232). Notice the shift here to "talk of t r u t h . " Frege has prepared us to be wary of this move to truth-talk - an account of "true," if it is confined to what we've called first-order uses, may tell us nothing about truth's involvement with assertion. So in what follows we will be interested not only in the adequacy of Brandom's treatment of the word "true"; we will also be interested in the relevance of that treatment to assertion.

Copyrighted Material

80

Dorit Bar-On and Keith Simmons

We turn first to B r a n d o m on "is t r u e " and truth-talk. B r a n d o m takes to heart the pragmatists' idea "that in calling something true, one is doing something, rather than, or in addition to. saying something" ( B r a n d o m 1994: 287). According to B r a n d o m . the pragmatists are motivated by the special relation that exists between the force of an act of taking as true and the force of a straightforward assertion (see Brandom 1994: 288). B r a n d o m makes the same observation as Frege: "In asserting 'it is true that p', one asserts that p, and vice versa. The force or significance of the two claims is the same" (Brandom 1994: 288). Frege drew the conclusion that "true" cannot be the mark of assertion. Brandom draws a different conclusion: "true" makes explicit what is implicit in the act of assertion. When we assert "Aardvarks amble," we take it as true that aardvarks amble. When we assert " A a r d varks amble" is true," we are asserting the same thing as before - but we are now making explicit our formerly implicit attitude of taking-true. B r a n d o m writes: "Semantic vocabulary is used merely as a convenient way of making explicit what is already implicit in the force or significance that attaches to the content of a speech act or attitude" ( B r a n d o m 1994: 82). This is not to endorse illocutionary deflationism: Brandom is not claiming that predicating " t r u e " of a proposition introduces or produces an assertion. It is just that what was already implicit is now m a d e explicit: nothing new is added. Now, according to B r a n d o m , " t r u e " is force-redundant. The assertion that p is true merely preserves the force of the assertion that p. As long as the t r u t h locution is freestanding, and not. say. the antecedent of a conditional, force redundancy is a central p h e n o m e n o n of truth-talk (see B r a n d o m 1994: 299). But this, B r a n d o m says, cannot be the whole story about "true." Recall Frege's observation: there are uses of truth locutions that are not assertoric consider, for example, the locution "'Penguins waddle' is t r u e " in the conditional "If "Penguins waddle' is true, then penguins waddle." B r a n d o m offers "a more general redundancy view that has the force r e d u n d a n c y of freestanding truth-takings as a consequence" ( B r a n d o m 1994: 299). This is the prosentential theory of truth, according to which " . . . is t r u e " is a prosentence-forming operator. 2 8 Prosentences are analogous to pronouns: just as "She s t o p p e d " differs from " M a r y s t o p p e d " in its explicit dependence on a token of " M a r y " as its anaphoric antecedent, so the prosentence "'Penguins waddle" is t r u e " differs from "Penguins waddle" in its dependence upon an antecedent (perhaps the token of "Penguins w a d d l e " that it contains) - but there is no difference of semantic content between the presentence and its a n a p h o r i c antecedent. According to the prosentential theory, truth locutions have the same semantic content as their antecedents whatever the context whether, for example, they are freestanding or antecedents of conditionals. Moreover, according to Brandom, content-redundancy entails force-redundancy: intersubstitutability of "it is true that p" and "p" in all occurrences, embedded or not. is sufficient to yield force redundancy in freestanding

Copyrighted Material

Assertion

and truth

81

uses as a consequence. If two asserted contents are the same, then the significance of asserting them in the same pragmatic context should be the same. ( B r a n d o m 1994: 300) This last claim of Brandom's seems correct: given content redundancy, force redundancy seems to follow. If "Aardvarks amble" and " A a r d v a r k s amble' is t r u e " share the same content, then the force of uttering one in a given context will be the same as uttering the other in that same context. But then how can content-redundancy yield the kind of connection between " t r u e " and assertoric force that B r a n d o m is claiming? That is, how can it be maintained that " t r u e " makes explicit what is already implicit in the force that attaches, specifically, to assertion? Suppose I conjecture that aardvarks amble. Given the content-redundancy of truth, and holding the context fixed, if we substitute "that 'aardvarks amble' is t r u e " for "that aardvarks amble," there will be no difference of force. The force will still be that associated with an act of conjecturing. But surely we should not now conclude that the addition of "is true" makes explicit what we are doing when we conjecture. T h a t is, the use of "is t r u e " here cannot be taken as signaling that we're putting p forward with the force of a conjecture. Given the prosentential account, " t r u e " is c o n t e n t - r e d u n d a n t and force-redundant. But if that is so, then there can be no privileged connection between truth and assertion, as Frege thought. We can no more say that " t r u e " makes it explicit what we are doing when we are asserting than we can say that " t r u e " makes it explicit what we are doing when we are conjecturing or supposing or entertaining a thought. On a content-redundancy view such as prosententialism, there can be no general link between truth and force, and in particular no link between truth and assertoric force. The point generalizes to other deflationary accounts - for example, disquotationalism and minimalism. Given the claimed equivalence of "Aardvarks amble" and "'Aardvarks amble' is t r u e " (or "Aardvarks amble" and " T h e proposition that aardvarks amble is true"), their intersubstitutability in a given pragmatic context will not affect the force in any way, whatever that force may be. Treated along these deflationary lines, the truth predicate is quite inert with respect to force, and it cannot function to make explicit what we are doing when we engage in this or that speech-act. But this breaks the link that Brandom claims to exist between my assertional act of taking-true and the word "true." In endorsing the pragmatic forcebased approach to truth for freestanding uses (as opposed, for example, to embedded uses in the antecedent of a conditional), Brandom writes: " O n the pragmatic line being considered, it is the practical significance or force of asserting that defines taking-true, and this sense of taking-true accounts for o u r use of ' t r u e " ' (Brandom 1994: 297-8). But if uses of " t r u e " are accounted for along prosentential lines, then those uses cannot be explained in terms of taking-true. For the prosententialist, "true" is content-redundant.

Copyrighted Material

82

Dorit Bar-On and Keith

Simmons

and thereby force-redundant. So the word " t r u e " has no link to assertoric force, or any other kind of force. And in the other direction, a prosentential account of " t r u e " has no tendency to provide a deflationary understanding of truth-taking. (Indeed, as we shall see below. Brandom's own account of truth-taking is far from deflationary.) Again, we should heed Frege warning not to assume too quickly that a deflationary account of " t r u e " will bear on truth's involvement with assertion. Deflationary accounts of first-order uses of " t r u e " will have the result that to conjecture that p is true is to do n o m o r e than to conjecture that p. to wonder whether p is true is to d o nothing other than to wonder whether p, and so on for the various available illocutionary forces. Content r e d u n d a n c y brings force-redundancy in its train. But this does nothing to address the explanatory claim that links assertoric force, specifically, to presenting as true (or to Brandom's own taking-true). To deflate that claim it is not enough to be able to substitute "p" for "p is true." We need to be able to substitute "presenting" for "presenting as t r u e " in our explanation, and we need to be able to substitute " t a k i n g " for "taking true." And we have argued that deflationary treatments of "is t r u e " have not shown us how to d o that. Thus, we may accept that the verbal separation of sentences or t h o u g h t s into the " t r u e " ones and the rest is not a substantial separation that aims to identify some c o m m o n feature shared by all and only items falling under "true." For all that, it may still be the case - and indeed we've argued that it is the case - that the separation of acts into truth presentings and the rest, or the separation of attitudes into truth takings and the rest, is a substantial one. 2 9 So let us now turn to the altitude of truth-taking or taking as true. Even if the prosentential theory of the word " t r u e " is correct - indeed, even if we can eliminate altogether all first-order uses of the word " t r u e " f r o m the language - still the notion of taking as true remains, and with it the concept of truth. In Brandom's phrase, truth here is "what one is taking, treating, or putting forward a claim as" ( B r a n d o m 1994: 202) when one asserts. At this point, then, the deflationist needs a suitably deflationary account of taking as true and the associated concept of truth. But this is not something that B r a n d o m provides. B r a n d o m is dismissive of an inflated respect for truth - but truth-taking is an altogether different matter: "Talk about the cardinal importance of concern with truth is a dispensable façon de parler". What actually matters is the pragmatic attitude of taking-true or putting forward as true, that is, judging or asserting" ( B r a n d o m 1994: 82). Here, as in many other places, B r a n d o m identifies taking-true with asserting and so Brandom's account of asserting is at the same time an account of taking-true or putting forward as true (recall Brandom's remark that a theory of asserting is a theory of taking-true). A n d this theory of asserting or takingtrue is not a deflationary one. What are we doing when we assert or put forward a sentence as true? Brandom's general answer is that we are u n d e r t a k i n g a certain kind of c o m m i t m e n t . The c o m m i t m e n t may be expressed in terms of the Sellarsian notion of the practice of giving and asking for reasons:

Copyrighted Material

Assertion

and truth

83

The kind of c o m m i t m e n t that a claim of the assertional sort is an expression of is something that can stand in need of (and so be liable to the d e m a n d for) a reason; and it is something that can be offered as a reason. ( B r a n d o m 1994; 167) It is a necessary condition of assertional commitments that they play the dual role of justifier and subject of d e m a n d for justification; assertions "are fundamentally fodder for inferences'' (Brandom 1994: 168). Brandom's account of asserting or taking as true goes forward in terms of commitments, inferences, entitlements, justificatory responsibilities - clearly this is not a deflationary account of taking-true. There is a phenomenalist aspect to Brandom's account (see B r a n d o m 1994: 29Iff.). Brandom explicitly rejects a metaphysical account of truth, and instead aims for a phenomenalistic understanding of truth. We are to start not with truth, but with takings-true (just as a phenomenalist might start not with what is represented but with representings of it). This strategy may lead to the rejection of any metaphysically robust property of truth serving as the extension of " t r u e " in its ordinary first-order uses. But it does not permit us to dispense with the concept of truth altogether. That concept still figures in. Taking-true is a distinctive attitude, a special kind of taking; presenting as true is a special kind of presenting. And our use of " t r u e " in these locutions indicates what it is that is distinctive about this kind of presenting. Truth is what we take or present a claim as when we assert, whether or not we think that truth is a special feature shared by all true claims. Truth is thus bound up with the attitude of taking as true - and we have seen that Brandom's treatment of that attitude is far f r o m deflationary. B r a n d o m may lay claim to metaphysical deflationism, and to a prosententialist deflationism about "true," but for all that he is not a conceptual deflationist. 3 0 As a theorist of language, he still needs to invoke the concept of truth to explain certain of our linguistic doings, even if there is no substantive property of truth shared by all and only things we (properly) call true, and even if the word " t r u e " in its first-order uses is r e d u n d a n t . Put together Brandom's metaphysical deflationism and his redundancy account of "true," and you still won't have enough to break the conceptual connections between assertion and truth, or to deflate the concept of truth. In treating truth phenomenalistically, B r a n d o m does not free himself of the concept of truth - indeed, neither does he free himself of the word "true." For what are we to make of the occurrences of "true" in such locutions as "to assert is to take as true or to put forward as true"? As we suggested earlier, in locutions like these we are making explanatory uses of "true," as opposed to first-order uses. These explanatory uses cannot be treated along prosentential lines. Here " t r u e " is not a prosentence-forming operator, and there is no anaphoric antecedent. In claiming that to assert p is to put forward p as true, we say something about the act of asserting a sentence; we

Copyrighted Material

84

Dorit Bar-On and Keith

Simmons

d o not explicitly or obliquely call some antecedent sentence true. Brandom does acknowledge that there are uses of the truth-predicate that his theory cannot handle. As examples, he cites uses of the substantive " t r u t h , " as in "Truth is one, but beliefs are m a n y " and "Truth is a property definable in the language of some eventual physics" ( B r a n d o m 1994: 323). But these uses of "true," B r a n d o m argues, are the result of a mistake: a false analogy is drawn between " t r u e " and ordinary predicates, with the result that a property of truth is hypostatized. B r a n d o m writes: "It is no defect in the a n a p h o r i c account not to generate readings of the fundamentally confused remarks that result" ( B r a n d o m 1994: 324). But the a n a p h o r i c account does not generate a reading of "When we assert p we put forward p as true," and B r a n d o m surely will not regard this claim as confused. Is there a way of maintaining both Brandom's account of assertion a n d conceptual deflationism? One option would be to deny that taking-true or putting forward as true have any connection to truth. But this seems at best implausible and at worst contradictory. It is clear that B r a n d o m himself does not endorse this option. He shares the pragmatist's c o m m i t m e n t to p h e n o m e n a l i s m about truth, that "once one understands what it is to take or treat something as true, one will have understood as well the concept of t r u t h " ( B r a n d o m 1994: 291): being true "is to be understood as being properly taken-true" ( B r a n d o m 1994: 291). 31 It is clear that B r a n d o m seeks an account of truth - a pragmatic account that proceeds f r o m the attitude of taking-true. There is no getting rid of the concept of truth: it is one of B r a n d o m ' s explananda. This point is reinforced by Brandom's treatment of assertion and knowledge. According to B r a n d o m , "assertions have the default status and significance of implicit knowledge claims" ( B r a n d o m 1994: 201). Consequently, says B r a n d o m . some account must be given of the truth condition on knowledge. The question then is: " W h a t is the socialdeontic attitude corresponding to the truth condition on attributions of knowledge?" (Brandom 1994: 202). The aim is not to deflate truth, but to explain truth's connection to knowledge within the pragmatic f r a m e w o r k of c o m m i t m e n t s and entitlements. 3 2 This project, we have argued, is quite distinct f r o m that of providing a deflationary account of "true." Brandom's pragmatic project could be seen as similar in some ways to that of the ethical noncognitivist who tries to explain o u r use of an ethical term such as " g o o d . " The noncognitivist wants to explain ethical discourse without invoking a substantive property of goodness a worldly feature (whether irreducibly moral or reducible to some natural feature or features). Toward that end the noncognitivist might invoke a h u m a n attitude of "taking-to-be-good" as the more basic explanatory notion, in terms of which our moral attitudes and behavior, as well as aspects of our moral discourse (such as o u r calling things good) may be explained. The noncognitivist may f u r t h e r try to reduce the attitude of taking-to-be-good to some other psychological attitudes or propensities. But if she were to do so, it doesn't seem to us that this would a m o u n t to a "conceptual deflation" of

Copyrighted Material

Assertion

and truth

85

the notion of goodness, or of taking-to-be-good. For the noncognitivist's reductive success might itself serve to reveal the richness of the relevant notions, and their connections to other ones in our conceptual scheme. (If the concept of goodness were merely a "thin" concept, whose understanding is exhausted by its mastery as a purely formal or grammatical device, the noncognitivist reductive or reconstructive task would be a very easy one indeed! 33 ) A n o t h e r option open to the conceptual deflationist is to embrace Brandom's account of assertion - the pragmatic account of assertion that proceeds in terms of commitments, entitlements, inferences, and social-deontic attitudes - and bypass any identification of assertion with taking-true. The idea is that if we take the account this way, we have an account of assertion that is entirely independent of truth. Since truth makes no appearance, the account may seem to be compatible with conceptual deflationism. However, there are at least three problems with this option. First, it may be objected that this is a purely verbal maneuver. We can withhold the claim that to assert is to take as true. But withholding the claim does not make it false. Perhaps all we have really done is to reconstruct in pragmatic terms the notion of taking-true under a different name, under the n a m e of assertion. At issue isn't the word "true," but an allegedly explanatory notion that we invoke using that word in certain claims about the use of language. Second, even if assertion is pulled apart from notions like taking-true, putting forward as true, presenting as true, those notions still stand in need of explanation. What is it to put forward a proposition as true? It seems for all the world that this is the speech-act of assertion. But suppose it isn't. Then it is a distinct speech-act that is characterized in terms of truth. A n d , as we have seen, truth's involvement in taking-true or putting forward as true is not explained by a deflationary account of the (first-order uses) of the word "true." As long as we need the concept of truth to explain certain speechacts, we must reject conceptual deflationism. Third, consider Brandom's account of assertion taken in a way that bypasses truth altogether. According to this pragmatic account, when I assert that aardvarks amble I take on certain c o m m i t m e n t s a n d responsibilities. But what is it about the act of asserting that generates these inferential c o m m i t m e n t s and justificatory responsibilities? Why are they not generated when I suppose, conjecture, or ask a question? The answer seems obvious: it's because when I assert that aardvarks amble, I'm putting forward that proposition as true, as the way things are. I'm not putting it forward as something to be assumed for the sake of argument, or as something to be questioned, or as something that is possible - I'm putting it forward as true. This is what distinguishes assertion from other speech-acts, and this is what generates the c o m m i t m e n t s and responsibilities identified by the pragmatic account. It is no accident that taking-true and its cognates figure so centrally in Brandom's account. We cannot make proper sense of the pragmatic account of assertion without appreciating the role that the concept of truth plays in our understanding of assertoric force.

Copyrighted Material

86

Dorit

Bar-On

and Keith

Simmons

In o u r i n t r o d u c t i o n we cited Williams's c l a i m t h a t , a c c o r d i n g t o deflat i o n a r y views, the f u n c t i o n of t r u t h talk is "wholly expressive, never explanatory" (Williams 1999: 547). If o u r a r g u m e n t against c o n c e p t u a l deflationism is right, t h e n t r u t h - t a l k does have a n e x p l a n a t o r y f u n c t i o n , contra t h e deflationist. It is n o t the case t h a t the distinctive f u n c t i o n of " t r u e " is only t o allow us " t o e n d o r s e o r reject s e n t e n c e s (or p r o p o s i t i o n s ) t h a t we c a n n o t simply a s s e r t " (Williams 1999: 547). We c o n c l u d e t h a t a t h o r o u g h - g o i n g d e f l a t i o n i s m is i n c o m p a t i b l e with a n a d e q u a t e a c c o u n t of a s s e r t i o n . T h i s is not to say t h a t we m u s t a b a n d o n a d e f l a t i o n a r y a c c o u n t of the w o r d " t r u e " - Frege h o l d s o u t t h e possibility t h a t we c a n explain a s s e r t i o n while e n d o r s i n g linguistic d e f l a t i o n i s m ( f o r first-order uses of " t r u e " ) . A n d it is n o t to say t h a t we m u s t a b a n d o n metaphysical deflationism B r a n d o m p r o v i d e s a rich a c c o u n t of a s s e r t i o n in t e r m s of t r u t h - t a k i n g while m a i n t a i n i n g b o t h m e t a p h y s i c a l a n d linguistic d e f l a t i o n i s m . But it is t o say t h a t we m u s t a b a n d o n c o n c e p t u a l d e f l a t i o n i s m . P e r h a p s t h e r e is n o r o b u s t p r o p e r t y of t r u t h , a n d p e r h a p s first-order uses of " t r u e " a r e r e d u n d a n t , b u t if w e a r e to u n d e r s t a n d w h a t we d o w h e n we a s s e r t , we c a n n o t d i s p e n s e w i t h the c o n c e p t of t r u t h . 3 4

Notes 1 2 3 4 5 6 7 8 9

Williams (1999: 547). emphasis in the original. See Bar-On et al. (2000). See Jackson et al. (1994). This "fact about truth" is considered by Horwich (1990: 23-4). The phrase is Horwich's; see Horwich (1990: 22ff.). Horwich (1990: 26: original italics). See Horwich (1990: 24). See Quine (1970: 12). For example, see Leeds (1978: 120-21 and n. 10). Field (1986: 58), Resnik (1990: 412). and David (1994: 107 and Ch. 4). 10 See Field (1999: 369). 11 "Truth, to coin a phrase, isn't a real predicate" (Grover et al. 1975: 97). 12 Hill (2002: 3-4) recommends a version of minimalism, and regards deflationism in general as a view that "truth is philosophically and empirically neutral, in the sense that its use carries no substantive philosophical or empirical commitments." 13 For influential contemporary versions of ethical noncognitivism, see. for example, Blackburn (1993) and Gibbard (1990). It could be argued that Ayer, an early and well-known proponent of emotivism. did have hopes of "deflating" the concept of goodness (and other ethical concepts). But Ayer's version of ethical noncognitivism is notoriously riddled with difficulties, in good part because of its failure to do justice to the role played by ethical concepts in our conceptual scheme. (Ram Neta has drawn our attention to a view of Geach's (1956) according to which "good" is merely a grammatical device a predicate modifier - which does not represent a substantive concept.) 14 The emphasis is Freges. Frege says the same thing at a number of places. For example: "Once we have grasped a thought, we can recognize it as true (make a judgement) and give expression to our recognition of its truth (make an assertion,)" (Frege 1979: 185).

Copyrighted Material

Assertion and truth

87

15 "The logician does not have to bother with mock thoughts, just as a physicist, who sets out to investigate thunder, will not pay any attention to stage-thunder" (Frege 1979: 130). 16 Dummett also urges a conceptual connection between assertion and truth, but stresses the reverse direction: "Without doubt, the source of the concept [of truth] lies in our general conception of the linguistic practice of assertion" (Dummett 1991: 165). 17 We have expressed the claim here in terms of presenting a thought as true - but one could equally well put the claim in terms of presenting a sentence as true, or a proposition as true. 18 We owe this suggestion to Dean Pettit. 19 Even if one can present ambling aardvarks in the absence of ambling aardvarks (perhaps via some form of "deferred ostension"), there is still the problem of identifying what has been conveyed in this way: is it the general claim that aardvarks amble or the particular claim that there are some ambling aardvarks about? There is also a question how to regard particular claims such as "Asserting that aardvarks amble is presenting aardvarks as ambling" and "Asserting that penguins waddle is presenting penguins as waddling," and so on. as instances of the general claim "To assert that p is to present p as true," since the particular claims are not the result of any suitable substitution for p. 20 Strawson's variant of the redundancy theory identifies a performative role for "true": we use "true" to perform speech-acts such as endorsing, agreeing, and conceding, as well as asserting. See Strawson (1950). 21 See for example Frege (1956) in Blackburn and Simmons (1999: 88): "the sentence 'I smell the scent of violets' has just the same content as the sentence 'It is true that I smell the scent of violets."' 22 Greimann's article provides an excellent discussion of Frege on the expressive, illocutionary, cognitive and explanatory functions of truth. 23 In a similar vein. Dummett (1991: 167), remarking on the explanatory role of the concept of truth, has suggested that truth belongs with "second-level concepts, used to comment on our employment of our language." 24 Ted Parent has pointed out to us that, if one endorses a deflationary account of first-order uses of "true" and accepts our claim that reflective or theoretical uses cannot be deflated, one may have to opt for a non-uniform account of the truthlocutions for any language that permits both types of uses. This raises issues that go beyond the scope of this paper. We are here only concerned to argue that even if we were to accept a deflationary treatment of first-order uses of "true." we would still have reason to reject conceptual deflationism. 25 A similar point can be made about the link between truth and meaning via the notion of truth-conditions. For an argument that any explanation of meaning requires appeal to truth-conditions, and a discussion of the issues this raises for a deflationist, see Bar-On et al. (2000). Dummett has famously argued that a deflationist about truth cannot appeal to the notion of truth-conditions in explaining meaning. For an early statement, see Dummett's "Truth" in Dummett (1978: 7). Patterson (2003) argues that deflationism can be sanguine about the explanatory role of truth-conditions, as long as the deflationist account is confined to the role of "true" in the object language. He argues, however, that Dummett is right about the incompatibility of a truth-conditional account with what he calls "metalanguage deflationism". While we find Patterson's overall line congenial, we think his main point is better understood in terms of our distinction between first-order uses and explanatory uses of truth-locutions. 26 We thank Matthew Chrisman and Thomas Hofweber for remarks that have prompted this clarification. We should distinguish the schematic form Φ-ing the thought that p is true and Φ-ing the thought that p as true, where Φ stands for a

Copyrighted

Material

88

27

28

29 30

31

32

Dorit Bar-On and Keith

Simmons

speech-act. The former is arguably equivalent to Φ-ing the thought that p. But, in sharp contrast, the status of the latter seems to shift according to the value of Φ. Although we won't pursue it here, for some values the result is nonsensical (consider questioning whether the thought that p as true). For other values, the result is the speciation of a genus - in particular, this is so for the cases of interest to us, presenting as true or taking as true. It is less clear to us whether there are well-formed values that are equivalent to Φ-ing the thought that p. See e.g. Brandom (1994: Ch. 5. V. Section 2). entitled "Semantic Deflationism." Brandom's linguistic deflationism takes the form of prosententialism. and he explicitly denies that there is a property of truth: see. for example, Brandom (1994: 325). Brandom departs from the original prosentential theory, according to which " . . . is true" is a syncategorematic fragment of a semantically atomic generic prosentence "that is true" (see Brandom 1994: 305). This goes back to the point of clarification offered toward the end of Section II (see p. 78). Brandom does not claim to be a conceptual deflationist - but then he does not separate it off from other forms of deflationism in the way that we have. His willingness to place the notion of taking as true and its cognates in so central a position might suggest that he would reject conceptual deflationism; on the other hand he embraces deflationism without any apparent reservation. Brandom interpretation aside, we are urging that Brandom's account of assertion is incompatible with conceptual deflationism. Brandom writes: "Being true is then to be understood as being properly takentrue (believed). It is this idea that is built on here, jettisoning the details of the classical pragmatist account of belief or taking-true, and substituting for it the account of assertion and doxastic discursive commitment introduced in Chapter 3" (Brandom 1994: 291). Brandom also ties truth to other notions, for example successful action (see

1994: 527-9). 33 But see footnote 1. 34 We wish to thank Matthew Chrisman. Thomas Hofweber, Ted Parent, and Dean Pettit for reading a draft of the paper and providing very helpful comments.

References Ayer, A. J. (1936) Language. Truth, and Logic. London: Victor Golancz. Bar-On, D., Horisk, C. and Lycan. W. (2000) "Deflationism, Meaning, and Truthconditions," Philosophical Studies. 101: 1-28. Blackburn, S. (1993) Essays on Quasi-Realism. New York: Oxford University Press. Blackburn, S. and Simmons, K. (eds) (1999) Truth. Oxford: Oxford University Press. Brandom. R. (1994) Making It Explicit. Cambridge. MA: Harvard University Press. David. M. (1994) Correspondence and Disquotation. New York: Oxford University Press. Dummett. M. (1978) Truth and Other Enigmas. Cambridge, MA: Harvard University Press. -------(1991) The Logical Basis of Metaphysics. Cambridge, MA: Harvard University Press. -------(1993) The Seas of Language. Oxford: Clarendon. Field. H. (1986) "The Deflationary Conception of Truth." in G. MacDonald and C. Wright (eds) Fact. Science, and Morality. Essays on A. J. Ayer's Language, Truth and Logic. Oxford: Basil Blackwell. pp. 55-117.

Copyrighted

Material

Assertion and truth

89

--------(1999) "Deflationist Views of Meaning and Content," in S. Blackburn and K. Simmons (eds) Truth. Oxford: Oxford University Press, pp. 351-91. Frege, G. (1879) "Begriffsschrift (Chapter /)," reprinted in P. Geach and M. Black (1960) Translations from the Philosophical Writings of Gottlob Frege, Oxford: Basil Blackwell, pp. 1-20. ---------(1892) "On Sense and Reference," reprinted in P. Geach and M. Black (1960) Translations from the Philosophical Writings of Gottlob Frege, Oxford: Basil Blackwell, pp. 56-78. (1956) "The Thought: A Logical Inquiry," reprinted in S. Blackburn and K. Simmons (eds) (1999) Truth, Oxford: Oxford University Press. --------(1979) Posthumous Writings, Oxford: Basil Blackwell. Geach, P. (1956) "Good and Evil," Analysis 17: 33-42. Gibbard, A. (1990) Wise Choice, Apt Feelings, Cambridge, MA: Harvard University Press. Greimann, D. (2004) "Frege's Puzzle about the Cognitive Function of the Concept of Truth." Inquiry, 47: 425-42. Grover, D., Camp, J. and Belnap, N. (1975) "A Prosentential Theory of Truth." Philosophical Studies, 27: 73-125. Hill, C. (2002) Thought and World, Cambridge: Cambridge University Press. Horwich, P. (1990) Truth, Oxford: Basil Blackwell. Jackson, F., Oppy, G. and Smith, M. (1994) "Minimalism and Truth Aptness," Mind, 103: 287-302. Leeds, S. (1978) "Theories of Reference and Truth," Erkenntnis 13: 111 -29. Patterson, D. (2003) "Deflationism and the Truth Conditional Theory of Meaning," Philosophical Studies 124, 3: 271-94. Quine, W. V. (1970) Philosophy of Logic, Prentice Hall, NJ: Englewood Cliffs. Ramsey, F. P. (1927) "Facts and Propositions," Proceedings of the Aristotelian Society, suppl. vol. 7: 153-70. Resnik, M. (1990) "Immanent Truth," Mind, 99: 405-24. Strawson, P. (1950) "Truth," reprinted in S. Blackburn and K. Simmons (eds) (1999) Truth. Oxford: Oxford University Press, pp. 162 82. Williams, M. (1999) "Meaning and Deflationary Truth," Journal of Philosophy 96: 545-64. Wright. C. (1992) Truth and Objectivity, Cambridge, MA: Harvard University Press.

Copyrighted Material

5

The assertoric use of the concept of truth1 Dirk

Greimann

In o r d e r to analyze the "expressive f u n c t i o n " of the concept of t r u t h , its role in c o m m u n i c a t i o n , we must identify, in a first step, the linguistic devices of expressing t r u t h , i.e. the linguistic structures that enable us to express that s o m e t h i n g is true. 2 T h e surface g r a m m a r of natural language suggests that we express that s o m e t h i n g is true by m e a n s of the word "true." Just as redness is expressed by the predication of "red." so t r u t h seems to be expressed by the predication of " t r u e . " A c c o r d i n g to Frege, however, natural language is misleading here. He writes: [I]t is by using the f o r m of the assertoric sentence that we express t r u t h [womit wir die Wahrheit aussagen], a n d to d o this we d o not need the word "true." Indeed, we can say that even where we use the locution "it is true that ..." the essential thing is really the f o r m of the assertoric sentence. (Frege 1983: 140) 3 This f o r m , which in traditional g r a m m a r is often called the "declarative m o d e , " is the bundle of syntactic a n d phonetic properties by which assertoric sentences are distinguished f r o m interrogative a n d imperative ones. It does not c o r r e s p o n d to any word or p h r a s e of natural language but is realized in verb position, p u n c t u a t i o n , i n t o n a t i o n and o t h e r m o d e indicators. Frege's analysis implies that the expressive f u n c t i o n of the concept of t r u t h c a n n o t be identified with the c o m m u n i c a t i v e services of the word " t r u e " because there is also an "assertoric use" of the concept of t r u t h , as I would like to call it, which plays an i m p o r t a n t role in m a k i n g assertions a n d kindred speech acts. In the present paper, my aim is to describe this use in m o r e detail. In Section 1, Frege's analysis of the linguistic devices of expressing t r u t h is elaborated a n d defended. This is d o n e a l o n g the lines indicated by Frege, but with a systematic intention a n d w i t h o u t any exegetical ambitions. 4 In Section 2. the role of the concept of t r u t h in m a k i n g assertions is explained. Section 3 is devoted to the task of distinguishing the assertoric use of the concept of t r u t h f r o m o t h e r uses with which it is easily c o n f u s e d . T h e latter consists in the use of the t r u t h - p r e d i c a t e "____ is t r u e "

Copyrighted Material

The assertoric

use of the concept of truth

91

and the use of the truth-connective "it is true that _ . " Section 4 finally gives a brief survey of the implications of the analysis given here with regard to the philosophical theory of truth.

1 The assertoric form as a device of expressing truth To see that the assertoric form is a " t r u t h expressing device" of natural language, suppose that we want to express the truth of, say, the prepositional content that snow is white. This can simply be achieved by using the sentence "Snow is white": in virtue of its assertoric f o r m , this sentence expresses that snow's being white is true. The interrogative sentence "Is snow white?" expresses that snow is white, too, but it does not express that this is also true. M o r e precisely, the interrogative sentence expresses the request to determine the truth-value of the content that snow is white, 5 and the assertoric sentence expresses that the truth-value of this content is the True. In natural language, assertoric sentences are thus used for two fundamentally different linguistics tasks: as a means of expressing prepositional contents and as a means of expressing that such a content is true. To be sure, an assertoric sentence, considered as a mere syntactic object, does not express anything at all. To say that an assertoric sentence "expresses" the truth of a prepositional content means, strictly speaking, that it is normally used by speakers to express this. There are, of course, contexts in which the form of a sentence does not coincide with the illocutionary category of its utterance. On stage, for example, assertoric sentences are used, not to make assertions, but to act as if one were making an assertion. 6 But this does not imply that Frege is wrong when he claims that truth is expressed by the form of the assertoric sentence. For, in normal contexts, this form is indeed used to express truth and this already justifies Frege's claim. There are also contexts in which the sentence " T h e weather is fine" is used to express that the weather is not fine, as, for instance, in ironic talk; nevertheless, it remains correct to say that " T h e weather is fine" expresses that the weather is fine, because in most contexts this sentence is actually used to express this. Since the prepositional content of the assertoric sentence "Snow is white" is identical with the prepositional content of the interrogative sentence "Is snow white?," the expression of truth contained in "Snow is white" must be conceived of as a constituent of its illocutionary meaning, i.e. the assertoric force with which it is normally uttered. To make this explicit, Frege introduced the so-called judgment-stroke which is an illocutionary operator whose task is, in the first place, to make the truth-valuation associated with the form of assertoric sentences syntactically visible. 7 The sentence "Snow is white" can accordingly be reformulated as "_ White(snow)," where " White(snow)" is an expression that corresponds to the that-clause "that snow is white," which expresses that snow is white without expressing that this is also true.

Copyrighted Material

92

Dirk Greinumn

It is normally taken for granted that in the process of language learning the concept of truth comes into play when the language L 0 which is learned at the early stages of language acquisition is extended to a language L1 by introducing the predicate "_is true.1 According to our analysis, on the other h a n d , the concept of truth is already used at the level of L 0 . To make this plausible, we have to analyze the structure of assertions m o r e closely. Historically, there are mainly two approaches of analyzing the structure of assertions: the Aristotelian and the Frege-Searle approach. The former analyzes the speech act of assertion on the model of the grammatical formation of 1 sentences like "Snow is white." Typically, such a sentence is formed by applying a predicate-expression (like "_ is white") to a subject-expression (like "snow"). To this grammatical operation corresponds the speech act "to say something of something" (ti kata tinos), which is considered to be the basic speech act. On this approach, the application of the concept of truth is not a constituent of the assertion that snow is white. For, according to it, we need only to predicate whiteness of snow in order to assert that snow is white. The assertoric form of a sentence does not express the truth of the sentence's propositional content, but expresses - in the simplest case of elementary sentences - the satisfaction of concepts and relations by objects. Consider, for instance, the sentence "Snow is white," which has the grammatical f o r m "S is P." T h e difference between this sentence and the mere list of words "Snow, is, white" is that the sentence expresses, in virtue of its form, that the object referred to by the subject-expression and the concept (or property) referred to by the predicate-expression stand in a 1ccrtain logical relation, namely, the relation of1falling under a concept. Similarly the sentence " R o m e o loves Juliet" expresses in virtue of this form that the objects referred to by the subject-expressions stand in the relation referred to by the predicate-expression. Generalizing this, we can say that an elementary assertoric sentence expresses that the object(s) denoted by the subjectexpression(s) satisfy the concept or relation denoted by the predicateexpression. Now, in order to assert that snow is white, we need to express that snow satisfies whiteness. But, to do this, we do not need to apply the concept of truth, according to the Aristotelian a p p r o a c h , because it already suffices to predicate whiteness of snow. The form of the assertoric sentence must hence not be considered as a device of expressing truth, but as a device of expressing the satisfaction of concepts and relations by objects. The Aristotelian approach is not acceptable, however, because it involves a confusion of predication and assertion. Since the predication of whiteness of snow is a constituent, not only of the assertion that snow is white, but also of the question whether snow is white, it does not, taken by itself, constitute an assertion. Thus, the interrogative sentence "Is snow white?" expresses that snow is white (or that snow falls under the concept of being white), too; otherwise, it would not differ from the mere list of words "snow.

Copyrighted Material

The assertoric

use of the concept of truth

93

is, white." Nevertheless, the interrogative sentence does not express that snow's being white is also a fact, i.e., it does not express that the predication expressed by it is true. To explain this more accurately, we must distinguish between three different senses of "predication": a syntactic, a semantic and a pragmatic one. [i] Predication in the syntactic sense is the syntactic operation of forming an elementary sentence by filling appropriate subject-expressions into the empty places of predicate-expressions. The sentences "Socrates is wise," "Is Socrates wise?" and "Socrates, be wise!," for instance, all contain a syntactic predication, because in each case the predicate-expression "_ is wise" is applied to the subject-expression "Socrates" to form an elementary sentence. [ii] Predication in the semantic sense consists in the application of the meaning of a predicate-expression to the meaning of appropriate subjectexpressions in order to express a proposition. This semantic act, which is a special case of what is usually called the "propositional act," is a constituent of the assertion that Socrates is wise as well as of the question whether Socrates is wise and the order that Socrates should be wise. It has become c o m m o n practice in large parts of linguistics and philosophy to construe predication as a proper constituent of propositional acts that corresponds to predicate-expressions. According to this usage, the propositional act contained in the assertion that Socrates is wise divides into two parts: the act of referring to Socrates and the act of predicating the property of being wise to him. T h e motivation behind this usage is to do justice to the fact that the propositional acts corresponding to the sentences "Socrates is wise," "Socrates is unmarried." "Socrates is a philosopher" have a c o m m o n property, namely, the reference to Socrates, and that the propositional acts corresponding to the sentences "Socrates is wise," "Aristotle is wise," " K a n t is wise" have also a c o m m o n property, namely, the ascription of the property of being wise. Since, however, it is not possible to predicate something in isolation, i.e., without performing a complete propositional act, it is properly speaking not legitimate to construe predication as a proper part of propositional acts - predication so understood is not an act at all. Rather, we must construe predication as a special case of propositional acts, namely, as the expression of an elementary proposition as opposed to a complex one that corresponds to elementary sentences like "Socrates is wise" a n d " R o m e o loves Juliet." According to this usage, semantic predication is the semantic counterpart of syntactic predication considered as the f o r m a t i o n of elementary sentences by filling in the empty places of predicate-expressions. In what follows, I will stick to this usage of "semantic predication." [iii] Predication in the pragmatic sense finally consists in the act of asserting something of something. Thus, the assertoric sentence "Socrates is wise" contains a pragmatic predication because it asserts of Socrates that he is wise. The interrogative sentence "Is Socrates wise?," on the other hand, does not assert of Socrates that he is wise, although it predicates of Socrates that he

Copyrighted Material

94

Dirk Greimann

is wise in the semantic sense. The main difference is that the assertoric sentence contains the information that the semantic predication is true. In order to predicate something of something in the pragmatic sense, it is necessary to predicate something of something in the semantic sense, but not vice versa. The latter is shown by clauses like "that Socrates is wise"; in contrast to "Socrates is wise," this clause does not have assertoric force, i.e., it does not contain the claim to be true. Pragmatic predication is what is usually called the "ascription" of a property to an object. It would surely be awkward to say that by asking whether Socrates is wise, the speaker "ascribes" the property of being wise to Socrates. For, by asking this, the speaker does not assert of Socrates that he is wise. The ascription of a property is always a special case of asserting something of something. N o t e that the Aristotelian locution "to say something of something" is ambiguous, because it can m e a n pragmatic predication as well as semantic predication. If we identify "to say something of something" with predication in the pragmatic sense, we must conclude that interrogative sentences like "Is Socrates wise?" and imperative sentences like "Socrates, be wise!" d o not say anything of anything at all. Rather, we must say that whereas the assertoric sentence "Socrates is wise" asserts something of something, the interrogative sentence "Is Socrates wise?" asks something of something and the imperative sentence "Socrates, be wise!" orders something of something, where these acts are all construed as elementary acts that do not have a c o m m o n constituent. In this case, however, we cannot account for the obvious similarity between these sentences with regard to predication. If, on the other hand, we identify "to say something of something" with predication in the semantic sense, we cannot consider the act of "saying something of something" as the basic speech act because predication in the semantic sense is not a full speech act at all. but only a constituent of such acts. Consider, for instance, the act of predicating being wise in the semantic sense, which may be accomplished by the utterance of the clause "that Socrates is wise". Obviously, the utterance of such a clause does not have an illocutionary point, i.e., it cannot be used to perform an illocutionary act like making an assertion or asking a question. For this reason, semantic predication cannot be considered as the basic speech act. however. The Frege Searle a p p r o a c h is motivated by the quest to account for the similarities between sentences like "Socrates is wise," "Is Socrates wise?" and "Socrates, be wise!." by separating assertion and predication. 8 It takes into account that with one and the same semantic predication (or more generally; with one and the same prepositional act) one can associate different illocutionary acts as, for instance, an assertion and a question. The general structure of speech acts must accordingly be analyzed as I L L ( P R O P ) . where P R O P is the prepositional and ILL is the illocutionary act. 9 This analysis implies that in order to assert that Socrates is wise, we have to d o essentially two things: we must express a proposition and present it as

Copyrighted Material

The assertoric

use of the concept of truth

95

a fact. As Frege notes, this duality is hidden by the syntactic structure of assertoric sentences. He writes: In an assertoric sentence two different mately bound up with one another: assertion of its truth. And this is why tinguished. However, one can express time putting it forward as true.

kinds of things are usually intithe thought expressed and the these are often not clearly disa thought without at the same (Frege 1997: 239)

Frege's point is that in order to assert that Socrates is wise, it does not suffice merely to express the proposition (or " t h o u g h t " ) that Socrates is wise; the proposition expressed must also be put forward as true. These tasks are related to the syntactic structure of the sentence "Socrates is wise" as follows: the words of which this sentence is composed serve to express the proposition, and its assertoric form serves to present the proposition as a fact. From this it immediately follows that the assertoric form must be regarded as a truth-operator, i.e. as a device of expressing truth. Surprisingly, the speech acts of c o m m a n d i n g , promising and questioning involve an implicit application of the concept of truth as well. 10 To see this, we must ask what knowledge the child has to acquire in order to understand these acts. When, for instance, the mother asks her child "Is this white?," the child must have a twofold knowledge to understand this question. It must first be able to identify the state of affairs (propositional content) with which the question is concerned, and second, must know that he is supposed to determine whether this state of affairs is a fact. And, if the m o t h e r orders the child "Get into bed, now!," then in order to understand this order the child must also know that he is being c o m m a n d e d to make the described state of affairs a fact. Finally, when the m o t h e r promises "I will cook you p u d d i n g tomorrow," then the child must k n o w that the m o t h e r is obliging herself to make the described state of affairs a fact. If we wanted to construct a language whose grammatical structure is isomorphic to the structure of speech acts, we would hence have to reformulate assertions, questions, c o m m a n d s and promises according to the following paradigm: we reformulate "Socrates is wise" as "That Socrates is wise, is true," "Is Socrates wise?" as "That Socrates is wise, determine the truth-value of this!," "Socrates, be wise!" as "That Socrates is wise, m a k e this true," and "I, Socrates, will be wise" as "That Socrates is wise, I will make this a fact."

2 The role of the concept of truth in making assertions It has often been stressed, by D u m m e t t and others, that in making assertions the concept of truth plays an important normative role in the sense that we aim to say something true when we perform this speech act. 11

Copyrighted Material

96

Dirk Greimann

A c c o r d i n g t o the a l t e r n a t i v e analysis I wish to p r o p o s e , the c o n c e p t of t r u t h plays m o r e o v e r a constitutive role in the sense t h a t the a p p l i c a t i o n of t h e c o n c e p t of t r u t h is a n essential c o n s t i t u e n t of m a k i n g a n a s s e r t i o n . For, in o r d e r t o assert t h a t p. it d o e s n o t suffice simply to express t h e p r o p o s i t i o n a l c o n t e n t t h a t p; we m u s t also p r e s e n t this c o n t e n t as true. T h i s analysis d o e s n o t rule o u t . of course, t h a t we are d o i n g several o t h e r t h i n g s as well w h e n we a r e m a k i n g a n a s s e r t i o n . T h u s , w h e n we a r e m a k i n g a n a s s e r t i o n , we are n o r m a l l y d o i n g also the f o l l o w i n g things: we a r e t r y i n g t o c o n v i n c e the h e a r e r of the t r u t h of the p r o p o s i t i o n expressed; we a r e expressing o u r c o n v i c t i o n t h a t the p r o p o s i t i o n expressed is true; we a r e c o m m i t t i n g o u r s e l v e s t o t h e t r u t h of the p r o p o s i t i o n expressed; a n d we a r e implicitly c l a i m i n g t h a t we a r e able t o d e f e n d t h e p r o p o s i t i o n expressed with good arguments. T h e r e is r o o m for discussion w h e t h e r these acts are constitutive for m a k i n g assertions. If, for example, we c o u n t u t t e r a n c e s of "Presumably, Socrates is wise" as assertions, we m u s t d r o p the a s s u m p t i o n t h a t by m a k i n g a n assert i o n we a r e always implicitly c l a i m i n g t o be able t o d e f e n d t h e a s s e r t e d c o n t e n t with g o o d a r g u m e n t s . For the s e n t e n t i a l a d v e r b " p r e s u m a b l y " is a n e p i s t e m i c o p e r a t o r by m e a n s of which t h e s p e a k e r signals that he d o e s n o t have c o n c l u s i v e r e a s o n s t o believe the p r o p o s i t i o n e x p r e s s e d . 1 2 T h e r e c a n n o t , however, be a n y d o u b t t h a t the act of p r e s e n t i n g t h e p r o p o s i t i o n as t r u e is constitutive, for t w o reasons. First, it is a n essential p a r t of the m e a n i n g of the p e r f o r m a t i v e v e r b " t o a s s e r t " t h a t by a s s e r t i n g t h a t p t h e s p e a k e r p r e s e n t s it as a f a c t t h a t p. A s W r i g h t (1992: 23 ff.) p u t s it: "[A]sserting a p r o p o s i t i o n - a F r e g e a n t h o u g h t is c l a i m i n g t h a t it is true. T h e c o n n e c t i o n is partially c o n s t i t u t i v e of t h e c o n c e p t s of a s s e r t i o n a n d of t r u t h . " S e c o n d , the " a s s e r t o r i c f o r c e " with which a s s e r t o r i c s e n t e n c e s are n o r m a l l y u t t e r e d consists exactly in the t r u t h - c l a i m t h a t is m a d e by asserting s o m e t h i n g . To have a s s e r t o r i c f o r c e a n d to claim t r u t h f o r w h a t h a s been a s s e r t e d a r e o n e a n d the s a m e thing. It c a n , m o r e o v e r , be s h o w n t h a t the act of p r e s e n t i n g the p r o p o s i t i o n a l c o n t e n t as t r u e must be c o n s i d e r e d t o be a n irreducible c o n s t i t u e n t of assertions. In this c o n t e x t t h e f o l l o w i n g a n a l o g y is instructive. In his t h e o r y of j u d g m e n t , Frege stresses the d i s t i n c t i o n b e t w e e n the cognitive a c t s of " j u d g i n g " a n d of merely " g r a s p i n g " a p r o p o s i t i o n , which a r e c o r r e l a t e s of t h e linguistic a c t s of a s s e r t i n g a n d of merely expressing a t h o u g h t . 1 3 W h e n , for i n s t a n c e , a child living in a tropical c o u n t r y a s k s itself w h e t h e r s n o w is w h i t e while leaving o p e n w h e t h e r this is a c t u a l l y true, it merely " g r a s p s " the t h o u g h t t h a t s n o w is white. In o r d e r t o j u d g e t h a t snow is white, it m u s t , in a d d i t i o n , d e t e r m i n e the t r u t h - v a l u e of this t h o u g h t , i.e. it m u s t j u d g e w h e t h e r o r n o t it is the case t h a t s n o w is white. O n e m i g h t t h i n k t h a t it is g r a s p i n g of a t h o u g h t by the g r a s p s the t h o u g h t of s e c o n d simply: iff x g r a s p s t h e t h o u g h t

possible t o r e d u c e j u d g m e n t to the m e r e f o l l o w i n g definition: x j u d g e s t h a t p iff x o r d e r t h a t the t h o u g h t t h a t p is t r u e (or t h a t it is t r u e t h a t p). It is, however, possible

Copyrighted Material

The assertoric

use of the concept of truth

97

to grasp the t h o u g h t of second order that the thought that p is true without judging that p. A child, for instance, may ask itself whether it is true that snow is white while leaving open the question whether this thought is also true. This shows that the application of the concept of truth in grasping a t h o u g h t of the form "It is true that p " must be distinguished from its application in judging: to judge something as true must be considered as an act sui generis that c a n n o t be reduced to the mere grasping of a thought. The same holds, mutatis mutandis, for the linguistic act of presenting something as true, considered as a component of assertion. It would, again, be natural to reduce this c o m p o n e n t to the mere expression of a propositional content of the form "It is true that p," by the following definition: x asserts that p iff x expresses the thought of second order that the thought that p is true (or simply: iff x expresses the t h o u g h t that it is true that p). Again, it is possible to perform the one act without performing the other. For instance, by asking "Is it true that snow is white?," we express the thought that it is true that snow is white without expressing that this thought is also true. From this it follows that we must not consider the act of presenting a t h o u g h t as true as a special case of semantic predication, namely, as the propositional act of expressing a proposition of the form "It is true that p," but must acknowledge it as an irreducible c o m p o n e n t of assertion. 1 4 In order to complete our analysis of the role of the concept of truth in making assertions, we must describe in more detail [i] the nature and [ii] the point of the act of presenting something as true. [i] If we consider this act as a special case of the ascription of a property (pragmatic predication), we are driven to the absurd consequence that any assertion refers to a proposition and ascribes truth to it. On this assumption, the assertoric sentence "Snow is white," for instance, does not ascribe whiteness to snow, but it ascribes truth to the proposition that snow is white. This c a n n o t be true, however, because by asserting that snow is white we are talking about snow, not about the proposition that snow is white. Furthermore, the assumption entangles us in an infinite regress, because according to it we must ascribe truth to the proposition that F(a) in order to ascribe the property F to the object a. This implies that in order to ascribe truth to the proposition that F(a), we must in turn ascribe truth to the proposition that the proposition that F(a) is true, and so on ad infinitum. In order to avoid these consequences, we must construe the act of presenting something as true as a constituent of the ascription of a property by means of which the corresponding semantic predication is provided with assertoric force. Its task is to determine whether the denotations of the subject-expressions actually satisfy the denotations of the predicate -expressions. In the case of the sentence "Snow is white," for instance, the assertoric form "S is P " expresses that it is true that snow is white, i.e., that snow really is white: while "that snow is white" expresses that snow is white without at the same time expressing that this is also true, the assertoric sentence expresses both. But this does not mean, of course, that the sentence refers to a proposition and ascribes

Copyrighted Material

98

Dirk Greimann

truth to it. The presentation of the content that snow is white as true does not consist in the ascription of truth to a truth-bearer, but in the truth-claiming ascription of whiteness to snow. [ii] The point of presenting something as true is to convert a merely semantic predication into a pragmatic one, by endowing the semantic predication with assertoric force, i.e. the claim to be true. 1 5 To explain this point m o r e fully, we must distinguish between two kinds of truth, a semantic and a pragmatic one. 1 6 Semantic truth consists in the truth of a proposition and pragmatic truth in the truth of an utterance. T h e truth of a proposition is obviously independent of whether or not the proposition is or has been asserted by someone. That snow is white, for instance, would remain true even if no one had ever asserted this. On the other h a n d , the truth of an utterance is dependent on whether or not it has assertoric force. Although, for instance, the proposition expressed by the question "Is snow white?" is true, the question itself is neither true nor false. If someone asks, "Is snow black?," it would obviously be absurd to reply "This question is false," because by making the question the speaker does not claim that the proposition expressed is true. An utterance is true or false only if the speaker claims truth for the proposition he expresses, and this is the case if and only if the utterance has assertoric force. When we are making an assertion, we normally want our utterance to have a truth-value. To achieve this, it does not suffice merely to express a proposition, but we must also claim the proposition expressed to be true. To see this, consider an utterance of the clause "that snow is white." A l t h o u g h the proposition expressed by "that snow is white" is true, the utterance of this clause is neither true nor false. Thus, someone who utters "that snow is black" cannot be criticized for committing an error, because he leaves open whether the propositional content is true or false. 17 The reason is that the utterance of sentences of the form "that p" lacks what Searle calls a "direction of fit." Let me explain. According to Searle's typology, there are five basic types of illocutionary acts: assertoric, directive, commissive, expressive and declarative acts. Assertoric acts serve to present a propositional content as a fact, directive acts to get the hearer to do something, commissive acts to commit the speaker to some course of action, expressive acts to express some psychological state, and declarative acts to bring about a correspondence between the propositional content and reality by declaring the correspondence to exist. Assertoric acts have the "word-to-world direction of fit," because their success depends on that the proposition presented as true actually is true. Since the task of directive and commissive acts is to get the world to change in such a way that it matches the proposition expressed, they have the "world-to-word direction of fit."18 Since expressive acts serve to express the speaker's feelings, they have the "null direction of fit." Finally, declarative acts have both the "word-to-world direction of fit" and the "world-to-word direction of fit," because their task is to make the proposition expressed true by presenting it as true.

Copyrighted Material

The assertoric

use of the concept of truth

99

By uttering "that snow is white," we are merely expressing a proposition without performing any illocutionary act. Consequently, the utterance of this clause does not have any direction of fit. In order to assert that snow is white, on the other hand, we must also fix the direction of fit our utterance is supposed to have, because otherwise o u r utterance would be neither true nor false. This is exactly the task of the constituent act of presenting the proposition expressed as true: its point is to endow our utterance with the word-toworld direction of fit, by claiming truth to the semantic predication expressed.

3 Other uses of the concept of truth In order to characterize the assertoric use of the concept of truth in more detail, we must distinguish it from other uses with which it is easily confused. These are [i] the "predicative use," which consists in the use of the truth-predicate " is true," and [ii] the "connective use," which consists in the use of the truth-connective "it is true that ." [i] The competing theories of the truth-predicate's expressive function divide into two main categories: "surface conservative" and "revisionary" ones. The former proceed from the assumption that its expressive function agrees with the surface g r a m m a r of sentences in which the word "true" occurs. Since, from the grammatical point of view, this word occurs as an adjective, these approaches construe the application of the truth-predicate as a case of predication. Examples of this are the traditional conceptions of truth as correspondence with facts, as coherence, as usefulness, etc. The revisionary approaches, on the other hand, claim that " t r u e " is a pseudo-predicate, i.e., that sentences containing the word " t r u e " have a deep-structure in which " t r u e " does not occur as a predicate. This assumption leads to a program that Richard K i r k h a m has termed the "deep-structure project." 1 9 Its aim is to make explicit the errors about the expressive function of truth that arise through the surface g r a m m a r of sentences containing the word "true." Thus, the disquotation theory aims to show that the sentence "'Snow is white' is true" has the deep-structure of "Snow is white," and that the speech act performed by using " t r u e " is hence not the ascription of a property, but the purely syntactic operation of removing the quotation marks. In what follows, I shall assume that at least in some cases the truthpredicate is used to ascribe a property to truth-bearers. This "predicative" use, as I shall call it, might not occur in everyday talk, but it surely does occur in philosophical talk. There are, to be sure, also other uses of the truth-predicate, for example a disquotational and a prosentential one. These latter uses may be ignored, however, because they are not similar to the assertoric use. In order to introduce the predicative use, two steps are needed. The first consists in the introduction of a category of objects that are apt to serve as truth-(value)-bearers. This is usually accomplished by "reifying" the propositional contents of assertions that are then termed "propositions." The

Copyrighted Material

100

Dirk Greimann

second step is the introduction of a corresponding truth-property which applies to the proposition that p if and only if the truth-value of the proposition that p is the True. This is usually achieved by laying down the convention that the proposition that p is a true proposition if and only if p. The primary function of the predicative use obviously is to ascribe truth, considered as a property, to propositions or other truth-bearers. Since the assertoric use of the concept of truth is a constituent of the act of ascribing something to something, the predicative use involves also the assertoric one. Consequently, the metalinguistic sentence "The sentence 'Snow is white' is t r u e " has the deep-structure "I- True('Snow is white')," where the trutho p e r a t o r " I " expresses that the property for which the truth-predicate stands actually applies to the sentence "Snow is white." In asking "Is the sentence 'Snow is white' true?," the predicative use does not serve to ascribe truth to a truth-bearer, but merely to predicate truth of a truth-bearer in the semantic sense. This sentence has the deep-structure " ? True('Snow is white')," where "?" is an operator whose c o u n t e r p a r t in natural language is the form of the interrogative sentence. Its task is to express the request to determine the truth-value of the proposition expressed. T h e main difference between the predicative and the assertoric use of the concept of truth is that the former is a special case of (semantic or pragmatic) predication. As we have seen, the assertoric use cannot be construed as a special case of predication, but must be acknowledged as an irreducible constituent of m a k i n g assertions. To present something as true and to predicate something as true must hence be considered to be fundamentally different speech acts. [ii] The connective use consists in the use of the truth-connective "It is true that " and its cognates like "It is a fact that " and "It is the case that ." The difference between the truth-connective a n d the truthpredicate is that the former can be applied only to sentences, whereas the latter can be applied only to names of truth-bearers; for, expressions like "It is true the proposition that snow is white" and "Snow is white is t r u e " are syntactically not well-formed. The connective use agrees with the assertoric one in that it does not serve to ascribe a property. Just as in "It is not the case that snow is white" the connective "It is not the case that " is not used to ascribe the property of falsity to a truth-bearer, so too in "It is true that snow is white" the truthconnective is not used to ascribe the property of truth to a truth-bearer. The latter sentence does not ascribe truth to the proposition that is white, but ascribes whiteness to snow. The main difference between the connective and the the former cannot be used to endow an utterance with Thus, by uttering the interrogative sentence "Is it true the speaker makes use of the truth-connective, but he proposition expressed as being true.

Copyrighted Material

assertoric use is that the claim to be true. that snow is white?," does not present the

The assertoric use of the concept of truth

101

What is, then, the communicative function of the truth-connective? Given the approach developed here, it seems to be plausible to say that its primary communicative function is to make the truth-valuation explicit that is implicitly contained in assertoric sentences. As we have seen, the truth-valuation contained in "Snow is white," for instance, is indicated, not by a special sign, but by its assertoric form. The task of the truth-connective in "It is true that snow is white" is to bring this valuation to the grammatical surface, by representing it verbally. The effect of this is that by uttering "It is true that snow is white" the speaker emphasizes the truth of what he asserts. Since, however, the proper bearer of expressing truth in "It is true that snow is white" is the assertoric form, the truth-connective is, strictly speaking, redundant with respect to the task of presenting something as a fact, and it has its use only as a means of drawing the attention of the listener to the actual obtaining of a state of affairs. The main doctrine of the so-called "redundancy theory of truth," as advocated by Ramsey and others, is not that the truth-connective is redundant in the sense that it merely repeats what has already been said in the sentence in which it appears. Ramsey did not mean that in sentences like "It is true that snow is white" the truth of something is expressed twice; rather, his claim was that the expression of truth is superfluous. According to our approach, on the other hand, the expression of truth is an essential constituent of assertions, and the truth-connective is redundant only because the proper bearer of expressing truth is the assertoric form. For, on our analysis, the sentence "It is true that snow is white" has the deep-structure " it is true that White(Snow)." and this implies that the sentence is a pleonasm because it expresses the truth of the content that snow is white twice: first, by means of its form, and second by means of the truth-connective. 4 Implications The main tasks of the philosophical theory of truth are to determine the nature, the content and the function of the concept of truth. With regard to this, the Fregean analysis of the linguistic devices of expressing truth has the following five implications: [i] It is commonly taken for granted that a person's understanding of truth consists in the linguistic knowledge underlying the competent usage of the word "true." This approach is based on the following picture of the acquisition of the concept of truth. 2 0 In order to acquire the concept of truth, the child must learn how to use the word "true." The decisive step is to learn that the word "true" fulfills the Tarskian truth-schema, i.e., the child must learn that it is entitled to assent to a sentence of the form "The proposition that p is true" just in case it is entitled to assent to the corresponding sentence "p." This picture suggests that a person's understanding of the concept of truth consists in, as Paul Horwich puts it, "his disposition to accept, without evidence, any instantiation of the schema [ . . . ] "The proposition that p is true if and only if p"' (Horwich 1990: 36, 38).21

Copyrighted

Material

102

Dirk Greimann

According to our a p p r o a c h , on the other h a n d , a person's understanding of the concept of truth consists in his knowledge of the meaning of the assertoric form. To characterize this knowledge m o r e fully, we must explain how it is acquired. O r d i n a r y concepts like being red are learned through induction, i.e. by learning the empirical criteria on which it depends when to assent to the question "Is this red?" and when to deny it. Can we not assume that the concept of truth is also learned inductively, by learning when to assent to the question "Is this true?" and when to deny it? We cannot, because according to our a p p r o a c h understanding truth is a precondition of the successful performance of the speech acts of assent and dissent themselves. Since these speech acts are basic to the acquisition of any knowledge, we must assume that our species is endowed with a certain innate "sense of truth." i.e. a disposition to u n d e r s t a n d the linguistic utterances of his fellow h u m a n s from the outset in the perspective of truth and falsehood, because otherwise the speech acts of assent and dissent could not be learned. [ii] We must consider truth as a f u n d a m e n t a l concept that does not admit of a non-circular explication. For, in order to explain to a person x what truth is, one would have to perform certain assertoric speech acts. However, in order to understand these acts, x must already have the concept of truth at his disposal, because x must already k n o w what is expressed by the form of assertoric sentences. [iii] In order to determine the content of the concept of truth, we must analyze what the f o r m of the assertoric sentence expresses. This is simply that the sentence's content is the case. Thus, the sentence "Snow is white" expresses, in virtue of its assertoric form, that snow is indeed (or "really") white. To be true is, according to this view, to be the case. 2 2 This ontological notion of truth, as I would like to call it. must of course not be identified with the semantic notion of a true proposition. The former corresponds to the form of the assertoric sentence and the latter to the truth-predicate. [iv] The explications of truth given by Aristotle and Tarski are circular in the sense that they presuppose a prior grasp of the ontological notion of truth. The Aristotelian explication is this: a true sentence is a sentence which says that the state of affairs is so and so, and the state of affairs indeed is so and so. What it means for a state of affairs to be indeed so and so is not f u r t h e r explained. For this reason, the Aristotelian explication presupposes a prior grasp of the ontological notion of truth and it achieves a positive explication only of the derivative concept of a true sentence. T h e Tarskian theory of truth, considered as a semantic conception of t r u t h . 2 3 is based on the intuition that truth is a relational, language and world connecting property that can be positively explained in terms of the concept of satisfaction. According to this theory, truth is a special case of the satisfaction of propositional functions by infinite sequences of objects, namely, the satisfaction of a closed propositional function by all such sequences. T h e core of this theory is the explanation of satisfaction along the lines of the following schema:

Copyrighted Material

The assertoric use of the concept (S)

x1 iff

x„ satisfy the p r e p o s i t i o n a l f u n c t i o n " F ( x 1 , ... F(x1

of truth

103

, xn)"

xn).

T h e d i f f e r e n c e t o t h e classical c o r r e s p o n d e n c e t h e o r y is t h a t T a r s k i c o n strues t r u t h , n o t as a r e l a t i o n b e t w e e n s t a t e m e n t s a n d facts, but as a relation b e t w e e n s t a t e m e n t s a n d infinite s e q u e n c e s of objects. However, a c c o r d i n g t o o u r a p p r o a c h , the n o t i o n of t r u t h is c o n c e p t u a l l y p r i o r to the n o t i o n of s a t i s f a c t i o n b e c a u s e t o say of a n o b j e c t x t h a t it satisfies t h e p r o p o s i t i o n a l f u n c t i o n F(x) is t o say t h a t it is the case t h a t x is F. T h e s a t i s f a c t i o n of a p r o p o s i t i o n a l f u n c t i o n by a n o b j e c t m u s t h e n c e be c o n s i d e r e d as a special case of t r u t h , n o t vice versa. T h e r e a s o n is t h a t , in o r d e r t o u n d e r s t a n d w h a t it m e a n s for a n o b j e c t to satisfy a given p r o p o s i t i o n a l f u n c t i o n , we m u s t a l r e a d y u n d e r s t a n d p r a g m a t i c p r e d i c a t i o n a n d this p r e s u p p o s e s in t u r n t h a t we a l r e a d y have a g r a s p of the o n t o l o g i c a l n o t i o n of t r u t h . R e g a r d e d as a n a t t e m p t to explain the n o t i o n of t r u t h , Tarski's definition of t r u t h in t e r m s of s a t i s f a c t i o n is, t h e r e f o r e , entirely circular. [v] In o r d e r t o a n a l y z e the expressive role of t r u t h correctly, we m u s t t a k e i n t o a c c o u n t t h a t the c o n c e p t of t r u t h is a b e a r e r of i m p o r t a n t i l l o c u t i o n a r y f u n c t i o n s . T h e r e d u n d a n c y t h e o r y of t r u t h is b a s e d o n a linguistic fallacy b e c a u s e we c a n n o t i d e n t i f y the expressive role of the c o n c e p t of t r u t h w i t h t h e c o m m u n i c a t i v e services of the w o r d " t r u e . "

Conclusion If Frege's analysis of t h e linguistic devices of expressing t r u t h is c o r r e c t , the linguistic t h e o r i e s of t r u t h that a r e c u r r e n t l y discussed are essentially defective b e c a u s e they i g n o r e the i l l o c u t i o n a r y services of the c o n c e p t of t r u t h . T h i s d o e s n o t m e a n , of course, t h a t these t h e o r i e s are entirely w r o n g , but only that they a r e i n c o m p l e t e : they a r e not t h e o r i e s of the t r u t h - e x p r e s s i n g devices in general, but only t h e o r i e s of the t r u t h - p r e d i c a t e . T h e a r g u m e n t s p r e s e n t e d h e r e t o u n d e r p i n Frege's analysis a r e f a r f r o m b e i n g conclusive, of course. A m o r e t h o r o u g h d e f e n s e w o u l d have t o s h o w t h a t the A r i s t o t e l i a n a p p r o a c h c a n n o t be m o d i f i e d in such a way that it a c c o u n t s for the similarities between speech acts like a s s e r t i n g that s n o w is white a n d a s k i n g w h e t h e r s n o w is white. T h i s task m u s t be left to a n o t h e r paper, however.

Notes 1 The analysis presented here is a strongly improved and slightly extended version of the analysis presented in my German paper Greimann (2002). I would like to thank Martin Grajner (Munich) and especially Adolf Rami (Dresden) for their comments on the German paper. 2 The expressive function of the concept of truth must be distinguished from its explanatory function, i.e. the role it plays in carrying out the explanatory programs

Copyrighted Material

104

3 4 5 6 7

8 9 10 11 12 13 14 15 16 17 18 19 20 21 22

23

Dirk Greimann

of philosophy, in particular those of logic, semantics and epistemology. This distinction is due to Williams (1999: 547). My translation. For a reconstruction of Frege"s views, see Greimann (2004). I am following Frege (1997: 346) here. This observation is also made in Frege (1997: 330). Cf. Frege (1983: 214). " ⊢ Δ " expresses in Freges first system that the content that Δ A "is the True." For more details, see the reconstruction in Greimann (2000: 215 ff.). Cf. Frege (1983: 201. 214. 273). Frege (1990: 232). and Searle (1969: 26. 122). Exceptions are W-questions like " W h o is coming?" and expressive acts like "Ouch." A similar observation is made in Searle (1969: 122 ff.). See, for instance, Dummett (1959) and Horwich (2001: 160). To take this into account, we could distinguish between two kinds of assertion, weak and strong. A similar proposal is made in Siegwart (unpublished: 318 ff.). See, for instance. Frege (1997: 354) and Frege (1983: 214). This is another reason why the reduction of force to prepositional content in Lewis (1970) does not work. Here and in what follows, I am confining myself to the simplest case of elementary assertions. This distinction is inspired by the similar distinction in Levinson (1985: 252). As Frege (1983: 58) notes, we can write down " 3- = 4" without falsity. Thus, the success of the order that Socrates should be wise depends on that Socrates makes the proposition expressed true. Cf. Kirkham (1995: 28 ff.). See also G u p t a (1993: 78). See also Soames (1999: 229. 23). This is confirmed by the linguistic fact that in natural language the locutions "it is true that p" and "it is the case that p " (and also "it is a fact that p") are completely synonymous. The theory presented in Tarski (1944) has two readings, a semantic and a deflationary one.

References Dummett, M. (1959) "Truth." Proceedings of the Aristotelian Society, 59: 141-62. Frege, G. (1983) Nachgehissene Schriften urnI Wissenschuftlicher Briefwechsel, vol. i. ed. by H. Hermes, F. Kambartel. and F. Kaulbach. 2nd edn. Hamburg: Meiner. (1990) Kleine Schriften. ed. by I. Angelelli. 2nd edn. Hildesheim: Olms. (1997) The Frege-Reader, ed. by M. Beaney. Oxford: Basil Blackwell. Greimann, D. (2000) "The Judgement-stroke as a Truth-operator: A New Interpretation of the Logical Form of Sentences in Frege's Scientific Language," Erkenntnis. 52: 213-38. (2002) "Grundriss einer sprechaktanalytischen Theorie der Wahrheit." Zeitschrift für philosophische Forsclning. 56: 23 -51. (2004) "Frege's Puzzle about the Cognitive Function of the Concept of Truth." Inquiry, 47: 425-42. G u p t a . A. (1993) "A Critique of Deflationism." Philosophical Topics. 21: 57-81. Horwich, P. (1990) Truth. Oxford: Basil Blackwell.

Copyrighted Material

The assertoric use of the concept of truth

105

—(2001) "A Defense of Minimalism," Synthese, 126: 149-65. Kirkham, R. L. (1995) Theories of Truth. A Critical Introduction, Cambridge, MA: M I T Press. Levinson, St. (1985) Pragmatics. Cambridge: Cambridge University Press. Lewis, D. (1970) "General Semantics," Synthese, 22: 18-67; reprinted in D. Lewis (1983) Philosophical Papers I. Oxford: Oxford University Press, pp. 189 232. Searle, J. R. (1969) Speech Acts. An Essay in the Philosophy of Language, Cambridge: Cambridge University Press. Siegwart, G. (unpublished) "Prädikation," ms. Soames, S. (1999) Understanding Truth. New York: Oxford University Press. Tarski, A. (1944) "The Semantic Conception of Truth," Philosophy and Phenomenological Research. 4: 341-75. Williams, M. (1999) "Meaning and Deflationary Truth," Journal of Philosophy, 96: 545-64. Wright, C. (1992) Truth and Objectivity, Cambridge, MA: Harvard University Press.

Copyrighted Material

6

Assertion as a practice1 Gary Kemp

Austin began his f a m o u s essay " P e r f o r m a t i v e U t t e r a n c e s " (1961) by a t t e m p t i n g to distinguish a class of u t t e r a n c e s which serve n o t t o say something, but to do s o m e t h i n g , such as w a r n , apologize, o r promise. By the end, however, he retracted the idea. O n the o n e h a n d , there is the plain fact t h a t saying is a kind of doing. O n the o t h e r h a n d , n o such c o n t r a s t is s u p p o r t e d by t h e idea of the explicit p e r f o r m a t i v e verb: just as one can promise s o m e t h i n g by using the p e r f o r m a t i v e verb "I promise t h a t . . . , " o n e c a n state s o m e t h i n g by saying "I state t h a t . . . . " In neither case is the f u n c t i o n of the p e r f o r m a t i v e v e r b to describe the act that o n e is p e r f o r m i n g ; rather it serves at least p a r tially to effect it, even if sometimes to make explicit what sort of act it is, in the way that a sweep of the a r m indicates w i t h o u t stating t h a t o n e is b o w i n g r a t h e r t h a n s t o o p i n g t o check one's shoes. T h u s the features of a certain restricted class of p e r f o r m a t i v e verbs which first attracted Austin's a t t e n t i o n t u r n out to be m o r e general. T h e conclusion is that all utterances - o r rather all speech-acts - c a n be d e c o m p o s e d into the expression of a c o n t e n t a n d the a c c o m p l i s h m e n t of a certain variety of action: besides the q u e s t i o n that h a s been very m u c h studied in the past as t o w h a t a certain u t t e r a n c e means, there is a f u r t h e r q u e s t i o n distinct f r o m this as to what [is] the force, as we m a y call it, of the utterance. . . . W h a t we need besides the old d o c t r i n e a b o u t m e a n i n g s is a new d o c t r i n e a b o u t all the possible forces of utterances, t o w a r d s which o u r p r o p o s e d list of p e r f o r m a t i v e verbs would be a very great help. (Austin 1961: 238) Implicit in Austin's conclusion is that, since there is n o t h i n g in principle less p e r f o r m a t i v e a b o u t assertion t h a n there is a b o u t o t h e r illocutionary acts (to a d o p t his later terminology), we should be o p e n t o the possibility that assertion is c o n v e n t i o n - b o u n d in ways that parallel those of promising, apologizing a n d the like. It is n a t u r a l to think otherwise. It is n a t u r a l , for example, to t h i n k that whereas in some circumstances one's inner intention is powerless to prevent an u t t e r a n c e of "I p r o m i s e " or "I swear" f r o m constituting a promise or

Copyrighted Material

Assertion

as a practice

107

oath, assertion is categorically unlike this. Assertion, we naturally suppose, is individualistic in the sense that the matter of whether or not I assert a proposition I express is entirely determined by my intentions. As Frege noted, no sign or other gesture, not even the predicate "is true," could be sufficient for assertoric force (Frege 1979: 251). There is no d o u b t that speaker-intention has more executive power over assertion than it has over promising; in typical circumstances, it is relatively m u c h easier to wiggle out of an apparent assertion than it is an apparent promise or oath. M y aim in what follows is to respect this fact, but to advance a conception of assertion that accords with what I take to be the implicit point from Austin just mentioned. According to this conception, assertion is a rule-governed practice: to assert is to say something in such a way as to be subject to certain constitutive n o r m s (roughly, a constitutive norm is one whose jurisdiction is partly definitive of a practice). I shall take pains to explain why this should be so, that assertoric force should not be conceived entirely in terms of the speaker's intention. Any novelty in the theory itself lies in the detail of its shape and content, not in the very idea that assertion should be conceived as a practice or as essentially subject to norms. By contrast with some recent writers, however, I do not think that any one account of the practice of assertion can be quite correct: assertion, I will suggest, is best understood not as a single practice, but as a family of practices.

1 Norms I begin with some remarks about what might be called the metaphysics of norms. This is partly because I wish to avoid any hint of the idea that something's being subject to such-and-such a n o r m might be a kind of primitive or preternatural fact about it. There is some complacency in recent philosophy on this point, and I would not want an analysis that appeals to normativity to be anything but thoroughly sublunary. But further, the precise meaning of the claim that assertion is a practice depends critically on how the variety of n o r m envisaged as entering into the practice is conceived. In philosophy, the notion of a normative utterance is meant to capture what is c o m m o n to utterances that express what one ought or ought not to do, what it would be correct or incorrect to do, right or wrong to do. 2 They take either an imperative form, as in " D o n ' t bow with your h a n d s in your pockets" or, by explicitly using either a prescriptive modal auxiliary such as " o u g h t " or "should," or a normative adjective such as "correct" or "right," a declarative form, as in "You shouldn't bow with your h a n d s in your pockets," or "It is incorrect to bow with your hands in your pockets." We can distinguish n o r m s from statements of value or utility, such as "Generosity is g o o d . " Henceforth by a proper norm I will understand a normative proposition (not an imperative) that is necessarily true, that is, whose jurisdiction does not depend on matters of contingent fact. I suppose this is what is often meant by the notion of a rule. I take them to be paradigmatically

Copyrighted Material

108

Gary

Kemp

expressed in the explicitly deontological form "it is (in)correct to ® " or "It is right (wrong) to Φ" where Φ is a suitably general action-type, or the modal form " O n e must (not) "One ought (not) to Φ" or "One should (not) (I will not stop to spell out the degree or kind of generality required of a rule). I assume that n o r m s need not be explicit. Rules of grammar, for example, may be implicit, binding without having been explicitly promulgated or even formulated (rules of games may also have this status). We can distinguish some different types of norms, whether proper or not: • • • • • • • •

Practical n o r m s Moral/Ethical n o r m s Teleological n o r m s Cognitive n o r m s Instrumental n o r m s Constitutive n o r m s Conventional norms Absolute n o r m s

Members of the first g r o u p of four are distinguished by subject-matter; m e m b e r s of the second by more general features of their logical, epistemological or metaphysical status. Some members of the first group may subsume or overlap others. I assume that every m e m b e r of a category in the first group is a m e m b e r of some category of the second group. T h o u g h n o t h i n g hangs on this. I assume also that the second group, but not the first, exhausts the universe of n o r m s (there are also legal norms, religious norms, aesthetic norms, etc.). We need to have some points in place concerning the second group. An instrumental norm is a n o r m that is explicable in terms of a value, p u r p o s e or goal. So for example a rule-consequentialist in ethics holds that the authority of specific ethical n o r m s derives from their tendency, if followed, to bring about a value such as happiness. Proper instrumental norms, by what was said above, are necessary truths, such as "it is right to maximize happiness"; but we typically take these as read, and say such things as "It would be wrong to allow the toxins to escape." An improper instrumental n o r m is a true normative proposition that follows from proper instrumental n o r m s together with non-normative matters of fact (there is no need to stop over the vexed question of which embeddings of normative propositions in sentential operators are themselves normative propositions). A constitutive norm N for action-type Φ is one such that it is conceptually necessary that agents performing Φ who fully understand what it is to Φ thereby understand themselves to be subject to N. In this sense, it is essential to Φ-ing that those who self-consciously Φ appreciate their being subject to N. It is not necessary that one obey or conform to the n o r m in order to perform the associated action-type: although there are surely limits to this, even a cheater, one who breaks the rules of a game, may still be playing the

Copyrighted Material

Assertion

as a practice

109

game. T h a t an action-type is governed by a constitutive n o r m may in one useful sense be what we mean by calling the action-type a practice: an actiontype Φ is practice when it is a conceptual truth that to regard an act as an instance of Φ is to invoke certain standards of correctness. More on this below. We will not invoke conventional norms in the sequel but we should take pains to distinguish them from constitutive norms. Conventional n o r m s are non-constitutive n o r m s that only contingently govern some activity or other d e p a r t m e n t of living, but which are not derivable from proper instrumental or absolute n o r m s together with non-normative facts. T h e most conspicuous examples are explicable as solutions to social coordination problems: the correctness of driving on the left in the United K i n g d o m , that of using " L o n d o n " to refer to L o n d o n , for example. Some examples are more subtle: One can eat spaghetti by twirling it o n t o the fork on the plate, twirling o n t o the fork in the spoon, slurping without twirling, or cutting before piling on the fork; but in Italy, only the first is correct. It is not a constitutive n o r m of eating spaghetti that one should eat it that way: one who eats it an entirely different way is still eating spaghetti (even correctly so, if not in Italy). It is at least arguable that a room full of spaghetti eaters eating it in the same way is aesthetically a less displeasing spectacle than one full of them eating it however they like, and similarly for other customs where aesthetic value or social cohesion is served by uniformity (or at least was once so served; some customs survive their justifications). The extensional b o u n d a r y between constitutive rules and conventions is no d o u b t vague - as is the identity of practices over time or from place to place - but the distinction between the concepts is clear enough. For example, a practice similar to ours of gift-giving in which the "giver" is free to take back the "gift" might be thought not to be genuinely the same practice as our practice of gift-giving; the question clearly hangs on whether or not the n o r m that precludes taking it back is constitutive. Absolute norms are non-instrumental, necessary, and non-constitutive. I am going to make the simplifying assumption that there aren't any. The reason is that the two categories instrumental norms and constitutive norms provide models of intelligibility: they provide two ways of explaining what a n o r m is, of accounting for its character and status, that leaves minimal metaphysical residue. Kant's categorical imperative, or the Ten C o m m a n d ments, on some readings of these dicta, would be supposed examples of absolute norms. I d o actually take the notion of an absolute n o r m to be unintelligible, subject to Mackie-style objections from queerness; it seems to me to a m o u n t to a kind of mysticism or superstition to suppose that there are (non-conventional) rules that are neither constitutive n o r instrumental. But the relevant philosophical point for our purposes is the less contentious one that it is theoretically desirable to locate n o r m s in the metaphysically less problematic categories of the constitutive and instrumental: to do so is to regard them as explicable. We should only with great reluctance recognize n o r m s that are neither constitutive nor instrumental.

Copyrighted Material

110

Gary

Kemp

2 Intention and convention In this section, I will try to explain why we should conceive assertion as a practice, not as a matter of the speaker's intentions. The argument will be partly indirect, and partly an argument to the best explanation. Of course, a lot of philosophical industry has been devoted to sustaining the intentional view. I am certainly not going to recount the history of that industry. But I think it will be agreed that the character of the difficulties it faces suggest that it is bound to be unremunerative; and I think I can say enough to m a k e it plausible that the practice-view, while not itself beset by those difficulties, explains why they should arise for the intentional view. The relevant principle of argumentation is this: if. amongst competing hypotheses A and B. the truth of B would explain why A should be beset by certain difficulties whereas A c a n n o t do the same for B. then, everything else being equal, that is a good reason to prefer B. Let us consider one plausible construal of the intentional view. Intuitively and naively, assertoric force is something like the verbal expression of belief, whereby the agent utters a sentence that expresses, at the context of utterance, the content of the real or p u r p o r t e d belief (we assume that sentences may be uttered elliptically as in answer "yes" to a question, and that there are non-vocal ways of uttering such as writing). But it is not simply that, since presumably it is essential to assertion that it be a vehicle of or attempt at communication. If one sits alone in one's chair, and intentionally says something aloud with the intention of saying something that one believes, it seems wrong to say that one thereby asserts something. Like promising to oneself, it lacks the objective onus that assertion necessarily incurs. The reason would seem to be that one is not trying to communicate, not trying, nor trying to appear to try, to tell someone something. The centrality of the other-directedness is attested to by the general sanction against insincere assertion; such a blanket prohibition would seem unduly intrusive as well as unenforceable if assertion did not essentially involve an envisaged social transaction. Putting these points together, then, we may consider the following definition. I-assertion: A I-asserts p iff for some declarative sentence s and context c. A intentionally utters s at c, A knows that s at c means that p, 3 and A intends that for some audience B. A succeeds by uttering s at c to bring it a b o u t by virtue of B's understanding s at c that B believes that p and believes that A believes that p. T h e basic idea is that to assert is to try to infect an audience with a belief, or to seem to be trying to that is what we mean by the idea of telling someone something: the objective onus derives f r o m the fact that success incurs responsibility for the cognitive states of others. N o t e that since occurrences of " B " are inside the scope of "A intends," one can make an

Copyrighted Material

Assertion

as a practice

111

assertion when there is no audience. T h u s despite the communicative intention, the idea remains individualistic: provided he utters a sentence whose meaning he knows, whether or not the speaker makes an assertion depends entirely on his intentions. 4 But there are two problem cases for such an account. First, the necessity is d o u b t f u l of A's intention to persuade B: can we not imagine A s asserting that p when he hasn't the slightest hope of persuading B? Admittedly, if A has really given up trying to persuade B that p, yet goes on saying it, then certainly A's aim is uncharacteristic of assertion. A is not trying to inform or persuade B that p so much as to impress B with his sincerity, or the strength or rationality of his conviction, or something of that sort. But even so, it is hard to see how else to categorize what A is doing, but that he is carrying on asserting. Second, the intention to seem sincere can be lacking. Suppose A wishes to cast suspicion on himself when questioned by police (perhaps he wishes not too transparently to protect a friend, or has a perverse wish to be sent to prison, or some such). Affecting a pronounced nervousness, he says that he was at work when the crime was committed, hoping that the policeman will take this for a lie. In such a case, not only is the agent's intention to persuade the audience of p absent, so is the intention to persuade the audience that the agent believes that p. Again, it is hard to categorize what A is doing, except as asserting; he would, for example, go down in the police log as having said that, or stated that, he was at work when the crime was committed. We can try to avoid such counterexamples by complicating the definition. But I want to dispose of this sort of strategy in principle. For its defect, as I see it, is this. Attention to cases suggests that assertion can be undertaken with a variety of aims; certain aims may be characteristic of acts of assertion. but they are not going to be essential to it. in the sense that a verbal act is one of assertion only if the agent has those aims. Famously, Grice attempted to formulate a more complex combination of reflexive intentions to cope with various intricate sorts of examples (Grice 1957, 1969). I am not going to prove that such an endeavor cannot succeed. I d o d o u b t that any such characterization will intuitively be counterexample-proof, and I do think it prima facie implausible that the concept of assertion should turn out to be so complex. But there are more f u n d a m e n t a l grounds for dissatisfaction with the Gricean project. Grice's original aim was to characterize a concept of an agent's meaning that p by a certain act without presupposing we know what it means to call the act linguistic. As we consider more complicated candidate cases for Gricean " n o n - n a t u r a l meaning," confidence in intuition tends to dwindle (as Grice himself acknowledges); further, insofar as we d o find definite matches and mismatches between intuition and Gricean definitions, they tend to be unenlightening. If anything they dispel rather than enhance the conviction that something of central importance is being revealed. O u r concern here is narrower - unlike Grice, we are taking for granted the notion of the linguistic expression of a content - but the character

Copyrighted Material

112

Gary

Kemp

of the issues is much the same: the qualifications needed to cope with ramifying counterexamples are conspicuously ad hoc. What I think is going on is that insofar as c o m m o n sense possesses a stable concept of assertion, of telling or saying something, it is essentially linguistic, in such a way that reasonable extensions to non-linguistic cases - if there are any - presuppose it, and depend upon it as its model of intelligibility. At the core of this t h o u g h t is that assertoric force exists only by virtue of a certain institution. The reason that the intentionalist strategy flounders or lapses into unpalatable and uninformative complexity is that assertion must first be understood as a practice governed by certain rules, such that an act of assertion is explained as one that engages in that practice and is thereby subject to those rules. If so. then the potential variability of assertoric intentions is exactly what we should expect, not a baffling obstacle to the analysis of assertion. It is no more surprising t h a n the variability of intentions with which one might play a given game, make a promise, or give a present. Roughly, and in the most general terms, what makes non-verbal Gricean n o n - n a t u r a l meaning possible even in cases where the means employed are completely novel - is o u r prior consciousness as language-using creatures of the general idea of conventionb o u n d saying and signaling, along with our readiness to extend the forms of such transactions to novel means. This will. I think, become clearer once we have identified the sense in which assertion is conventional, that is, a practice.

3 Practices What is a practice? What would be the correct shape and character of a theoretical account of a practice? For all the talk of practices amongst philosophers roughly since the ascendance of Wittgenstein's later philosophy, such questions have not received much attention except on a case-by-case basis. I am not going to defend general answers to these questions here. In any case, I don't suppose that the concept of a practice indicates something like a natural kind, or that the extension staked out by a substantive and philosophically useful definition of the notion of a practice will include all and only those things that might ordinarily be called a "practice." As opposed to the verb, the noun is not even a particularly ordinary one; in philosophy, we might regard it as technical or as a term of art were its use in philosophy not so vague. To account fully for a variety of act Φ as a practice, I will assume, requires (at least) three things. First, we must specify the circumstances under which an agent engages in the practice Φ for the practice of Φ-ing, we must produce something informative of the form x Φs only if Fx. This will generally involve specifying something we may call the realizing action for Φ-ing, that is, an action specified without using the concept T h u s for example the realizing action for the practice of bowing would be something like bending over at the waist. Second and here is the substance of calling Φ a practice we must specify its constitutive n o r m s in the sense described

Copyrighted Material

Assertion

as a practice

113

above: we must produce something of the form x must Φ (only) if p, where the modality of " m u s t " is taken as normative. So practices are varieties of action governed by constitutive n o r m s (I a m not supposing that that statement expresses a sufficient condition of practice-hood). Third, we must characterize the function or purpose of the practice. Practices are in general either intentionally devised or naturally evolved social artefacts, selected for utility, in the broad sense whereby utility is simply that which explains the growth or persistence of the practice (practices are much like Dawkins's memes). Typically, if not always, the function will be to bring about some real or perceived benefit to a population or proper part of a population, such as economic, procreative, social, political, cognitive or other benefits.

4 Dummett If assertion is a practice, then we need to specify (1) what sorts of actions c o u n t as realizing it, (2) its constitutive norms, and (3) its function. I shall approach the problem via some points from the f a m o u s discussion of assertion in D u m m e t t ' s 1973 book on Frege. 5 Frege said that truth is the aim or goal of j u d g m e n t , in something like the way that catching fish is necessarily the aim of fishing (Frege 1979: 128). D u m m e t t considers the analogous idea that truth is the aim of assertion, and compares this with the idea that winning is the object of a game. Of course if someone asks, of some board game, "What is the object of the game?," only a wag would answer "To win"; for in just the way that the aim of any action is to succeed - the winning state of affairs is by definition the object of any competitive game. What the questioner wants to know is how winning is defined for this particular game. But that winning is by definition the object of the game does remind us that in order to know what winning is for a given game, we need to k n o w what the players standardly understand themselves and each other as trying to do. This tells us something about the general concept of winning. Consider two people who agree to "play chess," but with the aim of being checkmated. If that is what they are competing over, if that is the object of the game they are playing, then, speaking strictly, we should not say that they are playing chess whilst trying to lose, but playing a variant of chess whereby the winner, if any, is the player checkmated (like "lowball" in poker). C o m p a r e the " C r e t a n " hypothesis that the speakers of some language always aim to speak falsely: they always assert the negation of what they think - for example that the cat is not on the mat when they think it is, that it is when they think it isn't, and so on. As Wittgenstein observed, to suppose such an hypothesis to be genuine is to misunderstand such concepts as truth, negation and meaning. If we were to translate the supposed language, then rather than translating what the native says on the basis of his belief that p into an English sentence that means that not-p, we should translate it into one that means that p, thereby understanding him as intending to speak truly. The point is not

Copyrighted Material

114

Gary

Kemp

simply that the two hypotheses arc empirically equivalent. T h e reason for translating them as speaking truly is that insofar as their beliefs are true, their assertions would carry information about the world and express the speakers' beliefs in exactly the same way as those of any n o r m a l languagespeakers. Their actual, successfully realized aim would be to correlate meaningful sentences systematically with states of the world, exactly as it is for non-Cretans. T h u s D u m m e t t writes: Any possibility of asking, "Which of these two classes is the class of true sentences and which the class of false ones?", naturally depends on presupposing a prior understanding of the terms " t r u e " and "false". N o w what does this prior understanding consist o f ? . . . Clearly the only principle available is that according to which the use of assertoric sentences consists in our trying to utter only true ones: the class of true sentences is the class the utterance of a member of which a speaker of the language is aiming at when he employs what is recognisably the assertoric use. It is just because this is the only possible principle on which to discriminate between the class of true sentences and the class of false ones that there is no genuine possibility of a linguistic activity consisting in the attempt to utter only false sentences. ( D u m m e t t 1973: 320) If we knew of a final position for a b o a r d game that it constituted "winning" for one player and "losing" for the other, but didn't know what these terms meant, then much, if not all of the point of calling them that could be conveyed by saying that the object of the game is to achieve the one called the "winning" one. If so, then if truth is the aim of assertion, there ought to be some analogous way of conveying what truth is. There are, however, two i m p o r t a n t infelicities in D u m m e t t ' s discussion. First, the idea cannot simply be that truth is to be explained or defined as that at which asserting speakers aim, as if an assertion were true just in case it achieved its aim: that would make truths of successful lies. Second, in explicating the idea about the aim of assertion. D u m m e t t runs it together with another, namely that the idea of truth is that of the correctness of an assertion: there is n o way of saying what truth-values are save that they are the referents of sentences, and no way of distinguishing one truth-value f r o m the other save by adverting to the use of sentences to make assertions, and to that feature of this practice which makes it possible to ask whether someone is right or wrong in what he asserts. ( D u m m e t t 1973: 321; D u m m e t t ' s emphasis) The idea now is that the truth-condition of a sentence are the conditions under which its assertion would be correct. 6 But that truth is the correctness of an assertion and that truth is its aim are very different points. If winning

Copyrighted Material

Assertion

as a practice

115

is the aim of playing the game, and truth that of assertion, winning is not a kind of correctness. N o r is it very plausible that whatever it is that constitutes the correctness of an assertion must constitute its aim. Nevertheless, the normative point can be harmonized with the intentional point, and it is i m p o r t a n t to see how. When we say that winning is the aim of playing the game, we do not mean that every player always plays to win: a person may, in the individual case, play a game with the intention of losing. Likewise, a person may, in the individual case, m a k e an assertion with the intention that it be false. But this does not show that the intention to speak truly should not figure in an account of assertion. For it is clear that in the game case, despite the fact that players may occasionally intend to lose, the winning state of affairs remains the object of the game. Of course, which state of affairs is the winning one for a given game is a matter of convention, in something like the way that the meaning of a particular sentence is a matter of convention. If the parallel holds, then we should understand truth-telling to be the aim, not of acts of assertion, but of the practice of assertion. D u m m e t t ' s dictum that truth is the aim of assertion must be thought of as characterizing the practice of assertion, not as directly specifying conditions necessary for an act to be one of assertion: the linguistic acts should be classified as conventional actions, not as the external expression of interior states. Assertion, for example, is to be explained in terms of the conventions governing the use of those sentences which are understood as having assertoric force, not as the utterance of a sentence with the intention of expressing one's interior act of judgement (or interior state of belief) that it is true. ( D u m m e t t 1973: 311; see also p. 302) there is a general convention whereby the utterance of a [declarative] sentence, except in special contexts, is understood as being carried out with the intention of uttering a true sentence. ( D u m m e t t 1973: 298; cf. p. 302). We should not think of sentential utterances as mere expressions of content, and then ask what makes some of them assertions; we should note that sentences in the declarative mood constitute a "conventionally demarcated" type, and then ask what is the convention for using those. The m o o d of a sentence a matter of convention rather than intention - bears a necessary relation to the variety of force which is standardly attached to utterances of whole sentences in that m o o d (the idea of m o o d is not syntactical: if the syntactic forms which actually constitute a given mood were to be swapped with those that constitute another, then the m o o d s would have changed their syntactical form, not the variety of force which they customarily express). 7 But in what sense is truth the aim of assertion? H o w can that be a convention? Unfortunately the quotations above do not make this clear. Having

Copyrighted Material

116

Gary

Kemp

had the crucial insight that assertion is a practice. D u m m e t t is c o m p a r a tively careless in saying what the practice is. The second q u o t a t i o n directly above suggests something like: D-Assertion: x D-asserts that p iff for some declarative sentence s and context c, x intentionally utters s at c, x knows that s m e a n s that p at c, and c is a context with respect to which a linguistically competent audience would be entitled to believe that x intends, by uttering s at that context, to say something true. But this c a n n o t be right. 8 First, it is far f r o m clear in what sense an audience is always entitled to regard someone recognized as m a k i n g an assertion as trying to say something true. N o t only can any such entitlement or obligation be overridden (by awareness of the speaker's dishonesty), the circumstances under which such an entitlement or obligation could be overridden are surely t o o multifarious to circumscribe non-trivially. Second, the formulation fails to distinguish assertion from some other speech-acts such as guessing, (certain kinds of) assuming, and the swearing of oaths; perhaps the first and third o u g h t to be classified as varieties of assertion, but perhaps not, and certainly not the second. Third, there is nothing in the formulation that requires the other-directedness or communicative dimension of assertion that we emphasized earlier. 9 Fourth, and most fundamentally, the falsity of an assertion, according to this account, does not violate any convention or norm that is constitutive of assertion. If I say "the cat is on the m a t " in such a way that you are entitled to regard me as trying to say something true, and the cat is not on the mat, I speak falsely, but I do not violate D u m m e t t ' s convention. In fact, even if I a m not trying to say something true, then, although I no d o u b t violate n o r m s of some kind. I do not thereby violate D u m m e t t ' s convention - I merely exploit it. D u m m e t t ' s actual formulations proscribe neither insincerity nor falsehood. 1 0

5 Constitutive rules: truth, belief and knowledge We can d o better if we now invoke the general conception of practices outlined in Section 3. In particular, the way to follow D u m m e t t is to regard truth as a constitutive n o r m of assertion: that truth is the aim of assertion in the sense that truth is what asserting speakers are supposed to be aiming at. The idea is not the asserting speaker's attempt to tell the truth, but that there is a rule or n o r m that that is what they should he doing: that if they are not, then they are in violation. (Similarly - though perhaps here the sanctions are not so severe - a person playing a game is supposed to try to win; that is the convention of competitive game-playing, and we are rightly annoyed when our supposed o p p o n e n t violates it.) Such a rule might take either of two forms. Since normally one attempts to say something true only if one intentionally says something the content

Copyrighted Material

Assertion

as a practice

117

of which one believes, we could reconfigure Dummett's idea as d e m a n d i n g , at the very least, the following normative constraint on assertion: The belief-rule: O n e must assert that p only if one believes that p. This rule, taken as constitutive, distinguishes assertion f r o m guessing and assuming. T h u s the practice of AssertionB may be partially characterized as follows (the reason for the subscript will emerge): (1) its realizing actions are (possibly implicit or elliptical) utterances of declarative sentences; (2) its constitutive rule is the belief-rule (we will return to the question of the p u r p o s e of assertion). To identify assertion with Assertion B , however, does not imply that untrue assertions are incorrect. 11 To deliver that result, we should have to assume the following as constitutive: The truth-rule:

One must assert that p only if it is true that p.

Let AssertionT be the practice that is just like Assertions except with the truth-rule written for the belief-rule; let Assertion TB be the practice with the two rules together as its only constitutive rules. I think there will not be any d o u b t that the belief-rule is true, and that if assertion is a practice in the sense I have described then the belief-rule is one of its constitutive rules. So let us d r o p Assertion T f r o m the picture. Is the truth-rule both true and constitutive of assertion? One d o u b t is that when a speaker speaks sincerely but falsely the kind of criticism he is in for is quite different f r o m what a mendacious speaker is in for. Unlike the mendacious speaker, the sincere speaker means well, and we d o not charge him with anything like cheating or abusing the sacred institution of assertion. His failure, we might think, is not a violation; it is more like that of the game-player who loses, the m a r k s m a n who misses the target, t h a n the rulebreaker. But this difference is perfectly consistent with the truth-rule, and is easily explained. The mendacious speaker is subject to the criticism that he has wilfully misused the practice of assertion for an immoral end, namely deception. The sincere teller of a falsehood is not subject to criticism of that character or vehemence. But not all incorrectness is moral incorrectness; not all n o r m s are moral ones. Failure to d o something right may be failure of skill, perception, reasoning or even of circumstance, as when a tickle in the throat prevents us from singing the correct note. M a n y kinds of incorrectness are excusable; circumstances, even if they do not justify them, mitigate our failures, reduce or eliminate our responsibility for them. T h u s the fact that we do not always criticize assertions merely for being false does not indicate that they are not, all the same, incorrect, in violation of constitutive norms. Henceforth I assume that the truth-rule is both true and constitutive. What is correct in Dummett's idea that truth is the aim of assertion is that truth-telling is necessarily what asserting speakers ought to be aiming at.

Copyrighted Material

118

Gary

Kemp

T h e r e are. however, f u r t h e r c a n d i d a t e s for c o n s t i t u t i v e rules of a s s e r t i o n . F o r one, t h e r e is: The justification-rule: belief t h a t p.

O n e m u s t assert t h a t p only if o n e h a s a justified

T h i s m i g h t be d e f e n d e d o n the g r o u n d s t h a t i n s o f a r as a s s e r t i n g s p e a k e r s a r e r e s p o n s i b l e for t r u t h , it is a v i o l a t i o n n o t to have t a k e n sufficient c a r e t o e n s u r e it, even if the a s s e r t i o n is p e r a d v e n t u r e true. We m a y t h u s define Assertion J B as just like AssertionTB except with the t r u t h - r u l e r e p l a c e d by the j u s t i f i c a t i o n - r u l e . A n d since the j u s t i f i c a t i o n - r u l e c o m b i n e d with t h e t r u t h - r u l e yields a r e q u i r e m e n t of justified-true-belief, we m a y similarly d e f i n e Assertion J T B W i l l i a m s o n (1996) h o w e v e r c l a i m s that the t r u t h - r u l e c a n itself deliver s o m e t h i n g close to the j u s t i f i c a t i o n - r u l e by a p p e a l to the following: (*) If o n e m u s t (Φ only if p), then o n e s h o u l d (Φ only if o n e h a s evidence t h a t p). A s W i l l i a m s o n p u t s it. a rule t h a t f o r b i d s Φ-ing unless p provides a r e a s o n not t o Φ w i t h o u t evidence t h a t p. T h u s if o n e m u s t assert t h a t the cat is o n t h e m a t only if the cat is o n the m a t , t h e n o n e s h o u l d d o so only if o n e h a s e v i d e n c e t h a t it is ( w h e r e a d m i s s i b l e evidence w o u l d i n c l u d e perceiving t h a t very state of affairs). (*) is n o t itself a n o r m , since its m a i n o p e r a t o r is a c o n d i t i o n a l , n o t a n o r m a t i v e t e r m . It is s o m e t h i n g like a n a p r i o r i , n e c e s s a r y t r u t h of rationality. Is the resulting evidential r e q u i r e m e n t e q u i v a l e n t t o t h e j u s t i f i c a t i o n - r u l e ? It d e p e n d s here o n how s t r o n g the evidential r e q u i r e m e n t is: strictly s p e a k i n g . (*) r e q u i r e s only t h a t o n e s h o u l d not Φ if o n e h a s zero evidence t h a t p. O n e c a n certainly have evidence t h a t p w h e r e o n e w o u l d n o t be justified in believing t h a t p. For simplicity a n d convenience, let us t a k e the n o r m in the s t r o n g e r f o r m a c c o r d i n g t o w h i c h " h a v i n g evidence t h a t p " m e a n s " h a v i n g evidence t h a t justifies believing that p." We will r e t u r n t o this issue. T h e c o n j u n c t i o n of (*) with the p r e m i s e t h a t the t r u t h - r u l e is c o n s t i t u t i v e d o e s n o t entail that the j u s t i f i c a t i o n - r u l e is c o n s t i t u t i v e (even if we r e a d " o n e s h o u l d assert that p only if o n e h a s evidence t h a t p " as a n o t a t i o n a l v a r i a n t of the j u s t i f i c a t i o n - r u l e ) ; it only entails t h a t it is true ( a n d b o t h necessary a n d a priori g r a n t e d that c o n s t i t u t i v e rules as well as (*) are n e c e s s a r y a n d a priori). In f a c t a c e r t a i n principle of e c o n o m y suggests t h a t if (*) is g r a n t e d , t h e n the j u s t i f i c a t i o n - r u l e is not c o n s t i t u t i v e : if a n o r m N * follows f r o m a n a priori necessary principle of r a t i o n a l i t y t o g e t h e r with a n o r m N t h a t is c o n s t i t u t i v e of a given practice, t h e n the N * is not itself c o n s t i t u t i v e of the practice. T h a t is t o say. d e f i n i t i o n s s h o u l d be m i n i m a l , a n d n o t include w h a t will a l r e a d y flow f r o m t h e m by v i r t u e of a p r i o r i n e c e s s a r y principles. But the spirit of W i l l i a m s o n ' s p o i n t r e m a i n s : g r a n t e d

Copyrighted Material

Assertion

as a practice

119

the principle, sufficiently reflective, rational agents will know themselves to be bound by the justification-rule when making assertions. Why does Williamson write " s h o u l d " in the consequent of (*) rather than "must"? One reason might to be to avoid a regress. If we read " s h o u l d " as " m u s t " in (*) then (*) entails that (*) If one must (Φ only if p), then one must (Φ only if one has evidence that one has evidence that p). . . . and so on. Perhaps this is acceptable on some readings of "has evidence," but it rules out that the relevant evidence should be basic, i.e., such that there cannot be evidence for it. In order to avoid the regress, we have to understand the modality expressed by " s h o u l d " as differing from that expressed by " m u s t . " Normally we hear " s h o u l d " as weaker than "must": it expresses something more like advice than a requirement. So long as we accept that there is some such difference whereby " s h o u l d " does not entail "must," regress is avoided. 1 2 An argument similar to Williamson's suggests that the truth-rule delivers the belief-rule. The following is no less compelling t h a n (*): (**) If one must (Φ only if p). then one should (Φ only if one believes that p). Suppose, for example, that there is a rule that one must not enter the club unless one is carrying at least £10 cash. Suppose you are well aware of the rule a n d accept it fully. If you enter the club without a belief one way or the other as to whether or not you're carrying £10 cash, but in fact you are carrying it, then still you have acted wrongly. You have not sought to obey the rule: you have knowingly acted as if you attached no importance whatsoever to a rule whose jurisdiction you fully appreciate. (**) characterizes the sort of violation this is: it proscribes the actions of an agent who does not seek to regulate his actions according to a specific rule of conduct, even where the o u t c o m e happens to c o n f o r m to it. One might think (**) is more basic than (*), on the grounds that its conjunction with the following principle entails (*): (***) One should believe that p only if one has evidence that p. . . . so long as evidence is understood to include self-evidence, including for example perceiving the very state of affairs that p. In that case the truth-rule alone would establish a justified-true-belief rule (provided there c a n n o t be prudential reasons for belief). In any case there is one more rule to consider. For the central claim of Williamson's important paper is that mere warranted true belief, even very high and amply warranted credence in a true proposition, is insufficient for

Copyrighted Material

120

Gary

Kemp

correctness of assertion. For example, suppose that in 1983 astronomers had detected a large asteroid on a path to a p p r o a c h E a r t h in 1998, a n d justifiably set the epistemic probability of its striking E a r t h at about one in 100,000. As it turned out, it missed. Still, if someone had said, u p o n hearing the a n n o u n c e m e n t , " T h e asteroid will not strike E a r t h , " the assertion would have been out of order, despite its high evidential probability a n d actual truth. The reason, according to Williamson, is that the speaker would not have known that the asteroid was not going to strike Earth; he would have lacked the epistemic authority required of the asserting speaker. T h e point seems to be corroborated by the many varieties of justified-true-belief without knowledge that have been described since Gettier's original discussion; attention to such cases suggests that correctness of assertion tracks knowledge, not merely justified-true-belief or warranted high credence in a true proposition. Williamson thus proposes the following as a constitutive n o r m of assertion: 1 3 The knowledge-rule:

One must: assert that p only if one knows that p.

We may thus define AssertionK in the obvious way. We have mentioned several candidates for norms constitutive of the practice of assertion, thereby partially characterizing several practices: Assertion T B , Assertion B Assertion T Assertion J T B , A s s e r t i o n k . There is much m o r e to say as why we might prefer some to others. First, however, we need to say something more general a b o u t what it is to make an assertion, granted some particular constitutive rule.

6 Defining the act A n act of assertion will involve uttering, perhaps only implicitly, a declarative sentence. What earlier I called the realizing action for assertion will be an act of sentential uttering, however sententious. Now, although I rejected the idea that assertion can be characterized solely in terms of the agent's intention, intention is not irrelevant. The balance to be struck between intention a n d convention is delicate. I propose that we can strike it in the following way, where " N " is replaced by a sentence expressing the satisfaction of the n o r m that is constitutive of assertion, and the subscript "v" is schematic: Assertion : x assertsv that p iff for some declarative sentence s and context c, x knows that s means that p at c, x intentionally utters s at c, and in doing so, x successfully intends that he must not do so if not-N. T h u s for example A s s e r t i o n T B a n d Assertion K come out as: Assertion TB'- X asserts TB that p iff for some declarative sentence s and context c, x knows that s means that p at c, x intentionally utters s at c,

Copyrighted Material

Assertion

as a practice

121

and in doing so, x successfully intends that he must not d o so if he does not have a true belief that p. AssertionK: x assertsK that p iff for some declarative sentence s and context c, x knows that s means that p at c, x intentionally utters s at c, and in doing so, x successfully intends that he must not do so if he does not know that p. Assertion that p must be intentional in a way that goes beyond the mere intention to utter a sentence that means that p. Assertion is necessarily subject to constitutive norms. The simplest way to draw these points together is to hold that the requisite f u r t h e r intention is simply the intention, successfully realized, to be subject to the n o r m . 1 4 Admittedly, even leaving aside the question of which n o r m s are really constitutive of assertion, such an account remains incomplete in one conspicuous respect. For example, there might be a custom of solitary recitation of declarative sentences before one goes to bed, in such a way that those undertaking the practice understand that they would be violating a constitutive rule if any of the uttered sentences expressed untrue propositions or propositions not k n o w n to the speaker. Such a practice, surely, would not be one of assertion: it would have nothing intrinsically to d o with communication, which, at some level, assertion surely does. In order to address this point we must admit that a complete account of assertion should specify the conditions under which a speaker successfully intends to be subject to the relevant constitutive rules. As at least a comp o n e n t of this, we should require of the speaker a belief perhaps a warranted belief - that some suitable audience is or might be privy to the utterance, where a suitable audience is one that would, upon appropriately perceiving the utterance and relevant circumstances, be entitled or obligated to believe that the speaker intends to be subject to the constitutive rule of assertion, and would hold him subject to it. 15 However, I think we should not expect an informative and complete specification of the circumstances under which an audience would be entitled or obligated to hold a speaker subject to the belief and truth-norms. Too much depends on the epistemic condition of the potential audience - on what they have to k n o w in order to be entitled or obligated to impose the n o r m s of assertion - which may be quite peculiar to the circumstances. For example, if an actor were to shout "Fire!" during a p e r f o r m a n c e - really meaning to alert the audience to a fire - the audience might well not be entitled or obligated to hold him subject to the rules constitutive of assertion; in order to be so entitled or obligated, they would have to realize somehow that this was not part of the play. Certainly conventions do not tell the audience how to respond in this and in any many other less acute situations that d o or might crop up. and we should not assume that there is always a fact of the m a t t e r as to what a speaker has done. In any case, as I urged at the end of the

Copyrighted Material

122

Gary

Kemp

section on I-assertion, the thing to resist is the temptation to think that what really settles the matter either way is the intention of the speaker. The question of whether a speech-act a m o u n t s to assertion is settled by whether or not he successfully intends to be subject to certain norms; whether or not that is so may turn out to be a relatively untidy business. M o r e on this in closing. There is one more quite general point that needs emphasizing about what is involved in conceiving assertion as a practice. Williamson characterizes a n o r m ' s being constitutive of an action-type simply as the action-type's necessarily rather t h a n contingently being subject to the norm, and does not explicitly require that self-conscious agents undertaking the action-type understand themselves to be subject to it. I think the requirement is obligatory, and worth m a k i n g explicit. The idea behind a constitutive n o r m the point of calling it "constitutive" is that the n o r m is definitive of the action-type, that as a matter of conceptual necessity, it is at least part of what individuates it (see Williamson 1996: 492). That an action type Φ is necessarily constrained by a n o r m One ought to Φ only if p does not capture the idea of a constitutive n o r m , since that proposition might simply be a necessary truth in its own right, having nothing specifically to do with ing. 1 6 The idea that Φ-ing is a practice is largely the idea that it is a social construction - that is, that the explanation for why its realizing action is under certain circumstances subject to the n o r m is that the social group imposes it, or understands such acts to be subject to it. It is not that certain acts just are, as a matter of inscrutable fact, subject to rules in some Platonic Heaven. This is not to suggest that for example the truth-rule might be merely parochial: its jurisdiction might be a necessary condition of there being a linguistic community in the first place.

7 Which practice? We now return to the question earlier left hanging: Which c a n d i d a t e - n o r m captures our actual practice of assertion? Let us begin with an objection to Williamson's knowledge-rule. We may grant. I think, that if the knowledgerule is true, then it is constitutive (there is room for supposing it be an instrumental but non-constitutive n o r m , but ignore that possibility for the m o m e n t ) . T h u s imagine a c o m m u n i t y that spoke a language just like English. Suppose they engaged in something that is either assertion or something very like it, but did not hold these acts of saying to be subject to the knowledge-rule. For example, suppose one of the natives buys a lottery ticket with a one-in-a-million chance of winning. In advance of the draw, and knowing nothing germane not known to the first native, a second native says to the first. "Your ticket didn't win." Assume the first native's ticket is in fact a loser, and that no native witness to the speech-act would find the second native's utterance out of order. According to Williamson, if the second native had thereby m a d e an assertion, then he would have violated

Copyrighted Material

Assertion

as a practice

123

the knowledge-rule, since despite being nearly certain of something true on good evidence, he does not know that the ticket didn't win. If we assume, as I think we should, that if no native would find the second native's utterance out of order, then it cannot have been out order, and the knowledge-rule could not be constitutive of the native practice (assume the natives are reasonable, self-conscious concerning their own practices, and so on). T h u s in order to hold on to Williamson's claim, we should have to deny that the native practice is one of assertion - in which case the second native did not assert that the first native's ticket didn't win. This seems quite wrong. Of course, "assertion" is not a very c o m m o n word; but it seems clear that the native would have said that the ticket didn't win; he stated it, made that claim. In order to address this kind of d o u b t about the knowledge-rule, we have to acknowledge and discharge a c o m m i t m e n t that we have so far held at bay that is incurred by characterizing assertion as a practice. Assertion, I have argued, must be understood as a practice in the sense described. It is the availability of such a practice - the fact that it is open to an agent to act so as to exemplify it - that enables the agent to make assertions with varying communicative or social purposes, so long as his intention is realized that he be subject to the n o r m s constitutive of the practice. But according to what we said earlier, practices - and their accompanying sets of conventions - necessarily have functions, and the functions that they have are essential to them. A theoretical understanding of a practice therefore requires an understanding of its function, its purpose 1 7 In order to understand the practice of bowing, for example, it is not sufficient to describe how correctly to d o it; we have to explain the role the practice plays in a society where the practice prevails. What is the purpose of the practice of assertion? According to o u r reconstruction of D u m m e t t , the asserting speaker is obligated to try to tell the truth, but that does not establish that truth-telling is the function of the practice of assertion. If the practice has an essential or characteristic aim, then the aim is something like communication, the transmission of information telling someone something, or letting them know what you think, and so on. 1 8 If Williamson is right, then we should go a step further and say that the function of the practice of assertion is the transmission of knowledge. (Which is not to say that knowledge might exist where there is no practice of assertion, as if the practice might consciously have been devised by knowledge-endowed beings in order to transmit knowledge.) If we grant that reflective participants in this practice grasp this, then they must understand asserting speakers as either endeavouring to transmit knowledge, or as endeavoring to deceive an audience into believing that they are endeavoring to do so. T h u s if the knowledge-rule is constitutive of assertion, it is so for the following reason. The function of the practice is the transmission of knowledge, and the way in which a certain kind of sentenceuttering performs that function is by the participants' holding it subject to the

Copyrighted Material

124

Gary

Kemp

knowledge-rule. The knowledge-rule is constitutive because assertion is a socio-functional concept, and the rule's being imposed by the c o m m u n i t y is precisely the m a n n e r in which the function of the practice is realized. But if assertion is a socio-functional concept, why should its function not be simply the transmission of justified-true-belief, or of mere true belief? Williamson's observations about the lottery-paradox, especially together with related observations concerning justified-true-belief without knowledge, d o suggest that the function of assertion is the transmission of knowledge as opposed to true belief or justified-true-belief. A n d m o r e general theoretical c o r r o b o r a t i o n can be adduced. As epistemically advanced beings, we d o in point of fact value knowledge over true belief. What is more, we value knowledge over justified-true-belief. T h e rationale for this preference is n o t easily explained, since it will depend on what is the right or best way to explain the difference between knowledge a n d justified-truebelief as well as an account of the nature of justification; we c a n n o t defend Williamson's position fully on this point without some heavy epistemology. But let us assume that this preference is b o t h actual and justified. If so, then arguably a linguistic practice whose function is to transmit knowledge is cognitively m o r e valuable than one whose function is to transmit justifiedtrue-belief. By evolutionary principles of social selection, practices along with their constitutive rules tend to reflect values, in which case the former practice is more likely to be the actual practice. Since this is corroborated by intuition, the case for Williamson's knowledge-rule is perhaps as compelling as can be expected, where this sort of issue is concerned.

8 Pluralism Even so, however - even if Williamson has correctly characterized our practice - the point about its p u r p o s e does not show that the native practice imagined earlier would not be one of assertion. T h a t the lottery assertion violates a constitutive rule does not establish that the knowledge-rule is the constitutive rule of assertion, because nothing has been said to establish that the concept of assertion is univocal, that is, that it stands for a single practice rather t h a n a family of practices, actual or possible. But the native case was only a softening-up exercise: I think that if we look f u r t h e r r o u n d our own practices, we find it hard to accept that everything rightly called "assertion" is subject to the knowledge-rule. This not because I think there is some other rule that is constitutive of assertion. I think it appropriate to sound a note of skepticism about the whole enterprise of defining assertion. In the course of this paper, for example, I have m a d e a great m a n y assertions. But I d o not believe that they are subject to the knowledge-rule. T h e point is not simply that I d o not k n o w very many of the propositions I've asserted to be true, and personally find it hard to accept that I a m therefore violating norms, despite my best intentions. The point rather is that I am not atypical in this respect, and it is not plausible that philosophy

Copyrighted Material

Assertion

as a practice

125

is typically in violation. That is because the function of philosophical discourse c a n n o t very plausibly be said to be that of transmitting knowledge, nor even that of truths. Of course it is ultimately if sometimes quixotically aimed at truth and knowledge, but the function of actual philosophical assertion - as opposed to the overarching aim of the whole enterprise - is not to transmit it: it is something like that of developing arguments, establishing pathways of reason through virgin or otherwise badly marked conceptual territory. I d o not care to try to characterize the nature of philosophy here, but it seems to me perfectly evident that the s t a n d a r d s governing what we say shift significantly when we enter the seminar room. This can remain so even if we grant that the ultimate aim of philosophy is knowledge. For a n o t h e r case, take scientific assertion. There are good reasons to deny that scientific theories are proper objects of knowledge; and even - especially where theoretical physics is concerned not proper objects of belief, but intuitively this does not imply that they should not be asserted. T h u s Patrick Maher's work on the acceptance of scientific theories suggests that the constitutive n o r m governing scientific assertion is not truth, not belief or evidence, but expected cognitive value a function of informativeness, probability and verisimilitude. M a h e r seems to hold that this holds for assertion generally, which in my view is not at all plausible, but the position is compelling with respect to his main target, scientific discourse. 1 9 Consider the case of the pupil under oral examination. If asked " W h a t is the atomic n u m b e r of Helium?," and the student answers "Two," it seems that he is, in this admittedly artificial circumstance, responsible for neither knowledge nor belief, but only truth; yet clearly he asserts that the atomic number of helium is two. Consider confession. It is hard to see how the car's having narrowly missed the child should impugn the confession of a man who sincerely and penitently says to the Priest, "I maliciously ran over that child." The language game here is different: only sincerity counts. Yet all confessions, it seems to me, are assertions. We may also wonder about talk of matters aesthetic, comic, and moral. The idea that we make assertions in these areas would survive, it seems to me, the discovery that there is no knowledge to be had in those areas. A related point concerns metaphorical assertion: if with Davidson we accept that they are typically false, this would not seem to show that they are not assertions or that they violate n o r m s of assertion; if we deny Davidson's claim, then still admission that successful ones do not express knowledge would not seem to disqualify them as assertions. Finally there are more idiosyncratic cases such as certain sorts of jokes. I serve my friend a cappuccino whom I'm in the habit of teasing about his weight. "It's m a d e with skimmed milk, I thought it might help with your weight," I tell him. I'm teasing. There are many ways in which one might reasonably describe what is going on in such a case, but it seems to me the most natural way is to say that I did say that the drink was m a d e with

Copyrighted Material

126

Gary

Kemp

skimmed milk; I told him so. What I was engaged in was a linguistic practice whereby I pretend to be sincere, but wherein my only actual responsibility is to avoid hurting the feelings of my friend or at least that any such disvalue should be outweighed by amusement value - and not to be either sincere or veracious. Is it assertion? It might seem unduly grave to call it that, but if assertion is telling, then since it seems I did tell my friend that the milk is skimmed milk. I did assert it. Again, I a m not going to defend a theory of teasing, of the comic or of the aesthetic let alone the moral or the scientific - and thus I am not hereby pretending to produce universally compelling counterexamples. But I hope it is at least plausible that some of these practices, these languagegames, t h o u g h they do d e m a n d justification of some sort from the agent, often serve purposes and involve standards that differ substantially f r o m those of knowledge-transmission. N o r am I claiming that there is no such thing as knowing that something is beautiful or funny. The point is to observe how thoroughgoing and wide-ranging the metaphysical a n d epistemological implications of Williamson's thesis are. There is, as far as I can see, no reason to accept the thesis unless all these implications can be m a d e good independently not conspicuously likely - or there is some general reason to accept that the concept of assertion must be univocal - which I c a n n o t see that there is. The conclusion is that assertion is something like a family of practices, the concept a family resemblance concept. Different practices serve different purposes. T h o u g h the central ones, no d o u b t , serve epistemic purposes of knowledge-transmission, belief-transmission and truth-transmission, they are distinct, and not exhaustive. We should also recognize that some forms of assertion may rest upon rules engendered ad hoc and only tacitly or implicitly between very small groups, even pairs of speakers. The fiction of Henry James, for example, is sometimes directed at exploring certain delicate forms of understanding that arise spontaneously between certain refined sorts of people which might well be analyzed in terms of the a d o p tion of idiosyncratic conversational norms, ones which define the languagegame. Finally, if we admit to pluralism in this way, then we should expect maverick cases or cases that slip between the cracks, in which there is no fact of the matter as to which set of constitutive n o r m s are in play, hence no fact of the matter as to which practice a speaker is engaged in even where we are rightly confident that what he is engaged in is assertion. For that may m e a n simply that each practice to which the speech-act might reasonably be assimilated is a variety of assertion (especially where the speaker is unclear or confused about what he is doing, has no coherent intentions; that is often how it is, I think, with the keeping of diaries). For my part. I find n o t h i n g counterintuitive or c o n t r a r y to experience in the idea that there should sometimes fail to be a fact of the m a t t e r as to what a speaker is doing, or that an agent may, due to lack of understanding of the practices in view, fail to engage in the practice he intends, or fail to engage

Copyrighted Material

Assertion as a practice

127

in a n y at all. In such cases we m a y always fall b a c k u p o n a d e s c r i p t i o n in t e r m s of t h e agent's a c t u a l i n t e n t i o n s , h o w e v e r c o n f u s e d o r i n d e t e r m i n a t e they m a y be.

Notes 1 I thank Chris Belshaw, Richard Holton, David Lumsden, Philip Percival, John Skorupski and Crispin Wright for comments either on earlier fragments or on the oral version given at the Scots Philosophical Club meeting of December 2003. 2 Outside philosophy, a norm is merely something normal, i.e.. typical or usual, as when we say "By the 1970s, electric typewriters were the norm." 3 An utterance constitutes an assertion only if the speaker understands the sentence or rather, less demandingly, that the speaker knows what the sentence means; otherwise it might turn out that the sentence one intentionally enunciates expresses a belief one has only by chance: one interprets the sentence incorrectly, but adventitiously says something one believes. I lake it that one may know what a sentence means without understanding it; one might have no idea of its grammatical structure, let alone of how its meaning is derived from those of its parts; still one could make an assertion by means of it. One does this using phrase-books. 4 One might suppose that the social requirement is too strong: we could imagine A being stranded on a desert island, keeping a diary, without belief in its ever being read. Might A not be thereby make assertions? Perhaps A could be making assertions if B is A: A's later self might be the intended audience for As assertions. But suppose not: suppose A does not intend to read it himself. We can coherently imagine the case where A merely hopes that the diary will be read, knowing the chances of its being read to be very low. This is consistent with the above formulation, so long as we read "intends" in such a way that one can intend to do something despite believing that the chances of success are low (we must allow quite liberal substitutions for "B," such as "the person or people who find the diary"). But what if he hasn't any such hope? Perhaps linguistic intuition simply doesn't decide the case; at best, a defender of the intentionalist view may insist that denying that the castaway's utterances are assertions is not demonstrably incorrect. This kind of case is more readily made sense of in terms of the practice-view of assertion presented later. 5 My use of Dummett's discussion is highly selective, and should not be regarded as an attempt to capture Dummett's intentions. 6 When but only when - we turn to a particular language, that idea decomposes into a descriptive taxonomy of linguistic acts together with rules for their correctness, leaning on a generic notion of correctness. Famously, Dummett holds that the idea of such correctness must be specified as the outcome of a decision procedure, an effective canonical means of verification. The importance of this theme in Dummett's discussion should be acknowledged, and some indication given of why it can reasonably be bracketed. First, one could accept that truth is correctness but deny that correctness must be verifiable - though of course one would have to answer Dummett's arguments to the contrary. Second, the antirealist consequence that Dummett ultimately draws does not entail that "truth is the aim of assertion" is not a substantive claim: if truth is correctness, then still it is non-trivial that correctness is the aim of assertion; also the point acquires further substance when the particular standards of correctness appertaining to assertion are filled in. 7 Frege tried to register this point by saying that only declarative sentences express thoughts (later, that only declarative sentences and yes-no questions do); at

Copyrighted Material

128

8 9

10

11

12

13

14

15

16

Gary Kemp

another stage by saying that assertoric force is "bound up with the predicate" (Frege 1979: 252). And like Russell he held that a logically perspicuous language would contain a sign of assertion. Compare Pagin's (2004) formulation of Dummett's view (Dummett 1973: 18). It fails to accommodate Dummett's insistence on the conventional nature of assertion. Actually. Dummett acknowledges this requirement in summing up a provisional attempt at characterizing assertion strictly in intentional terms: "we might try the following formulation: a man makes an assertion if he says something in such a manner as deliberately to convey the impression of saying it with the overriding intention of saying something true" (Dummett 1973: 300). But he regards such a formulation as mistaken, and the other-directedness of assertion fails to appear so explicitly in the formulations he advocates later. Presumably there are other norms that, in conjunction with Dummett's convention, would impose such constraints on the speaker. For example, there might be a moral rule, or even a socio-epistemic norm, that, except in special conditions, one should not act in a way that is likely to mislead others about states of the world or about one's other mental states. But such derived norms would not be constitutive of assertion that is, ones whose appreciation is essential to an understanding of the concept of assertion. If Dummett is right to think of an assertion's falsity as its incorrectness, and not simply that of the underlying belief, if any, then here is where the analogy between winning and truth-telling breaks down: although there is arguably a norm of game-playing that one should try to win. there is no rule or norm whereby it is incorrect not to win. Avoiding the regress is also a reason to not express (*) as the higher-order norm "One must: [ensure that: if one must (Φ only if p) then (one Φs only if one has evidence that p)]. Assuming that 'one must' distributes over the conditional as 'necessarily" does, this entails (*)." I am taking a slight liberty here, as Williamson does not himself extend the point beyond lottery-type cases. The existence of non-lottery-type cases makes it much more difficult to explain them in terms of only a truth-rule together with Gricean maxims of cooperation (for such an attempt that fails for this reason, see Weiner 2004). It might be necessary to add that the intention must be reasonable as well as successful, that is, that it be reasonably believed that the intention is successfully realized, or that it be reasonably believed that the probability of the intention's being realized is nonzero. This will depend on precisely what is involved in being subject to the relevant norms. I will not pursue this possible complication here. So a liar intends to be subject to the norm? Yes: the liar intends that others should correctly take him to be subject to the norm; but one is correctly taken to be subject to a norm iff one is subject to it. C o m p a r e the insincere promiser: it is precisely because the promiser expects to have made a contract that he can expect to gain the advantage afforded by the promisee's expectation that he will keep the promise. To intend to be subject to a rule, of course, is not to intend to be subject to the sanctions that underwrite it. Where Φ is a practice, it is part of the content of saying that an agent performs Φ that the agent is subject to the norm. It does not quite follow that if N is a constitutive norm of a practice Φ. then necessarily anyone who performs Φ at least implicitly knows himself to be subject to N. It is certainly not the case that we credit someone as engaging in a practice only if they know all its constitutive rules (if we did. then we would hardly ever credit anyone with playing baseball!). The question of whether a norm is constitutive, rather, should be whether a person undertaking a practice who understands fully what the practice is must thereby know himself to be subject to the norm.

Copyrighted

Material

Assertion as a practice

129

17 Some practices perhaps outlive their functions; but the idea of a practice that never had a function is at best the idea of a degenerate case, a social spandrel. Other practices change functions; ! have chosen to characterize the function of a practice as essential to it. but I see nothing substantive hanging on this. 18 Similarly, winning is not the aim of playing the game in the sense that it is the purpose of playing the game; winning may be the object of the game - in the sense that the winning position is what the competing player seeks and a player who displays not even the pretence of trying to win destroys the amusement or other satisfactions whose seeking normally motivate play; but to win is not standardly why a given player plays the game, and it is not the purpose of a particular game like chess, or of games generally, to constitute ways of achieving an overarching goal of winning. 19 Actually Maher's account concerns belief, not assertion. Maher identifies his concept of acceptance with (categorical) belief, and argues that high probability is neither necessary nor sufficient for rational acceptance of a hypothesis. I think that Maher errs in identifying rational acceptance with rational belief. For one thing, rational belief is closed under logical consequence; rational acceptance is not. For another, it is too far from evident that a theorist who advances a theory and rationally accepts it - must thereby believe it. If it is true, as Maher suggests, that Einstein probably would have bet against the truth of relativity theory, then perhaps we ought to conclude that that attitude towards theories that Maher so compellingly describes, and which is characteristic of the scientist, is not that of belief.

References Austin, J. (1961) Philosophical Papers, Oxford: Oxford University Press. Dummett, M. (1973) Frege: Philosophy of Language, Cambridge, MA: Harvard University Press. Frege, G. (1979) Posthumous Writings, Chicago. IL: Chicago University Press. Grice. P. (1957) "Meaning," The Philosophical Review, 66: 377-88. — (1969) "Utterer's Meaning and Intentions," The Philosophical Review, 78: 147-77. Maher, P. (1993) Betting on Theories, Cambridge : Cambridge University Press. Pagin, P. (2004) "Is Assertion Social?." The Journal of Pragmatics, 36: 833-59. Weiner, M. (2004) "Must We Know What We Say?," ms. Williamson. T. (1996) "Knowing and Asserting," The Philosophical Review, 105: 489-523.

Copyrighted Material

7

Truth, assertion and the sentence1 Wolfram

Hinzen

1 Introduction U n d e r s t a n d i n g the illocutionary role of t r u t h , it is said, is to u n d e r s t a n d the role of the concept of t r u t h in the p e r f o r m a n c e of the speech act of assertion. T h a t could m e a n that assertion is actually the source, a n d p e r h a p s the e x p l a n a t i o n , of o u r concept of t r u t h . In this paper, I will deny some such c o n t e n t i o n . For how tight is the connection between assertion, as a variety of h u m a n language use, a n d the concept of t r u t h ? These two n o t i o n s certainly seem inextricably linked. But are they? C o u l d we not have h a d some practice of assertion a n d c o m m u n i c a t i o n , but n o concept of t r u t h ? T h e answer, I will here suggest, is yes. Moreover, given that we have the n o t i o n of t r u t h , o u r grasp of it d o e s not seem exhausted by its role in the practice of assertion. For a l t h o u g h assertion in actual linguistic practice d e p e n d s on a n o t i o n of t r u t h what we assert in language is the t r u t h of s o m e t h i n g we d o n ' t c o n c l u d e t r u t h f r o m whatever was. is. or will be asserted. Both of these p o i n t s suggest that assertion, as a practice, a n d t r u t h , as a h u m a n concept, live at least partially independent lives, a n d have potentially indep e n d e n t evolutionary histories, which raises the question of what provides for their linkage, which clearly exists. T h a t linkage, I argue here, may be the h u m a n sentence a particular structural configuration that is uniquely evaluable for t r u t h , a n d that may have given rise to a f o r m of assertoric practice where assertion intrinsically is the assertion of t r u t h . 2 To anticipate some of the m o t i v a t i o n for this. I will claim t h a t there is n o t h i n g in a practice of assertion as such that would necessarily give rise t o a n o t i o n of t r u t h , if a linguistic c o m m u n i t y lacked that n o t i o n before. Such a c o m m u n i t y might for example be o n e where a language of a n o t h e r structural type evolved, in which there were n o sentences. In that c o u n t e r f a c t u a l case, there would be no t r u t h , which we uniquely ascribe to the c o n t e n t s of sentences, as o p p o s e d to the c o n t e n t of o t h e r syntactic structures. For example, it is simply not g r a m m a t i c a l to say that "the n u m b e r 2 is true." o r that " M a d o n n a " is.3 In essence, we need a sentence introduced by a so-called complementizer, such as " t h a t " or " f o r " in " t h a t G o d is dead is true." a n d " f o r that to be true, G o d has to be d e a d . " In this way, a distinction emerges

Copyrighted Material

Truth, assertion and the sentence

131

between syntactic objects suited for purposes of truth (or the predication of it), on the one hand, and syntactic objects suited for purposes of reference (noun phrases such as " M a d o n n a , " "the number 2," etc.), on the other. (That distinction will be scrutinized in Section 2.) But even though the concept of truth, at least in one of its grammaticalizations, is tied to the notion of (complemented) sentence, 4 and thus the counterfactual c o m m u n i t y just envisaged would lack truth, it is unclear whether assertion needs to be tied to the sentence in the same way. True, when engaging in the practice of assertion, we essentially use sentential structures. But then again, there seems to be no reason to d o u b t that our counterfactual c o m m u n i t y could more or less do with language what we do: say what happens, tell their opinions, incur a c o m m i t m e n t , tell jokes, chat over the latest events, etc. Indeed, it would seem that if language evolution had taken a slightly different course, with sentences not arising, this might be as a result of a sheer historical accident and could thus hardly be taken as a good reason for d o u b t i n g that we therefore would not have engaged in the above sorts of innocent activites, or have a f u n d a mentally different kind of mind. If that intuition held water, sentences would be no necessary preconditions for using language to do essentially what we do with it. But in that case, sentences, while being no precondition for a practice of assertion, will he a precondition for the ascription of truth, to the extent that truth is essentially applied to sentences in h u m a n language, which is what we just saw. But then, it follows that we must decouple truth from assertion to at least some extent: we could have had assertion and some practice of semantic evaluation that it will involve, but not sentences, and hence not truth, again to the extent that truth as a concept is necessarily and distinctively applied to sentences or the kind of information that they uniquely encode. But there is another option too. What would assertion in that counterfactual linguistic community really be like? For us, assertion is the assertion of truth. Would that be what assertion in the alternative c o m m u n i t y also is? Suppose not. Then to whatever extent we think it's not - and a priori any such d o u b t seems quite sensible - we would arrive at the surprising view that the contingent structural format of h u m a n languages - all of which figure sentences as f u n d a m e n t a l linguistic units 5 - actually played an instrumental causal role in the genesis of h u m a n truth, that is the specific form of semantic evaluation that h u m a n s and probably only them engage in. This is because the difference between assertion in the counterfactual community and o u r own kind of assertion would boil down to a purely structural or syntactic one. It is this striking conclusion that I will ultimately argue we cannot easily dismiss. That is, either we think that the evolution of the sentence is quite irrelevant to the evolution of truth, and the link between truth and assertion is quite independent (and may be the cause) of it. Or, we think the evolution of the sentence was quite instrumental to the evolution of truth, and hence to the origin of a practice of assertion that distinctly involves it. Clearly,

Copyrighted Material

132

Wolfram Hinzen

which of these options is right is a question at the borders of current inquiry into the roots of h u m a n nature. Yet. this paper aims to open up this issue, and to make some coherent suggestions that may merit f u r t h e r inquiry. It proceeds, in the following section, by problematizing the semantic distinction between the two f u n d a m e n t a l h u m a n semantic base notions, reference and truth, and asks why this duality evolved, instead of a single semantic base notion, which I will call, following Carstairs-McCarthy (1999), "applicability." The option is envisaged here that the distinction between reference and truth is at b o t t o m a syntactic one. In Section 3. I elaborate on the argument sketched above that the evolution of a practice of assertion and semantic evaluation would not as such have to give rise to the evolution of the sentence, or depend on that evolution. Accordingly, it could also d o without a distinction between truth and reference to which the evolution of the sentence seems crucial. Section 4 scrutinizes the notion of truth as a h u m a n cognitive universal itself, reiterates arguments for its independence from the notion of assertion, and argues for the plausibility of the view that the origins of truth are indeed internalist ones: truth's existence does not merely derive from action-theoretic factors, on functionalist grounds, from adaptive or representational interactions with an environment, or from causal relations in which we stand to it. Section 5 states the analogous internalist conclusion for the structural format of language and the universal syntactic categories themselves, to which we have seen the evolution of truth is linked: the sentence, that is, does not plausibly arise f r o m the way we semantically evaluate language, and the causal arrow more plausibly points in the opposite direction. Sections 6 and 7 bolster this conclusion by insisting, with reference to current accounts of the architecture of the language faculty, that the fact that we m a p particular syntactic structures to truth and others to reference is no explanation of why these m a p p i n g s occur: semantics as such does not seem to motivate the structural format of language that is the basis of these mappings. Section 8 relates these results to familiar philosophical considerations on socalled truth-bearers and truth-makers. Section 9 concludes that public acts of assertion do not hold the key to the explanation and origin of the h u m a n notion of truth. Strikingly enough, that notion may not only have an important structural basis in the evolved design of the h u m a n linguistic mind, but that design may also fail to have an externalist rationale. H u m a n external functioning, as when engaging in acts of assertion, may be the consequence of an evolved cognitive structuring more than its cause.

2 Applicability It is virtually impossible for us - being all speakers of a h u m a n language, hence of a language with a specific and unique structural set-up - to ignore the intrinsic syntactic difference between sentences and Noun Phrases (NPs), which correlates with the semantic distinction between truth and reference. It

Copyrighted Material

Truth, assertion and the sentence

133

is virtually impossible for us even to conceive of what some propositional thought or assertoric content might be, unless we identify it through some sentence. But such a dissociation between assertoric content and sentences is precisely what we must try, since it is that possibility of dissociation that would provide evidence for or against the idea that the connection between the assertion of truth and sentences is a necessary and intrinsic one. One way in which it might not be is that contrary to first impressions there simply is no clear semantic categorical distinction between reference and truth. For if there wasn't, there would be no reason to ascribe a certain semantic power or potentiality specifically to sentences - namely, truthbearingness - that we couldn't ascribe to other syntactic objects. T h u s truth wouldn't be intrinsically sentential, and it would relate to assertion as much as reference would, no m a t t e r whether assertion exploits sentences or not. To put this differently, for truth to relate intrinsically to assertion it would have to relate to it in a way that reference would not. And that is not possible if reference and truth are semantically non-distinct. So let us ask whether there really is this distinction between crucially two semantic base notions, or two ways in which language may "fit the world." Let us try supposing that, instead, there is only one such way, and call it "applicability": expressions, no matter whether nominal or clausal, are either applicable to the world or not. 6 On this approach, "the present king of France" and "Pegasus" would both not be applicable, for example, and "there is a highest prime n u m b e r " would not be either, in the same sense and for the same reason: non-existence. In turn, "the present queen of E n g l a n d " and "Queen Elizabeth II" would be applicable, and "there is no highest prime n u m b e r " would be, too. Semantic knowledge has now nothing intrinsically to do with truth, because the notion of t r u t h as employed in the project of a truth-theoretic analysis of meaning depends (at least in part) on its contrast with the notion of reference. There is no such contrast between truth and reference in the method of applicability above. 7 The contrast is paradigmatically explicated with respect to an intuitive difference between how N P / D P s 8 such as "the present king of F r a n c e " or "Pegasus," on the one hand, and sentences such as "there is a highest prime number," on the other, have meaning. But in the above a p p r o a c h , there is no such distinction. There is only a syntactic distinction between two types of expressions. So far this may seem like a challenge that is easy to meet by a semantic theory appealing to both reference a n d truth. For a defender of such a semantic theory could simply ask for a motivation for giving up a reference/ truth approach for a novel - and surely somewhat strange - applicabilitytheoretic approach. A f t e r all. there is an intuitive distinction between reference and truth, one of these notions having to do with what we refer to, the other with what is true of the things we refer to. For this reason, the objection would go, there is something wrong and a distinction would be lost, if we said, of an expression like "there is no highest prime number," that it is

Copyrighted Material

134

Wolfram Hinzen

applicable rather than true, and saying the same thing of "the present queen of E n g l a n d . " However, this reply that, in the one case, we say something true, while in the other we (merely) refer to (or identify), say, a person - just reiterates a traditional distinction. All the applicability theorist has to say in reply to the objection is that the reference/truth distinction is a mere shadow of the syntactic NP/sentence distinction, but that, semantically speaking, there is no such distinction. Hence the defender of the semantic relevance of the reference/truth distinction has to substantiate what exactly makes reference different from truth. And here, as we shall now see, there seem to be some serious difficulties. To begin with, it is not unintuitive (nor is it ungrammatical) to say that the n a m e " G o e t h e " is true of Goethe, that "the a u t h o r of Wahlverwandtschaften" is also true of Goethe, and that "There is no highest prime n u m b e r " is true of worlds which are characterized by mathematics as we k n o w it presumably all possible worlds. But then, in all three cases - that of a name, a definite description, and a full clause we can use the "true o f " relation, and the intuitive truth/reference distinction collapses, inasmuch as it depends on the distinction between things like N P s and verb-headed clauses, of the sort usually thought to express propositions. T h e semantic distinction in question is immediately blurred also by recalling the p o p u l a r intuition that while a N P refers to a thing, a sentence also refers to something, namely a fact, a proposition, or a truth-maker. In other words, an assimilation of a referential view of meaning in the case of sentence-meaning has not struck philosophers as terribly far-fetched. C o n sider further that it is not unintuitive to say that a sentence like "Sam kicked Shaun yesterday" is the description of an event, as in a Neo-Davidsonian semantics: indeed we are talking about the event of Sam's kicking Shaun yesterday here. Descriptions are NPs, however, just as in the case of [ N P the a u t h o r of Wahlverwandtschaften], and NPs were said to belong to the paradigm of reference. What then, we may insist, is the f u n d a m e n t a l difference between describing an event, as when using "Sam kicked Shaun yesterday," and referring to it. as when using a nominal expression that is not sentential, like "the event of Sam's kicking Shaun yesterday"? Are our intuitions in this regard in any way firm? Why should we not say that we refer to something when saying "There is a car crash outside"- namely, a car crash outside? Or. that we refer to its occurrence? Or. perhaps, that we refer to a truth (rather than a possibility, say), as in: "That there is a car crash outside is a truth, not a possibility?" 9 Or. finally, that we refer to the truth (namely, the truth that there is a car crash outside, as referred to in the assertion "The truth is that there is a car crash outside")? Philosophical meaning of the worth, do not expression that

tradition aside, my conceptual intuitions regarding the ordinary words "reference" and "truth," for what they are exclude any of these possibilities. But in all of them, an is standardly thought to be an exemplar of a sentential

Copyrighted Material

Truth, assertion and the sentence

135

expression that is true or false would be unmasked as an exemplar of reference, and the m e t h o d of applicability might be rethought, after all. (In Section 3 I will address the objection, touched upon in Section 1, that one could not use definite descriptions of the form "the event of Sam's kicking Shaun yesterday" to make assertions, and hence the idea that the semantic difference we are looking for lies right here.) N o t e f u r t h e r that in a sense, truth-theoretic semantics in its modeltheoretic f o r m formalizes something like an "applicability-theoretic" view by m a p p i n g any syntactic category to a particular semantic value. The notion of a "semantic value" is, after all, nothing other than a technical correlate for the intuitive notion of reference, and satisfaction and truth are both special cases of referential values. 1 0 The semantic values are different in the two relevant cases, to be sure, and N P s refer to set-theoretic objects, while verbs refer to set-theoretic functions and sentences to t r u t h values, say, but the semantic relation of satisfiability as such does not vary. All that said, I hasten to admit that there is of course an intuitive difference between the sentential "Sam kicked Shaun yesterday" and the nominals "Sam's kicking of Shaun yesterday" or "the kicking of Shaun by Sam yesterday." But then again, does that difference relate to some crucial semantic difference between reference and truth, or is it merely a difference in what we refer to? The latter option seems quite sensible: in "Sam kicked Shaun yesterday," we could say that we refer to a time, namely the time at which there was a kicking of Sam by Shaun and which was yesterday. By contrast, in "the kicking of Shaun by Sam yesterday," we refer to an event, the event of kicking. This is a difference in referential perspective, brought out by the fact that we have a tensed verb making reference to a time in the first instance, but not in the second. Could the semantic difference between truth and reference boil down in this fashion to a difference in mode of reference? In any case, at this point of our inquiry we must state that, on reflection, our intuitive conviction that truth is paradigmatically distinct from reference is open to question. A fortiori, a practice figuring some kind of semantic evaluation of what is said by its participants would not depend on the evolution of the sentence and the truth-reference distinction that comes with it.

3 Doing without sentences Let us now return to the earlier objection that the difference between "Sam kicked Shaun yesterday" and "The kicking of Shaun by Sam yesterday" is that one can use only the former expression to make an assertion that is subject to s t a n d a r d s of semantic and normative assessment (a "move in the language game," in Wittgensteinian terms), while the latter can only be used to refer. If this objection was valid, then it should be the case, in a hypothetical language exclusively having expressions that are nominal, that its expressions cannot be used to assert. 11 For if they can, this would show that one can have

Copyrighted Material

136

Wolfram Hinzen

assertions (and hence practices of semantic evaluation) without having specifically sentences to make them, and the difference between sentences and N P s is not, after all, the one between something that can be used to assert and something that can only be used to refer: nominals could be used to assert, too. But if this is not the relevant difference, what is it? Specifically, if the truth/reference was to boil down to a merely syntactic one (the one between sentences and NPs), the m e t h o d of applicability would once more a p p e a r to be what is empirically warranted. For then there would be syntax, with whatever structures it generates a n d their respective empirical differences, and these would then either apply to the world or not. The truth/reference distinction would be no m o r e than a s h a d o w of the sentence/NP distinction. This is not an implausible conclusion, perhaps, in light of the fact that when we explain what an assertion is, we typically use specifically a sentence. If assertive practice gave a rationale for certain uniquely truth-apt structures in the grammar, it should be identifiable independently of those structures. Hence, rather t h a n grounding our use of sentences, o u r sentencebased assertive practice might depend on the syntactic structure that a sentence is and that is available in all actual h u m a n languages. In a hypothetical language where only nominal expressions are available, an assertoric practice, if it were to exist, would of course not depend on the notion of sentence. But this point would not weigh very heavily, for the likely explanation of why we m a k e assertions with sentences is, not that an assertoric practice demands sentences, but that we are all and unavoidably speakers of a language of a certain species-specific type. Given that historical circumstance, there may indeed be no way for us to specify the contents of the assertions m a d e by means of non-sentences in those hypothetical languages, except by using a sentence. But this would not show that in those other languages, the notion of assertion was dependent of the (there non-existing) notion of sentence. Let us consider more concretely one such language, envisaged by CarstairsMcCarthy (1999, 2003), called "Nominalized English." 1 2 It has n o verbs, hence no verb-headed syntactic units such as clauses. Instead of saying "Archimedes fell at the capture of Syracuse," you would say "Archimedes' capture at the fall of Syracuse." Rather t h a n asserting "I have a heavy headache," you would say some such thing as "a fairly strong headache on my p a r t . " Now, clearly, saying the latter in English, where verb-headed structures are available, may lead us to the verdict that a N P of this sort could not really have an assertoric force. While that verdict may be plausible in the case of English, however, it seems it is not at all plausible in N o m i nalized English, where people's assertions would have to rely on things like N P s - unless we say that speakers of Nominalized English, despite being capable of predication, would be incapable of bringing anything they utter to carry an assertoric force. But need we not concede to these speakers that they may distil some of the needed information to recognize something as an assertion f r o m context? A f t e r all, this sort of p h e n o m e n o n would be just

Copyrighted Material

Truth, assertion and the sentence

137

what happens a lot in actual English as well, where m a n y features vital to h u m a n communication and the determination of speech content have to be distilled from context, without being coded in the linguistic structure itself. Indeed, if we denied speakers of Nominalized English the power to make assertions, it seems we must deny it English speakers as well. We ourselves may read the (non-sentential) heading "Victory for Bush" in a daily, and this would unfailingly yield an assertoric effect. 1 3 The daily would no d o u b t be held responsible for its accuracy. If the information was inaccurate, there would no d o u b t be no reason to say that it was merely a failure to identify an existing object or state of affairs (or a failure of reference). Clearly, it would simply be false. A nominal expression supposedly only usable to identify or refer can thus be shown to have the use, even in English, of an expression that is true or false. Equally in the case of an assertion of a speaker of Nominalized English of "a fairly strong headache on my part," there appears to be no basis, in that language, for a verdict that we have here merely a case of identification of an existing state of affairs, rather than an instance of an assertion or of telling the truth. 1 4 In sum, then, while there is a reference/truth distinction in English, it is at this point unclear what, in the assertoric use of language (as distinct from some identificatory or referential use), grounds it. This leaves us, once again, with suspecting that the distinction is a mere shadow of a syntactic idiosyncrasy of English, shared by all h u m a n languages, but not by other conceivable ones, spoken by organisms with other evolutionary histories. If, in a world where all languages are replaced by nominalized substitutes of them, even the most ingenious philosophers of language could spot no difference between reference and truth, a sensible conclusion appears to be that h u m a n syntax holds the key for understanding this distinction. 1 5 Their difference is then a difference in syntactic category, and this is all there is to the matter (though what that difference a m o u n t s to in syntactic terms would still have to be clarified).

4 Truth's independence from assertion The fact that there could have been an assertoric evaluative practice without a sentence/NP distinction has led me to question an intrinsic connection between assertion and sentences, and as a consequence of that, between assertion and sentential truth. But let us now consider the relation of assertion to the concept of truth itself, disregarding the connection to sentences. There is a sense in which the capacity for truth uniquely identifies our species: it is the capacity to ask - say, in moral contexts, or scientific ones, and quite independently of any practical concerns (such as survival, and possibly counter to such concerns) what really is true, and is not just what somebody is claiming is true. Agreed, the vision of h u m a n s as disinterested seekers of the truth has come to seem somewhat conservative and old-fashioned, but I am appealing to a quite basic and very plain observation

Copyrighted Material

138

Wolfram Hinzen

here: we are creatures distinguishing (mere) assertion from fact - we are creatures pondering, deliberating, and in fact agonizing over the truth, potentially disregarding anything anyone wants us to believe. We sometimes really want to know what really happened in war crimes in Bosnia, whether it can be morally justified to abort a child, or to kill a person. Certainly some of our closest relatives in the animal kingdom have a concern with what we might call accuracy of information - say, they punish defectors breaking a contract d e m a n d i n g the sharing of vital information (see e.g. Cheney and Seyfarth 1990). But the interpretation of the relevant animal data is unclear, and indeed it is u n k n o w n whether, for example, the signals that monkeys use to warn their conspecifics of certain predators are in any interesting sense precursors of h u m a n words and their referential properties, let alone larger linguistic units (Hauser el al. 2002). Moreover, such behavioral patterns, even if they were to involve sensitivity to something we wish to call "truth." may not require any concept of this good. Just as many other processes in our cognitive and physical economy are functional without us having any concept of these ongoings, there seems to be no reason why the animal practices in question require anything like a concept of what truth is. 16 But what may weigh more heavily in this respect is that o u r notion of truth actually appears to have relatively little to d o with the notion of accuracy of representation. Accurately representing environmental ongoings, for example, appears to have little bearing on a task such as deciding for oneself the truth about abortion. Why would accurately representing anything out there matter to the morality of a decision? Even on the rather extravagant view that there are moral facts out there for us to represent them, it is not clear why representing them accurately would as such warrant a moral decision (a j u d g m e n t would be required on our part regarding why these facts should be facts). Scientific reasoning, too, has been described as having little to do with "representing facts accurately," a m e t h o d that would seem to yield no scientific profit. Science is a creative activity, it depends on the interference with nature more than on its representation (Hacking 1983). This suggests that truth, as that notion figures in our cognition, is not quite a matter of what facts there are out there it is rather a matter of our judgments, these j u d g e m e n t s not being forced by any fact out there (indeed, being forced in this fashion would de-legitimize them as free and rational ones). As noted, we do not take it to be a simple matter of assertion or a consequence of it either. N o matter what people say, and how m a n y assert something, or what they say in the f u t u r e (or indeed what we say), we think that this need not tell us anything about the truth. It would seem for this reason alone that having the concept of assertion and engaging in the relevant practice does not as such give us the concept of truth. We have two different concepts here, neither of which reduces to the other. It is a corrolary of this that the concept of truth is also not r e d u n d a n t . 1 7 We should note a n d be puzzled about this presumably universal feature of h u m a n

Copyrighted Material

Truth, assertion and the sentence

139

cognition, that we think that there is a truth of the matter in many issues of h u m a n concern, independently of whether we will ever know it and our best efforts to arrive at the right judgments. The feature is not only striking but also somewhat embarrassing f r o m a naturalistic point of view: 18 we should ask how we came up with that in terms of the evolutionary process. An ecologist visiting f r o m M a r s used to studying how animals function and fit into their respective environmental niches, might well be surprised over the ways in which we value and p o n d e r the truth. T h o u g h the question needs much further investigation, it seems there is no rationale, in ecology, for asking the question of truth, if distinct from the question of well-functioning and proper representation, and if different from other notions that have a plausible functionalist rationale, say the notion of reliability. Inquiry in the sense of practical problem-solving in an ecological environment no d o u b t has an adaptive rationale. But whether scientific inquiry is of the same general sort, and whether there is anything intrinsically adaptive about having science at all, is unclear. If we wished to give a functional rationale for having a notion of truth, we might also point to o u r practice of rationalizing h u m a n acts, since here we standardly mention what we think the agent believed was true about the world. But again, truth, far from being required for such action explanations to work, may seem like an unnecessary flourish here. Instead of referring to truth in such contexts we might rather say that we act in the light of what we think will most probably realize the objects of our desires (Ramsey 1926). Why couldn't subjective probability not substitute for truth? It likely doesn't - in that probability equal to I and truth remain distinct concepts (see Levi 1980) - but if so, that merely expresses a difference while leaving us puzzling why it exists. M o r e recently, psychologists of reasoning emphasize "ecological rationality," the use of fast and frugal Darwinian mechanisms for dealing efficiently with environmental complexity (Gigerenzer 2000). Again this striking property of our thoughts, that they are true or false irrespective of h u m a n actions and functional concerns, appears to take the hindmost. Some insist that there is a need to invoke t r u t h at least in interpreting our scientific theories, but even if that was right, that won't tell us why this notion is there. Science itself is a surprising p h e n o m e n o n , about the cognitive f o u n d a t i o n s of which we know little. The line of reasoning in the previous section speaks to these questions, too. To the extent that truth is tied to the h u m a n sentence, and the sentence was an evolutionary accident (cf. Carstairs-McCarthy 1999), our notion of truth would itself be an evolutionary accident. Were it not for the intricacies of the h u m a n clause and the still ill-understood nature of the set of functional projections structuring its left "edge" (cf. Rizzi 1997), all of which make the h u m a n sentence the hierarchical structure that it is, we would not have had that notion. Its evolutionary origins would then tell us nothing about some special epistemological virtues going along with it, such as some particular kind of "access" to an objective realm of "facts." 1 9

Copyrighted Material

140

Wolfram Hinzen

Two conclusions suggest themselves at this point of our inquiry. First, maybe we have to get used to the fact that o u r knowledge of concepts such as truth or reference, whatever their origins, tells us little about the way in which we relate to the world semantically - whether there are "real facts" out there, for example, in addition to objects, in some ontological sense that does not depend on language - and what epistemological powers we have to "hit u p o n " such entities or to depict them correctly. Second, if our notion of t r u t h has an unclear functional rationale, is n o n - r e d u n d a n t , independent of and perhaps prior to the practice of assertion (which will not as such give rise to it and is compatible with non-sentential formats), we may have to search for an internalist and non-functional rationale for truth, presumably related to specific structural-syntactic resources that o u r language faculty m a d e available to our brains.

5 Explaining syntactic categories Since I have suggested to take seriously the idea that the truth/reference distinction may at b o t t o m be a syntactic one. we should now reflect on what the difference between the two relevant syntactic categories (sentences and NPs) really a m o u n t s to. Interestingly, it quickly transpires f r o m a look at the syntactic literature that this is a hotly contested question. M o r e precisely, it has been noted long since that there are in fact i m p o r t a n t structural uniformities between the nominal d o m a i n which are headed (projected f r o m ) nouns, and sentences which are " v e r b - h e a d e d " (see Bernstein 2001). To put this in a different way: the theory of syntax has aimed to show that there are more general structural principles that are insensitive to the categorial distinction between N o u n and Verb, while still accounting for the way these enter complex syntactic structures. If this research is on the right track, it might indeed explain why our semantic intuitions as regards the truth/reference distinction are vague to some extent, as noted above. Structural uniformities between the nominal and verbal d o m a i n s notwithstanding, we have to address the question what accounts for the sentence/NP distinction, which doesn't go away through these uniformities. Truth be told, there are very significant problems in general in grounding the distinction between syntactic categories in non-syntactic (in particular, semantic or notional-conceptual) distinctions. This is an old observation not only powerfully restated in Baker's (2003) recent theory of the lexical categories, which defines them in purely syntactic and internalist terms, but also in Jackendoff's (2002) very different, more functionalist f r a m e w o r k . Carstairs-McCarthy (1999), too. argues for an internalist rationale of syntactic categorical distinctions, for different reasons. He notes that the typical asymmetry of subject and predicate in h u m a n clause structure follows f r o m nothing in logic or in what sentences mean. To illustrate this asymmetry, Verbs take a so-called "internal" (deeper embedded) argument, the result of which is a phrase (a verb phrase. VP), which then is predicated as a whole of

Copyrighted Material

Truth, assertion and the sentence

141

the subject, the so-called "external" argument: ([NP s u b j e c t [VP [NP object ]]]). In turning to its explanation, Carstairs-McCarthy notes m o r e generally "the lack of a consistent relationship between certain aspects of syntax-as-it-is and any representational or communicative function" (Carstairs-McCarthy 1999: 77). 2 0 This observation motivates the suspicion that the sentence/NP distinction should indeed be expected to arise on internalist grounds, rather than being grounded in an independently given, "objective," truth/reference distinction. T h e latter distinction, too, then, does likely not owe its existence to some independently given distinction between events (depicted by verbs) and their participants (depicted by NPs), or between predicates and subjects. Both distinctions are in fact consistent with the absence of sentences. Nominalized English, in particular, allows for event descriptions as well (e.g. "the description of Syracuse by Caesar"), and it does for predication too. N o n e of this would be surprising if indeed it remains more generally mysterious what should be communicatively efficient or particularly functional in most of the universal structural design features of language that modern linguistics has unearthed over the last decades, such as locality conditions on movement, trace deletions and their licensing principles, and complex morphology, all or most of which would seem to make little sense to an artificial language designer who were to design an efficient c o m m u nicative tool. 2 1 In the biological world at large, structures in the organism d o not evolve for functional needs, but randomly, with the environment having the role of selecting a m o n g arising forms. For the case at hand, this means that we d o not expect sentences will spring into existence merely because h u m a n s were engaging in certain communicatively relevant acts, such as assertions. In this way, and although the argument requires much elaboration and qualification, Carstairs-McCarthy's internalist conclusion that a crucial design feature of h u m a n language, the sentence/NP distinction, is not grounded in semantic or external conditions, is both what we have come to expect more generally in linguistics, a n d what the failure of any easy functionalist logic in biological explanation suggests.

6 The concept of truth in human language E m b a r k i n g on an internalist course of explanation, then, it is natural to ask which role precisely, if any, truth-evaluable structures play in the workings of the computational system underlying h u m a n language (CS H L ). 2 2 CSHL, on standard assumptions, selects lexical items stored in long-term memory and combines them, so as to create complex structured units subject to specific interpretations by the sensory-motor systems on the one hand (speech articulators and perceptual mechanisms), and the semantic c o m p o nent, on the other. This raises the question of how exactly it is that specific categories computed in the course of a syntactic derivation get mapped to particular semantic ones, and indeed in the right way: NPs to objects, VPs to events, and sentences to truth.

Copyrighted Material

142

Wolfram Hinzen

Clearly, syntax and semantics cannot be entirely unrelated (or be entirely separate competences), since generative syntax has centrally shown how semantic facts (such as ambiguities, referential dependencies, thematic interpretation of verbs, etc.) pattern with syntactic ones (that is, different syntactic configurations and derivations, to which we can trace the differences intuitively classified as "semantic"). Still, just what mechanism relates syntax and semantics in the h u m a n mind we do not know. We deeply wish to say that N P corresponds to reference, while sentences correspond to assertion/truth (how could it be the other way around?). But the question is what this desired m a p p i n g follows f r o m . We have seen above that there is no necessity to this mapping, as in a hypothetical language which has no sentences but only nominais, 2 3 there might still be assertions, and perhaps kinds of truths (though they would not, of course, by speakers of that language, be identified by means of a sentence): indeed why should the truth that the number 2 is prime, say. relate in any particular way to h u m a n language at all. given that it is. presumably, a truth holding even in possible worlds where there are no language-using creatures? 2 4 In our own language, too. the m a p p i n g we so much think is right might, logically speaking, precisely be the reverse: in "John's car crashed," the nominal "John's c a r " might be m a p p e d to a truth in some context. say that car's existence in that context, while the predicate would be a function from a truth value to an individual, say the crashing event. In a language so interpreted, events might perhaps have names. 2 5 Of course, h u m a n language is not like that, and we know for sure how the m a p p i n g should go. But why we think that is precisely the question. O u r " t h i n k i n g " that it has to be that way has to be m a d e explicit, and ideally the m a p p i n g should be derived. Why, then, do we think that? If this question is an empirical one. which I intend it to be, its answer will be an answer about the structure of that part of the h u m a n language faculty in which h u m a n s c o m p u t e the meanings of expressions. Let semantics be the scientific theory of this semantic component, SC for short. I will say that the phonological component in the h u m a n language faculty, PC. regulates the sound of items taken from the lexicon, L E X , when put together by syntax, i.e. the CS H L - The SC regulates what items taken from LEX mean, once CSHL has applied its operations to them (a highly restricted notion of meaning, of course: it is restricted to those structural aspects of the meanings of expressions that the h u m a n linguistic system determines). The syntactic component is C S H L itself. When applied to several lexical items specified for their sound and meaning in LEX. it regulates how the complex sound of the resulting complex expression is paired with a specific complex meaning. 2 6 Given the lexical items it contains (specified for their meaning and sound in LEX), it sounds the way it does and " m e a n s " the way it does. A derivation then is the step-wise c o m p u t a t i o n of a complex sound-meaning pair, given a number of lexical items specified for sound and meaning in LEX, and it is systematic and lawful.

Copyrighted Material

Truth, assertion and the sentence

143

For example, the sound structure corresponding to "the portrait of Jay's," by contrast to the sound structure "Jay's portrait," can be used to refer to a portrait painted or owned by Jay, but crucially not to a portrait that has Jay on it. I claim this is an empirical fact about this particular complex expression. It pairs one particular sound with a (number of possible) meaning(s), excluding other logically conceivable meanings, in particular the impossible meaning just mentioned. That the sound pairs with the meanings it does, and not with the one it does not pair with, has in this case a complete explanation in terms of the workings of C S H L 2 7 In the workings of the C S H L , collections of syntactic features naturally group together in the way of influencing and depending on one another. T h u s it is still widely assumed that there is a "theta-system," regulating which argument of a verb gets into what syntactic position, a Case-system, an Agreement-system, and so on. Modular organization in this sense is an empirical finding,28 and it is possible that the SC has a m o d u l a r architecture in the sense in which CS H L is thought to have it, as well. By and large, however, the SC is a depressingly dark issue, and there is not remotely as m u c h clarity about which structures enter into the semantic interpretation of a structure c o m p u t e d by CSHL as there is about which structures enter into the derivation of such structures as such. Clearly, syntactic structures get accessed for use on an occasion by a h u m a n organism, but just how (or by what internal mechanisms) that happens is a vastly open question. Still the syntax and the SC clearly interact, and we define their interface to be the level of representation L F ("logical form"), a technical notion which we thus - qua syntactic interface see to have nothing much to do with logic, despite its historical label, a point to which I return in the next section. At the one side of this interface, CS H L delivers syntactic structures that have to obey certain conditions imposed by Universal G r a m m a r (UG). These structures correlate - or so we are now assuming with certain semantic interpretations in SC, as when NPs are m a p p e d to objects and sentences to truth. At the other side of this interface, the SC-side (where, we should note, we are still inside the mind), these structures get accessed for use on an occasion of language use. Having elaborated on this general framework for answering the empirical question of how different types of expressions get their different types of interpretation, the crucial question we can now formulate is this: Is there anything in the deliverances of CS H L on the syntactic side of the LF-interface that correlates in any intrinsic way with truth more than it does with reference? Does a truth-theoretic interpretation as familiar from logic-based semantic theories in some ways rationalize at least some of the constructs delivered at the interface, so that we know they exist because truth plays some crucial and particular role in the way we use language? The hope is that the constructs thus picked out would be sentences, for in that case, despite our verdicts that there is no intrinsic connection between assertion and sentences, at least our involvement with truth rather than reference in

Copyrighted Material

144

Wolfram Hinzen

relating language to the world would tell us why there is a distinction between sentences and NPs. In turn, if that distinction had nothing to d o with the fact that one of these constructs bears truth and the other does not, we would again have reached a dead end. Truth would not only not be grounded in assertion, but it would also not account for the existence of sentences as opposed to NPs: sentences would not be things which exist simply because truth but not reference can be assigned to them. Again, the right explanatory direction would seem the reverse: that sentences engender crucial aspects of truth. To say that prior to the evolution of h u m a n language in the m o d e r n sense a r o u n d 200,000 to 50,000 years ago there were t h o u g h t s that were true or false, and sentences arose to express them, is in this sense circular: it is unclear why, prior to the evolution of sentences, there would be any propositional truth, and designated constructs to attach it to. 2 9 I see little in the language evolution literature to contradict this conclusion. Pinker and Jackendoff (forthcoming) defend the idea that h u m a n language is an innate mechanism that evolved for the purpose of communicating complex propositions. But either these propositions are the sort of things expressed by sentences, in which case they depend on sentences, of which no functional rationale is known, or they are not, in which case their existence also would not explain the coming into existence of structures we specifically apply t r u t h to. A recent view defended by Hale and Keyser (2002) in fact implies that the syntax even of the argument structure of verbs - a crucial sub-part in propositional information - c a n n o t have arisen on the basis of evolutionarily prior non-linguistic conceptual structures or t h o u g h t s of the kind Jackendoff and Pinker envisage. For this view argues that these argument structures are, or fall out from, the syntax. N o t h i n g can explain the origin of syntax, if it is syntax. In short, while pre- and nonlinguistic creatures no d o u b t h a d some kind of " t h o u g h t , " it is not clear why it should be of the specifically h u m a n propositional format. Let us then, for the rest of this paper, explore the view further that o u r capacity for truth and propositional t h o u g h t post-dates the evolution of h u m a n language, and is rationalized by it, rather than vice versa.

7 Truth and LF: connected yet divorced The LF-interface of the g r a m m a r has, of course, as I mentioned, traditionally been the level of representation where syntactic structures exhibit logical properties, can be m a p p e d to propositions, and can be studied in a truth-theoretic fashion. N o t e however to begin with that one may perfectly well endorse a working truth-conditional semantics for natural language say, a Davidsonian one - without asking oneself once whether this theory is actually true of what goes on in the SC. 3 0 T h e point becomes clearer by a c o m p a r i s o n with syntax: an arbitrary syntax for some formal language need not in principle tell us anything a b o u t h u m a n syntax, and logical syntax (in some C a r n a p i a n sense, say) a n d U G are plainly not wedded to one another.

Copyrighted Material

Truth, assertion and the sentence

145

In principle, they may diverge at any point. Similarly a syntactic theory that works in the sense of generating all and only the "well-formed" expressions (by some criterion of well-formedness) need not be an empirical contribution to syntactic theory in the naturalistic and empirical sense of the generative or "biolinguistic" tradition (Jenkins 2000). I am simply assuming that exactly the same point can be made for semantic theory, as there appears to be no asymmetry here. W h a t would be evidence, however, that some semantic theory is actually true of the semantic c o m p o n e n t , and in particular whether truth does play a crucial role for the interpretation of LFs? Now, there are of course certain structural resemblances between LFs, on the one h a n d , and logical descriptions of expressions for which truth conditions can be defined, on the other. This indeed is why, historically, the label L F derives from "logical f o r m . " Thus, e.g., L F s have so-called "empty categories" where logical forms in the logical sense have logical variables; the former establish scope relations through syntactic movement up the syntactic tree whereas the latter represent it through the ordering of quantifiers, and so on. Consider then again the following stab at an explanation of why sentences m a p to truth: Because there are truth-conditional semantic theories for logical languages designed to encode reasoning, and h u m a n languages show structural similarities to these logical languages once their expressions are represented at the level of LF, this is evidence that there is some intrinsic connection between LFs and truth: as it happens, truth conditions are what is determined at the interface. But this line of t h o u g h t (again, an externalist one) is fallacious. L F would be what it is even if no truth-conditional semantic theory had ever been developed, and no theory of logic existed. Historically, the influence of logical semantics for the development of a theory of L F in the grammatical sense was clearly vital, but the rationale for the existence of LFs is nonetheless independent of logic or the theory of reasoning: its rationale lies wholly within syntax as such (apart f r o m general constraints, such as that the language system must be usable at all). Put differently, reasoning is a use that L F s have, but this use is not what defines them. Consider in particular the view of a m a j o r pioneer of the L F - n o t i o n in generative grammar, James H u a n g : ' " L F is not" to be equated with the level of semantic structure any more than P F [phonetic form] is to be treated as a level specifying the sound waves of any given utterance" ( H u a n g 1995: 128). P F s are internal representations defined to encode instructions to the speech articulators, hence to p e r f o r m a n c e systems inside the brain. But they do not mysteriously reach outside the organism, saying something directly about air movements, say, or m o t i o n s of molecules. In the same sense, I take H u a n g to suggest, L F s as such have no intrinsic connection with mind-external propositions, states of affairs, commitments of thought, etc. H u a n g goes on: It [LF] expresses only aspects of semantic structure that are syntactically expressed, or that are contributed by the g r a m m a r .

Copyrighted Material

146

Wolfram Hinzen T h e supposition that the meanings of sentences are not directly "read o f f " from their surface forms is based to a large extent on the combination of the following three facts: (a) that sentences with quantifiers and question words exhibit special semantic properties which distinguish them f r o m non-interrogative, non-quantificational sentences; (b) that these properties reflect syntactic generalizations that are best captured by reference to their structure at LF: and (c) that the derivation of L F representations f r o m S-Structure involves little or n o extra cost other t h a n what is already m a d e available by a proper theory of overt syntax. ( H u a n g 1995: 128)

In other words, L F is motivated by the finding that specific kinds of observations informally or pre-theoretically classified as "semantic" observations about inferential properties, truth conditions and referential properties of expressions - are actually captured by syntactic generalizations, or fall out f r o m the grammar. They have explanations in terms of operations that are assumed to be part of the g r a m m a r on grounds independent of the semantic facts for whose explanation they are now shown to be apt. Far from depending on a specific theory of what goes on in the semantic c o m p o n e n t , they derive various empirical observations about the meanings generated in it. As Uriagereka (2002: 212) crisply puts this viewpoint, syntax "is blind to semantic motivation, although it is not i m m u n e to semantic consequence." It is blind, because no syntactic mechanisms appear to run for "expressive needs," or because a derivation looks ahead into the useful and semantically interpretable result that will eventually come out at the end of the derivation. This is in the same sense in which nobody gets a flue because he desires so. H u a n g states a persuasive argument that semantics or truth provide no rationale for those interpretive properties of expressions that depend on their structure at LF. H u a n g notes that facts distinguishing quantified noun phrases from referential noun phrases with respect to generality and scope ambiguities are semantic facts a n d . by themselves, do not argue for the existence of a syntactic level of LF. Since L F structures are subject to interpretation, one may as well devise m a p p i n g rules that convert S-Structure representations directly into semantic structure, without the mediation of LF. N o appeal to semantics per se can provide a real argument for the existence of this level of syntactic representation. In spite of c o m m o n misunderstanding, L F is not motivated merely as a level of disambiguation. ( H u a n g 1995: 130) The structural similarities between L F s and truth conditionally interpreted structures in logical languages arc thus no reason in the end for

Copyrighted Material

Truth, assertion and the sentence

147

assuming an intrinsic connection between L F s and truth, for the LF-structures are what they are simply because the g r a m m a r has certain operations. As it happens, these lend themselves to use in theories of truth conditions, but it is not the truth conditions that we can use to explain why the CSHL has these operations. Truth conditions are a benefit, not a cause. Let us take a concrete example: the curious fact about the D P "the portrait of Jay's," noted above. Why on earth should we be forbidden to interpret this on a par with "Jay's portrait," or "portrait of Jay," both of which allow Jay to be the portraitee, contrary to the first? Clearly, we have a semantic or communication-theoretic motive for this interpretation. But the syntactic system proves insensitive to such motivations, and the syntactic explanation of our semantic observation turns on the fact that certain deletions are universally illicit in h u m a n languages, for some reason: concretely, deletions of elements that are not the top-most items in a complement of an overt, adjacent word (cf. Anderson and Lightfoot 2002: 47). On the reading where Jay is the object of the portrait, that is the portraitee, the structure of "picture of Jay's" as represented at L F would be "picture of [Jayi's [e, ei,]]." Here e is the position where "picture" is interpreted (even though it is moved and heard elsewhere) and e¡ would, if the impossible reading were possible, be the position of the portraitee, that is Jay (hence the co-indexing). This violates the UG-constraint just mentioned, since the adjacent position of e is not an overt word to which the deleted e¡ could attach. This then is the explanation for our initial semantic observation: deleted items c a n n o t cliticize to deleted items. T h e explanation contains no semantic terms. We could of course change our way of talking, write up truth conditions for "Picture of Jay's" and for "Jay's picture," and derive these, but it is not clear why we should. T h e move adds nothing to the syntactic law just stated. Whatever the semantics added, it will still be true that "picture of Jay's" c a n n o t mean what "Jay's p o r t r a i t " or "portrait of Jay" mean because elements that are not the top-most items in a complement of an overt, adjacent word c a n n o t delete. M o r e generally, talk of truth conditions gives us data, which we then must explain. Ideally, truth conditions fall out from the syntax: they are explananda, data that provide a test case for syntactic theory and are derived if the theory is successful, but they play no explanatory role. L F is syntax, then, but it is semantics, too. In fact, what basis d o we have in any case to keep these two academic labels apart? There is a basis for the distinction if we associate the label semantics with the study of l a n g u a g e world relations, for syntax in the generative tradition is an internalist science. But much of semantics is captured by the syntax of LF, and here the distinction between these labels has become increasingly opaque. Still what I have called the "benefit" of truth conditions is a benefit worth noticing. Why might this be? Clearly it cannot be a mere accident. Here then there is still our old intuition, an inner voice, as it were, that whispers to us the following:

Copyrighted Material

148

Wolfram Hinzen H u m a n languages have one primary use. This use is to assert things, to say that some proposition is true or false. The CS H L must thus be somehow tuned to this use. producing structures that are suitable to it. The needed structures, however, are sentences. Hence the LF-structures generated at the one side of the interface must be sentence-like, suitable to the expression of propositional thought. There is an intrinsic connection between LFs and truth for this very reason.

But this is just the old mistake, and the argument given above would on the contrary suggest that our intuition about what is a truth-bearing structure and what is not consists in our understanding of certain syntactic structures, their contrast with others, and the semantic consequences of these various structures. If it consists in them, it does not explain these syntactic structures. Syntax or L F does not answer expressive needs. While it is true that the prime kind of expression that we use to assertive purposes is the sentence, it is highly questionable whether there is any explanatory line running from that particular use to the existence of the kind of expression so used. The use of sentences in the practice of assertions or for the expression of "propositional t h o u g h t " does not speak in favour of any intrinsic connection between the notion of L F (the locus of meaning in the h u m a n language faculty), on the one hand, and the notion of truth, as relevant to our practice of assertion or the notion of proposition, on the other. Truth-theoretic semantic interpretation does relate to the level of LF. But L F is a creature of the syntax. This section asked whether we can show the concept of truth, which so prominently figures in our practice of assertion, to have a rationalizing role for the constructions that the syntax produces and delivers for interpretation. We have not found evidence that this is so. For all the development of the theory of L F in the biolinguistic tradition suggests, o u r intuitions about truth (and the structures that bear truth) might entirely consist in certain structural configurations that the syntax constructs, rather than explaining their existence. To rephrase this conclusion, we wanted to give the concept of truth - a paradigmatically externalist notion for most - some role to play in the internal workings of the semantic c o m p o n e n t (SC). Usually, it is brought into play because logical forms in the syntactic sense are " m a p p e d " to some proposition (with a "logical f o r m " in the logical sense), and these are things which, by definition, bear truth. But in this section I have argued that this philosophical notion of a truth-bearing proposition apparently plays no explanatory role for the workings of the C S H L ' or does not explain why they are the ones they are. As it happens, some structures the syntax generates - crucially including structures in which semantic properties of expressions are laid down - have an intuitive connection with t r u t h . But that connection is not. apparently - to the extent that the generative theory of syntax is correct - what motivates these structures. We basically do not know what motivates the connection. We have to return to the problem inaugurated before: we m a p syntax to semantics in a specific way - sentences

Copyrighted Material

Truth, assertion and the sentence

149

to truth and not to reference, while n o u n s to reference and not to truth. But we do not know what this mapping derives from.

8 Truth as predicate and as relation Above I have severed the connection between assertion (as a form of language use) and sentences, as well as between assertion and truth. In the previous section, I have f u r t h e r questioned the explanatory status of the notion of propositional thought in linguistic semantics. Let us now and lastly indicate some implications of these contentions with respect to more traditional philosophical concerns in the metaphysics of truth. I begin with the question of what truth attaches to - what it is predicated of - a question philosophical tradition has addressed under the label of the "truth-bearer." The question is somewhat dubious, in the following sense. Truth, I have here assumed, is to start with a (universal) h u m a n concept, like the concepts of "goodness," "mystery," or "love." Concepts are just that - concepts, lexicalized in LEX, non-structured by the CS H L (indeed I will assume they are " a t o m s " structurally). Being simply concepts in this sense, they are not intrinsically " b o r n " by anything. It is equally misleading to speak of "the truth-predicate": truth as a concept is no more intrinsically a predicate than any other concept is. If the question, on the other hand, is what the concept of truth when used as a predicate in a particular syntactic configuration is predicated of, o u r proper focus of attention should be precisely the structures themselves in which these truth predications occur: say the structures " W h a t you say has some t r u t h to it," " W h a t you say is true," or "[The truth of what you say] makes me wonder why I never listen to you." If we want to know what truth attaches to, that is, it seems sensible to suggest we should simply look at structures such as these empirically, a n d see what the concept is predicated of in these instances. Doing just that, and building on Uriagereka (2002: Ch. 9), Hinzen (2003) concluded that truth is a predicate of the subject of a so-called "small" that is, tenseless - clause which looks as follows: [small

clause

what you say [truth]].

This gives a concrete and empirical answer to the question of how our h u m a n concept of truth enters larger structural complexes, and gets attached to other constituents. 3 1 It is a consequence of this empirical proposal that the way in which our concept of truth is structuralized in h u m a n syntax - the way of, as Hinzen (2003) proposes, a partitive syntax - is the same, no matter whether we have nominal structures like " T h e truth of what you say," or sentences like " W h a t you say is true." 32 Here then we finally meet a sense in which not only the connection between assertion and truth is not intrinsic, but the connection between truth, as a h u m a n concept, and sentences, may not be either. That sense may relate to the deep structural

Copyrighted Material

150

Wolfram Hinzen

uniformities between nominal and verbal phrases that I alluded to before (see Bernstein 2001). T h a t is, maybe it didn't take sentences to turn us into truth-users and truth-seekers, but a more general structural pattern in our minds that cross-cuts the sentence/NP distinction. The conclusion that the concept of truth and propositional thought may not be something independent of the specific structures that our language faculty provides us with (and not be something evolutionary prior to them) remains. Having answered (or diffused) the question of the "truth-bearer," let us close by giving an indication for how an answer to the question of the truthmaker would go in the present f r a m e w o r k . We could ask, for example, what makes an assertion of the form " W h a t you say has some truth to it" true. 3 3 N o t h i n g goes wrong, I suppose, if we answer, in a somewhat truistic fashion: W h a t you say must (really) have some truth to it. Then, and only then, that what you say has some truth to it would be true. But what we notice here is the following. In the question of the truth-maker we were asked to identify the thing, X, that, if existing, would make some j u d g m e n t true. But at least in the present instance our only way to identify X is through the j u d g m e n t itself: X is a situation in which, well, what you say has some truth to it. In short, if there is a sense in which j u d g m e n t s of this sort are m a d e true by anything, what seems to be the correct choice for the t r u t h - m a k e r does not illuminate the t r u t h - m a k i n g relation. Truth does not "cancel o u t " in such instances. One might have hoped that we could have identified the situation X in non-circular terms. But this presupposes that the concept of t r u t h as structuralized in a j u d g m e n t like the above can be analyzed in terms of other concepts - which would work, for example, if the concept of truth was the concept of evidence plus perhaps an appropriate epistemic context and some f u r t h e r restrictions. I take it that no such decompositional analysis of our concept of truth has been given, and fully general arguments against lexical decomposition as in F o d o r (1998) apply to this particular case as well. It seems for this reason unclear how an illuminating or non-circular analysis of what makes t r u t h - j u d g m e n t s true could be possible. That leaves open whether the t r u t h - m a k e r of. say, "It's a fine day," where the concept of truth does not overtly occur, could be specified in non-circular terms, i.e. in terms not themselves involving the j u d g m e n t "It's a fine day." I do not know of any such specification. A situation in which it is true (for me) that "It's a fine day," is, well, just a situation in which I make that particular j u d g m e n t . It is clearly not the case that what makes the day fine is, say, the shining of the sun, or your believing the argument of this paper. Whatever list of such items we compile, and however long it gets, it always may describe, as long as we do not use the j u d g m e n t in question itself to describe it, a situation in which, for some reason. I do not make the j u d g m e n t in question. In this sense, it is simply inexpressible (non-circularly) when exactly a situation is such that all necessary and sufficient conditions obtain for me to make the j u d g m e n t "It's a fine day."

Copyrighted Material

Truth, assertion and the sentence

151

Searching for a non-trivial theory of truth-making leaves us thus in a dead end, no matter whether we consider explicit truth-judgments, or judgments in which the concept of truth does not overtly occur. The truthmaking relation depends on the meanings of the structures we consider, and does not explicate those meanings.

9 Conclusions I began by q u o t i n g a familiar statement relating the illocutionary role of truth to the practice of assertion. T h e r e seems to be a sense in which both t r u t h as such and assertion are things tied to the sentence, in a way that, without the sentence, o u r minds would be left with merely referring to things as opposed to being able to say anything true about them. On closer scrutiny, I argued, these intuitions crumble, in the following way. The truth/ reference distinction itself becomes somewhat obscure, if spelled out as a distinction in semantic category. In syntactic terms, the difference between sentences and N P s that corresponds to it is an empirically given and relatively solid one, and I have argued that we cannot dismiss the thought that whatever there is to o u r semantic intuitions regarding truth as opposed to reference, is ultimately explained by reference to our evolutionary endowment with a language patterning syntactically according to a particular type. O u r particular practice of assertion, and other externalist rationales, neither seem to ground or motivate that syntactic difference nor o u r notion of t r u t h as such, and perhaps they rather are, at least in part, a conscquence of both. A notion of assertion does not as such d e m a n d a sentential f o r m a t , n o r a notion of t r u t h of the kind that sentences uniquely support. Whatever the origins of propositional thinking and semantic evaluation in terms of truth, they don't seem to be the origins of the structural distinctions that we employ to these purposes. The assumption that linguistic constructions are tuned somehow to the use to which we put them does not seem b o r n out by the way in which generative g r a m m a r conceives of the semantic interface. While structures appearing at this interface are used for purposes of assertion and truth-ascription, I found little reason that these practices are causes of these structures rather t h a n their possible effects or benefits. When embarking on the internalist course of explanation that the present paper overall has been motivating, furthermore, the question of the truthbearer may cease to be formulable, and that of the truth-maker may not be non-circularly addressable. This point suggests more positively to study h u m a n truth j u d g m e n t s empirically as p h e n o m e n a in their own right, which, while being analyzable structurally and grammatically, c a n n o t be illuminated or explained externalistically by positing various relations to facts or things out there. N o t e that despite the n o n - r e d u n d a n c y of the concept of truth that I have defended (on what I believe to be common-sense grounds), no substantive or metaphysically "rich" notion of truth has been

Copyrighted Material

152

Wolfram Hinzen

p o s i t e d here, w h e r e n o r e l a t i o n a l (or f u n c t i o n a l ) n o t i o n s s e e m e d t o t h r o w light o n t h e n o t i o n of t r u t h .

Notes 1 I wish to thank the editor for his careful reading and comments. 2 Of course, I don't here use the term "sentence" in the trivial sense of a mere shape on paper, or an "empty form." 1 use the term in a more technical sense of contemporary linguistic theory, where it is a structure analyzable at different levels of description, including acoustic, phonetic, phonological, syntactic and semantic ones. It is particularly the latter two, more abstract levels of representation. where we crucially find hierarchical (phrase-structural) as opposed to merely linear organization. This hierarchical organization, with the syntactic categories that make it up. is what matters for the purposes of this paper. 3 In this respect, the post-verbal adjective "true" differs from others, for we can say that the number 2 is "prime", or that Madonna is "beautiful", where we also have adjectives in post-verbal position. So there is a constraint on which kinds of syntactic objects we can apply "true" to. 4 This is not true for other grammaticalizations of it. For example, one can say "God is the truth." where "God." syntactically speaking, is obviously not a sentence. 5 From a linguistic point of view, many qualifications would be appropriate here, but I ignore these here. For empirical evidence for the universality of the verbbased sentential format, see Baker (2003). 6 The "method of applicability" has been positively suggested, though somewhat in passing, by Carstairs-McCarthy (1999: 228-9). ' 7 One objection to this consists in saying that "applicability is truth when applied to sentences, and reference when applied to NPs" (as one referee put it). This objection presupposes the truth/reference distinction, which is what I aim to understand and explain. To do so. we must find out why a dual sort of semantics evolved in the human lineage, instead of an applicability-theoretic semantics alone, in which the difference between reference and truth is not a fundamental and semantic one, but essentially syntactic (see below). 8 Along with a standard assumption in grammatical theory I will assume that Noun Phrases (NPs), when fully analyzed, are complements to Determiner Phrases (DPs). hence that nominal expressions basically have the structure of DPs: [DP[NP]]. See Bernstein (2001). 9 It is true that in this example and the following any way of telling what truth we are referring to or talking about will make use of a sentence. But it seems unclear why the need to explicate a referential act in clausal terms makes it any less a referential act. 10 Basically, formal and model-theoretic semantics is a formal approach whose success and virtue reside in part in abstracting from the philosophical problem of reference and the question addressing its range of applicability. 11 Cf. Carstairs-McCarthy (1999) for the following line of argument, which is used by the author to highlight the fundamental underspecification of the structural features of human languages by their functional uses. 12 It is just one of many conceivable verbless languages that the author experiments with. Another is "Monocategoric." which is like a categorial grammar, albeit with only one basic syntactic type, the other being an n-ary operator type mapping the basic syntactic type to the same type. Expressions of Monocategoric would be perfectly capable of distinguishing argument roles by their order, say, so that "John encyclopedia Mary GIVE," in which "GIM: is an operator applied to three arguments,

Copyrighted Material

Truth, assertion and the sentence

13

14

15

16 17

18

19

20

21

22

23

24

153

would mean what in actual English we express sententially by saying "John gave the encyclopedia to Mary." To say that this expression really is sentential, though only covertly (i.e., not overtly in the actual phonetic form), does not seem an attractive proposal, as there should be syntactic evidence for postulating such a covert structure, independent of the desire to make the theory come out true that assertoric force requires the presence of sentential forms. Ludlow's (2005) argument that seemingly non-sentential assertions are sentential on a covert level of L F runs into just this problem. Note that the notion of an L F is not tied to the notion of a sentence (see Section 7). And if Ludlow's proposal was true for English, it would still not thereby be true for Nominalized English. One reader objects that "to be a sentence is to be used as a sentence." "Since in the hypothetical language names are used as sentences, they are not really names, but sentences." But, of course, no new cognitive competences arise from simply deciding to use old forms to a new purpose. I am here trying to understand the cognitive basis for engaging in certain kinds of uses in the first place. This is argued by Carstairs-McCarthy (1999), who provides empirical and convergent evidence that our employment of the syntactic sentence/NP distinction (and the truth/reference distinction, to whatever extent it is tied to it), is no more than an evolutionary accident, having to do with a dramatic change in the physiology of speech which evolution used to manufacture the h u m a n sentence. Carstairs-McCarthy (1999: Ch. 3) also extensively discusses the attempts of philosophers such as Plato, Wittgenstein, Frege, and Strawson to ground the syntactic sentence/NP distinction in semantic or pragmatic (hence non-syntactic) terms, and plausibly concludes that this distinction must have an internalist rationale. See Hinzen (2006) for discussion. In an analogous way, one may as well have a method for detecting music, without having a concept of music. Though this is not for the reason most frequently cited (e.g. in deflationism), namely that the word has the so-callcd "pro-sentential" use and is non-redundant for that reason. As Rorty (1986) notes, when he argues that it follows from an appropriate "Darwinian" thinking that we are not oriented towards the truth, but increased prosperity. The notion of a "fact" is given a content through its contrast with the notion of an object. But the fact/object distinction we make in semantics and ontology looks suspiciously close to the sentence/NP distinction itself. Would the speakers of Nominalized English think of the world as being structured into facts? Jackendoff (2002: 253) suggests that the Topic-Comment organization of information structure in discourse is the cause of the subject-object asymmetry. Carstairs-McCarthy (1999) discusses and rejects this option. See Uriagereka (1998), Baker (2001) or Hinzen (2006) for recent formulations of this challenge to a functionalist logic of thinking with respect to h u m a n syntactic structures and their origins. I will follow standard assumptions in assuming this computational system to be universal, so that differences between languages, as on accounts such as Chomsky (1995: Ch. 1), reduce to the morphological expression of the (very small) functional part of the lexicon (apart from "Saussurean arbitrariness," i.e the arbitrary pairing of phonetic features with given concepts in the mental lexicon). Of course, such a language is very likely not natively learnable as a first language, but this if irrelevant here. We wish to understand why languages that h u m a n s can natively learn have a specific structural setup. Consideration of hypothetical languages tells us what these specifics are. I owe this point to correspondence with Juan Uriagereka.

Copyrighted Material

154

Wolfram Hinzen

25 Cf. Uriagereka (2001: fn 8). The Löwenheim-Skolem theorem, moreover, proves that as long as we stick to a first-order language, no particular mapping from syntax to semantics follows from it. and semantics will essentially be arbitrary. 26 It is not true that any derived expression can mean just anything. 27 For the (standard, I believe) explanation, to which I return below, see Anderson and Lightfoot (2002: 50-51). 28 Recent work suggests, for example, that these modules obey common principles, in fact, as the same few grammatical operations apply to all these features, leading to a much simpler and more uniform grammar than envisaged in the 1970s and 1980s (Hornstein 2001). 29 No doubt there was some notion of satisfiability, of the kind we use to ascribe a semantic content to animal thoughts or thought procedures, say as used in navigation and foraging. 30 Davidson in particular famously offers an argument that it "does not add anything to this thesis [a descriptive model of the speaker's linguistic competence] to say that if the theory does correctly describe the competence of an interpreter, some mechanisms in the interpreter must correspond to the theory" (Davidson 1986: 438). But this conclusion sounds rather strange, as facts about brain architecture, evolutionary history, etc. may clearly constrain the truth of a given semantic theory (unless the question of how humans come to know the meanings of their expressions is argued not to matter). In the Minimalist Program in particular (see Chomsky 1995). much of the design of human grammars has to do with how syntactic computations meet interface representations between language and thought, on the one hand, and language and sound, on the other. That is, much of language design is constrained by how a cognitive system is embedded in the rich context of the modular architecture of the mind/brain. While the cognitive and physiological/anatomical internals of the mind or the organism are no constraints on setting up logical formalisms supplied with a Tarskian semantics, it seems bizarre to suggest on a priori grounds that they do not matter for a description of human (knowledge of) language. 31 This answer cannot speak to the discussion of whether truth attaches to sentences or propositions, though, since that dispute depends on an externalist notion of sentence (as a symbolic type in a public language). The generative or "I-linguistic" tradition has discarded this notion from its inception, viewing a sentence as a pairing of a sound with a meaning instead, where both the sound and the meaning are internally represented structures in the mind. Only if a sentence is conceived as a some external physical entity, it seems, which I have assumed no sentence in the linguistic sense is. can there be a temptation to say that properly abstract entities such as propositions must be the truth-bearers. 32 It is another interesting consequence that the concept of truth does not have a "denominalizing" function, of the kind frequently assumed in the philosophical literature, for example. Künne (2003: 35-6). On this view, which may be seen as supporting a variant of the redundancy thesis of truth, "is true" when added to a that-clause does nothing but to cancel out the effect of the nominalizer "that." As far as I know, however, the idea that -clauses are "nominal" is indefensible from a linguistic point of view to start with, given that "that" is, in the syntactic sense, a complementizer here (whose SC interpretation is presumably declarative force). If truth is predicated of a small clause subject which is such a complemented clause, it does not "cancel the complementizer out" (it would rather seem to depend on it. as a marker of sententiality). There also simply appears to be no known grammatical process of this kind, hence no linguistic ground for asserting a redundancy thesis of truth on the basis of familiar logical biconditionals such as "That the earth is round is true iff the earth is round." Hinzen (2003) in fact suggests that a self-standing assertion of "The earth is

Copyrighted Material

Truth, assertion and the sentence

155

r o u n d " has an underlying L F which is identical to the mainly phonetically and stylistically distinct "That the earth is round is true." In this sense, the former expression reduces to the latter, not the other way around, and no conclusions relating to the redundancy of truth can be drawn. 33 In Hinzen (2003) the mentioned form is an example for what is there called a truth-judgment.

References Anderson, S. R. and. Lightfoot, D. (2002) The Language Organ. Linguistics as Cognitive Physiology, Cambridge: Cambridge University Press. Baker. M. C. (2001) The Atoms of Language, New York: Basic Books. (2003) Lexical Categories, Cambridge: Cambridge University Press. Bernstein, J. B. (2001) "The D P Hypothesis: Identifying Clausal Properties in the Nominal Domain." in M. Baltin and C. Collins (eds) Handbook of Contemporary Syntactic Theory, Oxford: Basil Blackwell, pp. 536-61. Carstairs-McCarthy, A. (1999) The Origins of Complex Language: An Inquiry into the Evolutionary Beginnings of Sentences, Syllables, and Truth, Oxford: Oxford University Press. —(2003) "The Link Between Sentences and Assertion': An Evolutionary Accident?" in R. Elugardo and R. Stainton (eds) Ellipsis and Non-sentential Speech, Dordrecht: Kluwer. Cheney, D. L. and Seyfarth, R. M. (1990) How Monkeys See the World: Inside the Mind of Another Species, Chicago. IL: University of Chicago Press. Chomsky, N. (1995) The Minimalist Program, Cambridge, MA: M I T Press. (2000) New Horizons in the Study of Language and Mind, Cambridge: Cambridge University Press. Davidson, D. (1986) "A Nice Derangement of Epitaphs", in E. Lepore (ed.) Truth and Interpretation, Oxford: Basil Blackwell, pp. 433-46. Fodor, J. (1998) Concepts, Oxford: Basil Blackwell. Gigerenzer, G. (2000) Adaptive Thinking: Rationality in the Real World, Oxford: Oxford University Press. Hacking, I. (1983) Representing and Intervening, Cambridge: Cambridge University Press. Hale, K. and Keyser, S. J. (2002) Prolegomenon to a Theory of Argument Structure, Cambridge, MA: M I T Press. Hauser, M. D., Chomsky, N. and Fitch. W. T. (2002). "The Faculty of Language: What is it, who has it, and how did it evolve?," Science 298: 1569-79. Hinzen, W. (2003) "Truth's Fabric," Mind and Language, 18: 194-219. —(2006) Mind Design and Minimal Syntax, Oxford: Oxford University Press. Hornstein, N. (2001) Move! A Minimalist Theory of Construal, Oxford: Basil Blackwell. Huang, J.-T. (1995) "Logical Form," in G. Webelhuth (ed.) Government and Binding Theory and the Minimalist Program, Oxford: Basil Blackwell. Levi, I. (1980) The Enterprise of Knowledge, Cambridge, MA: M I T Press. Ludlow, P. (2005) "A Note on Alleged Cases of Non-sentential Speech," in R. Eluguardo and R. Stainton (eds) Ellipsis and Non-Sentential Speech, Dordrecht: Kluwer. Jackendoff, R. (2002) Foundations of Language, Oxford: Oxford University Press. Jenkins, L. (2000) Biolinguistics, Cambridge: Cambridge University Press. Künne, W. (2003) Conceptions of Truth, Oxford: Clarendon.

Copyrighted Material

156

Wolfram Hinzen

Pinker, S. and Jackendoff, R. (forthcoming) "What's Special about the H u m a n Language Faculty?" Cognition. Ramsey, F. P. (1926) "Truth and Probability." in R. B. Braithwaite (ed.) (1931) Foundations of Mathematics and other Logical Essays. London: Routledge. Ch. VII. pp. 156-98. Rizzi, L. (1997) "The Fine-structure of the Left Periphery," in L. Haegemann (ed.) Elements of Grammar Dordrecht: Kluwer, pp. 281-337. Rorty, R. (1986) "Freud and Moral Reflection," in J. H. Smith and W. Kerrigan (eds) Pragmatism's Freud: The Moral Disposition of Psychoanalysis, Baltimore, M D and London: John Hopkins University Press, pp. 1-27. Uriagereka, J. (1998) Rhyme and Reason. Cambridge, MA: M I T Press. (2001) "Review of Carstairs-McCarthy (1999)," Language, 77: 368-73. (2002) Derivations. Exploring the Dynamics of Syntax, London: Routledge.

Copyrighted Material

8

Declarative thought, deflationism and metarepresentation John Collins

1 Introduction T h e plausibility of a deflation of t r u t h has d o m i n a t e d recent philosophical discussion of the concept. A b s t r a c t i n g f r o m m u c h complexity, such a deflation proposes that the c o n t e n t s of truth predications are given by the contents to which t r u t h is predicated (see Collins 2002b). On this model, " P is t r u e " is flat with the c o n t e n t of P in the sense that an a d e q u a t e explanation of the semantic role the t r u t h predicate is here playing need advert to n o m o r e t h a n " P " - neither "P," nor its constituent parts, nor their values need be m e n t i o n e d in the explanation. T h e status of the various ways of spelling this out are currently the topic of m u c h debate. W h a t e v e r the merits of deflationism vis-à-vis the traditional disputes over t r u t h , it a p p e a r s t o offer n o g u i d a n c e when thinking a b o u t the illocut i o n a r y role of the t r u t h predicate. To use "true," even where it is disquotable, is not merely to be l o n g w i n d e d . O n the face of it, " t r u e " plays a m e t a r e p r e s e n t a t i o n a l role, allowing expression of a t h o u g h t about the r e p r e s e n t a t i o n of which it is predicated. Of course, by itself, this is not the least p r o b l e m a t i c for the deflationist, w h o m a y off-load such a role o n t o the utility of the truth concept as o p p o s e d to its c o n t e n t proper. T h e following will argue against such a move by presenting an a c c o u n t in which m e t a r e p r e s e n t a t i o n is a constitutive c o n d i t i o n , not a mere utile excrescence, of o u r t r u t h concept. So, o u r capacity to assert P involves the m e t a r epresentational capacity to think of P as true. On the deflationary model, it is as if assertion o r declarative t h o u g h t m o r e generally is antecedently available to us, a n d the provision of a t r u t h concept merely allows us to be c o m p e n d i o u s or otherwise oblique in o u r assertions. I think this a c c o u n t is t h o r o u g h l y b a c k - t o - f r o n t . To possess a conccpt of t r u t h in the guise of the capacity for m e t a r e p r e s e n t a t i o n a l t h o u g h t is what enables assertion o r declarative t h o u g h t in the first place. T h e account is a n i m a t e d by asking what the conditions are for a linguistic structure to be t r u t h - a p t . Let us begin there.

Copyrighted Material

158

Declarative

thought

and

metarepresentation

2 The conditions for truth Whatever else m a k e s for linguistic structures being t r u t h - a p t , it is essential t h a t we view them as being open for the expression of belief, assertion, conjecture, etc. In this section I shall elaborate what I take to be the m e t a r e p r e s e n t a t i o n a l role of such states. U p o n this basis we shall see how o u r concept of t r u t h as enshrined in our conception of t r u t h - a p t sayings is essentially a m e t a r e p r e s e n t a t i o n a l n o t i o n , not a flat n o t i o n that, as it were, directly takes us to the world. Pylyshyn (1978) a n d Perner (1991) have offered a useful a c c o u n t of metarepresentation as a characterization of propositional attitude states (belief, assertion, etc.): a representation is a metarepresentation when it represents a representation as a representation. In other words, it is not e n o u g h if one representation is e m b e d d e d in another, or q u o t e d ; the e m b e d d e d representation must be accessible as a representation t h r o u g h its being so e m b e d d e d . W h a t accessibility a m o u n t s to here is crucial. Let us u n d e r s t a n d representation tout court to be a structure that is t r u t h - a p t , or, m o r e neutrally, satisfaction-apt. 1 A n e m b e d d e d representation, therefore, is accessible as e m b e d d e d if a c o m p e t e n c e with the whole involves the speaker/hearer recognizing it as t r u t h - a p t . This does not m e a n that the speaker/hearer must have full c o m p e t e n c e with the e m b e d d e d structure i n d e p e n d e n t of its being e m b e d d e d . 2 T h a t is, m e t a r e p r e s e n t a t i o n is not necessarily a t r a n s p a r e n t relation for the speaker/hearer; a c o m p e t e n c e with the whole does not entail an equal c o m p e t e n c e with its parts. Still, if. for example, we take belief states to be m e t a r e p r e s e n t a t i o n a l , then their c o n t e n t s (as picked out by the p r o p o s i t i o n a l clauses of o u r a t t i t u d e ascriptions) must be u n d e r s t o o d as potentially i n d e p e n d e n t representations that are currently being represented as belief contents. In o t h e r words, the represented representation (the e m b e d d e d ) is assessable a l o n g a different dimension than the representing representation (the e m b e d d i n g ) . We might take the e m b e d d e d representation to be false, a n d the e m b e d d i n g representation to be true, or vice versa. For Perner, this potential divergence in values is the f o u n d a t i o n of o u r c o m m o n s e n s i c a l theory of mind. T h e developing child ceases to think of the m i n d as a t r a n s p a r e n t w i n d o w on the world, where a c o n t e n t representation would be u n d e r s t o o d as a f r a g m e n t of reality, a n d begins to view o t h e r m i n d s a n d its own as representational, i.e.. how things are for a subject mentally is not necessarily how things are anyhow. In this sense, m e n t a l states are not merely a way of t a k i n g in the world, as it were; to be in a mental state of the a p p r o p r i a t e kind o n e that is t r u t h - a p t involves an essential capacity for reflection. I w a n t to gloss this idea in a particular way. 3 T h e independent representationality of represented representations is functional, not ontological. That is, a structure (a mental state) counts as a representation just because it features as a represented representation: it is not independently representational, i.e.. t r u t h - a p t . Thus, a representation is

Copyrighted Material

John Collins

159

independent of any given attitude state, not of the operation of the system of states. For example, it is typical of attitude explanation that the same contents recur under different state descriptions. Smith desires that P; if Smith believes that if Q, then P, then, ceteris paribus, Smith will try to bring about Q. This explanation schema can be satisfied only if the same contents occur as values of " P " and " Q " respectively. It is in this sense that representations enjoy an independence from any particular mental state type. We may call this characteristic of metarepresentational states transfer independence: (Tl) T h e content of a mental state is independent of any given state it falls under; it can freely transfer from one state to another. 4 N o w such independence appears to entail that a set of representations are available to rational operations, where the representational properties of the representations are already determined. In other words, it is as if truth conditions (and so truth values, context sensitivity of such values apart) are already fixed, and the representations are chosen precisely because of their t r u t h conditions. A n d yet, trivially, sentences or structures do not make assertions, and words do not refer. Speakers make assertions with sentences, and refer to things with words. Just so, a structure counts as a representation because it is employed in some cognitive process; subtract the process and one is left with a functionally inarticulate structure, like marks in sand. We can be explicit about this: ( R E P ) A mental structure is a representation as a function of its employment in metarepresentational cognitive operations over which TI holds. A mental structure is not a representation because of any relation between it and reality independent of its cognitive employment. A n immediate dilemma appears to arise. On the one hand, the idea that representations only c o u n t as such because of their employment in cognitive processes seems to lead inexorably to a "functional role" account of content, where content just is some constrained sum of the relations between mental states. After all, if the representations only count as such because of their role in psychological processes, then from where else does content derive other than the relations between mental states? This position, however, appears to vitiate the very independence enshrined in TI. How can the same content recur across distinct mental states - let alone across different individuals - if content just is the summed effect of the relations under which the content falls? 5 This q u a n d a r y leads to the other extreme that content is a property possessed by structures independent of their playing a role in the system of mental states, i.e., to be representational is not an internal cognitive feature. 6 R E P may coherently stand outside of either of these extremes. The trick, as it were, is to see representations as gaining a u t o n o m y as

Copyrighted Material

160

Declarative

thought and

metarepresentation

enshrined in TI (i.e., not being a function of holistic conceptual role) precisely through their role in the system. Otherwise put, an initial structure type only gets to be a representation because it attains transfer independence. Let us be more concrete. A linguistic structure is a hierarchically organized arrangement of lexical items. T h e organization expresses semantically relevant information, such as binding options, scope interpretation (perhaps governed by a syntactic relation of c-command), argument role (who did what to whom, with what, how, where, when, etc.), elliptical filling-in. etc. The items also express information of "vertical syntax" inherent to them: e.g., the difference between abstract, mass, count and animate n o u n ; the difference between sortally reducible and sortally irreducible determiners (e.g., every and most respectively), verbal aspects (e.g., keeping time (tracking) is different f r o m keeping a car (possession) which is different from keeping M a r y (paying for)), etc. Such fine discrimination encoded in the linguistic structures is precisely what distinguishes one " t h o u g h t " from another. But what does it take for a structure to express a "proposition," be a content? - nothing at all other than its being used by those systems that govern our cognitive goals and general reasoning. It would be mistaken to think of the structures as propositional before their employment, simply because they do not exist prior to their final construction by the speaker: the mind does not contain an infinity of structures awaiting their time in the cognitive spot-light. So, the f r e e d o m there is accrues from the single generative system - the language faculty - being able to reproduce the same structure (type) for states of belief, d o u b t , denial, desire, etc. They count as the same structures because they posses their semantic features inherently as a function of the merging of lexical items that comprise them, but they count as propositional just because they are used - articulated in o u r ascriptional and assertoric practices. It may be helpful to think of the proposal here as the reverse of the usual position. Standardly, propositional attitude states, as the term suggests, are understood in terms of an antecedent notion of truth-aptness that applies, in the first instance, to the representational contents of the states. T h u s the states are taken to be defined over an independent realm of true or false propositions or sentences. T h e present position, on the contrary, views linguistic structures as being formed to be used by mental states, and they acquire their truth evaluability only as so used; they get to be representations (truth-apt) only as metarepresented. What, then, of truth? The t r u t h of sentence S depends on what S "says" and how the world is. T h e world part appears to be wholly absent from the above internalist story, under which representation is simply a notion that drops out of the interface a r r a n g e m e n t of the language faculty and systems of mental states and general reasoning, all of which we take to be functional aspects of the mind/ brain. But truth appears to involve an external relation between mind and world which thus either obtains or not, independent of the internal layout

Copyrighted Material

John Collins

161

of h u m a n representational architecture, or so it seems. Let us consider some likely rejoinders. " T h e account amounts to some kind of solipsism." This charge is mistaken. According to the story being sketched, thinking it so doesn't make it so; we are not burdened with the thought that beliefs, say, are self-guaranteeing or that the "world" is but a mental representation. The reason for this is simply that content representations are characterized by TI, and so, by the architecture of the system, the standing of a given representation is not determined by one's current assessment of it, as expressed by a particular attitude. Just so, the standing of the assessment of the attitude is not determined by the attitude itself, and so on up. In other words, because the object representations are taken to be transfer independent, that one, say, believes P provides no license whatsoever for P tourt court. There might well be a world out there anyway, but we don't need to factor any such notion into our representational wherewithal to answer the charge of solipsism. Solipsism arises only if the standing of an object representation is not independent of the particular metarepresentations we hold it under. The externalist intuition, however, is not so much grounded in any metaphysical claim, but the much more m u n d a n e thought that how things are is on a distinct axis f r o m how things are represented to be. So, the internal layout of thought does not decide which object representations are true or false; one's going right or wrong is not internally specifiable. To deny this is surely to fall into an egregious cognitive idealism. Quite so. Yet, though the independence or objectivity of truth is not be denied, it doesn't follow that a truth-apt system of thought is not internally characterizable. F r o m the above, we may say that the representational properties of a given object representation are not taken to be determined simply by the attitude we take towards it. It is as if, internally speaking, we are landed with representations that do not by themselves decide what attitude we should have towards them. The metarepresentational system allows given representations to fall under a multitude of opposing states; the representations may also go "offline" in the sense of being assumptions f r o m which j u d g m e n t is withheld, or feature in chains of reasoning, or be conditionally and counterfactually embedded. N o given representation, therefore, is marked as true or false in itself, no matter what role it plays. This follows f r o m TI. T h e architecture gives rise, therefore, to the independence of content from any given judgment. This independence, I submit, is not based upon a notion of truth, but is precisely what constitutes our conception of truth: truth is a concept only operative in such a system. It is not so much the world which is independent of us, but the contents available to us vis-à-vis any given j u d g m e n t . This picture may be brought into some relief by considering a different kind of putatively representational system. Imagine a system (perhaps a n o n - h u m a n animal) that is granted "representations" but is unable to metarepresent them under mental states - it

Copyrighted Material

162

Declarative

thought and

metarepresentation

7

lacks reflective j u d g m e n t . We may say that the representations are true or false in our assessment of the system's cognitive activities, but it makes no sense to think of the system itself as possessing a concept of truth precisely because it can't assess its own representations along a right/wrong dimension; it can't reject or c o m m e n d a given representation. T h e point translates to ourselves: truth is not a magical concept that underlies o u r grip on reality; it is a concept we get from being able to submit any representation to assessment and decision as to what attitude we should have towards it, or simply to suspend j u d g m e n t . "But," it will be protested, "what does it take for a representation to he true? Have we just not admitted that a simple, non-metarepresentational system can have true 'thoughts' because of its relation to the world independent of how things are for it internally?" Well, yes and no. We imagined a system which is able to generate structures relevantly similar to the o u t p u t profile of the h u m a n language faculty but which are not integrated, put to use, by a system of thought. This might be an empirical impossibility; indeed, if, following the Minimalist Program (Chomsky 1995. 2000. 2005), we take the faculty to be essentially "designed" to answer to legibility conditions of external systems with which it interfaces, then it may be that we just can't divorce structure from its use in the way imagined. 8 Be that as it may, the imagined system lacks the concept of truth but can be interpreted as tokening true representations. Here, the truths in question devolve o n t o o u r capacities to attribute thoughts, but there appears to be no substantive issue as to how representations m a n a g e to be true in themselves, for the system of which they are a part doesn't allow for such an assessment. To see the point here, consider an inscription in sand of " C h o m s k y is a linguist." It is idle to wonder whether the inscription itself is true or not. We may take it to be true because we can consume it, interpret it into our structure of belief and assessment, but the m a r k s themselves are just patterns in sand; just so for the imagined system. Unless, c o n t r a r y to the scenario, we credit the system with the capacity to employ its representations in a wider metarepresentational structure, no question arises about what it takes for representations in themselves to be true or not, for it counts as a representation just for us, not for the system. 9 Resistance to this position may be expected from the tempting idea that what is involved in the notion of truth is the image of t h o u g h t and world being side by side, such that an item on the thought side counts as true just if it fits with an item on the world side (otherwise it is false). 1 0 This image seems to depict our position as much as that of the simple non-metarepresentational system. The image, however, is misleading. We can put o u r t h o u g h t s side by side with the "world"; we d o so every time we express a belief, make a conjecture, etc. But this precisely involves metarepresentation: a representation of a representation as both that which is believed and as a representation that has a standing independent of our given attitude towards it. O u r thoughts are not side by side with the world anyhow, as it

Copyrighted Material

John Collins

163

were, as we might imagine them to be for the imagined animal. Rather, propositional attitudes generally are a form of self-reporting or reflection. It is incoherent to c o m p a r e t h o u g h t s with the world directly (whatever stuff there is); for the very act of comparison or assessment involves the representation of that with which we are making the c o m p a r i s o n . " Otherwise put, the "world" as that which makes our thoughts true or false is always already conceptualized by the very resources which enter into our object representational states. Unfortunately, this somewhat trivial point has been taken to signal a coherentist or pragmatist understanding of truth, which in turn, due to the familiar problems that beset such accounts, has encouraged the thought that our true thoughts do indeed track "the way the world is anyway." 12 Truth does indeed involve the notion of a comparison with reality free of epistemological or instrumental constraints (insofar as what we represent can be represented as true, independent of other considerations), but the crucial point remains that the kind of comparison enshrined in the "correspondence intuition," however elaborated, presupposes metarepresentational capacities on the part of the person doing the comparing. If there is no-one doing the comparing, then we merely have a metaphysical picture, with a relation only realizable when metarepresentational capacities are employed. It might be further protested that the above line does indeed conflate epistemological concerns of verification or assessment with semantic concerns: the truth of a representation doesn't require a f u r t h e r representation of the relation between thought and world: if the relation obtains, then the representation is true, independent of whether anyone knows it or not. This complaint misfires. The issue before us is whether a bare structure not employed in metarepresentational processes of belief formation or assertion, say, can serve as a representation, be truth-apt. This inquiry doesn't presuppose a denial of snow's being white before any agent had the thought that snow is white: all we need deny is a conception of the world under which it is structured to make our true thoughts true independent of any particular thought being had. On this view, the world does consist of "thinkables" (facts that at the same time are contents of thought (McDowell 1994)), but only from our cognitive, reflective perspective. From the optic of the question of how we can have a world in view that is structured so as to make our t h o u g h t s true, we cannot take the isomorphism as a given; it is something we must arrive at. So, it is not that we must think, still less confirm, that P is true before P can he true; rather, for P to be a thing that is so much as truth-apt, it must be open to metarepresentational processes. To think otherwise would be to hold that a structure in some sense points to a fact, its "truth-maker," as an inherent feature of its structure, independent of its employment in thought. 1 3 The thesis, then, is that the concept of truth arises with an internal metarepresentational system under which content representations have a freedom of transfer such that they can occur under belief or doubt, say, and also be held in abeyance of judgment, or simply be assumed for the p u r p o s e

Copyrighted Material

164

Declarative

thought

and

metarepresentation

of r e a s o n i n g . T h e a p p a r e n t w o r l d - i n v o l v i n g aspect of t r u t h e n s h r i n e d is simply t h e effect of t h e system's a r c h i t e c t u r e u n d e r which, b e c a u s e a r e p r e s e n t a t i o n is i n d e p e n d e n t of a n y given a p p l i c a t i o n of a m e n t a l state, it t a k e s o n a n aspect u n d e r w h i c h it m a y be e v a l u a t e d i n d e p e n d e n t of a n y given o p e r a t i o n of t h e system. A s we saw. however, this very idea p r e s u p p o s e s a system of m e t a r e p r e s e n t a t i o n : the w o r l d - i n v o l v i n g aspect of t r u t h is i m m a n e n t t o o u r cognitive system.

3 Beyond deflationism It m i g h t seem t h a t the a b o v e a c c o u n t of the c o n d i t i o n s f o r p o s s e s s i n g a c o n cept of t r u t h is flush with a positive a c c o u n t of t r u t h itself, viz., d e f l a t i o n i s m . A s earlier flagged, d e f l a t i o n i s m seeks t o e s c h e w m e t a p h y s i c a l issues in favor of c e r t a i n g r a m m a t i c a l - l o g i c a l c o n s i d e r a t i o n s , w h i c h putatively s h o w t h a t the t r u t h p r e d i c a t e e a r n s its c o n c e p t u a l keep, as it were, n o t t h r o u g h b e i n g t h a t which u n d e r l i n e s o u r r e p r e s e n t a t i o n a l c a p a c i t y in general, but t h r o u g h supp l e m e n t i n g such a c a p a c i t y in c e r t a i n necessary ways. T h e r e is i n d e e d s o m e c o n c o r d a n c e b e t w e e n d e f l a t i o n i s m a n d t h e m o d e l I p r e s e n t e d above, b u t t h e r e is a l s o a f u n d a m e n t a l difference. T h e d e f l a t i o n a r y c o n c e p t i o n of t r u t h is n o t m e t a r e p r e s e n t a t i o n a l in the a p p r o p r i a t e sense, a sense clearly e x h i b i t e d by c o m p e t e n c e with t r u t h . T h i s section will explain this difference. T h e result will be t h a t we c a n see o u r way t o a p p r o p r i a t i n g w h a t is right a b o u t deflat i o n i s m whilst d i s c h a r g i n g the a s s u m p t i o n that gives a unity to the collection of a c c o u n t s t e r m e d deflationist. " D e f l a t i o n i s m " is the t e r m given to a f a m i l i a r l y m o t l e y collection of p o s i t i o n s o n t r u t h . W h a t they share, I t h i n k , is a claim a b o u t the c o n t e n t of the t r u t h p r e d i c a t e the exhaustion thesis'. ( E T ) T h e c o n t e n t (/cognitive significance/sense) of the t r u t h p r e d i c a t e is e x h a u s t e d by the c o n t e n t of t h a t t o which it applies, o r a specification t h e r e o f . 1 4 E T is n o t so m u c h a thesis a b o u t w h a t " t r u e " m e a n s , as a severe s c h e m a t i c c o n s t r a i n t o n a c c o u n t s of w h a t " t r u e " m e a n s . Its severity rests in t h e c o n s e q u e n c e t h a t the t r u t h p r e d i c a t e m a k e s n o u n i f o r m c o n t r i b u t i o n t o t h e c o n t e n t s over which it operates. In o t h e r words, there is n o u n i f o r m a n s w e r to t h e q u e s t i o n " W h a t is t r u t h ? " I n s t e a d , t h e c o n t e n t of It is true that snow is white wholly devolves o n t o the c o n t e n t of the s u b o r d i n a t e clause - that snow is white: the c o n t e n t of It is true that the sky is blue similarly devolves o n t o the clause that the sky is blue', a n d so o n . A n y u n i f o r m i t y t h e r e is t o t h e c o n t e n t c o n t r i b u t i o n of the t r u t h p r e d i c a t e in t w o o r m o r e c o n s t r u c t i o n s is a n a c c i d e n t of the c o - o c c u r r e n c e of t h e s a m e lexical m a t e r i a l in t h e c l a u ses t o w h i c h t r u t h is a s c r i b e d (especially see Field (1994) a n d G r o v e r (1992)). Plainly, t h e n , such a c o n s t r a i n t rejects the t r a d i t i o n a l a n a l y t i c a l e n d e a v o r t h a t seeks a u n i f o r m i n s t a n t i a t i o n for the s c h e m a AT:

Copyrighted Material

John Collins

165

(AT) (∀x)[True(x) . . . x . . . ] There just is no story to be told about the independent content of "is true," for it doesn't possess any. The positive part to deflationism consists in a story about the truth predicate playing a logical/grammatical role consistent with ET (details vary between the various proponents). This aspect is often presented as an answer to a query: "If ET is correct, then why do we possess a concept of truth? It appears to be a mere indirection." For example, typically, it is held, via an extension of ET, that there is no more to be said about truth other than the instances of a deflationary equivalence schema DT: 15

(DT) TRUE(x") iff P, where the left flank is a stand-in for various types of truth predication, and the right flank is a stand-in for the clauses to which truth is predicated on the corresponding left-flank instances. The following are examples: (1) a. It is true that snow is white iff snow is white. b. "The sky is blue" is true iff the sky is blue. c. The proposition (/belief/assertion) that grass is green is true iff grass is green. Here, the employment of the truth predicate on the left flanks appears to be superfluous, for by the equivalence relations, one can directly assert the clauses to which truth is predicated without apparent loss of cognitive significance. The deflationist need not further claim that "truth is empty," or even that the flanks are synonymous. All that the deflationist need be committed to is the claim that a complete account of truth, cleaving to ET, is, in some sense, delivered by the instances of DT. If truth were only predicated of full clauses, then DT (or its instances) might well suffice as a complete account. But, clearly, truth can serve as a predicate in general. Here is where the positive aspect of deflationism comes to the foreground. The raison d'être of the concept of truth accrues from its enabling generalizations over clausal position, where the position is subject to quantification: (2) a. Something Bob says is true. b. Mary's last claim is true. c. Every instance of excluded middle is true. The idea here is that often we wish to assert or otherwise assent to a set of sentences too numerous - or indefinite - to articulate directly. The truth predicate enables us to telescope our intent by providing the means to assert compendiously whatever is covered by our term of generalization (typically, a

Copyrighted Material

166 Declarative thought and metarepresentation quantifier noun phrase). Expressed in first-order terms, "is true" is a predicate that rescues the stranded object variable in potential regimentations of the cases in (2). Witness: (3) a.*(∀x)[xis an instance of EM — x].

b. (∀x)[x is an instance of E M — TRUE(x)].

Yet, and this is crucial, such a service does not extend the concept of truth beyond DT or ET, for. when instantiated, each instance of a given generalization may give way to a "disquoted" form as mandated by DT. So, if Bob says "Snow is white," and one wishes to (universally) generalize over clausal position, the truth predicate enables one to do so via an indirect assertion of each instance of If Boh says P. then P. The "world aspect" to truth is provided simply by the disquotation (perhaps better, denominalization) of the clause featured in any given DT instance. For example, if one says that the truth of "Snow is white" depends on what it says and how the world is, then the second part is provided by the very sentence having the assertoric content it possesses, i.e., the use of the sentence articulates all one needs to know about the world for the sentence to be true: (4) It is true that snow is white iff snow is white. There is much to recommend deflationism; in particular, the approach's explicit rejection of the traditional analytical endeavor. Mind, some deflationists present truth as peculiar precisely because it does not give way under analysis, where this claim is thought to support the anti-metaphysical aspect of deflationism (cf. Horwich (1990)). Such reasoning is spurious. It is fair to say that no concept at all has given way under analysis (in the appropriate sense); there is no reason to expect truth to differ. Thus, no particular proposal about truth follows, or is otherwise commended, by our failure to find an adequate analysis of truth: the problem is with the very idea of analysis, not with the particular instance of truth (cf. Collins (2002b)). ET - the claim that the content of truth predications flattens onto the content of that to which truth is predicated is the central thesis of deflationism. The claim contains a presumption about nietarepresentation that is at least questionable. To see this, compare the following two notions of nietarepresentation: (OM) Metarepresentation = The representation of a representation as a representation. (TM) Metarepresentation = The representation of a representation as an aspect of the content of the host representation.

Copyrighted

Material

John Collins

167

The first notion we may call opaque metarepresentation: the represented representation is marked as a representation, but is not necessarily articulated within the (first) representation. In cognitive terms, one can metarepresent R via R + without the availability of R + entailing the availability of R to o n e s wider thoughts. This is what makes the metarepresentation opaque: one can pick out a representation as a representation without it being contentful for one. This might be achieved by recognition of its clausal f o r m . 1 6 It doesn't follow that O M involves mention as opposed to use: I assume that one can use a structure in an embedded context without it enjoying an independent cognitive life, as it were. Thus, metarepresentation is not simply a fancy way of speaking about quotation. 1 7 The second transparent notion says: to represent R via R + is ipso facto to represent R in such a way for it to be available to one's wider thoughts qua an aspect of the content of the host. Otherwise put, R is picked out in terms of its content, and so to employ R + is to understand R. 1 8 My use of the adjectives opaque and transparent is not intended to m a p o n t o the familiar division between contexts that are and are not open to co-referential substitution salva veritate. T h e adjectives are used in strict modification of " m e t a r e p r e s e n t a t i o n . " 1 9 Further, the difference between the two notions is not one of kind. O M is the general case, in which the kind of cognitive access a speaker has towards R is left open; T M is a restricted instance of O M . Hence, in general, one does not have to decide between the two. The issue is to decide where the more restricted notion applies. If we turn to truth in particular, O M appears to be operative, i.e., R may be metarepresented without the content of R being available. In effect, this m e a n s that ET c a n n o t be correct, insofar as the metarepresentational role of truth, if governed by ET, must be flat, content-wise, with the object representations, i.e., the metarepresentation must be of the restricted T M stripe. 2 0 To show, therefore, that the operative metarepresentation is in fact O M is to show that E T is at best incomplete. This will now be shown. First, the only restriction on a truth predication, under any of the disquotational schemata illustrated above, appears to be that the predicated structure has a clausal form. In particular, there is not a f u r t h e r restriction that a subject who understands the truth predication must have access to the content of the clause as an independent representation. To see this, consider the mathematically ignorant faced with some abstruse result, such as the "Taniyma-Shimuru conjecture." (5) appears to be a readily available thought: (5) It is true that every semi-stable elliptical curve is a m o d u l a r form. We may see this by embedding (5) into a wider discourse where understanding is explicitly denied.

Copyrighted Material

168

Declarative

thought and

metarepresentation

(6) It is true that every semi-stable elliptical curve is a m o d u l a r form; it is what Wiles proved on his way to establishing Fermat's last theorem, but heaven knows what it means. This would suggest that there is, at best, a pragmatic presupposition of u n d e r s t a n d i n g , not a semantic one. for we can cancel any such presupposition. That is, a l t h o u g h u n d e r s t a n d i n g is explicitly denied, the initial t r u t h claim remains coherent. In some sense, the availability of the t r u t h claim relies merely upon the clausal f o r m of the c o m p l e m e n t (its being truth-apt), not on the availability of its content. C o m p a r e this case with (7). (7) I believe that every semi-stable elliptical curve is a m o d u l a r form, but I don't k n o w what a semi-stable elliptical curve is, and heaven knows what it is for one to be a m o d u l a r f o r m . Cancellation here strikes me as incoherent: the self-ascription entails that one has access to the content of the clause: it doesn't suffice for one merely to recognize its clausal f o r m . Thus, one c a n n o t have a belief that every semistable elliptical curve is a m o d u l a r form without that content being otherwise available to one. This appears to reflect a general principle that, to use "belief b o x " talk, nothing gets in one's belief box unless it is independently understandable. Interestingly, this f u r t h e r contrasts with the third person case. Plug in Boh for just the matrix subject (only the first "I") of (7) and coherence is restored (with appropriate change of inflection on the matrix verb). T h a t is, we can ascribe propositional attitudes to one a n o t h e r without our having independent access to the contents ascribed. Again, recognition of clausal f o r m appears to suffice. 21 Second, the raison d'être of the truth concept offered by deflationism a p p e a r s to involve metarepresentation not of the T M variety. The case here essentially devolves o n t o the first one. A predication of truth to a quantifier n o u n phrase ( Q N P ) is taken to be a c o m p e n d i o u s way of asserting an indefinite n u m b e r of claims in which truth does not feature as a proper constituent. Thus, Everything Boh says is true gives way, content-wise, to: (8) If Bob says snow is white, then snow is white; and if Bob says grass is green, then grass is green; and . . . This approach brings the apparently a n o m a l o u s case of Q N P s in line with the direct predication of truth to clauses, and so in line with ET. For the purposes of argument, let us assume that some such unpacking is along the right lines, at least in the sense that the truth predicate allows for a metarepresentation over whatever Bob says. It seems clear that on this assumption the metarepresentation is not transparent. That is, for one to have

Copyrighted Material

John Collins

169

access to the thought that everything Boh says is true does not perforce give one access to any content represented by Bob himself (the extension of the Q N P ) . Pretty plainly, I should say, any such access one happens to have is an independent competence, not any aspect of the competence deployed in the understanding of the QNP. The metarepresentation, therefore, is of the more neutral O M variety. Of course, and this is the crucial point, it might well be, as we are assuming, that the kind of conjunction displayed in (8) is semantically related - for the subject - to the Q N P representation, but as just shown above, the conjuncts of (8) are not essentially characterizable in T M terms. To accept the Q N P representation is to accept the conjunction, but such acceptance doesn't mean that one has independent access to any of the embedded content representations. Again, a mere recognition of the clausal forms appears to suffice for competence with the conjuncts of (8). The consequent clauses, free of an explicit self-ascription, appear to work like third person reports. Thus: (9) If Bob says every semi-stable elliptical curve is a m o d u l a r form, then every semi-stable elliptical curve is a m o d u l a r form: Bob is always right about these things, but heaven knows what an elliptical curve is . . . It seems that the metarepresentation involved is the more neutral O M all the way down. Well, the importance of these two related cases is that E T the core deflationary claim depends upon a T M characterization of truth predication, not merely O M . This is because T M guarantees that the contribution of the truth predicate flattens o n t o that to which it is predicated, insofar as the content said to be true is available in the truth predication itself as an exhaustive determination of the host content. If the metarepresentation involved in truth predication is not transparent, then there just is no guarantee that, for the given speaker, an employment of truth devolves o n t o the employment, content-wise, of that claimed to be true. It might, in some cases, but there is no logical, conceptual, or empirical reason to think that it does as a matter of principle. So, the extent to which the above cases are recalcitrant to a transparent reading is the extent to which E T simply fails to characterize the behavior of the truth predicate as an aspect of our linguistic competence. Deflationism is, at best, a partial account. Further, because the assumption of general transparency is undermined, even where the content to which truth is predicated is transparent, the cases are not explained by ET. The question all of this leaves us with is: If we grant that truth competence arises from a cognitive architecture of metarepresentation, then what story, if any, arc we supposed to tell about the concept of truth itself? The next section will seek to supply an answer.

Copyrighted Material

170

Declarative

thought

and

metarepresentation

4 Truth T h e simple a n s w e r to o u r q u e s t i o n is t h a t t r u t h is as d e f l a t i o n i s m c l a i m s save t h a t it is essentially o p a q u e l y m e t a r e p r e s e n t a t i o n a l , n o t t r a n s p a r e n t l y so. T h i s m e a n s t h a t t r u t h , c o n t r a ET, d o e s express a relation. T h i s relation, however, is n o t a r t i c u l a t e d b e t w e e n t h o u g h t a n d w o r l d ; r a t h e r it is a n aspect of the p r e d i c a t i o n a l s t r u c t u r e of t h e o b j e c t r e p r e s e n t a t i o n itself. A given s t r u c t u r e m e t a r e p r e s e n t e d via t h e d e p l o y m e n t of the t r u t h p r e d i c a t e f o r m s a r e p r e s e n t a t i o n of t h e s t r u c t u r e as a r e p r e s e n t a t i o n , as s o m e t h i n g a p t t o f e a t u r e in a s s e r t i o n , j u d g m e n t , belief, etc. For e x a m p l e , t o t h i n k It is true that snow is white is t o r e p r e s e n t t h e c o m p l e m e n t - that snow is white - as a r e p r e s e n t a t i o n ( o n e need n o t fully u n d e r s t a n d - see below) as o p p o s e d t o a n a s p e c t of the w o r l d t h a t s t a n d s against y o u r r e p r e s e n t a t i o n , as it were. Just so, t o t h i n k It is fact that snow is white is n o t merely t o t a k e in t h e w o r l d , as it were, but to represent a r e p r e s e n t a t i o n t o yourself t h a t y o u t a k e to m i r r o r t h e way t h i n g s are. In this light, it w o u l d be m i s l e a d i n g t o say, as if it were theoretically m a n d a t e d , that It is true that snow is white j u s t m e a n s t h a t s n o w is white. A s we saw above, the d e f l a t i o n i s t c a n n o t get away with saying this in g e n e r a l i t y a n y h o w , for " t r u e " applies w h e r e u n d e r s t a n d i n g is a b s e n t . We s h o u l d simply decline f r o m a n s w e r i n g such q u e s t i o n s as " W h a t d o e s ' t r u e ' m e a n ? , " j u s t as the deflationist a d m o n i s h e s us t o decline q u e s t i o n s of t h e f o r m " W h a t is t r u t h ? " 2 2 In c e r t a i n cases, we w o u l d n o t g o w r o n g to c o r r e l a t e a t r u t h p r e d i c a t i o n with its d i s q u o t a t i o n , but f r o m such cases we c a n ' t assert t h a t to t h i n k t h a t P is true is merely t o t h i n k t h a t P, as if any P to w h i c h t r u t h m a y be p r e d i c a t e d is a n t e c e d e n t l y available to us f o r t h o u g h t o r a s s e r t i o n . T h e applicability of t r u t h p r e d i c a t i o n issues simply f r o m t h e r e c o g n i t i o n of the c o r r e c t clausal f o r m ( d e t e r m i n e d by the i n h e r e n t f e a t u r e s of t h e s t r u c t u r e - in simple t e r m s , a d e c l a r a t i v e ) r a t h e r t h a n a n y richer c o n c e p t i o n of w h a t the s t r u c t u r e says. 2 3 T h i s o p a q u e m e t a r e p r e s e n t a t i o n a l role of t r u t h b e c o m e s m o s t clear in cases of w h a t m a y be t e r m e d incorporation, i.e.. w h e r e we i n c o r p o r a t e a n alien s t r u c t u r e (one we d o n ' t u n d e r s t a n d ) i n t o o u r cognitive e c o n o m y u n d e r the t r u t h predicate. T h i s gives us a new r e p r e s e n t a t i o n as a m e t a r e p r e s e n t a tion of a s t r u c t u r e m a r k e d as t r u t h - a p t . even t h o u g h we can't use the struct u r e i n d e p e n d e n t l y - for us. it is a representation only as r e p r e s e n t e d . 2 4 T h e c o n c e p t of t r u t h here gives us access to b a r e clausal f o r m s in t e r m s that m a k e t h e m a p t t o be representations. T h i s m o s t obviously w o r k s in the discursive setting of being c o n f r o n t e d with sentences we d o n ' t u n d e r s t a n d (e.g.. the m a t h e m a t i c a l case above, foreign sentences, etc.). T h e c u r r e n t suggestion, however, is that this kind of i n c o r p o r a t i o n h a s its d o u b l e , as it were, as a f e a t u r e of h u m a n cognitive capacity. O u r possession of a c o n c e p t of t r u t h j u s t is o u r capacity t o represent s t r u c t u r e s for reflective assessment, with these s t r u c t u r e s n o t necessarily being o t h e r w i s e accessible as representations. So, my t h o u g h t is t h a t discursive i n c o r p o r a t i o n , while a p p a r e n t l y a m i n o r aspect of o u r t r u t h c o m p e t e n c e , in fact reflects a f u n d a m e n t a l f e a t u r e of

Copyrighted Material

John Collins

171

h u m a n thought. Attention to the discursive case serves to distinguish the current position from deflationary conceptions. F r o m the general principle above, it seems that one can't believe P unless one otherwise understands P. Yet the truth predicate precisely enables us to incorporate alien structures under belief, assertion, etc. Thus, it does indeed seem that we can understand P is true - it can go in the belief box, as it were - without understanding P itself. The distinctness of this incorporative position from the deflationary understanding may be highlighted by considering the pairs below: (10) a. b. c. d.

Bob believes that snow is white. Bob believes that it is true that snow is white. I believe that snow is white. I believe that it is true that snow is white.

In b, the employment of the truth predicate, which takes the matrix complement of a as its complement, makes no noticeable contribution: the two are equivalent in being opaque, i.e., the speaker need have no understanding of the content attributed to Bob. O n the other hand, there is a difference between c and d: the metarepresentation in c is transparent; in d it is opaque. In other words, for a speaker of c, the complement structure must be otherwise available for thought; d, in contrast, might still be true even if the speaker does not have independent access to the content of the complement of " t r u e " (cf. "I believe that it is true that every semi-stable elliptical curve . . . : but I haven't a clue what elliptical curves are."). So, in d, the truth predicate makes a noticeable contribution: it gives the subject access to a representation as metarepresented. It is crucial to note, however, that the truth predicate is formally behaving the same in the two cases: a structure is represented as a representation via the truth predicate. The difference turns on whether the subject has independent access to the structure such that it may serve as a representation independent of the truth predicate. If the subject has such access, as is entailed by self ascription of belief, then the metarepresentational function of the truth predicate is r e d u n d a n t , but essentially available - it is simply inherent to one's understanding of the structure. If the subject lacks such access, then the truth predicate can still provide a structure for thought opaquely represented, where the only content available is one that essentially employs truth. Similar reasoning holds for truth predications to QNPs. The clausal structure of quantificational subjects with truth predicates m a n d a t e s the inferences to each conjunction/disjunction of the form the deflationist takes to be compendiously expressed by the generalizations. 2 5 But truth does not flatten out on the object representations unless such representations are themselves independently accessible. Further, we may say that generalization over clausal position precisely involves one's relinquishing any claim to be able to think what is covered by the truth predication. We may think

Copyrighted Material

172

Declarative thought and

metarepresentation

of the m e t a r e p r e s e n t a t i o n here as being of the f o r m of a n o p e n s e n t e n c e "x is t r u e , " with " x " r a n g i n g o v e r the d o m a i n fixed by the n o m i n a l restriction of t h e q u a n t i f i e r . T h i s gives us access to a h o s t of p o t e n t i a l r e p r e s e n t a t i o n s a n y t h a t satisfy the restriction as m e t a r e p r e s e n t e d . T h e r e is n o g u a r a n t e e t h a t w e c a n d i s q u o t e , but t h e r e is a g u a r a n t e e that we c a n represent o r i n c o r p o r a t e the new s t r u c t u r e s as true.

Notes 1 Thus, we may include interrogatives, imperatives, etc. under the notion of representation. 2 Recognition of truth-aptness alone, of course, suffices for a certain grammatical competence. For example, one might not understand a declarative active sentence, but still be able to form from it a question (polar interrogative or wh) or the passive, or extraposition, etc. 3 By appealing to this notion of metarepresentation I'm not suggesting that Perner's "theory-theory" account of folk psychological development is correct or even likely to be. My interest is in the synchronic point about the separation of represented and representation, not the diachronic issue of just how the child develops to achieve this cognitive separation. Equally, some other conception of metarepresentation might support my conclusions, although certain constraints must be in play. 4 This is modulo selection restrictions. For example, an interrogative complement of know cannot transfer to be the complement of believe. for the latter selects for declaratives alone. 5 Here I mean what is sometimes referred to as narrow conceptual role (NCR) in contrast to wide conceptual role that factors in external conditions. For example, see Loar (1981), Harman (1982) and Block (1986). NCR is often understood as one component of content to be supplemented with an external notion of truth conditions, for example. Field (1978) and McGinn (1982). My concern is not to decide which, if any, of these notions is viable. For arguments against a range of conceptual role accounts of content, see Fodor and Lepore (1992). The present point is not that NCR cannot constitute some level of content, but that the bare idea of conceptual role individuation appears unable to account for truth in particular, independent of whether truth conditions enter into the individuation of content. 6 Fodor's (1975, 1987. 2003) language of thought (LOT) model fits this extreme. The LOT model has it that contents are expressed by structures whose parts express the contents' constituent concepts: the structure as a whole compositionally expresses the content. Such structures are the bearers of content, but their contents are somehow determined independently. Fodor has played around with a few ideas, especially a causal/informational link between tokenings of concepts and instantiations of properties which are the concepts' extensions. This, however, strikes me as a wholly metaphysical story. I can't envisage any naturalistic inquiry which would seek to find such links, either diachronically or synchronically. In other words, Fodor's naturalism is ontological/reductive, not methodological (cf. Stich (1996)). 7 1 use "reflective judgment" here as an echo of Kant's use of the phrase in the third Critique. Without too much abuse to Kant's understanding, we may think of reflective judgment as our power for metarepresentation, i.e.. the free deployment of concepts (universais) over a range of object representations. 8 By minimalism, the language faculty just is that system which integrates independent systems of sound articulation and conceptuality-intentionality. Thus,

Copyrighted Material

John Collins

9

10

11 12

13

14 15 16

17

18 19

20

173

whatever features we take the faculty to have, they are to be explicable in terms of the faculty's interface conditions or the general biological/physical features of the system (especially see Chomsky (2001)). These remarks are not to be taken as belittling the cognitive capacity of nonhuman animals. If an animal cannot assess its representations and reason offline, then a notion of truth would simply be an idle cog for it. It would be for us to determine what property, if any, we wish to posit to make sense of the integration of the animal's internal states with its environment. And here, for explanatory purposes, truth appears not to be a particularly useful one. for truth is only operative where we have a structure of grammatical complexity that makes it fit to support inferential reasoning of the kind (apparently) not performed by animals. The notion of "fit" may be left loose, for it is the very idea which is of present interest, not any particular realization of it. So, the points to follow should be taken to hold equally against, say, a purely formal notion of fit, as perhaps offered by Wittgenstein (1922), and a "naturalized" one explained in terms of causal relations. See Field (1978), Dretske (1981) and Fodor (1990). Cf. Putnam (1981), Hornstein (1984) and Stich (1990). See, for example. Devitt (1990) and Alston (1996) on Putnam. Putnam (1994) has latterly dropped the notion of a coherence/pragmatist "theory" of truth; cf. Davidson's (1990) rejection of both realist and anti-realist theories of truth. Myopia on this point no doubt has its source in the common appeal to propositions and sentences (defined as belonging to an interpreted language), which are understood to carry their meaning essentially. From the perspective of this paper, sentences/propositions only exist as abstractions from cognitive activities. Thus, the notion of a representation is not an independent 'something' the linguistic mind represents; it is an abstraction of the output of the language faculty. For arguments against the theoretical worth of sentences and propositions see Collins (2003). See Collins (2002b) for a full defense of ET as an appropriate encapsulation of the varieties of deflationism. For example. Quine (1970), Leeds (1978), Soames (1984), Horwich (1990), Field (1994). Cf. Higginbotham's (1989) position that a specification of truth conditions alone reveals a certain clausal form without ipso facto capturing any broader content, such as word meaning. Indeed, quotation itself may be a form of use. See McDowell (1980). In general, the use/mention distinction is not the same as the representation/metarepresented distinction. This understanding of TM corresponds to Recanati's (2000) notion of iconicity, which he understands to be a characteristic of metarepresentation generally. If we assume that propositional attitude contexts are referentially opaque, then the metarepresentational distinction cross-classifies the referential one. That is, referentially opaque contexts can be either transparent or opaque metarepresentations. Insofar as the familiar distinction depends on externalist notions of truth and reference, then it is inapplicable on the current assumptions. The intuitive data, however, remains. For an 'internalist' approach to the data see Collins (2003). At least the logic of this argument should be straightforwardly acceptable to the deflationist who follows Field (1986. 1994, 2002a, 2002b) and takes the content of truth, for a speaker S, to be relative to the conceptual resources of 5. Field, to my mind, is correct in thinking that deflationism as understood via ET entails such relativity (but see the discussion of incorporation below). Others are less comfortable, for example, Horwich (1999) and Soames (1999), but more register

Copyrighted Material

174

Declarative thought and

metarepresentation

such discomfort than explain how truth may be both flat with antecedently determined contents yet not be determined in its content by them. This is the basis of Gupta's (1993) dilemma for deflationism - either truth is relative to speakers or truth is a concept no one possesses. See Collins (2002a, 2002b) for discussion of the former horn. 21 Recanati (2000) argues that all propositional attitude ascription is iconic - in our terms, transparent metarepresentation. Recanati appears to assume that metarepresentation a priori involves transparent representation, so that what one is metarepresenting is always a thought otherwise available to one. Recanati's argument for this claim appears to be simply that one cannot entertain a representation without understanding it (at least up to a Kaplan-esque character): to be in the mind is to be interpreted, as it were. This clearly supports a restriction of metarepresentation to the transparent case only where first person ascription is at issue. That is, one cannot believe P unless P is a content otherwise available to one. The situation is more complex in cases of incorporation (see below), where a putative content is considered under an attitude without its being otherwise available to one's thought (e.g., it is not constituted by items from your lexicon; Lacanian dictums are Recanati's favored examples); here, Recanati appeals to a deferential operator whose argument is the structure under consideration, and whose value is a content an implicitly indicated agent would express by the structure in a similar context (see Collins (2002c) for a similar independent idea). But again, the case is first person, and doesn't generalize to belief ascription tout court. Still, even here, I cannot see why an understanding of clausal form would not suffice to support a truth predication. To a similar suggestion from Sperber (1997), Recanati (1997: 95) complains that it goes against semantic innocence (the truth predication would alter the meanings of the terms). This is simply false, unless one thinks that truemetarepresentation is always transparent over otherwise available representations. Since we are concerned with mental lexicons, not public words/sentences, there is no innocence to be retained other than what the individual speaker understands, and in incorporation cases the speaker understands merely the clausal form. Of course, one takes there to be something more than clausal form at issue, but the content another speaker would express with the clausal form need not enter into deferentially or otherwise - one's representation. Recanati's "deferentialism" is an essentially externalist notion. 22 In general, questions of the type "What does ' F ' mean?" are for the lexicographer, not the philosopher, psychologist or linguist. The felt equivalence of P is true with P is a datum; it is not a theory. The metarepresentational account of truth being spelt out seeks to shed light on this datum, and also the cases where a felt equivalence fails to hold. All this is perhaps a prolegomenon to a proper account of the structure of the concept of truth as an aspect of wider cognitive structure, which our common predicate perhaps only in part reflects, at least in clausal application. We just don't need theories of particular words beyond that which explains their syntactic distribution and any felt relations which mandate inference. 23 We may think of the declarative form as default. That is, interrogation, imperative, etc. require particular features of stress, focus, and movement; in general, they are marked morphologically. This is revealed by the fact that the syntactic declarative form can be employed in the expression of different moods with the supplementation of stress/focus. This is exemplified with English "echo questions." Compare (i)-(iii): (i) What did Bill steal? (ii) Bill stole WHAT? (iii) * Bill stole what?

Copyrighted Material

John Collins

175

(iii) exhibits the standard declarative syntax, with the object of the verb in situ preceded by the head of the verb phrase. It is, however, clearly anomalous. In (ii). with the wh-object receiving stress, however, an interrogative reading is acceptable, which (pragmatic features apart) corresponds to the standard interrogative form of (i), where the raised wh-object is similarly stressed. The same reasoning applies to polar (Yes/No) interrogatives. Compare: (iv) Is Bill tall? (v) BILL is tall? (vi) *Bill is tall? 24 The term incorporation is used by Field (2002a) in a similar way. However. Field employs this feature of truth as part of a defense of a general deflationary position. I have responded to Field elsewhere (Collins 2002c). What does seem clear enough is that where truth is being used for incorporation, then there is a representation (the meta one) available to a speaker/hearer that is not otherwise available, and so the same 'thought' is not available in the form of the object representation, and so the metarepresentation is not cognitively equivalent to the object one. 25 This is a strong (empirical) assumption for the purposes of argument. It is certainly true that we expect there to be a level of syntactic structure (often misleadingly referred to as 'logical form') that encodes those inferences that follow analytically from a generalization as instances, cf. Ludlow (2002). But whether this holds across all natural language Q N P s is at best moot. Determiners such as few, several, many, most, etc. appear not to mandate any inference without an independent determination of the number of 'things' (e.g., Bob's utterances) which are being quantified over.

References Alston, W. (1996) A Realist Conception of Truth, Ithaca, NY: Cornell University Press. Block, N. (1986) "An Advertisement for a Semantics for Psychology," Midwest Studies in Philosophy, 10:615-78. Chomsky, N. (1995) The Minimalist Program, Cambridge, MA: M I T Press. -- (2000) "Minimalist Inquiries: The Framework." in R. Martin, D. Michels, and J. Uriagereka (eds). Step by Step: Essays on Minialist Syntax in Honor of Howard Lasnik, Cambridge, MA: M I T Press. (2001) "Beyond Explanatory Adequacy," MIT Occasional Papers in Linguistics,

20: 1-28.

(2005) "Three Factors in Language Design." Linguistic Incquiry, 36: 1 -22. Collins, J. (2002a) "Truth or Meaning? A Question of Priority," Philosophy and Phenomenological Research, 65: 497 536. —(2002b) "On the Proposed Exhaustion of Truth," Dialogue, 41: 653-79. —(2002c) "Truth: An Elevation," American Philosophical Quarterly, 39: 325-42. (2003) "Expressions, Sentences, Propositions." Erkenntnis, 58: 233-62. Davidson. D. (1990) "The Structure and Content of Truth," Journal of Philosophy, 87: 279-328. Devitt, M. (1990) Realism and Truth, 2nd edn. Oxford: Basil Blackwell. Dretske, F. (1981) Knowledge and the Flow of Information, Cambridge. MA: M I T Press.

Copyrighted Material

176

Declarative thought and

metarepresentation

Field. H. (1978) "Mental representation," Erkenntnis, 13: 9-61. (1986) "The Deflationary Conception of Truth." in C. McDonald and C. Wright (eds) Fact, Science and Value. Oxford: Basil Blackwell. --(1994) "Deflationist Views of Meaning and Content," Mind. 103: 249-85. (2002a) "Postscript to Deflationist Views of Meaning and Content." in Truth andthe Absence of Fact, Oxford: Oxford University Press. (2002b) "Meaning Attributions," in Truth and the Absence of Fact. Oxford: Oxford University Press. Fodor, J. (1975) The Language of Thought. Cambridge. MA: Harvard University Press. (1987) Psychosemantics, Cambridge, MA: M I T Press. — ( 1 9 9 0 ) A Theory of Content and Other Essays. Cambridge, MA: M I T Press. (2003) Hume Variations, Cambridge, MA: M I T Press. Fodor, J. and Lepore. E. (1992) Holism: a Shopper's Guide. Oxford: Basil Blackwell. Grover, D. (1992) A Prosentential Theory of Truth. Princeton, NJ: Princeton University Press. G u p t a , A. (1993) "A Critique of Deflationism." Philosophical Topics, 21: 57-81. H a r m a n , G. (1982) "Conceptual Role Semantics." Notre Dame Journal of Forma! Logic, 23: 242-56. Higginbotham, J. (1989) "Knowledge of Reference." in A. George (ed.) Reflections on Chomsky, Oxford: Basil Blackwell. Hornstein. N. (1984) Logic as Grammar, Cambridge. MA: M I T Press. Horwich, P. (1990) Truth, Oxford: Basil Blackwell (1999) "The Minimalist Conception of Truth." in S. Blackburn and P. Simmons (eds) Truth, Oxford: Oxford University Press. Leeds, S. (1978) "Theories of Reference and Truth." Erkenntnis, 13: 111-29. Loar, B. (1981) Mind and Meaning, Cambridge: Cambridge University Press. Ludlow, P. (2002) " L F and Natural Logic," in G. Preyer and G. Peter (eds) Logical Form and Language. Oxford: Oxford University Press. McDowell, J. (1980) "Quotation and Saying That." in M. Platts (ed.) Reference, Truth, and Reality: Essays on the Philosophy of Language, London: Routledge and Keegan Paul. McDowell. J. (1994) Mind and World. Cambridge, MA: Harvard University Press. McGinn, C. (1982) "The Structure of Content." in A. Woodfield (ed.) Thought and Object, Oxford: Oxford University Press. Perner, J. (1991) Understanding the Representational Mind, Cambridge. MA: M I T Press Putnam, H. (1981) Reason, Truth, and History. Cambridge: Cambridge University Press. (1994) Words and Life. Cambridge, MA: Harvard University Press. Pylyshyn, Z. (1978) "When is Attribution of Beliefs Justified?," Brain and Behavioral Sciences, 1: 592-93. Quine. W.V. (1970) Philosophy of Logic, Cambridge, MA: Harvard University Press. Recanati, F. (1997) "Can We Believe What We Don't Understand?." Mind and Language, 12: 84-100. —(2000) Oratio Obliqua, Oratio Recta: An Essay on Metarepresentation. Cambridge, MA: M I T Press.

Copyrighted Material

John Collins

177

Soames, S. (1984) "What is a Theory of Truth?," Journal of Philosophy, 81: 411-29. (1999) Understanding Truth, Oxford: Oxford University Press. Sperber, D. (1997) "Intuitive and Reflective Beliefs," Mind and Language, 12: 67-83. Stich, S. (1990) The Fragmentation of Reason: Preface to a Pragmatic Theory of Cognitive Evaluation, Cambridge, MA: M I T Press. (1996) Deconstructing the Mind, Oxford: Oxford University Press. Wittgenstein, L. (1922) Tractatus Logico-Philosophicus, London: Routledge and Keegan Paul.

Copyrighted Material

9

The concept of truth and multiple facets of the speech-act equivalence thesis concerning "true"1 Bo Mou

In this essay I intend to argue for several points. First, t h o u g h providing a strong intuitive basis for o u r u n d e r s t a n d i n g of t r u t h , the speech-act equivalence thesis concerning " t r u e " in o u r language (E) To say of a sentence that it is true is equivalent to using that sentence is b o t h a m b i g u o u s a n d vague. (E) is s o m e h o w associated with some distinct equivalence theses as its (actual or alleged) multiple facets, which are underlain by a variety of distinct pre-theoretic u n d e r s t a n d i n g s c o n c e r n i n g t r u t h a n d / o r "true." O n e thus c a n n o t establish one's philosophical conclusion concerning t r u t h a n d / o r " t r u e " based on (E) without first clarifying (E) a n d spelling out exactly which equivalence one is talking a b o u t . Second, three (actual or alleged) variants of (E), the Tarski-style non-epistemic semantic equivalence, the Frege Ramsey style epistemic p r a g m a t i c equivalence, the Strawson style non-epistemic p r a g m a t i c equivalence serve their distinct p u r p o s e s with their different structures a n d contents; they thus should not be conflated with each other. Third, based on the preceding discussion, a n u m b e r of conclusions or d u e elaborations have been drawn c o n c e r n i n g the relation between the concept of t r u t h a n d the speech-act equivalence thesis (E) a n d c o n c e r n i n g some methodological strategies in treating the philosophical issue of t r u t h . In the following, my strategy is this. First, in Section 1, I explain how (E) a p p e a r s to be linguistically a n d conceptually compatible with, a n d actually subjected to, various e l a b o r a t i o n s into a n u m b e r of equivalence theses. Second, in Sections 2, 3, a n d 4, I respectively analyze three distinct equivalences, actual or alleged v a r i a n t s of (E). Third, in Section 5, I d r a w a n u m b e r of conclusions based on the preceding discussion.

1 Multiple facets of the speech-act equivalence thesis concerning "true" In the c o n t e m p o r a r y inquiries into the philosophical issue of t r u t h , the foregoing intuitive speech-act equivalence thesis, (E), is often resorted to, or based u p o n , for the sake of examining the raison d'etre and f u n c t i o n s of the

Copyrighted Material

Bo Mou

179

t r u t h predicate a n d u n d e r s t a n d i n g the concept of t r u t h . A l t h o u g h providing a strong intuitive basis for u n d e r s t a n d i n g of the truth predicate in o u r language a n d the concept of t r u t h , (E) is b o t h a m b i g u o u s a n d vague. It is clear that one should not, a n d indeed c a n n o t , convincingly establish one's philosophical conclusion concerning t r u t h w i t h o u t first clarifying (E) a n d spelling out exactly which equivalence o n e is talking a b o u t , especially when some of those actual o r alleged variants of (E) a r e related to (E) in a m o r e complicated way t h a n what a p p e a r s to be. It is n o t e d that, in so f a r as (E) itself is a colloquial expression in o u r o r d i n a r y language, it would not be a b n o r m a l that (E) has a certain degree of ambiguity a n d / o r vagueness. Nevertheless, the point is that, when seeking an intuitive basis f r o m some (E)-like equivalence for the sake of drawing a certain philosophical conclusion c o n c e r n i n g t r u t h , o n e needs to clarify exactly which equivalence o n e is really talking a b o u t a n d u n d e r s t a n d its due relation to (E). In the remaining p a r t of this section, I present three m a j o r , actual o r alleged, v a r i a n t s of (E). In the subsequent sections, I then f u r t h e r examine each of them. As f a r as its a m b i g u i t y is concerned (E) claims either: in cm assertoric context in which an assertion is m a d e , 2 that a speaker's speech-act of saying of a sentence x that it is true a n d her speech-act of just uttering the sentence x (to assert x or to claim the t r u t h of x) m e a n the same; or, in a non-assertoric context, that the two speech-acts have the same illocutionary force. (It is noted that, in such contexts, two speech-acts that have the same locutionary m e a n i n g - a m e a n i n g which does not depend on whether o r not the t r u t h predicate is used t o give a description o r m a k e a statement - m a y have different illocutionary f u n c t i o n s - f u n c t i o n s which d o d e p e n d on w h e t h e r or not the truth predicate is used to serve some illocutionary purpose.) 3 In the former case, (E) would be elaborated into the following Frege-Ramsey-style meaningequivalence thesis (Frege 1892: 64; R a m s e y 1927: 16-17): ( F R ) One's claim that p is true m e a n s n o m o r e than what one's utterance p means; or m o r e accurately, ( M ) For any u t t e r a n c e p t h a t a person X u n d e r s t a n d s , her claim that p is true m e a n s (for X) the same as her assertion p. where " p " is universally instantiated by the q u o t a t i o n n a m e of any sentence of a certain language a n d , in X's claim that p is true, the p r o p e r t y of t r u t h is ascribed to p. In the latter case, (E) would be elaborated into the following illocutionary-force equivalence thesis (Strawson 1950 a n d 1964): (I) For any u t t e r a n c e " q " that a person X understands, X's u t t e r a n c e "'q' is t r u e " has the same illocutionary force (for X) as X's utterance "q,"

Copyrighted Material

180

The speech-act

equivalence

thesis

where "q" is universally instantiated by any sentence of a certain language. N o t e that, in X's utterance "q is true," " t r u e " is regarded not as a descriptive term but as a performative term, playing a role in quite different speech-acts like agreeing, endorsing, conceding, etc. A n d , as far as its vagueness is concerned. (E) is sometimes conflated with the Tarski-style extensional equivalence thesis: (T) x is true in L if and only if p or taken to be a speech-act version of (T). 4 In the subsequent sections, let me begin with my examination of (T). and then (M) and (I).

2 (E) and the Tarski-style non-epistemic extensional equivalence (T) In this section, through examining the source, status and nature of the Tarski-style equivalence (T) in a way relevant to the current topic, I intend to argue: (1) (T) does not stem f r o m (E) but has its independent source; (2) (E) gives a pragmatic expansion of (T) by explicitly or implicitly referring to some pragmatic elements involved in speech-acts that would be irrelevant in the original semantic context of (T); (3) (T) and (E) are thus significantly distinct f r o m each other; (4) therefore any attempt that alleges to elaborate and defend the point of (T) by conflating (T) and (E) and somehow resorting to (E) to argue for the case not merely owes us an explanation of the conflation but is seriously mistaken. Perhaps the basic point about truth on which most people could agree is the following pre-theoretic "correspondence" understanding: a true sentence or statement describes things as they are. T h o u g h appearing to be naive, this point is often simply expressed in such uncontroversial and plain conditionals in ordinary language as: (RLT1) If snow is white, then "Snow is white" is true and (RLT2) If snow is not white, then "Snow is white" is not true (or false)

(RLT2') "Snow is white" is true only if snow is white), which I call "real-life truth-conditionals." For the sake of convenience of discussion, (RLT1) and (RLT2') are combined into one: (RLT) "Snow is white" is true if. and only if, snow is white

Copyrighted Material

Bo

Mou

181

which I call a "real-life t r u t h - c o n d i t i o n equivalence" o r "real-life T-sentence." O n e of the crucial f e a t u r e s of o u r p r e - t h e o r e t i c u n d e r s t a n d i n g of truth is t h a t a sentence, say, " S n o w is w h i t e " is made true by its truth condition of snow's being white, n o m a t t e r w h a t k i n d of o n t o l o g i c a l s t a t u s s n o w ' s b e i n g white h a s , 5 r a t h e r t h a n ascribed truth by an epistemie agent.6 T h i s f e a t u r e is m o r e o r less reflected by the c h a r a c t e r i s t i c s of the l e f t - h a n d side of ( R L T ) : w h a t " ' S n o w is white" is t r u e " m e a n s s h o u l d n o t be u n d e r s t o o d as w h a t is expressed by t h e p h r a s e "An epistemie a g e n t ascribes t r u t h (by a d d i n g the t r u t h p r e d i c a t e ) t o the s e n t e n c e ' S n o w is white"' (or "An epistemic a g e n t c l a i m s t h a t ' S n o w is w h i t e ' is t r u e " ) b u t as w h a t " ' S n o w is w h i t e ' is t r u e " m e a n s . In this sense, o u r real-life t r u t h - c o n d i t i o n e q u i v a l e n c e s a r e non-epistemic. it d o e s n o t involve a n epistemie agent's c e r t a i n e p i s t e m i e a t t i t u d e t o w a r d ' S n o w is white'. A n o t h e r f e a t u r e is t h a t such a n o n - e p i s t e m i c m a k i n g - t r u e relation is a non-linguistic relation, say, b e t w e e n the linguistic item ' S n o w is w h i t e ' a n d the extra-linguistic item, s n o w ' s b e i n g white. G i v e n t h a t s e m a n t i c s is the s t u d y of t h e n o n - l i n g u i s t i c relation b e t w e e n linguistic e x p r e s s i o n s a n d t h e extra-linguistic o b j e c t s f o r which the linguistic expressions s t a n d , o u r real-life t r u t h - c o n d i t i o n e q u i v a l e n c e s a r e semantic in n a t u r e . In this way, o u r real-life t r u t h - c o n d i t i o n e q u i v a l e n c e s a r e non-epistemic a n d semantic e q u i v a l e n c e s r e g a r d i n g non-linguistic t r u t h in t h e prec e d i n g senses. T h e t w o a f o r e m e n t i o n e d f e a t u r e s of o u r p r e - t h e o r e t i c u n d e r s t a n d i n g of t r u t h a c t u a l l y c o n s t i t u t e t w o c e n t r a l p o i n t s of the locut i o n a r y " c o r r e s p o n d e n c e " d i m e n s i o n of the m e a n i n g of t h e t r u t h p r e d i c a t e " t r u e " in o r d i n a r y l a n g u a g e . T a r s k i believes t h a t real-life t r u t h - c o n d i t i o n e q u i v a l e n c e s t o g e t h e r with their c o n s t r u c t i o n rules in o r d i n a r y l a n g u a g e c o n s t i t u t e a g o o d o r even t h e best i n f o r m a l c h a r a c t e r i z a t i o n of o u r p r e - t h e o r e t i c u n d e r s t a n d i n g of t r u t h . 7 In o t h e r w o r d s , T a r s k i t h i n k s t h a t , in this r e g a r d , real-life T - s e n t e n c e s a r e m u c h b e t t e r t h a n a n y " p h i l o s o p h i c a l " f o r m u l a t i o n in o u r n a t u r a l l a n g u a g e o r even b e t t e r t h a n A r i s t o t l e ' s w e l l - k n o w n f o r m u l a t i o n . 8 T o c a p t u r e o u r pret h e o r e t i c u n d e r s t a n d i n g of t r u t h in a c o n s i s t e n t l a n g u a g e f r e e of t h e s e m a n t i c p a r a d o x e s , T a r s k i follows the suit, o r the p a t t e r n of ( R L T ) , of reallife T - s e n t e n c e s in o r d i n a r y l a n g u a g e to c o n s t r u c t his s c h e m a (T) as a f o r m a l c o u n t e r p a r t of t h e generality of real-life T - s e n t e n c e s in a wellb e h a v e d first-order l a n g u a g e : (T) x is T r u e 9 in L if a n d only if p w h e r e " x " is replaced by a n y n a m e of a n y sentence of t h e l a n g u a g e L f o r which T r u t h is being d e f i n e d , a n d "p" is replaced by t h e translation of t h e s e n t e n c e in t h e m e t a - l a n g u a g e ML of L. Tarski believes t h a t T - s e n t e n c e s , t h o s e i n s t a n c e s of the s c h e m a (T), t o g e t h e r with their f o r m a t i o n rules w o u l d c a p t u r e t h e essential f e a t u r e s of o u r p r e - t h e o r e t i c u n d e r s t a n d i n g of t r u t h . T h a t seems t o be why t h e first c a n d i d a t e for a definition of t r u t h t h a t T a r s k i n a t u r a l l y c o n s i d e r s is a list-like definition (for a finite l a n g u a g e ) : it is t h e

Copyrighted Material

182

The speech-act equivalence

thesis

conjunction of all the relevant T-sentences (in the sense of "relevant" to be explained below). Only after consideration of some technical difficulties does Tarski turn to his recursive definition via satisfaction. 1 0 Then, for any definition which is intended to capture our pre-theoretic understanding of truth, Tarski provides the following criterion for its material adequacy, i.e., Convention T: such a definition is an adequate definition of truth if the definition has the following (logical) consequences: 1 1 all those instances of the Schema (T) where "x" is replaced by a structurally descriptive name of the sentence for which Truth is being defined and whose translation in ML replaces "p" (Tarski 1933: 187-8; Tarski 1944: 50). I call those instances that meet such conditions "the relevant T-sentences (built on the predicate 'is True')" t h r o u g h o u t this work. It is clear that the point of Tarski's Convention T is to establish the essential connection of an adequate definition of t r u t h with our pre-theoretic understanding of truth via a certain strong logical connection (logical consequence) of the definition with those relevant T-sentences. In such maneuvers what is interesting and significant is this: Tarski takes a direct a n d least metaphysically loaded a p p r o a c h in order to capture the essential features of our real-life notion of truth. T h a t is, Tarski directly generalizes people's ordinary and plain expressions of their pre-theoretic u n d e r s t a n d i n g of t r u t h in particular cases, the real-life T-sentences, in the f o r m of schema (T), instead of resorting to any traditional (quasi-) philosophical formulations which are either unclear or metaphysically loaded. Given that the real-life T-sentences are most uncontroversial, clear a n d plain expressions of our pre-theoretic u n d e r s t a n d i n g of truth concerning particular cases (via employing the truth predicate in natural language), and given that the generality of truth consists in those conceptual elements that universally or generally a p p e a r in all real-life T-sentences, Tarski's strategy of capturing (at least some of the crucial features of) our pre-theoretic understanding of non-linguistic truth via T-sentences a n d thus Convention (T), in my opinion, is tenable a n d in a correct direction. 1 2 In this way, it is evident that (T) stems from Tarski's intention to capture our pre-theoretic understanding of t r u t h as delivered in o u r real-life Tsentences which constitute the due source of (T) and thus Convention (T). (T) is a non-epistemic equivalence thesis serving some semantic p u r p o s e instead of some pragmatic purpose. In sum, (T) does not stem f r o m (E) but has its independent source as explained. Rather, (E) gives a pragmatic description, or, more accurately, a pragmatic expansion, of (T) by explicitly referring to the speech-acts per se ["to say of a sentence that . . . " a n d "using that sentence"], which are irrelevant in the original context of (T), a n d implicitly referring to the subject or agent of the speech-acts ["(for one) to say of a sentence that . . . " and "(one's) using that sentence"], which is also originally irrelevant in the semantic context of (T). In this way, it would be misleading to simply say that (T) is an elaboration of (E) in the sense that (T) gives some f u r t h e r details on the basis of those that have already been associated with (E) to

Copyrighted Material

Bo

Mou

183

serve some p r a g m a t i c purpose. Rather, one might as well say that, with its own semantic purpose, (T) " d e f l a t e d " (E) by leaving out irrelevant pragmatic elements to its semantic context. Indeed, the a f o r e m e n t i o n e d pragmatic expansion of (T), resulting in (E) or some of its truly associated variants, actually t u r n s (T) into some equivalence that is categorically distinct f r o m (T). To this extent, it would be seriously misleading to call (E) as "a . . . version of (T)" or consider (T) as a variant of (E), that is, s o m e t h i n g that essentially delivers the same message a n d serves the same p u r p o s e as (E) does. This point can be m o r e clearly seen when we f u r t h e r examine two genuine, instead of alleged, variants of (E), i.e., ( M ) a n d (I), which are distinctive pragmatic elaborations of (E) as two speech-act equivalences t h o u g h serving their distinct speech-act purposes.

3 (E) and the Frege-Ramsey style equivalence thesis (M) In c o n t r a s t to (T) as a non-epistemic semantic thesis whose u n d e r s t a n d i n g is based u p o n o u r pre-theoretic u n d e r s t a n d i n g of non-linguistic t r u t h , ( M ) For any u t t e r a n c e p that a person X understands, her claim that p is true m e a n s (for X) the same as her assertion p is an epistemico-pragmatic thesis, whose u n d e r s t a n d i n g is based u p o n o u r pre-theoretic u n d e r s t a n d i n g s of the following two dimensions of what Q u i n e calls " d i s q u o t a t i o n " use of the t r u t h predicate (in regard to p h o n e t i c natural languages like English), a n d which involves an epistemic a t t i t u d e t o w a r d s the sentence to which the t r u t h predicate " t r u e " is added. (1) A sentence is explicitly given by its structurally descriptive n a m e ( f r o m the l e f t - h a n d side of " m e a n s the same a s " in ( M ) to its right-hand side). In most situations, when saying something by a sentence, we believe it a n d take it to be true instead of merely expressing a t h o u g h t without acknowledging its t r u t h . T h a t is, with such a n epistemic attitude being involved, we assert the sentence. As Frege analyzes it, when explicitly giving a sentence by uttering its structurally descriptive n a m e x (e.g., " S n o w is white"), we can m a k e an assertion of the sentence simply by uttering the sentence with assertoric force w i t h o u t using the t r u t h predicate. 1 3 This is one dimension of the " d i s q u o t a tion" use of the t r u t h predicate. This dimension might as well be called "the presupposed d i m e n s i o n " of the " d i s q u o t a t i o n " use of the t r u t h predicate; for this dimension is presupposed by the manifest " d i s q u o t a t i o n " use. or the manifest dimension of such use, as given below. (2) A sentence s is not explicitly given (from the r i g h t - h a n d side to the left-hand side). As far as phonetic natural languages like English are concerned, when a sentence is not explicitly given but is instead given by a n o n displaying or non-structurally descriptive n a m e (such as "Tarski's favored sentence"), the direct assertion of s c a n n o t be realized simply by uttering the n a m e alone; for the assertoric force in question would be available only

Copyrighted Material

184

The speech-act

equivalence

thesis

when the sense of (or thought expressed by) s to be asserted is expressed explicitly. The truth predicate is then useful in this context: when it is added to the name, " t r u e " plays a syntactic role to restore the structure of a complete sentence by semantically ascending to talk about s, and also plays a semantic role as a device of u n d o i n g the effect of semantic ascent so as for X to assert s or claim that the world is as s says it is. 14 This is the manifest dimension of the M i s q u o t a t i o n " use of the truth predicate. It is clear that, although both sides 15 of (M) as an elaboration of (E) in an assertion-context convey the same information, the two sides of (T) convey essentially different pieces of information: the concepts needed to u n d e r s t a n d one side are not necessarily needed, or are simply not needed, to understand the other side. Consider such a T-sentence as " T a r s k i s favored sentence is true if and only if snow is white" (given that "Tarski's favored sentence" is a name of "snow is white"). On the one hand, to understand its right-hand side one has to understand the concepts indicated by such terms as "snow" and "is white," while to understand its left-hand side one does not need to understand them. On the other hand, at least from the point of view of our pre-theoretic understanding of truth, to understand its left-hand side one has to understand the sense of "is true," while to understand the right-hand side one does not need to understand the sense of "is true." In this way, generally speaking, the two sides of (T) are simply not synonymous. 1 6 For these two reasons, it is not the case that (T) and (M), and thus (T) and (E), are just each other's c o u n t e r p a r t s with all other things being equal except for different equivalence-connectors: they characterize different things and serve different purposes. In the remaining part of this section, let me highlight the significant distinction between (T) and (M) through analyzing how, in a representative deflationist argument, the conflation of both would result in some seriously mistaken philosophical conclusion. It is arguably right that the core idea of deflationism consists in the deflationary thesis about truth as follows. Truth is not substantive either in the metaphysical sense (the truth property, if any, c a n n o t exist independently of the logical or practical functions of the truth predicate) or in the explanatory role sense (truth does not play any significant explanatory role in philosophical inquiry, beyond and separate from the practical or logical roles of the truth predicate). To argue for this thesis, deflationists typically build up their cases on (E), a seemingly intuitive speech-act equivalence thesis regarding " t r u e " in our language: To say of a sentence that it is true is equivalent to using that sentence. 1 7 In some contemporary deflationists' writings, the intuitive (E)-like equivalence is somehow elaborated into the Tarski style equivalence: (T) x is true in L if, and only if. p; 18

Copyrighted Material

Bo Mou

185

(TP) T h e proposition that p is true if, a n d only if, p. T h o s e c o n t e m p o r a r y deflationists tend to argue for the deflationary thesis about truth on the basis of their linguistic observation of the uses of the truth predicate, which they claim are shown by (T). Let us examine one typical a r g u m e n t along this line. This a r g u m e n t is based u p o n the following linguistic observation a n d analysis of the logico-syntactic f u n c t i o n of "true": the truth predicate in o u r natural language is often used to restore the structure of a sentence or to express certain kinds of generalization. Let us look at a representative deflationary explanation, (1.1a) through (1.1 d) below, of how the t r u t h predicate plays its logical role. This explanation was presented by a leading deflationist in the literature (see Horwich 1995: 813). N o t e that the point of the deflationary analysis f r o m (1.1a) through ( l . l d ) is to argue neither that the first-order objectual quantification is m o r e natural t h a n one that q u a n tifies over sentential variables n o r that we c a n n o t eliminate the t r u t h predicate in ordinary language; it takes b o t h of these claims for granted. Its point is to provide an explanation of the raison d'être of the t r u t h predicate in o r d i n a r y language, i.e., why the truth predicate c a n n o t be eliminated in o u r ordinary language and how it plays its logical role. Suppose that you believe Einstein's claims about physics. Suppose that, u n k n o w n to you, Einstein's last claim about physics was that q u a n t u m mechanics is wrong. Exactly which proposition becomes the a p p r o p r i a t e object of your belief in his last claim about physics? W h a t is needed is something equivalent to the infinite conjunction: (1.1a) If what Einstein said was "nothing goes faster than light," then nothing goes faster than light, and if what he said was " q u a n t u m mechanics is wrong," then q u a n t u m mechanics is w r o n g , . . . and so on. However, if you want to get the effect of asserting all those conjuncts, you would have to quantify, replacing sentences with variables. But it seems that the above infinite c o n j u n c t i o n c a n n o t be summarized using the ordinary universal quantifier, "every," which generalizes over sentential variables, in this way: For every proposition, x, if what Einstein said = x, then x. For, in English, the quantification is considered to be most strued as first-order a n d objectual. 1 9 But, with the help of express the infinite c o n j u n c t i o n with the first-order, objectual According to disquotationalist deflationism, on the basis of (T), the initial c o n j u n c t i o n (1.1a) may be reformulated as:

Copyrighted Material

naturally con"true," we can quantification. the equivalence

186

The speech-act

equivalence

thesis

(1.1b) If what Einstein said was "nothing goes faster than light," then " n o t h i n g goes faster than light" is true, and if what he said was " q u a n t u m mechanics is wrong," then " q u a n t u m mechanics is w r o n g " is true, . . . and so on. And this can be summarized using the ordinary universal "every," which generalizes over objects:

quantifier,

(1.1c) For every object, x, if what Einstein said = x, then x is true. Or, stated in the following natural-language expression, ( l . l d ) What Einstein said is true. The above deflationary analysis is intended to explain why the truth predicate cannot be eliminated in ordinary language and how it plays its logical role. T h e logico-syntactic function of the truth predicate might be called "de-nominalization" (Horwich 1998: 5). Now. on the basis of the linguistic analysis of the logico-syntactic function of "true," disquotationalist deflationists f u r t h e r claim: first, "the truth predicate exists solely for the sake of a certain logical need" (Horwich 1998: 2); that is, the raison d'être of " t r u e " consists in its purely logical function: second, (T) just shows how this function is fulfilled (see Horwich 1998: 2. 5). Then, to establish the deflationary thesis about truth, they would argue this way: 2 0 (1.1) (From the above linguistic analysis) (T) just shows how the truth predicate plays its logical role; (1.2) (Truism) O u r basic pre-theoretic understanding of truth is characterized by (T); (1.3) (From (1.1) and (1.2)) Truth consists just in the logico-syntactic function of the truth predicate; the notion of truth does not play any explanatory role except for a certain logico-syntactic role of "true"; (1.4) (From (1.3)) The deflationary thesis about truth: truth is neither substantive in the metaphysical sense nor in the explanatory-role sense. In the following, I explain why and how. in this representative deflationist argument, the Tarski-style equivalence thesis (T) is mistakenly used to argue for the deflationary thesis about truth. As a result, although disquotationalist deflationists give a good analysis of the logico-syntactic functions of the truth predicate in our ordinary language, they fail to justify their philosophical deflationary thesis about truth by this argument. What we need to examine in the first step (1.1) of this argument is whether it is (T) as deflationists claim, or something else, which shows how the truth predicate plays its purely logical role; that is, we need to clarify exactly which one, a m o n g the two intuitive equivalences (M) and (T), is really involved in

Copyrighted Material

Bo

Mou

187

showing how the truth predicate plays its purely logical role. O n e difficulty with disquotationalist deflationism, in my opinion, is its conflation of (M) with (T) in (1.1) of this a r g u m e n t . Such a conflation is m a d e not merely by deflationism but also by some o t h e r writers for different reasons in various ways. Either (T) a n d ( M ) are conflated with each o t h e r by attributing some characteristic features of ( M ) t o (T) - this is the case of disquotationalist deflationism; o r they are taken to be almost the same in the sense that (T) a n d ( M ) are regarded as essentially each other's c o u n t e r p a r t s with all other things being equal except for seemingly different equivalence-connectors. 2 1 F o r the reasons I explained at the outset of this section, I think this conflation is incorrect. N o w let us look at where the a r g u m e n t becomes problematic. As explained before, the indispensable logico-syntactic f u n c t i o n of " t r u e " in o u r linguistic practice is characterized by the epistemie p r a g m a t i c equivalence ( M ) rather t h a n by the non-epistemic semantic equivalence (T). However, in the argum e n t , the characteristic features of ( M ) are mistakenly attributed to (T). As a result, (1.1b) would n o t say the same thing as (1.1a). Actually, disquotationalist deflationists could avoid this pitfall by explicitly resorting to ( M ) rather t h a n to (T) in the s u b - a r g u m e n t f r o m (1.1a) t h r o u g h ( l . l d ) . For any c o n j u n c t of (1.1a) essentially involves X's epistemic attitude: for example, the c o n j u n c t "If what Einstein said was ' n o t h i n g goes faster t h a n light,' then n o t h i n g goes faster t h a n light" really m e a n s that, if what Einstein said was " n o t h i n g goes faster than light," then " n o t h i n g goes faster than light" will be an assertion m a d e by X (or, then " n o t h i n g goes faster t h a n light" will be uttered by X with assertoric force); thus, (M), rather t h a n (T), fits. So, if (1.1a) is reformulated as (1.1b) by appealing to (M), (1.1b) would m e a n the same as (1.1a). However, in that case, deflationists would win the battle but lose the war. For, in that case, (1.1) would b e c o m e the thesis (1.1)* to the effect that (M), rather t h a n (T), just shows h o w " t r u e " perf o r m s its logical role; nevertheless, f r o m (1.1)* a n d (1.2), (1.3) a n d then (1.4) d o not follow. At this point, we can see that the conflation of (T) with ( M ) plays a crucial role in this deflationist a r g u m e n t , which otherwise would be invalid. However, when deflationists conflate (T) with (M), the a r g u m e n t becomes u n s o u n d . Because the characteristic features of ( M ) a r e incorrectly attributed t o (T), a n d because of the n a t u r e of (T) as characterized before, the a r g u m e n t fails to interpret properly the point of (T). In this way, a l t h o u g h the a r g u m e n t is valid, a n d a l t h o u g h one can agree to (1.2), (1.1) is untenable. Consequently, (1.3) a n d then (1.4) are also untenable. This representative deflationist a r g u m e n t is thus u n s o u n d . In this way, a l t h o u g h disquotationalist deflationism gives a g o o d analysis of the logico-syntactic f u n c t i o n of the t r u t h predicate in o u r language, the Tarski style equivalence thesis (T) is misidentified as implying the deflationary thesis about truth in the a f o r e m e n t i o n e d a r g u m e n t . T h e a r g u m e n t thus fails to provide a convincing reason for the central thesis of deflationism in terms of (T).

Copyrighted Material

188

The speech-act

equivalence

thesis

In sum, in this section I have explained why (T) and (M) are significantly distinct, a n d I have analyzed how, in one representative deflationist argument, the conflation of b o t h would result in some seriously mistaken philosophical conclusion regarding truth.

4 (E) and the Strawson-style equivalence thesis (I) As indicated in Section 1, (E) To say of a sentence that it is true is equivalent to using that sentence could be elaborated into saying, in a non-assertoric-context, that the two speech-acts (the one of saying of a sentence that it is true, while the other of simply using that sentence) have the same illocutionary force, based on their same locutionary meaning, on the occasion of a certain specific use which does not depend on whether or not the truth predicate is used to give a description (or make a statement), t h o u g h the two may have different illocutionary functions which d o depend on whether or not the t r u t h predicate is used to serve a certain illocutionary purpose. That is, in a non-assertoriccontext, (E) would be elaborated into the following illocutionary-force equivalence thesis: (I) For any utterance " q " [as specified before] that a person X understands, X's utterance "'q' is t r u e " has the same illocutionary force (for X) as X"s utterance "q". N o t e that, in X's utterance "q is true," " t r u e " is regarded not as a descriptive term but as a performative term, playing a role in quite different speech-acts like agreeing, endorsing, conceding, etc. It is arguably right that the illocutionary function or force of the t r u t h predicate " t r u e " that Strawson addresses a n d that is to be indicated in (I) does n o t constitute the raison d'être of the truth predicate and, m o r e generally speaking, of the linguistic expression of truth and (of the concept of truth) in the following sense. Those illocutionary purposes which Strawson claims are served by "true," generally speaking, do not have an intrinsic or necessary connection with the locutionary use of " t r u e " specified by the meaning-equivalence thesis (M), namely attributing the property of t r u t h to the truth-bearer in question in an assertoric context, which per se presupposes our pre-theoretic understanding of non-linguistic truth. T h o s e illocutionary purposes might be served, or are probably better served, by some other predicates or phrases, such as "I agree," "I accept," a n d "I concede," which would m o r e explicitly or directly give due locutionary contents based on their literal senses. There is n o evidence in our linguistic practice that, for the illocutionary purposes in question, people prefer using the truth predicate to using those other phrases. Consequently, there seems

Copyrighted Material

Bo Mou

189

to be n o convincing reason to say that those illocutionary uses of the t r u t h predicate " t r u e " captured by (I) constitute the raison d'être of the t r u t h predicate, much less the raison d'être of, generally speaking, the linguistic expression of t r u t h a n d (of the concept of truth). Indeed, the d u e philosophical concern with t r u t h is considered by m a n y as being essentially related to o u r reflective concern with some indispensable explanatory role played by the concept of t r u t h that, whether or not it is substantive, is somehow shown or delivered t h r o u g h the raison d'être of the linguistic expression of t r u t h a n d of the concept of truth. In this way, those w h o are really concerned with the non-epistemic semantic nature of the concept of truth a n d its indispensable substantive explanatory role would consider (T), rather t h a n (E) as (I), as genuinely relevant in their pursuits. A n d those deflationists w h o are concerned with the raison d'être of the t r u t h predicate (as indispensable but non-substantive means for some logical reason) would consider (E) as (M), also rather t h a n (E) as (I), as genuinely relevant in their pursuits. 2 2 In the preceding discussion, it has been presupposed that the genuine or intrinsic relevance to the raison d'être of the linguistic expression of t r u t h (and of the concept of truth) is a d u e criterion for where o u r reflective examination in this connection needs to focus. However, the questions remain: Why does one need to take this kind of "linguistic" a p p r o a c h t o the substantial issue of truth? To what extent is this strategy valid? W h a t is the relation between non-linguistic t r u t h , the linguistic expression of truth, a n d the linguistic truth predicate? These issues will be addressed in the next section.

5 The concept of truth and the speech-act equivalence thesis concerning "true" If my foregoing discussion is correct, a n u m b e r of due elaborations or conclusions can be d r a w n concerning the relation between the concept of t r u t h and the speech-act equivalence thesis (E). 1

2

Given that (T) provides one i m p o r t a n t a n d indispensable resource for o u r u n d e r s t a n d i n g non-linguistic t r u t h and its c o r r e s p o n d i n g concept, because (T) is neither a genuine variant of, nor based on, (E) but categorically distinguishes itself f r o m (E) and its genuine variants (M) a n d (I), any account of t r u t h or the concept of truth that is based exclusively on (E) a n d its linguistic analysis, either ignoring (T) or conflating (T) with (M), is d o o m e d to be incomplete. Given the n a t u r e of (M) and (I), the two genuine variants of (E) respectively in the epistemie a n d non-epistemic pragmatic contexts, (E) c a n n o t be taken as a p r i m a r y basis for o u r understanding of non-linguistic t r u t h a n d its concept. T h e reason is this: t o truly u n d e r s t a n d (E) one has to resort to o u r u n d e r s t a n d i n g of non-linguistic t r u t h a n d its concept,

Copyrighted Material

190

The speech-act

equivalence

thesis

rather than vice versa. As explained before, in most situations, when we say something by a sentence, we believe it and take it to be true, rather than merely expressing a t h o u g h t without acknowledging its truth. That is, with such an epistemic attitude involved, we assert the sentence. When explicitly giving a sentence by uttering its structurally descriptive n a m e x (such as its quotation name), we can make an assertion of the sentence simply by uttering the sentence with assertoric force, without using the truth predicate. As this familiar case shows, our pre-theoretic understanding of truth is internal to. or presupposed in, assertion. To make an assertion, even by simply uttering a sentence with assertoric force but without using the truth predicate, is to acknowledge the truth of the sentence. In this way, even if the truth predicate "be true" is not needed, and thus semantically r e d u n d a n t , in such a context, we still implicitly appeal to the substantive understanding of truth that is internal to our assertions. This understanding is substantive, because it is independent of any linguistic function of the truth predicate and because it goes beyond what deflationism tells us about truth. In this way, the substantive non-linguistic truth is internal to our assertions or judgments, although the truth predicate "be t r u e " is not. 3

A n o t h e r point related to the linguistic observation concerning (M) addressed in the preceding discussion is about some methodological strategy in treating the philosophical issue of truth. It is known that one d o m i n a n t philosophical m e t h o d in the linguistic-turn background of twentieth-century philosophy is the linguistico-philosophical m e t h o d , sometimes called "methodological nominalism." to the effect that all philosophical questions about the nature of things can be reduced either to empirical questions (thus to be turned over to science) or to questions a b o u t language, especially about the meaning of language. This linguistico-philosophical m e t h o d evidently bears on some methodological considerations in treating the philosophical issue of truth in c o n t e m p o r a r y philosophy. Some a u t h o r s have explicitly or implicitly assumed a methodological nominalism regarding truth that seems to consist of two related points: (1) the way the notion of truth is used and the way the truth predicate is used are considered as essentially the same; (2) consequently the n a t u r e or role of truth, if any, can be determined by the analysis of how the linguistic expressions of non-linguistic truth (if any) and its concept are used in our language. The foregoing examination of (E) via (M) has shown that, in the epistemic assertoric context, the disappearance of the truth predicate "be true" does not imply the disappearance of our pre-theoretic understanding of truth. The identityconditions of non-linguistic truth (property) and of its notion are actually independent of the linguistic truth predicate and its various linguistic functions (even its semantic function) in such a context. The way our pre-theoretic and reflective understanding of truth is exercised and the way the truth predicate "be t r u e " is used are thus not the same thing. 2 3

Copyrighted Material

Bo

Mou

191

To this extent, the truth predicate serves only as a volatile linguistic expression of the non-linguistic truth property and of its notion. One thus c a n n o t conflate the identity-condition of non-linguistic truth with the identity-condition of its volatile linguistic expression. So the claim (1) of the methodological nominalism regarding truth would be false. Nevertheless, (1) is only a sufficient, but not necessary, condition for (2). T h a t is, even if the way the notion of truth is used and the way the truth predicate is used are not essentially the same thing, it might still be right that the n a t u r e or role of truth, if any, can be s o m e h o w determined by the analysis of (the use of) the linguistic expressions of non-linguistic truth and its notion in o u r language practice. O n e might as well call (2) per se "semantic-ascent" a p p r o a c h to treating the philosophical issue of truth. T h e basic line of the "semantic-ascent" a p p r o a c h is this: instead of directly talking about non-linguistic truth (property) a n d its notion (including both its pre-theoretic understanding and its reflective concept), one is to a p p r o a c h the issue of truth concerning the nature and f u n c t i o n of truth (property) and its notion through examining how their linguistic expressions are used in o u r language practice. As an instrumental m e a n s in philosophical inquiry, this type of "semantic-ascent" a p p r o a c h per se can be both illuminative and effective, instead of a mere fashion. However, the question remains: How to effectively carry out this "semantic-ascent" approach without de facto r u n n i n g into the problematic strategy of (1)? 4

Can we implement the aforementioned "semantic-ascent" a p p r o a c h in treating the philosophical issue of truth by just focusing on the raison d'etre of the linguistic expression of truth and its notion in our language practice? Indeed, this focus can be reasonable. For, if the raison d'être of the linguistic expression of something in our language practice is really unavailable, one would have a strong reason to disregard the thing as something reflectively worthy or philosophically interesting and significant. Nevertheless, an application of this raison-d'être-focus strategy can become problematic if it focuses exclusively on the raison d'être of the linguistic truth predicate in an epistemie pragmatic context like that of (M), while ignoring the linguistic truth predicate and other significant linguistic expressions of non-linguistic truth and its notion in other significant kinds of linguistic contexts. This treatment reduces the raison d'être of the linguistic expression of non-linguistic t r u t h and (of its notion) in o u r language to the raison d'être of the linguistic truth predicate in a pragmatic context of language use. It limits the language practice, in which one can examine the raison d'être of the linguistic expression of non-linguistic truth (and of its notion), merely to the way of how to use the linguistic truth predicate in a pragmatic context. In the previous discussion (in Section 3) of one representative deflationist argument, one can see how this treatment does its work, via conflating (T) with (M), and how it can become problematic.

Copyrighted Material

192

The speech-act

equivalence

thesis

One can also find this kind of treatment in some earlier deflationists' a r g u m e n t for the r e d u n d a n c y thesis of substantive truth based on the semantic-redundancy thesis of the truth predicate. The semantic-redundancy thesis of the truth predicate claims that the truth predicate makes n o contribution to the semantic content of any context in which it appears as predicate-, so, (contemporary deflationists would add: if it weren't for certain logico-syntactic considerations,) the truth predicate can (could) be eliminated f r o m any context without any loss of semantic content. Then, according to the aforementioned methodological nominalism regarding truth, the redundancy thesis of substantive truth thus follows. O n e difficulty with this argument lies in the local or partial character of the argument for the semantic-redundancy thesis of the t r u t h predicate; for this local character compromises the complete coverage which the semantic-redundancy thesis of the truth predicate pretends to claim. The local character of the a r g u m e n t results f r o m the following facts. (1) T h e linguistic contexts under its consideration are actually limited only to some of the "first-order" contexts, i.e., the kind of epistemico-pragmatic contexts as highlighted by (M), in which one makes one's assertion either by assigning the truth predicate to an individual sentence or simply by uttering the sentence with assertoric force. (2) T h e linguistic contexts under its consideration include neither the " s e c o n d - o r d e r " or reflective ones, in which the truth predicate a n d its cognates are used to reflectively talk about the pre-theoretic understanding of truth in its reflective elaborations, nor such "first-order" contexts as the non-epistemic semantic ones which are given in the forms like (RLT1) a n d (RLT2) in natural language a n d reflectively presented by (T) in a m o r e f o r m a l language. True, a c o n t e m p o r a r y deflationist typically talks about (T); but, in such a talk, (T) is actually conflated with (M), as explained before. T h e same point can be put in a n o t h e r way: (T), which as a whole is to give a contextual characterization or definition of o u r (pre-theoretic and reflective) u n d e r s t a n d i n g of non-linguistic truth in view of an original or primary p u r p o s e of one constructing and using (T), is not supposed to be used as a resource for the equivalence replacement between "' . . . ' i s t r u e " (its lefth a n d side) a n d " . . . " (its right-hand side) in a pragmatic context; such a replacement function in our language practice is assumed by (M) instead of (T). (3) Given that there are similar pre-theoretic and reflective understandings of non-linguistic truth b o t h in the Western tradition and in a nonWestern tradition like the Chinese one, and given that the c o m m o n pre-theoretic understanding of non-linguistic truth and its reflective understanding have been linguistically expressed in some distinctive ways respectively in a phonetic language like English and in an ideographic language like (classical) Chinese (both in the "first-order" folk context and in the "second-order" reflective context), a linguistic observation limited to the former case is clearly not a complete one. To my knowledge and my j u d g m e n t , a comparative examination concerning how our pre-theoretic and reflective understandings of non-linguistic truth have been presented in distinctive philosophical traditions

Copyrighted Material

Bo Mou

193

via their distinctive linguistic m e a n s w o u l d be p h i l o s o p h i c a l l y illuminative a n d w o u l d b e a r o n a general p h i l o s o p h i c a l c o n c l u s i o n r e g a r d i n g t r u t h . A p a r t i a l d i s c u s s i o n of this issue is given in M o u (2006).

6 Summary In the preceding discussion I have tried to a r g u e for several points. First, the speech-act equivalence thesis r e g a r d i n g " t r u e " in o u r language, (E), which is c o n s i d e r e d by m a n y as providing a s t r o n g intuitive basis for o u r u n d e r s t a n d i n g of t r u t h is b o t h a m b i g u o u s a n d vague: (E) is s o m e h o w associated with s o m e distinct equivalences u n d e r l a i n by v a r i o u s distinct pre-theoretic u n d e r s t a n d i n g s c o n c e r n i n g t r u t h a n d / o r " t r u e " ; o n e c a n n o t establish one's philosophical c o n clusion c o n c e r n i n g t r u t h a n d / o r " t r u e " based o n (E) w i t h o u t first clarifying (E) a n d figuring o u t exactly which equivalence thesis that is talked a b o u t . Second, the Tarski-style n o n - e p i s t e m i c extensional equivalence (T), which might be c o n s i d e r e d as a n o n - e p i s t e m i c e l a b o r a t i o n of (E) in the sense explained, serves as o n e intuitive basis for the s e m a n t i c c o n c e p t of non-linguistic t r u t h to c a p t u r e o u r pre-theoretic " c o r r e s p o n d e n c e " u n d e r s t a n d i n g of t r u t h . T h i r d , the FregeRamsey-style equivalence, (M), as a n epistemically oriented p r a g m a t i c elab o r a t i o n of (E) serves as the intuitive basis for m e a n i n g equivalence in the assertoric context. F o u r t h , the Strawson-style l o c u t i o n a r y - m e a n i n g equivalence, (I), as a non-epistemically o r i e n t e d p r a g m a t i c e l a b o r a t i o n of (E) serves as the intuitive basis for v a r i o u s illocutionary p u r p o s e s in the n o n - a s s e r t o r i c context. Finally, b a s e d o n the f o r e g o i n g discussion, a n u m b e r of conclusions a n d d u e e l a b o r a t i o n s have been d r a w n c o n c e r n i n g the relation between the c o n c e p t of t r u t h a n d the speech-act equivalence thesis, (E), a n d c o n c e r n i n g s o m e m e t h o d o l o g i c a l strategies in treating the philosophical issue of t r u t h .

Notes 1 This essay is a substantial revision and extension of some of the ideas in two of the author's previous publications; see Mou (2000a) and Mou (2000b). I am grateful to Richard Feldman, Rolf Eberle, and Theodore Sides for their helpful comments and criticism of some of the basic ideas delivered in this essay and their related writings. I would like to thank Dirk Greimann and Geo Siegwart for their helpful comments on this essay. 2 Throughout this paper, in the Fregean sense, "assert" or "assertion" means claiming the truth of a sentence or making a claim as to how things actually are. See Frege (1918). 3 Compare David (1994): 67 9. He says that (E) as the speech-act principle is to describe the linguistic role [that the term "true" plays in our language]. This role . . . is described by the following rule: (UT) Whenever one utters (or otherwise produces) a sentence x that consists of the grammatical predicate "is true" attached to a (quoted) sentence y, one might as well utter (or otherwise produce) just the sentence y; and vice versa. (David 1994: 69)

Copyrighted Material

194

The speech-act equivalence

thesis

4 For example, David says that the speech-act equivalence thesis (E) is ''the speechact version of (T)" (David 1994: 67). 5 One can understand the ontological status of, say. snow's being white in various ways: in a solid realist way or in a radical anti-realist way or in some other way. That is, the truth condition specified by the right-hand side of (RLT) is ontologically neutral. For this reason, if by " t r u t h " we mean what the real-life truthcondition equivalence characterizes, the notion of truth does not exactly amount to an ontological notion; a theory of truth does not amount to an ontological theory. The so-called "realist pre-theoretic (correspondence) understanding of t r u t h " is actually a combination of our pre-theoretic (correspondence) understanding of truth and a realist ontological understanding or explanation of what counts as, say, snow's being white. 6 One might object this way: an anti-realist would also have no problem with accepting our real-life T-sentences and thus our pre-theoretic understanding of truth; but it seems to her that the fact of. say. snow's being white is not independent of the epistemic agent's epistemic attitude; so truth is ascribed to "Snow is white" by the epistemic agent or by a beliefs-system. Note that what we are talking about is our pre-theoretic understanding of truth rather than a certain understanding of the ontological status. As explained in the preceding note, no matter how one interprets the ontological status of. say, snow's being white, the relation between snow's being white and "Snow is white" in the pre-theoretic understanding of truth is a kind of making-true relation; such a relation per se does not involve the agent's epistemic attitude and so is non-epistemic in this sense and to this extent. 7 Tarski is impressed with the following result in a statistical questionnaire: in a group of people who were questioned only 15 per cent agreed that "true" means for them "agreeing with reality," while 90 per cent agreed that a sentence such as "it is snowing" is true if, and only if, it is snowing. Tarski concludes: "Thus, a great majority of these people seemed to reject the classical conception of truth in its 'philosophical' formulation, while accepting the same conception when formulated in plain words" (Tarsky 1944: 61). 8 Cf.. Tarski (1944: 49 -50. 61). The so-called Aristotle's formula is the formula in his Metaphysics (Book IV. Ch. 7. 1011b27): (A) To say of what is that it is not. or of what is not that it is. is false, while to say of what is that it is. or of what is not that it is not, is true. 9 Throughout this work, when the predicate "is true" is used with the lower case, "t", of the 20th letter in the English alphabet, it is taken either as our real-life truth predicate or as a Tarskian predicate whose definition meets Convention T. The same holds for its noun counterpart "truth." Note that, in the case of the schema (T). it is at issue whether or not what is defined captures the point of the real-life truth predicate before one can tell if the definition meets Convention T. To indicate this suspensive case, I mention the predicate "is True" with the capital case, "T", of the 20th letter in the English alphabet or use such sayings as "Truth predicate" or "truth-definition candidate" to refer to the predicate defined. 10 It is known that Tarski actually provides two alternative approaches to the definition of truth. One of them is an implicit list-like-definition (for a finite language) which consists of all those relevant T-sentences. The other is an explicit recursivedefinition via satisfaction (for an infinite language) to which, for the sake of some formal considerations. Tarski is forced to resort. For Tarski. there are technical problems with the notion of infinite conjunction of infinitely many sentences that

Copyrighted Material

Bo Mou

11 12

13 14

15

16

17 18

195

is needed for a list-like definition for an infinite language; such a definition cannot be explicitly formulated in the metalanguage because of the finite length of individual sentences in the metalanguage. Moreover, there is the problem regarding quantified sentences whose components are open sentences and quantifiers lacking truth value. See Tarski (1944: 50 I; 1969: 107). Tarski's solution is his recursive-definition via satisfaction. Nevertheless, for the purpose of this essay, I only consider his list-like-definition. For Tarski's strict characterization of logical consequences see Tarski (1935: 346). This claim amounts to saying neither that Tarski's (T) and Convention (T), as they are, successfully capture all the essential features of our pre-theoretic understanding of truth nor that Tarski's strategy is the exclusively correct approach. For example, Tarski's (T) and the Tarskian formal definition of truth is considered as merely having enumerative character and thus failing to capture the general character of our pre-theoretic understanding of truth. See Davidson (1990: 285-8); Field (1972: 356); Soames (1984: 419 -20). This is true to some extent. For a detailed discussion of this issue, see Mou (2001). Gila Sher argues that "[Tarski's] equivalence shema . . . only . . . provides a partial account of truth [in regard to logical structure or what is called 'logical factor']." (Sher 2004: 1415). I agree with this judgment, though it is worth further discussing where the partiality of (T) and Tarski's account is located. See Frege (1891: 34): Frege (1892: 64); Frege (1918: 7-8); Quine (1990: 81). See Quine (1970: 11-12; 1990: 80). Two points are to be noted here. First, it might be argued that, when one uses the truth predicate in this case, one might adopt a non-assertoric epistemic attitude rather than claim the truth of the sentence: one might simply accept or concede it. As 1 will discuss in Section 4 below those non-assertoric or non-truth-claiming epistemic attitudes cannot explain the raison d'être of the truth predicate; (M) is supposed to be a meaning-equivalence thesis in an assertion-context. Second, disquotationalist deflationism is inspired by Quine's disquotational account of truth. Nevertheless, Quines disquotational account is significantly different from the disquotionalist version of deflationism. Disquotationalist deflationism takes the "disquotation" function of "true" as a purely syntactic one while, in Quine's account, "disquotation" function of "true" primarily means a non-syntactic semantic function. (This point has been emphasized by Quine himself in this author's conversation with him at Harvard on 10 June 1997.) That is why I call the deflationary view in question here "disquotationalist" rather than "disquotational": I intend to distinguish it from Quine's original account. Again, the so-called "both sides" of (M) are relative to "means the same as." Note that, although "p" here is instantiated by the quotation name of a sentence, both sides of "means the same as" are not quotation names but "X's claim that p is true" and "X's assertion p". See Gupta (1993: 59-67). Gupta gives a good analysis of why it is problematic for deflationists to resort to (T) in their elaboration of the logical role of the truth predicate, pointing out the fact that both sides of (T) are not synonymous. See Ayer (1936: 89), Strawson (1964: 46) and Horwich (1998: 4-5). The schema (T) is sometimes called "thesis" below in a non-strict way. If one maintains that a thesis should be a complete statement at the meta-level rather than a schema at the meta-meta-level, then the Tarski style equivalence thesis (T) could be given as the following claim at the meta-level by means of substitutional quantification: (T/) For any sentence p, x is true in L iff p. where the universal generalization may be instantiated by any sentence of L whose name replaces "x".

Copyrighted Material

196

The speech-act equivalence thesis

19 Although, as indicated above, the deflationary analysis from (1.1a) through ( l . l d ) is not to argue that the quantification (in English) is most naturally construed as first-order and objectua! but takes it for granted, one can figure out the reasons why it is taken for granted. For one thing, the bound propositional variable x is grammatically like a singular term, but a single x without a predicate is not a sentence in English, and so it cannot be allowed to occupy sentence position in English. That suggests that a natural language like English is not a second-order language but a first-order language (otherwise, it would allow first-order sentence letters or proposition letters as variables which could stand alone without a predicate). For another thing, because, in usual cases, we use our natural language to talk about objects rather than mere linguistic expressions, and because the objectual interpretation of quantification appeals to the objects over which the variables range while the substitutional interpretation resorts to the expressions that can be substituted for the variables, the usual interpretation of the quantification in natural language is considered to be objectual. 20 This line of deflationary argument is typically presented by those contemporary deflationists like Horwich. See. for example. Horwich (1995: 813; 1998: 5-6. 1012, 36-7, 50 1). 21 This seems to be true of Quine partially because his extensionalist position regarding meaning makes him deny a sharp distinction between meaning equivalence and extensional equivalence. 22 For some representative deflationists' concern in this regard, see Horwich (1995: 813; 1998: 2, 5). For those deflationists" attitude towards (1). one can look at Horwich's following explanation: from the present perspective we are rejecting the idea due to Strawson . . . that the truth predicate is not used to give descriptions or make statements about the things to which it is applied, but that it is used instead to perform quite different speech-acts: endorsing, agreeing, conceding, etc. . . . N o doubt we d o perform all kinds of speech-act (such as agreeing and conceiting) with the truth predicate. But, . . . it is best to say that we do so by (not instead o f ) making a statement that is. by attributing the property, truth, to the proposition in question. (Horwich 1998: 38-40) 23 This point can be made not merely through examining such "first-order" contexts as (M) but also in view of how truth and its notion is linguistically delivered in the reflective "second-order" contexts. I have explained and illustrated this by examining two distinct versions, the semantic-ascent version and the paraphraseexplanatory-reduction version, of the thesis of truth as strategic normative goal in Mou (2006: 323-8. 332- 7).

References Ayer, A. J. (1936) Language, Truth and Logic. New York: Dover Publication. Blackburn. S. and Simmons, K. (eds) (1999) Truth. Oxford: Oxford University Press. David. M. (1994) Correspondence and Disquotation: An Essay on the Nature of Truth, Oxford: Oxford University Press. Davidson. D. (1990) "The Structure and Content of Truth," Journal of Philosophy. 87: 279-328. Field, H. (1972) "Tarski's Theory of Truth." The Journal of Philosophy. 69: 347-75.

Copyrighted Material

Bo Mou

197

Frege, G. (1891) "Function and Concept," in P. Geach and M. Black (eds) (1980) Translations from the Philosophical Writings of Gottlob Frege, 3rd edn, Oxford: Basil Blackwell. pp. 21-41. (1892) "On Sense and Meaning," in P. Geach and M. Black (eds) (1980) Translations from the Philosophical Writings of Gottlob Frege, 3rd edn, Oxford: Basil Blackwell, pp. 56-78. (1918) "Thought." in P. Geach (ed.) (1977) Logical Investigations, Oxford: Basil Blackwell. pp. 1 -30. G u p t a , A. (1993) "A Critique of Deflationism," Philosophical Topics, 21: 57 81. Horwich, P. (1995) "Truth," The Cambridge Dictionary of Philosophy, Cambridge: Cambridge University Press, pp. 812-13. (1998) Truth, 2nd edn, Oxford: Clarendon Press. Kirkham. R. (1992) Theories of Truth, Cambridge, MA: M I T Press. Mou, Bo (2000a) "Tarski, Quine, and 'Disquotation' Schema (T)," The Southern Journal of Philosophy, 38: 119-44. (2000b) "A Metaphilosophical Analysis of the Core Idea of Deflationism," Metaphilosophy, 31: 262 86. (2001) "The Enumerative Character of Tarski's Definition of Truth and its General Character in a Tarskian System," Synthese, 126: 91-122. (2006) "Truth Pursuit and Duo Pursuit: From Davidson's Approach to Classical Daoist Approach in View of the Thesis of Truth as Strategic Normative Goal," in Bo M o u (ed.) Davidsons Philosophy and Chinese Philosophy: Constructive Engagement, Leiden and Boston: Brill, pp. 309 -349. Quine, W. V. (1970) Philosophy of Logic, Cambridge, MA: Harvard University Press. —(1990) Pursuit of Truth, Cambridge, MA: Harvard University Press. Ramsey. F. P. (1927) "Fact and Proposition," in G. Pitcher (ed.) (1964). Truth, New Jersey: Prentice-Hill, pp. 16 -17. Sher. G. (2004) "In Search of a Substantive Theory of Truth," The Journal of Philosophy, 101: 5-36. Soames, S. (1984) "What is a Theory of Truth?," The Journal of Philosophy, 81:411 - 29. (1998) Understanding Truth, Oxford: Oxford University Press. Strawson, P. F. (1950) "Truth," in G. Pitcher (ed.) (1964). Truth, New Jersey: Prentice-Hill. pp. 32-53. (1964) "A Problem about Truth: A Reply to Warnock," in G. Pitcher (ed.) (1964) Truth, New Jersey: Prentice-Hill, pp. 68-84. Tarski, A. (1933) "The Concept of Truth in Formalized Languages," in A. Tarski (1983) Logic, Semantics, Metamathematics, 2nd edn, translated by J. H. Woodger. Indianapolis: Hackctt, pp. 152-278. — (1935) "Foundations of the Calculus of System," in A. Tarski (1983) Logic, Semantics, Metamathematics, 2nd edn, translated by J. H. Woodger, Indianapolis: Hackett, pp. 342-83. (1944) "The Semantic Conception of Truth and the Foundations of Semantics." in A. P. Martinich (ed.) (1990) The Philosophy of Language, 2nd edn, Oxford: Oxford University Press, pp. 48-71. (1969) "Truth and Proof," in Scientific American, 220: 63-77: reprinted in R. I. G. Hughes (ed.) (1993) A Philosophical Companion to First-Order Logic, Indianapolis: Hackett, pp. 101 -25.

Copyrighted Material

Part III

The normativity of truth

Copyrighted

Material

10 Inflating truth Wright's argument from normativity to propertyhood Adam Kovach

The intuition that truth is normative for thinkers and language-users appeals to our common sense. 1 Witness the interchangeability in many contexts of "true" with "correct" and "right. " Philosophers who take this intuition seriously are liable to object to any theory of truth, which fails to do it justice. This sort of objection has a respectable history. Failure to account for the normativity of truth is one of the criticisms James raised against the "copy theory. " If truth consisted merely in an idea's copying a reality, James asked, why would we be so sure that we ought to believe the truth? (James 1978: 112) The same intuition troubled Russell at about the same time, when he held the view that truth is an indefinable, intrinsic property of true propositions. He expressed his reservations to his own view with the remark that it seems to make "our preference for truth over falsity a merely aesthetic judgment" (Russell 1973: 75). More recently, Michael Dummett has criticized Frege's theory of truth and falsity as the references of sentences for its failure to account for the fact that "it is part of the concept of truth that we aim at making true statements" (Dummett 1978: 2). Still more recently. Crispin Wright has polished up the old projectile and aimed it at a new target, deflationary theories of truth (Wright 1992, 1996, 1999, 2001). Wright's work inspires the question addressed in this essay, which is whether or not there is a good inflationary argument (an argument against deflationism) from the premise that truth is normative. 2 At a first glance there is something surprising about the suggestion that the normativity of truth spells trouble for deflationism. The deflationists' main claim is that truth is not a property, at least not a property of an interesting sort, as opposed to being a mere shadow of a predicate. James and Russell thought that if truth is normative, truth cannot be a property of this or that particular kind. Wright's argument would appear to be that if truth is normative, then deflationism is wrong and truth is a property. If any concept is normative, the concept of good is. Compare: If goodness is normative, then goodness is a property. Wouldn't it be nice to settle the debate about moral realism so swiftly! Isn't the argument usually the other way 'round? Value irrealists are suspicious of normative concepts because they fear that the only properties such concepts could express would be "queer properties. "

Copyrighted

Material

202

Adam Kovach

This first appearance may turn out to be largely correct, but as so often, things are more complicated than they appear to be at first glance. The issues surrounding the question have to be clarified. That is the task of Sections 1-4, which discuss what is at issue between deflationists and inflationists, and the status of the claim that truth is normative. In Sections 5-8, I discuss the two-stage strategy of Wright-style inflationary arguments, criticize Wrights two most recent presentations of his argument, and develop a variation on Wright's theme. I identify points at which the strategy runs into difficulties, and ways in which deflationists can resist the arguments, but I also find the basic intuitions that go into this argument strategy attractive. So. I try to shore them up. The final section is about why an inflationist might hope to emerge someday from the thickets of this debate with fewer scratches than the opposition. 3

1 Standard deflationism There are two sides to standard deflationary theories of truth: a view about our competence with truth, and a view about truth as a property,4 Views about the meaning of the truth predicate and the main purposes for which we use it fall under the heading of competence. Standard deflationary accounts of truth begin by noting the role of the truth predicate in certain equivalences which competent speakers recognize as trivially true. For a theory about the truth of propositions, the relevant family is some collection of "central" instances of the equivalence schema (E). (E) The proposition (statement, belief etc. ) that p is true iff p. The deflationary claim is that central instances of (E) (the propositional tional of the truth predicate. The strategic words "in some way" are here to indicate that there are various different proposals for how to develop a definition from a schema like (E). The most obvious way is to take the totality of the propositional T-equivalences as an account of our concept of truth. Of course, no one can understand them all. and that is a truly massive theory of truth. The question arises whether there is a satisfactory way to make a finite definition out of (E), for example, by substitutionally quantifying over the open variables in the schema. The question of which definitional strategy works best merits a separate essay. Our purpose is expedited by pretending that the totality of instances of (E) is the standard deflationist's account of the concept of truth. It can be taken as a stand-in for some yet to be determined shorter-winded formulation, and there will be no objections raised along the lines of Sellars's old quip that some theories of truth are like the telephone directory (Sellars 1962: 33). In addition to explaining the meaning of "true, " the standard deflationary account of our competence also explains the reason why we have a predicative

Copyrighted

Material

T-equivalences,

or

From normativity

to propertyhood

203

concept that can be defined in this way. The standard explanation is that the truth predicate is a useful device for carrying out certain logical or expressive functions. In particular, it is useful for formulating certain generalizations, and also for expressing claims indirectly. "Everything you say is true" and "that last thing you said was true" are major time-savers. 5 The other side of standard deflationism is the claim that the truth predicate does not express any substantial property. This claim is not merely about our competence; it is about what there is, or rather, what there is not. The term "substantial" is in contrast to the weak "shadow of a predicate" sense of the word "property," according to which every predicate expresses a property. All it takes for there to be a property of truth, in the weak sense, is for the expression "true" to belong to the logical category of a predicate. Deflationists need not deny that truth is a property like that. In denying that truth is a substantial property, what deflationists mean is that truth does not have an underlying nature that can be described with an interesting degree of generality; to the extent that truth can be said to have a nature at all, it is radically disjunctive. To see what it means for truth to have a radically disjunctive nature, note how the T-equivalences assign truth conditions to propositions individually. Each T-equivalence provides a particular, unique account of what the truth of a distinct proposition would consist in. N o collection of these particular accounts gives any indication that there is a significant common feature of true propositions. Yet their totality adds up to one radically disjunctive account. An explicit partial definition makes this disjunctive nature apparent: x is true = df [x = the proposition that snow is white, and snow is white] or [x = the proposition that snow is black, and snow is black] o r . . . 6 For truth to have an underlying nature, which can be the subject of a general theory, there would have to be some relevant features that true propositions have in common to some degree and in virtue of which they count as true. The deflationist denies that there are any such features. If there are no such features, then neither can there be any interesting theory of truth as a property. So, standard deflationism can be summed up as two basic claims: (The competence thesis) The meaning and purpose of the concept of truth are completely determined by the equivalence schema. (The no-property

thesis) Truth is not a substantial property.

These two claims are logically independent. 7 It is possible to hold that truth is not a substantial property, and yet that the concept must be defined in some other way, or even that it is indefinable. Less obviously it is possible to hold that the equivalence schema explains our competence with truth, but that truth is a substantial property. The most that can be objected to the latter

Copyrighted

Material

204

Adam Kovach

view is that if truth has an underlying nature worth investigating, the concept should be defined so as to indicate what this nature is, and not to invite the suspicion that its nature is thoroughly disjunctive. However, a theory of a concept is one thing, and a theory of the property it expresses is another. Part of the difference is that a concept captures how we think about the world, even how we misconceive the world, as some concepts are spurious. Because the two basic claims are independent, critics of standard deflationism may reject one without the other, but just as importantly, a reason to reject one of these claims may fail to impact the other. Another point is that, of the two claims, the no property thesis is the more important. As we shall see, there are other ways of thinking about our competence with truth, but none count as deflationary unless they deny truth's propertyhood.

2 Varieties of inflationism Inflationism is the view that truth has an underlying nature, which can be investigated. It is an ontological thesis, and is compatible with different ideas about our competence with truth. Thus, the core disagreement between inflationists and deflationists is about the status of truth as a property. Inflationists reject the no-property thesis. However, not all inflationists are committed to investigating the nature of a single uniform truth-property, the same feature common to all true propositions, assertions or beliefs. This point is sometimes overlooked, but there are varieties of inflationism. Paul Horvvich, for example, characterizes the deflation-inflation controversy as a contest between standard deflationism and the project of devising a reductive theory, or analysis of truth, the ideal result of which is a definition of the form: a proposition (statement, belief, etc. ) is true iff it is F where F is filled in with a description that picks out a uniform property of true propositions. Horwich considers this project hopeless, and introduces deflationism as a reaction to a tradition for which the point of a theory of truth is to produce an analysis of just this form: Among the products of this traditional view there is the correspondence theory (x is true iff x corresponds to a fact), the coherence theory (x is true iff x is a member of a set of coherent beliefs), the verificationist theory (x is true iff x is provable, or verifiable in ideal conditions), and the pragmatist theory (x is true iff x is useful to believe). Deflationists reject all such accounts since "there is no prospect for an explicit definition, or reductive analysis, even a very approximate one" (Horwich 1998b: 239). It is easy to see why a deflationist must reject the attempt to analyze truth in this way. That project would undermine both of

Copyrighted

Material

From normativity

to propertyhood

205

the standard deflationist claims. By identifying a common feature of truths in virtue of which they count as true, an analysis conflicts with the noproperty thesis. In claiming that an account of this common feature is an analysis of truth, one denies that the T-equivalences determine the meaning and purpose of the concept of truth. If deflationism and reductive analysis were the only two options for a theory of truth, then doubts about the feasibility of one would count in favor of the other. Framing the issue in terms of this stark choice, however, sets targets too close for deflationism and too far for its critics. An inflationist does not have to expect an analysis in order to think that truth has an underlying nature that can be investigated, and deflationism requires more than the claim that no analysis is possible. To see why, it helps to compare truth with some other alleged properties, as we do in comparing questions such as the following: Q1 Q2 Q3 Q4

What What What What

makes makes makes makes

a a a a

proposition (statement, belief, etc. ) true? m a p accurate? plan of action effective? purposive device well-designed?

For example, how does being an accurate map stand with respect to propertyhood? Here are some possible scenarios, which fall short of there being a uniform property of map-accuracy. 1 There is one high-level property of being an accurate map, but it is realizable in different ways. Different maps can have this property in virtue of very different features. 8 2 There is no single property of being an accurate map. "Is an accurate map" is used to express a finite number of different properties, each of which must be explained separately. Satisfying the disjunction of these properties is a necessary and sufficient condition of being an accurate map. 3 While there is no finite set of properties (even disjunctive properties) possession of which is necessary and sufficient for being an accurate map, there is a significant pattern of resemblance among accurate maps in virtue of which they all count as accurate. Every accurate map is accurate in virtue of some properties. Some of the properties that make a map accurate are the same as or similar to some of the properties that make other maps accurate. In each of these cases, map-accuracy is a fragmented feature, but not so fragmented as to lack a complicated underlying nature, and not so fragmented as to be radically (infinitely) disjunctive. These proposals about ontology are compatible with a range of views about how to account for the concept of map-accuracy. In particular, the first is compatible with an analysis, and the last is compatible with the view that map-accuracy is a family

Copyrighted

Material

206

Adam Kovach

resemblance concept or that our understanding of map-accuracy depends on our grasp of certain exemplars or prototypes and a similarity metric. Supposing that any one of these cases were a plausible way to think about the example, map-accuracy would have a nature that might be described at some level of generality. Someone after a theory of map-accuracy might seek this generality in descriptions of map-making conventions, or of how various features of accurate maps correlate with aspects of terrain. The result would not likely look like an analysis. It wouldn't take the form "a map is accurate iff it is F. " However, it wouldn't look like a radically disjunctive account either. It wouldn't account for accuracy one map at a time, so to speak. The result might be a modest inflationism about map-accuracy. To complete the comparison with truth, there are modest inflationary views, which expect that there will be some degree of generality from a theory of truth, which may fall far short of a uniform analysis.

3 Inflationary arguments Having clarified how the debate between deflationists and inflationists is at base about the status of truth as a property, the next issue to consider is how inflationists make their case. Given the state of the art - some would say the history of failures at analyzing truth - that case cannot rest on the success of any of the usual suspects among the traditional approaches. The analogy between truths and accurate maps is a place to start. There are at least two reasons for believing that map-accuracy is a substantial property. One is the prima facie plausibility of there being a pattern of common features among accurate maps. Cartography requires that certain structural characteristics coincide between map and represented terrain. It deals in spatial isomorphism. In contrast, since the decline of logical atomism and picture theories of meaning, few have argued that true propositions, assertions or beliefs are related to whatever they are about (e. g. states of affairs) in any remotely similar way. Another reason to think that accurate maps share underlying natures is that the concept of being an accurate map enters into explanations of how we use maps to successfully guide ourselves where we want to go, and why on the whole people prefer and strive to achieve accuracy in map-making. For example, it is a good (partial) explanation of why a hiker found the way back to the campsite that (a) the hiker had an accurate map and (b) people with accurate maps tend to find their way. But premise (b) is a generalization about accurate maps, and it couldn't be of use in explaining the hiker's behavior unless map-accuracy is a substantial property. So the argument goes. Indeed the major source of resistance to deflationism is the belief that the concept of truth does have significant roles to play in philosophical explanations. For example, a recently debated case is the role of truth in the explanation of practical achievements. It has been urged that truth must be

Copyrighted

Material

From normativity

to propertyhood

207

a substantial property in order for the concept to enter into successful explanations using general explanatory principles such as these: 1 True practical beliefs tend to help us achieve our goals. 2 Truthful communication tends to promote successful coordination of group activities. 3 Approximately true scientific theories tend to yield successful predictions. A similar debate concerns the role of truth in theories of meaning, where it is argued that if there is any truth to the saying that the meaning of a sentence is determined by its truth condition, then truth must be a substantial property. Deflationists can respond to this sort of challenge in two ways. One is to argue that, contrary to appearances, truth plays no role in these claims beyond its familiar role as a device for forming generalizations. If that fails, it is open to deflationists to deny that explanations involving these principles are any good. For this reason, the debate over deflationary theories of truth is bound up with the debate over conceptual role theories of meaning as alternatives to truth-conditional theories of meaning. 9 The question of this essay is whether or not there is a good inflationary argument from the premise that truth is normative. At least at the outset, that is a distinct question from whether or not there is a good inflationary argument from the premise that truth has some particular role in philosophical explanations. The initial attraction (and genius) of Wright's inflationary argument strategy is that it appears to by-pass this sort of debate over the explanatory role of truth, however, mind the warning about first appearances. 4 The normativity of truth The intuition that truth is normative is an intuition about our competence. It is grounded in the recognition of a family of principles about how we ought to behave in making assertions, answering questions, and forming and revising beliefs. For example, this is a principle about assertion: one ought, other things being equal, to assert a proposition only if it is true. According to similar principles, one ought, other things being equal, to accept the truth, to revise one's beliefs so as to accommodate them to the truth, and, to seek the truth in inquiry. Such normative principles involving truth implicitly guide people's thought and language use, and, upon reflection, most people acknowledge them as obvious. Two remarks about these principles are in order. First, they have been characterized rather roughly. It is an interesting question just how precisely they can be characterized, but it is a good conjecture that any prima facie plausible normative principles about truth will be variations on the three basic themes of truth as a norm of assertion, inquiry and belief formation/revision. Second, such principles must be taken with a stiff dose of ceteris paribus clauses. Our concern with truth is one among many. The need to be truthful and the truth-aim compete

Copyrighted

Material

208

Adam Kovach

with other needs and other aims, and this must be reflected in the relevant norms. The normative principles involving truth are globally prevalent but locally easily trumped. This is to say that truthfulness is a general concern across cognitive and communicative contexts, so much so that it is hard to imagine any aim of thought and language use that could systematically compete with truth. Yet in particular cases it is easy to imagine a goal that is immediately more important, or to find reasons for breaching the principles. In a community that follows such normative principles, "truth" and "falsity" can function as terms of approval and disapproval, and be used for guiding behavior. So. it is not surprising that "true" is interchangeable with "correct" and "right" in many contexts. However, it is one thing to notice that truth has these uses, and another to hold that truth is a normative concept in the strong sense that the normative character of truth is part of the meaning of the truth predicate. One may admit the facts adduced so far as pragmatic facts about how we use the concept, but hesitate to take the extra step of concluding that some normative principles involving truth are analytic, or in any way constitutive of or derivative from the meaning of "truth. " It follows that there is a distinction to be drawn between a weak, merely pragmatic, thesis about the normativity of truth, and a strong, semantic thesis. (Weak normativity

thesis) The concept of truth enters into certain norms.

(Strong normativity thesis) The concept of truth is a normative concept, in that some claims involving "true" entail claims involving explicitly normative vocabulary. For example, one who accepts the weak thesis might recognize the prevalence of the principle that other things being equal, one shouldn't deny the proposition that p if it is true that p. while an advocate of the strong thesis might add that (1) entails (2). (1) The proposition that p is true. (2) One ought not to assert that not-p.

5 Wright's inflationary argument If the concept of truth delivers entailments of this sort, then there is a straightforward objection to standard deflationism. The standard deflationary competence thesis evidently conflicts with the strong normativity thesis. For how can truth be normative in the strong sense if the meaning and purpose of the concept of truth are completely determined by the equivalence schema (E)? The T-equivalences imply nothing about the normativity of truth. The point may be illustrated by considering a language fragment that

Copyrighted

Material

From normativity

to propertyhood

209

is free of all normative vocabulary. All the resources are there to introduce a mini-truth predicate for this fragment using nothing more than the Tschema, and yet by hypothesis, no normative claims can be entailed by claims involving that truth predicate. The objection here is that something is left out of the standard deflationary account of the meaning of "truth, " and not that there is something wrong with what goes into that account. There are two ways for deflationists to accommodate the intuition that truth is normative. Standard deflationists must deny the strong normativity thesis, and there is room for that maneuver. They can maintain that a weak normativity thesis is all that is needed to account for the normative character of truth (cf. Horwich 1996, 1998a). The other option is to improve the deflationary account of competence by cross-fertilizing standard deflationism with a strong normativity thesis. The way to do this with the least change is to simply conjoin the two. The resulting normative-deflationary hybrid theory consists of all the instances of schema (E) and some relevant normative principles involving truth, which jointly give the meaning of "truth. " On either approach, deflationists can be secure in the knowledge that the objection, whatever its merits as a criticism of the competence thesis, appears to have no bearing on the no-property thesis, which is, after all the main point of contention. Or is this false security? Crispin Wright's interesting contribution to the debate over deflationism has been to raise the possibility that the normativity of truth spells greater trouble for deflationism than it would seem. Over the past decade, Wright has plied his trademark inflationary argument through a number of his writings (Wright 1992. 1996, 1999, 2001). Putting details of presentation aside, Wright's strategy is two-staged. At the first stage a strong normativity thesis about the concept of truth is established. At the second stage it is argued that truth's being a normative concept is inconsistent with the deflationist's no property thesis. Because the spirit of Wright's argument moves me farther than its letter, I will discuss each of the two stages separately. At each stage I will start by reviewing and criticizing Wright's version of the argument, and then do what I can to make improvements. The crucial and problematic inference comes at the second stage, but Wright's arguments at the first stage are also interesting. The question of the next section is whether the strong normativity thesis can be made to stick. Sections 7 and 8 try to wring a conclusion out of the crucial inference from normativity to propertyhood.

6 Arguing for the strong normativity thesis Wright gives two different arguments that truth is a normative concept. One attempts to show that the deflationary account of competence itself along with some uncontroversial assumptions about the nature of assertion and epistemic justification leads to the conclusion that truth is a normative concept. The other argument starts with the observation that truth plays a normative

Copyrighted

Material

210

Adam

Kovach

role in the assessment of assertions, namely that it functions as a device for their endorsement. After discussing these arguments, I will suggest two other ways to make the case for a strong normativity thesis. 6. 1 Wright's first argument Wright attempts to show that standard deflationism must admit that truth is a normative concept: We begin with the observation that truth-apt contents, or sentences expressing such contents, demand a distinction between circumstances under which asserting them is warranted and those under which it is not. And competent thought and talk requires an ability to tell the difference: 1 need to be able to tell which assertions 1 am warranted in making in a given state of information and which 1 am not. So if I am warranted in asserting P. that fact will be recognizable to me, and I will thereby be warranted in claiming that I am so warranted. Conversely, if 1 am warranted in thinking that the assertion of P is warranted, I will be beyond relevant - that is, epistemic - reproach if I go on to assert it. But that is to say that I will be warranted in doing so. We accordingly obtain: There is warrant for thinking that [it is warrantedly assertible that P] iff there is warrant for thinking that [P], Given the Equivalence Schema, this will in turn yield: There is warrant for thinking that [it is warrantedly assertible that P] iff there is warrant for thinking that [it is true that P], And now, since warranted assertibility is, in a perfectly trivial sense, a normative property - a property possession or lack of which determines which assertions are acceptablc and which are not - it follows that truth is too. For by the above equivalence, to be warranted in thinking that P is true has exactly the same normative payload as being warranted in thinking that it is warrantedly assertible. Moreover, our finding is that truth, as characterized by the schemata, and warranted assertibility coincide in positive normative force. (Wright 2001: 755) The argument takes some unpacking. It appears to go as follows. The first step is to establish (I). (1) There is warrant for thinking that [it is warrantedly assertible that P] iff there is warrant for thinking that [P].

Copyrighted

Material

From normativity

to propertyhood

211

At the second step, (1) and the equivalence schema yield (2): (2) There is warrant for thinking that [it is warrantedly assertible that P] iff there is warrant for thinking that [it is true that P], At the third step, from (2) Wright concludes that, (3) Truth is a normative concept. ("Truth and warranted assertibility coincide in positive normative force. ") Two lesser points of criticism clear the path to big trouble. First, there is reason to quibble with (1). The argument for the right to left conditional ("If there is warrant for thinking that [P], then there is warrant for thinking that [it is warrantedly assertible that P]") rests on a strongly "internalist" assumption about justification. Wright assumes that anyone who is warranted in asserting that P must have reasons for saying "P, " and be able to become aware of those reasons. Thus one who is justified in saying " P " is justified in saying "'P' is justified, " etc. This smacks of a regress in the making, which is one reason why such strong assumptions about warrant are more controversial than Wright would have it. Second, the step from (1) to (2) requires not just the equivalence schema but also the assumption that substitution of "It is true that P" for "P" after "there is warrant for thinking that" is fair game. Blanket substitution of equivalents, even logical equivalents, in that context is not fair game. But here, Wright may take the step under the assumption that instances of the T-schema are common knowledge. If so, then arguably, "P" is warrantedly assertible just in case "it is true that P" is warrantedly assertible. Third, the big trouble is that (2) doesn't lead to the conclusion that truth is normative. This may be seen by comparison of (2) to another claim involving an iterated normative concept. Moore famously held that "good" iterates. If a state of affairs is good, then it is good that that state of affairs is good. Supposing Moore is right, (4) is right. (4) If it is good that [P] then it is good that [it is good that P], And if given the T-schema, one concludes (5): (5) If it is good that [it is true that P] then it is good that [it is good that P], What follows? That truth is normative? That it packs the normative force of the concept of good? Surely not. Good is a normative concept. Nothing follows from (5) about the nature of the concepts used in the sentences embedded after "it is good that. " By the same token, warrant is a normative concept, but nothing follows from (2) about the nature of the concept of truth, embedded as it is in (2) after "there is warrant for thinking that. "

Copyrighted

Material

212

Adam Kovach

6. 2 Wright's second argument In another recent version of his anti-deflationary argument, Wright arrives at the same conclusion from a different direction, beginning with the observation that one of the expressive functions of the truth predicate is its use as a device of endorsement (Wright 1999), If one wants to endorse, corroborate, affirm or commend what another person says, one can do so by saying "That's true" or "What you say is true. " 1 0 Wright thinks there is a lot submerged beneath this simple observation. To endorse something is to recommend or approve of that thing as meeting a certain standard or norm. It follows that if truth is a device for the endorsement of propositions, there is a relevant norm for the endorsement of propositions. Given that we believe and assert propositions, it seems that the relevant norm is a norm governing belief or assertion. So, to endorse a proposition as true, is, roughly, to recommend it or approve it as acceptable or appropriate for belief and for assertion because it meets some standard. And, according to Wright, this is a claim about the very meaning of "true. " He concludes that "affirmations of truth, are normative claims" (Wright 1999: 211). The conclusion is hasty. A strong normativity thesis does not follow directly from the observation that "true" is used to express approval according to a standard. For consider the case of a community, the Consumers, who take the fact that something is selling for a good price as an excellent reason to buy it. Consumers use "it's cheap" to endorse items for purchasing (and "it's expensive" as a form of discouragement). Surely it isn't so easy to turn "cheap" into a normative concept. All that a community of Consumers must do is share the belief that one ought to buy cheaply, and they are free to make off with our word and endorse away. The observation that we are not all consumers but we can't help being truth-mongers is to the point, but I don't think it can sway a skeptic about the normative character of the concept of truth. A weak normativity thesis remains an option.

6. 3 Arguments from the relation of truth to other normative concepts A more promising line of argument in support of the strong normativity thesis begins with the premise that truth is semantically related to other concepts, which are more obviously normative, such as rationality, epistemic justification and warranted assertibility. For example, supposing that concepts of justification are more obviously normative than the concept of truth, what distinguishes epistemic justification from other varieties of justification? Many epistemologists answer that epistemic justification must be defined so as to insure that this type of justification is "truth-conducive. " "Truth-conductivity" here means that having epistemically justified beliefs substantially increases the likelihood of having true beliefs (Bonjour 1985; Alston 1989, 1996). Assuming this to be the right epistemological view (and

Copyrighted

Material

From normativity

to propertyhood

213

there is debate about the subject), we can ask why epistemic justification is a normative concept. And it appears to be a good answer that it is because truth is normative, and truth occurs in the best account of epistemic justification. So epistemic justification is normative because truth is. Similar arguments might be made in the case of other normative cousin concepts of truth. Arguments that follow this strategy are arguments to the best explanation. As such, they are open to challenge, but they also can shift the burden of proof to those who doubt a strong normativity thesis. Why else would such concepts be normative? While it's no knock-down, this strategy can convince. 6. 4 Do we need a conclusive argument for the strong normativity

thesis?

Just how much support for a strong normativity thesis can we expect to find? There is good reason to think that lack of a conclusive argument is not a bad predicament for the advocate of the strong normativity thesis. After all, it should not be any easier to prove that truth is normative than it would be to prove that an obviously normative concept like goodness is normative. But the latter is not easy. Compare convincing someone who doubts that truth is a normative concept to convincing a strange skeptic who doubts that goodness and fairness are normative concepts, or still closer to the target, who doubts that thick value concepts like bravery and chastity are partly normative concepts. Fortunately no one wants to argue for weak normativity claims about these concepts. Indeed there may be nothing to gain by trying to prove that any concept is normative to someone who has contrary intuitions. What matters more is whether there are good semantic intuitions in support of the view that a concept is normative, and whether the view can be defended against objections. Elsewhere, I have argued that it is a defensible view that some truth claims entail normative claims, in that these alleged entailments pass standard Gricean tests (Kovach 2000). It would be surprising if the intuition that truth is a normative concept could be defended any more forcefully than this. In keeping with this verdict. I suggest we grant Wright his premise that affirmations of truth are normative claims and ask whether an inflationary conclusion can be drawn on this basis.

7 Wright's crucial premise The heart of Wright's inflationary argument is actually a simple inference, which can seem complicated because, as Wright presents it, it comes entangled with certain considerations about the difference between truth and epistemic justification. Having established that truth is a normative concept (as a consequence of the observation that truth is a device of endorsement) Wright insists, the deflationist clearly cannot a l l o w . . . that "true" when used to endorse. has the function of commending a proposition for its satisfaction of

Copyrighted

Material

214

Adam Korcich some distinctive norm which contrasts with epistemic justification and which only "true" and equivalents serve to mark. For if there were a distinctive such norm, it could hardly fail to be reckoned a genuine property of a proposition that it did, or ilid not comply with it. And if the norm in question were uniquely associated with "true" and its cognates, that would be as much as to allow that there would be a special property of truth - at which point the deflationary game would have been given away. (Wright 1999: 212, my italics)

Here Wright concludes that for the deflationist to accept the strong normativity thesis is to give away the game to give up the no-property thesis, that is. This conclusion rests crucially on the premise that if truth is a normative concept, then truth is a substantial property. This is the crucial premise to be examined, but first it is important to defend this way of understanding the argument, by disentangling the premise from the surrounding considerations about the difference between truth and epistemic justification, and showing how these are mere distractions from the main issue. The considerations about the difference between truth and epistemic justification snake their way into the argument, because Wright imagines a potential path of escape for the deflationist. As Wright points out. truth is not the only standard for the acceptability of beliefs. For various kinds of beliefs there are various local standards of justification and evidential support. For the sake of argument these standards may be brought together as varieties of epistemic justification. But then, worries Wright, perhaps the deflationist can say that the concept of truth expresses the property (or family of properties) that is already familiar to us as epistemic justification. The point of this maneuver would be to concede that the normativity of truth implies that there is some property relevant to the endorsement of propositions as true, but to soften the blow by claiming that we use "true" to endorse propositions as epistemically justified, a property all parties to the debate already admit exists (Wright 1999: 212). The maneuver fails, however. Truth is something distinct from epistemic justification. At least, this is something the deflationist must admit. The argument for this is as follows. Let "p" stand for some suitable sentence. It is possible that p even though one is not justified in believing that p. But the T-equivalence for " p " ensures that it is not possible that p even though one's belief that p is not true. So, "true" and "epistemically justified" diverge in intension, and by sad experience, we know they also diverge in extension. Wright concludes that the relevant norm for the endorsement of propositions as true is sui generis. "True" designates a distinctive standard distinct, that is, from epistemic justification. Wright's point that "true" and "epistemically justified" diverge in intension and extension merely serves to cut off the escape route of identifying truth with epistemic justification. It does not pose an independent problem for deflationism. How could it, since the difference between truth and epistemic

Copyrighted

Material

From normativity

to propertyhood

215

justification is a consequence of the T-equivalences? Furthermore, deflationists would be foolish to follow the escape route that Wright offers them. To accept the conclusion that truth is a property and to identify that property with epistemic justification is not to maintain a deflationary position. It is to adopt a crude epistemic theory of truth. Identifying truth with epistemic justification is giving up the deflationist's no-property thesis, pure and simple. Nor is there any gain for deflationism in noting that there are as many properties available as there are local varieties of epistemic justification. A fragmented theory of truth as a family of properties is still not a deflationary theory. Once these points are recognized, the difference between truth and epistemic justification can be seen for what it is, a distraction from the central issue, which is Wright's inference from normativity to propertyhood.

8 Normativity versus propertyhood Now for the central question: Does the inference from normativity to propertyhood go through or could truth be normative and nonetheless fail to express a property? As Wright makes the argument, the crucial premise comes without support. At least one deflationist, Paul Horwich (1998a: 143-4), responds that he finds "absolutely no reason" to accept Wright's inference in the absence of such support. Can the deflationist do better than this, and refute the crucial premise? Can the inflationist do better by supporting the crucial premise? Both questions are considered in this section.

8. 1 Hybrid normative-deflationary

theories

A good way to show that the normativity of truth does not entail propertyhood would be to construct hybrid deflationary theories of truth, and show that such theories can account for how the concept of truth functions normatively, in the absence of the assumption that truth is a property. Consider a hybrid deflationary theory, with two components, all the requisite instances of schema (E), and the principle (N). (N) The truth of a proposition makes it correct, which is, other things being equal, a reason for believing it and asserting it. It is to be understood that (N) is part of the meaning of the truth predicate. The question is whether a hybrid theory like this can explain our normative uses of the truth predicate without giving up the no-property thesis. It may seem that this isn't much of a challenge to the deflationist. For example, how does one use the truth predicate to endorse a proposition as true? The deflationist can reply as follows. To decide whether to endorse the assertion that snow is white as true, consult the T-equivalence for that assertion. The Tequivalence reveals that if one determines whether or not snow is white, one

Copyrighted

Material

216

Adam

Kovach

determines whether the proposition that snow is white is true. (N) ensures that if one applies the truth predicate to the proposition that snow is white, one makes a normative claim. T h a t is how we endorse the corresponding assertion and belief. Of course the theory of truth is silent about what one must d o to determine whether or not snow is white, but no theory of truth can be expected to include that knowledge. Apart from knowing how to determine what the facts are, the equivalence schema gives us all the guidance we need to use " t r u e " as a device of endorsement. N o assumptions about propertyhood are needed to explain how we accomplish this, and endorsements can be full-blown normative claims for all that. So deflationists are free to accept the strong normativity thesis, if they find it compelling, because a thesis like (N) can be grafted o n t o the standard deflationary account of competence while conceding nothing to inflationism.

8. 2 Reviving the inflationary

argument

It is time for a change of perspective. So far. following the deflationists and Wright, in discussing the normativity of truth, we have focused narrowly on only one normative role of the concept, by which people use the word " t r u e " to endorse, or give favorable criticism. However, contexts of endorsement are not the only ones in which truth might function as a n o r m . Let us distinguish two types of contexts in which a n o r m can operate. Contexts of performance or production are one. and contexts of assessment or criticism are the other. A diving judge knows the standards for scoring dives and applies them to the assessment or criticism of dives. A competent diver knows the same standards and applies them as best she can in diving. The diver performs. T h e j u d g e assesses. A professional egg-grader knows the standards that eggs have to meet in order to make grades B, A or A A . That's a context of assessment. Alas, the standards are lost on chickens, who work in a context of production. In its role as a device of endorsement, the truth predicate is used in contexts of assessment. Does the concept of truth also play a role in contexts of p e r f o r m a n c e and production? It certainly seems to. Wherever people strive to produce true assertions and beliefs, to give one another accurate information, or to discover true answers to their questions, they a p p e a r to be following n o r m s for the production of representations. If this is so, then truth functions normatively in at least two ways. As a device of endorsement the concept is used to guide behavior through the mediating influence of criticism. In contexts of production, however, norms play a more immediate role: there are norms governing the production of representations, and part of what is involved in understanding truth is understanding such norms. " It was argued that the deflationist has no endorsement use of "true. " which takes place about this second, hitherto neglected, role of To pursue this question we can proceed as

Copyrighted

Material

difficulty in accounting for the in contexts of assessment. How truth in contexts of production? before by considering a hybrid

From normativity

to propertyhood

217

theory of truth consisting of all the relevant T-equivalences, and some principle to capture the normative role of truth in contexts of production. It is helpful for the sake of argument to assume that this can be done with a single principle about representation (although the actual story of this aspect of our competence with truth may be more complicated). (R) will do: (R) Other things being equal, one ought to produce representations that are true. The question now is, can a deflationist account for our understanding of the principle (R) without assuming that truth is a property? The first point of the deflationist's answer is that (R) is a generalization. Of course the equivalence schema (E) is the basis for the deflationist's account of the use of the truth predicate as a device of generalization. According to the deflationist, (R) is a generalization of its instances, which are approximately of the following form: (R*) Other things being equal, one ought to produce a representation with the prepositional content that p only if the proposition that p is true. For each instance of (R*), schema (E) provides a T-equivalence, which is the basis for our understanding the role of truth in that instance of (R*). An instance of (R**) follows from each instance of (R*) by simple substitution of equivalents. (R**) Other things being equal, one ought to produce a representation with the propositional content that p only if p. So, according to the deflationist, one's grasp of instances of (R**) together with one's understanding of truth grounded in (E) accounts for one's grasp of (R). It is here that the critic of deflationism enters an objection. There is an important difference between understanding the principle about representation (R) and understanding any collection of instances of principle (R**). As a generalization about representations, (R) does not explicitly describe the content of any representation. Each instance of (R**), however, is about some specific representation with some particular content. In order to understand an instance of (R**), one has to understand the content of the specific representation that the instance is about. And, according to the deflationist, understanding (R) depends on understanding many instances of (R**). The critic charges that this type of understanding of the normative principle (R) falls short of our actual understanding of (R), because it gives an inadequate account of how understanding (R) guides us in contexts of production.

Copyrighted

Material

218

Adam

Kovach

The inadequacy can be brought out by an analogy. Imagine giving a would-be cartographer instructions for m a p p i n g terrain as follows. "If you find a stream, make a blue line like this, a n d if you find a hill, m a k e a round shape like that, and if you find a . . . . " O u r instructions continue in a list-like m a n n e r including conjuncts for how to depict as m a n y specific, different kinds of geographic features as we can think of. N o w can we imagine our cartographer setting out on cartographic adventures, strictly following the rules as given, never deviating from or going beyond the instructions on the list? If so, how will he behave when he comes across terrain features of a novel sort? Suppose we haven't thought of waterfalls in m a k i n g our instructions, and he finds a waterfall. Does our cartographer stop in his tracks muttering " N o w what?" or, after some head-scratching, does he invent a new convention, make some novel mark and go on his way? If it is the latter, then he has caught on to something about m a p - m a k i n g that goes beyond his instructions. He u n d e r s t a n d s some general n o r m of accuracy in m a p - m a k i n g that allows him to solve the problem of what to do with novel terrain. A n d if he gets that, he also understands that each time he represents some feature with some m a r k he is. in a way. doing m o r e of the same thing with the same aim. Returning to the case of truth, a radically disjunctive concept of truth can't be what we grasp in understanding (R). A person who understood (R) solely to the extent of understanding a collection of instances of (R**) would be like a cartographer who receives a list-like set of instructions, but fails to get the point of accuracy in m a p - m a k i n g . However, we aren't such hapless cartographers and our n o r m a l understanding of truth isn't like this. We grasp general n o r m s of truth in assertion and belief-formation beyond what the hybrid deflationary theory can convey with its interpretation of (R). It follows that anyone who u n d e r s t a n d s the concept of truth as we d o must be committed to the existence of some more unified feature (or group of features - just how fragmented a picture of truth our understanding permits has not yet been settled) than a radically disjunctive concept requires. T h e argument is s u m m e d up as follows. Part of what is involved in understanding the concept of truth is getting the point or aim of producing representations. We do, in fact, get the point, but having a massively disjunctive understanding of truth is not adequate for conveying the point. So there is more to our understanding of truth than even a hybrid n o r m a t i v e deflationary theory can capture. This seems to me the best line of argument that can be m a d e on behalf of inflationists who would follow Wright's lead in marching f r o m normativity to p r o p e r t y h o o d .

8. 3 Two

objections

Let us consider two ways of objecting to the argument. The first is to argue that deflationism accounts for all that we need to understand of a principle like (R). T h e second is to argue that even if the critic is right, and a hybrid

Copyrighted

Material

From normativity

to propertyhood

219

deflationary account does not capture our competence with truth, it does not follow that truth is a property. The deflationist can press the first objection as follows. Notice that the sole justification given for why a radically disjunctive understanding of map-accuracy or truth is inadequate is that it won't help us when we arrive on new terrain. But there are limits to what a cartographer can do with new terrain. He may not know how to map it. In such cases, knowing the norm can't help him a bit. The limits of what people can represent linguistically are the limits of their languages, just as one's understanding of the T-equivalences is limited to those instances one can formulate in one's own language. A disjunctive understanding of truth reflects the real state of our representational abilities: wherever one can successfully come up with an assertion, one can understand the corresponding T-equivalence. Once we see this, we can see how little understanding truth has to do with understanding norms for producing representations. Many abilities go into being able to produce a truthful representation. One has to know one's language (like the map-maker has to know rules of map-making). In addition, one has to be able to recognize and otherwise come to know the circumstances that one represents. What the critic of deflationism dramatically calls "understanding the point of producing truthful representations, " is really understanding the point(s) of doing many such things. Granting that being able to represent means grasping linguistic and cognitive norms, why exactly should understanding those add up to understanding truth? It is here that deflationist and critic might arrive at a deep disagreement about what it takes to understand the normative content of truth. The critic might insist that anyone who achieves a certain degree of linguistic and cognitive sophistication must acquire understanding of a general point or aim to producing representations, and that it is this aim that is meant by our concept of truth. The deflationist might simply deny this, and pending further argument, point to a hybrid theory as sufficient to capture the normative content of truth. There is no easy further argument to be made in support of the critic's view, but I expect that it will strike many as right just the same. Suffice it to say that progress past this point will be difficult, given deep theoretical divisions and the present state of knowledge about our linguistic and cognitive abilities. 12 To take up the second kind of objection, suppose it were conceded that our ordinary competence with truth involves understanding a normative principle that cannot be explained by a hybrid deflationary theory. In this case, understanding the ordinary notion of truth would involve commitment to some more unified feature(s) than a massively disjunctive concept requires. Does it follow that truth is a property? It does not. Any normative principle that belongs to our competence with truth which requires that there be a corresponding property, has the character of an expectation or demand placed upon the world. It is not up to us whether or not that demand is met. Those who believe that truth is normative may expect that truth is a more or less unified property, and their behavior may be governed

Copyrighted

Material

220

Adam Kovach

by norms, which make it appear to them that there is a common goal to their truth-directed cognitive and linguistic endeavors, but the world does not have to comply. If our demands are unrealistic, no degree of reflection on their internal nature can establish that they are met. It remains open to those who think our demands are unrealistic to argue that aspects of our understanding of truth are contaminated with conceptual error. This perspective is dramatized in Richard Rorty's (1995) tale about theistic archers, who aim not only to strike the mark but also to seek divine favor in striking the mark. They take their theistic aim seriously, to the point that it infects their competence with the locution "striking the mark" and gives these words religious significance. It does not follow that because the mark is struck, divine favor is also found. That would require collusion by God and the world. Similarly it does not follow from the fact that we think the concept of truth is normative that when we judge or assert truly, our moves instantiate a relevant property. It is possible to pursue an illusion. Truth could be normative yet stand for nothing more than "a mythical goal" of inquiry (Price 1988: 150). Wright is aware of this objection from error theorists such as Price and Rorty. His reply is that anyone who accepts the T-equivalences as true is bound to admit that some propositions are true. If some propositions are true, and it is part of what we mean by "true" that the true propositions meet a standard, then the standard is sometimes met. "It is not merely that our concept of truth calls for such a norm: the call is answered" (Wright 1999: 215). The trouble with this reply is that it fails to appreciate the nature of an error theory about truth. The view in question is not that truth is a purely normative concept. Like a thick value concept, truth has normative and also descriptive content. If the normative content implies that truth is a property, and there is no corresponding property, then it is the normative content alone which is in error. The error theorist about truth might deny that any propositions are true, in the full-blown normative sense, but won't deny that some propositions fall in the extension of "true" as defined by the T-equivelences. There is nothing stranger with this than there is with an error theorist about chastity denying that anyone is chaste without denying that some people are abstinent.

9 Conclusions The five main conclusions of this essay are as follows: 1 The intuition that truth is in some sense normative can be captured by either a strong normativity thesis, according to which the normativity of truth is part of the meaning of the truth predicate, or a weak normativity thesis, according to which the truth predicate enters into the formulation of certain norms, which may be important, but do not affect the meaning of "truth. "

Copyrighted

Material

From normativity

to propertyhood

221

2

N o argument has been uncovered which conclusively establishes a strong normativity thesis. It is nonetheless a plausible view that truth is a normative concept in the strong sense. The best evidence for this lies in semantic intuitions. There is not much that can be done to convince someone w h o doesn't share these intuitions. But in this respect, truth is no worse off than other putatively normative concepts. 3 If a strong normativity thesis is granted, it follows trivially that standard deflationism is inadequate as an account of the concept of truth. 4 Wright has claimed that it also follows from a strong normativity thesis that truth is a property. This is not so. A hybrid deflationary theory consisting of the union of standard deflationism and some normative principle(s) about truth of truth is possible. Such a theory can account for our intuition that truth is normative without surrendering the deflationist's no-property thesis. 5

Even if it were granted that our ordinary concept of truth is in part a normative concept that calls for the existence of a c o m m o n property (or collection of properties) of truth-bearers, it does not follow that there is such a property. Mere observations from the perspective of the conceptuser about the nature of our concept of truth c a n n o t rule out the possibility of an error theory of truth, according to which the normative content of our concept of truth is spurious.

This confirms the suspicion expressed at the beginning of this essay that it would be quite surprising if there were a successful inflationary argument f r o m the premise that truth is a normative. If some of the prcccding gives the impression of persistent attempts to milk a he-goat (or pick your own favorite K a n t i a n m e t a p h o r for elaborately futile philosophical argument) this is because there is more than one way to be a deflationist, and in this essay I have uncritically allowed alternatives to standard deflationism to bloom. However, before leaving the issue we should ask how plausible such views as hybrid normative-deflationary theories and error theories of truth really are. The hybridizing approach simply grafts whatever normative content one attributes to truth right on to the T-equivalences that fix the concept's descriptive content. It is not just that there is something ad hoc about this way of a c c o m m o d a t i n g the intuition that truth is normative. One wonders why we would have a concept like this at all, and beneath that lies the deepest question about the normative character of truth: Why is it that n o r m s of truth govern t h o u g h t and speech at all? Once this question is appreciated, it appears less important whether the strong or the weak normativity thesis is the right one. If truth is a genuinely normative concept, then it can be asked why we have such a concept, but if not, then it can just as well be asked why there are n o r m s of truth. For there is no question that truth does enter into pervasive n o r m s of thought and speech. The genetic question is even more pressing in the case of an error theory, because error theorists must have plausible explanations of the origin and

Copyrighted

Material

222

Adam

Kovach

endurance of widespread illusions. Rorty suggests that the origin of n o r m s of truth can be retraced through the history of a culture influenced by "Greek metaphysics and patriarchal m o n o t h e i s m " (Rorty 1995: 300). Price develops an ingenious speculative evolutionary account of the spread of belief in truth as a mythical goal, roughly due to the advantages of having such a n o r m in play as a catalyst for the coordination of beliefs across social groups (Price 1988, 1998). These simple explanations strike an inflationist like me as very implausible. I suspect that the story of why we follow n o r m s of truth and can't help but follow them will be hidden near the b o t t o m of a t h o r o u g h understanding of language, thought and meaning, which is yet very far off. But an inflationist can bet today that the explanation of why there are n o r m s involving truth will invoke properties that we will want to identify with truth. The wager is that if there is going to be an inflationary argument from the premise that truth is normative, it will be a version of the familiar inflationary argument from the explanatory role of truth - Truth itself will be a part of our best theory of why we have the normative concept of truth. 1 3

Notes 1 In this essay, I use italics to indicate concepts and double quotation marks to indicate words. The concept of truth is truth. "Truth" and "true" express that concept. And where "truth" appears in plain text it refers to the property, truth. 2 Wright's argument has drawn interesting and mixed criticism some of which is discussed here (Rorty 1995: Horwich 1996. 1998a; Price 1998). 3 Full disclosure: I am an inflationist. For a long time, I was under the spell of Wright's inflationary argument. This essay is about how I lost faith in the argument but not. as yet, in its conclusion. 4 In this section. I describe "standard deflationism, " in order to provide background for the arguments to come. More detailed presentations of this view can be found in Horwich (1990, 1998a) and Soantes (1999). My choice of a prepositional as opposed to a sentential equivalence schema as the basis for deflationism is only for expository purposes. Field (1999) and Hill (2002) develop the sentential alternative. The standard I describe captures features common to much recent deflationary thinking about truth. According to another deflationary theory of truth, the prosentential theory "is true" does not belong to the logical category of a predicate, but is an incomplete symbol, part of the expressions "it is true" and "that is true, " which belong to a logical category of sentential variables; see Grover (1992). Some difficulties for standard deflationism discussed in this article also affect the prosentential theory, but I will not discuss the prosentential theory here, and leave it to interested readers to consider how the arguments might be recast in the case of that theory. 5 See Quine (1970) and Horwich (1990. 1998a). 6 The partial definition is intended as a disjunctive account of truth, as an insubstantial property. It does not follow that this is a proper way to define the concept of truth. 7 Strawson (1949). with his early endorsement theory of truth accepted the noproperty thesis, but not the deflationist's competence thesis. Alston (1996: 4 1 5 1 ) holds the view that truth can be defined in terms of (E) and that it is also possible to develop a theory of truth as a property, perhaps along the lines of a correspondence theory.

Copyrighted

Material

From normativity

to propertyhood

223

8 Wright (1999) suggests that truth is a multiply realizable property. Lynch (2001) develops this view. 9 Horwich (1990, 1998a) argues that truth functions only as a device of generalization in these explanations. Gupta (1999) argues that in order for these explanations to be adequate, truth must be considered to be a substantial property; see also McGrath (1997a, 1997b) and Kovach (1997). My view is that the second kind of response is best for deflationists, and that while it may involve giving up on certain "realist" intuitions, for example, about how to explain the success of science, this is not so damaging since the alleged explanations are typically quite unilluminating. A tougher challenge is doing without explanatory concepts of truth and reference in a theory of meaning. 10 This observation forms the basis of Strawson's endorsement theory of truth, a deflationary view. Like the earlier redundancy theorists, Ramsey and Ayer, Strawson denied that truth is a property, but he recognized an additional feature of our competence with truth in its role as a device of endorsement. According to Strawson, to call a statement true is both to repeat it and to endorse it. 11 It would be absurd to suggest that people's understanding of truth somehow "tells them what to do, " or determines what they should say and believe, and that is not what has been suggested here. 12 The intuition that understanding the concept of truth comes with knowing a language is also, I think, closely allied to the intuition that the concept of truth is so basic that it is indefinable, but that is a large topic. Compare Davidson (1999: 309). 13 The difference between this argument-of-the-future and the kind of arguments considered in the rest of this essay is that the Wright-style arguments try to inflate truth from the inside, starting with assumptions about how our practices appear to us, while this new sort of argument starts from the outside, by trying to explain why we would have these truth-involving practices. It may be that Wright (2001: 757) himself is coming to see his old trademark inflationary argument in this way. In his most recent presentation of it, there is a final twist, which deserves attention. As noted above, "true" and "epistemically justified" differ in intension and extension. It follows that neither "There is no life on Mars" nor "The proposition that there is no life on Mars is true" can mean the same as "The proposition that there is no life on Mars is epistemically justified. " Wright challenges the deflationist is to explain this difference in meanings. The reason is that, according to Wright, "this distinction cannot be recovered from any contrast between the circumstances under which the two propositions are respectively warranted, s i n c e . . . there is none. The difference between them resides, rather, precisely in a difference in correctness conditions of another sort (whisper: truth conditions). " And there is the final twist: truth is an explanatory concept required for a truth-conditional theory of meaning. Wright chooses to make this point by contrasting the truth-conditional approach with the program of explaining meaning in terms of assertibility conditions (circumstances under which assertion is warranted). The latter, he claims fails to distinguish meanings adequately. Wright recognizes that this is not the end of the story (Wright 2001: 783, fn 12). He suggests that a deflationist can respond by explaining the difference in terms of some other theory of meaning, such as an inferential role theory. Indeed it would be surprising if an inferential role theory could not explain the difference in question. After all, the consideration that the biconditional "The proposition that Mars is uninhabited is warranted iff Mars is uninhabited" is false, suffices to show that the "The proposition that Mars is uninhabited is warranted" and "Mars is uninhabited" must have different inferential roles. So, an inferential role theory ought to assign propositions of the form "p" and "The proposition that p is warranted" different inferential roles in recognition of the different meanings.

Copyrighted

Material

224

Adam

Kovach

What is important for the purposes of this essay, however, is not the question of whether a deflationist about truth can successfully take an inferential role approach to the theory of meaning, nor whether truth has a role in the theory of meaning inconsistent with deflationism. The question is whether there is an inflationary argument from the premise that truth is normative. The argument that only inflationists can allow the concept of truth to play a necessary role in the theory of meaning, whatever its force, has nothing to do with the normativity of truth. It can be made quite independently. If this is what Wright's inflationary argument boils down to. then it is hard to see why he suggests that deflationists should worry about the normativity of truth in the first place. Yes, but it is worth noticing that Wright's new version of the argument represents a shift from what I have called inflating truth from the inside, to inflating it from the outside.

References Alston, W. P. (1989) Epistemic Justification: Essays in the Theory of Knowledge, Ithaca, NY: Cornell University Press. (1996) A Realist Conception of Truth. Ithaca. NY: Cornell University Press. Bonjour. L. (1985) The Structure of Empirical Knowledge, Cambridge, MA: Harvard University Press. Davidson, D. (1999) "The Folly of Trying to Define Truth, " in S. Blackburn and K. Simmons (eds) (1999) Truth, Oxford: Oxford University Press. Dummett, M. (1978) Truth and Other Enigmas. Cambridge, MA: Harvard University Press. Field, H. (1999) "Deflationist Views of Meaning and Content. " in S. Blackburn and K. Simmons (eds) (1999) Truth, Oxford: Oxford University Press. Grover, D. (1992) A Prosentential Theory of Truth, Princeton, NJ: Princeton University Press. Gupta, A. (1999) "A Critique of Deflationism. " in S. Blackburn and K. Simmons (eds) (1999) Truth, Oxford: Oxford University Press. Hill, C. (2002) Thought and World: An Austere Portrayal of Truth. Reference, and Semantic Correspondence, Cambridge: Cambridge University Press. Horwich, P. (1990) Truth, Oxford: Oxford University Press. — (1996) "Realism Minus Truth. " Philosophy and Plienomenological Research, 56: 877-81. (1998a) Truth, 2nd edn, Oxford: Oxford University Press. (1998b) Meaning, Oxford: Oxford University Press. James, W. (1978) Pragmatism, a New Name for Some Old Ways of Thinking, and The Meaning of Truth, a Sequel to Pragmatism. Cambridge. MA: Harvard University Press. Kovach, A. (1997) "Deflationism and the Derivation Game, " Mind, 106: 575-9. (2000) "Truth as a Value Concept. " in A. Gupta and A. Chapuis (eds) (2000) Circularity, Definition and Truth. New Delhi: Indian Council of Philosophical Research. Lynch. M. (2001) "A Functionalist Theory of Truth. " in M. Lynch (ed. ). The Nature of Truth, Classic and Contemporary Perspectives, Cambridge, MA: MIT Press. McGrath, M. (1997a) "Weak Deflationism, " Mind. 106: 69-98. (1997b) "Reply to Kovach. " Mind. 106: 581-6. Price, H. (1988) Facts and the Function of Truth. New York: Blackwell Press.

Copyrighted

Material

From normativity

to propertyhood

225

(1998) "Three Norms of Assertibility, or How the MOA Became Extinct," in J. Tomberlin (ed. ) Philosophical Perspectives. 12, Language, Mind, and Ontology, Cambridge: Blackwell Press. Quine, W. (1970) Philosophy of Logic. Englewood Cliffs, NJ: Prentice Hall. Rorty, R. (1995) "Is Truth a Goal of Enquiry: Davidson vs. Wright, " Philosophical Quarterly, 45: 281-300. Russell, B. (1973) "Meinong's Theory of Complexes and Assumptions" in D. Lackey (ed. ). Essays in Analysis, New York: Braziller. Sellars, W. (1962) "Truth & "Correspondence", " The Journal of Philosophy. 59: 29-56. Soames, S. (1999) Understanding Truth. New York: Oxford University Press. Strawson, P. F. (1949) "Truth, " Analysis, 9: 83-97. Wright, C. (1992) Truth and Objectivity, Cambridge, MA: Harvard University Press. (1996) "Precis of Truth and Objectivity, " Philosophy and Phenomenological Research. 56: 863-8. (1999) "Truth: A Traditional Debate Reviewed, " in S. Blackburn and K. Simmons (eds) (1999) Truth. Oxford: Oxford University Press. (2001) "Minimalism, Deflationism, Pragmatism, Pluralism, " in M. Lynch (ed. ). The Nature of Truth, Classic and Contemporary Perspectives, Cambridge. MA: MIT Press.

Copyrighted

Material

11 Norms of assertion Graham Oppy

Grice (1989: 26-8) provides a taxonomy of norms of assertion. ' Within the overarching framework provided by "the co-operative principle" - i. e. the principle that "one ought to make o n e s conversational contribution such as is required at the stage at which it occurs, by the accepted purpose or direction of the talk exchange in which one is engaged" Grice distinguishes four types of norms of assertion. First, there are norms of quantity, which relate to the quantity of information to be provided: (1) do not provide less information than is required for the purposes of the current exchange of information; (2) do not provide more information than is required for the purposes of the current exchange of information. Second, there are norms of quality. Grice assigns these norms an ordering: the "super" norm (1) do not assert that which is not true; and the "more specific" norms (2) do not assert that which you do not believe; and (3) do not assert that for which you have inadequate warrant. (Note that these three norms taken together might plausibly be taken to entail a derivative norm of quality, namely (4) do not assert that which you do not know. 2 ) Third, there is the norm of relation: be relevant. Fourth, there are various norms of manner-. (1) avoid obscurity of expression; (2) avoid ambiguity; (3) avoid unnecessary prolixity; (4) be orderly; and so forth. 3 Grice notes that there are many other types of norms - aesthetic, social, moral, etc. - that are typically observed by participants in talk exchanges, but that these other types of norms are not "specially c o n n e c t e d . . . with the particular purposes that t a l k . . . is adapted to serve and is primarily employed to serve" (Grice 1989: 28). Moreover, Grice also says that he would like to be able to show that anyone who cares about the goals that are central to conversation/ c o m m u n i c a t i o n . . . must be expected to have an interest, given suitable circumstances, in participating in talk exchanges that will be profitable only on the assumption that they are conducted in general accordance with the Cooperation Principle and the [norms]. (Grice 1989: 30)

Copyrighted

Material

Norms of assertion

227

Grice's norms of quality have frequently reappeared - at least inter alia in subsequent discussions of the norms of assertion (and often with a similar ordering assigned): consider, for example, Price (1998: 245-9), who claims that assertion is subject to the "weaker" norms (2) and (3) and to the "stronger" norm (1), or Williamson (2000: 238-69), who claims that assertion is subject to norm (4) and hence, derivatively, to each of the norms (1)(3). Much of the more recent subsequent discussion has been concerned with the question whether those who embrace minimalist conceptions of truth can consistently endorse Grice's "super" norm. However, before we proceed to a consideration of this more recent debate, I think that it will be worth spending some time considering in what sense - if at all - Grice's "super" norm really is a norm of assertion. The plan of the paper is as follows. Following time honoured tradition, I begin by drawing some distinctions, and adverting to some disputes upon which I will take up a stance without proper discussion. After briefly considering the distinction between norms that are distinctive of assertion and norms that are shared between assertion and other speech acts, I spend some time thinking about different ways in which norms can be classified (and, in particular, I consider ways in which the taxonomy of Wright (1992) can be improved upon). Then, after some brief consideration of the distinction between norms and secondary proprieties, I conclude with a rather short discussion of exactly what it takes for a speech act to be an assertion (as opposed to some other kind of speech act that is characteristically performed using declarative sentences). In the second section of the paper, I turn my attention to the claim that it is a norm of assertion that one ought not to assert that which one does not know. I argue that the arguments that Williamson (along with others) has offered in defense of this claim fail, and that there are good reasons for rejecting the claim that it is a norm of assertion that one ought not to assert that which one does not know. In the third section of the paper, I consider the claim that it is a norm of assertion that one ought not to assert that which is not true. I argue that the standard argument in favor of this claim fails, and that there is good reason to deny that it is a norm of assertion that one ought not to assert that which is not true. Moreover, I also argue that there is similar good reason to deny that it is a norm of assertion that one ought not to assert that for which one has insufficient warrant. However, I do not go on to reject all of Grice's norms of quality, for I also contend that it is the central norm of assertion that one ought not to assert that which one does not believe. In the fourth section of the paper, I consider some further reasons for supposing that the central norm of assertion is that one ought not to assert that which one does not believe. After noting that it is plausible to suppose that assertion is the sole speech act that has the expression of belief as its proper end, I argue that consideration of the Gricean ambition mentioned above - namely, the ambition to show that people should be expected to have

Copyrighted

Material

228

Graham Oppy

an interest in participating in conversational exchanges of this kind only on the assumption that they are conducted in general accordance with this norm supports the claim that the sole constitutive individuating norm of assertion is that one ought not to assert that which one does not believe. In the fifth section of the paper, I turn my attention to the consequences that the view that I have defended about the constitutive individuating norms of assertion has for a recent dispute about minimalist and deflationary theories of truth. Wright (1992) claims that considerations about the norms of assertion establish that deflationism is an inherently unstable view. Horwich (1998) denies that Wright's arguments establish the conclusion for which he argues. Price (1998) argues that, while Horwich is right to criticize Wright's arguments, it is nonetheless true that there are norms of assertion that deflationary theories of truth fail to capture. I argue that Price's argument fails to establish that there are norms of assertion that deflationary theories of truth fail to capture; and, moreover, I argue that Wright and Price get great mileage from mistaken views about the norms of assertion. In the sixth, and final, section of the paper, I provide the beginnings of a discussion of the connections that exist between norms of assertion, norms of belief, and norms of inquiry. In particular, I consider the impact that the adoption of the views of Kelly (2005) on the epistemic significance of disagreement has for discussion of the norms of assertion. If we suppose that disagreement between doxastic peers need give none of them reasons to change their minds, then I do not think that we should suppose that it is even a proximate aim of assertion, in general, to secure agreement between parties to a conversation. 1 Preliminary considerations (1) Grice's norms of assertion can be sorted in ways other than those that he mentions. In particular, it is worth noting that most of these norms apply to a wide range of linguistic phenomena, of which assertion is but one example. Plainly enough, the norms of relation and manner apply to giving orders, asking questions, and so forth, no less than they apply to making statements. (One should avoid irrelevance, obscurity of expression, ambiguity, unnecessary prolixity, disorder, and so forth, in asking questions or giving orders. ) Moreover, there is at least an analogue of the norms of quantity that applies in these cases as well. (One should not ask unnecessarily complicated or unsuitably simple questions; one should not give unnecessarily complex or insufficiently detailed orders. ) Consequently, if there are norms that are distinctive of assertion - i. e. norms that apply only to assertions and not to other linguistic phenomena - then plausibly these norms will be those assigned by Grice to quality. (2) There are many different ways of classifying norms. Wright (1992: 15-16) mentions three distinctions. First, he claims, we may distinguish between

Copyrighted

Material

Norms of assertion

229

descriptive norms and prescriptive norms. (A characteristic of moves in a practice is a descriptive norm iff participants in the practice are guided in their treatment of moves by whether moves possess that characteristic. A characteristic of moves in a practice is a prescriptive norm iff recognition that moves possess that characteristic provides defeasible reason for specified treatment of those moves, even if these reasons tend to go unacknowledged by participants in the practice. ) Second, we may distinguish between positive norms and negative norms. (Positive norms are tied to the selection/endorsement/permission of moves; negative norms are tied to the rejection/condemnation/prohibition of moves. ) Third, we may distinguish between constitutive and non-constitutive norms. (Roughly, a constitutive norm is a norm that "enters constitutively into the identity of the practice concerned" (Wright 1992: 16). On Grice's account, amongst the identified norms for assertion, only the norms of quality are so much as candidates for consideration to be constitutive norms of assertion. ) There are ways in which this account might be tidied up. Perhaps, for example, it would be neater to start with the distinction between positive and negative characteristics - supposing that that distinction can somehow be made out - and then frame the distinction between descriptive and prescriptive norms in terms of contrary pairs of characteristics. (So, for example, the pair of characteristics < true, not true > would be a candidate prescriptive norm of assertion. ) If one were to proceed in this way, then one would not need to follow Wright in first defining what it is to be a positive descriptive norm, and then adding on an account of what it is to be a negative descriptive norm. Moreover, there are other distinctions that require attention (and which, when noted, might lead us to reformulate the distinctions that we already have). For example, there is a distinction between what I shall call internal norms and external norms. An internal norm for a practice is a norm that a participant in the practice is guaranteed to be able to apply in the case of his own moves (provided that he is sufficiently attentive, intelligent, and so forth); whereas an external norm for a practice is a norm that a participant in a practice may not be able to apply to his own moves, no matter how attentive, intelligent, and so forth that participant may be. I suppose that Grice's norms of quantity, relevance, and manner are all internal norms: provided that I am sufficiently attentive, intelligent, etc., I am as well-placed as anyone to judge that my assertions satisfy these norms. However, I take it that at least the first and the third of Grice's norms of quality are external norms: it may well be that other people can see that my assertions are neither true nor adequately warranted, even though these facts are invisible to me (and could remain so even if I were more intelligent, attentive, and so forth). A second distinction to which attention might be paid discriminates between what I shall call explicit norms and what I shall call implicit norms. (As with my previous distinction, the choice of labels may not be entirely happy. ) An explicit norm for a practice is a norm that is explicitly and

Copyrighted

Material

230

Graham Oppy

consciously acknowledged to be a norm by the participants in the practice (and perhaps even is such that one cannot be a fully fledged participant in the practice unless one explicitly and consciously recognizes the norm in question); whereas an implicit norm for a practice is a norm that need not be explicitly and consciously acknowledged to be a norm by participants in the practice. A third distinction, related to the second. discriminates between what I shall call overt norms and what I shall call covert norms. An overt norm for a practice is a norm that is or must be - borne in mind by participants in the practice when they are carrying out that practice; a covert norm for a practice is a norm that is "conformed to" by participants in a practice but which is not borne in mind by participants in the practice when they are carrying out that practice. A fourth distinction, again related to the distinctions already mentioned, discriminates between what I shall call individual norms and what I shall call group norms. An individual norm for a practice is a norm that governs all of the individual instances of participation in the practice, whereas a group norm governs the operation of more significant parts of the practice as a whole. (It might be, for example, that, while it is not an individual norm that an agent ought not to believe that which is false, it is nonetheless a group norm that the group ought not to believe that which is false. ) Perhaps we might also wish to distinguish practice norms, which govern the operation of the practice as a whole: for example, it might be that, while it is not a group norm that a group ought not to believe that which is false, it is nonetheless a practice norm that inquiry ought not to settle into stable equilibrium on that which is false. A fifth distinction, also related in complicated ways to the distinctions already mentioned, discriminates between what I shall call self-regulating norms and what I shall call other-regulating norms. Roughly speaking, an other-regulating norm is a norm to which explicit appeal can only be made in order to try to regulate the behavior of other participants in the practice, whereas a self-regulating norm is a norm to which one can make explicit appeal in order to try to regulate one's own behavior. One might think, for example, that it is a self-regulating norm of assertion that one ought not to assert that which one does not believe, and (hence?) that it is only an otherregulating norm of assertion that one ought not to assert that which is not true. (A gap between belief and truth may be visible in the case of others, but not in one's own case. ) There are doubtless further distinctions to be drawn, and also doubtless ways in which the distinctions that I have drawn can be refined. However, we have already done enough to show that it is not a simple matter to ask, for example, whether it is a norm of assertion that one ought not to assert that which one does not know. For it may be that our question is concerned primarily with, say, internal, explicit, overt, individual, self-regulating, prescriptive, constitutive norms; and it might also be that the answer would be different if we were asking about, say, external, implicit, covert, group, other-regulating,

Copyrighted

Material

Norms of assertion

231

descriptive, constitutive norms. Given the complexities here, I shall mostly proceed to ignore the distinctions that I have drawn; however, it should be remembered that there is a tacit disagreement in the literature between those who suppose that we are primarily interested in the internal, explicit, overt, individual, self-regulating, prescriptive, constitutive norms of assertion, and those who do not suppose that these are the kinds of norms in which we have primary interest. (3) There is a distinction that has become more or less standard in the literature between primary norms and secondary norms (or secondary proprieties). Given that it is a primary norm that N, it is a secondary norm that one do one's best to conform to N. Moreover, provided that one truly does one's best to conform to N, one is not open to (much) criticism if one fails to conform to N. Nonetheless, while an actor may not be (much) criticized for her failure to conform to N, her action is nonetheless criticizable because it fails to conform to N. (See, for example, Williamson (2000: 257) and De Rose (2002: 180). ) There are clearly distinctions to be drawn here. It is, for example, one thing to conform to a norm, and a quite different thing to reasonably take oneself to have conformed to a norm. Similarly, as Williamson and De Rose observe, it is one thing to have taken every measure that one could be reasonably expected to take in order to try to conform to a norm, and another thing to actually succeed in conforming to the norm. However, it is worth noting that these kinds of distinctions have much more significance for some kinds of norms than they do for other kinds of norms. For example the gap between conforming to a norm and reasonably taking oneself to be conforming to a norm is much larger in the case of external norms than it is in the case of internal norms. If a norm is internal - i. e. is such that a participant in the practice is guaranteed to be able to apply the norm in her own case provided that she is sufficiently attentive, sensitive, intelligent, and so forth - then the gap between conforming to the norm, and doing everything that one could reasonably be expected to do in order to try to conform to the norm, is occupied only by cases in which agents are constitutionally unable to attain the levels of attention, sensitivity, intelligence, and so forth that are required in order to be able to apply the norm in one's own case (and perhaps there is no gap at all between conforming to the norm and reasonably taking oneself to be conforming to the norm). On the other hand, if a norm is external - i. e. is such that there is no guarantee that a participant in the practice is able to apply the norm in her own case, no matter how attentive, sensitive, intelligent, etc. she may be - then there is much more scope for agents to violate the norm while nonetheless reasonably taking themselves to be conforming to it. (4) Before we can discuss the standing of norms of assertion, we need to determine what is to count as assertion. This is by no means a straightforward

Copyrighted

Material

232

Graham Oppy

matter. Williamson (2000: 258) proposes that assertion is the default use of declarative sentences. It is not clear that this is acceptable. There are many things that can quite properly be done using declarative sentences: swearing, conjecturing, guessing, betting, suggesting, predicting, explaining, opining, story-telling, and so forth. Given the many and varied proper uses of declarative sentences, it seems that we need some further guidance in order to determine which uses of declarative sentences are the default uses. I think that it is a mistake to try to give an account of assertion in terms of a restriction on the class of uses of declarative sentences. Instead. I would prefer an account that ties assertion to belief: the proper function of assertion is to express belief. Since there is a range of mental states other than belief that are quite properly expressed using declarative sentences, we should resist any suggestion that assertion can be analyzed in terms of the default uses, or the proper uses, or the conforming-to-implicit-convention uses, of declarative sentences. (Of course, the view that I am suggesting here is enormously controversial, and requires much stronger defence than I can hope to supply. Nonetheless, I need at least to mention it. since it has some role to play in the subsequent discussion. ) Enough of these preliminary discussions! I turn now to an examination of the claim - arguably implicit in Grice's account of the norms of assertion, and recently defended with considerable vigor by Williamson - that one ought not to assert that which one does not know. 2 Assertion and knowledge Williamson (2000: 238-69) defends the view that there is just one constitutive individuating norm of assertion: one may assert that p only if one knows that p. To say that this norm is the sole constitutive individuating norm of assertion is to say that all other norms for assertion are the joint outcome of this constitutive individuating norm and other considerations that are not "specific to assertion. " Williamson introduces the technical term "simple" to describe any theory of the norms of assertion according to which there is just one constitutive individuating norm or assertion. Williamson acknowledges that it might well turn out that the correct theory of the norms of assertion is not simple: there might be several constitutive individuating norms of assertion, or there might be none; and he also offers nothing more than "theoretical satisfaction" as a reason for supposing in advance that the correct account of the norms of assertion is simple. However, many of the arguments that Williamson gives are arguments for the conclusion that it is at least a norm of assertion that one ought not to assert that p if one does not know that p; it is to these arguments that we shall be attending here. Williamson offers two main reasons for supposing that it is at least a norm of assertion that one ought not to assert that p if one does not know that p. First, he claims that the supposition that it is a norm of assertion that one ought not to assert that p if one does not know that p provides a

Copyrighted

Material

Norms of assertion

233

better explanation of the inadequacy of probabilistic grounds of assertion than do accounts which suppose that there is some other norm of assertion. Second, he claims that there are various kinds of "conversational patterns" that confirm the suggestion that it is a norm of assertion that one ought not to assert that p if one does not know that p. In support of the first claim described in the preceding paragraph, Williamson discusses the example of someone who asserts that a given ticket in a lottery with many equiprobable tickets did not win on the sole grounds that it is massively unlikely that that ticket won (and the generalized example of someone who asserts a proposition on the sole grounds that it is overwhelmingly likely that the proposition is true). Williamson (2002: 250) claims that an "obvious moral" of his discussion is that "one is never warranted in asserting a proposition by its probability (short of 1) alone. " Moreover, he contends: (1) that this "obvious moral" is well-explained by the observation that it is a norm of assertion that one ought not to assert that p if one does not know that p (essentially because "the standard of probability 1 on one 's evidence is no more demanding that the standard of knowledge"); and (2) that this "obvious moral" is not well explained by any "weaker" norms of assertion. Despite Williamson's claims to the contrary, it seems to me that the "obvious moral" is actually quite a skeptical one. Suppose, for example, that something like standard quantum mechanics is known to be true and that there are very small but non-zero - probabilities for events such as a stationary marble tunnelling through the wall to which it is adjacent. By ordinary standards, if I place a chair in an otherwise empty room, and then close the only door to the room behind me as I leave, the closing of the door - and the subsequent loss of eye-contact with the chair - does not bring it about that I do not know that there is a chair in the room. But. given standard quantum mechanics, once the door is shut, my sole grounds for believing that there is a chair in the room are probabilistic: there is a tiny chance that the chair has tunnelled from the room since I shut the door. So, Williamson will have us say that, if something like standard quantum mechanics is known to be true, then we do not know a huge number of the things that we ordinarily take ourselves to know; and, moreover, he will also have us claiming that, if something like standard quantum mechanics is known to be true, then we violate the norms of assertion when we say such things as that there is a chair in the unoccupied room next door. Since it is not out of the question that something like standard quantum mechanics is known to be true, I take it that these are not acceptable results: one might be - and perhaps even can be - warranted in asserting a proposition by its probability (short of 1) alone. But. if that's right, then - without considering the remaining steps in Williamson's argument we can conclude that considerations about lotteries and the like do not end up supporting the claim that it is a norm of assertion that one ought not to assert that which one does not know. 4

Copyrighted

Material

234

Graham Oppy

In support of the second of the two contentions mentioned above, Williamson offers several types of conversational evidence that he takes to confirm the suggestion that it is a norm of assertion that one ought not to assert that p if one does not know that p. (1) One standard response to assertions is to ask, "How do you know that?" That this question constitutes an implicit challenge to the assertion provides support for the claim that one ought not to assert that which one does not know (and is hard to explain if this claim is rejected). (2) A more aggressive response to an assertion is to ask. " D o you really know that?" That this question is more aggressive is well explained if it is a norm of assertion that one ought not to assert that which one does not know (and hard to explain otherwise). (3) There is something wrong with any assertion of the form "A. but I do not know that A. " That there is something wrong with any assertion of this form is easy to explain if it is a norm of assertion that one ought not to assert that which one does not know (and hard to explain otherwise). The support that is offered by this conversational evidence for Williamson's central contention is weak. I think it is clear that one can find lots of cases in which assertions are made in which it is proper to respond with "How do you know that?" and " D o you really know that, " because there are lots of cases in which assertions are properly construed as claims to knowledge. However, there are other central cases of assertion in which it would not be proper to respond with these kinds of questions, and in which it seems very odd to construe assertions as claims to knowledge. Consider, for example, the case of assertions about the future. Suppose, for example, that we are discussing the upcoming Test. I assertively utter the words "Australia will win. " This looks like a perfectly straightforward and normatively unobjectionable kind of assertion. But. if challenged, I would immediately concede that I don't know that Australia will win. Indeed, if pushed, I'd be quite happy to say: "Of course, ¡don't know that Australia will win, hut (nonetheless) Australia will win. " The general point here is that there is an important family of declarative sentences of the form "It will be t h a t . . . " of which it cannot plausibly be maintained that their default use is governed by the norm that one ought not to assert that which one does not know. Moreover, while one can argue about whether acts of swearing, conjecturing, guessing, betting, suggesting, explaining, opining, story-telling and so forth are assertions, it seems very hard to deny that standard uses of sentences of the form "It will be t h a t . . . " are assertions. Given the discussion to this point. I think that we are in a position to conclude: (1) that Williamson's arguments do not establish that it is a norm of assertion that one ought not to assert that which one does not know; and (2) that there are considerations that weigh heavily in favor of the claim that it is not, in fact, a general norm of assertion that one ought not to assert that which one does not know. Moreover. I think that we are also in a position to conclude that these very same considerations weigh heavily against some "modifications" of the position that Williamson defends, for

Copyrighted

Material

Norms of assertion

235

example, that it is at least a norm of assertion that one ought not to assert that which one does not believe that one knows, and that it is at least a norm of assertion that one ought not to assert that which one does not rationally believe that one knows. Considerations about assertions about the future (and perhaps also considerations about assertions based on merely probabilistic grounds) strongly suggest that one can quite properly make assertions of that which one does not believe that one knows, and of that which one does not rationally believe that one knows. 3 Assertion and truth Williamson supposes that the chief competitor to the view that he defends is the simple theory that maintains that the sole constitutive individuating norm of assertion is that one may assert that p only if it is true that p. While he supposes that it is, indeed, a (derivative) norm of assertion that one ought not to assert that which is not true, Williamson contends that this norm could not be the sole constitutive individuating norm of assertion, because there is no way of deriving evidential norms of assertion from this norm. Given that it seems plausible to suppose that there is no way of deriving the claim that one may assert that p only if it is true that p from evidential norms of assertion, one might think that, if Williamson is right to claim that there is no way of deriving evidential norms of assertion from the claim that one may assert that p only if it is true that p. then the argument of the preceding section gives us good reason to think that there is no correct simple account of the norms of assertion. However, even if Williamson is right to claim that there is no way of deriving evidential norms of assertion from the claim that one may assert that p only if it is true that p, but wrong to claim that it is a norm of assertion that one may assert that p only if one knows that p. it does still remain open that there is a correct simple account of the norms of assertion: for it might yet turn out that it is not a norm of assertion that one may assert that p only if it is true that p.5 It may seem obvious that it is, in fact, a norm of assertion that one ought not to assert that which is not true. After all, one might think, it is clear that we very often object to, or criticize, assertions on precisely these grounds. If, for example, someone asserts that p in circumstances in which one knows, and perhaps even knows that one knows, that not p, then - at least other things being equal - one will be entirely justified in pointing out to the person in question that they are mistaken because what they said is simply not true. How could one always be pro tanto justified in pointing out that one who has made a false assertion has made a mistake unless it is a norm of assertion that one ought not assert that which is not true? I do not believe that this is the decisive consideration that some take it to be. Certainly, it should not be denied that someone who makes a false assertion is subject to a certain kind of liability; but it does not follow from this fact that they have violated one of the norms of assertion. Suppose that

Copyrighted

Material

236

Graham Oppy

we agree that it is one of the norms of assertion that one ought not to assert that which one believes to be false. Suppose further that we are dealing with a case in which a speaker is conforming to this norm, so that when she makes the false assertion that p. this is because she has the false belief that p. Then it plainly remains open to say that, while this person violated no norm of assertion in expressing her false belief that p, she is nonetheless mistaken in believing that p. (If we were so inclined, we could go on to claim that she must have violated one of the norms for belief, namely that norm which enjoins that one believe that p only if it is true that p. I shall return to consider this proposal later. ) Perhaps we might go on to allow that there is a kind of secondary or derivative " n o r m " of assertion to the effect that one ought not to assert the false because one ought not to believe the false; 6 but this allowance is perfectly consistent with the proposal that the central norm of assertion is that one ought to assert only that which one believes. Indeed, for all that the considerations currently under examination can show, it may be that the sole constitutive individuating norm of assertion is that one may assert that p only if one believes that p. There is a general point here that deserves to be made explicit. The practice of assertion does not exist in isolation; rather, it is one amongst many practices that are related in various ways. Given the variety of relations that hold between assertion and other practices, it may be that what appear to be norms that are specific to the practice of assertion are not really as they appear to be: for the norms of assertion and the norms of other practices may jointly issue in what are merely apparent norms of assertion. Moreover, this general point does not rely upon the correctness of the example which I have used to illustrate it: it might be that there is something other than norms of belief to which one might appeal in an attempt to explain the merely apparent norm that one ought to assert only that which is true. If it is conceded that one might, in the way suggested, undermine the obvious argument in favor of the claim that it is a norm of assertion that one ought to assert only that which is true, then it seems to me that it ought also to be conceded that we cannot hope to determine the norms of assertion simply by thinking about general kinds of criticisms that are specific to assertions. Even if it is true that assertions are unique amongst speech acts in always being open to criticism on grounds of falsity, that in itself does not necessarily provide a strong reason for thinking that it is a norm of assertion that one ought to assert only that which is true. So far, then, perhaps, we have managed to neutralize the suggestion that there is a straightforward argument for the conclusion that it is a norm of assertion that one ought to assert only that which is true. But is there any plausible argument for the conclusion that it is not, after all, a norm of assertion that one ought to assert only that which is true? In order to answer this question, I think that we shall need to think a bit harder about what we take the main function of assertion to be.

Copyrighted

Material

Norms of assertion

237

There is widespread acceptance of the view that the primary function of assertion is to facilitate transmission of information from speaker to hearer: for example, the speaker is able to pass on knowledge to the hearer (Williamson (2000: 267)); or the speaker is able to pass on truths to the hearer (Grice (1989: 26-8)); and so forth. But, given that we have a transactional or transferential view of the primary function of assertion, it seems to me to be more or less inevitable that we shall end up supposing that it is a norm of assertion that one ought not to assert that which is not true. After all, it seems that nothing less would suffice to guarantee the quality of that which is transferred: transactional quality control seems to require that the information to be transmitted meets the standard of truth. We can certainly imagine communities in which the primary function of assertion would be to facilitate transmission of information from speaker to hearer. Imagine, for example, a world in which agents have a general need to be apprised of as many facts as they can be and in which agents are perfect detectors of facts to which they are directly exposed, but in which direct exposure to facts is rare. In communities of agents in this imagined world, it is plausible that the primary function of assertion would be to facilitate the transmission of information from speakers to hearers. But our world is very different, and the communities of agents to which we belong are also very different, from those that I have just imagined. In our world, agents have very imperfect and partial access to information, and very different background beliefs into which information must be accommodated. Moreover, and most importantly, all agents in our world are prone to form mistaken beliefs about very many different kinds of things (including, often, things that it is important not to be mistaken about). In our world, except in special circumstances - such as, perhaps, the elementary school classroom - transmission of information from speaker to hearer is not in any way a straightforward matter. However, having a practice in which agents put forward their beliefs for comparison with the beliefs of other agents is a plausible mechanism for improving the beliefs of all who engage in the practice. What assertion makes possible is debate, criticism, consideration of alternative perspectives that one would not have otherwise considered, and the like: and it makes this possible because it has the primary norm that one ought not to assert that which one does not believe. (Debate, criticism, and so forth could not lead to improvement in belief unless people actually put forward their beliefs for debate, criticism, and so forth. Wholesale departure from the norm that one ought only to assert that which one believes would undermine the role that the practice of assertion has in improving agents' systems of belief. ) If the claims that I have made in the previous paragraph are plausible then we do in fact have a reason for denying that it is a norm of assertion that one ought to assert only that which is true. True enough, there is a sense in which one's assertions go better for one if one non-accidentally asserts that which is true. But one's non-accidentally asserting that which is

Copyrighted

Material

238

Graham Oppy

true is not an indication that one is more closely tuned to the primary function of assertion than one who fails to non-accidentally assert that which is true. Insofar as we merely pay attention to the (primary) function of assertion, one does all that is required of one if one's assertions are expressions of one's beliefs. Of course, if one's beliefs are substandard, then there is a derivative sense in which one's assertions will be also: but this does not mean that one's assertions are failures qua assertions. The position that I have been outlining in the preceding paragraphs may well seem absurd. After all, not only have I denied that it is a norm of assertion that one should not assert that which is not true, but I have also (implicitly) denied that it is a norm of assertion that one ought not to assert that for which one has insufficient warrant. Moreover, I have not motivated these denials by focusing exclusively on the first-person perspective (even though, of course, one cannot now discriminate between those of one's present beliefs that are true, those that are false, those that are sufficiently warranted, and those that are not sufficiently warranted: if I believe that p, then I take it that it is true that p. and I also take it that I am warranted in believing that p). However, as I have already noted, I have insisted on a distinction between the claim that it is in some sense better that one's assertions are true and possessed of sufficient warrant and the claim that one ought not to assert that which is false or that for which one has insufficient warrant. 7 Provided that one conforms to the genuine norm of assertion - namely, that one ought not to assert that which one does not believe - then the sense in which one is better off if one's assertions are true and possessed of sufficient warrant is precisely that one's beliefs are true and possessed of sufficient warrant (and there is arguably no mystery about why this is good for one).

4 Meeting the Gricean constraint 1 Against the claim that the sole constitutive individuating norm of assertion is that one ought not to assert that which is not true Williamson (2000: 244) objects that assertion is not the only speech act to aim at truth. Consider conjecturing, for instance. While it is somehow good to conjecture the true and bad to conjecture the false, conjecturing that p is not a way of asserting that p. Given that conjecturing is not asserting, this does seem to provide a reason for rejecting the simple theory according to which the sole constitutive individuating norm of assertion is that one ought not to assert that which is not true. But there is no similar objection to the claim that the sole constitutive individuating norm of assertion is that one ought not to assert that which one does not believe. It does seem plausible to claim that assertion is the only speech act to have the expression of belief as its proper end. When one conjectures - or guesses, or opines, or suggests, etc. - one is not giving expression to that which one believes (and so examination of this consideration suggests that the simple theory which takes belief to be the

Copyrighted

Material

Norms of assertion

239

sole individuating constitutive norm of assertion does track the intuitive conception of assertion). 2 One of the objections to his own preferred account of the norms of assertion that Williamson (2000: 255ff. ) considers is that it is often reasonable to believe that p in circumstances in which one knows that one does not know that p; from which it follows, on Williamson's preferred account, that the sole constitutive individuating norm of assertion criticizes one for doing what it is reasonable for one to do. Williamson responds to this claim by drawing a distinction between that which it is reasonable for one to do and that which it is permissible for one to do: in his view, one may reasonably assert that p in circumstances in which one ought not to assert that p. More generally, he suggests that we should think of "the knowledge rule" as giving the condition on which a speaker has the authority to make an assertion. In his view, "the distinction between having a warrant to assert that p and reasonably believing oneself to have such a warrant becomes a special case of the distinction between having the authority to do something and reasonably believing oneself to have that authority". Moreover, when faced with the suggestion that "the knowledge account" seems to imply that speakers should always be at great pains to verify a proposition before asserting it (since the one constitutive individuating norm of assertion is that one ought not assert that which one does not know), Williamson responds that "when assertions come cheap, it is not because the knowledge rule is no longer in force, but because violations of the rule have ceased to matter so much" and goes on to insist that "when we are relaxed in applying the rule, we feel entitled to assert that p whenever we are not confident that we do not know that p. We still try to obey the knowledge rule, but we do not try very hard" (Williamson 2000: 257). The suggestion that one is entitled to assert that p whenever one is not confident that one does not know that p is. I think, in the same ballpark as the suggestion that one is entitled to assert that p whenever one believes that p. 8 Setting aside worries about the move from proscription to entitlement - it is, after all, one thing to suggest that one ought not to assert that p if one does not believe that p, and quite another to suggest that one is entitled to assert that p whenever one believes that p - we might plausibly conclude that Williamson recognizes or, at any rate, has the resources to recognize - a category of "cheap assertion" for which the one constitutive individuating norm is that one ought not to assert that p if one does not believe that p. Given that the view defended in the previous section has the resources to recognize categories of "strict assertion" - for example, a category of assertion that is governed by the norm that one ought not to assert that p if one does not know that p - one might be given to wonder whether there is really a substantive debate to be engaged in here. Perhaps there is something to be said in favor of the view which takes assertion to be governed by a fairly undemanding constitutive individuating norm, and then allows that there are specialized categories of assertion that

Copyrighted

Material

240

Graham Oppy

are required to conform to more demanding norms (e. g. knowledgeable assertion, authoritative assertion, carefully considered assertion, and the like). If a constitutive individuating norm is to govern every assertion, then it has to make sense that every act of assertion is criticizable in light of that norm. This requirement will be satisfied if the constitutive individuating norm for assertion is relatively undemanding: knowledgeable assertion, authoritative assertion, carefully considered assertion, and the like are all property criticizable in light of the norm that one ought not to assert that which one does not believe. On the other hand, if assertion is taken to be governed by a very demanding norm, then there is at least some reason to say that categories such as that of "cheap assertion" are not really species of assertion at all: it doesn't really make sense to criticize "cheap assertions" for failing to live up to the standard that one ought not to assert that which one does not know. Given that assertion is governed by a very demanding norm, "cheap assertion" is no more assertion than plaster ducks are ducks. Even setting this kind of consideration aside, I do not think that it is plausible to suppose that there is no substantive issue between those who favor undemanding constitutive individuating norms and those who favor very demanding constitutive individuating norms. For there are constraints that must be met by any candidate constitutive individuating norm for assertion. In particular, there is the constraint that is suggested by the remark of Grice's that I cited in the opening paragraphs of this paper: a constitutive individuating norm for assertion must be a norm which satisfies the condition that people should he expected to have an interest in participating in conversational exchanges of this kind only on the assumption that they are conducted in general accordance with this norm. If it is plausible to suggest that a less demanding putative constitutive individuating norm meets this constraint whereas a more demanding putative constitutive individuating norm does not, then that will be a strong reason for supposing that it is actually the less demanding putative constitutive individuating norm that actually governs practice. Now, I have, in effect, already given my reason for thinking that only the less demanding putative constitutive individuating norm meets the Gricean constraint. It is, of course, true that there are - or, at any rate, could be circumstances in which people only have an interest in participating in conversational exchanges that are conducted in accordance with the constraint that one assert that p only if one knows that p. However, in general, it is not true that assertoric exchanges are such that people are interested in participating in them only on the assumption that they are conducted in general accordance with this norm. Or. at any rate, so I contended in the previous section of this paper in which I argued that, in fact, assertoric exchanges are such that people have an interest in participating in them even if they are only conducted in general accordance with the norm that one ought not to assert that which one does not believe. Given the interest that people can be expected to have in debate, criticism, consideration of

Copyrighted

Material

Norms of assertion

241

alternative perspectives, and the like, one would expect people to have an interest in participating in assertoric exchanges whose sole constitutive individuating norm is that one ought not to assert that which one does not believe. The Gricean desideratum supports the view that 1 have developed (and counts against Grice's own claims about the further norms of quality that he postulates). 5 Consequences for a dispute about deflationism Wright (1992: 12) claims that considerations about the norms of assertion can be used to show that "deflationism about truth is an inherently unstable position. " His argument for this claim has two key planks: first, he defends the view that deflationism about truth is committed to the claim "that 'T' and 'warrantedly assertible, ' coincide in positive normative force" (p. 12); second, he argues that deflationism about truth is committed to the claim that "'T' and 'is warrantedly assertible' have to be regarded as registering distinct norms - distinct in the precise sense that although aiming at one is, necessarily, aiming at the other, success in the one aim need not be success in the other" (p. 23) From these two key supporting arguments, Wright concludes that "because deflationism, in holding that, modulo a flourish or two, the truth predicate is merely a device of endorsement of assertions, is thereby committed to the idea that warranted assertibility is the only norm operative over assertoric discourse, the finding is, as advertised, that deflationism about truth is an inherently unstable position" (p. 32). I do not propose to argue with the second of the two key claims that Wright defends: I agree that even minimalists about truth should agree that " . . . is true" and " . . . is warrantedly assertible" are not guaranteed to be co-extensive. However, I do not think that even minimalists about truth should agree that truth and warranted assertion provide the same norm for assertion; and nor do I think that even minimalists about truth should agree that truth and warranted assertion provide the only norm that is operative over assertoric discourse. Of course, given the arguments of the previous section of this paper, it is plain that / think that neither truth nor warranted assertion provides a norm for assertion; but at least one of the issues now before us is whether minimalists can endorse the reasons that I have given in support of the claim that neither truth nor warranted assertion is a norm of assertion. Let me begin with a rather extended quotation in which Wright (1992) gives his reasons for claiming that deflationists must concede that truth and warranted assertibility "coincide in positive normative force": Consider the practice of the sincere and literal use of the sentences in the range of the T-predicate. In order for these sentences to be determinate in content at all, there has to be a distinction, respected for the most part by participants in the practice, between proper and improper

Copyrighted

Material

242

Graham Oppy use of them. And since they are sentences with assertoric content, that will be a distinction between cases where their assertion is justified and cases where it is not. It follows that a norm, or complex of norms, of warranted assertibility will hold sway, both prescriptively and descriptively, over sincere and literal use of the sentences to which the T-predicate applies: prescriptively, because to have reason to think that a sentence is warrantedly assertible is, trivially, to have (defeasible) reason to assert it. or endorse its assertion - the 'moves' distinctive of assertoric linguistic practice: descriptively because (or so it seems plausible) unless participants in the practice for the most part try to respect the norms of warranted assertion which govern it, it is not clear in what the fact could consist that its ingredient sentences have the content which they do. But now, given the explanatory biconditional link effected by the Disquotational Schema between the claim that a sentence is T and its proper assertoric use. it follows that T ' is likewise, both prescriptively a n d . . . descriptively, a predicate which is normative of assertoric practice. T ' is prescriptively normative, because any reason to think that a sentence is T may be transferred, across the biconditional, into reason to make or allow the assertoric move which it expresses. And T ' is descriptively normative in the sense that the practices of those for whom warranted assertibility is a descriptive norm are exactly as they would be if they consciously selected the assertoric moves which they were prepared to make or allow in the light of whether or not the sentences involved were T. We already noted the plausible thought that a distinction between warranted and unwarranted assertion must be respected to a substantial extent by actual assertoric practice if assertions are to be determinate in content; accordingly, the biconditional dependence effected by the DS between predication of 'T' and warranted assertion ensures that, to that substantial extent, any actual assertoric practice will be just as it would be if T were a self-conscious goal. In fact, though, the conclusion we ought to draw is stronger than the claim merely that 'T' is normative of any assertoric practice. Say that two predicates coincide in (positive) normative force with respect to a practice just in case each is normative within the practice and reason to suppose that either predicate characterises a move is reason to suppose that the other characterises it too. Then what we may conclude is that 'T' and 'warrantedly a s s e r t i b l e ' . . . coincide in (positive) normative force. For reason to regard a sentence as warrantedly assertible is, naturally, reason to endorse the assertion which it may be used to effect, and conversely: and reason to endorse an assertion is, by the DS, reason to regard the sentence expressing it as T, and conversely. So reason to regard a sentence as T is reason to regard it as warrantedly assertible, and conversely. (Wright 1992: 18ff. )

Copyrighted

Material

Norms of assertion

243

Suppose minimalists agree - as I think they should - that there is good reason to hold that there is at least one constitutive individuating norm of assertion. Wright claims that this agreement alone forces minimalists to concede that truth is both descriptively and prescriptively normative for assertion, since minimalists accept all instances of the schema "It is true that p iff p. " There are several reasons why this claim of Wright's just seems wrong. 1 It is true - given the schema - that I shall take myself to have warrant to assert that p just in case I take myself to have warrant to assert that it is true that p. However - as we have already seen in our discussion of Williamson's views - there is a distinction to be drawn between that which is it reasonable for one to do (given the norms that govern a particular practice) and that which it is permissible for one to do (given the norms that govern that practice). Whether I am warranted in asserting that p is not the same question as whether I (reasonably) take myself to be warranted in asserting that p. Even if the T-schema does effect a plausible connection between (reasonably) taking oneself to be warranted in asserting that p and (reasonably) taking oneself to be warranted in asserting that it is true that p, there is no sense in which this shows that truth coincides with warranted assertion in positive normative force. 2 The established equivalence between "p" and "it is true that p " surely does nothing at all towards establishing that it is a norm of assertion that one ought not to assert that which is not true. Perhaps the easiest way to see this is to suppose, for the sake of argument, that the sole constitutive individuating norm of assertion is that one ought not to assert that which one does not believe, and then add to this the minimalist claim that the content of the truth predicate is entirely fixed by the disquotational schema. There is just no way of deriving the claim that one ought not to assert that which is not true from the disquotational schema together with the claim that one ought not to assert that which one does not believe. 3 Wright (1992: 24) claims that truth is a prescriptive norm of assertion because "any reason to think that a sentence is T may be transferred, across the biconditional, into reason to make or allow the assertoric move which it expresses. " But - as Price (1998) points out - there is something seriously wrong here: that a sentential operator S figures in a true biconditional of the form "Sp iff p " is plainly not sufficient to establish that there is a norm of assertion to the effect that one ought not to assert that p unless Sp. (Moreover, this same point applies even if the biconditional is necessary and a priori. ) While it is true that any reason to suppose that p is reason to suppose that it is true that p (and vice versa), this does nothing at all towards showing that it is a norm of assertion that one ought not to assert that p unless it is true that p.

Copyrighted

Material

244

Graham Oppy

For these reasons - and others - it seems that we are justified in rejecting Wright's claim that deflationists are committed to the claim that truth and warranted assertibility coincide in positive normative force. However, I do not think that the difficulties for Wright's view end here; for the claim that deflationists are committed to holding that warranted assertibility is the only norm operative over assertoric discourse also deserves to be treated with suspicion. Even if we allow, with Wright, that deflationism is (more or less) committed to the claim that "the truth predicate is merely a device of endorsement of assertions, " it remains open to us to insist that deflationists can allow that assertion is governed by the various norms of quality that Grice proposes. (Could Wright respond by saying that "the" norm of warranted assertibility includes proscription of assertion of that which one does not believe? I don't think so; that move would take us back to the confusion between having warrant and reasonably taking oneself to have warrant. ) So far, then, I have agreed with Horwich (1998) and Price (1998) that Wright fails to show that considerations concerning the norms of assertion establish that deflationism about truth is an inherently unstable position. However, even if 1 am right in supposing that Wright's argument fails, it would (of course) be a serious mistake to suppose that we are now in a position to conclude that there are no considerations concerning normativity that escape the grasp of the deflationist. For all that has been argued thus far, it may yet turn out that there are normative features of the concept of truth that the deflationist is simply unable to capture. Price (1998) argues for just this conclusion. He claims that it is a norm of assertion that one ought not to assert that p unless it is true that p, and that this is a fundamental feature of the concept of truth that is not captured by the deflationary theory. However, he goes on to insist that there is a sense in which the spirit of the deflationary theory is correct; for the function of the truth predicate is properly given a "minimalist" explanation which shares the anti-metaphysical ambitions of the deflationary theory. I think that there is at least some reason to claim that Price is wrong on all three of the matters just mentioned. First, 1 have already given reason for thinking that it is not a norm of assertion that one ought not to assert that p unless it is true that p. If this is right, then Price's argument falls over at the beginning. Of course, if it were a norm of belief that one ought not to believe that p unless it is true that p. then the same set of issues would arise. (Indeed, any norm of the form "one ought not to is structured such that •

3 is a non-empty set of

states: E is a reflexive and transitive relation (a pre-ordering) on 5R ç E is reflexive and quasi-convergent w.r.t E such that if i R j and i' R j then there exists some point h such that j E h and j E h and i 5ft h: and • V is a function that maps atomic propositions p, q. . . . into the powerset of 3 such that for i, j € 3: if i e IEj(p)and i' E j then j e V(p). The basic semantic notion of d-verifiability is then defined recursively:

• •

• • •

• •

p is d-verifiable at i (in M) M =, p iff i E V(p); M = (A & B) iff M A and M =, B: M or M (A V B) iff M M h B: M I=i, (A B) iff for every j such that iEj: if M =A then M (=, B: M = ' A iff there is no j such that iEj and M =,- A.

The logical constants thus receive an intuitionistic interpretation in the clauses for d-verifiability. Finally, truth is defined on the basis of d-verifiability:

Copyrighted

Material

On the Peircean

•

conception

of truth

335

A is true at i (in M) iff there is a point j such that i R j and M =¡A

If truth is instead defined relative to R* the has to be strengthened to strict convergence j, j € 5 if R j and i R then, for some h; e convergence in the case of R is necessary u n d e r modus

condition of quasi-convergence that states that for any points i, 3 : iRh, R h and j 5ft h. Strict to ensure that truth is closed

ponens.

I contend that Rabinowicz"s definition presents a model for a concept of truth that is susceptible to epistemic concerns. However, Rabinowicz's definition of intuitionistic truth is clearly tailored for the needs of intuitionistic logic, and the definition is indeed adequate, i.e. sound and complete, for intuitionistic logic, in the sense that a proposition A is true at every point in any Rabinowicz-model M = < , E, R> if and only if A is a theorem of intuitionistic (prepositional) logic. 33 How does this affect a pragmatistic conception of truth? The epistemology underlying intuitionism obviously differs in important aspects from the central ideas of pragmatism. Observe, for example, that it is assumed that the information that is employed to verify a proposition (at some point i) is always supposed to be correct, so that the information states are assumed to be error-free and the contraction of information states need not be considered. As a consequence, d-verifiability (at point i) implies truth at i, and d-verifiability is cumulative along f - p a t h s (and hence along R-paths): no information is false and no information is lost in the course of inquiry. To what extent this is or is not plausible for mathematical reasoning is of no concern of ours here, but it certainly does not conform to any sensible account of rationally controlled knowledge acquisition in empirical domains. Pragmatism, in particular, explicitly acknowledges the possibility of error as part of our cognitive lot. For the purposes of a pragmatic conception of truth, epistemic situations thus should not be supposed to be necessarily error-free, and therefore information is not simply additive. Unless any risk of importing error in the transition from one epistemic state to another is prevented, the possibility of contracting an informational state should be an option. And while Rabinowiczs account admits that the set of true propositions expands, at least when the definition of truth is based on the relation R instead of its ancestral R*, it seems preferable for a pragmatic conception of truth to have it static and time-invariant. But more importantly, the peculiarities of intuitionism eliminate any normative component in the concept of truth. Instead, they allow for a specific ordering of the epistemic states and the set of true propositions will expand along this ordering. The important point is, again, the absence of any possibility of genuine error. If no information will ever be erroneous, then successive epistemic states are supersets of their predecessors and the epistemic states are therefore ordered linearly by the subset relation. Due to the risk of error, however, the subset relation between epistemic states is, on the pragmatistic reading, not decisive for questions concerning their truth.

Copyrighted

Material

336

Ulrich

Metschl

What is crucial is rather the quality of an epistemic state: how it responds to the challenge of evidence and argument. For remember that a belief is true just in case it cannot be improved upon. And that seems equivalent to saying that a belief is true if it would be held in the best informational state that were ever available. But what does "available" mean in this context? And how are epistemic states to be evaluated at all? Any proposal has to observe a number of points. According to the pragmatistic account of belief and inquiry, information states are not necessarily reliable and, in general, call for improvement: information may turn out to be erroneous and contraction may become mandatory, just as information may be incomplete and expansion may be desirable. 34 For a formal definition of truth along the lines of Rabinowicz's definition of intuitionistic truth this would suggest a different conception of verification relative to the points in a model: verification cannot be cumulative along the paths of epistemic accessibility, and for a contingent proposition A neither A nor - A may be contained in some epistemic situation i'. In addition, it should be permitted that epistemic states are not only "underdetermined" but also "overdetermined," in the sense of "verifying" both A and A, for some proposition (and, hence, for any proposition A if classical logic is adopted). A suitable logic can indeed be found. Nuel Belnap's four-valued logic for "tautological entailment" is appropriate as a conception of informational states that may be incomplete as well as "overcomplete," i.e. inconsistent. 3 5 It is in fact motivated by a concern for such situations, modelled, e.g., by a computer that accepts inconsistent input. Belnap's logic results from equipping four truth values Positive, Negative, None, Both with two lattice structures simultaneously. 36 These structures are induced, respectively, by an "informational" pre-ordering with None as bottom element. Both as top element and Positive and Negative incomparably in between, and the "logical" pre-ordering with Negative as bottom element, Positive as top element, and None and Both incomparably in between. Intuitively, the first structure is an ordering from less to more information, while the second structure represents an ordering from falsehood to truth, in the sense of reducing falsehood and/or increasing truth. It is convenient to think of the four truth values as elements of the power set of the set [T. X] of classical truth values, so that the subset relation yields already the informational ordering. A valuation, then, is a function which marks each atomic sentence with one of the four truth values, and any valuation may be extended, respecting appropriate constraints in connection with the lattice structures, to a truth value assignment for all sentences (or propositions) of a (first-order) language. 37 For our purposes, further details can be safely ignored, and we will simply assume that we have, within this approach, a definition for "A is verified relative to the valuation V" - M =¡ A~. (for the appropriate language) which is supposed to hold precisely when A is marked Positive relative to V and which corresponds, roughly, to Rabinowicz's relation of verification in a model M |=i A. We thus think of a proposition's A being marked Positive

Copyrighted

Material

On the Peircean conception of truth

337

(relative to some valuation V) as a verification of A by the epistemic situation that is represented by the valuation V. Will this take us to a plausible concept of truth with a distinctively pragmatic ring? Note, in particular, that according to the remarks above a proposition A might be verified in state i although A is not true, because the epistemic state verifying A is not optimal. Could we then invoke deontic elements to define pragmatistic truth in a way similar to the definition of intuitionistic truth above? Consider a standard axiomatization for orthodox, unconditional deontic logic as it results from adding to the prepositional tautologies the following axiom schemata: •

(O1,) O(A)

•

( 0 2 ) O (A

.0(. B)

A)

( O (A)

0(B))

and the rules of inference •

(RPL) A1, ..., An f- B if B can be deduced from A1,, ... , A„ by classical prepositional logic

•

( R O ) A - O (A)

Despite the fact that these principles invite an epistemic, or rather alethic reading, the remarks in the preceding paragraph made clear that we should withstand the temptation to impose a naive deontic reading on the pragmatistic conception of truth. The pragmatistic view is not committed, and even is hostile to the assumption of a class of epistemically ideal situations in an agent-independent, Platonistic sense. The deontic logic sketched above, however, represents those models in which every possible world has a non-empty set of deontic alternatives. The guiding idea thereby is that it is obligatory in world i to bring about a certain state of affairs, corresponding to a certain proposition, just in case this proposition is true in all those worlds that are the deontic alternatives to i and which, thus, are understood as the morally perfect worlds relative to i. While the formal characterization of this logic - namely that in its models there exists for every world i a nonempty class of deontic alternatives - is indeed too weak to shed any light on a pragmatistic concept of truth, its intuitive motivation, namely that obligatory propositions are those that hold in the morally perfect worlds, seems to be too strong. The idea that there is a uniform set of epistemically ideal states, independent of any given situation, is entirely against the grain of pragmatism. What is needed, rather, is a preference relation on epistemic situations, or a choice function that selects the "optimal" alternatives from a non-empty set of options for any epistemic state. In general, choice functions are an extremely versatile tool for modelling rational deliberation. Their role is, more specifically, to select, for each nonempty set of available alternatives, a (non-empty) set of best options for the purposes at issue and under appropriate constraints. So abstractly conceived,

Copyrighted

Material

338

Ulrich

Metschl

choice functions are clearly compatible, and even seem congenial to Misak's formulation of pragmatistic truth as the selection of those beliefs which are "best" were inquiry carried as far as it would prove fruitful. 3 8 In a formal setting, Misak's proposal would amount to something like the following: a belief is true in an epistemic state i if it is verified or confirmed by those paths of inquiry or research developments (originating in i) that are optimal by the epistemic standards (in part evolving in the course of inquiry) and thus are selected by an epistemic choice function. Let us say that a path of inquiry is summarized by the epistemic state to which it leads up, and let us say that the relation of epistemic accessibility represents possible epistemic developments. Then Misak's formulation of the concept of pragmatistic truth seems to be captured in the following: A is true (at i) if and only if A is verified in those points j summarizing paths of inquiry beginning at i that are elements of the set of epistemically accessible points that is selected by the epistemic choice function. If it is assumed, in addition, that the epistemic choice function corresponds to an epistemic preference relation ≤, (relative to point i), and that models are based on the valuations of Belnap's four-valued logic such that M =, A means that A is assigned the value Positive at i (in M), we arrive at: A is true at i (in M) if and only if: if for all j such that iEj and h ≤¡ j for all h E 3 : M =j A i.e. if A is verified at all those points jE-accessible from i that are, among the E-accessiblepoints, (weakly) epistemically preferred from i's perspective, where a model M = < G, E. { ≤ Ee V > consists of a non-empty set 9 of epistemic situations, an epistemic accessibility relation on 3 (whose properties are here left open), a family of epistemic preference relations ≤¡ on 9 for each i e 9 and a valuation V that assigns to any atomic sentence one of the four truth values Positive, Negative, Both. None and which can be extended to a truth value assignment for all sentences of a prepositional language. The suggested formulation, although not spelled out in all its details here, is in spirit at least close to the pragmatistic conception of truth as it is defended by Misak. Observe, for example, what happens when a proposition A is not uniquely verified in all the epistemically undominated points accessible from i: then A is not true at i. This is in agreement, I contend, with the pragmatistic view that when A is not unanimously held in the "best" doxastic situations that can be reached from i, the dispute about A is not settled and doubts concerning A still prevail. An epistemic choice function or an epistemic preference relation can even integrate the role of science as the one method that most likely leads to true beliefs, for it might be claimed that it is scientific evidence on which the choice of epistemic developments is to be based.

Copyrighted

Material

On the Peircean conception of truth

339

What must remain controversial, however, is that any formulation in terms of epistemic choice functions or preference relations makes pragmatistic truth a matter of rational decision-making, and it is not clear whether Peirce or Misak would want to go that far. It is true that, to some extent, scientific reasoning can be fruitfully explained as a case of cognitive decision making, as the Bayesian paradigm amply testifies. 39 But for the purposes of an analysis of the concept of truth, this is compelling only to the extent that the specific constraints on rational choice with respect to epistemic situations are elaborated. Neither Peirce nor Misak are explicit about the way choices should be made between possible epistemic developments. 4 0 It is, in fact, doubtful that a concept of truth can be elucidated merely on the basis of the formal ingredients of theories of rational decision making, even when these are trimmed according to the needs of epistemic developments. Present-day accounts of rational choice conceive of decision making as something that is rooted exclusively in the subjective opinions of supposedly rational and autonomous individuals. Any evaluation of the available options is based on subjective preferences, either in an ordinal or in a cardinal representation, and on a subjective assessment of the situation that determines - in combination with each of the options - the outcomes, and is thus relevant to the decision (or action) at issue. 41 The implicit conception of preferences thereby is that these are the expression of individual tastes and values. But such a conception seems clearly inapplicable to an evaluation of information states in their relation to truth. Peirce, for example, thought of science as a social art, and the epistemic situations from which the truth is to emerge therefore reflect not individual mental states but rather public or collective epistemic situations. In the case of pragmatistic truth, any account that is confined to purely subjective constituents thus seems bound to fail. Therefore, Misak's conception of true beliefs as the "best" beliefs when all relevant evidence is taken into account must embrace substantial ingredients that say more on truth than the merely formal components of the theories of rational deliberations for individual agents are able to reveal. The "best" beliefs are, to some degree, objectively best, at least to that degree that allows truth to be independent from the vagaries of individual reasoning. But which conditions on the rationality of choices concerning possible epistemic developments are sufficient to guarantee an access to the truth? Consider, for example, two points i, j in a model M = < E, { ≤ i}i 6 , V > . Nothing so far prevents that the set of propositions true at i is different from those true at j. Arguably, we should allow that there are points from which the truth cannot be reached. But a notion of truth that admits that different propositions are true relative to different epistemic states, so that a consensus on factual issues seems precluded, admits that truth collapses into Peirce's thinking to be true. A conception that equates true beliefs with those that are "best" according to the relevant evidence and "cannot be improved upon" for the settlement of controversial issues, remains incomplete unless we are told by

Copyrighted

Material

340

Ulrich

Metschl

what standards the evaluation of beliefs as better or worse for the settlement of doubts and controversy is to be done. A pragmatistic concept of truth, that is supposed to be sensitive to evidence and inquiry, is backed by substantial assumptions that are to warrant that truth is markedly different from subjective opinion, and which so far await further philosophical elucidation. A formal definition of pragmatic truth might turn on a measure of informational content. The informational content of a proposition A, relative to some point of background information i, depends on A's "informativeness," i.e. how much information does it provide with respect to the matter at issue, as well as on its truth-likeliness. The trade-off inherent to the informational content is motivated by our desire to have at once reliable (i.e. true) and specific information. It is, however, an open question how this trade-off can be characterized in a sufficiently general manner and it is indeed not unlikely that different situations are characterized by different informational needs that might correspond to different measures of informational content. So far, at any rate, no standardized measure of informational content that would have gained uniform acceptance is known. 4 2 Be that as it may, this is certainly not the project upon which Misak has embarked. The context within which her description of true beliefs as those which cannot be improved upon is developed, reveals that her conception of pragmatistic truth puts faith in the consensual force of open, unrestrained dialogue. In that she comes close to an almost "Habermasian" belief in "communicative rationality" that is supposed to transgress the narrow confines of "strategic rationality," which in turn is ill-suited or even useless for the "legitimation" of practical or ethical norms or the verification in matters of fact. The belief in the rectifying and harmonizing power of dialogue is supposed to explain the fundamental political role that truth plays in Misak's account. In times when liberal virtues are more and more endangered and fundamentalism and bigotry are resurgent, Misak's endeavor deserves approval and support. All the more it is important to demonstrate the feasibility of an account of pragmatistic truth that integrates concerns as diverse as those envisaged by Misak. Perhaps, then, the idea of true beliefs as "best" beliefs in the light of the relevant evidence means something like the following: a belief on which we cannot improve results when an issue has been thoroughly considered from all angles that can contribute to an understanding of the question at issue. If all relevant perspectives are taken into account, then we may hope that the gap between the public opinion and what can be taken to be the reality is closed. This, however, hinges on a conception of reality that is similar to Peirce's own conception of reality according to which the object in the opinion ultimately agreed upon is "the real." But Peirce's position is no less optimistic than the one of Misak or Habermas. It rests on the hope that consensus, at least in vital matters, is not impossible. Levi has identified this hope as "messianic realism." 4 3 Messianic realism is committed to the view that "the ultimate aim of inquiry is to realize the true story of the world at

Copyrighted

Material

On the Peircean conception of truth

341

the end of days, regardless of what happened in between" (Levi 1991: 162). Levi points out that the proponents of this view owe us an explanation why avoidance of importing error is an aim that regulates each of the successive steps in inquiry. "Secular realism," in contrast, is concerned only with the avoidance of error for that change in the epistemic situation that is under consideration in the immediate presence of doubt. While truth as a regulative ideal plays a pivotal role in combination with messianic realism, its role in combination with secular realism is much more modest. But the secular realist is, unlike the messianic realist, not committed to the view that there is one uniform standard of assessing the epistemic quality of information states. The assumption that there is such a standard may afford us a pragmatistic definition of truth, although one that amounts to messianic realism. It is, however, clearly at odds with the fundamental principles of pragmatism which state that truth will not be independent of human practice and experience. Messianic realism, although implicit in some of Peirce's remarks, is hardly coherent with his pragmatism unless we are told how to get to "the true story about the world at the end of all days." The conclusion can only be that there is an unresolved tension in the pragmatistic conception of truth as it is identified by Misak. On the one hand, this conception sets out as a promising candidate for an account of the conceptual role of the notion of truth. This was found wanting insofar as alternative accounts of truth were seen to be largely unable to explain the role that the concept of truth plays in various forms of discourse, and in particular to those which are sensitive to issues of justification. On the other hand, however, the requirements for ending at a sufficiently precise account of truth were seen to be such that the pragmatistic conception cannot live up to them without becoming unfaithful to its own principles. Truth, as regulative ideal, requires a standard for objectivity that is beyond the reach of the pragmatist. The confidence, on which the hope is founded that the epistemically ideal situations which warrant truth are ultimately accessible by fallible human beings, is not justified by an argument. Rather, it is an article of faith. 4 4 Maybe then, truth remains a regulative ideal that is as elusive as all true ideals ever have been.

Notes 1 Thanks are due to William Hanson and Dirk Greimann for comments on earlier versions of this paper. 2 See, for example, Davidson (1996). 3 This, arguably, is the main thrust of "deflationary" or "minimal" theories of truth, cf. Horwich (1990). 4 An impressive attempt to defend a broad conception of truth against its postmodern deriders is advanced in Williams (2002). 5 Cf. Misak (1991) and Misak (2000). 6 For statements where Peirce comes at least close to such a view see his paper "The Scientific Attitude and Fallibilism", in Peirce (1955).

Copyrighted

Material

342

8

9 10 11

12

13

14

15

16

17

18

19

Ulrich

Metschl

7 A point that is emphasized forcefully by Misak. in particular in (2000) and (2004). Incidentally, it is this f act that Misak has tried to convert into a defense for liberal democracy as the only political order that fosters open dialogue and unrestrained public deliberation, see Misak (2000). See Read (1994: Chapter 1) for a discussion of these points and a motivation for a minimalist account of truth. See Horwich (1990: 44 and 65) for this line of reasoning. It should be emphasized that Horwich. in his remarks on a "so-called pragmatic theory of truth" equates this conception with the view that truth is utility (Horwich 1990: 10). a common, though nonetheless deplorable misunderstanding which has been fostered, however, by some pragmatists. It should be emphasized that this view is definitely not held by Peirce and will not be discussed in the sequel. In point of fact, we hope to rectify this situation somewhat by presenting a more serious conception of pragmatistic truth. For the complaint that the orthodox conception of truth is ill-suited to the phenomenology of a moral or political discourse where it is taken for granted that some attitudes are simply wrong (and others ultimately right), and where reasoned arguments and not acts of intimidation or violence decide about the Tightness of positions, see Misak (2000). A discomfort with the fact-value dichotomy is also at the heart of Putnam (1981). See Dewey (1929): Peirce's papers regularly contain more or less open attacks on a Cartesian-minded epistemology, but see in particular his famous paper "How to Make Our Ideas Clear" reprinted in Peirce (1923). Although it is admirable how some persons manage to make their experiences fit their expectations, this attitude can hardly be taken as the hallmark of rationality. For the question how it can be rational, from an agent's own point of view, to admit evidence that is inconsistent with a given body of belief, see Erik Olsson's discussion of Isaac Levi's epistemology; Olsson (2003). Peirce's f a m o u s four methods of tenacity, of authority, of the a priori, and of scientific investigation, with only the last one passing the test, are set out in his paper "The Fixation of Belief." For an exposition and a discussion of these methods, and in particular the method of science, see Misak (1991: esp. Chapter. 2). This kind of belief change is, according to a distinction that Isaac Levi has urged on us, only one in doxastic performance, in contrast to changes in doxastic commitments, see Levi (1991: 6ff.). As any account of rational belief and rational belief change will clearly be focused on doxastic commitments, shortcomings in doxastic performance will be largely ignored in the sequel. The qualification that p be seriously accepted is meant to take into account circumstances where p is accepted only "for the sake of the argument" or in any other form, if there be such, that deviates from honest and sincere belief such that no commitment to truth is involved. This term is due to Isaac Levi, see, for example Levi (1991). Levi identifies the belief-doubt model in the c o m m o n core of Peirce's and Dewey's pragmatism, and does not confine the label exclusively to Peirce's epistemology. This marks a difference to Popper's falsificationism, for under Popper's position a dispense of further research is never licensed as a rational move. Any "corroboration" of a hypothesis, as much as it would seem to work against the infection of doubt, can hold only tentatively while the obligation to attack a certain hypothesis by exposing it to ever further testing (until it is finally falsified) holds unconditionally. A suspense of research concerning an as yet unfalsified hypothesis may be due to a lack of resources but is. according to Popper's falsificationism, never mandated for scientific reasons.

Copyrighted

Material

On the Peircean

conception

of truth

343

20 See for the following Misak (1991) and (2000). 21 It is worth repeating that Misak does not claim that Peirce ever offered a biconditional like the one stated, or even came close to doing so. She explicitly writes: "Peirce does not want to offer a straightforward biconditional of the sort" [just stated, U.M.]. However, she continues: "But it turns out that he is well on the way to getting himself such an equivalence" Misak (1991: ix). For further remarks in that direction see also Misak (1991: 42f.). 22 Compare this with Oscar Wilde's well-known remark that "if one tells the truth, one is sure, sooner or later, to be found out." Wilde's pragmatistic leanings are, however, somewhat tempered by his belief that "a truth ceases to be true when more than one person believes in it." 23 See Misak (2000: 51). 24 See Misak (2000: 51). 25 For her critique of Wright's concept see Misak (2000: 65ff.). 26 There is, at this point, no attempt made to delineate this class of truths, and the logic to be favored can be left to the reader's discretion. 27 This formulation is to be found in Peirce's paper "The Scientific Attitude and Fallibilism", see Peirce (1955: 54). 28 Respect, in this connection, is meant roughly along the lines of Isaiah Berlin's distinction between contemptuous toleration and skeptical respect, where the latter means a willingness, at times not without reluctance, to listen to the dissenting views of others, while the former means only to refrain from prosecuting dissenters without opening one's mind to their views at all. See Berlin (1959) and, for an illuminating employment of this distinction, Levi (1997). 29 As Dummett has pointed out, this is not without some ambiguity as there may be different standards for proofs. Dummett, in particular, makes a distinction between canonical proofs and (more or less) informal demonstrations, the latter often being indirect in that they only indicate how a statement could be verified without themselves providing this verification; see Dummett (1973). 30 Rabinowicz (1985). 31 Formally a valuation is a function that assigns to every atomic sentence a subset of the set of points with the intended reading that i E V(p) iff p is verified at i. 32 As we will see, E is supposed to be a pre-ordering, i.e. a reflexive and transitive relation on the set of points, an assumption that corresponds to the usual requirement in the semantics for intuitionistic logic that the information states are pre-ordered. 33 In this context, it is convenient to view Rabinowicz's proposal as a refinement of the informational interpretation of intuitionistic logic in the tradition of Grzegorczyk (1964) and Kripke (1965), and Rabinowicz elaborates on the similarities and differences between his own semantics and Kripke's as well as Beth's for intuitionistic logic. For the use of informational interpretations for substructural logics see Wansing (1993). 34 As the theory of belief revision suggests, it is not so much contraction that is of interest but rather revision, where some informational item A is replaced by its opposite and this suggestion is in line with the Peircean conception according to which pure contraction would mean an uncompensated informational loss that must induce doubt where formerly an opinion was settled. It is convenient, however, to understand the revision of some epistemic state K w.r.t. A as the contraction of K by -¡A followed by a subsequent expansion of the contracted state K' by A, a procedure that is known in the belief revision community as Levi-identity. The formal as well as philosophical issues of belief revision are laid out in Gärdenfors (1988) and Levi (1991). 35 Belnap (1976) and Belnap (1977). Note that in systems of relevant logic the rule that from inconsistent premises anything can be concluded is suspended.

Copyrighted

Material

344

Ulrich

Metschl

36 For our exposition we have renamed Belnap's original truth values True, False, None, Both. 37 Recall that a logic, abstractly conceived, is a set of designated objects, the truth values, in combination with some lattice structure such that meets and joins at least for finitely many of the designated objects exist. Meets and joins capture conjunction and disjunction, respectively, and if in addition some kind of negation is to hold, a certain form of complementarity must also exist. 38 For the rationality of decisions it is necessary, however, that choices are made in a coherent manner, and the coherence of choices puts particular constraints on choice functions. Perspicuous formulations of conditions on choice functions that are meant to ensure the coherence of choices are due to A m a r t y a Sen: Let X be a (non-empty) set of alternatives and let S, T be subsets of X. Let C be any choice function: Sen's Property a: if S ⊆ T then: if x E S and x E C(7) then x 6

C(S).

Sen's Property (3: if S C T a n d both x and y are in C(S) then x € C(7) if and only if y E C(T). Sen's in Y then x E C(U Y). While property α defines a kind of " d o w n w a r d " coherence on choices, in that it requires that the set of "best" options remains unaltered when sub-optimal or dominated alternatives are removed from the set from which the choice is to be made, properties (3 and y each define an " u p w a r d " coherence, either by demanding that no discrimination can be made between optimal options when further alternatives come into play or by demanding that an option x that is a m o n g the 'best' ones can still be chosen when other options are considered that are dominated in other choices. An interesting and important result about choice functions states that any choice function that satisfies properties a and y corresponds to a binary preference relation. Further conditions on choice functions impose additional restrictions on the corresponding preference relations. If we think of preference relations as orderings, i.e. as reflexive, transitive and connected relations on the set of alternatives then a well-known result is that a choice function C induces an ordering Rc if and only if C satisfies properties a and |3, where x Rc y if and only if for some S ⊆ X: x E C(S) and y E S. The preference relation corresponding to a choice function which respects the properties a and can, in addition, be shown to conform to a further requirement that is often seen as essential for preference relations, in particular in the context of dynamic choice situations where the set of available options is subject to change. This requirement is known as "contextfree ordering" or "independence of irrelevant alternatives" (the latter name being used for a number of non-equivalent principles, including property and it says, intuitively, that the ranking of any two alternatives does not change when other alternatives are added or removed. It should be obvious how this relates to the properties of choice functions. The results mentioned in the text are proposition 8 and proposition 11 in Sen 1977). 39 Bayesianism, as a doctrine in the philosophy of science, concentrates on the concept of confirmation understood as an increase of the probability of a hypothesis by evidential input. For a detailed exposition of this idea see Howson and U r b a c h (1986), a more critical exposition is found in E a r m a n (1996). 40 This is not to deny that an account of cognitive decision making can be founded on Peirce's epistemology, or at least can take clues thereof. Levi (1991) is out-

Copyrighted

Material

On the Peircean conception of truth

41

42

43 44

345

standing among those who demonstrate how ideas of a pragmatistic epistemology may be applied in an account of cognitive decision making and the theory of belief revision. Strictly speaking, this is not quite correct. The von Neumann-Morgenstern approach is indeed based on objective probabilities. These, however, combine with a subjective utility function in the assessment of expected utilities which still are subjective. The notion of informational value is of central concern in Levi (1991), where a probability-based version is studied in some detail, elaborating on Levi's own earlier attempts; see in particular Levi (1991: 122ff.). Levi (1991: 158ff.). William Hanson has pointed out to me (personal correspondence) that the Bayesian approach holds out some hope for consensus, even when agents start with radically different beliefs (in the form of subjective probability assessments). As a formal result this is known as "merging of opinions" or "washing out of priors," a result that usually is cited as a piece of evidence that the "subjectivity" in subjective probability judgments is unproblematic because these will converge, by up-dating with new information, thus becoming indiscernible form objective probabilities. The problem, however, is that Bayesian up-dating comprises a rather narrow conception of confirmation, blamed by Levi for its "confirmational tenacity." Thus, Bayesianism seems to go against the pragmatistic grain, insofar as it precludes the option of changing epistemic (i.e. confirmational) standards. For a critical assessment of these and other aspects of Bayesianism from a pragmatistic perspective, see Levi (1980).

References Belnap, N. D. (1976) "How a Computer Should Think," in Contemporary Aspects of Philosophy, Proceedings of the Oxford International Symposium, 1975. (1977) "A Useful Four-valued Logic," in J. M. Dunn and G. Epstein (eds) Modern Uses of Multiple-Valued Logic, Dordrecht: D. Reidel. Berlin, I. (1959) "John Stuart Mill and the Ends of Life," in I. Berlin (1969) Four Essays on Liberty, Oxford: Oxford University Press, pp. 173-206. Beth, E. W. (1956) "Semantic Construction of Intuitionistic Logic," Mededelingen der Koninleiijke Nederlandse Akademie van Wetenschappen, Afd. Letterkunde, new series 19, 13: 357-88. Davidson, D. (1996) "The Folly of Trying to Fefine Truth," Journal of Philosophy, 93: 263-78. Dewey, J. (1929) The Quest for Certainty, in J. A. Boydston (ed.) (1969-91) The Collected Works of John Dewey, 1882-1953, vol. 4, Carbondale and Edwardsville, IL: Southern Illinois University Press. Dummett, M. (1973) "The Philosophical Basis of Intuitionistic Logic," in H. E. Rose and J. C. Shepherdson (eds) (1975) Logic Colloquium '73, Amsterdam/ Oxford/New York: North-Holland, pp. 4-40; reprinted in M. Dummett (1978) Truth and other Enigmas', London: Duckworth, pp. 215—47. Earman, J. (1996) Bayes or Bust, 2nd edn, Cambridge, MA: M I T Press. Gärdenfors, P. (1988) Knowledge in Flux. Modeling the Dynamics of Epistemic States, Cambridge, MA: M I T Press. Grzegorczyk, A. (1964) "A Philosophically Plausible Formal Interpretation of Intuitionistic Logic," Indagationes Mathematicae, 26: 596-601. Horwich, P. (1990) Truth, Oxford: Basil Blackwell.

Copyrighted

Material

346

Ulrich

Metschl

Howson, C. and Urbach, P. (1986) Scientific Reasoning. The Bayesian Approach, Chicago/La Salle, IL: Open Court, 2nd edn, 1993. Kripke, S. A. (1965) "Semantical Analysis of Intuitionistic Logic I," in J. Crossley and M. D u m m e t t (eds) Formal Systems and Recursive Functions, Amsterdam: North-Holland. Levi, I. (1980) The Enterprise of Knowledge. An Essay on Knowledge, Credal Probability, and Chance, Cambridge, MA: M I T Press. (1991) The Fixation of Belief and its Undoing, Cambridge: Cambridge University Press. (1997) "The Ethics of Controversy," in I. Levi The Covenant of Reason', Cambridge: Cambridge University Press, pp. 239-54. Misak, C.J. (1991) Truth and the End of Inquiry. A Peircean Account of Truth, Oxford: Clarendon Press. (2000) Truth, Politics, Morality. Pragmatism and Deliberation, London/New York: Routledge. (2004) Truth and the End of Inquiry. A Peircean Account of Truth, Oxford: Oxford University Press. Olsson, E. J. (2003) "Avoiding Epistemic Hell: Levi on Pragmatism and Inconsistency," Synthese, 135: 119-40. Peirce, C. S. (1923) Chance, Love, and Logic. Philosophical Essays, edited and introduced by M. R. Cohen. Lincoln, N E and London: University of Nebraska Press; reprinted 1998. (1955) Philosophical Writings of Peirce, selected and edited by J. Buchler, New York: Dover Publications, 1974. Putnam, H. (1981) Reason, Truth, and History. Cambridge: Cambridge University Press. Rabinowicz, W. (1985) "Intuitionistic Truth," Journal of Philosophical Logic, 14: 191-228. Read, S. (1994) Thinking about Logic, Oxford/New York: Oxford University Press. Sen, A. K. (1977) "Social Choice: A Re-examination," Econometrica, 45: 53-89; reprinted in A. K. Sen (1982) Choice, Welfare and Measurement, Oxford: Basil Blackwell, pp. 158-200. (1982) Choice, Welfare and Measurement. Oxford: Basil Blackwell. Wansing, H. (1993) "Informational Interpretation of Substructural Propositional Logics," Journal of Logic, Language, and Information, 2: 285-308. Williams, B. (2002) Truth and Truthfulness, Princeton, NJ: Princeton University Press.

Copyrighted

Material

Part IV

Truth and propositional meaning

Copyrighted

Material

Copyrighted

Material

16 Meaning and truth-conditions Richard Heck, Jr.

1 Opening Donald Davidson (1984e) famously proposed that what we want from a theory of meaning we may get from a theory of truth. What we want from a theory of meaning for a natural language, that being the case at issue, includes at least an account of the phenomena that give rise to the principle of compositionality: we want an explanation of how the meaning of a complex expression, such as a sentence, depends upon and is determined by the meanings of its syntactic parts. One central advantage of Davidson's proposal, as he saw it, is that we know from Tarski (1933) how to construct a theory of truth for languages with certain sorts of structures, and such a theory does help us to understand how at least some of the semantic properties of complex expressions are determined by those of their parts. If, as Davidson's practice suggests he may have hoped, all sentences of natural language could be assigned logical forms of the variety familiar from mathematical logic, then semantics would all but reduce to syntax: The only outstanding problem would be to uncover the "logical forms" of the sentences of, say, English. That hope does not now seem reasonable. But the methods available for investigating logical form - i.e., syntax - have become so powerful that it is not at all unreasonable to suppose that the syntactic part of the problem is solvable, at least in principle, and Tarski's work can be adapted to many of the resulting structures. The work that is left for the semanticist proper then becomes to solve the problems that syntax and Tarski do not solve. These include, for example, discovering what sorts of semantic values need to be assigned to novel semantic primitives, such as mass terms, and explaining the semantic significance of novel modes of composition, such as [vpV[cpS]].1 It is unfortunate that a proposal with such terrific advantages faces such grave difficulties. Indeed, it is probably fair to say that most philosophers of language, and perhaps even most working semanticists, regard truth-theoretic semantics as a non-starter, for the simple reason that it seems to have the wrong subject-matter. We will not get what we want from a theory of meaning from any theory that does not specify meanings. N o theory that

Copyrighted

Material

350

Richard Heck, Jr.

does not specify what the sentence "snow is white" means, for example, can possibly explain how this sentence's meaning what it does is determined by its parts' meaning what they do. But a theory of truth for English does not specify what even one sentence of English means: any such theory is consistent with infinitely many different suppositions concerning what the sentences of English mean. For example, a theory of truth that proves this T-sentence 2 for "snow is white" (1) "snow is white" is true iff snow is white is consistent with this sentence's meaning anything, so long as it means something true. A monolingual speaker of German who believed that "snow is white" was true iff snow was white could rationally believe as well that "snow is white" meant that grass is green. If "snow is white" did mean that grass is green, then, since snow is white iff grass is green, it would still be true iff snow is white. Since this sort of objection was first forcefully raised by John Foster (1976), it has come to be known as the "Foster problem." I intend here to defend a form of Davidson's proposal against this sort of objection. What most fundamentally distinguishes my view from Davidson's is that I accept a "cognitive conception of understanding". According to me, our actual ability to interpret the utterances of other speakers depends upon our knowing what their utterances, as made on that occasion, mean. I shall not argue for this conception of understanding here - see Heck (2006c) for the argument - but I shall not simply assume it, either. What I shall argue is that no solution to the Foster problem is available unless we adopt the cognitive conception and so that truth-theoretic semantics is committed to it. In particular, I shall argue in Section 2 that Davidson's own proposed solution rests, and rests essentially, upon his metaphysics of the mind, and I, at least, find that metaphysics unpalatable. In Section 3, I shall consider an attempt to resolve the Foster problem based upon a suggestion made by John McDowell in a somewhat different context. I shall argue that it too fails unless we embed it within the cognitive conception, towards which it therefore points, once properly understood. In Section 4, I shall explain and further motivate the cognitive conception and then, in Section 5, drawing on work of James Higginbotham and Ian Rumfitt, I shall argue that the Foster problem can be solved from within the cognitive conception.

2 Radical interpretation as a solution to the Foster problem Davidson addresses versions of the Foster problem in several places. His first suggestion was that a T-sentence cannot, in the relevant sense, "give the meaning" of the sentence it mentions on its left-hand side unless it is a consequence of a compositional truth-theory for the language to which that sentence belongs (Davidson 1984e: 25-6). That plausibly accounts for such outliers as

Copyrighted

Material

Meaning and truth-conditions

351

(2) "snow is white" is true iff grass is green since, as Davidson (1984e: 26, fn. 10) notes, it is hard to see how this sentence could be a (canonical) consequence of a correct compositional truththeory. By itself, however, this amendment is insufficient. As Foster (1976: 13) observed, one should be able easily to reformulate any compositional theory of truth that yields the desirable (1) so that it instead yields: (3) "snow is white" is true iff snow is white and grass is green. If the theory makes use of the axiom (4) "is white" is true of x iff x is white, for example, one need only replace it with (5) "is white" is true of x iff x is white and grass is green, which is true if (4) is. In later work, Davidson insists that T-sentences that purport to be interpretive "are empirical generalizations about speakers, and so must not only be true but also lawlike" (Davidson 1984e: 26, f n . l l ) . This suggestion, too, however, is insufficient, as an observation due to Scott Soames (1988: 187-8) shows. Soames observed that, if we replace (4) with (6) "is white" is true of x iff x is white and arithmetic is incomplete, then the resulting theory delivers (7) "snow is white" is true iff snow is white and arithmetic is incomplete, which is an empirical law if (1) is. Davidson's solution to the Foster problem must therefore be taken to lie elsewhere, namely, where it is usually taken to lie, in his discussion of radical interpretation. Davidson claims that "a T-sentence of an empirical theory of truth can be used to interpret a sentence . . . provided we also know the theory that entails it, and know that it is a theory that meets [certain] formal and empirical criteria" (Davidson 1984b: 139). The main formal criterion is just that the theory of truth in question should be compositional. The empirical criterion is that the theory should be verifiable by a radical interpreter. There are different ways we might understand this proposal. In the passage just quoted, Davidson says that one can use a theory of truth to interpret a language if one has certain sorts of knowledge about that theory. This suggestion is oddly meta-linguistic, however. Fortunately, there is another way to understand Davidson's suggestion, so

Copyrighted

Material

352

Richard Heck, Jr.

that what is required is not that one should know something about a theory of truth but, rather, that one's knowledge of a theory of truth should have a certain provenance. On this interpretation. Davidson is committing himself to the following thesis: (RI) Someone who knew a truth-theory for a language L, and knew that theory on the basis of the sort of evidence that would be available to a radical interpreter, could use T-sentences (canonically) derived from that theory to interpret speakers of L. I shall adopt this interpretation of Davidson's position here. Nothing in my criticism of his solution to the Foster problem will turn upon this choice, however. Surprisingly, Davidson offers no argument for (RI), either in "Radical Interpretation" (1984b) or anywhere else, so far as I know. He does, of course, raise the question whether it is correct. However, Davidson remarks only that his "idea is that what Tarski assumed outright for each T-sentence" - namely, that the sentence used on its right-hand side translates the sentence mentioned on its left-hand side - "can be indirectly elicited by a holistic constraint", namely, "that that totality of T-sentences should . . . optimally fit evidence" available to a radical interpreter. If so - that is, if the holistic constraint really does constitute a workable alternative to Tarski's appeal to translation - then "each T-sentence will in fact yield an acceptable interpretation" (Davidson 1984b: 139). But Davidson never argues that the "holistic constraint" actually is adequate for this purpose: In general, that is to say, Davidson nowhere argues that, if a theory of truth does optimally fit the evidence available to a radical interpreter, then it can be used to interpret; in particular, he nowhere argues that no theory of truth issuing in such T-sentences as (7) could meet his "holistic constraint." Radical interpretation is modeled on radical translation. We are to imagine someone traveling abroad to a country whose language is utterly alien to her and attempting to construct and confirm a theory of truth for that language on the basis of her observations of its native speakers. Why should this thought experiment be of any interest to a philosopher? 3 Davidson is, of course, free to engage in whatever speculation he wants, and if he wants to know what sorts of theories of truth could be confirmed by a radical interpreter, he is free to ask that question. But plainly, Davidson thinks the answer to this question matters: he thinks one can draw important conclusions from it. Davidson seems to think, for example, that one could establish a thesis with the same import as Quine's thesis of the indeterminacy of translation by showing that the evidence that is available to a radical interpreter will not, in general, support a unique theory of truth for a given language. But that makes it all the more important to ask why this thought experiment should be supposed to illuminate questions about meaning.

Copyrighted

Material

Meaning and truth-conditions

353

One answer might begin with Davidson's claim that "[a]ll understanding of the speech of others involves radical interpretation" (Davidson 1984b: 125). Surely, however, we do not actually confront other speakers from a standpoint of total ignorance about what their words mean and what their beliefs are: there is a clear sense, then, in which we are not all radical interpreters. On the other hand, however, in his very first paragraph, Davidson (1984b) explicitly disclaims any interest in the psychology of language-use, so he cannot have intended radical interpretation to model the psychology of ordinary communication. Perhaps, then, the clue to Davidson's intentions is the sentence immediately preceding the one just quoted, in which Davidson writes: "Speakers of the same language can go on the assumption that for them the same expressions are to be interpreted in the same way, but this does not indicate what justifies the assumption" (Davidson 1984b: 125, my emphasis). Perhaps what radical interpretation is supposed to model is not the psychology but the epistemology of ordinary communication. Although there is, as I've noted, some evidence for this reading, it does not satisfy. There are two main difficulties. First, the epistemology to which Davidson would hereby be committing himself is deeply problematic. It is all well and good to raise the question why we are entitled to assume, as we typically do, that what sounds like the language we know really is the language we know, but it simply does not follow that the epistemic situation of the ordinary speaker is comparable to that of a radical interpreter. If challenged to defend the assumption that, say, Ned Block speaks English, I would appeal to a great deal of evidence that Davidson would deny a radical interpreter. One might respond that this evidence could itself be called into question and that, ultimately, my evidence is only what would be available to a radical interpreter. But the occurrence of the word "ultimately" here is a sure sign of a kind of foundationalism that has not only gone badly out of style but is not the sort of epistemology one typically associates with Davidson. 4 More importantly, the epistemological construal of the thought experiment does not fit the use Davidson makes of it. As noted, Davidson agrees with Quine that meaning is, in a certain sense, indeterminate, and his reason is that the evidence that is available to a radical interpreter does not suffice to restrict the acceptable theories of truth to one. On the epistemological interpretation of radical interpretation, however, no such conclusion could possibly be forthcoming. All we could conclude would be that the evidence available to a radical interpreter was inadequate to distinguish among the competing theories; in so far as the radical interpreter models the epistemic situation of ordinary speakers, one could conclude that no ordinary speaker could justify one such interpretation over another. But it simply would not follow that there was no fact of the matter which interpretation was correct. The questions that are bothering Davidson are thus not epistemological but metaphysical. They are the questions Strawson asks at the beginning of "Meaning and Truth" (1971): In virtue of what are sentences meaningful?

Copyrighted

Material

354

Richard Heck, Jr.

In virtue of what does a given sentence have the very meaning it has? Davidson (1984c) argues that any answer to these questions must take due notice of the fact that the meaning of a sentence depends upon the meanings of its parts. He argues in Davidson (1984e) that only a theory of truth can give a substantive account of this dependence. For that reason, he thinks, a theory of meaning should take the form of a theory of truth. If so, then we may be tempted to reformulate the initial question what it is for a sentence to mean what it does as: In virtue of what is a theory of truth correct or incorrect? As Foster was the first to emphasize, however, if our interest is in questions about meaning, then that is not quite right: a theory of truth can be correct in the sense that all of the T-sentences it proves are true and yet not throw much light at all on questions about meaning. So the question that needs asking, as Davidson sees things, is rather: In virtue of what is a theory of truth interpretive? Davidson's answer to this question, I take it, invokes a form of what I have in Heck (2006d) called the Use-Meaning Thesis: the meaning of a sentence is determined by how it is used. So stated, the Thesis is entirely programmatic: until we are told what "meaning" and "use" are, the Thesis has very little content. It is clear enough, however, how Davidson understands these notions: meaning is to be understood in terms of truth-conditions; use is to be understood in terms of Davidson's notion of "holding true." So, in Davidson's hands, the Use-Meaning Thesis takes the following form: whether a theory of truth for a given language is interpretive is determined by the circumstances under which sentences of that language are held true by its speakers. Why does it seem obvious to Davidson that use in his sense determines meaning in his sense? Davidson is moved. I think, by much the same sort of thought that has attracted so many other philosophers to the Use-Meaning Thesis: if there is no difference in how two expressions are used, then there can be no difference in what they mean; nothing that does not somehow surface in the use of an expression can be part of its meaning. One might wish to quarrel with that idea, but in this form it is so programmatic that one might well just concede it until more has been said about what use is supposed to be: if it is a fact about how "snow is white" is used that it is used to say that snow is white, then we can all quickly concede the Thesis. The interesting question is why Davidson thinks use can and should be characterized as he characterizes it. Why, for example, should the conditions under which a sentence is "held true" by speakers of the language be included among the facts about how it is used? Why should semantic facts, such as facts about what expressions of the language are used to say, be excluded? Why should "detailed descriptions of the speaker's beliefs and intentions" (Davidson 1984b: 134) be excluded? This last question, of course, is the crucial one. 5 Let me say as clearly as possible what question is at issue here: Why should we think that any difference in what two sentences mean must

Copyrighted

Material

Meaning and truth-conditions

355

manifest itself in some difference in the conditions under which those sentences are held true by speakers? Why, in particular, should we think that there could not be differences in meaning that make no difference to what is held true when, but that do make a difference that is visible when we consider "the complex and delicately discriminated intentions with which the sentence[s are] typically uttered" (Davidson 1984b: 127)? The question is particularly pressing in light of Soames's question how Davidson's radical interpreter can distinguish the case in which "snow is white" means that snow is white from the case in which it means instead that snow is white and arithmetic is incomplete. The sentence could of course mean either thing. And since it is a necessary truth that arithmetic is incomplete, it is not clear how such a difference in meaning could make any difference to the conditions under which "snow is white" is held true. The point, of course, is that Davidson understands the conditions under which a sentence is held true to be external conditions. It is clear enough what kind of difference such a difference in meaning might make to the internal conditions under which the sentence was held true. As things are, many speakers who know nothing of logic hold "snow is white" true. If it meant that snow is white and arithmetic is incomplete, one would not expect that to be the case. Such differences of meaning are thus easily enough distinguished if we characterize use in terms of the mental states of speakers. But Davidson explicitly denies such information to the radical interpreter. 6 Davidson's reply would surely be not so much to concede these points as to insist upon them. On his view, the radical interpreter is not just trying to interpret her subjects' words but also to interpret her subjects' minds. Davidson denies the radical interpreter access to prior information about her subjects' minds precisely because he thinks that there is only the one, unified interpretive project: The evidence [available to a radical interpreter] cannot consist in detailed descriptions of the speaker's beliefs and intentions, since attributions of attitudes, at least where subtlety is required, demand a theory that must rest on much the same evidence as interpretation. (Davidson 1984b: 134) The difficulty is that, although this remark seems fair enough if it is intended as one about the epistemic predicament of a radical interpreter, the question why we should care about the epistemic predicament of a radical interpreter is still unanswered. If the radical interpreter were a good epistemological model of an ordinary speaker, then that would provide a reason: perhaps ordinary judgments about meaning would be vulnerable to skeptical attack if meaning were not determined by the evidence available to a radical interpreter. Indeed, Dummett (1993a: 177) is clearly moved by just such a worry, but I have already registered my doubts about the epistemology that lies behind it.

Copyrighted

Material

356

Richard Heck, Jr.

There is another option, namely, to take Davidson once again to be making not an epistemological claim but a metaphysical one. On this interpretation, Davidson is denying the radical interpreter prior knowledge of what her subject believes and intends not because one can only know what someone believes and intends once one is able to interpret her words, but because the facts about what someone believes and intends are not independent of the facts about what her words mean. The radical interpreter's epistemological predicament, on this interpretation, is thus not a model of an ordinary speaker's epistemological situation but rather a reflection of the metaphysics of intentionality. Some might suggest that we should have started with this interpretation, and I certainly am not claiming to have discovered some hidden commitment of Davidson's view. As is well known, the metaphysical view this interpretation ascribes to Davidson that meaning, belief, and intention are constitutively intertwined - is one to which Davidson explicitly subscribed (1984d). What 1 am doing here is simply emphasizing that Davidson's metaphysics of meaning - his view that whether a theory of truth is interpretive is determined by the circumstances under which sentences are held true - depends upon his metaphysics of intentionality. Our question is whether any difference of meaning must manifest itself in a difference in the circumstances under which sentences are held true. I have argued that, in so far as Davidson offers us reasons to believe that it must, those reasons depend crucially upon the claim that meaning, belief, and intention are constitutively intertwined. However, if we conjoin that thesis with the view it is being used to establish, then we thereby commit ourselves to a metaphysics of the mind that seems to me. at least, to be extremely unattractive. Radical interpretation, in Davidson's sense, is possible only if meaning supervenes on the sorts of facts that are available to the radical interpreter. But if meaning, belief, and intention are constitutively intertwined, then belief and intention must also supervene on the facts available to a radical interpreter: that is to say, what mental states one has, and with what contents, is determined by (i) the conditions under which one would hold various sentences true and (ii) the facts about which sentences one actually does hold true. This view is but half a step beyond behaviorism, and, in my opinion, it deserves the same fate. But it is undoubtedly a view Davidson (1984d: esp. 162) held: it is just an "interpretive" view of the mind - what one's mental states are is a matter of how one would best be interpreted - coupled with a particular view about what the basis of interpretation is. The argument here can perhaps be made a little clearer if we state it in the form of a challenge to Davidson. Davidson is claiming that what a sentence means is determined by the circumstances under which it and related sentences are held true. But suppose there were two situations in which the facts available to a radical interpreter - the facts about what is held true when - were the same, but in which the facts about speakers' mental states - and in particular the facts about the communicative intentions with

Copyrighted

Material

Meaning and truth-conditions

357

which they uttered various sentences - were different. Now, on Grice's view (1989b), meaning is determined not (just) by what is held true when but by facts about speakers' communicative intentions; on the view I hold, meaning is determined by what speakers believe about the meanings of their words (see Heck 2006b, 2006c). Of course, neither Grice nor I would hold that every difference in the psychological states of speakers entails a difference of meaning: supervenience is a one-many, not a one-one, relation. But we would both be open to the possibility that the meanings of sentences could vary while the facts accessible to a radical interpreter remained the same. The question for Davidson, then, is what reason he can give us to suppose that no such possibility can be realized. His answer would be to insist that radical interpretation is simultaneously both of a speaker's words and of her mind. And Davidson's point, once again, would not just be epistemological: it is not just that it is hard to tell when one is in such a situation. Davidson's claim, rather, is that mentality itself supervenes on the kind of evidence available to a radical interpreter - on what is held true when - just as meaning does, so that there could not be such a situation. Mental states cannot vary while the facts about what is held true remain the same. Davidson, then, may well have had a solution to the Foster problem that worked for him: if mentality supervenes on the kind of evidence that is available to a radical interpreter, as he thought it did, then, plausibly enough, a theory of truth known to be true on the basis of the kind of evidence available to a radical interpreter can always be used to interpret. What is more important, however - and what I have just been arguing - is that Davidson's metaphysics of mind is also necessary if radical interpretation is to provide a solution to the Foster problem. A theory of truth known to be correct on the basis of the kind of evidence that is available to a radical interpreter can be guaranteed to be interpretive only if mentality supervenes on what is held true when. But it doesn't - at least not according to me - and so Davidson's solution is not one that works for me.

3 Frege on meaning and truth-conditions Davidson's suggestion that a theory of truth may serve as a theory of meaning was arguably anticipated by Frege. One might, therefore, seek a solution to the Foster problem in Frege's writings. Part I of Frege (1962) 7 is devoted to the exposition of the formal system in which he proposes to work. As part of that exposition, of course, Frege must explain the formal language of that system. In Part 1.1 (spanning Sections 1-25), he specifies the denotations of the language's primitive symbols. Concerning the identity-sign, for example, he writes: (8) "T' = A" shall denote the True if T is the same as A; in all other cases it shall denote the False. (Frege 1962: Section 7)

Copyrighted

Material

358

Richard Heck, Jr.

Frege takes this clause to specify which function the identity-sign denotes: it denotes the function whose value for first argument T and second argument A is the True if T is the same as A and is the False otherwise. There is a similar clause for each of the other primitive symbols. Frege then argues, in part 1.2.i (spanning Sections 26-33), that these stipulations are sufficient to assign denotations to all well-formed expressions of his formal language. The core of that argument is contained in Sections 30 and 31. In the former. Frege argues that any complex expression correctly formed from denoting expressions denotes; in the latter, Frege argues that the simple expressions of the language denote. It follows that "the proper names [including sentences], and names of first-level functions, that we can form . . . out of our simple names . . . always have a denotation" (Frege 1962: Section 31). Frege then writes: [N]ot only a denotation, but also a sense, attaches to all names correctly formed from our signs. Every such name of a truth-value [that is, every sentence] expresses a sense, a thought. Namely, by our stipulations it is determined under what condition the name denotes the True. The sense of this name - the thought it expresses - is the thought that this condition is fulfilled. . . . The names, whether simple or themselves composite, of which the name of a truth-value consists, contribute to the expression of the thought, and this contribution of the individual component is its sense. (Frege 1962: Section 32) That is to say, Frege takes the stipulations he has made regarding his primitive signs to determine the truth-condition of each sentence of his formal language, and we may take the sentence's truth-condition to determine, or perhaps just to be, its sense. So, for example, the stipulations determine that the formula "0 = 1" denotes the True if, and only if, zero (which is, on Frege's view, the value-range of a certain function) is the same as one (the value-range of a certain other function). The sense of this formula is thus the thought that zero is the same as one. There is something obviously right about the quoted remark. Formal languages are introduced by philosophical logicians all the time. And when they are, an informal theory of reference is what is typically offered to explain the expressions of that formal language. Such explanations are typically intended to convey not just what the denotations of these expressions are but also what they mean, and the explanations often succeed. When I have properly appreciated such an explanation, I take myself to understand the formal language in question, that is, to know what its expressions mean. So I am, and am meant to be, reading the meanings of the expressions of the language off a theory of reference for it. just as Frege suggests his readers should. Now, according to Frege (1984: 159, op. 27), sense determines reference, but reference certainly does not determine sense, and the same plausibly

Copyrighted

Material

Meaning and truth-conditions

359

holds for any reasonable notion of meaning that might replace the notion of sense. 8 If so, however, then it can easily seem mysterious how a theory specifying what the references of certain expressions are could possibly determine the meanings of those expressions. If reference does not determine meaning, then specifying the reference of an expression simply cannot suffice to specify its meaning. That line of thought, however, misses a crucial aspect of the case, one nicely brought out by Dummett: When we are concerned . . . with laying down . . . what the interpretation of a newly introduced symbol is to be, the particular specification of the reference may be taken as conferring simultaneously a definite corresponding sense. . . . We may adapt here, in expounding a doctrine of Frege's, the famous distinction between saying and showing that Wittgenstein used in the Tractatus: the specification of the reference says what the reference is to be, and, by saying it in a particular manner, shows what the sense is to be. (Dummett 1991: 149) One's first reaction might be to wonder just how helpful any explanation can be that invokes the distinction between saying and showing. But Dummett's idea here is plausible enough. In saying what the reference is, Frege has had to choose among different ways in which the reference itself might correctly be specified. So, for example, there are infinitely many ways correctly to specify the reference of the function-symbol " = " in his formal language. Actually to specify what its reference is, Frege has had to choose one such way, and the choice he has made may be taken to indicate - that is, to show - what sense he intends the symbol to have. Similar remarks apply to Frege's other primitive expressions and to other formal languages. How its inventor chooses to state what the references of a new language's expressions are to be may reasonably be taken to indicate how s'he intends those expressions to be understood. It is important to note that Dummett's remark explicitly concerns a case in which one is stipulating the meaning of a previously uninterpreted symbol. Dummett is not, I think, as clear about this point as is necessary. In particular, the last line of the passage just quoted is importantly ambiguous. Dummett speaks of "the specification" of a symbol's reference as showing what the sense is to be, but a specification can be an act of specifying as well as a sentence used to make a specification. The sentence Frege uses to specify the reference of the identity-sign does not "indicate" anything about what the sense of this symbol is: Qua sentence, it says what the reference of the symbol " = " is to be, and that is all. What indicates what sense the symbol is to have is Frege's using a certain sentence to specify the reference of that symbol. It is not what Frege says, then, that shows what sense " = " is to have, but his saying what he says in a particular way or, perhaps, under particular circumstances. Showing, in Dummett's sense, is thus akin to (and may even

Copyrighted

Material

360

Richard Heck, Jr.

be a sort of) implicating. Only speech acts, we may thus conclude, show anything, in the relevant sense. This point does not affect Dummett's vindication of Frege, but it does affect any attempt to extend it to the semantics of natural language, where a similar problem arises. The problem I have in mind arises for anyone who accepts the apparently obligatory distinction between meaning and reference, but it arises particularly clearly for those of us who accept the distinction between sense and reference for proper names. John McDowell (1998d) proposed some time ago that this form of the problem can be solved if we appeal to the idea of Frege's 1 have just been expounding. I shall therefore explain the matter first for the case of proper names. Consider the name "Hesperus." One could correctly say what its reference is in a variety of ways. One could say that it denoted Hesperus, or Venus, or Phosphorous. To say what its reference is, then, one must choose one such way, and which way one chooses will, under the right circumstances, show what one takes the sense of the name to be. One might therefore suggest that a semantic theory for English need say nothing about the meaning of a proper name other than what its reference is. even if one accepts the distinction between sense and reference for proper names. The theory will show what the sense of a name is even if it all it says is what its reference is. It is clear that, if this solution works at all, then there is a similar solution to the Foster problem. Suppose we have a truth-theory for English that (canonically) proves the T-sentence (9) 5" is true iff p. It is true enough that, as far as the correctness of the truth-theory is concerned, one can replace "p" with any other sentence that has the same truth-value. But to say what the truth-value of a sentence is, one must choose among the infinitely many different ways of specifying it, and which way one chooses will show what meaning one takes the sentence to have. So, for example, in propounding a truth-theory that canonically proves (1) rather than (2) or (7), we thereby show what meaning we take the sentence "snow is white" to have, even if the theory itself says nothing that distinguishes it from any other true sentence. Dummett rejects this view: Although he thinks that a stipulation of what a new expression's reference is to be can convey what its sense is to be, he insists that "a meaning-theory is required to do more than merely show (to someone who understands the meta-language in which it is formulated) what the senses of the words of the object-language are" (Dummett 1991: 149). But unfortunately, Dummett does not, so far as I can tell, offer any reason for this claim, and so it might simply be rejected. 9 One might suggest that the parenthetical remark is the key to his thinking: the theory will convey what the sense of the name is only to someone who already grasps that sense by associating it with the relevant word of the meta-language. But the theory

Copyrighted

Material

Meaning and truth-conditions

361

will convey which object the name denotes only to someone already able to refer to that object by means of the relevant word of the meta-language. It is not clear what the relevant difference between these cases is supposed to be. The proper complaint is that theories do not indicate things. The fact that a particular theorist is proposing one theory of truth instead of some other may indicate what senses that particular theorist takes the proper names in the relevant language to have. But it is, all sides admit, no part of what the theory says that a particular name has a particular sense. The relevant clauses of the theory, qua sentences, do not indicate anything about what the senses of the names are. They say what their denotations are, and that is all; only speech acts indicate. If so, however, then if what one wants a semantic theory for English to do is to specify the meanings of its various sentences (as potentially uttered), then a semantic theory for English cannot consist simply of a theory of reference for English, not if one accepts the distinction between meaning and reference. The point is clearest in the case of proper names. It is not enough that, by propounding this theory of reference instead of that one, I might thereby show what I take the sense of "Hesperus" to be. One wants the theory to explain the phenomena, not its being propounded by a theorist. One way to appreciate the problem is to consider how a theory of truth modeled on McDowell's proposal attempts to respect the principle of compositionality. It does so by deriving a T-sentence like (10) "Hesperus is a planet" is true iff Hesperus is a planet from clauses like (11) "Hesperus" refers to Hesperus (12) "x is a planet" is true of x iff x is a planet, and, more generally, by deriving facts about the references (or semantic values) of sentences from facts about the references of their syntactic parts. In that way, the theory explains how the reference of a complex expression is determined by the references of its parts. The theory seems to offer no explanation, however, of how the sense of a sentence is determined by the senses of its parts. Certainly, in propounding a theory that has a consequence like (10), one indicates what one takes the meaning of "Hesperus is a planet" to be; and since the theory also includes (11) and (12), one similarly indicates, by propounding the theory, what one takes the senses of "Hesperus" and "is a planet" to be. Since the theory derives (10) from (11) and (12), then, it derives a theorem whose (canonical) provability in the theory one is propounding indicates what one takes the meaning of "Hesperus is a planet" to be from axioms whose presence in the theory one is propounding indicate what one takes the senses of "Hesperus" and "is a

Copyrighted

Material

362

Richard Heck, Jr.

p l a n e t " to be. But it simply does n o t follow that one has thereby explained why "Hesperus is a p l a n e t " has the sense it has in terms of the fact that " H e s p e r u s " a n d "is a p l a n e t " have the senses they do. T h a t need not hold even if the relation is not the weak relation of indication but the m u c h stronger one of implication. T h e problem can be seen f r o m this diagram: sense (A)

sense (6)

sense (C)

T

T

t

ref[A)

ref(B)

ref(C)

T h e double arrow " " symbolizes a relation of determination; the single arrow may be taken to m e a n implies, indicates, or shows, as one likes. N o matter which of these readings one gives it, there is simply n o reason to suppose there should be a double arrow in the top line, at least for all that has so far been said. Perhaps one could argue that for the special case in which " " means shows, the double arrow will always be present, but that very m u c h needs to be argued. A good argument would replace the metap h o r of showing with something m u c h more precise. I ' m skeptical myself that there is any such a r g u m e n t to be found. But if there is, then, I conjecture, what replaced the m e t a p h o r of showing would be precise e n o u g h to be e m b e d d e d in a theory of meaning, so there would n o longer be any need for the sort of solution to the Foster problem we have been considering.

4 Understanding and knowledge of meaning In o u r earlier discussion of radical interpretation, the central question was whether the "formal and empirical criteria" Davidson specifies are sufficient to restrict the range of acceptable truth-theories to those that are "interpretive." Now, Davidson endorses Quine's thesis of the indeterminacy of translation: he calls the suggestion that his criteria determine a uniquely interpretive theory " a b s u r d . " Davidson (1984b: 139) nonetheless insists "that any theory that passes the tests will serve to yield interpretations." This remark is more puzzling, I think, than it is usually taken to be: In what sense is a truth-theory supposed to "yield interpretations"? O r perhaps the right question to ask is a different one: W h a t does it m e a n to say that a theory of t r u t h can be used to interpret a language or a speaker? Suppose Fred utters a sentence, a n d I wish to interpret this utterance. Suppose f u r ther that I happen to k n o w that Fred's utterance is true iff Bedrock is cold. W h a t a m I supposed to d o with this information? O n e proposal would be that the radical interpreter is supposed to use her theory of t r u t h explicitly to assign meanings to utterances. T h e radical interpreter would thus need to make something like the following inference:

Copyrighted

Material

Meaning and truth-conditions

363

(i) S is true iff p. (ii) That fact is a consequence of a theory of truth I know to be correct on the basis of the sort of evidence available to me qua radical interpreter. (iii) Therefore, S means that p. If interpreting a speaker requires one to assign meaning to her utterances in this sense, however, then, as Soames (1992) says in a similar context, the theory of truth is simply an auxiliary device used to generate assignments of meaning. It does not seem likely that Davidson would have wanted the notion of meaning to emerge so crucially here at the end of the story. A different suggestion is that the T-sentence itself can be used to interpret the speaker. The difficulty is to say how, and part of the problem here is that the notion of interpretation itself is not particularly clear: What exactly does Davidson mean by "interpreting" an utterance, a speaker, or a language, whichever the right target might be? I really don't know, so I propose to reformulate the problem. I take it that what Davidson has in mind is that someone who knew a theory of truth for a language, and knew that it met the formal and empirical criteria he specifies, could use that theory to speak and understand that language. So the question I am going to discuss is this: How could one use a theory of truth, whatever its pedigree, to speak and understand the language it concerns? That, of course, depends upon what one means by "speaking and understanding the language." As is common, I am going to concentrate here on what I take to be one fundamental aspect of speaking and understanding a language, namely, using the language to communicate, that is, to exchange information with other speakers. To be able to speak a language at least involves being able to use it to acquire beliefs from other speakers and to convey one's own beliefs to them. 1 0 Our question, then, may be put as follows: In what sense, if any, can a theory of truth for a given language be used by a radical interpreter (or an ordinary speaker) to allow her to communicate using that language? The examples due to Foster and Soames may seem to suggest that the answer should be: It can't. But there is reason for hope. I argued earlier that Foster and Soames are right, as regards Davidson's actual position, that we have no good reason to believe that a theory of truth that is known to be correct on the basis of the sort of evidence available to a radical interpreter will always satisfy Tarski's Convention T - that is, that the sentences used on the right-hand side of the T-sentences that the theory (canonically) proves will translate the sentences mentioned on their left-hand sides. However, suppose a radical interpreter were nonetheless irresponsibly to settle for one of the theories the criteria admit and set about using it to interpret the speakers of the language anyway. Then it is tempting to say that, just as Frege's choosing to specify the reference of a term in a particular way shows us what sense he takes it to have, so the radical interpreter's using her chosen theory will amount to her choosing to regard sentences of the target language as having certain meanings rather than

Copyrighted

Material

364

Richard Heck, Jr.

others. That is to say, the same line of thought we considered in the last section seems to lead to the conclusion that a radical interpreter who chose to interpret using a theory that proved (1) rather than one that proved (7) would thereby have chosen to regard utterances of the sentence "snow is white" as meaning that snow is white rather than that snow is white and arithmetic is incomplete. Of course several questions remain: In what sense does using a particular theory of truth to interpret a sentence involve assigning a particular meaning to that sentence? Is there some analogue for ordinary speakers of our irresponsible radical interpreter's decision to use one theory of truth rather than another? Before I continue, I need to say something explicitly about the overall structure of the considerations to follow. As I said earlier, I accept a cognitive conception of (occurrent) understanding: understanding an utterance, according to me, is a propositional attitude (Heck 2006c). That view leaves open the question what the content of that attitude is, however - in particular, it leaves open whether understanding an utterance is knowing its meaning, or its truth-condition, or what have you - and that is the question I mean now to discuss. How, though, should we evaluate the various proposals that might be made here? The following conditional should be uncontroversial: (13) If B utters a sentence S, which in that context means that p. and if A understands B's utterance and meets whatever conditions a correct theory of testimony would require us to include at this point, then A is in a position to come to know, or at least justifiably to believe, on a certain distinctive kind of ground, that p. This conditional is uncontroversial because it is trivial. The correct theory of testimony just is whichever theory makes this conditional correct. It is controversial which theory that is, but we may ignore that issue here. We are not discussing the epistemology of testimony. What we are discussing is what constitutes understanding an utterance. And what I am suggesting is that, just as satisfying (13) is very nearly what it is for a theory of testimony to be correct, so satisfying (13) is very nearly what it is for a theory of understanding to be correct. To understand an utterance is to be in a position - modulo one's satisfaction of the sorts of conditions a correct theory of testimony would specify - to acquire a belief for which one would then have a certain distinctive sort of justification. If so, then theories of understanding may be tested by determining whether they satisfy (13). In some cases, how we evaluate a given example may depend upon controversial claims about what the correct theory of testimony is. But that need not be so in all cases. Particular examples may be neutral on what precise conditions a correct theory of testimony would require to be satisfied in the sense that all (reasonable) theories may agree that the relevant conditions are satisfied in that case.

Copyrighted

Material

Meaning and truth-conditions

365

There are some substantive assumptions about testimony to which I shall be appealing, but I take these to be relatively uncontroversial. I shall assume, for example, that to acquire knowledge by testimony, one must be appropriately sensitive - whatever exactly that may mean - to whether one's informant is sincere and reliable. 1 shall also assume, slightly more controversially, that the distinctive sort of justification testimony makes available has a certain kind of structure. In particular, there is a way in which the justification one might have had for the belief that p can be undermined that leaves one with a justification for the weaker claim that the speaker believes that p. In particular, one might be presented with strong evidence that the speaker is not reliable with respect to the question whether p. In the usual sort of case, that would deprive one of justification for the belief that p but not for the belief that the speaker believes that p . 1 1 So let us now consider this proposal: to understand an utterance is to know its truth-condition, that is, to know a T-sentence for it. Suppose that Fred asserts the sentence "Bedrock is cold." Suppose further that Barney meets whatever conditions a good theory of testimony would say he needed to meet. Let's just assume he has excellent reason to believe that Fred is sincere and excellent reason to regard Fred as reliable with regard to what he has said; any reasonable theory would regard that as sufficient. Then, if Barney understands Fred's utterance, he will be in a position to form a justified belief, indeed, a belief that might constitute knowledge, namely, the belief that Bedrock is cold. Now suppose that Barney knows (14) "Bedrock is cold" is true iff Bedrock is cold. Intuitively, knowing (14) seems insufficient for Barney to acquire knowledge, from Fred's utterance, that Bedrock is cold. The truth of (14) is, after all, compatible with the hypothesis that "Bedrock is cold" means that Bedrock is broke, so long as Bedrock is broke iff Bedrock is cold. But if what the sentence "Bedrock is cold" actually means is that Bedrock is broke, then it is not plausible that Barney could, in reaction to Fred's utterance, come to know that Bedrock is cold, at least not in the distinctive way in which one comes to know something by being told. One might want to object that if Barney has good reason to regard Fred as sincere and reliable, that constitutes good reason to suppose that Fred's utterance is true. That, together with (14), which we are assuming Barney knows, then gives him good reason to suppose that Bedrock is cold. But although Barney could come justifiably to believe, in this way, that Bedrock is cold, his belief would not be justified in the distinctive way in which beliefs acquired through testimony are justified. The reason is that, absent knowledge of what Fred's utterance actually meant, Barney's belief cannot be appropriately sensitive to the question whether Fred is reliable. To evaluate Fred's reliability, Barney must know what belief Fred is expressing.

Copyrighted

Material

366

Richard Heck, Jr.

Barney may regard Fred as authoritative with regard to climatology, say, but not with regard to municipal finance, and these attitudes may be justified. So if Barney takes Fred sincerely to have said something about the local climate, he may be prepared to believe it, and be justified in doing so, so that he might acquire knowledge about the climate from Fred. But Barney may not be prepared to believe what Fred says about municipal finance, and he would not be justified in doing so if he did, whence he cannot acquire knowledge about municipal finance from Fred. That "Bedrock is cold" is true iff Bedrock is cold does not, however, imply anything about what belief Fred is expressing. That is the fundamental point: absent knowledge of what Fred's utterance means, Barney is in no position to determine what belief Fred is expressing.' 2 The point is worth making directly. Suppose Barney understands Fred's utterance and justifiably regards it as sincere. Then even if he is not sure about Fred's reliability, he may conclude, and even come to know, that Fred believes that Bedrock is cold. But if Barney does not know what Fred's utterance means - if all he knows about it is that it is true iff Bedrock is cold - then he can draw no conclusion whatsoever about Fred's beliefs. And so Barney's knowing a correct T-sentence for the sentence Fred uttered cannot constitute his understanding it. The natural counter-proposal is that understanding consists not in knowledge of an utterance's truth-condition but in knowledge of what it means. 1 3 Perhaps surprisingly, however, even knowledge of what an utterance means is insufficient to support attributions of belief on the basis of what someone has said. Consider again Fred's utterance of "Bedrock is cold" and suppose that Barney knows that "Bedrock is cold" means that Bedrock is cold. Fred may not. Maybe he thinks it means that Bedrock is broke. Barney may even have reason to doubt that Fred thinks "Bedrock is cold" means that Bedrock is cold. If so, then Barney cannot justifiably conclude that Fred believes that Bedrock is cold. What Barney needs to know if he is to attribute a belief to Fred is not just what Fred's words mean but what they mean to Fred. The point should be obvious once stated. Communication with another speaker - in so far as it involves attributing beliefs to her and acquiring beliefs from what she says - depends not upon what her words mean, in some objective sense, nor even upon what one takes her words to mean oneself. It depends upon what she means by her words and upon what one takes her to mean by her words. Now typically, what one takes someone to mean by her words is what one takes them to mean oneself, and perhaps such a default assumption is in some sense necessary. Nonetheless, when I am responding to the words of another, the crucial question I must answer is what she means by her words. 14 I take it, then, that the following inference would be a reasonable one for me to make: (15) The speaker, N, has uttered the sentence 5.

Copyrighted

Material

Meaning and truth-conditions

367

(16) N takes her utterance of S, in the present context, to mean that p. (17) N's utterance of S is sincere. (18) Hence, N believes that p. What I am now going to argue is that the role played in such an inference by (16) may be played instead by something like: (19) N takes her utterance of S, in the present context, to be true iff p. The motivation for this suggestion can be explained as follows. We began with the proposal that the work being done by (16) could be done instead by (16') N's utterance of S is true, in the present context, iff p. In moving from (16') to (16), we have made two changes: we have replaced the reference to the utterance's truth-condition with a reference to its meaning, and we have replaced an absolute judgment about what the meaning of the utterance is with a judgment about what the speaker takes it to mean. My suggestion is that the latter change is the one that really matters and that it is, in fact, adequate by itself to resolve the problems faced by the initial proposal. Actually, that's not quite right: (19) will not quite do. But something much like it will do. Before I explain how it needs to be amended, though, let me emphasize two features of the proposal I am making. First, one might wonder whether it is enough to defend Davidson's suggestion that a theory of truth may serve as a theory of meaning to defend the view that the knowledge of meaning on which ordinary linguistic action depends consists in knowledge of truth-conditions. Questions about meaning, one might say, are one thing, and questions about what, if anything, ordinary speakers know, in virtue of which they are able to communicate, are another (Soames 1988). But what interesting questions would remain unanswered if it could be established - as I think it can be - that a speaker's ability to communicate depends upon, and is explained by, her knowledge of truth-conditions? I do not know of any. In any event, the view I am defending accepts a strong form of Dummett's oft-cited insistence that "a theory of meaning is a theory of understanding" (Dummett 1993d: 3). I take the primary notion to be how a speaker understands an uttered sentence, not what a word means in a language. One can reconstruct the latter notion if one likes in terms of the former notion, but I do not myself think it is of much independent interest. Second, my view, as I've stated it, does not require that there be any sense in which speakers know theories of truth for their languages. If understanding an utterance involves knowing a T-sentence for it, and if knowledge of that T-sentence is to be among one's reasons for reacting to the utterance

Copyrighted

Material

368

Richard Heck, Jr.

in a certain sort of way, then one's knowledge of the T-sentence must be conscious, in the sense that it is available for the rational control of action. It is agreed on all sides, of course, that ordinary speakers do not have conscious knowledge, in this sense, of semantic theories for their languages. So if speakers do, in some sense, know semantic theories for their languages, that knowledge is tacit, and tacit knowledge is, almost by definition, not available for the rational control of action. That said, however, the question nonetheless begs to be asked what the source of our conscious knowledge of truth-conditions is, if we really do have such knowledge. It is a natural, though not inevitable, proposal that the truth-conditions we consciously know are the upshot of subpersonal processes that draw upon the information contained in the axioms of a theory of truth and, indeed, that these processes implement an algorithm for deriving T-sentences from tacitly known axioms of the theory of truth. I think something along those lines is correct - see Heck (2006e) - but I will not defend that claim here. It is unclear to what extent this view differs from Davidson's. On the one hand, Davidson (1984b) insists that the question that interests the semanticist is hypothetical: What could we know that would allow us to interpret the words of other speakers? Regarding this question, he writes that it is not the same as the question what we do know that enables us to interpret the words of others. For there may easily be something we could know and don't, knowledge of which would suffice for interpretation, while on the other hand it is not altogether obvious that there is anything we actually know which plays an essential role in interpretation. (Davidson 1984b: 125) It is tempting to read this passage as rejecting the view I am defending. If, however, what Davidson means by "know" here is consciously know, and if what he has in mind is not just knowledge of T-sentences but knowledge of a theory of truth, then there need be no disagreement between us. Davidson could consistently agree that understanding another speaker requires conscious knowledge of what she takes the truth-conditions of her utterances to be while remaining agnostic about the empirical question what the source of this knowledge is. 5 Understanding and truth-conditions The view I am defending, then, is that, in so far as we rely in linguistic communication upon knowledge of meaning, knowledge of meaning may be taken to be knowledge of truth-conditions. More precisely I am claiming that the role that is played by (16) in this reasoning (15) The speaker, N, has uttered the sentence S.

Copyrighted

Material

Meaning and truth-conditions

369

(16) N takes her utterance of S, in the present context, to mean that p. (17) TVs utterance of S is sincere. (18) Hence, N believes that p. can instead be played by something similar to (19) N takes her utterance of S, in the present context, to be true iff p. I use the non-committal term "takes" here because I mean to leave it open what the relevant psychological state is. I have said already that it must be conscious, in the sense that it is available for the control of rational action, but nothing in what follows will depend upon whether it is knowledge, belief, or something else still. Embedding Davidson's proposal within the cognitive conception already gives us some headway with the Foster problem. On this view, after all, what one takes an utterance to mean is, in effect, a matter of what one believes the speaker takes its truth-condition to be, so not just any T-sentence one knows for the utterance will equally serve to interpret it. One might well know oneself, for example, that (20) "Bedrock is cold" is true iff Bedrock is cold and arithmetic is incomplete, but if one docs not think that Fred knows (20), one will not use one's knowledge of (20) to interpret his utterances of "Bedrock is cold." But, of course, that's far from sufficient progress. Perhaps one met Fred at a logic conference and so may presume that he does know (20). And even if Fred isn't a logician, maybe he was paying attention in his basic science courses, in which case he presumably knows (21) "Bedrock is cold" is true iff Bedrock is cold and water is H 2 0 . And even if he wasn't paying attention in science class, surely Fred must know (22) "Bedrock is cold" is true iff Bedrock is cold and water is wet. There are plenty of similar cases. 15 James Higginbotham has suggested that, though the Foster problem shows that "meaning does not reduce to reference, . . . knowledge of meaning reduces to the norms of knowledge of reference" (Higginbotham 1991: 274). These " n o r m s " concern what a competent speaker of the language is supposed to know in so far as she is competent. So, very roughly:

Copyrighted

Material

370

Richard Heck, Jr. To know what a sentence S means is: (i) to know that S is true iff p and (ii) to know that knowing that S is true iff p is what, and all, competence requires a speaker to know about it.

For example, to know what "Bedrock is cold" means is to know (14) and to know that (14) is what, and all, a competent speaker is supposed to know about it. If so, then since one is not, in general, expected to know (20), (20) does not, in any sense, "give the meaning" of "Bedrock is cold." It is unclear, however, what Higginbotham means by what one is expected to know "in so far as one is a competent speaker." Is one expected to know (23) "All bachelors are married" is true iff all bachelors are female in so far as one is competent? On some views, the answer would arguably be that one is. "All bachelors are unmarried", the thought might be, is analytically false: Anyone who understands it must know that it is false, and so perhaps one must know that it is false if one is competent. But "All bachelors are female" is also analytically false, so maybe one is expected to know (23) simply in so far as one is a competent speaker. Perhaps not, but it is not obvious there are no similar counterexamples. However, even if Higginbotham's proposal is extensionally adequate, it is not clear how much explanatory value it has. Most competent speakers of English do know (22), and they expect other competent speakers to know it, too. It is natural to say that they do not know (22) in the same way they know (14); that their knowledge of (22) depends somehow upon their knowledge of (14); and that their knowledge of (14) is "linguistic" in a way that their knowledge of (22) is not. We might summarize these intuitions by saying that competent speakers of English know (14), but not (22), simply as competent speakers of English (Higginbotham 1991: 274). But talk of what one knows simply as a speaker is obscure, and one would like to have these intuitions explained, not just labeled. Such an explanation is what I shall now attempt. Utterances are acts. As such, they are typically performed for reasons. Fred's uttering "Bedrock is cold" is an act, and one typical reason for uttering it might be to communicate, to Barney, that Bedrock is cold. But Fred's utterance of this sentence is rational, as a means to this end, only if he is assuming certain facts about Barney, for example, that Barney will take him at his word. Of special interest to us, however, is the need for Fred to assume that Barney will understand him as having said that Bedrock is cold. If Barney instead took him to have said that Bedrock is broke, Fred's attempt to communicate to Barney that Bedrock is cold would fail. So if, as I am assuming, Barney's understanding Fred's utterance consists in his knowing something about it, we may now ask: What is it that Fred is assuming Barney knows? Earlier I suggested that if an irresponsible radical interpreter were simply to settle upon a theory of truth and set about using it to interpret the natives,

Copyrighted

Material

Meaning and truth-conditions

371

her doing so would amount to her assigning certain meanings to their utterances. Suppose Fred regards Barney as similarly irresponsible: despite the fact that Barney knows that Fred knows many correct T-sentences for "Bedrock is cold," Barney has nonetheless decided to use a particular one of them to interpret Fred's utterances. If Fred believed that the one on which Barney had settled was (14), then Fred's uttering "Bedrock is cold" in an effort to communicate to Barney that Bedrock is cold would make perfect sense: if Barney takes Fred at his word, that is precisely the belief he will acquire. It may well be that Fred also expects that Barney believes (22), at least implicitly. But unless Fred were prepared to suppose that Barney would deploy that belief in interpreting his utterance, it would be irrational for him to utter "Bedrock is cold" in an effort to get Barney to believe, in response to his utterance, that Bedrock is cold and water is wet. Of course, Fred might suppose that, if he can get Barney to believe that Bedrock is cold, he can thereby get him to believe that Bedrock is cold and water is wet. But Fred's reasons for uttering the sentence, in this case, are just different from the sorts of reasons he has if he assumes that Barney will deploy (22) in interpreting his speech. The lesson that is supposed to emerge from this story is that one way to distinguish interpretive T-sentences, such as (14), from non-interpretive Tsentences, such as (22), is simply by reference to the roles they play in the production and comprehension of speech. A speaker treats a given T-sentence as "giving the meaning" of a sentence if she deploys that T-sentence in a certain distinctive way, both when speaking herself and when interpreting the speech of others. For example, we may reconstruct Fred's reasons for uttering "Bedrock is cold" roughly as follows: (24) I want Barney to know that Bedrock is cold. (14) "Bedrock is cold" is true iff Bedrock is cold. (25) Barney knows (14), and he will deploy that knowledge in interpreting my speech. (26) If I utter "Bedrock is cold", in the present context, Barney will know that I have spoken the literal truth. (27) If I utter "Bedrock is cold" in the present context, Barney will (be in a position to) come to know that Bedrock is cold. (From (25) and (26).) (28) I shall utter Bedrock is cold. (From (24) and (27).) That (14) "gives the meaning" of "Bedrock is cold", for Fred, simply consists in his using it in the way here illustrated, and in the following sort of way when interpreting Barney's utterances: (29) Barney has uttered "Bedrock is cold".

Copyrighted

Material

372

Richard Heck, Jr.

(14) "Bedrock is cold" is true iff Bedrock is cold. (30) Barney knows (14), and he expects me to deploy that knowledge in interpreting his utterance. (31) Barney intends to be speaking the literal truth. (32) Barney believes that Bedrock is cold. (From (29), (30), and (31)) One could obviously go on in this way for some time, exploring the role that knowledge of (14), and mutual expectations about its deployment, play in the many different forms of reasoning in which speakers engage. N o t e that Fred's expectation that Barney will deploy the belief that his utterance of "Bedrock is cold" is true iff Bedrock is cold is an expectation that Barney will suppose that Fred himself has deployed the same belief in speaking. So Fred expects Barney to expect him to have deployed this belief in speaking, and we could continue the iterations. The expectations of the participants in a given communicative exchange are thus, at least when all goes well, mutual, not just in the sense that they are shared but in the stronger sense that they are intertwined: each conversational participant's expectations concern, inter alia, the expectations of the others. As Higginbotham (1991: 274) notes regarding a similar feature of his own proposal, this fact reflects the same sort of phenomena that led to Grice's (1989a) emphasis on the overtness of communicative intentions and to Lewis's (1986) emphasis on common knowledge. The suggestion, then, is that linguistic knowledge is knowledge that is used in a certain distinctive way: it is the knowledge one deploys in speech and interpretation. One might well want to ask, then, what is meant here by how someone deploys her knowledge in speech and interpretation. A flatfooted answer is that a speaker has deployed the belief that S is true iff p in speaking if she uses it in the way Fred uses his knowledge of (14) in deciding to utter "Bedrock is cold"; an interpreter deploys her belief that S is true iff p in interpreting an utterance if she uses it in the way Barney uses his knowledge of (14) in interpreting Fred's utterance of "Bedrock is cold." I would certainly like to be more explicit than that. Unfortunately, I can't seem to be quite yet. 16 So let me instead suggest a way of approaching the issue from a slightly different direction. As mentioned earlier, there is a strong intuition that knowledge of (20), (21), and (22) depend in some way upon knowledge of (14), that is, that (14) is, in some sense, one's most basic piece of semantic knowledge regarding the sentence "Bedrock is cold." So one might hope to characterize linguistic knowledge in terms of this notion of a speaker's most basic semantic knowledge. One way to do so would begin with the thought that our most basic linguistic knowledge has a distinctive source. As I mentioned earlier, the question begs to be asked what the source of our linguistic knowledge is, be it knowledge of truth-conditions or

Copyrighted

Material

Meaning and truth-conditions

373

knowledge of meaning. The sorts of phenomena that are used to motivate the principle of compositionality suggest, though they do not of course imply, that, as Dummett (1993e: 36) puts it, "a speaker of a language derives his understanding of any sentence of that language from his knowledge of the meanings of the words" that compose it. Of course, such a derivation is not conscious but tacit, so the proposal would have to be that our conscious linguistic knowledge is the product of subpersonal computational processes that draw upon the information contained in tacitly known theories of truth. If so, then the knowledge we deploy in speech and interpretation is, in the normal case, delivered by the language faculty. It is, of course, an empirical question what knowledge is so delivered in any particular case. But it is not unreasonable to suppose that (14) is so delivered and that (22) is not. McDowell (1998a: 33Iff.) emphasizes throughout his writings on such topics that we do not simply hear the words other people utter and then add an interpretation in thought. Rather, in the normal case, our appreciation of what someone has said is a result of our perceiving not just the words she has uttered but also, in the strictest sense, our perceiving what she has said. If so, then one's interpretation of speech is not always a matter of what information one chooses to deploy but is typically a matter of what information just does get deployed in one's perception of speech. One should not, however, suppose that such an observation shows that higher cognitive states, like beliefs, have no central role to play in speech and interpretation. 1 7 If one happens to know that Jones uses a particular word say, "livid" - differently from how one uses it oneself, then, although one may still hear Jones as saying that Smith was pale when he utters "Smith was livid," one may choose not to deploy that information in interpreting him. Rather, since one knows, via some other route, that, when Jones utters "Smith was livid," he is not himself deploying a belief that it is true iff Smith was pale but rather a belief that it is true iff Smith was flushed, one deploys that information instead. Of course, I usually just proceed on the assumption that Jones's words mean to him what they mean to me. But so to proceed is, in effect, to take appearances at face value and so to act on the perceptually justified belief that Jones's utterance is true iff Smith was pale. Ultimately, or so I suspect, an adequate answer to the question what linguistic knowledge is will have to make reference both to facts about how that knowledge is put to use and to facts about its normal source. Unfortunately, again, I must confess that I am unable at present to say anything that is likely to satisfy someone skeptical that an adequate answer can be sewn from the materials we have been discussing. In my defense, I plead that the question what linguistic knowledge is is famously difficult. The proposal I am making is at least not already known to fail.

Copyrighted

Material

374 6

Richard

Heck,

Jr.

Conclusion

T h e f o r e g o i n g c o n s t i t u t e s a p r i n c i p l e d , if still i n c o m p l e t e , s o l u t i o n t o t h e F o s t e r p r o b l e m . W e d o h a v e r e a s o n t o believe t h a t a s p e a k e r w h o c o n sciously k n e w t h e t r u t h - c o n d i t i o n s o f s e n t e n c e s o f a given l a n g u a g e (as u t t e r e d , o r p o t e n t i a l l y u t t e r e d , in given c o n t e x t s ) a n d w h o w a s p r e p a r e d t o d e p l o y t h a t k n o w l e d g e in s p e a k i n g a n d i n t e r p r e t i n g o t h e r s , w o u l d u n d e r s t a n d , a n d b e a b l e t o use, t h a t l a n g u a g e . O f c o u r s e , t h a t d o e s n o t s h o w t h a t c o m p e t e n t h u m a n s p e a k e r s do h a v e s u c h k n o w l e d g e , a n d m a n y p h i l o s o p h e r s , s u c h as S o a m e s , h a v e f o u n d t h e s u g g e s t i o n i m p l a u s i b l e , since it c o m m i t s us t o t h e view t h a t all c o m p e t e n t h u m a n s p e a k e r s , i n c l u d i n g y o u n g children, possess a meta-linguistic c o n c e p t of truth, a claim he regards as itself i m p l a u s i b l e ( S o a m e s 1988: 189). A s it h a p p e n s , I d o n o t agree. B u t t h e issue is u l t i m a t e l y a n e m p i r i c a l o n e , a n d m y p u r p o s e h e r e h a s s i m p l y b e e n t o c o m b a t s o m e f a m i l i a r a priori a r g u m e n t s t h a t t r u t h - c o n d i t i o n a l s e m a n tics c a n n o t deliver w h a t w e w a n t f r o m a s e m a n t i c t h e o r y . 1 8

Notes 1 Of course, in actual research, semantics and syntax interact with one another, and work on semantic issues can inform work in syntax. It is, in fact, a nice question whether syntax is " a u t o n o m o u s " in the sense my remarks somewhat sloppily suggest it is. 2 By a T-sentence here I mean any sentence of the form: S is true iff p. In particular, no relation between the sentence mentioned on the left-hand side and that used on the right-hand side is required for a sentence of that form to be a T-sentence. I shall say that a T-sentence is correct if it is true, that is, if the sentence mentioned on the left-hand side has the same truth-value as the sentence used on the right-hand side. 3 The importance of this question was made clear to me by Michael Dummett's insistence upon a similar question regarding radical translation, and my remarks here are inspired by Dummett's. See D u m m e t t (1978). 4 For remarks along similar lines, although much better developed, see Burge (1999). See also McDowell (1998a) for some general reflections on the epistemology of understanding that support the point being made here. McDowell would not apply these points to Davidson, however, as he reads Davidson somewhat differently from how I do. Dummett, rather, is McDowell's target. For what it's worth, however, in my view, D u m m e t t is not really an appropriate target after 1978 and the emergence of the view presented in D u m m e t t (1993c). 5 1 have argued in Heck (2006d) that the answer to the second question - why semantic facts should be excluded - is just that their inclusion would trivialize the problem. With respect to the first, Lewis (1985) quarreled with Davidson's inclusion of what is held true. For all its differences, however, Lewis's treatment of radical interpretation is still vulnerable to the objections I am bringing against Davidson. That is because Lewis's view depends as heavily u p o n his metaphysics of mind - his analytical functionalism - as Davidson's does upon his. 6 C o m p a r e Higginbotham (1991: 277-78). 7 Translations are primarily taken f r o m Frege (1964), with some minor changes. All further references will be m a d e by section number and are to Volume I. (The numbering of the sections is reset in Volume II.) For more on the issues discussed here, see Heck (1998) and Heck (2006a).

Copyrighted

Material

Meaning and truth-conditions

375

8 This is at least true when we are considering sentences. 9 If Dummett (1993d) does offer a reason, it is equivalent to the insistence that a theory of meaning must, in his sense, be "full-blooded" rather than "modest." But it seems likely that this issue and the one we are discussing here are effectively the same issue, and McDowell (1998b), in any event, would reject this requirement. See Heck (2006d). 10 Ian Rumfitt emphasizes in Rumfitt (1995) that there is no single reason for which we utter sentences, a point some of my own previous work e.g.. Heck (1995) neglects. The present discussion is intended, however, to be compatible with Rumfitt's observations. 11 This last assumption entails that one can only have testimonial justification that p if one also has testimonial justification that one's informant believes that p. That could be, and has been, questioned, but it is weak enough, and intuitively plausible enough, that I am willing to rest my case upon it. (Note that I am not assuming that one's justification for the claim that p must rest upon one's justification for the belief that one's informant believes that p. That claim is one antireductionists might want to contest.) 12 T h a n k s to Crispin Wright for helping me understand the structure of my thought on these matters. 13 Cf. Richard (1992). 14 In fact, something similar is true even when I am speaking: the crucial question for me is how my words will be understood. See Heck (2006b) for more on this matter. 15 Note that since (22) is presumably c o m m o n knowledge if (14) is, it will not suffice to strengthen (19) by inserting a reference to common knowledge. 16 Crispin Wright suggested that, for the case of the speaker, one might appeal to some notion of the intention that is most immediately controlling the act of uttering the sentence. Once spelled out. that is probably close to a proposal of Rumfitt's. Adapted to the present framework, it is that the T-sentences that constitute understanding are those that occur in what he calls the "lowest practical syllogism" that occurs in practical reasoning of the sort we've been discussing (cf. Rumfitt 1995: 842.) 17 McDowell sometimes gives the impression that, epistemologically. it is of no particular interest that we do not just perceive what someone has said but also perceive her words. Tyler Burge has argued for a similar view, namely, that one's justification for the claim that someone has said that p need not involve any claim about what words she has uttered (Burge 1993). But the possibility of the sort of re-interpretation I consider here depends upon our perceiving someone's words, and we do not perceive someone's words and what she has said as one can perceive an apple and an orange. We perceive what has been said in the words somehow, and so our perception of speech is itself "semantic," in the sense that meaning binds to sound. Precisely what that might mean is a very nice question. See Heck (2006c) for more on this issue. 18 My work on these issues has benefitted enormously from conversations over the years with Jim Higginbotham, Michael Rescorla, and Jason Stanley. Material from this paper was presented during a series of seminars I gave at the University of St Andrews in January and February 2004 as a British Academy Visiting Professor and then again in a graduate seminar at Harvard University in Spring 2004. Thanks to all who attended and participated for their comments and encouragement, especially Jake Beck, Tony Corsentino, Bernard Nickel. Daniel Nolan, Agustin Rayo, and Crispin Wright. Thanks to the British Academy and to Arché, the A H R C Research Centre for the Philosophy of Logic, Language, Mathematics and Mind, for their support, which is much appreciated, and to Crispin for arranging the visit.

Copyrighted

Material

376

Richard Heck, Jr.

References Burge, T. (1993) "Content Preservation." Philosophical Review. 102: 457-88. (1999) "Comprehension and Interpretation." in L. Hahn (ed.) The Philosophy of Donald Davidson, Chicago. IL: Open Court, pp. 229-50. Davidson, D. (1984) Inquiries into Truth and Interpretation, Oxford: Clarendon Press. Dummett, M. (1978) "The Significance of Quine's Indeterminacy Thesis," in M. Dummett Truth and Other Enigmas. London: Duckworth, pp. 375-419. (1991) The Logical Basis of Metaphysics. Cambridge, MA: Harvard University Press. — -(1993) The Seas of Language. Oxford: Clarendon Press. Foster, J.A. (1976) "Meaning and Truth Theory." in G. Evans and J. McDowell (eds) Truth and Meaning: Essays in Semantics. Oxford: Oxford University Press, pp. 1-32. Frege, G. (1962) Gnmdgesetze der Arithmetik. Hildesheim: Georg Olms Verlagsbuchhandlung. (1964) The Basic Laws of Arithmetic: Exposition of the System, ed. and trans. M. Furth, Berkeley, CA: University of California Press. (1984) "On Sense and Meaning." in Collected Papers on Mathematics. Logic, and Philosophy, trans. M. Black, Oxford: Basil Blackwell. pp. 157-77. Grice. H. (1989) Studies in the Ways of Words. Cambridge. MA: Harvard University Press. Heck. R. (1995) "The Sense of Communication," Mind, 104: 79-106. (1998) "Grundgesetze der Arithmetik I §§29- 32." Notre Dame Journal of Formal Logic, 38: 437-74. (2006a) "Frege and Semantics," forthcoming in D. Greimann (ed.) Essays on Frege's Conception of Truth. (2006b) "Idiolects," in A. Byrne and J. Thomson (eds) Content and Modality: Themes from the Philosophy of Robert Stalnaker. Oxford: Oxford University Press, pp. 61 92. (2006c) "Reason and Language." in C. MacDonald and G, M a c D o n a l d (eds) McDowell and His Critics. Oxford: Blackwells. pp. 22-15. (2006d) "Use and Meaning." forthcoming in L. Hahn (ed.) The Philosophy of Michael Dummett, Chicago. IL: Open Court. (2006e) What is Compositionality? Draft. Higginbotham, J. (1991) "Truth and Understanding," Iyyim, 40: 271-88. Lewis, D. (1985) "Radical Interpretation," in Philosophical Papers, vol. I, Oxford: Oxford University Press, pp. 108 18. — (1986) Convention, Cambridge, MA: Harvard University Press. McDowell. J. (1998) Meaning. Knowledge, and Reality, Cambridge, MA: Harvard University Press. Richard, M. (1992) "Semantic Competence and Disquotational Knowledge." Philosophical Studies, 65: 37-52. Rumfitt, I. (1995) "Truth-conditions and Communication," Mind. 104: 827-62. Soames, S. (1988) "Semantics and Semantic Competence," in S. Schiffer and S. Stelle (eds) Cognition and Representation. Boulder. CO: Westview Press, pp. 185-207. (1992) "Truth. Meaning, and Understanding." Philosophical Studies, 65: 17-35. Strawson. P. (1971) "Meaning and Truth." in Logico-Linguistic Papers. London: Methuen, pp. 170- 89. Tarski, A. (1933) "The Concept of Truth in Formalized Languages," in J. Corcoran (ed.) Logic, Semantics, and Metamathematics. Indianapolis, IN: Hackett. pp. 2-78.

Copyrighted

Material

17 Meaning, truth and normativity1 Michael Williams

1 Introduction Paul Horwich is a prominent defender of a deflationary approach to truth. He calls his particular version of this approach "minimalism." 2 Roughly speaking, the minimalist thesis is that the meaning of the truth predicate is fixed by the equivalence-schema: (MT) The proposition that P is true if and only if P. In treating the equivalence-schema as fixing the meaning of the truth predicate, Horwich intends to make two claims. First, the biconditionals that are instances of the equivalence-schema are "epistemologically fundamental." That is to say, "We do not arrive at them, or seek to justify them, on the basis of anything more obvious or more immediately known" (Horwich 2001: 559). Second, this underived inclination to accept these biconditionals is "the source of everything else we do with the truth predicate" (Horwich 2001: 559). What makes minimalism so strongly deflationary is the idea that the rule of use it points to tells us all there is to know about the meaning of "true" or, if we like, the nature of truth. The account of the truth-predicate given by this idea is deflationary in two ways. In the first place, it is a version of theoretical deflationism, in that it explains the nature of truth using very restricted theoretical resources. But theoretical deflationism goes naturally with functional deflationism, a view about the uses to which the truth-predicate can and cannot be put. For Horwich, as for deflationists generally, the truth-predicate is important only in that it adds significantly to the expressive resources of our language (as a special kind of generalizing device, exploiting semantic ascent). But it plays no role in explanation, other than the expressive role already admitted. In this sense, "truth" is not a substantive concept and truth is not a substantive property. 3 As much as any, it is this latter feature of deflationism that provokes resistance. For according to many philosophers - Donald Davidson being the most prominent - there is one area of theory in which a robust concept

Copyrighted

Material

378

Michael

Williams

of truth is indispensable, namely, the theory of meaning. For Davidson, there is no hope of explaining the nature of meaning without making use of the concept of truth, because the meaning of a sentence is to be explained in terms of its truth-conditions. But if we need truth to play a serious explanatory role in the theory of meaning, a deflationary approach to truth is untenable. Rejecting functional deflationism, Davidson rejects theoretical deflationism too. To be sure, Davidson does not expect an illuminating theoretical analysis of the nature of truth. Indeed, to an extent, Davidson can agree with Horwich and other deflationists that our inclination to accept instances of some appropriate version of the equivalence-schema is "underived." But that is because the concept of truth is primitive, not because it is a mere device for semantic ascent, whose entire meaning can thus be exhaustively captured in the deflationist's stripped-down terms. 4 Horwich agrees that minimalism precludes his making any serious appeal to truth in the theory of meaning; and he agrees that this rules out explaining a sentence's meaning in terms of its truth-conditions. Thus he sees Davidson's argument as a challenge to deflationists to come up with an alternative approach. His suggestion is that we explain meaning in terms of use. 5 His position seems to be the mirror-image of Davidson's. Davidson thinks that to illuminate meaning we must make use of the notion of truth. As a result he finds deflationary accounts of truth untenable. Convinced by a deflationary account of truth, Horwich looks for another approach to meaning. This is about as stark an opposition as one could imagine. Or is it? I think that, on closer examination, the relation between Horwich's views and Davidson's is far from simple. Although, as we shall see, they are indeed at odds over fundamentals, it is not clear that what divides them has much to do with truth. Getting clear about what is really at issue will be a way of exploring the general question of what deflationists about truth can and should say about meaning. Exploration of this general question will lead me to investigate a number of more detailed issues. Assuming that deflationists cannot explain meaning in terms of truth-conditions, must their approach to the theory of meaning make use of the notion of (word or sentence) meanings? But is the assumption really correct that, for the purposes of giving a theory of meaning, the notion of truth-conditions is off-limits to deflationists? Or is there perhaps a sense, available to deflationists, in which a sentence's meaning is a matter of its truth-conditions? And does the notion of meaning have some kind of ineliminably normative dimension?

2 Meaning: Horwich and Davidson The question is whether we need a substantive concept of truth for use in the theory of meaning. But what is a theory of meaning? Let us begin with Horwich. For Horwich, a theory of meaning is first and foremost a theory of the nature of meaning. (I take the phrase "theory of

Copyrighted

Material

Meaning, truth and nor/nativity

379

the nature of meaning" from Gurpreet Rattan.) 6 In Horwich's view, such a theory should accomplish at least two things. It should be a theory of meaningfulness: that is, it should explain what makes a certain vocalization (or scribble) the utterance (or inscription) of a (meaningful) word, and not just a sound (or scrawl). 7 But a theory of (the nature of) meaning should also respect a principle of "non-revisionism." It should take as its starting point our (alleged) commonsense conviction that words have meanings.8 So a second goal of a theory of meaning should be to explain what meanings (of both words and sentences) consist in. That is, the theory should identify the facts that constitute a given word's (or sentence's) having the particular meaning that it has. For Horwich, the second question is fundamental. In explaining how a certain word comes to have a particular meaning (how "dog" comes to mean DOG), we automatically explain what meaningfulness in general consists in. But the questions are worth separating, since it is not obvious that we are compelled to proceed this way. Horwich agrees that, as a deflationist, he cannot approach meaning armed with a richer-than-deflationary notion of truth. But of course, the constraints he is working under are stronger even than this. He cannot approach meaning armed with concepts that would permit the construction of a richer-than-deflationary notion of truth. For example, he cannot help himself to a (non-deflationary) notion of reference.9 So what can he do? According to Horwich, a sound or scribble becomes a word by having a use. To have a use is to be governed by an appropriate kind of core regularity. 10 The core regularity is the use property that best explains the word's overall use. In this way, a word's particular use determines its particular meaning. It is thus an important part of Horwich's approach that there be certain special use-properties that are meaning-constitutive in the sense of "constitutive of particular meanings." This is how Horwich respects his principle of non-revisionism. For Horwich, use properties are acceptance-properties. He recognizes two broad kinds of meaning-constituting acceptance-property. For some words, the relevant acceptance-property, or use-property, consists in the fact that we accept certain specified sentences containing them. For example, the acceptance-property that governs a speaker's use of "and," thus fixing its meaning, is (roughly) his tendency to accept "p and q" if and only if he accepts both "p" and "q." Or again, as we saw, minimalism about truth holds that our understanding of true is fully explained by our conforming to a rule of use involving our acceptance of all non-paradoxical instances of the equivalence-schema: (MT) The proposition that P is true if and only if P. Accordingly, "true" itself, as accounted for by minimalism, provides a paradigm example of how a word can derive its meaning from a basic acceptance-property. 1 1 The use of "true" as a paradigm is a nice touch, for

Copyrighted

Material

380

Michael

Williams

it brings out how Horwich's approach to meaning reinforces his deflationary view of truth. Minimalism explains the meaning of "true" by pointing to our inclination to accept instances of the equivalence-schema. Letting < p > represent the proposition that-p, we might suppose that the minimalist account of truth is equivalent to the following definition: x is true ≡ [x = < dogs bark > and dogs bark; or x = < pigs fly > and pigs fly; or . . . and so on.] But now a dilemma looms. If the number of such instances is potentially infinite, the minimalist proposal implies that "true" will never be fully understood. Alternatively, if a speaker's understanding of "true" is exhausted by the instances of the equivalence-schema he can presently recognize, then its meaning will depend on and vary with whatever else is in his vocabulary. 12 But as Horwich points out, this objection assumes that the minimalist intends to offer a definition. This is not so. The minimalist thesis is not meant as either an explicit or contextual definition of "true." Rather, it is supposed to "specify the fact of usage that provides the truth predicate with its meaning"; and that fact of usage is simply "our underived inclination to accept instances of the equivalence-schema" (Horwich 2001: 567). This is not a defect in the minimalist account of "true" but an illustration of how meanings generally are to be explained. So far, so good. However, the account of acceptance-properties suggested by " a n d " and "true" cannot be thought to govern the use of every word we understand. The biconditionals governing the use of "true" are all sentences that we actually accept. In this sense, to understand "true" is to accept the postulates of the minimalist "theory" of truth. But we do not want to say, in parallel fashion, that to understand "phlogiston" is to accept certain key sentences such as "Phlogiston is released when a substance burns." Such an acceptance-property would be sufficient to fix the meaning of "phlogiston," but cannot be necessary. This is because we can understand "phlogiston" without committing ourselves to phlogiston theory. Accordingly we must distinguish two kinds of acceptance-commitments. On the one hand there is "the unconditional practice of using that word to formulate one's acceptance of certain substantive principles." On the other hand there is the conditional commitment "use that word to articulate those principles, if they are to be accepted" (Horwich 1998a: 91). This conditional acceptance is equivalent to acceptance of a conditional. Thus to understand "phlogiston," all we have to accept is some sentence like "If anything is phlogiston, it escapes during burning, has negative weight. . . . etc." (Horwich 1998a: 91). When our use of a word is governed by our acceptance of such a " C a r n a p conditional," the word expresses a concept that we possess but do not deploy. This distinction applies beyond the terms occurring in scientific

Copyrighted

Material

Meaning, truth and normativity

381

theories. For example, an intuitionistic mathematician will know what inference rules must govern "not" if it is to express classical negation. But he will not avail himself of such rules in his own theorem-proving. He will possess the concept of classical negation, while refusing to deploy it. Horwich sets considerable store by the contrast between possession and deployment. Although his idea of core, meaning constitutive use properties is a close relative (I would say version) of the analytic-synthetic distinction, by allowing meanings to be constituted by the acceptance of Carnap conditionals, Horwich frees the analytic distinction from any unwelcome, epistemological consequences. The sentences that we accept, just by virtue of understanding some term, are indeed apriori. But they are only conditional. The substantive postulates of a theory are never true by virtue of meaning alone. We can detach Q u i n e s comprehensive fallibilism from his skepticism about meanings. The use properties discussed so far all involve only the acceptance of sentences. In characterizing such properties, we make no reference to anything outside language. But not all meanings can be explained this way. For some words, meaning-constitutive use properties involve our being disposed to utter those words, or to accept sentences involving them, in circumstances characterized by using those words themselves. Words that occur in observation-reports are the paradigm instances here. Thus the meaning of "red" may be partially constituted by our tendency to accept "That's red" in the presence of things that are (visibly) red, "That's a dog" in the presence of dogs, and so on. Horwich sees his approach to meaning as a form of "non-nominalistic reductionism" (Horwich 1998a: 5). It is non-nominalistic because it makes no attempt to reduce meanings to non-semantic entities: "the concept D O G (i.e. the abstract property of doggyness) cannot be identified with some physical or mental object" (Horwich 1998a: 5). Rather, a concept can be characterized only indirectly by reference to "the circumstances in which a word or complex expression would indicate its engagement by the speaker's mind" (Horwich 1998a: 5). These circumstances are, of course, the speaker's acceptance-commitments, as manifest in (by way of being the best explanation of) his verbal behavior. But since these circumstances can be characterized in wholly non-semantic terms, Horwich's approach is still reductionist. To be sure, there are no reductive analyses of meaning entities. But there are reductive analyses of meaning properties, since we can specify in wholly non-semantic terms the conditions for a word's meaning A N D or DOG or whatever. This further clarifies what Horwich means by his non-revisionism: his commitment to the principle that words have meanings. In a way, Horwich does not countenance meanings: that is, he does not identify meanings with any things. Nor does he treat even meaning-properties as part of the ultimate furniture of the universe, since he thinks that such properties can be given reductive analyses. Nevertheless he takes meanings very seriously.

Copyrighted

Material

382

Michael

Williams

Meaning-properties have precise identity conditions. Two words have the same meaning if and only if they are constituted by the same acceptancecommitments. Horwich is untroubled by Quinean doubts about synonymy. The analytic-synthetic distinction is absolutely central to his account of meaning. A final clarification. Horwich often speaks of "rules of usage" and of "accceptance-commitments." I myself have spoken of acceptance-properties "governing" our use of a word. But none of these locutions is more than a façon de parler. The "rules" that constitute our meanings are simply regularities or, more precisely, underlying acceptance-properties that best explain manifest regularities. "Use" is a matter of what we in fact do, not what we ought to do. Rules are regularities that we do follow, not standards that we ought to follow (not even in virtue of committing ourselves to follow them). So Horwich's view is not only reductive but fully naturalistic. N o t only can meaning-properties be explained in wholly non-semantic terms, they can be explained in wholly non-normative terms. Normatively informed concepts have no fundamental theoretical role to play in the theory of meaning. So much for Horwich. Turning to Davidson, the first thing to notice is that his use of the phrase "theory of meaning" differs markedly from Horwich's. In Davidson's usage, a theory of meaning (for a particular speaker) is a representation of what we know (normally implicitly) when we are able to interpret that speaker's words. For clarity, I shall refer to what Davidson would call a "theory of meaning" as an "Interpretation" or a "Davidsonian theory of meaning." If I use the phrase "theory of meaning" without qualification, I shall be talking about a theory of the nature of meaning. Although Davidson's use of the phrase "theory of meaning" differs from Horwich's, Davidson is not indifferent to questions about the nature of meaning. Davidson's answer to the question "When is a sound a word?" is "When it is interpretable." For Davidson, explaining what this answer comes to requires us to do two things. We must decide how the results of interpretation are to be reported: in his terms, what form a "theory of meaning" should take. And we must explain the empirical and theoretical constraints with reference to which a theory of this form can be judged correct. Taken together, Davidson's accounts of the proper form of an interpretation and of the general methodology of interpretation yield an implicit theory of the nature of meaning. Davidson's views about the form and method of interpretation are justly famous, so I shall describe them only briefly. 13 According to Davidson, the proper form of an Interpretation (for a particular speaker) is a Tarski-style truth theory (for the language of that speaker). Such a theory will assign referents to names and extensions to predicates; then, given a way of determining the logical form of any given sentence of the language in question, the theory will enable us to prove theorems - T-sentences - of the form:

Copyrighted

Material

Meaning, truth and normativity

383

s is true-in-L if and only if p. Thus a Davidsonian theory of meaning, for a given language, reports the meanings of the sentences in a way that avoids all talk of meaning. We might wonder how this can be. Surely, as material biconditionals, Davidsonian T-sentences are too weak to capture meanings. However, Davidson thinks that when we see that interpretation requires us to generate T-sentences in a systematic way, under severe theoretical and empirical constraints, we realize that this worry is misplaced. Following Quine, Davidson thinks that we can identify the relevant constraints by considering the enterprise of radical interpretation. That is, we can ask how we would have to proceed to construct a theory of meaning for a speaker, none of whose words are antecedently understood. Radical interpretation demands that we proceed charitably. First of all we must take the person whose speech we are interpreting to be at least minimally rational: inclined to avoid egregious self-contradiction and to accept obvious logical consequences of acknowledged commitments. This is not a matter of being indulgent: if a person were grossly illogical, we could discern no patterns in his speech and would have nothing to work with. And the point is not merely epistemological: in the presence of gross illogicality, there would he no patterns to discern. Second we must take him to hold beliefs that are for the most part true (by our own lights). If we knew what his words meant, we could find out what he believes. If we knew what he believed, we could begin to interpret his words. Knowing neither, we must take him to accept at least the more obvious things that we ourselves hold true. And finally, we must pay special attention to occasion sentences, for these are the point at which the causes and contents of beliefs converge. Typically, "Dog?" will elicit assent in the presence of a dog, but not otherwise. Comparing Horwich's approach to meaning with Davidson's, there are three features of Horwich's view that I want to emphasize. 1 Horwich's theory explains meaningfulness by way of explaining meanings. The meaning of a word is constituted by certain core patterns typically acccptance of certain sentences in which it figures conformity to which gives the best explanation of the word's overall use. In short, Horwich's commitment to meanings leads him to embrace a version of the analytic -synthetic distinction. 2 Horwich's theory is fully naturalistic. Meanings are a matter of regularities to which our utterances in fact conform, not a matter of norms to which they ought to conform. To be sure, language-use is subject to all kinds of instrumental "oughts." If we want to make ourselves understood, we do well to use words in ways that accord with community-wide uses. But there is nothing intrinsically normative about meaning. 1 4 3 Because it explains word-meanings simply in terms of regularities, Horwich's theory has no need for a (substantive) concept of truth.

Copyrighted

Material

384

Michael

Williams

Davidson differs from Horwich on all three points, or so it seems: 1 Davidson's theory makes no appeal to "meanings." Material equivalences, systematically generated under the relevant empirical and theoretical constraints, can be taken to give the meanings of the sentences of the target language. But Davidson is not committed to identifying the sort of "core regularities" that are central to Horwich's theory. Provided we can generate sentence-to-sentence equivalences, under the appropriate constraints, we need not worry about which of our interlocutor's sentences are "meaning constitutive" and which express his substantive beliefs. Davidson has no use for the analytic-synthetic distinction or anything like it. 2

Davidson's theory is irreducibly normative. Precisely because of the holism of meaning and belief, implicit in rejection of the analytic-synthetic distinction, Davidsonian radical interpretation must proceed under the constraints of charity. But charity requires seeing a speaker we wish to interpret as rational. This means that, to have any hope of interpreting the utterances of another, we must avoid attributing to him egregiously inconsistent beliefs. Indeed, we should impute any kind of error only where error is (rationally) intelligible. Without rationality - a normative notion if any is there is no meaning. Pace Horwich, meaning is intrinsically normative. That is to say. normatively informed notions play an indispensable role in the theory of meaning.

3

Davidson makes use of the concept of truth both in presenting particular Interpretations and in his interpretative methodology. At the level of particular Interpretations ("theories of meaning" in Davidson's sense), the meanings of sentences are given by stating truth-conditions. To be sure, a particular Interpretation can be taken to define (recursively) a restricted truth-predicate (true-in-L), but our capacity to recognize the predicate thus defined as a truth-predicate presupposes a grasp of the general (and unanalyzed) concept of truth. As for Davidson's methodology of interpretation, the normative constraints of charity are naturally stated in terms of truth. At least, this is the official story.

What are we to say about all this? 3 Meaning, meanings and normativity Horwich thinks that his strategy of explaining meaning by way of meanings is mandated by his principle of "non-revisionism." According to Horwich, we should respect our commonsense conviction that words have meanings. However, this argument is just a piece of spccial pleading. Of course words have meanings: they are not just sounds or scribbles. But it does not follow that it is anyone's commonsense conviction that they have meanings in Horwich's highly theorized sense. And even if it did, why should we care?

Copyrighted

Material

Meaning, truth and normativity

385

For all 1 know, the commonsense conception of meanings is Lockean: meanings are ideas in the head. Would non-revisionism require us to respect this? Of course not. To see whether Horwich's principle is one we should accept, we need to examine how it works in particular cases. We have already met Horwich's paradigm examples of how meanings are constituted by use: "true," "phlogiston," and "dog." As we saw. the meanings of "true" and "phlogiston" are fixed entirely by our acceptance of certain meaning-constitutive sentences. As an observation-term, "dog" needs special treatment, so I shall leave it for later. Let us begin with "true." Horwich argues that the meaning of "true" is fixed by our commitment to a rule of use that involves our accepting all non-paradoxical instances of the equivalence-schema (MT) The proposition that-p is true if and only if p. This is just deflationism about truth in minimalist guise. However, although I am sympathetic to deflationism about truth, I am not inclined to suppose that "true" provides much of a guide for how meanings are constituted generally. There should be nothing surprising about this. After all, a central tenet of deflationism is that "true" is special. Since the truth-predicate is well behaved, we can say, if we like, that truth is a property. But it is not a "substantive" property: it has no underlying nature, and the concept of truth does no causal-explanatory work. This is why its use can be summed up in a compact way. But precisely because of its special character (at least as that character is explained by deflationists), the truth-predicate is dangerous as a paradigm of how meanings are constituted generally. The problems that bedevil analyticity arise most vividly in connection with concepts which, unlike that of truth, do play serious explanatory roles. Such concepts, which are always appearing in new beliefs, thereby forging new inferential connections, are not apt to have their behavior summed up by neat formulae. For such concepts, the line between what is or is not "meaning-constitutive" is pragmatic and context-sensitive subject to particular expository interests. There is no reason to suppose that it has any deep theoretical significance. Horwich makes it look as though the theory of meaning needs meanings - here in the guise of sharply delimited core regularities, embodied in our acceptance of sentences that are meaning-constitutive for the terms they contain - by choosing what he himself ought to see as a far-from-typical example. This puts a lot of weight on "phlogiston," which does express a substantive theoretical concept. As we noted, Horwich is interested in "phlogiston" because he wants to show that we can accept meaning-constitutive sentences involving that word without committing ourselves to the truth of phlogiston-theory. But again, the very reason for his taking an interest in this word points to a way in which it too is far too special a case to serve as a template for how meanings are constituted generally. "Phlogiston" belongs

Copyrighted

Material

386

Michael

Williams

t o a d e a d ( t h u s f r o z e n ) t h e o r y . In H o r w i c h ' s t e r m s , t h e c o n c e p t of p h l o g i s t o n is o n e t h a t we p o s s e s s b u t n o l o n g e r deploy. M o r e o v e r , even w h e n alive, t h e t h e o r y w a s n e v e r all t h a t w e l l - a r t i c u l a t e d . A s f a r a s I k n o w , t h e r e w a s n ' t m u c h t o p h l o g i s t o n b e y o n d its b e i n g p r e s e n t in c o m b u s t i b l e m a t e r i a l a n d released d u r i n g b u r n i n g . W i t h a t e r m like this, a t e r m t h a t is i m p o v e r i s h e d in its i n f e r e n t i a l c o n n e c t i o n s a n d n o w f r o z e n , it m a y i n d e e d be p o s s i b l e t o i d e n t i f y c o r e s e n t e n c e s t h a t " c o n s t i t u t e its m e a n i n g . " ( A l t h o u g h a l s o p e r h a p s not: even for a term f r o m a dead theory, the question of what we take to d e f i n e it m a y h a v e n o a n s w e r t h a t s w i n g s f r e e of p a r t i c u l a r h i s t o r i o g r a p h i c a l p u r p o s e s . ) B u t t h e f a c t , if it is a f a c t , t h a t w e c a n d o t h i s f o r " p h l o g i s t o n " gives us n o r e a s o n t o s u p p o s e t h a t t h e uses of living w o r d s ( e x p r e s s i n g actively d e p l o y e d c o n c e p t s ) c a n b e r e d u c e d t o rule. I c o n c l u d e t h a t H o r w i c h h a s given us n o r e a s o n t o t h i n k t h a t , in o r d e r t o e x p l a i n m e a n i n g f u l n e s s , w e m u s t c o u n t e n a n c e m e a n i n g s , o r even t h a t w e c o u l d find t h e m if w e w a n t e d t h e m . T h e p r o b l e m w i t h t h e a n a l y t i c - s y n t h e t i c d i s t i n c t i o n is n o t j u s t t h a t it t h r e a t e n s us w i t h u n w e l c o m e e p i s t e m o l o g i c a l c o n s e q u e n c e s : s u b s t a n t i v e c o m m i t m e n t s m a s q u e r a d i n g a s implicit d e f i n i tions. T h e s e c o n s e q u e n c e s c a n p e r h a p s b e a v o i d e d , as H o r w i c h a r g u e s . T h e p r o b l e m is t h a t t h e r e is n o p r i n c i p l e d w a y t o d r a w it in t h e first place. W i t h t h i s c o n c l u s i o n in h a n d , w e c a n t a k e u p t h e q u e s t i o n o f w h e t h e r m e a n i n g is somehow "normative." A c c o r d i n g t o H o r w i c h , t h e m e a n i n g - c o n s t i t u t i v e uses g o v e r n i n g s o m e words - "true" or "phlogiston" a r e e m b o d i e d in o u r r o u t i n e a c c e p t a n c e o f certain sentences c o n t a i n i n g those words. However, as we noted, observat i o n t e r m s , like " d o g " o r " r e d , " a r e i n v o l v e d in r e g u l a r i t i e s o f a q u i t e d i f f e r e n t k i n d . T h e s e r e g u l a r i t i e s i n v o l v e o u r d i s p o s i t i o n t o u t t e r t h e w o r d s in q u e s t i o n in c i r c u m s t a n c e s t h a t c a n be c h a r a c t e r i z e d o n l y b y u s i n g t h o s e v e r y w o r d s . So, r o u g h l y s p e a k i n g , t h e o b s e r v a t i o n a l c o m p o n e n t in t h e m e a n i n g of " d o g " c o n s i s t s in o u r b e i n g d i s p o s e d t o u t t e r " D o g " ( o r t o a s s e n t t o " D o g ? " ) in t h e p r e s e n c e o f d o g s . (I say " r o u g h l y s p e a k i n g " b e c a u s e H o r w i c h is q u i t e c l e a r t h a t this a c c o u n t of t h e o b s e r v a t i o n a l c o m p o n e n t in t h e m e a n i n g o f t e r m s like " d o g " is o v e r s i m p l i f i e d . ) T h e r e is a w e l l - k n o w n o b j e c t i o n t o a n y t h e o r y of this type, a n o b j e c t i o n t h a t is f a m i l i a r t o r e a d e r s of Sellars. 1 5 T h e o b j e c t i o n is t h a t a n e x p l a n a t i o n of t h e m e a n i n g o f o b s e r v a t i o n - t e r m s t h a t a p p e a l s only t o a s i m p l e d i s p o s i t i o n t o r e p o r t - i n - p r e s e n c e will n o t d i s t i n g u i s h genuine o b s e r v a t i o n - r e p o r t i n g , exhibiting c o n c e p t u a l mastery, f r o m the mere c o n d i t i o n e d responses o f t h e p a r r o t , trained to squawk " D o g " when the family canine enters the room. Anyone w h o presses this o b j e c t i o n is likely t o f o l l o w Sellars in s u p p o s i n g t h a t g e n u i n e o b s e r v a t i o n - r e p o r t i n g d i f f e r s f r o m " p a r r o t i n g " b y b e i n g e m b e d d e d in a m o r e w i d e - r a n g i n g p r a c t i c e of i n f e r e n c e a n d a c t i o n . O n t h i s view, o b s e r v a t i o n - r e p o r t s a r e essentially e v i d e n c e f o r f u r t h e r beliefs. T h u s it is e s s e n t i a l t o t e r m s w i t h o b s e r v a t i o n a l uses t h a t t h e y h a v e n o n - o b s e r v a t i o n a l uses a s well. I see n o r e a s o n t o p r e s s this o b j e c t i o n a g a i n s t H o r w i c h . F o r in H o r w i c h ' s view, o u r use of " d o g " also c o n f o r m s t o c e r t a i n c o r e r e g u l a r i t i e s o f t h e

Copyrighted

Material

Meaning, truth and normativity

387

" t r u e " a n d " p h l o g i s t o n " type, r e g u l a r i t i e s e m b o d i e d in o u r r o u t i n e a c c e p tance of such sentences as " D o g s are q u a d r u p e d s . " H o r w i c h never claims t h a t t h e m e a n i n g o f a t e r m like " d o g " o r " r e d " is exhausted by its o b s e r v a tional c o m p o n e n t . T h e real p r o b l e m f a c i n g H o r w i c h c o n c e r n s his a c c o u n t o f t h e o b s e r v a t i o n a l c o m p o n e n t itself. H o r w i c h is q u i t e c l e a r t h a t a s i m p l e a p p e a l t o " r e p o r t i n g - i n - p r e s e n c e " will n o t d o . T h e r e a s o n w h y is o b v i o u s e n o u g h : a n y d e s c r i p t i o n o f t h e r e p o r t i n g - i n - p r e s e n c e use of a n o b s e r v a t i o n - t e r m m u s t i n v o l v e a s u b s t a n t i a l ceteris paribus c l a u s e ( n o t e a s y t o spell o u t , if it c a n b e spelled o u t at all). W h e t h e r , in t h e p r e s e n c e o f a n a p p r o p r i a t e o b j e c t o r p r o p e r t y , a n o b s e r v a t i o n - r e p o r t i n g d i s p o s i t i o n results in a n o b s e r v a t i o n r e p o r t is a l w a y s t o s o m e d e g r e e d e p e n d e n t o n c o l l a t e r a l beliefs. W e a r e not a l w a y s inclined t o say " D o g " in t h e p r e s e n c e o f d o g s : if we have n o t g o t a g o o d l o o k at t h e a n i m a l , if w e t h i n k t h e d o g is a s t u f f e d d o g , a n d so o n a n d so o n , o u r d i s p o s i t i o n t o say " D o g " will be s u p p r e s s e d . B u t this is j u s t where Davidson's normative requirement t h a t we r a t i o n a l i z e a n o t h e r ' s s p e e c h , especially b y m i n i m i z i n g i n e x p l i c a b l e e r r o r - m a k e s its f o r c e felt. U n l e s s w e a p p r o a c h s p e a k e r s a l o n g D a v i d s o n i a n lines, w e will n o t find t h e regularities governing observation-terms that Horwich himself regards as e s s e n t i a l t o t h e c o n s t i t u t i o n of t h e i r m e a n i n g s . But if this is r i g h t , H o r w i c h ' s o w n t h e o r y h a s n o h o p e o f b e i n g t h e k i n d o f pure r e g u l a r i t y t h e o r y t h a t h e w o u l d like it t o be. H o r w i c h t e n d s t o c o n s t r u e t h e q u e s t i o n of w h e t h e r m e a n i n g is i n t r i n s i cally n o r m a t i v e r a t h e r n a r r o w l y . T h i s is n o t a c c i d e n t a l b u t r e s u l t s f r o m his c o m m i t m e n t t o m e a n i n g s . G i v e n this c o m m i t m e n t , t h e q u e s t i o n of w h e t h e r m e a n i n g is intrinsically o r o n l y i n s t r u m e n t a l l y s u b j e c t t o n o r m a t i v e c o n s t r a i n t is n a t u r a l l y i d e n t i f i e d w i t h t h e q u e s t i o n o f w h e t h e r t h e r e a r e local (i.e. w o r d - s p e c i f i c ) m e a n i n g - n o r m s . D a v i d s o n , w h o d i s p e n s e s w i t h m e a n i n g s altogether, can happily agree with Horwich that there are not. For Davids o n , n o r m a t i v e c o n s i d e r a t i o n s e n t e r at t h e g l o b a l level, in t h e c o n s t i t u t i o n o f meaningfulness. I n d e e d , t h e y e n t e r precisely b e c a u s e o f his m e a n i n g belief h o l i s m a n d its a s s o c i a t e d r e j e c t i o n o f e n c a p s u l a t e d m e a n i n g s . In i n t e r p r e t i n g p r i m a f a c i e p u z z l i n g u t t e r a n c e s , w e t r a d e o f f d i f f e r e n c e s in v e r b a l u s a g e a g a i n s t d i f f e r e n c e s in c o l l a t e r a l beliefs ( " D o e s h e n o t k n o w w h a t a n X is [what " X " m e a n s ] , o r is t h e r e s o m e r e a s o n w h y h e d i d n ' t r e c o g n i z e this o n e ? " ) , a n d a l w a y s in t h e c o n t e x t of a n overall p r o j e c t of r a t i o n a l i z a t i o n . H o r w i c h rejects n o n - i n s t r u m e n t a l m e a n i n g - n o r m s b e c a u s e h e r e c o g n i z e s t h a t a s p e a k e r c a n d e p a r t f r o m c o m m u n i t y uses a n d still m a k e sense. But as a soi-disant n o n - r e v i s i o n i s t ( " W o r d s h a v e m e a n i n g s " ) , his rejection of m e a n i n g - n o r m s focuses on normativity rather t h a n on m e a n ings. So, f o r H o r w i c h , t h e r e a r e m e a n i n g - c o n s t i t u t i v e uses, b u t t h e r e is n o t h i n g n o r m a t i v e a b o u t t h e m . By c o n t r a s t , D a v i d s o n a l s o rejects w o r d specific m e a n i n g - n o r m s . But h e d o e s so by r e j e c t i n g w o r d - s p e c i f i c m e a n i n g s . T h i s a l l o w s h i m t o give d u e r e c o g n i t i o n t o n o r m a t i v e a s p e c t o f m e a n i n g b y l o c a t i n g it w h e r e it really b e l o n g s : at t h e g l o b a l level.

Copyrighted

Material

388

Michael

Williams

4 Meaning and truth W h a t role d o e s t r u t h p l a y in all this? T h e a n s w e r m i g h t s e e m o b v i o u s : D a v i d s o n e x p l a i n s m e a n i n g s in t e r m s o f t r u t h - c o n d i t i o n s . In f a c t , h e identifies m e a n i n g s w i t h t r u t h - c o n d i t i o n s . H o r w i c h d o e s n e i t h e r o f t h e s e t h i n g s . S o D a v i d s o n n e e d s a s u b s t a n t i v e n o t i o n of t r u t h , w h i l e H o r w i c h d o e s n ' t . H o w e v e r , this l o o s e t a l k o f " e x p l a i n i n g m e a n i n g in t e r m s o f t r u t h - c o n d i t i o n s " is j u s t t h a t : l o o s e t a l k . We m u s t r e m e m b e r D a v i d s o n ' s special use of t h e p h r a s e " t h e o r y o f m e a n i n g . " A D a v i d s o n i a n " t h e o r y o f m e a n i n g " is a n I n t e r p r e t a t i o n . It gives t h e m e a n i n g s o f t h e s e n t e n c e s o f s o m e t a r g e t l a n g u a g e . B u t it is n o t a t h e o r y o f t h e n a t u r e o f m e a n i n g . It d o e s n o t e x p l a i n w h a t m e a n i n g is. A f o r t i o r i , it d o e s n o t e x p l a i n w h a t m e a n i n g is in t e r m s o f t r u t h - c o n d i t i o n s . Still, t h e q u e s t i o n o f t h e role p l a y e d by t r u t h in D a v i d s o n ' s a c c o u n t of m e a n i n g d e s e r v e s a closer l o o k . G i v e n D a v i d s o n ' s t h o u g h t t h a t m e a n i n g is w h a t e v e r i n t e r p r e t a t i o n c a p t u r e s , t h e r e a r e t w o w a y s in w h i c h t h e c o n c e p t of t r u t h m i g h t e n t e r his a c c o u n t of t h e n a t u r e of m e a n i n g : t h r o u g h his c o n c e p t i o n o f t h e p r o p e r f o r m o f an I n t e r p r e t a t i o n , o r via his m e t h o d o l o g y . N a t u r a l l y , t h e s e o p t i o n s a r e n o t exclusive. W i t h r e s p e c t t o t h e p r o p e r f o r m o f a n I n t e r p r e t a t i o n , it s e e m s t o m e t h a t a D a v i d s o n i a n t h e o r y o f m e a n i n g is " t r u t h - c o n d i t i o n a l " only in t h e sense t h a t it issues in m a t e r i a l b i c o n d i t i o n a l s . W h e n s y s t e m a t i c a l l y g e n e r a t e d b y a theory meeting the a p p r o p r i a t e methodological constraints, such Tarskian T - s e n t e n c e s c a n be t a k e n a s g i v i n g t h e m e a n i n g s o f t h e s e n t e n c e s o f t h e t a r g e t l a n g u a g e . I n t h i s way, as R a t t a n nicely p u t s it. t h e t h e o r e m s of a D a v i d s o n i a n t h e o r y of m e a n i n g convey a lot m o r e t h a n t h e y explicitly state,16 E v e n so. t h e w h o l e p o i n t of D a v i d s o n ' s s t r a t e g y is t o avoid a p p e a l t o a n y r o b u s t n o t i o n o f m e a n i n g . S o w h i l e it is t e m p t i n g t o d e s c r i b e D a v i d s o n ' s a p p r o a c h as o n e t h a t identifies m e a n i n g s w i t h t r u t h - c o n d i t i o n s , t h e t e m p t a t i o n is best resisted. D a v i d s o n d o e s n ' t identify m e a n i n g s w i t h a n y t h i n g : he d o e s a w a y w i t h t h e m . D a v i d s o n n e e d s " t r u e " precisely b e c a u s e he e s c h e w s " m e a n s . " D a v i d s o n d o e s n o t w a n t t o say: " L a neige est b l a n c h e " m e a n s t h a t s n o w is white. H o w e v e r , if we a r e t o r e p l a c e " m e a n s t h a t " w i t h a m a t e r i a l b i c o n d i t i o n a l , w e n e e d a s e n t e n c e o n t h e left side a n d " L a neige est b l a n c h e if a n d o n l y if s n o w is w h i t e " is n o t n o r m a l l y c o u n t e d g r a m m a t i c a l ( t h o u g h p e r h a p s it m i g h t be). T o solve this e x p r e s s i v e p r o b l e m , we a p p e n d " - is t r u e " t o a sentence-nominalization. D o w e u n d e r s t a n d s e n t e n c e s t h u s o b t a i n e d ? Well, c o n s i d e r h o w a d e f l a t i o n i s t t a k e s t h e a t t r i b u t i o n o f t r u t h ( o r falsity) t o a s e n t e n c e he d o e s n o t u n d e r s t a n d . H e takes such an attribution to imply that the sentence he does n o t u n d e r s t a n d is e q u i v a l e n t in m e a n i n g ( f o r p r e s e n t i n t e n t s a n d p u r p o s e s )

Copyrighted

Material

Meaning, truth and normativity

389

to some sentence he does understand and that he holds to be true (or false) in a deflationary sense. This shows how a deflationist should read the theorems of a Davidsonian theory of meaning. Of course, the theorems do not explicitly say anything about meaning. However, as we have seen, they convey more than they state. If I am right about this, Davidson's truth-conditional approach to meaning is fully available to deflationists about truth. But of course, this is just a way of saying that the approach is not seriously truth-conditional in the first place. Here, I think, Davidson misunderstands himself. Am I trying to have it two ways here, taking a Davidsonian theory of meaning to employ a prior understanding of truth to tell us about meaning and a prior grasp of meaning to tell us about truth (as applied to sentences not antecedently understood)? Not really. As I read Davidson, he uses the notion of interpretation to tell us something about meaning and something about truth. There is no circularity here. 1 shall return to this point below. A final point before moving on: I said that a Davidsonian T-sentence, appropriately arrived at, although it makes no explicit mention of meaning, can be taken as interpreting or giving the meaning of the sentence named to the left of the biconditional. Does this introduce meanings through the back door, by way of establishing their identity conditions? I do not think so. Nothing in Davidson's account of interpretation suggests that there is always a unique best way to interpret another speaker. Nothing excludes the thought that the best way to interpret a speaker, on a given occasion, may reflect the interests of the interpreter. Davidson is not committed to capturing a strong, objective notion of interpersonal synonymy. In this connection, it is important not to take the phrase "theory of meaning" (in Davidson's sense) too seriously. This is perhaps a further point of contrast with Horwich. Where Horwich thinks of his "acceptance-properties" or "core regularities" as a kind of more-or-less permanent substructure that explains the manifest features of a speaker's linguistic behavior, Davidson thinks of interpretation as an ongoing process of adjusting one' speech to that of another. What we implicitly entertain, with respect to another speaker, is not a more or less fixed theory of meaning but a sequence of "passing theories." Davidson sometimes puts this point in the form of the dramatic claim that there is "no such thing as language" (where "language" is thought of as constituted by just the kind of underlying structure of rules that Horwich is after). 1 7 If Davidson makes real explanatory use of truth, it must be in his methodology, which in any case carries most of the burden of his implicit theory of the nature of meaning. Certainly, Davidson's discussions of his methodology involve a lot of truth-talk. But this may not mean much. As we noted at the very outset, deflationists do not prohibit our making use of truth-talk in explanations, provided that the concept of truth plays only the expressive role they assign to it. Their objection is to the concept of truth's functioning as a theoretical primitive. So it is far from clear, to me anyway,

Copyrighted

Material

390

Michael

Williams

that the concept of truth plays any role in Davidson's account of his methodology that deflationists cannot happily countenance. Davidson appeals to consistency, and to occasion sentences and their causes: no need to invoke truth here. He also insists on the need to attribute lots of truths. However, this last requirement involves the notion of truth only in the generalizing capacity stressed by deflationists. Indeed, it could as well be stated in terms of agreement ("truth by our lights." as Davidson himself likes to say). I have referred from time to time to the theory of the nature of meaning implicit in Davidson's interpretationist stance. At this point, it will be useful to say more explicitly what I take Davidson's general picture to amount to. For Davidson, as for Quine, considering the methodology of radical interpretation is a way of getting at the nature of meaning. The factors we must attend to in constructing a theory of meaning, in the context of radical interpretation, are the factors that constitute meaning. From this stratospheric height, the differences between Horwich and Davidson do not seem so great. Indeed, the picture of meaning that emerges from Davidson's discussion of radical interpretation can quite naturally be thought of as a usetheory. Horwich and Davidson offer competing articulations of a view of meaning that Robert Brandom calls "broad inferentialism." 1 8 On this view, words derive their meanings - to whatever extent their meanings are determinate - from their contributions to the meanings of sentences. Sentences derive their meanings from their "conditions and consequences." These will include their relations to other sentences: what follows from them and what they follow from. But the notion of conditions must be broadened to include, for some sentences at least, environmental circumstances ("Dog" and dogs again). And consequences must include actions. (I respond to the vicious-looking dog by running away.) For Davidson, as for Brandom, interpreting someone's speech involves attributing desires as well as beliefs: a complete psychology. Seeing how closely Brandom's approach to meaning follows Davidson's should reinforce our sense that Davidson's own approach is not seriously truth-conditional. A seriously truth-conditional theory would take the notion of reference (or satisfaction) as basic, perhaps explaining it in causal terms. Davidson does not do this. Davidson sees reference-relations as abstracted from the broad patterns of use that constitute them. In consequence, Davidson's conception of meaning is thus a far cry from a "building-block" theory. It does not involve an explanatorily robust notion of truth-conditions, in the way that a formally similar approach incorporating a causal theory of reference would do. However, the holistic picture of meaning we find in Davidson might itself be thought to pose a problem for my attempt to narrow the distance between Davidson and the deflationists. This is what I take Davidson himself to hold. When Davidson says that we must take truth as primitive, he means that truth cannot be defined. Indeed, in Davidson's view, the lesson to be drawn from the failure of the Socratic quest for definitions is that no philosophically

Copyrighted

Material

Meaning, truth and normativity

391

i n t e r e s t i n g c o n c e p t c a n be d e f i n e d . B u t he d o e s n o t t a k e t h i s t o i m p l y t h a t philosophically interesting concepts cannot be illuminated. However, we d e e p e n o u r u n d e r s t a n d i n g o f s u c h c o n c e p t s b y b r i n g i n g t o light t h e i r links with other i m p o r t a n t concepts, rather than by reducing them to simple f o r m u l a e . In t h e c a s e o f t r u t h , t h e r e l e v a n t n e i g h b o r i n g c o n c e p t s a r e belief a n d m e a n i n g . D a v i d s o n ' s m e t h o d o l o g y o f i n t e r p r e t a t i o n ( a n d t h u s his i m p l i c i t t h e o r y of t h e n a t u r e o f m e a n i n g ) involves a t t r i b u t i n g lots o f t r u e beliefs t o t h e p e r s o n w h o s e s p e e c h we a r e i n t e r p r e t i n g . T o a p p r e c i a t e why this is n e c e s s a r y j u s t is t o g a i n insight i n t o t r u t h , by s e e i n g h o w h o l d i n g t r u e beliefs p l a y s a n i n e l i m i n a b l e role in t h e c o n s t i t u t i o n o f m e a n i n g . D a v i d s o n t h i n k s t h a t H o r w i c h m i s s e s all this precisely b e c a u s e , in his q u e s t t o u n d e r s t a n d t r u t h , h e is still a f t e r a s i m p l e rule, even if t h e rule h e p r o p o s e s is n o t exactly a d e f i n i t i o n . 1 9 All d e f l a t i o n i s t s m a k e this m i s t a k e . A d e f l a t i o n i s t will reply t h a t h e is n o t c o m m i t t e d t o d e n y i n g t h a t m e a n i n g s o m e t h i n g b y o n e ' s w o r d s involves b e l i e v i n g l o t s of t r u t h s . H i s p o i n t is t h a t this h a s n o t h i n g t o d o w i t h t h e n a t u r e of t r u t h . O n his view, t r u t h d o e s n o t h a v e a n i n t e r e s t i n g n a t u r e . R a t h e r , t h e essential f e a t u r e s o f " t r u e " a r e c o m pletely c a p t u r e d b y a rule o f use t h a t involves a c c e p t i n g all n o n - p a r a d o x i c a l i n s t a n c e s o f s o m e f a v o r e d f o r m o f t h e e q u i v a l e n c e - s c h e m a . It is in this spirit t h a t H o r w i c h c l a i m s t h a t t h e c o n c e p t o f t r u t h h a s " a c e r t a i n p u r i t y , " in t h a t " o u r u n d e r s t a n d i n g o f it is fairly i n d e p e n d e n t of o t h e r i d e a s . " But h o w is s u c h a c l a i m t o b e j u s t i f i e d ? W h y is H o r w i c h ' s r u l e of use c o n s t i t u t i v e o f t h e m e a n i n g o f " t r u e , " w h e r e a s t h e t h i n g s D a v i d s o n tells us a b o u t t h e role o f t r u e beliefs in t h e c o n s t i t u t i o n o f m e a n i n g a r e j u s t f u r t h e r i n f o r m a t i o n ? W h y d o n ' t t h e t h i n g s D a v i d s o n tells us i l l u m i n a t e t h e c o n c e p t o f t r u t h ? M o r e p o i n t e d l y , w h a t could u n d e r w r i t e t h e c l a i m t h a t t h e y d o n o t d o t h i s except t h e a n a l y t i c - s y n t h e t i c d i s t i n c t i o n ? It l o o k s as t h o u g h H o r w i c h ' s c o m m i t m e n t t o m e a n i n g s u n d e r l i e s n o t o n l y his a t t i t u d e t o n o r m a t i v i t y b u t his d e f l a t i o n a r y a t t i t u d e t o t r u t h a s well. T h i s is fine f o r H o r w i c h . b u t n o t f o r me. I w a n t t o agree with H o r w i c h a b o u t t r u t h but with D a v i d s o n a b o u t meanings. We m i g h t p e r h a p s c o n c e d e t h e p o i n t t h a t , in t r a c i n g t h e c o n n e c t i o n s b e t w e e n belief, t r u t h a n d m e a n i n g , we a r e p u t t i n g s o m e flesh o n t h e b a r e b o n e s o f d e f l a t i o n a r y t r u t h , w i t h o u t r i s k i n g m u c h d a m a g e t o t h e spirit o f d e f l a t i o n i s m . We c o u l d a f f o r d this c o n c e s s i o n b e c a u s e t h e w a y in w h i c h we w o u l d n o w be " i n f l a t i n g " t h e c o n c e p t o f t r u t h w o u l d n o t i n v o l v e o u r t r e a t ing it a s a n e x p l a n a t o r y p r i m i t i v e in t h e t h e o r y of m e a n i n g . D a v i d s o n ' s u s e o f " t r u e , " b o t h in p a r t i c u l a r I n t e r p r e t a t i o n s a n d in his m e t h o d o l o g y o f i n t e r p r e t a t i o n , w o u l d still b e expressive r a t h e r t h a n e x p l a n a t o r y . T h i s fact w o u l d n o t b e c h a n g e d b y o u r d e c i d i n g t o t r e a t t h e p o i n t t h a t we c a n m a k e sense of o t h e r s p e a k e r s o n l y a g a i n s t a b a c k g r o u n d of m a s s i v e a g r e e m e n t in beliefs ( " t a k i n g lots o f t h e i r beliefs t o be t r u e " ) as f u r t h e r i l l u m i n a t i n g t r u t h b y t r a c i n g its c o n n e c t i o n s t o belief a n d m e a n i n g . H o w e v e r , w h i l e we m i g h t , as d e f l a t i o n i s t s , b e able t o m a k e s o m e c o n c e s s i o n s a l o n g t h e s e lines, t h e r e is n o c o m p e l l i n g r e a s o n t o d o so. T h e d i f f i c u l t y a b o u t t h e " p u r i t y " o f t r u t h is m o r e a p p a r e n t t h a n real.

Copyrighted

Material

392

Michael

Williams

D a v i d s o n a g r e e s t h a t d e f l a t i o n i s t s d o n o t m e a n t o define t r u t h . B u t , as w e h a v e seen, he t h i n k s t h a t t h e y s h a r e s o m e t h i n g of t h e a t t i t u d e t h a t l e a d s t o a quest for a definition: the urge to take a philosophically interesting concept a n d s u m it u p in a n e a t f o r m u l a . H o w e v e r , d e f l a t i o n i s t s n e e d h a v e n o general p e n c h a n t f o r n e a t f o r m u l a e , f o r t h e y see " t r u e " as special. B e c a u s e it p l a y s a p u r e l y expressive role in o u r l a n g u a g e , t r u t h is a k i n d of logical c o n c e p t . It is u s e f u l h e r e t o c o m p a r e w h a t d e f l a t i o n i s t s say a b o u t t r u t h w i t h w h a t R o b e r t B r a n d o m says a b o u t t h e s e n t e n t i a l c o n n e c t i v e s . 2 0 A c c o r d i n g t o B r a n d o m , t h e i r use, t o o . is essentially expressive. T h e y a l l o w us t o e x p r e s s a s c l a i m s i n f e r e n t i a l c o m m i t m e n t s t h a t w e c o u l d o t h e r w i s e o n l y d i s p l a y in p r a c t i c e . F o r e x a m p l e , w e c a n r e a c t t o l e a r n i n g t h a t - p by t a k i n g o u r s e l v e s t o be c o m m i t t e d t o h o l d i n g t h a t - q : w e d i s p l a y o u r i n f e r e n t i a l c o m m i t m e n t b y i n f e r r i n g q f r o m p. But w i t h t h e c o n d i t i o n a l in h a n d , w e c a n e x p r e s s t h a t i n f e r e n t i a l c o m m i t m e n t a s a n explicit c l a i m : if p t h e n q. All w e n e e d is t h e m a t e r i a l c o n d i t i o n a l . L o g i c a l n o t i o n s , like t h e m a t e r i a l c o n d i t i o n a l , c a n s o m e t i m e s be s u m m e d u p in n e a t f o r m u l a e : t r u t h tables, o r G e n t z e n - t y p e rules. L o t s of o t h e r n o t i o n s c a n ' t . O b v i o u s l y . D a v i d s o n d o e s n o t o b j e c t t o s u m m i n g u p t h e b e h a v i o r of t h e m a t e r i a l c o n d i t i o n a l in s u c h a f o r m u l a . O n t h e c o n t r a r y , a D a v i d s o n i a n t h e o r y of m e a n i n g will c o n t a i n a r e c u r s i o n c l a u s e d o i n g j u s t t h a t . A c c o r d i n g l y , j u s t as d e f l a t i o n i s t s n e e d n o t be u n i f o r m l y d r a w n t o n e a t f o r m u l a e , D a v i d s o n i a n s c a n n o t be u n i v e r s a l l y h o s t i l e to them. In criticizing H o r w i c h , I a r g u e d t h a t t h e c o n c e p t o f t r u t h is t o o special t o s u p p o r t t h e c l a i m t h a t a t h e o r y of m e a n i n g m u s t i n v o k e s o m e g e n e r a l f o r m of t h e a n a l y t i c - s y n t h e t i c d i s t i n c t i o n . By p a r i t y of r e a s o n i n g , t h a t d i s t i n c t i o n n e e d n o t b e i n v o k e d by d e f l a t i o n i s t s in d e f e n s e o f w h a t t h e y see as t h e " p u r i t y " o f t h e c o n c e p t o f t r u t h . T h e special c h a r a c t e r of t r u t h is e n o u g h all b y itself.

5 Conclusion W h y d o e s D a v i d s o n t h i n k t h a t he h a s t o t a k e t r u t h as p r i m i t i v e ? O n e r e a s o n is t h a t he o v e r r e a c t s t o a n a r g u m e n t a d v a n c e d by D u m m e t t . 2 1 ( H o r w i c h f o l l o w s h i m in this.) A c c o r d i n g to T a r s k i , a " m a t e r i a l l y a d e q u a t e " t h e o r y of t r u t h , f o r l a n g u a g e L s h o u l d e n t a i l all i n s t a n c e s o f t h e s c h e m a : (T) S is t r u e - i n - L if a n d o n l y if p. We f o r m a T - s e n t e n c e b y a t t a c h i n g , t o t h e left of t h e b i c o n d i t i o n a l , t h e m e t a linguistic p r e d i c a t e " is t r u e - i n - L " t o a s t r u c t u r a l - d e s c r i p t i v e n a m e o f s o m e s e n t e n c e o f L. while t o t h e r i g h t we p u t a s e n t e n c e of t h e m e t a - l a n g u a g e t h a t is a c o r r e c t t r a n s l a t i o n of L. A c c o r d i n g l y . T a r s k i s e x p l a n a t i o n o f ( T ) a p p e a l s t o t h e n o t i o n of m e a n i n g , h e r e in t h e guise of " c o r r e c t t r a n s l a t i o n . " It foll o w s t h a t if we t a k e " C o n v e n t i o n T." o r a n y g e n e r a l i z e d f o r m of it. t o d e f i n e

Copyrighted

Material

Meaning, truth and normativity

393

truth, we must presuppose a prior grasp of the notion of meaning. Conversely, if we think that T-sentences can give meanings, we must take for granted a prior grasp of the notion of truth. Thus in modelling a "theory of meaning" on a Tarskian "theory of truth," Davidson faces a choice: either his theory defines truth or it explains meaning. But it cannot do both. 2 2 Davidson chooses to presuppose truth. The distinction between Interpretations ("theories of meaning") and theories of the nature of meaning should encourage us to look askance at Dummett's argument. As we have seen, Davidson doesn't explain meaning in any serious way in terms of truth, either at the level of particular Interpretations or in his methodology. But just because of this, we can take Interpretations to tell us something about meaning and something about truth. Interpretations tell us something about meaning in that they tell us what foreign sentences mean, though they do not tell us what meaning is. But since they tell us what foreign sentences mean by matching them with sentences that we already understand, and can thus hold true or false (in a deflationary sense), they also tell us something about truth. They take us from a primitive to an extended notion of (deflationary) truth. However, they take for granted a prior grasp of (deflationary) truth for the home language. Accordingly, they do not "tell us what truth is," any more than they tell us what meaning is. The problem of explaining truth in terms of meaning and meaning in terms of truth does not arise. 23 To sum up, we should side with Horwich about truth but with Davidson about meaning. We might call the result "deflationary Davidsonianism," an ugly name for an attractive position. Notes 1 An earlier version of this paper was presented at a symposium on truth at the American Philosophical Association Central Division Meeting. April 2002. I would like to thank my fellow symposiasts, the late Donald Davidson and Paul Horwich. as well as several members of the audience, for their stimulating comments and criticism. 2 See Horwich (1998b). 3 For more on the two aspects of deflationism, theoretical and functional, see Williams (2002). In my view, the claim that truth is "not a substantive property" is just another way of saying that the function of truth-talk is expressive rather than explanatory. It is not an explanation of why truth-talk has no explanatory uses. That claim is justified by Occam's razor: we don't need any richer notion of truth, even supposing that we know what one would look like. 4 For his attitude to deflationism, see Davidson (1996). 5 See Horwich (1998a). 6 See Rattan (unpublished). Rattan's paper responds critically to Williams (1999). Rattan's criticisms forced me to clarify my views concerning the compatibility of a deflationary view of truth with a broadly Davidsonian approach to meaning and were very helpful to me in writing this paper. See note 22 below. 7 See Horwich (1998a: 1). 8 See Horwich (1998a: 4).

Copyrighted

Material

394

Michael

Williams

9 Horwich's account of "refers" parallels his account of "true." Our understanding of "refers" reflects an underived inclination to accept instances of the schema "Tokens of *n* refer to n." (Refinements are required to take account of singular terms that lack referents and reference-talk as applied to terms from foreign languages.) The function of the concept of reference is communicative: we can acquire information from other people with whom we d o not share all beliefs, and who may therefore describe objects in different ways from ourselves. For details, see Horwich (1998a: 115ff.). 10 See Horwich (1998a: 44-6). 11 See Horwich (1998a: 45). 12 See G u p t a (1993). 13 See the essays in Davidson (2001). 14 See Horwich (1998a: 184-94). 15 See for example Sellars (1997: sects 30 1 and 34-5). 16 Rattan (unpublished). 17 See Davidson (1986). 18 Brandom develops his own version of this approach in Brandom (1994: 130ff.). 19 Davidson (1996). 20 Brandom (1994: 102-11). 21 See Dummett (1959). 22 Rattan (unpublished) calls this "Dummett's Dilemma." 23 Here as elsewhere, my argument depends crucially on the distinction between a theory of meaning (in Davidson's sense) and a theory of the nature of meaning. Rattan (unpublished) argues that I should make a parallel distinction between a theory of truth (for a given language) and a theory of the nature of truth. If 1 did. he claims, I would find that truth is just as substantive as meaning. I think that, at bottom, this objection is more or less the same as one 1 have already considered: that in tracing the connections between meaning, belief, and truth, we are putting flesh on the bones of the (minimalist) concept of truth. Accordingly I have already answered this objection in saying that we can think of a Davidsonian theory of meaning, viewed as a theory of truth, as a theory of extended disquotational truth. From the Davidsonian/deflationary standpoint I am recommending, in talking about interpretation, the nature of meaning and the nature of extended deflationary truth we are talking about the same thing.

References Brandom. R. (1994) Making It Explicit. Cambridge, MA: Harvard University Press. Davidson, D. (1986) "A Nice Derangement of Epitaphs," in E. Lepore (ed.) Truth and Interpretation: Perspectives on the Philosophy of Donald Davidson, Oxford: Blackwell. pp. 434-6. (1996) "The Folly of Trying to Define Truth." Journal of Philosophy, 93: 263-79; reprinted in M. Lynch (ed.) The Nature of Truth Cambridge, MA: M I T Press, pp. 623-40. (2001) Inquiries into Truth and Interpretation, rev. edn, Oxford: Oxford University Press. Dummett, M. (1959) "Truth." Proceedings of the Aristotelian Society. 59: 141 62; reprinted in M. Dummett (1978) Truth and other Enigmas, London: Duckworth, pp. 1 24.

Copyrighted

Material

Meaning, truth and normativity

395

G u p t a , A. (1993) "A Critique of Deflationism," Philosophical Topics, 21: 57-81; reprinted in M. Lynch (ed.) The Nature of Truth, Cambridge, MA: M I T Press, pp. 527-57. Horwich, P. (1998a) Meaning, Oxford: Oxford University Press. (1998b) Truth. 2nd edn, Oxford: Oxford University Press. (2001) "A Defence of Minimalism," in M. Lynch (ed.) The Nature of Truth, Cambridge. MA: M I T Press, pp. 559-77. Rattan, G. (unpublished) "Semantic Deflationism and Dummett's Dilemma," ms. Sellars, W. (1997) Empiricism and the Philosophy of Mind, Cambridge, MA: Harvard University Press. Williams, M. (1999) "Meaning and Deflationary Truth," Journal of Philosophy, 96: 545-64. (2002) "On Some Critics of Deflationism," in R. Schantz (ed.) Truth, Berlin: Walter de Gruyter. pp. 146-58.

Copyrighted

Material

Index of names

Alder, J. 277n Alston, W.P. 4. 10. 17, 19 23, 25-26. 29 30n. I73n. 213. 223n. 250-53, 277n. 315 Anderson, S.R. 147. I54n. Audi. R. 277n Austin. J.L. 9, 16-17. 106-7. 289 Ayer, A.J. 73. 86n, 195n. 223n. 302n Baker, M.C. 140. 152 53n. Bar-On, D. 4. 86-87n Barsalou, L.W. 303n Bell. D. 291 Belnap, N.D. 336. 343 44n Belshaw, C. 127n Berlin, I. 343n Bernstein, J.B. 140. 150 Beth. E.W. 343n Blackburn. S. 86-87n Blandshard, B. 309 Block, N. 172n BonJour, L. 315 Bradley, F.H. 309 Brandom, R. 79-86. 87n. 117. 210. 254-55. 258. 281, 310. 318n. 390. 392. 394n Burge, T. 275. 276n. 375n By bee. J. L. 300 Carstairs-McCarthy. A. 132. 139. 140 41. 152 -53n. Cheney, D.L. 138 Chisholm, R. 153n, 316-17 Chomsky, N. 154n. 162. 173n Chrisman, M. 88n Coady, C.A.J. 273 74. 277n Collins. J. 4, 157. 166. 173 75n David, M. 86n. 193 94n Davidson, D. 125, 154n. 173n. 223n. 253-54, 259. 262. 269. 308. 312. 318n. 341 n. 350-57. 362- 63. 369. 374n. 377-78. 382-84. 387-93. 393-94n

Copyrighted

De Rose. K. 231. 276n Dewey. J. 310-11. 318n. 326. 342n Devitt. M. 173n Dretske. F. 173n Dummett, M. 5n. 70-71. 87n. 95. 104n. 113 16. 118. 123. 127n. 201. 251-53. 374n. 276n. 286. 272-73. 277n. 291. 308-9. 311 13, 318n, 340. 343n. 355, 359-60. 367. 373, 375n. 392. 393. 394n Earman. J. 344n Einstein. A. I29n. 185-87 Faueonnier. G. 297 Field. H. 67. 86n, 164. 172 74n. 222n Fleishman. S. 300 Fodor. J. 150. 172-73n Foster. J.A. 351. 354. 363 Frege. G. I 5. 5n. 28-30. 69 70, 73 -75, 77-78. 81. 87n. 90-91, 95-96. l03-4n. 128n. 183. 193n, 195n. 250. 282.-3. 289. 290 94. 301. 302-4n. 308. 357.-62.374n Flicker. E. 275 Fuhrmann. A. 248n Gardenfors, P. 343n Gazdar. G. 247n Geach. P. 87n Gibbard. A. 86n Gigerenzer.G. 139 Ginet. C. 317. 319n Goldman. A. 315 Goldstone. R.L. 303n Graham. P. 277n Greenberg. R. 303n Greimann. D. 4. 76. 87n. 103-4n.. 193n. 291. 303n. 341n Grice. H P. 17-19. 39-40n. Ill 12, 226-29, 232. 237. 244. 247n. 276n. 357 Grover. D. 86n. 164. 222n. 280-81

Material

Index of names

397

Ludlow, P. 153 Lumsden, D. 127n. Lynch, M. 223n

Grzegorczyk, A. 343n Gupta. A. 104n. 174n. 195n. 394n Hacking, I. 138 Hale, k . 144 Hanna, R. 303n Hanson, W. 341n. 345n Hare, R.M. 260 Harman, G. 172n Heck, R. 4, 350, 354, 357. 364. 367-68, 370-71, 373, 374-75n Hempel, C. 309 Higginbotham, J. 173n, 350, 369-7, 374-75n Hill. C. 67, 86n, 222n Hinst. P. 57n Hinzen, W. 4, 149. 153-55n Hofweber. T. 88n Holt, L. 304n Holton, R. 127n Hornsby. J. 290 Hornstein, N. 154n.. 173n Horwich. P. 64-65, 71, 86n. 102, 104n, 166, 173n, 185-86. 195 96n, 204-5, 209, 215, 222-23n, 245. 281, 302n, 318n, 324-25, 341-42n, 377-83, 385-87, 391-93, 393 94n Howson, C. 344n Huang, J.-T. 145-46 Jackendoff. R. 140, 144, 153n. Jackson, F. 63, 86n James, W. 201, 310. 325, 332 Janich. P. 57n. Joachim, H.H. 309 Johnson, S. 277n Kant, I. 1-3, 5n, 109. 172n, 273, 275n, 280, 283-90, 294, 302-3n, 311 Kelly, T. 228, 246 47 Kemp, G. 4, 5n Keyser. S.J. 144 Kirkham, R. 99. 87n. 104n Korsgaard, C. 276n Kovach, A. 4. 213. 223n Krabbe, E. 254, 256-57, 276-77n Kripke, S.A. 343n

McDowell, J. 163, 173n, 50, 360-61, 373, 374-75n MacFarlane, J. 254-55 McGinn, C. 172n McGrath, M. 223n Maher, P. 129n. Meggle, G.58n. Metschl, U. 4 Misak, C.J. 322, 329-30. 338-41, 341-43n, Moore. G.E. 211, 324 Moser, P. 315. 318n Mou, B. 4. 193n, 196n Neurath, O. 298. 309 Nichols, S. 277n Nuyts. J. 300-301, 304n Olsson. E.J. 342n Oppy. G. 4, 232 Palmer. F.R. 300, 304n Papafragou, A. 304n Parent, T. 87-88n Patterson, D. 87-88n Peacocke, C. 277n Peirce. C.S. 310, 321-22. 326-29. 332, 339-41, 34 i —43n, 345n Percival. P. 127n Perner, J. 158 Pettit. D. 87-8811 Pinker, S. 144 Pollock, J. 15, 317, 319n Popper, K.R. 342n Price, H. 220, 222, 222n, 227-28. 243-46 Pryor. J. 277n Putnam. H. 173n, 280, 310-15, 317 18. 318n, 342n Pylyshyn. Z. 158 Quine, W.V. 86n, 173n, 183. 195-96n, 222n. 262, 273. 296, 313. 352. 362. 381. 383, 390

Künne. W. 154n Leeds. S. 86n. 173n Leite. A. 255 Lepore, E. 172n Levi, 1. 139, 340-41. 342-43n, 345n Lewis, D. 15-16. 28. 39, 277n, 374n Lightfoot. D. 47, 54n. Loar, B. 172n

Rabinowicz, W. 334-36, 343n, 352, 364, 383 Ramsey, F.P. 1, 66-67, 101. 139. 223n. 302n Rattan. G. 393-94n Rawls, J. 260-61, 276n Raz, J. 252-53, 260, 268. 276n Read. S. 342n

Copyrighted

Material

398

Index of names

Recanti, F. 173 74n Rescher, N. 57n Rescorla, M. 4. 375n Resnik. M. 86n Richard, M. 375n Ricketts, T. 291 Rizzi, L. 139 Rorty, R. 153n, 220. 222. 222n Rumfitt, I. 350, 375n Russell. B. 128n. 201. 307 317. 326. 318n Schantz, R. 4, 5n. 318n Schiffer, S. 17 19. 30n Searle. J. 4, 9. 15. 17. 98. 104n. 250-53. 260-61, 267-68 Sellers, W. 250. 254, 386. 394n Sen, A. K. 344n Serrano, M.J. 302 Seyfarth. R . M . 138 Sher, G. 4, 293 94. 296, 298-99, 302-3n Siegwart, G. 4. 57-58n, I04n. 193n Simmons, K. 4, 87 Soames, S. I04n. 173n, 222n. 351. 355, 363, 367. 374 Stanley. J. 375n Stich, S. 172-73n, 277n Strawson, P.F. 87n. 178, 188-89. 195-96n. 223n. 308. 318n. 353-54 Sweetser. E. 300 Swoyer. C. 304n Talmy, L. 303n Tanesini, A. 23. 26 28

Copyrighted

Tarski. A. 32-33, 102-3, 104n, 178. 180 84. 186-87. 193. 194-95n. 302n. 349. 388. 392-93 Tennant. N. 57n. Tolstoy. L. 299 Turner. M. 297 Urbaeh. P. 344n Uriagereka, J.146. 149. 153-54n Van Cleve. J. 303n Vanderveken, D. 15 Von Wright. G.H. 301 Walton. D. 254, 276-77n Wansing. H. 343n White. A.R. 306n Wiggins. D. 330 Williams. B. 39. 259. 276n. 304n. 341 n Williams. C.J.F. 302n Williams. M. 4. 61, 86n. 104n, 255. 382. 391. 393n Williamson. T. 118-20. 122-24. 126-27. 128n. 227, 231-35, 237- 39. 243. 247-48n. 250-53 Winch. P. 272- 74 Wittgenstein, L. 112-13. 135. 173n. 291. 317. 318n. 307, 359 Woleiiski. J. 290, 293 Wrisht. C. 63. 96,127n. 201. 207-16. 218. 220-21. 222-24n. 227-29. 241—43. 250. 308-9. 314. 318n. 332-33. 343n. 375n Wright. C D. 4. 304n

Material

r

a

Related titles from Rout ledge

Routledge Series in Contemporary

Philosophy

Fiction, Narrative, and Knowledge A sense of the world Edited by John Gibson, Wolfgang Huemer, and Luca Pocci

A t e a m of leading c o n t r i b u t o r s f r o m b o t h p h i l o s o p h i c a l a n d literary b a c k g r o u n d s h a v e been b r o u g h t t o g e t h e r in this impressive b o o k t h a t e x a m i n e s h o w w o r k s of literary fiction c a n be a s o u r c e of k n o w l e d g e . It a n a l y z e s the i m p o r t a n t t r e n d s in this c u r r e n t p o p u l a r d e b a t e , which h a v e never b e f o r e been collected t o g e t h e r in a single v o l u m e . T h e i n n o v a t i v e f e a t u r e of this v o l u m e is t h a t it mixes w o r k by literary theorists a n d s c h o l a r s with w o r k of a n a l y t i c p h i l o s o p h e r s t h a t c o m b i n e d t o g e t h e r p r o v i d e a c o m p r e h e n s i v e s t a t e m e n t of the v a r i e t y of w a y s in which w o r k s of fiction c a n e n g a g e q u e s t i o n s of worldly interest. It uses t h e p r o b l e m of c o g n i t i v e v a l u e to explore: • l i t e r a t u r e ' s c o n t r i b u t i o n to ethical life • l i t e r a t u r e ' s ability to e n g a g e in social a n d political c r i t i q u e • the role n a r r a t i v e plays in o p e n i n g u p possibilities of m o r a l , a e s t h e t i c , experience a n d s e l f h o o d . T h i s r e m a r k a b l e v o l u m e will a t t r a c t a b r o a d a u d i e n c e f r o m a r o u n d t h e w o r l d a m o n g p r o f e s s i o n a l s a n d s t u d e n t s with a g r e a t variety of interests.

I S B N : 978-0-415-70191-4 ( h b k )

A v a i l a b l e at all g o o d b o o k s h o p s F o r o r d e r i n g a n d f u r t h e r i n f o r m a t i o n please visit: www.routledge.com

V

J

Copyrighted

Material

Copyrighted

Material

r

a

Related titles from Rout ledge

Routledge Series in Contemporary

Philosophy

Einstein, Relativity and Absolute Simultaneity Edited by William Lane Craig and Quentin Smith

2005 m a r k e d t h e c e n t e n a r y of o n e of t h e m o s t r e m a r k a b l e p u b l i c a t i o n s in t h e h i s t o r y of science, A l b e r t E i n s t e i n ' s " O n t h e E l e c t r o d y n a m i c s of M o v i n g B o d i e s , " in w h i c h he p r e s e n t e d a t h e o r y t h a t later c a m e t o be k n o w n as t h e Special T h e o r y of Relativity. T h i s 1905 p a p e r is widely r e g a r d e d as h a v i n g d e s t r o y e d the classical c o n c e p t i o n s of a b s o l u t e t i m e a n d space, a l o n g with a b s o l u t e s i m u l t a n e i t y a n d a b s o l u t e length, w h i c h h a d reigned in physics f r o m the times of G a l i l e o a n d N e w t o n t o t h e d a w n of t h e t w e n t i e t h c e n t u r y . A s we e m b a r k u p o n a n e w c e n t u r y , t h e Special T h e o r y is n o w 100 years old, a n d a g r e a t deal h a s t r a n s p i r e d in b o t h p h i l o s o p h y a n d physics since its first p u b l i c a t i o n . T h i s v o l u m e is a timely r e a p p r a i s a l t h e t h e o r y ' s c e n t r a l claims, especially c o n c e r n i n g t h e e l i m i n a t i o n of a b s o l u t e time a n d a b s o l u t e s i m u l t a n e i t y . T h i s collection d r a w s t o g e t h e r essays by b o t h p h i l o s o p h e r s a n d p h y sicists, a n d reflects t h e c u t t i n g edge of research a n d t h o u g h t o n t h e q u e s t i o n of a b s o l u t e s i m u l t a n e i t y . T h e issues discussed in t h e b o o k include A s p e c t ' s c o n f i r m of Bell's t h e o r e m , D e B r o g l i e - B o h m ' s q u a n t u m m e c h a n i c s , t h e privileged c o s m i c t i m e series in a F r i e d m a n universe, L o r e n t z ' s ideas a n d n e o - L o r e n t z i a n t h e o r y a n d o t h e r r e l e v a n t issues. A l m o s t all t h e c o n t r i b u t o r s a r e c o n v i n c e d t h a t the received view t h a t s i m u l t a n e i t y is n o t a n a b s o l u t e r e l a t i o n is n o t only u n w a r r a n t e d b u t false, a n d it is h o p e d t h a t this collection will s t i m u l a t e discussion a m o n g b o t h p h i l o s o p h e r s a n d physicists c o n c e r n i n g t h e w a r r a n t f o r a n d p r o b l e m s with a s s e r t i o n s of t h e relativity of s i m u l t a n e i t y o n t h e basis of Einstein's theory.

I S B N : 978-0-415-70174-7 ( h b k ) I S B N : 978-0-203-70005-1 (ebk)

A v a i l a b l e a t all g o o d b o o k s h o p s F o r o r d e r i n g a n d f u r t h e r i n f o r m a t i o n please visit: www.routledge.com

V

J

Copyrighted

Material

Copyrighted

Material

r

a

Related titles from

Routledge

Routledge Series in Contemporary

Philosophy

A Pragmatic Philosophy of Democracy Communities of inquiry Robert B. Talisse

In recent years t h e r e h a s been a renewed interest in A m e r i c a n p r a g m a tism. In political p h i l o s o p h y , t h e revival of p r a g m a t i s m h a s led to a new a p p r e c i a t i o n f o r the d e m o c r a t i c t h e o r y of J o h n D e w e y . In this b o o k . R o b e r t B. Talisse a d v a n c e s a series of pragmatic argum e n t s a g a i n s t D e w e y a n d e m o c r a c y . P a r t i c u l a r l y , Talisse a r g u e s t h a t D e w e y a n d e m o c r a c y c a n n o t a d e q u a t e l y recognize pluralism, t h e fact t h a t intelligent, sincere, a n d w e l l - i n t e n t i o n e d p e r s o n s c a n d i s a g r e e s h a r p l y a n d r e a s o n a b l y o v e r m o r a l ideals. D r a w i n g u p o n t h e e p i s t e m o l o g y of the f o u n d e r of p r a g m a t i s m , C h a r l e s S. Peirce, Talisse d e v e l o p s a c o n c e p t i o n of d e m o c r a c y t h a t is a n t i - D e w e y a n but n o n e t h e l e s s p r a g m a t i s t . Talisse t h e n b r i n g s the P e i r c e a n view into critical c o n v e r s a t i o n with c o n t e m p o r a r y d e v e l o p m e n t s in d e m o c r a t i c t h e o r y , i n c l u d i n g d e l i b e r a t i v e d e m o c r a c y , R a w l s i a n political liberalism, a n d R i c h a r d P o s n e r ' s d e m o cratic realism. T h e result is a n e w p r a g m a t i s t o p t i o n in d e m o c r a t i c t h e o r y .

I S B N : 978-0-415-77088-0 ( h b k )

A v a i l a b l e at all g o o d b o o k s h o p s F o r o r d e r i n g a n d f u r t h e r i n f o r m a t i o n please visit: www.routlcdge.com

V

J

Copyrighted

Material

Copyrighted

Material