Lecercle - Deleuze and Language

Jean-Jacques Lecercle

Deleuze and Language

Language, Discourse, Society

General Editors: Stephen Heath, Colin MacCabe and Denise Riley

Selected published titles:

Norman Bryson VISION AND PAINTING The Logic of the Gaze

Elizabeth Cowie REPRESENTING THE WOMAN Cinema and Psychoanalysis

Theresa de Lauretis TECHNOLOGIES OF GENDER Essays on Theory, Film and Fiction

Mary Ann Doane THE DESIRE TO DESIRE The Woman's Film of the 1940s

Alan Durant CONDITIONS OF MUSIC

Jane Gallop FEMINISM AND PSYCHOANALYSIS The Daughter's Seduction

Peter Gidal UNDERSTANDING BECKETT A Study of Monologue and Gesture in the Works of Samuel Beckett

Piers Gray, edited by Colin MacCabe and Victoria Rothschild STALIN ON LINGUISTICS AND OTHER ESSAYS

Alan Hunt GOVERNANCE OF THE CONSUMING PASSIONS A History of Sumptuary Law

Ian Hunter CULTURE AND GOVERNMENT The Emergence of Literary Education

Jean Jacques Lecercle DELEUZE AND LANGUAGE

Colin MacCabe JAMES JOYCE AND THE REVOLUTION OF THE WORD Second edition

Jeffrey Minson GENEALOGIES OF MORALS Nietzsche, Foucault, Donzelot and the Eccentricity of Ethics

Laura Mulvey VISUAL AND OTHER PLEASURES

Christopher NorrisRESOURCES OF REALISMProspects for `Post Analytic' Philosophy

Denise RileyÀM I THAT NAME?'Feminism and the Category of `Women' in History

Jacqueline RosePETER PAN, OR THE IMPOSSIBILITY OF CHILDREN'S FICTION

Moustapha SafouanSPEECH OR DEATH?Language as Social Order: a Psychoanalytic Study

Moustapha SafouanJACQUES LACAN AND THE QUESTION OF PSYCHOANALYTIC TRAINING(Translated and introduced by Jacqueline Rose)

Stanley ShostakTHE DEATH OF LIFEThe Legacy of Molecular Biology

Lyndsey StonebridgeTHE DESTRUCTIVE ELEMENTBritish Psychoanalysis and Modernism

Raymond TallisNOT SAUSSUREA Critique of Post Saussurean Literary Theory

David TrotterTHE MAKING OF THE READERLanguage and Subjectivity in Modern American, English and Irish Poetry

Geoffrey WardSTATUTES OF LIBERTYThe New York School of Poets

Language, Discourse, Society Series Standing Order ISBN 0 333 71482 2 (outside North America only)

You can receive future titles in this series as they are published by placing astanding order. Please contact your bookseller or, in case of difficulty, write to usat the address below with your name and address, the title of the series and theISBN quoted above.

Customer Services Department, Macmillan Distribution Ltd, Houndmills,Basingstoke, Hampshire RG21 6XS, England

Deleuze and LanguageJean-Jacques Lecercle

# Jean Jacques Lecercle 2002

All rights reserved. No reproduction, copy or transmission of this publication may be made without written permission.

No paragraph of this publication may be reproduced, copied or transmittedsave with written permission or in accordance with the provisions of theCopyright, Designs and Patents Act 1988, or under the terms of any licencepermitting limited copying issued by the Copyright Licensing Agency,90 Tottenham Court Road, London W1T 4LP.

Any person who does any unauthorised act in relation to this publicationmay be liable to criminal prosecution and civil claims for damages.

The author has asserted his right to be identified as the author of this workin accordance with the Copyright, Designs and Patents Act 1988.

First published 2002 byPALGRAVE MACMILLANHoundmills, Basingstoke, Hampshire RG21 6XS and175 Fifth Avenue, New York, N.Y. 10010Companies and representatives throughout the world

PALGRAVE MACMILLAN is the global academic imprint of the PalgraveMacmillan division of St. Martin's Press, LLC and of Palgrave Macmillan Ltd.MacmillanT is a registered trademark in the United States, United Kingdomand other countries. Palgrave is a registered trademark in the EuropeanUnion and other countries.

ISBN 1 4039 0036 1

This book is printed on paper suitable for recycling and made from fullymanaged and sustained forest sources.

A catalogue record for this book is available from the British Library.

Library of Congress Cataloging in Publication Data Lecercle, Jean Jacques. Deleuze and language / Jean Jacques Lecercle.

p. cm. (Language, discourse, society)Includes bibliographical references and index.

ISBN 1 4039 0036 1

1. Deleuze, Gilles Contributions in philosophy of language. 2. Language and languages Philosophy. I. Title. II. Series.

P85.D45 L43 2002 401 dc21

2002025242

10 9 8 7 6 5 4 3 2 1

11 10 09 08 07 06 05 04 03 02

Printed and bound in Great Britain by Antony Rowe Ltd, Chippenham and Eastbourne

Contents

List of Figures vi

Acknowledgements vii

List of Translations viii

Introduction: Deleuze, Beckett, MeÃme Combat 1

1 The Problem of Language 8

Interlude 1 Images of Trains, Trains of Images Dickens, Deleuze, Representation 41

2 `Linguistics has done a lot of harm' 62

3 Events, Sense and the Genesis of Language 99

Interlude 2 Making Sense of Literature Joyce, cummings, Woolf 132

4 Another Philosophy of Language: the New Pragmatics 154

5 Another Philosophy of Language: Machines,Assemblages, Minority 174

Interlude 3 A Reading of Kipling's `Wireless' 202

6 Another Philosophy of Language: Style and Stuttering 218

Conclusion 247

Notes 258

Index 271

v

1

2

3

4

5

6

7

List of Figures

Correlation 56Signifying semiotics 82Semiotic square of communication 149Style 245Ambivalence towards language 254Stitching point 257Deleuzean reading 257

vi

Acknowledgements

I would like to thank the University of Cardiff, and in particular Cather-

ine Belsey and David Skilton, for giving me the time to research and write this book; Denise Riley, that paragon among editors; Emmanuelle Guattari, for the conversations that convinced me her father wasn't second best; Daniel Haines, who lent me books, and all the members of the Cardiff seminar, for their patience; NoeÈlle Batt and the Paris VIII seminar, for lending an indulgent ear to my Deleuzean disquisitions.

Part of Chapter 6 appeared in a slightly different version in L'Esprit creÂateur (Winter 1998) vol. XXX VII I: 4. I would like to thank Anne Tomiche for permission to reprint.

JEAN-JACQUES LECERCLE

vii

List of Translations

All references to works by Deleuze and Guattari are to the first French editions. However, I have provided page references to the English editions, whenever possible. The English editions are listed below in the order in which the books first appeared in French.

Proust et les signes Proust and signs, trans. R. Howard (New York: George Braziller, 1972).

Le Bergsonisme Bergsonism, trans. H. Tomlinson and B. Hab-

berjam (New York: Zone Books, 1988).

DiffeÂrence et reÂpeÂtition Difference and Repetition, trans. H. Patton (London: Athlone, 1994).

Logique du sens The Logic of Sense, trans. M. Lester and C. Stivale (London: Athlone, 1990).

L'Anti-údipe Anti-Oedipus, trans. R. Hurley, M. Seem and H. R. Lane (London: Athlone, 1984).

Kafka Kafka, trans. D. Polan (Minneapolis: Univer-

sity of Minnesota Press, 1986).

Dialogues Dialogues, trans. H. Tomlinson and B. Hab-

berjam (London: Athlone, 1987).

Mille plateaux A Thousand Plateaus, trans. B. Massumi (Minneapolis: University of Minnesota Press, 1987).

CineÂma 2: L'Image-temps Cinema 2: The Time-Image, trans. H. Tomlin-

son and R. Gulta (London: Athlone, 1989).

Foucault Foucault, trans. S. Hand (Minneapolis: University of Minnesota Press, 1988).

Le Pli The Fold, trans. T. Conley (Minneapolis: University of Minnesota Press, 1993).

Pourparlers Negotiations, 1972 1990, trans. M. Joughin (New York: Columbia University Press, 1995).

viii

List of Translations ix

Qu'est-ce que la philosophie? What is Philosophy?, trans. H. Tomlinson and G. Burchell (London: Verso, 1994).

Critique et clinique Essays Critical and Clinical, trans. D. W. Smith and M. A. Greco (London: Verso, 1998).

Introduction: Deleuze, Beckett, MeÃme Combat

Imagine a playwright who can no longer bear to write. Imagine a poet whose feelings for language have become so exquisite as to preclude their formulation in words. Imagine Hamlet so radically curtailed in its garrulousness that it is reduced to the dumb show. Imagine a state where language is exhausted in the two senses of the term, in that all the potentialities for expression have been realised and frozen into words, with the result that all potential speakers are too tired to speak.

Such is the situation in Beckett's late television plays, Quad, Ghost Trio, . . . but the clouds . . . and Nacht und TrauÈme, where language there is none: at best it is reduced to a voice off describing the props, at worst to a number of stage directions that will naturally vanish when the play is performed.

But the language hater is faced with a problem. He may quietly abandon language he may paint, he may compose music, he may dance, he may film silent films. He may thus demonstrate that certain things cannot be said, but only shown, that the best epiphanies are silent, that the capture and welcoming of the event as it emerges is best achieved without words. He may revel in the realisation that, in the realm of art at least, dumb is not daft. But in so doing he will not be actively hating language. He will not be expressing his contempt at the betrayal of his expressive needs by language, he will not be waxing eloquent at the dangers of eloquence, he will not be warning other potential speakers, in the familiar pragmatic paradox, not to speak. For of course the only way to express one's hatred and contempt for language, or one's distrust of it, is through language.

Beckett is not immune to this paradox, since his late dumb plays for television are in front of me, on my desk, in the shape of a book,1 a book that bears on its title page the indication `traduit de l'anglais par Edith

1

2 Deleuze and Language

Fournier', and where the plays are followed by a lengthy commentary by Gilles Deleuze, entitled `L'EpuiseÂ' (the exhausted). Words, words, words, and with reference to more than one natural language.

The paradox is reiterated in Deleuze's essay. The text naturally deals with the exhaustion of language in Beckett's television plays, which is to be taken, as it ought, as the point where language reaches its limit, vanishes, is no more: at the end of his trajectory, the famous playwright has finally managed to do without language altogether. But Deleuze constructs his commentary on this fascinating state of affairs around a theory of Beckett's language, even a theory of Beckett's three types of language, so that he deals with a text where only a vestigial form of language remains by constructing a theory of its (presumably absent) language.

I intend to use this paradox (the details of which I shall consider in a moment) as a parable, the parable of the problematic nature of lan-guage, of the love hate relationship its practitioners (I will not say its users) have with it, of the necessity to push it to its limits, into silence, and of the inexhaustible garrulousness that defeats exhaustion, of the constitutive effability that inhabits even the ineffable.

I shall take this paradox as my guiding thread through the philosophy of Gilles Deleuze.2 I shall treat language in his works not as a theme, but as a problem. As we shall see, Deleuze in many ways resists the linguistic turn that French and European philosophy in the 1960s and 1970s has taken, in a different but parallel form to the more famous linguistic turn taken by analytic philosophy several decades before. In other ways, especially if we take into account the specific contribution of Guattari to their common work, he is part of that turn. This is why the reader is under the impression that Deleuze always seeks to ground language in something else (sometimes it is the body, sometimes it is thought), to treat it, if not as an epiphenomenon, at least as a secondary phenom-

enon. But the same reader is also under the impression that Deleuze is always, obsessionally, returning to language. Language may be an effect of surfaces, a superficial field, but then Deleuze is the philosopher of surfaces and planes. His hostility to interpretation and to metaphor, two processes characteristically within the scope of language, is notorious. But equally notorious is his love of literature as the art of language for Deleuze, a high modernist in his approach to literature, literary texts (their literariness) are very much a matter of style and syntax. In short, we do not find in the works of Deleuze an explicit theory of language, not even a single book devoted to the question of language; but we do find a theory of meaning (in Logique du sens), a theory of poetic language

Introduction 3

and style (most explicitly in Critique et clinique), a theory of pragmatics (in Mille plateaux). The task of this book, therefore, will be to develop this paradox. As we shall see, this will involve constructing language in Deleuze as a problem, in the strict sense that he gives to this concept.

But let us return to our initial paradox (the poet that is a prey to the ineffable) and our parable (`Deleuze, Beckett, MeÃme Combat'). Language in Beckett's television plays is indeed a strange object. In Quad, the actors move on stage, but never a word they speak, so that language is restricted to the stage directions, and absent from the play itself. Its relation to the object of our admiration, the play, is that of the instruc-tions for use of the lawn mower we have just bought. In Ghost Trio, there is a voice speaking. That voice is female. The only character present on stage, however, is male. So the voice is a voice off, a murmur, mouthing the stage directions that could only be read in Quad, describing the props, which the audience can see anyway, and the movements of the figure. Only once does the voice leave its purely descriptive tone, and betray an expression of affect, when an Àh' of surprise (an affect duly noted in a stage direction) escapes her. By that time the play is nearly exhausted, and the Voice's contributions become rare and monosyl-

labic. In . . . but the clouds . . . (part of a line from a poem by W. B. Yeats not the sunniest of his poems by far) there is a male Voice and the stage

directions tell us it is the voice of the Man we sometimes see on stage although on stage he is silent. The voice describes the actions of the Man, and the expressions of a female Face, seen in close up, who at times appears on the screen. In an interesting development, the lips of the female Face are seen to utter lines from Yeats, except that, the stage directions maintain, her voice is ìnaudible'. Lastly, in Nacht und TrauÈme, language has completely disappeared. The male figure is dreaming, mostly of himself, while a male voice softly hums the last seven bars of the eponymous lied by Schubert.

It is clear, then, that at the end of his career (the plays were `written' from 1980 to 1982), Beckett has moved away from language, the stuff that drama is made of, in all possible ways. He does so by moving on to other media, music and the image; by moving towards silence as the climax, but also the substitute of language; by marginalising language, when it resists and insists by exiling it to the margins of the play, the voice off of commentary or the stage directions; by foregoing the ex-pression of affect, which is one of the functions of language, and one that is central to drama; and by renouncing interlocution, which is supposed to be what language is essentially for he only keeps language where it is useless, in its capacity awkwardly to translate visual images


into words. But who wants ekphrasis, if she can have visual access to the real thing (the plays for television make sorry and painful reading)? Even if he goes back to one of his obsessions, the systematic exhaustion of a set of logical possibilities, this exhaustion, in Quad, is visual and gestural, not linguistic as in Watt (one recalls the infinite permutations of Watt's linguistic quirks).

But when Deleuze sets out to write an essay on the plays, he carefully abstains from taking the obvious line of the vanishing of language in late Beckett. He begins by identifying, in typical Deleuzean fashion, the meaningful core of the text as a problem, the problem of exhaustion by which he means all types of exhaustion: physical exhaustion (the char-acters on stage are so tired that they can hardly move), logical exhaus-tion of possible combinations (one recalls not only Watt, but Molloy, his pockets and his 16 circulating pebbles), but also the exhaustion of a personal trajectory, of a writer's project (this exhaustion takes the form of the exhaustion of language). `L'eÂpuiseÂ' is not only the typical late Beckett character, but the author himself and, above all, language.

So, not so illogically, exhaustion in Beckett's late plays will be ac-counted for in terms of a theory of language, or rather three, with logical passage from one to the other. Language no. 1 is the language of names. It is already far from language

as we know it. Indeed, it hardly deserves the name. Language as we are familiar with it involves syntactic arrangement, the systematic construc-tion of meaning allowed by its double articulation, and proper refer-ence, as the sentences embody propositions. Not so with the language of names: name-words are disjunct atoms, their sequences form enumer-

ations or lists, not propositions, and their combination is algebraic, in the obsessional Beckettian manner, rather than syntactic. However, outlandish as this language appears to us, we shall regret its quaint familiarity when we proceed to the next one. Language no. 2 is the language of voices. The familiarity of language no.

1 is explained by the fact that there is still a subject, a speaker, who is in charge of naming. This mythical nomothetes, who lays down the linguis-tic law, offers the guarantee of a form of reference, albeit disjointed, non-propositional, to the world. Signification, or the putting together of signs in some form of logical order, is awkward; manifestation, the presence of a speaking subject, may be limited, but designation, the relation of reference between words and things, is still there. Not so with language no. 2. For in this form of language, which it is increas-ingly difficult to call `language', words have disappeared. `Doing away with words' is the task that is ascribed to it and which it fulfils. This

Introduction 5

vanishing of words has two immediate and interesting consequences. First, such language is no longer my language, since I fail not only to make sense out of it, but to grasp it as an expression of meaning. Second, meaning has of course absconded: such language, if a tongue at all, can only be a foreign tongue, one with which I am not conversant, uttered by the utterly Other. Deleuze takes advantage of this situation to sketch a non-trivial theory of the Other as possible world, whose only point of contact with the world of my reality is the Voice that no longer makes sense. Yet, foreign and alienating as this language no. 2 may be, we shall regret it when we pass on to the next and last stage, the stage of language no. 3. Language no. 3 is the language of images. Its qualification as `language'

can only be the product of a long drawn-out metaphor, for it seems to have abandoned all the characteristics we ascribe to ̀ normal' language. In language no. 2, odd as it sounded, there was still a sender, not me (as it was not my language); not even a speaker (the Voice being impersonal), but something like a source of linguistic sequences, of strings of words. In other words, where language no. 1 maintained a form of odd designa-tion, language no. 2 maintains a form of garbled manifestation. Not so language no. 3, which goes one step further in this process of askesis. We no longer have either series (the exhaustion of the logical combination of referential names, as in language no. 1) or stories (for language no. 2, being the Voice of the Other, is reminiscent and nostalgic, and sometimes inventive and fanciful). Language no. 3 consists of impersonal, singular images (both in sounds and pictures) it is the site for the emergence of images. Needless to say, this is no longer a language in any sense: signifi-cation having vanished with language no. 1 and designation with lan-guage no. 2, now it is the turn of manifestation, after which there is nothing but a heap of broken images. The obvious question is: why call it language at all? And the answer has something to do with what I have called Deleuze's high modernism. Language no. 3 has neither subject nor object, speaker nor referent. But it still has an addressee, the audience, and there is still something going on, the process of emergence of those images. This must be taken reflexively: that process, of dereliction and of emergence, is the process of language itself, when art takes it to its limits, moves it closer and closer to silence, to which it aspires, and which achieves it, in the French as in the English sense of the term. Language no. 3 is the language of the limits of language, when it turns into silence, or to another medium, music or picture.

According to Deleuze, such progress towards language no. 3 is idiosyncratic, it belongs to the development of the Beckettian oeuvre


(language no. 1 is characteristic of the novels, above all Watt; language no. 2 is to be found in novels and plays, and especially in the radio plays; language no. 3 was born in a novel, How It Is, but flourishes in these television plays). But the progress is also logical, it formulates a form of truth about language, states what is ànnoying' about it. For language as we know it, the language Beckett tries to get rid of, is not only deceitful in the ambiguity of its words, it is laden with all the paraphernalia of communication and interlocution: intentions, significations, memories and customs all of which freeze and poison our words, and stifle us, their speakers. The speaker is in constant danger of being burked by language: a wet blanket of signification smothers any attempt at expres-sion. Hence the necessity, as Deleuze phrases it, of `boring holes' on the surface of language to find out `what is hidden behind'. And only a change of medium, the combination of music and picture in Nacht und TrauÈme, can fully achieve this.

What I find fascinating in Deleuze's account of Beckett's television plays is its paradoxical aspect. We think we know what language is, how it works, on its phonetic, semantic, syntactic and pragmatic levels. And it is clear that none of the Beckettian modes of expression that Deleuze analyses can count as language. Yet this is the very term he chooses to describe them. Or again, which is another formulation of the same idea, the more language vanishes into thin air in the plays, the more its metaphysical range extends in Deleuze's commentary: metaphorical drifting and inflation compensate for literal evanescence.

The paradox is inscribed in the very nature of the text. Beckett is neither a musician nor a film director; here, he is doing his best not to be a writer, which accounts for the difficulty and pain one experiences when reading a text that is clearly not meant to be read. Yet, Deleuze's commentary chooses to address the text, not the television pro-

grammes, since it is printed in the same volume. He comments on a text that hardly deserves the name, being broadly confined to stage directions: not the best medium for indulging in the pleasures of writing and style.

But that is the whole point. The opportunity this `text' gives Deleuze is that of a reflection on the limitations of language, and the theorisa-tion of the consequent hostility to language. This is a familiar thread throughout the philosophical and religious traditions: language betrays thought, it is incapable of reproducing the emergence of the event, ekphrasis is a sorry substitute for the immediacy of the picture. Language does not, contrary to appearances and to its explicit function, foster the emergence and circulation of meaning: it freezes into common sense; it

Introduction 7

betrays the violence of the event in articulate discourse. And yet this pessimistic view of language is, at the same time, a paean to the necessity of literature, the art of language, which can capture and welcome the event, by pushing language to its limits, on that frontier where it meets other media. For there is a form of language, the language of poetry, that can, if it abandons `le vieux style', accommodate vision and music. There is, however, a condition to be fulfilled: such language must be as a foreign tongue within the material idiom language, but strange and out of kilter. Deleuze has a phrase to describe this, which is also Beckett's `non-style', his rejection of any attempt at writing: such language is `mal vu mal dit'.3

This, as you may imagine (this is where my paradox becomes a par-able) raises a number of questions. Is this merely a return to a concept of poetic language as eÂcart as divergence from normal language? Is there such a thing as poetic language, distinct in grammar and syntax from ordinary language? Are we going back to a Romantic concept of the prophesying-because-inspired poet uttering the unutterable? Is Deleuze, in his untimeliness, merely repeating the high modernist gesture half a century too late (the canon of his favourite authors might incline us to think so)? I shall argue that this literary question is not Deleuze's main concern: that its persistence in his works is only the symptom of a philosophical problem, of the problem of language, in its two, paradoxical aspects of the philosopher's necessary resistance and hostility to, but also obsession with, language. What we must begin with, therefore, is the construction of the problem of language in Deleuze's philosophy.

1 The Problem of Language

1 Incipit

The first lines of Anti-Oedipus are justly notorious:

It is at work everywhere, functioning smoothly at times, at other times in fits and starts. It breathes, it heats, it eats. It shits and fucks. What a mistake to have ever said the id. Everywhere it is machines real ones, not figurative ones: machines driving other machines, machines being driven by other machines, with all the necessary couplings and con-nections. An organ-machine is plugged into an energy-source-ma-

chine: the one produces a flow that the other interrupts. The breast is a machine that produces milk, and the mouth a machine coupled to it. The mouth of the anorexic wavers between several functions: its pos-sessor is uncertain as to whether it is an eating-machine, an anal machine, a talking-machine, or a breathing-machine (asthma attacks). Hence we are all handymen: each with his little machines. For every organ-machine, an energy machine: all the time, flows and interrup-tions. Judge Schreber has sunbeams in his ass, A solar anus. And rest assured that it works: Judge Schreber feels something, produces some-

thing, and is capable of explaining the process theoretically. Some-

thing is produced: the effects of a machine, not mere metaphors.1

The text immediately raises a question: what language-game are we playing? Am I reading a treatise of philosophy or a novel? I need points of comparison. Here are other famous first words:

1 The world is all that is the case. 1.1 The world is the totality of facts, not of things.2

8

The Problem of Language 9

It is some time ago now since I perceived that, from my earliest years, I had accepted many false opinions as being true, and that what I had since based on such insecure principles could only be most doubtful and uncertain; so that I had to undertake seriously once in my life to rid myself of all the opinions I had adopted up to then, and to begin afresh from the foundations, if I wished to establish something firm and constant in the sciences.3

An ancient English cathedral town? How can the ancient English Cathedral town be here! The well-known massive grey square tower of its old cathedral? How can that be here! This is no spike of rusty iron in the air, between the eye and it, from any point of the real prospect. What IS the spike that intervenes, and who has set it up? Maybe, it is set up by the Sultan's orders for the impaling of a horde of Turkish robbers, one by one.4

It is clear that our incipit does not belong to the same style of writing as Wittgenstein's Tractatus, with its objective, exaggeratedly scientific turn of phrase. It is also clear that it has little to do with the more human, because introspective, but still abstract incipit to Descartes's Meditations. And it is clear that the language-game being played is closest to the literary language-game of Dickens's Edwin Drood, even if the voice heard there is one crazed and fantastical through the inges-tion of opium. What is striking in the first words of Anti-Oedipus, as in Edwin Drood, is that a singular (even if dual) voice is speaking. I shall therefore for a while forget that the text purports to be the first page of a book of philosophy, and indulge in a literary reading of the pas-sage.

In a literary text (what about a text of philosophy?), the first word is often of crucial importance. Here it seems to be innocuous enough, being the pronoun ìt' (hardly more meaningful than the definite article that begins Wittgenstein's Tractatus), except that it is underlined a few lines later. And with it we encounter our first problem of translation, since the original word, in French, is the deictic cËa (`cËa fonctionne part-out . . . ' remember this is a literary analysis, and the original language of the text is not indifferent: you do not read Hamlet's monologues in translation, if you can help it). I am not criticising the translation, whose choice of ìt' is inevitable. But a deictic is not a personal pro-noun.

They have, however, common points, which justify the translator's choice, being both neuter: ìt' is what Benveniste calls a `non-personal


pronoun',5 `cËa' is the neuter form of the deictic. And they share some of their syntactic functions. Both are used for zero-topicalisation, when the sentence starts without presupposing any already known information (compare `cËa va?' and the universal ìt' of ìt is raining'). Both can be used for purposes of grammatical cataphor, announcing left-focus or extra-position (`c'est lui qui l'a fait', ìt is true that . . . '). Lastly both can be markers of simple anaphora, referring to a topic already broached (`cËa n'est pas le probleÁme' presupposes that the referent of `cËa' has been made clear in previous dialogue, and ìt is working' does the same with the referent of ìt'). The question is: which of these three syntactic functions does the `cËa'/ìt' of our text fulfil? The answer is that our occurrence seems to hesitate between the first and the third, between zero-topica-lisation (this is an absolute beginning) and anaphora (there is an impli-

cit co-text before the text, even if this is the first word of the text). In other words, our word is an occurrence of a well-known literary trope, the trope of false anaphora, where the text's beginning feigns not to be the beginning of the text. If my novel begins with the sentence `The old man was dying', the definite article suggests that the old man is already identified, and the story appears to begin in medias res.

My two initial words, `cËa' and ìt', also share one semantic function. They are both vague in their reference. As we know, vagueness, far from being a defect of reference in natural languages, is an asset.6 There are pragmatic situations where the speaker, lacking knowledge or feigning such lack, needs to refer vaguely. `Ca' and ìt' are perfect instruments for this. Here their vagueness is tropic, it marks the trope of suspense or delay (this is close to the syntactic function of zero- or empty topicalisa-tion: there is no previous information or, if there is, it is withdrawn). The effect on the reader is one of temporary exasperation: what are they talking about? They are talking about machines, of course, since the alleged reference of ìt' is given on line 4. But such definition clearly is not one (those machines are in the plural, where ìt' is singular; and the verbs of which it is the subject clearly demand an animate, not a machinic subject: machines do not fuck). So the tropes of vagueness and suspense are clearly at work.

This is also, however, where the translation problem obtrudes, where the paths of `cËa' and ìt' diverge. Ìt' is semantically empty, to the point of all-inclusiveness: (it is a pro-noun, but it can be a pro-predicate, even a pro-utterance; we have seen that it can even represent the whole of the situation, as in ìt is raining' (worried by this state of affairs, Greek grammarians interpreted the Greek equivalent of ìt rains', a subjectless verb, as short for `Zeus rains'). Not so the deictic `cËa', whose function, at


the very margin of language, is to inscribe within language a gesture of pointing (as its very name indicates), and whose referent, therefore, is strictly singular, if vague. `Ca' notes the immediate object of my percep-tion, an object so immediate that I am not yet able to name it. And such content, the word being what linguists call a `shifter', is different with every new occurrence, with every new utterance. The paradoxical effect of this quasi-gesture, on the threshold of the text again we are in the realm of tropes is to point towards the outside of language at the very moment when we are entering language. `Ca' is also the name of the hero of the piece, the subject of all the animate verbs that follow, although it is a vague, but also a non-animate, non-personal pronoun. So that the said hero, albeit obviously endowed with a body (it eats, etc.), is also obviously nobody, not a person, perhaps even not a body, since it is characterised, through what seems to be a metaphor, as a machine. Most of this, inevitably but sadly, is lost in the translation.

The first word of the text is a nexus of tropes. And tropes are a necessary, not sufficient characteristic of literary texts. Tropes are indeed abundant in the rest of the text. The personification of cËa is sustained, even if `cËa' ìs machines'. The narrative trope of false anaph-ora is followed by the rhetorical trope of anaphora, or repetition (the style of the piece is deliberately repetitive). The formulaic mode of expression (`For every organ-machine, an energy-machine') gives the text the poetic force of proverbs or slogans (in the sentence just quoted the elision of the verb contributes to the general terseness of the style). And that sentence also works through allusion (a fleeting allusion to Marx), as does the rest of the text, which explicitly refers to Bataille (A solar anus), Judge Schreber and, on the rebound, Freud. Here comes the second translation problem, as `cËa' is not only vague, but the site of a pun: `cËa' and `le cËa' are translated respectively by ìt' and `the id', which, again inevitably and sadly, loses the play on words on which the whole paragraph rests. Lastly, the text appears to be studded with metaphors, that cardinal trope. In the terms of Lakoff and Johnson,7 we encounter an ontological metaphor (`the id') and a sustained structural metaphor (the BODY is a MACHINE). True, the metaphorical nature of these metaphors is immediately, and strongly, denied. But the literary critic may well interpret such denial as an instance of Freudian negation, as the denied term also provides the chute of the text: it has and is the last word of the paragraph.

All this amounts to a style. An inimitable voice is heard here, as singular as the voice in the first words of Edwin Drood, as recognisable as a picture by CeÂzanne. This is indeed a purple passage, evincing all the


stipulative eloquence of the professionally persuasive, of the orator, with what the French call his effets de manche: Ànd rest assured that it works . . . ' one can sense the implicit exclamation mark. And yet, in this text, the speaker is absent. There is no direct answer to the question, `who speaks?'. Neither is there in Dickens, at least immediately, for the question insists, and is duly answered at the beginning of the second paragraph of the novel (`Shaking from head to foot, the man whose scattered consciousness has thus fantastically pieced itself to-

gether . . . '8). But here no voice of a narrator intervenes to temper the excesses of the text and guarantee that common sense will prevail. Perhaps we are reaching the limits of the literary language-game: that voice turns out to be as stipulative as Wittgenstein's, as contorted and sly as Descartes's even if the range of topics broached appears to be wider than in a philosophical text, and more characteristic of literature.

One consequence of this rhetorical turmoil is at least clear: this is one of those texts where language obtrudes, a text meant for loud and public utterance, au gueuloir, as Flaubert used to say, and not for the introspect-ive and quiet reading in the philosopher's study. Yet this obtrusive language is also explicitly made little of, even excluded, by the text. The text insists on the reality of those metaphorical machines, and, in its working as a machine, it excludes vital components of language: it breaks the rules of collocation of subject and verb, more or less excludes speaking from the activities of the ìt' at the very moment when ìt' is so obviously writing. This is no longer merely a literary gesture: this is a philosophical gesture.

So our literary text is a text of philosophy. And in that light we must reconsider its first word, `cËa', with its twin linguistic functions of zero-topicalisation and vagueness. In a philosophical text, the incipit is often the site of two, complementary gestures, a gesture of position and a gesture of exclusion.

Here are two examples of a gesture of position: `There is a reason in Nature why something should exist rather than nothing', and its daugh-ter, `Why are there beings rather than nothing?'9 What is posited here, in the form of a question, is of the most general kind, nothing short of `being' itself. And I am struck by the linguistic vagueness of Leibniz's `something', a fit equivalent for our `cËa'. Which means that our first word, too, inscribes a gesture of position. `Ca' is a name for whatever it is that is given, in other words yet another name for being, which it is the object of the rest of the book to specify. Or rather, and the linguistic operation of zero-topicalisation is of considerable help here, it is the substitute for the yet absent name of being, which will eventually


emerge in the analyses that follow (in Deleuze's work, this name varies: in DiffeÂrence et reÂpeÂtition, it is `the eternal return of the same'; in later works, it is `chaos'). What I am suggesting is that our initial word is the equivalent not only to Leibniz's `something', but to Heidegger's `there is' (a canonical marker of zero-topicalisation) that signals, even in the form of a question, the gesture of position.

However, the `cËa' is also a site for a foundational gesture, a gesture of exclusion. And here the vagueness of `cËa' is of great help: it enables the word to carry out the exclusion, en douceur, without even having to mention it. It also marks the immediacy of the presence of the phenom-

ena: that is how it is. In spite of the professed anti-Hegelianism of Deleuze and Guattari, they have read with due profit the first pages of the Phenomenology of Spirit, `Das Diese und das Meinen', where the para-doxical nature of the deictic, `diese', is noted: it expresses the most singular contents of the immediate apperception of the phenomena by my consciousness, but it does this through the most universal linguistic element, the shifter, empty of meaning in itself, but which can be filled with any meaning, and is therefore the one truly universal designator.10

The exclusion operated by the text is threefold. It starts with the exclusion of the Ì': as an inanimate deictic, `cËa' excludes, as subjects of the main verbs and objects of the text, all personal pronouns, and most spectacularly the first-person pronoun that represents the speaker. This is where the comparison with the incipit to the Meditations is relevant. In Descartes, although the speaker's formulations aim at universality (his concern is the foundation of science, nothing else: presumably not for himself alone), the text is pervaded with markers of the first person (when Ì' is not the subject of the sentence, `my' or `me' are conspicu-ously present). On n'est jamais si bien servi que par soi-meÃme: in the philosophical equivalent of what linguists call the `me-first principle', the analysis of phenomena, like charity, begins at home (this is the English equivalent of the French proverb). Not so in our text: the very first words do concern the most immediate phenomena, the `cËa' that is immediately given, but not through the immediacy of perception or of introspective consciousness. This is a beginning this is the text's first gesture of exclusion that does without any form of cogito: neither Ì think' nor even ìt thinks', `cËa pense'; all `cËa' does is be at work, function like a machine. Such a gesture radically rejects the phenomenological approach an operation as rich in philosophical as in literary shock value (texts, especially philosophical texts, are not usually concerned with fucking, and there is obvious pleasure at mentioning the unmen-

tionable).


With hindsight, it can be stated that the exclusion of the Ì', of the speaking subject, is characteristic of Deleuze's philosophy, perhaps the only philosophy that radically seeks to demote the subject from its central position. I shall briefly mention two other contexts. In Logique du sens, the event, in this case the event of death, is formulated through the impersonal phrase, on meurt, not il meurt.11 Deleuze calls such im-

personal formulas, in a phrase he borrows from Lawrence Ferlinghetti, `the fourth person singular'. The event is itself a singularity, but it is an impersonal and pre-individual singularity (the terms are borrowed from the philosophy of Gilbert Simondon, part of Deleuze's idiosyncratic philosophical canon12). The move from `je' to ̀cËa' takes us into the realm of the impersonal, the move from ìl' to ̀on' into the realm of the pre-individual.

In his Foucault, Deleuze, in order to account for Foucault's `theatre of utterances', with its absence of speaker (the subject of the utterance is a place or position variously filled according to the type of the utterance), uses the phrase òn parle' an anonymous murmur offering a place for possible subjects.13 The focus of the analysis is no longer the speaker, or the author, but the system of utterances (Foucault's brand of discourse theory is called `science des eÂnonceÂs'). Hence the anonymous murmur of the on. With language, Deleuze adds, there are usually three ways to begin. You may begin with persons (the sender, the receiver); you may begin with signifiers (the expression here would be cËa parle); you may begin with the originary experience of the world of phenomena, as embodied in the exchange between parole parlante and parole parleÂe (here, the world itself is speaking, addressing us). What Foucault does, according to Deleuze, is to suggest a fourth beginning, the holistic murmur of ìl y a du langage'. A whole philosophy of language is implicit here, in this radical bypassing of subjectivity.

Our text, however, practises a second form of exclusion, this time in an entirely explicit manner. It concerns the choice between the two mean-

ings of ̀ cËa' on which the text puns. The vague deictic, the nexus of tropes, actively exclude the Freudian id: `quelle erreur d'avoir dit le cËa'. There are two reasons for the exclusion of the Freudian ontological metaphor (which turns a mere pronoun, and a shifter at that, into a noun, denoting an entity). The first is the very abstraction on which the metaphor is based: an instance of false personification, of fetishism, whereby a multi-

plicity of processes is turned into one object. The underlined definite article suggests that the uniqueness of the object is what Deleuze and Guattari object to: whereas the singular `cËa', which is determined by the plural `machines' (Èverywhere it is machines') is constitutively


multiple and processual. The second reason is that the metaphor terri-torialises the `cËa', moves it from a temporal process to a spatial object, thus turning what should be the exploration of a map (flows and inter-ruptions: the mapping of a river) into a geology of fixed strata. Or again, thus turning those machines into a structure. In a famous passage of Tristes tropiques,14 LeÂvi-Strauss compares psychoanalysis and Marxism to geology: all three `sciences' are concerned with deeper strata inaccessible to the naked eye. Nothing is more alien to Deleuze and Guattari, philo-sophical cartographers for whom everything happens on the surface of a plane. Through this second exclusion, a correlation of key concepts is being set up.

Before we make the correlation explicit, however, let us note that the text is also the site of a third form of exclusion, already alluded to in the passage from Foucault. For the `cËa' in our text may work, breathe, heat, eat, shit and fuck, but there is one thing it signally fails to do: it does not speak (at least in the opening words: a `talking-machine' is mentioned later). This absence is all the more striking as `cËa parle' is the formula that condensates Lacan's notorious thesis, `the unconscious is structured like a language'.15 This formula is linked to his celebrated definition of the signifier, à signifier represents a subject for another signifier'.16

The exclusion, therefore, is double. Our text does not explicitly exclude `speaking' from the crucial activities of the `cËa', since that `talking-machine' is mentioned, but the strong implication is that eating, not to mention other bodily functions, is more important then speaking and it is true that one of the constant interests of Deleuze is the rela-tionship between the two, as the allusion to anorexia in our text shows: after all, the same orifice is concerned in both cases. The result of this is that language cannot be a centre, or a focus, of interest: there are more important human activities than speaking. Whereby our text explicitly resists the linguistic turn taken by structuralism (in the form both of the centrality of language and linguistics as the model science, and of a topology of the psyche, with an unconscious `structured like a lan-guage'). And our text, therefore, also excludes the signifier as that which structures the psyche. This, which is a constant theme in Anti-Oedipus, can be said to be one of Guattari's specific contributions, as we shall see (although it is already present in Logique du sens, where Deleuze is polite to Lacan). There are more `regimes of signs' than the one centred on the signifier the tyranny of the signifier is historicised, linked to the capitalist mode: those are important themes of Capitalism and Schizophrenia. They insist on the fact that, semiotically speaking, there is more to man than a speaking animal: Guattari's position on this


is that faces (or `facialities') and jingles (`ritournelles') are other, equally important, semiotic modes.

We have moved from a literary style to a style of (writing) philosophy, to a system of concepts. The term `system' is of course both unfortunate and misleading. Deleuze, a coherent philosopher, does not claim to be a systematic one. His favoured discursive tool is the correlation. He never uses the term himself, but the practice is prevalent in his work, under the name `series', or `synthesis', or even `line' an integral part of the philosophical geometry or geography characteristic of Deleuze. What, then, is a correlation?

Traditionally, and trivially, the work of the concept is conceived in Hegelian terms. A concept develops along the dialectic triad of thesis, antithesis and synthesis (this is Hegel for French sixth-formers), a pro-cess leading towards the mysteries of Aufhebung. One contemporary way out of this traditional conception is deconstruction, with its operation of inversion, displacement and reinstatement under erasure.17 Deleuze's anti-Hegelianism is more radical, in that he has recourse, instead of the dialectic development of the concept, to a correlation: two antithetical series, or rows, a series of columns structured through disjunction (`x, not y'), each member of a row thus having an opposite number on the other row. The cohesion of each row is not causal (`x, therefore x0'), but obtained through connective synthesis (`x and x0'). The result is neither a single concept (although each row may be taken as an addition of determinations) nor a dichotomy, as contrasts proliferate along the row, which turns into a line of flight. Because all this is rather abstract, I shall take a (reflexive) example. I shall produce the correlation of the correl-ation, as opposed (concepts do not come singly) to dialectical develop-ment:

1 2 3 4 5 6 A Correlation multiple anarchic open-ended relational line B Dialectic single fixed origin teleological essential spiral

7 A rhizome B tree of Porphyry

8 9 map rhythm triangulation closure

The diagram reads thus: where a correlation involves a proliferation of concepts, the dialectic involves a single one; where a correlation is anarchic, the dialectic involves a fixed origin or principle. Whoever has read more than ten pages by Deleuze has come across this type of reasoning, time and again. So a correlation is both anarchic and open-


ended, which means that it always begins in the middle (àlways begin in the middle' is one of Deleuze's more notorious prescriptions). There lies its main advantage over the single concept (whose limited range will be exhausted by a finite list of determinations) and over the single dichotomy (teleologically oriented towards the third term that will `relieve' it in the course of spiral development). The principle of the correlation, therefore, is the following: there is always another column, which preludes closure, encourages rhythm, fosters imaginative concep-tual jumping from column to column (the connections between members of a row being no stricter than Wittgenstein's family resem-

blance), forbids the systematic binary hierarchy that goes under the name of `tree of Porphyry', turns the whole setting into a rhizome, which can be explored like a map, but without the Cartesian coordinates that allow triangulation. The best example of a correlation I know outside the works of Deleuze is the correlation that structures Jakobson's essay on aphasia.18 It begins rather sedately with the two forms of organisation of the language chain, paradigm and syntagma. It correl-ates with this two types of aphasia, as failures in the (paradigmatic) selection or (syntagmatic) combination of linguistic elements. It goes on to correlate this to the two tropes of metaphor and metonymy, the two forms of the dream-work, condensation and displacement, and two styles in literature (Romanticism and Realism) and painting (Surrealism and Cubism). At which, of course, the mind boggles: one hesitates between indignation at the flippant facility of the connection and admiration for the inventiveness of the critic's imagination.

I seem to have forgotten my text. It is, of course, too short to allow a full correlation to expand. Nor does it take the explicit rhetorical form of a correlation. But we do find the adumbration of such a correlation in it, a correlation that, being possessed of hindsight, we can develop. The linguistic marker that signals the possibility of a correlation is negation. And it is not lacking in our stipulative text, the canonical sentence of which appears to be `x is the case, not y' (Èverywhere it is machines real ones, not figurative ones'). Sometimes the contrast that indicates the presence of a column is overt (as in the sentence just quoted), sometimes it is covert in such cases only the positive term is men-

tioned, as Deleuze claims to be a philosopher of affirmation, not nega-tive ressentiment.

If we reread the text, then, with the analytic tools familiar to the semanticist, or structural discourse analyst, the following overt oppos-itions come to our immediate notice: machines v. organs (your organs are in fact machines: the body is a machine); real v. figurative (there is


direct intervention of words in the world of things, rather than endless play between signifiers and signifieds); effects v. metaphor (`cËa' is all about concrete bodily effects, not the ideality of signs). These overt contrasts can be prolonged in covert, but easily retrievable ones: body v. logos (every word of the text, hence its shock value, insists on the gross materiality of the body as opposed to the ideality of language reason mind); and flows and cuts v. signs (signifier over signified, in the Saussurean diagram of the sign): a horizontal, linear, not a vertical order; lastly, the text implicitly contrasts the multiplicity of singularities (this is where the empty shell of the deictic, the fact that it is a shifter, helps) v. the unity, and universality of abstraction (this contrast is in fact explicit, although lost in translation, in the opposition between `cËa' and `le cËa', ìt' and `the id').19

All these contrasts form the beginning of a correlation. In the middle of it, there is one contrast that structures the text, accounts for its rhetorical rhythm and allows the correlation to proliferate. It is the general contrast between function and representation/interpretation. The text describes the `cËa' at work, a matter of functioning and produ-cing effects (`we are all handymen', the text says the French word is `bricoleur', with reminiscences not only of the DIY shop, but also of LeÂvi-Strauss), not of representing states of affairs and questioning, or inter-preting, such representations. And the text itself is, iconically, seen to be at work, an intervention in a philosophical conjuncture, an attempt to transform the world that other philosophers, so Marx says, have con-tented themselves with interpreting. We understand why the text ends as it does: `Something is produced: the effects of a machine, not mere metaphor.'

For the correlation goes on, of course. It is not limited to the incipit: a machine is not a structure (Anti-Oedipus is an anti-structuralist text); energy and force, intensities, are not signification; an assemblage (of enunciation, of desire) is neither a (speaking) subject nor an individual. This expanded correlation is the core of what I shall eventually call Deleuze's philosophy of language. It will be explored in the rest of this book. Meanwhile, I shall take leave of the incipit to Anti-Oedipus by insisting on the paradox that animates it, a paradox that marks language as problematic in Deleuze. The correlation as we have described it implies a strategy of displacing language from its centrality in the cul-ture of structuralism, of reinscribing it in its material site of emergence, the body. But such strategy is deeply paradoxical. The text, in its account of `cËa' at work describes flows of energy and interruptions, or cuts, as the flow of maternal milk is interrupted by the baby's ravenous mouth, a


breast machine and a mouth machine. But one of the best examples of this type of machine, if not the most natural, is the flow of air that goes through my vocal organs, and the interruptions that the said organs impose upon it to turn my heavy breathing or scream into articulated language, my phone into logos. Language demoted immediately has to be reinstated as the most obvious example of the process described: ìt' may eat and heat, and shit and fuck, but above all, cËa parle, in the text (for hundreds of inexhaustible, garrulous pages) and out of the text. The exclusion of language is an instance of Freudian denial. A fit beginning for a text entitled Anti-Oedipus.

2 Deleuze against language

The demoting of language must be taken seriously, even if language returns. The prognosis of Freudian denial is too facile, it smacks of the psychoanalyst's customary trick, which consists, Deleuze and Guattari say in a little-known but outspoken text,20 in providing the answers before one has had time to ask the questions. Little Hans is genuinely interested in horses, and in the little girl in the flat below: he is not allowed to explain or expand, he is immediately Oedipalised, triangu-lated. Deleuze deserves a better fate, if only because, in his hostility to language (the term is too strong, and inaccurate: it will have to be qualified) he is part of a tradition.

It is generally recognised that French philosophy too has taken a linguistic turn, albeit in a different form from its Anglo-Saxon equiva-lent. The contents of this turn can be summarised using the phrase Antoine Compagnon has coined to describe the classical moment of theory (by which he means French philosophy and social sciences in the late 1960s and 1970s): `the conjunction of formalism and Marxism'.21

`Formalism' means the centrality (Guattari would say `the dictatorship') of language and its science, linguistics. And it is striking that almost all the French theorists of that period not only have (or have had) a relationship to Marxism, but that they indulge in an even closer form of consort with linguistics. Remember not only Lacan's idiosyncratic attachment to the signifier and his preoccupation with metaphor, but the role of phonology in LeÂvi-Strauss, of structural semantics in Grei-

mas's narratology, of the distinction (borrowed from the Soviet linguist Saumjan) between genotype and phenotype, and between the semiotic and the symbolic in the early Kristeva. Remember Bourdieu's sociolin-guistics, Derrida's interest in bilingualism (he is, after all, a formidable close commentator of texts, in the Talmudic tradition, or in the French


academic tradition of explication de textes), Foucault's archaeology of knowledge as science des eÂnonceÂs. Remember Lyotard's celebrated, but highly creative, misprision of Anglo-Saxon speech act theory in Le

22DiffeÂrend. My last sentence suggests that there are links between the two linguis-

tic turns, the analytic and the Continental: they would make an inter-esting theme for cultural history. Roughly, and also tongue-in-cheek, one might distinguish four moments.23 The first (largely due to the fact that the Anglo-Saxon linguistic turn occurred several decades before the Continental) is blissful mutual ignorance. The second, when reciprocal notice was at last taken, can be called dialogue de sourds. The best example of this incomprehension, of the incapacity of each party even to read, let alone to understand, the works of the other, is the celebrated Royaumont colloquium of 1962.24 The third moment, when the rela-tionship is at last engaged, albeit in the military sense, may be called sceÁne de meÂnage: the canonical example of such active misunderstanding is the transatlantic exchange between Derrida and Searle, which culmin-

ated in Limited Inc.,25 and in which no punches were pulled. The last, and most interesting and fruitful, is the creative misprision of which Lyotard's book is a fine example. But another example of such import-

ation, critique and development of analytic philosophy is to be found in the pragmatics of Deleuze and Guattari. What Lyotard does to the second Wittgenstein, they do to Austin and Searle: they borrow the concepts, put them to work, like so many machines (Èverywhere it is at work . . . '), assess and criticise the results and take the concepts on entirely new and unexpected paths, but paths that offer brand new vistas.

So Deleuze is an integral part of the linguistic turn that characterises classical French theory, even if he can hardly be accused of `formalism'. But the paradox is that he is also the heir of a tradition of, if not downright hostility to language, at least deep distrust. In France, and this is especially important for Deleuze, one of the main sources of this tradition is Bergson. Here is a typical passage, culled from La PenseÂe et le mouvant:

Language is, I grant you, pervaded with science; but the scientific spirit demands that anything may be questioned at any moment, and language needs stability. Language is open to philosophy; but the philosophical spirit inclines towards endless renewal and reinven-tion, for that is how things work, whereas words have definite mean-

ing, a relatively fixed conventional value: they can only express the


new as a recomposition of the old. This is usually and perhaps im-

prudently called `reason', the conservative logic that rules communal thinking: conversation sounds very much like conservation.26

Bergson's distrust of language that is a theme close to the centre of Deleuze's thought is due to the fact that words freeze concepts, make them dependent on common sense. Language has a social origin: its aim is to create communication so as to foster cooperation. This means that its main function is industrial, commercial and military, always social. It fixes reality as it names it according to the needs of society. But the philosopher has little to do with such common sense: metaphysical questions go far beyond immediate social needs. As a result, the philo-sopher's task is to go beyond the words, against the grain of language, to discover the problem that he will formulate. A philosophical problem, Bergson says, is solved as soon as it is correctly formulated (ùn probleÁme speculatif est reÂsolu deÁs qu'il est bien poseÂ'27): language, in its `gross utili-tarianism', conceals and obfuscates; problems always lie beyond the reach of common words. This is where the philosopher is naturally sympathetic to the poet, who, like him, is trying to take language to its limit. The difference is that, for the poet, beyond the limit lies silence, towards which he paradoxically strives; whereas for the philosopher, there is life outside language, in problems and concepts.

That Deleuze shares this view, that he is in a sense a disciple of Bergson (he contributed to the revival of a philosopher whom the positivist tradition of structuralism despised) is clear. Like Badiou (who calls such people sophists), he intensely dislikes the Anglo-Saxon exponents of the linguistic turn in philosophy, most notably Wittgenstein (against whom he uses violent and uncontrolled words in his AbeÂceÂdaire28): for him the idea that all philosophical problems might be grammatical problems is anathema.

One of the recent commentators of Deleuze insists on his `pessimism and distrust' towards language: language for Deleuze, he claims, ìs the root of all illusion, of man's finitude and of his subjection'. Language is the source of the taste for opinion, the main source of evil in man.29 Yet, of course, the instrument and medium of philosophy, the sentences from which philosophy extracts its concepts, is language. This paradox is by now familiar. Let us see how it is negotiated in Deleuze's works, let us take notice of the absent presence of language in the Deleuze corpus he never actually wrote a philosophy of language, but Logique du sens comes close to it; he never constructed a fully-fledged theory of lan-guage, but he gives us a multiplicity of partial theories (of sense, of style,


etc.) and language is constantly present as a privileged point of applica-tion for theories or concepts that at first sight do not seem directly to concern it.

It is time to let Deleuze speak for himself. The following is a passage in Dialogues, taken from a chapter signed by Claire Parnet (but `Deleuze', as we have seen, is never an individual name, always a collective assem-

blage). We may take it as typical of the Deleuzean attitude:

It must not be said that language deforms a reality which is pre-existing or of another nature. Language is first, it has invented the dualism. But the cult of language, the setting-up of language, linguis-tics itself, is worse than the old ontology from which it has taken over. We must pass through [passer par] dualisms because they are in language, it's not a question of getting rid of them, but we must fight against language, invent stammering, not in order to get back to a prelinguistic pseudo-reality, but to trace a vocal or written line which will make language flow between these dualisms, and which will define a minority usage of language, an inherent variation as Labov

30 says.

This implies a more complex attitude to language, on the whole, than Gualandi ascribes to Deleuze. It appears to state, as background to a rather idiosyncratic distrust of language, the rather strong thesis that there is no prelinguistic reality: this is stated twice in this short passage. By which we may understand not that there is no outside language, but that the limits of our language are the limits of our world. There may be Things-in-themselves or the Real out there, but our reality is linguistic: the phenomena are not only pervaded with language, but linguistically constructed. This may be taken as an anti-Husserlian statement: Deleu-

ze's phenomenology appears to be language-driven. Hence two conse-quences, which the passage immediately draws. Language freezes reality, turns it into an object of common sense, named by the tyrannous `dualisms' (we have learned to call them, in Derridean parlance, `di-chotomies': nature and culture, body and soul, language and thought). Such dualisms must be struggled against, but they are inevitable. They are, in fact, the stuff that correlations are made of. And correlations are necessary for two opposed and paradoxical reasons: they revel in the dichotomies, which they conserve and glorify; and, by making them proliferate, they enable us to go beyond them, to turn their verbal architecture into lines of flight, their incipient trees into rhizomes. The second consequence is that the necessary struggle against language


no longer takes the form of ineffable intuition, of immediacy, but of stuttering. If the structure imposed by language upon thought is both necessary and nefarious, the philosopher's task is to trace a line between them, to negotiate a path like the sailor when he manage to avoid both Charybdis and Scylla the philosophical equivalent of Brave Soldier Schveik's tactics. Since we cannot overturn the dualisms that define not only reality but language itself (langue v. parole; signifier v. signified), let us turn their hierarchic superposition into a line, let us invent a minor, or stuttering use of language, one that will be both politically and poetically subversive. The very term `minor language' embodies the paradox of Deleuze's attitude to language: we are immersed in it; we are always dreaming of emerging out of it, on to the dry sand of a prelin-guistic reality that has no reality.

This is no blueprint for resignation, however. Stuttering and the min-

orisation of a major language are active strategies of resistance to what Deleuze sometimes calls `the imperialism of language'. The passage in Mille plateaux where this phrase appears is in fact devoted to the joint imperialisms of language (over all strata) and of the signifier (over language).31 Strata are the planes where phenomena occur: where ma-

terials are formed, intensities are captured, singularities fixed. They are defined by flows and codings, they are the sites for movements of territorialisation and deterritorialisation. Although their name comes from geology, they are at the heart of Deleuze and Guattari's plane geometry: a stratum is a plane of consistency.

The model for this stratification seems to be language. A flow of phonic substance, the phonemic coding, the plane of consistency of language where the linguistic sequence, the chain or line of articulated words meanders: the image of language enables us to make sense of what appears at first sight to be a quaint theoretical construct. But there are two dangers, which the passage just mentioned about the double imperialism of language points out. The first danger is the inversion of the causal order. By making language the canonical image, if not the source, of the theoretical construct, we put things the wrong way up, and repeat the error of Saussure, whose semiology is just a weak gener-alisation of the characteristics of the system of language. Language is not the source, but one of the effects of the semiotics of flow and coding already present on the first lines of Anti-Oedipus. The couple of concepts, flow and coding (the codes interrupt the flow by cutting it, and thus articulate it) describe a process upstream of language, far more general, the general process of the flow of energy and its capture in and by machines. There are other semiotics than linguistics, other codes than


the code of grammar. Not all planes of consistency drawn over chaos are planes of signification.

The second danger lies in the imperialism of the signifier, when the signifier reflexively repeats the imperialism of the linguistic coding of strata. Such insistence on the signifier impels us towards a concept of language based on representation and interpretation, which Deleuze and Guattari in their struggle against psychoanalysis wish to reject. In the passage I am referring to, the authors illustrate the dangers of the imperialism of the signifier by the example of the prison, borrowed from Foucault.32 The thing, they claim, `la chose-prison', is not so much linked to the word `prison' that names it, as to a concept, the concept of `delinquency', which (the concept and not the word) provides the said thing with its form of expression. For the `form of expression', a phrase borrowed from one of the few linguists Deleuze and Guattari approve of, Hjelmslev, is not necessarily, or as we might have believed, composed of words, of signifiers in this case it consists of a number of eÂnonceÂs, utterances (the exact words of whose formulation are irrelevant): they appear in the social field that is the relevant stratum for the crime-

machines. In other words, language as collection of signifiers is bypassed in two ways: because the relevant elements are utterances, not words; and because the machines in which they are captured, the assemblages in which they set to work involve far more than language: institutions, ways of seeing, material practices, etc. The analysis ends on a famous and striking sentence: Òn a beau dire ce qu'on voit: ce qu'on voit ne loge jamais dans ce qu'on dit' Ìt is in vain that we say what we see what we see never resides in what we say'.33 So, beyond language, but articulated to language, as to reality, in the same machines and assemblages, there is at least what Deleuze in his Foucault calls `visibilities', which he contrasts with `sayabilities'. We understand that his keen interest in literature should be equalled by his interest in the visual arts, painting (see his Bacon book) and film.

There is a consequence of this attitude of putting language in its place, which I shall just mention here, as it is the main object of the next chapter: Deleuze's hostility (here the word is not an exaggeration) to linguistics. One of the reasons why Deleuze's positions on language, in their complex and paradoxical way, are of paramount interest to the student of language, is that they outline a concept of language which has little to do with either structuralism or the Chomskyan research programme, which has been dominant in linguistics for several decades.

So there is in Deleuze a definite desire to move away from the central-ity of language. Such a move has to do with Deleuze's ingrained empiri-


cism, as the move away from language fosters an interest in sensation and in thought. Deleuze is one of the rare philosophers nowadays whose thought explicitly centres on a reflexive enquiry into thought not so much that which is betrayed by language as one element of reality, distinct from language and graspable through language but beyond language. As we have seen, the task of the philosopher is to extract concepts from sentences.

3 Against metaphor

The best instance of Deleuze's wary attitude towards language is to be found in his notorious hostility to metaphor, a strong formulation of which we have already found in the first paragraph of Anti-Oedipus. This is a constant theme in Deleuze's work. Here is perhaps the best known and most explicit passage. It contrasts metaphor and metamorphosis in the work of Kafka and quotes one of his aphorisms, `Metaphor is one of the things that make me despair of literature.'34 The text goes on to say that Kafka actively rejects, even eliminates (a word to be taken literally) any form of metaphor or symbol, any signification or designation. The instrument of such elimination is metamorphosis, the opposite of meta-

phor, which has the twin advantages of dissolving the barrier between words and things, as between the literal and the figurative: then words become things, and things are bundles of intensity on which words or sounds trace the paths of their deterritorialising lines of flight. Meta-

morphosis is the best example of an ìntensive, non-signifying use of language'.35 With it language is no longer representative, that is, separ-ated from the things and states of things on which it is articulated: it tends towards its own limits, and becomes non-subjective, as subjects (both the speaker and the grammatical subject, `sujet de l'eÂnonciation' and `sujet de l'eÂnonceÂ') are replaced by collective assemblages of enunci-ation.

All these terms are still mysterious. They will, I hope, be made clearer in the course of the book beginning with the interlude that follows this chapter. At the moment, we must be content to note that Deleuze's hostility to metaphor lies at the heart of his conception of language, and that the general move from metaphor to metamorphosis is fairly clear: it is a move away from a representational concept of language (with the attendant distinction of the literal exact or true representation and the figurative indirect or fictitious representation), towards a concept of a language immersed in the world of things, intervening among them, forming machines with them, capturing and distributing intensities,


being itself part of the general movement of deterritorialisation and reterritorialisation.

This is not as abstract as it seems. And it certainly involves the slogan `Reject metaphors!' applying first to Deleuze himself, a stylistic practice, outlined in Dialogues, where Deleuze gives an exoteric version of his philosophy. It starts in his usual brusque manner with a pun on `mot propre', the exact word, the word with the proper meaning.36 By con-trast, metaphors, he says, are `mots sales', dirty words. And he rejects both types of words, clean or dirty, because both, in their different ways, seek to be true (theories of metaphor often, but not always tend to claim a certain form of truth for the practice of metaphor, at least a tension towards it: after all, a metaphor can be apt, or a failure). Deleuze claims that we always use inexact words to designate things exactly, that we must push language out of the ruts of habit, and create extraordinary words for the most ordinary uses. In other words, in the metaphor that aptly sums up Deleuze's conception of language, we must make lan-guage stutter. In this, the philosopher is like the poet, one who speaks a foreign tongue in the midst of his maternal language.

I seem to have painted myself into a corner. I am expounding Deleu-

ze's theory of the necessary rejection of metaphor, and yet I have just identified the central metaphor of stuttering that characterises his whole conception of language. And truly this enemy of metaphor is as fine and prolific a creator, or importer, of metaphors as the most figura-tively inclined of poets could wish for. Deleuze, as he often does, borrows from astronomy the concept of `black hole', which he deterri-torialises from its original field and reterritorialises in the field of phil-osophy. Guattari is thinking about `white walls'. Put them together, Deleuze says, and you have a face: from this the concept of `visageÂiteÂ' (a dreadful coinage the English translation has `faciality' but an entirely valid concept) is born.37 The deliberately inexact transfer of a word from its original field into another is, of course, a reformulation of the standard definition of metaphor it is even true to the etymology of the word. Yet the denial, `nous parlons litteÂralement', on the same page of Dialogues must be taken literally: the structure is the same, but linguistic deterritorialisation is not a form of metaphor, it must be taken literally, and it involves the whole of the concept of language just men-

tioned. That the mouth on my face can be called a black hole is a blatant use of metaphor (and a dangerous one, since it seems to transfer to this homely organ some of the characteristics of a rather terrifying astro-nomical object). But taking it literally, demoting (or de-mouthing, if you pardon the terrible joke) my mouth involves, in the apt inexactness


of this violent dispossession of its name, a whole new conception of the face, of the body, of myself as a person a conception, or a problem, which the new concept of visageÂiteÂ names (the awful coinage may there-fore be a necessary one).

In short, what Deleuze objects to in metaphor, as a philosophical concept rather than as literary practice (although he does tend to admire authors, like Kafka or Beckett, who preferred to do without it) is the twin structure of truth and representation that the concept carries with it.

The metaphorical substitution, a paradigmatic substitution in the terms of Jakobson's correlation, allows the clash of images that produces illumination, and consequent affect of truth.38 Against this, Deleuze defends a syntagmatic, geographical, rather than a paradigmatic, geo-logical notion of transfer. And the transfer has to be taken literally: not merely a deliberate misprision of words whose deterritorialisation pro-duces effects of meaning, but a mapping of the plane, or surface, where words and things, propositions and states of affairs meet, with its lines of flight, without origin or primacy or hierarchy (as between the literal and the figurative), where singularities are produced in the midst of strange attractions and connections. If we take this `literally' literally, as we must, we are plunged in the middle of one of Deleuze's metaphysical themes: that the separation between subject and object, thought and matter, words and things, is an illusion of language. Only the last dichotomy (thoroughly deconstructed in Logique du sens) will directly interest us. But the deconstruction of the first two is not so outlandish: even if Deleuze is not a straightforward phenomenologist, he has obvi-ously read Husserl.

The second conceptual structure of metaphor that Deleuze criticises is the very structure of substitution, which implies a concept of language as representation: literal wording is an exact representation of the world, or of part of it; figurative wording is only indirectly apt, involving as it does an interpretative path. And Deleuze, who is against metaphor, is also against interpretation and representation. The link between those two themes, and the characteristics of representation to which Deleuze objects will be the subject of the Interlude that immediately follows this chapter.

4 The linguistic turn in Deleuze

So far, I have insisted on Deleuze's distrust of language, on his refusal to take the linguistic turn of Continental philosophy. But his attitude is more complex than this as, in seeming contradiction, he can also be said


to have taken an active part in this turn. The merest sketch, in anticipa-tion of the chapters that follow, will be sufficient here: reading the Deleuzean corpus gives the impression that indeed language is a source of illusion, which must be kept in its secondary place, and yet that Deleuze finds an overpowering interest in it and is always calling for it to be taken to its limits. And, with the collaboration of Guattari, he did take an active part in the linguistic turn, through the elaboration of a Continental brand of pragmatics.

As an illustration of this, the other side of Deleuze's attitude to lan-guage, I shall take an example: his treatment of Foucault's archaeology of knowledge as `science des eÂnonceÂs'. The canonical example of an èÂnonceÂ', an utterance, which Deleuze finds in Foucault, is the sequence of letters on a typewriter or computer keyboard, ÀZERT' (quite fittingly, the keyboard is a French one). Outside the keyboard, of course, such a sequence of letters is purely arbitrary, a matter of chance. On the key-board, however, it becomes necessary, being determined by the fre-quency of the letters in a given natural language and the ergonomics of typing. Furthermore, once installed on the keyboard, it acquires what Hjelmslev and Barthes call connotation. The string of letters now pro-claims: Ì am the beginning of the first row of letters on a French keyboard.' In so doing it becomes part of a complex assemblage involv-ing machines and technology, applied science, but also institutions, relations of production and rapports de force. In other words, it becomes an eÂnonceÂ. There is one aspect of the French word that the English ùtterance' loses, so that one is tempted to coin a linguistic monster, `the enunciated': it is a past participle, which insists on the facticity of the eÂnonceÂ, on the fact that it is always-already formulated, and does not belong to the realms of the possible or the virtual. This actuality of eÂnonceÂs implies that the proper way of dealing with them is not through syntactic or semantic analysis, but through a form of pragmatics (Where does the utterance come from? What type of subject was enabled to utter it? What relations does it have with the other utterances in the same conjuncture and the assemblages that produce them?) a prag-matics with a difference, in that it rejects the methodological individu-alism on which the Anglo-Saxon pragmatics of Austin, Searle and Grice is based (meaning here is not individual, and the calculus of implicature does not seek to reconstruct intention). Deleuze hails Foucault as the founder of this new pragmatics a state of affairs which Foucault himself indirectly recognises when he acknowledges the influence of speech-act theory on his thought and claims that his philosophy (or rather the philosophy he is calling for) is a pragmatics of power.39


Incidentally, in the same passage he distances himself both from Hum-

boldt's exaggerated view of the centrality of language (as sole source of the relationship between man and his world) and Bergson's devaluation of language as frozen, as betraying in its spatial ordering the experience of dureÂe (the celebrated concept under which Bergson thinks time). Anglo-Saxon philosophy of language, he claims, maintains that lan-guage neither reveals nor betrays: `le langage, cela se joue' language is in play and at play (whether language-games are cooperative games, or zero-sum games, is another matter).

There is a philosophy of language implicit in Deleuze's exposition of Foucault's theory of eÂnonceÂ, a new linguistic object, neither a propos-ition nor a sentence, with the three spaces it organises around itself (the collateral space of other eÂnonceÂs; the correlative space of the pragmatic structure, with its places and positions, it enters in order to set to work; the complementary space of the historical conjuncture political events, institutions etc. in which it appears, remains or is reinvented in active repetition). For eÂnonceÂs are not susceptible to either formalisa-

tion (like Chomskyan sentences) or interpretation (like propositions or texts: the archivist never supposes that her eÂnonceÂ says one thing and secretly means another). And they are not dependent on an authorial and authorised subject that controls their meaning and whose inten-tions subsequent interpretations strive to recover. The central concept is not `meaning', ìntention' or ìnterpretation', but `power'. Like Deleuze and Guattari's slogans, eÂnonceÂs ascribe places and inscribe relations of power (rapports de force). Rather than `je parle', the linguistic version of the cogito, we have òn parle' (in Les Mots et les choses), or ìl y a du langage' (in L'ArcheÂologie du savoir). As we can see, we are indeed in the middle of a fully-fledged philosophy of language, or rather at its beginning: those formulae are foundational, and use the same linguistic trappings as Leibniz's `Why is there something . . . ', or the first words of Anti-Oedipus.

I have taken the example of Deleuze's Foucault, but there is a multipli-

city of such examples in the Deleuzean corpus. Language being the privileged example for the application or illustration of theories that have their source in an entirely different field, one is constantly under the impression that Deleuze is on the verge of offering yet another fully-fledged philosophy of language. A reading of DiffeÂrence et reÂpeÂtition,40 for instance, with a view to extracting a concept of language from it, would yield a rich crop: a theory of esoteric words (p. 156), to be developed in Logique du sens, a concept of linguistic multiplicity (p. 230 language provides an apt examplification for the central Deleuzean concept of


multiplicity), the notion of a linguistic idea (pp. 262 3), which will inform the more explicit theories of language in Logique du sens and, Anti-êdipe. For a self-styled chastiser of language, Deleuze is strangely fascinated by the object of his distrust.

Perhaps his main contribution to the field, however, is not so much to the linguistic turn in philosophy as in the development, or conserva-tion, of a modernist or avant-garde aesthetics, as I have already hinted. Deleuze is indeed a willing scion of the modernist tradition, for whom literature is an exercise of language. Language for Deleuze can always be retrieved from its deleterious customs or occupations, provided it is made to stutter, pushed to its limits, made to work against itself as the minor uses of language subvert the major ones. For literature, the site of such liberation of language occupies a crucial place in Deleuze's thought, both in his theory and in his practice. Perhaps this is to be taken as another tribute to Foucault, who in his early work (most notably Les Mots et les choses, and a host of seminal essays, now collected in the first volume of Dits et eÂcrits41) ascribes the highest value to literature. Thus, in the second chapter of Les Mots et les choses, he defines literature as the expression of our desire to transcend the limits of the conjuncture, and the limits of the reigning episteme. He offers two accounts of this. The first is nostalgic (literature goes back to a view of language held by a defunct episteme) but the second is progressive: `literature' is that which best expresses the conception of language of the new, `modern', episteme, which has replaced the classical episteme based on representation: it is that which actualises the tendency inher-ent in the classical view of language towards full naming.42 In other words, for Foucault literature is the name of the `prise d'inconscience', of the temporality of the episteme, a name for the way we experience the temporality of thought, of knowledge and belief.

Deleuze's views on literature (they will be developed in Chapter 6) are somewhat different, and yet they share a number of characteristics with Foucault's: literature is language at its highest and most complex. As such, it must be taken seriously: it intervenes in the social and historical conjuncture, it exerts force. This involves a new concept of langue (the minor v. the major language), a new concept of syntax (the intensive line of syntax, at its best in the literary stuttering of language), a new concept of style, and ultimately a new philosophy of language, based on literature as the core of language (which will be anathema to empirical linguists, proud of their billion-word corpora, and keen to welcome all registers and all genres of language in a spirit of genial benevolence and egalitarianism).


5 Guattari

I have referred to the name `Deleuze' as denoting not an individual author but a collective assemblage of enunciation. Indeed, the texts I have used are sometimes signed by Deleuze alone, often by Deleuze and Guattari, or even by Deleuze and Parnet. This passage from single author to assemblage is a philosophical gesture in the spirit of `Deleuze': a dual gesture, of position and of exclusion.

It is a gesture of position. I wish to treat the totality of the Deleuzean oeuvre as a corpus, at the cost of somewhat reducing its heterogeneity. This reduction is imposed by my problem. Not all the books offer insights into the workings of language or into Deleuze's conception of them. So I prominently include works which he wrote on his own (from Logique du sens to Critique et clinique), and disregard others (his thesis on Spinoza, or his books on the cinema, which are at the centre of other readings of his philosophy43). And I also prominently include all the works of which Guattari is the co-author (from L'Anti-êdipe to Qu'est-ce que la philosophie?). Which means that the title of this book is unfair to Guattari.

There is also, however, a gesture of exclusion: not only of those books by Deleuze which I deem to be irrelevant, but of some books by Guattari himself, namely his late books, and most of all his last book, Chaos-mos.44 This book I exclude from consideration because I fail to under-stand it. Usually, when I fail to understand a book of philosophy, I apply a form of the principle of charity: if it is too difficult for my feeble wits, this disastrous state of affairs is due to my ignorance rather than to its author's incoherence. But Chaosmos, I must confess, places a strain on my benevolence: the jargon is so thick as to conceal the meaning, at least to me.

But there is a risk in my gesture of exclusion again, the risk of being unfair to Guattari, of treating him as a minor excrescence to the great philosopher, the Siamese twin that nature has made a parasite on his more vigorous brother. This is an attitude I do not wish to adopt, for two reasons.

The first reason is that there appears to be a doxa in the making in Deleuze criticism, at least in France: the great philosopher did not choose his friends wisely, and Guattari plays a non-philosophical Mr Hyde to the philosopher's Dr Jekyll. For the French academic tradition (which scarcely acknowledged him when he was alive), Deleuze is an easily assimilated, because profoundly respectable, character. He was a professional philosopher, he held a chair (albeit in a dubiously leftist


university), he wrote an old monster of a French thesis, a theÁse d'Etat (a rather eccentric piece of work, when read in the light of what such theÁses usually are). He is immersed in the French academic tradition of the history of philosophy, and the ever renewed Talmudic commentary of canonical texts (in spite of Deleuze's claims of scandalous originality, writing books on Bergson, Spinoza and Leibniz is an eminently respect-able achievement in a philosopher). And he is a technical philosopher who, again in spite of his claims, often writes for other philosophers (DiffeÂrence et reÂpeÂtition makes no efforts to reach out to the general public: such theÁses were often read only by the five-member committee that examined them). If we want to focus on this, the respectable side of Deleuze, it is normal that we should wish to exclude Guattari, a maver-

ick with no serious philosophical background, who led his friend away from the straight and narrow path of boring philosophy into the dubi-ous avenues of political, social and aesthetic intervention.

The problem with that doxa (I have caricatured a tendency that hardly dares to be explicitly formulated, at least in writing), is that it is false. There is, thank Heavens, a profoundly disreputable, and therefore im-

aginative side to Deleuze, who obviously wanted little incitement to intervene in aesthetics and politics (the conjuncture of the May events in 1968 naturally helped).

I could cite, as proof of the falsity of the doxa, the innumerable passages in Deleuze's interviews where he stresses the importance for him of his friendship and collaboration with Guattari. This is not sufficient: he might be lending a hand to a friend in need. Except that such tributes are not vague and general declarations about similarities in their views. They are usually more precise than that, and Deleuze is often careful to outline Guattari's specific contribution to the communal enterprise. Thus, he claims that Guattari brought him (a) an experience of politics (Guattari was for a time a member of the French Communist Party, which Deleuze never was; he was also, for a time, closely associated with various leftist groups or movements); (b) a strong hostility to psychoanalysis (there is a paradoxical flavour in this: Deleuze, like most progressive philosophers of his time, evinced a keen interest in psychoanalysis, as shown by the fact that he called his pre-Guattari Logique du sens a `psy-choanalytical novel'; it took the trained analyst and former Lacanian that Guattari was to make him see the error of his ways); and (c), most interesting for me, an informed if critical interest in the science of lin-guistics. Deleuze is most explicit about this in a passage in Pourparlers:45 I don't think, he claims, that linguistics is fundamental. It is Guattari who traced the development from phonology to syntax, and from syntax to


pragmatics, which transformed linguistics and turned it into an interest-ing field, allowing linguists to meet with novelists, philosophers and what he calls `vocalists', people who conduct research into sound and voice:46 Barthes, Nathalie Sarraute, Oswald Ducrot, meÃme combat.

But perhaps the best reason to distance oneself from the incipient doxa is that their collaboration is the only real case I know of philosoph-ical dialogue (Badiou's attempt at posthumous dialogue with Deleuze in his Deleuze book is a signal failure47). Theirs was a true eÂcriture aÁ quatre mains. Deleuze once said that, in their common books, one of them wrote the sketch of a chapter, and the other developed it. Whether this is fact or legend is irrelevant here: in both cases it marks the depth and seriousness of their collaboration.

All this is still not fair enough to Guattari. The problem is: having admitted they were Siamese twins of equal vivaciousness, how can we separate them, and discern, beyond the general indications given by Deleuze, Guattari's specific contribution? It would appear that we can-not, and no careful reading of L'Anti-êdipe and Mille plateaux will yield indications susceptible of proof, as opposed to vague feelings and arbi-trary decision, of what is to be ascribed to whom. But in fact, we can. One year before the publication of Mille plateaux, Guattari published, under his sole signature, a collection of essays, L'Inconscient machinique,48 which covers the same ground as several plateaux of the communal work. We are therefore in a position to assess Guattari's specific contribution. And it is central to Mille plateaux (Guattari does acknowledge that he wrote those essays in a period of dialogue with Deleuze; nevertheless, they are pub-lished under his sole name). We can even assess his difference with Deleuze, as the last section of the book is devoted to a reading of Proust, a strong reading, but one very different from Deleuze's own.49 Briefly, his specific contribution may be described under four headings.

The first, and most important, contribution is the foundation of a new pragmatics, at the centre of which lies a theory of assemblages (in other words, as we have seen, a pragmatics that avoids the methodological individualism of Anglo-Saxon speech-act theory). It begins with a dis-satisfaction with traditional, or mainstream, linguistics (Chomsky is naturally a prime target): the science of language seeks to capture nat-ural languages in its structures; fortunately, Guattari says, `les langues fuient'50 the vessels are leaking, and languages blithely disregard the `laws' that science seeks to impose upon them. He goes on to reject two of the central tenets of structuralist linguistics: the distinction between langue and parole (he would be comforted by the fact that the history of linguistics can be largely described in terms of a displacement of this


distinction: syntax was excluded from langue by Saussure, and is the centre of competence in Chomsky51); and the existence, if not of lin-guistic universals, at least of universal features in the scientific analysis of language (universal features which, for instance, will enable the analyst to tell a linguistic sequence from an instance of glossolalia: for Guattari, there is no separation, only transition, between the two).52

As a result, the pragmatics he advocates will not be a pragmatics of intention and communication, but a pragmatics of power. For him, grammatical markers are markers of power an analysis since de-veloped, with considerable success, by Bourdieu.53 A natural conse-quence of this is that the subject, the speaker, is not the author of her communication (expression reflecting intention) but at best an effect of the operation of collective assemblages of enunciation (agencements collectifs d'eÂnonciation). Guattari's contribution is the beginning of a theory of assemblages, of which he gives a first definition: their main characteristic is that they do not so much `speak of things' as `speak among things', àÁ meÃme les choses', in the midst of facts, states of affairs and subjective states.54 Such unholy mixture of the material and the ideal is a characteristic, as we have seen, of Deleuze's philosophy of language. Guattari also gives some examples of what such an assemblage might be, most notably the assemblage that produces the `sujet de l'eÂnonciation', the speaker, where it appears that the speaking subject and the psychological or psychoanalytical subject emerge in a field of social forces and institutions of power (institutions of masculine author-ity, repressive features of social `facialities' etc).55

The second contribution is that there is more to semiotics than mere linguistics. Guattari is the indefatigable proponent of a multiplicity of semiotics, only one of which, and not the most important, is a signify-ing semiotics. There are more codes than the signifying codes that govern conscious and deliberate utterance: the relationship between the machine and its driver, the courting rituals of certain birds that make extensive use of blades of grass for signalling purposes are also caught in regimes of signs, and have recourse to pragmatic markers in the same way as linguistic communication does. On the face of it, this is a mere extension of linguistics into semiology, of the type Saus-sure already called for. Except that it is not an extension, which takes language as the model, if not the source, of all semiological processes. Guattari's aim is to put language in its place, one process among many, without any privilege. Thus, he describes three important pragmatic fields, in terms of the assemblages of enunciation and semiotic compon-

ents involved, according to the following correlation:


Symbolic field territorialised assemblage icons and indices Signifying field individualised assemblage semiological triangle Diagrammatic field collective-machinic assemblages particle-signs.

The semiological triangle is the familiar triangle of word, meaning and thing, or, rather signifier, signified and referent. I will not go into his theory of `particle-signs', the signs emitted in machinic assemblages: my point is that this correlation, to be taken diachronically as a histor-ical progression, and synchronically as the exhaustion of logically pos-sible positions, precisely puts the signifying semiotic in its place, as a moment, both historical and logical, in the development of pragmatic fields. Hence Guattari's third and fourth contributions.

His third contribution concerns faces. Not ordinary faces, but faces as signs: the semiotics, or grammar, of faces ( for which he coins the concept of `faciality', `visageÂiteÂ'). The idea is venerable, from seven-teenth-century treatises on types of faces,56 nineteenth-century ac-counts of the facial expression of passions (I am thinking of course of Darwin57), abundantly illustrated with pictures of variously contorted or grinning faces (the faces of mad people naturally provoked peculiar fascination there is a fine series of visages de fous by the French painter GeÂricault), to the suggestion Chomsky once made that a strict grammar of faces, on the same basis as grammar proper, might be constructed. Guattari, of course, would welcome the suggestion, without the causal sequence implied: for him, the semiotics of faces is independent of language. Far from the characteristic features of a face, the signs that allow identity and recognition, being linguistic in origin or by analogy, it is the production of linguistic signs through the voice that is managed, guided, enhanced or commented upon, in other words determined, by the accompanying facial expressions. A short story by Raymond Carver, `The Father', deals with the nonentity of the father in the eyes of his wife and daughters. The story climaxes with the exclamation, by one of the daughters, as she enthuses in the usual way on the newborn son's resem-

blance to his various relatives (we all know the scene: various bits of the extended family's facial anatomy are optimistically projected on to the infant's blank and podgy face): `But who does daddy look like?' And answer there is none.58 Guattari develops these features of our daily experience into a theory of consciousness, which has nothing to do with the Cartesian cogito, contrasting the two aspects of consciousness as the territorialised recognition of the other's face (my identity is projected back on to me through the recognition of the other) and the subjective black hole that causes anguish (my face is the only one that I fail to


recognize the same is true of my voice from lack of experience: I cannot spend my whole life in front of a mirror).59 Individuality, for Guattari, is constructed around this paradoxical constitution of a con-science visageÂifieÂe. I find this materialisation of consciousness, with the co-occurrent idea that my unconscious is not inside me in the depths of my psyche, but outside, inscribed on the surface of my body or on the body of others (I would willingly add, but Guattari might not, inscribed in language), highly invigorating.

The fourth contribution concerns the temporal semiotics parallel to the spatial semiotics of faces: the semiotics of ritournelles, or jingles. The term describes time as experienced in the body, in the form of rhythm, and externalised as a musical beat, or jingle. It plays a similar semiotic role to faces, being associated with identity, the relation of the individual to the group, and therefore the territory (the jingle is the signal that home has been reached, a marker of territory, as in bird songs of which Guattari appears to have been a keen student), and movements of deterritorialisa-tion/reterritorialisation. Jingles, like faces, belong to what Guattari calls `micro-eÂquipements collectifs', instruments of socialisation: they give rhythm to our most intimate sense of time and allow us to externalise it, to make it public. Like faces, they allow us to construct our relationship to the landscape of our lives, to our Lebenswelt. Like faces, they allow redundancy, a term to be understood in its specialised linguistic sense of the necessary repetition that cancels noise and allows successful trans-mission of the message. Redundancy is the main feature of all semiotic systems, it is the main instrument of our construction of reality: faces and jingles are instruments of redundancy as important, if not more, for the said construction, as linguistic communication, which is itself an instru-ment of construction of identity through interpellation and recognition.

Guattari's reading of Proust is conducted in terms of faces and jingles: in this he complements Deleuze's own reading in terms of signs (lin-guistic signs, but all sorts of other signs as well). The developments on visageÂiteÂ and ritournelle will have a familiar ring to readers of Mille plat-eaux: but that is precisely my point, which is to stress the importance of Guattari's contribution to the common work.

6 The concept of a problem

The point of this chapter is double: to show that for Deleuze language is a problem (hence the paradox I have delineated in terms of distrust and fascination); and to extract language as the problem that provides a guiding thread through the whole of Deleuze's philosophy: in other


words to treat Deleuze as he treated Spinoza or Leibniz, to claim that in his work language is what he claims expression is in Spinoza or the fold in Leibniz.

The first point is easy enough to establish. The second is, precisely, more problematic, and needs a short detour. For `problem' is one of the concepts Deleuze has reinvented, and I wish to take it in the special sense he himself uses.

Deleuze's reaction to the first book ever published on his work was one of ill-concealed rage.60 There were reasons for this: the book came from the leftist circles of which he was a prominent member, it was written in the heady 1970s, when nobody (except Chairman Mao) was allowed to be a master, and it contains not only an assessment of Deleuze's concepts, but also remarks (not in a spirit of charity) on his academic position and the length of his nails. In his `Letter to a Harsh Critic',61 Deleuze adopts the same shocking language but he also tells us what writing a book of philosophy, especially a book that seemingly belongs to the genre of the history of philosophy, involves. The answer is simple: buggery. The philosophical commentator `fait un enfant dans le dos aÁ son auteur'. He does not respect his intentions of meaning (for what they are worth), he is not content with respectfully expounding the solutions his author presents. He does not treat the work he comments on as a `boõÃte de signifieÂs', a collection of ready-made meanings, to be reassembled for ease of exposition. What he does is extract a problem from the text, a problem that does violence to the text, of which the author himself may not have been aware, but which enables us to understand how the text works (for ìt', being everywhere at work, is also at work in philosophical texts): he calls this `reading a text inten-sively'.

So the point of a book of philosophy is not the exposition of solutions but the extraction of a problem. More precisely: to create a concept, or concepts, adequate to a problem. Find the problem, Deleuze says in his AbeÂceÂdaire, and you leave abstraction for the concrete.62 Thus, Leibniz invents the concept of a monad (a strange invention, but there is a little madness, as well as a lot of method, in the invention of a concept). And the concept expresses a problem which will wait for Deleuze, and the creation of a new concept, to be made explicit, to be unfolded. The problem that the concept of monad seeks to address, the problem of the subject that expresses the totality of the world, is that the world is a collection of things that are folded within one another. Le Pli is Deleuze's attempt to unfold Leibniz's problem: thus does the history of philoso-phy move forward, in a succession of concepts addressing a succession


of problems. And a problem in that sense has one immediately striking characteristic: it is interesting, rather than true or false.

What exactly, then, is a problem? Well, the short answer is that a problem is not a question neither is it, of course, a solution.

The problem with questions, Deleuze says in Qu'est-ce que la philoso-phie?, is their propositional nature.63 A question is a temporarily sus-pended statement, the bloodless or ghostly double of the proposition it calls as its answer, whereas a problem, the site of creative thought, of the creation of thought, is never propositional (it is formulated as a con-cept). Or again, this time the passage comes from DiffeÂrence et reÂpeÂtition, questions are concealed imperatives that impose their answer on the addressee, or butt, of the question:64 obviously, Deleuze believes that all questions are what linguists call closed questions, which presuppose their answer (Ìs it not the case that . . . ?') in which he agrees with Canetti, in a famous passage from Crowds and Power,65 and with the average policeman in television series: Ì'm asking the questions here!' A problem, of course, is what linguists would call an open question except that it does not take the form of an interrogation, addressed to the reader, and demanding an answer (even the most innocuous òpen question', even `What is the time, please?', is part of a pragmatic struc-ture that assigns places and through which power circulates this is Canetti's point: he is duly mentioned in Mille plateaux).

And a problem is not a solution. Deleuze's theory of the problem is 66most explicitly expounded in DiffeÂrence et reÂpeÂtition. A problem has

three characteristics: (1) it is different in nature from a solution; (2) it transcends the solutions it generates; (3) it remains immanent in the solutions that eventually come to replace or conceal it which means that a problem does not disappear when a solution is produced. This is the position of Lestrade (in The Adventures of Sherlock Holmes), who wishes to arrest somebody, whether he is the culprit or not, and of Dr Watson, who exults in the results achieved by his friend; this is not the position of Sherlock Holmes himself, even less of his elder brother Mycroft, the real genius, who mulls over the problems and never goes on all fours to collect fragments of ash in order to come closer to the solution and establish proof. As we all know, what is interesting in a Sherlock Holmes story is not the solution (which is always a disappoint-ment and announces the death of the text), it is the problem: why should a red-headed shopkeeper be employed to copy the Encyclopaedia Britannica for three hours every afternoon? The answer so that a gang of villains can bore their way through the wall of his shop into the vault of the bank next door is as trivial and uninteresting as can be: the


problem is brilliantly original and remains (the best parts of the stories are the moments when Watson, in two lines, evokes other cases which he never fully chronicled, like the case of the man who went raving mad when he opened a matchbox that contained a worm of a type nobody had seen before).

The reason why the problem insists and persists in the solutions which it needs (the existence of the problem is dependent on the search for solutions, as the adventures of Sherlock Holmes clearly show) is that the solution is a particular proposition, whereas the problem is ideal, is an Idea, which no solution can ever exhaust or solve away. Again, a comparison with literature may allow us to understand better what this means. It is a characteristic of a text, and most clearly of a persistent text, like a literary text, that it both transcends the interpretations to which it gives rise (and which purport to `solve' its meaning), and is immanent in them (in that there is no text without a reader and a reading and any reading is an interpretation, or a solution). But no interpretation, or solution, is true, although some are more just than others.67 No interpretation is so good as to preclude the emergence of an endless series of other interpretations even if the text is a detective story, and the interpretation is given within the text by the author in the form of the solution to the mystery. Even if Agatha Christie authorita-tively answers the question: `Who killed Roger Ackroyd?', nothing will prevent an astute and complicated reader from claiming her solution was a decoy and suggesting another whether the result be a postmod-

ern novel or an essay is immaterial.68

In a curious passage, Deleuze finds a linguistic marker for the presence of a problem (he does not claim its presence is necessary, merely indica-tive): what French linguists call the `NE expleÂtif ', because they are unable to explain it (àvant que tu ne viennes', `before you come': no negation is involved here). This modal particle, this èmbarrassment' to linguists,69

is the mark of non-being. And le non-eÃtre, far from being merely nega-tive, is of a problematic nature (this is part of the affirmative, anti-Hegelian aspect of Deleuze's philosophy).

It remains for me to show why I think language in Deleuze is a problem in that sense. I shall first note, en passant, that the concept of a problem is often determined and exemplified in terms of language, from the concealed imperative in questions to the `ne expleÂtif' that linguistically marks a problem. And there are at least four symptoms of a problematic treatment of language in Deleuze.

The first symptom is what I have called the paradox in Deleuze's attitude to language. The main characteristic of a true paradox is that


it insists in the solutions that attempt to put an end to it, and that it tends therefore to persist. The second symptom is the hostility to linguis-tics, as the science of systematic solutions to the problem of language, to the problem raised by language. What Deleuze reproaches linguistics with he is not the only one is that the corset of rules it seeks to impose on language gives an impoverished and often downright coun-terproductive view of the workings of natural languages. Hence his pre-ference, and mine, for the practice of poets who are both more respectful of language, in that they do not seek to manacle it, and less respectful of it, in that they take it to the limits of its potentialities, and sometimes into downright silence, as we saw with Beckett. The third symptom, which shows that the problem survives all attempts at a solution, is the proliferation of part theories (which will be duly examined in the next chapters) in Deleuze's works theories that never explicitly form a whole (the perilous attempt to extract some form of philosophy of language from them is left to me) and may not even be coherent. The fourth symptom, therefore, is the immanence of the problem of language in the multiplication of part theories, combined with the global tran-scendence of the general problem of language, through and beyond those theories at least in the idea that impelled me to write this book, the idea that there is a Deleuzean programme of research in the field of language, one that would not have to pay the heavy metaphys-

ical price to be paid for the Chomskyan programme. I have just sketched the plan of the rest of this book.

Interlude 1

Images of Trains, Trains of Images ± Dickens, Deleuze, Representation

1 `The train of his description'

In Graham Swift's Ever After, the Victorian hero, Matthew Pearce, de-scribes the opening of the Bristol and Exeter railway for the benefit of his future wife and her father, the local vicar:

Matthew does not overdo his scene painting, but his voice betrays a curious urgency, and he finds himself gratified, peculiarly enlight-ened, by the looks of astonishment and vague fear on his listeners' faces. As the train of his description is greeted by the cheering crowds at Exeter, a train of thought passes through his mind that fills him he cannot say how much it has to do with Miss Hunt's look of appealing vulnerability with a sudden gush of liberating relief.1

The train he is describing turns out to be a rather complex object. It is a real train a referent is uniquely picked up by the linguistic signs that make up the description, he is talking about the first train ever to reach Exeter. But it is also, already, a mythical train, carried by, and dragging in its train, a number of clicheÂs and discourses: this train is a symbol of progress and enlightenment, hence the ùrgency' in Matthew Pearce's voice and the `vague fear' evoked in the listeners. Its not quite con-trolled violence, a fine illustration of modernity, is also (again, already) a metaphor for the unleashing of affect, as at the end of the passage Matthew falls in love with Miss Hunt, with à sudden gush of relief', which can be seen as analogous to a burst of steam. In short, this train is already, always-already, caught in the intertextual and intratextual web of story and history, where history is both personal, with a small `h' (the twentieth-century narrator of the novel will soon learn that his real

41


father was a railway engineer, not the official diplomat) and political, with capital `H' (as the narrator rejects the demands of his colleague, the professional historian, who claims that only he possesses the historical knowledge necessary to be the editor of Matthew Pearce's newly dis-covered diary).

What interests me most in this passage, of course, is the mise en abyme that justifies the inversion in my title. It follows the metaphorical drift from `train' to `train of thought', which is in fact the inverse of the historical drift of linguistic change, as the phrase `train of thought' already occurs in Hobbes, whereas the first occurrence of `train of rail-way carriages' only goes back to 1830. More importantly, it contains a genitive, `the train of his description', which out of context is ambigu-

ous between a subjective and an objective interpretation. If the genitive is subjective, the description is a train, of thoughts, images or words, thus reminding us of the constitutive link between language, in so far as it is embodied in ordered linear sequences, and the word `train', whose earliest meanings are `proper sequence' (first used in 1528, according to the Oxford English Dictionary) and òrdered sequence of elements' (first used in 1610). If the genitive is objective (which in the context is the only actual interpretation my contention is that the subjective inter-pretation is present as a resonant virtuality of meaning), the train is the object evoked.

The ambiguity of the genitive phrase is important, because it is the site of a linguistic tension with wider implications. If the genitive is objective, there is an object, which is presented through language and represented by the signs of the text (a phrase that uniquely picks up a referent is what philosophers call a `definite description'). And there is also, around that object, a literary description, a lull in the narrative, a delay in the devel-opment of the plot (and ̀ delay' is also an archaic meaning of ̀ train', from the French verb `traõÃner', in the sense of `dawdle') the rationale for this moment of stasis in the dynamics of the plot being that it enables the describer to present and re-present the object. But if the genitive is subjective (and we may, of course, decide that it is both), if the train is the description, there is no object, only a sequence of words and/or ideas; and there is no stasis, no moment of reflective delay, but only the abstract progress of teleological development, as each element in the ordered sequence follows its predecessor, in a series of increasingly constrained choices, as in a Markov chain, which pre-Chomskyan linguists thought might provide a model for language.

The aim of this interlude is to explore this paradox, etymologically to explicate the ambiguity of the genitive. It will follow the path that

Interlude 1 43

the genitive, or the antimetabole in the title, indicates. In so doing it will go through many dark tunnels, possibly reaching the dizzying speed of 30 miles per hour. Whether it will go anywhere is another matter.

2 The paradox of representation

The general name of the genitive paradox is the paradox of representa-tion. If there is inevitable difference in repetition, how can there be a re-presentation, a second presentation that captures the immediacy of the first? This is not merely a Heraclitean train of thoughts, not merely a reflection on the passing of time it is also a question of a change of medium, as between the represented object or scene and its representing image or sign, there is an ontological gap, which language inscribes in another pun on the word `train'. `Train' does not mean only an ordered sequence (the order of isomorphism, of one series running parallel to the other, like a pair of rails), but also, in an archaic sense, a trick, a betrayal (this one is, according to the OED, a different word, derived from the Old French `traõÈne', hence `trahir' a case of antanaclasis, not metaphorical drift). This gap, the ideology of representation, which in literature culminates in the eminently nineteenth-century genre of lit-erary description, has always tried to bridge. The object of this interlude, therefore, is to attempt an analysis of the literary version of the myth of representation, spelling out its internal logic and its inevitable dissol-ution into another type of logic, the elements of which are already implicit in the aporia of representation. I have just sketched a train of argument, along which I shall progress in four steps or stations, calling at various texts about trains, culled in Dickens's Dombey and Son or Lewis Carroll's Through the Looking-Glass. As we shall see, all of my texts are àbout' trains, but the status of their representation will vary enor-mously, until another logic emerges, which I shall call, in Deleuzean terms, a `logic of assemblage'.

3 `The yet unfinished Railroad'

In the second chapter of Terry Coleman's history of the building of the railways in Britain, The Railway Navvies,2 there is a long quotation from Dombey and Son, the description of the cutting on Camden Hill, as the London to Birmingham line leaves London. This quotation is associated with plate no. 9 in the illustrations of the book, the legend of which says: `Building the retaining wall at Camden on September 17th, 1836,


probably the scene described by Dickens in Dombey and Son.' In spite of the slight uncertainty betrayed by the use of `probably' (where the aporia of representation makes itself felt), this is a classic case of the literary description as vignette, which implies (a) the immediacy of the link between the description and its referent (the impression that this is it, that the thing itself is present, before our eyes, or before the eyes of our imagination hence what Barthes calls èffet de reÂel'); and (b) the mutual translatability, even substitutability of the word-description and the image-description, the text and the picture. The text enables us to recognise the picture, which in turn confirms and guarantees the adequacy of the literary description. This is the epitome of the logic of representation: here, to pastiche Rorty's title, discourse is the mirror of nature. Here is Dickens:

The first shock of a great earthquake had, just at that period, rent the whole neighbourhood to its centre. Traces of its course were visible on every side. Houses were knocked down; streets broken through and stopped; deep pits and trenches dug in the ground; enormous heaps of earth and clay thrown up; buildings that were undermined and shaking, propped by great beams of wood. Here, a chaos of carts, overthrown and jumbled together, lay topsy-turvy at the bottom of a steep unnatural hill. There, confused treasures of iron soaked and rusted in something that had accidentally become a pond. Every-where were bridges that led nowhere; thoroughfares that were wholly impassable; Babel towers of chimneys, wanting half their height; temporary wooden houses and enclosures, in the most unlikely situ-ations; carcases of ragged tenements, and fragments of unfinished walls and arches, and piles of scaffolding, and wildernesses of bricks, and giant forms of cranes, and tripods straddling above nothing. There were a hundred thousand shapes and substances of incom-

pleteness, wildly mingled out of their places, upside down, burrowing in the earth, aspiring in the air, mouldering in the water, and unin-telligible as any dream. Hot springs and fiery eruptions, the usual attendants upon earthquakes, lent their contributions to the confu-sion of the scene. Boiling water hissed and heaved within dilapidated walls; whence, also, the roar and glare of flames came issuing forth; and mounds of ashes blocked up rights of way, and wholly changed the law and custom of the neighbourhood.

In short, the yet unfinished and unopened Railroad was in pro-gress; and, from the very core of this dire disorder, trailed smoothly away, upon its mighty course of civilisation and improvement.3

Interlude 1 45

The passage possesses all the characteristics of description according to Philippe Hamon's classic account.4 We find a number of existential sentences (`there were a hundred thousand shapes . . . '); the elements of the description are presented in order (`Here, a chaos of carts . . . ; there, confused treasures of iron soaked . . . . Everywhere were bridges that led nowhere'); the various sentences that make up the description are linked together through mere juxtaposition: paratax, says Hamon, is the natural language of description, as it is of counting (one and one and one); lastly, we may note the delayed appearance of the `pantonym', the sign whose direct referent is the object of the description: it occurs in the three-line summary at the end of the text (Ìn short, the yet unfin-ished and unopened Railroad was in progress'). As a result, the whole text gives the impression of being (a) a list of items and (b) ordered like a taxonomy, as the eye of the describer methodically surveys the scene, even if the scene is one of confusion indeed, this is the best illustration of the powers of description, in that the passage manages to construct a textual cosmos out of the chaos of referents: `there were a hundred thousand shapes of incompleteness, wildly mingled out of their places, upside down, burrowing in the earth, aspiring in the air, mouldering in the water, and unintelligible as any dream'. In spite of the insistent vocabulary of topsy-turveyness and shapelessness, the scene, caught in the neatly balanced rhythm of syntax, is perfectly in order, as the draughtsman provides form and order to the unruly topography that stretches in front of his eyes. Lastly, (c) the whole text gives the impres-

sion of a moment of stasis, as the curious onlooker stops in his walk to survey the scene, which delays the progress of the narrative. A textual window is opened in the middle of the plot, a place of rest for the reader to recover her breath, and yearn for the next disclosure.

The general effect of the description, therefore, is one of immediacy. But it will soon be apparent that such immediacy is mythical. There is a pane in that window, which deforms as much as it reflects. The name of the game is not immediacy but iconicity. The description is a series of rhetorical devices that mime reality rather than present it. Description, Hamon says, is an intertextual before it is a referential game: the indi-vidual description imitates textual models rather than a problematic reality. Even as John Martin's visions of the Apocalypse are partly in-spired by the sight of canals blasting their course through rock,5 Dick-

ens's description of the railway cutting is inspired by intertextual memories of apocalyptic scenes (`the first shock of the great earthqua-ke . . . Babel towers of chimneys, wanting half their height . . . ). This is particularly apparent in the abyme of the last paragraph, which confesses


that the description is also a train of words, in which the ordered sequence of the text reproduces the order of the railway train.

Yet the aporia of description is far from being total. The text exerts illocutionary force on the reader, a mythical effect is produced, which is one of presence. The text is so real as to become a primary source for social history, a purple passage for whoever wishes to celebrate Dickens's novel as a reflection of the railway age, on its way towards Zola's La BeÃte humaine. The contents of the myth of description as representation have the following six characteristics. (1) The text illustrates reality, with the same increment as in a photograph, but no more: reality is filtered through the choice of point of view, through the construction of com-

position, and there is always the possibility of an unexpected Barthesian 6punctum. (2) But reality is presented by the text, as it is: the filter selects,

but preserves the overall shape. (3) As a result, the text is an analogon of the picture. (4) The text has the property of surviving the reality it re-presents or supplants: this is the stuff that memories are made of. Not only do we `perceive', on reading Dickens, Camden as it then was and is no more, but the first readers were already presented with a defunct scene, since, later in the novel, we learn that the line has been com-

pleted. (5) The text makes reality public, that is, recognisable, through the individual means of style. Description gives the describer control over the phenomena, like a kind of extended ontological metaphor. (6) The comparison between description and ontological metaphor is not innocent. The filter of description does not only select, it also abstracts and simplifies. This is Camden railway cutting; this is also a railway cutting in the 1830s, an element extracted from a class all railway cuttings are susceptible of the same (or at least a similar) descrip-tion, hence the `probably' in the legend of plate no.9. Unless a signpost, with a date, is there (an unlikely occurrence), one can never be sure. The presence of the pantonym, itself an ontological metaphor, ensures this.

The six characteristics are incoherent. A general effect of representa-tion qua reflection is produced, but the generalisation of the description threatens to ruin the presentation into abstract symbolism: no longer a scene, but an archetype.

4 Retribution

He heard a shout another saw the face change from its vindictive passion to a faint sickness and terror felt the earth tremble knew in a moment that the rush was come uttered a shriek looked round

Interlude 1 47

saw the red eyes, bleared and dim, in the daylight, close upon him was beaten down, caught up, and whirled away upon a jagged mill, that spun him round and round, and struck him limb from limb, and licked his stream of light up with its fiery heat, and cast his mutilated fragments in the air.

When the traveller, who had been recognised, recovered from a swoon, he saw them bringing from a distance something covered, that lay heavy and still, upon a board, between four men, and saw that others drove some dogs away that sniffed upon the road, and soaked his blood up, with a train of ashes.7

This famous text describes the death of the villain, Carker. To the best of my knowledge, this is the first time that the locomotive, a God-like machine rushing in, ex machina, at the moment of retribution, is being used as the instrument of the villain's execution the first time, but not the last, especially in the cinema.

The text does not quite read like a description in Hamon's sense rather like a riddle. First, it consists in the diagrammatic evocation of an event, the plot rushing at full speed towards one of its climaxes, rather than a moment of stasis delaying the action even if we do have a scene, as in description proper, with a point of view, or rather two, Carker himself in the first paragraph, and Dombey in the second, who appro-priately swoons to spare the reader the goriest details. Secondly, and this is a defining characteristic of description, the core of the text consists in the notation of sense impressions: Carker `heard', `felt', `saw'. Thirdly, there is no pantonym, except in the last sentence, where it is concealed beneath an older sense of `train': Carker's blood is soaked up by a `train of ashes'. This is indeed a riddle, a description in which the referent is concealed and the pantonym withdrawn. The effect is welcome as Carker cannot know, at least at first, what is happening to him.

But the text moves even further away from description proper: it moves into myth. This is not only the train that killed Carker, not merely a train to London: this is a mythical object, an archetype. For that train is not even described: we have a frame of actions and emotions, but not the framed source of these emotions, as if the train came and left so fast that it could not be seen. This is the description of an absence, through metaphor and metonymy (for instance the synecdoche cum metaphor of the `red eyes'). The reason why such tropic displacement does not hinder recognition is simple: the tropes are heavily intertextual, the frame for the absent object is informed by our encyclopaedia. The reader always-already knows that the train is a metaphor for the devil (those


red eyes!), a modern embodiment of the Juggernaut, that the victim is drawn and quartered by the steam horse, and literally put to the wheel this is indeed an execution. And we have enough history to remember the fate of Huskisson, run over by a train in September 1830, on the day of the inauguration of the Liverpool and Manchester line he was killed exactly like Carker, and is said to have been dazed and numbed by the train rushing towards him as by a snake. Lastly, the detail of the dogs sniffing at the blood on the line turns Carker into a modern version of evil king Ahab, whose fate is told in 1 Kings 22:38.

This passage from description to myth is a passage from reflection to refraction. Instead of the illusory, unmediated relationship called re-presentation, between the described object and the linguistic signs of description, we have the mediated relationship of a work of refraction, as one talks of the dream-work. In his essay `The Word and its Social Function', VoloÂshinov gives an account of `the internal dialectic of the sign':

All of reality, the entire being of man and nature are not simply reflected in a sign, but also refracted in it. And this refraction of being in the ideological sign is determined by the crossing of differ-ently social interests within the parameters of the sign collective, that is, in the class struggle.8

I take the explicit Marxism of the end of the passage as a compen-

satory symptom for the displacement of the classic Marxist concept of reflection that the beginning operates. The train in literature is not a reflected object but a motif, a refracted sign, the embodiment of a multiplicity of conflicting discourses, of differently directed social inter-ests. As another illustration of this, we might read the following text, the epitaph of a navvy, culled from Coleman's book:

`The Spiritual Railway' The line to heaven by Christ was made With heavenly truth the rails are laid From Earth to Heaven the line extends To Life Eternal where it ends

Repentance is the Station then Where Passengers are taken in No Fee for them is there to pay For Jesus is himself the way

Interlude 1 49

God's Word is the first EngineerIt points the way to Heaven so clear,Through tunnels dark and dreary hereIt does the way to Glory steer

God's Love the Fire, his Truth the Steam,Which drives the Engine and the TrainAll you who would to Glory ride,Must come to Christ, in him abide

In First and Second, and Third ClassRepentance, Faith and HolinessYou must the way to Glory gainOr you with Christ will not remainCome then poor Sinners, now's the time

At any Station on the LineIf you'll repent and turn from SinThe train will stop and take you in.9

The date is 1845, three years before Dombey and Son, and the discourse about the railway already appears to be the site of a conflict between social interests, and also of its resolution. For if the class struggle, pace VoloÂ-shinov, is not directly present in the text (unless we read it into the allusion to first, second and third class I rather like the idea of meeting my Maker with a third-class ticket in my pocket), we can hear in it the echo of the social contradiction between the older values of religion and the new values of scientific progress and industry. We may also see the text as an attempt to deprecate the social danger represented by the godless, boozing and violent railway navvy, a terror to the respectable.

The epitaph shows that refraction works through metaphor. The skel-eton of the text is a structural metaphor, in Lakoff and Johnson's ter-minology, which moves from the linguistic tautology, à RAILWAY is a WAY' to the religious metaphor, `the RAILWAY is the PATH to GLORY'.10

Unless, of course, it is à PATH to PERDITION', as it is with Carker. The Pilgrim's Progress must be modernised: Christian reaches his goal by railway, and junctions replace crossroads. What the epitaph, in its naively methodical style, tells us, is that refraction is not analogical or holistic, but digital. It works through the development of parallel series. It also works through generalisation and abstraction: the metaphor of the railroad to glory goes from concrete (the train) to abstract. Indeed, it


is always-already abstract, as the train is merely the latest embodiment of one of the oldest abstract schemata: the path to Heaven is the path par excellence.

5 Imitative harmony: on our way to Leamington Spa

The move from reflection to refraction is one from the immediate to the mediate, from the concrete to the abstract, from the particular to the universal and the archetype. Yet we are still within the logic of repre-sentation. Perhaps one more passage from Dombey and Son will take us one step further.

In Chapter 20, Mr Dombey takes a train, on the line we saw under construction in Chapter 2, in order to go to Leamington Spa. At the station, he meets a member of the new railway proletariat in the person of Mr Toodle, who is the engine's stoker. Hurt in his pride when he realises that Mr Toodle dares to wear a piece of black creÃpe as a sign of mourning for Paul Dombey, his son, Mr Dombey, as the train starts, is a prey to sombre thoughts. The journey has no joy for him, as all his musings are about death:

He found no pleasure or relief in the journey. Tortured by these thoughts he carried monotony with him, through the rushing land-scape, and hurried headlong, not through a rich and varied country, but a wilderness of blighted plans and gnawing jealousies. The very speed at which the train was whirled along, mocked the swift course of the young life that had been borne away so steadily and so inexorably to its foredoomed end. The power that forced itself upon the iron way

its own defiant of all paths and roads, piercing through the heart of every obstacle, and dragging living creatures of all classes, ages and degrees behind it, was a type of the triumphant monster, Death.

Away, with a shriek, and a roar, and a rattle, from the town, burrowing among the dwellings of men and making the streets hum, flashing out into the meadows for a moment, mining in through the damp earth, booming on in darkness and heavy air, bursting out again into the sunny day so bright and wide; away, with a shriek, and a roar, and a rattle, through the fields, through the woods, through the corn, through the hay, through the chalk, through the mould, through the clay, through the rock, among objects close at hand and almost in the grasp, ever flying from the traveller, and a deceitful distance ever moving slowly within him: like as in the track of the remorseless monster, Death!

Interlude 1 51

Through the hollow, on the height, by the heath, by the orchard, by the park, by the garden, over the canal, across the river, where the sheep are feeding, where the mill is going, where the barge is floating, where the dead are lying, where the factory is smoking, where the stream is running, where the village clusters, where the great cath-edral rises, where the bleak moor lies, and the wild breeze smooths and ruffles it at its inconstant will; away, with a shriek, and a roar, and a rattle, and no trace to leave behind but dust and vapour; like as in the track of the remorseless monster, Death!11

As we can see, the first paragraph of the text develops a similar sort of structural metaphor as the epitaph: the course of life is a railway jour-ney. Not so much a metaphor, in fact, as a simile, since the compared elements are made explicit in the text. The railway engine, in a touch of dramatic irony that announces the death of Carker, is explicitly de-scribed as à type of the triumphant monster, Death'.

The next two paragraphs, however, are another matter: a famous purple passage of imitative harmony, where there is no image of a train, not even a train of images, but rather a train of thoughts and feelings, the feelings of Dombey, who is not thinking of the train or the landscape, but whose body subconsciously registers the rhythms of the journey. Such evocation is no picture, no description. True, railway tunnels and cuttings are mentioned, and at the end of the passage the railway is used as a metaphor for the opposition between the two nations a means of communication that passes by and fails to link up. Nevertheless, the main point lies elsewhere, for, ironically, this time the train is present, physically present, at the very moment when its description has vanished. It is present through other senses than vision, mostly through the sense of hearing, in the rhythm of the sentences, in the repetition of alliteration, in all the devices that make up imitative harmony. Again, we have a form of polyphony, as in the preceding section, where the voice of the text was echoed by a plurality of inter-textual voices. Here, the voices are two. There is the voice of the signi-fied, as the contents of Mr Dombey's `small perceptions', to speak like Leibniz, are conveyed to us. And there is the voice of the signifier, inseparable from the signified and yet singing its independent tune. Those two paragraphs are the acme of the logic of representation (as the object is at last present in person, not through the medium of a visual image); they are also its ruin: the train, present as never before, is also entirely absent from the text. But this absence may be an optical illusion due to its very centrality. What is the referent of the absent


subject in all those participle or prepositional clauses? Is it the traveller, Dombey, or is it the train itself? The only answer to this question is grammatical and rhetorical. The poetic function of language has taken over, in the form of a paratactic prepositional style, which has the extreme interest of being anapaestic (`with a shriek, and a roar, and a rattle, from the town . . . '), a naturally leaping rhythm, as witness Coler-

idge's famous definition: `With a leap and a bound the swift anapaests throng.' The result, of course, is a musical text, a tour de force, which is the linguistic equivalent of Honegger's Pacific 231.

This is why the logic of representation finds both its acme and its demise in this text. On the one hand, the train is physically embodied in the voice that reads the text aloud. The speed and noise of the train are felt directly, as in Honegger: evocation is sensuous translation. But, on the other hand, the train is totally abstracted. It is reduced to the linear sequence of the rhythmic beat, what you perceive at night when you lie awake in your sleeper. And in turn this symbolic reduction is symbolised by the sequence of anapaestic prepositional phrases, paratactically joined up like so many carriages. The process of symbolisation and abstraction is in fact so thorough that what we have is the essence of the train, the common element in a railway train, a train of thoughts, a sequence of words, a rhythmic musical beat: the hyperonym that organ-ises the metaphoric ballet of the meanings of the word `train' in the OED, in other words a Kantian schema, the schema of the path, one of those abstract archetypal schemata that Mark Johnson describes in The Body in the Mind.12

I propose to call the abstraction aspect of the in-between stage we have reached (in between abstraction and embodiment, as the Kantian schema is in between understanding and intuition) sublimation. The schema of the path sublimates the sensuous energy of the embodiment into an abyme of the abstract musical or linguistic sequence as pure linearisation of time. The inversion in my title unwittingly illustrates the potential contradiction between embodiment and sublimation: images of trains are also trains of images; thoughts about trains are trains of thought. The choice of the word `sublimation' of course draws on a Freudian analogy. In Freud, sexual libido, while remaining sexual, is transferred to a non-sexual cultural object. Here, the sensuous force of the embodiment is transferred, translated, through the mise en abyme, to a non-sensuous, abstract, semiotic pattern. Or again, the mimetic force of representation (the signified is an image, acoustic or visual) is trans-ferred through the sublimation of metaphor to a semiotic pattern of signifiers, which are non-representational. We go from imitative har-

Interlude 1 53

mony to a linguistic sequence, from trains of images to strings of words.

6 Àlice through queen's third (by railway)'

This station has proved to be the end of the line of representation, which dissolves in the contradiction between embodiment and sublim-

ation. The next question is: is there another station beyond, which we might reach, perhaps at the cost of changing the gauge of the railway? The answer is: there is.

Our next station, then, comes from Chapter 3 of Lewis Carroll's Through the Looking-Glass.13 In the game of chess that provides a pretext for the text, the chapter corresponds to Alice's first move: Àlice through queen's third (by railway) to queen's fourth (Tweedledum and Tweedle-

dee)'. The image of a train in Carroll is very different from what Dickens

gave us. Alice's journey is obviously undergone under the twin signs of confusion and mixture. Confusion indeed reigns supreme. We have ser-ious doubts about the reality of the scene (trains do not usually jump over brooks), about the logic of the sequence of events (the guard first looks at Alice through a telescope, then a microscope, then an opera glass). And we have doubts about the distinctions around which our reality is constructed. Thus, the infinitely large appears to be confused with the infinitely small; thus, Alice is labelled `lass, with care', thereby blurring the distinction between human and mineral; or again, Alice is treated like a telegraphic message, thus blurring all ontological distinc-tions. Such confusion is due to a state of unholy mixture, where what our reality keeps separate becomes confused. Ticket and ticket-holder have the same size, and seem to be on the same ontological level (the same happens with the playing cards in Alice's Adventures in Wonder-

land); words are no longer distinguished from ideas, etc. The centre of this confusion is a metaphorical drift, an incipient pun, on various senses of the word `train' perhaps, or of `communication'. The train is a means of communication, and communication as exchange of ideas seems to be difficult for Alice and her fellow passengers.

This is no longer an image of a train. Space has changed, it is no longer the form of intuition that frames our usual pictures, visual or verbal. It seems, for instance, to be expandable (`What a number of people there are in this railway carriage', Alice thinks). This is no longer the work of the logic of reflection, refraction or even schema, but a new logic, in which the contradiction of embodiment and sublimation has been


resolved. In short, this is the Deleuzean logic of unholy mixtures, an AND rather than an INSTEAD OF logic (ìnstead of' is the motto of representation). Such is the logic of the Deleuzean àssemblage of enun-ciation' as opposed to `representation of reality'. The canonical example is to be found in Mille plateaux:14 it is the feudal assemblage of enunci-ation, which is a chain, on the same level, of items belonging to various ontological regions, castles, knights and their horses, laws of land tenure, poems of courtly love etc.15 again, we have a sense of unholy mixture. It is not difficult to grasp those various elements in the same thought, but that is not what Deleuze and Guattari mean: we must envisage their material conjunction. And this is precisely what is happening in Carroll, where we find an image not of a `real' train, but of the Carrollian assemblage of enunciation, a series of machinic coup-lings, which include, on the same ontological level, engine and railroad, ticket and ticket collector, passenger and Bradshaw's guide, together with the iconic rhythm of the discourse. What we have is no longer an image in any sense, no longer a representation of a train, but a train of heterogeneous carriages, the whole of which constitutes a kind of text-ual machinery, through which the assemblage works.

The process I have outlined can be recapitulated thus. First, there is a train, a concrete object, a part of solid reality. Then, this train is reflected in a description that gives an image of it with minor displacement due to the choice of point of view: significant details are privileged, the overall form is stressed. Later, the train is refracted into a symbol, it becomes a mythical object in a tradition of mythical objects, from dragons to railway engines. Later still, it is embodied and sublimated into a schema, the schema of the path, as iconically represented by the linguistic and prosodic sequence. Lastly, with Carroll's text, the train is metamorphosed into an assemblage, where objects, ideas, institutions and persons freely mix, where we no longer have the two separate series of isomorphism, where metaphor becomes metamorphosis. Nor is this a logical, or chronological progression, but rather a circle. It can be argued that the train as real object is the product of a collective assemblage of enunciation. The logic of representation followed a Platonist railroad: it went in a straight line from the real thing to the icon (or true represen-tation) and then to the idol (false, or aesthetic, representation). The logic of assemblage is circular: not only are images of trains the products of trains of images, but trains themselves are produced by discourse, as much as they induce discourse.

At the Carrollian stage, the train is no longer a metaphor or symbol. It is both literalised (this train does not represent anything, not even the

Interlude 1 55

class of trains, it is just a train) and metamorphosed. This train is not only in thought (the tale is a dream), it is of thought, made of thought, roaming in the indeterminate space of thought, a pleasure train no longer conforming to the reality principle and the judgement of exist-ence: a desiring train, where middle-aged gentlemen meet little girls and steal kisses from them. The distinction between inner and outer world, mind and object, thought and referent, representative and represented, is blurred, and the result is a dream-like metamorphosis. This is how the passage ends:

In another moment she felt the carriage rise straight into the air, and in her fright she caught at the thing nearest to her hand, which happened to be the Goat's beard.

* * * * * * * * * * * * * *

But the beard seemed to melt away as she touched it, and she found herself quietly sitting under a tree, while the Goat . . . was balancing itself on a twig over her head.16

The goat, invisible up to this moment, is now the size of a chicken. Nonsense is the literary name for the assemblage of enunciation this is why nonsense texts are so often presented as the narratives of dreams.

7 Correlation

My recapitulation can take the Deleuzean form of a correlation between a mode of relationship involving the text and its object (the description of the train and the train it describes) and a literary genre or sub-genre (see Figure 1). The two bottom rows merely develop the two top ones, and need little comment, except perhaps to note that metaphor com-

bines imitation (one talks of `syntactic metaphor' in order to account for cases of iconicity) and abstraction. Lakoff and Johnson's òrientational metaphors', which metaphorise the world in terms of the speaker's body, are iconic; their òntological metaphors' are merely another name for what we call abstraction.17 And the effect produced on the reader by metaphors is one not of presentation, or abstraction, but of translation: embodiment transfers into another medium (the visual into the acoustic, the acoustic into the linguistic), and sublimation is a displacement in the quality of the effect involved.


1 2 3 4

Embodiment

I Mode of operation Reflection Refraction Assemblage

Sublimation

II Literary mode Vignette/Description Myth Trope/Schema Dream

Iconcity

MetaphorIII Textual object Image Symbol Metamorphosis

Abstract schema

IV Effect on Presentation Abstraction Translation Confusion and relationship mixture text/reader

Figure 1 Correlation

8 Logic of representation

The time has come to offer a more general account of the two logics at work, to deal with representation in general. The time has come to go back to Deleuze, and to his notorious hostility to representation.

The concept of representation is constitutively ambiguous. There is a metaphoric drift between a first acceptation of the term, where repre-sentation is another presentation, the return of the same as a new occurrence, and a second acceptation, where representation is replace-ment, where x represents y in y's absence: this acceptation focuses on difference, not repetition or identity.

From this, we may (this is where stipulation comes in) derive five characteristics of representation. The first is difference, the difference between representative and represented, which do not belong to the same order of being (in other words, the second acceptation overrules the first: representation as repetition strives towards identity, but can never achieve it). The second is separation: there is no continuity, or contiguity, between representative and represented, but on the contrary a break or gap. The third is replacement: the representative is present in the absence of the represented, because of its absence, which it palliates. The fourth is hierarchy, the hierarchy between representative and repre-sented. They are not merely different and separated, they are also differ-ently valued. The direction of the hierarchic domination varies, and the represented is not always deemed superior to the representative, quite the contrary. The fifth is generalisation or abstraction, as the representa-tive is abstracted from the represented.

Interlude 1 57

We can illustrate these five characteristics by considering the image as the archetype of representation. The image, on paper or in the mind, is indeed different from the object it represents (an image of a train is not a train, Magritte's pipe is not a pipe). It is also obviously separated from it (we could superimpose a hallucinated train upon a train, but they would remain separate: the carriages of our imagination cannot be hitched to the engines of reality). The image, in the guise of a photograph, a picture, a description or a memory, does replace the object, as our first text from Dombey and Son replaces a vanished Camden. The object is more real than the image, which in turn is more persistent than the object: whatever the sense of the hierarchy, object and image are not on the same rung of the ladder. Lastly, the image is always threatening to become an icon there are conventions (of composition, for instance) in photography as in painting, which tend to abstract and generalise the image from the phenomenal object.

It is possible to produce a declension of these five characteristics in the related fields of representation: the word as representative of its referent, metaphorical meaning as representative of literal meaning, symbol and symptom as representing the Freudian Ding, or even the political system of democratic (or undemocratic) representation. And it will be found that the characteristics do obtain, sometimes against the grain of ap-pearances. For instance, it might be objected that in the democratic system of representation, there is no difference between representative and represented, every citizen being eligible, and the member of parlia-ment elected being originally only a citizen among others: thus, my first characteristic seems not to apply. However, as everyone knows, this is mere appearance. A process of transmutation has to be undergone before one can successfully contest an election. One is first selected by the apparatus of a political party (an independent candidate can only get in by a fluke): only the elect can be elected. And, once elected, the MP is no longer an ordinary citizen. She enjoys legal privileges, and the difference between her and her electors is embodied in political think-ing as `the independence of the MP from her constituents'.

I shall spare you the complete declension, and shall concentrate in-stead on the case of metaphor, as the acme of the logic of representation and the point where it gives way to another logic. The first characteristic applies to metaphor in that it consists in a transfer of words which in turn evokes a transfer of images: the shift establishes a link between word-

representations, which in turn represents a link between object- repre-sentations. This amounts to saying that metaphor belongs to the species of the sign. The second characteristic, separation, is also present in that


the literal relationship being by definition true, the metaphorical rela-tionship is blatantly false: Richard is emphatically not a lion, nor is Sally an ice- cube. Metaphor is a false sign, separated from the true literal sign that picks up its referent. The third characteristic, replacement because of absence, also concerns metaphor. The evoked tenor of the metaphor (the literal lion that Richard is not) is doubly absent: as replaced by the true sign, as displaced by the false sign. Metaphor is the idol to the literal icon: it completely cancels the point of departure of the signifying process, which normally is to be found in the object. The fourth characteristic, hierarchy, has always been central to the theory of metaphor: the usual posture will incite us to state that the literal meaning is more important than the metaphorical one and precedes it. This tradition of hostility to metaphor (to be found, for instance, in Locke and Hobbes) has its oppos-ing tradition of the superiority of metaphor over literal meaning and its historical precedence (the names of Vico and Rousseau immediately come to mind). Lastly, metaphor, be it ontological, orientational or structural, abstracts and generalises. A concept is nothing but a com-

pletely general metaphor. We can understand Deleuze's hostility to metaphor (he would, for

instance, indignantly deny the characterisation of the concept I have just given). What he rejects is the logic of representation that underlies our usual conceptions of metaphor. This is why, instead of metaphor, he advocates metamorphosis. The contrast concerns each of the five char-acteristics. First, with metamorphosis, there is no parallelism in the links between words (taken literally or metaphorically) and their referents, because words and things are taken as being on the same ontological level, and a direct word-object linking, which short-cuts the sign pro-cess, becomes possible: words no longer represent objects, because they are themselves objects (they have material shape, they exert force, they mix with objects).18 Secondly, there is no separation through obvious falsity: in the world of Carroll's train, Richard is a lion (perhaps the lion that is sitting opposite Alice in the railway carriage). From this point of view, metaphor is seen as an attempt to deflate the violent potential of words, their ability to intervene, as objects, among objects: suddenly, like the werewolf of our favourite horror film, Richard roars in his rage, and claws his opponent to shreds. Thirdly, there is no replacement of tenor by vehicle. In metamorphosis, as in a dream, the chessboard actually turns into a railway and the train, not its image, is in the mind. Neither element of the transformation is allowed to stand for the other. Metamorphosis connects what metaphor, in spite of its revo-lutionary aura, carefully keeps apart, at a safe distance. Fourthly, since

Interlude 1 59

there is neither difference nor possible replacement, there is no hier-archy in the terms involved in a metamorphosis. What we have is not a hierarchic tree, whatever its direction, but an anarchic rhizome. The assemblage of enunciation, based on the operation of metamorphosis, is deeply disruptive of order. That is why it is so easily inscribed in a nonsense text. Lastly, metamorphosis fails to abstract and generalise, for reasons that by now are clear. And an assemblage of enunciation is made up of singularities, not abstractions.

The logic of assemblage has a psychological equivalent in the dream-

work, and a literary inscription in nonsense: I am suggesting an analysis of the dream-work that would avoid accounting for surface dream-

thoughts as representations of latent thoughts. I am suggesting that we think of dreams not in terms of enigma and meaning, according to the logic of representation, but of what Kristeva, following Plato, calls a chora, a melting pot of unordered semiotic elements.19 Indeed, in a dream all elements are treated alike, on a par, whether they be thoughts, objects or words. In a dream, everything is taken to the letter. There is no representation of the òther' in a dream: the dream, at least the literary dream of Carroll's text, brings together mimetically incompatible elem-

ents (a brook, a chessboard, a train, a joke), in the same linguistic, literary, psychological space. Working through confusion and mixture, the dream-work does not represent, it enacts it acts upon, it arranges incompatible elements together. Perhaps what I am suggesting is a non-Freudian theory of dreams such as the one evoked by Blanchot in L'Espace litteÂraire. According to him, dream is what blurs the distinction between day and night, what makes personal identity confused. It is based on similitude, not representation: `Dreams belong to a region of pure similitude. Everything in a dream is an appearance, every element is another element, is similar to it, and to yet another, even as the other element is in turn similar to another. You may look for the origin, for the model; you would like to be referred to a starting point, an initial revelation, but there is no such thing: dream is merely similitude that endlessly refers to similitude.'20

Dream and nonsense, therefore, as worldly practice and literary genre, have a natural link with the logic of assemblage. And now we are in a better position to understand the interest that draws nonsense and dream together: nonsense is the literary explication of the logic of assemblage, even as dream is its embodiment. This is why nonsense unfolds through or like a dream. The story might go like this: the logic of representation is what constitutes our reality (it conceives of langu-age as a system of signs; it conceives the functions of language as


communication, expression and reference) in short, it is the structur-ing process that constructs a liveable world around us. The result is the production of stable meaning and stable subjects to exchange it. Non-

sense, qua absence of meaning, destabilises this production of stable entities. By rejecting meaning, it opens up not so much to a multiplicity of meanings as to the phenomena themselves its function is to bridge the gap between word and world. The essence of radical nonsense lies there: in the destabilisation of our idealised represented reality by the shocking incoherence and brutal actuality of assemblages. The result is dangerous: reality as we know it is but an illusion, the hallucination of order through adequate representation; dream, with its apparent inco-herence, is the only reality. Naturally, it is difficult for nonsense texts to keep to such dizzy heights: Victorian nonsense is also the genre that makes sure that language, which temporarily opens up in nonsense, soon closes up again,.

9 Another logic

I have moved from images of railways to images and railways, indiscrim-

inately. The site where such mixture comes to exist is an assemblage of enunciation. Carroll's text is a good illustration; it is not the only one: trains can function in various assemblages, which always involve texts, symbols, patterns of behaviour, institutions and objects. The logic of the contemporary assemblage of enunciation centred on trains governs such everyday activities as catching a train to go to Cardiff, vast polit-ical-economic concerns, like Eurostar and the Channel Tunnel, risking one's life on British trains, and various myths and aesthetic productions, from the toccata in Villa-Lobos's second Bacchiana Brasileira to the extraordinary film by Lars von Trier, Europa. And the assemblage does not discriminate between them, does not establish hierarchies, as repre-sentation does causal chains go in every direction, and not only from object to sign. `J'eÂcrirai mon ideÂe sur le sol': I shall inscribe my idea on the landscape, wrote the banker Pereire about the Paris to St Germain line, the first railway in France. And he did.

The logic of assemblage, the logic of AND rather than INSTEAD OF, can be linked to the way language actually works. Paratax, for instance, as opposed to hypotax. Or tropes, like zeugma, which lies at the centre of Ryle's concept of `category mistake', to which he claims in The Con-cept of Mind to be able to reduce the constructs of Cartesian dualism21

(and in a way this interlude has been nothing but a non-Rylean reflec-tion on the rejection of Cartesian dualism, of the idea that the mind

Interlude 1 61

somehow represents the body). Language does allow us to conjoin terms belonging to different types, thus causing category mistake, as in `she came home in a flood of tears and a sedan chair'. True, here the con-junction is one of words the logic of the assemblage of enunciation is that it also conjoins words and referents (referents are the words of the language of action, and words have material existence and exert mater-

ial force, like objects). Literature (nonsense merely takes this possibility to its limit) is precisely that which makes use of this potentiality of language, which takes it literally. The true function of literature, if I may be allowed such portentous words, is to resist the logic of repre-sentation and open up to the logic of assemblage: not to replace or represent the world, but to act, within it and upon it, by mixing with it. Philosophers so far have been content with interpreting the world the task of literature is to metamorphose it.

In this way, literature will become not only non-representational but prophetic. The last text I wish to evoke is by Bruno Schulz. It is the first page of his short story Àt the Sanatorium Under the Sign of the Hour-

glass', which was adapted for the screen by Werner Haas.22 The train we are asked to board is less disturbing than Carroll's, but more sinister. It is clearly a train of thoughts, which seems `lost in meditation'. Let me suggest the subject of its prophetic musings not a meeting with the narrator's father in some strange lunatic asylum, as actually occurs in Schulz's story, but the anticipation of other, more terrible and more crowded train journeys, towards an apocalyptic fate, as part of another assemblage of enunciation, involving railway carriages normally used for goods and cattle, vast bureaucratic planning, the resources of chem-

istry and various experiments in medical science, a number of machine guns, whips and hanging nooses, and, not least, involving also a myth of exclusion at the core of which lies the representation of a whole people as having nefarious physical and cultural traits. This is what a collective assemblage of enunciation is, and that is what it can do. Schulz's story was written before the Second World War. It cannot in any sense `represent' the Nazi death camps. But it anticipates them, it mingles with them in advance, it is a literary intervention in an assem-

blage of enunciation which at the time had not materialised, even as the silent slow trains of the film Shoah intervene in the same assemblage in retrospect. Bruno Schulz himself, a Polish Jew, was never sent to Treb-linka in such a train. He was murdered by the Gestapo in 1943, in the small Polish town where he lived. My interlude is dedicated to him.

2 `Linguistics has done a lot of harm'

1 `La linguistique a fait beaucoup de mal'

In 1997, two years after Deleuze's death, Editions du Montparnasse 1published L'AbeÂceÂdaire de Gilles Deleuze, three video cassettes recording

an eight-hour interview with Deleuze by Claire Parnet (with whom he co-signed Dialogues), according to the A to Z format (À pour AmitieÂ', etc.). The cassettes were recorded around 1990, at a time when Deleuze was working on Qu'est-ce que la philosophie? He insisted they should not be made public before his death.

Viewing those cassettes is an extraordinary experience. We are pre-sented not with the distant, necessarily anonymous voice of the author of a book, but with the quasi-physical presence of a speaker. This is not a book but a series of lectures, and since Deleuze is always facing us, talking to me (Claire Parnet is a voice-off whose discreet presence can sometimes be glimpsed as a reflection in the mirror behind Deleuze), it is a series of private lectures: the great philosopher is talking to me alone.

The experience has three main aspects. The first is a voice and a face. The relevance of Guattari's insistence on the semiotics of faces (`facial-ities') and of sound (`jingles') becomes apparent. Language, even if this is an interview, is no longer the only means of expression. Our construc-tion of Deleuze's meaning no longer exclusively relies on sadly equivo-cal strings of words mourning for the inevitable disappearance of their original speaker; it can take every advantage of gestures, smiles and nods, and all manner of facial expressions. And what a voice it is, and what a face! The great philosopher, obviously a great teacher, and justly and touchingly proud of it (in the course of the interview he insists on his lifelong attachment to the art of lecturing) is also a great demagogue,

62

`Linguistics has done a lot of harm' 63

effortlessly practising all the tricks of the orator. The viewer cannot but be aware he is being seduced and, as with a great novel or film, he likes it. We understand why Deleuze's lectures at Vincennes had such a numerous and enthusiastic following: we also understand why he at-tracted the most diverse and strangest people. The paradox, of course, is that the cassettes are an instance of language extended into a compre-

hensive enterprise of seduction through communication, but that the extraordinary voice produces not only phone, but also logos: the main themes of Deleuze's philosophy are expounded here, which makes the cassettes a worthy addition to the Deleuzean corpus.

So the experience has a second aspect. Even as Lacan's Seminars are sometimes said to be Lacan in plain French, the AbeÂceÂdaire gives us the exoteric Deleuze. Deleuze is, as we all know, a difficult author a first reading will hardly exhaust the interest of DiffeÂrence et reÂpeÂtition; it took me, who is slow-witted, 30 years to begin to grasp Logique du sens. But not the AbeÂceÂdaire: the viewer is taken in hand by the most skilful of pedagogues, he immediately understands, he has even the gratifying impression that he is intelligent. Points are numbered, arguments are recapitulated, the necessary redundancy is here. Obscurer points of the doctrine are carefully bypassed: television audiences may not be deaf, but they are dumb Deleuze's elitism and contempt for the media are given free rein in the AbeÂceÂdaire but Deleuze shows that he can talk to us without talking down at us. So we have the extraordinary impression of witnessing the occurrence of an event: thought in progress, as it emerges and erupts into words and sentences. Except, of course, that we have read the rest of the corpus, and recognise the textual origin of every proposition.

The third aspect of the experience is due to the condition set by Deleuze, that the cassettes should be shown only after his death. Speak-ing from the grave gives him a total freedom of expression: the privilege not only of old age but of death allows him to be as rude as he pleases, and the word ÌmbeÂciles!' is heard more than once. Politeness or com-

promise are no longer necessary. This is the revenge of the teacher, after a lifetime of pedagogic responsibility, and consequent frustration (`Your objection is highly interesting, even crucial . . .', which, as we know, means: `You idiot, you haven't understood a word I said!'). Deleuze takes every advantage of this, to the point of being unfair: thus Eco is a know-all, Wittgenstein an assassin of philosophy, journalists are idiots, and the present cultural scene disastrous (there is more than a modi-

cum of the old duffer in such pronouncements: an old man coming to terms with approaching death) although, to be fair to Deleuze, his


enthusiasm for the subjects he loves (philosophy, teaching, art and friendship) is in no way impaired.

Among his frank and uncompromising remarks, we find a judgement passed on linguistics, which can be paraphrased thus:

In order to understand style, you must know nothing about linguis-tics. Linguistics has done a lot of harm. Why is this the case? Because there is an opposition between linguistics and literature, which do not go well together. According to linguistics, langue is a system in a state of equilibrium, which can thus be an object of science. All the rest, all the variations are set aside as belonging not to langue but to parole. But a writer knows well that a language is a system that is by nature far from equilibrium, a system in a perpetual state of imbal-

ance, so that there is no difference between a level of langue and a level of parole. Language is made up of all sorts of heterogeneous currents, in a state of multiple disequilibrium.

This is an extract from the letter S section, `S is for Style', and it is significant that in the AbeÂceÂdaire there is no `L is for Language' (L is for Literature) or `D is for Discourse' (D is for Desire). And it is obvious that the rejection of linguistics is of the most sweeping kind. Using an ambiguity peculiar to French, where `langue', before it denotes Saussure's system, means a natural language (`la langue francËaise'), Deleuze denies the dichotomy on which Saussure founded structural, and with struc-tural, scientific linguistics: the langue/parole dichotomy. This appears to be a position of deliberate regress, all the more so as Deleuze surprisingly keeps a term, `style', which, in French culture, is felt to be antiquated. So the strategic aim is to go back to the time before the founding moment of structural linguistics a vantage point from which the centrality of language, and therefore the role of linguistics as a model science, can be denied.

Another immediately interesting aspect of this outburst is that lin-guistics is not conceived independently of literature, even if they are sharply separated. Again, there is an underlying position here, which concerns their respective importance. Literature is not one of the regions providing texts from which the empirical linguist will democrat-

ically choose his examples, on a par with the language of advertising or cookery. It is not even a specific region of language, often called poetic language, the object of a specific sub-part of linguistics, called stylistics. Literature is conceived here as the other of linguistics, its opponent, that which puts it in its place (which is no place at all). This means that


literature is separate from linguistics, it is not a possible object for it, it is not under its dominance (this reverses the whole trend of structuralist literary theory, as exemplified by Barthes's first manner). It also means that, as practice, literature is on a par with linguistics, probably even superior in philosophical interest to it. (Is linguistics truly a science? Do empirical sciences in the field of language matter for philosophy? Do they provide the kind of knowledge of language that the philosopher, that all of us need? It is strongly suggested that they do not). So that in fact there is a choice to be made between literature and linguistics, and it is an exclusive choice. The two paths of linguistics and literature do not cross, and we are left in no doubt along which one we must saunter. The object of the exercise is in fact to deny the relevance of the Saussurean concept of langue and replace it with a concept that is adequate to the philosopher's needs, as far as language is concerned, the concept of style, the concept that problematises at one go the notions of language and literature. Such construction, as usual, takes the form of a correl-ation, which can be reproduced thus:

1 2 3 linguistics system homogeneity literature variation heterogeneity

4 5 6 equilibrium langue science disequilibrium style philosophy

(art)

Column 1 merely gives its name to the correlation. Column 2 questions the scientific foundation of linguistics, the core of the Saussurean revo-lution, or change of paradigm. For post-Saussure linguists, our attitude to language is scientific whenever we extract from the infinite variety of phenomena an object that is systematic. The canonical example of this is the infinite variety of sounds susceptible of phonetic description, as opposed to the limited system of phonemes that define a language phonemes that have no independent meaning but only distinctive value as part of a system independent of speaker, space or time (this is the synchronic solidarity of the system). We note, however, that Deleuze still uses the word `system': he is not, therefore celebrating the infinite variety of the phenomena, but insisting, against the solidarity and homogeneity of synchrony, on the multiplicity of variations. The object, language, can be rationally described, but as a multiplicity of heteroge-neous currents, each with its own `speed' or temporality, each following


its own line of flight. We also note that what the synchronic system of langue excludes is called `the rest', something I have elsewhere tried to account for under the concept of `the remainder'.2 This column, there-fore, stages two objections to `scientific' linguistics. The first is by now familiar: the gesture of exclusion by which the science is founded leaves out a remainder that comes back, like Banquo's ghost, to haunt the scientific linguist. The history of linguistics is that of a never entirely successful attempt to lay that ghost to rest. The second is that langue qua system ignores that which in language is in the nature of a process: it tends to exclude the `becoming' of language under the name of di-achrony (a worthy field of research, but one carefully separated from the central region of synchrony); and it tends to exclude, or relegate to the margins, the study of variation. The latter point is the more import-

ant for Deleuze: language, like any other object, is subject to historical change; but it can also be said to be in a state of constant, because constitutive, flux, so that becoming is not an effect, to be judged by its results, from the vantage point of the end of a long-term process, but the very life, the very nature of language. Such variation, then, is the centre of any study of language, hence the persistence of the term `system' in Deleuze's text: language does not vary haphazardly, but along lines of flight that can be charted, through movements of deterritorialisation and reterritorialisation that need to be carefully described.

Column 3 draws the natural consequences from this difference be-tween synchronic langue and Deleuzean language. The synchronic system is homogeneous: it is based on the combination and interplay of elements, phonemes for instance, that are of equal value and are treated alike. The system may distinguish between different levels, for instance the two levels involved in double articulation, but the relation-ships between such levels are fixed (thus, phonemes combine into units of the next level, morphemes). Not so Deleuzean language, which is made up of elements without a fixed position, which is always caught up in ongoing variations, the site of `differences of potential' (an elec-trical metaphor), caught in heterogeneous currents (another electrical metaphor) that produce temporary configurations of singular points. For Deleuze, this picture does not concern merely language, but the totality of the world, and of thought: the chaos from which everything starts is made up of potentials, from which the singularities that make up the world, or thought, or language, are generated.3

Column 4 develops this consequence. The synchronic coupe d'essence (I borrow the phrase from Althusser4) defines a state of equilibrium, the calm and stability necessary for observation etymologically, theory is


contemplation experiment and prediction. Let us freeze a drop of the biological soup on a slide and count the bacteria, or the phonemes. We need time, even in the age of computers, to register dialects, or registers. But time there is not: languages die before linguists manage to photo-graph them for posterity, registers change before their description can be published. Such delay and failure are not contingent. Because language is not homogeneous, because langue is an abstraction forced upon it, the stability of which is reached at the cost of artificiality, it is always moving beyond the grammar that seeks to freeze it into a system. This is what Deleuze and Guattari mean when they claim that `languages leak'. Only individual style, only literature can capture the change at the speed at which it occurs, because they intervene in it, because they are not so much reflections as agents of it. Only literature, not linguistics, can do justice to the disequilibrium of language.

This is why, in column 5, the opposite of langue is not parole, but style. The old dichotomy embodies an exclusion which is not a reflection of fact, but a stipulation the imposition of a theoretical decision. In its Chomskyan version, linguistics seeks to disguise this by naturalising the decision and locating it in the mind-brain. The metaphysical cost of this (all the philosophical paraphernalia of innate ideas and an antiquated philosophy of mind) is high. Deleuze of course flatly denies the oppos-ition: he does not want to celebrate literature as a form of parole (a tactic that smacks of linguistic imperialism: the dubious science of stylistics, a parent pauvre of linguistics proper, is then erected to deal with those epiphenomena), he hails it as the field where such distinction cannot make sense. Again, we note that Deleuze does not deny the possibility of a system, of thinking in terms of langue. What he denies is the onto-logical hierarchy, and separation from parole. A system there may be, but it is a strange one, which a systematic linguist would fail to recognise. There are constraints on grammar, but they are not laws, rather defeas-ible maxims; generalisations are possible, but they are only partial there are no universals of language; a system we may describe, but it is a curious one, one that is heterogeneous and constitutively imbalanced, a paradoxical system of singularities, the object of a science of the singular. Style for Deleuze is just that.

Column 6, my addition, brings the correlation to a temporary end (correlations, as we know, being rhizomatic, their columns have a ten-dency to increase and multiply). The terms I have picked out for the bottom row of the column are somewhat arbitrary (àrt' is a danger word, threatening grandiloquence, although Deleuze blissfully ignores the danger, and makes extensive use of it). It marks the abandonment of


the form of science as the mainstream knows it in the field of language. Language, or style, belongs, qua practice, to the realm of art, and qua reflexive creation of concepts, to the realm of philosophy. Only in those fields can the complexity of language, its life, be preserved.

All this naturally raises a number of questions. If we reluctantly accept, in a polite and watered-down form, that `linguistics has not done all the good that was expected of it', we still find ourselves orphans. Will literature actually do as a substitute for linguistics? We are entitled to express tremulous doubts. But Deleuze is never short of answers and blandishments. So here is another suggestion, which goes some way towards explaining why the term `style' has been chosen. Our task is not to extract a homogeneous system from language, but to produce a concept of language. And one of the ways of doing it is through genealogy. Not the old search for the origins of language, long aban-doned as undecidable, but the question of its genesis. As we shall see (in Chapter 3), Logique du sens offers a full-blown theory of this. But here is a similar passage from L'Anti-êdipe:

The sign is a position of desire; but the first signs are the territorial signs that plant their flags in bodies. And if one wants to call this inscription in naked flesh `writing,' then it must be said that speech in fact presupposes writing, and that it is this cruel system of in-scribed signs that renders man capable of language, and gives him a memory of the spoken word.5

This shows the characteristic flippancy and assertiveness of Deleuze and Guattari at their best (or their worst). It asserts that in the beginning was the sign. But not the Saussurean sign: the sign as they mean the term in ethology, the natural sign that inscribes itself on the body as symptom, or the elementary non-natural sign that marks a territory, be the signer man or beast. And, in a gesture reminiscent of certain psychoanalysts (I am thinking of Serge Leclaire6), such inscription of signs on the body of the earth or the subject's body is a form of Ur-writing, the postulated origin of the spoken word and human language. We may note that the langue/parole dichotomy has been substituted with a writing/speech op-position, and that this conception of `writing', which in this passage seems to be completely naturalised, is not so far from the eÂcriture of literature: such proximity, but also the distance that the choice of the term `style' involves, is a central theme in Deleuze's collection of late essays, Critique et clinique. In an interview collected in Pourparlers, Deleuze seeks to clarify his position about literature, which is, the interviewer


notes, èverywhere present in [his] work'. And he does it in terms of the conjunction of the critical and the clinical, and in terms of signs:

It's not just a matter of diagnosis. Signs imply ways of living, possibil-ities of existence, they're the symptoms of life gushing forth or draw-

ing away. You don't write with your ego, your memory, and your illness. In the act of writing, there's an attempt to make life more than personal, to free life from what imprisons it. . . . You write with a view to an unborn people that doesn't yet have a language. Creating isn't communicating but resisting. There's a profound link between signs, events, life, and vitalism: the power of nonorganic life that can found in a line that's drawn, a line of writing, a line of music.7

I do not offer this quotation as a solution to our present theoretical problems it raises more questions than it suggests solutions, especially in the appearance of the Bergsonian vocabulary of vitalism. But we are at the beginning of our enquiry. And what the quotation shows is that the link between language, natural signs, the body, writing and style is not contingent in Deleuze no mere philosophical gesticulation, but a cluster of concepts that are beginning to cohere.

2 What is wrong with linguistics? Developing Deleuze's critique

There are three aspects to Deleuze's critique of linguistics. In the whole of the corpus, there is a critique of the failed ambition of linguistics to be a science. In the texts he co-signed with Guattari, this critique takes the form of the repressed politics implicit in such scientistic attempt, whereas in the texts that bear his sole signature, what we find is rather a critique of the philosophy that underpins the proclaimed science. The most extended and explicit locus for the critique is Plateau no. 4 in Mille plateaux, which will be explored later. Meanwhile, I shall develop the three aspects, starting with a passage from Foucault. Foucault's archaeology, Deleuze claims, avoids the two main techniques so far used by archivists: formal-

isation and interpretation. In interpretation, the interpreter moves from the level of the manifest meaning of the utterance to another level, which he sees as secretly related to the first: ìn this way the initial inscription is doubled by a second inscription'.8 The archaeologist, on the other hand, does not interpret: the elements he isolates in the utterance9 are on the same level as what is said. He avoids both the deep structure of formalisa-

tion and the deeper meaning of hermeneutics.


The critique of linguistics qua science, therefore, starts with a critique of formalisation. At first sight, this might be taken to be a critique of the notion of metalanguage (the object of one of Lacan's most famous pronouncements). There is nothing wrong as such with metalanguage. We obviously need it there is no speaking about language without having recourse to technical terms. And what are, after all, philosoph-ical concepts if not metalinguistic terms: Aristotle's concepts in the field of language are now our grammatical common sense. It is true that the efflorescence of technical jargon in `scientific' linguistics is a matter of concern. And this may well go beyond the tendency towards jargon of individual linguists, as they stake their claims to scientificity: such proliferation may well suggest that meta-language is problematic, that language actively resists all attempts to capture and formalise it. The paradox is familiar, and need not detain us: the technical language about language is still part of language, in spite of brave attempts of the more mathematically inclined linguists to produce real formalisa-

tion for their discipline. But the thrust of Deleuze's critique lies else-where: he is not afraid of jargon, which he is apt to turn into imaginative metaphors. His object is the stratification that the recourse to metalanguage or formalisation implies. What he really is against is the geology of scientific linguistics, where a sentence is merely the sur-face trace of a deeper, formalised string of specialised symbols. The question Deleuze asks is: who is the master? Is it the linguist, who translates and interprets language by formalising it? (There is a deep complicity, even homology, between formalisation and interpretation: what formalisation yields is not the structure of language but an inter-pretation, with the added advantage of explicitness, and the usual risks of arbitrary imposition.) Or is it language, which ignores interpretation and resists it, and its servant, the stylist?

For Deleuze, linguistic formalisation has the twin defects of abstraction (or, rather, the wrong level of abstraction: it is too abstract not to betray the phenomena, and not abstract enough to reach the diagrammatic level of the abstract machines) and universalisation. In Kantian terms, linguistic judgements claim to be reflective: they claim to start from the phenomena, and abstract and generalise from them. The usual claim, explicitly made by Chomsky, for instance, is that a grammar, as theory of a natural language, aims at generating all acceptable utterances in the language and only those. Can one be closer to the actual phenomena? But in spite of such claims (to which classic structural linguistics was not faithful) linguistic judgements tend to be, on the contrary, determinative judgements, which Kant defines as judgements that start from abstract


and universal categories and subsume the phenomenal manifold under them. It would be interesting to study the politics of the asterisk and the initial question mark (which denote unacceptable and dubious utter-ances respectively) among Chomskyan linguists: one would soon realise that they are markers of power, designed to force the phenomena to comply with the universal laws the theory claims to state. So the main butt of Deleuze's critique is the architectonics of linguistics, where deeper layers formalise and interpret the surface of language, with the hierarchy of level this implies (deep is meaningful). Against this, Deleuze, in the field of language as in other fields, constantly advocates the older and apparently more frivolous practice of cartography: map-

ping treats language as a plane, of immanence and of consistency; it respects the heterogeneity and diversity of language, it does not freeze its currents of becoming into structures, it does not force its lines of flight into a hierarchy of channels. Most of all, it does not practise the exclusion on which `scientific' linguistics is founded, which at best restricts the study of language to what Bourdieu calls internal linguis-tics, and at worst leaves out a remainder where what is most alive and creative in the process of language finds refuge. As the mention of `vocalists' in Chapter 1 shows, Deleuze is fascinated with language not only in all its varieties, but in its variations. This is the exact opposite of the `scientific' linguist's attitude, whose sole concern is to clip the wings of language to keep it still and make it manageable for dissection. And this is not, on Deleuze's part, merely a neo-Romantic gesture of celebra-tion of the poetic virtues of language, this is a philosophy of language, an integral part of a metaphysical view, in which language, fascinating as it is, is merely of regional interest even if it is, as I am trying to demonstrate, Deleuze's specific problem.

As usual, the opposition can be captured in a correlation. The number of columns is almost indefinite. Here is a somewhat arbitrary selection:

1 2 3 scientific linguistics and logic formalism deep vs. surface structure philosophy cartography plane of immanence and

consistency 4 5

democracy of corpus structure literary text as manifestation of highest power of language machine

6 interpretation effects of meaning and lines of flight


Deleuze's doubts about the philosophical centrality of logic are loudly expressed in Qu'est-ce que la philosophie? Column 4 is one more expres-sion of the elitism of Deleuze's high modernist position: there is no equality among texts for him, and Proust tells us more about the work-

ings of language than the daily effusions of the television journalist: I happen to agree with this. Column 5, an old favourite, insists on the dynamic aspect of language, as well as on its materiality, as opposed to the ideality of the structure. Traces of the correlation can be found in the following passage from L'Anti-êdipe:

No one has been able to pose the problem of language except to the extent that linguists and logicians have first eliminated meaning; and the greatest force of language was only discovered once a work was viewed as a machine, producing certain effects, amenable to a certain use. Malcolm Lowry says of his work: it's anything you want it to be, so long as it works Ìt works too, believe me, as I have found out' a machinery. But on condition that meaning be nothing other than use, that it become a firm principle only if we have at our disposal immanent criteria capable of determining the legitimate uses, as op-posed to the illegitimate ones that relate use instead to a hypothetical meaning and re-establish a kind of transcendence.10

Having duly hailed the presence of the phrase, `the problem of lan-guage', we note that Deleuze and Guattari's tactic is more complex than I have made it out to be so far. For linguists and logicians are called upon as potential allies against hermeneutics, or the compulsive search for transcendent meaning. Immanence in Deleuze is to be taken in a much wider sense than in structuralist linguistics (the `principle of imma-

nence' is what makes linguistics an ìnternal' discipline). Yet Deleuze and Guattari, in their advocacy of a new pragmatics, are resolutely èxternal' in their approach: language is of the world and in the world: they seem to be making an alliance with the hard, distributionalist, American side of structuralist linguistics, whose description of linguistic structure is based on a radical exclusion of meaning an alliance, therefore, with formalism, against the softer European variety, always suspected of a return to hermeneutics, and which could be represented by the enunciation theory of Emile Benveniste. Such alliance, however, is purely tactical yet it points to a `structuralist' vein in Deleuze, most apparent in Logique du sens: is he not after all the author of that remark-

able pedagogic essay À quoi reconnaõÃt-on le structuralisme?'11 Lastly, we note with a certain ironical amusement that the reintroduction of


meaning in the second half of the passage does not only, as our correl-ation predicts, need a literary text, but that it has Wittgensteinian overtones of meaning defined through use.

`Scientific' linguistics is criticised for exclusion, the imposition of hierarchic structure, the twin sins of formalism and interpretation, the wrong level of abstraction and generalisation. Such positions, as the phrase `markers of power' shows, are hardly innocent, so the critique takes a political turn. The merest hint will suffice here, as this criti-que, characteristic of Deleuze cum Guattari, will be developed in the next sections. The line taken is not unexpected. In a classic move in polemics against structuralism, history is reintroduced into the picture. It is twice denied by the structuralist version of scientific linguistics, through the concentration on synchrony (I have already evoked the problems raised by what Althusser calls `coupe d'essence'), and on di-achrony, where history is present, but only as the history of language: Deleuze and Guattari's position on this is that there cannot be a history of language alone. So history returns with a vengeance. The ahistorical position of linguistics is historicised. It is noticed that its architectonics privileges a form of power; that its hierarchic construction makes gram-

matical markers markers of power; that its practice of exclusion (of variations if not of variants) strongly reminds us of those who practise the class struggle by denying there is such a thing as the class struggle. So linguistics is hounded out of its ivory tower. It becomes, if not straightforwardly a `bourgeois science' as in the old Marxist heresy, at least a part of the ideological structure of capitalism: the universal abstract langue it constructs is the view of language that capitalism needs, and the phrase `the imperialism of the signifier' is no exagger-ation. Interesting as it is (and it can lend itself to interesting develop-ments), this position is itself part of history: it smacks of the post-1968 political conjuncture, of what Deleuze called `le gauchisme ordinaire', `leftism-as- usual'.12

But there is also in Deleuze a much more central and much more sustained critique of linguistics: a philosophical critique. It rests upon the following theses. (1) There is a philosophy of language, implicit or explicit, behind every variety of `scientific' linguistics (the Chomskyan research programme, from this point of view, offers one considerable advantage: its philosophical underpinnings are entirely explicit). (2) The philosophies of language that support `scientific' linguistics all partake of the logic of representation. We remember, from our first inter-lude, the five characteristics of representation and the importance, for Deleuze, of the passage from metaphor to metamorphosis. Deleuze's


philosophical critique of linguistics is already present, almost explicitly so, in his critique of representation.

Let us take, for ease of exposition, the `word' as a loose equivalent for the linguistic sign, the theory of which linguistics formulates. We can easily show that the five characteristics of representation fully apply to it. The word does embody the difference between representative and represented: a word is not a thing. The linguistic sign improves on older conceptions of the sign by redoubling the separation, by going from designation (the word is separated from the thing, but directly points towards it) to denotation (the `pointing towards' takes place within the word, between signifier and signified, while the thing, the referent, is expelled into outer darkness). Deleuze's logic of assemblage is a protest against the lack of contiguity or continuity implied by such difference and separation. The word is a substitute for what it represents: so replacement is here too. As we know from the sad state of Laputa, such re-placement is a good thing: we cannot carry all our referents on our backs (although concepts at least do not weigh much). Formalism takes this replacement to extremes, when symbols not only replace but generalise. Deleuze's philosophy of language, on the other hand, is fond of unholy mixtures, where the word happily consorts with its referent and does not straightforwardly replace it. That hierarchy is present in the linguistic sign has already been made clear: the structure of immediate constitu-ents is one of syntagmatic trees. The hierarchy here does not obtain so much between representative and represented (the Member of Parlia-ment, in spite of his protestations, is of a higher order of being than the electorate in the matter of language, the absent thing is often endowed, qua absent, not only missing but missed, with mystical super-iority) as between signs themselves, which function only in so far as they are hierarchically ordered. Not so, of course, with Deleuze's rhi-zome. Lastly, abstraction (of the wrong sort) raises its ugly head. A `name' is an ontological metaphor, betraying the phenomena by freezing them in order to make them manageable; and the symbols that scientific linguistics uses to construct its Lego structure betray the process of variation that is constitutive of language. Against this, as we shall see in Chapter 3, Logique du sens offers a theory of the genesis of language, based on the articulation of divergent series and the circulation of sense: an adaptation of structuralism, but one based on a logic of circulation, not of representation. And the two volumes of Capitalisme et schizophreÂnie (L'Anti-êdipe and Mille plateaux) develop a logic of assemblage, what Deleuze calls a logic of AND rather than of INSTEAD OF. Here, the critique must be qualified, as Deleuze consistently evokes the logic of


AND against the logic of IS (he plays on the paronomasia ET/EST) which characterises the essentialism that has plagued philosophy throughout its history. So that if abstraction (of the wrong kind) remains a defect, allies can be made of difference and separation, provided they are taken as organising a system of relations, not of resemblance. Deleuze's atti-tude to structuralism is indeed ambivalent. The following, which admit-

tedly concerns Godard's films rather than the linguistic sign, is typical:

The key thing is Godard's use of AND. This is important, because all our thought's modelled, rather, on the verb `to be', IS. Philosophy's weighed down with discussions about attributive judgements (the sky is blue) and existential judgements (God is) and the possibility and impossibility of reducing one to the other . . . Even conjunctions are dealt with in terms of the verb `to be' look at syllogisms. The English and the Americans are just about the only people who've set conjunctions free, by thinking about relations.13

The AND, the conjunction that names the logic of assemblage, is the indication that the straightforward relation of representation gives way to the mixture of assemblages, and consequent stuttering of language. Language does not work according to orderly hierarchies or stable and sedate abstractions, it stutters along, on a line that is never straight, the active and creative line of flight. Neither the INSTEAD OF of representa-tion nor the IS of essence can deal with the intensive multiplicities that make up language. Those terms are still obscure: let us hope they will soon be clarified.

Meanwhile, it is no wonder that Deleuze's references to the great mainstream linguists, Saussure and Chomsky, are critical.

A passage in DiffeÂrence et reÂpeÂtition examines what Deleuze calls `the linguistic Idea' (a term which, contrary to appearances, must not be interpreted as Platonist: no transcendence is involved, the linguistic Idea is a structure).14 Deleuze gives a positive account of the importance of relations in the definition of structure of the differentiality of the elements and of their relations. But he criticises Saussure for accounting for such diversity of differenciation in the negative terms of opposition. When Saussurean discreteness and value are interpreted in terms of negativity (when the phoneme `b' is defined not only through its differ-ence with other phonemes, but in opposition to them, `b' is `not a vowel', `not ``p'' ', `not ``t'' ', etc.), the free play of language, what he calls `the throwing of the linguistic dice' (the reference is, again, literary, to MallarmeÂ's `coup de deÂ' but for Deleuze the concept that the phrase


names is characteristic of structuralism), is obscured. Against Saussure's negativity he evokes the positivity of the treatment of imperatives by the French linguist Gustave Guillaume, in terms of continuous process and the crossing of linguistic thresholds (as opposed to the opposition between discrete units).

The critique of negativity in Saussure's conception of language is taken up again in L'Anti-êdipe, where it is linked with the transcendence of the signifier.15 The system of oppositions, creating Saussurean `values', freezes the elements at a place, and makes them immune to variation. The plane of immanence thus created is subjected to the transcendence of the signifier: the signifier is the ultimate guarantee that the system functions, the signified is subordinated to it, tropes are dependent on it, and the whole of language can be accounted for through the metaphor of the game. Against such `linguistics of the signifier', Deleuze and Guattari defend a linguistics of `flows and codes' (it was announced, in nuce, in the incipit to L'Anti-êdipe), which they attribute to the Danish linguist Hjelmslev: in his work, signifier and signified are of equal importance, and there is no tran-scendence involved although we may wonder whether the charge of transcendence, coherent with Deleuze's constant adherence to a phil-osophy of immanence, concerns the linguistic signifier as defined by Saussure.

The third theme of Deleuze's critique of Saussure concerns the wrong level of abstraction in the linguist's abstract machines. In certain ways the linguistic system is too abstract, it excludes part of the phenomena, which it refuses to take into account (for instance, the region of lan-guage since dealt with by pragmatics). In others, it is not abstract enough: it fails to generalise and abstract up to the level (beyond the organon that accounts for language as organic, reintegrates it within the body from which it issues) where very abstract regimes of signs (Deleuze and Guattari call this level of abstraction the diagrammatic one) pro-duce, among other systems, such as faces and jingles, the flows and interruptions, the coding and the machines that make language work.16

Negativity, transcendence, the wrong level of abstraction. Some of these criticisms concern Chomsky. In his case, however, since he has produced not only a linguistic theory but also an explicit philosophy of language, the attack is more global, and concerns Chomsky's position as the product of an image of thought (this is why the main thrust of the critique is political, why syntactic markers are said to be markers of power). Images of thought roughly correspond in Deleuze to Foucault's concept of episteme. The image of thought that produces Chomsky is


sometimes called the State (as opposed to the nomadic this is the dialect of L'Anti-êdipe) image of thought, and sometimes the dogmatic image of thought (this is the earlier dialect of DiffeÂrence et reÂpeÂtition). In Dialogues, the State image of thought, with explicit reference to Choms-

ky's markers of power, is described in terms of universality (Platonic Ideas or human nature), question and answer (Plato's maieutics, and philosophical enquiry as dialogue), judgement (the natural form of the philosophical proposition), recognition (the unveiling of an already present essence, or truth) and just ideas (adequacy, not invention, is the cardinal virtue).17 In DiffeÂrence et reÂpeÂtition, the account of the dogmatic image of thought is much more thorough and detailed: it involves eight `postulates' (of the philosopher's good will, of common sense and the concordia facultatum, of recognition, of representation etc.),18 that act in the silence of the inexplicit and stifle creative thought under the weight of the logic of representation.

So Chomsky, far from being the main actor in a change in paradigm, of a Kuhnian scientific revolution, is just a ripple on the dominant image of thought. We understand why he did construct an explicit philosophy of language, and why such philosophy is so regressive, an almost intolerable dead weight imposed on his linguistic research pro-gramme.

One aspect of the image of thought that comes in for special criticism, because of its importance not only in Chomsky but for the whole of structuralist linguistics, is binarism. The binary structure (syntactic nodes tend to have two branches, and only two; phonetic features have a binary, plus or minus, yes or no organisation) tends to be pre-sented as natural, as a feature of reality a tendency considerably encouraged by the rise of computer science. But there is no necessity that language (or the whole of language) should be organised along binary principles. In syntax, for instance, other, non-binary models have been proposed not to mention the field of enunciation linguis-tics, or pragmatics, where binarism no longer applies. So binarism appears to be an ideological or broadly political imposition. It is linked to the conception that the main function of language is the conveyance of information (Deleuze is, as usual, brisk and somewhat brusque: ìnfor-mation is a myth, and language is not essentially informative'19). And the main feature of binary syntax, in a neat inversion of Chomsky's innatism, is not that it is part of the natural order of nature, but that it constitutes a structure easily taught in the form of commands: follow this, not that branch. Syntax, far from being a natural object, is a construction of power.


The result of the critique of binarism is a change of metaphor to account for the structure of language: not the tree, with its roots and branches, that is, with its hierarchic structure, but the rhizome, a liter-ally anarchic growth, without clearly ascribable root or origin, the de-velopment of which follows no perceptible plan, but seems to follow its own lines of flight. Language is no forest of conifers, it is a potato field. And the rhizome, of course, in spite of what I have just said, is not to be taken as a metaphor but literally: it is a concept. Like all of Deleuze's concepts, it develops along a plane of immanence, and offers a multipli-

city of characteristics. The theory of the rhizome is presented at the beginning of Mille plateaux (it was published independently as a short pamphlet, under the title Rhizome20): the rhizome is characterised by a number of `principles', of connection, heterogeneity, multiplicity, asignifying rupture (à rhizome may be broken, shattered at a given spot, but it will start up again on one of its old lines, or on new lines'21), cartography and decalcomania (a rhizome has no genetic axis or deep structure, it is a map, immune to the logic of tracing and reproduction). This theory of the rhizome is already a fully fledged, non-Saussurean and anti-Chomskyan philosophy of language, even if the rhizomatic organisation concerns a much wider field. The critique is all-embracing: tree-like structures are structures of order, but also of subjectification, the imposition on an individual of his or her place as subject and speaker. Against this, the theory of the rhizome evokes the possibility of conceiving language as an a-centred society of words (a phrase Deleuze and Guattari borrow from the semioticians Rosenstiehl and Petitot22). We understand two characteristics of the àssemblages of enunciation', which for Deleuze and Guattari are the relevant units for the study of language: they are collective, and they produce utterances in the sense of Foucault (`statements'), that is, utterances guaranteed by no transcendence, and in need of no interpretation, because they have no deep structure and yield no encrypted meaning.

3 The imperialism of the signifier

The core of Deleuze's critique of Saussure, as we have seen, concerns the signifier. But the critique is somewhat obscure and misleading, and, since it occurs mainly in the two volumes of Capitalisme et schizophreÂnie, seems to concern the signifier in Lacan (which can be and has been described as transcendent23) rather than the linguistic signifier proper. As is now notorious, the relation of Lacan's signifier to Saussure's is one of eponymy: they share a name, and little else; one claims to be the


offspring of the other, but the relevant tests have failed to establish paternity.24 So, what we are faced with is a projection, back on to an unwitting Saussure, of a concept of signifier that has only the slightest links with the linguistic signifier. There is an element of self-criticism in this, since Logique du sens offers a theory of sense, and of the genesis of language, with structuralist overtones. And, as we saw in Chapter 1, the critique of the imperialism of the signifier can be said to be Guattari's specific contribution to the common enterprise.

Let us begin with the exoteric Deleuze. In Dialogues, the signifier is described as the `dirty little secret', the source, cause and centre of the process of interpretation.25 In the Lacanian inversion, the signifier is no longer that which is to be accounted for, the ultimate element linguistic analysis yields, but that which interprets us (remember the already quoted Lacanian definition of the signifier). And the trouble with inter-pretation this kind of interpretation at least is that the answer is always given in advance: the signifier, Deleuze claims, always leads us to mum and dad, as is the case with little Hans. Hence the Deleuzean motto: always experiment, never interpret.

This will be of little relevance to linguists, as it clearly does not concern them. But the critique goes much further. What Deleuze rejects is the primacy of the signifier, the Saussurean idea that semiology, the science of signs, is an extension and a generalisation of linguistics, that the linguis-tic sign (signifier v. signified) is the model for all signs. But regimes of signs, what Guattari calls `semiotics', are innumerable, and only a theor-etical fiat can reduce them to language and its semiotics: such projection is at the heart of the scientistic programme of linguistics, and accounts for the insistent presence of facialities and jingles, in so far as they are semiotics different from linguistic semiotics, in Mille plateaux.

This is where the critique of the imperialism of the signifier reaches linguistics at last, and is no mere reflection of Deleuze and Guattari's well-known hostility to psychoanalysis. The critique goes in three steps, and we have the adumbration of a correlation: the linguistic signifier, like its Lacanian counterpart, works through transcendence, whereas language unfolds on a plane of immanence; the signifier, as concept, must be historicised, and appears to be a despotic, as opposed to a territorial, sign; and the binarism of the signifier, its bi-univocal rela-tionship to its signified (this no longer applies to the Lacanian signifier) represses the ambiguity, the polyvocal proliferation that is the real characteristic of language.

The linguistic signifier, therefore, also involves a form of transcend-ence. Deleuze and Guattari's tactic here is to take for granted the fidelity


of Lacan's signifier to Saussure and its filiation (later critics, as we have just seen, raised strong doubts). In psychoanalysis, an object, the des-potic signifier, is detached from the signifying sequence, made up of asignifying, detachable elements. This detached signifier gives its mean-

ing to the whole of the chain, in which it is lacking, makes it cohere by triangulating its elements (Deleuze and Guattari are fond of this carto-graphical metaphor). The phallus is that detached signifier that triangu-lates the chain of libidinal part objects.26 On the face of it, the case of the linguistic chain is vastly different: no single detached signifier seems to ascribe meaning, and freeze it, along the whole chain. But the organ-isation of the chain is dual: it is not only syntagmatic (what Jakobson calls the axis of combination), it is also paradigmatic (the axis of selec-tion), so that at any single point in the chain a virtual paradigm is reintroduced as missing, in absentia, an absent presence, which means that the whole of the system is a transcendent presence, through the negativity of opposition, to every actual sequence of signifiers. The role of transcendence in the system is the usual one: it triangulates every element in the chain, like the Lacanian upholstery button (or quilting point), it guarantees that meaning is achieved, that the value, in the Saussurean sense, of each variable, or word, is determined.

Hence the second aspect of the critique, that the signifier is a despotic as opposed to a territorial sign. Here, the signifier is obviously caught in the quasi-Marxist periodisation of L'Anti-êdipe. Its emergence is linked to the emergence of the state and what they call the imperial, or des-potic machine. And this process of emergence is one of abstraction, in which the very nature of the sign changes: the sign is no longer territor-ial, a concrete mark of belonging (the jingle, la ritournelle, is characteris-tic of territorial semiotics: it marks the subject's territory), it has become abstracted through deterritorialisation, no longer directly a sign, but à sign of a sign', `merely the deterritorialised sign itself', the sign not as warning or symptom but as letter. This is in fact the source of the transcendence of the signifier, in that original link with a long-vanished despot. Hence the famous pronouncement: Èven when it speaks Swiss or American, linguistics manipulates the shadow of Oriental despot-ism.'27

The third step of the critique, which is entirely coherent with the first two, concerns the `bi-univocality' of the signifier. The following passage is typical:

There is no linguistic field without biunivocal relations whether between ideographic and phonetic values, or between articulations of


different levels, monemes and phonemes that finally ensure the independence and the linearity of the deterritorialised signs. But such a field remains defined by a transcendence, even when one considers this transcendence as an absence or an empty locus, per-forming the necessary foldings, levellings (rabattements) and subordi-nations a transcendence whence issues throughout the system the inarticulate material flux in which this transcendence operates, op-poses, selects, and combines: the signifier.28

The effect of the bi-univocality of the linguistic sign and of the binar-ism of linguistic analysis is to repress the polyvocality of the phenomena. Linguistics always experiences the question of homonymy and ambigu-

ity as a problem, but not as a creative meaningful one as one it seeks to forget in immediate binary solutions. The polyvocality that other, older semiotics sought to capture is reduced, repressed, imposed upon by the semiotics of the signifier which is complicit with arrangements of power that impose meaning (signifiance) and subjectification on the speaker.29

What is wrong with the imperialism of the signifier is now clear: it imposes a subjective structure on the speaker; it interpellates her at a place at the very moment when, through her appropriation of language in her speech-acts, she believes she freely expresses herself. And this concerns the Saussurean definition of the signifier (a sign represents an object for a subject), as well as the Lacanian (a signifier represents the subject for another signifier).30 The signifier, a carrier of force, an instru-ment of power, is the main actor in the process of subjectification.

In Mille plateaux, Deleuze and Guattari sum up their critique of the signifier and it is now clear that this is a thoroughgoing critique of the concept of sign, on which the whole of linguistics, as a scientific endeavour, is based in the diagram shown in Figure 2.31 As is obvious, it involves the linguistic signifier, but other elements as well: faces, the political and religious organisation of the state, the ritual of the scape-goat etc. Language, in Deleuze and Guattari, is never separated, or separable, from the world in which it is spoken. It does not constitute such a world, it does not merely phrase or represent it, it is caught in it and intervenes in it as a cog in a machine, as a component in an assemblage of power. And since those are historical constructions, the type of semiotics in which language claims to play a central role in the form of the imperialism of the signifier is only one among several. The terms vary: in Mille plateaux, the four types of semiotics are called pre-signifying, signifying, counter-signifying and post-signifying. I am afraid there is an element of dramatic irony in this: through the use of


2

5

4

6

3

1

Figure 2 Signifying semiotics: (1) The Centre or the Signifier; the faciality of the god or despot. (2) The Temple or Palace, with priests and bureaucrats. (3) The organisation in circles and the sign referring to other signs on the same circle or on different circles. (4) The interpretive development of signifier into signified, which then reimparts signifier. (5) The expiatory animal; the blocking of the line of flight. (6) The scapegoat, or the negative sign of the line of flight.

prefixes in order to characterise the other three, the signifying semiotics regains the centrality the proliferation sought to deny it. This is yet another instance of the paradox on which this book is based: language is indeed a problem for Deleuze.

4 Another theory of the sign

Deleuze, so he claims in various interviews, discovered linguistics through Guattari (and did not like it). It is interesting, therefore, to go back to his earlier works to see his prelinguistic concept of sign at work. We do not have to go very far. He is, after all, the author of a strong reading of Proust entitled Proust et les signes.32 The concept of sign used in this book claims to be derived from Proust himself, and has nothing to do with the Saussurean sign: Deleuze deals with Proust as he deals with philosophers he extracts a problem from his works, and formu-

lates the problem in a concept. In a nutshell, the concept of sign thus produced is a concept of generalised, as opposed to strictly linguistic, signs. That there is more to signs than the strictly linguistic variety is obvious to anyone who has read Peirce, whom Deleuze relies on in his later theory of the 16 varieties of signs, expounded in his books on the cinema. Here, however, the developments, being directly inspired by Proust, are more idiosyncratic, although the attitude towards the multi-

plicity of sign-types is the same. As may be expected, Deleuze begins his reading of Proust with the

question of memory. But there is an immediate and unexpected twist: memory is all about signs. The Proustian character lives in a world of signs and is compelled to behave like an unwitting Egyptologist: the


world of Proust is a world where deciphering and interpretation are the central preoccupations, the most useful skills in order not only to make progress, but simply to survive. Some of the signs are linguistic: conver-sations do occur, and words are bandied about. But they are not the most important: to concentrate on the explicit, to fall into the illusion that words actually designate objects, is to miss the real signs, which need to be learnt, and demand a skilled interpreter. And where are those signs to be found? There are four regions of the world that teem with them: social life as envisaged by Proust, that is, not the whole of society, but only good society, practising its mondaniteÂs in the salons of the bourgeoisie or aristocracy; love (the lover, especially if he is still at the uncertain moment that precedes the declaration, is a compulsive interpreter of signs); impressions and sensible qualities (the celebrated madeleine is an archetypal sign); and art, the region of dematerialised signs (Vinteuil's `little phrase' is a good example). A number of criteria are offered, which enable us to recognise a sign: the material it borrows; the type of emission; its effect on the receiver; the relation between a sign and its sense (not a simple matter of reflection or representation: the signs of love, for instance, are often mendacious); the faculty, memory or intelligence, needed to interpret the sign; the temporal structures that apply to the sign; lastly, the sign's place in a gradation of signs at the top of which is enthroned the sign of art, in its capacity to reveal essence (I leave you to decide whether this is too conven-tional a theory of art).33 And signs are defined not only through their characteristics, but through their mode of operation, which involves complication and implication as well as explication there lies an early theory of sense, destined for more glorious developments in Logi-que du sens.

But there is also a theory of literature in the book, by no means a trivial one. Proust's vision of the world is centred on signs, and only literature is capable of interpreting them. For Proust is, as we might have guessed, neither a philosopher nor a scientist. Philosophy needs direct and explicit propositions, issuing forth from a mind in search of truth; natural science presupposes objective and unambiguous matter, and a mind aware of the factual constraints of reality. Not so the literary mind: it does not believe in facts, only in signs; nor does it believe in truth, only in the plurality of interpretation. Neither a metaphysician nor a physicist, the literary subject, author and reader alike, is, as we have seen, an Egyptologist. So literature is the site of the conscious and explicit experience of signs, and of the formulation of the principles of their interpretation: as such it is as important as philosophy or science,


because signs, the encounter with signs, are what jogs us into thinking and thinking is interpretation.

As the author of a book called Interpretation as Pragmatics, I rejoice at the good news about interpretation. This is the glorious dawn of Deleu-

zean metaphysics, before the encounter with psychoanalysis, when interpretation, far from being the archetypal form of thinking, will become the trivial disclosure of the dirty little secret, always the same, always mum and dad, a freezing of thought into preÃt-aÁ-penser, not a liberation. Meanwhile the non-centrality of language, and the non-linguistic nature of signs is inscribed, at the outset of the second part of the book, in what we may call the Proust correlation, which contrasts the literary mind (here, Proust) with the philosopher:34

1 2 3 4 Philosopher Observation Philosophy Reflection Proust Sensitivity Thought Translation

5 6 7 Explicit signification Conversation Word Implicit signs Silent interpretation Name

This correlation, of course, concerns Proust, and Deleuze's interpret-ation of him. Nevertheless, it explicitly contrasts the world of signs and what Deleuze calls `the world of logos', the philosopher's world. The terms are slightly displaced (the second part of the book is a later addition), and interpretation, which is relevant to my point, is now silent. We understand the coherence of Deleuze's treatment of the dere-liction of language in late Beckett from which this book started: as column 7 suggests, we find in Proust an adumbration of the language of names. And Deleuze's generalised signs are vastly different from linguistic signs: they are not articulated, do not work through oppos-ition, as a result of which they are endowed with meaning rather than with value, they are the vectors of epiphanic revelation, and, being the objects of constant and uncertain deciphering, they are encrypted rather than representational. Deleuze is obviously a supporter of the old theory of the sign as mystery, to be taken for wonders, rather than a vector of communication and information.

5 The postulates of linguistics

So far Deleuze's critique of linguistics has been conducted en passant, concerning details of the doctrine, in a haphazard manner. The fourth


plateau of Mille plateaux, on the other hand, contains a systematic cri-tique of what the authors call `the four postulates of linguistics'.

What can they mean by ̀ postulates' of linguistics? For a positive expos-ition of similar postulates, we shall make a detour through Jean-Claude Milner's L'Amour de la langue.35 For the Lacanian linguist langue has something to do with the Real (as opposed to the constructed reality we live by) in so far as its frontier is defined by a certain kind of impossibility (the typical mark of the Real): in order to know whether a string of words belongs to langue, is acceptable in that it abides by the rules that govern the system, what we must do, all we have to do, is pass the judgement: `this can/cannot be said (in English)'. Langue is defined negatively as that which organises the converse of the set of utterances that `cannot be said'. The linguist rather hubristically maintains that this proximity to the Real distinguishes linguistics, which has always existed (witness the Sanskrit grammar of Panini), from other social sciences, like sociology, which dwell in the realm of the imaginary and construct their object.

Here comes the postulate (Milner calls it, of course, an axiom): the Real of langue can be the subject of a calculus. Acceptance of this axiom (which, like all axioms, can be denied) distinguishes the scientific lin-guist from the amateur. From it, we can derive four maxims (they are couched in the form of commands), from which the whole of scientific linguistics can be deduced: they spell out the operation of the calculus. Maxim no. 1. Langue must be constituted as a Real, which, being Real,

is causa sui. Hence the arbitrary character of the sign: there is no neces-sity that such and such sign should designate this particular referent; in the same vein acceptable utterances are given, and we must accept them as such: we have, qua speakers, no possibility of intervention other than recognising they are what they are. Maxim no. 2. Langue must be constituted as an object of formalisa-

tion. It is an object for the form of symbolic writing that logic and mathematics practise. The concept that allows us to actualise this possi-bility is the Saussurean concept of the sign: langue is a system of dual signs (signifier/signified), acceptable utterances are the rightful combin-

ation of such signs. Maxim no. 3. Langue must reduce the speaker to what is needed to be

the bearer of the calculus: a mere sender, or emitter of signs, without any other characteristics (such as a body, an unconscious, a place in a social field or a historical conjuncture). Langue needs, and recognises, only angels. Maxim no. 4. Langue must reduce the community of speakers to what

is needed for the calculus: the exchange of messages between sender and


receiver, addresser and addressee, what we have learnt to call the `situ-ation of communication', inscribed in Jakobson's diagram of communi-

cation (see Chapter 2, note 18). From these four concepts (the arbitrary character of the sign, the sign

itself, the speaker, communication) all possible forms of linguistics can be deduced. Naturally, our linguist being also a Lacanian (and a lover of poetry: he is the author of a treatise of prosody36), he does not claim that langue is the whole of language, or that the scientific concept accounts for the production of poets: Lacanian lalangue, which is the real object of his book, accounts for parts of language that langue does not reach.37 Langue is what Galilean science abstracts from the messy phenomena that make up language, through a process of exclusion and separation that turns them into true phenomena in the Kantian sense, that is, makes them manageable for the calculus. There is a cost: in the matter of language, langue can never have the last word; but there is also the usual gain, of explicitness, clarity and the promise of cumulative knowledge.

Although, for obvious reasons, they never mention Milner (he was writing at the same time as they did), Deleuze and Guattari adopt the same approach to linguistics, by isolating four postulates with the difference, however, that such postulates are invalid.

The four postulates are formulated thus:

Postulate no. 1. `Language is informational and communicational.'38

The translation, unfortunately, is compelled to lose the conditional form that the French language can use (`le langage serait informatif . . .'), thus formulating the postulate and at the same time casting doubt on it. This postulate roughly corresponds to Milner's fourth maxim: language is reduced to the most abstract form of the situation of communication, where information is exchanged in the most neutral fashion, without reference to affect, needs or aspirations on the part of the speakers.

Postulate no. 2. `There is an abstract machine of language that does not appeal to any `èxtrinsic'' factor.'39 This corresponds to Milner's first maxim. Langue conforms to the structuralist `principle of immanence', it is causa sui, and as such any relations it has with the world are merely contingent or rather, they do not fall under the scope of linguistics, but of some other, secondary science. Linguistics inhabits an ivory tower, and langue is immune to worldly influences. The advantage of such constitutive separation is that language is able to be ìnformational and communicational'. Postulate no. 3. `There are constants or universals of language that

enable us to define it as a homogeneous system.'40 I am not sure that


this can be said to correspond to one of Milner's maxims, except indir-ectly to his second, which defines the sign as the building block of a system that must be homogeneous (that is made homogeneous by the very construction of the concept of sign). But the search for linguistic universals is as old as linguistics it can be said to inspire the Chomsk-

yan research programme, and the very definition of Galilean science makes it look for constants that will enable us to describe a system as homogeneous. Postulate no. 4. `Language can be scientifically studied only under the

conditions of a standard or major language.'41 This describes the whole of the scientific enterprise in a nutshell, and can be said to be a direct consequence of Milner's maxims nos. 3 and 4. The reduction of the speaker to an angel, and of the situation of communication to a polite conversation between angels in Heaven inevitably turns langue into a normal, what they call a ̀ standard', form of language: angels do not speak dialects, they do not have a social or regional accent or lexicon. And, Heaven being a place of unspeakable boredom, much in need of a serpent to make it interesting, angels are no poets, in spite of their lyres.

Deleuze and Guattari's position on these postulates is, not unexpect-edly, that such constructions fail to account for the real workings of language and that the parts they seek to exclude return with a ven-geance, like the Freudian repressed. Their opposition to Milner, there-fore, runs deep: they refuse to treat langue as belonging to the Real but then, it is obvious that they are not Lacanians. For them, the separation between what can be said and what cannot be said is not a given (this is how it is, there is nothing we can do about it) but an imposition of power, formulated in a `mot d'ordre': say this! Do not say that! They have no difficulty in showing that such injunctions vary in the course of history (Milner's maxims presuppose the synchronic coupe d'essence), that they are variously imposed on various sectors of the speaking population, etc. Hence the presence, in the fourth postulate, of the word `major', which is already in itself a critique of the scientific concept of langue, presupposing, as it does, the existence, with the major idiom, of minor ones, which will not be reduced to Milner's lalangue and `points of poetry'. But let us take things in order.

The critique of the first postulate is based on the reintroduction of pragmatics, naturally excluded from scientific langue, as a region of linguistic practice as worthy of study as the others, if not more. This, of course, profoundly alters our conception of language, because an utterance, from the point of view of pragmatics, is communicational


but rarely straightforwardly informative (the conveyance of meaning is indirect, and demands a calculus of presupposition or Gricean implica-

ture); and, more importantly, an utterance never merely transmits a message, but always also exerts force. What is `communicated' is not merely a message, but also an impulse (the French language allows this play on `communication', which Derrida uses in his celebrated essay `Signature eÂveÂnement contexte'42). But Deleuze and Guattari go further on this road than Austin and Searle: the utterance is not merely the locus of a speech-act (a promise, for instance), but of a social act, a mot d'ordre, a slogan. With them, pragmatics is moving from a legal model (did Diana actually say the right, legally binding words that made her Charles's wife?) to a political model. The canonical form of utterance is not declarative, an assertion, but imperative. And the origin of the slogan is not an individual, a master giving his orders, but social: the origin of the slogan is in a collective assemblage of enunciation, that mixture of bodies, instruments, institutions and utterances, which speaks the speaker. Hence the well-known assertions that all language is spoken in indirect style. And we understand why syntactic makers are markers of power.

The critique of the second postulate insists on the materiality of lan-guage. Scientific linguistics idealises language into disembodied langue. But an assemblage of enunciation is never only that: it is always also a machinic assemblage of bodies. Hence langue cannot be causa sui because utterances are immersed in a world which they express and in which they are uttered, but also for a deeper reason; they are ontologically an integral part of the assemblages that make up the world, they follow their lines of flight on the plane of immanence or of consistency that is our world. As we have already had plenty of opportunity to realise, Deleuze and Guat-

tari's ontology is based not on separation and abstraction, but on mix-

ture. This postulate, incidentally, also contains a critique of Marxism in the field of linguistics, as formulated by the pseudo-Stalin:43 whether we consider, like Stalin, that language is a neutral instrument, or that it is part of the superstructure, like Nikolai Marr, the Soviet linguist whom he criticised, we are still postulating a separation between the material base and the superstructures. Deleuze and Guattari's ontology rejects this separation, both at the level of the individual speaker (where the distinc-tion between the word and the scream is under attack) and of the collect-ive utterance, which is part of an ontologically complex machine. So there can be no independence of language.

The critique of the third postulate has a more familiar ring, as the denial of the separation between Saussurean langue and parole is a famil-


iar tactic. But it usually takes the form of an extension of langue at the expense of parole, moved by the scientistic hope that parole will ultim-

ately be resolved into langue. This is not Milner's position: for him such reduction is in principle impossible, and science can never capture all the linguistic phenomena; I have myself attempted to formulate such impossibility with the concept of `the remainder'.44 But Deleuze and Guattari go further on that road: for them the distinction dissolves altogether, as language is the site of continuous variation, of heteroge-neous currents, of the singularity of style, so that agrammaticality is neither an object of exclusion nor a remainder, but an integral part of language. Far from being laws of nature, linguistic `rules' are merely partial and temporary maxims, that is, attempts at imposing some sort of order on a language that does not care for such imposition and whose stuttering constantly subverts them. This also enables us to understand why the synchronic coupe d'essence is pointless: all we can do is map the variations, chart the heterogeneous currents of language.

The critique of the fourth postulate takes this lack of homogeneity seriously. Since speakers are no angels, they speak in an infinite variety of dialects, registers and jargons. This is where scientific langue is most blatantly the creation of a rapport de forces, the imposition of markers of power. Linguists working on the languages of former colonies have coined the twin concepts of linguistic imperialism and `glottophagy' to account for the linguistic struggle between dominant and dominated, colonised and colonising languages.45 But it is not merely a question of a clash between languages, of a language opposing another and eventu-ally liquidating it (that situation existed between English and Cornish, and still threatens in the case of English and Welsh), for the same type of relationship obtains within a language, where the major dialect (the dialect, for instance, of the WASP speaker) is àffected by continuous variations that transpose it into a ``minor'' language'.46 Deleuze and Guattari conclude: `there are not, therefore, two kinds of languages, but two possible treatments of the same language'.47 The interest of this internalisation of linguistic struggle is that it offers the speaker a modicum of hope. For if the speaker is subjected to linguistic markers of power, spoken by a collective assemblage, possibilities for expression are strictly limited: to use the same analogy as Deleuze and Guattari, she is as free when she speaks as when she votes (she is free to vote for whoever she wishes, provided the system, or the President of the United States, approves of the candidate). But if minor dialects are constantly `working' within and against the major language, it becomes possible to follow one's lines of flight, to make language stutter, to do what in


older parlance would have been called `poetically to inhabit language' except that such inhabitation, which Deleuze and Guattari call style, is not reserved for poets, even if they are better at it than the generality of mankind, even if they are the technicians of stuttering.

We are leaving the realm of critique, and moving towards a positive reconstruction of the analysis of language. This is how Deleuze formu-

lates its main features in an interview:

I don't think we, for our part, are particularly competent to pro-nounce on linguistics. But then competence is itself a rather unclear notion in linguistics. We're just addressing a number of points that we ourselves consider fundamental: first the part played in language by precepts; second, the importance of indirect discourse (and the recognition of metaphor as something that just confuses matters and has no real importance); third, a criticism of linguistic constants, and even linguistic variables, that emphasizes ranges of continuous vari-ation.48

We need not take the modest disclaimer on which the quotation begins seriously: it is merely the pretext for another attack on Chomsky. But the rest of the passage does sketch a new form of pragmatics one, however, with a strongly Continental flavour.

6 Towards another concept of language

Deleuze's critique of linguistics has at least a quality of comprehensive-

ness. It starts by demoting language from its central position in our view of the world at large, and of philosophy in particular. It goes on by demoting linguistics from its role as model, its central position in the social sciences. And it goes even further by denying it the status of a scientific discipline. Yet, unlike Badiou, Deleuze does not use this pos-ition for an overall condemnation of philosophers who indulge in their fascination for language as Sophists (for Badiou, Wittgenstein is the arch-Sophist, with Lyotard not far behind).49 For language, as we have seen, insists in Deleuze's work: it may be the object of a concept, or a series of concepts. The very critique of linguistics has already yielded a rich crop (sense, pragmatics, style, minor language, stuttering), which the next chapters will develop.

To a linguist, Deleuze's critique of linguistics appears somewhat facile: sweeping statements are not lacking. But, unfair as it sometimes is, or misguided, it is never ignorant. Deleuze (perhaps I should say Deleuze-


and-Guattari) actually reads the works of linguists. But he picks and chooses: even as he implicitly defends a literary canon, he has his linguistic favourites. His canon is often idiosyncratic, to the point of being sometimes quirky (the French linguist Henri Gobard, with his concept of `tetraglossia', is now entirely forgotten), but it also invokes linguists whose importance in the development of the discipline is not in doubt. Here are three examples, in decreasing order of importance: Deleuze makes considerable use of the arch-structuralist, Hjelmslev; he approves of linguists who practise external linguistics, for example Wil-

liam Labov; at times his views on language appear to be close to those of the French school of ènunciation linguistics', the main representative of which is Emile Benveniste.

On the face of it, Deleuze's attraction to Hjelmslev is strange. And we must avoid from the start a terminological misunderstanding. Deleuze is a philosopher of immanence, and for him language unfolds on a plane of immanence (what he perceives as the transcendence of the signifier is firmly rejected). But this does not prevent him from supporting, and to a certain extent practising, the external linguistics that situates language within the material world. Therefore, he seems not to approve of the `principle of immanence', which is central to structuralist linguistics, and notably Hjelmslev's version of it the principle that formulates the utter separation of language from anything that is not itself.50 Neverthe-

less, Deleuze consistently contrasts Hjelmslev, whose technical vocabu-lary he uses, with Saussure and he thinks he can detect in Hjelmselv's use of the term ìmmanence' not merely the separation to which he objects, but the rejection of all transcendence. It is Hjelmslev, in fact, who gives him his concepts of planes and strata; who insists on the relation of reciprocal presupposition between the two planes of expres-sion and content (which radically blocks any form of transcendence); who complicates the duality of the Saussurean sign by introducing a four-term relationship, where the form v. substance contrast is crossed with the expression v. content contrast (so that there is a form of con-tent, the morphemic structure, that informs the substance of content, the continuous flow of thought); who enables him to treat language as a machinic organisation of flow and interruptions by introducing the concept of matter, as that which is not even yet the substance that is destined to be formed a matter which is close to what Deleuze means by chaos; lastly, who provides Deleuze with the right level of abstrac-tion, the diagrammatic level, as his concept of langue is wide enough to cover not only natural languages, which are merely a subtype, but also a wide variety of entities possessing a similar structure the structure that


a form imposes on a substance it extracts from matter (the word Deleuze and Guattari use for this, `coding', has a strong Hjelmslevian ring). A passage in L'Anti-êdipe makes the contrast explicit, in the form of the following correlation:51

1 2 3 Saussure Linguistics of signifier Transcendence of the signifier Hjelmslev Linguistics of flows Algebraic plane of immanence

4 5 Signifier dominates signified Duality of sign Reciprocal presupposition Two crossed contrasts

The last column appears again in Mille plateaux, where Hjelmslev is hailed as a Spinozist linguist, because he breaks with the old signifier/ signified opposition and makes it more complex by introducing the concept of form (as opposed to substance) and of expression (as opposed to content): the double articulation of linguists is redoubled and allows the possibility of a more abstract conception of the system and of other, asignifying semiotics.52

The second linguist Deleuze reads and quotes is William Labov, the embodiment of external linguistics because what he practises is usually called `sociolinguistics'. Even as Hjelmslev was the clearer side of Saus-sure, Labov is the acceptable side of Chomsky. What Deleuze finds in him is a way out of the fixity of Chomskyan competence (remember that Chomsky goes to unacceptable extremes in order to naturalise it by making it a characteristic of the species and inscribing it in the brain): this way out goes through the concept of `continuous variation' or ìnherent variation'. Such concept enables Deleuze to conceive of a system, but one that is in a state of imbalance. Deleuze is not simply hostile to science rather to the positive version of science to be found in Chomsky. He denies natural competence, but welcomes the descrip-tion of a system of variations. And, indeed, Labov does not describe Black English merely in terms of free (phonetic) variables, to be culled by a linguist who is a mere collector: he wants to describe the system of what are not merely variables, but variations in the almost musical sense of variations on a theme. And he is not describing the mixture or clash of two different systems either (standard, or major English v. the minor dialect): he is describing a system in a state of flow, characterised not by contrasts or universals but by variation.53

The case of enunciation linguistics is more complex. This is a French theory, which has had little influence outside French-speaking coun-


tries. It displaces the centre of the linguist's attention from the utterance as result, the string of elements the structural linguist segments and classifies, or seeks to generate, to utterance as process (the contrast in French is between eÂnonciation and eÂnonceÂ): its attention, therefore, is focused on the speaker as utterer, and on the linguistic markers (shifters, deictics, markers of tense and modality) through which she makes her presence felt and inscribes the utterance-as-process within the utter-ance-as-result. Anglo-Saxon equivalents to this theory have been sought in speech-act theory and linguistic pragmatics, or, more convincingly, in semantic grammars such as Langacker's.54 The best-known French ex-ponents of enunciation linguistics are Gustave Guillaume, Emile Benve-niste and Antoine Culioli.55 Deleuze only mentions the first two, Culioli, who is not the least important of the three, having reached print much later.

Deleuze makes a rather conventional use of Benveniste's theory of pronouns (we have seen that, for Benveniste, third-person pronouns are non-personal pronouns: they refer to the absent, or silent member in the situation of dialogue that enunciation linguistics seeks to chart). Thus, in Mille plateaux, Benveniste is evoked when Deleuze and Guattari account for the semiotics of the plane of consistency in terms of linguis-tic markers: infinitive, proper names, articles and pronouns. This occurs in a passage where they seek to define the concept of haeccity (hecceÂiteÂ), a mode of individuation that is neither that of a subject or of a thing (thus, they say, a season, a date or a haiku are good examples of haec-cities). The plane of consistency is made up of rhizomatic haeccities whose lines are criss-crossing. The plane of consistency, in so far as it is the bearer of haeccities, has its own semiotics, with markers that indi-viduate the singular points where lines cross, without freezing them into things or subjects: infinitive verbs, proper names or indefinite articles are far from indeterminate, but they do not interpellate subjects (we remain in the anonymous position of the on): Ìndefinite article � proper name � infinitive verb constitutes the basic chain of expression, correlative to the least formalised contents, from the standpoint of a semiotic that has freed itself from both formal significances and per-sonal subjectifications.'56 One of the interests of Deleuze is indeed that he gives us a philosophy of the impersonal, the pre-individual and the non-subjective, a philosophy where the concept of subject plays a minor part, if it plays any part at all. But we must confess that his recourse to enunciation linguistics is puzzling, as Benveniste in particular places the subject at the very centre of attention: his is no longer a linguistics of langue, but of discourse, and discourse is the site of the àppropriation' of


language by the speaking subject. Deleuze and Guattari's use of Benve-niste in the course of their attempt to define a non- subjective site for discourse, the haeccity that is neither man nor beast, certainly goes against the grain of the text. Another passage in Mille plateaux combines a reference (a rare occurrence) to Althusser and his concept of interpel-lation with a reference to Benveniste's theory of pronouns, with the conclusion that there is no subject in language, but only collective assemblages of enunciation (we now know the source of the last and apparently barbaric term in the phrase). Subjectification (`subjectivation') is merely a regime of signs among others, which refers to a certain assemblage, rather than a general condition of the workings of lan-

57guage: Benveniste could never have agreed to such plundering and misprision of his pet concepts.

So references to enunciation linguistics (mainly to Benveniste: the few references to Gustave Guillaume are less controversial Guillaume is protected by his sometimes extravagant metalinguistic jargon) are of the order of creative misprision rather than respectful borrowing: the con-cepts of non- person (Benveniste) or of line (Guillaume) are deÂtourneÂs, taken through pastures new, in unexpected directions, in a word led astray. But Deleuze is past master at such conceptual piracy.

I shall take as an example of such reading of linguistics, where the critical combines with the constructive, the creative with the outra-geous, the concept of a Markov chain. A Markov chain is the product of an ideal computing machine known as a finite state automaton. The machine goes through a number of states, producing a symbol with each state: it is a syntagmatic machine that goes from position to position generating the linear, surface structure of the sentence by combining the two axes of choice (the vertical paradigm) and placement (the horizontal syntagma). If we add the measure of the probability of appearance of each symbol, we have a model for a grammar, which is a model of performance:58 it tells us how the speaker produces the sen-tence and organises its surface structure. An example will make this clear. When I utter a sentence, the choice of the first word is entirely free (each `state' of the machine generates a word), the second is con-strained by the first, and so on, until I come to the last word, which is maximally constrained, both syntactically and semantically. When I have produced the string of words `Pride comes before a . . .', there is little doubt how the sentence will end.

At an early stage in his career, Chomsky produced a critique of this type of grammar, which was taken as devastating. The proposition he advanced is simple: Ènglish is not a finite state language.'59 The gist of


the demonstration is that certain characteristics of the syntax of English are incompatible with a finite state grammar, as a result of which a vast number of English sentences are not Markov chains; for instance, all the sentences that have recourse to syntactic embedding, where the linear ordering of the grammatical sequence is provisionally interrupted to welcome another, embedded sequence. Such an elementary sentence as `The man who said that S2 is arriving today' (where S2 is any sentence you wish) cannot be generated by a finite state grammar, which cannot accommodate dependencies that cross over S2.

Deleuze makes a number of allusions to Markov chains, and they are always positive. The following passage is typical. Deleuze and Guattari are not speaking about language, but about DNA and genes:

It was possible to insist on a common characteristic of human cul-tures and of living species, as `Markov chains': aleatory phenomena that are partially dependent. In the genetic code as in the social code, what is termed a signifying chain is more a jargon than a language, composed of nonsignifying elements that have a meaning or an effect of signification only in the large aggregates that they constitute through a linked drawing of elements, a partial dependence, and a superposition of relays.60

A (rare) footnote refers the reader, in the matter of Markov chains, to Raymond Ruyer, a philosopher of biology, and his book, La GeneÁse des formes vivantes. As we can see, our authors' concern is entirely different from Chomsky's. They are concerned with a far wider range of semiotics, `language' being here either an example or a metaphor. And it is easy to see what attracts them to Markov chains: the string of symbols develops along a line (there is no hidden structure, no transcendent guarantee of meaning, only a surface organisation), and it offers a good example of a `line of flight', since the linear development is partly aleatory and partly dependent. To use my example again, if I feel a strong compulsion to finish the Markov chain `Pride comes before a . . .' with the word `fall', nothing actually prevents me from engaging in a creative exploitation of the proverb by replacing the expected noun with another one: `Pride comes before a vote of impeachment' so I was talking about Nixon all the time. And the word `jargon' in the quotation is to be taken posi-tively: the difference between what Deleuze and Guattari call jargon and a language is that the jargon produces open-ended and polyvocal chains, whereas the chains of `language' are characterised by closure and semantic fixity (here, obviously, `language' means something akin


to `logical language').61 Parodying Lacan (who is credited with having discovered the vast domain of a `code of the unconscious'), we might say that for Deleuze and Guattari, the unconscious is not `structured like a language', but `machined like a jargon'. In Foucault, Deleuze again refers to Markov chains in order to characterise diagrams (the right level of abstraction) and the relation between thought and chance. Here is the relevant passage:

The relations between forces, as Foucault understands them, concern not only men but the elements, the letters of the alphabet, which group either at random or according to certain laws of attraction and frequency dictated by a particular language. Chance works only in the first case; while the second case perhaps operates under condi-tions that are partially determined by the first, as in a Markov chain, where we have a succession of partial relinkings. This is the outside: the line that continues to link up random events is a mixture of chance and dependency.62

This passage expresses the three characteristics of Markov chains very clearly: they are linear (where Chomsky's deeper structures are embed-

ded and hierarchic), they are aleatory (whereas Chomsky's are necessary: they are produced by a form of creativity that is rule-governed and Deleuze, naturally, is interested in rule-breaking creativity) and they are partly dependent (Deleuze, as we have seen, does not advocate a form of ànything goes' in language: he wishes to describe a system, but a system of variations, characterised by partial dependency, maxims that are used as guidelines and meant to be flouted, rather than rules that look very much like laws of nature). So, agrammaticality will no longer be a failure of the system but a creative development, a tropic exploitation of the grammatical maxims. e.e. cummings's `he danced his did' is far more interesting, both to the lover of poetry and to the student of language, than `the man hit the ball' (in Mille plateaux, Deleuze and Guattari even go as far as saying that `the evaluation of degrees of grammaticality is a political matter'63).

Can we try to pursue Deleuze's Markovian intuition in order to sug-gest a provisional theory of the construction of meaning? I think we can, but only at the cost of introducing unwelcome semantic closure. My `Pride comes before a vote of impeachment' example, where mean-

ing was constructed along a Markov chain, step by step, is still depend-ent on the semantic closure of the last word (`so I was talking about Nixon after all'), in a move that Lacan neatly captures with the phrase


ùpholstery button', or `quilting point' (point de capiton), the moment when the independently floating chains of signifiers and signifieds are stitched together to produce meaning. For Deleuze, such closure would be a blocking of a semantic line of flight, characteristic of the signifying semiotic illustrated with his diagram. But this, of course, may be the price to pay in order to achieve meaning although we shall find a perhaps more satisfactory theory of meaning, or rather sense, in Logique du sens.

But Deleuze's interest in Markov chains and linear finite state pro-cesses does show that language may not actually work as Chomsky would have it. In particular, the celebration of Markov chains as partly aleatory, partly dependent processes subverts the positivist or naturalis-tic concept of rule (of grammar) Chomsky defends. And the sad, or rather the exhilarating, truth is that language does not care for Choms-

ky's, or anybody's, rules of grammar that syntactic `rules' are defeasible and meant to be flouted and exploited.64 Take the example of the `rule' that governs the appearance of a reflexive pronoun: a reflexive pronoun must (note the modal auxiliary) have an antecedent in the same clause. The necessity operates at the level of deep structure, which will account for a number of apparent counterexamples in surface sentences. A cer-tain amount of latitude in crossing over clause boundaries will even be allowed, so that, for instance, a sentence of the type `this book was written by Deleuze and myself' (Guattari, of course, is speaking), will be taken as an argument for the necessity of an underlying superordin-ate clause marking the speech-act (Ì state that this article was written by Deleuze and myself': the reflexive dependency holds over the boundary of a that-clause). But no amount of latitude in the system will account for the following extract from a detective novel, where the stationmaster is trying to placate the irate wealthy passenger who claims he has been insulted by the porter: Ìf you've got a bad heart, I should calm yourself, sir.'65 This `yourself' does, but cannot, refer to the `you' in the hypothet-ical clause (this is not the kind of boundary the rules allow you to cross, the if-clause being hierarchically dominated by the main clause). And the sentence, at least this is what I am told by my English-speaking colleagues, is profoundly normal: it is not even a joke, or a trope: the sentence offers an example of what grammars of English call the `semi-

emphatic' use of the reflexive pronoun (here, it works as a kind of honorifics: the stationmaster conforms to a principle of politeness and respects a social, not a grammatical, hierarchy). So, language does not care in the least about the rules of universal grammar, which it joyfully breaks to let complex meanings proliferate along the line of flight of a


Markov chain. A Chomskyan grammar predicts the sentence is simply agrammatical (which it is not). If we treat it as a Markov chain, the result is more interesting: the sentence turns a hierarchic construction of deep and surface levels (the main clause is structurally prior to the if-clause, which it dominates) into a linear construction, along a Markov chain, in which `yourself' coming after `you' in the linear sequence, can refer to it. In this sentence, the tree has become a rhizome.

The centre of Deleuze's hostility to linguistics, the reason why he thinks it has done a lot of harm, is now clear. It can best be expressed as a rejection of Milner's central tenet in his philosophical reconstruc-tion of the science of language: that there is a Real of langue, and that this Real is the object of a calculus. This is precisely what Deleuze denies, in the following theses that underlie his critique and criticism:

1 Language is not the model for all types of semiotics, which are taken to derive from linguistics, but one aspect of a much wider range of semiotic phenomena.

2 As a result, linguistics has no privilege among other disciplines; we may even go further and deny that the object it constructs, langue, is the right object. What Deleuze advocates is a return to `language'. Not, to be sure, a return to simple description of an empirical kind (he is not hostile, as we saw, to the notion of a system), but to an object which is ontologically heterogeneous and, although endowed with a form of immanence, not separated from the material world in which it intervenes.

3 Language, therefore, is one aspect of the behaviour of bodies, from which it cannot be idealised through the wrong kind of abstraction; the same goes with institutions. Such ontological mixture bodies, institutions, utterances is characteristic of collective assemblages of enunciation. For Deleuze (and Guattari), they are the relevant objects for the study of language, which can never be the study of language alone.

4 But there is a right kind of abstraction, and language is a subset of a set of semiotics captured under the concept of abstract machines, and all taking part in the most general process of all, the process of drawing planes over chaos in order to impose a form of order upon it. And since this process is what constitutes the world, language is irredeemably caught up in the world, of which it is an integral part.

3 Events, Sense and the Genesis of Language

1 Introduction

Logique du sens 1 is central to my concern, as it is the only book by Deleuze which can be claimed to be àbout language'. Do we not find in it a chapter (Series no. 26) entitled Òf language'? Does not its chute, its last `series of paradoxes' offer us a fully fledged theory of the genesis of language? Can we not say that, although its ostensible theme is a theory of `sense' (a notion to be carefully distinguished from linguistic meaning), the text is a reading of the works of Lewis Carroll, a literary author whose nonsense is steeped in language, and can even be said to consist mainly of intuitions about the workings of language?2

Yet we are immediately faced with a problem. In spite of its late appearance in English translation, the text is relatively ancient. It was published in 1969, only one year after DiffeÂrence et reÂpeÂtition, and belongs to the moment of structuralism (indeed, its Ur-text, as we shall see, is an essay entitled À quoi reconnaõÃt-on le structuralisme?'). Hence the in-sistent murmur, among Deleuze scholars, that it was a text the master himself later disapproved of as `too structuralist'. To be sure, it seems to have left few traces in the remainder of the corpus, and is hardly ever mentioned in the interviews. Its main concept, `sense', does not seem to have been productive, contrary to the concept of èvent', to which it is closely associated, and which is still present in the later books. Concepts which became central to Deleuze are, quite understandably, absent (I am thinking of the concept of assemblage), or when present, either as a mere seed (the concept of body without organs) or under a mask (there is a link between the concept of surface and the concept of plane, which became crucial to Deleuze and Guattari's philosophical geography: there are also significant differences).

99


One concept that plays an important part in Logique du sens, the concept of phantasm, later vanished without trace. No wonder: its origin is to be found in psychoanalysis it is borrowed from Melanie Klein, who contrasts conscious fantasies (as in day-dreaming) with un-conscious phantasies that structure the subject's psyche. So Logique du sens, a book in which Deleuze still draws on psychoanalysis and its concepts, is a pre-Guattari book lacking (whether this is an advantage or a disadvantage is an open question) what Guattari brought to Deleuze: a critical interest in linguistics and a definite hostility to psy-choanalysis. As a result of this, Deleuze's `book about language' is com-

pletely untainted by the science of language, and offers a theory of the genesis of language with definite psychoanalytic flavour: Deleuze him-

self calls the book a `psychoanalytic novel'. So the text can be said to be `too psychoanalytic' as well as `too

structuralist'. But there are certain advantages to such conjunction, not least the fact that language is central to the book, not only in a regional theory of sense which is broadly speaking structural, but in a philosophy of language, much influenced by psychoanalysis, in its Kleinian or Lacanian versions, and in which Deleuze's final philosophy of language, as sketched in my first chapter, finds an early formulation in two respects: (a) Logique du sens gives us a theory of the grounding of language in the body (an attempt distinct from, but parallel to, the Kleinian exercise); (b) a theory of language in which the poetic practice of language is central, since it is reached through the reading of a literary text. We shall see in conclusion that this early philosophy of language is also, on certain points, incompatible with the later one.

2 Preface

Deleuze's proposed image for the philosopher is that of the artisan. Not only is the philosopher in possession of a specific techne, essential to the well-being of the community, a techne the practice of which demands the use of specialised tools, but he makes his own tools: a system of concepts is a box of tools. Deleuze takes this conception seriously: each new book demands a new set of tools, or concepts; they are needed to extract the problem that is the object of the book. In the case of Logique du sens, the problem Deleuze extracts from the works of Lewis Carroll is double: (a) it is the problem of sense, that is, of the emergence of meaning out of a form of nonsense, of the deep-seated complicity between sense and nonsense (in other words, he takes the usual inter-pretations of Lewis Carroll aÁ rebours: nonsense is not the dissolution, but

Events, Sense and the Genesis of Language 101

the condition of possibility of meaning); (b) this complicity is demon-

strated through the association between sense and event: the event is not so much the brutal emergence of the utterly new as the circulation of sense. The problem extracted from Carroll, therefore, bears the name of the two concepts created to extract it: sense and event. The term `create' is used advisedly: not only because, for Deleuze, the philoso-pher's task is to create, to invent new concepts, but because the two concepts, although formulated in words that belong to ordinary lan-guage and have been the object of detailed philosophical elaboration, are in fact new. The meanings Deleuze ascribes to the terms, even if he presents them as part of a tradition (his concept of event goes back to the Stoics), are idiosyncratic. Thus, there is even a hint of reflexivity in the Deleuzean problem of the event: the event is a problem in Logique du sens, but problems in general, as conceptualised by Deleuze, are events. What he is dealing with in this book, therefore, is the problem of the problem. Hence the importance of the text, in spite of its author's later apparent forgetfulness.

The best way to take an overview of the text is to read the first two paragraphs of the Preface:

The work of Lewis Carroll has everything required to please the modern reader: children's books or, rather, books for little girls; splendidly bizarre and esoteric words; grids; codes and decodings; drawings and photographs; a profound psychoanalytic content; and an exemplary logical and linguistic formalism. Over and above the immediate pleasure, though, there is something else, a play of sense and nonsense, a chaos-cosmos. But since the marriage of language and the unconscious has already been consummated and celebrated in so many ways, it is necessary to examine the precise nature of this union in Carroll's work: what else is this marriage connected with, and what is it that, thanks to him, this marriage celebrates?

We present here a series of paradoxes which form the theory of sense. It is easy to explain why this theory is inseparable from para-doxes: sense is a nonexisting entity, and, in fact, maintains very special relations with nonsense. The privileged place assigned to Lewis Carroll is due to his having provided the first great account, the first great mise en sceÁne of the paradoxes of sense sometimes collecting, sometimes renewing, sometimes inventing, and some-

times preparing them. The privileged place assigned to the Stoics is due to their having been the initiators of a new image of the philoso-pher which broke away from the pre-Socratics, Socratic philosophy,


and Platonism. This new image is already closely linked to the para-doxical constitution of the theory of sense. Thus to each series there correspond figures which are not only historical but topological and logical as well. As on a pure surface, certain points of one figure in a series refer to the points of another figure: an entire galaxy of prob-lems with their corresponding dice-throws, stories, and places, a complex place; a `convoluted story'. This book is an attempt to develop a logical and psychological novel.3

A few words of commentary are in order. The first thing to note is that the text does not avoid the mention of

psychoanalysis. Carroll's works have à profound psychoanalytic con-tent', they belong to the numerous category of texts where `the marriage of language and the unconscious' is celebrated. It is all the more surpris-ing, therefore, that the translator chooses to translate the last words of our passage, ùn essai de roman logique et psychanalytique', as àn attempt to develop a logical and psychological novel'. (The only explanation I can find is that The Logic of Sense is a late translation, which appeared a full seven years after the translation of L'Anti-êdipe: it is difficult to present the notorious contemnor of psychoanalysis as the author of a psychoanalytical novel).4

The second thing to note is the presence in the text, even if it is reduced to a mere hint, of the philosophy of first beginnings, or the beginnings of philosophy. The interest we take in Carroll goes a long way beyond mere childish games with language: it concerns the serious play of sense and nonsense, a `chaos-cosmos'. So Carroll's work deals with the creation of cosmos out of chaos, it is situated in that murky original site, a form of Platonist chora, where being (whose name is Chaos) is formed into a world later this will be called, using Joyce's phrase, `Chaosmos' (the title of Guattari's last book). The task of the work of art, which explains its crucial importance for philosophy, is to recapitulate the process of constitution of the world, by drawing planes (of immanence, of consistency) over chaos. The surfaces on which events and sense circulate in Logique du sens are not yet a plane, properly speaking: but the fact that literature is a form of thought is already stressed; and we understand why `nonsense' is not merely the subver-sion of order: it is, on the contrary, a process of imposition of order over chaos.

It is natural, therefore this is the third thing to note that the theory of sense elaborated here should be a theory of nonsense: that sense should be determined through `series of paradoxes' (this is the strange


name Deleuze gives to his chapters). The term `series' has two interest-ing philosophical connotations. First, if the passage from chaos to cosmos consists in the institution of some form of coherence or order, the series is the elementary form of that coherence, in that it is sequen-tially ordered and allows a synthesis of the manifold (a theory of the three types of synthesis is duly provided in Chapter 24 in Logique du

5sens ). Secondly, the concept `series' is conceived in structural, even structuralist terms. The archetypal series is the linguistic chain or se-quence, with its three characteristics: (a) there are always two series, for instance of signifiers and signifieds; (b) one series, for instance the series of signifiers, is always in excess of the other: the two series are in a state of hierarchical imbalance; (c) the convergence of the two naturally divergent series, the passage from one to the other, is operated by a paradoxical element that circulates among them and clinches them together, like a zipper the èmpty square' that, itself devoid of value, gives their value to the elements of both series in its circulation.6 In this `paradoxical element', the word `paradox' takes on a technical meaning, but it has wider philosophical implications: it operates a distribution, if not of truth and opinion, at least of creative thought and common sense: paradox is the element in which sense qua nonsense emerges from common sense.

The fourth thing to note is that the text proposes Deleuze's version of the history of Western philosophy. In a philosophical gesture that is by no means original, it offers its own version of the overcoming of Platon-ism by playing the Stoics against Plato. That a philosopher of imma-

nence (even if the concept is not present in Logique du sens) should be anti- Platonist is to be expected: the philosophy of sense is a philosophy of surface effects and quasi-causes, not of transcendent causes. For Deleuze, the Stoics exemplify a new ìmage of the philosopher', a term in which we may choose to read an allusion to the concept of ìmage of thought' developed in DiffeÂrence et reÂpeÂtition. The subversive tradition that links the Stoics, Lewis Carroll and Antonin Artaud is the dark face of the dominant, or `dogmatic' image of thought associated with the name of Plato.

If the Preface (this is the fifth point to note) announces the movement away from the Platonist image of thought, we might expect to exchange a system of vertical metaphors (the opposition between appearance and reality described in terms of surface and height: the Platonist Ideas are up there, in their own heaven), for a system of horizontal metaphors, where there are only surfaces (the plane of immanence). But those surfaces are not yet, or not only, a plane drawn over chaos. The reference


to psychoanalysis still involves a contrast between surfaces (where lan-guage dwells) and depths (the dark realm of instinctual drives). This contrast will be abandoned later, with psychoanalysis: with Logique du sens we are at the stage of reversal, where height is demoted and depth is still present; the stage of cancellation, where everything is inscribed on a plane, comes later. So surfaces are the site where series are inscribed: a series is a sequence of singularities inscribed on a surface. A singularity an important concept in Deleuze, a philosopher of singularity as op-posed to the universal as he is a philosopher of immanence is a point on the surface; its main characteristic is not so much that it is unique (first meaning of `singular') as that it is peculiar (second meaning of `singular'): a singular point stands out, it organises around itself a `neighbourhood', in the mathematical as much as in the geographic sense. Thus, the figure of a triangle is defined by three singular points. We understand why the `figures' mentioned in our text are not only historical but `topological as well'. And we also note the presence, in the same sentence, of the MallarmeÂan phrase `dice-throws': it names the Deleuzean version of Democritus' clinamen, the element of chance that extracts forms out of chaos, through aggregation of atoms. It marks the necessity of arbitrary decision that jogs us into thought, that coalesces a number of thoughts into a problem. The link between event (the `coup de deÂs', a form of fiat, can be called an event), problem (the highest task of thought is the construction of problems and the creation of the con-cepts that formulate them) and sense, which is the object of the book, is sketched here.

So our text (this is the last point to note) ends on the characterisation of the whole book as àn attempt at a logical and psychoanalytic novel'. The first of those two terms is somewhat strange. Deleuze's distrust of logic (most forcefully expressed in Qu'est-ce que la philosophie?) is notori-ous, as is his hostility to analytic philosophy in so far as it is a logicist, as much as a logical, form of philosophy. The earlier collocation `topo-logical and logical' provides an answer: the theory of the constitution of sense will be synchronic, topological and logical, induced by the imma-

nent coherence of the structure; and it will also be a diachronic theory of the emergence of sense and of the genesis of language, through a psychoanalytic novel that may well be a romance (the Freudian `family romance' is called `roman familial' in French), that is a myth of the emergence of language, but one pregnant with displaced truth. Because I am interested in literary theory, and because this book of philosophy is ostensibly devoted to the reading of literary texts, I shall also suggest that the most interesting form of novel is a philosophical argument, and


that the best philosophical arguments are those embodied in a tale, albeit a tangled one that, in a formulation that owes little to logic, philosophy is the natural metalanguage of literature, and literature of philosophy.

3 Structuralist Deleuze

Logique du sens is a structuralist book. The concept of machine, central to L'Anti-êdipe is absent, the concept of structure pervades. Chapter 8 (Òf Structure') develops the sketchy indications of the Preface and spells out the conditions of a structure, in terms that by now are clear:

(1) There must be at least two heterogeneous series, one of which shall be determined as `signifying' and the other as `signified' (a single series never suffices to form a structure). (2) Each of these series is constituted by terms which exist only through the relations they maintain with one another. To these relations, or rather to the values of these relations, correspond very particular events, that is, singular-ities which are assignable within the structure. . . . A structure includes two distributions of singular points corresponding to the base series. And for this reason it is imprecise to oppose structure and event: the structure includes a register of ideal events, that is, an entire history internal to it . . . . (3) The two heterogeneous series converge toward a paradoxical element, which is their `differentiator'. This is the principle of the emission of singularities. This element belongs to no series; or rather, it belongs to both series at once and never ceases to circulate throughout them.7

This is a strictly structuralist approach to the notion of structure. The canonical example, as explicitly mentioned in Deleuze's text, is the system of phonemes that structures the linguistic plane of expression. Phonemes have no value in themselves, only in the contexts that oppose a phoneme to all the other phonemes in the structure; this value, therefore, is purely differential and needs a `differentiator' to emerge. The differentiator is what turns the virtuality of the system into the actuality of expression by circulating along the series of phonemes, occupying in turn every singular point in the series, that is, every phoneme. Because the differentiator is an empty square not a phoneme itself, but that which enables the phonemes to function as phonemes, it is a purely relational entity, a pure site for value. Thus, the phoneme /b/ is not intrinsically what its name suggests: it is the


conjunction, obtained when its place in the system is occupied by the empty square, of `not-/p/', `not-/g/', `not a vowel', etc.

This structuralist account also seeks to avoid a classic criticism levelled against structuralism: that it ignores or actively denies history. In what is an equally classic move, Deleuze reinscribes history within the struc-ture, as a series of events that are associated with the singularities of the structure. The interest of this for us is that it provides the first approach to the Deleuzean concept of event: the event is what occurs when the virtuality of a structure becomes actual (as when the system of langue is actualised into occurrences of parole). As a consequence an event is constitutively dual, has two faces, both of which are real (Deleuze, following Bergson, is careful not to confuse the possible v. real with the virtual v. actual contrast):8 an actualised face in the appearance of singularities within the structure, the singularities that will bear the actual meaning, producing values that are expressed; and an ideal (but nevertheless real) face that is contained in the structure qua virtuality.

There is still a certain slippage in this account. Thus, it is based on the canonical example of the phonic system. But it is not clear that in the case of phonemes two series are involved. Certainly the second series, if there is one, is not a `signified' series in the sense of the opposition of signifier and signified in the Saussurean sign: at best the second series is constituted of the physical sounds, among which the singular phonemes organise a series of `neighbourhoods' by selecting those sounds that will count as actualisations of a single phoneme. Nor is it yet clear whether the two-faced events described here belong to the actualisation of the phonemic structure into parole, into concrete utter-ances, or whether their passage from virtual-ideal to actual occurs within the structure itself, as the text seems to suggest.

In the hope of clearing those difficulties, let us turn to the Ur-text of Deleuze's structuralism, À quoi reconnaõÃt-on le structuralisme?'9 The text appeared in an eight-volume history of philosophy, written for students Deleuze also contributed the chapter on Hume to another volume.10 It was written in 1967, before Deleuze had published DiffeÂr-

ence et reÂpeÂtition, and Logique du sens, and it is a living testimony to Deleuze's talent as a pedagogue, being wonderfully clear and method-

ical. The starting point, not unexpectedly, is that you find structuralism

where you find a structure; and you find a structure only if language is involved: structural descriptions concern, not only primarily, but exclu-sively, languages. Since the proposition runs counter to our most elem-

entary intuitions, the concept of language must be extended to cover


instances of other languages than our verbal language: so there are non-verbal languages (symptoms such as body-language, the social language of kinship that the anthropologist studies etc.). I take this as the first formulation of Deleuze's ambivalence about language. Here, he defends what he later denied the structuralist position that articulated human language is the model for all `languages' but at the same time he insists on the idea that there is more to language than double articulation and the signifier v. signified contrast.

How do we recognise structuralism, then? Well, there are seven cri-teria: having defined the genus (`language'), we come to the species and its specific differences. Here is the list. The first is symbolicity. Struc-turalism avoids the customary opposition between real and imaginary by introducing a third region, the symbolic, on the site of which it situates itself. The real is unique in its truth (there is only one truth); as soon as we see double (images, fantasies etc.) we are in the imaginary; let us add one, and reach the number three: we are in the symbolic, with its triangle of signifier, signified and referent. This triadic structure is the sign of the presence of a structure, defined by a certain slippage, or circulation, of sense. The second criterion is positional: elements of the structure are not endowed with intrinsic meaning, only with relational value. This is why a system of phonemes is the preferred example of a structure, even if the presence of a second series is not clear: for such meaningless elements will acquire meaning only through combination with other, equally meaningless elements, so that the second, signified, series appears when single positions combine to form a signifying chain.

The third criterion is differential and singular: the differentiality of value (as in what linguists call a `minimal pair': `Did you say ``pig'' or ``fig''?') produces singular points, with their neighbourhoods: thus, the sound made by a lisping speaker will be attracted to the central /s/ sound that actualises the phoneme and will count as an `s', that is, be inter-preted as one. This criterion gives us the three characteristics of a struc-ture: (a) it is composed of symbolic elements; (b) they enter into differential relationships; (c) those relationships define singular points, or singularities. The fourth criterion, the presence of a differenciation, draws the natural consequences involved in the third: the structure itself is virtual, a network of differential (note the `t') relations. It is actualised in actual configurations (parole, attested systems of kinship, concrete symptoms etc.) through a process of differenciation (note the `c'). The difference, noted by Deleuze `differen t/c iation' expresses the contrast between the virtual and the actual, and also illustrates


the archetypal symbolic relation of a minimal pair (the smallest contrast in expression that produces contrast in content).

The fifth and sixth criteria are by now familiar: they concern the presence of two series (criterion no.5) and of a paradoxical element, or empty square (criterion no.6). The two series define a field of problems, and their terms are correlated through a relation subject to constant variation and displacement (the famous slippage between the chain of signifiers and the chain of signifieds, needing the episodic presence of quilting points). And the empty square zips the two series together, makes them converge where their constant variation encourages them to diverge. Deleuze gives it a Kantian name, the òbject � x', that which organises difference, being at the same time one too few (it is `missing in its place') and one too many (it is also, always, `supernumerary'). The examples given are esoteric words in Carroll, and, more clearly, the concept of zero phoneme. It is one of the distinctive features of struc-turalism that the absence of an element often counts as an element, the zero-element: thus, the English language makes use of a zero-article, on the same level as the definite and indefinite varieties. The last criterion concerns the subject: it is entirely relevant that it should come last, as structuralism demotes the classical and the phenomenological subjects from their positions of centrality. The structure knows only places, filled in its virtual state by symbolic elements, like Greimas's actants. In the process of actualisation, these places will capture their actors, the indi-viduals that will be turned into subjects. So subjects are marginal, being effects of the structure: we understand why Lacan and Althusser are as structuralist as LeÂvi-Strauss or Hjelmslev. This treatment of the subject is important for us, as it announces the insistence, in Logique du sens, on pre-individual and impersonal phenomena (singularities are never sub-jective, they are pre-individual and impersonal, but they may eventually capture subjects), and the later concept of haeccity.

As we can see, the main themes of Logique du sens, as exemplified in the Preface, are already present in the Ur-text, with the massive excep-tion of the two main concepts, èvent' and `sense'. Let us take them in order.

4 Two types of event

This is an event: our hero, on the road to a neighbouring town, meets God. His life, and ours, are irremediably changed by the encounter. But this also is an event: our hero is madly running; he no longer quite knows whether he is in flight or rushing into the fray; he knows that


around him a great battle is being waged, but he cannot experience or grasp the battle itself, only skirmishes, individual actions, fragments of experience for the battle, hovering like a mist over the field, is every-where, but nowhere to be seen, or grasped.

St Paul and the anonymous hero of Stephen Crane's The Red Badge of Courage are both involved in events, but of very different types. The event on the road to Damascus is a violent irruption, or intervention, in an established world. This irruption, which has the force of a volcanic eruption, is also an interruption: it cancels the time of the current situation and marks a new foundation of time; the only name that aptly accounts for it is `revolution'. Not so the event of the battle: a haze or mist rather than a volcanic eruption, it hovers out of grasp over the field, inscrutable and indifferent (or at least neutral) to the activities of the actors involved; yet it is what retroactively gives them sense (in other words, the event circulates, like the paradoxical element the event is a circulation of sense); and it is both involved in chronological time (the event has a date) and outside time, in a kind of eternity (the battle is as animated, yet as immobile, as the scenes sculpted on Keats's Grecian Urn).

My aim in this section is to contrast the two types of event, and the two concepts of event they have inspired in current theory: I shall call them the Badiou event and the Deleuze event for short. My primary interest is in the Deleuze event, but for reasons of contrast, I shall first make a brief detour through the Badiou event.

Badiou's theory of the event is expounded in his magnum opus, L'Etre et 11l'eÂveÂnement, and it informs the numerous books that followed, notably

L'Ethique.12 The concept of event comes as part of a system of concepts, of which it is the apex (there is a strong sense of philosophical architec-tonics in Badiou: he is a builder of systems, an occupation now rare among philosophers). It starts with a situation (what I just called àn established world' or à current situation'), with a language and an encyclopaedia (the set of beliefs and elements of knowledge that make such a world liveable). The situation is what we find ourselves in: it is a manifold of manifolds (the situation in which the French Revolution occurs is made up of an infinity of manifolds: classes, economic rela-tions, the court, a bunch of philosophers, etc. not forgetting a Bastille). The situation is punctured by an event, for instance the fall of the Bastille. The event is situated in the situation, it has a site in it, but it is not of it, it does not belong, it is supplementary to it. It comes and goes in a flash (it has no proper duration: its temporality is the retroac-tive temporality of after-the-event; and yet, as we have seen, it interrupts


and it founds), but it leaves traces, traces that allow an encounter with elements of the situation, who undergo a process of conviction, or conversion. Those elements become the militants (or witnesses) of the event, which has initiated a process of truth. The event is the site of the production of a truth (the article is important: Badiou claims that the concept of Truth is empty, but that events in particular fields pro-duce truths in the plural); yet it is not an illumination: a series of inquests (a term obviously borrowed from Chairman Mao) is necessary to ascer-tain the reality of the event, its truth-producing potential. Once con-firmed, the truth places the militants in a state of faithfulness (fideÂliteÂ), which makes them militants of this event. For the militant who has encountered the event, the situation is radically, and definitively, changed: she, in her faithfulness, becomes a subject not an individual subject, not a psychological subject, but the bearer of a process of truth. Thus, the subject of the truthful amorous encounter is not the individ-ual lover but the Two of them. The subject is that which makes the human animal escape her animality, that which gives her her intim-

ation of immortality. According to Badiou, there are four fields in which events occur and

truths are produced: science, art, politics and love. Or again, there are revolutions in politics, but also in science, and there are breakthroughs in art, even as there is such a thing as love at first sight (Badiou is the first and only philosopher of the coup de foudre, a good name for an

event). Of course, my very exposition has suggested there is a fifth, unnamed field involved, the field of religion, as the system of Badiou's concepts can easily be translated into Christian terms: for `faithfulness', read `faith' etc., and the effect of the event, as indeed was the case on the road to Damascus, is one of conversion (there is a snag in this: Badiou is a professed atheist, and his theology, if so it may be called, has no place for God, redemption and the meaning of life a striking characteristic of his concept of event is indeed that the event, being unnameable in the language of the situation, fails to make sense).

The Badiou event, as we have seen, has a specific form of temporality: it occurs in a flash, and interrupts the time of the situation; but it also founds another time, the time of the inquest, of the process of truth and faithfulness. This is where it can go wrong, and where Evil comes in (the theory of Evil can be found in L'Ethique). There are three ways in which Evil can ruin truth and/or annihilate the event. The first is the erection of an element of the old situation into a simulacrum of event: there are false events as there are false prophets, and the National Socialist `revo-lution', unlike the October revolution, was not the event it sought to


pass for. The second is betrayal: the militant, whose life is radically altered by the event, does not always have the courage and perseverance that such a sea change requires. He can renege on truth, abandon the event and actively deny it. History is full of such turnings of coats. Lastly, the militant may forget the exceptional, but also the meaningless nature of the event: he may attempt to make it make sense, to force its singularity on to the whole of the situation, to oblige elements of the situation to cohere with the odd-element-out. This is the Evil of terror, the historical archetype of which for Badiou is Stalinist terror, when a genuine event, Lenin's revolution, is forcibly imposed on the old Rus-sian empire we are only too aware of the consequences.

This is a very strong theory of the event, and it can be put to work, it can help us to understand literary texts. I have attempted elsewhere13 to show that it can inform a convincing reading of Mary Shelley's Franken-stein, in terms of èventuality' (what is the idea that comes to Victor, if not a world-shattering event?) and betrayal (what is the strange repul-sion towards his creature that seizes Victor at the very moment of his triumph, when the inert body of the monster comes to life, if not a violent form of betrayal?). And what is the monster, in his extreme benevolence that turns to the evil of terror, if not the incarnation, the trace of a Badiou event?

Lacan is one of the few masters Badiou acknowledges, and his theory of the event has a somewhat Lacanian flavour. As a point of comparison, here is Z Ïek on the event: Ïiz

The Event is the impossible Real of the structure, of its synchronous symbolic order, the engendering violent gesture which brings about the legal Order that renders this very gesture retroactively ìllegal', relegating it to the spectral repressed status of something that can never be fully acknowledged-symbolised-confessed. In short, the syn-chronous structural Order is a kind of defence-formation against its grounding Event which can be discerned only in the guise of a mythical spectral narrative.14

It is clear that this account of the event is close to Badiou's, and follows the straight and narrow Lacanian path: the event as presented here is the irruption of the Lacanian Real, that which it is impossible to symbolise, into reality, here called `structure'. There is the double effect that it shatters the old structure and cancels it, and that it founds a new structure, or new legal order. And we must note the paradoxical nature of this event. What it engenders (diachronically, we might think) is a


synchronous structure (here, access to the event is always retroactive, in the mode of the always-already, àfter-the-event') no wonder it should be a `violent gesture'. And the event founds a new legal order, but does not belong to it, it is itself illegal: there is no coexistence or contiguity between the event and either the order it destroys or the order it creates. Hence its mode of subsistence can only be that of the trace, whether it be memory, or the return of the repressed: the event is accessible to us only in the form of a ghost. We understand why, being a ghost, that is, belonging to the Real, it is outside the symbolic, and cannot enter language: it cannot be àcknowledged', or phrased, `symbolised', or represented, `confessed', or celebrated.

This strong concept of event is, however, open to criticism on two counts. In spite of Badiou's professed atheism, there is a risk of tran-scendence. There is no teleology involved, since the event is arbitrary, not part of God's eternal design, but the event still comes from outside, or above, the world it shatters, and it creates a retrospective teleology the teleology implicit in truth and faithfulness. This is a criticism Deleuze levels at Badiou in Qu'est-ce que la philosophie?: the Badiou event reintroduces transcendence in the form of the void in the situ-ation that is the site of its occurrence, and in the role ascribed to philoso-phy (a traditional conception of the `higher philosophy') which `seems to float in an empty transcendence, as the unconditioned concept that finds the totality of its generic conditions in the functions (science, poetry, politics and love)'.15

The second criticism concerns the use of the concept for the interpret-ation of literary texts. The Badiou event may be welcomed (but not represented) in the text, except that such `welcoming' will inevitably be a form of representation: Frankenstein is a staging of the irruption of the event, a transformation of this irruption into a powerful myth, that is at best an indirect attempt to acknowledge, symbolise and confess it. For the event, which is inexpressible in the language of the situation, can only be expressed in it, as there is no other language available. This is a new form of the paradox, from which this book started, of the language-hater who waxes eloquent. Even the most linguistically disruptive of avant-garde poets, even the lettriste poet who does not express himself in natural language, will only be able to welcome the event in the text, within language. The concept cannot conceive the event of the text, the text itself as event.

This is, of course, unfair to Badiou and Z Ïek, as this paradox is one of Ïiz

the hallmarks of high modernism, witness this passage from Blanchot's L'Espace litteÂraire:


The event of art is the mark of the blinding emergence of a unique event, which the comprehending mind will then attempt to grasp, as its point of departure, but which it will only grasp in so far as it escapes it: the event is incomprehensible, because it occurs in that region we can only name under the veil of negativity.16

On the whole, the concept of event used here is very similar to the Lacanian event in Badiou or Z Ïek: we have a blinding emergence, the Ïiz

retroactive grasping, through a kind of negative theology, of something of the order of the Real something which simply is and escapes comprehension. And the function of art is to express this inexpressible event: we understand why, in Blanchot, the language of the poem always strives towards silence as its limit. But there is also a significant difference: here the event is the work of art itself. It is not something the text represents, or even welcomes: art is the event, as the event cannot be captured in language and can only exist in it.

Badiou, who is a novelist and playwright himself, and a strong inter-preter of literary texts,17 would have no quarrel with this (since events do occur in the field of art although not every true poem is an artistic revolution). We will, however, find a more immediately convincing version of this in the Deleuzean concept of an event, which, as we have seen, is vastly different. To make the difference obvious, I only need to give two other examples of an event, as offered by Deleuze himself: a smile without a cat (an allusion to the vanishing of the Cheshire Cat in Alice's Adventures in Wonderland) and my death. The second example, in its grim inevitability, might also be taken as an example of a Badiou event (which means that the two concepts are not poles apart: I am insisting on their difference, perhaps exaggerating it) as the moment when the Real will eventually shatter my painfully constructed reality. But the first example cannot be taken as a Badiou event, and takes us on an entirely different path.

The concept of event in Deleuze goes back to the Stoics. In a passage of Dialogues,18 he goes back to his debt to the Stoics' pan-somatism: every-thing, according to them, is a body or a mixture of bodies (for instance the mixture of the knife and the flesh that it cuts); the world is the site of a gigantic corps-aÁ-corps. But the actions and passions of such mixtures have effects that are incorporeal: a mist hovering over the surface of bodies and states of affairs, neutral and motionless, which language can capture only in the guise of an infinitive verb (a verb that stands by itself and needs neither tense nor subject). Such is the mist of events. For the Stoics draw their philosophical line not between body and soul, or


sensibilia and the intelligible, but between physical depth (as the knife penetrates the flesh) and metaphysical surface (where the event of the wounding resides). The philosophical line of separation, therefore, runs between events and things. Hence the paradoxical nature of the event: it is the result of the mixture of bodies, it is produced by such mixture, but it is of a different order, outside the bodies, outside their actions and passions, outside the time in which they unfold. Where is the battle? Everywhere and nowhere. It is not in the actions and passions of com-

batants, yet it envelops them like a mist, brushes against them like a caress, always present and always elusive, ungraspable, both physically and intellectually. You might say that this is the position of the soldier, lost in the middle of the fray, not of the general, who, from the vantage point of a hill, directs the operations. But this picture is, precisely, a picture: the representation of the battle, after the event, from the im-

possible bird's-eye view of the painter, who reintroduces transcendence in order to dominate his subject. The event itself knows no such tran-scendence, it occurs on a plane of immanence, on a surface: it is not above, but on the interface, between words and things, between past and present, between the material and the ideal. Duality is its name, and paradox: an event is attributed to states of affairs, and expressed in prop-ositions. The event of the battle is attributed to a confused state of affairs on the battlefield, and it is also expressed in an infinitive verb: to-do-battle. Why an infinitive? Because it is a verb, marking a process (there is a certain obvious dynamic quality in the event); because it does not require a subject (as opposed to an indicative verbal form) the event has no subject, it is impersonal, Deleuze speaks of the ìt' (ìl') of the event; and because it has no tense, because it is both timeless (the battle, an absolutely contingent occurrence acquires, once it has occurred, an eternal retroactive necessity: so the event is not so much timeless as omnitemporal) and what distributes past and present (if death is an event, it has no time itself, the moment of death having no duration, but it effects the passage between `destined to die' and `dead', two states of affairs that are different from the eventual state of affairs itself, yet dependent on it).

At a later stage of Deleuze's career, the event, a concept with strong survival instinct, is defined thus:

This is what we call the Event, or the part that eludes its own actual-isation in everything that happens. The event is not the state of affairs. It is actualised in a state of affairs, in a body, in an experience, but it has a shadowy and secret part that is continually subtracted


from or added to its actualisation: in contrast with the state of affairs, it neither begins nor ends but has gained and kept the infinite movement to which it gives consistency.19

Some of this is familiar enough: the event is actualised in a state of affairs, yet it is different from it; and it enjoys the paradoxical temporal-

ity I have just outlined. There is, however, a significant difference: our event is now deemed worthy of a definite article and capitalised initial not an event (a chance occurrence) but `the Event', the part of ideality, or rather virtuality, that eludes actualisation in everything that happens. This is probably where the most striking difference with the Badiou event lies: Badiou events, being revolutions that shatter the established order of things, are, naturally, few and far between. The Deleuze event, on the contrary, is everywhere, it is present in everything that occurs: my death (not an indifferent occurrence), a battle (not a slight occur-rence, even if it is not the battle of Waterloo), but also the Cheshire Cat's smile, or the headache that composing this page has induced in me.

One thing, though, that the event has kept is its duality: in everything that happens, the accident (the mixture of bodies), which is actual, must be distinguished from the other of the occurrence, the event, which remains virtual. And we understand why the word is capitalised: virtua-lities form a realm of their own, with its own reality all events belong to one large metaphysical Event, another name for Fate. The concept of virtuality, borrowed from Bergson, is also one of the survivors in Deleu-

ze's metaphysics, hence its importance. In his Bergson book,20 Deleuze develops the opposition between the two contrasts, real v. possible and actual v. virtual. Not only are these contrasts distinct, so that the virtual is no less real than the actual, but they are opposed, in that the relation between possible and real is one of resemblance, between virtual and actual one of difference: the real is the image of the possible, whereas the actual does not resemble the virtuality it embodies. Actualisation is difference, is creation (hence the fact that the encounter on the road to Damascus, or death, are also events for Deleuze). Here is a new version of the battle event, in Qu'est-ce que la philosophie?

From everything that a subject may live, from its own body, from other bodies and objects distinct from it, and from the state of affairs or physico-mathematical field that determines them, the event re-leases a vapour that does not resemble them and that takes the battlefield, the battle, and the round as components or variations of a pure event in which there remains only an allusion to what concerns


our states. The event is actualised or effectuated whenever it is inserted, willy-nilly, into a state of affairs; but it is counter-effectuated whenever it is abstracted from states of affairs so as to isolate its

21concept.

Counter-effectuation, or amor fati (the term is probably coined on the Freudian term of anticathexis, contre-investissement in French) is what enables the subject interpellated by the impersonal event to enjoy a minimal amount of freedom: this is the moment when the actor, lost in the mist of the event, catches a glimpse of the sun rising over Austerlitz, and makes decisions (thus the cowardly flight of the hero of The Red Badge of Courage turns into heroic bravery).

So the Deleuze event, as opposed to the Lacan-type event, has three striking characteristics: it is incorporeal (inseparable from the states of affairs it envelops, yet ontologically distinct from them), impersonal (the term ìnterpellation', just mentioned, might give the impression that the event is an epiphany: God is calling me; but the Deleuze event does not care, even if actors are affected by it) and infinitive (subjectless and omnitemporal). Thus, Deleuze claims, the event of personal death has nothing personal about it: it is the one thing that utterly eludes me as a person at the very moment when it destroys me, it is in a sense the one moment in my life in which I cannot be involved; it drastically alters the state of my body, but it is not a body or a mixture of bodies itself; and it cannot be expressed in the present tense of the proposition that describes an actual state of affairs: it takes Poe's Mr Valdemar, in a fantastic tale, to utter the utterly unutterable sentence, Ì am dead.'

In Logique du sens (remember we are reading a `psychoanalytic novel'), those characteristics of the event are shared by another concept, which thus becomes a synonym of èvent': the concept of phantasy.22 Chapter 30 is devoted to an exposition of the concept, and of its three character-istics. (1) A phantasy is an event: it does not represent the actions and passions of the body but results from them, like a pure event. (2) It causes the ego to dissolve or vanish: phantasies produce à-cosmic', pre-individual and impersonal singularities, a neutral form of energy, the energy that circulates among Judge Schreber's propositions, in Freud's celebrated analysis, and the grammatical transformations between them. (3) It is expressed by the neutral infinitive verb that marks the pure event here Deleuze uses an early essay by Irigaray, in her Lacanian period, entitled `Du fantasme et du verbe',23 in which she analyses fantasies in terms of infinitive verbs that mark pure relations, before the distinction between subject and object is achieved. There is a sense


of insistence here: the Deleuze event, as it acquires more and more characteristics, is getting more and more involved with language, but a language devoid of subjectivity, the impersonal language of the `fourth person singular', a phrase borrowed, as we have seen, from Ferlin-ghetti:24 in Deleuze's non-subjective philosophy, the concept of subject is replaced by the concepts of singularity and, later, of haeccity.

But how do we go from event to language? The correlation of the Deleuze event, as summarised in the following diagram, will tell us which path to follow:

Attributed to states of affairs Expressed in proposition EVENT

PHANTASY SENSE

The event is both phantasy and sense. In so far as it is an attribute of a state of affairs, it has characteristics not only of a Freudian or Kleinian phantasy, but of a ghost, a phantasm (here the choice of the English translator is right), hovering over the mixtures of bodies. In so far as it is what is expressed in propositions, it becomes inseparable from the words that express it, from the thoughts the proposition is made of its name under that aspect is sense.

Let us, therefore, move into language. A detour through the paradox-ical duality of the event's temporality will make this easier. There are two times involved here, as we have seen: the time of the accident, which Deleuze calls Chronos, the time of chronology, in which the accident is dated, the time of the ever-present present sliding along time's arrow, the corporeal time of the mixture of bodies; and the time of the event, which he calls Aion, the time of the infinitive, of a present constantly divided into past and future, of a present, therefore, which is outside time's arrow, inscrutable and neutral. On the one hand the time of emergence, of irruption, of revolution; on the other the motionless time of surfaces, where no accident ever happens. But what has all this, (mildly) interesting as it may be, to do with language? The answer Deleuze gives is that it is the temporality of the event, which requires language and makes it possible. The only access we have to the pure event, to the battle, is through the proposition that expresses it, and that allows it to be attributed to state of affairs: our theory of the event rests on a theory of sense.

One last thing before we come to it. We now understand the differ-ence between the two concepts of event, Badiou's and Deleuze's, as far as literature is concerned. The Badiou event may be welcomed and/or represented in literature; but the text itself is no event, only the medium


through which the emergence of the event is testified to whereas the Deleuze event gives us access to, and makes us welcome, the pure event, but only in so far as the literary text is itself an event, or rather an inseparable part of the event it does not represent, but counter-effectu-ates. The general on his hillock may believe he is directing the battle: it escapes him as much as it does the infantry butchered on the field below. Only the novelist Tolstoy, Stendhal or Steven Crane can grasp the event of the battle by counter-effectuating it in a text: only in the eternal and motionless surface of the text, in its arrested time, can the battle at last be approached. The only site for the Deleuze event is a Grecian Urn; not even the actual urn, but the ode in which the counter-effectuation of the event, the passage back from the actuality of the accident to the eternal virtuality of the pure event, is achieved. We understand why Deleuze is a high modernist: the poem is the true site of the event, because the event does not make sense, but is sense.

5 A theory of sense

At the end of Chapter 2, I presented a provisional theory of meaning in terms of Markov chains and points de capiton. Such a theory has obvious limitations. A Markov chain, or the sequential construction of meaning, cannot achieve meaning in that it fails to clinch the two series of signifiers and signifieds, either because it cannot cope with secondary branchings (this is Chomsky's critique: Markov chains cannot accom-

modate the grammatical phenomenon of embedding), or because meaning, whatever the teleology implicit in the chain, can always escape in the end. Hence the necessity of the quilting point that finally clinches the series together and ensures meaning: without a regular quilting of such points the two naturally divergent series would fail to converge towards meaning. The quilting point as used here belongs to the early Lacan, the Lacan of the Ecrits. Later developments have given a wider and more portentous sense to the concept, as quilting no longer concerns sentences, but the subject himself (psychosis is due to the absence of quilting points, or signifiers that are `knot symptoms' Z Ïek's phrase in the subject's psyche), but such difference between Ïiz

literal and textual meaning is still insisted upon by Lacanian critics.25 So the point de capiton is now always was a master-signifier, the point where meaning is achieved, where the line of desire intercepts, and retroactively coheres the chain of signifiers. It is easy to conceive a Deleuzean critique of this: the quilting point reintroduces another level, a form of transcendence in the immanent chain, or plane, of


language (meaning that is `textual' in the literalness of the chain). And he would probably add that this is yet another instance of the imperial-

ism of the signifier, since the master-signifier is in a position of tran-scendence, as shown by Z Ïek on the French pragmatic linguist, Oswald Ïiz

Ducrot:

Oswald Ducrot developed the thesis that in our language all predi-cates are ultimately just reified argumentative procedures in the last resort, we use language not to designate some reality, some content, but to dupe the other, to win an argument, to seduce and threaten, to conceal our true desire . . . In ordinary language, the truth is never fully established; there are always pros and cons; for each argument there are counter-arguments; there is ànother side' to every point; every statement can be negated; undecidability is all-encompassing this eternal vacillation is interrupted early by the intervention of some quilting-point (Master-Signifier).26

Thus duly capitalised, the quilting point, or the signifier, has now little to do with language, having been called to operate on a higher plane. Deleuze's linguistic geography is more down-to-the-surface: his concern is how to make the two divergent series converge how to define and describe the paradoxical element that makes them converge. That element, which resides neither in words nor in things, states of affairs or propositions, and at the same time makes them inseparable, is now given a new name: the sense that circulates on the frontier between words and things, the event in so far as it concerns language, as it enables language to develop and makes it necessary. It is to be distin-guished from linguistic meaning: this is easier to achieve in English than in French, where `sens' has to do for both acceptations (but this ambiva-

lence is, of course, a philosophical asset: it marks the vaporous ungrasp-able nature of Deleuzean sense). The ambiguity in French is disambiguated by contrasting sens with bon sens and sens commun, `good sense' and common sense.

Let us go back, as usual, to the beginning. A commonsensical ap-proach will inform me that, in order to think and to express my thoughts, I need three things. I need an Ego (a subject, a speaker: cogito), I need a world (as object of perception and discourse) and I need God (as a guarantee that I do not fall victim to Descartes's malin geÂnie, and that my propositions are correctly formulated, my sentences are well-

formed). With those three elements I can produce propositions, distinct from things and states of affairs, and bearing on them. Thus is meaning


generated: in sentences that inscribe propositions. The structure of the proposition is binary in appearance (subject predicate), but in reality ternary (subject God predicate): God is in the hyphen that links subject and predicate, or in the copula (Ì am a man').

We can deduce from this a theory of meaning in the proposition, a theory that will account for the commonsensical fact that what I say (sometimes) `makes sense'. Deleuze gives it a form that can be summar-

ised in the following correlation (between types of meaning and the parts of speech that inscribe them in the sentence):

1 2 3 I manifestation denotation signification II Ì' (shifters) `this' (deictics) `therefore' (markers of cohesion

and syntactic structure)

`Manifestation' marks the presence of the subject in the proposition, of the speaker in her sentence: in order to `make sense', the proposition must be ascribed to an ego, who assumes responsibility for it as linguists sometimes phrase it, a proposition must be asserted. But an assertion cannot be empty, it has contents, it points towards an object: this is the moment of `denotation', or indication. And it must be well formed, it must be able to take part in Ducrot's arguments, it must be coherent, and a willing element in a coherent chain of argument. This is the moment of `signification'. And we understand why, when the prop-osition becomes a sentence, the three moments are incarnated in shifters, deictics and markers of cohesion (from individual words, like conjunctions and prepositions, `syncategorematic' elements in Aristo-tle's description of language, to the whole sentence in so far as it has syntactic structure).27

The coherence of the whole correlation is provided by the introduc-tion, late in Logique du sens, of a third line, the elements that construct meaning and are expressed by it:

1 2 3 III EGO WORLD GOD

This is the relevant passage:

If the self is the principle of manifestation, in relation to the propos-ition, the world is the principle of denotation, and God the principle of signification. But sense expressed as an event is of an entirely


different nature: it emanates from nonsense as from the always dis-placed paradoxical instance and from the eternally decentered ex-centric center.28

It is clear that my three concepts define a theory of meaning, not of sense, and that this meaning is the meaning of doxa, as opposed to paradoxical sense (as exemplified in this `decentered ex-centric center'). Such meaning is fixed, or received, or established, or resulting meaning, belonging to the realm of common sense. Hence a new line in the correlation, where Deleuze makes a distinction between bon sens and sens commun. This fourth line concerns the two variants of doxa:

1 2 3 \ /

IV common sense good sense

Manifestation and denotation inscribe common sense in the propos-ition, as signification does good sense. `Bon sens', good sense, puns on the French word `sens' in its meanings of both `meaning' and `direction'. Good sense, therefore, is the right direction (which is also a one-way street) it is the direction of right thinking (do not stray to the left!), and also of the inevitability, from left to right, of time's arrow. And, since Deleuze never resists a little wild conceptual jumping, it is responsible for the historical practice of enclosures (this is delightfully wild, but not entirely irrelevant: there is always a geography underlying Deleuze's concepts good sense signposts and stakes a territory). Common sense, on the other hand, is not a direction but a function, or an organ. Here Deleuze draws on the venerable history of the concept of common sense, from Aristotle onwards: common sense is the sixth sense that ensures the unity of the other five in my global perception of the world (we understand why it unites Ego and the World, and why good sense is God's gift to humankind). The function of common sense is to provide identity, identification and recognition: the unity of per-ception and various states of consciousness in the individual subject, or person; and the coherence of the objects of my perception into a world. The two types of doxa are complementary: they form the prepositional ark of alliance between God, the world and men. Common sense con-stitutes persons; good sense, God's gift and instrument, articulates lan-guage.

Need we go further? The structure is coherent, and appears to be stable. But some of us are atheists, and the structure is, in fact, rather


unstable. It is threatened from below and from before. For the meaning thus obtained in the proposition is a surface meaning, always threatened by the chaos of drives and primary phenomena in the depth of the body (remember we are dealing with a psychoanalytic novel and have yet to provide an account of the genesis of language). My Adamic dialogue with God, when I name the world in accordance with the laws He has established, is threatened by the dreadful screams, emanating from the depths of my body, for the repressed Devil will be heard. So the polite conversation between Lewis Carroll and a little girl may reach its moment of truth in the mad screaming of Artaud, the demented poet (who, several decades after Carroll's death, accused Car-

roll of having plagiarised him). Here we meet a first type of nonsense: not the surface nonsense of the Victorian teller of tales, but the dreadful nonsense of the sufferings of the lunatic. But there is also a surface nonsense, or sense, and doxic meaning is also threatened not from below, but from before. The problem here is the circularity, or mutual implication, of manifestation, denotation and signification. As Deleuze carefully demonstrates in the third chapter of Logique du sens, it is impossible to posit one of the three without presupposing the other two: each in turn occupies the position of origin, only to be demoted by the next dirty rascal. The self, the world and God are in a state of mutual presupposition, at least as far as their inscription of language is concerned (the contrast between langue and parole is seen by Deleuze as a desperate attempt to establish some form of transcendence in language).

But we need not resign ourselves to such circularity. The way out is into a fourth element, or moment, the element, precisely, that circulates and explains the circularity of the other three, the element of sense. It (chrono)logically precedes denotation (it is not concerned with truth and falsity), manifestation (it is impersonal) and signification (contrary to the good sense of signification, it goes in both directions at the same time: its element is paradox, not doxa). This fourth element, therefore, which precedes the usual trinity and dislodges it from its position as source and site of meaning, has a number of characteristics, which Deleuze details, notably in the fourth chapter (Òf Sense').

Being expressed by propositions and attributed to states of affairs, being therefore a version or aspect of the event, sense is incorporeal and circulates on the surface of things. Since it is incarnated in the paradoxical instance that makes sense of the series, it is characterised by a number of paradoxes. The four paradoxes (of regress Frege's paradox; of sterile division the paradox of the Stoics; of neutrality the paradox of the medieval philosopher: contradictoria ad invicem idem


significant; of the absurd, or the impossible objects Meinong's paradox) are duly listed in Chapter 5. Being neither the terms of the proposition, nor the state of affairs, nor the experience of the subject that expresses it, sense, neither word nor body, sensibilia nor concepts, is neutral: it does not exist, but subsists and insists, indifferent to the usual dichoto-mies (neither personal nor impersonal, particular nor general etc.). Its grammatical marker in the sentence is neither a shifter, nor a deictic, nor a syncategorematic word: it is the verb itself, as marker of process, which inscribes the pure event. Sense is indistinguishable from the surface nonsense Carroll captures in his tales and his paradoxes; it is the effect produced by the circulation of the paradoxical element as it zips the two series together (a dangerous metaphor, as it reintroduces the teleology of meaning in a bi-directional process without principle and goal, beginning or end). Lastly, like the event, its temporality is not that of time's arrow (the temporality of the Markov chain and of the construction of meaning), but the eternal and at the same time bi-directional temporality of Aion, not Chronos.

What all this amounts to is a series of characteristics, the series of sense, the singular line of flight of the concept in which a number of terms consist in order literally to make sense: sense nonsense event pre-individual and impersonal singularities counter-effectuation Aion surfaces. And the hyphens in this series are not the hyphens of God and signification, they are the hyphens of the AND, of the oper-ation of consistency through which the series is constituted. This series, which develops on the surface, has a certain fragility. As we shall see, it has a shadowy double in the depths of the body, where another form of nonsense is lurking, and threatens to erupt: not all events are Deleuzean

the Lacanian Real and the Badiou event, in the shape of disaster, may well introduce an element of disorder into the order sense makes of chaos (an order which is not a traditional world, but Joyce's chaosmos, a strange mixture).

6 Illustrations, and the final correlation of sense

Let us start with Deleuze's own illustration of the workings of sense. The context is Carroll's Alice's Adventures in Wonderland. In the episode of the Caucus-race and the Long Tale, the animals drenched in the Pool of Tears and a very wet Alice attempt to dry themselves by listening to the driest story the Mouse knows.29 As might be expected, the method is not particularly successful, but the Mouse's tale is indeed very dry, being borrowed from a primer in history. It contains the following half-


sentence: Ànd even Stigand, the patriotic archbishop of Canterbury, found it advisable . . . ' Upon which the Mouse is interrupted by the Duck, who asks what referent is denoted by ìt' (when I mean a thing, he claims, it is usually a frog or a worm so what does this ìt' mean?). Since the ìt' is merely a grammatical prop (it anticipates the yet unex-pressed infinitive clause that should follow àdvisable'), it has no refer-ent, and the Mouse is deeply embarrassed, so he resorts to verbal terrorism, and the story is never resumed. For a linguist, this ìt' testifies to the quality of Carroll's intuitions about language, since it had to wait a hundred years to receive a convincing account in the Chomskyan theory of transformations, in this case the movement of extraposition.30

But Deleuze gives it another interpretation. It is clear that for the Duck, who is the representative of common

sense, ìt' should be a bearer of denotation (this is helped by Deleuze's translation of ìt' as `ceci', a deictic). But the Mouse, even if he is not aware of the complexities of Chomsky's generative-transformational grammar, is aware that the function of this ìt' is not to denote, that it has no referent. Or, rather, that the only referent of this ìt' is not an edible object, but the event expressed by the proposition, incarnated by the verb, temporarily absent but coming later, as the Mouse insists on finishing, if not the story, at least this sentence: `The Mouse did not notice this question, but hurriedly went on, `` found it advisable to go with Edgar Atheling to meet William and offer him the crown.'' '31 So the event is expressed by a verb in the infinitive and the ìt' is the ambiguous term in which the two series of denotation and expression paradoxically diverge (there is expression but no denotation) and con-verge (the ìt', as the Duck is aware, suggests or maintains the possibility of denotation). Hence the definition of sense that immediately follows:

Sense is never only one of the two terms of the duality which con-trasts things and propositions, substantives and verbs, denotations and expressions; it is also the frontier, the cutting edge, or the articu-lation of the difference between the two terms.32

The question, of course, is: can we generalise from this singular (single and peculiar) example? Can we make sense appear in the most innocu-ous, common-or-garden utterance? The following sentence, inscribing a single proposition, makes no claim to either semantic complexity or literary excellence:

(i) This blackboard is black.


This is indeed a well-formed, subject predicate proposition, with God (or signification) duly present in the copula. It is also endowed with truth value, and therefore denotation (my `this' points, or fails to point, towards an object, and the blackboard in question is, or is not, black). And it is endowed with manifestation: I can easily recover an implicit clause that grammatically governs my sentence and reintroduces the subject: Ì assert that this blackboard is black.' I can do this, because there is a marker of manifestation in the original sentence, as the deictic, being a first-person deictic (contrast `this country' with `that country') does not mark denotation only. So my proposition is thick with good sense and common sense: it offers to give solid evidence of the claim it makes, and it is a perfectly valid move in the game of interlocution.

In order to make sense appear beneath or before the thick layer of meaning, a coup de force is needed: we need to offer violence to common sense and good sense in order to dislodge them from their established position, where meaning is solid because frozen. We need, before good sense, the fluidity of sense, and we may need, beneath or beyond common sense, the delirium that threatens the order of language. Let us do just that, and introduce an element of controlled madness in our proposition:

(ii) This blackboard is and isn't black.

This, a blatant flouting of the principle of non-contradiction, but one that requires little linguistic expense, takes us beyond truth-value and denotation: the blackboard is now an impossible object, the proposition an absurdity, the speaker an idiot. Note that in order to fall back on my feet, and turn such absolute falsity into necessary truth, all I have to do is change the ànd' into òr'. This provides a good illustration of the kind of nonsense sense consists in: we understand what Deleuze means when he says that sense is neutral and bi-directional. From the point of view of sense, contradictories are equivalent, contradictoria ad invicem idem signifi-cant. Since we have started playing the fool, let us become literal, and say:

(iii) This black board is black.

Between (i) and (iii), only the intonation changes. But our initial proposition is somewhat troubled by this new version, which makes an older layer of language, one concealed in the congealed word `black-board', re-emerge. Is this `blackboard' a board? Actually, this one is not. Why is it called so, then? Because, originally . . . etc. What emerges in (iii)


is an element of sense present in (i) only as a trace; whence the idea that sense belongs to the realm of the virtual: `sense' is the set of virtualities on which the proposition draws, and which it actualises only in part. Denotation, manifestation and signification seek to anchor the propos-ition in what lies outside language (self, world and God): sense remains within language, on the surface of language. Even more, it defines lan-guage: it is what makes language possible.

Let us explore the virtualities further, and produce two more propos-itions:

(iv) The blackbird is black. (v) The black bird is black.

I am indulging in paronomasia, in the exploitation of the phonetic resources of language. I am being playful, or poetic, abandoning my mastery as speaker and giving sense a free rein: for what is sense, if not a reservoir, or a treasury, of tropes as appears if I offer the following false contradiction:

(vi) The blackboard is green.

There is no contradiction in good sense and common sense here: what my university puts at my disposal under the name of blackboard is a green expanse of a material I cannot identify, except that its artificial nature is not in doubt. But there is a sense in which language (that is, sense) whispers to me that blackboards ought to be black, strawberries ought to grow on straw, and ladybirds should have wings that are feathery. Common sense despises such childish games, fit only for poets or Continental philosophers. But language, or sense, will play, for instance in those sentences-within-sentences rich in naughty innu-endo (take the song, Òh, sir Jasper, do not kiss me': if you repeat it, each time with the exception of the last word, the protesting damsel will eventually yield to the squire's advances). And since madness (with a good deal of method mixed in) has set in, why stop here?

(vii) Vis belle as qu'aÁ bord hisse belle laque.

This is an instance of traducson, worthy of Luis d'Antin Van Rooten's Ùn petit d'un petit / S'eÂtonne aux halles.'33 It is obviously pregnant with sense, perhaps at the expense of meaning. But do not ask me that mean-

ing, I am merely the author, that is the mouthpiece of language.


So sense is neutral, virtual, tropic, more than slightly mad, but in a delightful poetic way. It permeates the work of Lewis Carroll, an indefat-igable explorer of sense. And a reading of e.e.cummings's poetry would, I think, show that the poems are based on a systematic inversion of the traditional positions of sense and meaning: meaning, which habitually obtrudes, is simple, but carefully concealed; what is explicitly offered to the reader, on the surface of the text, is the complexity and playfulness of sense hence the apparent àgrammaticality' for which cummings is justly famous.

The theory of sense developed in Logique du sens is entirely coherent. As usual, its coherence can be shown by a correlation, where the two divergent series of sense and meaning converge:

1 2 3 I sense paradox event II good sense and common sense doxa state of affairs

4 5 6 I verb expression Aion II other parts of speech `denotation' Chronos

Only columns 4 and 5 need further commentary. Column 4 lists the parts of speech that inscribe the sense-event or serve for the construc-tion of meaning. And the contrast is double: on the one hand, the verb that marks sense is opposed to the shifters of manifestation, the deictics of denotation, the syntactic markers of signification; on the other hand, the verb, marking the pure event as process, contrasts with the nouns and adjectives that refer to things and states of affairs. Column 5 con-trasts sense as `the expressed of the proposition', as Deleuze calls it, and the general construction of meaning, here represented by the denota-tion that seeks to link the proposition to the state of affairs in a direct relation of reference and representation.

But in our exposition of the theory of sense another correlation has appeared: it is not as systematic as the first, and describes the emergence of surface sense out of the depths of the body. It can be formulated thus:

1 I sense as surface nonsense II deeper nonsense of body

3 I articulated language II scream/breath

2 surface depth

4 secondary phenomena primary phenomena


As column 4 shows, this correlation is psychoanalytic in inspiration: it describes the genesis of language, the emergence of articulated language out of the noises produced by the body, whose primary order, a first ordering of chaos, is complexified on the surface.

7 The genesis of language

There is a deep-seated link between Deleuze events (but not Badiou events) and language. Events are deployed in the medium of language and in turn they make language possible:

[Sense] is exactly the boundary between propositions and things . . . It is in this sense that it is an èvent': on the condition that the event is not confused with its spatio-temporal realization in a state of affairs . . . The event belongs essentially to language: it has an essential relationship to language.34

Or again, in a passage from Chapter 23 that provides a detailed account of Aion, `the pure empty form of time', where events dwell:

It is this new world of incorporeal effects or surfaces which makes language possible. For . . . it is this world which draws the sounds from their simple state of corporeal actions and passions. It is this new world which distinguishes language, prevents it from being confused with the sound-effects of bodies, and abstracts it from their oral anal determinations. Pure events ground language because they wait for it as much as they wait for us, and have a pure, singular, impersonal, and pre-individual existence only inside the language which expresses them.35

It appears that Deleuze has set himself a philosophical task of the most traditional kind, which can be expressed in Aristotelian terms: the task of a philosophy of language is to explain the passage from phone, the sound effects of bodies, the noises the human vocal tract produces in the same way as that of animals, to logos, both speech and rational dis-course, where the animal sounds are articulated, where they make sense. Or again, how do we move from the actions and passions of bodies in their mixtures to the incorporeal existence of logos as reason? Deleuze's theory of event and sense tells us the tale (or novel) of that passage: events it is that make language possible, he states at the outset of Chapter 26 (Òf Language').36 They are said to `ground' language,


except that their incorporeal nature makes such `grounding' highly paradoxical, or metaphorical. But Deleuze's problem, which he inherits not only from Aristotle but also (mainly) from the Stoics (remember the pseudo-paradox of Chrysippus, `when I speak of a chariot, a chariot goes through my mouth', which deliberately conflates and confuses word and referent) is of considerable interest to the student of literature. Deleuze insists on the materiality of language, on its origins in the primary processes of the body; he details the emergence of voice from noise, and of language from voice (this is the subject of Chapter 27, Òf Orality': the voice the Voice of God is the canonical example modulates phone in anticipation, and expectation, of the event that will make sense, and therefore language, of it). He thus accounts, which is demonstrated in the practice of the poet, for the materiality of sounds in their inseparability from the ideality of articulate language. In this novel of the genesis of logos, the two main parts are played by Carroll, the elegant poet of surfaces and the complexity of articulated language, and Artaud, the unrespectable, because demented, poet of screaming depths. In Logique du sens, Lewis Carroll appears to have the upper hand language must be grounded, the event made welcome in polite, if agonistic, interlocution. In subsequent books, however, he vanishes, and Artaud, who provides the essential concept of `body without organs', remains in sole possession of the field. There is more to this than a simple change in literary taste: with Carroll's demotion, it is the importance of language that is denied. My contention, of course, is that the ghost of repressed language unceasingly returns to haunt Deleuze.

You will not be surprised if I suggest that this quasi-psychoanalytic genesis of language is best expressed in a correlation. The correlation formulates a three-stage model of the genesis of language in which the reader may, or may not, recognise an echo of the Freudian topology of the Id (or depth), the Ego (or surfaces) and the Superego (or heights). As usual, the correlation, even if it does not take the form of a diagram, or table, is entirely explicit: in fact it quite fittingly provides the chute of Logique du sens, as it can be found in the last three pages of the last chapter.37

1 2 3 I Primary order Secondary organisation Tertiary arrangement II Pre-sense Sense Denotation/Manifes-

(noise into voice) (verb) tation/ Signification III Passions/actions Event Ego/World/God

of body (individual/person)


IV Depth Surface Height / \

bodily metaphysical V Freud/Artaud Carroll Plato

This three-stage view of the genesis of language can also be taken synchronically as a three-tiered stratification of language (as every speaker recapitulates in the ontogenesis of speech the phylogenesis of the human faculty of language). It also operates, in row V, a philosoph-ical distribution of tasks. And if language is not mentioned in the correlation, it is because it is everywhere: it circulates through the correlation and gives it its coherence. It makes sense, welcomes the event it expresses, links bodily actions and passions and metaphysical heights, tames primary order into secondary organisation, only to be tamed, or made respectable, in tertiary arrangement. Logique du sens, even if it is ostensibly devoted to a theory of sense and the event (or rather because it is), is indeed Deleuze's book on language.

8 Conclusion

The first conclusion to be drawn is that Deleuze's theory of language is of extreme potential interest to literary critics. The Deleuze event, being so clearly involved with language, enables us to decide that the highest task of literature is not to represent the event, to re-enact or reproduce in memory the flash or illumination of the encounter with the Lacanian Real, but to be the event itself. Literature is concerned with the event of language, far more than with the event in language. These slightly grandiloquent words need to be put to the test of concrete analysis: this shall be the object of the Interlude that immediately follows.

Since Logique du sens is a relatively early text in the Deleuze corpus (even if translated late), all there remains to do is give an idea of its influence on the rest of Deleuze's oeuvre. We have already seen that the concept of sense more or less vanished, but the concept of event flour-ished, so that Logique du sens is both an adumbration of Deleuze's later (and somewhat implicit) philosophy of language, and a dead end. In it we find a geography of surfaces, where events and sense dwell, and the concepts of singularity, of chaosmos, even (albeit fleetingly), of `body without organs'. But a surface is not a plane of immanence, in that it requires a vertical structure of depths and heights. And the bodily depths of primary processes are not a smooth body without organs. Lastly, but not least, the ideality of sense will later appear incompatible


with the materiality of forces that lies at the heart of Deleuze and Guattari's pragmatics. The materiality here ascribed to phone and noise will later rise up to the surface, as planes are drawn over chaos.

Interlude 2

Making Sense of Literature ± Joyce, cummings, Woolf

1 The box of tools

Deleuze likes to compare philosophical concepts to a box of tools. Having described the characteristics of a few of those event, sense etc. I am eager to put them to use. Ìt is at work everywhere', as the philosophers say: but what does ìt' say about literary texts?

A number of precautions must be taken. We must make sure that Deleuze's concepts in Logique du sens are adapted to literature, as the carpenter's tools are of little avail to the mechanic. True, since the book is a reading of Lewis Carroll as much as of the Stoics, there is a strong suspicion that they are. But the dangers of application are not limited to this. We might be content to illustrate the concepts by applying them to short passages of a text or episodes of the story it tells a practice both facile and devoid of interest. We might take the literary text as a pretext for an exposition of the doctrine, thus imposing a form of coherence on it, but one coming from the outside: such critical transcendence, answering questions before the text has had the opportunity to ask them, is also facile and deeply uninteresting. Our ambition must reach higher we must seek to use Deleuze in order to produce a new inter-pretation of the text.

I am confident that Deleuze's concepts can help us achieve this. If you will allow me a personal anecdote, I started reading Deleuze when Logique du sens was first published, being at the time engaged in a thesis on Victorian nonsense literature in general, and the work of Lewis Car-

roll in particular. I had spent a whole year reading the Carroll critical corpus and remember this was 30 years ago I had experienced a feeling of disappointment close to despondency: some readings were psychological and trivial; some were more joyful, but mad (for instance

132

Interlude 2 133

an interpretation of the adventures of Alice as an acid trip); some were heavy with theoretical premises, and ponderous and predictable in their interpretative results: thus Alice, who keeps changing size, was, in some psychoanalytic readings, nothing but a walking penis. I still remember the feeling of illumination that dissipated the doldrums when I opened Logique du sens: here was a reading which, at last, was both strong and original. Reading Carroll in the light of the Stoics was a strange idea, but one that opened vistas of sense. True, at the time I understood little of Deleuze's highly technical philosophical discourse (it has taken me a bare 30 years to get rid of that feeling), but it undoubtedly gave new life to Carroll's text.

There are still limits to the exercise. A philosophical reading of Carroll or any literary text in the light of Deleuze does not cancel sedimented interpretations (and 30 years later the Carroll critical corpus has some-

what improved). Although original, it does not provide an entirely novel interpretation of the text: pottering with Deleuzean tools will not neces-sarily (hopefully not) yield a global `Deleuzean reading' of the text. The claim must be slightly more modest: a Deleuzean reading of the text will take its place in the unlimited chain of interpretations by reaching parts of the text other interpretations have so far failed to reach.

I shall propose three attempts at using Deleuzean tools. In order of length: the incipit of a story, the first stanza of a poem, a whole short story.

2 The incipit to `The Dead'

The first sentence of Joyce's `The Dead' is justly famous:

Lily, the caretaker's daughter, was literally run off her feet.1

A grammatical analysis of the sentence, for instance in the light of Chomsky's early transformational model, will tell me that for this sur-face sentence there are two corresponding deep structures, which can be noted down (in a simplified version, that is, as if they were surface sentences rather than strings of symbols) thus:

1 Someone literally ran Lily off her feet. 2 Lily was the caretaker's daughter.

These two simple, active, declarative sentences will then be combined through a number of grammatical transformations. Sentence 2 will


become an elliptic appositive relative clause (`Lily, [who was] the car-etaker's daughter, . . . ') and the now complex sentence will undergo passivation, with cancellation of the agent noun phrase (` . . . was liter-ally run off her feet by someone'), since its contribution to the overall meaning of the sentence is minimal. Thus is the complexity of the surface structure explained by the simplicity of the deep structure and the transformational history that generates the surface.

However, such an analysis says little about the meaning of the sen-tence, and nothing about its striking character, which, surely, resides in the last six words ` . . . was literally run off her feet', and especially in the adverb. So, since depth analysis has obvious limitations, we might perhaps remain on the surface, and treat the sentence as a line develop-ing along a plane of immanence.

This is where Deleuze's theory of sense can help. In terms of Deleu-

zean doxa, the sentence is well formed (this is indeed, as the previous analysis has conclusively shown, a grammatical sentence). The sentence has denotation: the function of a name, especially if supplemented by a definite description, is to enable us unequivocally to pick up a referent. This, therefore, is Lily, and none other. But the sentence also has mani-

festation. Somebody, a speaker, a narrator, is manifested in such denota-tion (remember that each of the three aspects of meaning is presupposed by the other two: their definition is circular), and he is responsible for the choice of modality (`literally' is a modal adverb) and metaphor (`run off her feet' is a metaphorical phrase), even if, as we shall see, the words are not originally his. There, of course, lies the rub, because I am not sure who the original speaker is, or rather I have the impression that more than one speaker addresses me in the sentence. I feel sure that at the end of the sentence, the words are Lily's: she it is who exclaims, Ì am literally run off my feet!' those words are a kind of signature. But she cannot speak of herself in the third person, nor is she likely to add the precision, `the caretaker's daughter'. Who, then, is speaking? Gabriel, no doubt, the `hero' of the piece, except that at this stage he is not yet born to the fiction. In any case the sentence is also endowed with signification: it is, as I have said, syntactically coherent and meaning (`literal' meaning) can be construed, if not along the lines of a Markov chain, because of the embedding process that creates the elliptic relative clause, at least along the lines of Chomsky's syntactic structures: the rest of the text will confirm that Lily is indeed the caretaker's daughter. The trope of false anaphora, which is so frequent in incipits, and marked here by the use of the definite article, ensures an imaginary coherence with a previous context that does not exist, since this is the first sentence of the text.

- - - - - - - - - - - - - - - - - - - - - -

- - - - - -

Interlude 2 135

Those three strands of meaning make the sentence an expression of good sense and common sense. I can plot them on the following dia-gram:

Referent Denotation- - - - - - - -! I Sender- - Text Receiver Manifestion IÃ I - -Signification

Code

The diagram is asymmetrical, in so far as only three out of its four branches have been named: if denotation is the name of the relation between text and referent, if manifestation occurs between author and text and signification between text and code, what is the name of the relation between text and receiver, or reader? This branch of the diagram is not unimportant, since this is where interpretation may be expected to take place. But this is not a question of absence or lack, but of excess: this branch, which apparently is empty, is in reality too full. It is filled with the interpellation that produces an effect of good sense and common sense in the reader (through the combined effect of the other three branches). And, paradoxically, it is also filled, or so I shall contend, with the proliferation of sense that prevents the text from freezing into definitive meaning, that ensures the circulation and multiplication of interpretations.

For there is sense in this sentence, a sense neither original nor deep, but running along the surface of the sentence. We have already seen it at work in the in medias res beginning of the text (this is the beginning, and yet obviously it is not, because of the trope of false anaphora), and in the uncertainty about the voice that utters the text (Lily is clearly speaking, and yet somebody else must also be speaking).

The function of sense, on the fourth level where the text `makes sense', is to pre-empt the other three: sense is folded upon itself, it unfolds along lines on the plane of immanence, but its lines are never straight, and denotation, manifestation and signification become un-certain. Manifestation is the first that caves in. Because of the strange form of

indirect free style in the first sentence of the text, it becomes undecid-able. For the sentence relates what Lily probably said, but in words she cannot have used herself, as appears if we compare our sentence with another one uttered, a few paragraphs later, by Lily in the first person: `The men that is now is only all palaver and what they can get out of


you.' This is Lily's real dialect, and the differences, both syntactic and lexical, with our sentence are clear. In a famous analysis of this incipit, Hugh Kenner shows that `literally' belongs to Lily's dialect: she it is who says `literally' and means `figuratively', since the phrase that follows is a metaphor. Kenner calls the operation whereby Joyce places invisible inverted commas around the word the Ùncle Charles principle'.2 The result is that the sentence is spoken by more than one voice a naturally Bakhtinian text: Bakhtin compares an utterance with a monad in which a multiplicity of voices and points of view are always-already engaged in dialogue. Hence the idea that here, it is the text that speaks, and not Lily, nor Gabriel, nor the anonymous narrator (who may or may not be Gabriel), nor Joyce himself. Or rather every single one of those various voices, a Babel of voices convoked by the sense of the utterance: sense is the voice of the text. Signification fares no better. Although it is grammatically coherent, the

sentence uses a grammatical construction that is legitimate but curious. The intransitive verb `run' is used in the passive, which goes against its `normal' usage (intransitive verbs do not take the passive voice). The reason is that what we have is a composite, or portmanteau construc-tion, both causative and resultative. The sentence may indeed be glossed in the following manner: `they made Lily run so much that she was off her feet' (or, if Lily is speaking, Ì was made to run . . . '), where `made' marks the causative nature of the construction and `so much that' its resultative nature (a typical resultative sentence would be: `he drank himself under the table'). Although the syntax is correct (resultative constructions are part of the grammar of English; they do allow us to passivate intransitive verbs, and produce strange contexts for reflexive pronouns, as in the sentence just mentioned), the meaning it allows us to construct is problematic. I could literally be òff my shoes', hardly òff my feet'. Our sentence shows that denotation alone cannot institute meaning, and that signification does not provide full coherence: in the interstices of significative coherence the presence of sense makes itself felt call this metaphor, in an extended sense of the term that covers all tropic moves.

For metaphor does make denotation uncertain. If Lily is rightly de-noted (this is indeed Lily), the event remains indeterminate: `being off her feet' has no literal meaning and requires interpretation. Even if the metaphor, being long dead and frozen, easily allows conventional meaning in, the uncertainty persists, for the impossible literal mean-

ing returns like a ghost to haunt the conventional metaphorical meaning. Even as Alice in Wonderland, when she finds herself taller

Interlude 2 137

than a giraffe, thinks of communicating with her feet by post, Lily's feet appear to be running by themselves, independently of their owner.

We are running into paradox, the natural site for Deleuzean sense. We may conceive it not so much as stratum of meaning as in the guise of a fold that folds itself with the other three. The surface of the text, even if easily understandable, is not merely a site for doxa: there is always an element of sense in it, which reintroduces chaos in the semantic cosmos. In our sentence, the site of sense is the adverb `literally'. For it has modal value, in that it is the mark of the speaker's point of view in his/her sentence. Linguists call such grammatical markers `hedges': they mark hesitation, distance or commentary. In our case the adverb is an intensifier: Ì am literally dead' means Ì am very, very tired.' But the status of this intensifier is problematic, as its utterer's meaning contra-dicts its utterance meaning, its metaphorical meaning contradicts its literal meaning. `Literally' is literally a mark of literalness, and meta-

phorically a mark of metaphoricity. Our adverb points towards an im-

possible literal meaning that it maintains under erasure, the better to mark the metaphorical status of the phrase that immediately follows. The result is that the adverb operates a deconstruction of the meaning of the sentence in order to give free rein to its sense, as the following table shows:

1 Literal interpretation (causative): Lily was made to run. 2 Metaphorical exaggeration (resultative): She was run off her feet. 3 False literalisation and increased exaggeration: She was literally

run off her feet.

The sentence moves from the literal to the metaphorical (a passage that is common enough). Then it undergoes a false literalisation that is both an anti-metaphor (we are going back to the literal) and an exagger-ated form of metaphor (as the metaphorical force of the phrase in-creases). This `literally' is the best embodiment I know of Deleuzean sense: it embodies one of the paradoxes of sense that Deleuze formulates with the already mentioned quotation from the medieval philosopher Nicolas d'Autrecourt: `contradictoria ad invicem idem significant', strict contradictories have the same sense.

Nor do we have to stop here. The adverb not only deconstructs the meaning of the sentence, it deconstructs itself. HeÂleÁne Cixous once remarked that the name of Lily, the Vestal Virgin that stands at the threshold of the house to welcome visitors for the Christmas party, that lily of unsullied whiteness (see Lily's remarks on `the men that is


now . . . '), is anagrammatically present in the adverb: Lily li(teral)ly. Such stuttering (`literally' can be an instance of the name `Lily' being repeated with some difficulty) draws our attention to a number of phonemes (/l/, /r/, /t/ etc.) that are the source of an anagrammatic family of words: `later', `tear-l', `real-t', `relat(e)', `(G)reta' (the name of Gabriel's wife). I am cheating, of course, as my use of hyphens and brackets shows. But no more so than Saussure in his anagram notebooks: there was a remainder in his anagrams as well. That remainder, the letter that cannot be integrated into the new word, or that must be added in order to reach it, is the mark of anagrammatic re-analysis, the mark that the process of sense never ends, that its proliferation knows no limit (neither the archaic limit of literal meaning, nor the teleological limit of the right interpretation). If, therefore, I go on cheating, my cluster of phonemes, `teral', will yield `letter', `litter', and the whole of Joyce, if not the whole of the English language.

This, of course, is an instance of delirium, but in the etymological sense: de-lira, or how to leave the furrow. And such waywardness is not entirely arbitrary. Here I could introduce, in anticipation of the chapter that immediately follows, the concept of minor language, which aims at making us understand that language is no transparent instrument of information and communication, but the site of conflict, a conflict waged in and among signifiers, a conflict of voices and discourses, the result of which is the stuttering of language I have evoked. Then it be-comes apparent that the caretaker's daughter, with her Irish working-

class dialect, causes the language of British imperialism and of canonical literature to stutter (even if the text is now a central part of the literary canon: such recycling marks the workings of dominant ideology).

The logical path to follow here would be to recall Deleuze's hostility to metaphor: since I already followed that path in my first interlude, I move on to my second example.

3 The first stanza of a poem by e. e. cummings

Poem 29 of the collection 50 Poems is well known because of its a grammaticality, and has often attracted the attention of linguists. Here is the first stanza:

anyone lived in a pretty how town (with up so floating many bells down) spring summer autumn winter he sang his didn't he danced his did3

Interlude 2 139

At the very first glance this stanza, as indeed the rest of the poem, appears to be a kind of word-salad. And, as usual with cummings, the very second glance will reveal, if not obsessive structure, at least a considerable amount of coherence. But there is no denying the amount of agrammaticality that was seen by linguists, in the first glory of the Chomskyan system, as a challenge.

Let us take a closer look. On the first line, the adverb ̀ how' is used in the position of a determinative adjective, which it cannot rightly occupy. And the indefinite pronoun ànyone' is followed by the wrong verb. Ànyone can do this' sounds right, ànyone lived' is definitely the wrong verb and the wrong tense. The second line plays with the syntax of the sentence, etymologically understood as the right placement of each element. It seems to have shuffled the words like a pack of cards, and it is not difficult to put them back in their right order and to produce the perfectly normal sentence, ̀ With so many bells floating up and down' a metaphorical description of the ringing of bells that is not strikingly

original. The third line is hardly agrammatical: it simply does without, to speak like Aristotle, the syncategorematic elements of the sentence, prep-ositions or articles. We only need to reintroduce them to obtain, once again, a `normal' sentence: ìn spring, in summer, in autumn, in winter', or `whether it was spring, summer, autumn or winter'. The fourth line is justly famous in that the slots reserved for noun phrases are occupied by verbs, by the auxiliary part of the verb, to be precise. For verbs can be transformed into nouns (called `deverbal nouns') but on certain condi-tions: only the main stem can undergo nominalisation; all marks of tense and modality disappear; the right suffixes, in `-ing' or `-ion' for instance, must be used. Thus, we talk about `his coming of age', `the workings of language', etc. None of these conditions have been met here, since the suffix is absent, marks of tense and negative modality are present, and the form ̀ didn't' suggests that the verb ̀ do' here is the auxiliary, not the main verb (which can be nominalised: `strange doings'). And we note that, unlike lines 2 and 3, line 4 does not lend itself to an unambiguous construction of meaning.

The rest of the poem is pervaded with similar instances of agrammati-

cality. One in particular recurs like a leitmotiv, an extended use of the preposition `by', which gives anomalous phrases like `more by more' (easily converted into `more and more') or `when by now and tree by leaf' (which is less easy to interpret).

Not all is lost, however. A careful reading of the text, requiring no great ingenuity, will provide the eager reader with what I have called elsewhere a `tin-opener':4 a key to open the meaning of the text. The


reader will note the systematic recurrence of the grammatical anomaly of the first word, ànyone', in lines like `that noone loved him more and more', ànyone's any was all to her', `someones married their everyones', and òne day anyone died i guess / (and noone stooped to kiss his face)'. She will remember the story of Ulysses' escape from the Cyclops, greatly helped by the fact that he had given his name as `Nobody', so that the literally benighted Cyclops could only denounce nobody to his brother cyclopses. She may also remember a passage in Through the Looking-Glass, where the King asks Alice to tell him whether she can see his messengers on the road, and Alice truthfully answers: Ì see nobody on the road', upon which the King exclaims: Ì only wish I had such eyes . . . To be able to see Nobody! And at that distance too!'5 Our reader, then, realises that in this poem indefinite pronouns must be taken for names, and that the poem relates the story of Anyone (a male figure, anaphorically referred to as `he') and Noone (a female figure, referred to as `she' the consistency of such anaphors goes a long way to give the poem its coherence): of their meeting, of their marriage and their death, so that the poem's narrative line can be summarised by a sentence of Joyce's: `they lived and loved and laughed and left'.

The choice of names is fairly obvious: this is the story of Everyman, l' uomo qualunque in Italian, M. Untel in French, Mr So-and-so in English. The advantage is the systematic ambiguity the name-pronoun (which is never capitalised) provokes. When Anyone dies, `noone stooped to kiss his face': you can choose between an optimistic and a pessimistic read-ing of the line. In spite of such playful ambiguity, however, it is our sad duty to note that the story-line of the poem is utterly trivial, a Mills and Boon without a happy end. No wonder that the heroes' lives should be compared to the succession of the seasons (`spring summer autumn winter') or of the elements (`sun moon stars rain' the weather as objective correlative of the various moods of the poem). This is a poem thick with doxa: with the good sense of the inevitable progression of human life towards death (as inevitable as the succession of seasons, or the sentence's progression towards the full stop that puts an end to it); with the common sense of the affects, of joy and of grief, with which such progression is usually accompanied.

But this doxic meaning is what interpretation yields: it is not, as such meaning usually is, always already given at the first reading, it must be construed. What the surface of the text gives us is something more playful and less orderly. I shall try to show that it is not meaning, but Deleuzean sense. Hence a striking characteristic of cummings's style, which is apparent in more poems than this: he inverts the usual rela-

Interlude 2 141

tionships between doxic meaning and sense. In most texts, meaning is given straight away, and straightforwardly: it hardly requires the detour of interpretation. And sense is only present in traces, which it is the task of interpretation to detect and analyse. In cummings's texts, on the other hand, trivial meaning is never given, it is what interpretation yields, whereas the surface of the text sparkles with the rich texture of sense. Cummings's texts are Deleuzean avant la lettre, in that they en-courage us not to interpret, but to remain on the surface, to explore its playful lines, to experiment with them (we remember Deleuze's `never interpret, always experiment!').

Let us, therefore, reread the first stanza in the light of Deleuzean sense: a number of characteristics of sense will thus emerge. The first line begins with the deliberate confusion of pronoun and name, and with this appears the first characteristic of sense, its indeterminacy, its capacity not to make an exclusive choice, but to keep two incompatible strands of meaning together. As for the rest of the first line, it can be analysed thus: the ill-placed adverb (à pretty how town') can be replaced with an adjec-tive, which will force us to notice that `pretty' has two possible grammat-

ical values, that of adverb (`this town is pretty large') and that of adjective (`this large town is pretty'). And both values are virtually present in our line since the adjective slot is occupied by ̀ how', which does not allow us to make a choice. But where does this `how' come from, how can it be justified? Imagine it is a trace of the following `dialogues':

A: This is a large town. B: How large? A: Pretty large!

Or, which is perhaps better:

A: This is a pretty town. B: How pretty? A: Pretty how!

`How' has become an intensifier, like `literally', albeit an unconven-tional and playful one.

What the line embodies is another characteristic of sense, its virtuality: sense is a reservoir of meaning, as real as doxic meaning, and virtually present in the text, some of which will be actualised in reading and interpretation. The same occurs with the famous last line of the stanza, `he sang his didn't he danced his did,' which can also be analysed as an


instance of virtuality, in the guise of a portmanteau construction (the term is inspired by Carroll's `portmanteau words', which are central to Deleuze's analysis of sense in Logique du sens).Thus, if we remove the illicit auxiliaries from their noun slots, we may replace them by the `natural' complements of the verbs: `He sang his song and did his dance.' Or again, to keep closer to the text, and in particular not to forget the negative modality, we may suggest another rewriting: `He didn't sing his song, but he did his dance.' And soon, exhilaration getting hold of us, we will produce a third: `He sang about what he didn't do, he danced about what he did.' The advantage of sense, its virtuality, is that it allows us not to decide, nay, that it forbids us to do so. It accommodates all the lines of meaning that may be developed from the text, and all the meaningless ones. Sense, as we saw in the last chapter, is meaning not yet frozen into doxa; it is also before and beyond any meaning. And the text is telling us that interpretation is inevitably a freezing of sense. So that when we come to the last line of the poem, `reaped their sowing and went their came', there is no difficulty in understanding that they reaped what they had sown and either went their way or went the way they came, or, naturally, both. And there is a warning that this very lack of difficulty is the mark of a dereliction of sense, that the text is richer if we refuse interpretation.

The second line of the poem, as we have said, only needs reshuffling in order to be commonsensical. But why this strange order? The answer is simple: iconism, an instance of the fluidity of sense. It is true that the line as it is is much more striking than its doxic equivalent: the place-ment of the two adverbs, ùp' and `down', at the extremities of the line, with metrical beat, reproduces the movement of the clapper as it swings (or `floats up and down') to hit the bell.

The rest of the poem will provide us with instances of the bi-direction-ality of sense (the example we mentioned was the ambiguity in `noone stooped to kiss his face') and of its abstraction, as the preposition `by' becomes prevalent and a marker of the more abstract and indeterminate form of relation, a marker of Relation itself (a function which, in the English language, can be said to be fulfilled by òf'). Here are a few examples:

More by more When by now Tree by leaf Bird by snow Stir by still.

Interlude 2 143

The interpreter's task in each case is to give contents to the relation, of which the preposition gives only an abstract diagram. The task is easy in `more by more', less so in `bird by snow'. Being arch, our poem also provides an occurrence of the matrix of all such phrases, the only instance in which `by' is used grammatically: `little by little'.

Our poem, therefore, is the site of an inversion of the respective roles of sense and meaning in language. And it also provides us with a number of determinations of sense: its virtuality, its indeterminacy, its abstraction, its fluidity, its bi-directionality.

4 A reading of `Kew Gardens', by Virginia Woolf

From the very beginning of `Kew Gardens', the reader is plunged into a chaos of impressions:

From the oval-shaped flower-bed there rose perhaps a hundred stalks spreading into heart-shaped or tongue-shaped leaves half- way up and unfurling at the tip red or blue or yellow petals marked with spots of colour raised upon the surface; and from the red, blue or yellow gloom of the throat emerged a straight bar, rough with gold dust and slightly clubbed at the end. The petals were voluminous enough to be stirred by the summer breeze, and when they moved, the red, blue and yellow lights passed one over the other, staining an inch of the brown earth beneath with a spot of the moist intricate colour.6

The chaos is one of impressions: Nature presents itself in its cosmic luxuriance, in the rich profusion of its beings, a profusion so over-whelming that the perceiving subject (but we do not know which subject this is, or even indeed whether the perceiver, the point of view, is a subject all we have is a bird's-eye view of a flower-bed) can hardly construct an overall picture, lost as it is in a wealth of detail.

The situation has not changed in the last paragraph, at least at first sight:

How hot it was! So hot that even the thrush chose to hop, like a mechanical bird, in the shadow of the flowers, with long pauses between one movement and the next; instead of rambling vaguely the white butterflies danced one above another, making with their white shifting flakes the outline of a shattered marble column above the tallest flowers; the glass roofs of the palm house shone as if a


whole market full of shiny green umbrellas had opened in the sun; and in the drone of the aeroplane the voice of the summer sky murmured its fierce soul. Yellow and black, pink and snow white, shapes of all these colours, men, women, and children were spotted for a second upon the horizon, and then, seeing the breadth of yellow that lay upon the grass, they wavered and sought shade beneath the trees, dissolving like drops of water in the yellow and green atmos-

phere, staining it faintly with red and blue. It seemed as if all gross and heavy bodies had sunk down in the heat motionless and lay huddled upon the ground, but their voices went wavering from them as if they were flames lolling from the thick waxen bodies of candles. Voices. Yes, voices. Wordless voices, breaking the silence suddenly with such depth of contentment, such passion of desire, or, in the voices of children, such freshness of surprise; breaking the silence? But there was no silence: all the time, the motor-omnibuses were turning their wheels and changing their gears; like a vast nest of Chinese boxes all of wrought steel turning ceaselessly one within another the city murmured; on the top of which the voices cried aloud and the petals of myriads of flowers flashed their colours into the air.7

The question that can immediately be asked is, what has changed between the beginning and the end of the story? For we notice a strange effect of boucle: the story begins and ends by focusing on the flowers in the flower-beds, on their stalks and petals, on the coloured impressions that form themselves in the perceiver's eye. Yet differences there are. The most obvious is enlargement: the hundred flowers are now `myriads', and the perceiving eye has zoomed up, so that the object of perception is no longer only the flower-bed, but the whole of the gardens, of the city that surrounds them, and of the cosmos. There is, of course, a relation between the spaces thus delimited; they form concentric circles, or perhaps a nest of Chinese boxes, so that the flower-bed becomes, synechdochically, the centre of the cosmos, the original mythical um-

bilicus around which the cosmos is structured. But there is also another, even more significant difference. The impressions in the first paragraph are visual: light, shapes and colours. In the last these visual impressions are still present, but they are no longer dominant: what dominates is impressions of sounds. Voices are heard: wordless, but nevertheless breaking the silence. And the cranking of the gears of omnibuses, and the humming of their engines. What is dominant is the murmur of the city, the humming of the whole cosmos. A joyful humming, as those

Interlude 2 145

voices break the silence with `such depths of contentment'. And we note that the Chinese boxes of which the universe is composed, as they revolve around one another, murmur too. So that it is difficult not to perceive an allusion to the musical cosmos embodied in the Platonist phrase, `the music of the spheres'.

The object of the text, therefore, appears to be ambitious but simple: it turns a humble flower-bed into a representation of the cosmos, it con-structs a cosmos out of a flower-bed. No wonder therefore that, at the centre of the flower-bed, we should find a snail: that creature carries its house on its back, but its house is the same shape as the world, so the story can be seen as the unfolding of the spiral-like shell of the snail. We started with sense impressions, and now we have reached a symbol.

This snail is of paramount importance in the story. It is introduced immediately after the incipit I quoted, in the following terms: `The light fell either upon the smooth, grey black of a pebble, or the shell of a snail with its brown circular veins, or falling into a rain-drop, it expanded with such intensity of red, blue and yellow the thin walls of water that one expected them to burst and disappear' (p. 35). The snail therefore is, at this stage, the only living creature mentioned in this vegetable and mineral world. It is also the true hero of the tale, in that a number of human couples will be evoked, but only fleetingly: they will come within range of the camera, pass along the flower-bed, and vanish, whereas the snail, who is trying to cross the flower-bed it inhabits, remains, a persistent presence, at least until the last paragraph. So that the story has a new object: `the snail's progress', as the text describes it on the second page (p. 36). Our snail, therefore, is a pilgrim, his journey an Odyssey, or an allegory of human travail. No wonder, in the same paragraph, the snail should be referred to first as an ìt' (as is natural: Ìt appeared to have a definite goal', p. 36), but later as a `he': `Before he had decided whether to circumvent the arched tent of a dead leaf or to breast it there came past the bed the feet of human beings' (p. 36). The grammatical shift in anaphoric pronouns answers a double aim: it raises the snail to the status not only of the hero of the tale but of a symbol of humankind: his path is our journey through life; and it provides a natural transition with the human couples that come crowding on the stage.

There is, however, a serious problem with this allegorical reading. It is simple: the snail never reaches his goal, on the other side of the flower-

bed. This failure is neither tragic (the creature might realise the absurd-ity of his endeavour) nor comic (obstacles might multiply and distract him): he simply vanishes from the range of the perceiving eyes, which


no longer focus upon him, so that the living presence in the last para-graph, apart from the ghost-like human beings that are projected upon the screen, is not the snail, but `the thrush' (we can only hope that this false anaphora the bird has not been mentioned before is not an indication that our snail has come to a sad end) as if the bird's-eye view of the first paragraph was incarnated at last. But the failure of the snail to reach its goal entails consequences: a general impression of aimlessness pervades the story, and this cosmos appears to be singularly chaotic: couples drift about, like fallen leaves caught in the breeze, and the apparently purposeful snail fares no better, at least till the last para-graph, when the music of the spheres re-establishes some sort of cosmic order.

But this is only an impression, on the surface of the text, one we will have to discard when we look at the structure of the text, when we consider the text as a structure.

The first element of structure becomes apparent when we realise that the story is composed of a number of passages that can be assimilated to camera shots in the cinema. The camera alternately focuses on the flower-bed and on the human couples that saunter along or around it. And the alternation is entirely regular, to the point of being obsessive. If I call X a flower- bed shot and Y a shot about a couple of human beings, the story has the following structure:

X1 Y1 X2 Y2 Y3 X3 Y4 X4

This structure is known in rhetoric as a chiasmus. And since a chias-mus is a symmetrical figure, it has a centre of symmetry, the blank between Y2 and Y3 that should be occupied, were this a simple case of alternation ( . . . Xn Yn . . . ) by a flower-bed shot. The chiasmus is the reduction of a nest of Chinese boxes to a line, where each sphere is represented by two points (X1 X4; Y1 Y4; X2 X3; Y2 Y3). And the universe thus diagrammatised has a hub, an absent centre, the flower-

bed that is not there at the centre of the text and whose present absence, like the paradoxical element in Deleuze's structure, organises the circu-lation of the shots, or spheres. What I am saying, of course, is that the chiasmic structure of the text's sequences reproduces, en abyme, the form of the universe the text evokes or constructs: the text itself is a nest of Chinese boxes.

Nor is this all. Let us look more closely at the Y sequences, and at those human couples. Their characteristics vary, and no couple is like any other: but they do not differ randomly, as Table 1 shows. Y1 is a married

Interlude 2 147

couple, with children: they reminisce about the past and their first experience of love. Y2 is made up of father and son: the aged father is no longer in his right mind and his speech is delirious, while his son remains silent, and tries to prevent his father from coming to harm. Y3 is a couple of elderly women, garrulous old crones who chatter of this and that. Y4 is a courting couple in the first marvellous moments of a successful relationship: their speech is rather vague, but their spirits are high. From this very list, it appears that the choice of couples is an integral part of the construction of the text as structure: the story gives us, if not the ages of man, at least the stages of the family incipient (the young couple), established and fruitful (the first couple), dissolving (the perhaps widowed father and his son), non-existent or no longer in existence (are the two old crones widows, or spinsters? They may even be possessed of any number of grandchildren about whom they inces-santly and vacuously talk).

About the determinations the text ascribes to these couples, two things must be noted. First, either those determinations (sex, for in-stance, or quality of speech) place the four couples in a chiasmus, or they seem to engage, Beckett-wise, in the logical exhaustion of possible combinations. Thus, we have the ages of man, but also the temporal ek-stases evoked in the couples' speech: the young couple talk of their future, the settled married couple is nostalgic, the old crones live in the present only, while the old man, in his dotage, invents an imaginary, or mythical past. It is now clear that the text combines an apparent chaos with an obsessive sense of structure.

The second thing to be noted is that the determinations of the couple situate them briefly, according to sex and age,8 and concentrate on the contents of their speech, and the quality (or lack thereof) of the com-

munication it achieves. And this communication appears to conform to

Table 1 The characteristics of the four couples in `Kew Gardens'

Sex Age Speech Communication Speech contents Time

Y1: married couple

Y2: couple of men

Y3: couple of women

Y4: young couple

Different

Same

Same

Different

Chiasmus

Middle-aged

One old, one young Old

Young

Exhaustion

Rather silent

One garrulous, one silent Garrulous

Rather silent

Chiasmus

Few words, lots of meaning Communication fails No communication (clicheÂs) Few words, lots of meaning Chiasmus

Reality

Imaginary

Trivial reality

Reality

Exhaustion?

Past

Imaginary past

Present

Future

Exhaustion


a simple rule of inversion: those who talk fail to communicate, and those who communicate do not need to talk, or to talk at length. The quality of communication is in inverse ratio to the number of words used, as the following gradient shows:

No communication Full communication - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -! silence ranting garrulousness few words communion silence

The first stage is the silence of non-cooperation, here represented by the old man's son. He does not talk because he knows attempts at verbal communication are useless: his father will not hear, or, if he hears, he will not listen. The second stage is the ranting of the delirious father. The third is the trivial bavardage (the Gerede of Heidegerrian parlance) of the two elderly women, a form of speech so empty that it loses all grammatical coherence, as the text points out:

`Nell, Bert, Lot, Cess, Phil, Pa, he says, I says, she says, I says, I says ``My Bert, Sis, Bill, Grandad, the old man, sugar,Sugar, flower, kippers, greensSugar, sugar, sugar' (p. 38)

This `pattern of falling words' (p. 38) is exclusively concerned with the unavoidable, and inevitably trivial, subjects of daily life, the family and shopping for food.

The fourth stage is that of the established couple, whose love for each other no longer needs to be expressed in words, and whose thoughts run in parallel, only needing in order to communicate the rare quilting point of a single word. We have all seen, in a restaurant, such silent couples, whose silence is one of profound contentment, not boredom or hostility. This is why the next stage is communion. The young couple hardly need words. All they need to do is squeeze hands and gaze upon each other, hence the vague and ineffectual nature of their words, whose apparent triviality (` ̀ `Come along, Trissie; it's time we had our tea.'' ``Wherever does one have one's tea?'' she asked, with the oddest thrill of excitement in her voice, looking vaguely round . . . ' p. 40) conceals the most exquis-ite feelings. And this is why the last stage is silence again silence as that towards which the fullest communication strives, and which is men-

tioned in the last paragraph in the oxymoron `silent voices'. So the whole text can be described as the unfolding of a structure,

which I can represent, after Greimas, as a semiotic square (see Figure 3).9

Interlude 2 149

4

Speech Communication

1 3

Non-communication No speech

2

Figure 3 Semiotic square of communication

As usual the semiotic square joins contraries (which admit a middle term: white is the contrary of black, grey is neither black nor white) and contradictories (there is no middle term between white and non-white). `No speech' is the contradictory of `speech'. And, as we have seen, in the world of the story, `communication' is the contrary of `speech'. The four vertices of the semiotic square can be joined, two by two, to form four `positions', numbered 1 to 4, on which, or with which, the text plays, and which structure the text.

Positions 1 and 3 are occupied by the various couples: thus, the couple of elderly women speak, but their garrulous speeches are soliloquies, whereas the married and the courting couples have little need of speech in order to communicate. This is the horizontal, human axis. But there is also a vertical, cosmic axis: position 2, neither speech nor communi-

cation, is the position of the world in the first paragraph, all light and no sound, but no sense either; while position 4 is the position of the world turned cosmos in the last paragraph, as the colourful universe is now humming and silent voices sing their deep contentment. And we under-stand the structure of the text: the snail's progress takes him from 2 in the direction of 4, while around the flower-bed, on their horizontal plane, human couples wax and wane like planets.

You may well ask: where is Deleuze in all this? So far, he is nowhere. My belief is that he may allow us to go a little further. There is still an element of the text that defeats the obsessive structure just outlined: the snail vanishes before he reaches position 4, and in the last paragraph the point of view is that of the thrush the immanence of the crawling beast, stuck on the plane of consistency of the flower-bed, seems to have given way to the transcendence of the bird and the all-seeing eye of the zooming camera.


There is one detail about the snail we have failed to notice so far. It occurs in the third snail episode (X3 in my proposed structure):

The snail had now considered every possible method of reaching his goal without going round the dead leaf or climbing over it. Let alone the effort needed for climbing a leaf, he was doubtful whether the thin texture which vibrated with such an alarming crackle when touched even by the tip of his horn would bear his weight.10

On the face of it this is a rather touching description of a snail gingerly apprehending the world with his small horns. But I believe there is more to it than this and that our snail is an intertextual beast: `snail-horn perception' is a well-known phrase in Keatsian criticism. It comes from a famous letter to the painter Haydon: `The innumerable compositions and decompositions which take place between the intellect and its thousand materials before it arrives at that trembling delicate and snail-horn perception of Beauty `11

Let us decide therefore such coup de force is one of the privileges and pleasures of literary interpretation that this is an allusion and that ours is indeed a Keatsian snail. This changes the focus of our interpretation, which so far revolved around the revolving couples and their cosmic order. Now the focus is on the snail as poet, as creator of cosmos, the couples being mere satellites. In his progress through the flower-bed, the snail is drawing a plane over chaos, making up a world: he no longer needs to reach a teleologically fixed goal, pre-existing his endeavours, since the goal itself is the result of his progress and is not reached but rather achieved through the fusion or communion of snail and cosmos. Being the hub of the universe, our poetic snail is everywhere and no-where, he is the universe itself: his only goal, which he fully achieves, is, en abyme, the construction of the text as work of art.

There is still something slightly grandiloquent in this Romantic van-ishing of our poetic snail in the infinite vastness of the cosmos, as the camera zooms up and away from the flower-bed. True, it is a more satisfactory ending than our first, down-to-earth interpretation (the snail never reaches the other side of the flower-bed because it is eaten by the thrush). Deleuze, however, can help us keep our feet on the earth. I propose therefore not a Keatsian but a Deleuzean snail.

The text is made up of two series, interlaced in the linearity of the narrative chain: a series of couples, and a series of descriptions of the flower-bed. This is what, in the text, makes sense. But the two series, as in the Deleuzean concept of structure, must be quilted, or

Interlude 2 151

zipped together, by the circulation of a paradoxical element which, senseless itself, makes sense of the structure. The snail is that element: it is not a Bunyan snail, as its progress is meaningless and has no end; it is not a Keats snail, as the cosmos is a place of excitement and wonder, of silent communion but not strictly speaking of Beauty it is a cosmic, not an artistic cosmos. The function of the snail is to circulate, aimlessly because nonsensically, along the vertical axis of my diagram and thus make sense of the structure.

This does not mean that his role is trivial. It is, on the contrary, of the utmost importance: for the paradoxical element, by engendering sense on the margins of the structure, is also the catalyst of the events that occur in the structure. And the text is the site of at least two singular events, events that form those paroxystic `moments of being' (this is not only the title of Woolf's autobiography, this is also the name Woolf gives to the minor epiphanies that are the best illustrations I know of the Deleuzean event). The first, and clearer, is evoked in the reminiscences of the wife, in the married couple, as she tells them to her husband:

For me, a kiss. Imagine six little girls sitting before their easels twenty years ago, down by the side of a lake, painting the water-lilies, the first red water-lilies I'd ever seen. And suddenly a kiss, there on the back of my neck. And my hand shook all the afternoon so that I couldn't paint. I took out my watch and marked the hour when I would allow myself to think of the kiss for five minutes only

it was so precious the kiss of an old grey-haired woman with a wart on her nose, the mother of all my kisses all my life.12

In the intertextual doxa that evokes Monet's NympheÂas and turns the chaos of coloured impressions in the first paragraph of the text into a Chevreul-like impressionist experiment in the decomposition and recomposition of colour, there occurs something entirely unexpected, but precious not the Hollywood kiss of the doxa (`her first kiss' is a fit theme for a Mills and Boon novel), but a strangely asexual, entirely meaningless (and all the more important for this) occurrence, an acci-dent a crucial part of which resists and escapes its actualisation.

The second event comes to the courting young man, as he suddenly realises he is going to have tea with his beloved:

Oh, Heavens, what were those shapes? little white tables, and waitresses who looked first at her and then at him; and there was a bill that he would pay with a real two shilling piece, and it was real,


all real, he assured himself, fingering the coin in his pocket, real to everyone except to him and to her; even to him it began to seem real; and then but it was too exciting to stand any longer . . . 13

The event is of course present in the dash, on the last line of the passage. It is trivial or futile to anybody but the subject who experiences it, it is meaningless, endowed with both reality and a thrilling lack of reality (it has nothing to do with the Lacanian event as devastating irruption of the Real), but it endows the whole of the hero's world with an aura of sense, as in a Joycean epiphany.14

We understand why the series of couples are described in terms of speech and (non)communication. The couples who recall or undergo events also make sense: they hardly need words, but such words as they resort to have an aura of sense around them that allows real communi-

cation. Thus, the rare words of the young couple are described in the following terms:

The action and the fact that his hand rested on the top of hers expressed their feelings in a strange way, as these short insignificant words also expressed something, words with short wings for their heavy body of meaning, inadequate to carry them far and thus alighting awkwardly upon the very common objects that surrounded them, and were to their inexperienced touch so massive: but who knows (so they thought as they pressed the parasol into the earth) what precipices aren't concealed in them, or what slopes of ice don't shine in the sun on the other side?15

Their words are both ìnsignificant' and so charged with meaning that they cannot fly very far: this paradox is that of sense, before it is frozen into good and common sense, as it is in the copious but senseless speech of the elderly ladies,16 who talk exclusively in clicheÂs, and in the insane ravings of the old man, where the deeper and disturbing nonsense of primary processes makes itself heard.

The whole of Kew Gardens, therefore, as described in the last para-graph, is a surface on which events and sense circulate. The transcend-ence of the bird's-eye view is illusory: it is the false transcendence of the general who, surveying the field from the vantage point of a hill, be-lieves (not always rightly) that he is directing operations. But the battle itself, the event, is the vapour that hovers over the scene the vapour in which the couples are enveloped, which is the real object of the snail's progress, into which at the end of the story it is transmogrified:

Interlude 2 153

Thus one couple after another with much the same irregular and aimless movement passed the flower-bed, and were enveloped in layer after layer of green blue vapour, in which first the bodies had substance and a dash of colour, but later both substance and colour dissolved in the green-blue atmosphere.17

We understand the title of the story: Kew Gardens is to be taken as an analogon of the universe because it is the site where events and sense circulate. And the text is, en abyme, the site of a poetics of the event. It tells us that events are not necessarily shattering and brutal, that not all epiphanies are met on the road to Damascus. And it tells us in its own discreet manner. Only the briefest of allusions, in the old man's delir-ium, reminds us that this humming city is a city at war: `Heaven was known to the ancients as Thessaly, William, and now, with this war, the spirit matter is rolling between the hills like thunder' (p. 37). The technique is similar to that of `The Mark on the Wall', the sister story, if I may say so, to `Kew Gardens': the chute of the text interrupts the musings and reminiscences of the narratrix with two brutally uttered pieces of information that the country is at war and that the mark on the wall is not a nail but a snail.

5 Conclusion

It might appear that, as far as literary texts are concerned, a Deleuze-type analysis does not go very far, or at least far enough, towards a global and fully fledged interpretation of the text. On second thoughts, however, the fact that you cannot conduct a Deleuzean interpretation from A to Z, that Deleuze does not provide a tin-opener for the text,18 is a great advantage. We have all read interpretations inspired by strong, or vam-

piric theories, sometimes of the psychoanalytic type, where the keen reader interpreted the text, or rather cut it up, with the zeal of the forensic pathologist, at the end of which process it lay on the gurney of the page, dead as a dissected dormouse. What Deleuze allows us to do, on the other hand, is to read for sense to become aware of the faint traces the aura of sense that envelops the text disseminates in it. Such reading must be encouraged, for at least it shows respect for the text. It does not provide the answers before the text has had time to raise its questions. It allows me to potter around with the text, in the hope of gaining a few more insights, while the text laughs at my feeble attempts, and escapes the assignation of meaning that I, or my numerous succes-sors, will impose upon it.

4 Another Philosophy of Language: the New Pragmatics

1 Language reframed

I shall begin with a brief reminder of the problem, or the difficulties, raised by mainstream `scientific' linguistics, that is, post-Saussurean linguistics. As we saw in Chapters 1 and 2, the problem with this type of linguistics (which tends to comprise the whole of linguistics) is four-fold. (1) It is founded on the wrong philosophy of language (it deals with a language the main functions of which are information and communi-

cation). (2) It resorts to the wrong type of immanence and the wrong level of abstraction: it is based on the separation of language from the rest of the phenomena (let us call it `the world', for short) and on the separation within language of the relevant from the irrelevant phenom-

ena, thus using a `postulate of immanence' to construct an internal linguistics, and in so doing leaving out a remainder that only literature is able to capture. (3) It resorts to the wrong type of epistemology: it treats linguistics as a science, the task of which is to formulate the laws of language in the same way as physics formulates the laws of matter; this involves a treatment of exceptions (to the laws of language or rules of grammar) as temporary obstacles to be dissolved as scientific enlight-enment dissipates the cloud of unknowing. (4) Lastly, it seeks to defend the wrong type of hierarchy between related disciplines: it postulates the centrality of the science of language for all disciplines that study signs. Thus, Saussurean semiology is an extension of linguistics this move-

ment is characteristic of a scientific conjuncture, loosely called the `moment of structuralism'. As we have seen, Deleuze and Guattari appeal, against this centrality of linguistic structures and imperialism of the signifier, to the arch-structuralist, Hjelmslev, whom, in an inter-esting and fruitful reading, they treat, against the grain, as an exception.

154

The New Pragmatics 155

This is the linguistics that has done a lot of harm and whose best-known representative is Chomsky. The reason why the Chomskyan research programme can be treated as the epitome of mainstream lin-guistics is that it offers the considerable advantage of explicitness, and is not shy of formulating the philosophy of language on which it is based. This philosophy of language is, to be brief, characterised by a number of postulates. The first is the principle of reductionism: linguistic phenom-

ena are explained by their possible reduction to material connections in the mind-brain: that this remains a purely theoretical postulate, in no way founded on hard fact, is explained away by the present stage of the science of neurology, whose description of the workings of the human brain is, at the moment, not sufficiently detailed to make an independ-ent science of linguistics redundant. This principle is the rationale for Chomsky's notorious `return to Descartes' and advocacy of innatism. The second postulate is the principle of rule-governed creativity: there is no creativity that cannot be accounted for by a combination or compli-

cation of rules, or, rather, this is the only relevant sort of creativity. The rule-breaking form of creativity (with which literature is actively con-cerned) is thus expelled to the limbo of chance phenomena at worst, and at best to the domain of dubious unscientific disciplines like stylis-tics or rhetoric. The third postulate is the principle of methodological individualism (the concept is by no means restricted to linguistics: it is central to analytic reinterpretations of various social sciences, for in-stance the revision of Marxism attempted by Jon Elster and the school of ànalytical Marxism').1 This principle states that all mind-brains are wired in the same way, but that utterances are strictly individual: there is no such thing as a collective utterance, it is always the individual speaker that speaks, never `language' that speaks her or speaks through her. The fourth postulate is the denial of historicity: language as the object of the science of linguistics is immune to history, independent of the historical development of human society. The only èvent' that the science of linguistics needs to postulate is a mythical ahistorical event: the moment in evolution when the fully-wired human mind-brain emerged hence the already mentioned innatism. Language, being a constituent part of the hard disk of the human brain (this type of theory is fond of metaphors borrowed from computer science) is indifferent to history, out of reach of social influences: this is internal linguistics with a vengeance, that is, interpreted literally. One of the consequences, perhaps the most striking and deplorable one, of the postulate is the demotion of national languages (which are also called, in a phrase that becomes hopelessly inadequate, `natural languages') from their central


role in the discipline. There is little point, from the point of view of scientific linguistics, in constructing a grammar of German, or of Dutch (there may, of course, be a pedagogic point of great import) since such `languages', and English their cousin, are only separated by a few switches more, or a few switches less, activated in the mind-brain. Their constitution into separate entities is an epiphenomenon, of con-cern to historians of culture, not linguists. This is Saussurean synchrony, and the consequent demotion of diachrony, taken to extremes. And it also implies that there is no concept of poetry or of poetic uses or function of language, because poetic texts are not an object for the science of language, which is only concerned with the natural laws that spell out the workings of the mind-brain. Utterances only figure as basic data, in so far as they are the productions of native speakers in `normal' circumstances, that is, in anthropological-experimental sur-roundings (the model is the anthropologist eliciting the utterance `Gavaga!' from the eager native, at the sight of a rabbit but the conclusions drawn from this are vastly different from Quine's, from whom I borrow this celebrated example). From such basic data the linguist will reconstruct the laws of language (as long as direct study of the human brain is not detailed enough to yield them): language is thus thoroughly naturalised.

Deleuze and Guattari are strong opponents of every single one of the postulates that underlie the Chomskyan research programme. Rather than the Chomskyan `materialist' reductionism (the inverted commas are due to the fact that this `materialism' is largely a question of principle: a rather vague philosophical gesture), they prefer another form of materialism, the ontological mixture of event and sense, or of lines traced on the plane of immanence, distributing points of singular-ity. They celebrate the rule-breaking creativity of literary texts, and make this capacity the centre of their philosophy of language. Unimpressed by the methodological individualism of mainstream linguistics (which concerns not only Chomsky, but all speaker-based forms of linguistics, like the ènunciation' theory of Benveniste or Culioli) they take as the elementary units of the reconstruction of language not individual utter-ances nor individual speakers but collective assemblages of enunciation. Thus, they propose (a very rare and precious thing) a philosophy of language which almost entirely dispenses with a concept of subject (either the grammatical subject sujet de l'eÂnonceÂ or the utterer sujet de l'eÂnonciation) and in which the subject maintains only a vestigial presence and only plays a residual role. Being, in a bizarre and twisted way, Marxists, they reject the ahistoricity of the mainstream research


programme: they historicise not only language itself, and natural lan-guages, but also linguistics. The type of semiotics that is based on linguistics does not play a central part in their general conception of signs because it is a historical construct, linked, in their broadly Marxist periodisation not of modes of production but of regimes of signs, to the despotic State. Hence, they reject the demotion of national language to contingent stages on a linguistic gradient, as more switches are switched on. Closer to Bourdieu in this respect than to Chomsky, they make such historical constructs one of the main objects of their analysis, even if they interpret their unity (Ènglish' as a single entity) as superficial and the result of conflict, of rapport de forces: as we shall see, a national language is a conjunctural balance between major and minor variants engaged in everlasting struggle. Far from rejecting the poetic uses of language to the marginal field of stylistics, they proclaim the centrality of poetics, and of a renewed concept of style, for the study of language. Lastly, rather than naturalising language, they denaturalise it, practising external linguistics with unrestrained zest: with language, as with the rest of the phenomena, we find ourselves not bolstered up by the fixity of laws of nature, but in an ever-changing world of continuous vari-ations, of speeds, intensities and lines of flight.

As a privileged example of such thorough-going inversion (the Deleu-

zean programme may aim to overthrow Platonism in the field of phil-osophy: in the field of language, it certainly overturns the whole of linguistics) we might take the question of exceptions. In Chomsky, as we saw, exceptions are temporary, meant to be dissolved as illusions (due to the superficial complexity of phenomena) or reduced as more and more detailed laws of language are formulated. They are also a pretext for the exercise of mauvaise foi by Chomskyan linguists, who tend to make a tactical use of the asterisk that marks agrammatical utterances, or of the initial question mark that notes questionable ones, in order to protect the `laws' they have just formulated from the return of the remainder (a concept that strongly suggests that there is no `rule' of grammar which is not defeasible for expressive purposes, and actually defeated by language: hardly has the linguist formulated a rule than exceptions increase and multiply). This, of course, justifies Deleuze and Guattari's pronouncement that grammatical markers are markers of power, that judgements of grammaticality are judgements in the legal sense, that is, violent impositions.

Here is a brief example from the latest book by Andrew Radford, the main exponent of the Chomskyan research programme in Britain the textbook expounds the latest form of the model, called `minimalist


grammar' (the Chomskyan model is revolutionised on average every five years: since one can surmise that the structure of the mind-brain does not change that fast, this raises serious doubts about the `scientific' nature of the rules formulated). Radford is demonstrating that, in a grammatical structure, a moved constituent leaves an empty category trace behind it. There are reasons internal to the theory for this, and they are developed first. But there is also èmpirical evidence', a string of utterances, with liberal use of the asterisk for agrammaticality. Here is one example:

(11a) Will we have / *we've finished the rehearsal? (11b) Should I have / *I've called the police?

The asterisk supports the claim that in this sentence the auxiliary `have' cannot be reduced (unlike other sentences of the Ì've done it' type): `The sequence we've in (11a) doesn't rhyme with weave (in careful speech styles), since we have can be reduced to / wi]v/ but not /wiv/: similarly, I've does not rhyme with hive in (11b) . . .'2 I have no objection to the grammatical conclusion (that moved elements leave traces): but the èmpirical' evidence adduced is deeply flawed. First, the reduced forms are excluded as agrammatical through the use of the asterisk a blatant marker of scientific or academic power. Then their actuality in speech is conceded in a parenthesis: `(in careful speech styles)', which suggests that the mind-brain has special switches for social or dialectal registers, or that some forms of language (the standard dialect of the academic, for instance) are more attuned to the mind-brain than other, less respectable ones.

Since Deleuze and Guattari are not linguists, their position is less fraught with difficulty: they do not attempt to account for the fuzziness and incoherence of actual language. But their global account of lan-guages enables us to understand why language should be fuzzy and incoherent. Aware as they are of the ceaseless minorisation of the stand-ard dialect, they would not need such terroristic use of asterisks. In the matter of exceptions, they would be on the side of the Dutch linguist evoked by Z Ïek:Ïiz

Why are five-cent coins larger than ten-cent ones; why this exception to the general rule according to which volume follows value? Karel van het Reve, the famous Dutch linguist, literary scientist and Pop-perian criticist of psychoanalysis and deconstruction has formulated the logic of rule and its exception in the guise of what he ironically


calls `Reve's conjecture': in the domain of symbolic rules, Popper's logic of falsification has to be inverted that is to say, far from falsifying the rule, the exception one has to search for confirms it. Besides enumerating examples from a multitude of symbolic, rule-regulated activities (in chess, we have rocade as the exception, a move that violates the fundamental logic of other possible moves . . .), Reve focuses on linguistics: in grammar, a particular exception is needed in order to reveal (and thus to make us sensible to) the universal rule that we otherwise follow: À rule cannot exist if there is no exception against which it can distinguish itself.' These exceptions are usually dismissed as so-called deponentia, ìrrational' irregularities due either to the influence of some neighbouring foreign language or to remain-

ders of earlier linguistic forms.3

Deleuze and Guattari would have approved of such a positive role ascribed to exceptions, even if they might have liked to substitute a state of continuous (but systematic) variation for the notion of ùniversal rule'. And there is more to this than a modern version of the debate between anomalists and analogists in medieval grammar: what we have is two opposed philosophies of language, whose quarrel is about not only the way language changes, but the way it is constituted.

2 Which pragmatics?

Since Deleuze finds himself in total opposition to the philosophy of language which underpins mainstream linguistics, and since he is no linguist, he needs to find a vocabulary, understandable by linguists, in which to phrase the opposition. Such vocabulary he found in Anglo-Saxon pragmatics. In Dialogues, Deleuze ascribes this importation to Guattari (in Chapter 1, we saw that Guattari's L'Inconscient machinique was largely devoted to a pragmatic analysis of various regimes of signs). The claims made for what is usually considered at best a marginal sub-field of linguistics are large: `pragmatics is called to take upon itself the whole of linguistics'.4 And Deleuze, following Guattari, formulates four principles of pragmatics. (1) Pragmatics is all-important because it in-scribes the politics, even the micro-politics of language. (2) With prag-matics, there are no universals of language, no invariants and no such thing as Chomsky's competence. (3) Against the structuralist `postulate of immanence', there is no abstract machine internal to language (the Chomskyan algorithms), but only external abstract machines that in language produce assemblages (of enunciation, of desire). (4) There are


several languages in a language or tongue: every language is constitu-tively bilingual, so much so that it stutters, as a language is the site of lines of flight that disturb and dissolve its vocabulary and its syntax the name of this is the `pragmatic line', theorised under the concept of `style'.

This chapter and the next two will be devoted to making sense of this exhilarating but still vague programme. But we must note at the very start that the picture of pragmatics sketched here would be unrecognis-able to Anglo-Saxon practitioners of the subject, so that we must at-tempt to explain the misprision and trace this emphatically Continental form of pragmatics to its analytic origin.5 In the Chomskyan research programme, pragmatics is tolerated, even treated as an ally, because it is not a danger to the syntactical core of the theory. The phenomena it accounts for, the questions it addresses, are beyond the pale: the field of pragmatics is contiguous to the field of linguistics, but it definitely lies outside it, so that there is no pragmatic module, since pragmatics deals with performance, not competence. And there is even no attempt to order that performance, to draw it back into the realm of competence, as Chomsky so famously did with syntax (which for Saussure belonged to parole, not langue).

There are reasons why the Chomskyan research programme should keep pragmatics at a distance. The first is that the strategy of pragmatics is, for the Chomskyan, paradoxical, if not contradictory: it attempts to formulate rules of performance (which it prefers to call conventions or maxims rather than rules or laws), that is, it attempts to establish a form of order in the anarchy of performance, but it does so without excluding part of the phenomena. The utterances that the Chomskyan linguist excludes through the liberal use of asterisks are interpreted by pragmat-

ics as instances of flouting of the maxims, demonstrating their defeas-ibility, and frequent actual defeat. This, of course, amounts to a serious blurring of the langue/parole or competence/performance oppositions on which `scientific' linguistics is based: the problem with pragmatics is that, because of its irresponsible attitude towards exceptions, it makes possible a passage from a linguistics of rules to a poetics of agrammati-

cality and style such a passage is achieved with the new, Continental, pragmatics of Deleuze and Guattari.

The second reason is that pragmatics dilutes the principle or postulate of immanence that makes linguistics ìnternal'. By concentrating on the context of the utterance and the intentions of the utterer, pragmatics leaves the realm of the corpus of utterances and the system of langue. It wanders off into the wide world, like Alice exploring Wonderland, and


opens up the possibility of reintroducing other points of view within the study of language for the performativity of an utterance is obviously linked to cultural, legal, or political factors. There is a symptom of this in Austin's early essay on performative utterances. He gives the following imaginary example of performative infelicity: a dubious proletarian character snatches the bottle of champagne from the hands of the visiting royalty and throws it towards the hulk, while saying, loud and clear, Ì name this ship the Generalissimo Stalin'.6 This has the quaint charm of a period piece, but it shows that politics is always near at hand, lying in ambush, eager to seize the pretext for re-entering language that pragmatics is willing to offer it.

Nor is this all, for pragmatics dilutes the postulates of immanence in yet another way, and this is the third reason for keeping it at a distance: by blurring the separation between the ideal system of langue and the materiality of the world. Pragmatics is the universe of performative utterances and speech-acts. It deals with actions, with the exertion of forces over things, and even if it abstracts concepts in order to account for such an ontological mixture of words and things, they are concepts of ontological mixture, no longer at a safe remove from the world they describe, in the ghostly realm of representation and intentionality. Utterances, like events in Logique du sens, are in the midst of things. Unlike events, they are not incorporeal, they are no longer on an ideal separated plane (in Deleuze's terms, there is only one plane of imma-

nence), like Popper's third world. The last reason is that pragmatics has no use for the concept of natural

`laws' of language and replaces it with concepts borrowed from Kant's second critique, like `principles' and `maxims'. This move from the first to the second critique is not innocent: it removes the science of lan-guage from the realm of the exact sciences and makes it an exercise in `practical reason', that is, of human activity (this is the etymological sense of `pragmatics'), in other words not a `science' at all.

In short, pragmatics does provide the philosophical lexicon that Deleuze and Guattari need in their attempt to overthrow mainstream linguistics. Or, rather, it allows their specific vocabulary (of flows, of assemblages etc.) to creep in, at the cost of a distortion. We may take this insistence on pragmatics as a symptom of the problematic nature of language in the work of Deleuze. Pragmatics allows Deleuze and Guat-

tari to reframe language, to demote it from its position of centrality (in a sense, they rejoice at the marginality of pragmatics vis-aÁ-vis linguistics proper). But language thus reframed is still an object of interest, of concern, of fascination. Pragmatics does away with language as an


independent scientific object, the canonical object whence all semiotics derive. But Deleuze and Guattari's conversion to the discipline still expresses their fascination with language, at the very moment it enables them to dissolve language into thought, or into a general theory of action and thought. So there is a sense in which Deleuze hesitates, in which he not only overturns `scientific' linguistics, but takes it to its logical end, by making pragmatics the last extension of the programme initiated by Saussure, the last morsel of linguistic order to be extracted from chaos, the moment when the distinction between langue and parole is finally abolished, but only because the whole of parole is absorbed into langue. This hesitation (pragmatics is the instrument of the overthrow of linguistics; it is the ultimate form of linguistics its achievement in both senses of the term) is felt in Deleuze's attitude towards the concept of system. He is sufficiently faithful to Hjelm-

slev to keep it (by carefully distinguishing, for instance, between non-systematic variety and systematic variation), and his concept of agram-

maticality, which is central to his poetics, is still within the compass of the pragmatic flouting of a maxim, so that agrammaticality is not the symptom or the celebration of linguistic chaos.

Nevertheless Deleuze and Guattari's recourse to pragmatics is a blatant case of misprision of Anglo-Saxon pragmatics. I borrow the term from Harold Bloom:7 it has the advantage of denoting both betrayal and creative development, so that what Deleuze and Guattari give us is a new pragmatics, with a strong Continental flavour. In order to follow their own philosophical agenda, they must deeply transform a discip-line which is still dependent on the philosophy of language that under-pins the mainstream research programme. We may start by noting that the positive traits I have just indicated are mostly tendential: they are possibilities of development, which Deleuze and Guattari eagerly seize upon, but which Anglo-Saxon authors do not actualise. If the scandal-ous working-class militant in Austin's text is only a passing example of infelicity, a wink at the reader, it is because the universe of which he is the representative or the symptom (the universe of the class struggle) is absent from speech-act theory. Thus Anglo-Saxon pragmatics has four types of serious limitations for Deleuze and Guattari's new pragmatics.

The first limitation is that Anglo-Saxon pragmatics still believes in methodological individualism. This appears with conspicuous clarity in the theory of meaning associated with it, that is, Grice's theory of `non-natural' meaning as intention and recognition of intention.8 This theory requires two fully individual and conscious subjects, one to mean and one to recognise. Interlocution is a deliberate exchange (emission


and retrieval) of fully conscious meanings, each intimately associated to a speaker who intends to exchange it: meaning is what someone means. It is easy to compare this either to Wittgenstein's concept of meaning as use, within a language-game, or with traditional semantic theories of meaning as determined by the system of langue (holistic theories, for instance, as defended by Quine),9 to notice the difference: meaning for Grice is individual, neither collective nor fixed by the system of lan-guage. For a Continental philosopher or linguist, this is a philosophical regression, when compared with the achievements of structuralism.

The second limitation is a consequence of the first. A speech-act is an individual act. Even if it conforms to maxims and conventions, and can be subsumed under a type (declaratives are not verdictives10), it is not caught in a communal language-game, nor is it the product of a chain, which goes from institution to ritual, from ritual to practice, from practice to speech-act, as in Althusser's theory of interpellation.11 In the case of Austin's or Searle's theory of speech-acts, the first three links in the institution ! practice ! ritual ! speech-act chain are simply inexistent (or ignored).

The third limitation is that the principles and maxims, even if they are not laws of nature, are a ahistorical, eternally valid, and thus call for a concept of human nature, as indeed does Kant's theory of practical reason. Grice's `cooperative principle', which governs interlocution as exchange, and makes language the carrier of information and commu-

nication, is ahistorical: the maxims that derive from it also derive from the eternal necessity (or at least a necessity as old as humankind) of exchanging information. This reliance on human nature also implies a decision (on the nature of this human nature) which is debatable: exchange is not only eternal, it is also irenic. So the dubious decision to rely on a principle may be compounded by the choice of the wrong principle. Deleuze and Guattari insist on the importance of agoÃn, not irene in human dialogue. In Qu'est-ce que la philosophie?, they describe the Greek polis as constituted by the community of free men united by their rivalry (they call this `generalised athletics'). The polis is founded on the political, which is the art of debate, so that there is society because there is such a thing as politics, politics exist because man is a speaking animal (this is the famous introduction to Aristotle's Politics),

12and language is the vector or the expression of agoÃn. The fourth limitation is that in spite of its insistence on performativity

as involving action, on speech-acts as acts, Anglo-Saxon pragmatics shows no real interest in the materiality of language. The action that the speech-act accomplishes is not considered, as with the Stoics, as a


mixing of bodies, but as the source (and the result) of an interpretative calculus. We have left representation (where language, in a broad sense, `represents' the world) for interpretation: a speech-act does not stem from a material chain of institution, ritual and practice, but it is rather an intermediary between two forms of calculus, the rational calculus of intention before action, and the equally rational calculus of interpret-ation after action. Neither the materiality of collective rituals nor the materiality of words is an issue for the exertion of this strange type of force (the illocutionary force of a perfidious speech-act, depending on the reconstruction of the implicit meaning called ìmplicature', is inde-pendent of the words that convey it words which can indefinitely vary).

The direction of the misprision of Anglo-Saxon pragmatics is therefore clear. Their philosophy of language is meant to reintroduce the histor-ical context of utterances, the institutional source of the forces involved, the collectivism of rituals and practices in the course of which the speech-acts occur, the ontological mixture of bodies, utterances and practices that constitute the assemblages that are the proper objects for the study of language. We understand why Guattari claims that pragmatics is a micro-politics of language. The Gricean theory of co-operative dialogue and the calculus of implicature is ethical: Deleuze and Guattari's new pragmatics is resolutely political, more interested in Lenin's pamphlet on slogans and the Leninist practice of oratory and political speech-acts than in the irenic description of communica-

tional exchange, meant to protect language from the infelicities that threaten it.

Hence, from the critique of the postulates of linguistics in Plateau no.4 of Mille plateaux, which we analysed in Chapter 2, there emerges a cluster of positive concepts, which constitute a Deleuzean system, that is, a system of continuous variation. From the first postulate (language is meant for information and communication) emerge the positive con-cepts of force and interest, and the idea that the elementary type of utterance is the slogan, not the declarative sentence. From the second postulate (there is no abstract machine of language that is exterior to it), we derive a positive consideration of machines, assemblages and territory (deterritorialisation and lines of flight). Postulate three (language is a homogeneous system) a contrario yields the concepts of variation and stuttering, of style and abstract machines. While the last postulate (the object of linguistic science is the standard variety of language) produces the concepts of minority and, again, style. I shall devote the rest of this chapter and the next to a detailed account of these concepts. The


concept of stuttering, which lies at the centre of Deleuze's poetics, will be dealt with in the last chapter, where Deleuze's views on syntax and his poetics will be discussed.

I have, however, to note an absence in this list of concepts. The concept of sense, central to Logique du sens and to Chapter 3, appears to be missing: it appears to have been abandoned when Deleuze went from the `psychoanalytic novel' he claimed Logique du sens was to a critique of psychoanalysis. And it is true that, after Logique du sens, we find no occurrence of the concept, only occurrences of the term, in the usual acceptation of `meaning'. Thus, in Kafka, in a passage that is a commentary on the Proustian phrase Deleuze is fond of quoting `the writer must become a foreigner in his own language', Deleuze and Guattari oppose the poetic use of language, which achieves the Prous-tian ideal, and the system of langue, that blocks deterritorialisation and lines of flight, with `sens', that is, with meaning (with what in Logique du sens is called `good sense' and `common sense'): the term is still present, the concept no longer so.13 On the other hand, the concept of event, which in Logique du sens is closely associated with sense (we opposed Deleuze's `sense-event' and Badiou's `truth-event'14), is still present in the later works, and figures prominently as late as Qu'est-ce que la philo-sophie?. In L'Image mouvement, Deleuze's first book on the cinema, `sense' is absent but the same propositions are advanced on the subject of àffect' as were advanced on the subject of sense (the affect is expressed by the proposition, distinct from the state of affairs, etc.):15 under a different term, the concept surfaces again. We may take this as a symp-

tom of the shift towards another philosophy of language: the concept of `sense' was still contaminated with shades of `meaning' and ìnterpret-ation', that is, with the older philosophy of language Deleuze seeks to replace.

3 Force

In Anglo-Saxon pragmatics, the consideration of force (as we have seen, a distinctive feature of pragmatics) came from a keen awareness of the workings of ordinary language. It stemmed from the realisation that, in our most ordinary uses of language, we do not simply convey infor-mation, but exert force, we do things with words. The concept of force, however, remains unanalysed, left to our common understanding. Not so in Deleuze, where force is a philosophical concept, with a tradition and a history. And it is, of course, borrowed from Nietzsche. Deleuze shared Foucault's view of the central importance of Nietzsche. He


devoted two books to him, one of which is a presentation for students,16

the other a much more ambitious study, where Deleuze treats Nietzsche as he later treated Spinoza or Leibniz, by identifying a problem in his work and concentrating on it a problem named by the concept `force'.17

The world as Deleuze describes it in that book is made up of forces and relations between forces. In other words, the bricks that particular cosmos is built with are forces, as they are bodies in Logique du sens (which follows the Stoics) and images in CineÂma 1 (which follows Berg-son). Thus, the objects that make up the world are expressions of forces, the subjects that interact with them are `differential elements of forces'.18 The subject's body is the site of a rapport de forces, where dominant and dominated, that is, active and reactive forces clash. Nietzsche's will-to-power is an internalisation of forces, a principle for the synthesis of forces. And the essence of a thing is the force (or forces) that possesses it: the aim of Nietzsche's philosophical question is the discovery of this essence, of this force.

Although language is not a theme or an object of the book, this con-struction of a concept of force entails consequences for language. The higher form of language (language `poetically inhabited') is the site for the interpretation that assesses and measures the rapport de forces that defines the subject.19 The aphorism, typical of Nietzsche's philosophical style, is the instrument of such interpretation, and poetry is the art of evaluation, of the determination of forces and their values: the task of interpretation is to evaluate forces and relations between forces. We do not wonder that Nietzsche began his career as a philologist: Deleuze calls him àn active philologist'.20 This àctive philology' involves a certain conception of language as not merely a conveyor of force, but also a locus for the clash of forces. The speaker does not exchange information: she enacts language through the imposition of forces. This entails three consequences. (1) Language is the art of authorised, and authoritarian naming. (2) A word means what the speaker forces it to mean (we might call this the Humpty-Dumpty principle, as it is Humpty-Dumpty's con-viction that `when I use a word . . . it means just what I choose it to mean, neither more nor less'21); more generally, meaning (the meaning of a proposition or an utterance) is the embodiment or inscription of a rapport de forces. (3) Consequently, speech gives rise to real acts, acts within the real, that is, the world outside the subject.

There is a fourth consequence that does not concern only language but that can be generalised to art: art is the acme of the `power of the false' (la puissance du faux). It magnifies the world by the power of error,


it invents lies, creates figments of the imagination which, far from being ghostly and tenuous phantasms, exert affirmative force. Artists discover new truths in the construction of error and thus invent new possibilities of life.22 We remember Nietzsche's celebrated concept of truth as suc-cessful metaphor, that is to say, as successful lie. Such a conception not only makes language a locus for the clash and also the exertion of forces, but it ascribes to it the highest creative power. The link between lan-guage and art, which in Deleuze's new pragmatics is marked by the centrality of style and poetic language, is already apparent at this early stage of Deleuze's career (Nietzsche et la philosophie is his second book).

What occurs here is a shift from meaning to force as the centre of linguistic inquiry. The traditional programme in the philosophy of language, beginning with Aristotle (how do we proceed from sound to sense, from phone to logos?) is discarded and replaced. Another question emerges in its stead: what are the forces that are enacted and interpreted in and by language? A new philosophy of language is sketched, and three of its features become essential parts of Deleuze's new pragmatics.

The first feature is the shift from rule-governed to rule-breaking cre-ativity, a characteristic of poetic language, as central to the study of language. There is no democracy of registers for Deleuze as there is for the empirical linguist. Shakespeare's idiom is more important to us than the picturesque dialect of Mrs Mop. So the study of language is insepar-able from the study of art. The canonical speaker is not the indetermin-

ate angel of Milner's axiom (we met him in Chapter 2), but the poet, she who lets herself be possessed by lalangue or the remainder. Literature is the site where the workings of language are at their most subtle, but also at their most explicit: it is the site where the forces of language are enacted, by which we mean not only the forces expressed in or incar-nated by language (the illocutionary force of Anglo-Saxon pragmatics), but, reflexively, the forces that constitute language. In literature those forces are given free play, not frozen into meaning. The embodiment of the freezing of meaning is clicheÂ (the mark of good or common sense): it is the strongest carrier of forces, because it is also the greatest stifler of the forces of language.

The second feature is that the study of language is inseparable from the materiality of the body. There is a constitutive corporeality of lan-guage as forces come from and act on bodies. Thus, language issues forth from the body and exerts pressure on it: it expresses and embodies active forces, bears witness to reactive forces. This amounts to taking Anglo-Saxon pragmatics to the logical extreme it carefully fails to reach: if both object and subject are to be analysed in terms of forces, language is a


medium in which force as material impulsion is physically `communi-

cated' in the French sense of the term. It is also the medium in which such force is expressed, that is, assessed and measured so that the result is a differential relation of forces, a rapport de forces.

The third feature concerns the second type of materiality, the materi-

ality of the social, of institutions, rituals, practices and the speech-acts that are produced by this sequence of material elements. If language is the site not only for corporeal, but also for social forces, the study of language is inseparable from a politics of language . We understand the rationale behind the repeated anti-Chomskyan statement that gram-

matical markers are markers of power. Here is a typical passage from Mille plateaux:

Pragmatics should reject the idea of an invariant immune from trans-formation, even if it is the invariant of dominant `grammaticality'. For language is a political affair before it is an affair for linguistics; even the evaluation of degrees of grammaticality is a political

23matter.

One finds similar sentiments in Foucault, where the origin of speech is not in the subject-author but in what Deleuze calls `foyers de pouvoir' (`focal points of power')24, the òn parle' of the anonymous speaker in which forces are expressed.

The best expression of this philosophy of linguistic forces is to be found in Pourparlers, where Deleuze is discussing the films of Godard:

Language is presented to us as basically informative, and information as basically an exchange. Once again, information is measured in abstract units. But it's doubtful whether the schoolmistress, explain-ing how something works or teaching spelling, is transmitting infor-mation. She's intruding, she's really delivering precepts. And children are supplied with syntax like workers being given tools . . . We ought in fact to invert the scheme of information theory. The theory as-sumes a theoretical maximum of information, with pure noise, inter-ference, at the other extreme; and in between, there's redundancy, which reduces the information but allows it to overcome noise. But we should actually start with redundancy as the transmission and relaying of orders or instructions; next, there's information always the minimum needed for the satisfactory reception of orders; then what? Well, then there's something like silence, or like stammering, or screaming, something slipping through underneath the redundan-


cies and information, letting language slip through, and making itself heard, in spite of everything . . . So how can we manage to speak without giving orders, without claiming to represent something or someone, how can we get people without the right to speak, to speak; and how can we restore to sounds their part in the struggle against power? I suppose that's what it means to be like a foreigner in one's own language, to trace a sort of line of flight for words.25

In this passage, it is clear that communication is not meant as the transmission of information but as the impulsion or the imposition of force. To communicate is not to cooperate but to claim and ascribe places in a power game, an agonistic exchange. And the danger language presents is when it is content to represent in the two main senses of the term (the proposition `represents' the state of affairs, the delegate `rep-resents' his members, vis-aÁ-vis whom he is in a position of power). Also, we note that poetic language, language qua art, a mixture of stammering and silence, is the solution to the representational danger language contains: in a gesture characteristic of high modernist aesthetics, art is seen as the antidote of representation.

The philosophy of linguistic force is at its most explicit in Mille plateaux. It accounts for Deleuze's rejection of the Althusserian concept of ideology (which belongs to the realm of representation, since it is the expression of contents, rather than a relation between forces, as is found in the ontologically heterogeneous assemblage). It accounts for the demotion of the concept of subject, no longer the origin of speech, but the end-product of an assemblage of forces, when the lines of flight freeze into meaning (this concept of subject, incidentally, is not unlike the Althusserian concept of subject as product of interpellation).

But the most forceful, and notorious, expression of this philosophy of forces is the question of slogans, of mots d'ordre. For the mot d'ordre is indeed a slogan not an òrder-word'. The bizarre coinage has the advan-tage of keeping the word òrder', but it has two disadvantages: unlike the French word, it does not belong to everyday parlance; and it individual-ises the process, as orders are given by one individual to another or a group of others, but mots d'ordre, that is, slogans, are issued by a collect-ive (a political party, for instance) and followed, not obeyed. Political action consists in the issuing of mots d'ordre, and in the motions neces-sary to ensure that the slogan is followed by the masses and is, therefore, just. This is why Deleuze and Guattari's mention of Lenin on slogans is justified and important.26 In July 1917, while living clandestinely in a hut, to escape being arrested by the provisional government, Lenin


found time to write a short pamphlet on slogans and their function.27

What was at stake was nothing less than the success or failure of the revolution, in the form of the question: is the slogan Àll power to the Soviets!' still correct or just? Obviously not a question of truth (a slogan is not a proposition of science, which purports to be true, in so far as it expresses a meaning that is uniquely adequate and stable), but of just-ness, that is, the capacity of the slogan to be adapted to the conjuncture, to exert force in and on it. This is why Lenin keeps insisting that his object is `the concrete analysis of a concrete situation', that is, the capacity to capture the novelty of the event. The word `truth' does occur in the pamphlet (the masses must be told the truth, that is the reality of the class position of those in power, of the enemies of the revolution), but as second and subordinate to `justness'. Truth, in Nietz-

schean fashion, is the effect produced on the masses by the justness of the slogan that not only enlightens them as to their interests, but moves them to revolutionary action. Illocutionary justness produces perlocu-tionary truth as its effect. Deleuze and Guattari comment on this pamphlet in terms of the relationship between language and politics: politics works from within language or, which amounts to the same, language is worked through by politics.

And what exactly is a slogan? For Deleuze and Guattari, it is the elementary unit of utterance in this they markedly differ from main-

stream linguists, for whom the elementary unit is the declarative sen-tence that embodies a proposition or judgement. There are, within the broad ambit of Chomskyan linguistics, theories that are close to this: J. R. Ross, an early disciple of Chomsky, proposed the performative hypo-thesis, whereby every declarative sentence was, in deep structure, subor-dinate to a performative clause, which being performative, is closer to a slogan (`the cat is on the mat' originates in a sentence of the type Ì state that the cat is on the mat', where the governing performative verb is a marker of force);28Ann Banfield sought to add to the first node of Chomsky's rewriting rules a special node (E for expression, as opposed to S for sentence) which was meant to accommodate all èxpressive' utterances, insults, exclamations, perhaps even slogans;29 and J. C. Milner sought to develop what he called a `grammar of insults'.30 All three attempts share the same limitations: the analysis is still based on a vertical, depth-to-surface, hierarchy. Deleuze and Guattari's analysis is, on this point, closer to French enunciation linguistics: the speaker never speaks directly, she speaks in/is spoken by indirect discourse. Here, they draw on the concept of polyphony in the work of Oswald Ducrot, and on its origins in the theory of free indirect speech, in Voloshinov's


Marxism and the Philosophy of Language,31 which means that the speaker is always spoken by the collective slogan. We understand why there is no such thing as an individual speaker (a subject) or an individual utterance: the utterance is always already secondary, always already indirect because it is the output of a collective assemblage (conceived under the metaphor of the machine, hence the word òutput' except, of course, that the assemblage is not a metaphor of a machine, it is a machine). We also understand the immanent relations between utter-ance and act. There is no transcendent intention of meaning that pre-cedes the utterance, nor is the utterance transcendent to an act, which it represents: it is an act, on the same plane of immanence as other acts, of a more immediately recognisable material nature. This entails three consequences, none of which is indifferent for a student of language: language is not a code (transcendent to its realisations); consequently, there is no distinction between langue and parole; nor is there any separation between syntax and pragmatics, or rather syntax (the arrangement of markers of power) is absorbed into pragmat-

ics. A page in Mille plateaux spells out the characteristics of the slogan:

When we ask what faculty is specific to the slogan, we must indeed attribute to it some strange characteristics: a kind of instantaneous-ness in the emission, perception and transmission of the slogans; a wide variability, and a power of forgetting permitting one to feel absolved of the slogans one has followed and then abandoned to welcome others; a properly ideal or ghostly capacity for the appre-hension of incorporeal transformations; an aptitude for grasping language as an immense indirect discourse.32

The language is not entirely perspicuous, but those four characteristics can be translated into terms more immediately relevant to the philoso-phy of language. Thus, slogans are ìnstantaneous' in so far as their force is exerted in a flash there is illumination, rich in perlocutionary effect in the utterance of the right slogan in the right conjuncture. The `vari-ability' of slogans, their `power of forgetting', and of being forgotten, precisely alludes to the fact that they are issued in a given conjuncture: a slogan, like Àll power to the Soviets!' is just at a given moment, and best forgotten the next moment. Deleuze and Guattari make much of the fact that, for Lenin, the slogan was just from 27 February to 4 July. Exactly on 4 July, they insist, the power of the Soviets came to an end, and a military insurrection became the order of the day. Hence the


power of ìncorporeal transformation' that slogans exert: like the event in Logique du sens, which is incorporeal, a mist hovering over the battle-field, and yet which operates a transformation in the state of affairs in which it occurs. Naming the conjuncture with the slogan that, in the conjuncture, is just, is no metaphor it is an action which is the source of a revolution and engages a metamorphosis. Lastly, this performative function of slogans, which means they are not restricted to the field of politics and historical events, but are to be found everywhere in daily life (where slogans are often present in the elementary or degraded form of orders), enables us to conceive discourse as always already indirect (ìndirect discourse is the presence of a slogan within the word').33

Language is made up of sedimented slogans, which is obvious in the case of political language (the various acceptations of `democracy' bear witness to the history of the class struggle and the slogans that em-

bodied the various moments of the struggle hence the polyvocality of the word).

There is a considerable amount in this that derives from Voloshinov. In his work, Deleuze and Guattari found a critique of Saussure's àbstract objectivism', and, in the sixth chapter, the five following theses, which sum up Voloshinov's philosophy of language, and ring a familiar bell:

1 Langue qua system is merely an abstraction that fails to account for the concrete reality of language.

2 Language is a process of continuous evolution which is actualised in the verbal interaction of speakers.

3 The laws of linguistic evolution are sociological, not psychological laws: they do not take individuals as their object.

4 The creativity of language cannot be grasped independently of the ideological contents and values that are involved in it.

5 The structure of verbal interaction is a strictly verbal structure.

This is a programme for an external linguistics. Not only is language social-historical, not only is the speaker always already a collective speaker, not only is an utterance always an instance of indirect discourse, but consciousness itself is outside the subject, in the verbal interaction between speakers. I find the conclusion that the inside is outside, that consciousness is nothing but its actualisation in speech, gesture and intonation (Voloshinov-Bakhtin is one of the rare philosophers who have shown interest in the question of intonation), nothing but a `structured material expression' and as such àn immense social force', highly stimulating.


4 A provisional conclusion: `je t'aime'

Before I deal with the main Deleuzean concepts of machine, assemblage and style, I shall stop to provide an illustration, the utterance-cum-

slogan, `je t'aime', which Deleuze and Guattari evoke in Plateau no.5 of Mille plateaux. The plateau is devoted to `several regimes of signs' and they conduct their analysis of the utterance according to the various regimes of signs they identify. Thus, there is a pre-signifying Ì love you', which is collective, as à dance weds all the women of the tribe'; a counter-signifying Ì love you' that is polemical, and has to do with war and rapports de forces; a signifying Ì love you' that invites interpret-ation and uses it `to make a whole series of signifieds correspond to the signifying chain'; and, lastly, a post-signifying Ì love you', which is an expression of passion, and moves from one point of subjectification to another.34

If we concentrate on the third of these Ì love you', uttered within the signifying semiotics that is based on the centrality of language, the fixation of meaning and the production of speaker-subjects, we can conduct an analysis of the utterance qua slogan, for it is clear that the utterance exerts force, seeks to claim a place, for the addressor, and to ascribe another to the addressee, so that the utterance must be under-stood not as a statement but as a demand (for reciprocity) and a declar-ation (in the sense one speaks of a declaration of war). The dialogue in a celebrated French song by Serge Gainsbourg, `je t'aime' `moi non plus' (Ì love you' `neither do I') provides a good illustration of this point. But we will also conduct an analysis in terms of indirect discourse, because the utterance presupposes, and can even be said to quote, the collective conventions of such an encounter, so that Romeo and Juliet are always already involved in the most common and trivial of utterances.

I have outlined, in terms that are still general, but will become more precise in the next chapters, the new pragmatics Deleuze and Guattari propose. A short passage in Pourparlers provides the summary we need at this point. Speaking of Mille plateaux and his collaboration with Guat-

tari, Deleuze says:

We're just addressing a number of points that we ourselves consider fundamental: first, the part played in language by precepts; second, the importance of indirect discourse (and the recognition of meta-

phor as something that just confuses matters and has no real import-

ance); third, a criticism of linguistic constants, and even linguistic variables, that emphasizes ranges of continuous variation.35

5 Another Philosophy of Language:Machines, Assemblages, Minority

1 Introduction

Chapter 4 identified the main aspect of the new pragmatics in a shift from meaning to force. This happens in two stages.

The first stage is straightforwardly Nietzschean. It consists in the operation of what I have called the Humpty-Dumpty principle, a principle that must be generalised in order to divest it of individual aspects, individual to the point of megalomania a pride that as we know comes before a fall. Meaning may not strictly be what the speaker decides, because the speaker is not alone. But it is certainly situated in a differential relation between forces, a rapport de forces. The immedi-

ate consequence of this is a certain instability of meaning: rapports de forces are inherently unstable, they change as the forces that make them up overcome other forces or give in to them. We have defini-tively left the problematic of representation, which presupposes a cer-tain stability in the relation between representative and represented, a stability that meaning is supposed to capture. We are now in the realm of assessment, or valuation of a situation in a process of constant change.

Hence the second stage, where meaning dissolves, where it becomes so dependent on a definitely deciduous conjuncture that it is lost on the one hand in the forces that constantly struggle and on the other hand in the interpretation that strives to keep up with the changes and must be constantly altered and overturned. This is where Deleuze's Nietzschean position acquires Marxist, even Leninist, overtones. The Leninist `just' or correct slogan is, as we saw, an incorporeal transformation: it is just today, but wrong tomorrow, as on 4 July 1917 the revolutionary slogan Àll power to the Soviets!' (a fine figure of a slogan, quite capable of

174

Machines, Assemblages, Minority 175

capturing the imagination of the Romantic revolutionary) becomes counter-revolutionary. What is being dissolved in this Deleuzean move is a central tenet of the classic philosophy of language: that the study of language aims at providing a reasoned account of what and how utter-ances mean. But Deleuze is not interested in this: his philosophy of language, like his general philosophy, is bent on capturing the new, the creative, the innovative, all of which are other names for the event. In that context, as we saw in Chapter 3, a sharp distinction must be made between sense (which is one aspect of the event) and meaning, or good or common sense, into which sense freezes as the vivid colours of the event pale. This is, of course, a modernist stance: it erects silence as the unacknowledged climax of language not the silence of idiocy, but the pregnant silence of ultra-sense and the dissol-ution of fixed meaning, explicitly celebrated, for instance, by Bataille, and analysed by Derrida in his essay on Bataille.1

There is a tension here in Deleuze's conception of language, a tension that animates the new pragmatics and informs his poetics. It is the tension between a form of modernism that finds its origin in Nietzsche, and another attitude towards language, shaped by Guattari's more ex-plicit Marxism. The cluster of concepts that constitute the new prag-matics (interest, machines, assemblages) reflects this tension. This chapter will explore them; the next will seek to account for Deleuze's dual, and potentially contradictory, poetics, in terms of style and stut-tering.

2 Interest

Deleuze and Guattari's insistence on the role of slogans as the basic element of utterances, and on the importance of indirect speech, derives from Voloshinov. Yet, there is a point on which they sharply differ from him: the role of interlocution. Voloshinov, whose main interest is in dialogue, still wants to preserve the structure of communication, as formulated, for instance, in Milner's axioms. The consequence of this is that his concept of language is open to idealist revisions of the type of Habermas's generalised pragmatics. But Deleuze has strong reservations about dialogic communication. And his hostility to the media as the site of universal, and compulsory, communication is a constant in his inter-views (most noticeably in the AbeÂceÂdaire, where he no longer tries to be politically correct) this is a trait he has in common with Bourdieu, and also with Chomsky.2 Here is a typical passage, from a chapter in Pour-parlers aptly entitled `Mediators':


We sometimes go on as though people can't express themselves. In fact they're always expressing themselves. The sorriest couples are those where the woman can't be preoccupied or tired without the man saying `What's wrong? Say something . . . ,' or the man without the woman saying . . . , and so on. Radio and television have spread this spirit everywhere, and we're riddled with pointless talk, insane quantities of words and images. Stupidity's never blind or mute. So it's not a problem of getting people to express themselves but of providing little gaps of solitude and silence in which they might eventually find something to say. Regressive forces don't stop people expressing themselves, but rather force them to express themselves. What a relief to have nothing to say, the right to say nothing, because only then is there a chance of framing the rare, and ever rarer, thing that might be worth saying. What we're plagued by these days isn't any blocking of communication, but pointless statements. But what we call the meaning of a statement is its point. That's the only definition of meaning, and it comes to the same thing as a state-ment's novelty. You can listen to people for hours, but what's the point? . . . That's why arguments are such a strain, why there's never any point arguing.3

On the face of it, what we have here is a classic definition of meaning. The meaning of the utterance is its point, which makes it relate to the world, meant in the broadest possible sense: an utterance has a point if it succeeds in stating something about something else. We are back to the concept of meaning the Humpty-Dumpty principle sought to avoid. But the translation is misleading. The French text has ìnteÂreÃt', a term with connotations rather different from those of `point'. `Pointless' is `sans inteÂreÃt'. But an utterance that is sans inteÂreÃt is not pointless: it usually has a point, only it is too obvious, and therefore trivial. So we have another definition of meaning, but one by no means incompatible with the def-inition of meaning as rapport de forces. What the exertion of a force in language does is to create interest, to provoke an event whose novelty is interesting. Expressing and capturing forces makes the proposition interesting, and as such meaningful, provided of course that they are forces that are still alive, and not frozen into ready-made meaning, preÃt aÁ penser. The interest lies not in the fact that the utterance has a point, but that the point is kept pointed, and sharp.

So this ranting against the media goes way beyond the intellectual's (even the left-wing intellectual's) irritation with media constitutively dominated by either morons or rightist fiends, whose only interest is


the expression and diffusion of an (always reactionary) good sense that in turn creates an audience, with its common sense. What we have here is a philosophical position where interest seeks to replace truth. The meaning of a proposition is not linked to its capacity to state the truth (which does not preclude the possibility of error or lying), but with its novelty or interest. We have gone beyond even Lenin, whose slogans had to be just, that is, adapted to the moment in the conjuncture that they correctly captured (so that even if justness logically and chrono-logically preceded truth, an effect of truth was produced by the just slogan in the masses it set into motion): interest, a Nietzschean concept, is all. This is explicitly stated in Qu'est-ce que la philosophie?:

Philosophy does not consist in knowing and is not inspired by truth. Rather, it is categories like Interesting, Remarkable, or Important that determine success or failure.4

There is, in Deleuze, a consistent description of the task of the philoso-pher as making sense of the world not by seeking truth but by cultivating novelty and interest. It is not even a question of producing propositions that are just (in the sense I defended the idea, in Interpretation as Pragmat-

ics, that no interpretation is true, but some are false, and some are just):5

what counts is the encounter with novelty, the capacity to find new ideas, not to recognise old ones and judge their aptness (the just is uttered by a judge Deleuze's hostility to Kant, one of his main philosophical oppon-ents, is due to Kant's propensity to issue judgements). Rather than recog-nition of a fit between judgement and fact, the task of the philosopher is to seek lines of conjunction (hence the importance of what Deleuze calls the logic of ET, AND, as opposed to the logic of EST, the logic of attribu-tion and judgement), to try and capture the effect of lightning that differences between forces create. So the philosopher explores lines of flight, she creates new and unforeseen syntheses in other words she creates concepts, each concept being a singularity, ordering around itself a neighbourhood, a territory. Where ̀ communication' works on the level of opinion to create consensus, philosophy, its opposite, partakes of the `general athleticism' that is Greek agoÃn (as we saw in the last chapter, the origin of both democracy and philosophy).

This concerns philosophy, and apparently not the generality of mean-

ing. We may accept that a philosophical concept ought to be creative, without accepting that meaning is determined by interest. But Deleuze is in fact making a distinction here, or rather forcing a separation, between meaning as doxa, which has point but no interest, and meaning


as sense, as force, which has interest and may be without point (extreme novelty often appears to be absurd). What Deleuze is separating goes in fact far beyond philosophical propositions or even conceptions of meaning. He is separating two ìmages of thought', two ways of envis-aging thought, and the thinker's reaction to the world: the doxic image, and the interesting image, which seeks to overturn the philosophical tradition (we may call this by the generic term `Platonism') and establish a new rapport de forces.

The two images of thought are contrasted in Dialogues.6 The doxic image (sometimes called `dogmatic') has the following characteristics. (1) The philosopher's good nature or good will impels her to seek truth. (2) Common sense is praised as the harmony of faculties in the thinking subject (this is derived from the Aristotelian concept of common sense as the sense that unifies the five other senses). (3) The task of the philosopher is to reach knowledge through recognition (this time Plato is the obvious source). (4) Conversely, truth is the determinant criterion of knowledge: that which guarantees the correctness of solu-tions to problems, which ensures that answers correspond to questions. On the face of it, this appears to describe the common sense of the philosophical tradition: for who would willingly forbear to seek truth and struggle against error? A Nietzschean Deleuze does, when he de-scribes an image of thought based not on truth but on interest.

Such an image of thought, a counter-image, une image contre, has the following characteristics. (1) It is violence, not good will, that induces thought. We rarely think, Deleuze claims, and only do so when we are violently jogged into it (we understand why academic argument is of little help). (2) What is to be sought is not common sense, not the concord of faculties, but their discord, the source of the violence that jogs us into thought. (3) Knowledge is not the result of recognition (there is little interest in knowing again what we somehow always-

already knew: reminiscence as a criterion of knowledge is self-defeating) but an encounter: knowledge is produced by an event in thought. (4) The evil that plagues thought is not so much error (which is a by-product of necessary experiment and may be fruitful an error may jog us into creative, and productive thought) as stupidity, which is inherently conservative and blocks all novelty. (5) Rather than solu-tions, or questions and answers, we need problems and concepts that name them (this point is by now familiar).

It is now clear that the concept of interest has a Nietzschean origin. Deleuze shares the interpretation of the Nietzschean critique of truth he finds in Foucault:


The fundamental link between Foucault and Nietzsche is a criticism of truth, framed by asking what `will' to truth is implied by a `true' discourse, a will the discourse can only conceal. Truth, in other words, doesn't imply some method for discerning it but procedure, proceedings, and processes for willing it. We always get the truth we deserve.7

So `truth' is linked to techniques of knowledge, power and subjectiva-tion. It is an end-of-chain effect, the result of proceedings initiated by interest, or `will'. This, incidentally, is why the term is still metonymic-

ally used by Deleuze to name the whole of the chain: truth may be used as a positive term (there are truths), but only in so far as it creates and destroys other, established, truths. In that context, truth, under whose name the concept of interest is perceptible, partakes of the power of the false: `There's no truth that doesn't ``falsify'' established ideas. To say that ``truth is created'' implies that the production of truth involves a series of operations that amount to working on a material strictly speaking, a series of falsifications.'8 Let us not mistake this for a Popper-ian manifesto: error here is not what discarded truths are reduced to, but an entirely positive part of the creative process, which is dominated by `will', that is, force and interest.

Such a concept of truth-as-interest of course entails consequences for a conception of language. The page in Qu'est-ce que la philosophie? that evokes the `modern' image of thought in terms of a Nietzschean shift from will to truth to creation illustrates the new mode of thought by mentioning Kleist and Artaud, that is, the point where `thought as such begins to exhibit snarls, squeals, stammers; [where] it talks in tongues and screams, which leads it to create, or to try to'.9 Creativity is the characteristic of interesting thought: making thought stammer, that is, stammering or stuttering language (in the active and the passive senses of the phrase), in fact enables the thinker to move to the modern image of thought. It appears, therefore, that interest as a criterion of thought is linked to the violence of language: the stammering of adoxic because agrammatical language jogs us into thought. We understand Deleuze's hostility to interpretation, in his post-psychoanalytic (that is post-Logique du sens) period. In his acceptation of the term, interpretation aims at truth, it seeks to answer the question, to suggest a solution to the problem. (There is, of course, no need to limit ourselves to this accepta-tion of the term, to treat psychoanalytic interpretation as the model of all interpretation: I have attempted to do just this in Interpretation as Pragmatics). This is why interest is more important than truth: language,


grasped from the point of view of interest, language that is interesting, is so in so far as it is creative. Poetic language is interesting language (such language is by no means limited to literature, but thematised in it). Hence the nexus, typical of Deleuze's poetics, of agrammaticality and style: there is more to this than the high modernist striving towards silence. The ultimate consequence of this shift from truth to interest is that the object of our enquiry into the workings of language is to construct not an ethics of communication, but a poetics of style.

3 Machines

Deleuze's movement away from mainstream linguistics involves a shift from structuralism, from the intellectual universe of À quoi reconnaõÃt-on le structuralisme?' and Logique du sens. The main symptom of this is the replacement of structure with machine as the main operative con-cept. The influence of Guattari is, once again, decisive.

The first collection of essays by Guattari, Psychanalyse et transversaliteÂ10

has a preface by Deleuze, and an essay entitled `Machine et structure'. The essay was written in 1969, at the time when Guattari had just begun his collaboration with Deleuze, and L'Anti-êdipe had not yet been pub-lished. It was written for `L'Ecole freudienne de Paris', the Lacanian school (which suggests it was not accepted by it this is the time of Guattari's break with Lacan, his first master), and initially appeared in Change, the journal founded by Jean-Pierre Faye. The object of the essay is to contrast the two concepts from the point of view of the production of subjects: the interlocking of machine and structure is said to produce `positions of subjectivity'.

The text has two main theoretical references: Lacan's `subject of the unconscious' (the explicit object of the essay is to support and explicate this concept), and Deleuze's concept of structure, as expounded in the essay on structuralism mentioned above. In order to make place for a concept of machine, which at that time is not part of Deleuze's concep-tual apparatus, Guattari revises Deleuze's definition of a structure, by keeping the first two characteristics (a structure is composed of two heterogeneous series; its terms are defined by their relations they have only differential value) and redeploying the third, which now concerns machines (there is a paradoxical element circulating along the series and articulating them).

The contrast between the two concepts is systematic, and takes the form of what we have come to know as a Deleuzean correlation. A structure is a series of positions (it is a spatial form of organisation);


it has elements; it is itself an element in superordinate structures; and it produces a subject. In other words, a structure is synchronic and static. A machine, on the other hand, is dynamic and diachronic. It is a temporal form of organisation (a machine needs time in order to work); it is characterised not by its elements, or parts, but by the events (called `breaks') of which it is the site; it is a site, therefore, for the emergence of events, not a piece embedded in a larger organism; and it does not produce subjects in a machine, `the subject is always elsewhere'.11 But although they are in sharp contrast, machines and structures are interde-pendent. One has the impression that a synchronic analysis yields struc-tures (where machines are positioned one in relation to the other to form a complex organism, as in a Heath Robinson drawing), and that a dia-chronic analysis yields machines, flows and ruptures, a machine being the negation of its predecessor and being in turn negated by its successor (imagine a complex Heath Robinson machine beginning to creak and clank, and starting to work). If the moment of structure dominates, what we have, to assume responsibility of the structure and its organism, is a subject, an ego. If the moment of the machine is dominant, Guattari claims, what we have is the subject of the unconscious. Lacan is the thinker of the human psyche as a machine.

The contrast entails two types of consequences. Concerning language, it is inscribed in another contrast, between the structure of langue (this is only to be expected) and the machine of `voice', the speech-machine, a term which Guattari does not explicate in the essay, but whose choice we may, with hindsight, understand: the voice is linked to the expres-sion of desire, to the corporeality of the body as machine. Human beings, the essay adds, are caught in the interlocking of machine and structure, as the voice `cuts through' the structural order of langue and founds it the machine, therefore, comes first. The speaker is the result of this interlocking as signifying sequences criss-cross through the indi-vidual, whom they both constitute and tear up (we recognise here the Lacanian theme of the inscription of the letter).

The second consequence is political. When machines overcome struc-tures, we have a revolutionary situation. When structures stifle the dynamics of machines, we have a state of oppression. This aspect of the essay Deleuze comments upon in his preface, an unusually politic-ally inclined text, where he uses Guattari's concepts in an attempt to define the `Leninist break', the active transformation of the economic, social and political dissolution of the Tsarist state into a victory for the masses. When social and political machines are on the move, the revo-lution can occur; when they freeze back into political structures, be it


the structure of the Bolshevik party (which reproduces the structure of the bourgeois state apparatus it claims to have smashed), what we have is political terror, a social freeze, and the eventual collapse of the Soviet state (which Deleuze could hardly conceive at the time, but which he makes intelligible). The contrast here is described as a contrast between two types of group, the groupe assujetti (subjected group), which is prepared to do anything in order to ensure its survival, and the groupe sujet (subject group, in the active sense of the term), which is always working towards its own demise. Stalinism transformed the Bolshevik `subject group' into a groupe assujetti. And it is obvious that this political analysis is also a form of polemics within French Marxism in the 1960s that the object of Deleuze's attack is the quasi-structuralist Marxism of Althusser.

The first sketch of a concept of machine is still developed within a version of psychoanalysis, still in terms of subjects (even if they are `group subjects'). But it already provides a rationale for the break with structuralism, and for the link between language theory and politics, which is at the heart of the new pragmatics. The concept of machine as used in L'Anti-êdipe, a central concept in the book, as our commentary on the first page showed, is already present, in nuce, in the essay.

In order to grasp what the Deleuze and Guattari machine is (as op-posed to the early Guattari machine), we may start from an apparently paradoxical passage in Dialogues, where the machine is said to be neither mechanical nor organic.12 That a machine is not organic, we may easily conceive. But how can it not be `mechanical'? Deleuze defines a mech-

anism as a relation of dependency, each cog being dependent on the next; whereas the machine is a `neighbourhood' of heterogeneous inde-pendent terms, that need not be contiguous. In other words, a machine is a singularity, defined, as we remember, as that which organises a neighbourhood (`voisinage', a term borrowed from topology) around itself. And a machine is an elementary form of assemblage, a `neigh-bourhood set', gathering together man and beast, tool and thing, whereas the mechanic that operates the mechanism is exterior to it: he is an integral part of the assemblage of the machine. The canonical example of a Deleuze and Guattari machine is Kafka's celebrated execu-tion machine (a fact which did not escape our authors), or the `machine' depicted in Heath Robinson's `Frantic Efforts to Amuse Himself and Relieve his Loneliness by the Last Man Left at the Seaside', where a manically grinning character is splashing in the water, surrounded by a group of ill-disguised tailor dummies, while other dummies recline on deck-chairs.13


L'Anti-êdipe provides more precise descriptions of what a machine is. It is essentially characterised by a flow of energy and a series of cuts, of breaks, of ruptures that give the flow form by coding it. Thus, the best example of a linguistic machine is a Markov chain (but with a code that is not so much a langue as a jargon, being unsystematic and polyvocal).14

In a section of the first chapter entitled `Machines', they list three general characteristics of machines: a system of breaks, the detachment of segments, and remainder-subjects produced by the machine. The interest of such a definition is, of course, that the subject is produced by the signifying chain as its remainder the beginning of the demotion of a concept of subject, which is still, at this stage, indebted to the Lacanian elaboration. The Deleuze and Guattari machine has two aspects, sometimes distinguished as two types of machines, the desiring and the social machine. Or again, a social-technical machine is a con-glomerate of desiring machines in a determinate historical conjuncture

the social machine is the molar regime of the molecular desiring machine.

As we can see, the theory of machines, which occurs at a certain moment in the development of Deleuze's philosophy (it is present in L'Anti-êdipe, in Kafka, in Dialogues, but afterwards tends to fade out), is a first version of the theory of assemblages. In Kafka, Deleuze and Guattari insist on the link: a technical machine is a constituent part of a social assemblage. The two aspects of the machine, the desiring and the social, prefigure the two aspects of an assemblage: the machinic assemblage of desire and the collective assemblage of enunciation. Although the dif-ference between the two concepts is clear: what is lacking in a machine, which therefore can only be a constituent part of an assemblage, an instrument for its operation, is the discursive aspect of the assemblage. A machine does not involve statements.

The strategic role of the concept of machine in L'Anti-êdipe is clear. The importance of machines refers us back to Marx, as the thinker of technical aspects of the world. By having recourse to it, Deleuze and Guattari accept the Marxist sea change in economics: the substitution of the point of view of production for the point of view of circulation. But a conventional Marxist would balk at the introduction of desiring ma-

chines, at the description of the flow of energy that machines code as a flow of libido. We are closer here to Bataille than to Marx.15 As in the early essay, the main target of the attack is Althusser's structuralist version of Marx. He is accused of reducing the machine (as a Marxist, Althusser cannot but be aware of the importance of production, and therefore of machines) to a structure, by adopting the central metaphor

16


of the theatre, which lies at the heart of the logic of representation: representation produces subjects as places and functions in a struc-ture. The Althusserian play on representation as Vorstellung and Darstel-

lung operates this reduction of molecular machines to the molar structure.

So far, the importance of the concept of machine for Deleuze's phil-osophy of language is not entirely clear. But important it is. It marks the beginning of a drift towards what could be a Marxist philosophy of language: the adoption of the point of view of production is such a mark. It announces the shift away from the idealist, or vulgar material-

ist, postulates of mainstream linguistics. It allows Deleuze and Guattari, for instance, to propose another theory of the symbol, where its mean-

ing is less important than its operative use (a move strangely reminis-

cent of the later Wittgenstein, except that this form of `meaning-as-use' insists on the social nature of meaning) and its positional functioning.17

A symbol is a machine: it is not defined by what it means, but by what it does again, we understand the reason for Deleuze's hostility to inter-pretation as search for meaning, not operation. And we understand the gains that the concept of linguistic machine brings to the new pragmat-

ics: a sense of dynamics (hence the insistence on the historicity of language, and on variation rather than on a static system); a sense of the materiality of machines (there is a materialism of technique and mechanics, inherited from Marx); a sense of the importance of language as social practice (the shift to production compels a shift away from individualism and the centrality of the subject qua speaker); lastly, an insistence on the concrete workings of language, as opposed to the abstract ideality of the system the most important aspect is no longer the potential combinations of the elements (le jeu du systeÁme), so much as the actual dice-throws of linguistic pragma (le systeÁme mis en jeu). And we also understand why Markov chains are so important as a model of language, why their mention, and praise, is recurrent in Deleuze: they help us negotiate the move from the closed system of langue to the open jargon of machinic language.

Important as it is, the concept of machine is a transitional concept, a step towards the formulation of the concept that lies at the core of the new pragmatics, the concept of assemblage.

4 Assemblages

The concept of machine, after L'Anti-êdipe, developed in two directions: towards ever more abstract machines, and towards assemblages.


Mille plateaux mentions machines that are abstract rather than mech-

anical (in the sense in which the French language talks of algorithms as machines logiques computer theory is the field of logical machines). Logical machines are situated at the right level of abstraction: they turn the plane of immanence into a plane of consistency, by fostering pro-cesses of territorialisation and deterritorialisation, and ordering the strata that form the plane (more like the French cake known as mille feuilles, which is made of fine layers of pastry, than a plane proper). This abstraction of the machine from its mechanical origin offers a transition towards assemblages: a machinic assemblage is the concretisation of an abstract machine.

So, the main development of the concept of machine is its absorption into the concept of assemblage. The machine as concept is historically situated: it is too technical (limited to a stage in the history of technol-ogy which is definitely `pre-high tech'), it smacks of nineteenth-century Marxism, of certain forms of the organisation of labour (Fordism is best described as a social machine). Assemblages avoid such associations: they avoid the sharp separation of base and superstructure that provides the coherence of some classical Marxist analysis. Indeed, the main characteristic of an assemblage is not that it is collective, in other words directly social (a machine is more than a simple tool in that it involves a form of collective organisation), but that it is ontologically mixed. The first meaning of this phrase, which I have had many oppor-tunities to use, is now clear: assemblages are neither machines nor apparatuses (as in Althusser's theory of ideology): they combine the abstract materiality of utterances and institutions and the concrete materiality of objects. In them, base and superstructure are irredeemably mixed.

The most explicit definition of an assemblage occurs in L'AbeÂceÂdaire, in the `D is for Desire' section. Deleuze is formulating a non-Freudian, far more optimistic theory of desire, around three theses: (a) the uncon-scious is not a theatre (this rejection of the logic of representation is by now familiar); (b) delirium is not about mother and father (there is a collective aspect, contrary to appearances, to delirium: its object is the whole of history and the cosmos); (c) desire has neither a single object nor a single subject: it is multiple, machinic and collective, and as such inseparable from assemblages. An assemblage is what desire flows through, or again, you always desire within an assemblage. (We under-stand the relevance of the slogan, `Never interpret!', that is, never seek to reduce desire to a relationship between a subject and an object, whose complexities need to be interpreted: the converse slogan is Àlways


experiment with assemblages!') Having thus established the essential function of an assemblage, Deleuze lists its constituent elements: things and states of affairs; utterances, in the two senses of process and result (eÂnonciation and eÂnonceÂ); a space, in other words territories on which desire circulates; and movements within this space, of deterritorialisa-tion and reterritorialisation. We can draw an interesting conclusion from this. It appears that an assemblage is not only machinic (hence the aptness of the translation of àgencement' as àssemblage'), but also geographic (hence the agencement is a territorial organisation, an ar-rangement or array of elements) an assemblage is a chart or a puzzle as much as a machine. The metaphor of `fit', much used by Searle in his Intentionality18 (there is a `direction of fit' in the relationship between an utterance and the world) might be useful here, provided we do not consider the result (à fits b'), but the process (`fitting a and b together'): an assemblage is an active process, and of course not a metaphor (Deleuze is practising his usual tactics of taking an ontological metaphor literally, that is, actualising or materialising it).

Since an assemblage is a process before it is a result or an entity, a comparison with the sense-event of Logique du sens is in order. Like the sense-event, the assemblage has two sides, one which involves bodies and states of affairs, and the events that occur in them, the other which involves utterances, and the sense they ascribe to the event. More precisely, an assemblage is an unholy mixture of events and territory. An assemblage begins by extracting a territory out of the environment, and it has two aspects, or planes (a plane of expression, or regime of signs, and a plane of content, in the shape of a pragmatics of actions and passions).19 So an assemblage is always dual, the combination of two aspects, or two sub-assemblages: a machinic assemblage of desire, and a collective assemblage of enunciation. An assemblage combines what one does and what one says, the objects of the world and their groupings, the utterances in the world and their enunciation. This duality of course reminds us of the dual nature of the event-sense, of which the assem-

blage is an unacknowledged development. The ontological mixture that characterises assemblages is not only a critical dissolution of the Marxist opposition between base and superstructure, but also a development of the Stoic mixture which their pan-somatism allows. The new element in relation to the event-sense is the insistence on the territorial grounding of assemblages (whereas the event, a mist hovering over the battlefield, involves no territory, being a purely incorporeal transformation, even if the bodies concerned by the state of affairs occupy a territory). This is made clear in a passage in Kafka, where an assemblage is described as a


mode of segmentation (it territorialises and reterritorialises desire in the shape of power) and as a site for lines of flight (it also deterritorialises desire).20

The time has come to wonder about the interest the concept has for the new pragmatics. It offers two main advantages: it avoids the imperi-

alism, or tyranny, of the signifier; and it avoids the tyranny of the subject.

For Deleuze and Guattari, the centrality of the signifier is merely a pretext for indulging in interpretation. Here is Guattari, in the course of a joint interview on L'Anti-êdipe:

We've no use for signifiers. . . . If our criticism of the signifier isn't terribly clear, it's because the signifier's a sort of catch-all that projects everything back onto an obsolete writing-machine. The all embra-

cing but narrow opposition of signifier and signified is permeated by the imperialism of the signifier that emerges with the writing-

machine. Everything comes to turn on the letter. That's the very principle of despotic overcoding. What we're suggesting is this: it's the sign of the great Despot (in the age of writing) that, as it with-

draws, leaves in its wake a uniform expanse that can be broken down into minimal elements and ordered relations between those elem-

ents. The suggestion does at least account for the tyrannical, terroriz-ing, castrating character of the signifier.21

In this unfavourable description we recognise the Lacanian signifier at the centre of his thought in his first, structuralist, period. Deleuze and Guattari unexpectedly appeal, against Saussure and Lacan, to Hjelmslev and his concept of the reciprocal presupposition of the two planes of expression and contents. We have two parallel, reciprocally determined, flows, without any superiority of one plane over the other, that is, without one plane being in any way the representative of the other (this is, as we saw in Interlude 1, the core of the logic of representation). Language develops in the shape of flows on the two parallel planes, intersected and joined together by assemblages. So the question is not, `What does the word or utterance mean?' (Ìn what way exactly are its signifiers the representatives of its signifieds?'), but `How does the utter-ance work?', `What does it achieve?', Ìn what assemblage is it inserted?', `How is it, as part of a collective assemblage of enunciation, conjoined to a machinic assemblage of desire?' Hence the well-known provocative statement on the same page: `For us, the unconscious does not mean, nor does language.'22 It is difficult to state more clearly the shift from


representation and meaning to production and working, that is to a broader sense of `pragmatics'.

So the collective assemblage of enunciation is indifferent to the oppos-ition, constitutive of much contemporary linguistics (Benveniste, Culioli, and the various forms of enunciation linguistics) of sujet de l'eÂnonceÂ and sujet de l'eÂnonciation, grammatical subject and subject qua speaker. A collective assemblage is non-subjective, the concept does away with the centrality of the subject as the central notion of the philosophy of language (such centrality is common, in various guises, to Chomsky, Anglo-Saxon pragmatics and theories of enunciation). There are two ways in which the collective assemblage rids us of the subject.

The first is obvious. The assemblage is collective even when it is incarnated in or emanates from what Deleuze and Guattari call the `celibacy' of the artist, his solitary singularity (the case in point is, of course, Kafka). An assemblage never refers to a subject, not even to two subjects (as in interlocution): there is no subject who is the sender of the utterance, no subject whose utterance is reported. The source of the utterance is a collective, whether social, national or political. The celi-bate artist is merely the mouthpiece of the community. Thus, the as-semblage, which Deleuze sometimes calls `the minimal unit of language' (a position he refuses to grant to the word, or concept), involves multi-

plicities of various kinds: population, territories, becomings, affects, events. There is an interesting consequence to this, one we have already briefly evoked: being collective, the utterance is not within the speaker, but always already outside her, floating among the constituent parts of an assemblage as the event floated above the battlefield. So a writer is invented by an assemblage at the very moment when, in his originality, he is inventing it. As a result of this the term àuthor' in the field of literature is not a good concept at best, it is the name of a function, as in Foucault; or rather, the actual celibate àuthor' forms an assemblage with the virtual community (in the case of Kafka, `the people'), of which he is a part, and which he brings to life in his utterances. The most individual type of utterance, the literary utterance, is always the product of a collective assemblage of enunciation. This is why Deleuze and Guattari are able to claim that `there are no individual statements, there never are. Every statement is the product of a machinic assem-

blage, in other words, of collective agents of enunciation. . . . The proper name (nom propre), does not designate an individual: it is on the con-trary when the individual opens up to the multiplicities pervading him or her, at the outcome of the most severe operation of depersonalisa-tion, that he or she acquires his or her true proper name'.23


But the collective assemblage of enunciation gets rid of the subject, or puts it in its proper, subordinate place, in yet another way. There are indeed such things as subjects, but only as end products of processes of subjectivation: the most important aspect is the process, not the result. This is explicit in Deleuze and Guattari's critique of the Althusserian concept of interpellation, with its dialectics of Subject and subject, and of Benveniste's linguistic anthropology, based on the historicity of the subject.24 For them, subjectivation, far from being an ahistorical con-stant (the Althusserian dialectics of interpellation is eternal like ideol-ogy), is the product of a collective assemblage characteristic of one regime of signs (among several others), which they call the post-signifying regime of signs. Subjectivity, therefore, is not a constitutive, or an internal condition of language: it is a logical-historical configur-ation. Certain assemblages produce subjects, others not: individual sub-jectivity is not the beginning and end of the human condition.

The time has come to illustrate the concept by mentioning examples of assemblages. We remember that the main characteristic of an assem-

blage is that it is an ontological mixture, and that it has two faces, or parts, a machinic assemblage of desire, and a collective assemblage of enunciation. There is a problem here, which Deleuze and Guattari do not really solve: how can we separate the machinic from the enuncia-tive, since these two assemblages are in fact only aspects of the one assemblage? So the dominant characteristic is the ontological mixture, the formula for which is to be found in Mille plateaux: àn assemblage of enunciation does not speak `òf'' things; it speaks on the same level as states of things and states of content'.25 The French phrase is àÁ meÃme les choses', which does not mean simply òn the same level as' but also ìn the midst of'. So the assemblage of enunciation is inextricably mixed with the machinic assemblage of desire, which is not of course a produ-cer of goods (like the Marxian machine) but of states of bodily mixtures in a given society.

The canonical example of an assemblage is the feudal assemblage:

Taking the feudal assemblage as an example, we should have to consider the interminglings of bodies defining feudalism: the body of the earth and the social body; the body of the overlord, vassal and serf; the body of the knight and the horse and their new relation to the stirrup; the weapons and tools assuring a symbiosis of bodies a whole machinic assemblage. We should also have to consider state-ments, expressions, the juridical system of heraldry: all the incorpor-eal transformations, in particular, oaths and their variables (the oath


of obedience, but also the oath of love, etc.): the collective assem-

blage of enunciation.26

In this case, the difference between the machinic assemblage and the assemblage of enunciation is fairly clear, although too sharp a separ-ation will make the status of institutions (and their specific form of materiality) dubious: an institution involves spaces and `things' (build-ings, for instance), as much as utterances. Deleuze and Guattari overlook this point, and the other (rare) examples they provide do not enable us to clarify it. Thus, in Dialogues, Freud's Little Hans is described in terms of àssemblage' (`house street the warehouse next door an omnibus horse a horse falls a horse is being beaten'),27 but without more specification: the list just quoted does seem to mingle `statements', `things' and `territories'.

If we concentrate on the plane of expression, that is, on the assem-

blage of enunciation, we note that the `Little Hans' assemblage com-

bines objects and spaces not so much with a `statement' or even an ùtterance' as with a sentence, susceptible of grammatical transform-

ations, as in the famous case of Freud's analysis of fantasy, À child is being beaten.' So, like Freudian fantasy, or the fantasy analysed in Irigaray, which revolves round a verb in the infinitive,28 the assemblage of enunciation is inscribed in a sentence (whereby it is linked to the concept of haecceity, which in Deleuze does the philosophical work usually entrusted to the concept of subject). Thus, in Dialogues again, the `vampire' haecceity is also an assemblage of enunciation: A VAM-

PIRE TO SLEEP DAY AND TO WAKE UP NIGHT.29 This is how a haecceity is defined: `there is a mode of individuation very different from that of a person, subject, thing, or substance. We reserve the name haecceity for it. A season, a winter, a summer, an hour, a date have a perfect individuality lacking nothing, even though this individuality is different from that of a thing or a subject'.30 Haecceities are the contents of the plane of consistency, whose plane of expression is filled by a form of semiotic, a semiotic that is composed of assemblages of enunciation composed of sentences of the form Ìndefinite article � proper name � infinitive verb' the basic chain of expression, freed from `formal sig-nificance' and personal subjectivation.31

So we have a nexus of concepts that define the new pragmatics and can, as usual, be inscribed in a correlation:

SUBJECT SYSTEM UTTERANCE AS REPRESENTATION OPPOSITION OF DEPTH AND SURFACE (HIERARCHY)


HAECCEITY ASSEMBLAGE UTTERANCE AS DETERRITORIALISA-

TION PLANE OF IMMANENCE, OR OF CONSISTENCY.

Because the illustrations Deleuze provides are sketchy, and the con-cept of assemblage is still a little abstract, I shall attempt a brief illustra-tion.

Towards the middle of the nineteenth century, a new genre appeared in English literature, the sensation novel. Under the generic name, a number of novels, I should say a great variety of novels, were grouped: the novels of Wilkie Collins, those of Mrs Braddon (Lady Audley's Secret, Aurora Floyd), East Lynne, Mrs Henry Wood's bestseller, but also Rhoda Broughton's Not Wisely But Too Well.32 The question, of course, is: why were those novels, in spite of their many differences, grouped into a genre? And an immediate answer would be historical and philological: we could look for the first occurrence of the word `sensation' in this sense in the OED. But such an answer would be contingent. What we are looking for is a concept in the Hegelian sense, a retrospective ordering that makes global sense of apparently disparate elements.33 The concept of assemblage enables us to construct a concept of `sensation novel' in the Hegelian sense. And it has the advantage not only of providing a rationale for the grouping of the novels, but of accounting for this grouping in terms not merely of the internal development of literature (that is, not only in terms of literary history) but of the whole historical conjuncture, involving as it does machinic assemblages of desire as well as collective assemblages of enunciation.

Which is the assemblage of desire involved in the sensation novel? Since the novel uses the archetypal erotic plot (who marries whom), the immediate answer seems to be: the couple. A, a man, enters in an assemblage of desire with B, a woman. But this (which nevertheless excludes homosexual love) is hardly original: even if we are not dealing with the eternal attraction between the sexes but with a historical construct (with the hierarchy of man as subject and woman as object of the plot), it seems to cover the whole history of the novel from its eighteenth-century origins onwards. The collective assemblage that dis-tributes women (by deciding which are marriageable and which are not) is as old as the hills: already in the eighteenth century Pamela and Clarissa Harlowe seriously question it.

So what is specific to the sensation novel? The assemblage it embodies is linked to a more specific historical conjuncture, to a more precise moment in the history of the heterosexual couple: not so much the moment of emergence of women as fully independent individuals


(this is already realised in Pamela) as the moment of the breakdown of the hierarchic marriage. The sensation novel is the novel of the age of the divorce court.

So we have a specific machinic assemblage of desire, composed of bodies (men and women of marriageable age), of buildings (the upper-class mansion where the heroine lives, but also the cottage where eccen-tric characters Ann Catherick in The Woman in White or the groom in Aurora Floyd live, their natural habitat), of a certain organisation of space (not only the mansion v. the cottage, but my lady's boudoir v. the official drawing room, or the man's smoking room Lady Audley's boudoir is an object of considerable fascination), which is also a social organisation (of wealth, of power, etc.).

This machinic arrangement has a correlate on the plane of expression in a collective assemblage of enunciation. We may begin with those indeterminate (because mixing objects and utterances) institutions: law courts and the laws that are enforced in them (the laws of inherit-ance, the law of divorce) and the assize court when murder (in which the sensation novel, an ancestor of the detective story, likes to dabble) has been committed. This, however, is not specific enough yet we need to identify the sensation novel as a specific assemblage of enunciation. There are indeed features common to the novels that justify the inven-tion of the genre, and they are linked to the laws and customs just outlined.

The sensation novel is characterised by a certain conception of what women are: the plots are often based on a link between woman, truth and secrecy (not only is truth feminine, as in the famous opening to Nietzsche's Beyond Good and Evil, but the heroine in the sensation novel, because, like Lady Audley, she conceals a secret, is the object par excel-lence of the enquiry into truth). This mystery that woman conceals (sometimes to the point of committing murder in order to preserve it) has a specific content. It concerns a dysfunctioning of the couple: where two should be company enough, there is always a crowd, in other words a third party. So the woman is, unwittingly, bigamous, as Bathsheba almost is in Far From the Madding Crowd and of course the first hus-band, presumed dead, never fails to return from the grave. Or the right man is always married (as in Not Wisely But Too Well). Simple adultery is usually not enough: it takes bigamy, or an invalid marriage (as in Col-

lins's No Name) to do the trick, that is, to unleash the violence latent in the state of matrimony (a social form that no longer expresses real relations between man and woman). Hence the core of the genre, its archetypal plot. A mystery from the past re-emerges in the present, as in


the return of the foreclosed in the shape of hallucinations. Except that, if dreams and hallucinations are indeed liberally resorted to (think of Collins's Armadale), the figure from the past that haunts the present is all too real, and must be dealt with through the real act of murder. And, this is the specificity of the sensation novel, it is the woman alone, as incarnation of the secret, who must bear the brunt of this re-emergence of the Real, and do man's work, that is eliminate it (this is the plot of Lady Audley's Secret). The woman is not only the object of the mystery, but also the subject of the action it provokes.

This archetypal plot is inscribed in a certain type of sentence (as we have seen, the Deleuzean assemblage, like Freudian fantasy, is also an assemblage of words, a sentence). In spite of obvious differences in style among different authors, there is a certain community in the presence of both the sensation sentence, expressing sensuality (women are mys-

terious not only because they hold the key to the past but because they are, in Freud's terms, a `dark continent'), and the sensational sentence, conveying mystery and suspense. The second type of sentence is in-herited from the Gothic novel, but not the first.

What I have described constitutes an assemblage of enunciation (typical sentences, the same encyclopaedia represented by similar clicheÂs, cer-tain similarities in the plots of the novels there are, of course, vari-ations on the archetypal plot and common unconscious philosophical underpinnings, witness my allusion to Nietzsche). This enables us to trace the genealogy of the genre in Dracula (a veritable orgy of sensation cum sensational sentences), in Hardy (who does enjoy a little melo-

drama) or in the detective story (Agatha Christie is fond of the return of the mysterious past, to the point of obsession). And this assemblage is collective: the individual author, whatever his individual talent or lack of it (Wilkie Collins is a better novelist than Mrs Henry Wood) is spoken by the genre, which means that with his concept of assemblage, Deleuze is offering us a new theory of what a literary genre is a gift not lightly to be declined.

5 Minority

In L'AbeÂceÂdaire, `G comme Gauche', Claire Parnet asks Deleuze whether for him Marx is still a reference, and whether he, Deleuze, is a revolutionary. All she elicits, of course, is an ironic answer: you must be pulling my leg (in the mid-1990s, the vocabulary of the 1960s can only be taken with a pinch of salt). But Deleuze, who has always claimed to belong to the left, proceeds to explain what it takes to be on the left (at the same time


insisting that there is no such thing as a left-wing government). To be on the left, then, is (a) to be aware of the horizon, not the centre, in other words to eschew egoism, or the egocentricity of identity claims, and (b) never cease to become a minority. Majority, the pretext for democracy, is for Deleuze not a quantity, a mass of people, but a standard: the acro-nym WASP describes the American version of majority. Against such majority there is a multiplicity not so much of minorities (with fixed boundaries and identities) as of becoming-minor: thus there is a becom-

ing-woman of men (but no becoming-man of women, masculinity being a constituent part of the standard). Minority, therefore, is linked to becoming, a combination of active forces, of forces for change, whereas majority is a nexus of static, reactive forces. There is also a certain creativity in minority, due to its capacity to struggle for change. Of course, the majority, being a model, can be less numerous, and usually is, than the multiple minorities, which, having no model, are in constant process. If we take the argument to the limit, nobody is a member of the majority (hence its famed `silence'), as everybody is caught in some sort of becoming-minor. And, contrary to appearances, the left's appeal to `the people' is not an appeal to the majority (only petty bourgeois illusion makes us incensed at the fact that George W. Bush polled fewer votes and yet found a `majority'), but an appeal to the potential for change of an array of creative minorities. This is why great artists, Deleuze adds, appeal to the people, why Kafka was a self-styled àrtist of the people': artists are constitutively participants in minorities, and art can only be defined as resistance (this, of course, is only another formulation of Deleuze's high modernist or avant-garde aesthetics).

It is logical, therefore, that the book in which the concept of minority plays a major part should be Kafka. Deleuze and Guattari's interpret-ation of Kafka largely relies on their definition of him as a great, because minor, artist (this, of course, is an inversion of the usual phrase, à minor classic', which they blissfully ignore, as it has no equivalent in French: they are not attempting to put Metamorphosis in the same category as Lorna Doone). The Kafka book attempts a strict definition of `minor literature'. As a preamble, it must be noted that minor literature is not the literature of a minor language (thus, Kafka did not write in Yiddish), but the literature a minority imposes on a major language. There are problems with this preamble, and accusations of Eurocentrism will soon be made: it seems to work for Kafka, a Czech Jew writing in German, but what about the claims to greatness of Czech and Yiddish literature? And what about African literature when authors refuse to write in


English or French (the case of Ngugi, the Tanzanian novelist, immedi-

ately comes to mind)? However, before indulging in a thorough rejec-tion of the concept, we might as well see how it works.

The concept has three characteristics. A minor literature is (a) deterri-torialised (in its texts, the major language is deterritorialised); (b) dir-ectly and entirely political; and (c) collective (all the values that inform it are collective values). In spite of the awkwardness of the preamble (which we may ascribe to ignorance of the postcolonial problematic in literary criticism, which had hardly emerged when Deleuze and Guattari wrote), the concept, it seems to me, has extraordinary progressive, even revolutionary, potential: it firmly places great literature outside the field of identity and identity claims (literature that is worth the name is about de-, not reterritorialisation); and it firmly places it outside the realm of ethics and the eternity of human nature (the problems discovered and negotiated in literature are political, not ethical-individual problems not the `dirty little story' of the author's own problems); it equally firmly ascribes a minor role, if any role at all, to concepts of author, originality or genius: a collective assemblage is speaking through the àuthor's' voice. As Deleuze and Guattari bluntly say: `there isn't a subject: there are only collective assemblages of enunciation'.34

The consequence of this is that minor literature is an expressive ma-

chine, rather than the cardinal instance of the significant use of lan-guage. Its aim is not to foster or extract meaning, but to give rise to intense, and intensive, expression: not unexpectedly, for this is a topos of modernism, the authors mentioned, apart from Kafka, are Joyce and Beckett. The problem the `great author' faces (the paradox is that those collective assemblages nevertheless form a canon: there is assemblage and assemblage, and some are more collective than others . . .) is how to prise a minor literature from its own major language, how to make it follow lines of flight, how to prevent its reterritorialisation into mean-

ing (for there is danger in language, in its representational and extensive use). Thus, minor literature may be defined as an intensive asignifying use of language. From this point of view also, we understand Deleuze's hostility to metaphor. Metaphor makes meaning increase and multiply, it is a device for generating ever more meaning: but the point of minor literature is not to make recognisable sense, but to express intensities, to capture forces, to act. This is why Deleuze claims to support metamor-

phosis against metaphor. The origin of this definition of literature as constitutively minor is in

that formula of Proust's, which Deleuze frequently quotes. Here is a typical passage:


We can see more clearly the effect of literature on language. As Proust says, it opens up a kind of foreign language within language, which is neither another language nor a rediscovered patois, but a becoming-

other of language, a minorization of this major language, a delirium that carries it off, a witch's line that escapes the dominant system.35

This witch's line, a more picturesque name for a line of flight, is also, as we shall see in the next chapter, the intensive line of syntax. This becoming-other of language in literature must be understood as a call for agrammaticality, that is, for linguistic and more particularly syntac-tic, creativity hence the three aspects of literature that this essay, aptly entitled `Literature and Life', spells out on the next page. Literature (a) deconstructs or decomposes the natural language in which it is written, that is, it invents a new language through syntactic creation; (b) takes language beyond the system that supposedly rules it, pushes it out of the straight furrow, into etymological delirium; and (c) makes it break away from its usual expressive limits, not into silence (this would be the modernist argument of Blanchot or Bataille) but into other media, what Deleuze calls `visions and auditions'.

The minorising of language has one potential consequence of consid-erable importance. Since the collective, `the people', is speaking, rather than the individual preÃte-nom that signs the text, all linguistic expression is an at least potential form of art. For even Markov chains can go mad (in another essay in the same collection, Deleuze describes the eroto-maniac delirium of Sacher Masoch in terms of stammering and stutter-ing). Other examples of apparently usual occurrences that become literature by minorising language (which is, therefore, never òrdinary') is the scream (in Artaud's `cris-souffles') or the exclamation (as used by CeÂline). For language, as we now know, is never a homogeneous system, but always the site of polyphony. What minor literature does to it is to turn this polyphony into polylingualism: such polylingualism actively prevents language from becoming homogeneous, it keeps it in a state of constant imbalance, and thus makes it creative. We understand the importance of Beckett for Deleuze, of his `translation' practice, a con-stant source of imbalance, of his experiments with language, most notoriously in Watt, but also in the plays for television we evoked in the Introduction.

If the minor treatment directly concerns literature, it has direct con-sequences, as we have just seen, for language as a whole, whose poten-tial for creativity is thus focused upon. We must, therefore, look again at the theory of minor language put forward in the critique of the fourth


postulate of linguistics in Mille plateaux.36 There, we learn that when a language becomes a major language, a historical cum political process involving phenomena of linguistic domination or imperialism, perhaps even glottophagy, it is also, in an apparent paradox, worked through by continuous variation that transcribes it in a minor mode. For linguistic domination is never as simple and straightforward as political domin-

ation: sometimes the language of the victor supplants the language of the vanquished, but sometimes the reverse is true (think of Latin in relation to the languages of the Barbarian invaders). This minorisation can occur in three ways: through contact between a major language and a subjected minor language (this is the case of English and Welsh, where Welsh is stifled by English, but puts up a fight); by the development of new Englishes (which collectively have now far more speakers than British English); or through the multiplication of dialects and registers within the standard dialect (English is prone to such minorisation to a far greater extent than French).

The role of minor language, or dialect, or mode, is therefore clear: it keeps the major language alive, it turns its potential freezing into a stream, an ongoing process. The minor `treatment' of a major language must be understood in the medical sense of the term. This is obviously the case of the literary treatment but this is also the case of the treatment through minor dialects and registers. English literature, as is well known, is peopled by Scots, the Irish and the Welsh: Under Milk Wood is an example of an extraordinary minorisation of English it is through such minorisations that the language lives.

This places the English language, as the archetypal major language, in a contradictory position. It inspires the same feeling that capitalism inspires in Marx: a grudging admiration, together with a feeling of urgency in the struggle against its domination. This is particularly ap-parent in Dialogues. A whole chapter is devoted to a paean to Anglo-American literature. It is entitled `The Superiority of Anglo-American Literature'. The geographical obsessions of such literature (from Hardy to Kerouac), its constant becoming-revolutionary, its breadth of scope, that is, its capacity to eschew the pettiness of the individual hero and to foster lines of flight, are much praised, and a canon (Kerouac and Hardy, but also Melville and Whitman) duly erected. (That this is a French invention of American literature is obvious: even as the French also invented the American film noir). But when it comes to the English language, its tendency to hegemony, even to imperialism, is duly noted, together with the paradoxical consequences this entails:


It is a hegemonic, imperialistic language. But for this reason it is all the more vulnerable to the subterranean workings of languages and dialects which undermine it from all sides and impose on it a play of vast corruptions and variations. Those who campaign for a pure French, uncontaminated by English, are in our view posing a false problem which only has any validity in the discussions of intellec-tuals. The American language bases its despotic official pretensions, its majoritarian claim to hegemony, only on its extraordinary cap-acity for being twisted and shattered and for secretly putting itself in the service of minorities who work it from inside, involuntarily, unofficially, nibbling away at that hegemony as it extends itself: the reverse of power. English has always been worked upon by those minority languages, Gaelic-English, Irish-English, etc., which are all so many war-machines against English.37

Like capitalism, and for the same reasons, the imperialist language is also the site for the linguistic revolution. There is, of course, a certain amount of optimism in such a description, which must be understood as a call for political action. For linguistic imperialism does not impose its domination by sheer force, and the Gramscian term `hegemony' (much elaborated in the work of Ernesto Laclau38) must not be taken lightly: the Irish and Welsh `war machines' are not so much held in check as absorbed by hegemonic English. Anglo-American literature reflects both aspects, the revolutionary aspect of the subversion of the imperialist language, and the hegemonic aspect through which the major language comforts its majority in dialectic interaction with the dialects or regis-ters that minorise it.

There are lessons for another philosophy of language to be drawn from the concept of minority. It makes three important contributions to the new pragmatics. (1) The study of language is inseparable from the study of the art of language, literature. The point of entry into the intricate workings of language is the literary text, neither the artificial but inexistent `correct' sentence of `the man hit the ball' type, nor the flood of actual sentences in computerised corpuses. (2) Politics is central to language, not only in the sense that there is a politics of languages, of their contacts, of their mixtures, of their rapports de forces, but in the sense that there is a politics of becoming: minorisation makes language alive, it incites in it a process of becoming, it incites it to break its own frontiers. Here the poetics of modernism can be generalised to linguistic activity as a whole, as minorised language is worked through by the event it seeks to capture. (3) The concept deals the last blow to the logic


of representation and the main form it takes as far as language is con-cerned, the problematic of identity and subjectivity: the becoming-

minor is an asubjective process, it does not claim or comfort identity, but on the contrary blurs fixed identities in its collective process of variation. It is all the more surprising, therefore, that literary minorisa-

tion should be accounted for by Deleuze under the concept of style.

6 Conclusion

The new pragmatics is, at the same time and indissolubly, a new poetics. This is a source of tension within Deleuze's philosophy of language, perhaps even of contradiction. But it is clear that for him there is no democracy of corpuses. The detailed study of an everyday discussion or telephone conversation yields trivial and uninteresting results, for such everyday exchanges are fully functional from the point of view of com-

munication, and more often than not irenic. And they do have a point, to be reached and negotiated as swiftly and economically as possible, as when I book my ticket on the Eurostar. But there is hardly any novelty involved, even if (especially if?) the conversation becomes personal and garrulous. As a result, we have a series of utterances without interest (in the sense in which I have used the term at the beginning of the chapter), a static talking machine (some of those conversations can be literally mechanised: Ìf you wish to check your bank balance, press one'), an assemblage of enunciation so frozen that it is no longer perceived as such, a mode of language so standard no process of minorising can reach it (there lies the danger to English as well as to the dominated language

of the global domination of the language of imperialism). So the point of entry into language, in its problematic aspect, before

common sense has had time to answer all the questions, is the literary text, and the new pragmatics is a poetics. The problem here is that it is two poetics rather than one, an obvious source of tension.

On the one hand, we have what I have called the `high modernist' or àvant-garde' Deleuze. The terms make no claim to historical accuracy and will no doubt be indignantly rejected by historians of literature. They point to a literary and cultural conjuncture which goes far beyond Deleuze, in which he is, like a number of his equally illustrious contem-

poraries, caught. Thus, Deleuze, the friend of Foucault, is, like his friend, an admirer of Blanchot and Bataille, and of their strong views on the nature of the literary work of art. He shares with his friend, but also with the Tel Quel group, the early Kristeva and the incipient Derrida (but not, for instance, with Althusser), a canon of high literature the names


(Artaud, Joyce, Kafka, Beckett, meÃme combat) have recurrently figured in these pages. This amounts to an elitist view of literature (`high modern-

ism' is the name for a now faded avant-garde), with a strong interest in the subversion of language and in its limits, a view where poetry will always be preferable to the novel, and where, if we must indulge in fiction, the experimental will always be preferred to the mainstream variety hence the importance of a broad concept of agrammaticality. The high point of such a view of literature is the striving for silence as the perfection of literary art. This view, as we shall see in the next chapter, is encapsulated in the concept of style.

But Deleuze is not only Foucault's friend, he is also Guattari's. So there is another Deleuze, and another poetics, a quasi-Marxist one, a poetics of machines and assemblages, with interesting characteristics, such as the essentially political nature of all utterances (hence the interest in Lenin), their consequently collective nature, and their historical, or con-junctural aspect (meaning, being the result of a rapport de forces, is conjunctural: assemblages are strongly historicised, and there is a his-torical succession of regimes of signs). Even if this analysis of language is not conducted in terms of ideology, apparatuses and the primacy in the last instance of the economic and social base, even if Althusser is the frequent object of implicit, and sometimes entirely explicit criticism, this picture of the workings of language, based on Bourdieu's external linguistics, is not very far from a straightforwardly Marxist analysis, witness a concept of subject as end-of-chain product (I have repeatedly alluded to the Althusserian `chain of interpellation', institution ritual practice speech-act). Like Althusser or RancieÁre, Deleuze and Guattari deal with subjectivation, not subjectivity or intersubjectivity. And the ontological mixture of assemblages corresponds to a form of material-

ism closer to historical materialism than to the reductive biological materialism of Chomsky. Such materialism even opens up interesting possibilities of conceiving, against psychoanalytic orthodoxy, the psyche not as mere interiority but as exteriority: the unconscious is not deep within us, but outside us, in language, in ontologically mixed assemblages (similar hints can be found in Voloshinov, and in the work of Denise Riley).39 Hence the demotion of concepts like àuthor' in Deleuze and Guattari's poetics: hence also a possible contra-diction with the elitism of canon building. What we have in this quasi-Marxist poetics is assemblages and/or language-games, not judgements of excellence.

This second poetics is, in the work of Deleuze, typical of the period of his collaboration with Guattari. In his last book, Critique et clinique, the


high modernist poetics of stuttering and the limits of language is again prevalent. Nevertheless, there is an attempt at overcoming the tension between the two poetics: it resides, unexpectedly so, in the concept of style, the object of the next chapter.

Interlude 3

A Reading of Kipling's `Wireless'

1 A story

`Wireless', first published in 1904, in the collection entitled Traffics and Discoveries,1 is an extraordinary story. It starts rather innocuously in a chemist's shop in a seaside resort on the south coast of England. The narrator, a doctor and a friend of the chemist, visits the shop because, in the back kitchen, the chemist's nephew is trying to operate an early Marconi wireless set and contact nearby Poole. While the nephew is setting up his apparatus and vainly trying to make contact, the narrator sits in the shop with the chemist's assistant, one Shaynor, a serious young man obviously dying from tuberculosis. To comfort him, the narrator, from the chemist's drugs and potions, concocts a strong and probably intoxicating drink (à few cardamoms, ground ginger, chloric ether, and dilute alcohol' the result is à new and wildish drink', pp. 184 5), of which Mr Shaynor liberally partakes. After a visit from his girlfriend, a red-headed beauty by the name of Fanny Brand, Mr Shaynor goes into a trance and starts, if not sleepwalking, at least something very much like sleeptalking. At first, it seems he is quoting Keats, but, little by little, the narrator realises that he is in fact not quoting but writing Keats, in the throes of painful inspiration: he is working his way, from hesitant first draft through emendation towards some of the most famous lines of `The Eve of St Agnes' and even Òde to a Nightingale'. As he is on the brink of producing that masterpiece (which is also his masterpiece), he wakes up from his trance, and, in answer to the narra-tor's eager questioning, confesses he has never heard of Keats (Ì haven't much time to read poetry, and I can't say that I remember the name exactly. Is he a popular writer?', p. 198). So he was not remembering the lines but actually composing them if a reminiscence at all, then a

202

Interlude 3 203

Platonist form of it. Meanwhile, the nephew, who so far has still failed to reach Poole, captures the disconsolate exchanges between two dread-noughts out at sea, whose sets emit but do not receive: ` ̀ `K. K. V. Can make nothing of your signals.'' . . . ̀ `Disheartening most dishearten-ing'' ' (p. 199). In the last paragraph of the story, when the nephew at last succeeds in receiving messages from Poole (` ̀ ` Here's Poole, at last clear as a bell'' ', p. 199), the narrator, who initially came to the shop for this, decides to go to bed, as he feels, he claims, à little tired' (p. 199).

2 Communication

The word that constitutes the title of the story, `wireless', is in inverted commas: the real object of the story is not the literal wireless that is the pretext for the narrator's visit to the chemist's shop, but a metaphorical wireless, not a Marconi, but a Keats set. And the object of the story, therefore, is not successful communication and exchange of messages, but its failure (in the case of the dreadnoughts) or its unearthly and unorthodox success in the case of Mr Shaynor. The last line, the chute of the story, makes it clear that common-or-garden communication, even through the brand-new technological channel of a wireless set, is deeply uninteresting. And the moral of the story appears to be that scientific communication is both aleatory and trivial, while telepathic or poetic communication is splendidly successful and thoroughly fas-cinating.

The historical and cultural context in which the story was written makes such a moral less extraordinary than it seems today. This was the time when tables were both automotive and garrulous, when Madame Blavatsky was a cultural force, Conan Doyle believed not only in fairies but in communication with the dead and Helen Smith impressed Saus-sure with her mediumnic knowledge of Sanskrit, if not Martian. The world of drab reality was surrounded by more exotic worlds peopled by fairies and spirits. That Kipling felt in tune with this cultural atmosphere is clearly indicated by another story in the same collection, `They', where the narrator, driving through the countryside in an early auto-mobile (we note the same preoccupation with ultra-modern technology, also present in Dracula, with its telegraph wires and recording ma-

chines), loses his way and eventually reaches a manor house inhabited by phantom children.

However, there is much more to `Wireless' than such a reflection of a Zeitgeist that today evokes at best an indulgent smile in us for a simple reason: the story never suggests that Keats, the spirit or ghost of Keats, is


speaking through Mr Shaynor in his trance this is not a spiritualist seance. Kipling's specific genius is that he treats this potentially spiritu-alistic situation as a situation of communication, on a par with normal situations of interlocution and new-fangled technological communica-

tion. The story is indeed, from beginning to end, and in the most emphatic manner, about communication, its predictable failures and strange and extraordinary successes. We can chart the development of the text as a series of situations of communication, for the most part analysed in terms of machines and assemblages of machines. Here is, as briefly as I can, the list of such situations.

The first does not belong to the main story-line and at first appears to be an irrelevant digression. The narrator, who is a doctor, seeks redress from a blundering chemist who made a mistake in concocting one of his prescriptions and refuses to apologise. On the advice of Mr Shaynor's boss, the old chemist, he resorts to Apothecaries' Hall, the ruling body of the profession, with swift and surprisingly successful results: Ì did [report him to Apothecaries' Hall], not knowing what djinns I should evoke, and the result was such an apology as one might make who had spent a night on the rack. I conceived great respect for Apothecaries' Hall' (p. 182). Here, a situation of failed communication (the narrator demands an apology, the chemist fails to comply) is turned into success-ful communication by going through another, mediating circuit (the narrator writes to Apothecaries' Hall, who write to the chemist, who writes to the narrator). Communication is achieved, but at the cost of indirectness. And the secondary circuit that allows communication is a circuit of authority: communication, in order to be successful, must be authorised. But we also note, of course, the phrase, `what djinns I should evoke', which is endowed with strong dramatic irony: authority works by means similar to those of magic.

The second situation also seems to be irrelevant to the story-line. Mr Shaynor tells the narrator how he is capable of serving a client while going on with his private train of thought and mentally perusing a page from Christy's New Commercial Plants, which he is studying for his exams. This, which places Mr Shaynor dangerously close to the irrespon-sible chemist in the first anecdote, makes for a situation of interlocution where communication appears to be successful (the client gets her prescription) but at the cost of inauthenticity: the outer circuit (Mr Shaynor's dialogue with the client) is purely automatic, while the inner circuit (Mr Shaynor recalling a page from Christy's textbook) is purely private. This is in fact a situation of non-communication, be-tween the circuits and within the circuits.

Interlude 3 205

The third situation takes us further from trivial interlocution or tech-nological communication. After a visit to the shop by Mr Shaynor's fiery red-headed girlfriend, the narrator notices a resemblance between the young woman and `the seductive shape on a gold-framed toilet-water advertisement' (p. 186), which Mr Shaynor treats as à shrine', and which he censes by burning pastilles in front of it àn established ritual which cost something' (p. 187). This is the archetypal situation of communication with the Gods, in the shape of sacrifice: the worshipper addresses the absent addressee, who is supposed to receive the message or offering, although she never responds, except through cryptic signs. The interesting point is that this episode occurs only one page after a short dialogue between the narrator and the nephew, who is still fid-dling with the knobs on his set, and claims to reach the Gods through his instrument, thus instrumentalising them: `That is the thing that will reveal to us the Powers whatever the Powers may be at work through space a long distance away' (p. 185). We note, of course, the deliberate use of the metaphor of power(s), in all its ambiguity: electrical power as much as the Gods as Powers. The distance between scientific communication, where God is replaced by an instrument, and magic or poetic communication, where God is the unresponsive Addressee, is now firmly established. This becomes even more apparent if we consider the next two instances of `communication'.

The first is an instance of poetic communication. Mr Shaynor is about to drink the narrator's concoction when he suddenly observes: Ìt looks, . . . , it looks those bubbles like a string of pearls winking at you rather like the pearls round the young lady's neck' (p. 188). The young lady in question is the image that he worships. But the interesting point is that the simile he uses to describe the bubbles of the drink, `like a string of pearls winking at you', is an allusion to Keats the first but not the last. In the second stanza of the Òde to a Nightingale', we find the following lines:

O for a beaker full of the warm South,Full of the true, the blushful HippocreneWith beaded bubbles winking at the brim.2

Since Mr Shaynor has never read Keats, he is not quoting him: he is rehearsing the poetic notions that made Keats revive an archaic sense of the verb `wink' (`to gleam or flash intermittently or fitfully', `to twinkle', but also `to close one's eye in association with drinking off at a draught'). Thus, he too revives the metaphoric drift that links the act of winking


(as perhaps indulged in by the seductive young lady in the advertise-ment towards the prospective customer, or Mr Shaynor, who worships at the shrine) with a visual quality of the bursting bubbles not only an unwitting allusion but an induced pun. The situation of communica-

tion thus evoked (in the sense that one evokes a ghost) is indirect and complicated. Neither Mr Shaynor nor the narrator know that Keats is speaking through Mr Shaynor to the narrator. The addressor is present only in the retrospective knowledge of the narrator, who will eventually realise (as allusions become more frequent and more explicit) what is happening a situation which, through the narrative trope of metalep-

sis, also concerns the reader, who notices the allusion to Keats in hind-sight, when he, like the narrator, has become aware of the workings of the Keats wireless set. So, in a classic case of metalepsis, we have a text (a word, `wink', which inscribes a metaphor and potential pun) with three authors at three different ontological levels: Mr Shaynor, present in the text but fictional; Keats, absent from the text but present in the hypotext, and real; and Kipling, only present in the text through his narrator, but manipulating the whole game with only too real effi-ciency.

On the next page, the sixth situation of communication is placed in strong contraposition to the preceding one. Mr Shaynor having fallen asleep through the effect of the drink, the narrator goes to the nephew, still fiddling with the knobs, and has him explain the workings of the Marconi set. The nephew first explains what induction is: an electric current in a circuit causes a magnetic field which charges a second circuit with electricity. In the case of the Marconi set, the Hertzian waves produced by an electrical circuit in Poole induce an electric charge in a coherer, several miles away. This coherer, activated by induc-tion, is linked to an electrical circuit that operates a Morse machine, turning the impulsions into lines and dots. We understand why I have used the metaphor of the circuit to describe communication. The whole story rests on the paronomasia that links `circuit' and `circulation' (of information), a paronomasia that only makes the metaphorical drift in the word `circuit' apparent: from a road circuit, or compass (hence a `Circuit Judge') we go to an electrical circuit (an established metaphor) and, through a new metaphor, to the circuit of communication (already virtually present in the word `circumlocution'). The interesting point is that the separation between the two types of communication, the poetic and the technological, now appears to be a parallelism: both methods are complex, in that they involve two different circuits, a relation of induction, and a coherer, for communication to be established. The

Interlude 3 207

Keats circuit induces words in the Shaynor circuit in the same way as the Poole circuit induces magnetism in the nephew's coherer.

The following pages, the centre of the story, develop this by describing the work of the Keats wireless set. For Mr Shaynor goes into a trance, and starts producing lines of poetry (beginning with Ànd threw warm gules on Madeline's young breast' Mr Shaynor is obviously dreaming of his young lady, whose name, Fanny Brand, proves to be highly significant, as it is very close to Keats's own Fanny Brawne): first mere clusters of words, broken and hesitant, then, more confidently, whole lines, and eventually and triumphantly, a whole stanza, às it is written in the book' (p. 195). A first explanation immediately comes to (the narrator's) mind. `The Eve of St Agnes' is one of the most erotic of Keats's poems, and Mr Shaynor's train of thought is as easy to guess as Keats's: the narrator primly notes that there `followed without a break ten or fifteen lines of bald prose the naked soul's confession of the physical yearning for its beloved unclean as we count uncleanliness; unwholesome, but human exceedingly' (p. 194). But Mr Shaynor does not stop at this. He attempts the Òde to a Nightingale', that is, sheer poetic magic (`Re-member that in all the millions permitted there are no more than five

five little lines of which one can say: ``These are the pure Magic. These are the clear Vision'' ', p. 196). Nor is this all: not only is he spoken by Keats's voice, not only does he become the poet himself, since we witness the trials and errors and eventual triumphs, the work of com-

position in progress, but he goes beyond Keats since, at one point, he improves on him. Thus, Keats, in the first line from `The Eve of St Agnes' produced by Mr Shaynor, has `Madeline's fair breast', à trite word', the narrator notes (p. 194), who `finds himself nodding approval' to Mr Shaynor's more vigorous `young breast'. The Keats wireless set is a re-sounding success, unlike the Marconi set, as the last two situations of communication demonstrate.

The first of these is the episode of the dreadnoughts: here, communi-

cation fails because the circuit is interrupted and the circulation of the message made impossible. Both ships manage to emit, but fail to receive. The second presents the inverse case: here, the failure of communica-

tion is due to its very success, as, when the machine works at last, the narrator loses all interest, no trivial interlocution, of the type Ì receive you', `Roger!', etc., being able to rival the communicative intensity of Keats's best lines.

We can draw a few conclusions from this list. The most obvious derives from the very length of the list: the story is not merely a quasi-fantastic story of possession (although I shall come back to this) but a


systematic exploration of the arcana of communication, in its authentic and in its inauthentic forms. And the interest of the text lies in its paradoxical attitude towards the relation between poetic and techno-logical communication. On the one hand it rehearses their similarity, and on the other it stresses their dissimilarity. Similarity first: communi-

cation is never direct, it requires two circuits, and a relation of induction between them, so that the speaker's words are never her words, but always induced by other words we are spoken not by ghosts but by Powers (òur unknown Power', the nephew calls it, p. 191); as a result, communication always occurs in and by a machine, or assemblage of machines, and as such it is always liable to failure, whenever and for whatever reason the flow or current is interrupted. But there is also strong dissimilarity involved. Both processes of communication are long and difficult: communication is delivered through painful labour. But the babies are vastly different. One is pug-nosed and ugly, like its parents in other words, everyday words are too trivial to communicate much, as the ironic non-dialogue between the dreadnoughts testifies: ` ̀ `K. K. V. Can make nothing of your signals.'' A pause. ``M.M.V. M.M.V. Signals unintelligible. Purpose anchor Sandown Bay. Examine instru-ments tomorrow'' ' (p. 199). The irony is, of course, compounded by the presence of the listening Tom, the nephew, who, in a symmetrical position to that of the dreadnoughts, can receive but does not emit. The other baby, however, is resplendent, she even outshines her glorious parentage: poetic communication, with its inspiration-inducing rituals, provides the highest level of communication. There is a certain triviality in this position. But, fortunately, Kipling's text is more complex.

3 Possession

The role of the narrator in the story is essential. Not only does he provide a background to the story by reminiscing about apparently irrelevant episodes; not only is he the initiator of the main chain of events by producing the wildish drink, the composition of which is as haphazard as that of Dr Jekyll's drug; not only does he play the part of Mr Shaynor's Watson by faithfully noting his vaticinations down to the slightest detail; not only does he touchingly try not to influence Mr Shaynor in his trance by inwardly reciting `The Rime of the Ancient Mariner', while Shaynor is labouring at the Òde to a Nightingale', in order not to prompt him telepathically; but he comments on the strange occurrences, invents various hypotheses, tests them and even produces a theory of what has happened, together with an ars poetica.

Interlude 3 209

When the narrator realises what is happening, his first reaction is àn over-mastering fear' (p. 194), which aptly marks the uncanny nature of the phenomenon. But being a doctor, like Dr Watson, even if he is a less positive spirit than his illustrious colleague, he immediately attempts a rational explanation (`he has read Keats') or rather two, the second being somewhat less rational than the first: the law of like causes produ-cing like effects means that if Mr Shaynor, a chemist's assistant, in love with a young woman called Fanny Brand, suffering from tuberculosis, starts spouting poetry, it must be, by induction, the poetry of Keats. A tall story, of course, but one which is eventually proved to be right when Mr Shaynor confesses he has never read Keats.

Strengthened by such inferential triumph, the narrator proceeds to construct a theory of poetry. This time, however, the theory combines a core of positive common sense with a capacity to trivialise the extraordin-ary. The production of poetry is an exercise in sublimation. Mr Shaynor starts with the grossest prose, the raw material of his instinctual drives producing a result which is more or less pornographic. This is then sublimated into the highest poetry. Poetry, then, is heightened prose, but only reached through the bodily travail, or pangs, of inspiration and composition. And it aptly ends in a climax. When Mr Shaynor has reached the perfect line, he collapses:

From head to heel he shook shook from the marrow of his bones outwards then leaped to his feet with raised arms, and slid the chair screeching across the tiled floor where it struck the drawers behind and fell with a jar. Mechanically, I stooped to recover it. (p. 197)

We are familiar with this scene: this is the scene where the medium, unable to bear the exquisite feelings caused by the materialisation of the spirits, wakes up, as the dreamer wakes in order to escape from his dream. And it is difficult not to sense a certain irony in this description: Mr Shaynor is too obviously possessed, too obviously adopting the Romantic pose of the vaticinating poet, too ready to illustrate the clicheÂs of a sub-genre (think of Conan Doyle's stories of mediums and spirits) and to confirm commonplace ideas about the superiority of poetry over prose (in a text which is, of course, a prose story).

This is where we realise that the whole story is carefully stage-managed, if not by a Dr Watson of a narrator, at least by Kipling. What we realise is twofold: the world of the story is, from the beginning, a world of signs; and it is also a world of machines, which are in stark


contraposition with theories of possession and inspiration, with the trivial ars poetica the narrator appears to defend.

4 Signs

I have reached a watershed in my reading. I have interpreted the text as staging two types of situation of communication. They are presented as similar, in so far as both are complex (two circuits are needed to achieve contact and communication, and operations of induction and cohe-sion: in both cases those circuits form machines or `sets'), and both are fragile (communication easily breaks down when one circuit is inter-rupted or messages fail to cohere). But they are also presented as dissimi-

lar, in so far as poetic communication is authentic, and everyday conversation, relayed by technology, is inauthentic (my experience of other people's mobile phones in public places confirms this). As a result of such dissimilarity a hierarchy is introduced, and is explicitly endorsed by the narrator in his theorising: one type, the poetic, is interesting, a source of inspiration and wonder, the other trivial. From this hierarchy the narrator derives a theory of art (as sublimation of erotic impulses) and an ars poetica. This position, which expresses a post-Romantic com-

monplace, is duly ironised: the narrator is too explicit, too insistent, too obviously manipulative, for us not to suspect that the whole exercise is conducted tongue in cheek. What remains to be done, therefore, is to assess the consequences of this irony.

We now fully realise that in spite of his Watson-like position of obser-ver and chronicler of Mr Shaynor's trance and the nephew's scientific experiments, the narrator is the real centre of the piece, the literal protagonist: the initial Watson is the eventual Mephisto. Or, the narra-tor as apparent witness and arch manipulator. Does he not, by moving back and forth from the shop to the back kitchen provide a physical link between the two situations? Does he not encourage Mr Shaynor in his psychic wanderings, even if this means sharply rebuking the nephew's calls for attention (` ̀ `Mr Cashell, there is something coming through here, too. Leave me alone till I tell you.'' ``But I thought you'd come to see this wonderful thing Sir,'' indignantly at the end. ``Leave me alone till I tell you. Be quiet'' ', p. 192)? And in his commerce with Mr Shaynor the narrator inscribes both situations. It begins with a trivial conversa-tion about the state of Mr Shaynor's health and his professional pro-spects, about the shop and its custom, not eschewing the naughty joke reserved for Victorian postprandial male company (Mr Shaynor chuckles as he relates the unwelcome consequences of the nephew's experiments:

Interlude 3 211

he set his apparatus in a hotel, unwittingly electrifying the plumbing, upon which the lady customers, on entering their baths, experienced a literal shock this is the opening of the story). And it ends in an inspir-ational conversation with the Poet. For the possessed Mr Shaynor, being in the position of the medium, is merely an intermediary. The real conversation occurs between the poet (both Keats and Shaynor his unwitting prophet) striving towards perfection and the narrator, who has now taken the position of the literary critic, who mentally encour-ages, and passes judgement.

But perhaps there is even more to the story than all this. A careful reading of the text shows that Mr Shaynor's possession by the voice of Keats is merely the tip of an iceberg of signs of all description. For not all signs (this is a Deleuzean note) are linguistic. What is truly mysterious in the story is not so much the sudden possession of Mr Shaynor as the fact that Keats's texts pervade the whole scene, and not only the words uttered in it. A few examples will show this.

The night is windy and cold. Mr Shaynor and the narrator are looking at the butcher's shop across the street:

`They ought to take these poultry in all knocked about like that,' said Mr Shaynor. `Doesn't it make you feel fair perishing? See that old hare! The wind's nearly blowing the fur off him.' (p. 193)

This combination of chill night and hare is, of course, a Keatsian echo of the very first lines of `The Eve of St Agnes':

St Agnes' Eve Ah, bitter chill it was!The owl, for all his feathers, was a-cold;The hare limped trembling through the frozen grass.3

But what to literary critics or the editors of the text are echoes of Keats (one editor notes that Èchoes of [`The Eve of St Agnes'] and other Keats quotations in the story can be found in the descriptions of Shaynor, the shop, and its environs'4) in the world of the story are signs. Being blaseÂ aesthetes, we will tend to read the text as a riddle testing our know-

ledge of Keats and our ability to recognise or guess echoes. But this dead hare (Keats's hare, though trembling with cold, is still alive) is a portent. A number of `Keats signs' make the text cohere: the physical appearance of Mr Shaynor's girlfriend and her somewhat authoritarian manner are reminiscent of Keats's Fanny Brawne. The church she suggests as her goal for her walking out with Mr Shaynor is dedicated to St Agnes. And


the allusion to Keats in the winking of the bubbles is, consequently, not only an allusion but a sign.

Once we become aware that we are plunged in a world of signs, where objects speak as garrulously as subjects, those signs multiply. For dra-matic irony also is a sign. Thus, when Mr Shaynor exclaims: `Brr! But it was cold out there! I shouldn't care to be lying in my grave tonight' (p. 186), his early death is speaking through him, a fact which does not escape the narrator, who immediately interprets the symptoms, another variety of sign: `[I wondered], while he spoke, into what agonies of terror I should fall if ever I saw those bright-red danger signals under my nose'(p. 186) (he is alluding of course to the blood spots on Mr Shaynor's handkerchief). We understand why the narrator is a doctor, a profes-sional reader of signs, and why it is that, at the beginning of the text, `for reasons of my own, I was deeply interested in Marconi experiments at their outset in England' (p. 183). We also understand the importance of the shrine, when Mr Shaynor censes the image of his beloved. The intercessor has become a worshipper, the whole of nature has become a book in urgent need to be read and interpreted, but also worshipped in wonder. This is why technological communication is so trivial, because it instrumentalises the Powers of nature the better to ignore their true significance, why it is only interesting when it fails (the blind dread-noughts out at sea recapture a sense of mystery, perhaps of awe, if not of `dread'), why when all becomes `loud and clear' the narrator goes to bed, because the Gods have departed.

There is only one further point to make clear: what exactly is a sign? Not a trivial question once we have understood that the story which on the surface is about communication, is also, and primarily, about signs and that the two cannot be equated. For our contemporary common-

sense concept of sign is linguistic and Saussurean: archetypal signs are those used for linguistic communication. This is where Deleuze is going to help us. Not only does he consistently insist that there are non-linguistic semiotics or regimes of signs, but he provides in his Proust et les signes5 a non- or pre-Saussurean concept of sign, where signs have little to do with linguistic communication, and which therefore may help us to read our text.

The bare lineaments of the theory are the following. (1) There are four fields (Deleuze calls them `worlds') where signs are encountered: the worlds of mondaniteÂ (`worldliness' in the sense one says that So-and-so is à man of the world'), of qualities and sense impressions (the exquisite feeling caused by ingesting a madeleine or listening to a musical phrase), of love (being in love turns one into a compulsive interpreter of signs)

Interlude 3 213

and of Art (the word is capitalised, as this is the climax in a progression towards perfect and dematerialised signs). (2) The reading of signs is a process of recherche, of search for a truth reached through an epiphanic encounter with the sign: for such encounter is an event, there is a certain violence in the sign that forces us to interpret it for truth. (3) Conversely, linguistic practice, as embodied in ordinary conversation, is decried as the best way not to encounter signs: language, at least ordin-ary, non- artistic (non-poetic) language causes us to miss signs. (4) The object of the encounter, on the other hand, is to reveal essences, from which emanates a meaning that lies beyond both the object that emits it and the subject that receives or captures it. Such reaching for the essence is achieved through a process of interpretation, which is also a process of learning: a sign is what we must learn to interpret. (5) The outcome of such learning is the creation of a style, through which the artist meta-

morphoses his material (the encountered signs) into essences a world is born whenever an essence is disclosed.

This is the earliest version of Deleuze's concept of style. It emphatic-

ally partakes of an aesthetics that I have called `high modernist', and will be discussed at more length in the next chapter. But it does inform our reading of ` ̀ `Wireless'' ', helps to make it coherent.

We can start with the status of everyday conversation as a stifler of signs: when words become loud and clear, signs vanish. On the contrary, the `conversation' of the narrator with a dreaming Mr Shaynor and a ghostly textual Keats is the culmination, the true encounter, of a com-

merce with a world peopled with signs the moment when signs are taken for wonders. And the signs that people this world are not merely the signs of Art: they are the signs of love (hence the importance of the young lady, and of the worshipping at the shrine) but also the signs of sense impressions. We understand why Kipling insists so much on the atmosphere surrounding the chemist's shop, on the wind and the chilly night. They are not only portents of Mr Shaynor's death, they are signs in their own right.

The shop, by the light of the many electrics, looked like a Paris-diamond mine, for Mr Cashell believed in all the ritual of his craft. Three superb glass jars red, green, and blue of the sort that led Rosamund to parting with her shoes blazed in the broad plate-glass windows, and there was a confused smell of orris, Kodak films, vul-canite, tooth-powder, sachets, and almond-cream in the air. Mr Shay-nor fed the dispensary stove, and we sucked cayenne-pepper jujubes and menthol lozenges. The brutal east wind had cleared the streets,


and the few passers-by were muffled to their puckered eyes. In the Italian warehouse next door some gay feathered birds and game, hung upon hooks, sagged to the wind across the left edge of our window frame. (p. 183)

This is not only a network of literary echoes (to Keats, to Maria Edge-worth's Rosamund, perhaps even to Dickens's east wind), this is the narrator describing a world of signs in so far as they take the form of sense impressions, affecting not only sight (the bright colours), smell (the rich and confused smell of the chemist's shop) and hearing (the howling of the east wind), but even, in true pre-Proustian fashion, taste (the sucked lozenges). This is Dickens's old curiosity shop turned into Ali Baba's cave.

Such profusion of signs, once they have been recognised for what they are, induces a process of compulsive interpretation and of learning. This is what the central episode, Mr Shaynor's trance and possession or inspiration, is about. The climax of this process is the creation or recre-ation of a style, Mr Shaynor's or Keats's, the object of which is to capture or disclose essences, which inscribe the unity of sign and meaning, which make sense of life through the immaterial signs of art, here the words of poetry.

We understand the mystic tone in which the narrator relates such strange phenomena: and we also understand that there is much more to them than mere possession by a spirit or the vulgar triviality of the spiritualist's seance. For it is in those terms that the chemist's positive nephew describes the failure of communication between the two dread-noughts: `Have you ever seen a spiritualistic seÂance? It reminds me of that sometimes odds and ends of messages coming out of nowhere a word here and there no good at all' (p. 199).

The nephew, however, has a point. He is not only a positive chemist, a scion of the great line of M. Homais, he is also an operator of machines. My interpretation of the story in terms of signs and wonder has forgot-ten and overlooked the machinic aspect of the phenomena. But here too, Deleuze is going to help.

5 Machines

The title of the story refers, if not to a machine, at least to a mechanical process or operation. Because it is placed between inverted commas, the process undergoes a metaphorical extension: the text is based on the consistent parallelism between the Marconi and the Keats wireless sets.

Interlude 3 215

The text itself is pervaded with mechanical metaphors of electricity and power, obviously, but not only those. Thus, when Mr Shaynor is almost at the climax of his trance, the narrator naturally uses a scientific simile: `Here again his face grew peaked and anxious with that sense of loss I had first seen, when the Power snatched him. But this time the agony was tenfold keener. As I watched it mounted, like mercury in a tube' (p. 197).

On the face of it, there is nothing unexpected in this: human com-

munication is normally metaphorised in terms of machine, in the shape of a circuit, of transmitter and receptor. In the commonplace `conduit' metaphor, one crams as much meaning as one can in one's words, which are then sent along a circuit, like a telegram or letters through the post. So it is no surprise that if the nephew is trying to listen in to Poole with the help of his apparatus, the narrator should establish contact with Keats with the help of a human transmitter. But this is too simple, even if, in the world of fiction, the metaphor of human transmission has become a metamorphosis: Mr Shaynor has literally become a wireless set.

In order to understand why this is too simple a picture, let us ask the question that comes first in all philosophical accounts of the world: what is this world made of? Sugar and spice and all that's nice? Or, in more Keatsian style, `candied apple, quince and plum and gourd' (this is Mr Shaynor repeating Keats, p. 195)? The answer is energy. The magnetic energy that makes the nickel filings cohere in their glass tube, the electrical energy of the battery that operates the secondary circuit and the Morse machine, but also tickles ladies in their bath, and the libidinal energy that moves an unconscious Mr Shaynor to wax Keatsian. This energy it is that turns primordial Chaos (the vague untapped Power that is òut there', waiting to be captured, harnessed into some sort of order) into a Cosmos, where the intensities of the `cËa' (which I described at length in my first chapter) are, if not straightforwardly domesticated, at least made significant (which means: through the active fabrication of signs). This is the world of flows and breaks or cuts with which we are familiar from L'Anti-êdipe. Flows of libidinal energies are interrupted and coded into significant order. This is what literally occurs in our text, both with the Marconi, and the Keats, set. The coherer is the archetypal energy machine, archetypically incarnated in the human coherer, Mr Shaynor.

So the world of our text is a world of machines in the Deleuzean sense. We saw in the last chapter that Deleuzean machines are not mechanical, not systems of interdependent cogs, but rather the organisation of a


neighbourhood, a micro-assemblage, the best illustration of which is a Heath Robinson machine. The comic version of such a machine is given in the first page of our text: it involves an aerial, a battery, a circuit of pipes that earth the electric current, and the naked bodies of various ladies entering their baths. In its more serious version, as operated by the nephew, it involves the organisation of a neighbourhood (again, various apparatuses, aerials, etc.), which coheres space into a plane of consistency, since contact is made not only with Poole but also àcross half South England' (p. 199) with Portsmouth and the dreadnoughts at sea. In its most striking version, the Keats wireless set, the machine becomes a true assemblage: there is always more than one machine, and the assemblage of machines has two aspects, a Machinic Assemblage of Desire (the libido of Mr Shaynor, combined with `the Hertzian waves of tuberculosis') and a Collective Assemblage of Enunciation.

Let us take a closer look at this last point. The Marconi set hardly produces individual or subjective utterances: the sentences are utterly conventional and/or trivial. When they are out of the ordinary (`Dis-

heartening most disheartening', p. 199 this exclamation is not what we expect from a naval officer), they are ascribed not to an individual speaker, but to a personified collective, the dreadnought. What is en-gaged in the overall process of communication is in fact a network, rather than individual speakers, who, in true Deleuzean fashion, are not users but parts of the machine. This is made entirely obvious when-

ever the circuit breaks, and some emit but fail to receive, while others do the opposite.

In the case of the Keats set, the fact that we have not an individual but a collective assemblage is entirely explicit. What the narrator's theory of the occurrence (which is proved to be right when Mr Shaynor wakes up) does is to describe the various elements of the assemblage, in their usual ontological mixture. In order for the Keats set to work properly a rather extraordinary happening a complex combination of elements is neces-sary: bacteria and illness (tuberculosis), bodies (a young man and his girlfriend), institutions (the profession of `druggist', as the narrator calls it), the subconscious substrata of the human psyche, striving for expres-sion, at the cost of distortion or sublimation, and a body of immaterial lines physically inscribed on paper by Mr Shaynor's mediumnic hand. This is a typical Deleuzean assemblage, in every way as complex as the feudal assemblage that provides him with his canonical example.

We understand that the metaleptic relation between Mr Shaynor and Keats (two speakers for the same text) is not a simple relation of posses-sion (which would resolve the paradox by providing a single author),

Interlude 3 217

but, in the world of the story, a more scandalous and paradoxical relation: they àuthor' the same text because they are parts of the same assem-

blage and are both spoken by it. This is why Kipling's text makes such extraordinary reading: because the solution suggested to the riddle of the story, even if it is ironic, is scandalous in so far as it is incompatible with common sense, even with the marginal common sense of `psychic phenomena'.

6 Conclusion

Of course, Kipling's text is not entirely faithful to such scandalous conclusion. It phrases it, but it also rejects it through its recourse to the banal diction of poetic inspiration and the magic of poetic genius. Indeed, the text develops around this tension, or even contradiction.

In Deleuzean terms, the text hesitates between the point of view of immanence (communication, even poetic communication, occurs through the coding of flows, the tracing of a plane of consistency on the plane of immanence, the construction of machinic assemblages) and the point of view of transcendence (the unknown Power that governs Hertzian waves and poetic inspiration is not so much òut there' as ùp there', on another level of being, in the realm of Platonic essences). As a result, the text constantly hesitates between a description of the occurrences in terms of machines and in terms of spirits. Is Mr Shaynor a different sort of coherer, but a literal, not a metaphorical one, or a medium? Ironically, Mr Shaynor, a positive chemist like the nephew, holds strong views about mediums: ` ̀ `But mediums are all impostors,'' said Mr Shaynor. . . . ̀ `They do it for the money they can make. I've see 'em'' ' (p. 199). Such positivism unwittingly supports the scandalous interpretation of the occurrences. This is the revenge of materialism and technology over the idealism of inspiration. But the narrator, of course, is laughing up his sleeve, having seen, and heard, Mr Shaynor in his mediumnic trance.

So this world is dual: it is a world of signs (intuition is better than technological communication), and it is a world of machines, which in Deleuzean terms are not merely mechanical (intuition is merely another form of machinic enunciation). This is the true meaning of the irony of the text, not a mere reversal (poetic intuition is so much superior to everyday communication), but a hesitation, or paradox, or Freudian denial: I know that poetic inspiration is the product of a machinic assem-

blage, like common-or-garden communication, but all the same . . .

6 Another Philosophy of Language: Style and Stuttering

1 Introduction

The theory of assemblages ought, at first sight, to be the climax and the end of Deleuze's philosophy of language: it is the most original of his various views on the workings of language, the one that makes the most decisive break with mainstream linguistics. But àssemblage' is not the last word in Deleuze's philosophy of language the last word, which is also the first, is style. And we reach not so much the acme of his thought about language as its point of highest tension, something, to use a term Deleuze never uses himself, of a contradiction.

The contradiction can be summed up thus. On the one hand, we have a quasi-Marxist concept of language, in terms of rapports de forces and collective assemblages, characterised by their dubious ontological mix-

ture a materialist but also a libidinal concept of language, which owes a lot to Guattari's contribution. Assemblages, as sites where forces com-

pete and coalesce, tend to increase and multiply, without privileging one genre or language-game over another. Such a theory is extremely powerful, not only in that it definitively breaks with the logic of repre-sentation and the dogmatic image of thought, but in that it finally eschews the centrality of the subject and firmly places the `depths' of the human psyche outside, in language, in its assemblages.

But, on the other hand, we also have an elitist view of literature, conceived as the acme of linguistic practice, together with a high-modernist, or avant-garde, position on literature as not merely a form of language, but as being mainly reflexively concerned with lan-guage.

The concept of style in Deleuze names this tension and aims to resolve it. It is doubtful whether it does resolve the tension (and we know

218

Style and Stuttering 219

enough about Deleuze to be aware that problems resist their solutions), but the attempt produces interesting results.

I shall concentrate on the concept of style in Deleuze's late work there is a chronological as well as a logical order in my three chapters on the new pragmatics but the concept has been with him throughout his philosophical life, with a first appearance in Proust et les signes.1 The core of the concept of style to be found in that book can be expressed in a formula of Deleuze's, offered in obvious contrast to Buffon's `le style, c'est l'homme': `le style, c'est l'essence'. Since the formula is rather cryptic, the chain of argument which it sums up must be briefly rehearsed.

Deleuze's reading of Proust (a strong reading if there is one) defines the `problem' of the Proustian oeuvre as the problem of signs (signs, as we saw in Chapter 2, appear in several types of world: the madeleine, or la petite phrase de Vinteuil are canonical signs). The signs that primarily interest Proust, and Deleuze reading him, are of course the signs of art. Like all signs, they initiate a process of learning, in that they need interpretation, but what they reveal is not merely objects (the referents of the signs) but essences. Essences are what art enables us to reach, or rather, the privilege and aim of art is to enable us to reach them. Essences are variously defined by Deleuze as absolute and ultimate differences, and as monads. An essence is a point of view on the world, each one expressing the world in an entirely different way, and creating subjects in the process. From that it appears that such essences must not be taken in their transcendent, Platonist form, but must be read as immanent (or the essence as the èssential oil' that gives the perfume its fragrance).

The question is, of course: in what form are essences embodied or revealed in the work of art? And the answer is, in the materials (matieÁres) the work of art is made of certain colours, or sounds, or words. We see what I have called Deleuze's `high modernism' creeping in: a work of art is not characterised by its contents or theme, but by the materials it is materially made of. And this is where the concept of style appears: style is the treatment of the materials that turns the work into a work of art; it is the establishment of unknown or unexpected relations between objects, through the organisation of materials. And it is defined in two moments: it works through metaphor (in so far as it establishes unknown relations), but those metaphors are èssentially' metamorphoses, when the objects related in style exchange their determinations and their names. Thus are a new world, a new point of view on the world, an essence reached. So the style is not the man, but the essence, and the chain of argument goes like this:


Signs ! signs of art ! materials ! style ! metaphor ! metamor-

phosis ! essence.

The àbsolute privilege of art', embodied in style, is the final adequacy between sign and meaning, essence and transmogrified material. At first sight, this theory of signs, art and essences seems totally incompatible with the theory of assemblages. Yet Deleuze tries to establish a link, precisely through the concept of style.

2 Style

In its traditional acceptation, for instance in Buffon, the concept of style involves notions of individuality (style is the original mark of the indi-vidual writer: it is idiosyncratic), of subjectivity (style is a deliberate construction: you work at it, you improve it, you master it, and this `you' cannot but be a subject conscious of her aims and of the means of her endeavour), of authorship (a true author is recognised by his indi-vidual style). All this, of course, as in Buffon, presupposes a concept of humankind that is singularly absent in Deleuze's philosophy, which is certainly not a version of humanism, not even an anthropology.

Why, then, since extreme distance must be taken, keep the term and extract a concept out of it on the face of it a strange and unex-pected decision? My hypothesis is that Deleuze remained faithful to the old-fashioned term in order to avoid its main competitor in con-temporary French philosophy, eÂcriture, writing. As used by Barthes, the concept of eÂcriture is, for Deleuze, probably too closely linked to the moment of structuralism, to the domination of the letter and the tyr-anny of the signifier. And as used by Derrida (in, for instance, L'Ecriture et la diffeÂrence),2 it is too language-centred. The advantage of style, since the centrality of language is one of the structuralist tenets Deleuze wishes to discard, is that it can also be used for painting (Rembrandt's style) and even to describe general behaviour (a lifestyle, a stylish dress).

In its post-Guattari definition, the Deleuzean concept of style is closely linked to concepts of assemblage and becoming. A style is no longer the privileged access to essence, it is an assemblage of enunci-ation (and as such always collective, even if associated with a proper name and a `celibate' author), and a becoming we remember that the main object of philosophical enquiry for Deleuze, more Hegelian in this than he would like to admit, is not the single `thing', or the static structure of objects, but always a becoming, a series of acts èxpressed


in a certain style'. So style is a name not for a form of diction (the choice of the proper, or the metaphorical, word), not for a structure of signi-fiers, not for a deliberate organisation of language, not even for the result of spontaneous inspiration, but for the discord, the disequilib-rium, the stuttering that affect language at its most alive. There is a clear link with the concept of minority: style it is that makes you a stranger in your own language, that opens up for you, as speaker or writer, the lines of flight that will allow thought to visit your utterance.

This does not mean that Deleuze is unaware of the word èÂcriture', or unwilling to acknowledge its existence: the word actually occurs in his texts, and frequently at that, but rather as a descriptive notion than a concept. Deleuze seems to prefer the verbal form of the word, èÂcrire', which suggests an activity, a process, and is ready to capture the dur-ation that becoming and the following of lines of flight demand. When one `writes', he claims, one reaches towards, or is captured by, style. Thus, Virginia Woolf's èÂcriture feÂminine' is not a separate form of writing, the expression of feminine identity, but the inscription, and at the same time the fostering, of a process of `becoming woman', which men as well as women have to undergo (for Deleuze and Guattari immediately insist that such a process is also very much present in the most phallocratic of male writers like Miller and Lawrence).3 In Virginia Woolf, this `writing', which amounts to a style, is characterised by qualities of impersonality, indiscernibility, imperceptibility:4 hardly the characteristic usually ascribed to modernist writing, or style, but qualities that have one point in common the capacity to take writer and reader along the lines of flight of becoming.

It would seem, then, that in Deleuze, `writing' is the site of a problem (`the problem of writing' is the object that Deleuze, in his preface to Critique et clinique, undertakes to tackle in that collection of essays),5 and style is the concept that names the problem. In L'AbeÂceÂdaire, in the È comme Enfance' section, the problem is indeed formulated as one of becoming: to write is to become, but not to become a writer (an obvious allusion to Proust); it consists in pushing language to its limits, not in recovering the trivial memories of one's childhood. Again, the very concept of style is a protest against theories of the individuality, and originality, of the author.

But what does style actually consist of? As one reads the Deleuze corpus in the order of writing (which is not the order of its translation into English), one can see the concept evolving, and the new concept forming, until it finds its final determinations in L'AbeÂceÂdaire and Cri-tique et clinique. A transitional formulation can be found in L'Anti-údipe,


in a passage which is (again) a eulogy of Anglo-Saxon literature, this time in terms of capacity for delirium, violence done to syntax, indul-gence in nonsense, the deliberate destruction of the signifier and the fostering of polyphony those Anglo-Saxon writers definitely have style.6 In the passage, Deleuze and Guattari oppose a form of literary criticism that attempts to read the text for its ideological positions (and they deplore, once again, the vagueness and confusion of the notion of ideology). Against those ideology-inspired readers, presumably Marx-

ists, and probably Althusserians like Macherey, they refer to Engels reading Balzac not as a politically reactionary author, not even as a mirror of French society, not (in other words) in terms of representa-tion, even indirect or distorted: a great writer, like Balzac, is one who allows out flows of energy that break the despotic (in the case of Balzac, the catholic) signifier and enter a revolutionary machine. And they add: `That is what style is, or rather, the absence of style asyntactic, agrammatical: the moment when language is no longer defined by what it says, even by what makes it a signifying thing, but by what causes it to move, to flow and to explode desire. For literature is like schizophrenia: a process, and not a goal, a production and not an expression.'7

The elements of a definition of style are already present here, together with the characteristic equivalence, or ambivalence, to which we shall have to return: style � non-style; style is best described as absence of style. We may take this apparent contradiction as the symptom of the constitutive tension in Deleuze's concept of style tension that does not disappear in later definitions, but animates the concept up to its last versions. In L'AbeÂceÂdaire, however, which has an `S is for Style' section, the characteristics ascribed to style are thoroughly positive: style means an original syntactic treatment of language, called stuttering or stam-

mering, and the capacity to take language to its frontiers with silence, but also with other media, notably music. This casts new light upon the strategy of Deleuze's reading of Beckett's television plays: the movement from the first to the third language (from a language which is still vaguely a language to a language which is no longer a language at all I seem to be speaking like the narrator in Watt) is, on Beckett's part, a movement towards style, the very movement of style with the para-doxical consequence that, as it reaches the acme of style, language dissolves altogether. We understand why style is also non-style.

In order to sum up Deleuze's thinking on style, I shall quote a passage from an essay in Critique et clinique that I shall analyse in more detail later:


When a language is so strained that it starts to stutter, or to murmur or stammer . . . then language in its entirety reaches the limit that marks its outside and makes it confront silence. When a language is strained in this way, language in its entirety is submitted to a pressure that makes it fall silent. Style the foreign language within language is made up of these two operations, or should we instead speak with Proust of a nonstyle.8

Style as a straining of language: a fascinating definition, but still somewhat vague. But even at this stage, there are three important consequences of the definition, which make it more precise.

1 Deleuze shares a belief with Chomsky although with a rather differ-ent import: he believes in the centrality of syntax. `The writer uses words, but by creating a syntax that makes them pass into sensation, that makes the standard language stammer, tremble, cry, or even sing: this is the style . . . .9

2 Deleuze insists on the impersonality of style, a tenet he shares with Foucault, the exact opposite of the commonsense conception of style. In one of the interviews in Pourparlers, while praising Foucault, his person and his behaviour as much as his work (style, we remem-

ber, is not simply a matter of language), he writes:

Even in a trivial situation, say when he came into a room, it was more like a changed atmosphere, a sort of event, an electric or magnetic field or something. That didn't in the least rule out warmth or make you feel uncomfortable, but it wasn't like a person. It was a set of intensities. It sometimes annoyed him to be like that, or to have that effect. But at the same time all his work fed upon it. The visible is for him mirrorings, scintillations, flashes, lightning effects. Language is a huge `there is', in the third person as opposed to any particular person, that's to say an intensive language, which constitutes his style.10

This passage makes us understand the paradox of style: how it can be perceived as the most individual inhabitation of language, to use Hei-

degger's metaphor, and yet not be ascribed to a person: there is a subject of style, but this subject is the end-product of a process of subjectivation (thus, the subject is not the origin, but the effect of her style: the author does not have style, it is style that has an author, that is inscribed, and in a way embodied, in an author's name), and this subject, both individual


(an ìnimitable' style) and collective (an assemblage is speaking) is in no way reducible to a person not even Foucault, charming as he was, was one, at least in Deleuze's portrait of him: ìf there is a subject, it is a subject without identity.'11

3 The linguistic characteristics of style in Deleuze are due to an aspect of language on which, after Hjelmslev and Labov, he tirelessly insists: language is in a state of continuous variation. Rather than various func-tions of language, he claims, we should speak in terms of regimes of signs, combining flows of expression and flows of content, and crys-tallising in assemblages, of desire or of enunciation. This continuous variation creates heterogeneity and imbalance: language is not a stable system. Style uses this aspect of the workings of language and thema-

tises it this is what in language it takes to the limit, on the arc that goes from stuttering to silence. In this matter, Beckett is the arch stylist:

Beckett's procedure . . . is as follows: he places himself in the middle of the sentence and makes the sentence grow out from the middle, adding particle upon particle . . . so as to pilot the block of a single expiring breath . . . . Creative stuttering is what makes language grow from the middle, like grass; it is what makes language a rhizome, instead of a tree, what puts language in perpetual disequilibrium.12

We shall have to come back to the intensive line of syntax. I believe we can now understand why the concept of style is essential

to Deleuze's philosophy of language. It is a node in this rhizome of concepts where many roots converge, where many offshoots grow. The concept inseparably links Deleuze's philosophy of language (flows and breaks; segments, deterritorialisation and lines of flight; variation, imbalance and the centrality of syntax; collective assemblages and impersonality) and something that is emerging as an independent poetics, whose internal tension is becoming more and more percep-tible: on the one hand, we have that modernist poetics of agrammati-

cality and silence; on the other, a poetics of impersonality and assemblage.

I would like to end this on an illustration. It would be too easy to demonstrate the relevance of Deleuze's analysis of style by quoting a text from his modernist canon. So let us attempt a more difficult case, away from avant-garde writing. This is an extract from one of the


forgotten masterpieces of English literature, and I hope the chute of the passage will provoke as much mirth in you as it does in me:

Somewhat under lower middle height, my father, even as a boy, had been inclined to corpulence, a characteristic, inherited by myself, that he succeeded in retaining to the end of his life. Nor did he ever lose or not to any marked extent either the abundant hair that grew upon his scalp, his glossy and luxuriant moustache, or his extraordinarily powerful voice. This was a deep bass that in moments of emotion became suddenly converted into a high falsetto, and he never hesitated, in a cause that he deemed righteous, to employ it to its full capacity. Always highly coloured, and the fortunate possessor of an exceptionally large and well-modelled nose, my father's eyes were of a singularly pale, unwinking blue, while in his massive ears, with their boldly outstanding rim, resided the rare faculty of inde-pendent motion.13

The hero of this autobiography, and author of this extraordinary description of his father, is a worthy descendent of Mr Pooter.14 There is apparently no question here of taking language to its limit in silence: not only is Augustus endowed with the same sonorous vocal organ as his genitor (as he himself would phrase it), but he is also afflicted with self-satisfied garrulousness, and, to put it vulgarly, will not shut up. But the author of the introduction to the Penguin edition is right in noting that in Augustus Carp `language is itself the character'. For this language is not merely the string of clicheÂs one can expect from a pompous nonentity, it shows style, it has a life of its own, a kind of excess, that is, a form of verve that constantly threatens to make it overflow into nonsense. And that is what stuttering is about, and a certain form of taking language to its (here, hilarious) limit. And it also appears that the centrality of syntax to style is not a vain claim, as Augustus Carp is created by a number of syntactic idiosyncrasies, not least the `solecism' known as the ùn-attached participle', on which the last sentence of our passage begins. Here is another example: `Born at half-past three on a February morning, the world having been decked with a slight snow-fall, it was then that my mother's aunt, Mrs Emily Smith, opened her bedroom door':15

Augustus is not telling us about the birth of his great-aunt, but about his own.

A description of stuttering is, therefore, the next logical step in our account of Deleuze's concept of style.


3 Stuttering language

3.1 Incipit

Let me begin this section with an unashamed value judgement, which is also a confession. The best incipit I know is the first sentence of Beckett's Murphy:

1 The sun shone, having no alternative, on the nothing new.16

This is indeed a gem: a king in incipitdom, a paragon of incipithood, the climax of a progression in incipitity. How easily it outshines the trivial opening of `Longtemps, je me suis coucheÂ de bonne heure.' Who wants to bother about going to bed early when the sun is shining so bright, when the revelation of the poetic event leaves the reader dazzled?

Yet, this incipit, which is clearly the best, is also the worst: the monu-

ment is a ruin. Trivial as it appears at first sight, the humble going to bed does an excellent job. It contains, as a word within a word, the crucial word ̀ temps', thus announcing a ̀ recherche', and of some length too, since that time is `long'. In other words, it initiates a narrative programme, which it will take many tomes to fulfil; it is redolent with presupposition (the tense of the verb indicates that the process is no longer current but why?) and duly provokes questions and expectations.

Not so with Beckett's incipit. Because it is a gem, it is entirely self-sufficient, cut off from any context (it could be the incipit to a parody of the adventures of Tarzan). It has all the qualities of a self-quotation: it is not the first link in a narrative chain, but a fragment, an aphorism, what the French call ùne sentence', that is, a maxim. It stands alone at the threshold of the text, but not actually as a threshold, not even an outpost, rather a statue, independent of the rest of the edifice. Or again, to change the metaphor, it does not engage the maculate concep-tion of a fictional world: it is not, as a good incipit always is, the annunciation to an expectant reader of the imminent birth of a narra-tive universe. It does indeed seem to start with the optimism of the narrative incipit, through a confident thetic statement, `the sun shone', that is, not with etymological adumbration, but with the promise of a bright new world (this is not the beginning of the creation scene in Frankenstein: Ìt was on a dreary night of November . . . '). But it immedi-

ately takes it back, through the immemorial trick of negation. Yet this negation is not used to create atmosphere, as in the famous incipit to Jane Eyre: `There was no possibility of taking a walk that day.' But it has


semiotic value, being an abyme commentary on the incipit itself: `no alternative' implies no choice, therefore no Saussurean value, therefore no meaning. This is far worse than Jane Eyre: this meta-narrative neg-ation scuttles the ship of words that has just been launched a self-destroying incipit that immediately ruins the expectations it evokes.

Yet, I have not changed my mind, and I would like to maintain that this is the best incipit ever. In order to grasp both aspects of this para-doxical reaction, let me explicate the three paradoxes implicated in the incipit.

First, an incipit is supposed to introduce the theme of the narrative by answering as many as possible of the following rhetorical questions: quis, quid, ubi, quando, cur, quomodo. This our incipit signally fails to do. It only answers one, quando, and in terms so vague and trivial (`the sun shone') as to be totally uninformative. Even the incipit to Tess of the d'Urbervilles, Òn an evening in the latter part of May . . . ,' does better than that. Yet, paradoxically, the very triviality of this quando becomes an asset when through the work of figurality it becomes a quis, a per-sonification, an allegory; this sun which has no alternative may be as sulky as it is sultry it is the hero that one expects to find in the first sentence of a novel. By `figurality' I mean what Laurent Jenny calls `le figural', the linguistic twist that jostles the reader's attention (his canon-ical example is the phrase, which he found in a translation from the Chinese, ùne terrasse en printemps' one normally says èn hiver' and èn eÂteÂ', but àu printemps').17 In the case of our sentence, the figurality resides in the illicit combination of the [ human ] (the sun) and [�human ] (being faced with an alternative) semantic features. And, of course, this sun, turned into a hero of the piece by a turn of phrase, is no hero: no identification, no interpellation takes place here. The reader is not convoked at a place (as are the readers of Proust, BronteÈ and Hardy) she is kept at a distance, with deliberate flippancy.

I have not mentioned the second half of the incipit yet. But it is perhaps what makes the sentence so memorable, what turns it into an apophthegm. The paradox, here, is by no means new but we have always-already known that there is nothing new under the sun: it is the paradox of the negative object, of the denied world. The purpose of an incipit is to create, and begin to furnish, a fictional world. The metaphor of furnishing is essential: what distinguishes a fictional world from the possible worlds of the logician and the philosopher is not merely that it is never complete, and is not always consistent, but also that it is `furnished', that it is the object of a positive and selective creation (it is not given en masse as the range of a function). But the fictional world


here created has hardly received its first piece of furniture, a sun, when it is immediately dissolved through negation. Linguistic analysis of the enunciation variety tells me that `the sun' is determined by deictic pinpointing (fleÂchage situationnel), an operation made possible by the existence of only one possible referent for the phrase. What we have here is an incipient cosmos, and a mimetic one at that: this is indeed a world, since it revolves around a sun, and it is very likely our world, since the sun is addressed with the easy familiarity of acquaintance. But this incipient cosmos is immediately shattered by the non-existence of the said world: what is described is the `nothing new', by which I am at liberty to understand `nothing'.

Only if I somehow force the phrase, though. Hence my third paradox. There are two interpretations of the end of the incipit. Either it denotes inexistent objects (`the nothing new'), thus sketching a negative cosmol-

ogy, rich in Meinongian quasi-beings (like square circles and Fortunate Isles: a fictional world, you see, does not have to be consistent), or it marks the triviality, to the point of absurdity, of too familiar objects (`the nothing new'). And there is no doubt that the second interpretation would be the preferred one, at least if the clicheÂ were not upside down. For the figurality of this incipit does not only reside in the illicit combination of semantic features, but also in the unexpected twist given to a well-

worn phrase, thus releasing new potentialities of meaning. This twist is what the Deleuzean phrase, `stuttering language', seeks to describe.

How does the stuttering work? We start with the hoary clicheÂ, which has the shape of an existential sentence:

2 There is nothing new under the sun.

An existential sentence has one interesting semantic feature: the structure does not involve presupposition. It is linked to the èxtraction', that is, the presentation value of the indefinite article. `There is a cat on the table' implies that this is the first mention of that cat (compare with `There's the cat, on the table', which is not an existential sentence, but involves anaphoric mention and deictic pointing). My existential sentence can be transformed into an equally presupposition-free vari-ant:

3 Nothing is new under the sun.

The difference is that instead of zero-topicalisation (the meaningless `there'), we have topicalised the negative indefinite pronoun, `nothing',


a topic which has the semantic value of a focus. But suppose I transform the sentence once more by displacing the copula:

4 Nothing new is under the sun.

This dislocation cancels the existential value of the sentence and intro-duces judgement and presupposition. It presupposes the prior existence of an entity (albeit a self-contradictory one), the `nothing new'. A whole negative ontology is generated by the simple dislocation of the copula.

Of course, the phrase `the nothing new' is problematic, its syntax dubious. (Article � Pronoun � Adjective), in that order, is not a licit combination, and requires reanalysis. The problem is that at least three are possible:

i (Article � (Pronoun � Adjective) � (Adjective � Noun)), or `the new that is nothing'.

ii (Article � (Pronoun � Adverb) � (Adjective � Noun)), or `the in-no-way new'.

iii (Article � (Pronoun � Noun) � (Adjective � Adjective)), or `the nothing that is new'.

Interpretation (i) is too out-of-the-way to be a real contender, as a result of which the stuttering takes place between (ii) and (iii) a variation from (ii) to (iii), whereby the nihil, by being spoken, acquires a type of existence (one recalls Max MuÈller's theory of the origin of myth in overextended metaphor, or the ontologising of linguistic expres-sions). This in turn relies on a general characteristic of human language, the centrality of negation the capacity of human language to state not only what is but also what is not.

This stuttering, which makes us put on an equal footing interpretations (ii) and (iii) (both are syntactically dubious, (ii) is semantically closer to the preferred meaning of the original sentence), is comforted by the operation of Jakobson's `poetic function' of language. Our incipit has all the marks of poetichood: the alliteration in `n', the incremental repeti-tion of ̀ no' as ̀ nothing', and the surprising focus that plays on the expec-tation created (`the sun shone . . . upon the world') and immediately disappoints it (` . . . on the nothing new'). This device was already used, in a less radical fashion, by Lewis Carroll in `The Walrus and the Carpenter':

The sun was shining on the sea. Shining with all his might:


He did his best to make The billows smooth and bright And this was odd, because it was The middle of the night.18

Carroll's personification is pleasurable and playful. Not so Beckett's. This is no Freudian negation, meant to let us reinstate some positive being through interpretation: this is the position of non-being as being a metaphysical statement which, of course, does not so much concern `the world' as, en abyme, discourse, and especially literary discourse: `there is nothing new under the literary sun' is itself a hoary clicheÂ. This, therefore, is a self-fulfilling and self-ruining clicheÂ. The writer seems to distance himself, in a position of ironic disenchantment, from an exhausted tradition, and yet he is an integral part, merely another link, in that tradition. Or again, the incipit is original, it is new in so far as no one had twisted the clicheÂ that way before, and yet utterly clicheic, as it consists of the mere twisting of a long-dead clicheÂ. And this clicheic aspect is in itself the site of a paradox, in the form of reflexivity. This is a second-order clicheÂ, one that marks itself as a clicheÂ, and thus escapes the clicheÂ aspect that it thematises, even as the cre-ation by an incipit of a fictional world is a creation of nothing (of something that merely exists in words, not in deeds and facts), and yet a creation of the nothing, a true creation: the demiurge does not utter fiat lux, but fiat nihil, et nihil fuit. Did I not say that negation lies at the heart of human language?

This ponderous metaphysical game is played through stuttering. Stut-tering language is the name of the work or play of negation at the linguistic level, which entails consequences, in the shape of variation and the blurring or subversion of constraints at the rhetorical or stylistic level, and eventually points towards the utterly Other which human language, at its limit, enables us to glimpse, beyond the world, at the metaphysical level. But one question remains to be answered, which conditions the whole progression: what is the source of this stuttering of language, or again, who or what is stuttering? Does the stuttering occur primarily at the linguistic, or at the stylistic level?

3.2 Stuttering language

The concept `stuttering' is present throughout Deleuze's oeuvre: it makes a brief appearance in Logique du sens and in Mille plateaux. But it plays a major role in his later works, and is at the centre of the essay `BeÂgaya-t-il'.19 The essay starts with the attempt by second-rate novelists to pro-


duce variations on the `he said' that frames their characters' speech. What really interests Deleuze, of course, is cases where the speech itself, and its contents, are affected by stuttering, where stuttering is shown, not told, in other words where the writer herself, and not merely the characters, becomes a stutterer, `beÁgue de la langue'. The canonical for-mulation is nominal, `le beÂgaiement de la langue', the stuttering of lan-guage, which has the advantage of being ambiguous between a subjective and an objective interpretation of the genitive: someone makes language stutter/language stutters. The English equivalent, `stut-tering language', offers a similar ambiguity, between a present participle (`language that stutters'), the licit or preferred interpretation, and a gerund (`someone/something stutters language'), an illicit interpret-ation, the ghost of which the work of the figural, or of stuttering, nevertheless evokes. This work of figurality I have attempted in the title of this section, an illicit turn of phrase, which Deleuze himself never uses, but which has the advantage of being an example of what it states: `beÂgayer la langue', a grammatical monster which language nevertheless produces without the least effort, is itself a perfect example of stuttering language. And it explicitly raises the question of the recov-ery of the absent subject of the infinitive: who stutters? Is it a he or an it? Language, or the poet? As we shall see, Deleuze sometimes points at a `he' I prefer the ìt' myself.

Yet Deleuze's answer is by no means the traditional answer that she, the poet, as author of the text, is the source and mistress of a deliberate stuttering. He does formulate his answer in terms of style, but only after a long detour, where he sketches nothing less than a new non-Saussur-ean, non-Chomskyan path for linguistics, a path with which we are now familiar. His point of departure is indeed that stuttering does not occur at the level of parole, of individual discourse, but of langue, the collective `system': it is no mere playful or flippant misuse of language, but something that runs much deeper, and affects both the workings of langue and the status and function of langage. In fact, he seems to replace the langue/parole dichotomy on which, since Saussure, mainstream lin-guistics has been based, with a langue/langage connection or mixture. Stuttering, he claims, is an affect of langue, not an affection (that is, something that affects or afflicts) of parole the word àffect' is used deliberately to reintroduce the tension of desire within langue: langue not only stutters, but also trembles; it is not the cold, neutral (neutered) and indifferent system of the synchronic coupe d'essence. As a result of this tension, of this stuttering, langue reaches out towards the limit of language, that is, towards the partition where language gives place to its


other, silence. In other words, Deleuze moves from a linguistic question (how can we extract the system of langue from the diversity of parole?) to a metaphysical question (why is there language rather than nothing, that is, than silence?), a movement which, incidentally, was already that of Beckett's incipit. The role of poetry, according to Deleuze, is to allow this passage: when langue is the same thing as parole, as he claims, it is because parole is poetic, reaches out towards silence.

The concept of language (rather than langue) that emerges from the essay is definitely non-Chomskyan, and corresponds to the other phil-osophy of language, or new pragmatics, expounded in the last two chapters. Language in this essay is, unsurprisingly, rhizomatic rather than arborescent, minorised and variable rather than systematic and stable. Two other characteristics, however, deserve further development.

The main one is, precisely, stuttering, which takes two forms. This is an acknowledgement of the minority of stuttering: it presupposes (in order to subvert) the main mode of structuring of systematic language, the opposition between vertical paradigm and horizontal syntagma. The paradigmatic axis works according to exclusive disjunction (`Did you say ``pig'' or ``fig''?,' as the Cheshire Cat says to Alice), and the syntagmatic axis according to progressive connections: if I start a sentence with the word `pride', my next choice is slightly restricted, but only slightly: by the time I have reached, by a series of choices, the string `pride comes before a . . . ', my last choice is no longer a choice it would appear that, in spite of Chomsky's attack, the syntagmatic sequence is a Markov chain. We understand once more why Deleuze ascribes considerable value, as we have seen, to this usually despised mode of syntactic analysis. And we know there is a qualification: the final meaning is given retrospectively, from the vantage point of point de capiton. In spite of the pressure of clicheÂ, I can always end the above string by a word other than `fall', thus giving the set phrase a tropic twist.

Since stuttering introduces slippage or subversion within systematic langue, there will be two kinds of stuttering: the paradigmatic stuttering of inclusive disjunctions (where `pig' and `fig' are no longer in contrast) and the syntagmatic stuttering of reflexive connections, where points de capiton multiply and combine, to create words within words or sen-tences within sentences. Here is the crucial passage where Deleuze defines the two stutterings:

Language is subject to a double process, that of the choices to be made and that of the sequence to be established: disjunction or the selection of similars, connection or the consecution of combinables.


As long as language is considered as a system in equilibrium, the disjunctions are necessarily exclusive (we do not say `passion,' `ration,' `nation' at the same time, but we must choose between them) and the connections progressive (we do not combine a word with its own elements, in a kind of stop-start or forward-backward jerk). But far from equilibrium, the disjunctions become included or inclusive, and the connections reflexive, following a rolling gait that concerns the process of language and no longer the flow of speech. Every word is divided, but into itself (pas rats, passions rations): and every word is combined with itself (pas passeÂ passion). It is as if the entire language started to roll from right to left, and to pitch backward and forward: the two stutterings.20

The examples come from one of Deleuze's favourite poets, Gherasim Luca, a Romanian writing in French: in the poem, the speaker literally stutters for several pages before uttering his point de capiton, `je t'aime passionneÂment'. In the same vein, one of the mock-exam questions in 1066 and All That reads: `Tory acts, factory acts, unsatisfactory acts'. The consequence of this dual stuttering is, of course, that the contrast it uses as its starting point, between paradigm and syntagma, is dissolved. When language stutters, inclusive disjunctions are not more distinct from reflexive connections than `pig' is from `fig'. We are left not with a tree-like hierarchy, but a rhizomatic proliferation of grammatical paths and semantic virtualities.

The result of such dissolution is another conception of syntax, not as a hierarchy of structures, an architecture of modules, a system of prin-ciples and parameters, but as a line, a `ramified variation', with its curves, its bends, its deviations. A dynamic line worming its way through the plane of immanence of language, slaloming, if you pardon the rather bathetic metaphor, through disjunctions and connections. A syntax neither `superficial' nor `formal', but en devenir, a creative syntax that gives birth to the foreign dialect of minority within the standard dialect. This syntax is not defined through the stability of a system of rules, but through its tension towards silence, towards the ineffable, towards the limits of language. This, Deleuze claims, is the syntax of cummings's too famous `he danced his did', of Melville's Ì'd prefer not to' (to which he devotes an essay, `Bartleby, ou la formule', in the same collection), of Artaud's mots-souffles or mots-cris. This, of course, has nothing to do with mainstream linguistics, as framed, for instance, in Milner's axioms. With Deleuze, the poet-speaker has nothing of the angel (she is very much involved in her Lebenswelt), diachrony returns within synchrony


with a vengeance, rules are based on continuous variation, they are not merely fuzzy or locally defeasible, arbitrariness is replaced by intensity, and the link between signifier and signified is no longer stable. In poetic discourse, but also in dialects, jargons, slogans, pathos and ethos return to stutter the logos from which mainstream linguistics sought to exclude them.

The conclusion is that the essential part of the study, not merely of language but even of langue, is no longer directly syntax (the Chomsk-

yan thesis of the centrality of syntax is, if not abolished, at least modi-

fied) but syntax as mediated by style, in its three characteristics: (i) style is the name of the foreign dialect of minority within the native tongue; (ii) its first operation is the tension, the intensifying of langue, which causes the two stutterings; and (iii) its second operation is the reaching out for the limit of langage, where speech is confronted with silence.

Stuttering for Deleuze is, therefore, an equivalent of poetic language. The problem with this is that it is a metaphysical answer to a metaphys-

ical question (why is there language rather than nothing?). The hero of stuttering, the philosophical character needed to support the concept, is the exiled poet, who subverts langue and aims at the noble form of silence, the silence of the ineffable, as opposed to the mute inglo-rious silence of stupidity. And the problem with this, of course, is that it is a Romantic pose. This induces me to explore the other pole of the `he'/ìt' contrast. Let us forget for a moment that poets stutter language (even if Deleuze forcefully demonstrates that they do). Let us wonder what happens when language itself stutters.

3.3 Colemanballs

I am looking for unpoetic style, for speech without style, writing degreÂ zeÂro: I need it to satisfy myself that the danger of authorial poeticity (which would take me back to Deleuze) is avoided. I believe I have a chance of finding this in the speech of sports commentators, where language may well be at its most ordinary, clicheic and commonsensical. We all remember those interminable lists of football results on the Satur-day night sports news: Blackburn Rovers, three, Plymouth Argyle, nil.

But in fact I am wrong, as the sports commentary given on the spot, dans le feu de l'action, is full of the pathos of excited speech and is, consequently, albeit clicheic, highly metaphorical. If I may risk another confession, my taste for metaphor, poetic and otherwise, originated before I was ten, when I was a keen reader of the accounts of the Tour de France published by the newspaper L'HumaniteÂ by their sports com-

mentator, one Abel MicheÂa. That man was a folk poet; he did not merely


possess the gift of the gab, he also had a gift for florid metaphor, which turned each boring stage of the Tour into material for a canto from the Iliad. It would appear that even if the situation is as matter-of-fact as can be, pathos disrupts the fixity of systematic langue and of clicheÂ.

So let us add another parameter, to be rid of folk poets and of style. Let us consider only involuntary productions, involuntary in the sense that the reader is entitled to decide that the speaker is not aware of exactly what he is saying, and would not have said it had he been. In other words, I am trying to restrict my corpus to the sports commentator's slips of the tongue. But here I find myself on fairly familiar ground. Such slips of the tongue, in Britain at least, have a name: Colemanballs. Terry Coleman was a BBC sports commentator, who talked into the microphone for hours on end in the Saturday afternoon `Grandstand' programme a situation where slips of the tongue are as unavoidable as they are understandable. They became famous, were collected in a special Private Eye column, published in booklets (there is even a linguis-tic study in a learned volume21). As its patronising name indicates, a Colemanball is a bloomer, an involuntary stupidity or platitude, of which the speaker must have been ashamed, or proud, afterwards. The front cover of the collection I have consulted22 prints the three following examples:

5 With the very last kick of the game, McDonald has scored with a header.

6 If Tchaikovsky was alive, he'd turn in his grave.7 Only one word for that: `Magic darts.'

Only the fact that they have actually been uttered, that they have acquired the force of established fact, can save what would make rather paltry jokes. Colemanballs are not instances of wit, or of poetic lan-guage. Nor are they, at first sight, instances of stuttering. There is no intensity in, or disruption of, syntax here. In (5) a clicheÂ metonymy (meaning `the end') clashes with a synecdoche, equally clicheic (a header is a type of goal): this is a simple case of mixed tropes. In (6) we have a mixed clicheÂ, with simple substitution:

i If Tchaikovsky was alive, he'd be shocked. ii If Tchaikovsky heard this, he'd turn in his grave.

All you need to do is combine the protasis of (i) with the apodosis of (ii): the common syntactic structure encourages the semantic slippage.


In (7) the second part of the sentence, which consists of two words, gives the lie to the first. The general name of the game is the clash between syntax as a Markov chain (a tendency encouraged by the `set' quality of the clicheÂ as one speaks of `set phrases') and the retrospective ascrip-tion of meaning from point de capiton. Having said this, I must admit that it corresponds to Deleuze's definition of stuttering as inclusive disjunction (two contradictory elements of the same paradigm are pre-sent in the same syntagmatic chain), or reflexive connection (the whole fails to cancel the independence of the first part of the sentence, as a result of which there is a semantic clash between first and second part, or between part and whole).

Consider the following list of Colemanballs, which all have the same structure:

8 It's especially tense for Parker who's literally fighting for a place in an overcrowded plane to India.

9 Tavare has literally dropped anchor. 10 His tail is literally up. 11 A wicket could always fall in this game, literally at any time. 12 Bulgaria were quite literally not at the match. 13 John Wayne a man literally larger than life. 14 We are literally giving Zeta One movie cameras, together with a

money-back guarantee.15 And Alex has literally come back from the dead.

Phrases (8) to (11) concern cricket, (12) football and (15) snooker (the latter example is illustrated by a skeleton playing snooker). It is easy to understand how these Colemanballs work. In each case, a blatantly metaphorical turn of phrase (so much so that, without context, it would be impossible to know what the statements are about) is intensi-fied by an adverb, `literally', which, as we saw in Interlude 2, is the worst possible intensifier since (a) it is used metaphorically itself; (b) its meta-

phorical use is self-contradictory; and (c) it ruins the metaphor it is supposed to intensify. Take (14), for instance, where `give away', a metaphorical gift, is brought back to the world of hard cash by `money back', and where the use of `literally' only makes the contradiction more acute and visible. It is the clash between those contradictory meanings, and the tension they imply, that causes mirth in the audience.

And of course we remember

16 Lily, the caretaker's daughter, was literally run off her feet.


This is not a Colemanball, but a canonical example of high literature. In this complex use of reported speech, language stutters: the line of speech is wayward and sinuous, characterised by bifurcation and ten-sion, as the minor working-class dialect threatens the educated major dialect, as the inclusive disjunction of an intensifier which literally means `literally' and metaphorically means `metaphorically' subverts the requisite stability of meaning. Even the resultative construction involves a curving and serpentine syntax rather than the established values of the system.

This short detour through the incipit to `The Dead' enables us to go back to our corpus of Colemanballs with a new attitude. And we will have no difficulty in finding examples of our four characteristics of Deleuze's non-Chomskyan concept of language.

i Language as continuous variation.

17 It's so true to life it's hardly true.

Here, the semantic field of `true' is not determined by the couple of antonyms `true'/`false', but by variation over a range of values, captured by semantic theories such as prototype theory or Culioli's concept of notion (where the gradient goes from the true to the not really true, to the really not true, to the untrue).

ii Minor v. major dialect.

18 I have other irons in the fire, but I am keeping them close to my chest.

19 When it's all said and done, that's the moment when the talking has to stop.

Phrase (18) is a case of mixed metaphor, (19) of mixed clicheÂ. Can this not be taken as incipient polyphony, a tension of dialects, a still small voice producing a contradictory trope to counteract the resounding platitude of the first half of the sentence?

iii Exclusive disjunction and reflexive connection.

20 And how long have you had this lifelong ambition?21 Well, Wally, I've been watching this game both visually and on TV.


22 Even when you're dead, you shouldn't lie down and let yourself be buried.

Phrase (20) (my favourite) is an excellent case of inclusive disjunction, since the very form of the question ought to exclude the adjective `lifelong', and (22) (which is the same thing) is a case of reflexive connection, as the second clause, by developing the isotopy of death, looks back reflexively on the first. As for (21), it is close to a mot d'esprit (as indeed (20) could be an instance of irony) a derogatory remark on the evils of television: compare Carroll's Snark, who is charged with àiding, but not abetting'.

iv Intensive syntax.

23 He just can't believe what's not happening to him.

This is very much in the vein of Beckett's incipit: this is the intensive line of serpentine syntax (a famous example of which is the `Duchess's sentence' in Alice in Wonderland), where multiple negation provides an image within langue of the limit that separates language from the silence of the ineffable.

24 I would advise anyone coming to the match to come early and not leave until the end, otherwise they might miss something.

This, of course, is straight out of Beckett's Watt, and reminds us of the advice the King of Hearts gives the White Rabbit: `Begin at the begin-ning, proceed to the end, then stop.' Here the sinuosity, not so much of syntax as of narrative, works against semantic coherence by subverting the hierarchy of the constituents: once again the point de capiton ruins the Markovian progress of syntax. And it appears that Deleuze's stutter-ing, far from being an effect of poetic language, is an integral compon-

ent of language. In paradigmatic stuttering, the whole paradigm (or rather too much of it) is actually present but is it not somehow always present in absentia, as a virtuality of meaning? It is this presence of the absent that stutters language. In syntagmatic stuttering, the Mar-

kov chain organises the string of words into a coherent whole, and the point de capiton deconstructs this organisation to reconstruct it, or not. Such inevitable but unnecessary reanalysis also stutters lan-guage.

I leave the last word to a Colemanball:


25 For people who like that sort of thing, that is the sort of thing they like.

I only wish to note that this is exactly what Miss Jean Brodie, in Muriel Saprk's eponymous novel, has to say about girl-scouts. This withering comment is the acme of wit: the platitude of tautology is always recov-erable, first as lapalissade, then as irony and wit. Having added Spark to Joyce, Carroll and Beckett, I can only conclude that Terry Coleman finds himself in good company.

3.4 Monstre de langue

I am not trying to pass Terry Coleman for a poeÁte maudit: I am trying to define the function of the poet, in his relation to language, and to use for this purpose the philosophy of language that Deleuze has be-queathed to us. I am trying to go from his subjective `x fait beÂgayer la langue', where the poet is granted the classic status of author, to the monstre de langue, `stuttering language', an example of double syntax, which ascribes to the poet a different status, that of an actor occupying a place designated by a complex pragmatic structure (the ALTER structure I described in Interpretation as Pragmatics). The poet is interpellated at the place of speaker by language, in the shape of the text she writes she counter-interpellates, in the freedom of linguistic and poetic impos-

ture, the language that interpellates her. To use another metaphor, of which Deleuze was fond, the poet stutters the language that stutters him, because he is like the surfer who rides the wave that plays with him, so that one never knows whether a surfer attempts, or tempts, his wave.

4 From stuttering to style: the construction of a concept

The `BeÂgaya-t-il' essay is the site of the most explicit, and most compre-

hensive, attempt by Deleuze to construct a concept of style, that is, to resolve the tension inscribed in the `style � non-style' equivalence. So I shall look in a little more detail at the structure of the argument, in the light of Deleuze and Guattari's theory of the concept, as expounded in Qu'est-ce que la philosophie? We remember that for them there are three moments in a concept: the determination of the plane of immanence, where the concept frolics and thrives; the enumeration of a list of characteristics, or determinations of the concept; and the meaning of the conceptual character that acts the concept out. My question is: how is this programme fulfilled in `BeÂgaya-t-il'?


There are three movements in the essay, or rather three strands, woven together. The first, conducted through covert allusion or explicit mention, consists in an excursus through the specialised field of linguis-tics. This reminds us of the problematic nature of language for Deleuze and it is the first indication that the plane of immanence on which the concept of style is deployed, and which it helps to trace, is the field of language. The second, which confirms the problematic aspect of the plane, is the construction of literary canon: it demonstrates that lan-guage is too serious a problem to be left to linguists; it also names the conceptual character that corresponds to Deleuze's concept of style: the fou litteÂraire or poeÁte maudit, the hero of Deleuze's avant-gardist poetics. The third lists no fewer than eleven characteristics of the concept of style, which can be neatly arranged on a cline, the cline of style that goes from language to silence.

The essay begins, as we have seen, on an indirect note: a reflection on the desperate effort of the second-rate novelist not to repeat himself when reporting dialogue, by finding variations on the `he said'/`she said' phrase, such as `he whispered', `he stammered', etc. The interesting moment, of course, comes when the author does not tell us about the stuttering but shows it to us, and this not only in the reported speech of the character, but in her own language: when she makes language stutter. Then indeed do we go from saying to doing without saying and from this to the limit, where we reach style, where saying is doing

where words themselves are made to stutter, when the writer becomes à stutterer in language' (`beÁgue de la langue', p. 135). This obvious allusion to Austin's seminal lectures (How To Do Things With Words23) nevertheless goes further than classic Anglo-Saxon pragmatics: it goes from affectation (a character's verbal signature, like Mr Micawber for instance) to affect, and to the idea that grammar is a carrier of affects and intensities, or language as an intensive line. Language, as envisaged by pragmatics, no longer represents the world, but enacts forces here the enacted forces carry affect.

The workings of this effectuation of affect are described through another allusion to linguistics, this time to the hyper-structuralism of Hjelmslev. As we saw, what interests Deleuze and Guattari in Hjelmslev is a Spinozan parallelism between the two planes language is made of: the plane of expression and the plane of contents. Here the saying of the `he stuttered' notation, on the plane of expression, can only work if it is paralleled by a certain àtmospheric quality', a doing, which on the plane of contents signals that the affect has indeed been evoked. But the two planes coalesce in the third moment, when doing and saying


coincide, in direct effectuation into a line. This line of style is further qualified through an explicit allusion to Gustave Guillaume, the French ènunciation' linguist, where the line of stuttering is described as a line of continuous variation; and it is assimilated, in the field of Deleuze and Guattari's own pragmatics, to a process of minorisation.

The excursus through linguistics ends in a quotation, and immediate subversion, of the central tenet of linguistic structuralism, best ex-pressed by Jakobson, that the linguistic sequence develops along two axes, the horizontal axis of syntagma and the vertical axis of paradigm. This is where the two stutterings, the rolling and the pitching, or the inclusive disjunction and the reflexive connection are described.

The plane of immanence of language is a linguistic battlefield: the object of the essay is to leave structuralism through the concept of linguistic stuttering, that is, to play Austin and pragmatics, Guillaume and enunciation theory, Hjelmslev and hyper-structuralism, or even the external linguistics of minorisation, against the orthodox version em-

bodied in Jakobson's famous essay on the two types of aphasia. But the path away from structural linguistics has always already been cleared by the practice of avant-garde poets, as the second movement of the essay demonstrates.

Literature is not the only site of the minorisation of the standard language or dialect, but it is the site par excellence where this process occurs. Hence a certain type of canon, which moves from prose writers like Melville or T. E. Lawrence to poets and to fous litteÂraires.

A brief list will be enough, in order of appearance in the essay. The successful candidates have not been selected at random: each contrib-utes a characteristic form of style. Taken together, those forms constitute the determinations of the concept. Thus Gherasim Luca, the Romanian poet, is the champion of reflexive connections, or words within words, or sentences within sentences (the French language is singularly conducive to this; the sentence `la chatte de ma cousine Julie miaule quand elle a faim' contains another sentence suggesting that my cousin Julia mews when she is hungry). Beckett is selected because of Watt's linguistic idiosyncra-sies: in a Hjelmslevian analysis, Watt's linguistic quirks are a form of expression, the parallel form of contents being his physical quirks, for instance his strange way of walking. Charles PeÂguy is notorious for his use of the rhetoric of repetition; Roussel, at least in the poem Nouvelles impressions d'Afrique, for his overindulgence in digression. Melville, of course, owes his presence in the list to Bartleby's notorious rejoinder, Ì'd prefer not to', an utterance that has long fascinated Deleuze and in which he mistakenly sees an instance of syntactic agrammaticality (he


claims that the sentence is agrammatical, which it clearly is not) this felix culpa, however, ought to draw our attention to the fact that for Deleuze style is (a) a matter of sentences (what kind of sentence is the Proust, or the Woolf sentence?) and (b) a matter of syntax (to what extent is the syntax of the sentence under strain?). The importance of such agrammaticality becomes entirely obvious when cummings's inev-itable `he danced his did', a constant joy to linguists, makes due appear-ance in our list. From syntax we go to affect and intensity. Antonin Artaud represents not only the climax of agrammaticality, the moment when language becomes irrecognisable as such, but its replacement with breaths and screams ( mots-cris and mots-souffles). In the same vein, CeÂline's style is captured through his intensive use of syntactic markers, which demonstrates that syntax too, and not merely individual words, can be iconic. With CeÂline and iconicity the canon is complete.

As the canon is constructed, and throughout the rest of the essay, the concept of style-as-stuttering is determined: it receives no fewer than eleven characteristics. Another list is in order.

1 Disequilibrium. Language stutters, and style encourages it to do so, because it is not a stable or fixed system, but a system in a state of constant disequilibrium. Inclusive disjunctions and reflexive connec-tions embody this state of affairs.

2 Variation. Disequilibrium does not mean that there is no system of language, although the Saussurean contrast of langue and parole has been superseded. But the system is a system of variations. In the first characteristic, language literally stuttered because the same element occupied more than one place, through the repetition that prevented stasis. Here, no element is fixed in its place, each element occupies a zone of variation, a place the position of which is not constant.

3 Vibration. Style makes language shiver, or tremble. The zone of vari-ation is a zone of vibration. Language vibrates like a musical instru-ment or hums like a machine. We are moving from an extensive account of language (as system of places creating semantic values) to an intensive account (where the linguistic sequence is a site for affects, where iconicity is as important as system). Stuttering is the name for this affective or intensive quality of language.

4 Line. I have mentioned the sentence, or the linguistic sequence. This is the material with which style works: language as a line, language `stretched along an abstract and infinitely varied line'. We no longer have hierarchy and depth (as we still had in Logique du sens, with the secondary organisation and tertiary arrangement of language); we


have a semi-aleatory trajectory on the surface, a line of infinite variation. The Markov chain is turning into Hogarth's serpentine line.

5 Minority. Stuttering is a mark of the minorising of standard language. This is a point which is now familiar. This is where the poetics of assemblage merges with the avant-garde poetics of high literature; this is where Kafka, a `celibate author', is nevertheless the champion of the people, for whom he writes. This minorisation is not conceived here as a clash between languages (Welsh, for instance, against im-

perial English) or between dialects or registers, but as the tracing of lines of flight within the standard language. The intensive line of syntax is a line of flight, which Deleuze, through a strange and novel metaphor (which he leaves unexplained) calls a `witch's line' (`ligne de sorcieÁre'). This seems to point towards a stylistic defamiliarisation through magic or incantation: not distortion (not style as eÂcart), but disequilibrium, modulation and bifurcation, as the jazz melody be-comes barely recognisable under the modulations of improvisation.

6 Inclusive disjunctions and reflexive connections. This is the central char-acteristic, right in the middle of the list, and no longer in need of elucidation. The next two characteristics can be treated as examples of the two stutterings.

7 Repetition. This is, as we saw in PeÂguy's contribution, the signature of his style: the obsessive, incantatory use of repetition turns the poem into a prayer, as a result of which words become `disjointed and decomposed members'. Here again, we have literal stuttering: through exaggerated repetition, language is taken to its limit (a device much used by Gertrude Stein, as by Beckett's Watt).

8 Digression. This is Roussel's contribution. The stutterer never finishes her sentence because she has always-already started another one, and yet another one. Here, style as stuttering works against the determin-

acy and fixity, the teleology of meaning. It enforces the open-ended-ness of sense, prevents the Markov chain from reaching its preordained end, indefinitely delays the point de capiton. It suggests the possibility that the quilting point will never be reached and mean-

ing never be achieved. Language is pushed to its limit in that the sentence, like Zeno's arrow, never reaches its goal.

9 The intensive line of syntax. This is not so much another characteristic as a summary:

This is therefore a ramified variation of language. Each variable state is like a point on a ridge line, which then bifurcates and is


continued along other lines. It is a syntactic line, syntax being constituted by the curves, rings, bends, and deviation of the dynamic line as it passes through the points, from the double viewpoint of disjunctions and connections. It is no longer the formal or superficial syntax that governs the equilibriums of language, but a syntax in the process of becoming, a creation of syntax that gives birth to a foreign language within language, a grammar of disequilibrium.24

As we saw, Deleuze seems to hold Chomsky's main tenet, the cen-trality of syntax. But with two non-Chomskyan, or even anti-Chomskyan qualifications: syntax is expressed by a line, not a hier-archy of dichotomies; and it is iconic rather than arbitrary a rhythm, rather than a formal code, an algorithm.

10 Rhythm. Deleuzean syntax is indeed endowed with rhythm, for which Artaud's mots-souffles and mots-cris provide the canonical example. This is where the line of syntax becomes intensive, ex-pressing the tensions of language. Àrtaud's deviant syntax, to the extent that it sets out to strain the French language, reaches the destination of its own tension in those breaths or pure intensities that mark a limit of language.'25 There is an echo in this of the constitutive paradox of the notion of rhythm, which etymologically denotes both flow and form:26 Deleuze's contribution is to attempt to conceive a form of language that would not be static but dy-namic.

11 Limit. The last characteristic is aptly named. It is perceived through the two aspects of CeÂline's language:

A limit of language that subtends the entire language, and a line of variation or subtended modulation that brings language to its limit. And just as the new language is now external to the initial language, the syntactic limit is not external to language as a whole: it is the outside of language, but not outside it.27

The limit of language is doubly paradoxical. First, it is always there, and yet it must be reached by straining language. Second, and this is the usual paradox of the limit (what lies beyond the limits of the world?), it is both the outside and yet not outside: the frontier itself. And in the case of language, there is an outside: either other media, images and sounds, as language turns to music or what Foucault calls `visibilities', or silence. `The outside of language' is the moment of the maximum striving of


what are still words, but hardly so (as in Beckett's second and third languages in the television plays), towards the images and sounds they will eventually become (when iconicity dominates meaning), or to-wards the silence in which they will vanish, like the silence at the end of the slow movement of Schubert's sonata in A (D 959) certainly not the silence of idiocy, but the moment when `words create silence' (`les mots font silence', p. 142: the French phrase enjoys the benefits of ambi-

guity, as it means both `words create silence' and `words fall silent'). We have now reached the limits of stuttering, when words fail, or fall

silent a highly creative achievement. And the progress of stuttering is also the progress of style, on the cline from meaning to silence. The cline can be represented in Figure 4.

We understand why `style is the economy of language' (p. 142), why style is equivalent to non-style: it has nothing to do with style as ornament, as expression of the writer's linguistic talent, and everything to do with the self-development of language, as it strives towards its limit.

The tension that the concept of style embodies cannot be resolved. There are two Deleuzes: one is political, the friend of Guattari, who shares his friend's revolutionary beliefs and with him constructs a theory of assemblages that equally concerns all types of utterances, as is the case with Foucault's archives; the other is artistic, and elitist at that, the defender of an avant-garde form of high literature, intent on preserving the primacy of the kind of linguistic practice that is charac-teristic of literature, as language moves on the gradient from meaning to sense and silence. In the `BeÂgaya-t-il' essay, the tension is felt in the contrast between a linguistic excursus and the construction of a canon. It is formulated in the felicitous ambiguity of the phrase `stuttering language': it is language that stutters, and the poet makes language stutter.

MEANING STYLE SILENCE

Disequilibrium Trembling Minorisation Repetition Line Limit (zone of of syntax

Continuousvibration) Line Digression Rhythm

variation STUTTERING

Rolling and pitching

Figure 4 Style


We could leave the scene here, and decide that such insolvable ten-sion expresses the problematic nature of the concept of language in Deleuze's philosophy. But we have come a long way since the first chapter, and should take the attempt at resolution seriously. What Deleuze's theory of assemblages and avant-gardist poetics have in common, as Figure 4 shows, is the concept of minority, through which the two strands of his thinking about language are articulated. This is another case of felicitous ambivalence, as the term has meaning both in the field of politics and in the field of art, and provides a link between the two domains.

What the concept does is to allow Deleuze definitively to leave the world (or the logic, or the image of thought) of representation (a world in which he still dwells in his Proust book a world of artistic essences, of metaphors, of interpretations, all coalescing in style), for the world of force and intervention, the world of the new pragmatics developed with the Kafka book and with Mille plateaux: a world where the concept of style is articulated with the concepts of force, of plane (of immanence and of consistency), of metamorphosis and assemblage. In this new world, the continuing presence of the concept of style (at the cost of an obvious shift in its determinations) can be expressed by parodying a famous feminist slogan: the artistic is political. This is, on the face of it, an avant-gardist position: the avant-garde artist as revolutionary. But Deleuze takes this conjunction entirely seriously, and pulls it in unex-pected directions: not towards the megalomania of the artistic genius who believes his elitist practice will save the world (there `the style is the man'), but towards style as impersonal and asubjective, where language, not the individual speaker, strives towards silence. The later concept of style, therefore, embodies this move away from the logic of representa-tion. With Deleuze we have, perhaps for the first time, a political concept of style.

Conclusion

1 A critique of language

Contemporary philosophy is generally said to have taken a linguistic turn several decades ago. It was first charted, in the case of analytic philosophy, by Richard Rorty in a famous anthology.1 And it is also said to have occurred, albeit much more recently, in the case of Contin-

ental philosophy, with deconstruction, the generalised pragmatics of Habermas and Apel, and Lyotard's Wittgensteinian period (mostly in Le DiffeÂrend).2 With their new pragmatics, what I have called their òther philosophy of language', Deleuze and Guattari can be said to belong to this trend. But this statement, we now know, must be replaced in its context. For it can equally be said that Deleuze's philosophy (this is an obvious difference from Guattari) is characterised by an anti-linguistic turn, that he is not to be included among the populous company of what Badiou calls `the modern Sophists' those philosophers who place language at the centre of their thought.

And it is fair to say that Deleuze, from his first to his last book, consistently seeks to think against the doxa in the philosophy of lan-guage (as well as the doxa that makes the philosophy of language central to philosophy). He seeks to overturn the image of thought such position presupposes and to construct an alternative one. Thus, he rejects the point of view of linguistic and philosophical structuralism, with its focus on signifiers, its erection of structures as the objects of all social sciences, and of linguistics as the model for the said sciences, a form of linguistics that is ìnternal' in that it relies on its own `principle of immanence' (the only relevant phenomena for the science of language are purely linguistic phenomena: language must be extracted from the world if it is to become a fit object for science). Not that the

247


philosophical fashion has not affected Deleuze (in the late 1960s): but it is merely a momentary influence, soon to be discarded, together with Lacanian psychoanalysis.

But Deleuze also rejects what in the philosophy of language seems to be the opposite of structuralism: the insistence on communicative action, interlocution and conversation, resulting in a form of philoso-phy of language inextricably mixed with ethics, which is the hallmark of Anglo-Saxon pragmatics and of Habermas's philosophy of communica-

tion. Deleuze is almost unique in his hostility to communication and dialogue, in which he sees a waste of time and a caricature of thought. So that even if the Internet is a rhizomatic, and therefore a Deleuzean entity, with its insistence on speed and irrepressible sprouting, the philosophy that underlies it is alien to Deleuze, who thinks that far from not communicating enough, we already communicate too much, and who ascribes the highest aesthetic value to silence. Such hostility to Anglo-Saxon pragmatics, from which he borrows a number of concepts, at the cost of strong misprision, is of course due to his rejection of the philosophy that underpins it: methodological individualism, an idealist theory of meaning, and a general reliance on conversational common sense and universal ethical maxims.

The outcome of this anti-linguistic turn could be appraised in Kantian terms: Deleuze's aim is to put language in its place, in all the senses of the term. In Deleuze, language is generally, and systematically, demoted: the focus of his philosophy is on thought (which is not assimilated to lan-guage nor is it a form of `mentalese', as with cognitive scientists), on intensity and speed, on a geography of planes and strata. Signs multiply, to a surprising extent, but they no longer necessarily possess the three main characteristics of the linguistic sign (arbitrariness, discreteness and double articulation). What is at stake here is the rejection of the image of thought dominant throughout the philosophical tradition: an image of thought characterised by what Deleuze sometimes calls the `logic of representation', based on the relationship between representative and represented in its five modes of difference, separation, substitution, hier-archy and abstraction.

This movement away from the dominant image of thought and the logic of representation can be charted, in Deleuze's oeuvre, through the internal history of the concept of sign. It has four stages, or moments. The first occurs in Proust et les signes. Here, the sign is a pre-Saussurean sign, devoid of the characteristics of the linguistic sign: a sign that con-notes before it denotes, to speak like Barthes. This means that the most important characteristic of a sign is not the object to which it refers and

Conclusion 249

which it represents, but the aura of signification with which it is sur-rounded, and which demands deciphering. Hence the first question that must be asked about a sign is not, `What does it stand for?', but, Ìn what world, the world of mondaniteÂ, of love, of sensible qualities or of art, does it appear?' In the context of their world of emergence, signs are objects of learning (of learning the ways and means of that particular world), and must be interpreted, a deciphering which in the case of signs par excel-lence, the signs of art, yields truth and reaches essences. The second moment should be the structuralist episode in Deleuze's career, the moment of À quoi reconnaõÃt-on le structuralisme?' and of Logique du sens. This should be the moment of the Saussurean, or the linguistic sign. But it is not, because in those texts the concept of sign is indeed conspicu-ous, but by its absence. The signifying chain is not constituted by the combination of individual signs (duly selected on the vertical, or para-digmatic, axis); what we have is two parallel series, of signifiers and signifieds, along which a paradoxical element, or empty square, circu-lates, distributing meaning at quilting points (points de capiton) and gen-erally making sense. This structure is closer to Lacan's concept of sign (where the dash that separates signifier from signified is a bar, preventing the division of the sequence into individual signs) than to Saussure's. In that context, the sign is a residual concept, defined in an appendix to Logique du sens as `what flashes across the boundaries of two levels, between two communicating series'.3 What really counts at that stage is the two series, not the sign. The third moment is the moment of Deleuze's collaboration with Guattari, the moment of Capitalisme et schizophreÂnie. At this stage, signs reappear. The ontology of flows and cuts is detailed in terms of coding, a literally signifying process. But the Saussurean sign's appearance is a modest one: it is quite lost in a proliferation and diversity of signs. There are as many types of signs as there are regimes of signs, and the imperialism of the signifier is enthusiastically defeated. Saussur-ean semiology is no longer a model or centre, it is drowned in a sea of non-signifying semiotics. So we have visual regimes of signs (with, for instance, the analysis of `facialities') or systems of sound signs (with the analysis of ritournelle, the refrain). The fourth and last moment is the mo-

ment of the two books on the cinema. The multiplication of signs reaches its paroxysm in Peircean efflorescence: Deleuze seeks to out-Peirce Peirce, whose taste for proliferating jargon is notorious: so the books lovingly detail and name no less than fifteen different types of signs.

There is a guiding thread through the four moments. From the begin-ning to the end of Deleuze's oeuvre, signs, rather than being discrete and combinable, are mysterious and in need of deciphering. As late as the


AbeÂceÂdaire, Deleuze defines friendship in terms of the reception and deciphering of signs.4 And in an essay the title of which seems to announce a philosophical joke, Àn Unrecognized Precursor to Heideg-

ger: Alfred Jarry' (Jarry was an early absurdist playwright, and the creator of that cult figure, le peÁre Ubu), he seeks to place Jarry (and consequently Heidegger) in the category of fous litteÂraires by again distinguishing between three levels of language, as in the case of Beckett's television plays. And in this context the concept of sign makes its last appearance:

Such is the response: language does not have signs at its disposal but acquires them by creating them when a language1 acts within a language2 so as to produce in it a language3, an unheard of and almost foreign language.5

We recognise the familiar definition of minority, and of style, and the allusion to the constantly quoted passage in Proust's Contre Sainte-Beuve, which provides an epigraph for the whole of Critique et clinique: `Great books are written in a kind of foreign language.'6

The result of this movement away from the logic of representation, which involves a radical critique of mainstream linguistics, is that it allows readers of Chomsky to proclaim that the emperor is naked, and that the metaphysical weight of the philosophy of language underpin-ning the Chomskyan research programme is unbearable. The combinat-

ion we find in Chomsky of gross materialism (the physical reductionism of the `mind-brain') and unabashed idealism (the recourse to innate ideas) is indeed crying out for an account in terms of what Althusser calls à spontaneous philosophy for scientists'.7 The importance of this critique cannot be overestimated: Deleuze's critique of language is first and foremost a critique of a form of `scientific' linguistics based on the concept of a universal grammar.

Important as it is, however, the quasi-Kantian critique of language has its limits. As we know, Deleuze is not a Kantian philosopher: Kant has always been an opponent for him, not an ally. What he dislikes in him is his taste for passing judgement: the typical Kantian gesture is to express philosophical propositions in the form of judgements. Deleuze does not take kindly to such niggardly constraints, to such a proliferation of precautions that ends up stifling thought. As a result, his critique of language can never be simply critical, and as a result of this result, his attitude to language, which will always take an affirmative turn, becomes deeply ambivalent. Language, to quote Corneille, is for Deleuze l'unique objet de son ressentiment: it is also an object of fascination, if not of desire.

Conclusion 251

We can find a symptom of this in the second chapter of CineÂma 2. The chapter is devoted to the relationship between image and language (it is entitled À Recapitulation of Images and Signs'). In the first section, Deleuze is rejecting the structuralist analysis of films (typically to be found in the early work of Christian Metz) in quasi-linguistic terms, an analysis for which the cinema is a form of language, and images are utterances, to be accounted for in terms of double articulation, para-digms and syntagmata. Against this, Deleuze bluntly states that `the cine-image is not an utterance'.8

Deleuze refers to a text where Pasolini defines the cine-image, `the movement-image', as forming an ìdiom of reality' (une langue de la reÂaliteÂ),9 which is not at all a language: we do not find the usual two axes, rather a process of differentiation and specification: not discrete units, but rather continuous variations and intensities (we cannot help noticing that those terms are applied by Deleuze and Guattari, in Mille plateaux, to the description of language). Here is the relevant passage:

These components of the movement-image, from the dual point of view of specification and differentiation, constitute a signaletic ma-

terial which includes all kinds of modulation features, sensory (visual and sound), kinetic, intensive, affective, rhythmic, tonal, and even verbal (oral and written). Eisenstein compared them first to ideo-grams, then, more profoundly, to the internal monologue as proto-language or primitive language system. But, even with its verbal elements, this is neither a language-system nor a language. It is a plastic mass, an a-signifying and a-syntactic material, a material not formed linguistically even though it is not amorphous and is formed semiotically, aesthetically and pragmatically. It is a condition, anter-ior by right to what it conditions. It is not an enunciation and these are not utterances. It is an utterable.10

So what we have with the cine-image is definitely not a form of lan-guage: a matieÁre signaleÂtique, a sign-matter or material, with `modulation features', a long list of which is given in our passage. That list makes fascinating reading. We note that (a) linguistic features come last, in what looks like an afterthought; (b) the general tonality of the list is organic: it concerns the affects of the body, not the abstract ideality of the linguistic system the cine-image has more to do with what Deleuze, in his Bacon book, calls `the logic of sensation', than with the logic of representation, as the cine-image is sensory and organic, and does not primarily represent


or narrate; (c) if another art form is evoked in the list, as a point of comparison with the visual art form of the cinema, it is music rather than poetry: those modulation features are rhythmic and tonal, not articulated; (d) the most important term in the list is ìntensive', and intensities is what language envisaged as a structure, concerned with positions and differential values, cannot capture. Lastly, (e), the list en-ables us to understand what Deleuze means by ̀ modulation': the world is a dynamic entity, in a state of constant and continuous variation the cine-image captures such variations directly, without having to go through the screen of language and representation. Hence the apparent paradox of Deleuze's conception of the cinema: the cine-image, appar-ently a kind of photograph endowed with movement, is not a representa-tion, and the cinema is not a representational art, and only contingently a narrative art. Exit Hollywood, pursued by a Deleuzean bear.

Our passage quotes another authority on film theory, Eisenstein. Pasolini's ìdiom of reality' is reinterpreted in terms of Eisenstein's theory of ìnterior monologue'. For Eisenstein, the cinema, not the novel, is the site where we catch a glimpse of an externalisation of our inner speech, the old logos endiathetos, the nature of which preoccupied Greek and medieval philosophers from Aristotle onwards.11 Except that, for Eisenstein and Deleuze after him, logos endiathetos is not a form of mentalese (the prevalent solution to the problem of inner speech in the philosophical tradition), nor is it internalised natural language (the minority position), but an accretion of non-linguistic material, a `masse plastique,' a shapeless but informable (and deformable) mass of asignifying and asyntactic matter a mass not yet linguistically, but already semiotically, aesthetically and pragmatically formed. In other words, Eisenstein's interior monologue, as transcribed in the cinema, consists of a mass of material that emits signs (but not linguistic signs). Those multiple and various signs produce aesthetic affects and induce action. The general lesson of this is the usual one: there is more to signs than merely linguistic or Saussurean signs.

We come to the centre of our passage: such non-linguistic matter, emitting signs, is in fact a prelinguistic matter not utterances, not a process of enunciation, both of which directly involve language, but still an utterable, ùn eÂnoncËable', liable to be uttered in a process of uttering. The text goes on to add that it is only when language appropriates such non-linguistic materials, when narration, for instance, is introduced, that utterances come to replace images and prelinguistic signs. `Lan-guage', Deleuze claims, èxists only as a reaction to a non-linguistic material, which it transforms.'12

Conclusion 253

There is, however, a problem with such demotion and dilution of language in the cinema, which could easily be generalised into a theory of language. For Deleuze's formulation is bizarre. Far from offering a solution to the question of the relationship between cine-images and language, as he claims (an anti-structuralist solution, one that strongly contradicts Metz), he renders such a relationship highly problematic, because of the terms he chooses to formulate it. Calling the material that the cine-image is made of an ùtterable' makes it prelinguistic: such ùtterability' sounds like Aristotelian potentia expecting to be actualised in utterances, in a process of enunciation. Even if it is deemed to be asignifying, asyntactic, and generally non-linguistic, this material is still, at least virtually, determined by language. There appears to be no independent way of talking about images (even if, in his Bacon book, Deleuze attempts to construct a series of concepts to do just that). So we come to our usual conclusion, the position that has produced this book: language is, for Deleuze, a problem indeed.

2 A celebration of language

Were Deleuze a Hegelian philosopher which he emphatically is not there might be a natural progression, hailed as progress, from mute and inglorious prelinguistic materials to their appropriation by language and metamorphosis into glorious, if garrulous, utterances: a progress from etymological infancy to fluency. But Deleuze, of course, would reject this narrative as too pat. His object is not to make language the apex or climax of an inevitable progress, but to put it in its place, a humble and dependent one. His object is to treat language as a sequel or epiphenom-

enon to images. This, however, is not an easy task. We already encountered, in

the Introduction, the paradox of the language-hater who waxes elo-quent in his denunciation. Hence the semantic tension that is felt in the word ùtterable', a movement towards actualisation in utteran-ces and enunciation, and its immediate freezing, with consequent re-treat except that Deleuze, who is not a Hegelian, is not an Aristotelian either: he follows Bergson in ascribing a central place to virtuality. This utterable is not merely possible, it is endowed with virtuality, which makes it as real as the actually uttered. We have indeed moved into a realm other than the realm of the abstract ideality of the linguistic system, or the realm of actual utterances, or statements, which corpus linguistics or the study of Foucauldian archives take as their starting point.


It is this move into virtuality, not abstract ideality, which enables Deleuze to do away with structuralism and offer another philosophy of language. What we have is a dissolution of the Saussurean sign which is both a dilution, as semiotics multiply because types of signs proliferate, and a destruction, as the characteristics of the linguistic sign, and the tyranny of the signifier, go by the board. Such a dissolution is no mean feat. It is also a profoundly affirmative gesture, no less than a theory of the sign, which can be summarised in Figure 5. The diagram describes Deleuze's ambivalence towards language: language is diluted among a proliferation of sign systems, yet it still serves as point of reference (if no longer as model). And the interesting sites in language are its margins: the margin of the prelinguistic utterable, which language appropriates in order to produce meaning; and the frontier between language and what is beyond it, where style dwells.

Because Deleuze is an affirmative philosopher, with little taste for the work of negation, his attitude to language cannot remain merely crit-ical. At worst, he might simply ignore language. But this he clearly cannot do: language admittedly a large subject crops up at almost every line of his work. So, when Deleuze is at his best about language, we find in his work a joyous affirmation of the creativity of language, embodied in the concept of style, and the construction of an alternative, and entirely positive, philosophy of language.

This I have attempted to describe in the last three chapters of this book. The time has come to summarise. Such summaries usually take the form of a correlation. This one will be no exception, except that the correlation will be given in a number (16 in all) of contradictory theses, which present the main concepts of Deleuze's philosophy of language, and their antonyms, or opponents:

1 Immanence, not transcendence.2 Becoming, not structure.

SIGNS

Prelinguistic Language Postlinguistic signs

material I I (music, image, etc.)

(utterable) I I

MEANING STYLE

Figure 5 Ambivalence towards language

Conclusion 255

3 Force, not signification.4 Speed, not stasis.5 Variations, not positions.6 Rhizomes, not tree-like hierarchies.7 Lines and planes, not depth v. surface.8 History, not synchrony.9 External, not internal.

10 Creativity, not rules.11 Maxims, not laws.12 Assemblages, neither subjects nor system.13 Minority, not standard or major language.14 Slogans, not propositions.15 Literature, not a democracy of corpuses.16 Agrammaticality, not acceptability and syntax as power.

At this stage, the 16 contrasts should require little comment: they delineate Deleuze's `philosophy of language'. This phrase, which the master himself would probably have rejected, insists on the range of the problematic of language in Deleuze, which the two columns of this implicit correlation illustrate. And `problematic' here refers to Deleuze's concept of problem: in this correlation, language appears as a highly problematic affair. Taken critically, the two columns show that in Deleuze, language is always in danger of being diluted in a larger meta-

physics of life, thought and bodies. But taken affirmatively, they show that the problematic of language tends to absorb the whole of Deleuze's first philosophy, that in Deleuze language is everywhere.

Taken in groups, the contrasts enable us to understand the variety and importance of Deleuze's contribution to the philosophy of language. The concepts in lines 1 to 7 are at the heart of Deleuze's vitalist meta-

physics, which makes him the odd philosopher out in a conjuncture dominated by structuralism and post-structuralism. The position owes as much to Bergson and the French tradition as it does to Nietzsche (it would be interesting to look at the role played by intensive qualities and force in the analysis of habit by Ravaisson, a philosopher whom both Bergson and Heidegger admired13). The concepts in lines 8 to 11 enable us to abandon the generality of research programmes in linguis-tics and construct a new, external form of linguistics. The concepts in lines 12 to 14 indicate the proximity of Deleuze's new linguistics to the Marxist tradition: not only is Voloshinov one of the authorities Deleuze is fond of quoting, but the collection of essays by Pasolini he refers to in the cinema books contains an essay entitled Àppunti (en poeÁte) per


una linguistica marxista'. This apparently strange combination of the poetic and the political is at the heart of Deleuze's (and Guattari's) view of language. It explains why the concepts in the last two lines (15 and 16) indicate the central role literature, a reading of literary texts, must play in the construction of a new linguistics and another philoso-phy of language. And it explains why the concept that Deleuze con-structs to name and explicate the problem of language is the concept of style.

We understand now why the concept of style is a permanent feature of Deleuze's work: it is already present in Proust et les signes, where it is defined (strangely, for anyone who has read the rest of the oeuvre) in terms of reaching for essences, interpretation, metaphor and metamor-

phosis; and it is at the heart of Deleuze's last book, Critique et clinique, where it has the well-known characteristics of stuttering language and reaching for its limit. And we also understand why this apparently passeÂ concept was preferred to the new-fangled concept of writing: because it has the requisite ambivalence. It seems to refer primarily to the art of language (`stylistics' is a part of linguistics) and yet it replaces language within a wider range of semiotics, visual or musical for instance (CeÂzanne and Schumann have style).

Lastly, we understand what happens when Deleuze reads other phil-osophers, as well as works of art, when he produces a strong reading, by extracting a problem and constructing a concept, often against the grain of the text, in other words against the established tradition of interpret-ation (he begins his book on Proust by stating that if there is one thing A la recherche du temps perdu is not about, it is memory): he reads for style. He extracts the problem and constructs the concept that will express the specific philosophical gesture of the text, its philosophical and literary style (there is a baroque style of Leibniz, captured by the concept of the fold). For Deleuze's problematic way of reading provides a form of qui-lting point for the text. A diagram, borrowed from the first stage of Lacan's Byzantine `bottle-opener' diagram, will chart the assemblage of corpus, problem and concept (see Figure 6). In Lacan, the first stage of the diagram14 describes the donation of meaning as the effect of point de capiton on a Markov chain. The Markov chain moves steadily from A to B, as a chain of signifiers. The second series, the series of signifieds, is not parallel to it: it goes from C to D, and intersects the chain of signifiers at points E and F, where meaning occurs, but retroactively, against the normal flow of the sequence of signifiers.

In the case of Deleuze's strong, problematic, reading, the diagram has a similar shape, as shown in Figure 7. The reading (a retroactive one,

Conclusion 257

A F E

BD C

Figure 6 Stitching point

A

bPro lem

F E

Reading

D C B

Concept Corpus

Figure 7 Deleuzean reading

against the grain of the corpus) selects a part of the corpus (A to B), which it reads by extracting a problem out of it (E to F is both a subsec-tion of the corpus, a reading of it, and a subsection of C to D, the problem that the reading causes to emerge and which it explicates). The reading has a goal, an end in both senses, in the concept that names it. This diagram accounts for Deleuze's readings of various texts, and I hope, en abyme, for my reading of Deleuze.

Notes

Notes to the Introduction

1 For reasons that will soon become obvious, I am using the French version: S. Beckett, Quad et autres pieÁces pour la television, suivi de L'EpuiseÂ, par Gilles Deleuze (Paris: Minuit, 1992). For the English translation of Deleuze's essay, see G. Deleuze, Essays Critical and Clinical, trans. by D. W. Smith and M. W. Greco (London: Verso, 1998) pp. 152 74.

2 I am using the phrase `Gilles Deleuze', which is only fair, not as the name of a person, even of a great philosopher, but of an àssemblage of enunciation', which of course includes Felix Guattari as one of its major components. This will be further explained in Chapter 1, Section 5.

3 Beckett, Quad, pp. 104 5. The phrase comes from Beckett: see S. Beckett, Mal vu mal dit (Paris: Minuit, 1981).

Notes to Chapter 1

1 G. Deleuze and F. Guattari, L'Anti êdipe (Paris: Minuit, 1972) p. 7 (Engl. trans., pp. 1 2).

2 L. Wittgenstein, Tractatus Logico Philosophicus (London: Routledge & Kegan Paul, 1961) p. 7.

3 This is the beginning of Descartes's first meditation. See R. Descartes, Discourse on Method and the Meditations (Harmondsworth: Penguin, 1968) p. 95.

4 Charles Dickens, The Mystery of Edwin Drood (Harmondsworth: Penguin, 1974) p. 37.

5 E. Benveniste, `La nature des pronoms', in ProbleÁmes de linguistique geÂneÂrale, vol. 1 (Paris: Gallimard, 1966) pp. 251 7.

6 J. Channel, Vague Language (Oxford: Oxford University Press, 1994). 7 G. Lakoff and M. Johnson, Metaphors We Live By (Chicago: University of

Chicago Press, 1980). 8 Dickens, Edwin Drood, p. 37. 9 G. F. W. Leibniz, À ResumeÂ of Metaphysics', in Philosophical Writings

(London: Dent and Dutton, 1973) p. 145; M. Heidegger, An Introduction to Metaphysics (New Haven, Conn.: Yale University Press, 1959) p. 1. Ralph Mannheim, the translator, has `Why are there essents rather than nothing'. The coinage, which is glossed in a translator's note, has not caught on.

10 For an analysis of Hegel's views on shifters and negativity, see G. Agamben, Il Linguaggio e la morte (Turin: Einaudi, 1982).

11 G. Deleuze, Logique du sens (Paris: Minuit, 1969) p. 178. The English translation has `they' (p. 152).

12 G. Simondon, L'individu et sa geneÁse physicobiologique (Paris: Millon, 1995); 1st edn (Paris: PUF, 1964). On Simondon's philosophy of individuation and technology, see M. Combes, Simondon: Individu et collectiviteÂ (Paris: PUF, 1999).

13 G. Deleuze, Foucault (Paris: Minuit, 1986) pp. 62 3. The English translation (p. 55) has òne speaks'.

258

Notes 259

14 C. LeÂvi Strauss, Tristes tropiques (Paris: Plon, 1955) pp. 58 9. 15 See J. Dor, Introduction aÁ la lecture de Lacan, vol. 1: L'inconscient structureÂ comme

un langage (Paris: Denoel, 1985). It must be noted that Deleuze and Guattari's exclusion bears on this formula, not the more interesting `the unconscious is the speech of others', to be found in the `Discours de Rome'.

16 For an account of this maxim, see my Interpretation as Pragmatics (Basingstoke: Palgrave Macmillan, 1999) pp. 79 80.

17 J. Culler, On Deconstruction (London: Routledge, 1983) pp. 86 8. 18 R. Jakobson, `Two Aspects of Language and Two Types of Aphasic Disturb

ances', in Selected Writings, vol. II (The Hague: Mouton, 1971). The correlation is reproduced in my The Violence of Language (London: Routledge, 1970) p. 131.

19 It does not escape me that the practice of inevitable mistranslation goes further than I have made out so far. For the original of the Freudian ìd' is a personal pronoun, ès', not a deictic. It must become a deictic in French, as the French language does not have neuter for pronouns. But I am commenting on a French text, which takes philosophical advantage, in creative misprision, of the inevitable failings of translation

20 G. Deleuze and F. Guattari, Politique et psychoanalyse (Paris: Des Mots Perdus, 1977).

21 A. Compagnon, `L'exception francËaise', in OuÁ en est la theÂorie litteÂraire? (Paris: Textuel, 2000) p. 43.

22 J. F. Lyotard, Le DiffeÂrend (Paris: Minuit, 1983); C. LeÂvi Strauss, Anthropologie structurale (Paris: Plon, 1958); A. J. Greimas, SeÂmantique structurale (Paris: Larousse, 1966); P. Bourdieu, Ce que parler veut dire (Paris: Fayard, 1982); J. Kristeva, Semeiotike (Paris: Seuil, 1969) and La ReÂvolution du langage poeÂtique (Paris: Seuil, 1974); J. Derrida, Le Monolinguisme de l'autre (Paris: GalileÂe, 1996); M. Foucault, L'ArcheÂologie du savoir (Paris: Gallimard, 1969).

23 For further developments, see my `Philosophies du langage analytique et continentale: de la sceÁne de meÂnage aÁ la meÂprise creÂatrice', in L'Aventure humaine, 11 (Paris: 2000) pp. 11 22.

24 The papers were published in Cahiers de Royaumont, La Philosophie Analytique (Paris: Minuit, 1962).

25 J. Derrida, Limited Inc. (Evanston, Ill.: Northwestern University Press, 1988). 26 H. Bergson, La PenseÂe et le mouvant (Paris: PUF, 1975; 1st edn 1938) pp. 88 9

(my translation). 27 Ibid., p. 51. 28 G. Deleuze, and Claire Parnet, L'AbeÂceÂdaire de Gilles Deleuze (Paris: Editions du

Montparnasse, 1997) letter W. 29 A. Gualandi, Deleuze (Paris: Belles Lettres, 1998) p. 107. 30 G. Deleuze and C. Parnet, Dialogues (Paris: Flammarion, 2nd edn, 1996)

pp. 42 3 (Engl. trans., p. 34). 31 G. Deleuze and F. Guattari, Mille plateaux (Paris: Minuit, 1980) p. 84 (Engl.

trans., p. 65). 32 Ibid., p. 86 (Engl. trans., p. 66). 33 Ibid., p. 87 (Engl. trans., p. 67). 34 G. Deleuze and F. Guattari, Kafka (Paris: Minuit, 1975) p. 40 (Engl. trans.,

p. 22). 35 Ibid., p. 41 (Engl. trans., p. 22).

260 Notes

36 Deleuze and Parnet, Dialogues, p. 9 (Engl. trans., p. 3). 37 Ibid., pp. 24 5 (Engl. trans., p. 17). 38 I have defended a theory of metaphor that involves just this, even if I

carefully avoided the word `truth': see J. J. Lecercle, The Violence of Language (London: Routledge, 1990) chap. 4.

39 M. Foucault, Dits et eÂcrits, vol. 3: 1976 9 (Paris: Gallimard, 1994) p. 541. 40 G. Deleuze, DiffeÂrence et reÂpeÂtition (Paris: PUF, 1968). 41 M. Foucault, Les Mots et les choses (Paris: Gallimard, 1966); Dits et eÂcrits, vol. 1,

1954 69 (Paris: Gallimard, 1994). 42 The passage I am commenting on occurs on pp. 53 4 of Les Mots et les

choses. 43 See for instance the excellent introduction to Deleuze's philosophy by John

Marks, Gilles Deleuze (London: Pluto Press, 1998). 44 F. Guattari, Chaosmos (Paris: GalileÂe, 1992). 45 G. Deleuze, Pourparlers (Paris: Minuit, 1990) p. 43 (Engl. trans., p. 28). 46 On this, see J. ReÂe, I See a Voice (London: Flamingo, 2000). 47 A. Badiou, Deleuze (Paris: Hachette, 1997). 48 F. Guattari, L'Inconscient machinique (Paris: Recherches, 1979). 49 G. Deleuze, Proust et les signes (Paris: PUF, 1964). 50 Guattari, L'Inconscient machinique, p. 10. 51 On this, see my The Violence of Language, chapter 1. 52 Guattari, L'Inconscient machinique, pp. 25, 131. 53 Bourdieu, Ce que parler veut dire. 54 Guattari, L'Inconscient machinique, p. 12. 55 Ibid., p. 44. 56 See J. J. Courtine and C. Haroche, Histoire du visage (Paris: Rivages, 1988). 57 C. Darwin, The Expression of the Emotions in Man and Animals (Chicago, Ill.:

University of Chicago Press, 1965). 58 R. Carver, `The Father', in The Stories of Raymond Carver (London: Picador,

1985) pp. 40 4. 59 Guattari, L'Inconscient machinique, p. 82. 60 M. Cressole, Deleuze (Paris: Editions universitaires, 1973). 61 Deleuze, Pourparlers, pp. 11 22 (Engl. trans., pp. 3 12). 62 Deleuze and Parnet, L'AbeÂceÂdaire de Gilles Deleuze, letter H for `History of

philosophy'. 63 G. Deleuze and F. Guattari, Qu'est ce que la philosophie? (Paris: Minuit, 1991)

p. 192 (Engl. trans., p. 204). 64 Deleuze, DiffeÂrence et reÂpeÂtition, pp. 252 5 (Engl. trans., p. 195). 65 E. Canetti, Crowds and Power (Harmondsworth: Penguin, 1973) pp. 331 7. 66 Deleuze, DiffeÂrence et reÂpeÂtition. The relevant passages are on pp. 210 3, 232,

244, 261 (Engl. trans., pp. 162, 179, 189, 202). 67 On this, see my Interpretation as Pragmatics. 68 This is indeed what Pierre Bayard does in Qui a tueÂ Roger Ackroyd? (Paris:

Minuit, 1998). 69 F. Brunot, La PenseÂe et la langue, 3rd edn (Paris: Masson, 1936) p. 525.

The relevant passage in DiffeÂrence et reÂpeÂtition is on p. 261 (Engl. trans., p. 202).

Notes 261

Notes to Interlude 1

1 G. Swift, Ever After (London: Picador, 1992), pp. 106 7. A first version of this interlude was given as a lecture at the second European Society for the Study of English (ESSE) conference in Bordeaux.

2 T. Coleman, The Railway Navvies (Harmondsworth: Penguin, 1968), pp. 37 8. 3 C. Dickens, Dombey and Son (Harmondsworth: Penguin, 1970), pp. 120 1. 4 P. Hamon, Introduction aÁ l'analyse du descriptif (Paris: Hachette, 1981). 5 See F. D. Klingender, Art and the Industrial Revolution (London: Paladin, 1972;

1st edn, 1947). 6 R. Barthes, La Chambre claire (Paris: Gallimard/Seuil, 1980). 7 Dickens, Dombey and Son, p. 875. 8 A. Shukman (ed.), Bakhtin School Papers, Russian Poetics in Translation, no. 10

(Oxford, 1983), p. 147.9 Coleman, The Railway Navvies, p. 175.

10 A short story by Nathaniel Hawthorne, `The Celestial Rail road' provides an ironic illustration of this metaphor. See N. Hawthorne, Tales (New York: Norton, 1987), pp. 131 43.

11 Dickens, Dombey and Son, p. 354. 12 M. Johnson, The Body in the Mind (Chicago, Ill.: Chicago University Press,

1987). 13 L. Carroll, The Annotated Alice (Harmondsworth: Penguin, 1965), pp. 217 19. 14 G. Deleuze and F. Guattari, Mille plateaux (Paris: Minuit, 1980), pp. 112 13

(Engl. trans, pp. 89 90). 15 One of Deleuze's disciples, J. C. Martin, has attempted a systematic account

of such an assemblage, the Romanesque assemblage, in his Ossuaires (Paris: Payot, 1995), a book derived from a thesis supervised by Deleuze.

16 Carroll, The Annotated Alice, p. 221. 17 G. Lakoff and M. Johnson, Metaphors We Live By (Chicago, Ill.: University of

Chicago Press, 1980). 18 The theory of series in Logique du sens offers a half way house: there are two

distinct series, of words and things, for instance, but they are joined together by the paradoxical or nonsensical element that circulates among them. See Chapter 3 of this book.

19 J. Kristeva, Polylogue (Paris: Seuil, 1977). 20 M. Blanchot, L'Espace litteÂraire (Paris: Gallimard, 1955), pp. 365 6 (my trans

lation). 21 G. Ryle, The Concept of Mind (Harmondsworth: Penguin, 1963). 22 B. Schulz, The Fictions of Bruno Schulz (London: Picador, 1988). The descrip

tion of the train is on pp. 237 8.

Notes to Chapter 2

1 L'AbeÂceÂdaire de Gilles Deleuze (Paris: Editions du Montparnasse, 1997; three videocassettes).

2 J. J. Lecercle, The Violence of Language (London: Routledge, 1991). 3 See Deleuze, AbeÂceÂdaire, `Z is for Zorro'.

262 Notes

4 Althusser's critique of the concepts of synchrony and diachrony is to be found in `L'objet du capital', in L. Althusser et al., Lire le Capital, vol. 2 (Paris: Maspero, 1965), Engl. trans. in L. Althusser and E. Balibar, Reading Capital (London: New Left Books, 1970). For a commentary on this see my The Violence of Language, pp. 201 5.

5 G. Deleuze and F. Guattari, L'Anti êdipe (Paris: Minuit, 1972), p. 170 (Engl. trans., p. 145).

6 S. Leclaire, Psychanalyser (Paris: Seuil, 1968). 7 G. Deleuze, Pourparlers (Paris: Minuit, 1990), p. 24 (Engl. trans., p. 15). 8 G. Deleuze, Foucault (Paris: Minuit, 1986), p. 24 (Engl. trans., p. 15). 9 Following the choice made by translators of Foucault, the translator of

Deleuze's Foucault translates èÂnonceÂ' as `statement'. He is right to follow an established practice. The original translation, however, I find unfortunate: a statement is made by a speaker. An eÂnonceÂ, an utterance, is just a text, in which the anonymous murmur of the on can be heard.

10 Deleuze and Guattari, L'Anti êdipe, p. 130 (Engl. trans., p. 109). 11 G. Deleuze, À quoi reconnaõÃt on le structuralisme?', in F. Chatelet (ed.),

Histoire de la philosophie, vol. 8: Le XXe sieÁcle (Paris: Hachette, 2000; 1st edn, 1973), pp. 299 335.

12 The political critique of linguistics is by no means restricted to Deleuze and Guattari, of course. Apart from Bourdieu's èxternal' linguistics (Ce que parler veut dire, Paris: Fayard, 1982), we find it in Kress and Hodge (G. Kress and R. Hodge, Language as Ideology, London: Routledge, 1979), Fairclough (N. Fairclough, Language as Power, London: Longman, 1989), and Andersen (R. Andersen, The Power of the Word, London: Paladin, 1988).

13 Deleuze, Pourparlers, p. 65 (Engl. trans., p. 44). 14 G. Deleuze, DiffeÂrence et reÂpeÂtition (Paris: PUF, 1968), p. 265 (Engl. trans.,

p. 204). 15 G. Deleuze and Guattari, L'Anti êdipe, p. 287 (Engl. trans., pp. 241 2). 16 G. Deleuze and Claire Parnet, Dialogues (Paris: Flammarion, 2nd edn 1996),

p. 137 (Engl. trans., p. 115). 17 Ibid., p. 20 (Engl. trans., p. 13). 18 Deleuze, DiffeÂrence et reÂpeÂtition, pp. 210 17 (Engl. trans., pp. 162 7). 19 Deleuze and Guattari, Dialogues, p. 30 (Engl. trans., p. 22). 20 G. Deleuze and F. Guattari, Rhizome (Paris: Minuit, 1976). 21 G. Deleuze and F. Guattari, Mille plateaux (Paris: Minuit, 1980), p. 13 ff. (Engl.

trans., p. 7 ff.). 22 Ibid., p. 27 (Engl. trans., p. 18). 23 See F. Dosse, Histoire du structuralisme, vol. 1 (Paris: Hachette, 1998), p. 290,

1st edition, Paris: La DeÂcouverte, 1992). 24 See M. ArriveÂ, Linguistique et psychanalyse (Paris: Klincksieck, 1985). 25 Deleuze and Guattari, Dialogues, p. 58 (Engl. trans., p. 46). 26 Deleuze and Guattari, L'Anti êdipe, p. 87 (Engl. trans., p. 73). 27 Ibid., p. 244 (Engl. trans., pp. 206 7). 28 Ibid., p. 245 (Engl. trans., p. 207). 29 See Deleuze and Guattari, Mille plateaux, p. 221 (Engl. trans., p. 180). 30 For an account of this maxim, see my Interpretation as Pragmatics (Basing

stoke: Palgrave Macmillan, 1999), pp. 79 80. 31 Deleuze and Guattari, Mille plateaux, p. 169 (Engl. trans., p. 135).

Notes 263

32 G. Deleuze, Proust et les signes (Paris: PUF, 1964; 2nd edn, 1970). I shall be using the second edition, which is much enlarged.

33 Ibid., pp 104 8 (Engl. trans., pp. 84 8). 34 Ibid., pp. 128 9 (Engl. trans., p. 105). 35 J. C. Milner, L'Amour de la langue (Paris: Seuil, 1978); Engl. trans. by A.

Banfield, For the Love of Language (London: Palgrave Macmillan, 1989). 36 J. C. Milner and F. Regnault, Dire le vers (Paris: Seuil, 1987). 37 For an account of this aspect of Milner, see J. J. Lecercle, Philosophy Through

the Looking Glass (London: Hutchinson, 1985), pp. 81 2, and The Violence of Language (London: Routledge, 1990), pp. 32 42.

38 Deleuze and Guattari, Mille plateaux, p. 95 (Engl. trans., p. 75). 39 Ibid., p. 109 (Engl. trans., p. 85). 40 Ibid., p. 116 (Engl. trans., p. 92). 41 Ibid., p. 127 (Engl. trans., p. 100). 42 J. Derrida, `Signature eÂveÂnement contexte', in Marges de la philosophie (Paris:

Minuit, 1972), pp. 365 93. 43 J. Stalin, Marxism and Linguistics (New York: International Publishers, 1951).

It is regularly suggested that the real author of the pamphlet was the Soviet academician V. V. Vinogradov.

44 See my The Violence of Language, passim. 45 See L. J. Calvet, Linguistique et colonialisme (Paris: Payot, 1974). 46 Deleuze and Guattari, Mille plateaux, p. 130 (Engl. trans., p. 102); àffected'

translates the French `travailleÂ', which has far wider implications. 47 Ibid. (Engl. trans., p. 103). 48 Deleuze, Pourparlers, p. 44 (Engl. trans., pp. 28 9). `Precepts' here translates

`mots d'ordre': a translator's note dismisses Brian Massumi's coinage of òrderwords' as conveying `none of the everyday resonance of the phrase' (ibid., p. 189).

49 A. Badiou, Conditions (Paris: Seuil, 1992), p. 60. 50 See, for instance, J. P. Corneille, La Linguistique structurale (Paris: Larousse,

1976). 51 Deleuze and Guattari, L'Anti êdipe, p. 287 (Engl. trans., pp. 241 2). 52 Deleuze and Guattari, Mille plateaux, p. 58 (Engl. trans., p. 43). 53 See, for instance, ibid., p. 118 (Engl. trans., p. 93). 54 R. Langacker, Foundations of Cognitive Grammar, 2 vols (Stanford, Cal.: Stan

ford University Press, 1987, 1991). 55 G. Guillaume, Langage et science du langage, 3rd edn (Paris: Nizet, 1984);

E. Benveniste, ProbleÁmes de linguistique geÂneÂrale, vol. 1 (Paris: Gallimard, 1966) and vol. 2 (1974) (Engl. trans. of vol. 1, Problems of General Linguistics, Miami: Miami University Press, 1971); A. Culioli, Cognition and Representation in Linguistic Theory (Amsterdam: John Benjamins, 1995); Pour une linguistique de l'eÂnonciation, vol. 1 (Gap: Ophrys, 1990) and vols 2 and 3 (1999).

56 Deleuze and Guattari, Mille plateaux, p. 322 (Engl. trans., p. 263).57 Ibid., pp. 162 3 (Engl. trans., p. 130).58 On this, see N. Ruwet, Introduction aÁ la grammaire geÂneÂrative (Paris: Plon,

1967), p. 93. 59 N. Chomsky, Syntactic Structure (The Hague: Mouton, 1957), p. 21. 60 Deleuze and Guattari, L'Anti êdipe, p. 343 (Engl. trans., p. 289).

264 Notes

61 Ibid., p. 46 (Engl. trans., p. 38). 62 Deleuze, Foucault, pp. 92, 125 (Engl. trans., pp. 86, 117). The passage quoted

is p. 125 (Engl. trans., p. 117). 63 Deleuze and Guattari, Mille plateaux, p. 174 (Engl. trans., p. 140). 64 I have defended this thesis in The Violence of Language. 65 R. Hull, Excellent Intentions (Harmondsworth: Penguin, 1949; 1st edn, 1938),

p. 15.

Notes to Chapter 3

1 G. Deleuze, Logique du sens (Paris: Minuit, 1969). 2 See J. J. Lecercle, Philosophy of Nonsense (London: Routledge, 1994). 3 Deleuze, Logique du sens, p. 7 (Engl. trans., pp. xiii xiv). 4 The same sentence contains another slip: Deleuze's ùne histoire embrouil

leÂe', as the inverted commas show, is a quotation of Carroll's title À Tangled Tale': à convoluted story' loses the allusion.

5 Deleuze, Logique du sens, pp. 203 5 (Engl. trans., pp. 174 5). 6 Ibid., p. 54 (Engl. trans., pp. 39 40). 7 Ibid., pp. 65 6 (Engl. trans., pp. 50 1). 8 On this, see G. Deleuze, Le Bergsonisme (Paris: PUF, 1966), pp. 99 101 (Engl.

trans., pp. 96 8). 9 G. Deleuze, À quoi reconnaõÃt on le structuralisme?', in F. Chatelet, Histoire de

la philosophie, vol. 8: Le XXe sieÁcle (Paris: Hachette, 1973; 2nd edn, 2000), pp. 299 335.

10 G. Deleuze, `Hume', in ibid., vol. 4 (Paris: Hachette, 1972; 2nd edn, 1999). 11 A. Badiou, L'Etre et l'eÂveÂnement (Paris: Seuil, 1988). On Badiou, see J. J. Lecer

cle, `Cantor, Lacan, Mao, Beckett, MeÃme Combat: the Philosophy of Alain Badiou', Radical Philosophy, 93, January/February 1999, pp. 6 13.

12 A. Badiou, L'Ethique (Paris: Hatier, 1993), Engl. trans. A. Badiou, Ethics (London: Verso, 2001).

13 See my article in Radical Philosophy, op. cit. Ïiz14 S. Z Ïek, The Fragile Absolute (London: Verso, 2000), p. 92.

15 G. Deleuze and F. Guattari, Qu'est ce que la philosophie? (Paris: Minuit, 1991),p. 144 (Engl. trans., p. 152).

16 M. Blanchot, L'Espace litteÂraire (Paris: Gallimard, 1955), p. 299 (my translation).

17 See his Beckett (Paris: Hachette, 1995); and his Petit manuel d'inestheÂtique (Paris: Seuil, 1998).

18 G. Deleuze and C. Parnet, Dialogues (Paris: Flammarion; 2nd edn 1996), pp. 77 80 (Engl. trans., p. 62 4).

19 Deleuze and Guattari, Qu'est ce que la philosophie?, pp. 146 7 (Engl. trans., p. 156; I have slightly modified the translation).

20 Deleuze, Le Bergsonisme, pp. 98 101 (Engl. trans., pp. 95 8). 21 Deleuze and Guattari, Qu'est ce que la philosophie?, p. 150 (Engl. trans.,

p. 159). 22 The English translation of Logique du sens has `phantasm' for the French

`phantasme'. I am not certain the choice is the right one, as it occults the

Notes 265

Kleinian origin of the term, in the contrast between fantasy (conscious) and phantasy (unconscious).

23 L. Irigaray, `Du fantasme et du verbe', in Parler n'est jamais neutre (Paris: Minuit, 1985), pp. 69 80.

24 Deleuze, Logique du sens, p. 125 (Engl. trans., p. 103). 25 See J. A. Miller, `Language: Much Ado about What?', in E. Ragland Sullivan

and M. Bracher (eds), Lacan and the Subject of Language (London: Routledge, 1991), pp. 21 35. Ïiz26 Z Ïek, The Fragile Absolute, p. 139.

27 On the concepts of `manifestation', `designation' and `signification', see Deleuze, Logique du sens, pp. 22 4 (Engl. trans., pp. 13 15).

28 Ibid., p. 206 (Engl. trans., p. 176). 29 L. Carroll, The Annotated Alice, ed. M. Gardner (Harmondsworth: Penguin,

1965), p. 47. 30 For a detailed analysis, see J. J. Lecercle, The Violence of Language (London:

Routledge, 1990), pp. 14 18. On Lewis Carroll's linguistic intuitions, see my Philosophy of Nonsense (London: Routledge, 1994).

31 Carroll, The Annotated Alice, p. 47. 32 Deleuze, Logique du sens, p. 41 (Engl. trans., p. 28). 33 Luis d'Antin Van Rooten, Mots d'Heures: Gousses, Rames (London: Angus &

Robertson, 1968). This obscure but moving line, obviously the work of an avant garde French poet, can be glossed thus: `you have a beautiful screw [a kind of crane] with which to hoist a fine cargo of lacquer on board'. The ghosts of Pierre Loti and Victor Segalen, unless it is Paul Gauguin, are clearly evoked.

34 Deleuze, Logique du sens, p. 34 (Engl. trans., p. 22). 35 Ibid., p. 194 (Engl. trans., p. 166). 36 Ibid., p. 212 (Engl. trans., p. 181). 37 Ibid., pp. 286 8 (Engl. trans., pp. 244 6).


1 J. Joyce, Dubliners (Harmondsworth: Penguin, 1957; first published 1914), p. 173. A first version of this section appeared in J. J. Lecercle, `Ce sieÁcle ne sera pas deleuzien', TLE, 14 (Paris: Presses Universitaires de Vincennes Saint Denis, 1996), pp. 129 49.

2 H. Kenner, Joyce's Voices (London: Faber, 1978), pp. 15 16. AndreÂ Topia drew my attention to this text.

3 e. e. cummings, Complete Poems 1904 1962 (New York: Liveright, 1994), p. 515. For a linguistic analysis of the poem, see J. R. Levin, `Poetry and Grammaticalness', in Reprints of Papers for the Ninth International Congress of Linguists (Cambridge, Mass., 1962); J. P. Thorne, `Stylistics and Generative Grammar', in Journal of Linguistics 1, 1 (1965); N. Ruwet, `ParalleÂlismes et deÂviations en poeÂsie', in J. Kristeva et al. (eds), Langue, discours, socieÂteÂ (Paris: Seuil, 1975).

4 J. J. Lecercle, Interpretation as Pragmatics (Basingstoke: Palgrave Macmillan, 1999).

5 L. Carroll, The Annotated Alice, ed. M. Gardner (Harmondsworth: Penguin, 1965), p. 279.

266 Notes

6 V. Woolf, `Kew Gardens', in The Haunted House (Harmondsworth: Penguin, 1973; first published 1944), p. 35.

7 Ibid., pp. 41 2. 8 Stephen Heath (personal communication) has suggested that they also

embody positions of class: the old crones are working class, whereas the old man and his son explicitly belong to the upper classes.

9 A. J. Greimas, SeÂmantique structurale (Paris: Larousse, 1966). 10 Woolf, `Kew Gardens', p. 39 (the italicisation is mine). 11 J. Keats, Letters of John Keats (Oxford: Oxford University Press, 1970), p. 83. 12 Woolf, `Kew Gardens', p. 36. 13 Ibid., p. 40. 14 J. Joyce, Èpiphanies', in Poems and Shorter Writings (London: Faber, 1991),

pp. 161 200. 15 Woolf, `Kew Gardens', p. 40. 16 At least as presented in the text: it could be argued, of course (perhaps it

ought to be argued) that the parallel utterances of clicheÂs make perfect sense and perfect communication: there is always another, positive side to both doxa and clicheÂ. I am indebted to Denise Riley for pointing this out.

17 Ibid., p. 41. 18 The tin opener technique of interpretation is analysed in my Interpretation as

Pragmatics, Chapter 1.

Notes to Chapter 4

1 See J. Elster, Making Sense of Marx (Cambridge: Cambridge University Press, 1985); J. Roemer (ed.), Analytical Marxism (Cambridge: Cambridge University Press, 1986).

2 A. Radford, Syntax: A Minimalist Approach (Cambridge: Cambridge University Press, 1997), p. 113.

3 S. Z Ïek, The Ticklish Subject (London: Verso, 1999), p. 99. The reference to ÏizReve is K. van het Reve, `Reves Vermutung', in Dr Freud and Sherlock Holmes (Hamburg: Fischer Verlag, 1999), pp. 140 51.

4 G. Deleuze and C. Parnet, Dialogues (Paris: Flammarion, 1977), p. 138 (Engl. trans., p. 115).

5 I have tried to do this in J. J. Lecercle, `The Misprision of Pragmatics: Conceptions of Language in Contemporary French Philosophy', in A. Phillips Griffiths (ed.), Contemporary French Philosophy (Cambridge: Cambridge University Press, 1987), pp. 21 40; J. J. Lecercle, `Philosophies du langage analytique et continentale: de la sceÁne de meÂnage aÁ la meÂprise creÂatrice', in L'Aventure humaine, 11 (Paris: PUF, 1999), pp. 11 22.

6 J. L. Austin, `Performative Utterances', in Philosophical Papers (Oxford: Clarendon Press, 1970), p. 239.

7 H. Bloom, The Anxiety of Influence (New York: Oxford University Press, 1973); A Map of Misreading (New York: Oxford University Press, 1975).

8 H. P. Grice, `Meaning', in Essays in the Way of Words (Cambridge, Mass.: Harvard University Press, 1975), pp. 213 23.

9 W. V. O. Quine, Word and Object (Cambridge, Mass.: MIT Press, 1960). 10 J. L. Austin, How to Do Things with Words (Oxford: Clarendon Press, 1962).

Notes 267

11 On this, see my Interpretation as Pragmatics (Basingstoke: Palgrave Macmillan, 1999) p. 156.

12 G. Deleuze and F. Guattari, Qu'est ce que la philosophie? (Paris: Minuit, 1991), pp. 10, 14.

13 G. Deleuze and F. Guattari, Kafka (Paris: Minuit, 1975), pp. 37 8 (Engl. trans., p. 21).

Ïiz14 The terms are borrowed from Z Ïek's discussion of Logique du sens in The Metastasis of Enjoyment (London: Verso, 1994), and his discussion of Badiou in The Ticklish Subject.

15 G. Deleuze, CineÂma 1: L'Image mouvement (Paris: Minuit, 1983), pp. 145 ff. 16 G. Deleuze Nietzsche (Paris: PUF, 1965). 17 G. Deleuze, Nietzsche et la philosophie (Paris: PUF, 1962). 18 Ibid., p. 7. 19 Ibid., p. 60. 20 Ibid., pp. 84 5. 21 L. Carroll, The Annotated Alice, ed. M. Gardner (Harmondsworth: Penguin,

1965), p. 269. 22 Deleuze, Nietzsche et la philosophie, p. 117. This theme remained a constant

preoccupation for Deleuze: a chapter of his CineÂma 2: L'image temps (Paris: Minuit, 1985) is entitled `Les puissances du faux'.

23 G. Deleuze and F. Guattari, Mille plateaux (Paris: Minuit, 1980), p. 174 (Engl. trans., pp. 139 40).

24 G. Deleuze, Foucault (Paris: Minuit, 1986), p. 26 (Engl. trans., p. 17). 25 G. Deleuze, Pourparlers (Paris: Minuit, 1990), pp. 60 1 (Engl. trans., pp. 40 1). 26 Deleuze and Guattari, Mille plateaux, p. 105 (Engl. trans., p. 83). 27 V. I. Lenin, Òn Slogans', in Collected Works, 25 (London: Lawrence &

Wishart, 1964), pp. 183 90. 28 J. R. Ross, Òn Declarative Sentences', in R. A. Jacobs and P. S. Rosenbaum

(eds), Readings in English Transformational Grammar (Waltham, Mass.: Ginn & Co., 1970), pp. 222 72.

29 A. Banfield, Unspeakable Sentences (London: Routledge, 1982). 30 J. C. Milner, De la syntaxe aÁ l'interpreÂtation: quantiteÂs, insultes, exclamations

(Paris: Seuil, 1976). 31 See, for instance, O. Ducrot, Le Dire et le dit (Paris: Minuit, 1994) and O.

Ducrot (ed.), Les Mots du discours (Paris: Minuit, 1980); V. N. Voloshinov, Marxism and the Philosophy of Language (Cambridge, Mass.: Harvard University Press, 1986).

32 Deleuze and Guattari, Mille plateaux, p. 107 (Engl. trans., slightly modified, p. 84).

33 Ibid., p. 106 (Engl. trans., slightly modified, p. 84). 34 Ibid., p. 183 (Engl. trans., p. 147). 35 Deleuze, Pourparlers, p. 44 (Engl. trans., p. 29). In this translation, `mot d'ordre'

is translated as `precept'.

Notes to Chapter 5

1 J. Derrida, `De l'eÂconomie restreinte aÁ l'eÂconomie geÂneÂrale', in L'Ecriture et la diffeÂrence (Paris: Seuil, 1967), pp. 385 6.

268 Notes

2 P. Bourdieu, Sur la teÂleÂvision (Paris: Liber Raisons d'agir, 1996); Contre feux (Paris: Liber Raisons d'agir, 1998); S. Halimi, Les Nouveaux chiens de garde (Paris: Liber Raisons d'agir, 1997); N. Chomsky, Necessary Illusions (London: Pluto Press, 1989).

3 G. Deleuze, Pourparlers (Paris: Minuit, 1990), p. 177 (Engl. trans., pp. 129 30). 4 G. Deleuze and F. Guattari, Qu'est ce que la philosophie? (Paris: Minuit, 1991),

p. 80 (Engl. trans., p. 82). 5 J. J. Lecercle, Interpretation as Pragmatics (Basingstoke: Palgrave Macmillan,

1999). 6 G. Deleuze and C. Parnet, Dialogues (Paris: Flammarion, 1977), pp. 31 2

(Engl. trans. pp. 23 4). 7 Deleuze, Pourparlers, p. 159 (Engl. trans., p. 117). 8 Ibid., p. 172 (Engl. trans., p. 126). 9 Deleuze and Guattari, Qu'est ce que la philosophie?, p. 55 (Engl. trans., p. 55).

10 F. Guattari, Psychanalyse et transversaliteÂ (Paris: Maspero, 1972). 11 Ibid., p. 241. 12 Deleuze and Parnet, Dialogues, p. 125 (Engl. trans., p. 104). 13 The watercolour appears on the cover of J. Hamilton, William Heath Robinson

(London: Pavilion Books, 1992). 14 G. Deleuze and F. Guattari, L'Anti êdipe (Paris: Minuit, 1972), p. 46 (Engl.

trans., p. 38). 15 See G. Bataille, La Part maudite (Paris: Minuit, 1967). 16 Deleuze and Guattari, L'Anti êdipe, pp. 365 6 (Engl. trans., pp. 306 7). 17 Ibid., p. 214 (Engl. trans., p. 181). 18 J. Searle, Intentionality (Cambridge: Cambridge University Press, 1983). 19 G. Deleuze and F. Guattari, Mille plateaux, (Paris: Minuit, 1980), p. 629 (Engl.

trans., pp. 503 4). 20 G. Deleuze and F. Guattari, Kafka (Paris: Minuit, 1975), p. 153 (Engl. trans.

p. 86). 21 Deleuze, Pourparlers, pp. 34 5 (Engl. trans., p. 21). 22 Ibid., p. 35 (Engl. trans., p. 22). 23 Deleuze and Guattari, Mille plateaux, p. 50 (Engl. trans., p. 39). I have already

voiced my doubts about the translation of èÂnonceÂ' as `statement'. 24 Ibid., pp. 162 3 (Engl. trans., pp. 130 1). 25 Ibid., p. 110 (Engl. trans., p. 87). 26 Ibid., p. 112 (Engl. trans., p. 89). 27 Deleuze and Parnet, Dialogues, p. 98 (Engl. trans., p. 79). 28 L. Irigaray, `Du fantasme et du verbe', in Parler n'est jamais neuter (Paris:

Minuit, 1985), pp. 69 80. 29 Deleuze and Parnet, Dialogues, p. 114 (Engl. trans. p. 95). 30 Deleuze and Guattari, Mille plateaux, p. 318 (Engl. trans., p. 261). English

translations of Deleuze use two forms for the word `hecceÂiteÂ': `haecceity', or `haeccity'.

31 Ibid., p. 322 (Engl. trans., p. 263). 32 For a study of the genre, see L. Tallairach's thesis, `Le Secret et le corps dans le

roman aÁ sensation anglais au 19e sieÁcle', University of Toulouse (2000). Ïiz33 On this, see Z Ïek's discussion of the concept `film noir', in J. Butler, E. Laclau

Ïizand S. Z Ïek, Contingency, Hegemony, Universality (London: Verso, 2000), pp. 242 5.

Notes 269

34 Deleuze and Guattari, Kafka p. 33 (Engl. trans., p. 18).35 G. Deleuze, Critique et clinique (Paris: Minuit, 1993), p. 15 (Engl. trans., p. 5).36 Deleuze and Guattari, Mille plateaux, pp. 130 8 (Engl. trans., pp. 100 10).37 Deleuze and Parnet, Dialogues, p. 72 (Engl. trans., p. 58).38 See E. Laclau and Mouffe, Hegemony and Socialist Strategy (London: Verso,

1985); also Butler, Laclau and Zizek, Contingency, Hegemony, Universality. 39 D. Riley, The Words of Selves (Stanford, Cal.: Stanford University Press, 2000).


1 R. Kipling, `Wireless', in Traffics and Discoveries (Harmondsworth: Penguin, 1987; first published, 1904), pp. 181 99.

2 J. Keats, The Complete Poems (Harmondsworth: Penguin, 1973), p. 346. 3 Ibid., p. 312. 4 R. Kipling, Mrs Bathurst and Other Stories, ed. Lisa Lewis (Oxford: Oxford

University Press, World's Classics, 1991), p. 287. 5 G. Deleuze, Proust et les signes, 2nd edn (Paris: PUF, 1970) (Engl. trans., Proust

and Signs, New York: George Braziller, 1972).

Notes to Chapter 6

1 G. Deleuze, Proust et les signes (Paris: PUF, 1964; 2nd edn, 1970) (Engl. trans., Proust and Signs, New York: George Braziller, 1972).

2 J. Derrida, L'Ecriture et la diffeÂrence (Paris: Seuil, 1967). 3 G. Deleuze and F. Guattari, Mille plateaux (Paris: Minuit, 1980), p. 388 (Engl.

trans., p. 276). 4 Ibid., pp. 342 3 (Engl. trans., pp. 280 1). 5 G. Deleuze, Critique et clinique (Paris: Minuit, 1993), p. 9 (Engl. trans., p. iv). 6 G. Deleuze and F. Guattari, L'Anti údipe (Paris: Minuit, 1972), p. 158 (Engl.

trans., p. 133). 7 Ibid. 8 Deleuze, Critique et clinique, p. 142 (Engl. trans., p. 113). 9 G. Deleuze and F. Guattari, Qu'est ce que la philosophie? (Paris: Minuit, 1991),

p. 166 (Engl. trans., p. 176). 10 G. Deleuze, Pourparlers (Paris: Minuit, 1990), p. 157 (Engl. trans., p. 115). 11 Ibid. 12 Deleuze, Critique et clinique, p. 141 (Engl. trans., p. 111). 13 Anon., Augustus Carp, Esq., by himself (Harmondsworth: Penguin, 1987; first

published, 1924), p. 3. 14 G. and W. Grossmith, Diary of a Nobody (Harmondsworth: Penguin, 1965;

first published, 1894). 15 Augustus Carp, Esq., by himself, p. 4. 16 S. Beckett, Murphy (London: Picador, 1973), p. 5. A longer version of this

section was published as `BeÂgayer la langue Stammering language', in L'Esprit creÂateur, xxxviii (1998), pp. 109 23. I have changed the title to conform to the translation of `beÂgayer' in the English translation of Critique et clinique.

17 L. Jenny, La Parole singulieÁre (Paris: Belin, 1990).

270 Notes

18 L. Carroll, `The Walrus and the Carpenter', in M. Gardner (ed.), The Annotated Alice (Harmondsworth: Penguin, 1965), p. 233.

19 Deleuze, Critique et clinique, pp. 135 43 (Engl. trans., `He stuttered', pp. 107 14).

20 Ibid., pp. 138 9 (Engl. trans., p. 110). 21 P. Simpson, `The Pragmatics of Nonsense', in M. Tooler (ed.), Language, Text

and Context: Essays in Stylistics (London: Routledge, 1992). 22 Private Eye's Colemanballs (London: Private Eye/AndreÂ Deutsch, 1982). 23 L. J. Austin, How to do Things with Words (Oxford: Clarendon Press, 1962). 24 Deleuze, Critique et clinique, pp. 140 1 (Engl. trans., p. 112). 25 Ibid., p. 141 (Engl. trans., p. 112). 26 See P. Sauvanet, Le Rythme grec d'HeÂraclite aÁ Aristote (Paris: PUF, 1995). 27 Deleuze, Critique et clinique, p. 141 (Engl. trans., p. 112).

Notes to the Conclusion

1 R. Rorty (ed.), The Linguistic Turn: Recent Essays in Philosophical Method (Chicago: University of Chicago Press, 1967).

2 J. F. Lyotard, Le DiffeÂrend (Paris: Minuit, 1983). 3 G. Deleuze, Logique du sens (Paris: Minuit, 1969), p. 301 (Engl. trans., p. 261). 4 G. Deleuze, L'AbeÂceÂdaire (Paris: Editions de Montparnasse, 1977), `F comme

FideÂliteÂ'. 5 G. Deleuze, Critique et clinique (Paris: Minuit, 1993), p. 124 (Engl. trans.,

p. 98). 6 Ibid., p. v. The French has `Les beaux livres . . .', which sounds slightly less

banal. 7 L. Althusser, Philosophie et philosophie spontaneÂe des savants (Paris: MaspeÂro,

1974; first published 1967). 8 G. Deleuze, CineÂma 2: L'Image temps (Paris: Minuit, 1985), p. 41 (Engl. trans.,

p. 27). 9 P. P. Pasolini, Empirismo eretico (Milan: Garzanti, 1972).

10 Deleuze, CineÂma 2, pp. 43 4 (Engl. trans., p. 29). 11 C. Panacio, Le Discours inteÂrieur (Paris: Seuil, 1999). 12 Deleuze, Cinema 2, p. 45 (Engl. trans., p. 30). 13 F. Ravaisson, De l'habitude (Paris: Rivages, 1997; first published 1894),

pp. 78 9. 14 For a remarkably pedagogic exposition of the complexities of the bottle

opener diagram, see S. Z Ïek, The Sublime Object of Ideology (London: RoutÏizledge, 1989), pp. 100 14.

Index

Althusser, L., 73, 108, 163, 169, 183 4, 185, 189, 199, 200

AND, 75, 123, 177 Antin van Rooten, L. d', 126 Aristotle, 70, 121, 128, 139, 163, 167,

178, 252, 253 Artaud, A., 103, 122, 179, 196, 200,

233, 244 assemblage, 34, 54, 60 1, 88, 159, 164,

183, 184 93, 195, 199, 216, 218, 224, 255

Austin, J. L., 28, 161, 162, 163, 240 Autrecourt, N. d', 137

Badiou, A., 21, 33, 90, 109 13, 117, 123, 128, 165

Balzac, H. d', 222 Bakhtin, M., 136, 172. Banfield, A., 170 Barthes, R., 46, 220, 248 Bataille, G., 175, 183, 196, 199 Beckett, S., 1 7, 27, 84, 147, 195, 196,

200, 222, 224, 226 30, 238, 239 Benveniste, E., 9, 72, 91, 93 4, 156,

188 Bergson, H., 20 1, 29, 106, 115, 253,

255 Blanchot, M., 59, 112, 196, 199 Bloom, H., 162 Bourdieu, P., 19, 34, 71, 157, 175,

200 Braddon, M. E., 191, 192, 193 Bronte, C., 226, 227 Broughton, R., 191, 192 Buffon, G. de, 220 Bush, G. W., 194

cËa, 9 12, 13, 18 Canetti, E., 38 Carp, A., 225 Carroll, L., 43, 53 5, 99, 100 2, 103,

113, 122, 123 4, 127, 129, 132 3, 140, 229 30, 239

Carver, R., 35 CeÂline, L F., 196, 242, 244 chaos, 102, 103, 215 chiasmus, 146 7 Chomsky, N., 24, 29, 33, 40, 70, 71, 75,

76 7, 78, 92, 94, 96, 97, 98, 118, 124, 133, 139, 155 8, 159, 160, 168, 170, 175, 188, 200, 223, 231, 232 4, 244, 250

Christie, A., 39, 193 Chronos/Aion, 117, 128 Cixous, H., 137 Coleman, T., 43, 48 Colemanballs, 234 9 Collins, W., 191, 192, 193 Compagnon, A., 19 correlation, 16 17, 55 6, 71 2, 84, 92,

120, 127, 129 30, 254 5 counter effectuation, 116 Crane, S., 109, 116, 118 Culioli, A., 93, 156, 188 cummings, e.e., 96, 127, 138 43, 233

Democritus, 104 denotation/manifestation/

signification, 120 2, 126, 134 7, 185, 186, 187, 195

de /reterritorialisation, 80, 164, 165 Derrida, J., 19, 20, 35, 88, 175, 199 Descartes, R., 9, 12, 13, 60, 119, 155 desire, 185 Dickens, C., 9, 43 8, 50 2, 214 Doyle, A. C., 203, 209 Ducrot, O., 119, 120, 170

eÂcriture, 220 1 Eisenstein, S., 252 Engles, F., 222 eÂnonceÂ, 28 Elster, J., 135 error, 166 7, 179 event, 104, 105, 108 18, 130, 151, 153,

175, 186

271

272 Index

faciality, 35, 62 force, 164, 185 73, 253 Foucault, M., 14, 24, 28 9, 30, 69, 78,

165, 188, 199, 200, 223, 224, 244 Freud, S., 190

Gainsbourg, S., 173 glottophagy, 197 Gobard, H., 91 Greimas, A. J., 148 9 Grice, H. P., 28, 162 3, 164 Gualandi, A., 22 Guattari, F., 15, 31 6, 62, 83, 102, 175,

180 2, 200, 218, 247 Guillaume, G., 76, 93, 94, 241

Habermas, J., 248 haeccity (haecceity), 93, 190 1 Hamon, P., 45 Hardy, T., 192, 193, 197, 227 Heath Robinson, W., 181, 182, 216 Hegel, G. F. W., 13, 16 Heidegger, M., 13, 223, 255 Helmslev, L., 24, 76, 91 2, 108, 154,

187, 224, 240, 241 Hobbes, T., 92 Holmes, S., 38 9 Honegger, A., 52 Humboldt, W. von, 29 Hume, D., 106 Husserl, E., 27

ideology, 169, 185 image, 57 image of thought, 76, 103, 178, 214,

247, 248 immanence, 91, 154, 217, 247, 254 imperialism of signifier, 24, 78 82,

119, 249 infinitive, 114, 116, 190 interest, 164, 175 80 interpretation, 185 6, 187 Irigaray, L., 116 17

Jakobson, R., 17, 27, 241 Jarry, A., 250 Jenny, L., 227 Johnson, M., 52 Joyce, J., 133 8, 152, 195, 200, 239

Kafka, F., 25, 27, 59, 182, 188, 194, 200 Kant, I., 70 1, 161, 163, 177, 250 Keats, J., 109, 118, 150, 202 17 Kenner, H., 136 Kerouac, J., 197 Kipling, R., 202 17 Klein, M., 100, 117 Kristeva, J., 19, 199

Labov, W., 91, 92, 224 Lacan, J., 15, 19, 63, 78 80, 81, 96, 100,

108, 111, 113, 116, 118, 123, 152, 180, 183, 187, 248, 249, 256

Laclau, E., 198 Lakoff, G., and M. Johnson, 11, 49 lalangue, 86, 167 Langacker, R., 93 langue/parole, 64, 66, 67, 85, 88 9, 91,

98, 106, 181, 231 2 Lawrence, D. H., 221 Lawrence, T. E., 241 Leclaire, S., 68 Leibniz, G. W., 12, 13, 29, 87, 166, 256 Lenin, V. I., 169 70, 174, 177, 181 LeÂvi Strauss, C., 15, 19, 108 literature, 30, 61, 64 5, 83, 105, 130,

195 6, 199, 255 Lowry, M., 72 Luca, G., 233, 241 Lyotard, J. F., 20, 90

Macherey, P., 222 machine, 164, 180 4, 185, 189, 195,

214 17 MallarmeÂ, S., 76, 104 Markov chain, 42, 94 7, 118, 123, 134,

183, 184, 196, 232, 236, 238, 243, 256 7

Marr, N., 88 Martin, J., 45 Marx, K., 183 4, 185, 189, 193, 197 Masoch, S., 196 Meinong, A., 228 Melville, H., 197, 233, 241 metalanguage, 70 metamorphosis, 25, 58 9, 219, 246 metaphor, 25 7, 51, 54 5, 57, 59,

136 8, 195, 219, 236 Metz, C., 251, 253

MicheÂa, A., 234 Miller, H., 221 Milner, J. C., 85 6, 87, 98, 167, 170,

175, 233 minority, 23, 89 90, 138, 164, 193 9,

221, 237, 241, 243, 255 modernism, 2, 199 200, 218, 219 monad, 37 mot d'ordre, 87, 164, 169 72, 234, 255 Muller, M., 229

Nietzsche, F., 165, 166, 170, 174, 175, 178, 192

Ngugi, 195

Panini, 85 paradoxes of sense, 122 3 Pasolini, P. P., 251, 252, 255 6 Paul, 109 PeÂguy, C., 241, 243 Peirce, C. S., 82, 249 phantasy, 116 17 plane (of immanence/of consistency),

102, 103, 114, 130, 149, 185, 187, 217, 241, 246, 255

Plato, 59, 77, 102, 103, 157, 178, 219 Poe, E. A., 116 point de capiton, 97, 118, 119, 236, 243,

249 Popper, K., 161, 179 postulates of linguistics, 84 90 pragmatics, 33 4, 72, 87 8, 159 65,

168, 175, 188, 198, 199 problem, 2 3, 7, 21, 36 40, 72, 100,

104, 255

Quine, V. W. O., 156, 163

Radford, A., 157 8 RancieÁre, J., 200 rapports de force, 28, 89, 166, 173, 174,

175, 178, 198, 200, 218, 255 Ravaisson, E., 255 remainder, 89, 167 representation, 43, 51 2, 54, 56 60,

73 5, 214, 248, 251 rhizome, 78, 255 Richardson, S., 191 2 Riley, D., 200

Index 273

ritournelle, 36, 62, 80 Rorty, R., 247 Rosenstiehl, P. and J. Petitot, 78 Ross, J. R., 170 Roussel, R., 241, 243 Ruyer, R., 95 Ryle, G., 60

Saumjan, S. K., 19 Saussure, F. de, 34, 64 5, 75 6, 78 80,

81, 82, 85, 88, 91, 106, 138, 154, 156, 160, 162, 172, 212, 227, 248, 249, 252

Schulz, B., 61 Searle, J., 20, 28, 186 semiotics, 34 5, 81 2, 254 sensation novel, 191 3 sense, 100, 102 3, 117, 118 23, 130,

134 5, 140 3, 153, 165, 175 series, 103, 105, 106, 108, 150 1 Shakespeare, W., 167 Shelley, M., 111 sign, 69, 82 4, 210 14, 217, 219 20,

248 50, 254 Simondon, G., 14 singularity, 104, 105, 106, 182 slogan see mot d'ordre Smith, H., 203 Spark, M., 239 Spinoza, B., 31, 37, 92, 166, 240 Stalin, J., 88, 161 Stein, G., 243 Stoics, 101, 103, 113 14, 132 3, 163,

166, 186 Stoker, B., 193, 203 structure, 165 8, 180 2, 254 stuttering, 138, 159, 224, 228, 230 4,

235, 236, 239, 240, 256 style, 64, 67, 68, 164, 180, 219, 220 5,

234, 239, 242 6, 256 subject, 81, 108, 114, 181, 182, 183 9,

255 Swift, G., 41 2 syntax, 223, 233, 234, 237, 238, 242,

243, 255

Thomas, D., 197 Trier, L. von, 60 truth, 170, 179

184

274 Index

variation, 65, 237, 242, 243, 253, 255 Villa Lobos, H., 60 visibilities, 29, 244 Voloshinov, V., 48, 170 1, 173 3, 175,

200, 255

Wittgenstein, L., 9, 12, 20, 63, 90, 163,

Wood, Mrs H., 191 Woolf, V., 143 53, 221 word, 74

Zizek, S., 111, 112, 113, 118, 119, 158 Zola, E., 46

Lecercle - Deleuze and Language

Documents

reader language

order isbn

standing order

psychoanalysis theresa

england deleuze

published titles

society series

social order