METAPHOR AND KNOWLEDGEMETAPHOR AND KNOWLEDGE

KNOWLEDGE AND LANGUAGE

Volume In

Metaphor and Knowledge

Edited by

F. R. ANKERSMIT Department of History, University of Groningen,

The Netherlands

and

J.J.A. MOOIJ Department of General and Comparative Literature, University of Groningen,

The Netherlands

SPRINGER SCIENCE+BUSINESS MEDIA, B.V.

Library of Congress Catalogue card number: 92-14226

ISBN 978-94-010-4814-9 ISBN 978-94-011-1844-6 (eBook) DOI 10.1007/978-94-011-1844-6

Printed on acid-free paper

AII Rights Reserved © 1993 by Springer Science+Business Media Dordrecht

Originally published by Kluwer Academic Publishers in 1993 No part of the material protected by this copyright notice

may be reproduced or utilized in any form or by any means, electronic or mechanical, including photocopying, recording or by any information

storage and retrieval system, without written permission from the copyright owner

TABLE OF CONTENTS

ACKNOWLEDGEMENTS Vll

PREFACE IX

F. R. ANKERSMIT AND J. J. A. MOOIJ / Introduction 1

PART I: METAPHOR AND TRUTH

ARTHUR C. DANTO / Metaphor and Cognition 21DAVID E. COOPER / Truth and Metaphor 37MARY B. HESSE /Models, Metaphors and Truth 491. 1. A. MOOIJ / Metaphor and Truth: A Liberal Approach 67SAMUEL R. LEVIN / Poetry, Knowledge, and Metaphor 81KUNO LORENZ / On the Way to Conceptual and Percep-tual Knowledge 95

PART II: THE USES OF METAPHOR

RICHARD WOLLHEIM / Metaphor and Painting 113SANDRO BRIOSI / The Confused God: About a Meta-phor in Literary Semiotics 127

JAN PEN / Economics and Language 137MARIA LUISA BARBERA / Metaphor in 19th-CenturyMedicine 143

FRANK R. ANKERSMIT / Metaphor in Political Theory 155

LIST OF CONTRIBUTORS 203

INDEX OF NAMES 205

INDEX OF SUBJECTS 209

ACKNOWLEDGEMENTS

At this occasion the editors would like to thank all those involvedin the organization of the Conference on Knowledge and Language.Without the help and enthusiasm of many members, both staff andstudents, of the departments of General Linguistics, General andComparative Literature, and History, of Groningen University, thisconference, and hence this publication would not have been possible.In particular we would like to thank the other members of theorganizing committee, Jan Koster and Henny Zondervan. In allmatters of organization and planning, Liesbeth van der Veldenprovided invaluable help, and so did Marijke Wubbolts. We wouldalso like to acknowledge the cooperation of the then Dutch DefenseMinister Frits Bolkestein and his staff in the organization of thepublic debate with Noam Chomsky on The Manufacture of Consent,as well as the contributions by the chairmen and panel members.The conference was characterized by lively and fundamentaldiscussions. At this point we would like to thank those whocontributed to that atmosphere by their presentations, and who forvarious reasons could not submit their contribution for publication,notably, Manfred Bierwisch, Denis Bouchard, Melissa Bowerman,Gisbert Fanselow, Sascha Felix, Johan Galtung, Alessandra Giorgi,Giuseppe Longobardi, David Pesetsky, Dan Sperber, MichaelTanenhaus and Hayden White.We also wish to express our gratitude for the financial support

by grants received from Kluwer Academic Publishers, Wolters­Noordhoff Publishing Company, NCR Handelsblad, the Royal DutchAcademy of Sciences, the British Council in The Netherlands, theStichting H.S. Kammingafonds, the Stichting Groninger Univer­siteitsfonds, the Faculty of Arts and the Executive Board ofGroningen University.We are also very grateful to the anonymous reviewers for encour­aging comments. Finally, we thank Kluwer Academic Publishers,especially Irene van den Reydt and Martin Scrivener, for their enthu­siasm and support during the preparation stage of these volumes.

The Editorsvii

F. R. Ankersmit and J. J. A. Mooi} (eds), Knowledge and Language,Volume III, Metaphor and Knowledge, vii.© 1993 Kluwer Academic Publishers.

PREFACE

This volume is one of three which emerged from the Conference onKnowledge and Language, held from May 21-May 25 1989, at theoccasion of the 375th anniversary of the University of Groningen.The aim of this conference was to investigate the role of concep­

tual structure in cognitive processes, exploring it from the perspec­tives of philosophy of language, linguistics, political philosophy,psychology, literary theory, aesthetics, and philosophy of science.The themes of these three volumes reflect the themes of the con­

ference: Knowledge and Evidence and Knowledge and Metaphor.The volume on OrwellsProblem and Plato's Problem discusses

various issues concerning the acquisition of linguistic and non-lin­guistic knowledge.Plato's problem is how we can know so much even when the

evidence available to us is so sparse. Inborn knowledge structuremay enhance acquisition, as in the case of the baffling rate at whichthe child, on the basis of scant evidence, acquires all it needs to knowin order to speak its mother tongue. Orwell's problem is why weknow and understand so little even when the evidence available tous is so rich. Perhaps in some domains inborn structure may havethe effect of impeding the acquisition of knowledge, as in the caseof human political history, where little seems to be learned and anabundance of evidence apparently does not suffice to stop repeti­tion of identical errors and blunders.Current research on syntactic parameters and language acquisition

directly bears on solving Plato's problem in the domain of linguisticknowledge.The volume on Lexical and Conceptual Structure addresses thenature of the interface between conceptual and linguistic structure.I.e., the question of how properties of concepts are syntacticallyreflected. This issue is of the utmost importance in order to under­stand how the language faculty relates to the other cognitiveabilities.The volume on Metaphor and Knowledge addresses the second

theme. It concerns the cognitive status of figurative and metaphoric

IX

F. R. Ankersmit and J. J. A. Mooij (eds), Knowledge and Language.Volume Ill, Metaphor and Knowledge, ix-x.© 1993 Kluwer Academic Publishers.

x PREFACE

use of language. Metaphor has the air of paradox. It is generallyunderstood as not rule governed but free. But, if it is free from rules,how is it possible for people to converge in their interpretations? Itmight seem that only a common conceptual structure, limiting therange of interpretations would facilitate this, but this leaves open thequestion of how precisely interpretation proceeds. Metaphoriclanguage is often associated with improperness, or at least impreci­sion. Yet, it has an important function in scientific texts. How canfigurative language be true or false? How is it possible that metaphoris often used rather to enhance precision? Again, this seems torequire an independent conceptual structure.

F. A. ANKERSMIT AND J. 1. A. MOOIJ

INTRODUCTION

I. METAPHOR IN DISCUSSION

There has not been one and only one mainstream in the debate onmetaphor which has flourished during the last decades. Rather therehave been several important tendencies, approaches or positions. Wewant to characterize them, if only in a very incomplete manner, inorder to give a historical background to the essays collected in thisvolume.The first position is marked by the notion that metaphors work,

essentially, through interaction between two schemes, more specif­ically through a projection from one scheme of concepts, ideas orcommonplaces onto something not pertaining to that scheme. Thenotion was introduced by Max Black in the fifties, though there wereacknowledged and unacknowledged forerunners such as I. A.Richards (1936), Wilhelm SHihlin (1914) and Karl BUhler (1934).Black's version of this approach was for many years the primarylocus of reference for most studies of metaphor (whether or not oneagreed with Black). Thus, the projection approach to the study ofmetaphor may be considered, although not the mainstream, at leastthe central stream. The more so because it gradually gained in forceby assimilating some other approaches. For instance, part of thetradition that viewed metaphor as working through the deletion ofirrelevant semantic features, which was for a long time the standardlinguistic approach and also the heart of M. C. Beardsley's conno­tation view, came to be combined with the projection approach. Andthe recent book by Eva Feder Kittay, Metaphor (1987), which clearlybelongs to the projective tradition, also shows the traces ofapproaches that were its rivals in the past.!However, not all rivals have been silenced or incorporated. One

of them seems to be more vital than ever. About ten years ago DonaldDavidson gave it new impetus. We are alluding to the tradition in thetheory of metaphor which denies the usefulness of a semanticalanalysis. Through the ages philosophers have insisted on the cogni­tive nullity, or void, of metaphor. Metaphors, at the very best, were

1

F. R. Ankersmit and J. J. A. Mooij (eds), Knowledge and Language,Volume lll. Metaohorand Knowledfle. 1-17.© 1993 Kluwer Academic Publishers.

2 F. A. ANKERSMIT AND 1. 1. A. MOOIJ

thought to have an emotive or a persuasive or a decorative function,because they were believed to be incapable of embodying anythought of their own. Accordingly, a metaphor would be, strictlyspeaking, either nonsensical or nothing but a substitute for a literalexpression as the bearer of meaning in that context. And so eitherthere is no interesting meaning at all, or there is no metaphoricalmeaning. In both cases there is no place for a semantics of metaphor.Together they make up the "no-semantics" approach.Recently, in 1978, Donald Davidson revived this tradition through

his essay 'What Metaphor Means' in Critical Inquiry, later publishedin the collection On Metaphor (1979) and elsewhere. Davidsonargued that there are no metaphorical meanings, although he wasfar from believing that metaphors are superfluous or worse. For intheir pragmatic dimension metaphors can be very powerful indeed,inspiring hearers and readers to make sense of normally absurd andfalse utterances.The third approach we would like to distinguish is radically

opposed to both the first and the second approach. This is the tra­dition of viewing metaphor as the primary, original vehicle ofmeaning. According to the two first-mentioned traditions, this ishardly a significant presumption. Indeed, the first as well as thesecond approach view metaphor as a phenomenon happening withinan existing environment of literal meanings, either by creating ametaphorical meaning out of them, or by finding its path betweenthem in the mind of the language users. On that basis, the idea thatliteral meanings can develop out of metaphorical meanings is fun­damentally incomprehensible. Nonetheless, one can find that ideain Nietzsche and, more recently, in H.-G. Gadamer. Fixed, literallanguage is then seen, in one way or another, as the result (or thelimiting case) of a long process (or a rich reservoir) of free, creative,ever changeable and changing linguistic communication. It can becalled the "primacy of metaphor" approach.This third position searches for the origin of language or the fun­

damental nature of language. So does the fourth, which can be calledthe anthropological approach. It attempts to find the origins oflanguage within man's place in the world: the place of the mind aswell as that of the body, the role of man's beliefs as well as that ofman's actions. Ernst Cassirer was one of its central figures in thefirst half of this century. As far as the recent period is concerned,the inspiration may be Wittgensteinian (especially as regards the

INTRODUCTION 3

Philosophical Investigations and related works) or Heideggerian(especially as regards Unterwegs zur Sprache and related works) orFreudian or Quinean or still otherwise. Thus, the fourth approach weare considering is in itself a broad spectrum of philosophical viewsand theories. Generally, however, it will have a strong savour of phe­nomenology, and often it will be close to the third approach. Indeed,the idea of the primacy of metaphor can be the result of an anthro­pological analysis of the origin of language, although it need not beso. One special reason why this tradition deserves to be mentionedseparately in this connection is that it often treats the nature ofmetaphorical language as a key issue; witness Paul Ricoeur's Lametaphore vive (1975) as well as Lakoff and Johnson's MetaphorsWe Live By (1980).2Lastly, one should not forget the age-old comparison view, i.e.,

the idea that the mechanism of metaphor is largely that of a com­parison, albeit in a more or less implicit way. This idea came heavilyunder attack from Max Black, his followers and others, but thoseattacks were often exaggerated and unreasonable. For many studentsof metaphor it became the focal competing approach, which wasgiven the villain's part and cried down. It was also used as a meansto disqualify all those other approaches considered to be unavowedkinds of the comparison view. We think it is only right that someonehas very recently come to the defence of this much abused approach.We are not implying that the comparison view is the right view afterall; however, it should not be rejected for the wrong reasons but,rather, defended for the right reasons in so far as it can be defended.So we welcome the book by Robert J. Fogelin, Figuratively Speaking(1988), as a contemporary representative of the comparison view.

If these five positions or approaches or traditions are the back­ground against which most theories on metaphor (like the contribu­tions to this volume) can be placed, the debate on the relative meritsof these theories has not always been an easy one. The difficulty inthis debate was - and is - to find a matrix for discussion that willsatisfy the following two conditions. First, this matrix should be suchthat each of the five positions can be articulated in terms of it.Secondly, articulation in terms of the matrix should not result in aloss of content and clarity of any of the five positions involved.Surveying the long and complicated debate on metaphor, one canargue that 'cognitive truth' is the matrix most likely to fulfil thesetwo requirements. Indeed, ordinarily, though not always, it has been

4 F. A. ANKERSMIT AND 1. J. A. MOOIJ

the domain of cognitive truth where the adherents of each of the fivepositions preferred to meet and to contest each other. Implied orexplicit positions with regard to the cognitive truth ofmetaphor weregenerally felt to be the best way of confronting one view ofmetaphorwith another. As might be expected, this debate about the cognitivetruth of metaphor (when seen from the perspective of each of thefive positions) could take one of two forms. In the first place theproblem could be dealt with from the point of view of metaphor ingeneral. As opposed to such a theoretical approach, one might preferto regard metaphor as the origin or conveyor of truth in individualfields, i.e., the sciences, the arts and the humanities. The essays inthis volume have been ordered accordingly. Similarly, this intro­duction will begin with an exposition of the essays on metaphor ingeneral and next discuss the contributions about the role of metaphorin the sciences, the arts and the humanities.

II. METAPHOR AND TRUTH

In his contribution to this volume Arthur Danto comes quite closeto the second, Davidsonian position. However, Danto's resistanceto the idea that metaphor can directly convey cognitive truth is basedon arguments different from those used by Davidson. Davidsonargues that metaphors mean what the words in their common literalinterpretation mean, and nothing more. He leaves no room formetaphorical meaning and is, therefore, unwilling to consider seri­ously the idea of metaphorical truth. Put differently, Davidson deniesmetaphor a semantic dimension other than the one based on literalmeanings (although it may have important functions, cognitive aswell as non-cognitive, in the field of pragmatics) and thereby rele­gates it to the sphere of the literal statement. Danto comes to similarconclusions, albeit by choosing the other horn of the dilemma. Witha strategy that is diametrically opposed to Davidson's, he enclosesmetaphor within itself. Danto does so by emphasizing the intensionalcharacter of metaphor, and, in this way, he further develops theviews on metaphor expounded in his The Transfiguration of theCommonplace (1981). Just as Philipon represented Louis Philippeas a pear in the well-known caricature, just as his predecessor LouisXIV has been represented as the Sun, so metaphor presents us withwhat the metaphor is about as represented by the phrase that is usedmetaphorically. It follows that metaphor is indifferent to the issue

INTRODUCTION 5

of cognitive truth. Metaphor primarily has to do with how truth ispresented, with what representation is chosen for presenting it, ratherthan with truth itself.Because of his doubts about the contribution of metaphor to cog­

nitive truth, Danto is of the opinion that what seems to be a metaphorin science, is, in fact, no metaphor at all. Take for instance the state­ment 'the heart is a pump'. Obviously, this seems to be a metaphor.But in Danto's view the statement is a theory or a statement of factand therefore literally false if the heart is not a pump.'Truth and metaphor' is both the title and the topic of Cooper's

contribution to this volume. Taking up the position he had expoundedin his Metaphor (1987), Cooper argues against traditional attemptsto explain or express metaphorical meaning in terms of literalmeaning. Such attempts will always put us in the following dilemma.Either metaphor can be translated into other expressions - but thenthese expressions cannot be literal - or we take those other expres­sions in which metaphor is translated to be literally true - but thentruth cannot be transmitted from metaphor to these expressions. Inorder to avoid the dilemma Cooper proposes to look for a broaderand deeper notion of truth, of which metaphorical and literal truthcan be seen as variants. For the development of such a wider optionof truth, Cooper appeals to Kierkegaard's conception of 'being intruth'. This conception enables us to conceive of truth as not merelya correspondence between language and the world, but as deriva­tive from a relation or 'comportment' of people to the world. Truthis embedded in an existential context from which truth grows.Cooper does not attempt to give a full specification of this 'beingin truth' or of the kind of 'comportment' involved in it. Instead hegives us three examples of how metaphor can effectively exemplifythis deeper notion of truth. What these three examples have incommon is that in each case metaphor is both part of and tool for(gaining truth about) the world the metaphor is about. And this putsCooper close to the second tradition in the theory of metaphor, thetradition questioning the usefulness of a semantical analysis ofmetaphor.Mary Hesse, whose publications on metaphor have already beenwidely influential, takes up the third position mentioned above. Inher essay she wishes to defend what she calls 'thesis M': 'metaphoris a fundamental form of language, and prior (historically andlogically) to the literal'. In arguing her case she polemicizes with

6 F. A. ANKERSMIT AND 1. J. A. MOOIJ

Cooper. In his Metaphor Cooper had mentioned several objectionsagainst thesis M, all clustering around the idea that metaphor cannothistorically precede the literal statement. For if we only havemetaphor, metaphor cannot be recognized as such nor function as aconveyor of meanings. Metaphor presupposes the literal statement.In Hesse's opinion the example of science demonstrates that

Cooper's rejection of thesis M is unjustified. The main idea is thatthe concepts used by science are ordinarily used in a metaphoricalway. Since no scientific model can ever accurately capture the orderof the natural world, concepts in science cannot possess the fixitywe customarily associate with the literal use of language. This isespecially true if we think of (the evolution of) scientific debate. Themain problem now is to uphold the possibility of meaningful debatein the sciences in the absence of this fixity. Hesse's argument followstwo tracks. First, she shows that thesis M is not incompatible witha moderate realism (so there remains an external reality to serve asa touchstone for scientific theories). Secondly, Wittgenstein's notionof family-resemblances - if applied to the sciences - gives us thelatitude to conceive of a process of concept-formation that is stillrule-governed, as apparently is the case in the sciences. By followingthese two tracks, Hesse is able to show that standards for correct­ness can be built into an interpretation of concept-formation in thesciences avoiding reliance upon non-metaphorical, literal meaning.Much of Hesse's argument is an amplification of Gadamer's

defence of thesis M. This is interesting for two reasons. First,Gadamer was Cooper's main target in his attack on thesis M. Thismight justify the heuristic conclusion that Gadamer's theory ofmetaphor is an appropriate background for a discussion of theprimacy of metaphor. Second, since Gadamer was mainly concernedwith the 'Geisteswissenschaften', Gadamer's usefulness is an implicitcomment on the relation between the sciences and the humanities.Mooij approaches the problem of metaphorical truth from both a

historical and a systematic point of view. Mooij ironically commentsthat the prestige that metaphor has acquired over the last one or twodecades is no less due to a growing scepticism with regard to thepossibility of rigorous standards for truth than to an increasing aware­ness of the cognitive resources of metaphor itself.Mooij's essay is an attempt to set forth a qualified defence of

metaphor's capacity to convey literal truth. Nevertheless, he beginsby giving us three good reasons for doubting the notion of the truth

INTRODUCTION 7

of metaphorical descriptions. But precisely when within the frame­work of this argument a Davidsonian rejection of the idea ofmetaphorical truth seems to have become inevitable, Mooij makesclear that he wishes to avoid both such a categorical rejection anda wholesale acceptance of metaphorical truth. Mooij has two argu­ments for such a 'liberal' and pragmatic position in the middle. First,we must recognize that some of the arguments levelled againstmetaphor can also be directed against the literal statement. And,secondly, like Cooper, Mooij also broadens our notion of truth bypointing out that our requirements for truth are always context­dependent. Moreover, it would be socially undesirable to denymetaphor the capacity to express truth since that would condemnmuch of what we say and write to a limbo of obscurity and irrele­vance. Because, then, of his qualified assent to the notion ofmetaphorical truth - a truth that is dependent on context and on thecircumstances under which a metaphor is being used - Mooij's viewof metaphor can be seen as an elaboration of the first position withregard to metaphor that was mentioned in the preceding section ofthis introduction.Levin repeats the by now familiar strategy of softening up tradi­

tional, ascetic notions of truth. Poetry will not often convey Tarskiantruth, but it may nevertheless comprise many intimations about howwe relate to the world and to what the world contains. Quite natu­rally, Levin associates this kind of poetic truth or knowledge withthe realm of Kantian aesthetic judgment. Next, Levin turns tometaphor in order to elucidate the character of this kind of poetic'knowledge' - a strategy implying that metaphorical truth is, in fact,paradigmatic for the kind of poetical truth with which his argumentstarted.In answering the question as to the nature of metaphorical (or

poetic) truth. Levin follows a suggestion made by Nietzsche whenthe latter wrote that metaphor often requires us 'to refashion theworld'. The unexpected and original twist Levin gives to Nietzsche'shint is that when we encounter a metaphor, interpretation shouldnot take the form of asking what the metaphor might mean, but,rather, of asking ourselves 'where could that happen?' Metaphor isnot an invitation to displace meaning but to displace contexts. Take,for example, Disraeli's remark after he was appointed prime minister:'I have climbed to the top of the greasy pole'. One should not startpondering about some 'metaphorical meaning' hidden in these ten

8 F. A. ANKERSMIT AND J. J. A. MOOIJ

words uttered by Disraeli about himself, but invoke the right contextof that remark, that is, the context of many years of political strugglethat were finally crowned with supreme success. Hence, we shouldnot make language (metaphor) comply with reality (as we know it),but we are asked to imagine a reality that would fit with what themetaphor literally says.Levin insists that one of the merits of such a view of metaphor isthat we can now account for the newness of the insights we so oftengain from metaphor: reality itself and not an idle reshuffling ofalready existing linguistic meaning is the real issue in the use ofmetaphor. Hence, although we cannot derive knowledge in the strictsense of the word from metaphor, our knowledge of reality can cer­tainly be deepened by it. And this, clearly is once again reminiscentof the second, the 'no semantics' approach to metaphor.The foregoing may be sufficient evidence that discussions ofmetaphor and the cognitive status of metaphor will often take theform of (or even imply) a debate about the relation between scienceand the arts. It seems natural to analyze the cognitive status ofmetaphor against the background of claims to knowledge that wemay attribute to both the arts and the sciences.This problem is addressed in Lorenz's contribution to the volume.Lorenz draws his inspiration from Goodman's characteristically crisppronouncement that contemporary philosophy does not deal with thestructure of the world (as in pre-Kantian philosophy), nor with thestructure of the mind (Kant's project), nor with the structure ofconcepts (as became customary from C. I. Lewis onwards), but withthe structure of symbols and signs. Semiotics is the proper form ofcontemporary philosophy. This series of transitions in philosophyhas served to relativize the seemingly self-evident distinctionbetween language and the world. Philosophy in its semiotic phasecan only be interested in a purely functional account of either whatit means to be an object (in the world) or what it means to be a sign(language). Two consequences, of special interest for Lorenz, followfrom this rearrangement. First, because of the emphasis on functionand the concomitant dissolution of the traditional world vs. languagedistinction, ontology and epistemology can now be seen as two sidesof the same coin. Second, the distinction between action (world) and'sign-action' (language) must now also be relativized. The result isa 'naturalization of language' and a 'symbolization of the world' -

INTRODUCTION 9

a result interpreted by Goodman as demanding a rapprochementbetween science and art.In the main section of his paper Lorenz attempts exactly such a

rapprochement with the help of the novelist and philosopherHermann Broch (1886-1951). Broch's notion of the so-called'RealiHitsvokabel' ('single expression of reality') is, more or less,the equivalent of the sign-action: like the latter it generates a situa­tion. It denotes a 'poetic action' and is a determination of what isnot yet a sign (as in the sciences) but can still become a sign (as inart). Lorenz mentions a number of dualisms in terms of which thedual aspect of the 'Realitatsvokabel' can be clarified: perception vs.conception, knowledge by acquaintance vs. knowledge by descrip­tion, symptom vs. symbol, philosophical poetry vs. epistemology,style vs. ideal language, 'Erleben' vs. 'Erkennen' and, lastly exem­plification vs. representation (as in Goodman).This last dualism between exemplification (art) vs. representa­

tion (science) is used by Lorenz in order to flesh out the relationbetween art and science. In this context Lorenz appeals to Peirce'snotion of a sign that contains the explanation for its use in itselfand is, therefore, a sign of itself. It is in this Peircian notion of theself-explanatory sign that art and science meet. In all fairness itshould be added, however, that this meeting place belongs to thedomain of art rather than to that of science.

III. THE USES OF METAPHOR

This part of the volume opens with Wollheim's discussion ofmetaphor in painting. Can we really speak of metaphor in painting,Wollheim asks. Since metaphor is a play with rules and conventions,metaphor necessarily relies on rule-governed and conventionalmeaning. This poses a problem. For it is not easy to see what thepictorial analogue would be of that breach of linguistic rules orconventions that always announces linguistic metaphor. Wollheimargues, however, that we ought to distinguish between what merelyannounces a metaphor (linguistic deviance) and metaphor itself. Sothe problem referred to is not yet a sufficient reason for doubting thepossibility or existence of pictorial metaphor.There is a more serious problem. Starting from Davidson's theoryof metaphor, Wollheim argues that even in the case of linguistic

10 F. A. ANKERSMIT AND 1. J. A. MOOIJ

metaphor we cannot speak of 'metaphorical meaning' as opposedto literal meaning. The difference between a metaphorical and aliteral use of language does not belong to the semantics but to thepragmatics of language. Linguistic metaphor distinctively does some­thing rather than says something. Now, according to Wollheim, wecan only speak of pragmatics if we have to do with the tokens (orinstantiations) of a type that can be used differently under differentcircumstances. No doubt this is the case with linguistic signs. Butsince pictures are ordinarily thought of as individuals, not types,the distinction between semantics and pragmatics - and hence theallocation of metaphor to pragmatics - is impossible for pictures.Pictures are not instantiations of types (of signs).Wollheim now makes a bold move and urges us to think of pic­

torial metaphor as belonging to semantics rather than pragmatics.Whereas linguistic metaphor is not a form of meaning, pictorialmetaphor is. After this courageous step we may wonder whether weare still entitled to speak of pictorial metaphor at all: the 'no prag­matics, no metaphor' mechanism seems effectively to rule out thenotion of pictorial metaphor. In order to argue his case, he returnsonce again to Davidson's account of metaphor, which he usesthroughout the essay as his yardstick for measuring either linguisticor pictorial metaphor. Wollheim appeals here to Davidson's claimthat metaphor (contrary to what is often thought) does not involvea change of meaning in the metaphorizing term (e.g., the word 'sun'in 'Julia is the sun'). According to Wollheim, such a shift of meaningis effectively avoided since in pictorial metaphor the picture itself,the picture taken as a whole, is the metaphorizing term. In the picturethere is no friction, semantic or otherwise, that would involve, implyor effect such a shift of meaning. To put it differently, Wollheim'semphasis on the picture as a whole leads to a certain evanescenceof what the picture or pictorial metaphor is a metaphor of. Thesubject of the pictorial metaphor is only indirectly revealed. Yet thatdoes not mean that the subject is absent. In many, if not most cases,the metaphorical subject will be the human body that is metaphori­cally 'transfigured', to use a term of Danto, both by how the humanbody is represented and by other aspects of the painting. Twoexamples illustrate Wollheim's claim. The fact that the two examplesare chosen from two entirely different periods in the history of art ­Titian and Willem de Kooning - lends extra support to his argument.Briosi's contribution investigates a specific metaphor that is often

INTRODUCTION 11

used in critical theory: the spatial inside-outside metaphor. Hisinterest for precisely this metaphor already demonstrates Briosi'saffinity with the anthropological, or phenomenological account ofmetaphor. The metaphor in question is often relied on for definingthe 'place' of author, narrator or reader with regard to the worldthat is presented in the novel. Briosi gives an exposition of the dif­ficulties that the metaphor is apt to run into. His exposition takesthe form of an attack upon Genette's well-known view that withregard to points of view we should distinguish between 'voice' and'mode' (the person who speaks vs. the person who sees). It followsfrom his argument that a neat systematization of point of view anda clearcut formalization of what it means to be either 'inside' or'outside' the world of the novel will inevitably fail. Human beings,when telling a story, will always be both 'inside' and 'outside' theirstories.Briosi invokes phenomenology in his attempt to deconstruct the

inside-outside metaphor. Phenomenology distinguishes between theperceiving subject (corresponding to the 'inside') and the subjectusing language (corresponding to the 'outside' perspective). Sincewe cannot doubt that narration always takes place halfway betweenperception and the (referential) use of language, it will never bepossible to fit the novel within the inside-outside metaphor. Next,the phenomenon of poetic creation also strongly suggests thepresence of both a certain objectivity, or an 'outside', and subjec­tivity. On the one hand, the words seem to come from somewhereoutside, as if they were dictated. And there is a sense or experienceof 'rightness' connected with them that can only be accounted forwithin the perspective of the 'outside'. On the other hand, the novelbelongs to the sphere of the 'inside', insofar as the novel is the artist'sown creation.The world created by the novelist is never like a city enclosed

within its walls, so that we can never unambiguously state when weare 'inside' or 'outside' that 'narrative city'. To put it differently,we could only attribute to God a point of view that has completelydetached itself from the 'inside'. The novelist, however, is always'a confused God', and we might add that all the interest of the novelfinds its origin in its ambiguity with regard to the perspective of the'inside' and the 'outside'. Needless to say, this ambiguity should berespected when the novel is read and interpreted. It follows thathermeneutics with its movement of a perpetual oscillation between

12 F. A. ANKERSMIT AND J. J. A. MOOIJ

the two perspectives (a movement denoted by the notion of thehermeneutic circle) will be more successful in helping us to under­stand the novel than structuralist literary theory.Economists hope to free their language from all the impurities that

obscure natural language; it is their dream and ideal to discuss eco­nomics in the clear and timeless language of mathematics. But, writesPen, this dream is disturbed by several factors. First economistsdiscover that politicians and journalists also discuss economics andthat these amateurs are not afraid of using rhetoric, narrative andmetaphor; worse still, the economist is aware that, whether he likesit or not, he will have to reckon with the pollution of economiclanguage resulting from that. Second, the economist cannot deny thathis arguments have political implications and are, perhaps, eventhemselves infected by political assumptions. It is true that for atime economists tried to ignore the political dimension to their work.But the problem refused to go away and became all the moremanifest with the rise of different schools within economic theory.Attempts to reconcile these schools within one synthesis provedfutile and economists had to accept that schools were here tostay.This has opened the eyes of (some) economists to the rhetorical

and metaphorical dimension of their professional language. It caneven be argued that there is a metaphor at the heart of each of themajor economic theories. Neo-classical economics likes to compareeconomic reality to clocks, steam engines, mechanical regulators oreven to nature itself; Keynesians rely upon the metaphor of economicreality as a hydraulic system, whereas Marxists see society as abattlefield. Pen emphasizes the unavoidability of metaphor and isconvinced that the growing awareness of 'the rhetoric of economics'(to quote the title of McCloskey's influential book on the issue) isa valuable gain for economic theory. When arguing this way Penproves himself to be no less a defender of the 'primacy of metaphor'view than is Mary Hesse. Nevertheless, Pen remains sensitive to thedangers of the use of metaphor: it is only too easy to get carriedaway by strong and suggestive metaphors. In the end scientificdebate will have to decide on the merits of the metaphors that havebeen proposed, and that means that scientific debate could be seenas the objective 'metaphor-free' forum that decides on the fate ofspecific economic metaphors.

INTRODUCTION 13

The two last essays in this volume are applications of certain(medical and political) metaphors to the human sciences rather thanattempts to put forward novel theories on metaphor and on metaphor­ical meaning. Barbera discusses in her contribution what could bestbe described as a 'double metaphor': the (human body as a politicalmetaphor and the body politic as a medical metaphor). Here medicineand politics metaphorize each other. With a wealth of historical detailBarbera shows that in the beginning of the nineteenth century, whenorganicist metaphors were widely used in political theory, the linkbetween politics and medicine was closer than ever before or after.Barbera emphasizes that the political use of the metaphor was notwithout its complications. For if in political thought the term'organic' is opposed to the terms 'atomistic', 'individual', 'revolu­tionary' or 'critical', as the followers of Saint Simon and AugusteComte were apt to do, then organicism unavoidably took on the char­acter of a political program and in doing so became tainted by thesame artificiality that these political theorism always sought tocontest.But the interaction between medicine and politics was not

restricted to the organicist metaphor. Barbera discusses the metaphorof tissue (introduced by Bichat), the metaphor of illness as disorder,which clearly is suggestive of political chaos and, lastly the metaphorof exchange. With the last metaphor we leave politics for medicine.For the notion of (economic) exchange was used in medicine as ametaphor of nutrition and of metabolism. It is the main purpose ofBarbera's essay to demonstrate that this (political) metaphor ofexchange stimulated the nineteenth century physiologist R. Virchowto defend new and original insights in his Die Kritiker der Cellular­Pathologie. Thanks to a political metaphor Virchow developed a newconception of the cell of how the cell determines metabolism andbodily growth. And the amazing conclusion follows that, at least inthis case medicine owes more to politics than political thought tomedicine.Another political metaphor, Plato's metaphor of the ship of State,

is the point of departure of Ankersmit's essay on metaphor in polit­ical theory. He argues that this metaphor has inspired a large partof Western political thought. And the nature of this inspiration canbe defined as a transposition of epistemological patterns of thoughtto the domain of politics. The metaphor of the ship of State, with

14 F. A. ANKERSMIT AND 1. J. A. MOOIJ

its suggestion of a well-defined center from which society is ruledand organized, is the political analogue of the Cartesian and Kantiantranscendental self.H can be shown, next, that both the realities of contemporary

politics and a certain dynamic in political thought itself underminethe kind of transcendentalist political thought that is suggested bythe metaphor of the ship of State. Benjamin Constant's novelAdolphe is used to explain and clarify an increasing entanglementof State and civil society and it is precisely this increasing en­tanglement that robs the metaphor of its significance and utility. Onemight ask what kind of political reality awaits us if the mutualindependence of State and civil society is gradually lost. The sug­gestion is that the Japanese political and social system, character­ized by the notion of the 'empty center', provides at least part ofthe answer.

IV. CONCLUSION

What conclusions can we draw after having read the eleven essaysin this volume? Since metaphor is studied from so many differentperspectives in these essays, it would be idle and naive to expect thatsome bold inferences can be derived from them. Yet one cannot failto be struck by the tone of caution and reticence that pervades mostof these contributions.

If we recall the five positions on metaphor that were outlined atthe beginning of this introduction, we must recognize that Davidson'saccount of metaphor is best suited for discouraging any high hopesone might wish to cherish with regard to metaphor. Indeed, of thesefive it undoubtedly is the 'coolest' view of metaphor. The evidentpopularity of Davidson's views is paradigmatic of the cautionrecommended in most of the essays. How much Danto mightdisagree in the end with Davidson, it is the latter's view of metaphorthat forms his point of departure. This is even more true ofWollheim's analysis of pictorial metaphor. And we cannot fail tonotice the similarities between Davidson's and Levin's positions: inopposition to most theories on metaphor both insist that we oughtto retain the literal meaning of the metaphorizing term. Cooper isprepared to consider metaphorical truth only on condition thattraditional conceptions of cognitive truth are substantially weakened.He would therefore agree with Davidson's rejection of metaphor's

INTRODUCTION 15

claim to offer cognitive truth. Lastly, although Mooij in his conclu­sions is more optimistic about the cognitive merits of metaphor thanDavidson, the scepticism that dominates a substantial part of hisargument runs parallel to that of Davidson.The same picture emerges if we look at the essays about the usesof metaphor. Wollheim appeals to Davidson in concluding that wecan only speak of pictorial metaphor if we are prepared to acceptan evanescence of the metaphorical subject. Briosi's contributionis, in fact, a sustained attack on one of the most natural and con­ventional metaphors in critical theory. Ankersmit argues thatmetaphor in political theory will obscure rather than help us under­stand contemporary political reality. Barbera is divided: in the caseinvestigated by her, metaphor proved fruitful for medicine but notfor politics. The only unqualified approval of metaphor is found inHesse's and Lorenz's contributions. Hesse's defense of thesis Mplaces metaphor right in the center of scientific practice. AndLorenz's rejection of any clearcut distinctions between science andart (metaphor) has similar implications. At first sight it may seemsurprising that science apparently stimulates a so much less scep­tical attitude towards metaphor than the arts and the humanities. Theexplanation for this counter-intuitive state of affairs might be - fol­lowing both Hesse's and Pen's suggestions - that science, unlikethe arts and the humanities - has some built-in guarantees againstthe misuse of metaphor.Perhaps this attitude of caution and reticence with regard to

metaphor is indicative of the (many) ironies of the relativism andscepticism so characteristic of the contemporary intellectual climate.In an initial and previous phase relativism taught us to take seriouslythose aspects of knowledge and of science (like metaphor) whichresist a reduction to the fixed algorithms of an older and more con­fident philosophy of science. As a result, metaphor came to be seenas a valuable asset in our attempt to understand the world. But nowanother step has been taken: like those algorithms of a logical-pos­itivist past, metaphor has now fallen victim to scepticism as well.

NOTES

1 For further details on this approach, its history and its relation to otherapproaches, see also Mooij (1976), Chs. 3,6, 7 and 12. See especially pp. 73-80for comments on the work of Stlihlin and Buhler.

16 F. A. ANKERSMIT AND 1. J. A. MOOD

2 The views on metaphor of Ricoeur and of Lakoff and Johnson are close to theprojection approach and could have been mentioned in connection with that tradi­tion. This again testifies to the central position of that approach. What makes ussee them as representatives of the anthropological approach, however, is their beingpart of an encompassing view of man's relation to the world we live in. With them,the mechanism of metaphor seems to be subordinate to a larger theory about man,culture and the world.

BIBLIOGRAPHY

Beardsley, Monroe C.: 1981, Aesthetics. Problems in the Philosophy of Criticism,2nd edition, Kackett Publishers Company, Indianapolis and Cambridge; seeespecially Ch. III, section 10 and Postscript 1980, section 4.

Black, Max: 1955, "Metaphor', Proceedings of the Aristotelian Society 55,273-294; widely reprinted.

Buhler, Karl: 1934, Sprachtheorie: Die DarsteUungsfunktion der Sprache, GustavFisher, Jena.

Cassirer, Ernst: 1925, Sprache und Mythos, Leipzig - Berlin, Susanne K. Langer,(tr.), Language and Myth, Dover Publications, New York, 1946; see especiallyCh.6.

Cooper, David E.: 1987, Metaphor, Blackwell, Oxford.Danto, Arthur c.: 1983, The Transfiguration of the Commonplace, HarvardUniversity Press, Cambridge, Massachusetts.

Davidson, Donald: 1987, 'What Metaphors Mean', Critical Inquiry 5,31-47, alsoin Sacks (1979), 29-45, and in Johnson (1981), 200-220.

Fogelin, Robert J.: 1988, Figuratively Speaking, Yale University Press, New Havenand London.

Gadamer, Hans-Georg: 1960, Wahrheit und Methode. Grundzuge einer philosophis­chen Hermeneutik, 1. C. B. Mohr (Paul Siebeck), Tubingen; see especially PartIII, section 2c; (tr.), Truth and Method, Sheed and Ward, London; 1975.

Goodman, Nelson: 1968, Languages of Art. Approach to a Theory of Symbols,The Bobbs-Merrill Company, Indianapolis; see especially Ch. II.

Hesse, Mary B.: 1966, 'The Explanatory Function of Metaphors', in M. B. Hesse,Models and Analogies in Science, Notre Dame University Press, Notre Dame,Quebec; pp. 157-177.

Hesse, Mary B.: 1980, Revolution and Reconstruction in the Philosophy ofScience,Indiana University Press, Bloomington, Indiana.

Johnson, Mark (ed.): 1981, Philosophical Perspectives on Metaphor, Universityof Minnesota Press, Minneapolis.

Kittay, Eva Feder: 1987, Metaphor. Its Cognitive Force and Linguistic Structure,Clarendon Press, Oxford.

Lakoff, John and Johnson, Mark 1980, Metaphors We Live By, The University ofChicago Press, Chicago and London.

Levin, Samuel R.: 1977, The Semantics of Metaphor, Johns Hopkins UniversityPress, Baltimore and London.

INTRODUCTION 17

Mooij, J. J. A.: 1976, A Study of Metaphor. On the Nature of MetaphoricalExpressions, with Special Reference to Their Reference. North-Holland,Amsterdam.

Nietzsche, Friedrich: 187311903, 'Uber Wahrheit und Luge im aussermora1ischenSinn', in Friedrich Nietzsche, Werke in drei Biinden; herausgegeben von KarlSchlechta, Wissenschaftliche Buchgesellschaft, Darmstadt; 1966, Dritter Band,pp.309-322.

Ortony, Andrew (ed.): 1979, Metaphor and Thought, Cambridge University Press,Cambridge, Massachusetts.

Richards, Ivor A.: 1936, The Philosophy of Rhetoric, Oxford University Press,Oxford; see especially Chs. 5 and 6.

Ricoeur, Paul: 1975, La metaphore vive, Editions du Seuil, Paris; Robert Czemyet al. (tr.), The Rule of Metaphor. Multi-disciplinary Studies of the Creation ofMeaning in Language, University of Toronto Press, Toronto and Buffalo.

Sacks, Sheldon (ed.): 1979, On Metaphor, The University of Chicago Press,Chicago and London.

Stlihlin, Wilhelm: 1914, 'Zur Psychologie und Statistik der Metaphem. Einemethodologische Untersuchung', Archiv fur die gesammte Psychologie 31,297-425.

PART I

METAPHOR AND TRUTH

ARTHUR C. DANTO

METAPHOR AND COGNITION

Discussing diction in Poetics, Aristotle writes: "the greatest thing byfar is to be master of metaphor", which is "the one thing that cannotbe learned from others; and it is also a sign of genius.'" Even ifgenius means something considerably less than it came to mean inromantic times, if Aristotle is at all right here then making metaphorscannot form part of ordinary linguistic competence, if only becausewe do acquire from others our knowledge of language - it is aparadigm of something taught and learnt. Moreover, it is widelyaccepted that linguistic competence entails a symmetrical capacityto form and to understand sentences, where no such symmetry isimplied in metaphoring activities, in which you presumably do nothave to be marked by genius to grasp metaphors once made. Isuppose it must be roughly parallel to humor, in that making jokesis a gift of a certain order whereas getting jokes is ordinarily not.Still, there is evidently room for education, as Aristotle's discus­sion in the Rhetoric, Book III, suggests, where he lays down anumber of principles that might be thought of as refining metaphor­ical talent, or as the principles of criticism, as if making metaphorswere parallel in certain ways to the exercise of taste. We learn thedifferences between good metaphors and bad ones, or between inap­propriate and appropriate ones: it is almost like a discourse amongmakers of haiku poetry. But this again has little to do with linguisticcompetence: it is, rather, more a matter of verbal etiquette, orlearning how to modulate metaphors in order to achieve desiredeffects, something the rhetorician is anxious to learn. Aristotle saysthat metaphors, like epithets, must be "fitting" - he explicitly usesthe analogy of the kinds of clothing it is fit for one to wear; and itis instructive to think out why, to use his example, a young man'scrimson cloak would not suit an old man, as though costume itselfwere metaphorical, or nearly so.I am anxious to enlist Aristotle's help in driving a wedge between

metaphor and linguistic competence, partly because any good theoryof metaphor would want to be able to explain why an article ofclothing could in certain circumstances have the power of a

21

F. R. Ankersmit and J. J. A. Mooij (eds), Knowledge and Language,Volume Ill, Metaphor and Knowledge, 21-35.© 1993 Kluwer Academic Publishers.

22 ARTHUR C. DANTO

metaphor, or must at the very least take into account the existenceof pictorial metaphors as well as verbal ones. Metaphor must operateat a level common to the two chief modes of representation avail­able to us (pictures and words), as well as to many minor ones, likeclothing and architecture; and cannot be narrowly construed as lin­guistic at all. And even when verbal metaphors are discussed, asseems almost always to be the case with the ancients (perhaps artcriticism in Greece, concerned with mimesis, had no room for aconcept of pictorial metaphor), it may be with analogies to whatmakes a pictorial metaphor that we will want to explain the differ­ence between "Men are animals" as metaphorical and literal descrip­tion of men, even if surface grammar will not discriminate. (One badtheory is that the literal is the metaphorical use gone stale, for while"Men are animals" is in fact a fairly stale metaphor, a cliche in effect,it remains a metaphor by contrast with the descriptive "Men areanimals" taken as the assertion of biological fact.) And this will putus in a position to approach the question that brings us here, of thecognitive dimension of metaphor, pictorial or verbal, if there is acognitive dimension. For there may be a deep reason why Aristotle'sdiscussions of metaphors are in his Rhetoric or the Poetics, ratherthan somewhere in the Organon, or why he thinks there is a con­nection between riddles and metaphors - "Metaphors imply riddles,and therefore a good riddle can furnish a good metaphor"2 - whereyou have to guess the right answer already known by the riddler.The rhetorician uses the metaphor to drive the hearer's mind wherehe wants it to go (which is why "metaphor" is not, Derrida notwith­standing, itself a metaphor: it is a matter of moving the mind whereone wants it to go, which is quite literally what takes place).Metaphor belongs to the theory of manipulation, in effect to thepolitics of the mind. So let's begin with political cartoons.Consider a famous caricature by Charles Philipon of Louis

Philippe as a pear, printed in his magazine, La caricature in 1831."Louis Philippe is a pear" is not, I think, an obvious metaphor, noteven if "une poire" had the slang meaning at that time of a "sim­pleton", for there is no easy way in which we can understand whatespecially inspired the slang expression in the first place, what storycan be told that explains, like a myth, the transit from simpleton topear. (Think of a riddle "Why is Louis Philippe like a pearT' whereour incapacity to guess marks a rhetorical failure on the part ofsomeone who wants us to see Louis Philippe a certain way by calling

METAPHOR AND COGNITION 23

him a pear, or showing him as one.) Philipon's drawing takes usthrough four stages, the first in which the King is drawn in a rec­ognizable way, with perhaps some slight exaggerations, as befits acaricature; then redrawn with those exaggerations further reenforcedwhile others are weakened, until finally only those lines remain thatthe original head had in common with a pear, with a few vestigialfacial features we might not even see as such outside the sequencein which the drawing occurs. It is as though Philipon had seen a pearinscribed in Louis's head, and taught us to see it through steps ­taught us to see what had always been visible to the genius of thecaricaturist who has, in Aristotle's word "an intuitive perception ofthe similarity in dissimilars." Moralists sometimes say we finallyearn our faces, that our faces finally have our moral historiesinscribed in their shapes, as if each of us displayed a metaphor ofour moral reality as a kind of signature. And when we see the pearin Louis's face, we instantly get the point Philipon slyly makes. Itis important, first, that all the pear contributes to this reduction ofLouis itself is its shape - not its taste or color - but that shape,large at the bottom and narrow at the top, implies fat jowls and atiny brain, hence someone stupid and gluttonous at once. (If mon­goloids had pear-shaped heads, that would explain the slang and thefact that no one not familiar with mongoloids would understand it,even if it got to be smart to call people pears when one wanted toinsult them.) A man with a pear-shaped body would have his bellydisproportionately large and his head by that measure small, con­noting through the identical bodily build the unfortunate conjunc­tion of negative attributes Philipon deftly draws out of Louis's mug.Philipon's visual metaphor was at odds with the one the king

himself undertook to project through his costume which, in his case,really was metaphorical even if, or just because, he dressed himselflike any prosperous bourgeois, not as an aristocrat but as MonsieurEveryone, domestic, paterfamilial, comfortable. Instead, Philiponmade him reprobate and ludicrous at once, like a vicious clown, andtook an artistic revenge for Daumier's imprisonment for havingdepicted the king as Gargantua, swallowing material goods (and ret­rospectively periform). The metaphor was immediately understoodand widely employed: Daumier used it over and over to the pointwhere he could simply have drawn a pear and everyone would haveknown who was meant, even though, to the uninformed, it might bepuzzling as to why there was a still-life in the middle of a magazine

24 ARTHUR C. DANTO

of pictorial satire (just as it is puzzling to see the peacock in a famousmontage in Eisenstein's film October, where we wonder what a birdis doing in the Kremlin when Eisenstein meant us to see thatKerensky was vain as the peacock). To those who understood,however, it would have a meaning that the drawing of a mere pearwould not have, and that, when worked out, would give us the insightrequired into why "Men are animals" is a metaphor when its indis­cernable counterpart is not. It must, on the other hand, be said thatthe metaphor was so successful that it degenerated into a meresymbol or emblem of the July Monarchy, acknowledged as such bythose ignorant of its history as we are of why New York is emblem­ized by an apple, or whether "New York is an apple" ever was ametaphor and why.I want two further examples before proceeding to matters of struc­

ture. There is an amusing photograph by the American photographer,Alice Boughton, an evidently rather forbidding looking woman, whowas famous in her own time for her portraits of the famous. Shedepicts herself as Queen Victoria, to whom she bore a certainlikeness. It is important that we understand this to be a picture ofAlice Boughton as Queen Victoria, rather than a picture of QueenVictoria as represented by Alice Boughton. In the latter case, shewould be merely the model for the old queen, dressed appropriatelyin the sorts of laces and ribbons we see in official late portraits ofQueen Victoria as an elderly woman. Pictures which use models arerarely pictures of models. In the famous self-portrait of Vermeerpainting the muse of history, we see in the painting within thatpainting the muse of history herself, while we also see, standing infront of Vermeer, whose back is turned to us, the model he happensto use, dressed in laurel leaves and carrying an elaborate horn ­studio props, as it were. Vermeer is not painting her - he is paintingthe muse of history, for which she stands, even if a painting of themuse of history and a painting of that model, whoever she was, mightlook exactly alike. And both those paintings might look exactly likea third, this one structurally equivalent to Alice Boughton as Victoria,of that particular woman as the muse of history. We might, in brief,imagine three distinct paintings, which we might as well imaginelooking exactly alike, though they have quite different representa­tional structures. These differences arise of course because modelsstand for what they are models of, and so are vehicles of represen­tation in their own right. Imagine Alice Boughton, who shot so many

METAPHOR AND COGNITION 25

of the world's famous, regretting that she never photographed QueenVictoria, and someone said why not dress up like the queen yourself,why not pretend to be the queen. Had that been the story, her's wouldhave been a picture of Queen Victoria, in the sense that the queenwould be the subject of the work, with the artist as the model. Butthe queen is not the subject of this work - Alice Boughton is ­dressed up as Queen Victoria. It is a self-portrait. And it is, moreover,a pictorial metaphor. It is Alice as Queen, almost as in LewisCarroll's story. It is Alice seen under the representation that weidentify as a Queen-Victoria-representation. It tells us volumes, Ithink, about how Alice Boughton saw herself that she should havechosen to depict herself this way, things we could not know froman ordinary self-portrait, and things we certainly could not knowfrom a picture in which Alice Boughton used herself as a model fora picture of the queen. Think of the profound difference betweenwhat we learn about Rembrandt from the fact that he portrayedhimself as the aging King Saul, a metaphor for his life which risesto the level of a tragic statement that even kings, in their golden vest­ments on their precious thrones, age and die, by contrast with whatwe would learn from a painting Rembrandt did of King Saul forwhich he happened to use himself as the model. We learn of coursethat he chose to paint a picture of a king, but this is pretty shallowin comparison with the deep, one might almost say philosophicallesson of the Self Portrait as Saul.One last example, this time closer to Louis Philippe as pear than

to Alice Boughton as Queen Victoria. Let us consider a beautifulsmall work of Zurbaran of a dead lamb which we know to be apainting of Christ, even if there were no visual evidence that it isChrist that is being metaphorically depicted and that it is not merelya painting of a dead lamb, as a painting by Albert Pinkham Riderof a dead bird is perhaps just a painting of a dead bird and not ofsomething else for which the dead bird is a metaphor. In Zurbaran,of course, there is the metaphysical blackness against which the lambis depicted, a touch of Caravaggism, which may carry the metaphor­ical meaning of "the darkness of the land" at Matthew 27, or maynot. Still, it is a most moving picture, and moving in a way thatcannot be the same as that in which one might be moved from thethought of a slaughtered lamb. One feels, in Spanish still-lifes of thatera, a kind of mysticism of commonplace things, almost as if every­thing were metaphorical if we but knew how to read it, and that in

26 ARTHUR C. DANTO

any case the artist is telling us something other than the visual truthof bottles and gourds if we but knew how to read it. In many casesthe meanings are lost, or can at best be recalled through scholarlyinvestigation, when the assumption would be that there was anaudience for whom Zurbaran painted who saw Christ in the lamband saw, for all I know, the holy spirit in a waterjug, just as therewas a community for whom it was plain that Louis Philippe was apear, and a much smaller community for whom Alice Boughton wasQueen and Empress: that system of meanings defines a communitywe, here, in every case are outside of, I think. The metaphors crys­talized into vivid images a body of beliefs and feelings that can nolonger be ours, as the figures themselves, in the case of LouisPhilippe and Alice Boughton, no longer engage anyone at all. It sayssomething about us, on the other hand, that Rembrandt and Jesusdo move us, and that the metaphors in the works I mention are alive.Dead and living metaphors are a function of subjects that themselvesare living or dead for the persons who perceive the metaphors. I usethe word "moving" in connection with Zurbaran's and Rembrandt'swork in part because the image really does move the mind of theviewer, as metaphors are literally supposed to do when they are suc­cessful. That is why they are called tropes. Metaphors got studiedas rhetorical cases, and rhetoric exists because of the tropismaticcharacter of the mind. In a small way metaphors achieve whatcatharsis is supposed to achieve in the experience of tragic art.Metaphors exist because of the truth that the mind is moved by rep­resentations, which explains why they were regarded dangerous inthe hands of rhetoricians by Plato, or why art is regarded by him asdangerous as a kind of entrapment of the mind. If we lose sight ofthe psychological power of the metaphor, and think of it merely asa figure of speech - or, in the case of pictures as a kind of manner­istic conceit - we have lost sight of something central in metaphorsand in ourselves.Obviously, the formula "moved by representations" must be care­

fully qualified. There was a time when Englishmen were moved bypictures of Queen Victoria, but only because they were moved byVictoria herself, perhaps a bad example for my point since she herselfwas a metaphor for the British Empire and all the values condensedin her name: stability, the sacredness of the home, morality itself.Still, the representations - photographs, say, or official portraits ­were transparent in that they sent the viewer directly to the subject.

METAPHOR AND COGNITION 27

Those who were moved by Alice Boughton as Queen Victoria weredoubtless moved by Alice, but then they would have been moved inthat transparent way by a straightforward picture of her because ofwho she was. Once more, the representation moves because of itscontent, or subject, and not because of any properties it has in itsown right as a representation. But someone might be moved by therepresentation as Queen Victoria - not necessarily by the fact thatshe portrayed herself so, and one were moved by the fact that shehad a sense of humor after all - but by the way the representationcaused one to think of her. In effect, the photograph is a represen­tation of a representation in this case, and it is this fact that beginsto account for the most interesting logical feature of metaphors, tomy way of thinking, namely their non-extensionality. And this I shallendeavor to account for by saying that part of what a metaphoricalrepresentation is about - one of its truth conditions, so to speak ­is a representation.Consider the difference between a portrait of Le Roi-Soleil,

namely Louis XIV, and a portrait of Louis XIV as Le Roi Soleil.Since Louis and Le Roi Soleil were identical, any royal portrait ofLouis would perforce be one of Le Roi Soleil. Still, you could notany more infer from even an awesome depiction of Louis byHyacinthe Rigaud that he was The Sun King than you could that hewas the third Bourbon monarch. Presumably a portrait of him asthe Sun would enable a visual inference if the portrait were a visualmetaphor. In fact I know of none, but I can easily imagine a portraitof Louis as The Sun: it would show him as radiant and luminous,and perhaps, if an allegory, in some dominant locus in the heavens,among the planets and stars. When the Virgin is represented as theQueen of Heaven, she is shown treading the Moon. What I am sayingis that such metaphorical portraits contain representations: a portraitof Louis XIV, which of course is equally a portrait of Le Roi Soleil,is a representation of the king, but it does not contain a representa­tion. A portrait of the King as Le Roi Soleil is a portrait of him,which contains a representation, the sun itself, perhaps. A metaphor­ical portrait of Louis XIV which reduces him to his metaphoricalattribute - as a caricature by Daumier of Louis Philippe as Pear mightshow simply a pear, leaving it to the informed audience to recog­nize it as of Louis Philippe - might be indistinguishable from apicture of the Sun. But it would be a picture of Louis as tacit subject,in which the sun appears as his representation. If it were a picture

28 ARTHUR C. DANTO

of the Sun, it would be a picture of the intensely hot, self-luminousbody of gases the density of whose center is 100 times that of waterand whose temperature is about 15,000,000 oK, just as a picture ofLouis XIV would be a picture of the husband of the Spanish PrincessMarie-Therese. But in fact though it looks like a picture of the Sun,and hence a picture of a body of gases, it is a picture of a king whohappens to be married to a Spanish princess. It is not a picture ofthe Sun - it is, if you wish, a Sun-picture which denotes, or standsfor a king. And this is one mark of metaphors, pictorial or otherwise:Juliet is metaphorically the sun, but not metaphorically an intenselyhot self-luminous body of gases. In describing her metaphorically asthe Sun, Romeo - a genius of love - is in effect saying: See her underthis representation. The metaphor is a relationship between anindividual and a representation, and the non-extensionality of themetaphor is due to the fact that a sun-representation is not a hot­body-of-gases-representation. It is not for the same reason that theexpression "The sun" and the expression "A hot body of gases" arepatently different expressions. Or for the same reason that "The Sun"is what Romeo said and he did not say "A hot body of gases". Apicture of Alice Boughton as Queen Victoria, since not a picture ofQueen Victoria, is not a picture of the wife of the Prince Consort.Nor does Alice show herself as the wife of the Prince Consort sincethat would be a representation of a different sort than the QueenVictoria representation she in fact uses.In my discussion of metaphors in The Transfiguration of the

Commonplace, I offered the thesis that metaphors have the logicalform of intensional sentences, and that intensional sentences as aclass are marked by the fact that they involve reference to - andusually a characterization of - a representation amongst their truthconditions. 3 Thus "s believes that P" has the usual marks of inten­sionality because it is about P - and not about what P itself is about- and as a (sentential) representation, P has what properties it hasas a specific representation. The paradigm for this analysis is thequotation, or direct-discourse ascription, S says "P", where weactually display a token of the same sentence that issued forth fromS's mouth when he said P. Here S designates the individual who said"P" - but "P" abbreviates, as Carnap was careful to distinguish, thesentence uttered. A name does not abbreviate the individual named,so "P" here is not the name of the sentence uttered but that verysentence, abbreviated.

METAPHOR AND COGNITION

Consider Oedipus:

Upon the murderer I invoke this curse ...may he wear out his life in misery.

29

Now, since Oedipus is the murderer, we can replace his name withthe co-referential "Murderer of Laius". But since we are displayingthe sentence Oedipus said, we cannot replace "the murderer" with"Oedipus" - for that is not what Oedipus said. Indeed, there is somereason to suppose that if he knew he were the murderer, he wouldnot have cursed himself. In any case, in quoting Oedipus we canliterally point to the sentence Oedipus said, which is what gives adegree of plausibility to Donald Davidson's thesis that "that" is ademonstrative in "On Saying That".4 Replacing "the murderer" with"Oedipus" would make it a different sentence, generating a differentset of tokens. So a quotation is about a speaker and a sentence,respectively designated and displayed.It is not implausible to extend this account to belief, just as Carnap

himself endeavors to do: the intensional "Charles thinks that A"yields to the extensional "Charles thinks 'A'" (where, I suppose,the quotation marks serve as demonstratives, pointing to the sentencealso displayed in abbreviated form). So belief, too, can be regardedas a relationship between an individual and a sentence; and itsdescription has again the logical form of a two-place predicate"Believes (A, S)" where again A names the believer and S displaysthe sentence believed. Just as there is a convention that allows us todisplay a spoken sentence with a written one, so we may display athought sentence with a spoken or a written one. Thus the murdererof Laius believes he has cursed the murderer of Laius, but does notbelieve either that he has cursed Oedipus or cursed himself. Andagain, the reason is that these would be different sentences than theones we correctly display. The non-extensionality of belief-sentences,or propositional attitudes in general, derives from this fact: they referus to the specific sentences we display in ascribing beliefs (or pro­positional attitudes) and not to some other sentence which mightbe gotten by replacing terms with co-referential terms or evensynonyms. They have the specific features of the displayed sentenceas among their truth-conditions, quite as much as does a direct quo­tation. It is, indeed, from this semantical feature of such sentencesthat I have sought to deduce the thesis that in ascribing a belief toa person, I am asserting of that person that he is in a sentential state

30 ARTHUR C. DANTO

which I can display a token of in saying what he believes. I canactually show you the content of the believer's mind.So let us tentatively propose that a metaphor involves an implicit

demonstrative - a 'that'- and a representation, so that Philipon ineffect is displaying a pear and saying: Louis Philippe is that; andthat Romeo, though he uses words, does so in the sense in whichthe word displays what it ordinarily designates, so that it is as if hewere to display the sun and say: Juliet is that. Metaphors are in thissense reductive of the individuals they designate: they reduce themto those features of themselves which the metaphors displayed makesalient, Le., Louis Philippe as fathead. In effect, the metaphor is aninjunction to see the individual as consisting merely of the attrib­utes made salient by the image, or as consisting of them essentiallyor fundamentally, or as being nothing but them. It confers upon theindividual a limited identity, and metaphors are inevitably of use tothe rhetorician whose enterprise it is to get us to see what he istalking about in the way carried by the representation, by that rep­resentation and not some other that might be otherwise coextensiveor even synonymous with it. Suppose the rhetorician says of a risingyoung politician that he is the morning star. In his poem "On MayMorning", Milton writes "The bright morning star, day's harbinger."In the Apocalypse, it is written "I am the root and offspring of David,and the bright and morning star." The evening star may be just asbright (it would have to be if the morning star is the evening star),but it would not replace morning star, not merely because that isnot what these texts write but because that is not what the writerswould have written: Tennyson writes, in "Crossing the Bar", "Sunsetand the evening star, and one clear call for me", where "morningstar" would not do at all. Morning star means celestially promis­sory, a new dawn, so "George Bush is the morning star" (forgive me)would hardly yield to "George Bush is the evening star." The "brightparticular star" of Shakespeare is almost certainly the morning, notthe evening star. So the metaphor fixes the image of the individualin the minds of the audience as the poet - or the rhetorician - wouldhave intended.The English "as" has the force of "like" and hence of similarity

or resemblance, which may then mark the difference in force at leastbetween similes and metaphors in that a simile merely remarks uponsimilarities whereas the metaphor, with its tacit demonstrative, iden­tifies the essence of the thing. Louis Philippe may be pear-like in

METAPHOR AND COGNITION 31

body-build, but that leaves it possible that the similarity does notpenetrate his essence, and that he would be the same individual ifhe were, by diet and strenuous exercise, to acquire the lean bodyline. But if Louis is a pear, the diet leaves him essentially unchanged.And metaphor accordingly involves a distinction between essentialand not essential (or accidental) attributes. Metaphors get to the heartof things, accordingly, and it is this as much as anything that musthave tormented Louis Philippe, who felt, rightly, that there was moreto him than that. When Alice Boughton depicts herself as QueenVictoria, she is telling us what she essentially is. Christ as lamb isnot simply innocent of taint, he is innocence itself, the very embod­iment of that. The Sun King incarnates luminousness and nobility.This, I think, may be a cognitive dimension for metaphorical repre­sentation - for representation as - on the view that cognition trafficsin essences and reductions. And I suppose, to the degree thatmetaphoricity characterizes art, it accounts for the cognitive dimen­sion of art as well. But this goes well beyond anything I can hopeto demonstrate here. Let me instead settle for some reflections onhow such essentialistic propositions are communicated, rememberingalways that the metaphor is a rhetorician's tool, and hence a devicefor moving minds. And metaphors can be cognitive only if the personor thing designated really is that - only if its essence is as shownor displayed. And anciently the rhetoricians were famous for makingthe worse appear better and better worse.The paradigm of the rhetorician's instrumentarium is the rhetor­

ical question. It is a question the answer to which is supposed to beso obvious that the answer is given by the audience almost mechan­ically. The rhetorician puts the question and the audience may indeedalmost shout the answer out as one voice. And the psychology is thatin answering the question themselves, participating, as it were, inthe process, the audience is convinced in a way it would not havebeen had the rhetorician instead given them an answer. It is like anenthymeme, in that the missing premiss must be a platitude or atruism, an unusual requirement for something in a supposedly logicalstructure, leaving it clear, I think, that the enthymeme itself is arhetorician's device for getting an audience to furnish the missingpremiss and connect the conclusion with the offered premiss ­moving the mind, as it were, along certain tracks, and convincingin consequence of this movement. My sense is that the metaphormust work in this way: the rhetorician demonstrates whatever it is

32 ARTHUR C. DANTO

that he wants the audience to believe whatever he is talking aboutis - and for this Peason the connection must be as obvious as thesuppressed premiss of the enthymeme or the stifled answer to therhetorical question. But this puts certain constraints on what therhetorician can use as a representation. It has to be familiar, for onething, which means that the metaphor implies a community ofreference, a group of individuals who may be expected to know whatthe reference is. It would be a very refined community indeed withinwhich I could use "is an enthymeme" as a metaphor. Even "is anangel" fails as a metaphor in a Chinese society, in which angels arejust unknown. Pears, the sun, define a very wide community, as dolambs: Queen Victoria one much less wide, but one in which picturesof the Queen would be familiar from perhaps postage stamps and thelike.But there is another dimension of familiarity less easy to state,where it will not be simply the reference, but those of its propertiesthat are at issue. In the case of a picture of candy, it will doubtlessbe the taste, whereas with the picture of a pear taste would bepuzzling. My sense is that here is where another application of pro­totype theory might be found: the metaphor must appeal to the wayinformation is stored, viz. what people spontaneously think of firstwhen presented with one or another stimulus. The work of EleanorRosch and her associates strikes me as profoundly relevant in iden­tifying the prototype structures of concepts, which in turn predictsuch things as the relevant frequency in our language of terms con­nected to the same prototype, the order with which these terms areacquired, and a host of other matters. Thus asked to name the firstanimal to come to mind, subjects are more likely to name dogs orcats than kangaroos or wildebeests, and "dog" and "cat" are morelikely to occur in printed texts than "kangaroo" or "wildebeest".Asked to name the first property of dog that comes to mind, peopleare more likely to say that they bark than that they lack sweat glands,as they are more likely to identify milk as white than full of calcium.The words a child first learns are likely to be of medium abstract­ness - the child is more likely to learn "dog" than "animal" (a higherorder concept) or to learn "cat" before she learns "Persian". And,as Jerry Fodor observes, not only do these words have a certain onto­genetic priority, they seem to be most easily taught through osten­sion. It is difficult to teach "animal" or "Persian" by example, but"cat" is gotten from picture books. Our prototypes define our world,

METAPHOR AND COGNITION 33

and it is by and large a shared world, on the basis of frequency andspeed of response. The rhetorician counts on speed and commonknowledge in making metaphors: He counts on his audience thinking"shape" when he says "pear", or thinking "crown" when he says"Queen" and so forth. Not merely thinking that, but thinking thatfirst and in a way that seems automatic and beyond the hearer'scontrol.Philosophers - I am thinking of Goodman - who think of resem­

blance simply as the sharing of properties, observe that everythingresembles everything else, since each thing shares properties witheverything else. But this logical truth overlooks the psychologicalreality to which resemblance appeals. The resemblances we seeare connected with the prototypes through which our world isconceptualized, and it is these that it is the great skill of the rhetori­cian to be able to activate, knowing how to move the mind as hemust. So Louis Philippe would be widely known in the societyin which metaphors about him were possible and indeed urgent, aswere the vehicles of metaphorical transformation he undertook touse and which Philipon did. That is, his clothing was metaphoricalbecause it was the familiar garment of the bourgeois paterfamilias- beaver hat, polished boots, waistcoat, cravat; and the pear wasnearly the prototypical fruit, a close second to the apple. The geniusAristotle appeals to consists in bracketing a king with a waistcoat,yielding the metaphorical "The King is a plain man" and the Kingwith a pear, yielding the metaphor that the king is greedy. The geniusconsists in putting two ordinary things together in a surprisingway. Both metaphors seek to reduce the king to his metaphoricalrepresentation, and if the king is what he is represented as, hisessence and his image are one. This is a consequence of theintensionality of the metaphor which in the end is made true by arepresentation.I am uncertain of the cognitive contribution of metaphors, but I

incline to the view that while they serve in a powerful way to fixour images of things, powerful because of the essentializing andreductive character they have, I am uncertain they ever, as metaphors,tell us something we do not know. For this reason I am indisposedto view certain claims in science as metaphors, even if they appearto be that. Consider, for example, the proposition that the heart is apump. This is, if a metaphor, a striking one because the heart wasknown long before the pump was, and there must have been a very

34 ARTHUR C. DANTO

narrow community indeed in which the pump was familiar enoughto serve a metaphorical office, when the pump did first come intobeing. Aristotle viewed the heart as a furnace, for example, and intruth today that would be a metaphor were someone to use it, justbecause it is known that it is not after all the office of the heart toheat the animal spirits and send them through the various tubes ofthe body to the distant limbs. So "My heart is a furnace" todayperhaps uses the word "heart" itself as a metaphor for the seat ofpassion. It is similarly possible for "My heart is a pump" to be ametaphor, but in its scientific use it is not: it is, rather, a theoryinitially and a fact finally of what the heart is. The test is inten­sionality. If the heart is scientifically speaking a pump, then whateveris true of pumps is true of the heart, at least at a sufficiently abstractand one might say essentialist view of pumps. "Louis Philippe is apear" is a compelling metaphor even if false (and difficult to shakeif it is false, which is why caricature is so terrifying to politicians).But "The heart is a pump" loses all interest if it is false, as does "Thebrain is a telephone exchange". If the latter were a metaphor, it wouldnot be defeated by its being false. So my sense is that metaphor hasno place in science.Nor has it a place in philosophy except when philosophy is liter­

ature, as it rarely is. "The world is a cave" is a powerful metaphorbut Aristotle's "The mind is a wax tablet" is not: for him a lot thathe wants to explain about the mind goes through if the brain reallyhas the properties of a wax tablet. "The body is a machine" is not ametaphor in philosophy, though it is not difficult to see how it couldbe one in life, as were I to say my body is a machine which workstirelessly and mechanically. My sense is that the concept of metaphorhas been expanded well beyond its importance, at least in cogni­tion, though perhaps insufficiently even now appreciated for itsimportance in politically defining reality for the mind.

NOTES

I Barnes (1984), p. 2334 =Aristotle, Poetics, 14593 5- 8.2 Barnes (1984), p. 2241 = Aristotle, Rhetoric III, 1405b 5- 6.3 Danto, A. C. (1981), pp. 188- 189.4 Davidson, D. (1968).

METAPHOR AND COGNITION

BIBLIOGRAPHY

35

Barnes, Jonathan (ed.): 1984, The Complete Works ofAristotle, Vol. II, PrincetonUniversity Press, Princeton.

Danto, Arthur C.: 1981, The Transfiguration of the Commonplace, HarvardUniversity Press, Cambridge, Massachusetts.

Davidson, Donald: 1968, 'On Saying That', Synthese 19, 130-146.

DAVID E. COOPER

TRUTH AND METAPHOR

'Metaphor', writes one literary critic, 'is a medium for fuller, riperknowing''! He is giving voice here to one of the many claims whichhave been made on behalf of metaphor as a special purveyor orvehicle of knowledge. If any of these claims are to be made out, eventhe less exaggerated ones, then some metaphors must be true. Ametaphor may, without being true, reflect, hint at, or inspire a searchfor, knowledge. But to count as a vehicle or medium of knowledge,a metaphor would have to be a bearer of truth. If the large claimsabout metaphors as vehicles of knowledge are to be defended,moreover, the metaphors in question must include challenging and'deep' ones. For it is these, and not hackneyed or trivial ones, withwhose contribution to knowledge the claimants have been concerned.In this paper, I examine the prospects for treating metaphors as

vehicles or bearers of truth. As the remarks at the end of the previousparagraph may suggest, there are two different issues involved: onepertaining to 'conventional' or 'established' metaphor; the other to'fresh' or 'live' metaphor. This distinction is a real and importantone despite the existence of many metaphors which we would notbe happy to assign to one rather than the other category. These bor­derline cases no more erase the distinction than the existence of her­maphrodites erases the distinction between the two sexes.I begin with a few words on truth and 'conventional' metaphor.

Since such sentences as 'I'm feeling down today' or 'I'm on top ofthe world' can be true, then conventional metaphorical utterancescan be truth-bearers. Indeed, they can have truth-values in thestraightforward ways that literal utterances can. Precisely becausethey can, some people argue that they are not any longer metaphor­ical at all. The metaphoricality of, say, 'down' in the above sentenceis now dead, and the word is straightforwardly ambiguous, like'bank' .2 In my view, there is reason to retain the notion of conven­tional metaphor. Much of what we describe in this way retains apower to generate, in a fairly systematic manner, novel and freshmetaphors, and in that sense has its metaphoricity preserved. 'PeterO'Toole went over the top' proves its continuing metaphorical life

37

F. R. Ankersmit and J. J. A. Mooij (eds), Knowledge and Language,Volume Ill, Metaphor and Knowledge, 37-47.© 1993 Kluwer Academic Publishers.

38 DAVID E. COOPER

by allowing the film critic to continue 'and having gone over, heliked what he saw and has stayed there ever since'.Although conventional metaphors owe their capacity as truth­

bearers to the very same feature as literal utterances - namely, theirestablished meanings - it would be a pity to overlook one distinc­tive aspect of the relation to truth which some of them display, whenconsidered diachronically. Erstwhile fresh metaphors may, as theycatch on or take root, help to effect a change in a concept. Thischange may result in, or indeed be constituted by, the possibility ofuttering new truths. A 'modern' concept of love as something whichdemands effort and compromise sounds very different from that ofthe troubadours or Lord Byron. In part, the development of theconcept may have been promoted by new metaphors which struck achord in people whose sensibilities were suffering from a surfeit ofromantic imagery of love. I have in mind metaphors like 'Love is acontract', 'We need to work at our relationship', or 'Love is a coop­erative enterprise'. If once fresh metaphors do playa role in theprocess of conceptual change, then here is a modicum of truth,perhaps, in Shelley's hyperbolic claim that metaphor 'creates anewthe universe'.We do not, here, have examples of metaphors which, when fresh,were true: but of ones which, through their appeal, contributed toconceptual revisions as a result of which new beliefs have come tobe accepted. They provide no evidence, therefore, for the claim thata fresh metaphor can itself be true: and this is the claim which thosewho speak of metaphor as a vehicle of knowledge and truth haveintended. It is to this claim which I now turn.

Most familiar accounts of metaphor provide room for the idea thata fresh metaphor can be true. They do so, moreover, in the same verygeneral way. A metaphor will be true when it is suitably related toother propositions which are literally true. Truth is, so to speak, trans­mitted from these propositions to the metaphor. Consider, forinstance, the suggestion that a metaphor is elliptical for a literal state­ment of comparison. If that statement is true so, necessarily, is themetaphor. Or consider the popular view that the metaphoricalmeaning of an utterance is identical with the literal meaning of aproposition which the speaker intended to convey, indirectly, by hisutterance. In that event, the utterance is metaphorically true just incase the intended proposition is literally true.3 If 'John is a dustbin'

TRUTH AND METAPHOR 39

is intended to convey, simply, that John eats almost anything, thenit will be true provided that 'John eats almost anything' is literallytrue.I have argued elsewhere that these, and many related accounts,

furnish mistaken analyses of metaphorical meaning.4 I shall notrepeat my criticisms here, but I do want to indicate the generaldifficulty facing any attempt to construe metaphorical truths asderivative from, or vicarious of, literal truths. What such accountsrequire is that a metaphor stands in a certain relation to other propo­sitions, and that the literal truth of these propositions confersmetaphorical truth upon the metaphor. The problem, as I see it, isof the Scylla and Charybdis variety. Either the metaphor is relatedto other propositions in a way which permits transmission of truthfrom them to it - in which event, unfortunately, those propositionsare themselves not literally true, so that the issue of non-literal truthremains. Or, the propositions to which the metaphor is related areliterally true - in which event, unfortunately, the relation is not onewhich permits transmission of truth from them to it.Let me give some illustrations of the first horn of the dilemma.

Accept, for the sake of argument, the ancient idea alluded to earlierthat a metaphor is equivalent to a simile of which it is the ellipsis.Now equivalence is a truth-transmitting relation, so if the simile istrue, so is the metaphor. But a simile qua simile, is not literally true.'Mrs. Thatcher is a bulldozer' is true, we grant, if 'Mrs. Thatcher islike a bulldozer' is true. But the latter sentence, if true, is not liter­ally so, in which case no progress has been made in deriving ametaphorical truth from a literal one. (To be sure, the Prime Ministermay, like the machine, flatten what lies in her path and dispose ofall obstacles - but she does such things only figuratively.) Orconsider the suggestion that a metaphorical description of, say, thesea is equivalent to a description of how the speaker experienced orsaw the sea as being.s Malcolm Lowry's ocean was a 'violetporridge' if that is how he experienced it. The trouble here is that,even construed as a description of the experience, the descriptionremains figurative. Not even Lowry, dipsomaniac as he was, liter­ally saw the ocean as a violet porridge. Consider, finally, the sug­gestion by one writer that 'in a scheme in which cities are thoughtof as women, "The Venice of the Renaissance was a noblewoman"would be true'.6 Here the suggestion is that a description will bemetaphorically true, relative to a scheme, if it describes, not how

40 DAVID E. COOPER

something actually is, but how it is thought of in that scheme. Thetrouble is that 'John thinks of cities as women' is not a literal descrip­tion of John - not unless he is hopelessly insane. Once again, then,one fails to reduce a metaphorical truth to a literal one.Steering towards Charybdis, we can accept that a metaphor is

importantly related to true, literal propositions - for example, by'prompting' or 'spurring' hearers into entertaining true thoughts.These relations, however, are not truth-transmitting. That a metaphor'prompts' me to have a true thought is insufficient for ascribing truthto the metaphor itself. A perceptive question, after all, can be a spurto the acquisition of true beliefs, but cannot itself be true. More per­tinently, we should not be overmodest and underestimate the con­tribution we make in turning a metaphor in the direction of truth ­when, for instance, we excize the silly or trite thoughts that it cansurely elicit. For instance, when I hear of the state as an organism,I tend to have the image, inspired no doubt by the frontispiece tothe Leviathan, of a huge man the scales of whose skin are individualcitizens. Fortunately I do not permit this image to influence myassessment of the metaphor. In putting the metaphor to work, I filterout the absurd thought of the state as a giant human being. To callthe metaphor 'true' when we have deployed interpretative imagina­tion and industry in hanging truths upon it is to give too much creditto it and too little to ourselves. Some metaphors may prompt prettyuniform thoughts in most hearers, and it will then be relativelyharmless to speak of them as true or false. Most of the exampleswhich get cited are, however, of more-or-Iess conventional metaphor,carrying an inherited load of established interpretation. Certainlythey cannot include those challenging, inspiring and variously inter­pretable metaphors on behalf of which people from Shelley to PaulRicoeur have proclaimed an essential contribution to truth.

These remarks are too brief to establish the failure of all attemptsto construe metaphorical truth in terms of a relation to literal truths,but they do suggest that a different direction might be morepromising. Instead of beginning with literal truth as the sole, or log­ically primitive notion, perhaps we could construct a broader, deepernotion of which literal and metaphorical truth would be among thetypes it covers. Metaphorical truths will not then be derived fromliteral ones: though since literal propositions will presumably remainthe paradigmatic truth-bearers, the new notion must demonstrate a

TRUTH AND METAPHOR 41

significant affinity between assessments of literal and metaphoricalutterances as true. Otherwise there would be no reason to take theexpression 'metaphorical truth' seriously.Construction of this broader, deeper notion would, if feasible at

all, be a complex undertaking. I begin by indicating some of thethoughts which must be taken seriously if construction of this notionof truth is to have either point or some prospect of success. Thefirst of these thoughts is that an adequate concept of truth mustexplain why truth is something people aim at. It is, so the thoughtgoes, part of the very meaning of 'truth' that people should gener­ally aim to make true, rather than false, assertions. This is whyTarskian accounts of truth have been criticized for telling us whattruth is only in the sense of providing a method for stating the con­ditions under which sentences are true, thereby remaining silent overthe role played by the concept of truth in our intellectual and com­municative activities. Michael Dummett once compared Tarski'saccount of truth to an account of winning a game which consistedin no more than stating the conditions under which the game is won.?An adequate account of winning should do more than tell us that,for example, a game of chess is won when a King is checkmated. Itmust, at the very least, explain that and how winning is the point ofgames: and this will require a description of the place which gameshave within our lives, a place which gives sense to notions ofwinning and losing. What is required, ultimately, is a philosophicalanthropology within which activities like games are assigned anintelligible position. The concept of truth, the thought goes, mustsimilarly be located within a philosophical anthropology whichdescribes the activities that lend the concept its normative role. Thepossibility would then arise that truth, so located and conceived, isnot an aim which it is the privilege of literal propositions alone toaccomplish. To suppose that it was might be like supposing thatwinning is the privilege of people engaged in ball-games only.This leads to a second thought. We should not follow blindly the

practice, evident in Chapter One of most books on truth, of dis­missing as 'secondary', 'figurative' or 'irrelevant' the ascription oftruth to objects such as paintings. Nor should we be dismissive ofuses of 'true' which are paraphrasable by words like 'authentic' and'genuine' - as in 'true friend', 'true emotion', etc. It is begging ques­tions to rule such ascriptions and uses out of court: to regard themas representing concepts which can be safely ignored by 'serious'

42 DAVID E. COOPER

theorists of truth who are concerned only with a 'central', 'primary'sense of the term.Among the items which get described, in everyday talk, as'authentic', 'genuine' and so 'true', are people and their relationsto the world. This leads on to the third, and most difficult, thought:the possibility that the truth ascribed to particular truth-bearers, suchas propositions and paintings, presupposes that those who producethese bearers - speakers, painters - stand in an 'authentic' relationto the world, a relation which itself must, to use a favourite expres­sion of Kierkegaard's, be 'in the truth'. Propositions and works ofart, this thought has it, will fail to be true if they are the emissionsor products of people who themselves are 'out of truth' in their com­portment towards the world. Suppose, for example, that people whoact, speak and paint as if there are objective moral certainties standin a distorted relation to the moral world. We may well, despite this,continue for everyday purposes to talk of true versus false moraljudgements and depictions, according to whether we subscribe tothem or not. But we will not regard such talk as being 'in the truth'.The confident assertion of a moral principle or a dogmatic work ofSocialist Realism, even if we go along with the speaker's or painter'ssentiment, will betray a wrong relation of speaker or artist to themoral world. Or suppose, to take another example, that one comesto think that distinguishing between actions for which people aredeemed responsible ('free' actions), and those for which they arenot, reflects an incoherent complex of practices, attitudes, beliefs,and emotions. Then one will be unable to hold, in all seriousness,that 'John is responsible for x' is true: not because one thinks thathis action was compelled, but because one is unable to concur inthe set of practices etc. reflected in such judgements.Crucial to examples like these is the thought that a comportment

towards the world (and to each other) is not to be fully representedby propositions or beliefs held to be true. A comportment belongsto a 'form of life' - a set of practices, natural responses, etc. - whichunderpins, and is not therefore constituted by, particular beliefs. The'distorted' stances towards the moral world and to the actions ofpeople, if 'distorted' is what they are, are not therefore to be artic­ulated as sets of false propositions. This is why it is more apt tospeak of such stances, not as being false, but as failing to be 'in thetruth'. The importance of this consideration will emerge when we

TRUTH AND METAPHOR 43

try to indicate the distinctive manner in which metaphors might beregarded as true.

If the thoughts adumbrated over the last few pages are taken seri­ously, then there are serious prospects for developing an account ofa broader, deeper notion of truth from which the modes of truthbelonging to, inter alia, literal propositions, metaphors, and paint­ings would be derivative. Such an account would be a descriptionof comportments or 'forms of life' which are 'in the truth'. The par­ticular products of comportments (propositions, etc.) will count astrue, in the final analysis, only to the degree that the comportmentsthemselves are 'in the truth'.How might such an approach cater for the concept of metaphor­

ical truth in particular? A metaphor will not, any more than apainting, have truth in virtue of its being equivalent to, entailing, oreliciting literally true propositions. The truth it will be derivativefrom will be that which characterizes people's relation to, comport­ment towards, the world. But this is equally the case with literal truthand so tells us nothing about a distinctive mode of metaphorical truth.Nor would it be sufficient for this purpose simply to furnish criteriafor distinguishing metaphorical from literal propositions. We wouldwant also to understand people's urge, or need, to express themselvesin metaphor, and the distinctive role of such expression as a dimen­sion of comportment towards the world - just as we would want tounderstand such matters in the case of painting or dance.A tempting suggestion might be the following. Not all of a com­

portment towards the world can be articulated as a set of proposi­tions held true, but human beings have, it seems, an inveterate urgeto try to articulate what is as yet unarticulated. Metaphors - those,that is, for which the vocabulary of truth is apt - are imaginativestabs at articulation. Some of them, the 'true' ones, are destined,once polished and refined, to pass into sedimented articulation. Butthis suggestion elides a distinction we should retain. Finding Hobbes'account of human psychology in terms of 'attractions' and 'repul­sions' naive and extremely crude, we may react by calling it'metaphorical'. But Hobbes, I think, intended it as a literal hypoth­esis. No one could similarly regard Hofmannsthal's description ofthe mind as a dovecote, and it is reasonable in my view to reservethe term 'metaphor', unless further qualified, for descriptions not

44 DAVID E. COOPER

intended by their authors as literal hypotheses, however imaginative.(Borderline cases will, of course, be plentiful.)Metaphors which are in, or out of, the truth are not embryonic

articulations of a comportment, and the test of their truth is not theirlater sedimentation into received theory. Such metaphors do,however, belong to the enterprise of bringing underlying features ofa comportment to the surface, where they may then, with luck, besurveyed and articulated. Theirs is the preparatory work of elicitingimages, evincing moods, triggering analogies which those intent on'straight' articulation may exploit and benefit from. Metaphorsmanifest, without themselves articulating, a comportment. (This, itmay be said, hardly distinguishes the work of metaphors from thatof the paintings and dances for which talk of truth is apposite ­unless the distinction is simply that the medium of metaphor iswords. But it is unclear to me that a distinction should be pressedtoo hard. Personally I feel no great reluctance in speaking of somepainting and dance as metaphorical in character.)It is, of course, incumbent on someone sympathetic to this

approach to metaphorical truth to remedy its vagueness and to spellout in a more concrete manner how a metaphor might 'manifest' or'express' a comportment, and so be in, or out of, the truth. Oneshould, however, note an obvious difficulty in spelling this out intoo much detail. It will not be possible to provide a full articulationof a comportment and then point to certain metaphors as an expres­sion of it, since on the approach in question no such complete artic­ulation is feasible. If it were, metaphors would not have the roleassigned to them by this approach, and could not be 'in the truth'in the suggested way. Not a few theorists of metaphorical truth spoiltheir insistence that there is something unarticulatable in the truthswhich metaphors reflect by then going on to state what these truthsare.sI have no very clear idea how to spell out what it would be, in

general, for a metaphor to 'manifest' or 'express' a comportment,and so share in whatever truth the latter may be 'in'. I would, though,like to cite some examples of metaphors regarded by their authors(or borrowers) in those terms, and which anyone keen to spell outthose terms would be advised to reflect upon.

Consider, first Wordsworth's sonnet, 'The World'. In this poem, heis doing three things. He indicates his sympathy for a view of reality

TRUTH AND METAPHOR 45

in which gods, Nature, and human beings are indissolvably united;makes clear that this view cannot consist simply in a cosmologicaltheory or set of beliefs; and provides some metaphors intended togive voice to that view. Regretting our contemporary inability tosee in Nature what is 'ours', Wordsworth wishes he were 'a Pagansuckled in a creed outworn'. But this 'creed' is less a doctrine thana relation to the world, a capacity to experience it in certain ways ­to 'have sight of Proteus rising from the sea;/Or hear old Tritonblow his wreathed horn'. And when, among the things with whichwe are now 'out of tune', Wordsworth lists 'this sea that bares herbosom to the moon;/The winds ... up-gather'd now like sleepingflowers', he is offering metaphors that sound the 'creed' which heis trying to recapture. These metaphors, if they are 'in the truth' atall, are so not because they naively describe what a sophisticatedcosmological theory would literally state, nor because they accu­rately describe what Wordsworth sees the world as. If they describeat all, it is more in the sense that Wittgenstein speaks of a pianist'sperformance describing the music. Such 'descriptions' are 'in thetruth' to the extent that they exhibit a relation, to Nature or to themusic, that is itself 'in the truth'.Consider, second, Roland Barthes' interpretation of the message

of Georges Bataille's book L'histoire de l'oeil. 9 In that book, Bataillepresents a bizarre array of metaphors in which eyes, testicles, eggs,planets and other more or less spherical objects get spoken of interms of one another. The point, suggests Barthes, is to convey thesense of a modern society, especially with respect to its sexual prac­tices, where identities are uncertain, boundaries shifting, rules trans­gressed, and categories dissolving. The relation or comportment ofpeople towards this modern world which Bataille is attempting toexpress is not fully articulatable: since the very concepts one wouldemploy in such an articulation are themselves subject to processesof dissolution and transgression of traditional boundaries. Themetaphors, therefore, are not stabs at what might later be literallyarticulated. Rather, they mimic, in their crazy cross-categorizations,the loss of stability and moral security which any authentic com­portment towards the modern world must recognize. The metaphorsare 'in the truth' to the extent that this comportment to the worldthey mimic is itself 'in the truth'.Finally, consider Heidegger's discussion of the truth of metaphors.

(Or, rather, the truth of what we would call metaphors, for Heidegger

46 DAVID E. COOPER

does not like the word - believing it to be encumbered by misleadingnotions, like that of 'transference' of meaning.)IO Particular truthslike 'That car is red', says Heidegger, 'disclose' particular phe­nomena to people, the colour of a vehicle, say. This is only possible,he argues, because people stand in a 'disclosive' relation to theworld. Truth, he suggests, attaches in the first instance to thisrelation, the precondition for structuring and interpreting the world,and only derivatively to the propositions which ensue from a wayof disclosing. The 'disclosive' relation is only an authentic one if itis self-consciously 'open': if, that is, people refuse to become fixatedin anyone way of interpreting their world. In Heidegger's termi­nology, people must hold themselves open for hearing various 'callsof Being'.This provides Heidegger with a reason for regarding some

metaphors as having a truth that others do not. One such metaphoris Holderlin's description of words as 'flowers of the mouth', whichis to be preferred to characterizations of words as tools or instru­ments. 11 It is not that, at a level of literal comparison, words are moresimilar to flowers than they are to hammers, but because the lattercharacterization betrays an attitude towards language which is atodds with an authentically 'open' comportment towards reality. Thisis the hubristic, technological attitude in which everything, wordsincluded, are treated as so much equipment at men's disposal forordering the world. Metaphors like Holderlin's, on the other hand,encourage us to 'hear the sound of language rising like the earth';to receive words, with patience and gentleness, as the intimationsthrough which the world opens itself up to us. The metaphor of wordsas tools also discloses an aspect of language, and has its element oftruth: but Holderlin's participates in that privileged kind of disclo­sure which draws attention to itself as a disclosure. The metaphordoes this through contributing to a sense of a world which remainspartly hidden beneath a surface, whilst permitting aspects of itselfto break through that surface for those who are patient enough towatch and listen. The metaphor thereby shares in the truth whichbelongs to this authentic relation between men, their language, andthe world.

I pass no comment on the plausibility or otherwise ofWordsworth's,Bataille's, and Heidegger's accounts of the comportments which theytake to be 'authentic' or 'in the truth'. I have rehearsed those

TRUTH AND METAPHOR 47

accounts as examples of the approach to metaphorical truth indicatedearlier: metaphorical truth as participation in, manifestation or reflec­tion of, a comportment towards the world which is 'in the truth'. Iwould like, one day, to be able to say something more precise aboutthis participation or manifestation. For the present, I content myselfwith offering the above illustrations as the kind on which reflectionis required if that precision is to be sought. Perhaps the search willnot progress very far: in which case given my negative remarks onmore familiar treatments of metaphorical truth, the conclusion wouldhave to be that metaphorical truth is a notion we do best to abandon.

NOTES

1 Wheelwright, P. (1954), p. 97.2 See, for example, Davidson, D. (1984).3 See, for example, Searle, J. (1979).4 Cooper, D. E. (1986), Ch. 2.5 See, for example, O'Hear, A. (1988), Ch. 5.6 Kittay, E. Feder (1987), p. 313.7 Cf. Dummett, M. (1958/9).8 Cooper, D. E. (1986), Ch. 4.9 Cf. Barthes, R. (1964).10 Heidegger, M. (1978), pp. 88ff.11 Heidegger, M. (1971), pp. 99ff.

BIBLIOGRAPHY

Barthes, Roland: 1964, 'La metaphore de l'oeil', in R. Barthes, Essais critiques,Seuil, Paris, pp. 238-245.

Cooper, David E.: 1986, Metaphor, Blackwell, Oxford.Davidson, Donald: 1984, 'What Metaphors Mean', in D. Davidson, Inquiries into

Truth and Interpretation, Oxford University Press, Oxford.Dummett, Michael: 1958/9, 'Truth', Proceedings of the Aristotelian Society 59.Heidegger, Martin: 1971, 'The Nature of Language', in M. Heidegger, On the

Way to Language, Harper & Row, New York.Heidegger, Martin: 1978, Der Satz vom Grund, Neske.Kittay, Eva Feder: 1987, Metaphor: Its Cognitive Force and Linguistic Structure,Oxford University Press, Oxford.

O'Hear, Anthony: 1988, The Element of Fire: Science, Art and the Human World,Routledge, London.

Searle, John R.: 1979, 'Metaphor', in A. Ortony (ed.), Metaphor and Thought,Cambridge University Press, Cambridge.

Wheelwright, Philip: 1954, The Burning Fountain: A Study in the Language ofSymbolism, Indiana University Press, Bloomington.

MARY B. HESSE

MODELS, METAPHORS AND TRUTH

1. A MODERATE SCIENTIFIC REALISM

In the English-speaking tradition of philosophy of language it hasgenerally been taken for granted that the ideal rational language isliteral and univocal and has a unique relation to truth. Its relation tothe real world is atomistic, that is to say, small portions of language,whether words, phrases, or sentences, attach themselves to the worldby some kind of correspondence or truth conditions, in a way thatis essentially independent of linguistic context. The presence ofmetaphors and other tropes in language is a deviation from rationalsense. As Hobbes put it, "such speeches are not to be admitted",and metaphors are abuses of speech by use of words "in other sensethan that they are ordained for; and thereby deceive others". Literallanguage in its relation to truth is held to be the proper vehicle forscience; it permits objective, testable, piecemeal accumulation ofknowledge and expression of belief. Metaphoric language on theother hand is ambiguous, holistic in meaning and context-dependent,and in this view fit only to express subjective attitudes and emotions.This primacy of literal language is closely connected with the

analysis of science, both in its 17th-century origins and in its theo­retical justifications. The early empiricists took scientific experienceand scientific knowledge to be models for all experience and knowl­edge. Thus the epistemological analysis of science became theanalysis of all language, of all expression of knowledge and beliefin every day contexts. But we now have the situation that the empiri­cist presuppositions that originally grounded this view have been dis­credited; on the other hand the philosophical bias towards thepresumed needs of science has become so deeply entrenched instandard analytic philosophy, that it has so far proved impossible toshift. It is time for philosophy of science to take the lead again, asit did in the 17th-century, and to demonstrate that a radical shake­up of analytic philosophy of language is required in the light ofmodern revisions of scientific epistemology. One of the motivationsand results of such a shake-up will be a better understanding of the

49

F. R. Ankersmit and J. J. A. Mooij (eds), Knowledge and Language.Volume III, Metaphor and Knowledge, 49-66.© 1993 Kluwer Academic Publishers.

50 MARY B. HESSE

nature and function of metaphor in natural language itself. In thispaper I shall argue not just that metaphor is as apt as literal languageto convey knowledge, but more radically, that metaphor properlyunderstood has logical priority over the literal, and hence that naturallanguage is fundamentally metaphorical, with the "literal" occur­ring as a kind of limiting case.I shall first outline, without too much argument, the case for a

moderate scientific realism that seems to me to emerge from recentcritique of scientific knowledge.! I shall then examine a controversybetween Cooper and Gadamer about a similarly realistic interpreta­tion of metaphor, and show how analogous arguments to those inphilosophy of science can be used to defend Gadamer's view of theprimacy of metaphor against Cooper's objections.The realistic view of scientific theory has generally postulated a

continually accumulating set of statements, based upon observation,and converging under ideal conditions upon the comprehensive truthabout the natural world. This is a strong version of realism that haslately been undermined by both epistemological and historicalarguments. First there is a problem about how to express the obser­vational data: the so-called problem of theory-Iadenness. Every"observation statement" describing empirical data has to be ex­pressed in some language or other, and every language containsgeneral descriptive predicates. Every set of predicates in a descrip­tive language implies a classification of the contents of the world.Aristotle believed that a natural classification of kinds or essencesis given by "intuition", but this view has turned out philosophicallyand scientifically untenable, not least because the intuitions of naturalkinds in different societies are very various, and most of them haveturned out to be unacceptable if taken as a basis for natural science.Modern versions of essentialism therefore tend to rely on science

itself for the discovery of natural kinds: natural kinds are those thatconform to the best scientific classification and the best system oflaws and theories. For example, the true classification of the materialelements awaited the delineation of chemical laws by Lavoisier andhis successors, and this, it is claimed by essentialists, is continuallybeing improved upon by better and better approximation to a truesystem of laws in later chemistry and physics. But the originalargument is now proceeding in a circle. We began with the accu­mulating data as one element in what was claimed to be a conver­gent process leading to true theory. But we have now found that in

MODELS, METAPHORS AND TRUTH 51

order to express even the data in the form of true propositions, wehave to presuppose that they are expressed in a language based ontrue natural kinds, and true natural kinds can only be identified if weknow what is the true theory. In reality we have to accept that thishermeneutic circle is consistent with a variety of different theoret­ical interpretations, even in natural science, and therefore that uniqueconvergence is illusory.The second problem about convergence concerns the particular

theoretical language and conceptual structure that has to be presup­posed to make strong realism at all plausible. Theories are imagi­native structures not wholly determined by data, and in order tochoose among an indefinite number of theories that more or less fitthe data, some principles of simplicity and plausibility need to beassumed. This is seen most strikingly as soon as some theory or spec­ulation is put forward which lies outside the generally acceptedframework tacitly assumed in the scientific community. The theoryis either rejected out of hand (one thinks of some discussions of para­psychology or alternative medicine), or at best it has to struggle formany years against entrenched paradigms before ultimate acceptance(one thinks of Copernicus, Galileo, and the Newton-Leibniz andEinstein-Lorentz controversies). It soon becomes very clear thatobjections to theories that transgress framework rules are hardlymitigated by how well they fit the facts. The officially received storythat science is wholly empirically based is inaccurate, and shown tobe so by scientists' own behavior.

It has sometimes been argued that the extra-empirical principlesneeded to restrict theoretical possibilities can be found a priori inconditions necessarily satisfied by the real world or by our percep­tions of it. But attention to the history of science shows that everyset of metaphysical or regulative principles that has been suggestedas necessary for science in the past has either been violated by sub­sequent acceptable science, or is such that we can see how possibledevelopments in our science would in fact violate them in future.For example, almost every traditional principle of space, time, matterand causality has been violated in modern physics.Scientific theories are models or narratives, initially freely

imagined stories about the natural world, within a particular set ofcategories and presuppositions which depend on a relation of analogywith the real world as revealed by our perceptions. Thus in the 17th­century the generally accepted model of the universe was mechan-

52 MARY B. HESSE

ical: everything was conceived as being like pieces of clockwork.Subsequently the mechanical categories fossilized, encouragedby Kant's attempt to found them upon a priori reasoning. But the20th-century revolution in physics required another story, anotherfundamental reclassification of the elements of nature, and of thecategories of space, time, matter, motion, action and causality. Therelation between one story and the next is not one of progressiveextension and refinement, and therefore not one of convergence, butis radical replacement, or in Kuhn's terms, radical revolution.This having been said, however, there is still some sense in which

science does progress through theoretical revolutions. Historicalexamples help to clarify what that sense is, and to show how far theconvergence postulate does capture something that is intuitively feltto be progressive about science. Consider the particular sequenceof conceptual revolutions in the history of the chemical elements,from the Aristotelian four elements through the phlogiston theory,to Lavoisier's list of elements. Each mode of classification of matterdisrupts and reorganizes the classification that went before, and thereis no obvious sense in which convergence of concepts can be main­tained. After Lavoisier, however, through Dalton, Avogadro, and intomodern physical chemistry, the list of elements proposed byLavoisier is not radically reformulated, but now chemistry has ceasedto be an autonomous science, and the underlying theories of physicshave been subject to radical revolution: the oxygen atom describedtheoretically by Lavoisier is quite different from that described inquantum theory.In spite of such radical revolutions, however, natural science is

instrumentally progressive. There is progress in the sense that wehave vastly increasing pragmatic possibilities of predicting and con­trolling empirical events by means of experiment and theory-con­struction. The human species employs a test-and-feedback methodas a natural extension of animal learning, in order first to survive inits environment, and later to change and exploit it. The method yields"objective" knowledge in this pragmatic sense. Progress takes placenot only in the sense of a numerical approximation of one theory tothe next in its detailed predictions, but also, as in the case of the listof chemical elements, because the concepts of one theory are useablewith sufficient approximation at certain levels of accuracy even inthe context of the next, conceptually different, theory. When Priestleyreferred to "phlogiston", his theory was false, and his conceptual

MODELS, METAPHORS AND TRUTH 53

scheme inappropriate, because it is not as he thought the same sub­stance that is involved in the phenomena of rusting and respiration,is a constituent of the atmosphere, etc. Nevertheless Priestley knewmany local truths about what we call "hydrogen", which he mistak­enly identified with phlogiston, and these are the truths that remainpart of our positive knowledge of observable objects.Such pragmatic progress does not entail or require convergence

of the conceptual framework towards universalizable truth. Pragmaticsuccess and approximate laws are always relative to particular, local,phenomena. Even if these phenomena extend to the galaxies, theynever encompass the possibly infinite universe, and however exten­sive they are in space, they are necessarily very local in time. Thewhole of the data is never in, and there is always room for furtherconceptual revolutions, however accurate the current theory may befor current purposes.Local approximation does not entail universalizability of concep­

tual frameworks, so it follows that truth claims at a given stage oftheory are strictly speaking non-propositional. Expressions of whatwe know pragmatically in one conceptual framework do not ingeneral preserve truth or meaning in another. There is no transi­tivity of entailment for sentences that are only approximately true,and no transitivity of "correct" application of general terms that areonly justifiably applied within limited domains of phenomena.In this way we arrive at what might be called a moderate realism,

lying between strong realism and instrumentalism. Such a moderaterealism of scientific knowledge turns out to be particular rather thangeneral, local rather than universal, approximate rather than exact,immediately describable and verifiable rather than theoretically deepand reductive. It is not the theoretical frameworks as such thatvalidate the claim of science to be a distinctive and reliable body ofknowledge, but rather the way in which they are used to further thefeedback method of successful prediction and control. How then cantheoretical models be regarded as significant in indicating the real,even though they cannot be taken as strongly realistic? The relia­bility of models for prediction depends on non-propositional ana­logical relations which they exhibit rather than state. What can bestated depends on the classificatory resources already present in thelanguage, and any scientific language is theory-laden with thatimplicit classification and never captures the world in its essence.Ideally the model says as much as can be said within the linguistic

54 MARY B. HESSE

resources available, and is enough for local applicability and pre­diction. This view is not a non-empirical idealism, because it doesnot deny that there is a real structure in the world, of which scienceprogressively reveals more and more. In this sense the view is realist.But it is not a strong realism in the sense that science can explicitlycapture this structure in an isomorphism of true natural categoriesand true descriptions.

2. THE PRIMACY OF METAPHOR

The moderate realistic view of scientific language can be extrapo­lated to a theory of natural language in general. The working of ascientific language in this view is a model for, and example of, theworking of metaphor in natural languages. Just as observationlanguage is theoretical and analogical through and through, but isstill the basis for realist descriptions and cognitive inference, so Ishall now argue that natural language is metaphorical through andthrough, and yet has cognitive meaning. I am talking here about whatmight be called descriptive metaphor - metaphor that has putativelysomething like "truth-value", that is, is cognitive. I am not talkingabout metaphor that is as the anti-rhetoricians used to say "purely"decorative or purely false or misleading. The question is whether"descriptive" metaphor is the empty class, or on the other hand,whether it is ever purely decorative. I am going to argue thatmetaphoric meanings have logical priority over literal meanings ina way analogous to the priority of theoretical over "observational"meanings in science. Let us call the thesis of the primacy of metaphorThesis M:

Metaphor is a fundamental form of language, and prior(historically and logically) to the literal.

The first point to be made is that the sense of metaphor here cannotbe that of a semantic trope or deviant form of language, since thatwould presuppose the prior existence of a literal language fromwhich the meanings of metaphors deviate. In other words, the usualsense of "metaphor" is opposed to "literal", and presupposes the verymetaphoric/literal distinction that is here in question. This does notof course exclude the other sense of metaphor, by means of whichwe make semantic distinctions within a working language, between

MODELS, METAPHORS AND TRUTH 55

what is standardly called "metaphoric" and what is called "literal".What it does entail, however, is a recommendation to understand thisdistinction in a non-standard way when an account of language sat­isfying Thesis M is in place.

It follows that there is one common form of argument againstThesis M that is not cogent. This argument is that since the standardsense of metaphor is only understood in contrast to the literal, andonly applied to linguistic utterances which are identified as not beingliteral, the literal must be prior to the metaphoric, which is depen­dent on it. This argument simply begs the question of the existenceof a literal/metaphoric distinction as a fundamental feature oflanguage, by presupposing the standard sense of "metaphoric" and"literal" which it is precisely the aim of Thesis M to question andreplace.I shall examine Thesis M by looking first at some recent argu­

ments by David Cooper. In his book Metaphor he has attempted toexpand Thesis M into a theory of language, which theory he thengives reasons to reject. He chooses Gadamer as his foil, but hisversion of Thesis M seems in fact to owe more to Vico's theory oflinguistic origins than to Gadamer, and I shall later suggest an alter­native development of Gadamer's ideas in terms of which to replyto most of Cooper's objections.Cooper's interpretation of Thesis M depends on the idea that

language develops in two stages (which seem to be intended as his­torical stages, since he speaks of dating a moment of change).2 Firstthere is an Ursprache which is wholly metaphorical in the sense thatit unselfconsciously expresses fundamental similarities between par­ticulars, and is independent of the classification of things undergeneral concepts to which words refer. This is intended to reflectGadamer's account of the "natural formation of concepts", aboutwhich Gadamer says:

there is no explicit reflection on what is common to different things, nor does theuse of words . .. regard what they designate as a case that is subsumed under auniversal)

Like Wittgenstein, Gadamer takes a novel application of a word tomodify the concept previously expressed by it, in such a way thatthe novel application cannot itself be dictated by the previous conceptby automatic application of rules. For Gadamer, the new applications

56 MARY B. HESSE

are not, however, arbitrary, as Wittgenstein and Nietzsche both seemto suggest, but are based on perception of similarities.In the second stage of language as described by Cooper, things

are classified into genera referred to by universal terms, and theseconstitute concepts and the "meanings" of words. This is the stageof our standard literal language, but according to Thesis M it has tobe prepared for by the "logical advance work" (Gadamer's phrase)of the earlier stage.4 It follows for Cooper that in the Urzeit there isno truth or falsity in the sense of standard semantics, since thisdepends on having a literal language with fixed meanings. Cooperalso suggests that in the Urzeit there can be no "correctness" or"incorrectness" of the "natural" application of words either. Hisargument for this extension of the no-truth-value condition is that acrucial feature of metaphor is that there are no rules of correctnessfor the metaphoric transfer of meaning, because new meanings arefreely created or stipulated. Hence Thesis M implies that theUrsprecher can have no such rules.sThus set up, the "Gadamer" theory is immediately open to cogent

objections. Cooper makes three main points. First, and most impor­tant, if metaphoric talk in the Ursprache is neither true nor false,correct nor incorrect, it cannot constitute a language. Rules oflanguage-use do not have to be explicit (even in literal language theyare often not), but they do have to be available, otherwise it wouldfor example be impossible to teach the language to children, whohave to be inducted and reinforced in correct usage.6 In the termsof Cooper's discussion this is a knock-down argument. I shall replyto it later in terms of what I take to be Gadamer's real intentions;here it is sufficient to notice that the argument depends on theassumptions (1) that there are no rules (even implicitly) ofmetaphoric talk, and (2) that the truth or falsity, correctness or incor­rectness of metaphoric talk would have to be understood in termsof standard philosophy of (literal) language. I believe Gadamerwould reject both these assumptions.Cooper's second objection is that metaphoric talk requires aware­

ness by speakers that it is metaphoric. If the Ursprecher have nocategories in terms of which they can reflect upon the language, itthen follows that they cannot speak metaphorically. Cooper'sargument seems to be this. As Quine has shown, understanding alanguage, whether of a foreign clan or one's own, requires a theoryof meaning that enables a distinction to be made between what is

MODELS, METAPHORS AND TRUTH 57

uttered and what is said. This is the case for all natural languagespeakers, who are liable to make mistakes, to utter ambiguities andinconsistencies, and sometimes to engage in maverick talk includingmetaphor. The theory of meaning consists of conditionals of the form"If X utters S, then, cet. par., he is saying that p", and "If X saysthat P, and if his utterance is not maverick, then, cet. par., he believesthat P". Now according to Cooper's earlier account,7 metaphoricutterances do not have the dominant aim of expressing the speaker'sbeliefs about how things are, that is, metaphors are not primarilycognitive. Quine's clansmen must therefore distinguish betweenmetaphoric and normal talk, for if not "they will find themselvesattributing beliefs to one another which are obviously not held....The thesis of the primacy of metaphor must be wrong; for it cannotbe that, unknown to the Ursprecher themselves, they are employedfull-time in speaking metaphorically".8This argument occurs on the last few pages of Cooper's book. The

most obvious thing to be said about it is that he seems here to haveabandoned his own previous allowance that whatever Gadamermeans by "fundamental metaphor", it cannot be metaphor in thesense of "maverick", or of the metaphoric/literal distinction. Theargument thus seems to be clearly circular if taken to be an objec­tion to Thesis M. There is, however, an important point lurkingwithin it, namely the assumption that normal, standard, or literal talkhas the purpose of conveying beliefs, whereas metaphoric talk doesnot. For one who espouses Thesis M it is therefore important to askwhether and in what sense metaphor can have the function of con­veying beliefs and stating how things are. This is what I shall attemptto do in the next section.A further, minor, difficulty that Cooper finds in Gadamer's account

lies in its dependence on transfer of meaning by similarity.9 Thiscannot, he says, be a privilege of the Ursprecher, that is, it cannotbe a sufficient characterization of their metaphoric talk, because weall from time to time, in literal language, extend meanings bysimilarities. Once again, however, this argument is circular, becauseif all talk is fundamentally metaphorical as required by Thesis M,transfer of meaning by similarity in so-called literal language willbe just what we should expect. In accordance with Thesis M such"literal" language has to be interpreted as some form of metaphor­icallanguage, certainly not distinguishable from it in terms of depen­dence on similarities.

58 MARY B. HESSE

3. GADAMER ON GADAMER

Gadamer's assumptions, methods and forms of expression are notthose of analytic philosophy, and it is often difficult to extract fromhis writings any sharply defined theses or arguments. Cooper's is theonly account I have come across that tries to do the analytic bestfor him. The attempt does, however, have the appearance of settingGadamer up for expedient demise, and therefore ultimately fails. Letus go back to Gadamer and see if some other type of rigorous sensecan be made of his discussion of metaphor.The discussion in Truth and Method to which Cooper refers occurs

in the context of a section called "The emergence of the concept oflanguage in the history of western thought",1O It is important to noticethat the historical development Gadamer is talking about is not ahistorical myth about the origin of language (he nowhere mentionsan Ursprache), but the successive accounts of language in the whollyaccessible classics of western philosophy. Gadamer is not concernedwith the history of language, mythical or not, but with "a constantprocess of concept formation by means of which the life of alanguage develops". Lest there be any doubt about this, Gadamerspeaks favorably of Aristotle's account of the convention of meaning:"Like every question of origin, [it] is for him a way to the knowl­edge of being and value ... in regard to language [this account]characterizes its mode of being and says nothing about its origin"(my italics),11When this is understood it becomes clear that Gadamer's thesis

is more radical than Cooper's version: it is not just some Ursprache(a "language of Adam")12 that is fundamentally metaphorical, butour own natural language. This at once makes it more urgent to givea radically revised account of what our own natural language is,and at the same time undermines those objections to Thesis M thatdepend on our already having a clear and true account of what it isin terms of its literal character, standard meaning and truth condi­tionals, and the standard accounts of understanding and communi­cation.Gadamer is quite clear that the sense in which language has a fun­

damentally metaphoric character is not the sense of metaphor asdeviance: "it is important to see that it is the prejudice of a theoryof logic that is alien to language if the metaphorical use of a wordis regarded as not its real sense".13 His theory goes beneath the emer-

MODELS, METAPHORS AND TRUTH 59

gence of logic in western philosophy to analyse what he calls"natural concept formation". He starts with the assumption that noultimate order can be apparent to finite minds, and that there is afundamental inexactness of all human knowledge. The structure oflogic that has been superimposed on natural language since Aristotlecannot capture the essences of things or of meanings, and thereforehas only relative truth. Aristotle has made "the logical ideal ofthe ordered arrangement of concepts. .. superior to the livingmetaphoric nature of language, on which all natural concept forma­tion depends."14 It does not follow, however, that no ultimate orderof things exists. Gadamer expresses this possibility of ineffable exis­tence with reference to the Christology which entered philosophywith, for example, Aquinas and Nicholas of Cusa: "Christologyprepares the way for a new philosophy of man, which mediates ina new way between the mind of man in its finitude and divineinfinity".15This reference is obscure to a modern secular philosophy, but it

may be glossed by comparison with the similar problem aboutultimate order in philosophy of science that was discussed in the firstsection. Just as no scientific model, or conceptual structure of a sci­entific theory, can capture accurately the order of the natural world,so no concepts embodied in the general terms of a language cancapture the order of things in general. But in science this does notimply a non-realist scientific theory because there is always a con­straint from experience and experiment upon the detailed working­out of any particular conceptual structure. This constraint is whatGadamer calls the "what is" that "is not this or that thing, but 'whatcannot be done away with"',16 and the knowledge of "what is" isproperly gained from experience, prior to conceptualization.Gadamer's image of Christology, the theory of divine incarnation,is for him the guarantee that ultimate divine order exists and ismediated to humans, not in an ideal language, but in thedivine/human nature of Christ. Gadamer is a Christian realist, not arelativist; there is ultimately a "unity of truth",!7Gadamer goes on to argue that the logic of induction and abstrac­

tion is misleading if it claims prematurely to capture that order infinite terms, and for human language the claims are always prema­ture. That is the fundamental reason why Gadamer espouses ThesisM, namely that "all language is metaphorical". Concept formationis always both a creative and corrective process of applying previ-

60 MARY B. HESSE

ously learned general terms to new particulars, and extending andmodifying them in the process. For Gadamer there are always generalterms in the language which have pre-established meaning (that is,which function according to rules), but their functions and meaningsare continually changing with experience and use. This is why it isa serious distortion of Gadamer's view to represent him as postu­lating a historically prior stage of language in which there are nocategorizations and no rules. It remains, of course, to show moreclearly than Gadamer does that such a rule-governed but unstablelanguage is possible.Gadamer's only references to the type of "rules" involved are to

the perceptions of natural similarities, but he does require that the"logical advance work of language" (before explicit categorizations)should be given "scientific justification". In order to understand howthis works we need to look at his general account of experienceearlier in Truth and Method. There he argues that the epistemolog­ical analysis of experience in the philosophical tradition has beenwholly oriented to science: to the attempt to make experience repeat­able and "objective" as between different observers. But not all expe­rience is related to the goals of science, valid as those are in theirproper place. Sometimes creative insight, confirmed by positiveinstances of experience, should be valued as highly as falsifiabilityand correction by negative instances. Indeed both positive andnegative moments are required: the positive to create concepts pro­visionally embodied in language; the negative to correct and modifythese by further experience.It is this dialectical process18 that is the permanent condition of

language, and it is not unlike the constant conjecture, refutation andreplacement of scientific theories, if these are interpreted in the mod­erately realistic way I have described. The dialectical process is alsowhat Gadamer calls the fundamental hermeneutic experience ofunderstanding and communicating with other persons in the guess­test-and-modify conditions of every human conversation, and whichalso grounds every attempt at understanding in the human sciences.i9

Can the possibility of a language based on this account of expe­rience be demonstrated more clearly than by vague reference todialectics? Witttgenstein's use of family resemblances gives us theclue that is needed to flesh out Gadamer's account of perceptual sim­ilarities.20 For Wittgenstein, objects may form a class to whosemembers a predicate P is correctly ascribed in a given stage of

MODELS, METAPHORS AND TRUTH 61

language, without assuming that there is any universal "P-ness"realized by the object. Instead we assume that in a family resem­blance class (for example "game"), the members of enough pairs ofobjects in the class resemble each other in some respects relevantto P, and are appropriately and sufficiently different from objects putin the class not-Po The resemblances can form a chainlike structurethrough a given class in such a way that there are relatively clearcases of objects falling within it, and relatively clear cases of thosethat do not. There may of course also be borderline cases where weare uncertain.New experiences give further information about the relative sim­

ilarities and differences between objects already known and aboutnew objects. This enables, as it were, a new computation to be doneto produce the best classes with maximal internal similarities andmaximal differentiation from complementary classes. As the processcontinues the extensions of classes will generally be modified, some­times at the edges and sometimes radically, as in a Kuhnian con­ceptual revolution. "Meanings" now become meaning relationsbetween objects, which are functions of perceptions of similarity anddifference, and of the rules of classification in the light of these per­ceptions. Classifications, and therefore meanings, may change withchanging experience, and with changing purposes for which the clas­sification is made, but such instability of meaning does not implythat there are no rules for the applications of words to things. Therules are those of the computation of classes based on experience.No doubt Wittgenstein would intensely dislike the notion of

reducing his family resemblances to rules and computations, but itis perfectly possible to see his conception as rule-governed. Indeedcomputer simulations of such processes are frequently applied toall kinds of taxonomic problems, and there is some evidence thatsuch simulations also model brain processes more or less adequately.The suggestion that this is the new kind of semantics that is requiredto understand linguistic universals and meaning relations, is a radicalone. But its possibility and partial realisation in several empiricalfields is enough to show that Gadamer's notion of "natural conceptformation" can be rule-governed. Of course this is not to say thatlanguage speakers "know" how the brain is programmed, any morethan most of them know explicitly the rules of Aristotelian grammaror logic. The process is a "natural", habitual one, which is taughtand learned by example and positive and negative reinforcements

62 MARY B. HESSE

within the language community. How this is done in detail is aquestion for empirical study in psychology, physiology, sociologyand artificial intelligence. It is most unlikely that it is done whollyby having an explicit Tarskian model of literal language conditionalsin the brain.That the rules of classification by similarity are not explicitly

known to speakers is the point made by Gadamer when he says that,in natural concept formation, there is no explicit reflection on whatis common among particulars. The explicit abstraction of classes,essences, universals, rules, "comes later", logically rather than his­torically later, because reflection on rules presupposes the existenceof rules. Again it must be emphasized that what Gadamer is doingthrough these pages on metaphor is not a history of the origins oflanguage, but a history of human thought about language. Thisthought, he believes, has been distorted by the need to analyzescience, and has led to a simplistic literalism of everyday languageas well as of science, forgetting that natural language has other formsthan the scientific.Reflection on classification processes is one thing, the process

itself is another. With regard to the natural process itself Gadameris quite clear that "things can be articulated in different ways ...according to the correspondences and differences")l Languages maytherefore exhibit varied forms of the logic of natural experience,depending on human needs and interests. Such recognition of culturaland historical diversity is something that causes problems forstandard theory of language, and leads to polarization of viewsbetween "realism" and "relativism". In this interpretation ofGadamer's view, however, these difficulties need not arise, becausethe natural brain-semantics suggested may be common to alllanguage users, without dictating any ideal form of language asnecessary for human communication. Different orders can be createdby different languages in the same way that different scientifictheories may be viable as adequate but provisional accounts of thenatural order. How communication can take place between differentlanguages on this view is something I return to below.Before leaving this interpretation of Gadamer we must give some

account of how it can accommodate the distinction of metaphoricaland literal which undoubtedly does have application in everydaylanguage, and which depends on the "deviant" sense of metaphor.Gadamer does not explicitly address this, but it is not difficult to

MODELS, METAPHORS AND TRUTH 63

see what must be said. For Gadamer, logic (and by implication literallanguage also) have only relative truth. That is to say, they can haveapplication in natural language only after the "logical advance work"of fundamental categorization implied in the general terms of aspecific language. Meanings only have stability so long as the naturalcategories are frozen at a particular stage or for particular purposes.The pragmatic and instrumental purposes of science constitute onesuch special case - one may compare with Kuhn's alternation ofthe relatively stable paradigms of normal science and the revolu­tionary episodes of category and meaning change. According toGadamer the mistake of positivist theories of language which areover-influenced by science, has been to take the stability of "normallanguage" as norm and as realized ideal, implying that words becomemere signs which mayor may not have their ghostly Platonic formsin non-linguistic "meanings" and "universals":

It can be stated as a fundamental principle that wherever words assume a mere signfunction, the original connection between speaking and thinking, with which weare concerned, is changed into an instrumental relationship. This changed rela­tionship of word and sign is at the basis of concept formation in science and hasbecome so self-evident to us that it requires a special effort of memory to recallthat, beside the scientific ideal of unambiguous designation, the life of languageitself continues unchanged.22

The realist/nominalist controversies about language are thereforesuperseded by a more fundamental understanding of natural conceptformation.

4. REPLY TO OBJECTIONS

Finally we can return briefly to Cooper's objections to his ownversion of Thesis M. These reduce to two major arguments. The firstis that no Ursprache nor any other natural language can both bemetaphorical through-and-through, and constitute a proper language,because metaphorical talk implies an absence of standards ofcorrectness. We have seen that this premise is false: there can bestandards of correctness built into the metaphorical development ofconcepts and the natural learning of a language. These standardsare habitual, not explicit, and for their rationale we must look to theevolution of the human brain as it has come to cope with its naturaland social environment. Tarski-like truth and meaning conditions are

64 MARY B. HESSE

but the tip of the programmed brain, visible only when a relativelystable language is frozen and reflected upon.Cooper's second major objection is that metaphoric talk requires

awareness by speakers that it is metaphorical, and hence awarenessof metaphor as distinct from literal usage. The primary reason headduces for this is that literal utterances have (cet. par.) the functionof communicating beliefs about what is the case, whereas metaphoricutterances do not. I have already pointed out the circularity of thisargument, since it presupposes that literal and metaphoric languageare distinguished by the important characteristic of expressing or notexpressing beliefs. From this of course it would immediately followthat no wholly metaphorical language could function as a naturallanguage does. But if Thesis M is to be maintained at all, it is anecessary rule of strategy that the "metaphoric" must be so reinter­preted and understood as to subsume the characteristics that so-calledliteral language clearly has, among which are its subjection to rules,its relative meaning stability, and its use for expressing beliefs. Thisis a strategy Cooper has ignored in his discussion. We have seen thatwholly metaphoric talk can follow rules, and can yield relativemeaning stability in special cases. We now have to ask whether it isindeed possible for metaphoric talk to express beliefs.Our Wittgensteinian model has shown that relative meaning sta­

bility may be produced by freezing a natural language at some stageof development, or by the relative rarity of new experiences and newpurposes of language use. Under these circumstances the model canfunction like a "literal" language, and therefore among other thingsit can be used to express beliefs. More fundamentally, however, wecan conceive of a dynamic language (one that is "more metaphor­ical") also expressing beliefs, although these may be of a morespeculative and ambiguous nature, and may require negotiationbetween speakers to draw out their possible "meanings". Newton,for example, expressed his beliefs about "force" and "mass" bysetting out some experiments, and his axioms and laws of motion.Both "force" and "mass" were used metaphorically, that is, not inaccordance with contemporary custom, but the extensions andcorrections of meaning involved were implicitly shown within thestructure of the theory itself - the theory was a recommendation toreclassify, to reject the necessary connection of force with push-pull,and to recognize all material bodies in the universe as "masses"within the meaning of the theory. The metaphors were not immedi-

MODELS, METAPHORS AND TRUTH 65

ately understood or accepted. They led, among other things, to a pro­tracted controversy with Leibniz, a negotiation cut off by Leibniz'sdeath, but gradually settled in Newton's favor in the analyticmechanics of the 18th century.Examples of expressions of belief by metaphorical shifts of

meaning need not be multiplied, because the literature on metaphor(including Cooper's book) is a rich source for them. But we usuallyhave to exercise a metaphorical shift of vision to see this. Wherephilosophers of metaphor have sought to translate metaphor intoliteral terms, whether by comparison or substitution or what Coopercalls the "vicarious truth" of literal paraphrase,23 we have to read thisdifferently, as the explication of some metaphors in terms of others:that is to say they are new, creative shifts of meaning in terms ofparaphrases which use old familiar categorizations which werenevertheless themselves metaphors. These paraphrases will gener­ally take up more space, just as the introduction of Newton's theoryto a schoolboy brought up on Aristotle in 1687 would have involveda great deal of time and talk in Aristotelian language. Even then theparaphrases and explanations will never be exact or guaranteed tosucceed in conveying Newton's beliefs to the schoolboy mind. Acreative "grasp of the concept" is needed on the part of hearer aswell as speaker, hopefully to be confirmed or corrected by subse­quent talk and experience in just the way Gadamer describes. It isthis feature of inexactness and insufficiency of definition in theunderstanding of metaphor that has led other analysts to rejectaltogether the notion of metaphor as conveying belief, and hence toconsign metaphor to expression of non-cognitive attitudes, emotions,etc. Of course metaphor does have functions other than the cogni­tive, but no argument yet adduced has shown that a language whichis metaphorical through and through cannot have cognitive functionsalso in the strong sense of being capable of expressing beliefs. Toreject this possibility is to have an unrealizable ideal of the exact­ness and univocality of cognitive language. The strong version ofconvergent scientific realism with which I began is just an applica­tion of this unrealizable ideal, and I hope to have shown that its rejec­tion paves the way for a more adequate theory of natural languagein general.

66 MARY B. HESSE

NOTES

1 For more details see Hesse, M. B. (1992).2 Cooper, D. E. (1986), p. 269.3 Gadamer, H. -G. (1975), p. 388.4 Cooper, D. E. (1986), p. 259.5 Ibid., pp. 267-268.6 Ibid., pp. 272-274.7 Ibid., p. 106.8 Ibid., pp. 278-279.9 Ibid., p. 265.10 Gadamer, H. -G., (1975), p. 366.11 Ibid., p. 391.12 Ibid., p. 396.13 Ibid., p. 389.14 Ibid., p. 391.15 Ibid., p. 388.16 Ibid., p. 320.17 Ibid., pp. 396-397.18 Ibid., pp. 317, 390, 393.19 Ibid., p. 324.20 I have developed this idea in Hesse, M. B. (1974), Ch. 2.21 Gadamer, H. -G. (1975), p. 394.22 Ibid., p. 392.23 Cooper, D. E. (1986), p. 200.

BIBLIOGRAPHY

Cooper, David E.: 1986, Metaphor, Blackwell, Oxford.Gadamer, Hans-Georg: 1975, Truth and Method, G. Barden and 1. Cumming(transl.),Sheed & Ward, London.

Hesse, Mary B.: 1992, 'Science Beyond Realism and Relativism', in DiederickRaven, Lieteke van Vucht Tijssen and Jan de Wolf (eds.), Cognitive Relativismand Social Science, Transaction Publishers, New Brunswick, pp. 91-106.

Hesse, Mary B.: 1974, The Structure of Scientific Inference, Macmillan, London.

J. J. A. MOOIJ

METAPHOR AND TRUTH: A LIBERAL APPROACH

I

Metaphors can have different functions in various contexts of humancommunication. Three main functions of metaphor are: the emotive,the persuasive, and the cognitive function. In the first, a metaphorserves to express and transfer emotional attitudes; in the second, itserves to persuade the audience with respect to a course of actionor a point of view; and in the third, it serves to express cognitiveinsights.Striking examples of the emotive function of metaphor are

metaphorical terms of abuse or praise, like "beast" or "angel".Examples of the persuasive function can easily be found in politicaland commercial propaganda. Metaphors with a cognitive functionabound (or seem to abound) in philosophy, criticism, the humanities,and even in science.In practice, the first and the second function will often overlap,

as will the second and the third.However, whether the cognitive function is really an independent

function of metaphor has been a matter of debate for centuries. Manyphilosophers have been of the opinion that, even when metaphorsseem to express cognitive content, they are basically employed toevoke a vivid, expressive picture of the relevant state of affairs. Inother words, metaphors were basically considered to be vehicles ofemotion or persuasion, detrimental to the progress of knowledge.This was the traditional view of empiricism, inaugurated by its keyfigures, Hobbes and Locke. More recent critical assessments oflanguage (by Fritz Mauthner and others) have given further impetusto the attack on and suspicion of the cognitive value of metaphor:in principle, all metaphors are then held to be falsifications ofreality.Even more ancient than the empiricist view was the notion,

embodied in much rhetorical thought, that ultimately metaphors hada merely decorative function. Accordingly, their cognitive contentwas considered to be equivalent to their literal paraphrase, and

67

F. R. Ankersmit and 1. 1. A. Mooij (eds), Knowledge and lAnguage,Volume III, Metaphor and Knowledge, 67-80.© 1993 Kluwer Academic Publishers..

68 J. J. A. MOOIJ

although this would not make them necessarily detrimental to knowl­edge, it would not give them any positive cognitive function either.To be sure, other philosophers in the past have acknowledged

metaphors as very useful or even essential vehicles of truth.According to their view, metaphors can help to reveal aspects ofthe world that would otherwise have remained hidden. Metaphorsare then thought to extend the range of objects that can be discussed,described, characterized, understood, and explained - briefly: known.The history of this standpoint is really very complex, because itincludes rather sober views in an empiricist vein of thought as wellas highly speculative ideas about the ability of metaphor to expressmetaphysical or religious truths - and all kinds of attitudes inbetween.!This positive view has lately won ground in philosophy, whereas

the two traditions mentioned before are now largely on the defen­sive. Notes of warning against the falsifying character of metaphor­ical descriptions are now seldom heard in philosophy (the philosophyof science not excluded). It seems symptomatic that the most widelyknown recent attacks on metaphor should not stem from philosoph­ical quarters (in a formal, official sense), but rather from the side ofliterary and social criticism; I am referring to Susan Sontag's Illnessas Metaphor (1978) and Aids and Its Metaphors (1988). In SusanSontag we find the continuation of the negative view of metaphor,the stress on its undesirable or even dangerous effects; but althoughthe former book contains a telling warning against the use ofmetaphor (there being no clear boundary-line between use andabuse), it has not had any sobering influence on the generally positiveview of the cognitive function of metaphor prevalent today.This is not meant to ironically discredit the cognitive function of

metaphor. I strongly believe that metaphor can serve the progressas well as the dissemination of knowledge. I should even like toargue that metaphorical statements may be true, in the strict senseof that term. But I also feel that this optimistic assessment of thecognitive value of metaphor has gained favour because of ratherradical changes in the general view of knowledge and science them­selves. It seems to be connected with the use of models and per­spectives and accordingly with a somewhat idealistic epistemology.Moreover and even more significantly, it is also dependent on acertain blurring of the difference between cognitive discourse andpersuasion. In so far as the belief in a pre-existent, mind-indepen-

A LIBERAL APPROACH 69

dently structured world as the unshakable object of knowledge haslost ground, persuasion has gained a legitimate foothold in the cog­nitive enterprise and, indeed, in science itself. If the structure of theworld is not simply there to be discovered and described, but if it ispartly a human construction, not only in the Kantian sense but alsoas a matter of human decision and preference, then the choice of amodel or a theory or a conceptual framework may partly be the resultof agreement instead of being exclusively dependent on a corre­spondence with reality. And such agreement may, in its turn, partlybe the result of mutual persuasion.This may explain why the optimistic assessment of the value and

force of metaphor should at the same time often be an uneasy one.A rather extreme example of this tendency is Donald Davidson'sargument that metaphors, although highly useful in human commu­nication and far from generally detrimental to knowledge, only meanwhat they literally mean, and that they should therefore, strictlyspeaking, be considered typically false. As far as their metaphoricalmeaning is concerned, they could not be true, since there is no suchmeaning. However, their purport or suggestive force can be trueenough, their very absurdity can be inspiring.2 Kindred opinions havebeen brought forward by F. C. T. Moore and David S. Miall, andvery recently also by David Cooper in his book Metaphor (1986).3Even Max Black, whose influence in upgrading the cognitive

import of metaphor has been dominant now for about thirty years,hesitated to call any metaphor, even a revealing one, really true.4

Ted Cohen has urged that philosophical respectability should not beequated with the possibility of stating the truth. S Arthur Danto aswell as David Cooper warn us not to expect too much from the cog­nitive value of metaphor. To be sure, they do not throw doubt onthe capability of conventional metaphors to state the truth. Evenfresh, new, creative metaphors may say something that is true.According to both of them, however, this would normally only be atruth we know already. New metaphors would be powerful meansto express an old truth, not to express a new one. And it would onlybe, as Cooper argues, by finding a new concept of truth (or byrenewing an old concept) that a fruitful connection between metaphorand truth can, possibly, be reinstigated.Perhaps one might conclude that several representatives of the

positive view of the value of metaphor have assimilated certain ratherunobtrusive elements of the other (primarily negative, suspicious,

70 1. J. A. MOOIJ

or radically critical) traditions. I do not want to repeat their argu­ments, but I must admit that there are yet further reasons for hesi­tation in ascribing truth to metaphorical statements, even when suchstatements are accepted as illuminating and informative.These further reasons apply only to fresh or original metaphors.Indeed, with conventional metaphors there is no serious problem(even according to persons who are rather sceptical about the cog­nitive possibilities of metaphors, as it appeared above). Let me justmention some examples. A person may be called an elephant or aserpent, or may be said to speak in a cutting tone or to be drunk withjoy. Or consider a passage, taken from The Observer of April 23,1989, speaking about "Mr. Kinnock's wholehearted embrace of theEuropean Community earlier this year" (p. 1). Most readers orhearers will easily understand what is meant, and such statementscan be perfectly true; of course, they may also be false, and occa­sionally their truth value may be difficult to determine, due to theirvagueness or whatever. All this is exactly as it should be, and thesituation differs in no way from the situation with regard to descrip­tive statements in literal terms.6Before concentrating on the difficulties connected with originalmetaphors I want to discuss an interesting case which often corre­sponds to an intermediate situation between old and original. Here,too, there is conventionality, but it is not necessarily the conven­tionality of the metaphor itself. It is a certain conventionality in theinterpretation of the metaphor. I am referring to situations, wherethe metaphor must be construed with the aid of a salient structureof literal meanings.Any language, I take it, contains many clusters of words or phrasesdenoting certain systems of things or properties or events, e.g.colours, directions, movements, social groups, relatives, etc. The useof one element out of such a cluster in a metaphorical sense orcontext may presuppose the implicit metaphorical application of thecluster as a whole, or at least a relevant part of it. Think of themetaphorical use of words like "white", "forward", "high", "parent",where a simple cluster, viz. an opposition, may be the basis of themeaning (white vs black, forward vs backward, high vs low, parentvs child). In order to understand the metaphor one has to pick outthe relevant opposition and to give it an appropriate metaphoricalapplication. Such metaphors are often proportion metaphors, as

A LIBERAL APPROACH 71

described by Aristotle. They need not take the form of a simple two­fold opposition, though. The structure can be three-, four- ormanifold, but if it is a clear and salient one, one may characterizethe interpretation based on it as a kind of conventional interpreta­tion. The outcome is often clear enough (even when the metaphoris new, which it need not be, of course) and the ascription of a truthvalue is often no problem.Now, such "scheme-dependent" metaphors (if I may call them

such) are a very important type of metaphor, and "schematic inter­pretation" is an important kind of metaphorical interpretation.Sometimes the suggestion is even made that the scheme situation isthe typical situation of metaphorical language. In his book Languagesof Art (1968) Nelson Goodman argues as if this were the case;moreover, he applies this view to all major kinds of figurativelanguage.7 In her impressive recent book Metaphor: Its CognitiveForce and Linguistic Structure (1987), Eva Feder Kittay develops akindred account of the semantics of metaphor, which is more detailedas well as more general than Goodman's analysis. Her account isbased on the notion of a "semantic field". She argues that metaphor- typically or always - consists in the application of a field ofconcepts (the field to which the vehicle pertains) to a strange domain(the domain to which the topic pertains). To be sure, this not onlyreminds one of Goodman but also of Max Black; the idea of inter­action is minimalized, however, and the discussion suggests a decid­able (although possibly complex) procedure.8

Is this view as an overall account of metaphor correct? I doubtit. Sometimes it is not at all clear what the relevant field would be.If the word "prison" is used metaphorically, we may wonder ifthe field applied is a field of buildings or one of institutions; butoften, it seems, only the opposition between prison and non-prisonwill be relevant. This could also be described as the exclusive rele­vance of the salient features of a prison, according to MerrieBergmann's analysis in her essay "Metaphorical Assertions".9 Inother cases, several fields seem to be relevant at the same time, orthe field is so blurred or complicated that the idea does not explainvery much. No doubt, semantic fields are operative in very manymetaphors, but the idea that this notion makes the phenomenon ofmetaphor perspicuous depends on a narrow conception of a semanticfield. On the other hand, a realistic and broad conception does not

72 J. J. A. MOOIJ

help much to assess a metaphor. Eva Kittay's book itself shows verywell that clear-cut cases of schemes and fields shade into complexor ambiguous and, as a matter of fact, intractable ones.In fact, a reader may be at sea, that is to say, he or she may be in

doubt about the meaning of a given metaphorical utterance. Even ifconventional metaphors and (semi-)conventional metaphor schemesare important parts of metaphor tout court, and show that one shouldnot dismiss the truth problem too quickly as irrelevant or as wronglyconceived (i.e., as simply a matter of falsity, or of neither-truth-nor­falsity), we must also pay attention to more difficult, more abstrusecases. The use of really original, fresh, creative metaphors in anattempt to describe the world, or human experience in particular,gives rise to serious problems, problems which have to be consid­ered now. And these problems force me to raise the topic ofmetaphorical truth again.

II

Suppose a metaphor, apparently meant to describe an aspect of theworld, should need explication. The speaker (the person who intro­duced the metaphor) or the hearer then has recourse to one or anotherout of a number of strategies, all of which throw doubt on the truthof the metaphorical description.To begin with, the speaker may weaken the force of his or her

original metaphor by adding a formula like "so to speak" or "as itwere". This may enlarge the area of possible interpretation, while itsimultaneously impairs any previous claim to truth. It may be a signthat the original metaphor was not meant very seriously from a cog­nitive point of view. This "so to speak"-procedure is, I think, con­ventionally connected with the use of metaphor; it is a conventionalright that has no analogon in the domain of literal statements, whereit is overruled by conventional obligations. (It is possible, of course,to withdraw or change a literal statement in the course of a conver­sation or discussion, but this is a different matter.)I should like to add that the above is different from the view, held

by Dorothy Mack and others, that to use a metaphor in a declara­tive sentence is to change the kind of speech act in a radical way.In that view, when using a metaphor in a would-be statement, oneis held no longer to be performing the act of stating or asserting,but rather an act like suggesting or proposing or inviting; the act of

A LIBERAL APPROACH 73

"metaphoring", to be precise (the nearest thing would be "uneinvite").l0 This view is much stronger than the idea I have just for­mulated: the conventional right to weaken a truth claim is by nomeans equivalent to the absence of any truth claim to begin with.In this light, many metaphorical utterances can be statements,although many are not - depending, of course, on the circumstances(kind of situation, verbal context, intention, emphasis, etc.). One alsohas a right, I would maintain, to have one's metaphorical sentencestaken as true or false assertions, and not merely as interesting sug­gestions or whatever (even if the metaphor should be changed intoa quasi-simile). Finally, the audience has a right to assess their truthvalue.There are further difficulties, however. It should be acknowledged

that in discussions about the truth or falsity of a metaphorical state­ment there is a tendency to fall back on literal statements. If thereshould be, or could easily be, a misunderstanding as to the importand, therefore, the truth or falsity of a metaphor, a speaker or writermay be inclined to use literal expressions as a common ground orfixed frame of reference. Even such "metaphor-happy" philosophersas Quine and Nelson Goodman follow this procedure now and then.For instance, when Goodman says, metaphorically: "Metaphor, itseems, is a matter of teaching an old word new tricks", he adds the(almost) literal explanation: "of applying an old label in a newway")1 Remarkably enough, however, the switch is sometimesmerely from metaphor to metaphor. Somewhere in his essay on "TwoDogmas of Empiricism" Quine announces that he will clarify anearlier passage without making use of metaphor anymore. He thenswitches from one metaphor to another and to a third one: from"nearness" to "germaneness" to "recalcitrance".12Indeed, one important function of literal expressions is to facili­

tate mutual understanding, and we would be rather helpless withoutthem. This, too, can be prejudicial to an original truth claim: themetaphorical sentence may appear merely to be an introduction to,or summary of, a set of literal statements. But again, it need not beso, for in itself, this procedure (which one might call "metaphordeletion") does not diminish the ability of metaphorical sentencesto state the truth. On the contrary, it may, if necessary, furnish a basisfor this ability.Let me discuss yet another, a third phenomenon relevant to the

truth problem, a phenomenon also pertaining to the understanding

74 J.1. A. MOOIJ

of difficult, complex metaphors. In certain cases, a hearer or readermay doubt whether a metaphor has any meaning at all, whether it isnot simply a misuse or abuse of language. Think, for instance, ofvarious (would-be) metaphorical descriptions of life. Human life maybe called (and as a matter of fact has been called) a tragedy, acomedy, a book, a school, a journey, a madhouse, a masked ball. Inthese examples the import is not too difficult to catch, although asto further details it may strongly depend on the context, of course.Moreover, the hearer or reader can assess the truth of the descrip­tion. The above statements are vague indeed, and some of them arepompous and/or one-sided, but they don't seem to be moreintractable than literal characterizations of human life like: life isinteresting, or mysterious, or ridiculous. Such-like metaphorical orliteral statements may primarily appear as expressions of a mood,but they may also state a partial truth.However, what if life should be called a sandal, a swimming-pool,

a carrot, a pebble, a nail? What is the point? Certainly, one canconstrue a context in which any of these examples acquires ameaning. Somebody wrote a poem based on "life is a carrot",13 And"life is a nail in one's coffin" sounds like an intriguing paradox.But if such a context is absent? Say, if the existing context is anormal one? Are these sentences then senseless, or are they simplyfalse? We may wish to conclude that the putative metaphor is not agenuine metaphor at all, but a meaningless utterance.This possible conclusion suggests a certain tendency not to

acknowledge the existence of metaphorical statements that are mean­ingful but false. Due to this tendency, it might seem that the meaningof bold metaphors depends on their being true, at least partially so,or by way of a first, important approximation. Should they be false,then they collapse into meaninglessness, into nonsense. But this isnot as it should be if one is concerned with truth: first and foremostthere should be the meaning, or the truth-conditions, and only thenthe truth value.One might call this tendency the "no truth-no sense" mechanism.

How strong it really is, is difficult to say. The answer would needmuch empirical research. It has been identified by several writerson the use of metaphor;14 but neither Nelson Goodman nor Eva FederKittay or Robert J. Fogelin, in their recent books Metaphor (1987)and Figuratively Speaking (1988), seem to accord it much impor­tance, since they speak freely about false metaphors,15 Indeed, one

A LIBERAL APPROACH 75

may deny that, metaphorically speaking, life is a comedy, just as onemay deny that, literally speaking, life is nice, or futile, or ridicu­lous. Yet, up to a point, this "no truth-no sense" mechanism seemsto exist, especially with extremely strange or surprising utterances.We should not forget that something similar is at work in the caseof opaque literal discourse: the question "What can this mean?" isoften connected with the question "What reading would make thisstatement true?". There are various charity principles active inhuman communication. In the case of literal statements, this isperhaps mainly a matter of reference, and with metaphorical state­ments a matter of meaning - a difference that would sustain aspecific difficulty in ascribing truth to full-blown metaphorical state­ments.

III

It is doubtful, however, whether all this must lead to the conclusionthat metaphorical statements cannot be true (sensu stricto). I havealready made a number of qualifying remarks on the three phe­nomena discussed before: the "so to speak"-procedure, the proce­dure of "metaphor deletion", and the "no truth-no sense" mechanism.My qualifications add up to the view that, whereas in many cases ametaphorical putative statement is to be taken as a suggestion, a hint,in many other cases it should be taken as a real assertion - and Ihave several reasons for holding this view.To begin with, there is a continuity from dead to fresh metaphors.

Dead metaphors may be brought to life again, witness manyexamples in poetry (W. H. Auden, M. Nijhoff, and many other poets),as also in journalism. In The Observer of April 23 (the issue men­tioned earlier) there is an editorial on the greenhouse effect and aboutMrs. Thatcher's moves and initiatives with regard to it, concludingwith the sentence: "Governments that live in greenhouses should notthrow stones". On the other hand, fresh metaphors can be as trans­parent as dead ones, and hence equally suitable to express an asser­tion.Consider the following example, taken from Quine: ". .. our state­

ments about the external world face the tribunal of sense experi­ence not individually but only as a corporate body".I6 This may betrue or false, or something in between, but has that difficulty anythingto do with the metaphorical character of the statement? Another

76 1. 1. A. MOOIJ

example, taken from Goodman: "Truth, far from being a solemnand severe master, is a docile and obedient servant".!? Intended asa description of natural science, among other things, this could havebeen true but is, I take it, false; in science, truth plays a differentrole from the one put forward by Goodman. Anyhow, the sentencehas a truth value notwithstanding the metaphor.

"It is obvious that truth in general depends on both language andextralinguistic fact", Quine says; for "the statement 'Brutus killedCaesar' would be false if the world had been different in certainways, but it would also be false if the word 'killed' happened to havethe sense of 'begat"'.18 That is obvious indeed, but it is far fromobvious that J. L. Austin was right in asserting somewhat morespecifically: "A statement is said to be true when the historic stateof affairs to which it is correlated by the demonstrative conventions(the one to which it 'refers') is of a type with which the sentenceused in making it is correlated by the descriptive conventions"19 ­for why that strong stress on conventions? Of course, conventionsplay an indubitable role, but breaks of convention do so as well. Toquote Goodman once again: "Whereas falsity depends upon mis­assignment of a label, metaphorical truth depends upon reassign­ment".20 And elsewhere: "Metaphor requires attraction as well asresistance - indeed, an attraction that overcomes resistance" (i.e.,in order to be true).21Attraction that overcomes resistance, or adequate reassignment:

such expressions suggest that the important distinction with respectto the truth problem is not that between dead metaphors and freshones, but that between transparent statements and opaque ones. Thisforms the second reason for my point of view. Note that the latterdistinction cuts through the set of literal statements as well as throughthe set of metaphorical statements! Although metaphorical statementsare not determined by convention or guided by rules, they can betransparent and decidable no less than literal statement can be so (butneed not be so).Let me once more refer to Donald Davidson's position, which is

very remarkable indeed. In his essay on "Communication andConvention" he is extremely generous in granting us the capacityto interpret utterances without any knowledge of the relevant con­ventions. " . .. [A]s interpreters we are very good at arriving at acorrect interpretation of words we have not heard before, or of wordswe have not heard before with meanings a speaker is giving them."22

A LIBERAL APPROACH 77

In principle, that is, we can understand what a speaker is up to bydiscovering his intention. Knowledge of the conventions is a veryimportant practical crutch, "but a crutch which, under optimum con­ditions for communication, we can in the end throwaway, and couldin theory have done without from the start".23 For, "convention isnot a condition of language".24This would seem, among other things, to clear the way for non­

conventional metaphorical meaning and metaphorical truth.However, the application of the above view about the theoreticalirrelevance of conventions within human communication to metaphoris blocked. It is blocked by Davidson's resolute denial that "associ­ated with a metaphor is a cognitive content that its author wishes toconvey and that the interpreter must grasp if he is to get themessage",25 This, Davidson says, is false. And of course it is falseas a universal statement. But is it not true that many metaphors infact do have such a cognitive content? Apparently, the human imag­ination has sufficient scope (or can easily be trained to have suffi­cient scope) to allow for the reassignment of concepts (or terms orlabels, as Goodman wants to have it). Very often one can see thepoint of a metaphor, and assess its truth value accordingly. Certainly,a lot of socialization or even indoctrination may be involved in theability to understand (and assess) a particular metaphor, but it wouldseem that the ability as such is fundamental, given the fact that somany metaphors go unnoticed and are processed unconsciously.(How many philosophers realize that Goodman and even Quine aregreat masters of metaphor, comparable in this respect with Nietzscheand Sartre rather than with Carnap and Husserl?)I should like to go yet further - for even the assessment of rich

and complex, and therefore opaque, metaphors is not always impos­sible. They may be used to tell the truth - but now the question maybe raised whether this does not ask for too lenient a conception oftruth.This is bound up with my third reason for not tabooing the notion

of truth in the analysis of metaphor too quickly. The degree of readi­ness to allow statements a truth value (as distinguished from con­sidering them illuminating, misleading, or something of a similarnature) is partly a matter of choice and preference. To be sure, theconcept of truth can be applied very strictly, so that only literal state­ments together with statements that can easily be translated intoliteral ones can have a truth value. This boils down, I think, to the

78 J.1. A. MOOIJ

choice of a context-free language (or segment of language) in whichto describe what is the case and to formulate the relevant truth-con­ditions. It follows, then, that a sentence X is true if and only if theworld is in accordance with those truth-conditions, or, to use Tarski'spopular and paradigmatic example of everything that should beimplied by a complete semantic definition of "truth": "Snow iswhite" is true if and only if snow is white.But one might object that, ultimately, truth is a matter of usingand accepting a sentence as an adequate description of a state ofaffairs. Thus, "Snow is white" is true if and only if we are preparedto use and accept that sentence to describe a property that snow26 infact has. "Snow is black" is a misassignment; it is false, because weare not prepared to use and accept that sentence as a description ofsnow. But with "Snow is marble" we may begin to hesitate; perhapsin certain circumstances, it is a true metaphorical description. AndI have no doubt about the metaphor used by Novalis, and later onby Karl Popper and Carl Hempel: "Theories are nets". In manycontexts this is used and accepted as a correct description of the func­tioning of scientific theories)7Use and acceptance are matters of choice, convenience, anddecision. One cannot rule out certain kinds of decision beforehand.Certainly decisions may be extravagant and unpracticable, but onemay be more or less liberal without being less or more sensible atthe same time. There is no inverse proportionality between liberalityand discretion or intelligence. It seems to me that a liberal approachwith regard to the question of granting a possible truth value tometaphorical statements is very sensible indeed, and for two reasons.The first is the social point of view: it is socially desirable toaccept vivid metaphorical language without at the same time refusingin principle to consider the statements in question as true or false.Speakers and writers should be granted the right to use metaphorswithout losing the possibility of having their statements accepted astrue - or rejected as false.The second reason is the cognitive point of view, yielding thereverse side of the issue: it is desirable to be able to tell the truthwithout at the same time being prohibited from the use of fresh (new,bold, creative) metaphors. Metaphors can be very useful in statinginsights into the way the world is, or may be taken to be, and shouldtherefore not be banned from cognitive discourse aiming to statethe truth. Nor should the utterer be given, in advance, the right towithdraw or disavow all claims to truth, for metaphors are too useful

A LIBERAL APPROACH 79

and too dangerous for that. A defence of a metaphorical statementas true and a critical exposition of a metaphorical statement as falseshould not be made illegitimate from the start.

NOTES

I The most detailed description of the history of the study of metaphor still isStutterheim (1941).2 Cf. Davidson (1979).3 Cf. Moore (1982), Miall (1979) and Cooper (1986).4 Cf. Black (1979), especially pp. 40-41.5 Cf. Cohen (1979).6 See also Binkley (1981).7 See Goodman (1968), Ch. II, section 8.8 Cf. Kittay (1987), especially Chs. 6 and 7.9 Cf. Bergmann (1982).10 Cf. Mack (1975).II Goodman (1968), p. 69.12 Quine (1961), pp. 43-44.13 The author was Ellen Spolsky, a participant in the conference on Knowledge &Language.14 Cf. Elliott (1967) and Kipp (1973). For a critical reaction, see Mew (1971).15 See, e.g., Goodman (1968), pp. 70-79; Kittay (1987), p. 313; and Fogelin (1988),pp.87-92.16 Quine (1961), p. 41.17 Goodman (1978), p. 18.18 Quine (1961), p. 36.19 Austin (1970), p. 122.20 Goodman (1968), p. 70.21 ibid., pp. 69-70.22 Davidson (1984b), p. 277.23 ibid., p. 279.24 ibid., p. 280.25 Davidson (1979), p. 44.26 i.e. What we call "snow": I leave the literal use of that word intact.27 I want to maintain this claim, although Theo Kuipers argued with me that thequoted metaphorical statement is false; "languages are nets", according to him,would be acceptable. I should like to thank him for his comments, some of whichI have followed.

BIBLIOGRAPHY

Austin, John L.: 1970, ' Truth', in J. L. Austin, Philosophical Papers, secondedition,1. O. Urmson and G. 1. Warnock (eds.), Clarendon Press, Oxford, pp.117-133.

80 J. J. A. MOOH

Bergmann, Merrie: 1982, ' Metaphorical Assertions', Philosophical Review 91,229-245.

Binkley, Timothy: 1981, ' On the Truth and Probity of Metaphor', in M. Johnson(ed.), Philosophical Perspectives on Metaphor, University of Minnesota Press,Minneapolis, pp. 136-153.

Black, Max: 1979, ' More About Metaphor', in A. Ortony (ed.), Metaphor andThought, Cambridge University Press, Cambridge, pp. 19-43.

Cohen, Ted: 1979, ' Metaphor and the Cultivation ofIntimacy', in S. Sacks (ed.),On Metaphor, The University of Chicago Press, Chicago and London, pp. I-tO.

Cooper, David E.: 1986, Metaphor, Aristotelian Society Series, Vol. 5, BasilBlackwell, Oxford.

Davidson, Donald: 1979, ' What Metaphors Mean', in S. Sacks (ed.), On Metaphor,The University of Chicago Press, Chicago and London, pp. 29-45. [With somerevisions also in Davidson D. (l984b).]

Davidson, Donald: 1984a, ' Communication and Convention', in D. Davidson (ed.)Inquiries into Truth and Interpretation, Oxford University Press, Oxford andNew York.

Davidson, Donald: 1984b, Inquiries into Truth and Interpretation, OxfordUniversity Press, Oxford and New York.

Elliott, R K.: 1967, ' Poetry and Truth', Analysis 27, 77-85.Fogelin, Robert J.: 1988, Figuratively Speaking, Yale University Press, New Havenand London.

Goodman, Nelson: 1968, Languages of Art, The Bobbs-Merrill Company,Indianapolis, Indiana.

Goodman, Nelson: 1978, Ways of Worldmaking, The Harvester Press.Kipp, David: 1973, ' Metaphor, Truth and Mew on Elliott', The British Journal of

Aesthetics 13, 30-40.Kittay, Eva Feder: 1987, Metaphor. Its Cognitive Force and Linguistic Structure,Clarendon Press, Oxford.

Levin, Samuel R: 1977, The Semantics ofMetaphor, The Johns Hopkins UniversityPress, Baltimore and London.

Levin, Samuel R: 1988, Metaphoric Worlds. Conceptions of a Romantic Nature,Yale University Press, New Haven and London.

Mack, Dorothy: 1975, ' Metaphoring as Speech Act', Poetics 4, 221-256.Mew, Peter: 1971, ' Metaphor and Truth', The British Journal of Aesthetics 11,189-195.

Miall, David S.: 1979, ' Metaphor as a Thought-Process', Journal of Aestheticsand Art Criticism 38, 21-29.

Moore, F. C. T.: 1982, ' On Taking Metaphor Literally', in D. S. Miall (ed.),Metaphor: Problems and Perspectives, Harvester Press, Brighton, pp. 1-13.

Quine, Willard V. 0.: 1961, From a Logical Point ofView, second edition, revised,Harper & Row, New York and Evanston.

Sontag, Susan: 1978, Illness as Metaphor, Farrar, Straus & Giroux, New York.[1979, Allen Lane, London.]

Sontag, Susan: 1988, Aids and Its Metaphors, Farrar, Straus & Giroux, New York.Stutterheim, Comelis F. P.: 1941, Het begrip metaphoor. Een taalkundig en wijs­

gerig onderzoek. Paris, Amsterdam.

SAMUEL R. LEVIN

POETRY, KNOWLEDGE, AND METAPHOR

If for some theoretical purpose it is required that the categories ofmetaphor and knowledge be juxtaposed and the relation in whicheach stands to the other be examined, then it seems to me obviousthat the sense of one or the other of these two notions must bemodified. Moreover, depending on the general background againstwhich the juxtaposition is to be effected, it might be the one or theother of them whose sense it would be necessary to adjust. If, ashere, the question is to be approached against the background of lit­erature, more particularly poetry, then it seems clear that needing toundergo modification is not so much our understanding of metaphorbut our idea of what constitutes knowledge. The fact is that the veryidea of knowledge when raised in connection with poetry has aboutit an air of the adventitious. For indeed, it is not clear that the readingof a poem causes us to 'know' anything. And this alleged deficit orincapacity of poetry has obvious implications when it is the relationbetween metaphor and knowledge that is the question at issue.

It is not part of my purpose - or competence - to go into a detaileddiscussion of knowledge - to consider its nature, its grounds, its vari­eties. Essentially, knowledge is a function of beliefs that we haveabout the composition of states of affairs. These beliefs may bedevelopments from personal experience of our natural surroundingsor have arisen in consequence of our having read or heard statementswhich purport to describe contemporary or historical conditions.Thus we may believe that snow melts, and we may believe that Platowas a Greek. If the statements expressing these beliefs correctlydescribe the states of affairs to which they refer - if, in other words,the claims they make are true - then we say we have knowledge ofthose states of affairs. It is as a rule required also that the beliefs,being given that they are true, be justified; that is, the grounds under­lying them must be legitimate, there having intervened in theadoption of those beliefs nothing in the nature of illusion, halluci­nation, deductive error, reportorial bad faith, or any other conditionthat might compromise the belief's validity.On this account, for a sentence to function as a possible source

81

F. R. Ankersmit and J. J. A. Mooij (eds), Knowledge and Language,Volume Ill, Metaphor and Knowledge, 81-93.© 1992 Kluwer Academic Publishers.

82 SAMUEL R. LEVIN

of knowledge it must make a factual claim, the truth of which isthen to be determined. But as is generally acknowledged, the sen­tences or statements in a .poem make no such claim. If, moreover,our theory of language is such that it is speakers and not sentenceswhich are held to make such claims, the same conclusion applies.As Sir Philip Sidney said, 'The poet ... nothing affirmeth'. By thishe meant to refute the charge that poets are habitual liars. We canalso understand him, however, to be denying that poets make state­ments, i.e., that they go surety for the claims made by the sentencesoccurring in their poetry. From all this it appears that if the conceptof knowledge is taken in a sense at all approaching strictness thenany examination of the relation it bears to poetry, and the functionof metaphor in poetry, will not yield much in the way of fruitful con­sequences.But of course for our present purposes there is no necessity to take

the concept of knowledge strictly. We may recall, in the first place,that there are many varieties of knowledge, most of which relax insome measure the requirements imposed by the standard. One hearstalk of innate knowledge, personal knowledge, tacit knowledge, spir­itual knowledge, declarative and procedural knowledge, knowing thatand knowing how, certitude (as well as certainty), and other vari­eties as well. I myself have spoken of conceptual knowledge. Whatthe existence of all these varieties suggests is that what we can besaid to 'know' comprises a good deal more than the set of experi­ences and descriptions with which truth can be unequivocally cor­related. There is in fact an entire other universe of experience,produced both directly and through the medium of language, bywhich we are affected and from which we learn things about theworld and about ourselves. This is a universe of subjective responses- of impressions, affections, insights, intuitions, and different typesof emotion - a collection of responses whose issue is perhaps notknowledge but whose collective yield plays a role in our epistemo­logical economy whose significance is not the less important for thatreason.We hear a bird singing on a clear summer morning, and we are

impressed by its pluck and affected by the cheeriness of the tune thatit warbles. We suddenly feel sanguine about the day in prospect,imagine agreeable possibilities. Perhaps we think of Shelley's skylarkor Hopkins' windhover and the sentiments we associate with them.We read a short poem by Emily Dickinson:

POETRY, KNOWLEDGE, AND METAPHOR 83

To make a prairie it takes a clover and one beeAnd reveryThe revery alone will doIf bees are few.

We read these lines and feel that they express a deep insight aboutrelations between the real and the imagined world. We are set tothinking about that relation, about how nature works and how themind works. Our immediate reaction to these lines might be, 'Howtrue', a reaction which is intended, however, not as an assessmentof a truth claim, but as a mark of appreciation for the compactnessand power of the sentiment expressed. Responses of these kinds,although they may not impart to us anything by way of knowledge,comprise intimations that affect crucially our relation to objects,persons, and facts in the world. Nor should we conclude that thesignificance of such experiences is a mere matter of feeling andimagination, devoid of cognitive significance. It is true that theremay emerge from such experiences no definite cognitions. Typically,however, they engender what we might call inchoate cognitions,inconclusive efforts to subsume the feelings and affections arousedby the experience under some concept or idea.We thus find ourselves in that region of epistemological indeter­

minacy that Kant endeavored to systematize in The Critique ofJudgment. In that Critique, when Kant comes to discuss the facultyof mind that constitutes genius, he introduces the notion of aestheticidea, the ability to form which is the characteristic function of thegenial mind. By an aesthetic idea, Kant says, he means 'That re­presentation of the imagination which induces much thought, yetwithout the possibility of any thought whatever, i.e., concept, beingadequate to it, and which language consequently can never get quiteon level terms with or render completely intelligible'.I In anotherformulation he says 'the aesthetic idea is a representation of theimagination, annexed to a given concept, with which, in the freeemployment of the imagination, such a multiplicity of partial re­presentations are bound up, that no expression indicating a definiteconcept can be found for it ... '.2The concepts of which Kant in the above passages speaks and of

which it is here a question are of a characteristic nature: they arenot like the pure rational ideas of God, freedom, or immortality;neither are they empirical concepts or categories of the under­standing. They represent, rather, a mixed type, abstract notions of

84 SAMUEL R. LEVIN

which we may experience instances or aspects. By way of examples,Kant offers the concepts of death, envy, love, and fame.3 Thesenotions can be represented in terms of their logical or descriptivepredicates, and for certain purposes such representations are quitesufficient. To certain 'poetic' natures, however, these concepts maypresent themselves not in purely discursive terms but accompaniedby intuitions that derive from and recall aspects of their personalexperience. By definition, there will exist on these occasions an epis­temic space between such a concept and the attendant intuition(s).It is with a view to closing this space that aesthetic ideas areproduced. In the nature of the case the effort so to close it must fail.However, although the space cannot be closed, an attempt can bemade to fill it, and the aesthetic idea, as poetically elaborated, re­presents such an attempt.In this Kantian framework let us consider an actual example.Keats's poem 'The Nightingale' is a dramatization of Keats'sstruggle to transcend the dispiriting conditions of ordinary life andachieve in the contemplation of nature a state of existence carriedon at a pitch of sustained and intensive fulfillment. This is the kindof general concept that counts for Kant as serving to stimulate thegeneration of an aesthetic idea. The nightingale functions in thispoem as what Kant calls an aesthetic attribute, something whichstands in a symbolic relation to the general concept but is never­theless concrete and of which, consequently, one can have an intu­ition. Onto and around this concrete object we can then aggregateother intuitive and imaginative 'particles', and in this way buildup an aesthetic idea. The poem, or parts of it, then represents whatI earlier referred to as the filling in of epistemic space, a space whosetermini in the present instance are, on the one hand, the generalconcept of nature's capacity to induce a state of intense imaginativesatisfaction and, on the other, the intuition of the nightingale. I willnow quote some lines from stanza 7 of the poem, these linesto be understood as representing a portion of the 'filling' process:

Thou wast not born for death, immortal Bird!No hungry generations tread thee down;

The voice I hear this passing night was heardIn ancient days by emperor and clown:

Perhaps the self-same song that found a pathThrough the sad heart of Ruth, when, sick for home,

She stood in tears amid the alien com;

POETRY, KNOWLEDGE, AND METAPHOR 85

If now we should ask what kind of knowledge is conveyed to us bythese lines, we would be thrown back onto our earlier conclusionsabout the compromised nature of the statements made in a poem.Keats in these lines does not convince us, i.e., make us hold fortrue, that one can find in nature escape and relief from the caresthat beset our ordinary lives; he does, however, present intimationsto that effect, intimations which the poem, as a crucible of artisticcreation, has concentrated, refined, and crystallized. Serving toconvey these intimations are not any empirical facts or logical argu­ments, but subjective impressions and personal associations. Thelines I have read are part of a perhaps unconscious effort to persuadethe reader that these impressions and associations have a universalvalidity. The appeal made by these lines, and by the poem in general,is thus not to our reason, but to our sensibilities. To that end Keatshas introduced and adapted the Biblical story of Ruth; on readingthese lines we empathize in the sadness of Ruth, we feel the loneli­ness, the sense of loss, of longing for a forsaken homeland; weperhaps indulge the fancy that the nightingale's song would lift thespirits of Ruth; perhaps find in them also a presage of Ruth's latermarriage and happiness. Responses like these, which are inducedby the poem, do not constitute knowledge in the strict sense of theword, but they serve to deepen our understanding of the human con­dition, and they represent for that reason a not unmeaningful com­ponent in the epistemological scheme of things.

Ifwe turn our attention away from the general relationship holdingbetween knowledge and poetry and consider now the relation thatobtains specifically between knowledge and metaphor, our criticalvantage point is not much improved: we find ourselves still con­fronting a relation of highly problematic character. In this case,however, it happens that there has been a good deal of discussionwhere precisely the question of the epistemic value of metaphor hasbeen the subject under consideration. Part of this discussion has dealtwith the role played by metaphor in language generally, part withthe role that it plays in poetry. At the center of all the discussion,however, has been a concern with the way metaphor figures in theincorporation and communication of knowledge.

It has been argued by a number of important commentatorsthat language is at bottom, and unavoidably, metaphoric. In its mostextreme form this argument has been made by Nietzsche. ForNietzsche the metaphoric character of language arises from the very

86 SAMUEL R. LEVIN

fact that human beings are not content simply to experience life,but that they have a need also to describe that experience; that theylive not merely as individuals, but also as members of social groups- in other words, that they need to develop and use a language.Given that for Nietzsche reality is in a constant state of becom­

ingness or flux, as the will to power unceasingly exerts itself in itsvarious manifestations, every individual experience of that realitywill be unique and momentary. It is the inescapable commitment oflanguage, however, to generalize over those experiences. A word like'leaf' , for example, expresses a concept; in that it stands for a classof referents, and in this way levels out the differences amongindividual leaves, it is for Nietzsche thereby a metaphor. Of course,given Nietzsche's view of the nature of reality, or I should say, ofour experience of reality, it is hard to see how language could beotherwise than nonrepresentative. Reality for Nietzsche is flux;language fixes that flux and is for that reason and to that extentmetaphoric.

If we ask what relation there can be between metaphor as con­ceived by Nietzsche and knowledge, the answer is not very positive.Since, according to Nietzsche, our language is congenitally and per­vasively metaphoric, we necessarily in using it misrepresent reality;and statements we might make, therefore, cannot be true and cannot,consequently, represent knowledge.On this approach the relation between metaphor and knowledge

is problematized originarily. If knowledge is assumed to be expressedin language, and if language is necessarily metaphoric, then the veryquestion of knowledge is rendered moot. We may believe that wehave knowledge of certain truths expressed in language, but alreadyin the transition between the facts and the expression of those facts,a 'metaphoric' turn has supervened - so that the very foundationson the basis of which we assume that we know something are under­mined at the outset. On this general approach to the problem,moreover, where metaphor is presumed to be integral to languageitself, metaphor in poetry would be a kind of second-order metaphor,and we might expect that its relation to knowledge would be evenmore attenuated; indeed, the question itself would now appear in amuch more problematic light.As it happens, Nietzsche himself offers us a way out of the

impasse. In the same essay in which he establishes to his ownsatisfaction the fundamentally metaphoric nature of language,

POETRY, KNOWLEDGE, AND METAPHOR 87

Nietzsche indicates that it is in myth, or art in general, that thelanguage, metaphoric though it may be, may yet be used metaphor­ically. We have to bear in mind that on the analysis provided byNietzsche metaphor is not the rich, complex, expressive linguisticdevice that it is held to be in literary studies. Although absolutelynecessary in order that norms of individual, social, and intellectualbehavior be established, metaphoric language, inasmuch as itexpresses general concepts and not the personal intuitions of indi­vidual experience, allows for only an impersonal, utilitarian type ofcommunication. Paradoxically, the only way to break out of this insti­tuted conformity and express something fresh and original is by theuse of metaphor. Nietzsche writes:

The drive toward the formation of metaphors is the fundamental human drive, whichone cannot for a single instant dispense with in thought, for one would therebydispense with man himself. This drive is not truly vanquished and scarcely subduedby the fact that a regular and rigid new world is constructed as its prison from itsown ephemeral products, the concepts. It seeks a new realm and another channelfor its activity, and it finds this in myth and in art generally. This drive continu­ally confuses the conceptual categories and cells by bringing forward new trans­ferences, metaphors, and metonymies. It continually manifests an ardent desire torefashion the world which presents itself to waking man, so that it will be ascolorful, irregular, lacking in results and coherence, charming, and eternally newas the world of dreams.4

The way to be artistically metaphoric, Nietzsche is saying, is tosubvert the language. One does this by bringing forward new trans­ferences - metaphors and metonymies - and in this way confusing,i.e., transgressing, the conceptual categories. By way of illustratingsuch transgressions Nietzsche instances the Greek myths, in whichtrees speak as nymphs, gods turn themselves into animals, andvarious other 'miraculous' events are described and which the Greeksbelieved.On the performance of one of these subversive manoeuvres, the

linguistic consequence is a sentence that is semantically deviant, thusa sentence that displays a characteristic property of poetic metaphor.Shakespeare begins one of his sonnets (33) with the lines:

Full many a glorious morning have I seenFlatter the mountain tops with sovereign eye,Kissing with golden face the meadows green,Gilding the pale streams with heavenly alchemy.

88 SAMUEL R. LEVIN

These lines attribute a number of unusual capacities to morning: itflatters mountain tops, it has a (sovereign) eye, it has a golden face,it kisses the meadows, and it gilds pale streams with heavenlyalchemy. In the face of these untoward attributions, these sort-cross­ings, we might try to achieve an interpretation of these lines by com­paring what they say to characteristics of morning as we know it.In accommodating the burden of these lines to our personal experi­ence' we might recall mornings in which we found ourselves in asetting of natural surroundings and how, in that setting, the sun'slight seemed to have about it a restorative quality, how its raysseemed to diffuse health and well-being through the entire sur­roundings; we might recall how the sunlight, playing on the moun­tains, fields, and streams, aroused in us a sense of existentialcontentment, perhaps quickened in us feelings of spiritual affinity.Whatever might be the merits of the preceding description,

however it might bear on the interpretation that we might finallyassign to the lines and to the sonnet, one thing seems clear: if wewere here to speak of knowledge in connection with metaphor, itwould have to be of the same compromised and (to speak theoreti­cally) degenerate sort as was the knowledge we spoke of when thesubject under discussion was poetry tout court; - knowledge in thecontext of metaphor is again an amalgam of feelings, impressions,intuitions, and assorted subjective responses - nothing, in short, likejustified true belief.This sameness of result ensued, however, because in our discus­sion of both poetry and metaphor we followed essentially the sameprocedure: in our respective analyses we attempted to accommo­date meanings to the world as we know it. In the case of the linesfrom Keats, the accommodation was comparatively straightforward;ignoring, as we did, a few metaphoric turns, we used the acceptedmeanings of the words to arrive at an interpretation; put another way,the language in Keats's lines made sense in the world as we knowit. And although when we dealt with the lines from Shakespeare, ourfocus was on the metaphoric expressions, we in effect followed thesame procedure; in reporting a set of possible responses, I as it weredisregarded what was literally asserted by the lines and disposedinstead of modified, nonmetaphoric meanings: the lines, after all, saythat the morning flattered the mountain tops, that it is fitted with aneye, that it bestows kisses, and so on - it being these predications

POETRY, KNOWLEDGE, AND METAPHOR 89

that confer on the lines the semantic deviance that I referred toearlier. In my discussion of these lines, however, I implicitly neu­tralized this linguistic property when I described a reaction that wasentirely consistent with everyday experience. This strategy of natu­ralizing the meanings of metaphoric expressions, i.e., construingthem so as to make claims that are congruent with the nature of theworld, is more or less standard. Theories of metaphor embodyingwidely different substantive and methodological assumptions agreein reducing the frequently fantastic and counterfactual claims madeby metaphoric expressions to conditions in the actual world.At this point I would like to go back to the remarks of Nietzsche

that I invoked earlier. In those remarks Nietzsche spoke of the irre­pressible human drive to transgress conceptual categories by bringingforward new transferences, new metaphors. We have found preciselysuch transgression and transferences in the lines from Shakespeare.But the results, to judge from my description, would seem by nomeans to have the overriding significance that Nietzsche assigns toartistic metaphor. The mere crossing of predicates does not, we see,suffice to, as Nietzsche puts it, 'refashion the world'. To accom­plish that end, it would appear, it is necessary not merely to trans­gress conceptual categories, it is necessary also to take seriouslythe consequences of those transgressions. It is a commitment of thistype - a commitment which entails the entertaining of strange andother-worldy ontological designs - that the argument of Nietzschereally calls for. What this means in practice is that (under certain cir­cumstances) when we encounter a semantically deviant expressionin a poem, our interpretation of that expression should not take theform of asking (implicitly) what can it mean, but of asking, rather,where could that happen. Put another way, the strategy takes the formnot of displacing the meaning, but of displacing the context. Facedwith a deviant expression, we do not proceed to contrive for it ameaning which would consist with conditions in the actual world;we proceed, rather, to conceive of a world in which the expression,taken at face value, would describe a possible state of affairs.sBefore discussing how this strategy may be used in the treatment

of semantically deviant metaphors, I will discuss its relevance andapplicability to a class of metaphors whose characteristic and sig­nificant feature is precisely the fact that there is in their linguisticcomposition no semantic deviance. Thus, coming in for considera­tion in theoretical discussions of metaphor there is a class that might

90 SAMUEL R. LEVIN

be referred to as 'equivocal' metaphors. Metaphors of this type areusually introduced precisely in order to show that semantic devianceis not a necessary condition for metaphor. Searle's example illus­trating this type is Disraeli's remark on being appointed PrimeMinister, 'I have climbed to the top of the greasy pole' ,6 MichaelReddy offers, 'The rock is becoming brittle with age' ,7 and Mooijhas instanced the sentence, 'There is a dog hidden in old newspa­pers'.8 None of these examples incorporates semantic deviance; allmake perfect sense taken literally. However, they do not make literalsense in the context of their use. When Disraeli made his remark,l.~lere was no greasy pole in the vicinity, nor had he just performedan act of climbing; in Reddy's example, a professor and not a rockis the reference of the sentence; Mooij's example was used in acampaign to collect and sell old newspapers, the proceeds to be usedto buy guide-dogs. In order for these, and similar expressions, tofunction metaphorically, a displacement of the general type describedabove has to be effected: one has to imagine a context which com­prises features such that for the expression to make sense in thatcontext it must be construed metaphorically. In order to make senseof Disraeli's remark, for example, one must invoke as context thegeneral background of his election as Prime Minister and the yearsof political struggle that preceded that election. Given the contextsin which these sentences are actually produced, contexts in whichtheir use makes no sense, equivocal metaphors may be said toinvolve what we might call pragmatic deviance. And in summoningup the conditions in the context of which the expression makesmetaphoric sense, we in a manner of speaking do refashion the world.When this strategy of contextual displacement is applied to

metaphors expressed in semantically deviant language the conse­quences are rather more complex, the refashioning considerably moredrastic. Applied to the lines from Shakespeare, for example, the pro­jected context is not a simple variant of the actual world, a state ofaffairs that we can be said to have experienced and thus to know.The crossing of predicates, if their semantic consequences are takenat face value, projects a state of affairs that we cannot possibly haveknowledge of. We have never seen (or heard) the morning flatterthe mountain tops, nor have we ever observed it to have an eye. Whatepistemological purpose then can be served by adopting the con­textual displacement strategy in such cases?

In speaking of knowledge, I think it is important to distinguish

POETRY, KNOWLEDGE, AND METAPHOR 91

between what we know and what we come to know, that is, betweenold and new knowledge. There is an important sense, I believe, inwhich the acquisition of a fresh piece of knowledge is more signif­icant, weighs more in the epistemic scales, than do fifty or a hundredregistrations of previously acquired knowledge. In fact, whereas anewly-acquired piece of knowledge, if it is to qualify as such, mustobviously undergo registration (as knowledge), we may questionwhether, when we merely reenact an experience, the fact of ourknowing something as a consequence of that experience is even(re)registered. When I go out on a summer's day and feel the warmthof the sun, this is a reenactment of an experience on the basis ofwhich I at one time learned, and now know, that the sun projectswarmth. It does not seem to me that when I reenact this experienceanything very significant in the way of knowledge is entrained. Onthe other hand, were I now to learn how long the sun's rays take toreach the earth, cognitive registration of the fact would be a signif­icant epistemic event. Obviously, in drawing this distinction, timingis of the essence; what is new knowledge today is old knowledgetomorrow. What I wish to stress, however, is that newness or novelty,as a property of knowledge, can playa significant role in determiningthe value that such knowledge may have for us. In fact, it is mycontention that the novelty of the experience, even if that experienceis not one of pure cognition and thus does not eventuate in knowl­edge, may produce significant intellectual consequences just becauseit is novel.Where poetic metaphor is concerned, displacement of the context

means that we try somehow to conceive of the states of affairs thatthe metaphor, taken literally, describes. If the metaphor says thatthe meadows are laughing, that the sky is angry, or that the morningflatters the mountain tops, we try to conceive of just those states ofaffairs. It seems clear that if we take literally such metaphors andsucceed in bringing to mind an impression of the states of affairsdescribed by them, that the mental content of that effort will benovel. I am assuming, of course, that we have come across themetaphors for the first time. The question then is what is the natureof these mental exercises - these conceptions, as I will call them ­and what relation do they bear to knowledge.We have come down now to a contrast between cognitions and

conceptions - cognitions being mental functions supported bymatters of fact, conceptions being mental projections from poetic

92 SAMUEL R. LEVIN

metaphors. It is in the nature of the case we are considering that theconceptions are novel, and I wish to claim a certain significance inthe epistemological scheme of things for such conceptions. If thephysical conditions that obtain in such conceptions are such as topreclude our having knowledge of them, nothing prevents us fromlending credence to the possibility of their occurrence. This attitude- one of conceptual tolerance - is in fact the one prescribed forapproaching poetry in general. It is the attitude expressed inColeridge's well-known phrase, 'the willing suspension of disbelief',and it is commensurate also with Keats's notion of negative capa­bility, according to which the poetic temperament should be 'capableof being in uncertainties, mysteries, doubts, without any irritablereaching after fact and reason'. My particular suggestion, i.e., tocountenance the possibility that ontologically bizarre or untowardconditions might somehow obtain, perhaps puts the principle of con­ceptual tolerance to the test, but I do not think that it exceeds theprinciple's limits.With the countenancing of such a possibility, a previously

unthought of conceptual relation has been entertained. If, in con­sidering Shakespeare's lines we take the words at their face value,then the conception induced by the lines is one that is completelyalien to our accustomed mode of thinking. It has never previouslyoccurred to us that the morning might engage in flattery, for example.In conceiving of such a thing we do not construe the meaning ofthe words in Shakespeare's metaphor; instead, we construe, i.e.,rethink, what we know about the world. In this process our customarymode of thinking is reorganized so as to allow for the possibility thatan insentient being may indulge in a conscious activity. Seen in thisway, metaphor is not a different way of saying something; it is asaying of something different. It is an imperfect but pregnant wayof saying something which, because it has never before been thought,cannot be expressed by combining words in any of their accustomedarrangements. These are novel thoughts; they lie at critical removesfrom ordinary thought - in metaphoric space. And though they maynot constitute knowledge, in that there correlates with them no facts,they assume a significance and importance in our conceptualeconomy simply on the basis of what they cause us to think about.To sum up the results of this discussion: the conclusion appears

to be that neither from poetry nor from poetic metaphor can we besaid to derive knowledge. At the same time, no one would deny that

POETRY, KNOWLEDGE, AND METAPHOR 93

our understanding of life can be deepened by poetry, and our con­ception of the world be enlarged by metaphor.

NOTES

1 Kant (1952), pp. 175-176.2 Kant (1952), p. 179.3 Kant (1952), p. 176.4 Nietzsche (1979), pp. 88-89.5 For extensive discussion of this viewpoint, see Levin (1988).6 Searle (1979), p. 97.7 Reddy (1969), p. 242.8 Mooij (1976), p. 27.

BIBLIOGRAPHY

Kant, Immanuel: 1952, The Critique of Judgment, J. C. Meredith (transl.), TheClarendon Press, Oxford.

Levin, Samuel R.: 1988, Metaphoric Worlds: Conceptions of a Romantic Nature,Yale University Press, New Haven and London.

Mooij, 1. J. A.: 1976, A Study of Metaphor, North Holland, Amsterdam.Nietzsche, Friedrich: 1979, 'On Truth and Lies in a Non-Moral Sense', inD. Breazeale (ed. and tr.), Philosophy and Truth: Selections from Nietzsche'sNotebooks of the Early 1870's, Humanities Press, Atlantic Highlands, NewJersey.

Reddy, Michael J.: 1969, 'A Semantic Approach to Metaphor', in Papers fromthe Fifth Regional Meeting of the Chicago Linguistic Society, Department ofLinguistics, University of Chicago, Chicago, pp. 240-251.

Searle, John R.: 1979, 'Metaphor', in A. Ortony (ed.), Metaphor and Thought,Cambridge University Press, Cambridge, pp. 92-123

KUNO LORENZ

ON THE WAY TO CONCEPTUAL AND

PERCEPTUAL KNOWLEDGE

In his introduction to Ways of Worldmaking Nelson Goodman hasgiven a characterization of some important changes which philos­ophy underwent during history in order to assign a proper place tohis own theory of symbols. It belongs, he states, to "that mainstreamof modern philosophy that began when Kant exchanged the struc­ture of the world for the structure of the mind, continued when C.I. Lewis exchanged the structure of the mind for the structure ofconcepts, and that now proceeds to exchange the structure ofconcepts for the structure of the several symbol systems of thesciences, philosophy, the arts, perception, and everyday discourse."And indeed, in the light of this characterization it is easy to see

that pragmatics has become the modern heir of ontology with semi­otics being its counterpart as the heir of epistemology. Of course,both disciplines have to be understood in the sense of C. S. Peirce,that is not as just two newly established empirical sciences, but asways of investigation where empirical procedures are united withphilosophical or reflexive procedures.Using this broader perspective both actions and sign-actions are

not only treated as the objects of research and representation but alsoas a means of research and representation. You not only observeand describe these entities according to certain standards but youalso produce them in a perspicuous fashion in order to arrive at somekind of approximating reconstruction of what you take to be avail­able, already.Wittgenstein, as is well known, has used the term 'language-game'

for this kind of activity which aims at disclosure of what is goingon by providing tools of comparison. Hence, the productions servecognitive purposes in the sense of delineating the very areas ofobjects you proceed afterwards to investigate in the more usual way.A language-game may count as a paradigm case of perceptual knowl­edge, insofar as its significative function works by being an icon inthe sense of Peirce. You have found an area of internally structuredobjects by inventing a prototype.

95

F. R. Ankersmit and J. J. A. Mooij (eds), Knowledge and Language,Volume III, Metaphor and Knowledge, 95-109.© 1993 Kluwer Academic Publishers.

96 KUNO LORENZ

And it becomes obvious that even the distinction of action andsign-action, a special case of the basic and embarrassing distinctionbetween world and language, has to be relativized in view of a purelyfunctional account of both what it means to be an object and whatit means to be a sign (of an object). In fact, it belongs to one of thebasic tenets of Goodman's approach that the seemingly clear-cutdivision of world and language - nonverbal language of courseincluded - as a division between the given and the constructed,between that which is found and that which is made, between thefact and the artefact, is outdated, and you may convince yourself thatit has even been challenged once and again since the times of thePre-Socratics. But only rarely is history looked at in this way. Anymatter we are concerned with, Goodman tells us, is dependent onsome manner as the means by which we deal with it. So worlds arebut versions and worldmaking begins with one version and ends withanother. The message we should learn, and which you are surelyaware of, runs thus: "never mind mind, essence is not essential, andmatter doesn't matter".He goes on in claiming that we choose the facts as much as the

framework, though this statement should better be split into two com­plementary statements: we produce the facts as much as the frame­work and we experience the framework as much as the facts.Productions, when serving cognitive purposes, are always repro­ductions.,'A'YEtv and 1t<xcrXEtv, the last two Aristotelian categories which

seemed forgotten throughout most of modern philosophy in theDescartes tradition will enjoy a lively comeback as the two wayswe are concerned with when doing something: you do it yourself(active) and you call others to do it (passive), where the latter versionmay also be framed 'you >see<l others doing the same'.I will return to these two ways a bit later under the labels: per­

formance and recognition (of actions). Now, no doubt, we are ableto live with, or in, different versions: no difficulty for anyone of usto say both 'the sun rises in the morning and sets in the evening' and'the earth rotates around its axis'. The problem which arises is notone of selecting the right version and identifying it by its being lit­erally true whereas the wrong version is at most metaphorically true.We should rather ask two more general questions.(a) If there is - by way of comparison with the one world it isclaimed we all live in - no chance to characterize one version

CONCEPTUAL AND PERCEPTUAL KNOWLEDGE 97

uniquely as the true one (apart from the fact that only sententialversions can be true, others like pictorial versions cannot be sub­jected to a truth claim), what can be the criteria to distinguishright versions from wrong ones?

(b) If there exist different, even conflicting, right versions, how isit possible to live as >the same person< in different worlds or,to have communication between persons living in differentworlds?The issues connected with (a) which essentially boil down to the

issue how to identify the objects (colors, feelings, lines, perceptualactions, things, experiments, etc.) we succeed and fail constructing,are dealt with extensively in Goodman's Ways ofWorldmaking whererightness of rendering is basically declared to be matter of fit to whatis referred to, and this is nothing but ultimate acceptability.The issues connected with (b) which in turn are focussed on the

issue how to secure the identity of the subjects living in possibly dif­ferent worlds have not yet been considered, because only the signi­ficative ('referential' is Goodman's term) and not the communicativefunction of sign-actions - my term for the elements in the processof worldmaking - is dealt with by Goodman. Hence, an explicit con­sideration of their interdependence is left out.To arrive at some clarification in this respect we have to go back

to Peirce. Some of his ideas I will deal with in my last, the third,section. For the moment it is enough to feel convinced of Goodman'ssuccess to have shown that the reorganizations of world views andthe constructions used to achieve them make the Cartesian separa­tion of ontology and epistemology obsolete - a move put forwardby Peirce when he characterizes the methodology of pragmatism byjust this claim: ontology and epistemology are but two sides of thesame coin.2

Furthermore, one should be prepared to take seriously the fact thatthere are many engaged in very different activities through times andspaces with obviously limited capacities to understand each other.To tie acquisition of knowledge basically to individual persons andjustification of knowledge basically to social communities exercizingcontrol over the individual idiosyncrasies would itself be dependenton the Cartesian topos of individual persons starting to digest thesensory input each for him- or herself and only afterwards con­fronting each other with possibly different experiences which, then,have to be >unified<.

98 KUNO LORENZ

But which criteria are available, as socially one doesn't have theexperience but only their linguistic - not necessarily verbal ­representations? The two moves open, and nowadays adopted bydifferent parties, to conventionalism, being an instrumentalist versionon the level of signs on the one side, and to ontologism which is alsocalled metaphysical realism, being a realist version on the level ofobjects assumed to be the same for all individual persons on the otherside, actually lead nowhere, since they have to beg the question.The only way out, I think, is Goodman's move to relativize thedistinction between objects and signs of objects. This leads quitenaturally to the construction of another version out of every daydiscourse using the concept of sign-action in such a way that we startby looking at verbal language, or other symbol systems in use, astypes of actions like eating and sleeping, and by looking at non­linguistic, especially non-symbolic, objects as parts in a web of inter­related and interdependent actions.Hence, what we do initially is to naturalize language includingother symbol systems, and to symbolize the world by paying atten­tion to that feature of actions which is underlying Goodman's treat­ment of exemplification as a tool to tie actions together with symbols,and which Wittgenstein has achieved by introducing language-games.Actions are both performed (the natural side) and understood (thesymbolic side) which in the first case results in being able to producetokens of a type, and which in the second case results in being ableto identify different tokens as belonging to the same type. Here, theequivocation of 'identification' is intentional: I mean both to rec­ognize two tokens as belonging to the same type (epistemologicalidentification: the type is recognized or determined) and to set twotokens equal, since they are tokens of the same type (ontologicalidentification: the type is constituted by setting all tokens equal toone another). Both readings are type-oriented, hence you lookthrough them at actions >universally<. On the other hand, the pro­duction of tokens as a kind of practical attempt to exhaust an action­type is token-oriented, you look at actions >singularly<, thus makingit possible to treat the tokens as parts of a whole in the making.

It should be added that the move to naturalize symbol systems ona par with symbolizing the world was not meant as a move to makethe steps of theory, Le., the steps of devising world versions,amenable to moral philosophy. Goodman stated quite concisely: "myargument that the arts must be taken no less seriously than the

CONCEPTUAL AND PERCEPTUAL KNOWLEDGE 99

sciences is not that the arts >enrich< us or contribute somethingwarmer and more human, but that the sciences as distinguished fromtechnology, and the arts as distinguished from fun, have as theircommon function the advancement of understanding".3Sign-actions may be the concern of many and shall both in their

natural or pragmatic aspect (performance) and in their symbolic orsemiotic aspect (recognition or identification) be viewed at withrespect to enlarging or refining men's abilities and not with respectto serving men's needs, though, of course, this can be done, too. Inthis context, the abilities on the semiotic side, only, we call under­standing, where beside conceptual abilities also perceptual abilitieswhich in turn are served, e.g. in painting, by skills with or withoutexternal tools, should be included.In my second section I want to supplement the ideas of Goodman

- the few I have presented - with ideas of his almost one genera­tion older contemporary Hermann Broch in order to shed more lighton the distinction of perceptual and conceptual knowledge, that is,of knowledge in the arts and in the sciences, we are looking for, byclosing the gap between objects and signs (of objects).I want to show that the philosopher-artist Broch and the philoso­

pher-scientist Goodman proceed in a peculiarly dual manner. Theyboth make claims concerning the epistemological relation of (cog­nizing) subject and (cognized) object which on the part of Broch isa conceptual one: There is one world underlying the various semioticrepresentations. On the part of Goodman we have been aware of aperceptual claim: There are many ways of worldmaking by the sameagents.Now, the difference between conceptual and perceptual knowl­

edge is characterized by Broch with the following words: "es ist dieAufgabe der wissenschaftlichen Erkenntnis [... ], zur TotaliUit derWelt in unendlich vielen, unendlich kleinen rationalen Schrittenvorzudringen (it is the task of scientific knowledge [... ] to reachthe totality of the world by means of infinitely many and infinitelysmall rational steps)",4 and further: "[in einem einzigen] Kunstwerk[Hisst ein Akt der Erkenntnis] die TotalWit der Welt erstehen ([in asingle] work of art [an act of cognition] gives rise to the totality ofthe world)".5In both cases we operate on the level of signs: Scientific proce­

dure is a process of gradually increasing the internal differentiationof a representation of the world until it approximates reality. Broch

100 KUNO LORENZ

uses the terminology of contemporary philosophy of science as hehad learned it from the logical empiricists in Vienna and speaks of>setting up a rational mode1<6 in the effort to approach a >map oftotality<.7 Artistic procedure, on the other hand, is with each suc­cessful instance a repetition of creation by means of multifariouslyarticulated symbols, a symbolic cosmogony eventually producing a>symbol of the world<.8He states: "das Wertziel des Dichterischen, die kosmische

Unendlichkeit, erftillt sich in der einzigen Realitatsvokabel eineslyrischen Gedichtes (the value aim of poetization, cosmic infinity,is satisfied by the single expression of reality of a lyric poem)".9Here, the term 'Realitatsvokabel' ('single expression of reality') isused for a linguistic expression which, as material in a literaryartefact, thereby generates a situation. For contemporary readers,Wittgenstein's quite similar idea in his Philosophical Investigations(part II, section XI) comes into mind, where a rabbit-situation isgenerated by the exclamation 'A rabbit'. I quote part of the relevantpassage:

I look at an animal and am asked: "What do you see?" I answer: "A rabbit." - Isee a landscape, suddenly a rabbit runs past. I exclaim "A rabbit!"Both things, both the report and the exclamation, are expressions of perception andof visual experience. But the exclamation is so in a different sense from the report[... ]It is related to the experience as a cry is to pain.

In the report case I proceed from something perceptually known tosomething known conceptually - a case of knowledge by descrip­tion. In the exclamation case something becomes known perceptu­ally, using language as a means of perception, not of conception ­a case of knowledge by acquaintance.If we call the elements of processes leading to perceptual knowl­

edge 'poetic actions' or poiesis, the concern of poiesis is nothing buta determination of what can become a sign and how: the result, some­thing invented, is, if successful, mimesis, something found. I amgoing to give a more elaborate explication a bit later. And philo­sophy as understood by Broch (he calls it 'value-theory') is just semi­otics combined with pragmatics in the Peircean sense, though in themain still based on abilities of single grown-up individuals and notyet on their gradual development by dealing with each other.Now, research and representation of value-positing activity, or phi-

CONCEPTUAL AND PERCEPTUAL KNOWLEDGE 101

losophy, will appear in two complementary garbs both of which arecharacterized by a restauration of the unity of object and method inthe sense of dealing with the object. Broch shows through analysisof many examples in history that separation of object and method(of treating it) yields a growing appearance of mutual independenceamong primordially interrelated ways of (scientific and artistic)activity. The sciences become >positive< fields which determinethemselves >objectively< - even treating their respective methodsas something specific for them, their >meta-objects<. They, thereby,gradually lose their ability for self-reflection. The arts, on the otherhand, turn into disciplines which define themselves by theirtechniques - even reducing their respective objects to technicalproblems concerning the specific materials - and thus end atactivities of mere >l'art pour l'art< which are unable to symbolize>totality<.The first, conceptual, garb of philosophy is epistemology insep­

arably tied to philosophical critique, the second, perceptual, garb ofphilosophy is philosophische Dichtung (philosophical poetry), thatis, an art-work which is simultaneously an instance of art criticism.I0

Broch confines his term to literary art works, though he discussesextensively, e.g. its equivalents in music.Conceptual knowledge is bound to an >ideal< language - Broch

calls it 'platonic' and at the same time the language of God, it iswithout >style<,ll whereas perceptual knowledge is gained by theuse of a style, that are those properties of an art-work - in the literarycase an art-work in the verbal medium - which determine its beingan articulated symbol. To uphold the unity of object and method, itis in both cases essential to realize that knowledge is concerned withthe relation of >form< or >structure<, and >content<. Hence, neitherthe move of the autonomously acting sciences to treat problems ofmethod as problems of higher order content, nor the move of thelikewise self-sufficient arts to turn problems of content into problemsof more refined techniques is left open. Broch knows that attemptsto characterize true cognition in the sciences internally, for exampleby following standards of rationality, do not suffice; they have tobe complemented by focussing upon what he calls the >irrationalroots< or >irreducible remainders<12 which show up when one paysattention to the way scientific activity is growing out of daily lifeproblems. It is likewise necessary, Broch insists, in order to graspthe cognitive value of the arts, to pay attention to the basic struc-

102 KUNO LORENZ

ture of a symbol - its determination by an >inseparable connectionof archetype and logoS<.13 The unity of knowledge is dependent onrealizing the "Totalitat des Erkennens und Erlebens" ("totality ofknowing and experiencing").I4

With these two terms 'Erkennen' and 'Erleben', we have againarrived at that basic distinction which I labelled earlier, usingBertrand Russell's terms, 'knowledge by description' and 'knowl­edge by acquaintance'. In fact, 'Erkennen' and 'Erleben' are almostprecisely the German equivalents of the two Russellian terms, andcan be characterized roughly in the following way, and I would liketo turn your attention again to Wittgenstein's rabbit-example whichI have quoted earlier for better visualization.Knowledge by acquaintance is knowledge completely dependent

on the situation of speaking and acting - one knows what the excla­mation 'rabbit' means only within a rabbit-situation -, I call it'object-competence'. On the other hand, knowledge by descriptionis knowledge independent of that situation - one knows what thereport 'rabbit' means also outside that situation, by previous expe­rience, so to speak -, I call it 'meta-competence'.The level of language belongs to both sides, to the level of objects

and to the level of signs, depending on the circumstances. Iflanguage, or another sign-system, is constitutive of the objects, i.e.,if it is applied object-competence, it displays symptomatic, object­like, features, if it is descriptive of the object, i.e., if it is appliedmeta-competence, it displays symbolic, representational features.Throughout his life Broch ponders systematically and historically

on the common roots of science and art. He treats both as contin­uous rational >organizations< (Formungen) of the unorganizedwhich, in fact, contains always previous rational organizations. Boththe sciences and the arts are reorganizations of world views quite intune with Goodman's idea that worldmaking consists in deriving aspecific new version out of an older one. Now, scientific (re)orga­nizations take place on the level ofmeta-competence of second order,with their two branches of giving descriptions about the knowledgeby acquaintance - this is called (scientific) research - and of givingdescriptions about the knowledge by description - this is called (sci­entific) representation; you may think of rendering the descriptiveknowledge axiomatically.On the other hand, artistic (re)organizations take place on the levelof object-competence of second order, that is, an ability to construct

CONCEPTUAL AND PERCEPTUAL KNOWLEDGE 103

signs (semiotic objects) showing thereby that you are acquainted withstructured objects, and also that you have descriptive knowledgeabout them. With respect to showing a knowledge by acquaintance- which is dependent, as we have seen, on at least partial presenceof the object - the artistic activity is called poiesis. What counts iswhat you do with semiotic objects like pictures, music-pieces, words,etc., not what they may signify; yet, with respect to showing a knowl­edge by description, the artistic activity is called - with a terminvented in antiquity - mimesis. What counts is that which thesemiotic objects signify. And it is this object-competence of secondorder which we should call perceptual knowledge, in correspondenceto meta-competence of second order being conceptual knowledge.Only here, on the level of second order, we are confronted withwhole structures of pieces of knowledge by acquaintance and ofknowledge by description, which usually is implied when referringto perceptual or conceptual knowledge.Broch treats myths as >first< rational organizations, though, I

think, this should be understood as restricted to artistic organiza­tions, everyday discourse being the candidate for first scientificorganizations. He describes myths psychologically in terms of a fightagainst the anxiety and loneliness of death. Here again, a surprisingcoincidence with another artist-philosopher, Albert Camus, may beobserved. Broch remains within the frame of individual psychologyin his use of the extremes of panic and ecstasy for the possibilitiesof new experience, whereas Camus resolves these extremes dialog­ically into a common knowledge about one's individual loneliness,a knowledge you experience through the limited solidarity of revolt.But both Broch and Camus use the same metaphor to convey thesuccess of >the will to live<, Broch's term, or the >yes to life<, theterm of Camus: it is balance.The rational organization must >fully balance<15 the unorganized

which it mirrors. An art-work must >balance the totality of theworld<.I6 Camus prefers the related term of keeping the >rightmeasure<,17 because each rational organization of oneself finds itsboundary at a rational organization of the other. What is understooddelimits what is not understood and must balance it. Camus drawsthe borderline between different individuals, Broch between an indi­vidual and that which no longer, not yet, or never, belongs to him.I 8

Broch has even tried to sketch a reconstruction of the primaryepistemic relation between a cognizing subject and the cognized

104 KUNO LORENZ

object which I would like to call to your attention. He starts withthe smallest syntactical unit, a sentence, and he asks in which waya cognitive unit (his term is 'Eidos-Einheit') can be made visible andaudible through it. 19

To answer this question he states that each cognitive unit is relatedto a section of reality which he calls 'Elementarsituation' (anexample he discusses is Flackerndes Licht [Flickering Light]). Inpresent day systematic terminology as derived from Peirce and thelater Wittgenstein the eidos is an object-type (or object-schema), andthe related elementary situation is one of its tokens (or actualiza­tions) such that the primary epistemic relation can be described asunderstanding a situation as actualization of a schema. The eidosis universal, the situation something singular, and the two are strictlycorrelated to each other.With the next step, the articulation of a type serves to show the

correlation of type and token by representing the cognitive unitthrough the syntactic unit. In contemporary terminology this meansthe following: An elementary sentence, e.g. 'this is flickering light' ,states that an elementary situation (a token) instantiates a schema,here: the >eidos< Flickering Light. Every such sentence, if utteredin an appropriate situation, is an accompanying part of that situa­tion which uses >the verbal sense< (looking at that situation wouldadd to it by using the visual sense), thereby making the eidos per­ceptible. We would say that by uttering such a sentence the situa­tion is understood as an instance of a schema, or: by uttering sucha sentence the situation is schematized, i.e., imbedded in a possiblesequence of like situations.But Broch does not stop here. He continues by declaring that the

grammatical subject, i.e., the demonstrator (or logical demonstra­tive) 'this' in our elementary sentence, should be understood as >aprojection of the speaker<.2o To proceed in this way has conse­quences which reach far beyond common logical or even linguisticanalysis which stops at saying that 'this' refers to an object givenin the situation of utterance. Broch says, and here his sophisticationmatches the one of Peirce, that the demonstrator 'this' is within theutterance an index of the construction of the object, about whichsomething is stated, out of its various subjective perspectives.Every object, an instance of a schema, splits into the indefinite set

of modes of presentation with respect to the perceiving subjects, andit can actually be identified with the totality of these perceptions.

CONCEPTUAL AND PERCEPTUAL KNOWLEDGE 105

Within our example one could articulate the perception which is usedwhen uttering 'this' in 'this is flickering light' by 'that which I justsee'. Broch calls the perceptions >identifications of the speakingsubject with its surrounding<; hence, through them there appear thecognizing subjects, whereas the cognized object is, according toBroch, defined by the function to construct the invariant out of themany perceptions.The consequences for the eidos, with the elementary situations

as its actualizations, can easily be drawn. Through articulation eacheidos gets split into perspectives each of which is a sign of the sameeidos, whereas the signs themselves are appearances of a subject.While performing an articulation both aspects are completely amal­gamated: a section of reality is, by virtue of the various signs whichare used for the same eidos, simultaneously present and interpreted.In this way Broch has succeeded in presenting a construction

where object and method, ontology and epistemology, appear asmerely two sides of the same coin in such a way that conceptualknowledge which searches for invariance of the represented is>objective< , and perceptual knowledge which aims at variance ofthe representing means is >subjective<.I have to leave out how his procedure works in more complex

cases beyond essentially one-term-articulations of a situation asinstances of an eidos, though it would be interesting to see how hewould treat Wittgenstein's Tractatus, and Joyce's Ulysses as com­paratively pure examples of a complex conceptual knowledge and acomplex perceptual knowledge, respectively.21A sentence, or better the correlated predicative expression,

becomes a symbol in Broch's sense (systematically it should becalled a 'symptom') not by simply being a sign of a schema but byexemplifying the schema.The difference between exemplification and representation as two

ways of reference tied to perceptual and conceptual procedure,respectively, was introduced by Nelson Goodman to account for thedifference between artistic and scientific cognition. And it is sur­prising how close Broch's description of the functioning of verbalsymbols in >philosophische Dichtung< (philosophical poetry) comesto Goodman's exposition. Exemplifying a schema refers to theschema not by convention but by certain internal or external prop­erties which the exemplifying predicative expression has in commonwith the schema. By implication the schema is itself not any more

106 KUNO LORENZ

a simple eidos but an eidos with an internal (and external) struc­ture. The cognitive claim of poetic worldmaking is substantiatedaccording to Broch by a work of art being a (structured) symbol ofreality as a whole. So it is tied to art as mimesis, whereas Goodmanhas tied the cognitive claim of poetic worldmaking to art as poiesis.Both are right, obviously, but there is an important relation to beaware of. Poiesis, the >style< of Broch, is the means in order to makethe mimetic function discernable.What is left, is to give an explanation of the relation between the

two types of reference, exemplification and representation, if indeedthey are exploited heavily in the arts and in the sciences, respec­tively. For this purpose I turn to my third and last section, where Iwant to sketch the Peircean idea to derive signs out of objects as itis contained in the late essay 'Meaning' of 1910:

If a Sign is other than its Object there must exist either in thought or in expres­sion some explication or argument or other context, showing how - upon whatsystem or for what reason the Sign represents the Object or set of Objects that itdoes. Now the Sign and the Explanation together make up another Sign and sincethe Explanation will be a sign it will probably require an additional Explanationwhich taken together with the already enlarged sign will make up a still larger Sign;and proceeding in the same way, we shall, or should, ultimately reach a Sign ofitself, containing its own Explanation and those of all its significant parts; andaccording to this Explanation each such part has some other part as its Object.According to this every Sign has, actually or virtually, what we may call a preceptof Explanation, according to which it is to be understood as a sort of emanation,so to speak, of its Object. 22

The argument calls for something which is a sign of itself, that is,which combines object and sign-features, or better: which functionsboth ways. The argument is itself a section of an open sign processon the level of reconstruction (the conceptualization of a Wittgen­teinian language-game). Now, the descending sequence of interpre­tants ends with an ultimate logical interpretant23 which is identifiedas a habit-change, Le., in contemporary terminology, as the acqui­sition of an action schema such that all the ways to deal with theobject in that respect which is signified by the initial sign areincluded.For further clarification it is useful to turn to the Peircean reading

of the semiotic triangle as it is contained in the following quota­tion: "A sign [ ... ] stands in such a genuine triadic relation to itsobject as to be capable of determining its interpretant to assume the

CONCEPTUAL AND PERCEPTUAL KNOWLEDGE 107

same triadic relation to its object in which it stands itself to the sameobject."24Each interpretant (cognitions of a mind, i.e., mental interpretants,

are some of them) is itself a sign for the same object, hence, it isgenerating a new interpretant, and so on. The sequence of interpre­tants which gets started in that way may be called a sequence ofgrowing understanding of the object by supplying more and moredifferentiated determinations. It should not be forgotten that Peirceinsists upon ever new, i.e., logically new, not empirically new, sign­users connecting the items of the sequence (therefore >quasi-minds<and not >minds<):

Signs require at least two quasi-minds, a quasi-utterer and a quasi-interpreter; andalthough these two are at one (i.e. are one mind [the dialogically constituted generalsubject!] in the sign itself, they must nevertheless be distinct. In the sign they are,so to say, welded.25

Going back to habit-changes, that is, to the acquisition of an actionschema, as the candidates for something which is a sign of itself,and thus the end of the descending sequence of interpretants, we mayconclude that a verbal sign of an object signifies a range of possi­bilities to deal with that object. Even more general, deleting thedummy term 'object', we may say that to understand a sign-action(a symbolic action) is tantamount to knowing, by that very action,of a whole range of further actions which should be called beingsignified by the symbolic action.The sign-action is signifying symptomatically and not yet

symbolically, if the sign-action itself is part of the range of actionsit signifies (in case of verbal symptomatic sign-actions you have noseparation of word and object: their relation is external, e.g. causal,and not internal, or symbolic). The same idea of explaining howsymbolic actions, i.e., verbal sign-actions, symbolize can be used toexplain how ordinary actions function as (not yet verbal) sign­actions. We, then, arrive at the following equivalence: to understandan actualization/a performance of an action, >knowing what one isdoing<, is knowing ways of dealing with it. In a more general way,you may say that to know an object is the same as treating this objectas a sign of its distinctions, that is, the ones you are able to make.And again, this is nothing but treating an object as a sign of itsinternal structure, a structure which gets exhibited in an open-endedsign process.

108 KUNO LORENZ

The simplest structure you start with, is, of course, the one whichshows up when you are able to >read< an actualization as a sign ofits type/schema. And this works when you treat such an ability asconveyed by a >language-game<, the same one which we know,already, under the label of an acquisition process of action compe­tence in an elementary dialogue-situation: the agent is performingthe action, and the patient at the same time is recognizing it, role­switching included.We are at the primitive stage where you have an icon of an object,

an action both as a pragmatic entity and as a semiotic entity, whereas,making a further step, you learn to split the type into parts, an actu­alization of one of them becoming a sign of the whole. This, then,is the stage where you have arrived at an index of the object: "Ifthe sign be an index, we may think of it as a fragment torn awayfrom the object, the two in their existence being one whole or a partof such a whole."26These are the symptomatic sign-actions. The rest is well known,though, of course, there are many details to cope with which I cannotdiscuss here.I only hope that I have succeeded in convincing you of at leastone point when looking at the three authors together in the way Idid: Using language may always occur on two levels, like a per­ception and like a conception. Perceptually, language is used as away of presenting an object which is a case of object-competence ­whereas, conceptually, with language you are referring back to otherterms used perceptually which is a case of meta-competence. In theWittgensteinian example of a rabbit-situation, the 'rabbit' -exclama­tion is a case of perceptual use, and the 'rabbit' -report a case ofconceptual use.There is always a natural and a conventional relation betweenwords and objects, a result which could now be used as a securestarting point for a critical evaluation of the debate on literal andmetaphorical meaning of signs including verbal signs.

NOTES

I The signs '>' and '<' are used for marking improper usage in distinction to apos­trophes which are used for naming.2 Cf. Questions Concerning Certain Faculties Claimed for Man (1868), CP 5.257.3 Goodman's answer 'Credence, Credibility, Comprehension' to Putnam's 'Reflec­tions on Goodman's Ways of Worldmaking', in: The Journal of Philosophy 76(1979), p. 619.

CONCEPTUAL AND PERCEPTUAL KNOWLEDGE 109

4 Denkerische und dichterische Erkenntnis (1933), KW 9.2, p. 48.5 Gedanken zum Problem der Erkenntnis in der Musik (1934), KW 10.2, p. 243.6 Entwurf fUr eine Theorie massenwahnartiger Erscheinungen (1941),KW 12, p.43.7 Das Weltbild des Romans (1933), KW 9.2, p. 116.8 Ibid.9 Das Bose im Wertsystem der Kunst (1933), KW 9.2, p. 136; cf. Die mythischeErbschaft der Dichtung (1945), KW 9.2, p. 204.10 Theologie, Positivismus und Dichtung (ca. 1934), KW 10.1, p. 234; cf. DerKunstkritiker (1920), KW 9.2, p. 39.11 Ibid.12 Cf. Die sogenannten philosophischen Grundfragen einer empirischenWissenschaft (ca. 1928), KW 10.1, p. 131.13 Cf. Einige Bemerkungen zur Philosophie und Technik des Ubersetzens (1946),KW 9.2, p. 68.14 Denkerische und dichterische Erkenntnis, KW 9.2, p. 46.15 Cf. Mythos und Altersstil (1947), KW 9.2, p. 215.16 Die mythische Erbschaft der Dichtung, KW 9.2, p. 210.17 Cf. L'homme revoIte, chap V, sect. 'Mesure et demesure'. In: Essais d'AlbertCamus, Paris 1965, Gallimard, pp. 697-704.18 Cf. Autobiographie als Arbeitsprogramm (1941), KW 10.2, p. 201.19 In his essay 'Uber syntaktische und kognitive Einheiten' (1946), KW 10.2, pp.246-299.20 Cf. op. cit., p. 252.21 For a more extensive treatment cf. my essay 'Broch's Concept of"Philosophische Dichtung"', in: S. D. Dowden (ed.) (1988), Broch, H., Literature,Philosophy, Politics. The Yale Symposium, Camden House, Columbia SC, pp.303-314.22 CP 2.230.23 Cf. CP 5.476 (ca. 1906).24 CP 2.274 (ca. 1902).25 CP 4.551 (1906).26 CP 2.230.

BIBLIOGRAPHY

Broch, Hermann: 1974-1981, Kommentierte Werkausgabe I-XIII, P. M. Liitzeler(ed.), Suhrkamp Verlag (KW), Frankfurt am Main.

Goodman, Nelson: 1978, Ways of Worldmaking, The Harvester Press, Hassocks(Sussex).

Peirce, Charles S.: 1931-1958, Collected Papers of Charles Sanders Peirce l­VIII, C. Hartshorne and P. Weiss (eds.), (A. W. Burks (ed.), VII-VIII), HarvardUniversity Press, Cambridge, Massachusetts.

Wittgenstein, Ludwig: 1953, Philosophical Investigations, Blackwell OxfordlMacMillan, New York.

PART II

THE USES OF METAPHOR

RICHARD WOLLHEIM

METAPHOR AND PAINTING

1

Nowadays the term 'metaphor' enjoys a considerable popularity inthe criticism of the contemporary visual arts. In part the explana­tion may lie in the fact that, since the decline of figuration, or atany rate the break-up of its monopoly, there has developed the unar­ticulated feeling that metaphor may be the route along which thevisual arts can reclaim some of the meaning that this process has lostthem. However this feeling - if I am right in positing it - has notled to what one might expect: that is, a systematic attempt to workout what metaphor might be in the field of the visual arts. In thisessay I set out to make good this deficiency, particularly for painting.In doing so, I shall draw upon certain ideas developed in Paintingas an Art.!

2

An adequate account of pictorial metaphor would have to meet thefollowing requirements:a) it would have to be internally consistent;b) it would have to be in the spirit of the best available account wehave of metaphor in that area where we all can identify clearand undisputed examples of the phenomenon, Le., language;

c) it must ultimately fit into whatever general theory we believe tobe correct of how painting acquires meaning;

d) there would have to exist some body of painting that is metaphor­ical in the sense suggested by the account and that is illuminatedby being thought of as such.

I believe that the account I shall prepare satisfies these four require­ments.

3

That my account meets the first requirement - or is consistent - is

113

F. R. Ankersmit and J. J. A. Mooi} (eds), Knowledge and Language,Volume III, Metaphor and Knowledge, 113-125.© 1993 Kluwer Academic Publishers.

114 RICHARD WOLLHEIM

something that I shall not attempt to prove. I have done my best tosecure it, and I can only wait for objections. As to the fourth require­ment - or that there is a body of painting that fits the account - Ishall, as far as this paper is concerned, rely entirely on what I haveto say about two painters when I come to apply my account in thelast section.I turn then to the second and third requirements: that the account

should be in keeping with what we think characteristic of linguisticmetaphor and that it should fit into the general theory of pictorialmeaning. Now initially there might appear to be a tension betweenthese two requirements. More specifically tension might seem toarise if we accept, as I argue in Painting as an Art that we should,that pictures acquire meaning in a way that is radically different fromthat in which language does. The fundamental difference is this:linguistic meaning is, in a broad sense, rule- or convention-governed,pictorial meaning is not. Pictorial meaning (I argue) requires whatI call a 'psychological' theory: that is, a theory that invokes, overand above the look of the picture, appropriately established, onlyinternal or mental factors: in particular, the fulfilled intentions of theartist and the experience of the adequately sensitive, adequatelyinformed, spectator. If pictorial meaning is grounded in such a dif­ferent way from linguistic meaning, how can it possibly be that lin­guistic metaphor provides the model for pictorial metaphor?This problem is, I believe, illusory. Reflection shows there to be

no real tension between, on the one hand, thinking that linguistic andpictorial meaning arise, or accrue to their bearers, in different waysand, on the other, asserting the paradigmatic character of literarymetaphor. And that is because an account of what linguistic metaphoris, is not as such an account of how linguistic metaphors gainmeaning. It is an account of how we identify certain stretches oflanguage as metaphorical, and we may consider it incidental thatsome thinkers have held that what is identificatory of these stretchesof language is how some of their crucial components gain theirmeaning. A parallel may clarify the point. A linguistic portrait gainsits referential meaning, or is of this person rather than of that person,for reasons altogether different, or so I maintain, from those whichsecure a pictorial portrait its meaning, or make it a portrait of thisperson rather than of that person. Nevertheless it is plausible to thinkthat why we believe each is a portrait has a common rationale.However in this case, in the case of the portrait, we are likely to

METAPHOR AND PAINTING 115

think that, if we want to grasp this rationale, it is best to take thepictorial portrait as paradigmatic.Once we have in the metaphorical case similarly sorted out the

question how we identify certain pictures as metaphors from thequestion how pictorial metaphors gain their meaning, we can see thatthere is no a priori objection to learning what we can from how weidentify certain stretches of language as metaphor. Indeed in the caseof metaphor the attractions of starting at the linguistic end are great.For there the term has an application that is universally conceded.Undoubted examples of linguistic metaphor are

Juliet is the sunReligion is the opium of the peopleNo man is an island unto himself.

What makes them metaphors?

4

The account of linguistic metaphor that I favour is that advanced byDonald Davidson.2 There are three features of this account that needto be brought to the fore.a) Linguistic metaphor does not require that the key words, or thewords that pick out what I call the 'metaphorizing' term - in myexamples 'the sun', 'opium', 'island' - change their meaning.Indeed reflection suggests that, if the metaphors are to work, ifthey are to convey the life-giving quality of Juliet, or the stulti­fying effect of religion, or the interdependence of humanity, thenthe sense of these words must not alter.

b) Linguistic metaphor does not require that there should be anypre-existent link, or any link that pre-existed the making of themetaphor, between the metaphorizing term and the subject of themetaphor, or what I call the 'metaphorized' term - in theseexamples Juliet, religion, man. It is conceivable that the metaphorcould survive such a link, but knowledge of the link would almostcertainly weaken the metaphor.

These two, basically negative, features of metaphor serve to bringout its essentially unregulated or improvisatory nature. The thirdfeature takes us to its essence.c) Linguistic metaphor distinctively does something rather than sayssomething. Better: in virtue of what it says, it does something,

116 RICHARD WOLLHEIM

and it says what it says in order to do this thing. As the result oflinking two terms, it casts light on one of them, and that is itsaim. The light is generated solely through conjoining the twoterms, and just what we are enabled to see by the light of themetaphor, or how the metaphor asks us to conceive of themetaphorized term, is something that we can only very approx­imately express in words. Metaphor is fundamentally ineffable.

Now, if this account is right, anything that is properly thought of aspictorial metaphor should surely preserve these three features to ashigh a degree as possible. It is in this sense that linguistic metaphoris paradigmatic. An appropriate rewriting of the features that makelinguistic metaphors gives us the following requirements, whichI further believe a certain number of pictures meet. A pictorialmetaphor is such thata) the ordinary mechanisms by means of which pictorial meaningis generated are in no way suspended or modified;

b) there need be no pre-existent link between anything to do withthe picture and that which the picture metaphorizes;

c) what is distinctively metaphorical about the picture is that it caststhe metaphorized term in a fresh light through conjoining it withsomething else: once again, a metaphorizing term. It presentsthe metaphorized term under a special conception, and this con­ception we can say something about, but we cannot totallycapture it.

5

Nevertheless there are big differences between linguistic metaphorand what, on the basis of their satisfying these requirements, we areentitled to think of as pictorial metaphors. These differences in turnspring from the large differences between the ways in which meaningaccrues to language on the one hand and to pictures on the other. Ibelieve that there is one difference in particular between linguisticand pictorial meaning that proves to be significant here.On the account I have recapitulated, linguistic metaphor may be

said to belong, not to semantics, but to pragmatics. Indeed puttingthe matter like this is a good way of bringing out how linguisticmetaphor depends on meaning without being a form of meaning.However this claim cannot be carried over bodily to pictorialmetaphor, for the reason that the distinction between semantics and

METAPHOR AND PAINTING 117

pragmatics, or between the meaning of a sign (to use the most generalterm available) and different uses to which it can be put on differentoccasions in virtue of its meaning, does not hold in any satisfactoryway for pictures.Why is this?There may be various reasons, but one reason seems to be that

this distinction, if it is to have powerful application, presupposesanother distinction which lacks application to pictures. The furtherdistinction is one between the sign itself, now conceived of as a type,and instantiations of the sign, now conceived of as its tokens, so that,at any rate in the first instance or fundamentally, it is the type thatpossesses meaning, whereas it is the tokens that are put to use.Meaning adheres to the type, use to its various tokens, essentially.The example of language suggests this, and much else seems toconfirm the suggestion. Now, if this suggestion is right, then the factthat pictures are naturally thought of as individuals, not types, andthus are without instantiations of themselves or tokens, explains whythe distinction between semantics and pragmatics, hence the alloca­tion of metaphor to pragmatics, is inapplicable within the pictorialdomain. In consequence, some independent thinking about pictorialmetaphor is called for.

It is to be noted parenthetically that sometimes, or in certaincontexts, we do think of pictures as types. This happens with highlyschematic pictures: so we can think of the picture of a train, or thepicture of a falling rock, or the picture of a deer. And, when we do,we not merely think of these type-pictures as having their tokens,but we are ready to contrast the meaning that belongs in the firstplace to the type-picture and the uses to which the various token­pictures, which inherit this meaning from their types, can be andare put. So, for instance, some tokens of the type-picture, the pictureof a falling rock, or some tokens of the type-picture, the picture ofa deer, will, as a result of their placement by the side of the road,serve as warnings: they will serve as warnings that - and this iswhere the meaning inherited from the type comes in - falling rocksconstitute a constant danger or running deer may be encountered.However, if it now starts to look as though, after all, the distinctionof semantics and pragmatics can be applied to pictures, two thingsmust be pointed out. In the first place, these are only marginal cases,and secondly, and more to the point, even in these cases the dis­tinction sits very superficially on the facts. For the use that adheres

118 RICHARD WOLLHEIM

to the token-pictures does not do so essentially. Remove these tokenpictures from the side of the road, and they are no longer warnings.By contrast, particular utterances addressed by me on a particularoccasion to a particular person, of the form 'There's a rock fallingin your direction' or 'There's a deer about to run across the road',are, if ever warnings, always warnings. There is no way in which aparticular token sentence, identified by its time and place of utter­ance, could take on, posthumously as it were, a new or different prag­matic force. Of course, a different token of that very same sentence,uttered at a different time or place, could be uttered with differentpragmatic intent and therefore have a different pragmatic force.

6

One way of recording the differences just discussed is to say that,whereas in language metaphor is not a form of meaning, withpictures it is. Pictures can have, say, representational meaning,expressive meaning, and metaphorical meaning. I personally favourthis way of putting the matter, though it is crucial to remember that,in saying this, we are not saying that with pictures their beingmetaphorical is part of their semantics.However, if we do say that pictures can have metaphorical

meaning, the point of the claim that in the case of language there isno such thing as metaphorical meaning should not be lost in theoverall account of pictorial metaphor. It should turn up somewhere.For it to do so, we have got to be clear what the point of the claimis. And the point of that claim is not just to emphasize that inlanguage there is a distinction between semantics and pragmatics.The claim, of course, makes use of that distinction, but its point, asI see it, is to insist that meaning-rules or conventions do not shiftwhen metaphor appears on the scene. Indeed, not merely do they notshift, they must not shift. So the question arises, How is this truthto be registered for pictorial metaphor, or in an area where, if I amright, meaning depends not on rules or conventions, but on the expe­riences that marked surfaces generate?I have a suggestion to make, which is not only bold, but far bolder

than it need be to satisfy what is required of it. However it seemsto me to do justice to independent intuitions we have about picto­rial metaphor. The suggestion is this: that, with pictorial metaphor,the metaphorizing term is not something referred to by the metaphor-

METAPHOR AND PAINTING 119

ical picture or some component of it in the way in which with lin­guistic metaphor the metaphorizing term is something referred toby the metaphorical sentence or some component of it. In the caseof a metaphorical picture, it is the picture itself, the picture taken asa whole, with all the ordinary mechanisms of meaning in place, thatis the metaphorizing term. Hence it is the conjunction of the pictureexperienced as a whole with the metaphorized term that casts thelatter in a new light, or presents it under some fresh and revealingconception.To this suggestion, another associates itself as a natural corol­

lary: and that is that the metaphorized term also is not referred toby the metaphorical picture. What the picture is a metaphor for isnot directly revealed in the picture. It may be indirectly revealed: inthat, other things being equal, what the picture metaphorizes can beinferred from what the picture is well suited to metaphorize. In pointof fact - though I have no arguments for this last piece of my accountof pictorial metaphor - I believe that there is just one thing that all(at any rate) great metaphorical paintings metaphorize: and that isthe human body. This, I must emphasize, is not part of a generalclaim about metaphorical art: it does not profess to hold good forall the arts. It is a belief about one particular art, painting, and thereason why it holds for this art is, I suspect, connected with itsmateriality. Painting, if not the most material of the arts, is asmaterial as any. Pigment invokes the body and its constituents orproducts.

7

What remains to be done is to fit this account of pictorial metaphorinto the theory of meaning that I believe to be correct for painting.This is what I call a psychological theory, and I have already set itout in outline. It is a theory that invokes, over and above the lookof the picture, the intentions of the artist in so far as these are ful­filled in the picture and the experience that the picture will give riseto in the mind of an adequately sensitive, adequately informed, spec­tator.

It is not difficult to see how the account of pictorial metaphor Ipropose fits within the theory of pictorial meaning I believe to beright.Something like this seems to be correct: A picture is a metaphor

120 RICHARD WOLLHEIM

for M, moreover it casts M in a certain light, or presents it undera certain conception, if and only if it has a certain overall look anda) this look derives from an intention on the artist's part that, bylinking the overall look of the picture with M, an adequatelysensitive, adequately informed, spectator should come to thinkabout and feel towards M in the desired way; and

b) the look is such that an adequately sensitive, adequately informed,spectator, who takes in the look, will be led by the experiencethis causes in him to think about and feel towards M in the desiredway.

Though this is stated in terms of necessary and sufficient conditions,it is obvious that refinements are called for if accuracy is to beattained. I shall not attempt such refinements here.However, with the formulation as it stands, certain comments are

in order.a) the term 'intention' is, as I insist throughout Painting as an Art,to be understood in a broad sense. As well as the artist's desirethat the spectator should be led to a certain way of experiencingthe metaphorized term, there will, characteristically, be emotions,feelings, phantasies on his, the artist's, part directed on to themetaphorized term out of which this desire is generated. Somesuch further set of psychological factors are required if the pictureis to count as a genuine or sincere metaphor and not as (say) anexercise in metaphor. It will count as a mere exercise in metaphorif the artist's desire to transform the spectator was unmotivated;

b) the experience that the appropriate spectator can be expected tohave in front of the metaphorical picture ensures that the artist'sintention is fulfilled: the two conditions in my account are inthis way linked;

c) insistence upon the second condition does more than ensure thefulfilment of the first condition. It also makes it clear how thefirst condition is to be fulfilled. Given that these are conditionsupon pictorial metaphor, the artist's intention must be fulfilled,or the transformation of the spectator effected, through a certainkind of experience that the spectator has, and has because theoverall look of the picture causes it in him. It is, for instance,out of order that the spectator should come to think about andfeel towards M in the way that the artist desires just throughlearning that this is what the artist intended. However that hemust do so through having a certain experience of the picture as

METAPHOR AND PAINTING 121

a whole is in turn compatible with this experience of the picturebeing dependent on - though not, of course, dependent solelyon - his knowledge of what the artist intended of him;

d) the experience that the appropriate spectator must be capable ofhaving through looking at the picture is grounded in perception;indeed it is partly perceptual; but it also transcends perception.It is partly affective.

I intend to conclude by illustrating my account from the work of twogreat painters of the western tradition. But, before I do so, let meinsist on a distinction. If it is lost sight of, then my account has nochance of seeming plausible. The distinction is this. I have beenoffering an account of pictorial metaphor, or of pictures that aremetaphors. But, alongside pictures that are metaphors, there arepictures that have as their content, or contain, metaphors. Thesemetaphors, presumably not themselves pictorial, are probably lin­guistic, and examples of the kind of picture I have in mind areChardin's House ofCards or the series of paintings by the Americanartist, Thomas Cole, illustrating the various stages of the 'journeyof life'. Chardin's painting contains the metaphor 'Life is a houseof cards', Cole's pictures contain the metaphor 'Life is a journey'.Such pictures could also be, if the conditions I have talked aboutare satisfied, pictorial metaphors. But to them in their own right, orthought of as pictures that are about, or that contain, metaphors,nothing of what I have been saying necessarily applies. In Paintingas an Art I proposed that we should regard them as having metaphorsas their textual content in the same way as other pictures have moraldoctrines, or cosmologies, or proverbs, or religious dogmas - thatis, in each case something with a propositional content - as theirtextual content.

8

I now return to the two great metaphorical painters of the westerntradition from whose work I intend to illustrate my account. Sincethey are great painters, you will not be surprised to find my believingthat their paintings metaphorize the body. They cast the body in anew light, they present it under a highly personal conception. Thepainters are Titian and Willem de Kooning. In this section I can dono more than hint at what it is about their paintings that allows themto serve along the lines I have indicated as metaphors for the body.

122 RICHARD WOLLHEIM

The property that does this, which is, of course, an overall propertyofthe picture, I call 'corporeality'. Corporeality depends upon certainoptical effects that the pictures that have it produce, and, though,since the days of the great Heinrich Wolfflin, the standard instru­ment of art-history has been the slide, there is no reason to believethat corporeality is preserved in a slide. What I want to ask, for bothTitian and de Kooning, is how each artist generates corporeality, andwhat, in each case, is the specific conception of the body deployed.I start with Titian, and with two early paintings painted close

together in time, both of the very highest quality: the ConcertChampetre (formerly ascribed to Giorgione) in the Louvre, and TheThree Ages of Man, on longterm loan to the National Gallery ofScotland.Titian makes his paintings metaphorical of the body in two stages.

First, in the way he represents the body: and then, secondly, in theway he transfers characteristics from the represented body to thepicture as a whole. Throughout his career, but most vividly in theearly pictures, Titian envisaged the body as the locus of vitality andemotion, intimately connected on the one hand with pleasure andon the other hand with the inevitability of death. The principal meansthat Titian exploits to exhibit the body, the represented body, underthis conception is, as I see it, the pivotal mechanism underlying rep­resentation: what I call 'twofoldness', or the way in which the spec­tator is simultaneously made aware of a marked surface and discernssomething in front of (or behind) something else. Titian exploitstwofoldness in that he simultaneously makes it look as though thatpatch of the marked surface in which we see a body is spreadingoutwards and as though the body we see in the patch is springinginto action or sagging back into repose.With this central conception of the body established, Titian then

uses a variety of devices to shift the effect that sustains it to thepicture as a whole. The devices are varied, they cannot be identi­fied outside the experience of the pictures that contain them, andhow they advance corporeality admits only of visual demonstrationin front of these pictures. The devices includea) a series of correspondences Titian establishes between the bodyand the other objects represented in the picture: between fleshand stone in the Concert Champetre, between young skin and skyin the Three Ages of Man;

b) the simplification of colour, and the use of near-complementaries,

METAPHOR AND PAINTING 123

which have a singularly binding effect on the picture as a whole;C) the benign neglect of perspective, so that what has been called'the gouging-out' of space, characteristic of much Florentinepainting, is avoided;

d) the anonymity of the represented figures - a device inherited fromGiorgione, but receiving at Titian's hands a further emotionalcharge - so that there is little in the way of distinct personalityto serve as an obstacle to the transfer of Titian's conception ofman from them to the picture as a whole. Interestingly enough,Titian preserves what we may think of as this protean view ofman where we would least expect it - in his portraits, at leastthe finest of them.

What is singularly striking in Titian's work is how he conserves rightinto old age, indeed into very old age, both his original conceptionof the body and his resolve to produce paintings metaphorical ofthe body under this conception. It is true that the conception isdeepened and saddened by experience. Vitality is, by the last phase,connected less with joy and the bitter sweetness of transience, morewith death, murderousness, and the will to snatch triumph out ofdefeat and physical humiliation. In the unbearably tragic Flaying ofMarsyas Titian's conception of the body is itself in extremis.What goes along with the persistence of Titian's original con­

ception of the body is the persistence of the original means wherebyhe achieves corporeality: that is to say, of the means whereby hetransfers something that characterizes the represented body to thepicture itself. Here too persistence proves compatible with develop­ment. For both the way in which colour knits together the pictureinto a whole and the way in which Titian establishes correspondencesbetween body and non-body evolve so as to give rise to the amazingperformances of the last years. Words cannot capture those effects,though I have tried to do so in Painting as an Art.The other metaphorical painter I want to consider is de Kooning.

There are many surface dissimilarities between de Kooning andTitian, and many deep similarities. A deep dissimilarity is this: that,whereas Titian's metaphors for the body are grounded in the repre­sentation of the body, de Kooning's are not. Even in de Kooning'sfigurative paintings, even in those paintings of his in which heactually represents the body, this aspect of the painting makes, as faras I can see, no contribution to the metaphorical meaning of thework. This has in large part to do with the fact that the conception

124 RICHARD WOLLHEIM

under which de Kooning metaphorizes the body is one connectedwith our earliest attitude towards and experience of it, predatingany organized visual schema of it. The body that de Kooning's paint­ings capture is supremely the locus of sensation. It is the locus ofthose sensations which arise in the course of our very first contactswith the world. I quote from myself when I list them as

sucking, touching, biting, excreting, retaining, smearing,sniffing, swallowing, gurgling, stroking, wetting)

Now these sensations - sensations, in each case, closely associatedwith some activity of ours - are sensations that occur to - or perhapsbetter within - a fragile, primitive self, whose boundaries only withgreat difficulty contain them. The archaic self, who is their owner,is in constant fear of disintegration under the impact of the aggres­sion they express and the excitement they generate. It is my sug­gestion that, in so far as we can distinguish between the primitiveself and its contents, de Kooning equates the former, the self, withthe near-square edges of the picture and the latter, the contents ofthe self, with what lies on the surface that they frame, and thedramatic confrontation that he then sets up between the great tur­bulent brush-strokes that squelch across the support and the formatwhose integrity they appear to threaten metaphorize the early lifeof what Freud called 'the bodily ego'. A series of pictorial resolu­tions correspond to different phases or positions in our early psycho­biography. Sometimes the massive furrowed strokes swerve so asto avoid collision with the edge. Sometimes this cannot be done, andone crashes into the other, and then we observe the tiny fragmentssucked back into the interior of the picture, like the frothy undertowof a spent wave. And sometimes the paint works its way forward andforms a protective barrier around the periphery of the picture, therebysuggesting the way in which benign sensations side with the selfagainst its more unruly contents.As we first look at these pictures, our attention is likely to be

monopolized by the insurgency of the paint. We pay no attention tothe format, which seems merely to mark where the paint stops. Thenthere is a reversal of attention, and we come to recognize that whatthe paint attests to, even at its most insurgent, is the frame that keepsit in place. It is now the victory, the hard-won victory, of the pictureas a whole over its sumptuous detail that engrosses us. Translated

METAPHOR AND PAINTING 125

into terms of self and sensation, these pictures tell us, metaphori­cally, of - better, they show us - the earliest triumphs of the self.

NOTES

1 See Wollheim, R. (1986).2 See Davidson, D. (1978).3 Wollheim, R. (1986), p. 348.

BIBLIOGRAPHY

Davidson, Donald: 1978, 'What Metaphors Mean', Critical Inquiry 5 (1); reprintedin Sheldon Sacks (ed.), On Metaphor, Chicago University Press, Chicago,Illinois, (1979), pp. 29-45, also in Donald Davidson, Inquiries into Truth andInterpretation, Oxford University Press, Oxford (1984), pp. 245-264.

Wollheim, Richard: 1986, Painting as an Art: The Andrew W. Mellon Lectures inthe Fine Arts 1984, Princeton University Press, Princeton, New Jersey; also pub­lished in 1987, Thames & Hudson, London.

SANDRO BRIOSI

THE CONFUSED GOD: ABOUT A METAPHOR

IN LITERARY SEMIOTICS

A recurring metaphor in general discourse as well as in specializedcriticism describes the phenomenon of narration in terms of theauthor's position in a "world" with which he maintains certain rela­tionships. It is a spatial metaphor and seems quite clear, but itsapparent clarity turns opaque as soon as we try to derive furtherheuristic conclusions from it. We regard it as quite obvious, forinstance, that the author of an autobiography takes up his stanceinside his world, to the point of being the very incarnation of hisnarrator - as opposed to the realistic writer, who views his worldfrom the outside, or, should he have any populistic or paternalisticleanings, even from above. Yet experience has shown how easilysuch a classification may fall apart. Verga, the great novelist of theItalian type of realism known as verismo formulated the theory ofthe total detachment of the author, of his being reduced to a neutraland "impersonal" recording instrument. It is precisely this imper­sonality, however, that places the author in a position of "weakness"relative to the Sicilian world he is describing. The very effort oferasing all preconceived notions and feelings, of wiping his innerslate clean, so to speak, carries a "risk": the author's personality maybecome wholly dominated by the outside world. And when thathappens, the author who wants to eliminate all temptation of inter­vening in the story - and who therefore somehow attempts to detachhimself from it - will paradoxically find that his desire to stay outsideit makes him view it entirely from the inside; so that the literarywork as Verga wanted it to be seems "to have come into being ofits own creation."Yet the use of the metaphor has not stopped spreading. The various

types of idealistic aesthetics, for instance, assume the existence ofa critic who sees "through the author's eyes." Whatever he may do,the latter is part of the world he is describing and cannot go outsideit: poetry elevates the literary work to the level of a rounded-off,non-communicating whole, devoid of anything external to it. Weshould remember that according to Croce the basis of the poeticuniverse is precisely its "cosmic" character.

127

F. R. Ankersmit and J. J. A. Mooij (eds), Knowledge and Language,Volume 1II, Metaphor and Knowledge, 127-136.© 1993 Kluwer Academic Publishers.

128 SANDRO BRIOSI

Nearer our time, the semiotic approach to narrative texts repre­sents an attempt to rearrange the system of concepts created in thewake of the metaphor of inside and outside. Lector in Fabula byUmberto Eco is one of the many examples in which the concept ofa "Possible World" is called on for aid. Within this "Possible World,"the author outlines certain "reading routes" (here we have the spatialmetaphor again), routes whose combinations and transformationsmake up the matter of the narrative. But abandoning the psycho­logical meaning of the metaphor for the sake of a meaning of asemiotic kind is not without its consequences: the psychological,historical, empirical author - the Author - cedes his place to his owncounterpart, the Narrator, and therefore we will no longer be tryingto establish his position in the existential or historical reality of theperson who wrote the work, but rather inside the metaphor of theNarrated World, taken at face value. Hence, though the expulsionof the author in favor of the narrator does explain the terms of theobject of critical analysis more clearly and defines them more strictly,it equally increases the danger of "abusing" the metaphor (Turbayne),whose rigidly metaphorical origin we tend to forget. Contrary to allappearance, in the universe created by the narrative's semiotics wehave not yet definitely avoided the pitfall of confusing inside andoutside, but have merely skirted it. In this respect, G. Genette'snarratology offers a case in point; some remarks on Figures III willgive us our starting point towards another type of spatial metaphor,undoubtedly more effective for solving the problem facing literarycriticism: how do author and critic "move" in their paper world?As we know, among the heuristic tools of Genette's narratology

the voice concerns the speaker in the story, whereas the mode, aparadigmatical spatial and visual metaphor, has to do with the"viewing" of the matter narrated by "focusing" on it, either fromthe inside or from the outside. The metaphor of focusing suggeststhe possibility of regarding the various modes of vision as forminga continuum, like the motions of a zoom lens; nevertheless, a cleardistinction is maintained between the narrator who views mattersfrom the inside, "through the eyes" of the protagonist, the narratorwho views them from the outside, and finally the one who does notfocus at all, because he knows more about it all than the protago­nist - that is to say, the omniscient, demiurgic narrator, whom wemay not only consider to be external to the situation, but also externalto the entire universe in which he has been placed and isolated.

THE CONFUSED GOD 129

Hence, we speak of an "internal focus" when the narrator "looks"through the eyes of the protagonist, but in fact the same notion mayapply to diametrically opposed situations: to a case of extreme sub­jectivity (the interior monologue, for example), or to one of extremeobjectivity (the "ecole du regard"). Indeed, the reader of Robbe­Grillet's La Jalousie, for instance, may well wonder how he can besure that the author views the story with the eye of the jealousprotagonist rather than with that of a god. No doubt he will find ananswer in the particular point of view from which things and situa­tions are looked at in the course of the novel - but can one trulyview things like a god, otherwise than by looking at them from aparticular point of view? In the same way, a reading of L'errangerby Camus as expressing a feeling of total alienation from the worlddoes not stop the spatial metaphor of narratology from working inat least three directions. Firstly, we "view" everything through theeyes of the stranger, Meursault (internal focus). Secondly, there isno focusing (zero focus or the omniscient point of view); to justifythis interpretation, we need only decide that the absurdity of theworld is not the effect of the strangeness of the person looking at it,but rather its intrinsic nature. In other words, the narrator knows alland sees all - in fact, there is nothing to be understood in the worldbeyond what one sees. Thirdly, and finally, what is there to stop usfrom taking up again the hypothesis (worked out, for anotherpurpose, by F. Ferrucci in his Il mondo creato, a novel in the shapeof an autobiography of God) of the point of view of a god who hasforgotten the sense of his creation, of a god who knows everythingwhile understanding nothing, hence the hypothesis of a zero focuswhich at the same time is an external one?But for Genette, these distinctions remain valid; in cases where

he is forced to admit that they are jumbled, he blames the author.This allows him to criticize as "incongruous" the following quota­tion from Proust: "Je descendis dans la prairie afin d' aller revoirl'Indre et ses isles, la vallee et ses coteaux dont je parus etre unadmirateur passionne" ("I went down into the meadow in order torevisit the Indre and its islands, the valley and its hillsides whosepassionate admirer I seemed to be").! Genette feels that since he islooking through the eyes of the focusing protagonist, the narratorcannot see the latter and say of him that he "seemed to be anadmirer," but the reply to this is that there is nothing to show thatsomeone looking at the world from a particular point of view cannot

130 SANDRO BRIOSI

simultaneously be seen - indeed, Merleau-Ponty has proved that itis even impossible to "see" without being "visible" at the same time.On the other hand, Genette himself is often aware of the limita­

tions of his own system: about another phrase by Proust, for instance,he remarks "that the ambiguity of the text will not allow us to decidewhether the 'perhaps' forms part of the indirect speech, nor, there­fore, whether the hesitation it denotes issues from the hero alone".2Yet it must be said that the terms "ambiguity" or "hypocrisy," appliedto devices "that allow the narrator to state hypothetically what hecould not state positively without stepping outside the internalfocus," are far from convincing.The fact is that all narration is "ambiguous" or "hypocritical" by

definition; even, nay, above all, in narratives with zero focus, wherethe narrator views things from above. If omniscience were reallypossible, it would lead to silence. We cannot broach a subject withoutfocusing on it, without seeing it as a whole against the backgroundof the rest of the world, in short, without distinguishing it; it isprobably because he knows all that God remains silent, or He maywell feel it is not worth the trouble.... But a mere human being, assoon as he starts telling a story, is always both inside and outsidehis tale. And it is precisely in order to come to terms with this basicambiguity of human consciousness that he employs narrative devices,so as to alleviate his incapacity of seeing without being seen. Theyare used in order to create the illusion of an alternative betweeninside and outside, an illusion that in fact captures on paper thevarious forms of human reality - a reality condemned to live outsideitself as awareness-of-something, and hence in the world, inHusserl's words, but without ever achieving in it the self-annihila­tion required to penetrate that world completely.

It is in phenomenology that we find the concepts that can help usalong towards conceiving more effective spatial metaphors, basedon the awareness of the profound distinction separating the per­ceiving subject from the subject using language. Indeed, one caneasily regard the former as comprising the world of the internalfocus: from that point of view, things-as-the-subject-sees-them arenot distinguished from the awareness-of-something, Le., from theawareness-of-the-subject. Conversely, the person using languagedenotatively or referentially, whether in speech or in writing, findshimself at the heart of a world one might well define as non-focused,a world in which the objects and events are transparent occurrences

THE CONFUSED GOD 131

of the referents foreseen by the differential system of the linguisticcode; a world in which the speaker feels himself to be the caretakerof an omnipotent linguistic and conceptual system, in which every­thing that could possibly occur would find its proper place.But the person telling a story, particularly in writing, is not really

interested in a system of unequivocal concepts rendering him respon­sible for their truth value or for their correspondence to experientialdata; nor, for that matter, does he position himself in the domain ofperception, where the subject can neither lie nor speak the truth,because all things are in him and he is in all things. Narration takesplace somewhere halfway between the domain of perception and thatof reference: contrary to the perceiving subject - who can, becausehe is in all things - neither author nor reader can any longer believein the reality of what they are looking at; they cannot even believein the truth of what they are saying, in the correspondence of theirideas with anything external. The narrator is simultaneously insideand outside.To clarify this mixed and ambiguous metaphor even further, we

can also think of the many writers - and painters, too - who haveborne witness to the fact that the construction of a work of art isachieved by way of a series of attempts and alterations whose aimis certainly not the adaptation of the parts to any preexisting idea:in fact, for the author each modification or improvement transformsthe total image of the work, a work that in this respect remains aunity, a dynamic structure in which everything holds together. Often,a line of verse, an image, a word form the starting point for the poet'swork; the traits of a character, the guidelines of a sequence for thatof the storyteller. Around these first givens, at once a horizon ofmeaning comes into being, to use a spatial metaphor borrowed fromphenomenology; and then this horizon of meaning determines thesearch for the words to follow, the bringing into play of othernarrative structures. Before it has any bearing on the interpretation,circularity - from the whole to its parts and vice versa - has alreadyhad its effect on the actual process of producing the text, because itis the earliest beginning of this circular motion that, perceptible fromthe very first words, gives rise to the correspondence between thestatements and the meaning they convey.It is well-known that a feeling of surprise, and even the impres­

sion of having received a "gift," accompanies the choice of the"right" shape and colour or the "right" words; and it is with this

132 SANDRO BRIOSI

notion that we link the ancient conception of poetic invention asinspiration. The words seem to come from somewhere outside, as ifthey were dictated, and yet the author immediately grasps their right­ness, because they correspond to an idea and a feeling deeply andintimately his. Consequently, for the author that idea and that feelingbecome irrevocably linked with those words - and this gives rise tothe impression of receiving a gift: though the words in question areas they should be, and are his, because he made them up, yet, inanother sense, they actually do not originate with the author, butspring from the possibilities offered by language, by literary tradi­tion, and sometimes even by chance. Thus, the idea and feeling inquestion do no longer form part of his life alone: torn, as Sartrewould say, from "the stickiness of what exists," they bring to lightthat potentially intersubjective character of reality that is theirhorizon of meaning, but without losing their roots planted in theauthor's individual experience. Committed to a system of words,such feelings acquire the status of referents, of an autonomous realitythat is in certain ways external to the existential world of the personwriting. However, in the non-fortuitous, even necessary accordanceof the symbol and the referent, they preserve their character of uniqueevent, inconceivable elsewhere than inside that particular world.It is phenomenological analyses like this one that may help us

out of the impasses created by a metaphorical system of oppositions,of inside or outside, that presuppose the Cartesian notion of thesubject, of the "res cogitans" capable of "moving" from one positionto the next. The teachings of Merleau-Ponty remind us that in per­ception it is the mutual correspondence and the ambiguity of thosetwo terms, inside and outside, that are at work. In cognition 'co­naissance', the knowledge of the subject with respect to the objectis in reality cognation, i.e., the springing from the same origintogether; and it is the latter that is imitated, that is reproduced, inartistic activity. The object recognized as the occurrence of a referentforeseen by the linguistic codes is to refer to the "tacit" meaningwhich encompasses and exceeds it in the concrete act of perceptualintent, what the things described or recounted by the writer are totheir horizons of meaning - a meaning inseparable from them andfrom the way in which they form a unity, a meaning that encom­passes them and justifies them. With respect to the things, situa­tions and characters he gives permanence to on the page, the writeris - we keep coming back to it - always outside, the main reason

THE CONFUSED GOD 133

being that they are givens belonging to everyone's reality. Never­theless, the author is also always inside, for it is he and only he whogives them their meaning. The existence of codes, of norms thatgovern and structure the world and the story's reception, puts theauthor outside his work; but the deviation he imposes on the normsproves that he knows more about it all than anyone else, and putshim back inside.So the "distancing" proper to poetic language does not affect the

supposedly objective existence of things, veiled by the habit of per­ception, as the Russian formalists thought. Actually, the impressionof objectivity is precisely brought about by that habit, by reducingthings to the status of transparent occurrences of referents. What this"distancing" allows us to discover is the subjective dimension ofexperience, as well as the possibility of making it communicableby giving it permanence in the open space between the non-fortu­itous symbol and its meaning. Telling a story, expressing a feelingpoetically, signify that one experiences them, or experiences themagain, "from the inside," but without for a single moment losing thefeeling that they might be, might have been, may be a story and afeeling experienced by someone else.Entirely unique, like an event lived through, and at the same time

clenched tightly upon itself and discrete as a unity of meaning, thework of art allows a rapport between mankind and the world thatseems contradictory. But it seems contradictory only because we takethe metaphors of inside and outside at face value, and because wethink that the world - real or "possible" - is really enclosed like atown within the circle of its walls, and that the narrator is really aCartesian entity, complete and alien, and able to enter and leave bythe town gates. This is why the writer, although he may sometimesthink so, will never be able to see things as God sees them.In fact, even in novels with zero focus, the reader notes hints of

the presence of an author, hints that are sufficiently clear even whenthe narrator remains "invisible." And if the Narrator enters the work,in the form of a narrating protagonist, he simply announces hispresence, giving shape to the awareness (that is always with him)of the impossibility of seeing without being seen, the impossibilityof giving things a meaning without at the same time regardinghimself as the origin - not the absolute origin, but the historical andpotentially intersubjective origin - of that meaning.

It is this basic ambiguity of the Author's condition that charac-

134 SANDRO BRIOSI

terizes the fictional narrative. The adventures of the Narrator, of hiswords and his voice, are merely the reflection of this ambiguity ­an ambiguity that we should not, therefore, reduce to clear anddistinct elements. I feel unhappy about having to use ambiguous for­mulations like inside and outside, but it is not my fault: it is the fault,rather, of the basic ambiguity of the human condition, itself thesource of the ambiguity of the narrative. According to me, the realfault lies in the "abuse" of the paradigmatical metaphor (Turbayne),causing the rigidity of that metaphor and making it responsible foran ontological correspondence to the reality of the world. In that waymy metaphor, my oxymoron of inside-and-outside, turns into adevice of language for the solution of the problems which that samenatural language may cause if one forgets its intimately meta­phorical character in order to hide in the illusory certainties of an"unequivocal" language comparable to the artificial language offormal logic, obviously needed, but only in its legitimate field ofapplication.On the other hand, if we eliminate ambiguity, it is inevitable that

it should creep back into the text at once. Or as we would say inItalian, ambiguity, turned out "by the door," will inevitably comeback in through a "window" that always remains open in the wallof the Possible World of the text.It should suffice to think once more of U. Eco's "possible world",3

a world which he describes as "furnished, but not substantial."Furnished, filled with objects that delimit a space as definite andmeasurable as real space. But those objects are not material ones,they do not really exist otherwise than thanks to the author who hasimagined them. Hence that author is in a certain way part of thatworld, doing duty as a piece of its furniture, and always aware thathe is not a god; part of the world, but just as much outside it, ableat any moment to intervene in order to shift and rearrange the fur­niture, which, hardly given its new place, slips through his fingersin order to settle again into its own unity and autonomy. So we maycompare the Narrator to a god, but a god who can only create worldsthat continually escape him, and never wholly a god. An ambiguousgod: for him, pantheism is out of the question, because he is not trulyomnipresent and cannot even live in his Empyrean in omniscientself-contemplation. Who knows - perhaps the narrator is the onlygod possible in our world, somewhat similar to the absent-minded,confused god of Ferrucci's novel. ...

THE CONFUSED GOD 135

In any case, he is proof of the fact that it is impossible for thewriter to annihilate himself in the object of his own creation, impos­sible even to be a real creator, one who is able to extract things andtheir meaning from outside, Le., from the Void, rather than cullingthem from the concrete intersubjectivity that is history. That is why,when we shift from the angle of creation to that of interpretation,once again what works best is an ambiguous metaphor. Thus notthe structuralists' metaphor, that of "constructing" and "decon­structing" the text, in which the clockmaker-critic clearly remainsoutside; nor the metaphor of "recreating" the work in accordancewith the author's point of view - but the hermeneutic metaphor ofoscillation, of a circular movement that goes from the centre to theperiphery and back to the centre.Definitely, we cannot free ourselves of metaphors, or do without

their ability to structure reality; in this paper I have merely brieflycorroborated the obvious. Most debates are based on the oppositionof metaphorical systems trying to stick to their object in the mosteffective way, each system arising from a particular philosophy.

It is clear that in our case the philosophy that forgets that that thetwo terms in question originally came into being together, the phi­losophy of inside or outside, may be defined as Cartesian thinking,based on the conviction that all reality, that of the possible world ofthe work of art included, can be reduced to discrete, oppositeelements, whereas the philosophy of phenomenology underlying thepresent paper is based on the notion that the world cannot be dis­sociated from the subjects living in it; on the notion that the possibleworld of the work of art (the "world" of the "text" as defined byRicoeur) cannot be dissociated from a relationship with its realauthor, with respect to the latter's freedom as a subject who is alwaysboth inside and outside the story of which he forms part and there­fore also of the story he is narrating.

It has often been said, and has been repeated again during thisconference, that metaphor does not yield true knowledge, sincemostly it only "suggests," vaguely and ambiguously, lines of researchalong which one may later discover new ideas. What this amountsto, however, is identifying "true" knowledge of something with thepossibility of reducing it to clear and discrete elements, to distinctand unequivocal units, to concepts. But if the thing in question ishuman reality and its fundamental ambiguity, this reduction oftenfalsifies knowledge of the matter - and that is why metaphor, with

136 SANDRO BRIOSI

its possibility of allowing us to say two different things at oncewithout disgracing ourselves, may sometimes come to hand as anindispensable tool for making statements about the reality of humanlife and human history; about a reality in which the mental struc­tures concerned in cognition are largely determined by social realityand by the literary or "correct" referents it imposes on the originalrelationship, the relationship of "co-naissance"4 between the subjectand the world; but a reality, too, in which the subject can alwaysrecover that original relationship, can always (thanks, among otherthings, to metaphorical knowledge) create a new arrangement ofthings, beyond the arrangement arising from the structures - whethersurface structures or "deep" ones - of either the natural language or"cognitive discourse," and can always make the presence of itsliberty felt.

NOTES

1 Genette (1972), p. 210.2 Genette (1972), p. 217.3 Eco (1979), p. 122.4 MerIeau-Ponty (1945), Chs. I, II.

BIBLIOGRAPHY

Camus, Albert: 1942, L'etranger, Gallimard, Paris.Croce, Benedetto: 1902, Estetica come scienza dell'espressione e linguisticagenerale, Laterza, Bari.

Eco, Umberto: 1979, Lector infabula, Bompiani, Milano.Ferrucci, Franco: 1976, Il mondo creato, Mondadori, Milano.Gadamer, Hans G.: 1960, Wahrheit und Methode, J.C.B. Mohr, Tiibingen.Genette, Gerard: 1972, Figures III, Seuil, Paris.HusserI, Edmund: 1950, Ideen zu einer reinen phiinomenologischen Philosophie

I, Nijhoff, Den Haag.Merleau-Ponty, Maurice: 1945, Phenomenologie de la perception, Gallimard, Paris.Ricoeur, Paul: 1975, La metaphore vive, Seuil, Paris.Turbayne, Colin M.: 1962, The Myth of Metaphor, Yale University Press, NewHaven.

Verga, Giovanni: 1880, 'L'amante di Gramigna', in Vita dei campi (ed. 1968),Mondadori, Milano.

JAN PEN

ECONOMICS AND LANGUAGE

Economists have a dream. It is about the world, that is society, andabout their own role in that world. The world is silent, no voicesare heard (apart from the fast talking of one special character: theauctioneer, but one of the recurrent themes in contemporary eco­nomics is that the auctioneer is missing). And the economist himself(he is a male) is almost silent too. He is seeing things. In fact, hesees himself as an observer who is capable to see things that cannotbe seen by others: the Invisible Hand, the Demand Function SlopingDownward, the Real National Income as a flow of goods, the LafferCurve drawn on a paper napkin in front of the President of the UnitedStates (who, all of a sudden, sees the deeper truth about Reagono­mics). The economist sees the true nature of market equilibrium, orif he happens to be a keynesian of the hydraulic type, he sees theflow of income and expenditure stagnating in pools of liquidity. Inhis neo-classical dream he visualizes economic growth as a forest,expanding slowly and naturally, and he sees the natural rate ofgrowth which can be described by algebra. For that is the econo­mist's dreamlike language: algebra. Words are superfluous.The human beings in the economist's dream are quiet people like

himself. They are utilitymaximizers, getting their signals from themarket in the form of prices - they are calculating actors, taking carethat their indifference planes are tangent to their budget planes. Novoices are heard, no discussions are going on, and even the mumblingauctioneer is often absent. The market is purring softly. And so isthe economist, who sees the underlying network of prices and quan­tities and who writes, in his dream, an article containing 80% algebra,10% graphs, 5% tables and 5% words. The article is accepted bythe American Economic Review, and then the economist wakes up.The notion that language should be avoided or at least minimized

is not often explicitly defended, but there is among economists animplicit "Look Ma, no words" mentality. Talk economics is a pejo­rative expression, in particular in The Netherlands. And there is atleast one economist of repute, Vilfredo Pareto, who said: "Social sci­entists can derive no advantage from words. The words can, however,

137

F. R. Ankersmit and J. J. A. Mooij (eds), Knowledge and Language,Volume 1II, Metaphor and Knowledge, 137-142.© 1993 KIIlWer Academic Publishers.

138 JAN PEN

incur great harm, whether because the sentiments that words arouseor because the existence of a word may lead one astray as to thereality of a thing it is supposed to represent".lNow dreams are there to be disturbed. Disturbance number one:

voices. When the economist wakes up he discovers that out there arepeople who do quite a bit of talking. They are arguing. They areusing rhetoric. They are using metaphors. They are trying to convinceothers. Now this observation is not always upsetting to the econo­mists: some of these people are just having a conversation, which isa leisure activity, some are poets, and some are even arguing pro­fessionally - the legal profession for instance. Lawyers are pleadinga case and this is none of the economist's business - that is what hethinks in his dream. More disturbing are the politicians and the jour­nalists. They talk, they use language, and when the economist, in thereal world, listens to their argument he discovers that the politiciansand the journalists are arguing about the same issues that he isdealing with. This is confusing, but the economist has a way out: thepoliticians and the journalists are operating in a political arena, theirbusiness is partly irrational and it is the economist's task to instillrationality in the discourse. Rationality means Cartesian logic,algebra, numerical results of econometric research, cost-benefitanalysis. This is the economist's scientific mission among the bar­barian's and the representatives of interest groups. Yet this mis­sionary work means that the dream is disturbed, because theeconomist has to use words; and a good many words are needed toget his message across. He will even have to use rhetoric. PaulSamuelson had a column in Newsweek and Milton Friedman writesbooks like Free to Choose (1979) - strong narratives, heavily loadedwith metaphor. Now the pure economist of the anti-language schoolmay accept grudgingly that these activities exist but he may stillcling to the silence of his dream - which is about science, not aboutpolitics.But then he discovers that another disturbing idea creeps in: his

own views - scientific views - contain a political element. Of coursethe profession has always known this but the information has beensuccessfully stowed away in a corner of the literature, where method­ology is discussed and this is not a popular corner of economics.Gunnar Myrdal wrote Das politische Element in der national­okonomischen Doktrinbildung (1931); its main thesis is that theeconomic observer is relatively free to choose his own theory,

ECONOMICS AND LANGUAGE 139

according to his ideological preferences. This thesis was acceptedin principle and forgotten in practice. It implies that economic the­orists have to argue among themselves, as if they were politicians!The counterargument is that economic science itself limits thefreedom of the economist, because there exists a continuous con­vergence towards greater objectivity. The technical, quantitativeefforts of the economic experts contribute to a greater consensus.This view was generally accepted in the fifties and the sixties. Therewas a great synthesis going on, absorbing classical, keynesian andmonetarist theory. The mainstream was widening. Progress was data­driven and only a few mud puddles (marxism, institutionalism) werestagnating on the side. There was hardly a need for great debates ­there was an urgent need for research, for better models and morestatistics. Language is unavoidable but should be kept at a minimum.It should be as clean and technical as possible.Unfortunately, a new disturbance was around the corner. It had

political origins in the real world - oil shock, stagflation, unem­ployment, imbalances in international trade, floating rates ofexchange, lack of political control - but it led to new discussionsamong scientific economists. The idea of a natural convergencetowards scientific consensus was discredited. A new school of youngeconomists with high technical expertise mounted an attack on theneo-classical-neo-keynesian synthesis. This was the rational expec­tations school, holding that the keynesians had it all wrong, andmoreover that economic policy of almost any kind is doomed to fail.At the same time another school, the post-keynesians, came intobeing, less high ranking in mathematical one-upmanship but with acertain political influence. One consequence of these new debateswas that economists became aware of the fact that schools are hereto stay. They tell different stories. It follows that discussion,argument, discourse, whatever you call it, have a place withineconomic science and not only at the borders where economists talkto strangers. A second consequence is that the scientific economistcannot do without rhetoric. Even when nobody listens you have totell your story well to make it credible. To yourself. This awarenessis the end of the wordless dream.Once the economist is awake he discovers that, in the real world,

the proponents of the rhetoric of economics have never been absent.I mention Benjamin Ward, whose book was called What's Wrongwith Economics (1972). His answer was that economists try to avoid

140 JAN PEN

the story and that this makes their reasoning dry and barren andunhumanistic. Ward's success was very limited. His position wasweakened by the fact that the great economists had always told greatstories. From Adam Smith to David Ricardo to Karl Marx to JohnMaynard Keynes strong narratives, full of metaphors, were amongus. Therefore Donald McCloskey's view was much more convincing.In a famous article in the prestigious Journal ofEconomic Literature(McCloskey 1983) he told the profession that all economics is deeplyrhetorical, that economists deny the rhetorical character of theirwork, and that they should not do this.2 Instead the profession shouldwelcome rhetorics, because it makes for better speaking, betterwriting, better public relations with other disciplines and a better dis­cussion among themselves.The reactions to McCloskey's view were mixed. The profession

was not overflowing with joy. Some economists put up a defenseagainst what they saw as an accusation: I never use metaphorsbecause I define my concepts carefully; I never use rhetoric, at leastnot in my serious work; I am aware of the fact that others use rhetor­ical tricks and I do not deny that I do it sometimes myself but thepractice should be avoided; a little bit of rhetoric and a fewmetaphors may be helpful in discussions and we should certainlypractice the art if necessary but we should beware of its impact onour scientific work. Obviously, McCloskey's normative positionlends itself to heated disagreement, but according to me his diag­nosis of the economist's literary style can hardly be disagreed with.As I see it, his article and the subsequent discussion have stimu­lated a new awareness and a new sensitivity with regard to theliterary forms of economic analysis.Speaking for myself, the game of uncovering economic metaphors

is easy. The powerful images of economics are visual metaphors: theinvisible hand, the depression, the natural rate of growth, the equi­librium of the market. The great theories can be classified accordingto the type of metaphor they use, and in the main there are three ofthem (with many subdivisions and auxiliary constructions).The neo-classical analysis depicts society as orderly and self­

regulating. The metaphors are referring to clocks, steam engines,mechanical regulators, or to nature: the natural order, the natural rateof growth (lends itself to precise mathematical description), thenatural rate of unemployment (should not be tampered with), the

ECONOMICS AND LANGUAGE 141

natural outcome of the market progress according to Darwinian prin­ciples. The keynesian analysis points to disorder, to depression andinflation, and uses the metaphors of streams that sometimes flowslowly and orderly but are often in turmoil. These flows may stagnateand even freeze (frozen unemployment situation). When keynesiansare referring to nature they speak of the depression - the economyis like the weather or the mind; vulnerable. In post-keynesian theorythere is the butterfly effect which is the metaphor of chaos. The worldis pretty unstable and unpredictable. Marxists like to describe societyas a battlefield, and everything that happens under the conditions oflate capitalism - government, newspapers, sports, culture - is seenin this light. This Marxist world is full of noise. There are voiceseverywhere, language is used to conceal and deepen the exploita­tion of the masses. The three great stories of economics are charac­terized by the consistent and cumulative use of metaphors - but onlyso if the narrator knows his craft. If he does not, metaphors maybecome confused and contradictory. This happened to keynesiantheory, where the initial idea of flows in turmoil (it has been calledthe hydraulic variant) was supplemented by the false-signal metaphorof disequilibrium analysis; the result was an unfortunate discussionat cross purposes between different types of keynesians. Metaphorsmay confuse the mind.But this is of course not the main criticism of metaphors in

science. As I understand Mary Hesse, the criticism is that they nibbleat truth. Now that would not be my first worry. When somebody tellsan economist that the Real National Income, or GNP (gross nationalproduct) is a metaphor, he knows already that the "truth" of thisfigure is a very feeble kind of truth. There exists a procedure tocalculate the GNP, which is practiced in all civilized countries andis recommended by the United Nations, but the outcome of thiscalculation is an index number, which does not correspond to anyreal phenomenon in the outer world. It represents a "basket of goods"but when this "basket" is changing in composition, which it is doingall the time, the correspondence between figure and reality becomestenuous. Every economist knows this, so "nibbling at truth" is nota very serious accusation. My own problem with the metaphor andthe strong narrative is that the economist may be carried away withit. Rhetoric gives the economist's mind great wings, to fly high overthe daily business of pedestrians, and to perceive the underlying

142 JAN PEN

structures that remain unseen by others - but at the same timerhetoric may become a net in which the eagle is entangled. It robsthe mind of its freedom.My conclusion is partly simple: economists are great users of

language, some of them tell great stories. It is worthwhile to studythis phenomenon. But the normative conclusion is not simple. Storiesmay be misleading and yet difficult to falsify. The profession shouldbeware - but beware of what? Of false arguments, of course. Ofexcessively seductive metaphors but which metaphors are exces­sively seductive and which are helpful and creative? The real worldis more complicated than the economist's dream and so is science.

NOTES

1 Pareto (1935).2 McCloskey (1983).

BIBLIOGRAPHY

McCloskey, Donald N.: 1983, ' The Rhetoric of Economics', The Journal ofEconomic Literature 21, 481-517.

Pareto, Vilfredo: 1935, The Mind and Society (Trattato di sociologia generale,Florence, 1916), A. Bongiorno and A. Livingston (trs.), 4 vols., Harcourt &Brace, New York.

MARIA LUISA BARBERA

METAPHOR IN 19TH-CENTURY MEDICINE

In the political thought of every age, references to the body, espe­cially to the human body, are essential. All societies, even religiousones, have always been regarded as a body, a person, at least sinceMenenius Agrippa spoke his apologue in front of the Romanpopulace to convince them to give up their revolt. Since then, themetaphor of the body has ruled undisputedly over the many othermetaphors in political thought, so that its history could be studiedin terms of the history of all the corporeal images which appear init.The extreme tangibility, familiarity and concreteness suggested by

the image of the body can at first sight explain its success as a socialmetaphor (and as a metaphor in general). In fact, if it is generallytrue that by using metaphors we tend to structure less concreteconcepts in terms of more concrete ones, then the body, in view ofits comfortable daily presence, should be the metaphor "par excel­lence". And yet, although solid and concrete, it seems to have oneparticular aspect which makes it unworthy of being called ametaphor: instead of structuring the context in which it is containedin its conceptual terms, it always structures itself in terms of thecontext. This peculiar characteristic of the metaphor of the body isnot something negligible: it is, rather than concreteness, which is thebasis of the success of the body metaphor in politics. In fact, con­creteness is not, as we might think, at the basis of the metaphoricaluse of the body, but on the contrary, it is a consequence of that use.In other words: each person only regards his own body as concrete,and not just any body. His own body, that is to say, a body consid­ered in terms of dominant categories in his own cultural context, inharmony, not in contrast, with them. So, in periods of social changes,i.e., periods of variation and subversion of imaginative categories,the problem of constructing a new corporeal image, correspondingto a new society, to which this new society can refer to as a metaphorof itself, becomes relevant.At the end of the 18th century the idea of the importance of

medicine, of its knowledge and methods, for the whole of society,

143

F. R. Ankersmit and J. J. A. Mooij (eds), Knowledge and Language,Volume Ill, Metaphor and Knowledge, 143-154.© 1993 Kluwer Academic Publishers.

144 MARIA LUISA BARBERA

gradually pervades politics with almost obsessive intensity. TheFrench physician, politician and philosopher Cabanis maintainedthat, even if medicine could not cure any disease, "it is worthy ofgreat attention all the same, as it is the basis of any sound rationalphilosophy". According to him, medicine is so deserving because itis able to find the "foundation of the rights and duties of man in theeternal laws of Nature." These ideas, found in the Preface of Dudegre de certitude de la Medecine,l are really the expression of theauthor's personal certainties: that is, the structure of the body incar­nates the Declaration ofRights of revolutionary France. In this workmedical practice is clearly defined as a magistrature and physicianand lawmaker are put on the same level, as "in their eyes .. , every­body is equal".2 This means that the political task of the physicianis no less important than the task of the lawmaker, as the formercan offer the latter the corporeal image he needs to refer to as ametaphor of the revolutionary equality he is going to create. Perhapsit is no coincidence that the French revolutionary government, at acertain point, took the name of Committee for Public Welfare; justthe same expression used by Cabanis, when he wished physiciansto become "guardians of morality, as they are in charge of publicwelfare" (salut).3 Just as it is no coincidence, perhaps, that in thetexts of the Constitutions that followed one another in France after1789, the body is so often named. In the 1789 Declaration ofRightswe can read that: "The representatives of the French People ... havedecided to state the natural, sacred and inalienable rights of man ina solemn declaration, so as to constantly remind all the members ofthe social body of their rights and duties". This body, the productof a free association, has no definite structure yet, but it alreadyclearly serves to deny and substitute the other body (the Guild, i.e.,corporatio from Latin corpus), a symbol and a synthesis of a realitythat is both bound and binding expressly mentioned later in art. 3:"The principle of every sovereignty is essentially held in the Nation.Neither bodies nor individuals can exert any authority that does notexpressly come from it".In any case the link between medicine and politics is very close

in this period, as is confirmed by the trend of medicine to structureitself in systems. Each medical system (such as those of Brown inScotland, Rasori in Italy, Broussais in France, Hahnemann inGermany) has its own language, in some case even a jargon, andeach of its concepts is only meaningful within itself. In each systemthe image of the body is exclusive and self-consistent, just like the

METAPHOR IN 19TH-CENTURY MEDICINE 145

model of society which refers to it. In the second half of the 19thcentury many of these systems disappeared completely; however,their collapse did not mean renunciation of the exclusivistic per­spective - inevitable and typical prerogative of a society that is notyet stable. Medicine did not get over its particularism and becomeunified by absorbing heterogeneous elements on the basis of theirmutual compatibility or of experimental verification, as it happenedfor other sciences in that same period, but simply by making thosesystems (and worlds), different from the one system (and one world)that could impose its powerful superiority, not only outdated but evenlogically unthinkable and morally unacceptable. Cabanis, forexample, gets rid of all previous or simply different medical theoriesby accusing them of being expressions of "brigandage", "deraison","fureur". Different medical opinions cannot (and should not) be dis­cussed, they should be "unmasked"; physicians who disagree withofficial, i.e., state medicine, are called simply "charlatans". Thereforeit is the duty of the new physician to denounce rather than toconfute.4

The close link between medicine and politics in this period alsoshows up in the use, sometimes even the misuse, of politicalmetaphors in medicine and, obviously, of biologico-medicalmetaphors in political thought. Of this latter group we will onlyanalyze one significant example: the organism. For modern polit­ical theorists (Hobbes, Locke and also Rousseau) society is not aprimary event, but is the result of a deal or contract. In any case itis the product of the artificial overtaking of a primary dispersion.The single person, isolated by terror or by a vagrant independence,is first its foundation, then its constituting element. The corporealimage this society refers to as its metaphor assumes the existenceof anatomy and - at the same time - contributes to its justificationas a rational science. Like society, biological life is born from a frag­mented reality and goes on including it afterwards as well. This par­ticular organism, whose parts ideally pre-exist, has now become sofamiliar to us that we could hardly imagine another. Incidentally,we owe the very conceivability of medical specialization to that, aswell as the idea of the possibility of organ transplantations; althoughthere was a time - the time we are examining - when this very ideawas so far from being natural and self-evident that some propagatedand defended it as a faith, while others denounced it as monstrousand unnatural.At the beginning of the 19th century, the romantic writer Mary

146 MARIA LUISA BARBERA

Shelley describes the creation of a monster by her famous character,Frankenstein, in this way: "After having spent some months in suc­cessfully collecting and arranging my materials, I began . .. Thedissecting room and the slaughterhouse furnished many of my mate­rials".5 Mary's monster is, properly considered, a biological metaphorof society as expressed by the Enlightenment (wasn't Hobbes'sLeviathan a monster too?) which the author opposed, as did manyother intellectuals of the same period. Frankenstein builds themonster's body - but already the French National Convention, 1793,spoke of Constituted, i.e., built, Bodies - not a mechanicalautomaton. Frankenstein uses pieces belonging to other living beings,subjecting them to the physiological laws that he knows. He doesnot use any devices expressly built with the purpose of being assem­bled, and so only coming temporally and practically before thewhole. If it were an automaton, it would not be a monster, becauseprimary dispersion would only be a technical necessity, not a tragedy,a violent annihilation of any pre-existing links, followed by a forcedrefunctionalisation. But above all Frankenstein's monster, if it werean automaton, would not be a wretched creature. It is exactly itsbeing so wretched (not its bestiality or its lack of reason) that is,according to the author, the main characteristic of a being which,by its origin, represents an atomized society and, by its existence,the psychological result of such atomization.I have chosen Frankenstein, a novel, not a treatise, as an example

of the point of view that considers the organism as something orig­inary, because the difference between the two meanings of thismetaphor shows up more distinctly than in any philosophical descrip­tion. But in this anti-Enlightenment controversy Mary Shelley is notalone: according to such philosophers as Saint-Simon and Comte,too, in spite of their differences in thought, organic is the directopposite of "atomistic", "individual", "revolutionary", "critical".Organism is together symbol and synthesis of their romantic pro­gramme. But precisely because their organism is a political pro­gramme, a social project, it is closely connected with the artificialitythat they wanted to dispel and that they considered the foundationof the corruption of the modern world. Paradoxically, just like theiradversaries, according to whom organism is the final result of aprocess, their organism is the final result of a process of "re-orga­nization" of an irremediably dispersed reality.6 Even if we do notchoose to regard, as Marx did, organization as synonymous withthe "division of labor", and so as the cause itself of that dismem-

METAPHOR IN 19TH-CENTURY MEDICINE 147

berment to which it claims to be an antidote,7 we should admitthat Comte's organic age is not the first mythical age at all, but thelast.s

The use of biological and medical metaphors in politics corre­sponds to the use of political metaphors in medicine. Of course, wecan find metaphors taken from other fields, such as the metaphor offorce (like vital force), which came to medicine from Newtonianphysics, but whose anthropological origin, perhaps in spite of itself,clearly showed in physics as well. Another non-political metaphoris very frequent in the medical textbooks of the first half of the 19thcentury: that of tissue. Invented and circulated by the famous physi­cian Bichat,9 it allowed one to interpret the body according to theexperience of an ancient technique, made richer and more produc­tive in those very years by the very fortunate conjunction withmechanics, while, at the same time it advanced the idea that knowl­edge of the body could be technical rather than theoretical. Tissuewas bound to supplant force as a privileged metaphor in the inter­pretation of the body. In any case in this period any debate for oragainst dynamics in medicine seems just a coded discourse aboutpower, meaning that the choice of an ideal of health as a balance offorces or as an integrity of tissues involves the priority of politicalpower or of wealth in the control of society.Furthermore, in every medical theory in this period, "Nature" is

constantly dispossessed by the therapeutical doctrine of the adver­saries and restored by the theory's own doctrine; vascular andnervous systems compete for primacy - indeed, two opposite systemsderive from the primacy of one or the other, i.e., "humorism" and"solidism". According to the French physician Broussais, thestomach is the King of the living economy, which is subjected to it,1Obut some others maintain that "intimate relationships, conspiracies,consents, sympathies, all demonstrate the constitutional monarchyof the animal machine")l However, the pressure of politics onmedicine is perceived and criticized by contemporaries themselves:"If my voice could be effective, I would ask to leave any questionsof primacy and independence to politics, instead of increasing thenumber of our neologisms . .. The scarce propriety of our languagehas produced such daringly metaphorical technology that has neverbeen greater before",12 But in vain: a disease is still nearly alwaysa disorder or discordance (health is, obviously, order and harmony);but, if pathological disorder is defined as a "riotous insurgency ofdependent parts", 13 the physician's task will not be that of reordering

148 MARIA LUISA BARBERA

a puzzle or returning a musical instrument, but a real policeman'sduty. If health is a question of public order, those responsible for itsdisturbance have to be detected first, then rendered harmless. Perhapsit could be maintained that, without this politico-juridical view ofthe problem, medicine would hardly have found the way leadingfrom the symptom (which is the expression of a generalized illness)to the exclusive consideration of the lesion (involving a patholog­ical localization), and perhaps it would never have asserted themicrobic origin of many diseases.But disease means transgression - although criminal anthropology

tells us that, on the contrary, transgression means disease14 - pre­cisely because health means obedience to the law. The metaphor oflaw is perhaps the most important political metaphor that we meetin all natural science, not only in medicine. It may be that the ideaof law was introduced in medicine through chemistry or physics, orthat it was directly taken from the idea of social organism, aboutwhich we have ¥ready spoken. At any rate the introduction of thenotion of law i~'medicine, that is, the birth of medical physiology,(moulded, on the basis of a paradoxically pre-existing "social phys­iology"), was not an easy event. It was not a painless event either,since it nearly made the existence of a real medical science impos­sible by making a scientific pathology impossible. For in otherwords, if science means having knowledge of laws, somethingmeaning transgression, deviation, lawlessness, cannot be consideredthe object of scientific research, and illness itself becomes impos­sible (physiological law has absolute power - another politicalmetaphor! - and is therefore capable of ruling the organism evenduring an illness, which is then not a real illness completely differentfrom health). For a few years medicine had the same meaning asphysiology and so abandoned any therapeutical pretension: if I don'tknow pathological laws, but only physiological ones, my interven­tion in the pathological disorder may be either blind or useless.The politico-juridical direction suggested to him by its metaphors

obliges the 19th century physician to balance himself, with great dif­ficulty, between the two opposite characters of policeman and spec­tator.At any rate, in the period we are examining, there are other, even

more significant metaphors. Strictly speaking, they are not reallymetaphors, at least not in form. They are, without any doubt, medico­biological expressions which, after having been used for a long time

METAPHOR IN 19TH-CENTURY MEDICINE 149

as metaphors in political thought, still retain its particular, exclu­sive and totalizing attitude. The medico-biological literature of thoseyears shows such definitions of life as reaction, consensus, associ­ation, obedience to laws, which are evidently political metaphors.But one also reads that life is nutrition, or circulation, or organiza­tion: terms which have the same totalizing aspiration (as if life couldbe defined through one of these characteristics only) as the earlierones, but which are different because they are strictly medical terms.In his Lehre der Nahrungsmittel, 1. Moleschott, by maintaining theabsolute primacy of nutrition in life, states that: "Life consists of anexchange of matter . .. of an eternal restless circulation" .15 Theextension of the nutritional pattern to every biological process isthe consequence of an over-evaluation of the concepts of exchangeto which nutrition was reduced. In other words, the concept of nutri­tion as exchange can only be thought of if we deny the existence ofany real difference between the living and the inanimate world, ormore generally, if we state that there is no real difference at all andwe assume the absolute primacy of equivalence: "As commerce isthe soul of any human relationship", writes Moleschott in DerKreislauf der Lebens, "so the eternal circulation of matter is thesoul of the world",!6 Moleschott's physiology, by generalizing thecirculatory pattern typical of humoristic medicine is a metaphor oftrade and its main elements, circulation and exchange.The theory of nutrition as exchange was opposed to that of nutri­

tion as assimilation (which in a certain sense it replaced). That theorystressed, by admitting the basic estrangement between the organicand non-organic worlds, the work of assimilation, that is, the workof translating the non-living element into terms of the living one,by means of the digestive apparatus. But the work of the digestivesystem was the pattern ofwork in general, as it was not only capableof changing its object's shape, but also of altering its substance bycreating new properties which were originally absent. In this waysome economists, such as the French economist J. B. Say, justifiedthe existence ofprivate property by showing it to be a result of workdone by the individual, in so far as it is capable of adding theproperty of utility to the other properties of matter'!? Perhaps thereis no need to stress the metaphorical nature of the properties ofmatter (in a sense matter, too, is an owner!), but it could be inter­esting to quote the development of this same theme by homeopathicmedicine. According to its main theorist, Samuel Hahnemann, it is

150 MARIA LUISA BARBERA

possible through the grinding and dynamization of substances togive them properties which do not exist in nature. Only these aretrue properties; chemical and physical, that is, natural, properties,are preferably to be called characters and, according to thehomoeopathists, have no pathogenetic and therefore no therapeuticaleffect. The new therapy assumes elaboration as its own conditionby making human (and mechanical) work capable of extendingNature's possibilities or, which is the same, by making Nature ahuman product. 18

Before it was defeated by its opposite metaphor, nutrition asassimilation had been a model for the solution of social conflicts:the Italian philosopher Gioberti, in his Prolegomeni, written before1848, stated that the main, rather difficult task of the "middle class"was that of "digesting" the "populace" by a process of assimilationof social extremes into itself.i9 The kind of society implied by thisimage of nutrition was a tendentially closed society, so that anyexternal factor could only delay or stimulate internal processes,which remained indigenous and somewhat preordained accordingto a strictly individual programme (this was also valid for the nation,which is the collective individual).This society was metaphorized by the theory of preformation,

according to which the forms of the future development are antici­pated in the germ, and cannot be changed or suspended. This imageof the body, both in medicine and in politics, obviously only accom­modates general diseases, as the cure is general in any case. In fact,a disease which is localized or particular is unthinkable within animage of the body (even a social one) which involves the possibilityof sympathies, that is, transversal relationships between organs, inde­pendent of their specific function and also of the relationship betweeneach one of them and the whole. The concept of sympathy itself orig­inally belonged to medicine, but it was translated (and confined) asa metaphor in the moral and political thought of the Scottish School.It then returned, this time as a metaphor, in romantic medicine andwe can also find it in authors such as the above-mentioned Broussais,friend and disciple of A. Comte, nearer in thought to some physiol­ogist physicians, but evidently still worried about finding a place forintermediate forms of aggregation among men in the social body.The same opinion was expressed by Tocqueville when he complainedabout the disappearance of intermediate or 'secondary' social bodies.This disappearance made people free, Le., capable of conducting

METAPHOR IN 19TH·CENTURY MEDICINE 151

their lives in an unpreordained and apparently almost unforcedfashion, but allowed them only occasional relationships with theirpeers. Consequently, the individual will be without defense againstcentral authority and, hence, without individual freedom. 21 In short,the refusal to reduce the complicated picture of organic (social andbiological) processes to the single measure of functionality, corre­sponds to these authors' refusal to reduce all relationships amongindividuals to the single measure of exchange.Although the German physiologist R. Virchow, father of cellu­

larpathology, shared many opinions with the positivist physiologists,he always disagreed with them about vitalism, whose validity he alsomaintained when it had become so discredited that it was consideredto be outside the bounds of science. It would be interesting to estab­lish, in this expulsion, the contribution of new developments inchemistry, which had annihilated the clear dividing lines betweenthe organic and non-organic and the contribution of the inherentincompatibility between vitalism and the idea of equivalence, whichis the basis of liberalism as an exchange-philosophy. In fact, ifvitalism means admitting something irreducible to the mutual con­vertibility of things, the actual problem is about exchange rather thanabout life.According to Virchow, nutrition as exchange could not be made

universal, as it was only connected with the idea of reproduction,so with the idea of the stillness of the whole. Virchow understoodand denounced the semblance of transformation which the exchangeinvolves, as exchange is only a surrogate for transformation anddevelopment: in fact, as the movement of exchange, if exchange isreally exchange, takes place between equivalents, it leaves theprevious situation practically unchanged and moreover has no limits(a further exchange is always possible). Virchow's opinion abouttransformation and development was quite different from that of thepartisans of absolute exchange, when he wrote that: "Any exchangeis essential to life which would cease to be if exchanges had nolimits. Such limits assume the existence of particular moderating andregulating devices", which he identified with the membrane andnucleus of the cell. The cell is the keeper of the "function" whichis capable of undergoing changes.22 Moreover, he thought that thecell was the first and foremost element of organic and social life.While solidists and humorists, as we read in Die Kritiker derCellular-Pathologie, consider the body "a despotic and oligarchic

152 MARIA LUISA BARBERA

unity", "cellulism . .. regards every part of the body as an elementnecessarily connected with the others, every one dependent on theother and linked by communal need" and the body itself as a "com­munity".23 Starting from the cell, Virchow came to refuse theomnipotence of exchange - or, perhaps, he arrived at the cell becauseof a theoretico-political need to counter-balance the indiscriminatefluidity of the economic competition.

In any case, the "federal disposition" of the organism, whose"demonstrative" and scientific nature he maintains by disagreeingwith the "axiomatic" nature of "unity" at all costs,24 contributed anew aspect of 19th-century pathology, which before Virchow hadbeen limited to a mannered vitalism, capable of only asserting life,or a mechanical physiologism, incapable of understanding patho­logical (and generally biological) differences, unless in a negativesense.What I have been saying up till now suggests the importance of

the contribution of political metaphors to the creation of new medicalconcepts (such as that of the cell). Metaphors (political or otherwise)have played an important part in re-establishing some traditionalterms of medical language by giving them new meaning.25 Bycontrast, the corporeal metaphors used in political thought do notseem to have had this kind of heuristic capacity, as they limited them­selves to a tautological and persuasive, that is, rhetorical, function.This is generally true, but we should not forget that terms can acquireotherwise unthinkable meanings, precisely owing to their tautolog­ical and merely rhetorical sojourn as metaphors in other contexts.Today the metaphor of the body has been transferred to ecology,

after leaving society to its own devices. The body is now an imageof the Earth itself as a sick body. I don't know if this new metaphorhas a rhetorical or heuristic function. I only know that the sick bodyof the Earth is, all the same, a metaphor of our most serious problem.

NOTES

1 Cabanis (1798), pp. 404 ff.2 Ibid., pp. 522n, 527.3 Ibid., p. 526.4 Ibid., pp. 519, 551.5 See Shelley (1818), pp. 314-315.6 Saint Simon (1814), pp. 153-248. See also Comte (1854), pp. 1-221.7 Marx (1844), pp. 429-434; and idem (1867-1894), pp. 272,297.

METAPHOR IN 19TH-CENTURY MEDICINE 153

8 Comte (1830), pp. 1-180.9 Bichat (1800), pp. 15, 138, and passim.10 Fodera (1821), p. 20; Broussais (1828), pp. 336, 361-364.11 Turchetti (1858), p. 21.12 Baroni (1853), p. 56.13 Turchetti (1858), p. 30.14 See Lombroso (1876).15 See Moleschott (1853); in the Italian translation pp. 60 and 42.16 See Moleschott (1852); in the Italian translation p. 38.17 Say (1815), pp. 9,44,89; see also idem (1828), Vol. I, p. 133; Vol. III, p. 197.18 Hahnemann (1828), pp. 1-16; Simon (1836), pp. 59,235.19 Gioberti (1847), p. 19,67.20 Broussais (1828), p. 290. On this work see Comte (1828).21 See Tocqueville (1856), Vol. II, 1.22 Virchow (1856), pp. 3-55.23 Virchow (1860), pp. 1-14.24 Ibid.25 On this subject see Canguilhem (1968), pp. 61-74.

BIBLIOGRAPHY

Baroni, Agostino: 1853, Sulla restaurazione ippocratica proposta da G. Franceschie sull' empirismo razionale, G. Lana, Fano.

Bichat, Marie Fran~ois Xavier: 1800, Traite des membranes, Richard, Paris.Broussais, Fran~ois Joseph Victor: 1828, De ['irritation et de la folie, LibrairiePolymatique, Paris-Bruxelles.

Cabanis, Pierre Jean Georges: 1798, 'Du degre de certitude de la medecine', inPJ.G. Cabanis, Oeuvres completes de Cabanis, pp. 404ff.

Cabanis, Pierre Jean Georges: 1823, Oeuvres completes de Cabanis, Vol. I, F.Didot, Paris.

Canguilhem, Georges: 1968, 'La philosophie biologique d'Auguste Comte et soninfluence en France au XIXe siecle', in Etudes d'historie et de philosophie dessciences, Vrin, Paris, pp. 61-74.

Comte, Auguste: 1822, 'Plan des travaux scientifiques necessaires pour reorgan­iser la societe', in A. Comte, Systeme de politique positive, pp. 47-136.

Comte, Auguste: 1828, 'Examen du traite de Broussais sur l'irritation et la folie', inA. Comte Systeme de Politique positive, pp. 216-228.

Comte Auguste: 1830, 'Convergence progressive des principales evolutionsspontanees de la societe modeme vers l'organisation finale d'un regime rationnelet pacifique'. in A. Comte, Cours de philosophie positive, Vol. VI, pp. 1-180.

Comte, Auguste: 1830b, Cours de philosophie positive, Everart, Paris.Comte, Auguste: 1854, Systeme de politique positive, Appendiece general L.Mathias, Paris.

Fodera, Michele: 1821, Histoire de quelques doctrines medicales comparees acelledu docteur Broussais, J.B. Bailliere, Paris.

Gioberti, Vincenzo: 1847, Prolegomeni, in Opere di V. Gioberti I, Bocca, Milano.

154 MARIA LUISA BARBERA

Hahnemann, Samuel: 1828, Die chronischen Krankheiten II, Arnold, Dresden­Leipzig.

Lombroso, Cesare: 1876, L'uomo delinquente, Bocca, Milano.Marx, Karl: 1844, 'Oekonomisch-philosophische Manuskripte', in Marx andEngels, Gesamte Ausgabe, Abt. I, Bd. 2, pp. 429-434.

Marx, Karl: 1867, 'Theilung der Arbeit und Manifaktur' , in Das Kapital, Marx andEngels, Gesamte Ausgabe, Abt. n, Bd. 5, pp. 272, 297.

Moleschott, Jacob: 1852, Der Kreislaufdes Lebens, (Italian transl.) Brigola, Milano.Moleschott, Jacob: 1853, Lehre der Nahrungsmittel, (Italian transl.) Treves, Milano.Saint-Simon, Claude Henri de: 1814, 'Reorganisation de la societe europeenne',in Oeuvres de Saint-Simon et d'Enfantin XV, Paris, pp. 153-248.

Saint-Simon, Claude Henri de: 1867-76, Oeuvres de Saint-Simon et d' Enfantin,Anthropos, Paris (anast. 1976).

Say, Jean Baptiste: 1815, Catechisme d'economie politique, Bossange pere, Paris­Londres.

Say, Jean Baptiste: 1828, Cours complet d'economie politique pratique I et III,De Casimir, Paris.

Shelley, Mary: 1818, Frankenstein, 'Three Gothic Novels', Penguin Books,London-N.Y. 1982.

Simon, Leon: 1836, Lerons de medecine homoeopathique, De Indenne, Bruxelles.Tocqueville, Alexis de: 1856, L'Ancien Regime et la Revolution, in Oeuvres com­

pletes II, 1, Gallimard, Paris.Turchetti, Odoardo: 1858, Colpo d'occhio sullo stato attuale medica patologia,manuscript, Milano.

Virchow, Rudolf: 1856, 'Alter und neuer Vitalismus', Archiv fiir pathologischeAnatomie und Physiologie und fUr klinische Medizin IX, 3-55.

Virchow, Rudolf: 1860, 'Die Kritiker der Cellularpathologie', Archiv fiir patholo­gische Anatomie und Physiologie undfiir klinische Medizin XVIII, 1-14.

FRANK R. ANKERSMIT

METAPHOR IN POLITICAL THEORY

I. INTRODUCTION

Perhaps no field is richer in metaphor than political theory. Plato,who is often seen as the first political philosopher in history, incor­porated numerous metaphors in his Republic and in his Laws. Noless metaphorical are the books written during the Golden Age ofWestern political philosophy. Hobbes, Locke, Rousseau and so manyothers all founded the political order on an original contract orcovenant entered into by its members. Often they admitted that thiscontract was a mere historical fiction; but that was of no consequenceas long as the ideas of the original contract could function as ametaphor in terms of which the legality of the existing political ordercould be analyzed. Especially in the political writings of that fierceopponent of metaphor, Thomas Hobbes,l metaphors abound. Forwhat is the notion of 'the body politic' other than an organicistmetaphor exploited as such throughout Hobbes's work; and is thecommonwealth, the State or that 'artificiall Man' (another metaphor)that forms the subject of Hobbes's book really the monster whichJob and the Psalmist refer to by the name of Leviathan? Coming toour own age, we are struck by the force of Foucault's metaphor of'the capillaries of power' and by the Rawlsian metaphor of 'the veilof ignorance' behind which our conception of the just society isformed. Most important of all, in these cases even a superficialawareness of the function of these political metaphors forces us torecognize that they do not merely touch the surface but form theessence of the political theories in which they are proposed.Metaphor is no mere ornament or didactic device: the convictioncarried by the argument itself depends upon an open or tacit accep­tance of the metaphor in question. If metaphor is eliminated theargument of the political philosopher degenerates into a meaning­less chaos.2 Metaphor is the heart that pumps the lifeblood of polit­ical philosophy.Yet political philosophers have only rarely and only recently beenprepared to recognize the metaphorical nature of their enterprise. In

155

F. R. Ankersmit and J. J. A. Mooij (eds), Knowledge and Language,Volume Ill, Metaphor and Knowledge, 155-202.© 1993 KJuwer Academic Publishers.

156 FRANK R. ANKERSMIT

fact, one of the main goals of this essay will be to clarify why andhow the perspectivism of metaphor has run counter to the univer­salist presumptions of nearly all of political theory. But in all fairnessI must begin by enumerating a few exceptions to the general andtraditional condemnation of metaphor by political philosophers.When William Connolly uses Gallie's notion of the so-called 'essen­tially contested concepts' for his analysis of political concepts like'democracy' or 'the State' - and when he points out that suchconcepts aim at 'describing a situation from the vantage point ofcertain interests, purposes or standards' ,3 Connolly is already quiteclose to recognizing the central role played by political metaphor.For it is only a small step from Gallie's 'essentially contestedconcepts' to metaphor. The 'essentially contested concept' has withmetaphor in common a shift from description or reference to a'seeing as'.Even more explicit was Terence Ball in his recent book on polit­

ical discourse.4 In contrast to Connolly, Ball does not hesitate toemphasize the metaphorical character of Gallie's 'essentially con­tested concepts' if used in a political context. Moreover, with thehelp of some felicitous examples, Ball explains why metaphor reallyis an indispensable instrument for the political philosopher. Politicalphilosophers are required to deal with 'new' situations that havearisen in political reality and it is the semantic 'newness' so oftenclaimed for metaphors that enables them to do so. The 'newness' ofthe political problem has its linguistic counterpart in the 'newness'characteristic of metaphor. Hence the predominant place which Ballawards to the notion of 'conceptual-cum-political' change. Anexample may be helpful. Ball demonstrates the metaphorical usemade by seventeenth-century political philosophers like Hobbes ofthe Galilean notion of mechanical force: political power as well asits causal consequences were modelled on the Galilean axiom ofinertia.s And, to mention a more concrete example, when in thecourse of the eighteenth century political parties began to form indifferent European countries, those who applauded the developmentsought a theoretical justification for this new kind of political asso­ciation. One of their most successful strategies was to replace theold word 'faction' by the word 'party' as term of reference. Ori­ginally the word 'party' was exclusively linked to the sphere ofjuridical ligitation. In other words, the unfavorable connotations ofthe terms 'faction' and 'factionalism' were neutralized by an implied

METAPHOR IN POLITICAL THEORY 157

metaphorical comparison of the political association with the partyas it was already known from the perfectly respectable and acceptedworld of juridicalligitation.6 In this way metaphor enabled politicalphilosophers to account for and to justify an important new polit­ical phenomenon.While Connolly, then, suggests rather than accepts the metaphor­

ical dimension of political theory, while Terence Ball uses the recog­nition of this dimension for a better understanding of politicaldiscourse and its metamorphoses through the centuries, yet a thirdstep has been taken by James Boyd White. For White the recogni­tion of political theory's metaphorical dimension not only gives usa better understanding of politics. According to him this recogni­tion must also have its implications for actual political practice itself.White, a lawyer by profession, illustrates his thesis with the exampleof lawgiving. Statutes, orders and laws are not, at least not primarily,rules for how the citizen should behave, but they 'establish a set oftopics or paradigms or metaphors for how to conceive of socialaction'.7 That is, a law or statute is a paradigm or metaphor in termsof which we try to see the interaction(s) between citizen and citizenor between the citizen and the State. It follows that lawgiving mustnot aim at 'implementing some apriori principles of (distributive)justice but take into account what room it makes for the officer andthe citizen each to say what reasonably can be said from his or herpoint of view about the transaction (... ) that they share'.8 Heremetaphor truly becomes a normative option for lawgiving.Acceptance of this option would, needless to say, have far-reachingconsequences for the aims and nature of lawgiving. It would, aboveall, require the lawgiver to guarantee that anything litigants mightsystematically wish to incorporate into the stories they tell abouttheir case will be reflected in the formulation of the law or statute.And, conversely, what they systematically exclude from theirmetaphorical points of view must also be kept out of the law and itsapplication. I suppose that this would result in a kind of 'democra­tization' of the law and in a legal system that has more in commonwith arbitrage than with existing legal procedures. More specifically,it would mean the end of the predominant role of the State as anindependent agent in legal procedure: for the State, in most cases,embodies a point of view that is completely alien to that of thelitigants.Considering these views of Connolly, Ball and White, we may

158 FRANK R. ANKERSMIT

conclude that some political philosophers now seem inclined toaccept the metaphorical dimension of political theory, but that noconsensus yet exists about what the role of metaphor is or ought tobe. And, opinions differ even more widely about the location ofmetaphor in political theory. In this kind of situation, where so littleseems certain or accepted, a fresh start appears advisable.

II. PLATO'S SHIP OF STATE

To begin with the beginning is not a bad strategy if one wishes tomake a fresh start. So let us go back to Plato's political writingsand find out what metaphors he uses and what their function is inhis work. Like most of his work, Plato's political philosophy is par­ticularly rich in metaphors (Plato, the sworn enemy of rhetoric, isHobbes's most serious rival in combining a prolific use of metaphorswith an utter condemnation of them). Especially when Plato embarkson a discussion of the ideal State, of how the ideal State can berealized and of the role of the statesman, his reliance upon metaphorbegins to determine even the substance of his argument. Tendenciesin his political thought, tendencies that Plato may not even have beenfully aware of himself, stand out if we concentrate on the metaphorshe uses. When Plato discusses the statesman and his duties vis avisthe State the message of his political philosophy is most clearlyexpressed by the metaphors which he uses and the common elementwhich unites them. The metaphors in question are those of thestatesman as a painter (who has to start with a clean canvas), builder,craftsman, cobbler, physician or shipwright. What all thesemetaphors have in common is their suggestion that the statesmanmust be in the possession of technocratic knowledge (techne ) whichmost other people do not have.9 This obviously accords with all theelitist tendencies of Plato's political philosophy. That tendency ismost pronounced in Plato's metaphor of the State as a ship steeredby a captain-statesman. This metaphor is even more suggestive ofthe necessity of technocratic knowledge than the ones I mentioneda moment ago, since, unlike the painter, the cobbler, etc., the captainof a ship really controls the destiny of fellow human beings whoselife and security depend on his seamanship. Since this metaphor ofthe State as a ship seems to be the most felicitous and suggestiveone from Plato's perspective, I shall focus my attention on it not onlyin this section but in the whole of this essay.

METAPHOR IN POLITICAL THEORY 159

Plato's metaphor has irritated many of his modern commentators.Thus Michael Walzer resents the technocratic undertone of themetaphor and the implied disregard for all interaction between therulers and the ruled - in Plato's metaphor the captain does not botherabout the wishes, if any, of his crew or passengers. to In a similarvein Bambrough complained that normally the passengers or theowner of the ship determine its destination, whereas Plato seems toassign this crucial task to the captain as well. The captain's techno­logical knowledge apparently also gives him the right to choose theship's destination and it is characteristic of Plato that he so easilymakes the transition from 'techne' to goal. l1

Surely, the technocratic implications of Plato's metaphor aredamaging to some extent and it is quite true these are at odds withour modern, democratic notion of the State and of its relation withthe citizen. There is, however, another and more important dimen­sion to Plato's metaphor that still harmonizes with how evennowadays we customarily conceive of the State and with the veryessence of how the State has always been perceived in the Westernpolitical tradition. I have in mind here the simple but decisive factthat the metaphor invites us to see the State as an entity with a certainunity that can and ought to be steered in the right direction by a com­petent and responsible statesman,12 I do not think that critics of Platolike Walzer or Bambrough would wish to disagree with this elementof the metaphor. Even more so, as we shall see, an intuition likethis has been behind the greater part of the Western political tradi­tion. If this intuition is abandoned, large parts of that tradition losemuch of their meaning and practical significance.We must be grateful to W. 1. Witteveen for drawing our attention

to this fact. In his introductory essay to a volume that takes Plato'smetaphor as its Leitmotiv Witteveen has given us a competent andconvincing account of how the complexities of modern government,of bureaucracy and of the (in-)effectiveness of government-policycan be fruitfully analyzed from the point of view defined by Plato'smetaphor of the State as a ship.13 Witteveen is clearly implying thatprecisely because Plato's metaphor has always been a tacit assump­tion behind most of Western political theory, practice and science,it provides us with the ideal background for gaining a better per­ception of the problems addressed in these areas. For as soon as wesay that the State has the obligation to solve social and politicalproblems and to strive for a better and juster society, regardless or

160 FRANK R. ANKERSMIT

how we define progress, we are unavoidably presupposing that theState is like a ship that can be set on a certain course. We can alreadydraw an important conclusion at this early stage of our argument. Ifwe nowadays encounter strange and unnerving obstacles on our routeto a better and juster society, this need not mean that our techno­cratic knowledge is insufficient, or that we should now heed thehighly impractical treatises of German philosophers on so-calledpractical philosophy, or that government agencies have become toounwieldy for effective action. Itmay mean, rather, that there is some­thing wrong with the metaphorical background against which politicshas commonly been defined in the Western tradition.With a view to what I wish to do in the remainder of this essay,

it will be necessary to dissect Plato's metaphor. The metaphor iscomplex, as we have seen already, and can only be fruitfully dealtwith if its constituent elements are scrutinized separately. This pro­cedure is not without risk, since metaphors often lose by being takenapart. But I believe that if the following three elements are discernedin Plato's metaphor no serious injustice will be done to it. Themetaphor's first suggestion is that the ship of State ought to followa certain course towards a certain goal- obviously here Plato's idealState. Whether the ideal state is in fact an attainable goal, or whetherit depends on some sudden epiphany of political truth or whether,as was suggested by Julia Annas, the Platonic ideal State functionsrather as a compass for existing States, is a matter of no consequencein this connection. 14 Secondly, the metaphor, by separating thehelmsman from both the ship itself and from its crew, implies adistance between the centre of decision-making and the people forwhom and on behalf of whom these decisions are made. In Plato'sideal State this distance found its expression in the way the phylakeswere separated from the other classes in the State. But usually thenotion of distance is not emphasized by such radical social divisions.All I wish to maintain is that there must be some location, place orperspective separate from either the ship or the State itself whichallows the captain or the statesman to make his decisions about thecourse that the ship will follow. Without this notion of distance, orsome variant of it, the whole metaphor collapses into absurdity.Thirdly, and lastly, like the ship the State must have a certain coher­ence and unity: one cannot navigate a chaotic farrago of boards, ropesand sails. Plato himself used to be quite explicit about the unity ofthe State. He considered the unity of the polis to be its paramount

METAPHOR IN POLITICAL THEORY 161

property15 and even compared his ideal State with the human indi­vidual16 which certainly implies a very high degree of integration.I can now outline what I intend to do in this essay. First I hope

to demonstrate the plausibility of Plato's metaphor insofar as themetaphor entails the three ideas of goal, distance and unity I men­tioned a moment ago. My argument will be, roughly, that politicalaction is metaphorical and, next, that as such political action involvesan acceptance of these three notions. Hence, my argument will besomewhat indirect. Instead of starting with Plato's metaphor in orderto show its appropriateness, I shall begin with political action andtry to deduce Plato's metaphor by showing that the universe of polit­ical action is defined by the three dimensions of goal, distance andunity. I hope to explain next why the metaphorical nature of polit­ical theory and practice has been so often and so consistently ignored.Having reached this stage, my argument will move into a different,apparently even opposite, direction. For I will go on to discuss theconsiderations that seem to militate against Plato's metaphor. Suchconsiderations find their justification in the changing boundaries ofthe political in the modern political order. These changing bound­aries seriously diminish the room left for the three notions of goal,distance and unity that we have associated with Plato's metaphor.This makes us wonder what it might be like to have a politics withoutmetaphor and a State that can no longer be compared to a ship.Japanese politics suggests an answer to this question.I am well aware of attacking no small problem. But I think it is

wrong to avoid big problems as so many contemporary philosophersare apt to do. I suppose that if there is such a thing as philosophicaltruth that makes the efforts of the philosopher worthwhile, truth must,metaphorically speaking, be a fairly sizeable thing that is best foundby investigating big problems. However, both the procedures and thetopology of much modern philosophy appear to be fed by theopposite intuition that truth is rather a particularly elusive thing thatprefers to hide in the most secret and unlikely places. It is as if onewere looking for elephants under a microscope. In my opinion polit­ical philosophy, of all philosophical disciplines, is least likely tosurvive this minimalist approach and I hope that this essay will,among other things, be an argument in support of that conviction.

162 FRANK R. ANKERSMIT

III. THE DEDUCTION OF PLATO'S METAPHOR

'For we all of us', George Eliot wrote in Middlemarch, 'get ourthoughts entangled in metaphors and act fatally on the strength ofthem'.J7 If we wish to understand the metaphorical basis of action,especially of political action, it is best to start with Donald Schon'sso-called 'generative metaphor'. In political action we must distin­guish, Schon writes, between problem setting and problem solving.Schon emphasizes the priority of problem setting over problemsolving. After all, we do not act in a void but always on the basisof a more or less clear idea of the situation we wish to improve:problem solving is thus logically dependent on problem setting.According to Schon, then, problem setting is mainly a matter offinding the appropriate metaphor. 'My point here', he writes, 'is notthat we ought to think metaphorically about social policy problemsbut that we already do think about them in terms of certain perva­sive generative metaphors',18 Schon proposes the term 'generativemetaphor' since the kind of metaphor in question 'generates newperceptions and inventions' .19Two proposals for the renewal of a slum area may illustrate

Schon's view of generative metaphor. According to the first proposalthe slum area is metaphorically seen as a 'diseased body', or as a'cancer' within the larger city and it is not hard to imagine whatpolicy and solution for the slum area follow from this way of settingthe problem. On the other hand, if the slum area is metaphoricallyseen as a 'natural community' whose natural growth has been stifled,a different policy is called for. All this sounds fairly convincing andI should like to add the following comment. Metaphor, as conceivedby Schon, is the 'missing link' between the 'is' and the 'ought'. AsSchon examples make clear, metaphors in this context will logicallyhave the character of proposals: they invite us to see part of (social)reality from a certain point of view which automatically leads to acertain kind of action. In other words, apart from their capacity tocharacterize an existing social situation, metaphors predispose usin favor of a specific line of action and it is because metaphorsembody proposals that they produce this effect. Thus, the goal direct­edness we defined as the first of the three elements in Plato'smetaphor is more or less thrown in our lap by Schon's analysis.Lastly, one can point out that the generative metaphor is neitherdescription nor action itself, so that it truly can be said to be the

METAPHOR IN POLITICAL THEORY 163

'missing link' between the 'is' and the 'ought', while it should notbe identified with either.Now I come to the heart of Schon's analysis. Schon says that the

two metaphors used for the problem setting of the slum area owetheir origin to a story or history that is or can be told of the slumarea. Generative metaphors are condensed histories.2o The firstmetaphor condenses a story of 'blight and renewal' and the other astory of 'the natural community and its deterioration' .21 Differentstories, different metaphors and vice versa. Governments character­istically act when they recognize that something has gone wrong(or is perceived to be wrong) and such recognition necessarilyrequires a quasi-historical assessment of the problem involved. Ona far larger scale than Schon has in mind, this fact is more strik­ingly illustrated by all nineteenth and twentieth century ideologies,which until recently were the most powerful sources of politicalaction. The social and political problems these ideologies attemptedto solve were always defined in terms of a historical interpretationof Western history. There has always been considerable overlapbetween what one might somewhat pompously call 'the logic ofpolitics' and 'the logic of the writing of history'. Efforts to excludehistory from political philosophy are like attempts to have a dinnerwith empty dishes.Schon's thesis of the similarities between metaphor and narrative

as expounded above is quite useful and will enable us to increaseunderstanding of both metaphor and narrative and also, in passing,of the nature of rhetoric.I shall begin with the question of how metaphor can be used for

a better grasp of narrative. If we have the metaphor 'x is a' theremust be a prima facie incompatibility between x and a - without thisincompatibility we would not have a metaphor. I shall naivelyassume that we are right in attributing to metaphor the capacity tooffer a more or less precise description of part of reality. I shall callthis the descriptive meaning of metaphor and in the case of themetaphor 'x is a' the descriptive meaning is expressed by the literalstatement 'x is b'. Most discussions of the semantics of metaphorconcern this difficult and complex translation problem and I there­fore do not want to commit myself to anything more than the rathertrivial and uncontroversial claim I made just now. In the presentcontext it is sufficient to notice that the metaphor 'x is a' and itsdescriptive meaning 'x is b' are not equivalent. This is made clear

164 FRANK R. ANKERSMIT

by negating the metaphorical statement. The metaphorical statementthen turns into a literal statement (e.g.: 'John is not a pig') and willbe perfectly compatible with the original metaphor's descriptivemeaning 'x is b'. Now, no two statements can have an identicalmeaning if the negation of the former is compatible with the latterand that implies that the meaning of the metaphor and its descrip­tive meaning cannot be identical. If we inquire next where and whythe metaphor and its descriptive meaning are not equivalent, we willnaturally turn to the capacity of the metaphor to individuate a pointof view. It is this capacity that is lost when we make the transitionfrom the metaphorical statement to its descriptive meaning and wherethe two are not equivalent.Because of the logical similarity between metaphor and narrative

established by Schon, we can make a similar claim for (historical)narrative. Narrative consists of statements; when taken separatelythese statements describe socio-historical reality and embody thedescriptive meaning of the narrative in question. When takentogether, these statements individuate the metaphorical point of viewfrom which the historian or politician invites us to see part of socio­historical reality.22 It should be observed that the narrative's pointof view arises out of the descriptive meaning of the narrative's state­ments - apparently the metaphorical character of narrative lies closerto descriptive meaning than that of metaphor itself. If metaphor isthe 'missing link' between the 'is' and the •ought' , narrative is the'missing link' between descriptive meaning - in other words, the 'is'- and metaphor.Next, and more importantly, we may ask what the logical simi­

larity between narrative and metaphor can add to our understandingof the latter. Aristotle defined metaphor as follows: 'metaphorconsists in giving the thing a name that belongs to something else' ,23and this is not in conflict with most modern definitions ofmetaphor.24

The customary idea is that the metaphorical statement 'x is a'requires us to look at x from the point of view of a. And this onlymakes sense if both a and x are more or less familiar to speaker andhearer prior to the formulation of the metaphor. This is the case ofSchon's generative metaphor. The slum area is compared to a canceror to a living thing whose growth has been stifled. An even morestriking example is Schon's account of how difficulties with a newtype of paintbrush were successfully solved by seeing the paintbrushas a kind of pump. Now, this state of affairs is not wholly unknown

METAPHOR IN POLITICAL THEORY 165

in (historical) narrative. The historian may see part of the past as arenaissance or as an Indian summer. But most often in historio­graphical metaphor, the a in 'seeing x as a' is not previously givenand the metaphorical image has to be created ad hoc for the purpose.In historiography metaphor is to be identified with the point of viewfrom which the historian proposes to see the past and this point ofview owes both its origin and its identity to the set of statementsused by the historian for writing his narrative. Obviously, in themajority of cases in historiography, metaphor at the same time pre­supposes and develops the point of view from which the past is seen.And because problem setting precedes meaningful political action,the same can be said for political action.But if we give credence to Derrida this already has its precedent

in the writings of Aristotle himself. Commenting on Aristotle's so­called 'heliotrope', the metaphor par excellence,25 Derrida quotesAristotle as follows: 'thus to cast forth corns is called "sowing"(speirein); but to cast forth its flames, as said of the sun [and herewe have the heliotrope EA.] has no special name'. And Derrida addsthe following rhetorical question: 'when has it ever been seen thatthere is the same relation between the sun and its rays as betweensowing and seeds?')6 Just as in the case of narrative metaphor wedo not have here two more or less well-known entities that are seenmetaphorically in the light of each other. For prior to the heliotropethe sun's sowing its flames was still a 'nameless act', to use thewords of Derrida.We observe here an essential difference between the heliotrope

and narrative metaphor on the one hand and models on the other.The heliotrope does not urge us to see, for instance, the pump as amodel for the paintbrush or, generally speaking, a less familiarsystem in terms of a better known system. Here the metaphor createsits own subject. Much is therefore to be said for Derrida's contentionthat we should bestow on the heliotrope - and therefore also on nar­rative metaphor - the honor of being the metaphor of metaphors.2?For the 'newness' of the relation between the metaphors's tenor andvehicle, to use Richards's terminology,28 is here not restricted tothat relation only, but pervades the domains of both tenor and vehiclethemselves. We do not merely have here - as with the model- a newview of an 'old' thing: in the heliotrope and in narrative metaphorwe have a new word for a new thing. But, if all this is true, we mayask, what still differentiates the heliotrope from literal language? For

166 FRANK R. ANKERSMIT

if the heliotrope introduces new words for new things we shall lookin vain for the semantic tension that is so characteristic of allmetaphor. Derrida considers this objection, insofar as he deals withmetaphor and the heliotrope throughout his essay, from the per­spective of 'usure'. As Alan Bass has pointed out the French word'usure' is ambiguous and refers to both use and to a deteriorationthrough use. 29 The suggestion is, clearly, that the heliotrope is theuse of metaphor which tends to obscure its metaphorical character- a tendency that is, of course, even more clearly visible in narra­tive metaphor than in Aristotle's heliotrope itself. In metaphor isoperating a mechanism which aims at hiding the metaphoricity ofmetaphor and in the heliotrope this aim has been realized. What,then, is right and what is wrong in the objection I mentioned amoment ago becomes clear if we contrast the heliotrope and its'usure', as explained just now, with catachresis. With regard to thetension between metaphoricity and literality we can say that thepseudo-literality of catachresis arises from our tendency to forgetabout its (original) metaphoricity, whereas in the case of theheliotrope we have not yet even come to recognize its metaphoricity(once again, narrative metaphor is an even better illustration than theheliotrope). In sum, the heliotrope, this metaphor of metaphors, is,paradoxically, the metaphor that most wishes to make us forget aboutits metaphoricity (think of the objectivist eros that has alwaysinspired both historians and philosophers of history). And theheliotrope tries to achieve this goal of self-occultation by bathingeverything in a clear and even light that makes us forget that thelight must have a source. The heliotrope is therefore all too easilytaken as guaranteeing a literal and realist presentation of the world.As we shall see in the next section, this is why metaphor in polit­ical theory has always presented itself in the guise of the heliotrope.I shall now add a comment on rhetoric and its relation to narra­

tive metaphor. Since rhetoric plays such a prominent part in politics,this excursion seems pardonable. To begin with, we can distinguishbetween two traditional conceptions of rhetoric. Both conceptionsare hinted at in the opening sentences of McCloskey's and Megill'spathbreaking volume on what they call the rhetoric of inquiry. Theirintroduction begins as follows: 'scholarship uses argument andargument uses rhetoric. The 'rhetoric' is not mere ornament ormanipulation or trickery. It is rhetoric in the ancient sense of per­suasive discourse' .30 According to one conception, rejected by Megill

METAPHOR IN POLITICAL THEORY 167

and McCloskey, rhetoric is mere adornment and adds nothing sub­stantial to the argument itself. According to the other conception,adopted by both authors, rhetoric is a deep structure underlyingargument itself; a deep structure that has hitherto been underplayed,ignored or repressed in the Freudian sense of the word. It should beobserved that although both conceptions may differ, both agree thatrhetoric is distinct from argument.Our analysis of metaphor, however, permits us to see a third con­ception of rhetoric that does not dissociate rhetoric from argument.All that is needed is a readiness to cast rhetoric in the mould ofmetaphor. No one doubts that metaphor is prominent in rhetoric, soin the absence of this readiness I am happy to restrict my argumentto rhetoric only insofar as it makes use of metaphors. Because ofthe affinity of rhetoric and metaphor the following can also be seenas a statement about metaphorical argument. We saw a moment agothat the narrative metaphor proposed by the historian or the politi­cian is defined or individuated by the statements that have been usedfor writing the narrative. It follows that statements about what state­ments a narrative metaphor does or does not possess are all analyt­ically true or false: it is part of the notion of a narrative metaphorthat it contains certain statements and that these statements can beanalytically derived from this notion)) Thus the construction ordeconstruction of narrative metaphors has all the characteristics ofan argument - even of analytical argument. On the other hand, nar­rative argument, the individuation of narrative proposals, shares withrhetoric the desire to convince by persuasion: not by having recourseto fact but by making use of a linguistic construct. Persuasion hereis aesthetic rather than factual, though not necessarily in conflict withfact.In short, from the perspective of narrative metaphor, rhetoric isneither mere embellishment of language nor its hidden deep struc­ture: rhetoric is argument itself and vice versa. The positivist tradi­tion has conditioned us to associate truth and argument primarilywith factual truth and deductive argument and this has made us insen­sitive to the kind of constructive argument one finds in rhetoric ormetaphor and in disciplines like history or politics which make useof constructive, aesthetic argument. Nevertheless, as this statementsuggests, recognition of the possibility of rhetorical argument doesnot rule out the possibility of distinguishing between rhetorical andrational argument. Rhetorical argument is always related to meta-

168 FRANK R. ANKERSMIT

phorical or narrative constructs and these constructs only appear andonly support rhetorical argument against the background of a specificculture containing these constructs. Rational argument may convincea God or a Neanderthal man if only these beings are rational.Rhetoric and rhetorical argument, on the other hand, if they are tobe effective, presuppose a knowledge of the metaphors that are inthe mind of whom one wishes to convince. In contrast to rationalargument rhetorical argument presupposes both self-knowledge andempathy with others.Finally, we have to consider in this section the three elements of

Plato's metaphor, goaldirectedness, unity and distance, and try toderive them from the nature of metaphor. Having dealt withgoaldirectedness in the beginning of this section, I turn to the notionof unity. That metaphor provides unity and coherence is fairly undis­puted.32 The main idea is that metaphor is not knowledge but ratherthe organization of knowledge while unity and coherence are due tothe organizational patterns, or 'Gestalts', proposed by metaphor.From the point of view of political action an interesting complica­tion can be observed here. The unity of the 'Gestalt' stands outagainst a diffuse background lacking this unity. Where, then, shouldwe situate political reality? If we grant reality the connotation of'externality' ordinarily associated with it, the background must bereality from a cognitive view and the 'Gestalt' must be reality fromthe view of political action. The background embodies externalitywhen political metaphors are to be constructed; such metaphors,however, embody externality when one wishes to act on reality. Itseems natural that political action will tend to avoid this realistdilemma as much as possible and this can be achieved by concen­trating on the borderline between the world of the 'Gestalts' and thatof the background rather than on either of these two worlds them­selves. Hence politics will not feel at home in spheres that are eithercompletely narrativized and metaphorized or not narrativized at all.Both the lifeworld, according to philosophers like Ricoeur or Carrthe richest source of narrativity, and nature therefore tend to falloutside the scope of politics. This means that it will be particularlydifficult for politics to deal effectively with ecological problems sincethese problems are far removed from the borderland where politicsis most at ease.That leaves us, finally, with the notion of distance. Above all we

must recognize with Danto the self-referential character of meta-

METAPHOR IN POLITICAL THEORY 169

phor,33 The metaphorical meaning of the statement of a historical orpolitical narrative lies not in what these statements express aboutsocio-historical reality but in what they contribute to the proposedmetaphor itself. Reference to or correspondence with reality obstructsmetaphor's striving for self-reference and thereby weakens metaphor.The greater the distance between the metaphor or its point of viewand the world metaphorized by it, the more successful metaphor is.In its most perfect form metaphor becomes part of a substitute realityin the way Gombrich's hobby-horse is a substitute for a real horse.Insofar as political action is inspired by metaphor it will demonstratea similar affinity with distance. It is surely no coincidence that inthe West the paradigm for effective action is the action of God whois farthest removed from our world and who looks at us, so to speak,from the perspective of the sun. We find a secularized version inLouis XIV, 'Ie roi soleil', who wished to emphasize the majesty ofhis absolutist kingship by moving from Paris to Versailles. 'This wasabsolutism in its most perfect form', writes Kossmann, 'it had cutits roots. It did not associate itself with any particular class or groupof the French population. It withdrew from society and from insti­tutions which had been created by the French monarchy. (... ) Butto be alone and superior the king must also withdraw from the centerof the country and build, at Versailles, not just a palace but the verysymbol of absolutism',34 If one has achieved, like God or Louis XIV,this maximal distance, one's actions seem no longer inspired by ques­tionable metaphors but by the nature of political reality itself. Herethe absolutism of Louis XIV and natural law philosophy mutuallyreinforce one another. This then, is another sense we may give to thepolitical heliotrope.Having found that the three notions of goaldirectedness, unity and

distance that embody the essence of Plato's metaphor are also essen­tial characteristics of political metaphor in general, we may concludethat Plato's metaphor is a metaphor of political metaphor. Smallwonder that political practice and political theory have always beeninspired by the idea that the State can be seen as ship, nor shouldwe be surprised by the fact that so few political philosophers haveseen this: presuppositions can only function as such so long as weare not aware of them.

170 FRANK R. ANKERSMIT

IV. WHY METAPHOR HAS BEEN IGNOREDBY POLITICAL PHILOSOPHERS

Let us listen to how Rawls defines his notion of 'the veil of igno­rance' behind which we decide on matters of distributive justice:'in justice as fairness the original position of equality correspondsto the state of nature in the traditional theory of the social contract.The original position is not, of course, thought of as an actual his­torical state of affairs, much less as a primitive condition of culture.It is understood as a purely hypothetical situation characterized soas to lead to a conception of justice. Among the essential featuresof this situation is that no one knows his place in society, his classor social status, nor does anyone know his fortune or the distribu­tion of natural assets and abilities, his intelligence, strength and thelike. I shall even assume that the parties do not know their concep­tions of the good or their special psychological propensities. Theprinciples of justice are chosen behind a veil of ignorance',35 Themetaphorical character of the Rawlsian notion of the veil of igno­rance will be obvious to anyone. It can even be argued that we haveto do here with a complex interweaving of two metaphors: themetaphor of the veil of ignorance and the time-honoured metaphorof natural society that is presented here under the guise of the'original position'.With its mixing of the metaphor of the original position and the

metaphor of the veil of ignorance, Rawls's political philosophy isthe best and most illustrative example of how Western political phi­losophy has always attempted to obscure and to neutralize itsmetaphorical nature. Unlike his predecessors, the natural law philoso­phers of the seventeenth and the eighteenth centuries, Rawls is notcontent to see contemporary society metaphorically in terms of astate of nature. Rawls also attempts to free the state of nature or theoriginal position of all perspectivism necessarily present in the useof metaphor. Rawls's two metaphors are juxtaposed in such a wayas to cancel out each other's metaphoricity. It is as if Rawls wantsto tell us, 'look here, I know that the original position is not histor­ical but a metaphorical fiction. But with my veil of ignorance I amable to show that the original position is a metaphor acceptable toeach rational person, and, therefore, in practice no different fromreality itself or from deductive argument. So you may be sure thatnothing has been constructed or "invented" here'. There is a uni-

METAPHOR IN POLITICAL THEORY 171

versalism in Rawls's argument that one will not even find in Hobbesor Rousseau, where it is always softened by the quasi-metaphoricalspecificity of some psychological or sociological assumptions. Rawlswants to do without any such assumptions and it is this wish whichnecessitated the introduction of some notion like the veil of igno­rance. The less a political philosophy has empirical content, the moreit is inclined to eliminate political action from its scope (and Rawls'sbook is completely devoid of political action), the more it will beforced to deny explicitly its metaphorical character. In sum, Rawls'sveil of ignorance, this metaphor which has no other purpose than towipe out metaphoricity, is a perfect instance of the politicalheliotrope as we defined it in the previous section, following on fromDerrida. If in general metaphor attempts to hide its metaphoricity,if all metaphor strives for the status of literal truth, then Rawls's 'veilof ignorance' perfectly exemplifies this mechanism for the case ofpolitical metaphor.To explain this fascination with the political heliotrope we mustturn to the transcendentalism of Western epistemology. Kress andGunnell already noticed the affinity of transcendentalism with whatwe have called here the political heliotrope: the quest for a solidfoundation of the political order as we find it in the writings ofRawls, Dworkin or Ackerman, is the political counterpart of thatquest for absolute foundations of knowledge which we find in theWestern transcendentalist tradition since Descartes,36 And it hasalways been the transcendental self in which such absolute founda­tions were located.In a crucial passage Kant characterized the transcendental self as

an entity which accompanies each of our thoughts but of which wecan have no knowledge as such. Kant argues that our thoughts canindeed be said to be the predicates of this transcendental self, butsince we have only access to these thoughts themselves or what theyare about, the transcendental self itself remains an unknown x, anindefinite peg for hanging our thoughts onto,3? Hence, right at theheart of the knowing subject lies a transcendental self that must nec­essarily remain completely unspecific since specificity is to be asso­ciated exclusively with the manifold of experience and thought andsince the transcendental self is ex hypothesi never tainted by thisspecificity. Thus the (metaphorical) point of view we would naivelyassociate with the transcendental self can never function as such,since the latter is devoid of all specificity and rather anybody's point

172 FRANK R. ANKERSMIT

of view. An even more striking example can be found inWittgenstein's Tractatus: 'wo in der Welt ist ein metaphysischesSubjekt zu merken? Du sagst, es verhiilt sich hier, wie mit Augeund Gesichtsfeld. Aber das Auge siehst du wirklich nicht. Dnd nichtsam Gesichtsfeld liisst darauf schliessen, dass es von einem Augegesehen wird')8 This is truly the exact analogue of Rawls's veil ofignorance. Of course we have to do here with a metaphor - themetaphor is even a metaphor of the point of view, which is a mostconvenient circumstance in this context. But the very purpose ofthe metaphor is to eliminate metaphor since the whole drift ofWittgenstein's transcendentalist argument here is to rob themetaphorical point of view of its capacity to function as a point ofview in the proper sense of the word. For we can only see what iswithin our scope and that excludes the eye - the point of view ­itself. Just as in Rawls's veil of ignorance, the point of view is onlyintroduced so as to be eliminated again. The contrast with narrativemetaphor could not be greater. For where the Kantian transcendentalself remains a bare peg for hanging thoughts onto, narrative meta­phor and its point of view completely identify themselves with thenarrative statements that individuate them. Narrative metaphor is asspecific as the unique set of perhaps thousands and thousands ofstatements that make up a specific historical narrative. Here (narra­tive) metaphor is truly radically opposed to the transcendentalist orpolitical heliotrope.The Rawlsian citizen who has retired from the historical world

behind a veil of ignorance is thus the political analogue of theKantian transcendental self. Given Rawls's explicit approval ofKant's demand that ethics should be completely free from empir­ical considerations, the affinity between Kantian transcendenta­lism and Rawls's veil of ignorance need not astonish us. Yet this isonly part of the story. For the transcendental seduction has also givenus the modern notion of the sovereign State. Perhaps no politicalphilosopher since Bodin has contributed more to the development ofthe modern notion of the sovereign state than Hobbes - so let usconsider Leviathan for a moment. One should resist a protototali­tarian reading of Leviathan, though the well-known frontispiece ofthe book seems to invite us to do so. For what is united in and byLeviathan, the modern sovereign State, are not individual citizensbut their points of view. The difference is far from being academicsince it marks the distinction between collectivism and the possi-

METAPHOR IN POLITICAL THEORY 173

bility of pluralism within an acceptance of the modern State. Thisis not hard to see. For to pursue exclusively one's own egoistic, anti­social interests is not at all necessarily incompatible with the accep­tance of a united point of view in the form of the modern, sovereignState. On the contrary, it may well be that we can best serve our own,egoistic interests and organize our self-directed actions most suc­cessfully if we are prepared to look at ourselves from that elevatedpoint of view. In fact, this is precisely the essence of Hobbes'sargument: the mere introduction of this unified point of view, whileleaving everything else, our egoism or aggression, etc., unchanged,will transform initial social chaos into a well-ordered body politic.It must be emphasized that the political heliotrope certainly does notexclude pluralism. If one should wish to attack the politicalheliotrope, the accusation of collectivism decidedly is beside thepoint. The State is, in fact, nothing but the institutionalization ofour readiness to look at ourselves from the top down, from the per­spective of the State, and it is immaterial whether we do so foregoistic or for altruistic purposes. However, in the course of time wehave become so accustomed to looking at ourselves and our inter­ests from the top down that we find it very difficult to conceive thatthings might be otherwise. But, as we shall see, there actually aresocieties where the top 'is blown off', so to speak, and that arenevertheless no less successful in dealing with the challenges of themodern world than our own.Thus, the Hobbesian State in which we have lived until at least

quite recently requires an indivisible point of sovereignty that iscreated through a unification of individual points of view. (Onceagain, this does not at all imply that what is seen from this unifiedpoint of view will in all cases be exactly identical, so that no roomis left for pluralism.) This unification is only possible by denyingthe citizens' point of view its specificity, which we cannot even beginto do as long as we are aware that a non-specific point of view is acontradictio in terminis. However, a political variant of transcenden­talism, which had elsewhere already succeeded in planting a com­pletely unspecific entity right in the heart of the individual, permittedthis crass denial of the metaphorical point of view's specificity. Themodern State is our shared political, transcendental self, not onlymine, but also yours and anybody's. This is why metaphoricity hasalways been both present and absent in Western political philoophyand why the heliotrope could guide the development of the latter.

174 FRANK R. ANKERSMIT

The affinities between transcendentalism and the politicalheliotrope invite a last comment that will only be developed inoutline here. Several philosophers, most notably the Wittgensteinof the Tractatus, have noted the propensity of transcendentalism forsolipsism. The idea is, roughly, as follows. Ifwe 'go transcendental'there is no obvious top demarcating the self from the non-self onthe route from the empirical self to the transcendental self. The bor­derline between the self and the non-self becomes an arbitrary choiceand the same will obviously be true for each proposal to distinguishbetween outward reality and the self. Solipsism and pure realismtend to coincide: 'hier sieht man dass der Solipsismus, strengdurchgeftihrt, mit dem reinen Realismus zusammenflillt. Das Ich desSolipsismus schrumpft zum ausdehnungslosen Punkt zusammen, undes bleibt die ihm koordinierte Realitat' .39 That political transcen­dentalism has this same tendency will be obvious if one recalls, aswe saw a moment ago, that the State is anybody's political tran­scendental self. Here the distinction between the State and what isnot the State is as arbitrary as in the case of solipsism. Now, solip­sism, despite its spectacular appearance, is in fact a most harmlessand innocuous philosophical doctrine. Probably no philosophicaldoctrine has fewer consequences, whether one subscribes to it or not.It is like changing from summertime to wintertime: there are newdenominations for everything but for the rest everything remains asit was. As far as I can see the same applies to the solipsism of polit­ical transcendentalism. There is, nevertheless, one consequence thatdeserves our attention from the perspective of intellectual history.As in the case of solipsism, political solipsism tends to paralyze ourinstincts and intuitions about demarcations - in this case, aboutdemarcations between the State and what is not the State, betweenthe State and the citizen, etc. In other words, political transcenden­talism has the tendency to create a discourse in which one can movearound with a surprising ease. The really amazing amount of polit­ical positions that have been developed in the Western world sincethe seventeenth century and the fact that individual people havesometimes moved from the extreme left to the extreme right withoutblinking an eyelid can perhaps be explained with an appeal to thesolipsism of political transcendentalism. Similarly, if the anti-tran­scendentalism that has recently become fashionable will also affectpolitical thinking, the broad spectrum of political opinions thatwe used to have may be expected to become much narrower. Surely,

METAPHOR IN POLITICAL THEORY 175

this anti-transcendentalism or anti-foundationalism is bound to causesome intellectual heavy weather for Plato's ship of State.

V. THE DISINTEGRATION OF PLATO'S METAPHOR:

THE ANATOMY OF INDEPENDENCE

Nevertheless, we ought to be realists and recognize that if Plato'sship is threatened it will rather be by social or political than by philo­sophical storms. That will be the topic of this section. First of all,the political heliotrope was never been in complete harmony withthe realities of politics. The indivisible point of view of the heliotropewas never realized. Sovereign political power was always dividedconstitutionally and in actual political practice. Moreover, it is hardto reconcile political representation with the aims and requirementsof the heliotrope, though this may not be impossible and an argumentcan be devised that turns representation rather into an ally than anenemy of the heliotrope.4o

But, more importantly, a number of political developments in mostWestern societies over the last few decades all suggest the gradualdisintegration of Plato's ship. Postmodernist distrust of the grandmeta-narratives that always gave to the ship of State its directionundermine the notion of goal-directedness. State and civil societyare anything but a unity and tend to dissolve into a number of moreor less independent islands, and little distance seems left betweenthe State and civil society nowadays. In the remainder of this sectionI shall concentrate on the third development - the dissolution of theState in civil society - since it is not difficult to extrapolate fromthis development to the other two.

If there is one fact contemporary political philosophers and polit­ical scientists agree about, it is that the traditional boundariesbetween State and civil society are disappearing. 'The distinctionbetween the State and civil society seems to have largely collapsed'writes Maier in his volume on the changing boundaries of the polit­ica1.4! Keane even speaks of 'a "magma" of overlapping hybrid insti­tutions no longer discernible as either "political" or "social entities".42Though this may be an exaggeration, it cannot be gainsaid that theState is no longer the unitary agent it used to be in the nineteenthcentury; it has lost much of its freedom of movement to transnationalorganizations on the one hand and to a multiplicity of subgovern­ments with their own systems of decision-making on the other.43

176 FRANK R. ANKERSMIT

And Claus Offe, Habermas's discussion partner in so many socialand political debates, has brilliantly shown in a recent analysis thatthe increasing entanglement of State and civil society and theincreasing impact of the State on the citizen has spawned an entirelynew political culture. This 'new paradigm' for politics as charac­terized by Offe has materialized in the counterpolitics of the 'greenparties', with their emphasis on ecology, their insistence that we mustundo what Habermas called 'the colonization of the lifeworld' etc.Two things must be noticed about this new political paradigm. Ittruly is a new paradigm because both in form and in content the alter­native politics of the 'greens' and their like cuts across the wholespectrum of traditional political ideologies and parties. On the otherhand, as Offe emphasizes, the new paradigm does aim at a reinvig­oration or even restoration of politics and its former preeminence.According to Offe the new paradigm is decidedly modern in its'evident belief that the course of history and society can be changedand created by people and social forces determined to do so' .44 Offe'sclaim about the greens is so interesting because it makes clear thatthe greens still have a lot in common with their neo-conservativeopponents like Samuel Huntington who also want to re-establisheffective government.45 Both the greens and the neo-conservativeslament the disintegration of Plato's ship and by doing so betray theirshared indebtedness to the political heliotrope. Both are reactionariesin the sense that they want to 'react' against what seems to be theirreversible course of contemporary political history.Insofar as the complexities of contemporary political life are both

the expression and the result of the increased entanglement of Stateand civil society, these complexities have been widely discussed bynumerous political scientists. I shall refrain from reviewing their dis­cussions here. I am no expert in this field. And there is a more fun­damental reason to be reticent. When political scientists discuss theseproblems they do so from the outside, as it were: from a safe distancethey discern two previously given entities, the State and civil society,and then they try to analyze how and why these two entities becomeensnared in one another. No less than the greens or the neo-conser­vatives, therefore, political scientists look at the problem from thepoint of view of the heliotrope. The approach of the political scien­tist is thus handicapped by presupposing what is really part of theproblem. I therefore prefer to approach the problem from 'the inside'- that is to say, by taking the notions of the State and of civil society

METAPHOR IN POLITICAL THEORY 177

themselves and not what is denoted by them as my point of depar­ture. I shall inquire next what dialectics governs the relation betweenthe two. This approach compels us to follow the path of politicalphilosophy rather than that of political science. And if we decide tofollow the former path the first political philosopher likely to be ofhelp is Benjamin Constant. For Benjamin Constant (1767-1830) notonly gave us the first but also the clearest definition of the conceptsof the State and of civil society. Moreover, as we shall see, hiswritings contain a surprising analysis of the very dialectics that weare looking for. No political philosopher has surpassed Constant'sanalysis of the relation between State and civil society in depth andsubtlety. The fact that in both his personal and public life Constanthad an almost neurotic obsession with all the problems this relationmay give rise to - especially where freedom and independence areconcerned - may explain the penetration of his insight and why hestill is the best thinker on the subject.The concepts that do most of the work for Constant are the

concepts of freedom and independence. The latter is perhaps themore important of the two since it gives the right flavor to the notionof freedom and since we can also apply it, unlike freedom, to insti­tutional spheres like the State and civil society. The central role offreedom and independence (or freedom as independence) inConstant's political philosophy is already exemplified by his defin­ition of the State and civil society in terms of freedom and inde­pendence. In contrast to Constant, modern writers on State and civilsociety do not make the notions of State and civil society concep­tually dependent on other notions and that may partly explain theirhelplessness. This conceptual relation is defined by Constant in thefollowing way. In his treatise on the contrast between ancient andmodern liberty, in which all the threads of Constant's political phi­losophy are adroitly woven together into one powerful intellectualtexture, Constant pointed out that ancient liberty or what we nowcall 'political liberty' consisted in the citizen's right to participatein the process of policy-making. Modern or 'civil liberty', on theother hand, is the freedom of the citizen from immixture of the Statein his affairs - it thus is primarily an independence from the State.Ancient or political liberty is best suited to the small State of theclassical polis, whereas modern or civil liberty is required for thelarge States of modern Europe. '11 resulte de ce que je viensd'exposer, que nous ne pouvons plus jouir de la liberte des anciens,

178 FRANK R. ANKERSMIT

qui se composait de la participation active et constante au pouvoircollectif. Notre liberte, a nous, doit se composer de la jouissancepaisible de l'independance privee' .46 The following qualificationmust be added. Constant does not wish the citizen of the modernEuropean State to renounce his political freedom: on the contrary,political liberty as exercized within a system of representative gov­ernment is a necessary condition for safeguarding civil liberty. Wecan therefore say that Constant's distinction between ancient andmodern liberty is analogous to the distinction between direct democ­racy and representative government - a distinction that was the pointof departure for most political debates in the days of Constant. Butif we consider how the two forms of liberty were defined by Constantwe shall see that the association of political liberty with the Stateand of liberty as independence with civil society must follow as amatter of course. Independence is the crucial concept here: bothspheres are defined in terms of their being independent of each otherand civil society can even be seen as the very incarnation of inde­pendence itself. Lastly, although Constant never explicitly says sohimself, the notion of civil freedom as used throughout his workoften has the connotation of 'security': the citizen is 'secure' whenand insofar as he is independent from the State. This is interestingfor the following reason. One can say that the modern welfare Stateconsiders its main task to lie in organizing the security of the citizenwithin a complex system of welfare facilities. What used to belongto the domain of civil society - security - has now become the veryraison d'etre of the State. So here we meet for the first time theconceptual dialectics between State and civil society that willdemand our attention for the rest of this section.In a most remarkable way Constant dramatizes the mutual inde­

pendence of the State and civil society by also assigning to each adifferent regime of truth. Constant's surprising thesis is that deceit,dissimulation, pretext, in short, hypocrisy may in many cases proveto be an unqualified good in the public sphere of the State47 - a thesisthat earned the young Constant even the honor of a debate with Kant.In the words of Holmes, Constant's most recent biographer, 'dis­simulation and fraud are an integral part of modern freedom' .48Constant's thesis is all the more surprising because, as HannahArendt has shown, late eighteenth-century intellectuals unanimouslyagreed that hypocrisy was the supreme vice in politics. Theirargument was that hypocrisy teaches us not only to playa role toward

METAPHOR IN POLITICAL THEORY 179

others but toward ourselves as well; thus the last court of appeal forthe cause of truth and justice would be corrupted. The authenticityand sincerity which form the ultimate basis for the State and goodgovernment would be undermined by hypocrisy.49 When Constantpermitted the State to yield, under certain circumstances,50 to theseductions of hypocrisy, he gave the following argument. What isat first only hypocritically feigned by the State may, in the end,become a true part of its character. Hypocrisy may therefore beinstrumental in the education, or 'Bildung', of the State. If the Stateimproves its character, so to speak, this does not happen overnightbut only after an intermediate phase during which the State merelyfeigns that it has bettered its ways. Because of his insistence on theeducational value of hypocrisy, Constant's defence of State deceitshould decidedly not be put on a par with Plato's 'noble lie' or withMachiavelli's eulogies of pretext and lying.From the perspective of the entanglement of State and civil society

- the sphere of the individual citizen - it is interesting that Constanttranslated his proposition about the hypocritical State to the indi­vidual citizen. And, as we might expect from that perspective, hisargument is that politics even intrudes upon the privacy of thecitizen's inner sanctuary. The hypocrisy of the State has its coun­terpart in our hypocrisy towards ourselves. In the mental strugglesof the characters in his two novels Constant is out to demonstratethe impossibility of discovering or laying bare our truest or inner­most feelings. Constant's own experience was the source of thisinsight into the insincerity of sincerity. For example, when writinghis diary, Constant observed that even if he did his utmost to besincere about himself he felt an irresistible urge 'to speak for thegallery' .51 There is in Constant.an almost Derridian resistance to 'ametaphysics of presence': even if we wish to get to the truth aboutourselves, we have no privileged access to ourselves that might guar­antee such a metaphysics of presence. Each attempt to penetrate intothe recesses of our mind chases away something else and so disturbsthe psychological mechanism we wish to grasp. Constant is here farsuperior to the psychological utopia of Rousseau's Confessions(although Rousseau's urge 'to speak for the gallery' was undoubt­edly even stronger than that of Constant). Thus the dialectics of theinsincerity of sincerity makes us aware that even in our most privateinner world, which we always believed to be the foundation andincarnation of our independence, this independence is successfully

180 FRANK R. ANKERSMIT

challenged by the political and the social in the form of an absent'gallery'. In Constant's rejection of the possibility of being sincereto oneself we encounter for the first time his recognition of theintrinsic weakness of independence.But it is only in his novel Adolphe that Constant gave a full

account of this weakness and showed why the desire for indepen­dence is self-defeating. This bleak and bitter story of the lovebetween Adolphe and Ellenore has been interpreted in many differentways over the last century and a half. To the dismay of both Mmede Stael and of Constant contemporaries saw the novel above all asan autobiographical account of their love affair - and this is certainlypart of the truth. In a more recent interpretation Todorov proposedto read Adolphe as proof of the immense performative power ofwords and speech.52 That already brings us closer to our goal. For Ishall discuss the novel for what it most manifestly is: a thesis aboutthe anatomy of independence and of power and freedom. In doingso I follow Holmes's approach when he says that the novel must beseen 'as a commentary on the emptiness of negative freedom'53 ­that is, of freedom as independence. It cannot be doubted thatindependence is the book's guiding theme; the word makes itsappearance at every crucial phase of the story. Adolphe is a highlypolitical text, no less political than Rousseau's novels orDostoyevski's Demons.The tone of the novel is set by the manner in which Adolphe loses

his independence. Bored by the social life at the tiny German courtof Dm and intrigued by the ecstatic stories of a friend's love affair,Adolphe decides that starting a love affair himself might dispel hisennui. He talks himself into a passion for Ellenore and (typicallyConstantesque!) is so successful that he soon believes his whole life,future and happiness depend on the favors of Ellenore. The messageis, clearly, that our independence is something we seldom have veryprecise ideas about and if we have them it is usually after we havealready lost it. Independence is an active factor in our life rather inits absence than when we possess it. Knowledge about our inde­pendence is characteristically a wisdom after the event.Almost as soon as Adolphe succeeds in winning Ellenore's heart

their relationship becomes strained and unsatisfactory to Adolphe.The mistress of Count P"', Ellenore was barely respectable beforeshe began her affair with Adolphe. But after this relationship hasstarted she becomes a social outcast and as a result is now wholly

METAPHOR IN POLITICAL THEORY 181

dependent on Adolphe. Paradoxically, this does not enlarge but onlyfurther restricts Adolphe's freedom and room for manoeuver: fromnow on it is he who is entirely responsible for Ellenore. ThusEllenore's dependence on Adolphe generates Adolphe's dependenceon Ellenore. There is apparently no law of conservation of energyin human relations: all the independence there was in the beginningmay turn into mutual dependence. And in the case of Adolphe andEllenore the result is an impasse that lasts for the greater part of thenovel and ends only with the tragic death of Ellenore, after she hasread the letters sent to her by Baron de T··· about Adolphe's decisionto leave her.The clue to the novel can be found in that chilly but lucid letter

Adolphe's father writes him after his departure for Poland togetherwith Ellenore: 'je ne puis que vous plaindre de ce qu'avec votreesprit d'independance, vous faites toujours ce que vous ne voulezpas' .54 This is truly the paradox in the story of Adolphe's life asrelated by Constant: his passion for independence merely results independence and in situations where he has to do what he neverwanted to do. But is it really a paradox? The main lesson ofConstant's novel is indeed that it is not a paradox; on the contrary,we are mistaken if we suppose that Adolphe's (or anybody's) desirefor independence should ever result in his freedom to do what hewants to.

If we ask how Constant's novel teaches us this counter-intuitivelesson, we must note, above all, that all action in the novel origi­nates in Adolphe's strained relationship with Ellenore. There is noaction outside this relationship and its history. Adolphe's life beforehe came to know Ellenore was an irrelevancy and after Ellenore'sdeath Adolphe failed to become the successful man his father andBaron de T··· had expected and hoped. Instead, he became thatshadow of a man encountered somewhere in the South of Italy bythe supposed source of the novel's text. Adolphe's regained inde­pendence did not turn out to be the necessary condition for new inde­pendent action on his part. And this is not because Adolphe is a weakman, though that is true as well. It is rather the reverse: because heis a weak man Adolphe became so passionately interested in his inde­pendence. Adolphe says so himself at the beginning of the novel:he explains there how his weakness and inherited timidity kindledin him 'un desir ardent d'independance' .55 So we must concludethat the kind of action (or, rather, inaction) arising from and asso-

182 FRANK R. ANKERSMIT

ciated with the desire for independence can only be found inAdolphe's relationship with Ellenore and not outside or beyond it.And that implies in turn that independence only plays its role withina context of dependence. Independence is like a fish swimming ina fishglobe of dependency. Independence is not an autonomous factorbut rather something to be regretted or to be restored and in bothcases independence presupposes the presence or a background ofdependence. Quite revealing here is what Adolphe says much laterabout his state of mind immediately after the death of Ellenore:'combien elle me pesait, cette liberte que j'avais tant regrettee.Combien elle manquait amon coeur, cette dependance qui m'avaitrevolte si souvent' .56 Adolphe always regretted his independence, butas he has now come to realize, what really lay behind this regret washis desire for dependence. Independence and dependence are noteach other's opposites at an equal level; independence is rather afeeling about, or a reaction to the wish for dependency and therebypresupposes the latter.It is not hard to spell out what this implies for the relationship

between the State and civil society - another very untidy love affair.When in the field of force between dependence and independencethe latter systematically proves to be the weaker of the two, themutual independence of the State and civil society cannot last forlong. All the more so since one of the two, civil society, is even thepolitical embodiment of independence and therefore by its verynature destined to undermine their initial relationship. That it hasnevertheless been possible to make the distinction between the Stateand civil society at all and that the distinction was reasonably clearfor a century and a half partly finds its explanation in the absolutismthat preceded the political world we live in since the beginning ofthe last century. Absolutism, as defined by 'droit divin' theorists likeKing James I, Robert Filmer or Bossuet, only recognized the rulersand the ruled and it had no use for a system of dependence andindependence (that is where it differs both from the modern worldand from feudalism). The terms dependence and independence arenot part of the vocabulary of absolutism. Admittedly, there is someconceptual overlap between the dichotomy of rulers versus ruled andof dependence versus independence, but this overlap covers only partof their respective meanings. Being ruled differs from dependence.My life may depend of the airpilot, but I am not ruled by him;Voltaire was ruled by Louis XV but it would be odd to say that

METAPHOR IN POLITICAL THEORY 183

Voltaire was dependent on this monarch. And think of independence.Are we not at first inclined to say that the ruler is independentbecause he is independent of those he rules? Yet, throughout thissection we have learned to associate independence with civil society,the sphere of those who are ruled. If we wish to capture the differ­ence between the two dichotomies, we might point out that the dif­ference between ruler and ruled far exceeds that between dependenceand independence - not in the last place because of the dialecticsbetween the notions of dependence and independence expoundedabove. Needless to say, when I use here the word 'distance', I takeit to refer to the metaphorical distance we encountered in our analysisof Plato's metaphor, the distance that proved to be a condition forthe manoeuvrability of the ship of State.As from the nineteenth century, however, the system of depen­

dence and independence was grafted onto that of the rulers and theruled. The inconsistencies between the two systems were concealedby that most obscure concept of all political philosophy, the conceptof popular sovereignty. This concept conveniently confounded thetwo systems in such a way that the politicians and political philoso­phers who used it could at all times produce the outcome requiredby the circumstances. For example, because in representative gov­ernment the rulers are to some extent dependent on the electorate thesophistry became now possible that not those who effectively ruledbut those who are ruled are the real rulers. The major shortcomingof political thought since the nineteenth century is that it has alwaysfailed to distinguish between the system of rulers and ruled and thesystem of dependence and independence. As we shall see in amoment, that has obscured from view a specific form of politicalpower. In any case, the absolutist system of rulers and ruled has accu­mulated in the course of time what we might call a large 'capital of(metaphorical) distance'. For a century and a half Western politicshas lived comfortably on this 'capital of distance' and it is this capitalthat long made the distinction between the State and civil societyboth acceptable and realistic. The capital of distance was, however,gradually consumed by the system of dependence and independence.And when the capital of distance was exhausted by the dialecticsbetween dependence and independence, the latter had to gain theupper hand since the system of rulers and ruled now could no longerslow it down. This meant both the dissolution of the boundariesbetween the State and civil society and rudderless ship of State)?

184 FRANK R. ANKERSMIT

If we want a closer look at the results of the dialectics of depen­dence and independence, the story of Adolphe is once again aninstructive guide. We should not forget that a great deal happens inthe novel: there are outbursts of emotion, there is a passionate loveaffair, an ensuing concatenation of miseries and, finally, the deathof Ellenore and Adolphe's dwindling into a mere shadow of hisformer self. Yet all these things happen without anybody seemingto be in charge. A great deal of power is implied in the drama actedout by Ellenore and Adolphe, yet this power does not seem to findits origin in either Ellenore or Adolphe. Their relationship is in manyrespects similar to that depicted by Hegel in the famous chapter ofthe Phenomenology on the master-slave relationship (in which Hegelalso made use of the terminology of dependence and independence).For in both cases power relations that originally seem clear andunambiguous are subverted and changed beyond recognition. Thusone could plausibly argue that Adolphe holds all the winning cardswhereas Ellenore is condemned to a passive role. But just as stronga case could be made for the opposite view that Ellenore is in thestrongest position. Is not the story of the novel impressive proof ofthe power she exerts over Adolphe? Yet both views would be incor­rect and fail to do justice to that peculiar dialectics of dependenceand independence that governs the relationship between Adolphe andEllenore. It is perhaps the supreme achievement of Constant's novelthat it successfully demonstrates how even in the tiny microcosmof a relationship between two individuals power can already becomeanonymous. Power - and as I said a moment ago, power has beenimmensely effective in the novel - has been swept from its base offoundations in the will, the desires or the goals of the principal actorsinvolved. Power freely circulates here between Adolphe and Ellenoreand although it occasionally absorbs in its course some externalinfluences like that of Adolphe's father or of Baron de T···· it largelyremains a self-propelling and self-perpetuating factor that is innobody's control.We might therefore discover in Adolphe a third paradigm of

power; a paradigm of power we can only begin to perceive owingto the dialectics of dependence and independence. The first paradigmof power is that of the heliotrope, the power of the 'roi soleil' to rulehis country. Right at the opposite end of the power spectrum wefind Foucault's 'capillaries of power', that insidious kind of powerthat conceals itself in disciplinary discourse. Foucault's power is a

METAPHOR IN POLITICAL THEORY 185

subtle kind of power, felt or even noticed by a few, but neverthelessquite real. It was undoubtedly Foucault's insistence that we still 'haveto cut off the King's head' that made him proceed to this otherextreme in the power spectrum without considering any intermediateform of power. Between the power of the heliotrope and disciplinarypower we can, however, situate a third paradigm of power. This isan anonymous and unpredictable kind of power, without traceableorigins or foundations, freely circulating around much like thoseimmense and elusive clouds of money moving around the financialworld in a way that worries so many contemporary economists; apower whose existence was first suggested by Constant with hisanatomy of independence. This form of power is no less real but agreat deal more visible than Foucault's disciplinary power. In fact,we have already been aware of this kind of power for a long timealbeit only as an absence, as powerlessness. But the powerlessnessof the modern State I have in mind here, its inability to steer the shipof State, should not be seen as mere powerlessness, as an absenceof power, as degree zero on the powerscale - rather, there is another,anonymous kind of power at work whose operations are still largelya secret to us. In the next section I shall try to lift a corner of theveil which has covered the mechanism of this anonymous power thatseems to lack foundations.

VI. THE EMPTY VERSUS THE SOVEREIGN CENTER

The gradual dissolution of the boundaries of the State, the increasingentanglement of State and civil society and the resulting loss of theState's manoeuvrability are considered by both politicians and polit­ical philosophers to be perhaps the most alarming evolution withincontemporary political practice. The intuition underlying their alarmis, obviously, that these and related developments are unnatural inthe sense of being at odds with how we traditionally conceive thenature of the State and of politics itself. The major problem for acorrect grasp of this dissolution of the boundaries of the State isthat we find it difficult to imagine a political order where is naturalwhat we consider as unnatural and as a degeneration of all politicalclarity and rationality - and, surely, as long as we accept theheliotrope we cannot do otherwise. Yet such an 'unnatural' politicalorder already exists. In many ways this political order and the societysupporting it are remarkably successful and effective, perhaps even

186 FRANK R. ANKERSMIT

more so than our own. We all know this society. It is what has beencalled 'the Japanese system' .58 To put it succinctly, the absence ofa political center, of a sovereign center, in short of the heliotrope,is not accidental to but the very essence of the Japanese politicalsystem. Over against our metaphor of the sovereign center, fromwhich the ship of State is governed, Japan places its metaphor ofthe empty center.Before entering into a discussion of the Japanese empty center Ishould comment on some differences between Western and Japanesevalue systems insofar as these differences are relevant to our subject.Within the Western theological and philosophical tradition some kindof 'pay-off' has always been expected from the discussion and theensuing identification of universal values. However sceptical, anti­theoretical or anti-foundationalist we may have become, we stillbelieve that such universal values are necessary for determininglarge-scale political action since they function, so to speak, as thecompass for the ship of State. Japan, on the contrary, has little orno use for universal values. A most striking example is given byEdward Seidensticker when he reports that a student, after havingbeen asked what values she accepted, came up with the followinglist: 'Marxism, non-Marxist socialism, liberalism, humanism, prag­matism, anarchism, nihilism, existentialism, nationalism, hedonism,ideology-free' and, last but not least, 'others' .59 Of course theexample is extreme but other inquiries confirm the impression thateven educated Japanese simply have no instinct for universal values.And we should not take that to be their universal value - the Japaneseare no nihilists. The explanation is, rather, that they do not wish toplay our game of seeing the political and social order in terms ofsuch universal values. They do not see political reality from the pointof view of such values, nor can they believe that these values createpolitical reality. In his magisterial book on Japanese political practiceVan Wolferen situates the crucial factor in the exercise of politicalpower in Japan in this 'near absence of any idea that there can betruths, rules, principles or morals that always apply, no matter whatthe circumstances. Most Westerners as well as most Asians whohave stayed for any length of time in Japan will be struck by thisabsence; and some Japanese thinkers also have seen it as the ultimatedeterminant of Japanese public behavior' .60 The Japanese are notmotivated by universal values but by values that are radicallyparticularist, that is, values that are directly related to the wish torealize some specific goa1.61 This leads to a preference of 'per-

METAPHOR IN POLITICAL THEORY 187

formance' to 'maintenance' since maintenance is rather a back­ground for discerning particularist goals than such a particularist goalor value itself.62 This preference of performance over maintenancemay partly explain the dynamism of Japanese society and thesurprising ease with which it adapts to new and changing circum­stances.Yet, to a certain extent this picture is misleading. For Japan is

not completely without universalist values. In his book on the historyof Japanese political thought Najita describes this thought as a con­tinuous search for the balance between bureaucracy ('kanryoshugi')on the one hand and universalist idealism ('ningensi' or 'kokoro')on the other.63 Right from the beginning of the Tokugawa period, sofor two to three hundred years longer than the West, the Japanesehave been discussing the phenomenon of bureaucracy and the valuesfrom which it should draw inspiration. However, and there's the rub,Japanese universal values are, in comparison with those of the West,peculiarly empty and devoid of positive content. Such values are, forexample, filial piety, devotion to the Emperor, a striving for 'wa',that is, harmony and consensus, or for striking the right balancebetween 'giri' and 'ninjo' or between 'on' and 'hoon'. These lastterms refer to the extreme sense of obligation Japanese are apt tofeel towards anybody who has done them a service, however small.This sense of obligation sometimes takes on the features of a kindof social original sin.64 All in all, such Japanese universal values asthere are rather have the character of loyalties than of a commit­ment to what society ought to be like from the point of view ofcertain ethical or political goals. The emptiness of Japanese valuesis notably illustrated by an often cited anecdote about a samurai whofound himself in the difficult position of having to choose betweentwo such conflicting loyalties. Because of the very emptiness of thetwo loyalties there was no way for him to decide between them andhe had no choice but to commit suicide. The story reminds one ofthe breakdown of a computer that has been asked to solve somelogical contradiction.65

But perhaps the emptiness of Japanese universal values is bestillustrated by Shinto, Japan's more or less official religion. Shintois a religion without dogma's, holy books, ethical prescriptions, etc.;it is what we should nowadays call a typically 'anti-foundationalist'religion. Now, what one does find in a Shinto-shrine, in its holiestof holies, is not an image of the deity, a relic, or some sacred text,but, of all things, a mirror. A mirror, moreover, that neither priest

188 FRANK R. ANKERSMIT

nor worshipper is permitted to look at or in. The mirror has thishonored function because, as Ono explains, 'it symbolizes the stain­less mind of the kami [the kami is the spirit which inheres in allthings EA.]. (... ) Everything good and bad, right and wrong, isreflected without fail. The mirror is the source of honesty becauseit has the virtue of responding according to the shape of objects. Itpoints out the fairness and the impartiality of the divine will' .66 It isprecisely the emptiness of the mirror which permits it to reflect whatis, without adding anything of itself. The mirror gives us a doublingor reality - one is reminded of the 'dedoublement' which RolandBarthes believed to be so characteristic of the Japanese use of thesign67 - rather than that it gives us a perspective on what is. Themirror is 'literalist' and 'anti-metaphorical'.Most writers on Japan agree that Japanese culture, religion and

politics are very closely related; so much is already suggested by thefact that the word used for referring to politics, 'matsurigoto', orig­inally means religious observance.68 Yet politics is the sphere wherethe notions, practices etc. of culture and religion are, in a way, reca­pitulated.69 Politics is all-pervasive in Japan. We may thereforeexpect that Japanese politics has an analogue to the empty center inethics and religion. This is in fact the case, according to the mainthesis of Van Wolferen's book on how power functions in the'Japanese system'. In Japan, writes Van Wolferen, there is 'a hier­archy, or rather a complex of overlapping hierarchies. But it has nopeak: it is a truncated pyramid. There is no supreme institution withultimate policy-making jurisdiction. Hence there is no place where,as Harry Truman would have said, the buck stops. In Japan, the buckkeeps circulating',70 The irony about Japan is that it so much makesthe impression on outsiders of being an economic giant deliberatelyand singlemindedly bent on the economic conquest of the world. But,there is no masterplan behind it all for the very simple reason thatJapan lacks the political center for developing such a masterplan.71

Nor should it be thought that this political center has abolished itselfand delegated its right to and its capacity for policy-making to otherinstitutions. For exactly the same pattern emerges when we lookclosely at the bureaucracy of the world of trade and industry.Whether we have to do with the relations between the State, thebureaucracy and the world of trade and industry or whether we haveto do with each of these spheres individually, decision-making andpolicy-making in Japan differs from what we are accustomed to in

METAPHOR IN POLITICAL THEORY 189

the West in that it always seems intent on blurring the lines of respon­sibility and accountability; it is as if hierarchy only exists in orderto conceal and to raise an impenetrable mist around what really goeson (or should be going on according to Western premises).72 Thus,it has been found that the process of coming to a conclusion ordecision is in Japan effectively rounded off by deliberately creatinga 'final ambiguity with respect to authority and responsibility for theproposal' .73 If the hierarchy's function is to hide what really goeson - and as such it is a vital necessity! - it follows that Japan hasno use whatsoever for strong charismatic leaders. A Napoleon wouldbe utterly inconceivable in Japan74 and there can be no greatermistake than to believe that the Japanese leaders during the SecondWorld War had anything in common with Hitler or Mussolini.The prime example of all this is the Japanese Emperor and his

function in the Japanese political universe. One is inclined to put theJapanese Emperor on a level with Europe's remaining constitutionalmonarchs since there is in both cases a discrepancy between hierar­chical status and the possession of real power. But the apparentsimilarity is misleading and, generally speaking, the outward appear­ances of Japanese parliamentary democracy must not be misinter­preted as evidencing the existence of a real parliamentary democracy.For one thing, whereas the ancestors of the European constitutionalmonarchs once possessed real power, this has never been the casewith the Japanese Emperor in Japanese history. There never has beena Japanese 'roi soleil' (even if the Emperor is believed to be adescendant of the sun or the Goddess of the sun Amiterasu). Thepowerlessness of the Japanese Emperor already amazed Chinesechroniclers in the third century AC,75 The powerlessness of theEmperor is not an ingenious solution for a constitutional problem ­in the way that Benjamin Constant came to circumscribe the role ofthe constitutional monarch as 'un pouvoir neutre' with the task ofarbitrating between Montesquieu's three powers76 - but is paradig­matic for how the Japanese wish to screen the origins and exerciseof effective power. That the powerlessness of the Japanese Emperoris paradigmatic of the anonymity of Japanese political power ratherthan the result of a constitutional arrangement is also clear from thefact that exactly the same phenomenon is repeated at lower levels:official dignitaries below the Emperor often functioned as no lessvisible but in fact powerless screens behind which real power wasexercized.77 And yet the institution of Imperial rule was never abol-

190 FRANK R. ANKERSMIT

ished and such an abolition was even never contemplated in thewhole of recorded Japanese history. We are thus left with the con­clusion that the institution of the Emperor served a hidden but vitalpurpose: namely the creation of an empty political center and of ananonymous power without foundations since these foundations soobviously could not be where they seemed to be.The full sovereign political center typical of the West, howeverthis center may be defined, exists because we are prepared to lookat ourselves in a political context from the top down. As we foundwhen discussing the Western political heliotrope and the Hobbesianconcept of the State, we in the West are required and inclined toidentify with the point of view of the sovereign center even if wehave only our own selfish interests at heart. In Japan the movementis exactly opposite. It is a movement that the Japanese refer to bythe word 'rengo-sei', that is, 'subordination':78 the superior movesdownwards, so to speak, to the subordinate and by doing so sanc­tions the latter's actions. The superior is required to identify withthe point of view of the subordinate rather than the other way round.Thus the Emperor has been described by Najita as 'a constantspiritual presence accessible to and, indeed, coincidental witheveryone' .79 It will be clear - and this is the key to the secret - thatthis coincidence of the Emperor with everyone is only possible tothe extent that the Emperor remains a completely empty presence:only thanks to this emptiness is such a 'dissemination' of hispresence conceivable. Any concreteness of Imperial will would upsetthe movement of 'rengo-sei' and put an end to the anonymity ofpolitical power.In the West the sovereign is situated, as it were, 'ahead of us':

the citizen and the civil servant follow the sovereign by identifyingthemselves with his heliotropic point of view (that they may con­tribute to its material definition does not alter this). In Japan thesovereign is, rather, 'behind' the citizen or civil servant, for the sov­ereign is prevented by his emptiness from giving actual guidance.This does not mean that the Japanese equivalent of the Western sov­ereign legitimates ex post facto the actions of the civil servant. Forthe notion of legitimacy has its origin in the heliotrope that doesnot apply to Japan. Even more beside the point would be the sug­gestion that the Japanese exercise of power is more democratic thanthat of the West owing to the absence of a sovereign will in the emptycenter. It is not illuminating to see the differences between the West

METAPHOR IN POLITICAL THEORY 191

and Japan in terms of democracy; these differences rather concernthe notion of sovereignty itself and that is an issue prior to ques­tions of democracy. It would be no exaggeration to say, instead, thatthe Japanese metaphor of the empty center is, from the Western pointof view, a formal denial of the notion of sovereignty. For Japanesepolitics only needs a supposedly sovereign center so that it can denythis center the quality of being the source, origin or foundation ofpower - though as such the center is absolutely vital. It is likeMagritte's painting of a pipe with the caption underneath saying thatthis is not a pipe - and, indeed, the painting is not a pipe. Surely,all this is mind-boggling for us who are accustomed to Westernpatterns of power; but a grasp of the 'Japanese system' can also helpus understand the new kind of political power that is slowly emergingfrom the familiar constitutional constraints of our own Westernpolitics.The exercise of political power does not take place in a vacuum.

The psychology of a people and its sociological characteristics arenot only a reflection but a part of power. A political philosophyneglecting psychological and sociological determinants must neces­sarily remain, as was stressed some years ago by Maurice Sandel inhis debate with Rawls,so a purely academic game only of interest toa few professional philosophers. I therefore turn now to the psy­chological explanation of the Japanese empty center given by TakeoDoi in his The anatomy of dependence.S) Central to Doi's argumentis the concept of 'amae' and the related verb 'amaeru'. The word'amae' is the center of a complicated web of other words whosemeanings lack exact analogues in Western languages and, even moresurprisingly, suggest different translations that actually contradictone another.s2 The Japanese mind really seems to be incommensu­rable here with that of the West. The Daigenkai, the Japanese dic­tionary, translates the word 'amae' as 'to lean on a person's goodwill's3 and 'amaeru' means to behave self-indulgently in the safe andreassuring confidence that the other will love you all the more forit.S4 'Amae' has its origin in the nursery and is closely related to theprocess during which the child separates itself from its mother andbegins to give itself an identity of its own. Doi clarifies 'amae' bycontrasting positive and negative love and by associating 'amae' withthe latter. Positive love is a reaching for the loved object; negativelove is creating, by showing 'amae' behavior, a kind of emotionalvoid or suction that will attract the loved object and make it reach

192 FRANK R. ANKERSMIT

for you. With its 'amae' behavior the child displays a specific kindof dependence aimed at making the mother identify with the child.In negative love and in 'amae' we encounter once again the wish tomake the outside world identify with the subject, a wish we discusseda moment ago in the form of 'rengo-sei'. In both cases there is a psy­chological or emotional inversion.

It is Doi's conviction that the concept of 'amae' is essential for aproper understanding of the nature of Japanese politics and he backshis claim by presenting the Emperor as the embodiment of thenation's desire to 'amaeru'. What 'amae', taken politically, aims atis establishing a relation of dependence - hence the title of his book.But it is a peculiar kind of dependence which is in some respectsdifferent from the notion of dependence we talked about in connec­tion with Constant's Adolphe. This kind of dependence has nothingto do with, for example, the relation between a medieval knight andhis feudal lord. It is, writes Doi, the kind of dependence we mustascribe to the Japanese Emperor. The Emperor depends entirely onhis subordinates who shoulder all actual responsibilities and fulfilall the actual tasks of government: 'the Emperor is in a position toexpect that those about him will attend to all matters great and small,including, of course, the government of the country. In one sensehe is entirely dependent on those about him, yet status-wise it is thoseabout him who are subordinate to the Emperor. When his degree ofdependence is considered, he is no different from a babe in arms, yethis rank is the highest in the land, a fact which is surely proof ofthe respect accorded to infantile dependence in Japan' .85 TheEmperor is the embodiment of the Japanese desire for 'amae': thatis, for combining complete dependence with being the center ofeverything. And that is why we find at the center of Japanese politicsan emptiness that is occupied by a person as helpless and depen­dent as a newborn baby; indeed, it is precisely this utter helpless­ness and dependence which qualifies him for this position. How farare we removed here from the 'roi soleil' and his modern succes­sors in the West!Political power in the West owes its legitimacy, its visibility and

its origins to the fact that it is reducible to a well-defined center, beit God, God's lieutenant on earth, the absolute monarch, the peopleor, in constitutional practice, that point where parliament and gov­ernment meet. Through this reducibility political power in the Westis always its own double: it always has - or at least ought to have

METAPHOR IN POLITICAL THEORY 193

- its shadow or counterpart in its origins or foundations. Power inthe West has a fondness for the cloak of representation; Westernpower always prefers to represent something else, God, a socialorder, the people, history, the structure of the human being or ofsociety. This is where Western power differs from that of Japan.Japanese power has no double, is not reducible to certain origins orfoundations: it is what it is. Yet Japanese power is no less well­organized, no less subtle than that of the West. It would be a gravemistake to interpret this kind of anonymous power, as we find it ona large scale in Japan, with chaos or mere arbitrariness. So it mayappear if we forget to abrogate our Western, heliotropic presuppo­sitions. But anonymous power is very well capable of organizing acomplex social and political world.No one has expressed this better than Roland Barthes, who usedthe map of Tokyo as metaphor for the Japanese empty center. Tokyois a highly complex city without a clear and well-defined plan; thestranger has the greatest difficulty in finding his way about. Roadsigns are often lacking and only elaborate maps drawn for thepurpose by inhabitants show the stranger how to get from one placeto another. The only obvious landmark in the city is the center. Butit is a secret and forbidden center, since it is the palace of theEmperor. 'Tokyo possede bien un centre, mais ce centre est vide.Toute la ville tourne autour d'un lieu ala fois interdit et indifferent' .86And yet this ill-ordered city, turning around an empty center that hasnothing to hide, this urban metaphor of the Japanese 'system ofpower' has presently become the economic capital of the modernworld.

VII. CONCLUSION

My point of departure in this essay has been the controllability ofthe State by the statesman - or of civil society by the State, if, unlikePlato, we decide to distinguish between the State and civil society.We have seen that political metaphor is instrumental in securing thiscontrol. More specifically, metaphor creates the distance between thepolitical helmsman and social and political reality required for allmeaningful political action. Distance stabilizes political reality. Thus,metaphor assumes the form of the heliotrope, as it so often has inthe history ofWestern political thought; an initially diffuse politicalreality will harden into an objective, external reality like the physical

194 FRANK R. ANKERSMIT

reality investigated by the scientist. Political metaphor - or ratherthe heliotrope - and political reality are each other's counterpartsand presuppose one another. Hence the permanent tendency ofWestern political thought to develop political systems pretending tobe accurate reflections of the actual nature of political reality andits no less permanent ambition to claim for these political systemsthe capacity to put an end to all politics. For if a political systemcorrectly reflects the nature of political reality it can show us theway to a situation in which future political action will no longer beneeded. This is obviously true for Plato's ideal State, for the naturallaw philosophies of the early modern period and for their subsequentnineteenth-century historicizations as found in Hegel or Marx.8?Machiavellianism and 'raison d'etat' doctrines are the exceptions tothis rule.Thanks to this metaphorical distance most Western political

philosophies imply and emphasize referentiality: they always aretheories about a supposedly theory-independent political reality andfrom this metaphorical distance they tell us how best to deal withit. Here I would like to recall the tendency of Western political phi­losophy we observed in the previous section to see political poweras representative of something outside itself (whether it be God, thenature of the body politic or the people). In other words, the distancebetween political language and a heliotropically defined politicalreality is in the Western tradition not merely a fine epistemologicalpoint about the relation between words and things: above all thisdistance is the birthplace of the kind of political power we know inthe West. Political action and political power only become possibleafter the two have been separated in this way.88 By seeing wordsand things as mutually independent in the sphere of politics, Westernpolitical theorists implicitly defined the kind of political power thatwould rule our world. Political power in the West is the power ofthe heliotropical word that presents itself as independent of aheliotropically defined political reality. Political power is the powerof language - a thesis, of course, that has been elaborated adnauseam by countless theorists of political ideology and byFoucault's followers.However, in contrast to the power of language we find in 'the

Japanese system' the language of power. As the Japanese counter­part to the political metaphor of the West we may above all think ofthe 'haiku'. I immediately grant that haikus, these unpretentious,

METAPHOR IN POLITICAL THEORY 195

homely and often rather trivial three-line poems, seem very farremoved from the sphere of politics. But if our analysis of politicalmetaphor has taught us anything it is that we must put aside ournaivete with regard to the political significance of linguistic phe­nomena. Moreover, if the haiku seems irrelevant to us from a polit­ical point of view, that may well be because the haiku prevents usfrom recognizing its political import by being the very opposite ofpolitical metaphor.Indeed, I propose to consider the haiku here for no other reason

than that its contrast with political metaphor is so striking. I followhere Roland Barthes. In a few brilliant pages Barthes argued thatthe haiku effects 'une exemption du sens' ,89 an abstention fromgiving meaning to reality; or, at least, the haiku suggests 'un sensObStrue'.9O Where political metaphor gives meaning to reality or evendefines it, the haiku seems to have no other purpose than to takemeaning back again. For the haiku possesses no clear, developed,rhetorical meaning as does poetic language in the West; the haikudoes not aim at expressing something deep or universal about eitherthe world or ourselves. Whereas language in the West wants to eraseitself by being transparant with regard to an underlying reality towhich it refers and that is represented by it, the haiku is merely sug­gestive of an interplay of language with itself. As Barthes put it,language in the West has the aspiration 'de suspendre Ie langage,non de Ie provoquer' - whereas the aim of the haiku is this provo­cation of language.91 And, most importantly, the haiku achieves thisprovocation of language by its having no pronounced sense ormeaning itself. Barthes here compares the haiku to a self-effacinghost who invites his guests to make themselves as comfortable aspossible.92 Precisely because the haiku is a kind of semantic void, asemantic emptiness, it can provoke a circulation of language andmeaning. Because of its high degree of readability, its almost plati­tudinous openness, the haiku allows any number of associations orinterpretations. The haiku is like an exchangeable coin going fromhand to hand, which everyone can use for his own purposes.According to Barthes this emptiness of the haiku's semantic centeris even typical of the Japanese use of the sign in general. Is Japanthen, Barthes asks, 'l'empire des signes?' and he answers: 'oui, siI' on entend que ces signes sont vides et que Ie rituel est sans dieu' .93So it is with political power. Power does not represent in Japan;

it is not the substitute for, or the double of another reality behind

196 FRANK R. ANKERSMIT

which it tries to hide itself. Power is just what it is. Power goesfrom one hand to the other like money or the haiku and its move­ments do not reflect a deeper structure of social or political reality.Thus, in the transition from the political metaphor of the West tothe Japanese haiku we witness the transition from a Foucauldianpower of language to the language of power, to a power with nodouble. This power, which is rather a speaking than a speaking about,we may expect to find in societies without deep social cleavages andwhere even the scars of earlier social strife and conflict tend to dis­appear. For social cleavages and conflicts were always the paramountproducers of political metaphor. It is an uncertain kind of power inthat it no longer urges us to follow the path of history as definedwithin some grand historical meta-narrative. But in spite of its uncer­tainty it is quite capable of, and even ideally suited to, structuringthe post-industrial societies of the present and of the near future.No less than language itself the language of power is a subtle andrefined structure that permits us to tell any number of stories withoutpresupposing anyone of them.

NOTES

1 Hobbes recognizes four 'abuses' of speech, of which the second is metaphor. SeeHobbes, T. (1970), p. 13.2 This view of the role of metaphor in scientific discourse has inspired McCloskey.See McCloskey, D. (1985); see also the article by Pen in this volume.3 Connolly, W.E. (1983), p. 23.4 Ball, T. (1988).5 Op. cit., pp. 83ff. A similar idea is the main thesis of Spragens, T. (1973).6 Ball, T. (1988), pp. 22-47.7 White, lB. (1987), p. 310. See also White, lB. (1985), Ch. 2.8 White, J.B. (1987), p. 316.9 If we wish to understand Plato's political philosophy, according to Klosko, 'theproper view is that the philosopher king is given an end at which to aim while hispolitical task lies in devising the proper means'. See Klosko, G. (1986), p. 171.Louis conjectures that it was Aeschylus' Seven against Thebes where Plato dis­covered the metaphor of the State as a ship. See Louis, P. (1945), pp. 155, 156.10 Walzer, M. (1983), pp. 284-287.II Bambrough, R. (1970), p. 105: 'he [Plato FA] obscures the fact that, in politicsas well as at sea, the theoretical knowledge and the practical ability of the navi­gator do not come into play until the destination has been decided upon'.12 There is a difficulty here that cannot be ignored. The difficulty presents itselfmost clearly when Witteveen explains Plato's metaphor as follows: 'the agent func­tioning as the steering agent within the approach is "the State" and the ship that

METAPHOR IN POLITICAL THEORY 197

is steered is "civil society'" (my translation). See Witteveen, W,J. (1985), p. 25.Plato did not, of course, distinguish between State and civil society. So we haveto be careful if we transpose Plato's metaphor to modem times. Bambrough explic­itly embraces Plato's metaphor insofar as the metaphor suggests that the State (orcivil society) must be steered. See Bambrough, R. (1970), p. 105.13 In a most illuminating essay Van Gunsteren explores the implications of Plato'smetaphor for our conception of the State's capacity to learn from political experi­ence. See Gunsteren, H.R. van (1985).14 Annas,1. (1981), p. 179.15 Op. cit., pp. 103, 104.16 Op. cit., p. 179.17 Eliot, G. (1965), p. 111.18 SchOn, D.A. (1979), p. 256. See also SchOn, D.A. (1983), pp. 182-185.19 Schon, D.A. (1979), p. 259.20 Op. cit., p. 255. See also Ankersmit, F.R. (1983), p. 216 and Ankersmit, F.R.(1981), p. 464.21 SchOn, D.A. (1979), pp. 262ff.22 Ankersmit, F.R. (1983), Ch. 5.23 Aristotle (1951), pp. 77, 78.24 See, for example, Burke, K. (1969), pp. 503, 504.25 Derrida,1. (1986), p. 243.26 Ibid.27 Ibid.28 Richards, LA. (1971), pp. 96, 97. For a comment on Richard's terminology,see Black, M. (1962), p. 47.29 See especially the last section of Derrida's essay referred to in note 25 and thetranslator's note 2 on p. 209.30 Nelson, J.S. (1987), p. 1.31 Ankersmit, F.R. (1983), p. 134-140.32 One may think of the influential work of Hayden White and Paul Ricoeur onthe role of metaphor in the humanities.33 Danto, A.C. (1983), pp. 165-209. The issue can best be elucidated with refer­ence to historical representation; see Ankersmit, F.R. (1988) and Ankersmit, F.R.(1989).34 Kossmann, E.H. (1987), pp. 134, 135.35 Rawls, J. (1972), p. 12.36 Kress, P. (1983); Gunnell, J.G. (1986), pp. 36ff.37 Kant, I. (1956), p. 374.38 Wittgenstein, L. (1971), section 5.633.39 Op. cit., section 5.64.40 At first sight Plato's metaphor seems incompatible with political representation:I here call to mind the objections of Walzer and Bambrough to the metaphor men­tioned in section II. We should note, however, that the notions of distance and unitysuggested by Plato's metaphor are in harmony with how political representationhas most often been conceived since the end of the eighteenth century. This con­ception of political representation as defended for example by Burke and Sieyes,is called the mandate view of representation. Within this conception the represen-

198 FRANK R. ANKERSMIT

tative is not the delegate, but the mandate of those he represents: the autonomy thatis thereby accorded to the representative embodies the 'distance' between the rep­resentative and the represented. Next, as Burke already emphasized, since the rep­resentative is not tied to those he represents, the representative represents the nationand not merely those who voted for him. This is where representation gives unity.See Pitkin, H.F. (1967); Ankersmit, F.R. (1987).41 Maier, C.S. (1987), p. 11.42 Keane, J. (1988), p. 6.43 Ibid.44 Offe, C. (1987), p. 91.45 Huntington, S.P. (1975).46 Constant, B. (1980), p. 501. The distinction can already be found in a manu­script written by Constant around 1796: 'il y a au contraire une partie de I'exis­tence humaine qui, de necessite, reste individuelle et independante, et qui est dedroit hors de toute competence sociale. La souverainete n'existe que d'une manierelimitee et relative'. See op. cit. p., 271. The distinction was anticipated by sucheighteenth century authors as Montesquieu, Hume, De Lolme and, above all, byFerguson; when Constant was in Edinburgh in 1787 he may have become aquaintedwith Ferguson's ideas on the subject. See Holmes, S. (1984), pp. 29-34 and Keane,J. (1988), pp. 35-73.47 The subject is explored in an illuminating way in Holmes, S. (1984), Ch. 4.48 Op. cit., p. 24.49 Arendt, H. (1975), pp. 102ff.50 Constant's acceptation of pretext and hypocrisy is certainly not unconditional.In his major work, De l'esprit de conquete et de l'usurpation, Constant harshlycondemns the political hypocrisy to be expected from bellicose governments: iftheir mentality becomes universal: 'Ie genre humain reculerait vers ces temps dedevastation qui nous semblaient l'opprobre de l'histoire. L'hypocrisie seule en feraitla difference; et cette hypocrisie serait d'autant plus corruptrice que personne n'ycroirait. Car les mensonges de I'autorite ne sont pas seulement funestes quand ilsegarent et trompent les peuples: its ne Ie sont pas moins quand its ne les trompentpas'. See Constant, B. (1980), pp. 135-136.51 Holmes, S. (1984), p. 167.52 See Todorov, T. (1977); this chapter is more convincing about the performa­tive character that speech may have than much that has been written in the wakeof Austin. If only for this reason one may readily agree with Mrs Ainslie'sview that: 'Constant a ecrit dans les cent pages d'Adolphe un livre dont aucumcritique n'epuisera les multiples significations'. See Ainslie, A. (1981), p. 78.53 Holmes, S. (1984), p. 13.54 Constant, B. (1957), p. 86.55 Op. cit., p. 37.56 Op. cit., p. 116.57 If we think of Plato's metaphor we shall recognize that with the dissolution ofthese boundaries the unity of the civil society will also be threatened. Here onemay observe a link between Plato's metaphor and the metaphor that is dominantthroughout Constant's writings. It has been pointed out (e.g. by Paul Delbouille)that, on the whole, Constant's clear, Voltairian prose is inhospitable to metaphor.Nevertheless, Markus Winkler has demonstrated that the metaphor of dust ('pous-

METAPHOR IN POLITICAL THEORY 199

siere' or 'sable') recurs again and again in Constant's writings and, moreover,that it often has a political connotation. To cite one example: 'quand chacun estson propre centre, tous sont isoles, it n'y a que de la poussiere. Quand l'oragearrive, la poussiere est de la fange'. See Winkler, M. (1981), p. 10.58 See Wolferen, KG. van (1989), pp. 5ff. Some of the main themes of this excel­lent study, which is of as much interest to the political philosopher as to the studentof Japan, can already be found in Wolferen, KG. van (1982), 'Reflections on theJapanese system'. Survey 26.60 Wolferen, KG. van (1989), p. 241; see also pp. 9, 10.61 Richardson, B.M. (1974), pp. 230ff., especially, p. 234.62 Bellah, R.N. (1957), pp. 14ff.63 Najita, T. (1974), pp. 2ff., 147, 148.64 Doi, T. (1981), pp. 31ff.65 Bellah, R.N. (1957), p. 122.66 Ono, S. (1962), p. 23.67 Barthes, R. (1970), p. 58. The spirit of Saussure is all-pervasive in this remark­able book on Japan; perhaps Japan is the kind of society we may expect if theSaussurian notion of the sign becomes socialized or politicized.68 Bellah, R.N. (1957), p. 87. Ono, S. (1962), p. 76.69 This is one of the leading ideas in Van Wolferen's book; see for example Ch.8. Bellah also emphasized that in Japan religion, politics and the social order areintertwined in a way inconceivable in the West. See Bellah, R.N. (1957), pp. 104,192.70 Wolferen, K.G. van (1989), p. 5.71 Op. cit., pp. 48, 49.72 Nakane, C. (1970), p. 69.73 Hsu, F.L.K (1975), p. 212.74 Pempel, TJ. (1982), p. 3.75 Wolferen, KG. van (1989), p. 27.76 Constant, B. (1980), p. 280.77 Wolferen, KG. van (1989), p. 28.78 Nakane, C. (1970), p. 65.79 Najita, T. (1974), p. 117.80 At the end of this book Sandel sums up his doubts about the deontologicalapproach of political problems as found in Rawls in the following way: 'toimagine a person incapable of constitutive attachments (... ) is not to conceivean ideally free and rational agent, but to imagine a person wholly without char­acter, without moral depth. For to have character is to know that I move in ahistory I neither summon nor command, which carries consequences none theless for my choices and conduct. It draws me more closer to some and moredistant from others; it makes some aims more appropriate, others less so'. SeeSandel, M. (1982), p. 172. It is interesting that Sandel defines these 'constitu­tive attachments' which are systematically ignored by the deontologists, in termsof dependency. Surely, Adolphe could never retire behind Rawls's 'veil of igno­rance' .81 See Doi, T. (1973), where the theme of 'amae' is related to the distinctionbetween appearance and reality ('tatemae' and 'honne').82 The verb 'hohitsu' is closely related to 'amae' as explained by Doi. When Doi

200 FRANK R. ANKERSMIT

translates 'hohitsu' as 'to assist', he does so in a context that is ambiguous as towhether the positive associations of 'assistance' lie in giving assistance or in beingassisted. And the context suggests that this ambiguity is precisely what 'hohitsu'(and 'amae') is about. See Doi, T. (1981), p. 58.83 Op. cit., p. 72.84 Op. cit., p. 29. The infantile origins of 'amae' are discussed in DeVos, G. (1985),especially, pp. 147-167.85 Doi, T. (1981), p. 58.86 Barthes, R (1970), p. 45.87 Ankersmit, F.R (forthcoming).88 Ankersmit, F.R (1987), p. 375.89 Barthes, R (1970), p. 11O.90 Op. cit., p. 93.91 Op. cit., p. 93.92 Op. cit., p. 91.93 Op. cit., p. 148.

BIBLIOGRAPHY

Ankersmit, Frank R: 1983, Narrative Logic, Martinus Nijhoff, The Hague.Ankersmit, Frank R.: 1981, 'Een moderne verdediging van het historisme',

Bijdragen en mededelingen betreffende de geschiedenis der Nederlanden 96,453-474.

Ankersmit, Frank R: 1987, 'Politieke representatie. Betoog over de esthetischestaat', Bijdragen en mededelingen betreffende de geschiedenis der Nederlanden102, 358-379.

Ankersmit, Frank R.: 1987, 'Referendum of representatie?', Namens 2, 17-25.Ankersmit, Frank R: 1988, 'Historical Representation', History and Theory XXVII,205-229.

Ankersmit, Frank R.: 1989, The Reality Effect in the Writing of History, Noord­Hollandsche, Amsterdam.

Ankersmit, Frank R: (forthcoming), 'In Search of the Political Object: Stoic andAesthetic Political Philosophy'.

Annas, Julia: 1981, An Introduction to Plato's Republic, Clarendon Press, Oxford.Arendt, Hannah: 1965, De revolutie, Het Spectrum, Utrecht.Aristotle: 1951, Aristotle's Theory ofPoetry and Fine Art, S.H. Butcher (tr.), DoverPublications, New York.

Ball, Terence: 1988, Transforming Political Discourse, Clarendon Press, Oxford.Bambrough, Renford: 1970, 'Plato's Political Analogies', in Peter Laslett (ed.),

Philosophy, Politics and Society I, Basil Blackwell, Oxford, 98-116.Barthes, Roland: 1970, L'empire des signes, Flammarion, Paris/Geneve.Bellah, Robert N.: 1957, Tokugawa Religion, Free Press, Glencoe.Black, Max: 1962, Models and Metaphors, Cornell University Press, Ithaca.Burke, Kenneth: 1969, A Grammar of Motives, University of California Press,Berkeley.

Connolly, William E.: 1983, The Terms ofPolitical Discourse, Princeton UniversityPress, Princeton.

METAPHOR IN POLITICAL THEORY 201

Constant, Benjamin: 1957, Adolphe, Cecile. Le cahier rouge, Gallimard, Paris.Constant, Benjamin: 1980, De la liberte chez les moderns. ecrits politiques, Livrede poche, s.l.

Danto, Arthur C.: 1983, The Transfiguration of the Commonplace, HarvardUniversity Press, Cambridge, Massachusetts.

Derrida, Jacques: 1986, Margins ofPhilosophy (Alan Bass tr. with additional notes),Harvester Press, Brighton.

Devos, George: 1985, 'Dimensions of the Self in Japanese Culture', in AnthonyJ. Marsella, George Devos and Francis L.K. Hsu (eds.), Culture and Self. Asianand Western Prespectives, Tavistock Publications, London, 141-185.

Doi, Takeo: 1971, The Anatomy of Dependence, Kodansha International, Tokyoand New York.

Doi, Takeo: 1973, 'Omote and ura', Journal of Nervous and Mental Disease 157,258-261.

Eliot, George: 1976, Middlemarch, Penguin Books, Harmondsworth.Fairlie, Alison: 1981, Imagination and Language, Cambridge University Press,Cambridge.

Gunnell, John G.: 1983, 'In Search of the Political Object', in John Nelson (ed.),What Should Political Theory Be Now?, State University of New York Press,Albany, 25-52.

Gunnell, John G.: 1986, Between Philosophy and Politics. The Alienation ofPolitical Theory, University of Massachusetts Press, Amherst.

Gunsteren, Herman R. van: 1985, 'Het leervermogen van de staat', in Marcus A.P.Bovens and Willem J. Witteveen (eds.), Het schip van staat, Tjeenk Willink,Zwolle, 53-75.

Hobbes, Thomas: 1970, Leviathan, Everyman's Library, London.Holmes, Stephen: 1984, Benjamin Constant and the Making ofModern Liberalism,Yale University Press, New Haven.

Huntington, Samuel P.: 'The United States', in Michael Crozier (ed.), The Crisisof Democracy, New York University Press, New York, 135-157.

Hsu, Francis L.K.: 1975, Iemoto, The Heart of Japan, Halstead Press, New York.Kant, Immanuel: 1956, Kritik der reinen Vernunft, Felix Meiner Verlag, Hamburg.Keane, John (ed.): 1988, Civil Society and the State, Verso, London.Klosko, George: 1986, The Development ofPlato's Political Philosophy, Methuen,London.

Kossmann, Ernst H.: 1987, 'The Singularity of Absolutism', in Ernst H. Kossmann,Politieke theorie en geschiedenis, Bert Bakker, Amsterdam, 127-139.

Kress, Paul: 1983, 'Political Theorizing in the Late Twentieth Century', in JohnNelson (ed.), What Should Political Theory Be Now?, State University of NewYork Press, Albany, 104-127.

Louis, Pierre: 1945, Les metaphores de Platon, Imprimeries Reunies, Rennes.Maier, Charles S.: 1987, 'Introduction', in Charles S. Maier (ed.), Changing

Boundaries of the Political, Cambridge University Press, Cambridge, 1-24.McCloskey, Donald: 1985, The Rhetoric of Economics, University of WisconsinPress, Madison.

Najita, Tetsuo: 1974, The Intellectual Foundations of Modem Japanese Politics,Phoenix Press, Chicago.

Nakane, Chie: 1970, Japanese Society, Penguin Books, Harmondsworth.

202 FRANK R. ANKERSMIT

Nelson, John, Megill, Allan and McCloskey, Donald (eds.): 1987, The Rhetoric ofthe Human Sciences, University of Wisconsin Press, Madison.

Offe, Claus: 1987, 'Challenging the Boundaries of Institutional Politics: SocialMovement Since the 1960s', in Charles S. Maier (ed.), Changing Boundariesof the Political, Cambridge University Press, Cambridge, 63-107.

Ono, Sokyo: 1962, Shinto. The Kami Way, Sadao Sakamoto, Tokyo.Pempel, Theodore J.: 1982, Policy and Politics in Japan. Creative Conservatism,Temple University Press, Philadelphia.

Pitkin, Hannah F.: 1967, The Concept ofRepresentation, University of CaliforniaPress, Berkeley.

Rawls, John: 1972, A Theory of Justice, Clarendon Press, Oxford.Richards, Ivor A.: 1971, The Philosophy of Rhetoric, Oxford University Press,Oxford.

Richardson, Barnard M.: 1974, The Political Culture of Japan, University ofCalifornia Press, Berkeley.

Sandel, Maurice M.: 1982, Liberalism and the Limits of Justice, CambridgeUniversity Press, Cambridge.

Schon, David A.: 1979, 'Generative Metaphor and Social Policy', in AndrewOrtony (ed.), Metaphor and Thought, Cambridge University Press, Cambridge,254-284.

SchOn, David A.: 1983, The Reflective Practitioner. How Professionals Think inAction, Basic Books, New York.

Spragens, Thomas: 1973, The Politics of Motion. The World of Thomas Hobbes,University Press of Kentucky, Lexington.

Todorov, Tzvetan: 1977, 'Speech According to Constant', in Tzvetan Todorov, ThePoetics of Prose, Clarendon Press, Oxford, 120-154.

Walzer, Michael: 1985, Spheres of Justice, Basil Blackwell, Oxford.White, James B.: 1985, Heracles' Bow: Essays on the Rhetoric and Poetics of the

Law, University of Wisconsin Press, Madison.White, James B.: 1987, 'Rhetoric and Law: The Arts of Cultural and CommunalLife', in John Nelson, Allan Megill and Donald McCloskey (eds.), The Rhetoricof the Human Sciences, University of Wisconsin Press, Madison, 298-319.

Winkler, Markus: 1981, 'Constant et la metaphore de la poussiere', AnnalesBenjamin Constant 2, 1-24.

Witteveen, Willem J.: 1985, 'Dokteren aan het schip van staat', in Marcus A.P.Bovens and Willem J. Witteveens (eds.), Het schip van staat, Tjeenk Willink,Zwolle, 23-53.

Wittgenstein, Ludwig: 1971, Tractatus logico-philosophicus, Suhrkamp Verlag,Frankfurt am Main.

Wolferen, Karel G.: 1989, The Enigma of Japanese Power, Macmillan, London.

NOTES ON CONTRIBUTORS

Frank R. Ankersmit is Professor of Intellectual and TheoreticalHistory, Department of History, University of Groningen.

Maria Luisa Barbera, Professoressa Associata di Storia dellaStoriografia Filosofica, Universita di Siena, Italia.Sandro Briosi, Professore Ordinario di Letteratura Italiana, Univer­sita di Siena, Italia.

David E. Cooper is Professor of Philosophy at the University ofDurham.

Arthur C. Danto is Johnsonian Professor of Philosophy at ColumbiaUniversity in New York University.

Mary B. Hesse, PhD, is Emeritus Professor of Philosophy ofScience, Department of History and Philosophy of Science,University of Cambridge, England.

Samuel R. Levin, PhD, is Distinguished Professor Emeritus ofEnglish and Linguistics, The Graduate School, City Universityof New York.

Kuno Lorenz, Dr. Phil., is Professor of Philosophy, UniversWit desSaarlandes, Saarbrticken, Germany.

Hans Mooij is Professor Emeritus of Comparative Literature andPhilosophy of Literature, Department of Comparative andGeneral Literature, University of Groningen.

Jan Pen is Professor Emeritus, Faculty of Economics and LawSchool, University of Groningen.

Richard Wollheim, MA (Oxon.) is Mills Professor of IntellectualPhilosophy and Civic Polity, University of California, Berkeley,and Professor of Philosophy and the Humanities, University ofCalifornia, Davis.

203

INDEX OF NAMES

Ackerman 171Aeschylus 196Agrippa, Menenius 143Ainslie, A. 198, 200Ankersmit, F.R. 13, 15, 155-202Annas, J. 160, 197,200Aquinas, St Thomas 59Arendt, H. 178, 198, 200Aristotle 21, 33, 34, 50, 52, 58,59, 61, 65, 71, 96, 165, 197,200

Auden, W.H. 75Austin, J.L. 76, 79Avogadro, A. 52

Ball, T. 156, 157, 196, 200Bambrough, R. 159, 196, 197,200Barbera, M.L. 13, 15, 143-154Barnes,1. 34,35Baroni, A. 153Barthes, R 45, 47, 188, 193, 195,199, 200

Bass, A. 166Bataille, G. 45, 46Beardsley, M.C. 1, 16Bellah, RN. 199, 200Bergmann, M. 71, 79,80Bichat, M.F.X. 13, 146, 153Binkley, T. 79, 80Black, M. 1, 3, 16, 69, 71, 79, 80,197,200

Bossuet,1.B. 182Boughton, A. 24-28, 31Briosi, S. 10, 11, 15, 127-136Broch, H. 9,99-106, 109Broussais, FJ.V. 144, 147, 150,153

Brown, 1. 144BUhler, K. 1, 15, 16Burke, K. 197, 198,200Burks, A.W. 109

205

Bush, G. 30Butcher, S.H. 200Byron, G.G. Lord 38

Cabanis, PJ.G. 144, 145, 152, 153Camus, A. 103, 129, 136Canguilhem, G. 153Caravaggio 25Carnap, R 28, 77Carr, D. 168Carroll, L. 25Cassirer, E. 2, 16Chardin, J.B.S. 121Cohen, E. 69Cohen, T. 79, 80, 121Colemann, W. 153Coleridge, S.T. 92Comte,A. 13,146,150,152,153Connolly, W.E. 156, 157, 196, 200Constant, B. 14, 177-182, 185,189, 192, 198-201

Cooper, D.E. 5-7, 14, 16,37-47,50,55-58,63,65,66,69,79,80

Copernicus, N. 51Croce. B. 127, 136Cusa, Nicholas of 59Czerny, R 17

Dalton, J. 52Danto, A.C. 4, 5, 10, 14, 16,21-35,69,169,197,201

Darwin, C. 141Daumier, N.V. 23,27Davidson, D. 1,2,4, 7, 9, 10,14 16,29,34,35,47,69,76,77,79, 80, 115, 125

Delbouille, P. 198Derrida,1. 22,165, 166, 171, 179,197, 201

Descartes, R 14, 171

206 INDEX OF NAMES

Devos, G. 201Dickinson, E. 82Disraeli, B. 7, 8, 90Doi, Takeo 191, 192, 199, 201Dostoyevsky, F. 180Dowden, S.O. 109Dummett, M. 41,47Dworkin, R. 171

Eco, U. 128, 134, 136Einstein, A. 51Eisenstein, S. 24Eliot, G. 162, 197,201Elliott, R.K. 79, 80

Ferguson, A. 198Ferruci, F. 129, 134, 136Filmer, R. 182Fodera, M. 153, 154Fodor, J. 32Fogelin, RJ. 3, 16, 74, 79, 80Foucault, H. 155, 184, 185Frankenstein 147Freud, S. 3, 124, 167Friedman, M. 138

Gadamer, H.-G. 2, 6, 16, 50,55-63, 65, 66, 136

Galileo 51Gallie, W.B. 156Genette, G. 11, 128-130, 136Gioberti, V. 150, 153, 154Giogione 122, 123Gombrich, E.H. 169Goodman, N. 9, 16, 33, 71, 73, 74,76, 77, 79, 80, 95-99, 102, 105,106, 108, 109

Gunnell,1.G. 171, 197,201Gunsteren, H.R. van 197,201

Habermas, 1. 176Hahnemann, S. 144, 149, 153, 154Hegel, G.W.F. 194Heidegger, M. 3, 45-47Hempel, C.G. 78Hesse, M.B. 5, 6, 12, 15, 16,49-66, 141

Hobbes, T. 43,49, 67, 145, 146,155, 156, 158, 171-173, 190, 196,201

Hofmannsthal H. von 43HOlderlin F. 46Holmes, S. 178, 198, 20 IHopkins, G.M. 82Hsu, EL.K. 199,201Huntington, S.P. 198, 201Husserl, E. 77, 130, 136

James I 182Johnson, M. 3, 16, 80Joyce, J. 105

Kant, I. 7, 8, 14, 52, 83, 84, 93,95, 171, 172, 178, 197,201

Keane,1. 175, 198,201Keats, 1. 84, 85, 88, 92Keynes, J.M. 12, 137, 139, 140,141

Kierkegaard, S. 5, 42Kinnock, N. 70Kipp, D. 79, 80Kittay, E.F. 1, 16,47, 71, 72, 74,79, 80

Klosko, G. 196,201Kooning, W. de 10, 121-124Kossmann, E.H. 169, 197,201Kress, P. 171, 197,201Kuhn, T. 52, 61, 63Kuipers, TAP. 79

Lakoff, I. 3, 16Langer, S. 16Lavoisier, A.L. 50, 52Leibniz, G.F.W. von 51,65Levin, S.R. 7, 8, 14, 16, 80, 81-93Lewis, C.1. 8, 95Locke,J. 67,145,155Lolme, de 1.H. 198Lombrosco, C. 153, 154Lorenz, K. 8,9, 15,51,95-110Louis, P. 196, 201Louis Philippe 4, 22, 23, 26, 27,30,31,33

Louis XIV 4, 27, 169

INDEX OF NAMES 207

Louis XV 182Lowry, M. 39

Machiavelli, N. 194Mack, D.S. 72, 79, 80Magritte, R. 191Maier, C.S. 175, 198,201Marie-Therese, Empress 28Marsella, AJ. 201Marx, K. 12, 139-141, 146, 152,154, 186, 194

Mauthner, F. 67McCloskey, D.N. 12, 140, 142,166, 167, 196, 201, 202

Megill, A. 166,202Merleau-Ponty, M. 130, 132, 136Mew, P. 79, 80Miall, D.S. 69, 80Milton, J. 30Moleschott, J. 149, 153, 154Montesquieu, c.L. de 189, 198Mooij, U.A. 6,7, 15, 17,67-80,90, 93

Moore, F.C.T. 69Moore, F.C.T. 79, 80Myrdal, G. 138

Najita, T. 190, 199, 201Nakane, C. 199, 201Napoleon 189Nelson, J.S. 197, 201, 202Newton, Sir L 51,64,65,147Nietzsche, F. 2, 7, 17,56, 77, 85-87,89,93

Nijhoff, M. 75Novalis 78

O'Hear, A. 47Offe, C. 176, 198,201Ono, S. 188, 199, 202Ortony, A. 17,80,202Orwell, G. VB

Pareto, V. 137,142Pempel, TJ. 199,202Pen, J. 12, 15, 137-142, 196Philipon, C. 4, 22, 30, 33

Pierce, C.S. 9,95, 97, 100, 104,106, 107, 109

Pitkin, H.F. 198, 202Plato vii, 13,26, 63, 81, 101, 155,

158-169 passim, 175-185 passim,193, 196, 197

Popper, Sir K. 78Priestley, J. 52, 53Proust, M. 129, 130Putnam, H. 108

Quine, W.v.O. 3, 56, 57, 73, 75,77,79, 80

Rasori, G. 144Rawls, J. 155, 170--172, 191, 197,202

Reddy, M. 90, 93Rembrandt 25, 26Ricardo, D. 140Richards, LA. 1, 17, 165, 197, 202Richardson, B.M. 199, 202Ricoeur, P. 3, 16, 17,40, 135,136, 168, 197

Rider, A.P. 25Rigaud, H. 27Robbe-Grillet, A. 129Rosch, E. 32Rousseau, U. 145, 155, 171, 179

Sacks, S. 16, 17,80Saint-Simon, C.H. de 13, 146,154, 162

Samuelson, P. 138Sandel, M. 191, 199, 202Sartre, J.P. 77Saussure, F. de 199Say, J.B. 149, 154SchOn, D. 162-164, 197, 202Searle, J.R. 47,90,93Shakespeare, W. 30, 87-90, 92Shelley, P.B. 38,40, 82Shelley, M. 145, 146, 152, 154Sidney, Sir Philip 82Sieyes, EJ. 197Simon, L. 153, 154Smith, A. 140

208 INDEX OF NAMES

Socrates 34Sontag, S. 68, 80Spolsky, I. 79Spragens, T. 196, 202Stael, Mme A.L.G. de 180Stahlin, W. 1, 15, 17Stutterheim, C.F.P. 80

Tarski, I. 7, 41, 62, 63, 78Tennyson, A. Lord 30Thatcher, M. 39, 75Titian 10, 121-123Tocqueville, A. de 150, 153, 154Todorov, T. 180, 198, 202Truman, H.S. 188Turbayne, J. 128, 134, 136Turchetti, O. 153, 154

Urmson, J.O. 79

Verga, G. 127, 136Vermeer, J. 24Vico, G. 55Victoria RI 24-28, 31, 32

Virchow, R. 13, 151-154Voltaire 182, 183, 198Vos, G. de 201

Walzer, M. 159, 196, 202Ward, B. 139, 140Warnock, GJ. 79Weiss, P. 109Wheelwright, P. 47White, J.B. 157, 196, 197, 202Winkler, M. 198, 199,202Witteveen, WJ. 159, 196, 197,202

Wittgenstein, L. 2, 6, 45, 55, 56,60, 61, 64, 95, 98, 100, 102,104-106, 109, 172, 174, 197,202

Wolferen, K.G. van 186, 188, 199WOlfflin, H. 122Wollheim, R. 9, 10, 14, 15,113-125

Wordsworth, W. 44-46

Zurbaran, F. de 25, 26

INDEX OF SUBJECTS

absolutism 182acceptance 78action 95, 96performance of 96poetic 100recognition of 96

AdoLphe (book) 180-182, 192aesthetic idea 83idealistic 127

algebra as language 137amae 191, 192, 199analogy 51anthropological approach tometaphor 2

anthropology, philosophical 41approximation, local 53art, science and 9assertion 72metaphor as real 75

author, narrator vs. 128awareness 64

being in truth 5, 42, 43, 46belief 64body, metaphor of the 143

cartoon, political 22 ff.category, Aristotelian 96cellulism 152center, empty vs sovereign185-193

Christology 59classification 62cognition, conception vs. 91metaphor and 21-35

cognitive content 77function 67value 68

comparison, metaphor as 3

209

concept formation, natural 59conception, cognition vs. 91conceptual category, transgression of89structure vii

Concert Champetre (painting) 122context, displacement of 91projection of 91

convention 76, 77conventionalism 98corporeality 122

death 84decidability 76Declaration of Rights (French) 144decorative function 67Dichtung, philosophische 101, 105difference 61disbelief, suspension of 92discussion, metaphor in 1-17disease is disorder 147is transgression 148

disorder, disease as 147distance, political 160, 161Du degre de certitude de La

Medecine (book) 144

economics 12language and 137-142

economist as observer 137ego, bodily 124eidos 105Eidos-Einheit 104ellipsis, metaphor and 38, 39emotive function 67Emperor, Japanese 189 ff.enthymeme 31,32envy 84epistemological indeterminacy 83epistemology, ontology vs. 97

210 INDEX OF SUBJECTS

Erkennen 102Erleben 102essentialism 50explanation 72

fame 84Flaying of Marsyas (painting) 123force 147Frankenstein (book) 146freedom 177French National Convention (1793)146

Geisteswissenschaft 6Gestalt 168goal, political 160, 161green politics 176Gross National Product 141

haiku 194, 195, 196heliotrope 165, 166, 172, 176, 185,190, 193, 194

hermeneutic circle 12, 51experience 60

hermeneutics 11homeopathy 150House of Cards (painting) 121humorism 147hypocrisy 178, 179

icon, sensu Piqerce 95idea, aesthetic 83identification, epistemological 98ontological 98

II mondo creato (book) 129independence 177anatomy of 175-185

indeterminacy, epistemological 83information storage, metaphor and32

insight, metaphor as 78intention 120invitation 72

Japanese society 186 ff.journalism 12

Journey of life (painting series)121

knowledge, acquaintance 102conceptual 95-110description 102metaphor and Vll

objective 52perceptual 95-110poetry, metaphor and 81-93types of 82

L'etranger (book) 129La Jalousie (book) 129language, context free 78economics and 137-142game, sensu Wittgenstein 95,98, 106

literal 49metaphor as 50medical 144metaphor as fundamental form of54

metaphorical, social desirabilityof 78

natural, fundamentallymetaphorical 58

naturalization of 8, 98rule governed 60value laden 138

law as metaphor 157lesion 148lexical structure viilife, metaphors for 74literal comparison 38logical empiricism 100love 84metaphors for 38

meaning, literal 70meaning relation as 61metaphor as primary vehicle of

2metaphorical 10pictorial 114psychological theory of 119representational, pictorial 118

INDEX OF SUBJECTS 211

medicine, metaphor in 143-154metaphorizes politics 13

metaphor, abuse of speech 49anthropological account of 11approach 2Aristotle's view of 21artistic 87as comparison 3as fundamental form of language

5as literal statement 4as paradox Vlll

as prompt 40awareness of 64cognition and 21-35cognitive nullity of 1, 2conventional 37descriptive 54economic 140, 141ellipsis and 38, 39epistemic value of 85equivocal 90exemplifies deeper truth 5false 74falsity of 69generative 162in discussion 1-17in painting 9indifference to cognitive truth4, 5

information storage and 32intensional character of 4knowledge and VII

linguistic 115, 116literal truth and 6, 7literary semiotics and 127-136logical form of 28medical 143-154models, truth and 49-66narrative 167, 168naturalization of 89negative view of 68painting and 113-125pictorial lO, 21-28, 116spectator and 120poetry, knowledge and 81-93political theory and

155-202primacy of 2, 54projection approach toproportion 70real assertion 75reality and 8riddle and 22scheme dependent 71scientific 64semantically deviant 89"that" implicit in 30truth and 4, 37-47liberal approach 67-80

understanding complex 74uses of 9weakening force of 72

metaphorized term 119metaphorizing term 118, 119mind, politics of 22model, metaphors, truth and 49-66monster 145, 146

narration 131, 133, 134ambiguity of 130

narrative metaphor 163, 167, 168narratology 128narrator, author vs. 128naturalization of metaphor 89Nightingale (poem) 84, 85no truth no sense mechanism 74,75

novelist, world of 11nutrition 149-151

Oedipus (play) 29ontologism 98epistemology vs. 97

organic age 146organism 145, 146Orwell's problem viipainting, metaphor and 9, 113-125pathology, cellular 151persuasive function 67phenomenology 11, 130philosophy 34of language, analytic 49semiotic phase 8

212 INDEX OF SUBJECTS

phylaks 160physician, political task of 144picture is not a type 10metaphorical meaning 118

Plato's problem viiplausibility, principle of 51poet, temperament of 92poetic creation 11poetry, knowledge, metaphor and81-93

pOlesIS 100, 103, 106polis, unity of 160political philosophy, ignorance ofmetaphor in 170-175theory, metaphor in155-202

politics 12green 176metaphorizes medicine13

pragmatics, linguistic metaphor and116

pragmatism 53prompt, metaphor as 40property, private 149proposal 72prototype theory 32

rationality of economic discourse138

Real National Income 141realism, moderate 53scientific 50

RealiUitsvokabel 9, 100reality is flux 86metaphor and 8

reassignment 76res cogitans 132rhetoric 31,166,167economist's use of139-142

riddle, metaphor and 22Roi Soleil 27,31, 169, 192rule, language 60

schema 104, 105, 107schematic interpretation 71

science 34art and 9instrumentally progressive 52not empirically based 51

scientific debate, evolution of 6realism, moderate 50

self, transcendental 171, 172semantic field 71semiotic representation 99triangle 106

semiotics, literary, metaphor in127-136

sentence, declarative 72intensional 28

Shinto 187ship of state 13, 158-161 passimdisintegration of 175-185

sign action 95, 96, 98, 99similarity 61perceptual 60principle of 51

social contract 145organism 148

society, civil, State and 177, 178Japanese 186 ff.

solidism 147solipsism 174Sonnet 33(William Shakespeare) 87

space, metaphoric 92spectator, pictorial metaphor and120

State 173civil society and 177, 178dissolution of 185-193

statute as metaphor 157suggestion 72suspension of disbelief 92symbol system, naturalization of 98symptom 148

taxonomy 61theory as model 51theory ladenness 50thesis M 5, 6, 15, 54-59, 63, 64Three Ages of Man (painting) 122tissue 146

INDEX OF SUBJECTS 213

token, picture as 117, 118type vs. 117

Tokyo as metaphor 193tolerance, conceptual 92transgression, disease as 148transparency 76trope 26truth claim, weakening of 73cognitive 3, 4indifference of metaphor to4, 5

context dependence of 7everyday description of 42literal, metaphor and6.7,40,41

metaphor and 4, 37-47liberal approach 67-80

models and 49-66non propositional 53strict notion 77Tarskian 7, 41vicarious 65

type, picture as 117picture not instantiation of 10token vs. 117unity, political 160, 161universal, Platonic 63Ursprache 55-58, 63Urzeit 56usage 78

value theory 100vehicle of truth 68veil of ignorance 155. 170. 172vitalism 151

word, symbolization of 8world. comportment towards 42.43possible 134refashioning of 89

zero focus 129, 130