Structuralism, in various disguises, has been stalking literary studies in England and America for some time. But it usually stalks as a movement, and not in the name of any one of its practitioners. One of the interesting things about Jacques Derrida, who is generally regarded as a post-structuralist, is that he has found a following, at least in America, as an individual figure. It’s not easy to measure the extent of his influence. Is it confined to a handful of English and French departments, or is it more widespread? It does seem fairly considerable in any case, and it raises interesting questions, not least because Derrida is not a literary critic but a philosopher. He teaches at the Ecole Normale Supérieure in Paris and also, for a brief spell each year, at Yale.

Almost all of Derrida’s work records and re-creates acts of reading, not explications de textes so much as interrogations of texts: texts by Plato, Hegel, Heidegger, Husserl, Rousseau, Saussure, Lévi-Strauss, Artaud, Bataille, Austin, Genet, Freud, Lacan, and others. “I write and read at the same time,” Derrida said in an interview: “slowly, taking pleasure in writing lengthy prefaces to each term.” Derrida’s name for his method of reading, when it tackles the long conspiracy which Derrida sees in Western thought, is deconstruction. He doesn’t deconstruct his texts, he asks them to help him in the deconstruction of the philosophy in which they are implicated.

Deconstruction has one or two rather lurid strategies, which I’ll return to, but its principal feature, as Derrida practices it, is a patient and intelligent suspicion, which falls less on the meanings and definitions of words than on their associations and affiliations, notably their complicity in the vast metaphysical plot running from Plato to Hegel, or, taking an even wider arc, from the pre-Socratics to Heidegger. The plot is a doctrine of presence, a faith holding that immediacy is value and indirection is evil, and Derrida uncovers it not only in all the predictable places (notions of an immanent God, self-consciousness as the guarantee of identity) but even in the unlikely region of linguistic philosophy, whose ideas of “context” and “ordinary language” conceal “behind a certain amount of confusion, very specific philosophical presuppositions.” There are obvious virtues in a plot you can find everywhere, especially if you enjoy your suspicions, but Derrida’s thinking does at times remind me of those Hollywood movies which insisted on confronting nothing less than the whole human condition.

Derrida attacks the great metaphysical conspiracy through what he calls logocentrism: the belief in the primacy of speech over writing, the claim that speech is in some way a “natural” or direct form of language. The doctrine of presence regularly resorts to a praise of speech or a vocabulary of the voice, and to a scorn, often quite virulent, of writing. In the beginning was the word, and the word was a logos, and logos for the Greeks was a spoken word. Even a science of writing, Derrida discreetly notes, has to call itself a grammatology. Writing in the West has been “debased, lateralized, repressed, displaced,” but like all repressed material, Derrida argues, continues to exercise “a permanent and obsessive pressure from the place where it remains held in check.” Even writers against writing know that writing is what they are doing, and this seems only to foment their fury. This is the situation which Of Grammatology sets out to explore. The bulk of the book consists of a detailed interpretation of Rousseau’s Essai sur l’origine des langues, with prior glances at Saussure and Lévi-Strauss, and a certain amount of general theory of writing.

In the Phaedrus Plato has Socrates (who didn’t write) call writing a drug, a means of ruining memory by offering to help it out, and Lévi-Strauss, in Tristes Tropiques, says “writing and perfidy” enter the unlettered jungles of Brazil hand in hand. Rousseau thinks that politically the only free people are those who can talk to each other directly (“Now, I say that any language with which one cannot make oneself heard by the assembled people, is a servile language; it is impossible that a people remain free, and speak that language”), and sees our “northern” (that is, French, German, and English) dependence on writing as a sure sign of our decline, our falling away from nature: “Our languages are more effective when written than when spoken, and people read us with more pleasure than they listen to us.” Saussure, who himself says that language is a system of signs “comparable to a system of writing,” insists that writing exists “for the sole purpose” of representing spoken language, and gets very angry at what he sees as writing’s “usurpation” of the role of speech:


People attach even more importance to the written image of a vocal sign than to the sign itself. A similar mistake would be in thinking that more can be learned about someone by looking at his photograph than by viewing him directly.

Saussure goes on to suggest that writing is not a guise for language but a disguise, a travesty (pas un vêtement, mais un travestissement), speaks of a “tyranny of writing” and finds our occasional practice of allowing spelling to dictate our pronunciation “really pathological.”

These are Derrida’s prime instances of logocentrism, and several things need to be said. First, Plato was not simply attacking writing through Socrates, as Derrida himself demonstrates in a brilliant essay in another book. He was playing off speech against writing in a very complicated way. Socrates himself, after his onslaught on written texts, which don’t answer when you talk to them (“they go on telling you the same thing, over and over”) and always need their “father” when they’re in trouble, speaks of the highest truths as “inscribed in the soul,” and Plato was not likely to be making a slip here—or even, as Derrida suggests in Of Grammatology, moving over to another sense of writing. Second, one might wish to argue, as Paul de Man does in Blindness and Insight, that Rousseau is not taken in by the fictions he so enthusiastically throws up, and that he is not therefore, even partially, a victim of the logocentrism he seems to proclaim.

More important, it seems possible that Derrida has got the whole argument upside down. Writing is not repressed in the West, but on the contrary incessantly celebrated, not least in the work of Derrida himself, and the examples of logocentrism he uses—Plato, Rousseau, Saussure, Lévi-Strauss—make clear, by their sparseness and dispersion, that they don’t represent the main stream of metaphysics, but are rather sporadic protests against a clearly felt supremacy of writing. Where Derrida sees repression, Saussure saw an alarming prestige, and there is no reason for us to think Saussure simply wrong.

None of this really damages Derrida’s case, though, at least in its more insidious form, because even upside down the ingredients of the question remain the same. I leave the larger metaphysical plot to the judgment of philosophers, but it is clear that in the quarrel between speech and writing, the irony of Plato and Rousseau, however complicated or undeceived, surrounds, but in no way cancels, the passion of their preference for speech; and that this passion is the same in Saussure and Lévi-Strauss; and the same, I will add, in most of us. I trust I am not alone in still feeling a sort of rage at the mere suggestion that speech is not more natural or more direct than writing. Even Joyce’s Ulysses, that monument of written words, implies in its very splendor that there is an even more splendid, verbal, Irish life beyond the book.

It is true that literature, in practice, is as de Man says “free from the fallacy of unmediated expression”; that it can’t be, as literature, the unmediated expression of anything. But this, as Derrida would be quick to say, is because there is no such thing as unmediated expression, because the phrase is a contradiction in terms, and literature, in this as in other respects, constantly dreams of what it can’t have. Literature is crowded with yearnings for unmediated expression and Tristram Shandy is the only literary text I can think of that is entirely comfortable with its condition as book, as not-speech, that thoroughly enjoys the chain of indirection which consists of writing a work which is taken to a printer who delivers it to a bookshop whence it goes to readers whose separation from you is a large part of the fun.

And of course our rage, if that is what it is, is precisely Derrida’s point. I would even suggest that if we happen to feel that writing is superior to speech, then Of Grammatology is not really for us, it will only flatter our prejudices. For that is not what Derrida wishes to argue. He wishes to argue, not that writing is not secondary, but that speech is not primary, that we are fooling ourselves if we think that speech is not, already and always, hopelessly indirect, enmeshed in a language which is much older and heavier and more intricate than we care to believe. Derrida wishes to show, not that writing is innocent while speech is guilty, but that writing is not something which “befalls” an innocent language:

It is not a question of rehabilitating writing in the narrow sense, nor of reversing the order of dependence when it is evident…. I would wish rather to suggest that the alleged derivativeness of writing, however real and massive, was possible only on one condition: that the “original,” “natural,” etc. language had never existed, never been intact and untouched by writing, that it had itself always been a writing.

The point, Derrida suggests, is to stop thinking about speech and writing ethically, and to see writing, as he rather too eagerly puts it, as “beyond good and evil.”


The more lurid strategies of deconstruction I mentioned earlier are what Derrida calls inversion and displacement. He wants to be exorbitant, he says in Of Grammatology, making play with ex and orbis and orbita, but he also wants to attack the orb from within: to throw bombs at philosophy’s building, but also to mine the foundation. And he does this by an elegant double maneuver. Since a structural opposition—life/death, speech/writing, or whatever—is never perfectly balanced but always values one of its elements over the other, an inversion of the values will not simply turn things round but will displace the whole system. And this is what Derrida is trying to do in Of Grammatology. If the argument looks upside down, that is because it is upside down, although not in the sense I first identified.

We think writing is a means of representing speech, but Derrida is telling us that speech is a form of unrecorded script, because it has all the essential characteristics of writing. It consists of signs which are repeatable, quotable, and above all, repeatable and quotable in the absence of an original speaker, and even in the absence of an original listener. Speech, like writing, leaves traces, it writes in the memory, and if it didn’t, not only would it not be speech, it would not be language at all. Writing, despised and rejected by passionate linguistics, is not the enemy of the spoken word but a picture of everything which makes the spoken word possible.

It is important to notice what is exorbitant here as well as what is simply true. Derrida borrows his logic from Nietzsche (who else, except possibly Freud, contemplating the harmony of a Greek scene, would take that harmony as an index of all the madness it was designed to contain?), and he is not really interested in what language does, except as an accomplice in the metaphysical conspiracy. He is interested in the conditions of language’s existence, and this is made very clear in Derrida’s intricate essay on J.L. Austin, where the same sort of upside-down logic is at work.

Austin in How To Do Things With Words wants to exclude from his present consideration a whole range of what he calls “infelicities,” things that can go wrong with speech acts:

…a performative utterance will, for example, be in a peculiar way hollow or void if said by an actor on the stage, or if introduced in a poem, or spoken in soliloquy.

Austin speaks of these special circumstances as falling under the “doctrine of the etiolations of language,” and literary folk have always blinked a bit at the thought of poems and plays as etiolations. Derrida pounces to the rescue and suggests that the so-called etiolations may be the foundations of language:

In other words, does the general risk admitted by Austin surround language like a sort of moat, an external place of perdition into which speech could always not emerge, which it could avoid by staying at home…? Or on the contrary, is this risk its internal and positive condition of possibility? is this outside its inside? the very strength and the law of its coming into being?

The attraction of this logic is that it holds things together, allows us to think of harmony and madness, speech and writing, ordinary language and King Lear, in a single thought, in Yeats’s phrase. It allows us to deal in relations rather than exclusions, and this is often a real advantage, even a liberation. We don’t have to forget comedy, for example, in order to talk about tragedy: on the contrary we have to remember it. The trouble with this logic in Derrida’s case is that his preoccupation with the “general space” of possibility often leaves all the particular spaces of practice to their own devices. Even when we know that speech is writing, we still have to talk, and Derrida will not help us here. His exorbitant readings, in Of Grammatology and elsewhere, are frequently brilliant, full of local insights. But his flight into generality is so persistent, his yearning to stand aside from philosophy while still doing it is so strong, that his whole enterprise becomes marked with a touch of dilettantism. Derrida’s interminable and sometimes complacent deconstructions of a metaphysical edifice that will never, quite, fall down begin to look a little like shadow-boxing.

Another way of getting at this aspect of Derrida would be to say that he can stand aside from philosophy only by enclosing himself in a text, in a special state of language. A text is a piece of writing perceived as writing, and it always has a certain obliquity for Derrida. The rules of its game, as he says, are not secret but they are not immediately perceptible either. The text is language lifted out of the illusion of immediacy. It becomes a form of utopia, an escape from the rawness of history and biography, a zone where contradictions can be seen to be cancelled, where opposites are united in what Derrida calls the coherence of desire. Very like a dream, and this is the comparison Derrida uses toward the end of Of Grammatology. The text, for example, is the place where Rousseau can be both for and against writing, where the spoken word can become not what it usually is in reality but what it might be in another dispensation. There is truth as well as polemic in Michel Foucault’s harsh comment that Derrida’s method is a recipe for reducing discursive practices to textual traces, and for eliding events so that only material for reading remains. Foucault claims that Derrida has invented a “neat little pedagogy,” une petite pédagogie historiquement bien déterminée, which allows the teacher simultaneously to tell his students that there is nothing outside the text and to keep repeating the text itself indefinitely.

Yet even here Derrida can be rescued from Derrida. Certainly he says there is nothing outside the text: il n’y a pas de hors-texte. And again: il n’y a rein hors du texte. Indeed he says this is the “axial proposition” of Of Grammatology, and it is this Derrida whom Paul de Man and J. Hillis Miller seem to be following when they speak of “the fallacy of reference” (de Man) and of deconstruction as an activity which “undermines the referential status of the language being deconstructed” (Hillis Miller).

But all this, to borrow a distinction which Derrida himself uses for Rousseau, is what Derrida declares, while what he describes is something else. What he describes, what a good deal of his work seems to rest on, is not a situation in which the text refers to nothing, but a situation in which there is nothing but “texts,” tissues of meaning awaiting interpretation. The text of Rousseau’s essay, for example, belongs to the “text” of Rousseau’s life, which reflects Rousseau’s nationality and language, which are marked by the eighteenth century, and so on: a chain of “texts.” “The writing subject [that is, the single, subjective mind behind a piece of writing] does not exist,” Derrida says in a lecture on Freud, “if by that we mean some sort of sovereign solitude of the writer.”

Of course there is something mischievous about calling “the age of Rousseau” a text; mischievous and modish, even if Derrida did more or less start the mode. But the word does insist on proliferating relations and doesn’t necessarily enclose us in a world of books or even, in the narrow sense, of texts. What Derrida is suggesting is not that language (or literature) doesn’t refer to reality, but that reality is so unstable a category that nothing can simply “refer” to it; that reality itself is a fabric of references, a web of signs which point to each other and not to a God or to a “real” reality behind the appearances. This view is obviously sound enough in its way, but it will seem disturbing, I think, only if you hanker for a theological certainty, however subtle or secular; if you long for an end to the play of signifiers, for a meaning which will finally and simply be the meaning.

I must add that in his writing Derrida doesn’t give the impression of being enormously anxious to confront texts in anything other than the perfectly ordinary sense of the word, and that even the larger view, inherited from Peirce and Saussure, the notion that life is literally infiltrated by signs, is a thin and rather mean perspective, once its therapeutic values have been recognized. It will cure us of simplicity about language and reality, but it will not nourish us beyond the cure; and even simplicity may be better than a complication we can’t escape from. We have all had experiences we would hate to see pictured as a text, as a cluster of signs; for which such a picture would not seem merely an option, but an insult, a horrible impoverishment. So that if Derrida’s work is, as Paul de Man says, “one of the places where the future possibility of literary criticism is being decided”—and I hope it is, since criticism cannot afford to ignore a reader as vigilant as Derrida—I should like to think that that possibility is also being decided in one or two other, rather different places; places where the pressures of life on literature are more acutely felt, and where the immersion of literature in the world is celebrated rather than half-denied.

In any case, as Harold Bloom remarks in a forthcoming book on Wallace Stevens, it’s not always easy to see what deconstruction means when applied to particular works of literature, and Geoffrey Hartman, in The Fate of Reading, wonders whether he may not himself be accused of “deconstructing without a license.” Clearly we can track down the metaphysical conspiracy in plays, poems, novels, and elsewhere, and this is being done very ably—Paul Bové has shown, for example, how much metaphysics there was in the New Criticism. But when Derrida, in Of Grammatology, demonstrates that Lévi-Strauss’s praise of the Brazilian Indian is marked by precisely the ethnocentrism it sets out to deplore, this seems to be more of a demystification than a deconstruction, and the emulation of Derrida could also lead this way: toward the unmasking of hidden contradictions wherever they are found.

It is true that Roland Barthes has already taught us how to do this—he called such contradictions myths—but Barthes is not a philosopher, and literary criticism at the moment seems to lean on philosophy much as it used to lean, and in some places still leans, on psychoanalysis. Derrida’s main appeal, perhaps, even more than his glorification of the text and his practice of reading as a form of mistrust, is his authoritative invitation to a dismantling of old gods and a deposing of old fathers. He says himself that his work resembles a negative theology, parfois à s’y méprendre. Deconstruction: the very word has an austere sound to it which makes it some sort of sign of our timid and disabused times.

Gayatri Chakravorty Spivak’s introduction to her translation of Of Grammatology—she has translated de la as “of” rather than “on,” she says, “against expert counsel,” because it suggests “a piece of” as well as “about”—is slightly pompous, infected by Derrida’s own rather solemn playfulness, and it indicates how easily Derrida’s practice of suspicion can slip into fussiness and self-consciousness. But her situating of Derrida among his precursors—Nietzsche, Freud, Heidegger, Husserl—and contemporaries—Lacan, Foucault, and the elusive animal known as structuralism—is very lucid and extremely useful.

The translation itself has some surprising errors—opiniâtrement, for example, is given as “opinionatedly” instead of “stubbornly,” and beaucoup seront tentés is rendered as “much will be attempted” rather than “many will be tempted.” But the chief faults of the translation are inextricable from its principal virtue, which is a persistent literalism. This means that for a reader prepared to imagine a sort of shadow-French behind the text (and I suppose we can do this whether we know French or not—I do it all the time with translation from Russian), a good deal of Derrida’s tone is kept and some of his finest phrases come through. Rousseau is said to have thought of language as “eaten by writing,” its accents “gnawed through by consonants”; and the relation between imagination and pity, both crucial concepts for Rousseau, is nicely defined: “Imagination is the becoming-human of pity.” Similarly, we see the wit of Derrida’s formulation when he says that Rousseau conceives of the origin as “the beginning of the end, as the inaugural decadence.”

On the other hand, the following passage, not unusually difficult in French, does seem daunting in English:

Perhaps patient meditation and painstaking investigation on and around what it is still provisionally called writing, far from falling short of a science of writing or of hastily dismissing it by some obscurantist reaction, letting it rather develop its positivity as far as possible, are the wanderings of a way of thinking that is faithful and attentive to the ineluctable world of the future which proclaims itself at present beyond the closure of knowledge.

To say in French that tout graphème est d’essence testamentaire is not an abstruse way of saying that every piece of writing has something of the quality of a will about it, but Spivak’s “All graphemes are of a testamentary essence” may lose more than a few English-speaking readers. Again, Derrida’s important, albeit perfectly conventional definition of literature as that form of writing which is more than the message it is supposed to deliver—une écriture qui ne s’épuise pas dans le message—becomes quite obscure, I think, when it emerges as “a writing which would not be exhausted by the message.” But I say all this to warn the reader, not to warn him off. The translation is a noble job, and we should be grateful to have this distinguished book in our hands.

Glas is in many ways Derrida’s most ambitious work to date, an attempt to foil the linearity of language, to write a book that is not a book (Derrida associates the book with the notion of stable, “natural” meanings), but a text; more precisely, a network of texts, a place where texts can talk to each other, a landmark of what has come to be called intertextuality, the haunting of writing by other writing.

The pages of Glas are divided into two columns. In the left-hand column runs an essay on Hegel, peppered with quotations and pulling out all kinds of implications of Hegel’s idea of the family. In the right-hand column there is a quirky essay on Genet and the role of names and flowers in his writing. The book—it is a book, as Derrida knows—begins and ends in the middle of two sentences, and perhaps we are meant to cross over when we finish, and run the last trailing sentence about Genet (“Aujourd’hui, ici, maintenant, le débris de“) into the book’s first sentence about Hegel (“quoi du reste aujourd’hui, pour nous, ici, maintenant, d’un Hegel?“), and vice-versa. Our reading would thus potentially go on for ever. If I add that the words here and now in both columns are already an allusion to Hegel, and that the first words of the Genet column are the title of a work by Genet (“ce qui est resté d’un Rembrandt déchiré en petits carrés bien réguliers, et foutu aux chiottes“) which evokes questions about remnants, shreds, survivals of the past into the present (“A text always has several ages,” Derrida says in Of Grammatology), you will begin to see how complicated the game is.

Glas means knell, and what is being sounded here is the knell of the book that won’t die (in a crack at McLuhan in Of Grammatology Derrida remarks that the death of the book seems to manifest itself mainly by a “compulsive proliferation of libraries”). But it is also the knell of what dies when books are written, when a book is death. Toute pensée émet un coup de dés, Mallarmé said. “Every thought is a throw of the dice.” Derrida modifies this mournfully to tout texte émet un coup de glas, “every text is the tolling of a knell.”

On the other hand, such ceremonies have their compensations. Earlier Derrida had often spelled the word différence as différance, in order to pick up the three meanings that are in the verb in French, but not in the noun (that’s different, that’s where we differ, this matter will have to be deferred). But when he did this, on at least one occasion, in a lecture, the crucial change of letter could not be heard, and became a humorous instance of the richness of writing. Puns too are tokens of the same richness, since many of them, perhaps most of them, rely on the fact that writing registers differences where speech confuses them. This begins to account for Derrida’s merciless punning in Glas.

In an introduction to a book called Marges (“Margins”), Derrida shared his space between his own text and a piece by Michel Leiris, and the introduction itself raises the question of what a margin is. Is Derrida here, on this split page, writing in the margin of Leiris’s text, or is it the other way around? When Hegel and Genet divide a page of Glas equally between them, does the margin disappear, or conversely does everything become a margin? Glas is a version of the scholarly commentary as work of art, a simultaneous parody and resurrection of the form, with footnotes and marginal glosses crowded into a text which is already annotating other texts. The joke is slightly academic, but what scholar has not dreamed of a book that would be all footnotes?

Derrida is reported to have said that all his own texts can be read as a footnote to Finnegans Wake, and the chief charm of Glas, for me, is that it delivers us over so completely to language. To take a simple example from the first page, the French usually pronounce Hegel as Hègle, which sounds just like aigle, eagle. If Derrida were making this pun on his own, it would be trivial, but if generations of Frenchmen have made it, it gathers a sort of history to itself. How can it not matter that people were saying eagle every time they said Hegel?

The actual writing of Glas strikes me as rather conventional, and my own feeling is that renovations of writing, if they are to come, must come at the level of style and the sentence—in the syntax, say, as with Mallarmé—and not in typography or blocks of quotations, or forms of collage or the piling of pun upon pun. Even in the matter of puns, Glas is a long way from being the sustained elaboration of a world-in-language that Finnegans Wake is. But then, modestly enough, Derrida says he’s writing a footnote to that work, and I find Glas, which presents the theoretician of writing as a writer himself—even if he is, still, writing about his reading—a brave and engaging gesture. I can’t give any real sense of the activity of internal reference in the book (a quoted dictionary entry on catafalque, for example, relates the word to the Germanic balk, which takes us to balcon, which leads us to…Genet; the title of the Slovene translation of Derrida’s La Voix et le Phénomène is Glas in phenomen), but perhaps Derrida’s own metaphor may help.

He thinks of himself as dredging texts, as one dredges the sea. He is thinking of Genet—he is dredging the Oeuvres complètes—but we can widen the picture. Everything we read joins the stock of everything we have read—and in a perspective which doesn’t seem to interest Derrida greatly, it joins the stock of much of what we have experienced outside books. Writing is among other things a way of dredging a flooded memory, and Glas is a portrait of the pleasures and insufficiencies of this operation.


Derrida’s principal works are: La Voix et le phénomène, 1967; L’Ecriture et la différence, 1967; De la Grammatologie, 1967; Marges de la philosophie, 1972; La Dissémination, 1972; “L’Archéologie du frivole,” 1973; Glas, 1974. More recently he has published “Le Facteur de la vérité,” 1975; “Economimesis,” 1975; “Où commence et comment finit un corps enseignant,” 1976.

Two books by Derrida are available in English: Of Grammatology, reviewed here, and Speech and Phenomena, translated and edited by David B. Allison, 1973. Essays from L’Ecriture et la différence have appeared in The Structuralist Controversy, 1970, and Yale French Studies, 1972. Essays from Marges have appeared in Philosophy and Phenomenological Research, 1969, New Literary History, 1974, and The Georgia Review, 1976. An English version of “Le Facteur de la vérité,” appeared in Yale French Studies, 1975.

This Issue

March 3, 1977