Navigation – Plan du site

AccueilNuméros1.1Articles “hors thème”The Vicissitudes of Text

Articles “hors thème”

The Vicissitudes of Text

Jonathan CULLER

Texte intégral

  • 1  This entry for texte in the appendix sounds as though it were written by Julia Kristeva (at least (...)

1The concept of text, which has been central to literary studies, has undergone many mutations, as it has traveled from the work of classical philologists, for whom it was and is the object of a powerful disciplinary formation, to postmodern theorists of the text, for whom, the concept might be summed up by the title of a fine book by John Mowatt: Text: the Genealogy of an Antidisciplinary Object. Of course, the interesting thing about a travelling concept is not that it travels — travelers, tales about their travels can be quite boring and wholly unprofitable — but what it reveals through its travels. One very striking point in the itinerary of “text,” though — unparalleled, to my knowledge in the travels of other concepts, is that in Oswald Ducrot’s and Tzvetan Todorov’s Dictionnaire encyclopédique des sciences du langage of 1972, texte has two contradictory entries. In the main body of the dictionary it is defined as an organization of the utterance beyond the sentence: “le texte peut coincider avec une phrase comme avec un livre entier; il se définit par son autonomie et par sa clôture (même si, en un sens, certains textes ne sont pas ‘clos’).” (375) But then the appendix, which seeks to take account of recent developments that had challenged the idea of a “science du langage,” also contains an entry for texte — the only concept thus honored — under the heading “le texte comme productivité.” Citing the recent work of Jacques Derrida and Julia Kristeva among others, this entry tells us that “par opposition à tout usage communicatif et représentatif — donc re-productif — du langage, le texte y est défini essentiellement comme productivité.” Texte involves the “mise en place d’une procédure qui fasse jouer à plein son pouvoir génératif. Le texte a toujours fonctionné comme un champ de transgression au regard du système selon lequel sont organisées notre perception, notre grammaire, notre métaphysique, et jusqu’à notre science : système selon lequel un sujet, situé au centre d’un monde qui lui fournit comme un horizon, apprend à déchiffrer le sens supposé préalable, voire originaire, au regard de l’expérience qu’il en fait.” (443)1

2The double inscription of text, as a key concept in the reorganization of the human sciences around the linguistic model and in the almost immediate and gleeful critique of the possibility of a scientific model, is a measure of the importance, of the pivotal nature of this concept. Today one no longer sees as many books as one once did with text or textual in their title, and “textualism,” as it has come to be called, is often considered a sin, or at least an insult. I don’t think, though, that stopping talking about text and textuality is going to solve any of the problems of literary and cultural study. If we don’t engage with the problems clustered around the notion of text, we are going to be programmed by our own unexamined assumptions, for the problem of the text is always with us.

3Let me briefly sketch what I take to be the principal travels of text as a way of taking up major issues we approach through it. In philology the notion of the text — as in the idea of “the establishment of the text — is already dual. Textual critics or textual editors contrasted the object before them, a text, with the text in the putatively perfect state in which it left the author’s hands and to which the editor aspired to restore it. The text is thus both the pure origin, the manifestation of the final intention of the author, and an object marked by a history of material practices of transmission, which bring corruption. Textual scholarship was and indeed still is a process of reconstruction, based on methods which themselves are much open to debate, but it is only recently that the idea of the text as the corrupted material form in which the original intentions must be divined has been challenged. Modern textual scholars such as Jerome McGann have sought to reconceive the text as social act and to focus on the materiality of the social practices of transmission and publication, from inks and papers to book prices and editorial practices, and to think of these things as socially significant. (12) McGann thus tries to move away from the idea that the material practices involving texts are above all so many possible forms of corruption that may befall the final authorial intention, of which the text ought to be the materialization.

4Through these recent developments, the idea of text in textual scholarship may come to intersect, here and there, in its travels, with other modern ideas of text, but we have yet to see, I believe, how far this may go or how much of a rapprochement there will turn out to be. What I would like to stress here, then, is that in the idea of text in traditional textual scholarship we see a duality that will reappear in different forms. The notion of text is that of a material object but also of the very form of the work, in its original, ideal state. Thus the term text gestures towards matter, manifestation, indeed, the colloquial notion of text is this: the text is the writing that you see before you, “the text of this law,” for instance. But texts are of interest above all because of that which is or ought to be carried by or manifested in the material text, and thus seldom identified unreservedly with what appears on paper.

5In Anglo-American New Criticism, the notion of “the text itself” comes into its own in its most useful form — in an opposition. The text itself, the aesthetic object of literary study, as opposed to what is it said to mean or reflect or manifest: as opposed to history, biography. Students were enjoined to pay attention to the text itself, to cite evidence from the text itself, to set aside what is said about it, or about the author. They are urged to focus on what the text says or better, does, as opposed to what it is supposed to say. “The text itself” — the emphatic pronoun so often accompanies it — is a complicated positivity: words on the page. But despite the banning of the intentional fallacy (confusing what something means with what someone is supposed to have meant by it), and the rule that arguments about meaning are not settled by consulting the oracle (i.e., asking the author, directly or indirectly) (Wimsatt 18), the new critical notion of the text is not wholly divorced from authorial intention, which takes the form of a powerful posited teleology. The text is the words on the page, yes, but these words are presumed to be organized as a complex whole — otherwise we wouldn’t speak of the text but just of writing.

6Michael Riffaterre, who despite his theorization of intertextuality inherits and articulates this concept of the text, writes, “By textuality I mean the complex of formal and semantic features that characterize a self-sufficient, coherent, unified text, and legitimize its forms, however aberrant they may be, by removing any hint of the gratuitous.” (1) The textuality of a text (its essence as text) is the complex internal organization that sets it off from any context. If aesthetic objects are, in Kant’s phrase, purposive wholes without purpose, it is the artistic intention embodied in the text that warrants our expecting that the parts will be related to each other, that obscurities will have their reasons, and that everything will contribute to the effect of the whole. Having said this, we have to stress that the work itself reveals that artistic intention in ways that no information about the writer and his or her plans or intentions can, so that the text itself, though subtended by an artistic purpose, is separate from any other kinds of information, which can all be regarded as ancillary and set aside by oppositions.

7It was in the inherited context of this idea of the text — an autotelic whole governed by a powerful aesthetic teleology — that Derrida’s lapidary formula, “il n’y a pas de hors-texte” was interpreted to mean something like “everything outside the actual text or texts we are considering is irrelevant and doesn’t really exist,” whereas in fact, it means something like the opposite, that there is only text, since you can’t get outside of text. But I am getting ahead of myself and my text.

  • 2  For discussion, see Jonathan Culler, Structuralist Poetics (Routledge, 1976)
  • 3  Apparently Roger Shattuck, in a book that calls for a return to traditional criticism Candor and P (...)

8The structuralist revolution, which is the turning point in the interdisciplinary fortunes of the text and thus the source of the interest of the notion for us today, consisted first of all in considering human activities as so many languages, sign systems whose functioning needs to be explained — how is it that their products have the meanings they do. Just as the task of the linguist is to describe the system of rules, conventions and practices that enable human beings to produce and understand sentences, so it is the task of the structuralist or semiotician to reconstruct the other sign systems through which culture takes place.2 The result is that the products of numerous semiotic activities are considered in similar ways, as products of systems of signification and thus as texts. The first consequence is the equivalence, through the notion of sign systems and of text, of different cultural products, whether literary works, fashion captions, advertisements, films, or religious rituals: all can be considered as texts. In the structuralist/semiotic perspective, anything can be a text. The second result, I think, is that text can even come to be a neutral term: to refer to “Strindberg’s text” is for many of us today non-committal, in that it does not specify a genre (play, comedy, etc.) nor decide between the written form and its performance, or between the purely verbal and the visual. Text can seem both a technical term that carries a lot of theoretical weight and the most neutral term possible to designate a cultural production. Of course, this usage is not in fact wholly non-committal, because it implies a framework in which the idea of text is basic, but text is here opposed to some sort of generic specification or reference to a particular medium.3

9It is also opposed to the idea of objects that do not require interpretation or do not depend on conceptual frameworks or sign systems. This is a point that needs to be developed.

10The structuralist moment is doubtless the most important in the travels of text, both for interdisciplinarity and for the general projects of the humanities and social sciences. As Fredric Jameson writes in an important essay, “The Ideology of the Text,” first published in 1975 but greatly expanded for publication in The Ideologies of Theory in 1987,

11Textuality may be rapidly described as a methodological hypothesis whereby the objects of study of the human sciences are considered to constitute so many texts that we decipher and interpret, as distinguished from older views of these objects as realities or existents or substances that we in one way or another attempt to know. (18)

12Now the advantages of this concept are, of course, greatest in non-literary disciplines, where I think we could say that the concept of text does three things:

13 (1) it suggests that the items under consideration should not be taken as given and that one should consider how they come to be produced, isolated, presented to attention.

14 (2) One needs to think about meaning itself as a problem not a given, so that

15 (3) the analyst’s methods have to be considered, not just prior to the inquiry to decide what will be done, but in the process of considering the objects of study themselves.

16Jameson writes, for instance, “the notion of textuality, whatever fundamental objections may be made to it, has at least the advantage as a strategy of cutting across both epistemology and the subject/object antithesis in such a way as to neutralize both, and of focusing the attention of the analyst on her own position as reader and on her own mental operations as interpretation.” (18) Clifford Geertz, for instance, who did much to establish the idea of text in anthropology, explains “To treat a cockfight as a text is to bring out features of it... that treating it as a rite or a past-time, the two most obvious alternatives, would tend to obscure: its use of emotion for cognitive ends.” (449)

17The concept of text, as the product of a sign system that must be interrogated, has been extremely productive, it seems to me, first for the sort of interdisciplinary cultural study inaugurated by structuralism and carried on (once we are said to have entered a post-structuralist age) by a cultural studies which inclines to be grumpy about or ignorant of its theoretical antecedents.

18One region into which this concept of text traveled and enjoyed many picaresque adventures was that of film studies. I have neither the time nor the expertise to undertake to describe the vicissitudes of the idea of the filmic text or textual analysis, in the work of Christian Metz, Raymond Bellour, Marie-Claire Ropars, and others. John Mowatt provides an interesting overview in his chapter, “The Textual Analysis of Film,” in his book Text. (141-176) I am sure that there are other sources as well for debates about treating films as texts, what this brings to film studies, what are said to be the disadvantages. I will return to one example of the interdisciplinary functions of the notion of text at the end of this paper.

19The best-known essay about the concept of the text from this period is no doubt Roland Barthes’s “De l’œuvre au texte” of 1971. Although the ideas it draws on were certainly in the air none of the characterizations of the text were at all surprising — Barthes’s distinctive articulation of therm is highly idiosyncratic and not conducive, as Barthes himself might be the first to admit, to the advancement of methodological clarity or of an analytical prograrn in literary and cultural studies. Let me say first, though, that Barthes locates the rise of the concept of texte in the context of interdisciplinary encounters, as both a product of this situation and a response to interdisciplinarity, and he insists that we are not dealing with a radical mutation, but with a “glissement épistémologique plus qu’une véritable coupure.” (71) Barthes describes the notion of the text through an opposition between text and work on a number of different parameters. The persistence of this opposition gives his essay its clarity and force: the text is always being opposed to the work. But what Barthes seems to wish to resist above all is the idea that the concept of text can be one that replaces the concept of work, for him, it isn’t a matter of changing our view of objects previously treated as works and conceiving them instead in a new way, tempting though this idea might be. It is as though the text is a new thing, previously unglimpsed. What constitutes the text, for instance, is “sa force de subversion à l’égard des classements anciens.” Text is “ce qui se porte à la limite des règles de l’énonciation, (la rationalité, la lisibilité, etc.)” (73) It is irreducibly plural, a practice or play of the signifier generating the infinite deferral of the signified, it is not consumed by the reader but solicits collaboration on the part of the reader and functions not as an object of consumption yielding plaisir but as a practice of disruptive and self-disruptive jouissance.

20What is it that I object to here? Well, Barthes distinguishes, as his title tells us, œuvre from texte. But his claim is not that literary studies used to operate with one notion of its object, that of the work, which now it has reason to contest, so that it now thinks of the objects of its study as texts, with myriad consequences. On the contrary, he wants to insist that there are indeed œuvres, things that we continue rightly to describe and analyze as such, which are locatable, describable, and then there is texte. Although he speaks here and there in the essay of texts in the plural, he insists that texts are not countable, computable, locatable, and he generally speaks of le Texte, singular, with a capital T and maintains that there is “du Texte” to be located here and there in oeuvres. The texte en soi, pure text, is not something that can be found, analyzed, so that we can’t say writings in this group are oeuvres and these others are texts. And he insists from the outset, knowing that this is how we are inclined to interpret his essay, “Il ne faut pas se laisser aller à dire : l’œuvre est classique, le texte est d’avant-garde.” (72) We mustn’t because there can be “du Texte dans une œuvre très ancienne, et bien des produits de la littérature contemporaine ne sont en rien des textes.” But of course this argument reinforces the idea that the œuvre is something like “normal literature” and text is something avant-garde, just so radical that it can’t be pinned down.

21In an essay of 1972 entitled “Jeunes chercheurs”, Barthes writes, “quand nous disons le Texte, ce n’est pas pour le diviniser, en faire le déité d’une nouvelle mystique, c’est pour dénoter une masse, un champ, obligeant à une expression partitive, et non numérative: tout ce qu’on peut dire d’une œuvre, c’est qu’il y a en elle du Texte.” (107) But this usage certainly does make Texte an honorific concept, if not quite a God. And indeed, my principal objection to Barthes’s formulations in “De l’œuvre au texte” is that while remaining within a logic of opposition, they work to generate a mystique of the text: it is something so radical, disruptive, indeterminate, that it is not even an object but a practice or process, at best identifiable in certain moments. If you can show that there is “du Texte” in works of the past, you have shown that they are radical, exciting, worthy of attention. Barthes describes accurately, I think, what was actually happening. Jeunes chercheurs, as he indeed notes, see it as their task to “repérer ce qu’il peut y avoir de Texte dans Diderot, dans Chateaubriand.” It is a matter of finding “ce qui, dans l’œuvre ancienne, est Littérature et ce qui est Texte”. (108) Barthes is always up front and undefensive, to the point of undermining his own concepts and projects and those of his students. His description does indeed evoke a lot of work done in this domain — finding what one might call disruptive, indeterminate, or perhaps “postmodern” moments or elements in works of the past. My quarrel is not with this sort of critical writing itself, of which I have done my share, and which may, like any other sort of critical writing, be surprising or predictable, dubious or convincing. My quarrel is rather, with the theoretical framework.

22Is this a good way to proceed?

23In general, there seem to me two defensible perspectives here: the first would distinguish between two different ideas of literary and cultural objects, that of the work and that of the text, two different ways of conceiving them. First, one might say: people used to treat Madame Bovary or “Sarrasine” as an oeuvre, product of an authorial intention, with a meaning that had to be sought, and aesthetic unity to be valued, etc. But we now see that such things can and perhaps should be treated as texts, which means as products of sign systems and intertextuality, instances of indeterminate functioning of language, part of historical processes of production and reception. Thus work and text would be two different concepts of the object of study. It is not that some writings would be one and some the other, though obviously jeune chercheurs would find it more interesting to argue that writings celebrated as œuvres should really be treated as texts than to continue to treat them as œuvres. And note that within this perspective the idea of text could be charged with all the radical potential that Barthes wants to give it. The claim would be that we should stop reading work in relation to the idea of l’œuvre and conceive of thern according to this impossible model of le texte.

24Alternatively, œuvre and texte could be two different classes of objects (roughly the traditional and the avant-garde). Barthes rejects this conception more vigorously than the first sort, perhaps because this is what it is easiest to understand him to be saying: avant-garde products are radical, disruptive, indeterminate, good, and thus merit the appellation texte, whereas the old one’s don’t, though some of them have some textual good in them. When Barthes insists that lots of contemporary writing are in fact nothing but œuvres, not texte, it certainly strengthens the idea that this is above all an honorific distinction.

25The problem here is one which runs through S/Z with its analogous distinction between le lisible and le scriptible, which Barthes won’t let us take either as two types of writing or as two ways of thinking of writing. But in S/Z, I am inclined to say, this constitutes an interesting paradox in the methodological framework of his engagement with Balzac’s novella, one which enriches the book and illuminates some of the paradoxes of critical procedure, as when the supposed refusal of interpretation becomes an interpretation. But here, in a brief expository essay that does nothing other than propose to illuminate a methodological framework, it becomes an unavoidable problem, not to say methodological incoherence. While opposing work and text, Barthes refuses to let text and work be concepts that operate at the same level or in the same way.

26One consequence of this is that while Barthes’s account of the distinctions helps students find du Texte in older works, it doesn’t help much for dealing with avant-garde works, which always fall short of the radical ideal, and which are not much illuminated by accounts showing them to fall short. The asymmetry helps animate the classic texts, in which one finds du moderne, but doesn’t help much with the texts on which Barthes’ scheme claims to set highest value.

27In short, although all of his determinations of the text are familiar parts of the idea of the text, Barthes’s insistence that this is not just a methodological shift but that there are indeed works (which sometimes contain du texte), makes the idea of the text seem something of a fetish, an ideal object so incredibly radical and disruptive etc. that no actual discourse is adequate to the idea (while of course works really do exist). The concept of text that he develops might function well in a univocal model, where we now read works as texts, but he insists on maintaining the binary model where there are still œuvres. Barthes’s essay thus seems to invalidate the idea of text that it claims to advance and that it presents so vividly.

28The way to take Barthes’s essay, it seems to me, is to try to reduce the asymmetry of the concepts of work and text and to note that each is a model, a conception of how things might in principle function. It is important to remember that the idea of the work as the realization of an authorial intention and as an organic whole is also something of a impossible ideal, and thus a goal of analysis, so that it may function symmetrically to that Barthian ideal of the text which is systematically set against it.

29Barthes’s essay, as I say, is unfortunately the most celebrated and had a significant effect on the idea of the text. Others had tried to define the idea more perspicuously. Julia Kristeva, for instance, in a number of places put the methodological concept of text in opposition to notions of representation and communication — they belong to different methodological networks. But she avoided any idea that old-fashioned writings were not texts (she wrote about medieval narrative, for instance) while other writings were. Kristeva emphasized the idea of treating language in general through the concept of text, which stresses that discourse is not simply expression by a subject but produces the subject, and that it is fundamentally intertextual, related to other discourses. She writes, for instance,

Si on admet que toute pratique signifiante est un champ de transpositions de divers systèmes signifiants (une intertextualité), on comprend que son ‘lieu’ d’énonciation et son ‘objet’ dénoté ne sont jamais uniques, pleins et identiques à eux-mêmes, mais toujours pluriels, éclatés. (60)

30The text discloses the mechanisms through which language produces its effects and dislodges speakers from strictly representational relations to language, thus threatening the identities of speakers by treating them as constructed, not given.

31Jacques Derrida, in remarks such as the famous il n’y a pas de hors-texte, which I cited earlier, articulates a conception which might be used to rescue the notion of text from the characterizations Barthes had given it, to make it more centrally normative of language in general. In the afterword to Limited Inc, a text called “Vers une éthique de la discussion,” Derrida writes:

le concept de texte que je propose ne se limite ni à la graphie, ni au livre, ni même au discours, encore moins à la sphère sémantique, représentative, symbolique, idéelle ou idéologique. Ce que j’appelle “texte” implique toutes les structures dites “réelles”, “économiques”, “historiques”, socio-institutionnelles, bref, tous les référents possibles. Autre manière de rappeler une fois encore qu’il n’y a pas de hors-texte. Cela ne veut pas dire que tous les référents sont suspendus, niés ou enfermés dans un livre, comme on feint ou comme on a souvent la naïveté de le croire et de m’en accuser. Mais cela veut dire que tout référent, toute réalité a la structure d’une trace différentielle, et qu’on ne peut se rapporter à ce réel que dans une expérience interprétative, Celle-ci ne donne ou ne prend sens que dans ce mouvement de renvoi différantiel. That’s all. (273)

32The argument that the referent has the structure of a differential trace, that it is textualized and not something of a different nature, has been conducted in many places. My favorite comes in Derrida’s discussion of Rousseau’s Confessions, where Rousseau characterizes writing and signs in general as “suppléments” to the thing itself but in fact shows that his experience is, in Derrida’s words, “celle d’un enchaînement infini, multipliant inéluctablement les médiations supplémentaires qui produisent le sens de cela même qu’elles diffèrent: le mirage de la chose même, de la présence immédiate, de la perception originaire. L’immédiateté est dérivée. Tout commence par l’intermédiaire...” (De la grammatologie 226)

33The more these texts want to tell us of the importance of the presence of the thing itself, the more they show the necessity of intermediaries. These signs or supplements are in fact responsible for the sense that there is something to grasp. What we learn from these texts is that the copies create the idea of the original, and that the original is always deferred — never to be grasped. Experience is always mediated by signs and the “original” is produced as an effect of signs, of supplements.

34For Derrida, Rousseau’s texts, like many others, propose that instead of thinking of life as something to which signs and texts are added to represent it, we should conceive of life itself as suffused with signs, made what it is by processes of signification. Writings may claim that reality is prior to signification, but in fact they show that “Il n’y a pas de hors-texte” — “There is no outside-of­text”: when you think you are getting outside signs and text, to “reality itself,” what you find is more text, more signs, chains of supplements. Derrida writes,

Ce que nous avons tenté de démontrer en suivant le fil conducteur du “supplément dangereux”, c’est que dans ce qu’on appelle la vie réelle de ces existences “en chair et en os” (...) il n’y a jamais eu que de l’écriture, il n’y a jamais eu que des suppléments, des significations substitutives qui n’ont pu surgir que dans une chaîne de renvois différentielle, le “réel” ne survenant, ne s’ajoutant qu’en prenant sens à partir d’une trace et d’un appel de supplément, etc. Et ainsi à l’infini, car nous avons lu, dans le texte, que le présent absolu, la

nature, ce que nomment les mots de “mère réelle”, etc. se sont toujours déjà dérobés, n’ont jamais existé ; que ce qui ouvre le sens et le langage, c’est cette écriture comme disparition de la présence naturelle. (228)

35This does not mean that there is no difference between the presence of someone and her absence or between a “real” event and a fictional one. It’s that her presence turns out to be a particular kind of absence, still requiring mediations and supplements.

36It is this interweaving of signs and supplements, of language with what we call real life, that provides the most elementary rationale for the notion of text: if language or logos were a separate stratum that were founded on something else, then it might indeed be conceivable, as traditional accounts wish, to set aside signs or discourse so as to accede directly to that other thing, whether it be thought, action, or reality itself. But this is what is not possible. One level or stratum interacts with the other and cannot be separated except provisionally or artificially. This is why the notion of the text as woven of different strands is much better than the geological metaphor of strata or levels: the text is not a series of layers but the interweaving of language with other threads of experience. Derrida writes in one of his early texts on Husserl, “La forme et le vouloir-dire”, where he notes Husserl’s own distrust of the metaphor of layers or strata, with which Husserl is nevertheless inextricably implicated:

L’entrelacement du langage, de ce qui dans le langage est purement langage, et des autres fils de l’expérience, constitue un tissu. (...) Si la couche du logos était simplement fondée, on pourrait la prélever et laisser apparaître sous elle la couche sous-jacente des actes et des contenus non-expressifs. Mais puisque cette super-structure agit en retour, de manière essentielle et décisive, sur l’Unterschicht [la couche sous-jacente], on est bien obligé, dès l’entrée de la description, d’associer à la métaphore géologique une métaphore proprement textuelle : (car tissu veut dire texte. Verbeben ici veut dire texere.) Le discursif se rapporte au non-discursif, la “couche” linguistique s’entremêle à la “couche” prélinguistique selon le système réglé d’une sorte de texte. (282)

37To say that “la texture du texte est en un mot irréductible”, is to say that “la description ne fait pas apparaître un sol absolument et simplement fondateur de la signfication en générale”, that “un sol intuitif et perceptif ne fonde pas le discours dans la présence originairement donnée de la chose même”. (283)

38Here, or in Derrida’s concept of “the general text,” a textuality that underlies and makes possible particular texts, the focus is not interdisciplinary inquiry, but the idea of a general textuality certainly encourages thinking of all cultural objects as texts and hence a measure of interdisciplinarity. Sometimes people claim that this idea of text is an instance of literary studies seeking to extend its empire by seeing everything as text, and that therefore what we have is not so much interdisciplinarity as the imperialism of literary studies. But the extension of the idea of text has transformed literary studies as much as, if not more than, any of the other disciplines, so it its not the case, for instance, that other disciplines have been assimilated to or had imposed on them a traditional idea of the literary. (That would, in any event, involve the idea of l’œuvre rather than texte.) Moreover, insofar as the idea of the text has challenged positivistic models in anthropology and sociology or straightforward representational models in history or art history, that is not so much the imperialism of the literary as a good in itself, a major intellectual advance.

39I want to conclude by pursuing the interdisciplinary potential of the idea of the text in a surprising place (at least surprising to me): Paul de Man’s Allegories of Reading. The second half of this book is devoted to Rousseau, and as we near the end, de Man undertakes a complex reading of Le Contrat social. There he describes what Rousseau calls the double rapport in which an individual is engaged: “chaque individu (...) se trouve engagé sous un double rapport, à savoir comme membre du souverain envers les particuliers, et comme membre de l’État envers le souverain” ; (290) on the one hand, as one helping to constitute general will and thus sovereign authority of the state and on the other as a member of the community subject to the sovereign authority of the state. “En effet, chaque individu peut, comme homme, avoir une volonté particulière contraire ou dissemblable à la volonté générale qu’il a comme Citoyen.” (291) As part of the general will, the citizen is alienated from the particular desires and interests that animate him as an individual, and this double relationship extends throughout political life: a piece of land, for instance, may be considered as part of the state or as private property. The same estrangement that separates the citizen as contributor to the sovereign authority of the state from the citizen as indivisual separates political rights and laws, on the one hand, from political action and history on the other.

40Working through Rousseau’s remarkably perspicuous accounts of these relations, de Man writes that “the structure of the entity with which we are concerned (be it as property, as national State, or any other political institution) is most clearly revealed when it is considered as the general form that subsumes all these particular versions, namely as legal text,” and after discussion of the problematical relation of the generality of law and its particular applications (a point Rousseau stresses), he writes,

we have moved closer and closer to the “definition” of text, the entity we are trying to circumscribe (...) The system of relationships that generates the text and that functions independently of its referential meaning is its grammar. To the extent that a text is grammatical, it is a logical code or machine (...) But just as no text is conceivable without grammar, no grammar is conceivable without the suspension of referential meaning (...) grammatical logic can function only if its referential consequences are disregarded. (320, 322)

41A law must be general, without reference particular to individuals — only to the empty chacun. But he continues, “no law is a law unless its also applies to particular individuals.” And Rousseau writes that the general will functions only because

  • 4  All these quotations are from the first version of Du Contrat social but in these passages the sec (...)

 il n’y a personne qui ne s’approprie en secret ce mot chacun et qui ne songe qu’à lui-même en votant pour tous. Ce qui prouve que l’égalité de droit et la notion de justice qui en découle dérive de la préférence que chacun se donne et par conséquent de la nature de l’homme. (306)4.

42 Rousseau stresses that “la volonté générale, pour être vraiment telle, doit être générale dans son objet ainsi que dans son essence et qu’elle perd sa rectitude naturelle sitôt qu’elle tombe sur un sujet individu et déterminé,” yet is it this mechanism of secretly referring chacun to oneself that allows the general will to function. “There can be no text without grammar,” writes de Man, “the logic of grammar generates texts only in the absence of referential meaning, but every text generates a referent that subverts the grammatical principle to which it owed its constitution.” (269) It is this contradictory or duplicitous structure that relates text and law: in Rousseau’s passage, it is the duplicitous production of a referent (“en secret”) that bridges the gap between the elaboration of the law and its application. De Man then links the general model of figurative language, where there is a gap between grammatical and referential meaning to Rousseau’s account of the state. He writes,

In the description of the structure of political society, the ‘definition’ of a text as the contradictory interference of the grammatical with the figural field emerges in its most systematic form (...) We call text any entity that can be considered from such a double perspective: as a generative, open-ended non-referential system and as a figural system closed off by a transcendental signification that subverts the grammatical code to which the text owes its existence. (270)

43Further, he writes,

the definition between the figural and grammatical language is duplicated in the differentiation between the State as a defined entity (état) and the State as a principle of action (souverain), or, in linguistic terms, between the constative and performative function of language. A text is defined by the necessity of considering a statement at the same time as performative and constative.

44In de Man, thus, we find that the text has become the name for and the model of a pervasive structural relationship to which we have in recent memory given many names, both in thinking about language (performative/constative, langue /parole) and in thinking about the most basic matters of action, identity, and institutions: the relationship between structure and event for example. The concept of text thus presents the possibility of functioning as the basis for wide-ranging interdiscisplinary study — a possibility tempered by the fact that the problematical structure may seem easier to grasp in individual cases, as when I spoke about the double rapport of the individual to the state, than through the concept of text. But of course this could change if this concept of the text were to become more widely disseminated and explicated.

45The notion of text seems to me to serve above all, as the varied examples I have adduced suggests, to foreground the complexity of the semiotic productions that we undertake to study. Preserved in most of these travels is stress on the etymological connotations of wovenness — multiple strands which can be pursued and whose relation needs to be considered — whether these strands are considered to be things like the codes that Barthes describes or the inextricability of language with the reality that it performatively helps to structure. The idea of text also foregrounds, especially in the case of non literary materials, the fact that the objects we are considering are themselves products, whether of specific discursive and artistic activities such as those of the writer, painter, or filmmaker, or of scripted behavior and conceptual frameworks, as in the identification of social facts for consideration as texts.

46But the fact that the notion of text can function both as a highly-charged impossible ideal, as in Barthes’s much-cited essay, as a relatively neutral, uncommitted description, a way of naming objects of inquiry prior to saying anything about them, indicates that the concept of text is not going to solve our problems for us by getting anything straight once and for all, but that is it itself a text, ceaselessly made and unmade in our uses of it and directing our attention above all, perhaps, to the constitutive iterability of language, which works by repeating without repeating, so that we must keep paraphrasing and theorizing in our pursuit of the text or, more productively, of particular texts or textual structures.

Haut de page


Barthes, Roland. “De l’œuvre au texte.” Le Bruissement de la langue: Essais critiques IV. 1971. Paris: Seuil, 1984. 71-80.

—. “Jeunes chercheurs.” Le Bruissement de la langue: Essais critiques IV. 1971. Paris: Seuil, 1984. 103 -110.

—. S/Z. Paris: Seuil, 1970.

Culler, Jonathan. Structuralist Poetics: Structuralism, Linguistics, and the Study of Literature. London: Routledge, 1976.

De Man, Paul. Allegories of Reading: Figural Language in Rousseau, Nietzsche, Rilke, and Proust. New Haven: Yale University Press, 1979. French translation: Allégories de la lecture: le langage figuré chez Rousseau, Nietzsche, Rilke, et Proust, Paris, Galilée, 1989.

Derrida, Jacques. De la grammatologie. Paris: Minuit, 1967.

—. Speech and Phenomena. 1967. Trans. David Allison. Evanston: Northwestern

University Press, 1973.

—. “La forme et le vouloir dire,” Revue internationale de philosophie 81 (Sept. 1967): 282.

—. “Vers une éthique de la discussion.” Limited Inc. Paris: Galilée, 1990. 199-285.

Ducrot, Oswald, and Tzvetan Todorov. Dictionnaire encyclopédique des sciences du langage. Paris: Seuil, 1972.

Geertz, Clifford. The Interpretation of Cultures. Basic Books: New York, 1973.

Jameson, Fredric. “The Ideology of the Text.” The Ideologies of Theory Minneapolis: University of Minnesota Press, 1987. 17-7 1.

Kristeva, Julia. “La productivité dite texte.” Sémiotikê: Recherches pour une sémanalyse. Paris: Seuil.

—. La Révolution du langage poétique. 1969. Paris: Seuil, 1974.

McGann, Jerome. The Textual Condition. Princeton: Princeton University Press, 1991.

Mowatt, John. Text: the Genealogy of an Antidisciplinary Object. Durham: Duke University Press, 1992.

Riffaterre, Michael. “Textuality: W.H. Auden’s ‘Musée des Beaux Arts,’” in Mary Ann Caws, ed. Textual Analysis. New York: MLA, 1986. 1-13.

Rousseau, Jean-Jacques. 1760. Du Contrat social, ou Essai sur la forme de la république (first version) Œuvres complètes. Ed. Bernard Gagnebin & Marcel Raymond. Pléiade. Paris: Gallimard, 1964. Vol. 3, 281-346. For later version, Du Contrat social, ou Principes du droit politique (1762), see 349-471.

Wimsatt, W. K., and Monroe Beardsley. “The Intentional Fallacy.” The Verbal Icon. Lexington: University of Kentucky Press, 1954.

Haut de page


1  This entry for texte in the appendix sounds as though it were written by Julia Kristeva (at least it presents idiosyncratic concepts of hers as normative) and for anyone interested in this topic is well worth reading.

2  For discussion, see Jonathan Culler, Structuralist Poetics (Routledge, 1976)

3  Apparently Roger Shattuck, in a book that calls for a return to traditional criticism Candor and Perversion (New York: Norton, 2001) offers “Nineteen Theses on Literature,” of which number 8 reads: “Let us eschew the freestanding word text. Let us take advantage of terms like novel, poem, play, essay.” (Review by John Sutherland, TLS, 28 April 2000: 26)

4  All these quotations are from the first version of Du Contrat social but in these passages the second version is identical (362, 363, and 373 respectively), except for the suppression of “en secret” from this last passage.

Haut de page

Pour citer cet article

Référence électronique

Jonathan CULLER, « The Vicissitudes of Text »E-rea [En ligne], 1.1 | 2003, mis en ligne le 19 juin 2003, consulté le 08 décembre 2023. URL : ; DOI :

Haut de page


Jonathan CULLER

University of Cornell
Professor of English and Comparative Literature
Jonathan Culler is the author of Structuralist Poetics: Structuralism, Linguistics, and the Study of Literature (1975), Saussure (1976), The Pursuit of Signs: Semiotics, Literature, Deconstruction (1981), On Deconstruction: Theory and Criticism after Structuralism (1982), Barthes (1983), Flaubert: The Uses of Uncertainty (1985), Framing the Sign: Criticism and Its Institutions (1992), Literary Theory: A Very Short Introduction (1997).

Haut de page

Droits d’auteur


Le texte seul est utilisable sous licence CC BY-NC-ND 4.0. Les autres éléments (illustrations, fichiers annexes importés) sont « Tous droits réservés », sauf mention contraire.

Haut de page
Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search