University of Virginia Library

Does "Text" Exist?
by
Louis Hay [*]

"No it doesn't": on this note, Jacques Petit concluded the first debate dealing with the production of the text and with literary manuscripts, ten years ago.[1] Looking back, what strikes us about this comment is not so much its provocativeness (or rather resignation?) as simply its date. From this point onwards, the text started generating new questions. At the same time appeared a textual approach, called today genetic criticism ("critique génétique"), and it is interesting to take a closer look at the rapport between these two events. But in order to speak of both genesis and text, we need first of all to redefine these terms, or rather to understand their original meaning. And so we must go back much further than ten years.

Historically, the notion of text has undergone a rather remarkable development, beginning with a long period of stability which was followed recently by a brief period of multiple mutations. The former period takes us back to the monastic Middle Ages, when by the 13th century in all European languages the word textus, meaning fabric, became text with the present accepted meaning for that term. At the beginning of the modern age, its meaning had remained unchanged. In the 1786 edition of the French Academy dictionary, it is defined as: "an author's own words, as opposed to notes, commentaries or glosses" ("les propres paroles d'un auteur, considérées par rapport aux notes, aux commentaires, aux gloses"), and the following example is given: "The text of Holy Scripture. It is the pure, formal text. To restore a text." ("Le texte de l'Ecriture Sainte. C'est le text pur et formel. Restituer un texte.")[2] Two contrasting notions are thus sharply distinguished: the text ("the author's own words") as opposed to the gloss ("notes, commentaries"). Each of these notions has precisely defined functions. The text

is "pure and formal"—its very form guarantees the pureness of its meaning, and the conditions of social transmission ensure its spiritual legitimacy. At the same time the functions of gloss are defined: phiology guards over the literal accuracy, it has to "restore the text"; the commentary is in charge of meaning. The reference to Holy Scripture reminds us that what is implied in such an arrangement is a distinction between the sacred and the profane. The critical edition, even secularised, has always remained faithful to its original calling—to this day it claims to carry the word of righteousness, and to reinforce the legitimacy of text.

The definitions were all the more firmly rooted, given that for centuries they formed the accepted tradition in all European cultures. The German Adelung dictionary echoes its contemporary, the French Academy dictionary: "The words of an author as distinct from their interpretation (. . .) in this way the passages of the Bible underlying the sermon are most specifically considered texts". ("Die Worte eines Schriftstellers, zum Unterschiede zu der Auslegung derselben (. . .) in welchem Verstande besonders die biblischen Stellen über welche gepredigt wird, Texte heissen.").[3] Charles Richardson's dictionary in Great Britain, though of later date, maintains the definition without noteworthy change: "A composition in writing, opposed to the notes or annotations." An example is given from Chaucer: "in plain text, withouten nede of glose", and the authors specify: "Text is technically applied to any passage quoted from the text of Scripture, as a subject of discourse or sermon."[4] This canonical understanding persists unchanged throughout a long period of cultural revolution. More than a century after the French Academy dictionary, Pierre Larousse uses a virtually unchanged definition, despite the generally distinctive perspective of his dictionary: "An author's own writing as opposed to the commentaries (. . .) The text of Holy Scripture (. . .) To restore the text." ("Propres paroles d'un auteur par opposition aux commentaires (. . .) Le texte de l'Ecriture Sainte (. . .) Restituer un texte.")[5] As the only signs of the times, as it were, it cites "the text of Plato" alongside the Bible, and mentions "translations" in contrast to the text. At around the same time, the monumental Grimm Brother's Wörterbuch gives the definition: "the original, fundamental, 'ur'-text, as opposed to the translation (. . .) the substantive words of a piece of writing as opposed to commentaries or interpretations: in the strict sense verse reference (Biblical text) from a sermon or speech." ("der Original-Grund-Urtext im Gegensatz zur Übersetzung (. . .) die Hauptworte einer Schrift im Gegensatz zu den Erläuterungen und Anmerkungen, im engeren Sinne der Grundspruch (Bibeltext) einer Predigt oder Rede.")[6] And leaping ahead one more century

to the present day, the same strict frame of reference remained unchanged up until a decade ago. "An author's own words, a law (as opposed to the commentaries)" ("Propres termes qu'on lit dans un auteur, une loi (par opposition aux commentaires)")—thus the definition in the 1964 edition of the Grand Larousse.[7] And in 1973, in the Brockhaus: "the words of a piece of writing or a lecture; Biblical passage underlying a sermon." ("Wortlaut eines Schriftwerkes, eines Vortrags; die der Predigt zugrunde gelegte Bibelstelle.")[8]

To the best of my knowledge, the first dictionary to mark a change is Le Robert, where as early as 1966 "literary work" ("oeuvre littéraire") is included in the definition, and "a well-written text" ("texte bien écrit") is used as an example.[9] Such uses are carefully identified as neologisms, however. A good indication of the rapid subsequent development can be seen by comparing the timidity of this entry with Julia Kristeva's imperious statement barely six years later: "We can no longer speak of 'literature' in general, but of the text."[10] However the entry in Le Robert is interesting because of its date: 1966 also saw the appearance of Structural Semantics (Sémantique structurale) by A. L. Greimas, George Poulet's Three Essays (Trois essais) and Gérard Genette's Figures I. It is remarkable, too, since this new dictionary was the work of a team of young linguists and semioticians, and the influence of the theories they represented was to be the motor of impending developments. As Roland Barthes later remarked: "It was at the height of structural linguistics (around 1960) that new researchers, themselves often the product of a linguistic training, began to formulate a critical approach to the sign and a new theory of the text."[11] At the same time, the direction of research was to vary from country to country, thus causing the sense of the term to fragment. It was no longer possible to discuss "the text" in an international forum without first establishing one's own definition of the term.

In Germany, and to a lesser extent in the United States, a novel assessment arose around editorial enterprises. Attention shifted from the gloss to the text when editors faced the problems of textual genesis by means of the apparatus of variants. It was from this angle that critics like Beda Alleman and especially Peter Szondi in Germany, or Vinton Dearing in the United States and Philip Gaskell in Great Britain, turned their interest to the problems of text production. But at the time there was no exchange with other theoretical approaches—neither the Anglo-American "new criticism" nor the German "Textlinguistik". In France, matters took a somewhat different turn, since a new current of criticism developed based largely on linguistic models. As Barthes put it: "A wide literary field has gone over to linguistics under the heading of poetics (a

displacement which Valéry had recognised as necessary)."[12] In this critical approach, the text itself is the focus of attention, situated between "the author's own words" and the eyes of the reader, but considered above all as an autonomous object. The aim of the critic is to build up the text as a scientific object to analyse with precision its constitutive system of forms, significations and functions. Despite its relatively limited lifespan (essentially from the late fifties to the early seventies), this innovative critical movement was to have long-felt consequences. And although, on the level of discourse, systems of textual syntax or semantics could not be established on the analogy of models developed by formal linguistics, a certain number of concepts, such as sign, structure or function, became increasingly articulate in theoretical works that brought about an irreversible renewal in literary criticism.

Yet somehow this widespread movement remained closely tied to its first origins, almost as by an umbilical cord. The new critical theory still assumed the classical notion of text, the "pure, formal text", incommensurable and opposed to anything not itself, albeit this opposition was no longer founded on a religious tradition but on an epistemological exigency: the closure of the text. As Michel Arrivé remarked (in 1978, and thus with a certain distance): "In structuralism, the text is considered as a finished product (. . .) categorically distinct from the pre-text and the post-text which remain peripheral (though they are both to be found within the work which thus seems to include the text, the pre-text and the post-text)." He also pointed out amongst the numerous ambiguities of such a postulate the "intricate question of the relations between such terms as discourse, text and narrative".[13] In today's light it is easy to discover the indication of a fundamental problem within this terminological ambiguity: that is to say the application of a new theory to an old object. Once again Roland Barthes was first to recognise the need for a second theoretical step designed to create a new object. In the early seventies, this was the goal of the theory of the text, which takes into account the social dimension by means of historical materialism, and the personality, by means of psychoanalysis. The interchange between these two fields creates a vast theoretical network whose pattern cannot be examined in this paper. One result however was the emergence of the double notion of pheno-text/geno-text, borrowed from Soviet linguists by French semioticians. In 1972 Julia Kristeva was the first to present the pheno-text as "a finished product: an utterance with a meaning". As such it is the empirical manifestation of the geno-text, "an infinite syntactic and/or semantic generation (. . .) which cannot be reduced to the generated structure."[14] In this construction, the deepest reality of the text lies in its productivity, "the limitless possible operations", not in

the product itself. With Barthes' neat comment "The Work is in your hand, the text in the language", the terms as originally canonized were completely inverted. The boundaries disappeared, not only between pretext, text and post-text, but also between writing and criticism. Both became subject to infinite significations. Due perhaps to this circularity, the theory of the text provoked a new problematic rather than concrete applications, since it no longer assigned a specific function to criticism.

*

This situation is in sharp contrast with a second critical approach which also took form at the beginning of the seventies, known as genetic criticism. This second current is contemporary with the theory of the text, and alike deals with the relationship between the text and its genesis, and with the mechanics of text production, the activity of the writing subject. But there, all resemblance ends. Both the method and the object of study of genetic criticism are quite distinct. Its method is the result of extensive empirical work dedicated to authors' manuscripts. It appeared progressively that these documents, under certain conditions, allowed to reconstruct the genesis of writing. Genetic criticism retains from its origins an inductive approach, which builds up general models from a series of concrete observations. As for its object, it has a dual status: it is both a material given, as a document of study, and an intellectual construction, as a pre-text. From the graphic pattern of writing immobilised by the pen and scattered over the page it is possible to reconstruct the process of creation and thought through a fully conceived sequence of analytic operations: deciphering, establishing the chronology, seizing the writing in its moves.[15] In this way, genetic criticism has to consider first the manuscript, then the writing, before finally getting back to the text on a new level.

The principal merit of manuscripts is that they demonstrate the limitations and possibilities of genetic criticism. From the outset there are important material limitations: it is impossible to study a nonexistent manuscript. In this regard, Monsieur de la Palice reminds us of the fragile process of text transmission which today we tend to overlook. Our notion of European literature would be radically different however were it not for the fortuitous survival of such unique manuscripts as Pascal's Thoughts (Pensées) or the Urfaust by Goethe, Lucien Leuwen or Kafka's great novels. Likewise, how different would it be had we inherited the great works which have disappeared, whose ghost-list we can barely establish? Coupled with the caprices of history are the imponderables of the human mind. Even the most detailed and well-conserved documentation reveals but a fraction of the complicated mental processes

to which it bears witness. The ink on the page is not the writing itself. One need only meet a contemporary writer, with evidence in hand from his manuscripts that one has had a chance to study, to be effectively cured of any presumptions one may have had to the contrary.

But manuscripts also give a new power to literary critics. They make it possible to examine how the pen works in its irrefutable material presence. In this way they manifest a level of reality to which no speculative interpretation can penetrate and possess a material richness that no effort of analysis can hope to exhaust. This becomes even clearer when we realise that manuscripts by their singular properties force us to change our habits of thought. They force us to take into account the unpredictable, since our knowledge changes every time an important document is discovered or a new technology gives access to previously unknown information. Likewise we must come to grips with their heterogeneity, since they are diverse by nature: sometimes they are the testimony of the original stimulation, sometimes the record of the remote memory like notes, notebooks or diaries; sometimes they document early operations like projects, workplans or scenarios, sometimes they are the instruments of revision such as sketches, early versions and most often rough drafts. Their polymorph structure is yet another challenge, as manuscripts have no respect for the convention of linearity, overflowing the page into multiple spaces. The ways in which the text is laid out on the page, with marginal notations, additions, cross-references, deletions, alterations, in different handwriting styles, and with drawings and symbols, texture the discourse, increase the significations and multiply the possible readings. In extreme cases, a single word isolated by the writing (Jean Levaillant called them "keystones" in reference to the manuscripts of Paul Valéry: to look, shine, incest)[16] can attract the meaning of an entire page. Or in the notebooks for Finnegans Wake, for example, a single iconic mark may determine the theme, the character, or the mythical entity which will govern the function of a certain set of words.[17] It is easy to see how thoroughly our analytic habits are shaken by the very contact with the material of genetic criticism.

This is even more the case when we pass from the original material to an intellectual construction. The constitution of the pre-text from handwriting which is both solidified and polymorphous implies a new type of reading which must take into consideration the totality of semantic and semiotic significations to be found in a page of writing. In this way a truly dialectic relationship is established between the document and the pre-text: each transcription of the manuscript is shaped by the vision one has of it, but at the same time this vision has to be modified by the reality of the object to be turned into an adequate representation

of it. This relationship and this polarity are evident in the diversity of the methods used to make textual genesis readable. Henri Bonnet and Bernard Brun, in dealing with Proust's manuscripts, try to clear the ground for a continuous reading, thus offering a spectrum of pre-texts in contiguity.[18] On the other hand, Dietrich Sattler attempts to reconstruct the depth of field of genesis in Hölderlin's manuscripts, and presents the witnesses to the stages of revision in their chronological sequence.[19] For Joyce or Kafka, Hans Walter Gabler or Gerhard Neumann both try to embrace these two aspects simultaneously, thus publishing the genesis in its process and the final state of the text side by side or alternately.[20] Each of these choices testifies to the diversity of manuscripts as well as to that of the theoretical options of the editor. But they all mark their distance from the traditional apparatus of variants, abandoning the viewpoint of pure erudition for the problematic of the pre-text.

This problematic is obviously multi-faceted. In defining the pre-text as a constructed object, one must accept the existence of a variety of possible constructions—a fact true both for writing theories and for editorial practices. An examination of the scope of such theoretical possibilities would necessitate a general overview of genetic studies, which is not the object of this paper. Rather, in order to clarify the status of the text, alone in question here, we will satisfy ourselves with a look at the two prime analytic models which deal with the two aspects of writing, production and product.

Text production brings forth two new types of approach. The first is strictly analytic, attempting to identify and describe the combination of transfers, substitutions, extensions and reductions manifest in manuscripts, in order to apprehend the whole gamut of genetic operations and put them into a system: programmation, textualisation, transformation. The second approach is inductive. Its aim is to trace back these operations to the dynamic forces which actuate them: affective impulses, representations of imagination, linguistic or rhythmic values. . . . The study of the product has more substantial support. Each in turn, new criticism, werkimmanente Interpretation and French structuralism have offered models for examining the forms, significations and effects of the text. The new problem, the implications of which remain unknown, concerns the relationship between the genetic approach and the textual approach, a problem which we will deal with as modestly as possible, beginning with a return to terminology.

In 1974 Jean Bellemin-Noël created the immediately successful term "avant-texte", or "pre-text", to encompass the great variety of documents which until then had had no specific nomenclature.[21] Being in full accordance

with the creative singularity of the French language, this term gave birth, by symmetry, to a number of subsequent terms, including the still surviving "après-texte", or "post-text". But by creating the pair "text" and "pre-text", the old binary opposition between text and non-text was rekindled, and a theoretical problem articulated. Trying to define this problem, Jean Bellemin-Noël commented in his Essays on Generative Criticism (Essais de critique génétique (1979)): "The difference between The Text (finished, in other words: published) and the pre-text is that the former offers itself as an entity spellbound in its destiny, whereas the latter holds and reveals its own history."[22] It is clear that this definition, both dense and subtle, favours objective definitions: the text is considered achieved when published, and its own determination—meaning both integrity and durability—is the result of this social destiny. Other factors can also be distinguished in the background: the author's intention which becomes manifest in the act of publication, and the internal coherence of the work, its "unity". We might simply add the act of reading which each time brings the text to existence—and in so doing we rediscover the criteria according to which the text is normally defined: the author, the work, the reader, the society. Each of these elements is the object of much debate, which I shall not attempt to reproduce here. But the difficulty of ascribing a general function to any of them can be understood by a brief look at certain ranges of evidence.

Invoking the social aspect of the text means placing it within an historical context, which at the same time brings into question the fluctuation of cultural norms and the variations of our own criteria. We now accept as texts many works which in the 19th century or even at the beginning of this century would have been regarded as mere files (such as the work of Arno Schmidt), rough drafts (the writing of Francis Ponge, for example), or collages of quotations (such as the early work of Philippe Sollers)—not to mention the archetype of Finnegans Wake which was called a "cross-word puzzle fanatic's Bible" by its contemporaries before becoming a prototype for present-day writing. And it is clear that a look at periods other than our own would give similar results.

Next to consider the function of reading is even more telling, since it does not exclusively rest in the accidents of history. By its very nature, the act of reading converts its object into a text. It is well-known how Mallarmé conceived of reading in absolute terms, considering any visual focal point as text including the countryside one might see in lifting one's eyes from one's book. But remaining within the written world, we easily mark that it is the act of reading which allows us to go from the manuscript to the pre-text, and which sometimes weakens the distinction between

manuscript and text. Such is especially the case when rough drafts are presented for reading—a practice which, with the recent publication of literary manuscripts in editions, now concerns the general public.

In turn, the contingencies of history and of reading call into question the structural coherence of the text. This is an objective criterion in so far as it is based on the laws which govern the language. Yet these laws, and even proper grammar, are far better respected in the rough sketches of Flaubert and Proust than in the final texts of Céline or Joyce. Such a consideration does not nullify the criterion of coherence, but rather acts as an incentive to place it in relation to the other variable determining factors. The last among these to be questioned is the writing subject, that is, the author, whose decision it is to cut the umbilical cord of the genesis, ushering the pre-text into the text.

The writing subject, of course, is the most evident and perhaps the surest of the criteria for textuality, in that it conjoins an individual determination and a social reality; in any case it is the easiest to handle. But it is not an absolute criterion. Its two constituent parts are far from being always related. Either the author's intention is defeated by economic or political obstacles ("A whole brood of birdlings chirping in my head / Books good for confiscation", as Heinrich Heine put it, making himself the mouthpiece of many of his contemporaries in Germany: A Winter's Tale), or on the contrary, the work is published against the author's desire, or posthumous desire as was the case with Kafka. But even when such is not the case, the relationship between what is non-published and what is published is as diverse as are the ways of writing: from Claude Simon who destroys his manuscripts after publication, to Julien Gracq who guards them under lock and key; from Aragon who gave his manuscripts to researchers, to Francis Ponge who closes the circle by publishing his rough drafts, thus transforming the pre-text into text.

*

Clearly none of the criteria of textuality is in itself a constant certain factor, nor can such certainty be established by putting them all together. The history of the text, its internal coherence, the act of reading and the author's design do not constitute a system. We have no four-faced prism which allows us infallibly to consider a given literary work as a text, and the search for such an instrument would undoubtedly be as futile as was that for the ultimate criterion of literariness. Must we simply conclude that the text does not exist? It seems to me that it should be sufficient to agree that there is no absolute definition. And the above-mentioned criteria remain operative as long as we accept them as parameters in a

variable field registering ever-differing realisations of the act of writing. Not The Text, but texts. And so I prefer to modify the concept and vocabulary in order to take a new look, not at the opposition between pretext and text, but at the relationship between writing and the written work.

Today, this relationship strikes us as being profoundly contradictory, undoubtedly because of the difficulty which we still have in understanding it. As we have just seen, it is impossible to separate categorically the act of writing from the written work. But likewise, it is impossible to confuse them one with the other. Between creative process which operates through repetition, contiguity, extensions and retractions, and the deployment of the text, there is no possible correspondence. On the contrary: experience has shown that the process of writing cannot be conjectured by studying the text to which it nevertheless gave birth. The trip backwards, from the text to the manuscript, is full of discoveries and often astonishment: one's expectations are foiled. The process of writing is of a different nature than the written word.

This difference is not disturbing as long as one seeks simply to describe the genetic process. Such an approach is well suited to dealing with heterogeneous conditions, since it functions from a base of displacements, substitutions and transformations, not stopping short at the text itself which is considered only as a stage of the genetic process. It matters little that the text should be seen as the outcome of preceding stages or considered merely within a set of many alternative versions. The difficulty arises when one tries to integrate the generative dimension within the study of the text. It is easy enough to say that the writing is always active within the written work, that "the ghosts of successive books", as Julien Gracq puts it, continue to inhabit the finished work which remains situated "in their light".[23] But how does the writing shed light on the text? And what can the one tell us about the other? Faced with this problem, I will once again make a slight detour by making two comments.

The first concerns our reflection—or should I say our lack of reflection?—on the question of the author. The writing subject has little place in modern criticism, fallen into disrepute first because of the banality of biographical commentaries, and subsequently removed from the text by the strict theoretical approach of formal analysis. Yet he resurges today as the subject of new questioning. In dealing with writing, criticism inevitably encounters the moment of the writing itself. It is stretched out between the author's life and the sheet of paper like a drumskin on which the pen beats its message. The echo we receive is but incomplete, yet reveals the complexity of the act itself, not to mention its contradictions. These contradictions come not only from the fact that writing tends alternately

(and sometimes simultaneously) towards communicating and holding back, but also from the contiguous use of imagination and meditated calculation. This dynamic opposition creates a polarity configuration which seems to be inherent in writing, a state which influences as much the concepts we use for dealing with the text as it does our understanding of the text itself.

To begin with, our concept of the work. For Kafka, for example, we possess a series of notebooks, of "writing" in which he has attempted to delimit and form entities—in other words, "works"—yet with great difficulty, judging by the successive presentation of about fifteen stories, other units of text that remained unpublished, and yet other textual segments for which hesitations even as to their actual status (narrative? aphoristic?) are indicated. Comparably, though in different ways, both the files from which Leiris composes his writing and the manuscripts which Aragon arranged in order to create Théâtre / Roman bear witness to such a conflict between the flow of the writing and the fixity of the work itself. From here we can undoubtedly advance in our consideration of the cohesiveness or fragmentation of an oeuvre—that is to say, the total production of a given author. And of course the concepts of intertextuality or historicity can also be evaluated through an author's working in that it helps us to elucidate the documents, observations and experiences which fed it.

Nevertheless this dialectic appears in each author in a different way. Writing subjects are always unique, and this uniqueness constantly causes entanglements in the generalising tendencies of literary criticism. Which means, undoubtedly, that we must diversify our approach in order to understand our subject better—not only in order to take the author into consideration once again, but also in order to consider his writing in a new way. This is the point of my second comment.

The analysis of meanings, forms and effects tends to treat the text as a system resulting from one single law, in such a way that none of its parts can be changed without affecting the whole. With this in mind, let us take a look at Paul Eluard's famous poem:

       

Sur mes cahiers d'ecolier	On my school notebooks
Sur mon pupitre et les arbres	On my desk and on the trees
Sur le sable sur la neige	On the sand and on the snow
J'écris ton nom	I write your name

It is easy to see how this text is triggered off and kept in movement, almost magnetised by one word which appears only as the incipit and excipit of the long poem of 21 stanzas:          

Et par le pouvoir d'un mot	And by the power of one word
Je recommence ma vie	I begin my life again
Je suis né pour te connaître	I was born to know you
Pour te nommer	To name you
Liberté.	Liberty.

What happens when we learn from the manuscript that this word is the result of a late correction (and that the author reveals to us the name of his beloved which was in its place)? Once again the answer will vary according to our approach to the text. For the historian, nothing changes: it was the theme of liberty which accounted for the poem's national success at the time of the Liberation, and no manuscript could change the phenomenon of collective enthusiasm which now belongs to literary history. For the Eluard specialist, interpretation of this work is clarified and verified. The substitution of the words attests to the relationship between the themes—and in this case it is the rapport between life and political commitment which was to determine the course of this poem. From a textual point of view things are more complex. It is true that the laws of structure are upheld: by changing a single word the entire poem is altered. Liberté is neither a variant nor even another version of the poem composed for Nusch—it is a different poem. But at the same time the perspective of genesis shows us that this first, distinct work was one of the possibilities of the text, though it was neither integrated nor subsumed in the second work. In other words, the writing is not simply consummated in the written work. Perhaps we should consider the text as a necessary possibility, as one manifestation of a process which is always virtually present in the background, a kind of third dimension of the written work. In this open (or half-open) space, the work is fatefully tossed between impetuous forward movements and calms of exhaustion, between stammerings and lacunae, from interruptions to unachievements that keep bringing us off course. The text is not annihilated by the weight of its possibilities but rather it stands out as an object which is far more complex than our former conceptions and far more aleatory than our modern ones. The effects of the upheaval caused by genetic criticism are only beginning to be felt. They will undoubtedly be consequential to research in the years to come.

Notes